Drug Discovery & Development

Drug Discovery

Development

Neal G. Simon, Ph.D. Professor

Department of Biological Sciences

ebook

+91-9884350006www.pubrica.com [email protected]

Disclaimer:

DRUG DISCOVERY

AND DEVELOPMENT

Volume 1: Drug Discovery

Edited by

MUKUND S. CHORGHADE

“Epothilone” cover art by Doug Scard

www.sputniknewmedia.com

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.

Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrival system, or transmitted in any form or by any

means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under

Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the

Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Centre,

Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at

www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions

Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011,

fax (201) 748-6008, or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in

preparing this book, they make no representations or warranties with respect to the accuracy or completeness

of the contents of this book and speciﬁ cally disclaim any implied warranties of merchantability or ﬁ tness for a

particular purpose. No warranty may be created or extended by sales representatives or written sales materials.

The advice and strategies contained herein may not be suitable for you situation. You should consult with a

professional where appropriate. Neither the publisher nor author shall be liable for any loss of proﬁ t or any other

commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer

Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or

fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be

available in electronic formats. For more information about Wiley products, visit our web site at

www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Drug discovery and development/edited by Mukund S. Chorghade.

p. cm.

Includes bibliographical references and index.

ISBN-13: 978-0-471-39848-6

ISBN-10: 0-471-39848-9 (cloth : v. 1)

1. Drug development. I. Chorghade, Mukund S. (Mukund Shankar)

[DNLM: 1. Drug Design. 2. Chemistry, Pharmaceutical–methods.

3. Drug Evaluation, Preclinical–methods. QV 744 D79334 2006]

RM301.25C488 2006

615'.19–dc22

2005021297

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

CONTENTS

Contributors xiii

Preface xv

1 From Patent to Prescription: Paving the Perilous Path to Proﬁ t 1

Richard J. Pariza

1.1 Introduction, 1

1.2 A Simple Solution to a Complex Problem, 3

1.3 An Intriguing Patent Problem, 8

1.4 Another Structural Insight, 10

References, 15

2 Medicinal Chemistry in the New Millennium: A Glance into the Future 17

Paul W. Erhardt

2.1 Introduction, 17

2.2 Practice of Medicinal Chemistry, 19

2.2.1 Emergence as a Formalized Discipline, 19

2.2.2 Early Developments, 23

2.2.3 Present Status, 26

2.2.4 Examples Involving Site-Directed Mutagenesis, 27

2.2.5 Latest Trends, 31

2.3 Evolving Drug Discovery and Development Process, 35

2.3.1 Working Deﬁ nition for Medicinal Chemistry, 35

2.3.2 Immediate- and Long-Term Roles for Medicinal Chemistry, 36

2.4 Pursuing Efﬁ cacy, 40

2.4.1 Gathering Positive, Neutral, and Negative SARs During HTS, 41

2.4.2 Example Involving Multidrug Resistance of Anticancer Agents, 42

2.4.3 Compound Libraries: Example of Working with Nature to Enhance

Molecular Diversity, 45

2.5 Assessing and Handling Molecular Conformation, 46

2.5.1 Chemoinformatics, 46

2.5.2 Obtaining Chemically Correct 3D Structures, 49

2.5.3 Inﬂ uence of Biological Environments: Example Involving Drug

Metabolism, 50

2.5.4 Dynamic Energy Relationships: Example Involving a Small Ring

System, 52

2.5.5 Druglike Properties and Privileged Structures, 54

2.5.6 Tiered Structural Information and Searching Paradigms, 55

2.6 ADMET Considerations, 57

2.6.1 Assuring Absorption, 57

2.6.2 Directing Distribution, 58

2.6.3 Herbal Remedies: Example of Working with Nature to Discover

ADMET-Related Synergies, 59

2.6.4 Brute Force HTS to Uncover Multicomponent Synergies, 62

2.6.5 Controlling Metabolism: Example Involving a Soft Drug Strategy, 63

2.6.6 Optimizing Excretion, 65

2.6.7 Avoiding Toxicity, 65

2.6.8 Weighting Decision Criteria from Efﬁ cacy and ADMET SAR, 67

2.7 Process Chemistry Considerations, 70

2.7.1 Cost and Green Chemistry, 70

2.7.2 Deﬁ ning Stereochemistry: Example Involving Benzylamine Chiral

Auxiliary Synthetic Reagents, 71

2.8 Analytical Chemistry/X-ray Diffraction, 74

2.8.1 Latest Trends, 74

2.8.2 Examples Involving Dopamine Receptors, c-AMP Phosphodiesterase

Enzymes, and the Dynamics of Protein Folding, 75

2.9 Summary, 78

2.9.1 General Points, 78

2.9.2 Attributes of Drug Discovery Libraries, Compound Hits, and Lead

Compounds, 81

2.9.3 Formalized Instruction of Medicinal Chemistry, 81

2.9.4 Intellectual Property Considerations, 83

2.9.5 Knowledge Versus Diversity Paradox, 84

Acknowledgments, 85

References and Notes, 85

3 Contemporary Drug Discovery 103

Lester A. Mitscher and Apurba Dutta

3.1 Introduction, 103

3.1.1 Getting Started, 103

3.2 Characteristics of a Suitable Lead Substance, 104

3.2.1 Potency and Selectivity, 105

vi CONTENTS

3.2.2 Structure–Activity Relationships, 107

3.2.3 Toxicity, 107

3.2.4 Changing Appellation of the Best in Series: Analog Attrition, 108

3.3 Some Criteria That a Hit Must Satisfy to Become a Drug, 108

3.3.1 Level of Potency, 109

3.3.2 Comparison of Potency and Efﬁ cacy, 110

3.3.3 Druglike Character, 110

3.3.4 Efﬁ cacy Following Oral Administration, 110

3.3.5 Lipinski Rules for Oral Absorption, 112

3.3.6 Injectable Medications, 113

3.3.7 Distribution, 113

3.3.8 Serum Protein Binding, 114

3.3.9 Metabolism, 114

3.3.10 Distribution, 114

3.3.11 Excretion, 115

3.3.12 Patenting, 115

3.3.13 Pharmaceutical Properties, 115

3.3.14 Idiosyncratic Problems, 115

3.3.15 Summary, 115

3.4 Example of Drug Development That Illustrates Many of the Aforementioned

Considerations, 116

3.4.1 Control of Blood Pressure with Drugs, 116

3.4.2 Historical Background, 116

3.4.3 Finding a Starting Point: A Clue from Nature, 117

3.4.4 Renin–Angiotensin–Aldosterone System, 117

3.4.5 Attempts to Inhibit Renin, 119

3.4.6 Attempts to Inhibit Angiotensin-Converting Enzyme, 119

3.4.7 Peptides Make Poor Orally Active Drugs, 120

3.4.8 Analoging Studies of Pit Viper–Inspired Peptides, 120

3.4.9 Peptidomimetics, 120

3.4.10 Adaptation to Inhibition of ACE, 121

3.4.11 Success Inspires Competition, 123

3.4.12 Taking a Different Approach, 124

3.4.13 Analoging to Enhance Absorption, 124

3.4.14 Clinical SAR, 126

3.4.15 More Recent Work, 128

3.4.16 Résumé, 128

3.5 Conclusions, 128

Additional Reading, 128

4 Combinatorial Chemistry in the Drug Discovery Process 129

Ian Hughes

4.1 Introduction, 129

4.1.1 The Birth of Combinatorial Chemistry, 130

4.1.2 Development of Screening Strategies for Libraries, 131

4.1.3 From Peptides to Small Molecule Synthesis, 132

4.1.4 Beyond Solid-Phase Chemistry, 133

4.2 The Role of Combinatorial Chemistry in Drug Discovery, 135

CONTENTS vii

4.3 Designing Combinatorial Libraries, 137

4.3.1 Describing and Measuring Diversity, 137

4.3.2 A More Focused Approach, 139

4.4 Tools for Synthesis of Combinatorial Libraries, 141

4.4.1 Nonautomated Tools, 141

4.4.2 Mix-and-Sort Systems, 143

4.4.3 Automated Synthesizers, 143

4.4.4 Postsynthesis Processing, 144

4.5 Managing the Combinatorial Process, 146

4.5.1 Speciﬁ cation of Combinatorial Libraries, 146

4.5.2 Controlling the Automated Workﬂ ow, 146

4.6 From Specialist Discipline to Standard Tool, 148

4.7 Application of Combinatorial Chemistry in Drug Discovery, 149

4.7.1 Case History 1, 150

4.7.2 Case History 2, 150

4.7.3 Case History 3, 151

4.7.4 Case History 4, 152

4.8 The Future of Combinatorial Chemistry, 154

4.8.1 Dynamic Combinatorial Libraries, 154

4.8.2 Miniaturization, 154

4.9 Conclusions, 155

References, 156

5 Parallel Solution-Phase Synthesis 169

Norton P. Peet and Hwa-Ok Kim

5.1 Introduction, 169

5.2 Ahead of Our Time, 169

5.3 Recent Reports of Parallel Solution-Phase Synthesis, 172

5.4 Solid Supported Reagents, Scavengers, and Catalysts, 178

5.5 The Future, 191

References, 191

6 Timing of Analog Research in Medicinal Chemistry 199

János Fischer and Anikó Gere

6.1 Introduction, 199

6.2 Early Phase Analogs, 199

6.2.1 ACE Inhibitors, 199

6.2.2 AT

Antagonists, 200

6.2.3 Proton Pump Inhibitors, 200

6.2.4 Insulin Sensitizers: Glitazones, 200

6.2.5 HMG-CoA Reductase Inhibitors, 202

6.2.6 Antimigraine Drugs, 202

6.3 Drug Analogs, 202

6.3.1 Metoclopramide Analogs, 203

6.3.2 Azatadine Analogs, 205

6.3.3 Miconazole Analogs, 205

6.3.4 Nifedipine Analogs, 206

viii CONTENTS

6.3.5 Propranolol Analogs, 207

6.3.6 Clodronate Analogs, 207

6.4 Summary, 208

Acknowledgments, 210

References and Notes, 210

7 Possible Alternatives to High-Throughput Screening 213

Camille G. Wermuth

7.1 Introduction, 213

7.2 Analog Design, 214

7.2.1 Deﬁ nitions, 214

7.2.2 Pharmacophere-Based Analog Design: Scaffold Hopping

or Scaffold Morphing, 215

7.2.3 Natural Compounds as Models, 216

7.2.4 Emergence of New Activities, 216

7.3 Physiopathological Hypotheses, 217

7.3.1 Discovery of Levodopa, 217

7.3.2 H

-Receptor Antagonists, 219

7.3.3 Rimonabant and Obesity, 220

7.4 Contributions from Clinical Investigations, 221

7.5 New Leads from Old Drugs: The SOSA Approach, 223

7.5.1 Rationale, 223

7.5.2 Examples, 223

7.5.3 Discussion, 226

7.6 Conclusion, 228

References, 229

8 Proteomics and Drug Discovery 233

Susan Dana Jones and Peter G. Warren

8.1 Introduction, 233

8.2 Drug Discovery Process, 234

8.2.1 Process Overview, 234

8.2.2 Motivation for Improvement, 236

8.3 High-Throughput Screening Approaches to Drug Discovery, 236

8.4 Emerging Technologies and Approaches: Scale and Speed, 237

8.5 Genomics, 237

8.6 Proteomics, 238

8.6.1 Functional Areas of Proteomics, 239

8.6.2 Fractionation and Puriﬁ cation, 239

8.6.3 Identiﬁ cation, 240

8.6.4 Quantitation, 242

8.6.5 Characterization, 243

8.7 Protein Chip Technology, 248

8.7.1 Issues Addressed, 248

8.7.2 Current State of the Technology, 249

8.8 Proteomics Data Analysis: Computational Biology and

Bioinformatics, 253

CONTENTS ix

8.9 Proteomics and Drug Discovery, 256

8.9.1 Target Identiﬁ cation, 256

8.9.2 Target Validation, 258

8.9.3 Screening for Hits, 259

8.9.4 Lead Optimization, 261

8.9.5 Pharmacology and ADME-Tox, 262

8.9.6 Clinical Trials: Biomarkers and Pharmacogenomics, 263

8.9.7 Case Study, 265

8.10 Conclusions, 266

Acknowledgments, 267

References, 267

Appendix: Public-Domain Software Tools and Databases, 269

9 Using Drug Metabolism Databases During Drug Design

and Development 273

Paul W. Erhardt

9.1 Introduction, 273

9.2 Historical Perspective, 275

9.3 Present Status, 276

9.4 Future Prospects, 280

9.5 Summary, 287

References and Notes, 288

10 Discovery of the Antiulcer Drug Tagamet 295

C. Robin Ganellin

10.1 Historical Background, 295

10.1.1 Prologue, 295

10.1.2 Pharmacological Receptors, 296

10.1.3 Peptic Ulcer Disease, 296

10.1.4 Search for New Antiulcer Drugs, 298

10.2 Search for an H

-Receptor Histamine Antagonist, 298

10.2.1 Histamine Receptors, 298

10.2.2 Biological Approach to a Histamine Antagonist at Non-H

Receptors, 299

10.2.3 Chemical Approach to an Antagonist: Generating a Lead, 300

10.2.4 Lead Optimization, 301

10.2.5 Validating the Research Program, 303

10.3 Development of a Clinical Candidate Drug, 305

10.3.1 Dynamic Structure–Activity Analysis, 305

10.3.2 Imidazole Tautomerism and Sulfur Methylene Isosterism, 306

10.3.3 Isosteres of Thiourea and the Discovery of Cimetidine, 307

10.3.4 Cimetidine: A Breakthrough in the Treatment of Peptic Ulcer

Disease, 308

10.4 Summary and Further Observations, 309

References, 310

x CONTENTS

11 Discovery of Potent Nonpeptide Vasopressin Receptor Antagonists 313

Bruce E. Maryanoff

11.1 Introduction, 313

11.2 Genesis of the Vasopressin Receptor Antagonist Project, 315

11.3 Vasopressin, Its Receptors, and Disease, 315

11.4 The Game Plan, 317

11.5 Novel Chemotypes: Variations on a Theme, 319

11.5.1 Azepinoindoles, 319

11.5.2 Bridged Bicyclic Derivatives, 322

11.5.3 Thiazino-, Oxazino-, and Pyrazinobenzodiazepines, 324

11.6 Epilogue, 332

Acknowledgments, 333

References and Notes, 333

12 Discovery and Development of the Ultrashort-Acting

Analgesic Remifentanil 339

Paul L. Feldman

12.1 Introduction, 339

12.2 Discovery of Remifentanil, 340

12.3 Chemical Development of Remifentanil, 344

12.4 Human Clinical Trials with Remifentanil, 349

Acknowledgments, 350

References, 350

13 Discovery and Development of Nevirapine 353

Karl Grozinger, John Proudfoot, and Karl Hargrave

13.1 Introduction, 353

13.2 Lead Discovery and Optimization, 355

13.3 Chemical Development and Process Research, 357

13.4 Mechanism of Action, 360

13.5 Clinical Studies, 361

Acknowledgments, 362

References, 362

14 Applications of Nuclear Imaging in Drug Discovery and Development 365

John W. Babich and William C. Eckelman

14.1 Introduction, 365

14.1.1 Process and Challenges of Drug Development, 365

14.1.2 Role and Contribution of Position Emission Tomography, 366

14.2 Principles and Evolution of Technology, 366

14.2.1 Introduction to PET Principles, 366

14.2.2 Suitable Targets, 367

14.2.3 Suitable Animal Models, 367

14.3 Role in Drug Discovery, 368

CONTENTS xi

14.3.1 Target Validation and Drug Design, 368

14.3.2 Preclinical Studies, 371

14.3.3 Clinical Studies, 373

14.4 Summary and Outlook, 376

References, 377

15 Polymeric Sequestrants as Nonabsorbed Human Therapeutics 383

Pradeep K. Dhal, Chad C. Huval, and S. Randall Holmes-Farley

15.1 Introduction, 383

15.2 Polymers as Speciﬁ c Molecular Sequestrants, 384

15.3 Sequestration of Inorganic Ions in the GI Tract, 385

15.4 Polymeric Potassium Sequestrants: A Nonabsorbed Therapy for

Hyperkalemia, 385

15.5 Polymeric Drugs for Chronic Renal Failure, 386

15.6 Polymeric Iron Sequestrants for the Treatment of Iron Overload

Disorders, 389

15.7 Sequestration of Bile Acids: Polymers as Cholesterol-Lowering

Agents, 392

15.8 Sequestration of Pathogens: Polymeric Anti-infective Agents, 396

15.9 Sequestration of Toxins, 397

15.10 Polymeric Antimicrobial Agents, 400

15.11 Conclusions and Outlook, 401

References, 402

16 Botanical Immunomodulators and Chemoprotectants in Cancer Therapy 405

Bhushan Patwardhan, Sham Diwanay, and Manish Gautam

16.1 Introduction, 405

16.2 Immunomodulation, 406

16.3 Ethnopharmacology and Botanical Immunomodulators, 406

16.4 Adaptogens or Adjustive Medicine, 407

16.4.1 Botanicals with Adaptogenic Activity, 407

16.4.2 Rasayana Botanicals as Adaptogens, 408

16.5 Chemoprotection, 409

16.5.1 Drug Targets and Current Trends, 409

16.5.2 Chemoprotectants for Antimetabolites, 410

16.5.3 Thiol-Based Chemoprotectants for Cisplatin and

Oxazophosphorine-Based Alkylating Agents, 411

16.5.4 Chemoprotectants for Anthracyclines, 414

16.5.5 Botanical Immunomodulators as Chemoprotectants, 414

16.6 Radioprotection, 417

16.6.1 Radioprotectants from Botanicals, 418

16.6.2 Botanical Immunomodulators as Antitumor Agents, 418

16.7 Conclusions, 419

References, 420

Index 425

xii CONTENTS

CONTRIBUTORS

John W. Babich, Molecular Insight Pharmaceuticals, Inc., 160 Second Street, Cambridge,

MA 02142, USA

Pradeep K. Dhal, Genzyme Corporation, 153 Second Avenue, Waltham, MA 02451, USA

Sham Diwanay, Department of Microbiology, Abasaheb Garware College, Pune 411004,

India

Apurba Dutta, Department of Medicinal Chemistry, Malott Hall, 1251 Wescoe Hall

Drive, Kansas University, Lawrence, KS 66045-7582, USA

William C. Eckelman, Molecular Tracer, LLC, Bethesda, MD 20814, USA

Paul W. Erhardt, Center for Drug Design and Development, The University of Toledo

College of Pharmacy, 2801 West Bancroft Street, Toledo, OH 43606-3390, USA

Paul L. Feldman, GlaxoSmithKline Research and Development, Research Triangle Park,

NC 27709, USA

János Fischer, Gedeon Richter Ltd., H-1475 Budapest 10, Hungary

C. Robin Ganellin, University College London, Department of Chemistry, Christopher

Ingold Laboratories, 20 Gordon Street, London WC1H 0AJ, UK

Manish Gautam, Bioprospecting Laboratory, Interdisciplinary School of Health

Sciences, University of Pune, Pune 411007, India

Anikó Gere, Gedeon Richter Ltd., H-1475 Budapest 10, Hungary

Karl Grozinger, Boehringer-Ingelheim Pharmaceuticals, 900 Ridgebury Road,

Ridgeﬁ eld, CT 06877-0368, USA

Karl Hargrave, Boehringer-Ingelheim Pharmaceuticals, 900 Ridgebury Road,

Ridgeﬁ eld, CT 06877-0368 , USA

xiii

S. Randall Holmes-Farley, Genzyme Corporation, 153 Second Avenue, Waltham, MA

02451, USA

Ian Hughes, GlaxoSmithKline Pharmaceuticals, New Frontiers Science Park (North),

Third Avenue, Harlow, Essex CM19 5AW, UK

Chad C. Huval, Genzyme Corporation, 153 Second Avenue, Waltham, MA 02451, USA

Susan Dana Jones, BioProcess Technology Consultants, Inc., Acton, MA 01720, USA

Hwa-Ok Kim, CreaGen Biosciences, Inc., 25-K Olympia Avenue, Woburn, MA 01801,

USA

Bruce E. Maryanoff, Johnson & Johnson Pharmaceutical Research and Development,

Spring House, PA 19477-0776, USA

Lester A. Mitscher, Department of Medicinal Chemistry, 4010 Malott Hall, 1251 Wescoe

Hall Drive, Kansas University, Lawrence KS 66045-7582, USA

Richard J. Pariza, Cedarburg Pharmaceuticals, 870 Badger Circle, Grafton, WI 53024,

USA

Bhushan Patwardhan, Bioprospecting Laboratory, Interdisciplinary School of Health

Sciences, University of Pune, Pune 411007, India

Norton P. Peet, CreaGen Biosciences, Inc., 25-K Olympia Avenue, Woburn, MA 01801,

USA

John Proudfoot, Boehringer-Ingelheim Pharmaceuticals, 900 Ridgebury Road,

Ridgeﬁ eld, CT 06877-0368, USA

Peter G. Warren, Independent Biotechnology Consultant, Lexington, MA 02421, USA

Camille G. Wermuth, Prestwick Chemical, Inc., Boulevard Gonthier d’Andernach,

67400 Illkirch, France

xiv CONTRIBUTORS

PREFACE

The pharmaceutical sector has traditionally been a vibrant, innovation-driven, and highly

successful component of industry at large. In recent years, a conﬂ uence of spectacular ad-

vances in chemistry, molecular biology, genomics, and chemical technology and the cog-

nate ﬁ elds of spectroscopy, chromatography, and crystallography have led to the discovery

and development of numerous novel therapeutic agents for the treatment of a wide spec-

trum of diseases. To facilitate this process, there has been a signiﬁ cant and noticeable effort

aimed at improving the integration of discovery technologies, chemical outsourcing for

route selection and delivery of active pharmaceutical ingredients, drug product formula-

tions, clinical trials, and reﬁ ned deployment of information technologies. Multidisciplinary

and multifunctional teams focusing on lead generation and optimization have replaced the

traditional, specialized research groups. To develop a drug from conception to commer-

cialization, the biotechnology and biopharmaceutical industries (which have been highly

entrepreneurial) have reached out and established global strategic partnerships with numer-

ous companies.

Currently, there is no single book in the market that provides an overview of strategies,

tactics, milestones, and benchmarks in the entire sequence of operations involved in discov-

ering a drug and delivering it to the armamentarium of clinicians and medical practitioners.

A large number of advanced texts dealing exclusively with medicinal chemistry have been

published; process chemistry has not received the attention it deserves (the journal Organic

Process Research and Development is a useful and overdue step in this direction). Stra-

tegic in licensing, virtual company interactions and related topics have hitherto not been

chronicled in books on drug discovery. There is usually a great gulf between the medicinal

and process chemists in industry; neither has the opportunity to delve into the disparate

literature of the other. This book is designed to bridge this gap and provide greater under-

standing of the target areas.

Conversely, the book is not designed to be a treatise or an encyclopedia. Its scope pre-

cludes complete coverage of any deﬁ ned area. Ideally, it is envisioned to be an advanced-

level monograph with appeal to active researchers and investigators in the entire gamut of

operations comprising the drug discovery and development process. This two-volume text

will be useful to a broad community of academic and industrial chemists. An overview of

several recent developments is presented; this will make it valuable as a reference primer.

The topics and the extent to which they are summarized are based on decisions by the

editor and authors. Each contributor has achieved international distinction in the relevant

ﬁ elds.

The introductory chapter in the ﬁ rst volume, by Dr. Richard Pariza, delineates all the

essential elements that comprise the development process, from the initial conception of

a program to the successful marketing of a new drug. A time line for making critical deci-

sions, conducting pivotal studies, and the approximate duration of different activities is

described. The time line helps to put the entire developmental process into perspective for

the reader and serves as a conceptual index that uniﬁ es all the contributions. Dr. Pariza

elaborates on these concepts by describing some fascinating aspects of the work done on

commercially successful analogs of erythromycin.

Professor Paul Erhardt describes the competition in the pharmaceutical industry to be

“ﬁ rst to the market” in a chosen therapeutic area and the strategies currently being pursued.

These include research in combinatorial chemistry, collaboration with biopharmaceutical

and “virtual companies,” and strategies in the licensing of drug candidates, among others.

Increasingly, the large pharmaceutical corporations have turned to the establishment of stra-

tegic links with small biotechnology and biopharmaceutical companies for in-licensing of

drug candidates and enhancement of drug portfolios. The author takes a futuristic look at

what medicinal chemistry is expected to be in the new millennium. Dr. Erhardt is chairman

of the Division of Chemistry and Human Health of the International Union of Pure and Ap-

plied Chemistry; his insights gleaned from expertise and experience constitute a valuable

lesson.

Professor Lester Mitscher, an internationally renowned academician and expert, and

Professor Apurba Dutta take us through the next critical phase of the drug discovery pro-

cess: detailed studies of the absorption, metabolism, and excretion of potential drug can-

didates. Such studies are of pivotal importance in determining the suitability of a new

compound for further clinical evaluation. His chapter on contemporary drug discovery

presents a broad overview of the successive steps in the progression of a drug from mind

to marketplace.

Combinatorial chemistry has played a highly visible role in the drug discovery effort

in several companies; numerous new companies have been set up to partner established

companies in the discovery of new molecular entities. The strategic focus in this ﬁ eld is

continually shifting; Dr. Ian Hughes reviews the state of the art with selected examples

from his own research at GlaxoSmithKline. This is followed by an excellent exposition by

Drs. Norton Peet and Hwa-Ok Kim regarding efﬁ cient design and development of paral-

lel solution-phase synthesis. Speciﬁ c examples of lead identiﬁ cation and optimization are

presented.

Dr. János Fischer and Dr. Anikó Gere delve into the important area of the timing of

analog research in medicinal chemistry. This work is a remarkable synthesis of knowledge

of drugs and their functional congeners and has formed the basis of a major IUPAC proj-

ect. Professor Camille Wermuth presents fascinating examples of speciﬁ c new drugs being

derived via the functionalization of old drugs. This approach uses the old drugs as new

scaffolds and derives beneﬁ t from new molecules already having a propensity to be “drug-

like.” Professor Wermuth has worked at the academia–industry interface for collaboration

in drug discovery.

xvi PREFACE

Drs. Susan Dana Jones and Peter Warren focus on the impact of proteomics on the

discovery of drugs: newer methods for efﬁ cient, economical, and safer production, and

the development of novel targets and assays for the application of traditional medicinal

chemistry methods. A brief survey of novel therapeutic concepts such as gene therapy,

antisense, transgenic animals, and pharmacogenomics that have opened new vistas in

drug development are surveyed. The authors have familiarized readers with several newer

biology-based technologies. Next, Professor Paul Erhardt introduces the concept of using

drug metabolism databases during the drug discovery and development process.

Professor C. Robin Ganellin exempliﬁ es the discovery of Tagamet using classical

structure–activity relationships and modeling of pharmacophore receptors. This drug was

the ﬁ rst “billion-dollar drug.” The research work by Sir James Black and Robin Ganellin

has long been considered to be a tour de force in modern medicinal chemistry.

The art and science of medicinal chemistry is exempliﬁ ed and epitomized clearly in

the next few chapters. The exponents of the art are highly distinguished and proliﬁ c in-

dustrial researchers whose work spans the gamut of the therapeutic spectrum. Dr. Bruce

Maryanoff brilliantly summarizes research into the discovery of potent nonpeptide vaso-

pressin receptor antagonists. The work is a great tribute to the perseverance and persistence

of researchers. Valuable insights are presented into the discovery process: A key idea is

followed through despite initial adversity. Dr. Paul Feldman presents an informative case

study on the discovery of Ultiva (remifentanil). This is an ultrashort-acting analgesic used

as an adjunct to anesthesia. Dr. Paul Feldman introduces the rationale for its discovery and

discusses how remifentanil ﬁ ts into the anesthesia drug regimen. The desire to discover

an ultrashort-acting analgesic, the group’s medicinal chemistry efforts, and the structure–

activity relationships are discussed. The divergent syntheses of analogs and the ﬁ nal pro-

cess route are described. Finally, the clinical trial data and clinical uses are incorporated in

the chapter to give a complete picture of Ultiva. Drs. Karl Grozinger, John Proudfoot, and

Karl Hargrave discuss the discovery and development of nevirapine. This drug was a key

ingredient in our efforts to combat AIDS, and the success of the researchers is an object

lesson in creativity and how various skills were brought to the forefront of research.

Drs. John Babich and William Eckelman present insights into the applications of

nuclear imaging in drug discovery and development; the work is technologically complex

and involves radiopharmaceuticals. An increasing number of biopharmaceutical compa-

nies are involved in this activity; readers will ﬁ nd this to be a new and exciting domain of

expertise.

Drs. Pradeep Dhal, Chad Huval, and Randal Holmes-Farley take the reader into a new

and somewhat unexplored area of polymer therapeutics. The exciting idea of using a poly-

mer as an active pharmaceutical ingredient was introduced in the 1990s and led to the

discovery of drugs such as Renagel and Welchol. A large-molecular-weight polymer when

used as a drug manifests its action in the gastrointestinal tract by adsorbing and removing

unwanted analytes. The drug is not systemically absorbed in the blood and therefore does

not generate any hazardous metabolites or lead to any toxic effects. It is also unnecessary

to do long-term toxicity tests. This leads to a signiﬁ cant acceleration of the time required

to introduce a drug to meet unmet medical needs.

Professor Bhushan Patwardhan and his collaborators demonstrate the utility of botani-

cal immunomodulators and chemoprotectants in cancer therapy. Much of this work has its

genesis in the Indian medicine systems of ayurveda; this turns pharmacology “on its head.”

It starts with plant extracts that have been used extensively in medicine in Asia and identi-

ﬁ es the active ingredients from a complex mixture of ingredients. There is considerable

PREFACE xvii

scientiﬁ c debate and discussion about whether the active moieties exhibit their pharmaco-

logical action in tandem or singly.

A detailed introduction to the second volume will be presented in its preface; given here

are glimpses of what is to come to whet the reader’s appetite. Drs. G. N. Qazi and S. Taneja

provide a unique perspective on the therapeutic action of bioactive molecules in medicinal

plants. Their group has several years of experience in prospecting natural products in plants

and following up with the isolation, characterization, and structure elucidation of natural

products.

Professor Steven Ley and his collaborators at Cambridge University enlighten read-

ers is to how natural products have served as inspiration for the discovery of new high-

throughput chemical synthesis tools. A salient feature of this masterpiece is the creative

use of polymer-supported reagents.

Drs. Braj and Vidya Lohray elaborate on the role of insulin sensitizers in emerging

therapeutics. A noteworthy feature of this work is that it was done entirely in India and

represents a fast-growing trend: the discovery of new chemical entities in that country.

Drs. Raymond McCague and Ian Lennon at Dowpharma next discuss the criteria for

industrial readiness of chiral catalysis technology for the synthesis of pharmaceuticals.

They exemplify how and why stereoselective reactions are invented for pharmaceutical re-

searchers: The methodology is applicable in both the discovery and development phases of

a drug in making analogs rapidly and by scalable transformations. Dr. Mukund Chorghade

then introduces readers to the ﬁ eld of process chemistry: the quest for the elucidation of

novel, cost-effective, and scalable routes for production of active pharmaceutical ingre-

dients. The medicinal chemistry routes used in the past have often involved the use of

cryogenic reactions, unstable intermediates, and hazardous or expensive reagents. A case

study of the development of a process for an antiepileptic drug is presented; readers will

also see how problems in the isolation, structure elucidation, and synthesis of metabolites

were circumvented.

Drs. Mukund K. Gurjar, J. S. Yadav, G. V. M. Sharma, P. Radha Krishna, C.V. Ramana,

Yatendra Kumar, Braj and Vidya Lohray, and Bipin Pandey have each made seminal con-

tributions to process chemistry. They have invented commercial processes for key phar-

maceuticals that have resulted in signiﬁ cant economies in cost and minimization of waste,

and have engineered “green chemistry” and the development of eco-friendly processes.

These scholars describe their work in the next few chapters with case studies of speciﬁ c

compounds. The work is an eloquent testimony to the collaboration and cooperation inher-

ent in the strategic triad of academics institutions government, and industry. The work is

applicable to the synthesis of both agricultural and ﬁ ne chemicals.

Over the last few years, an increasing number of pharmaceutical and biopharmaceutical

companies have resorted to outsourcing activities in chiral synthesis, process development,

and manufacturing. Dr. Peter Pollack demonstrates this strategy, provides useful pointers

about the do’s and don’ts, and beautifully elaborates the risks and rewards inherent in out-

sourcing in the pharmaceutical industry.

Dr. Shrikant Kulkarni exempliﬁ es solving regulatory problems via thorough investiga-

tions of processes and processing parameters. Dr. Peter Pollack delineates the fascinating

impact of specialty chemicals on drug discovery and development, providing further illus-

tration of the power and utility of outsourcing in drug manufacture.

Chemical engineering plays a central and pivotal role in scale-up operations. Dr. Andrei

Zlota discusses chemical process scale-up tools, mixing calculations, statistical design of

experiments, and automated laboratory reactors.

xviii PREFACE

Dr. Richard Wife explains how some novel initiatives will lead to rescue of “lost

chemistry and molecules,” how the net will make research results accessible to the entire

chemical world, and how information sharing will lead to better and more efﬁ cient re-

search. Thought-provoking and novel studies aimed at predicting compound stability are

presented.

In the concluding chapter, Dr. Colin Scott describes some general principles and prac-

tices in drug development. A brief review is presented of the history of the requirements

for clinical studies leading to the registration of a drug prior to being marketed. This is

followed by a discussion of ethical issues related to clinical studies, the phases of drug

development, and clinical trial design features. The support operations necessary for the

initiation of clinical trials and optimization of results are described. Finally, a global devel-

opment plan, accelerated development opportunities, international regulatory procedures,

and postmarketing requirements are summarized.

There are few courses in academic chemistry departments that deal with drug discov-

ery and development. Graduating students typically have scant exposure to the fascinating

world of industrial chemistry. I am conﬁ dent that the material will excite students inter-

ested in careers in the pharmaceutical industry. A salient feature of the book is the inclusion

of several case studies that exemplify and epitomize the concepts detailed in each chapter.

An instructor interested in developing a course in pharmaceutical chemistry will ﬁ nd the

book useful as a teaching text for a one-semester course.

Dr. Raghunath A. Mashelkar, Director General of the Council of Scientiﬁ c and Indus-

trial Research, has stated: “Rapid paradigm shifts that are taking place in the world as it

moves from superpower bipolarity to multipolarity, as industrial capitalism gives way to

green capitalism and digital capitalism, as information technology creates netizens out of

citizens, as the nations move from ‘independence’ to ‘interdependence,’ as national bound-

aries become notional, and as the concept of global citizenship gets evolved, will see a

world full of new paradigms and new paradoxes; there is no doubt that the rapid advance of

science and technology will directly fuel many of these. The global pharmaceutical and, in

particular, the contract R&D organizations have seen a dramatic change in their capabilities

and sophistication. International pharmaceutical companies should now be ideally poised

to seek collaborations to bring innovative drugs to the consumers at an affordable price.”

Finally, I wish to thank my wife, Veena, my son, Rajeev, and my parents for

their encouragement, emotional support, understanding, and love. They have helped

immeasurably during this endeavor.

UKUND S. CHORGHADE

PREFACE xix

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

FROM PATENT TO PRESCRIPTION:

PAVING THE PERILOUS PATH TO PROFIT

RICHARD J. PARIZA

Cedarburg Pharmaceuticals

Grafton, Wisconsin

1.1 INTRODUCTION

A research director at a major pharmaceutical ﬁ rm used to tell the new scientists in his

company that there was no nobler career than to discover and develop a drug that would

help alleviate human suffering or cure a deadly disease without causing serious side effects.

Many others have doubtless said the same, and added that the complexity of this adventure

can be compared to landing people on the moon and getting them home safely to Earth.

Notice that safety is paramount in both endeavors. Although we must at ﬁ rst do no harm, our

drugs must also do some good. Ethical drug companies spend millions of dollars studying

new drugs over many years to determine both safety and efﬁ cacy, in order to legitimately

promote new chemical entities and formulations to physicians, and more recently directly

to the public. Even with enormous research expenditures and careful regulatory scrutiny,

safety issues with blockbuster drugs are frequently in the news. Patients do not all respond

adequately to existing drugs or even drug classes, and new agents are regularly needed to

ﬁ ght infections caused by microorganisms that become resistant to available antibiotics. So

how do we get started along this path to better and safer drugs?

First, a target must be identiﬁ ed. This is a medical and marketing exercise, where a

problem is recognized that could be treated with a pharmaceutical drug that ﬁ ts into a

company’s portfolio. It is necessary to assure that adequate ﬁ nancial and human resources

will be available for this daunting task. Once the commitment is established, teams of sci-

entists must determine how a chemical could possibly be used to help patients. After all,

pharmaceuticals are chemicals, and pharmaceutical companies sell chemicals.

2 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

Biochemists, molecular biologists, physicians, pharmacologists, and others team up

with synthetic chemists to determine a strategy to attack a disease. Often, these scientists

are in what might be considered a virtual team: not in the same company, not on the same

continent, nor even working in the same decade. By following the medicinal literature

carefully over many years, often in ﬁ elds seemingly unrelated to their own, scientists can

gain insight into possible treatments and apply their own unique talents to come up with

a new drug. There is an enormous amount of information available online, on the World

Wide Web and various scientiﬁ c databases, and modern search engines make it easy to ﬁ nd

both obvious and obscure relationships. A small well-equipped startup company with the

right mix of desire and talent can make breathtaking strides only dreamed of a few decades

ago. They need to understand biology and chemistry, law and economics. To do so, they

must seek the wisdom from the past that often made success achievable even without these

modern tools. Wisdom translates knowledge into understanding.

Very sophisticated approaches are often envisaged that involve inhibiting complex en-

zyme pathways, preventing invading microorganisms or invasive cancer cells from multi-

plying, replacing natural hormones that are lacking in the body, or a host of other possible

ways to treat medical conditions. Chemists are involved in every phase, from planning to

execution of the research, from the laboratory to the clinic. The resulting product sold will

be a chemical, a pure chemical, or a well-deﬁ ned mixture, often a single enantiomer. It

must be stable enough to ship to pharmacies and consumers, who will store it, dispense it,

and use it. It must be safe to handle and have unambiguous safety and a predictable side-

effect proﬁ le once administered. These days especially, it must be cost-effective, offering

worthy advantages over cheaper generic drugs, often helping a patient avoid an expensive

hospital stay and getting him or her back to work sooner. There is always competition to

deal with, so the patent literature must be studied carefully, and risks must sometimes be

taken when working in areas where other companies may have also begun research, be-

cause earlier priority dates may already have been secured. As you will see below, you may

be sowing the seeds for a future partnership by doing research in a crowded ﬁ eld.

The chemical that will become the drug substance or API (active pharmaceutical ingre-

dient) will often be chosen by a process of screening thousands of contending structures,

with various attributes evaluated at each stage. Any structural insights that scientists have

in the early stages can help enormously to abbreviate this development. Rules of thumb re-

garding stability, solubility, and toxicity are ubiquitous, and the successful team will know

these well. ADME (adsorption, distribution, metabolism, and excretion) concepts must be

studied and applied to the drug candidates and their biochemical targets.

Modern approaches that can gain real advantages often involve computer-assisted mod-

eling of potential drug molecules and the sites of their activity. If an x-ray structure of a

target enzyme is known, especially with an inhibitor molecule ﬁ rmly docked, computer

modeling can be used to determine what other drug candidates may also bind strongly with

that site. NMR techniques are also used to screen and assess the interactions of hosts and

potential drugs. With this ﬂ ood of new technologies only now becoming available to me-

dicinal chemists, it is amazing indeed that so many powerful wonder drugs were discovered

and developed in the antediluvian days of the recent past.

At ﬁ rst the cost of producing samples for early testing may not be a major factor, but it

will become more and more important as larger quantities are needed for testing and prog-

ress is made toward clinical trials and commercialization. It is also essential that chemists

and engineers use the most cost-effective syntheses and modern approaches as early as

possible along the drug development time line so that when scale-up issues arise, as they

always do, the best options are available to solve problems quickly. In 2001, the top 16

pharmaceutical companies spent $90 billion to manufacture their products.

Manufacturing

costs have become more than twice the cost of R&D and nearly as much as marketing and

administrative costs. This is due partially to the enormous regulatory and quality issues,

which can lock inefﬁ ciencies into a manufacturing process very early in the ﬁ ling strategy.

Detailed process information, equipment speciﬁ cations, testing protocols, and storage and

stability programs must all be put in place long before clinical studies on a new drug are

completed and reviewed by the Food and Drug Administration (FDA). This is caused by

concerns that any process changes may lead to new impurities or higher levels of extant

impurities, or may make a product that will decompose more quickly and lose potency or

develop harmful by-products. Companies must choose between delaying a ﬁ ling, which

could allow competitors to move ahead or could lose precious patent life, or must submit a

ﬁ ling with a less than ideal manufacturing process. Because of the enormous proﬁ t incen-

tives to get a drug onto the market quickly, the latter is often the course chosen. The drugs

discussed in this chapter each sell at least $1 million to $3 million per day, so any delays

requested to investigate new chemical processes, even for a few weeks, may be considered

too costly. This often plays into the hands of the generic companies, which can start reﬁ n-

ing the manufacturing processes years before the innovator’s patents expire. Chemistry is

always on the critical path.

A glance into the past may convince the reader that human ingenuity, recognizing the

essential features of a problem, and applying Occam’s razor

can often lead to success.

Such cleverness may sometimes be rewarded with a dash of serendipity as well. Genius

transforms understanding into beauty.

1.2 A SIMPLE SOLUTION TO A COMPLEX PROBLEM

Erythromycin was introduced into the clinic in 1952, and although it was a useful antibiotic

with an excellent safety proﬁ le, allowing its use even in children and pregnant women,

blood levels were erratic and there were often annoying side effects, such as nausea, upset

stomach, and diarrhea. In fact, in 1984, the director of antibiotic sales for Abbott Laborato-

ries announced in a meeting with scientists that if they could come up with “[a compound

identical to] erythromycin, but without the belly ache,” he could triple the sales. Newer

formulations of erythromycin were tried but had only limited success in reducing this rela-

tively benign but market-limiting side effect.

It was recognized very early on that acid instability in the digestive track could be a

major cause of these problems. Although the mechanism of the acid-catalyzed degradation

was explored in a one-page publication by Abbott chemists in 1971,

the acid degradation

of erythromycin is not as simple as ﬁ rst envisaged. It was known that erythromycin A (IA)

formed enol ether (IIA), as did erythromycin B (IB). Due to the ⫺OH group on C-12 of

IIA, a further reaction can take place to form anhydroerythromycin A (III), a spiroketal

that is nearly devoid of antimicrobial activity. A paper published in 1986

corroborated this

idea, and a more detailed kinetic study in 1989

suggested that there is equilibrium between

IA and IIA. This equilibrium was conﬁ rmed by the very simple deuterium labeling study

shown in Scheme 1.1.

Work continues on this intriguing system.

Erythromycins A and B (IA and IB) were treated with anhydrous CH

D to form

IIA and IIB, with the ⫺OH’s exchanged for ⫺OD’s. When IIB was treated further with

D in D

O, a more acidic medium, erythromycin B was recovered, with deuterium

A SIMPLE SOLUTION TO A COMPLEX PROBLEM 3

4 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

incorporation and some epimerization

at C-8. It was known that under similar protic con-

ditions, IIA would convert to a single epimer of III.

However, when, after exchanging

the ⫺OH’s for ⫺OD’s, IA was treated directly with CH

D in D

O, within a few min-

utes the anhydroerythromycin A (III) that was formed contained about 50% deuterium at

C-8, as analyzed by

C-NMR. No deuterium was detected at C-10. Furthermore, when

naturally labeled (III) was treated similarly with CH

D in D

O, deuterium was slowly

incorporated at C-8.

Physiologically active compounds often have emergent properties that are due to the

unique spatial arrangement and interactions of their functional groups. For example, the

macrolactone (macrolide) ring appears to have a hydrophobic and a hydrophilic side in

its low-energy conformations, perhaps accounting for the amphiphilic nature of the mol-

ecule, with the OH at C-6 sticking out on the hydrophilic side. IR spectra of erythromycin

A indicate that there is one OH that is not involved in a hydrogen bond. The x-ray struc-

tures as well as molecular modeling show that the OH on carbon 6 is the only one in the

molecule not involved in an internal hydrogen bond with a neighboring polar functionality

(see Scheme 1.2).

OMe

DOAc

IA = erythromycin A, R = OH

IB = erythromycin B, R = H

DOAc/D

HOAc/H

IIA = ery A enol ether, R = OH

IIB = ery B enol ether, R = H

III = anhydroerythromycin

(spiroketal)

DOAc/D

100%

D-incorporation

expected

X = 52% H, 48% D

Scheme 1.1

Scheme 1.3 shows that the three secondary hydroxyl groups in erythromycin A (2', 4",

and 11) can readily be differentiated chemically. The most reactive ⫺OH group is on the

desosamine sugar moiety by virtue of the 3'-dimethylamino group acting as an intramo-

lecular catalyst. Thus, erythromycin A can easily be converted to its 2'-acetate (IV) in

dichloromethane by reacting with acetyl chloride and sodium bicarbonate as base, or acetic

anhydride and triethylamine, rendering the amino group nearly two pK units less basic,

due to the neighboring group interaction.

Further reaction when DMAP is present leads

to acetylation on the cladinose sugar at the 4"-hydroxyl group, and the hydroxyl group

at C-11 can be acetylated only after heating. The lone pairs of electrons on the oxygen at

C-4" are more readily accessible to the reagents than those at C-11, which is involved in a

hydrogen bond. However, if IV is treated with strongly basic conditions capable of fully

deprotonating an ⫺OH on the molecule, such as sodium hexamethyldisilazide in THF at

⫺78⬚C, and acetic anhydride is added, the ⫺OH at C-11 is preferentially acetylated over

the ⫺OH at C-4". This can be understood by stabilization of the C-11 alkoxide by the

neighboring proton on the C-12 hydroxyl, while an alkoxide at 4" is relatively less stable.

Interestingly, compound V was shown to be a 12,9-hemiacetal by NMR, and in the pres-

ence of water–deuterium oxide mixtures undergoes slow hydrogen–deuterium exchange of

the proton at the hydroxyl at C-6 on the NMR time scale. A similar 12,9-hemiacetal was

reported to be a major contributor to the equilibrium structures of erythromycin A (IA)

itself.

Although many scientists were studying these issues, a small group of scientists work-

ing at Taisho Pharmaceuticals in Japan read Abbott’s brief 1971 publication in Experientia

and thought of a very simple solution to the acid instability of erythromycin. In a meeting

with Abbott executives in the 1985, the lead Taisho chemist, Dr. Yoshiaki Watanabe, in

somewhat broken English, thanked Abbott for this one-page revelation. The head of the

Abbott delegation then mumbled to his scientists, on his side of the table, something to

the effect that they were never going to publish another (expletive deleted) paper. Thank

2′

4′′

Scheme 1.2 Internal hydrogen bonds of erythromycin A.

A SIMPLE SOLUTION TO A COMPLEX PROBLEM 5

6 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

goodness he was joking! A partnership was soon born, bridging time and space, and has

ﬂ ourished. Both clarithromycin and azithromycin, discussed below, have achieved annual

sales around $1 billion.

By the amazingly simple idea of blocking the hydroxyl group on carbon 6 with a

methyl group, these Japanese chemists were able to prevent formation of the enol ether

(IIA) or anhydroerythromycin A (III). The compound they ﬁ rst made in 1980,

now

sold as clarithromycin VII (Scheme 1.4), not only has superior acid stability, but pro-

duces less stomach irritation, a major drawback to the widespread use of erythromycin

itself. This serendipitous result may be due at least in part to the inability of clarithro-

mycin to form the enol ether (II), which was later shown in animal studies to increase

gastrointestinal motility to a much greater extent than the parent structure. This effect

was seen even after intravenous administration, so it is not simply the result of contact

of the drug with the stomach. Abbott had developed a gastrointestinal motility assay to

screen new drug candidates. Pressure transducers were attached along the outside of

the GI tract in an anesthetized beagle dog, and peristaltic contractions were recorded

after administering an erythromycin analog. Clarithromycin not only demonstrated a

OMe

OAc

Erythromycin A 2′-acetate

3′-Dimethylamino group:

2′O-ester: p

= 7.1;

2′-OH: p

= 5.2

O, Et

DMAP

Na(TMS)

THF, –78 °C,

4′′

2′

Erythromycin 2′,4′′-diacetate (VI)

Erythromycin 2′,11-diacetate-

12,9-hemiacetal (V)

3′

Scheme 1.3

reduction in the recorded contractions relative to erythromycin, but fewer belly aches

were reported in the clinic.

Clarithromycin has also improved absorption from the

GI tract and enhanced blood levels, coupled with lower intrinsic minimum inhibitory

concentrations (MICs) against important pathogens. Thus, this second-generation semi-

synthetic macrolide is a better antibiotic overall than the direct fermentation product

from which it is made.

The original synthesis

of VII involved protecting both the 2'-OH and the 3'-

dimethylamino functions on the desosamine sugar with benzyloxycarbonyl groups

(Z-groups). This was a method that had been used by many others and results in the loss

of one of the methyl groups on the nitrogen. This step was then followed by simple and

somewhat selective methylation of the 6-OH with methyl iodide. The Z-groups were then

removed by hydrogenolysis, and a methyl group was put back on the nitrogen by reduc-

tive amination with formaldehyde. However, as mentioned above, the mere presence of an

ester functionality on the 2'-OH, such as an acetate, renders the neighboring nitrogen group

much less basic and much less nucleophilic. Therefore, the ﬁ rst process used to prepare

larger quantities of VII was simply to methylate IV, erythromycin 2'-acetate, a compound

that is much more easily prepared and subsequently deprotected. This chemistry is shown

in Scheme 1.5, where each structure is purposely drawn with a different convention taken

from contemporaneous literature, to illustrate how information, even accurate information,

does not always lead to clarity!

A highly crystalline product resulted from the meth-

ylation of IV, albeit in low yield, which could be puriﬁ ed sufﬁ ciently for early studies.

In these early studies large supplies of drug were more important than the efﬁ ciency of

the manufacturing process. The only signiﬁ cant impurity was the 6,11-dimethylated com-

pound, similar to what was seen with the more onerous Z-group protection–deprotection

scheme. Dissolving IV in methanol, and allowing the methanolysis to take place at room

temperature overnight, quantitatively removed the acetyl group. It is interesting to realize

OMe

Scheme 1.4 Clarithromycin (VII).

A SIMPLE SOLUTION TO A COMPLEX PROBLEM 7

8 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

how close many scientists were to this novel second-generation macrolide when they were

working with the simple 2'-esters of erythromycin many years earlier. This is something

to keep in mind when working with a readily available derivative of an active compound:

What new chemistry can you do with it?

1.3 AN INTRIGUING PATENT PROBLEM

A different solution to the acid instability and erratic blood-level problems of erythromy-

cin was found with another analog. Ironically, this new compound has such low blood

levels that at ﬁ rst it seemed to some researchers that infectious disease physicians would

not trust it. The old paradigm was that an antibiotic had to exhibit blood concentra-

tions above the MIC for the particular strain of bacteria causing the illness. In fact, the

OMe

OCH

CCH

OMe

AcO

NMe

OMe

room temp.

Erythromycin A 2′-acetate

(IV)

3′-Dimethylamino group:

2′O-ester: p

= 7.1;

2′-OH: p

= 5.2

NaH in

DMSO–THF,

–5

°C, CH

4′′

2′

Clarithromycin (VII)

-Methyl erythromycin

2′-acetate

3′

methanol,

Scheme 1.5

infectious organism is often compartmentalized in particular tissues such as the tonsils

and the prostate gland. An antibiotic that can penetrate and sustain therapeutic levels in

those diseased tissues would actually be more useful than one that was largely in the blood

serum. This concept is also true of cancer chemotherapy agents, which need to accumulate

in the tumor cells rather than in the bloodstream or healthy tissue. Scientists at Sour Pliva

in Zagreb, in what was then Yugoslavia,

and at Pﬁ zer in Groton, Connecticut,

were

able, almost simultaneously and concurrently to solve the tissue penetration problems and

acid instability issues by cleverly adding an additional basic nitrogen atom in a Beckman

rearrangement process followed by reduction. Almost simultaneously—and the resulting

blockbuster drug, azithromycin (VIII), is the subject of two U.S. composition of matter

patents! Although Pliva ﬁ led their patent more than a year earlier, Pﬁ zer’s patent was issued

seven months sooner. It turns out that the two companies drew their new structures and

named their compounds using quite different conventions. They even numbered the macro-

cyclic ring differently than the classical structures shown in Schemes 1.1, 1.3, and 1.4. Due

to this confusion, the U.S. Patent and Trademark Ofﬁ ce thought the groups were claim-

ing two distinctly different compounds. Pliva had pioneered work with the ring-expanded

macrolides,

and since they ﬁ led their patent ﬁ rst, Pﬁ zer had to negotiate the rights to a

compound that it had discovered and patented independently (Scheme 1.6 and Table1.1). In

today’s competitive world, the Patent Cooperation Treaty requires publication of patents 18

months after ﬁ ling, or earlier claimed priority date: for example, from a provisional patent

application. Had this process been in place in the early 1980s, it would have allowed Pﬁ zer

scientists to see the Pliva application much sooner (the Pliva application would have been

published four months after the Pﬁ zer ﬁ ling), and the Pﬁ zer experts would doubtless have

realized that the compounds they both claimed were identical.

AN INTRIGUING PATENT PROBLEM 9

N(CH

)

OCH

N(CH

)

11-Methyl-11-aza-4-O-cladinosyl-6-O-desosaminyl-

15-ethyl-7,13,14-trihydroxy-3,5,7,9,12,14-hexamethyl

oxacyclopentadecane-2-one

1′

2′

1′′

4′′

2′

4′′

N-Methyl-11-aza-10-deoxo-10-

dihydroerythromycin A

(

)(

)

Scheme 1.6 (a) U.S. patent 4,517,359, May 14, 1985; (b) U.S. patent 4,474,768, October 2, 1984.

10 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

The ﬁ rst generic formulations of azithromycin were approved by the FDA on Novem-

ber 14, 2005 for two companies, Teva Pharmaceuticals and Sandoz, to sell this block-

buster drug for a wide variety of indications, permitting Pﬁ zer and Pliva almost exactly a

20.5-year head start, due to the extension of 3.5 years granted May 20, 1993 for the Pliva

patent. At that time, patents expired 17 years after issuing rather than 20 years from the date

of ﬁ ling, as is now the case (unless extensions are granted).

Note that in Scheme 1.7 azithromycin is drawn in two additional ways, as shown cur-

rently in SciFinder

(clockwise numbering) and the Merck Index

(counterclockwise

numbering), reversing the order of the sugars. The Physicians’ Desk Reference

draws

azithromycin in a fashion related, but not identical, to the Pﬁ zer patent (Scheme 1.6b),

which depicts the stereochemistry of C-6 as S when in fact it is R. Clarithromycin and

erythromycin are drawn in a format similar to that used in SciFinder, except inverted

(so the numbering runs traditionally counterclockwise); there are some ridiculously long

bonds, so the drawings don’t overlap; and all the necessary centers are reversed to maintain

the correct stereochemistry. It is highly unlikely that anyone has ever looked at structures

presented in that fashion and gained any useful insights. How could anyone be expected to

see the crucial interaction of the C-6 OH group with the carbonyl at C-9 in such a rendi-

tion? [Compare erythromycin (R ⫽ H) in Scheme 1.8b with Scheme 1.1 and 1.2.] The enor-

mous confusion caused by following such diverse conventions when drawing and naming

signiﬁ cant compounds has restricted an understanding of the literature to the few experts

who take the time to become familiar with the structures and conventions. It is fortunate

that modern desktop computer programs can recognize instantly that these structures are

equivalent, calculate empirical formulas, assign stereochemistry, and even predict NMR

spectra.

The use of such powerful computer routines, CAS Registry numbers, and other

modern library tools can save time for the expert and be invaluable to the uninitiated. It is

hoped that the confusion over structures such as these will become a relic of the past.

1.4 ANOTHER STRUCTURAL INSIGHT

Recently, it has been widely reported that the new class of wonder drugs called COX-2

inhibitors exhibit serious cardiovascular side effects, and several of these drugs have been

withdrawn from the marketplace. Meanwhile, another class of blockbuster drugs, the statins,

may not only be safe and effective in their intended role of lowering cholesterol, but may

have a plethora of other potentially valuable properties. Cancer, Alzheimer’s disease, dia-

betes, osteoporosis, high blood pressure, multiple sclerosis, and macular degeneration are

Patent Filed 18 months

Issued Extended

Sour Pliva,

U.S. 4,517,359

9/22/81 3/22/83 5/14/85 (43.5 months) 5/20/93 (42.2 months)

Pﬁ zer

U.S. 4,474,768

11/15/82 5/15/84 10/2/84 (23 months)

TABLE 1.1 Comparison of Paths to Patents

Under current regulations the patents would have been published 18 months after ﬁ ling.

1,267 days (3.5 years).

OCH

NMe

OMe

(

)

(

)

(

)

(

)(

)

(

)

(

)

(

)

(

)

(

)

(

)

(

)

(

)

(

)

(

)(

)

(

)

(

)

(

)(

)

Scheme 1.7 Azithromycin: (a) SciFinder, registry number 83905-01-5; (b) from the Merck Index.

OCH

N(CH

)

(a)(b)

Scheme 1.8 Structures similar to those in the 2002 Physicians’ Desk References; (a) azithromycin, pp. 2739, 2743,

2748; (b) clarithromycin, R ⫽ CH

, pp. 403, and erythromycin, R ⫽ H, pp. 454, 456.

among the diseases that the statins may ameliorate.

It has often been said that drugs are

discovered in the clinic. In this sense the clinic consists of the patients using these drugs in

the general population. Observational studies on the millions of people taking these drugs

revealed the problems of COX-2 inhibitors and the additional potential indications for the

statins. Another example of this in a much smaller population can serve as an illustration.

In the late 1960s, Pﬁ zer ﬁ led patents on an α

-adrenergic blocker that came to be known

as prazosin.

The structure is shown in Scheme 1.9, next to a very similar compound,

terazosin, patented in 1977 by Abbott.

Both of these compounds are effective in lowering

blood pressure and have beneﬁ cial effects on the plasma lipid proﬁ le. Dr. Marty Winn, a

chemist at Abbott, looked at a drawing of the structure of prazosin, and realizing that its

failings included problematic intravenous formulation and short duration of action, thought

that a similar molecule with higher water solubility might be more effective. He knew

that furan is only sparingly soluble in water, whereas tetrahydrofuran (THF) is completely

miscible with water. He concluded correctly that simply saturating the furan ring in prazo-

sin might lead to a much more soluble compound. In fact, his ﬁ rst samples were made by

direct hydrogenation of prazosin, leading, of course, to a racemate. This was not a problem

since in those days the FDA did not require compounds to be pure single enantiomers. In

the 1990s, Abbott considered making a new compound as the single enantiomer, a chiral

switch, but did not pursue the issue. It turns out that the base form of terazosin is 25 times

more water soluble than prazosin, and its elimination half-life is about three times greater,

permitting once-daily administration of the new drug. The difference for the correspond-

ing hydrochloride salts is even more dramatic. The terazosin salt is over 500 times more

soluble than the corresponding prazosin salt!

Since terazosin was projected to be a relatively low volume (about a ton per year) high-

potency (10 mg/day) drug, the cost of manufacturing was not deemed a big issue. Note that

one patient would take 3.65 g per year, so 1 metric ton of API is enough to treat 274,000

patients per year. At a price of $1.00 per day, the annual sales would be $100 million. In

the late 1980s a clever process chemist

in the pharmaceutical division suggested ways to

streamline the process and save as much as $500,000 per year. Management decided not

to pursue the new chemistry because it was estimated that it would cost over $2 million

to run several successful manufacturing scale batches, place the API on stability studies,

manufacture tablets, put the tablets on stability for a year, ﬁ le all the data with the FDA,

and wait for approval, before being able to switch over to the new process. The FDA would

have to be convinced that the new API made tablets that were identical to those made by the

old process, and the review process could take many months. Little did Abbott realize that

the market for terazosin was about to increase dramatically, so the new process, despite the

costs of implementing it, could have saved them many millions of dollars in manufacturing

the drug over the long term.

Soon after the introduction of terazosin into the marketplace as an antihypertensive, it

was noticed anecdotally that men with symptomatic benign prostatic hyperplasia (BHP)

who were given the drug to treat high blood pressure began reporting relief of their urethral

pressure and bladder outlet problems. Sales of terazosin increased slowly as word got out

of this promising new treatment for BPH, as it was evidently being prescribed for off-label

use. Physicians have the authority to prescribe drugs for conditions other than those pro-

moted by the pharmaceutical companies, so this is quite common: for example, with anti-

cancer drugs. However, Abbott had to conduct costly clinical studies and get FDA approval

to advertise and market the drug for this new use. Once the new indication was approved by

ANOTHER STRUCTURAL INSIGHT 13

R,S

(a) (b)

Scheme 1.9 (a) Prazosin; solubility of the hydrochloride salt in water (pH ca. 3.5) at ambient temperature (mg/mL): 1.4. (b) terazosin;

solubility in water at 25⬚C (mg/mL): 29.7-hydrochloride salt (mg/mL): 761.2 (544 times more soluble!).

the FDA for detailing, an unexciting drug that had been third or fourth tier for hypertension,

selling much less than $100 million per year, became a major seller at about $500 million

per year worldwide. Terazosin was becoming a very proﬁ table drug just as its patents be-

gan to run out. It quickly became a very attractive target for generic drug manufacturers.

Abbott was able to make deals with the generic competitors to keep them off the market

temporarily, extending its very proﬁ table franchise for about four years after the patent ran

out. They paid several companies millions of dollars per month to keep them off the market

with their cheaper generic version of terazosin. However, faced with an antitrust investiga-

tion in 1999, they canceled such arrangements with potential competitors.

Abbott’s sales

of terazosin fell 70% in the next year alone, to $141 million.

The insight of a chemist who looked at a structural drawing, and spotted the Achilles’

heel of the compound represented, and so elegantly corrected it is astounding. A clearly

drawn chemical structure can reveal the beauty of subtle emergent properties of the func-

tional groups to the astute imagination of a skilled scientist. The fact that this increase in

solubility makes terazosin a superior drug for BPH makes a mark on the positive side of the

ledger of unintended consequences. Serendipity can convert beauty into proﬁ ts!

We should learn never to ignore the clues that any piece of the puzzle is revealing: by

itself or as part of the emerging picture. Each fact that builds toward further understanding

can be exploited, but since our fellow scientists, competitors, or future partners may be far

from us in time and space, our discoveries must be published with clarity. Abbott’s brief

paper in the 1970s paved the way for its future Japanese partner to solve a long-standing

problem, and Pﬁ zer wisely joined with Pliva when it found success in a research area that

the Yugoslavians had pioneered. Despite the dazzling advances in the tools of modern sci-

ence that you will see in the following chapters, there is no substitute for the wisdom of

good common sense, along with painstaking attention to basic details and the occasional

ﬂ ashes of genius that reveal the true beauty of nature. If we only learn from mistakes, then,

after we have made all the mistakes possible, we will ﬁ nally do things right. Discovering

wisdom from the past is a much more efﬁ cient process!

REFERENCES

1. Leila Abboud and Scott Hensley, Wall Street Journal, September 3, 2003, p. A1.

2. The English philosopher William of Occam (1300–1349) propounded Occam’s razor: “Entities

are not to be multiplied more than necessary.” This is especially appropriate when trying to keep

microorganisms or cancer cells from dividing.

3. Paul Kurath, P. Hal Jones, Richard S. Egan, and Thomas J. Perun, Experientia, 27 (4), 362

(1971).

4. Paul J. Atkins, Tristan O. Herbert, and Norbert B. Jones, International Journal of Pharmaceu-

tics, 30(2–3), 199–207 (1986).

5. T. Cachet, R. Hauchecorne, J. Hoogmartens, G. Van den Mooter, and C. Vinckier, International

Journal of Pharmaceutics, 55(1), 59–65 (1989).

6. Richard J. Pariza, and Leslie A. Freiberg, Pure and Applied Chemistry, 66(10–11), 2365–2358

(1994).

7. Yong-Hak Kim, Thomas M. Heinze, Richard Beger, Jairaj V. Pothuluri, and Carl E. Cerniglia,

International Journal of Pharmaceutics, 271(1–2), 63–76 (2004).

8. Under these conditions, erythromycin B epimerizes to about a 1 : 1 mixture of epimers at C-8 in

a day or two.

REFERENCES 15

16 FROM PATENT TO PRESCRIPTION: PAVING THE PERILOUS PATH TO PROFIT

9. The Merck Index entry for erythromycin (Monograph 03714) gives the pK

value as 8.8; for

erythromycin propionate (Monograph 03719) the pK

value is 6.9. Merck & Co., Inc., White-

house Station, NJ, 2001–2005. pK

⫽ 14 – pK

10. J. Barber, J. I. Gyi, G. A. Morris, D. A. Pye, and J. K. Sutherland, Journal of the Chemical Soci-

ety, Chemical Communications, 1040, (1990).

11. Yoshiaki Watanabe, Shigeo Morimoto, and Sadafumi Omura, Novel erythromycin compounds,

U.S. patent 4,331,803, May 25, 1982; ﬁ led May 19, 1981. Merck Index, Monograph 02362,

Merck & Co., Inc., Whitehouse Station, NJ, 2001–2005.

12. Although it is imperative that animal models be validated in the clinic, demonstrating that the

responses in the animal model translate to the human condition, the real proof is often found only

after large populations have taken the drug.

13. During a plenary lecture on his monumental total synthesis of erythromycin at the 1977 ACS

National Organic Symposium in Morgantown, West Virginia, R. B. Woodward had a cartoon

on a slide depicting a tombstone with words to the effect: “Cahn–Ingold–Prelog R.I.P.” He was

annoyed that each time he changed the protecting groups on his intermediates the R, S nomen-

clature changed, so it appeared that stereogenic centers had been inverted when they had not. The

problem of conveniently conveying data by use of various nomenclature conventions has plagued

even Nobel laureates in this ﬁ eld! Thank goodness for computers.

14. Gabrijela Kobrehel and Slobodan Djokic, 11-Methyl-11-aza-4-O-cladinosyl-6-O-desosaminyl-

15-ethyl-7,13,14-trihydroxy-3,5,7,9,12,14-hexamethyloxacyclopentadecane-2-one and deriva-

tives thereof, U.S. patent 4,517,359, May 14, 1985; ﬁ led September 22, 1981.

15. Gene M. Bright, N-Methyl-11-aza-10-deoxo-10-dihydroerythromycin A, intermediates thereof,

U.S. patent 4,474,768, October 2, 1984; ﬁ led November 15, 1982.

16. Gabrijela Kobrehel, Gordana Radobolja, Zrinka Tamburasev, and Slobodan Djokic, 11-Aza-10-

deozo-10-dihydroerythromycin A and derivatives thereof as well as a process for their prepara-

tion, U.S. patent 4,328,334, May 4, 1982, ﬁ led March 28, 1980; cited in ref. 15.

17. SciFinder, registry number 83905-01-5, American Chemical Society, Washington, DC, 2005.

18. Merck Index, Monograph 00917, Merck & Co., Inc., Whitehouse Station, NJ, 2001–2005.

19. 2002 Physicians’ Desk Reference, 56th ed. Medical Economics Company, Inc., Montvale, NJ.

Since the data in the PDR are identical to the package inserts from the drug companies, it is

surprising that Pﬁ zer would approve an incorrect rendition of its own drug.

20. Programs such as ChemDraw Ultra 9.0 from CambridgeSoft Corporation, Cambridge, MA.

21. Ronald Kotulak, Chicago Tribune, December 25, 2005, p. 1.

22. Hans-Jurgen E. Hess, 2,4,6,7-Tetrasubstituted quinazolines, U.S. patent 3,511,836, May 12,

1970; ﬁ led December 13, 1967.

23. Martin Winn, Jaroslav Kyncl, Daniel Ambrose Dunnigan, and Peter Hadley Jones, Antihyperten-

sive agents, U.S. patent 4,026,894, May 31, 1977; ﬁ lled October 14, 1975.

24. Bruce W. Horrom, who started at Abbott in 1946 and became the longest-serving Abbott em-

ployee ever when he retired with more than 50 years of service. He came up with this new pro-

cess idea after having been at Abbott for over 40 years.

25. New York Times (National Edition), July 23, 2000, p. 1.

26. Pharmaceutical Business News Incorporating Biotechnology Business News, February 7, 2001,

p. 11.

MEDICINAL CHEMISTRY IN THE NEW

MILLENNIUM: A GLANCE

INTO THE FUTURE

PAUL W. ERHARDT

The University of Toledo College of Pharmacy

Toledo, Ohio

Who of us would not be glad to lift the veil behind which the future lies hidden; to cast a

glance at the next advances of our science and at the secrets of its development during future

centuries?

—David Hilbert in 1900, as quoted recently by R. Breslow

2.1 INTRODUCTION

Given the highly interdisciplinary nature of medicinal chemistry and the potential for its de-

ployment across a myriad of future life science research activities, this review seeks to high-

light only those possibilities that stand out upon taking a broad purview of the ﬁ eld’s most

prominent trends. From this vantage point, however, at least a glance will have then been cast

toward some of the more noticeable of the exciting opportunities seemingly in store as me-

dicinal chemistry moves forward into the new millennium.

1,2

An overview of the document’s

several sections is provided below. It lists what topics are covered as well as those that are

not covered and indicates the reasoning behind these choices. The overview also describes

the consistent tone that was sought while attempting to elucidate the numerous technologies

that necessarily become encompassed by the variously highlighted activities.

While initially contemplating how medicinal chemistry might continue to evolve as both

a basic and an applied science, it became apparent that it would ﬁ rst be useful to consider

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

This chapter has been used by permission of the International Union of Pure and Applied Chemistry (IUPAC) with

only minor modiﬁ cation from its original form as published in Pure and Applied Chemistry 74, 703–785 (2002).

18 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

where medicinal chemistry has been and how it has come to be what it is today. Thus,

toward quickly establishing a context from which the future might be better appreciated,

and perhaps even seen already to be repeating itself among a new set of players and tech-

nologies, in Section 2.2 a short discourse about medicinal chemistry’s emergence as a for-

malized discipline, its early developments, and its present status is provided by considering

how medicinal chemistry has been practiced across jumps of about 25-year increments.

This section does not include a chronological list of medicinal chemistry’s many contribu-

tions, nor does it highlight the many accomplishments of its noted investigators. Both of the

latter can be found elsewhere as part of more traditional historical treatments.

From the backdrop provided in Section 2.2, medicinal chemistry’s near- and longer-

term futures are considered in Section 2.3 relative to several of today’s trends that are

already having a major impact on the drug discovery process. A working deﬁ nition for

medicinal chemistry is recited at the opening of this section so that medicinal chemistry’s

immediate and future roles can be ascertained more clearly. Section 2.3 also sets the stage

for a later consideration of where several drug development topics may be headed in the

near and longer terms.

Gene therapy, vaccines, and biotech-derived therapeutic agents are not discussed in

Section 2.4, which addresses medicinal chemistry’s continued pursuit of efﬁ cacy. The

aforementioned topics reside primarily within the domains of other disciplines. Readers

are, however, encouraged to consult other reviews offered for these areas in order to appre-

ciate how their advances are sure to have a dramatic impact on future life science research

and its interface with medicinal chemistry (e.g., refs. 4 to 6, respectively). Alternatively,

because assessing molecular conformation is such an integral part of practicing medicinal

chemistry along any venue, several aspects of this key topic are considered in Section 2.5.

In particular, the handling of chemical structures in database settings (e.g., chemoinformat-

ics) is discussed in detail.

Several drug development topics are regarded as critical factors that will have a pivotal

inﬂ uence on medicinal chemistry’s continuing evolution in the new millennium. Each of

these topics is addressed brieﬂ y in Section 2.6. These key topics include assuring absorption,

directing distribution, controlling metabolism, assisting excretion, and avoiding toxicity

[i.e., the traditional absorption, distribution, metabolism, excretion, and toxicity (ADMET)

studies that previously have been undertaken by pharmaceutical companies during the

secondary stages of preclinical drug development]. As an important extension of the

ADMET discussions, nutraceuticals considered in parallel with pharmacological synergy

are also addressed in Section 2.6.

Issues pertaining to medicinal chemistry’s future roles in pharmaceutical intellectual

property (IP) and to trends associated with process chemistry are raised in Section 2.7.

With today’s highly publicized emphasis on genomics and proteomics, at least an abbrevi-

ated discourse about process chemistry is included at this juncture so that this fundamental

aspect of medicinal chemistry’s link with synthetic chemistry remains appreciated. Thus,

toward providing just some of such coverage, the unmet need for large-scale stereoselec-

tive synthetic methodologies is discussed brieﬂ y in Section 2.7. As part of this discussion,

an example is cited that concludes by conveying the medicinal chemistry logic that was

encompassed as a critical component of the example’s investigations.

Although it is beyond the scope of this review to discuss the impact that progress in

each of several analytical methods is likely to have on medicinal chemistry,

x-ray dif-

fraction has been selected to provide a representative discussion in Section 2.8. As is often

acknowledged by researchers from various disciplines, “science moves forward according

to what it can measure,” and presently, there appear to be numerous promising advances

among various analytical techniques that can be used to study drug–receptor interactions.

For example, readers are encouraged to seek other reviews in order to appreciate the poten-

tial impact that anticipated developments in nuclear magnetic resonance (NMR inclusive of

LC-NMR and high-ﬂ ow-through techniques), mass spectrometry (MS inclusive of LC-MS

and LC-MS/MS), microcalorimetry, and surface plasmon resonance may have on medicinal

chemistry’s future (e.g., refs. 7 to 10, 11 to 13, 14 and 15, and 16 and 17, respectively).

In Section 2.9, practical implications that stem from some of the earlier discussions are re-

visited in what also serves as an overall summary for this document. After restating medicinal

chemistry’s anticipated roles in future life science research, concerns pertaining to the training

of medicinal chemists, inventorship, and the interplay of patent trends and future research within

the context of the IP arena have all been reserved for comment in the concluding summary.

The document’s running dialogue has been developed from future possibilities suggested

by the current medicinal chemistry and drug discovery literature, as well as from general

observations afforded while consulting in both the private and public sectors. Descriptions of

speciﬁ c research projects have been interspersed throughout so that real case examples, along

with their chemical structures, could be conveyed explicitly. A concerted effort has been made

to keep hype to a minimum. Alternatively, jargon has been used whenever it was thought that

such terms portray the mind sets that were important for a given period, or because a par-

ticular term or phrase appears to be taking on an enduring signiﬁ cance. Some of the more

technical of these terms are listed in Table 2.1 along with a short deﬁ nition in each case.

Since several acronyms have been used for repeating phrases, an alphabetical listing of all

acronyms and their deﬁ nitions is provided in Table 2.2 to assist readers as they move deeper

into the document. Numerous references to secondary scientiﬁ c/primary news journals have

been cited because these journals are doing an excellent job of both reporting the most recent

trends and forecasting the potential future. In several cases, a single citation has been used to

list many of the informational Web sites that can often be found for a given topic.

Topics are considered into the future only to about one-half the distance that has been

summarized for medicinal chemistry’s past: namely, for about 75 years, with the ﬁ rst 25

being regarded as near term, and the next 50 being regarded as long term. The speculation

that has necessarily been interjected throughout the review has been done with the thought

that one of the goals for this type of chapter is to prompt the broadest contemplation pos-

sible about the future directions that medicinal chemistry might take. Finally, as medicinal

chemistry is shown to move forward in time, it has been considered as both a distinct pure

science discipline and, equally important, as a key interdisciplinary applied science col-

laborator seeking to mingle with what should certainly prove to be an extremely dynamic

and exciting environment within the life sciences arena of the new millennium.

2.2 PRACTICE OF MEDICINAL CHEMISTRY

2.2.1 Emergence as a Formalized Discipline

Medicinal chemistry’s roots can be found in the fertile mix of ancient folk medicine and

early natural product chemistry: hence its name. As appreciation for the links between

chemical structure and observed biological activity grew, medicinal chemistry began to

emerge about 150 years ago as a distinct discipline intending to explore these relationships

via chemical modiﬁ cation and structural mimicry of nature’s materials, particularly with

PRACTICE OF MEDICINAL CHEMISTRY 19

20 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Term Deﬁ nition

Ab initio calculations Quantum chemical calculations that use exact equations to

account for the complete electronic structure of each atom in a

molecule.

AMBER Molecular mechanics program commonly used for calculations

on proteins and nucleic acids.

Basis set Set of mathematical functions used in molecular orbital

calculations (e.g., 3-21G*, 6-31G*, or B3LYP when used in

ab initio calculations refer to the type of mathematical function

that was deployed).

Bioinformatics Application of computer science and technology to address

problems in biology and medicine.

Bioisostere Broadly similar atoms or groups of atoms in terms of

physiochemical or topological properties that can be used as

replacements in a biologically active compound to create a

new structure that retains all selected features of the parent

compound’s biological properties.

Chemoinformatics Handling of chemical structure and chemical properties in

database settings; when related to biological properties, this

ﬁ eld becomes a subdivision of bioinformatics.

Combinatorial chemistry Synthesis, puriﬁ cation, and analysis of large sets of compounds

wherein sets of building blocks are combined at one or more

steps during their preparation.

Combinatorial library Set of compounds prepared by using combinatorial chemistry.

Comparative molecular ﬁ eld

analysis (CoMFA)

A 3D-QSAR method that uses statistical correlation techniques

for the analysis of the qualitative relationships between the

biological activity of a set of compounds with a speciﬁ ed

alignment of their 3D electronic and steric properties.

Additional parameters such as hydrophobicity and hydrogen

bonding can also be incorporated into the analysis.

Druglike Structural motifs and/or physicochemical proﬁ les that can be

associated with providing an overall ADME behavior that is

conducive to effective use in humans via the oral route.

Electrostatic potential Physical property equal to the electrostatic energy between the

static charge distribution of an atomic or molecular system

and a positive unit point charge. Used in 3D-QSAR, molecular

similarity assessment, and docking studies.

e-Research Research asset management via in-house computer networks and

across the World Wide Web.

Extended Hückel calculations Low-level semiempirical molecular orbital calculations.

Gaussian programs Type of mathematical function used during ab initio calculations

to describe molecular orbitals. Numbers refer to year of

program updates.

Genomics Used herein as the study of the chromosomal and

extrachromosomal genes in humans, in particular their

complete sequential characterization.

High-throughput screening (HTS) The level of this activity is moving rapidly from 96- to 384-well

microplates.

436

In silico Tasks undertaken via computer (e.g., virtual screening).

Isosteres Molecules or ions of similar size containing the same number of

atoms and valence electrons.

TABLE 2.1 Selected Terms and Abbreviated Deﬁ nitions

Term Deﬁ nition

Ligand-based drug design Design of new structural arrangements based on their

resemblance to at least a portion of another compound that

displays a desired property or biological activity.

Metabophores Structural features residing in substrates that prompt their speciﬁ c

metabolic conversions.

MNDO calculations Semiempirical molecular orbital calculations that use a modiﬁ ed

neglect of diatomic (differential) overlap approximation.

Molecular mechanics Calculation of molecular conformational geometries and energies

using a combination of empirical force ﬁ elds.

Molecular orbital calculations Quantum chemical calculations based on the Schrödinger

equation, which can be subdivided into semiempirical

(approximated) and ab initio (nonapproximated) methods.

Multivalent ligands Molecular displays that provide more than one set of binding

motifs for interaction with more than one area on one or more

biological surfaces.

Nutraceuticals Dietary supplements purported to have beneﬁ cial therapeutic or

disease preventive properties (e.g., herbal medicines).

Pharmacogenetics Study of genetic-based differences in drug response.

106

Pharmacophores Structural features needed to activate or inhibit speciﬁ c receptors

or enzyme active sites.

Privileged structures Molecular frameworks able to provide ligands for diverse

receptors.

264,265

Prodrug Compound that must undergo biotransformation (e.g.,

metabolism) before exhibiting its pharmacological effects.

Proteomics Study of protein structure and function.

438

Quantum mechanics Molecular property calculations based on the Schrödinger

equation that take into account the interactions between

electrons in the molecule.

Semiempirical calculations Molecular orbital calculations using various degrees of

approximation and using only valence electrons.

Single nucleotide polymorphisms

(SNPs)

Used herein as single-base alterations in the human genome

that occur in speciﬁ ed percentages of distinct portions of the

population.

Soft drug Compound that has been programmed to be biodegraded (e.g.,

metabolized) to predictable, nontoxic, and inactive metabolites

after having achieved its therapeutic role.

Structure-based drug design Design of new structural arrangements for use as drugs based

on protein structural information obtained from x-ray

crystallography or NMR spectroscopy.

65,474

Toxicophores Structural features that elicit speciﬁ c toxicities.

481

Transportophores Structural features that prompt a speciﬁ c compound’s transport.

Ultrahigh-throughput screening

(UHTS)

The level of this activity has moved rapidly from 384- to 1536-

well microplates with peak throughput rates of over 100,000

compounds per day. It is estimated that this ﬁ eld will soon be

closing in on producing 1200 data points per minute.

478

Xenobiotic Compound that is foreign to a given organism.

Adapted from deﬁ nitions provided by the IUPAC.

432

Various deﬁ nitions can be found

433

; this particular deﬁ nition has been cited because it appears to be used most

commonly by the U.S. National Institutes of Health (NIH).

434

Adapted from deﬁ nitions provided by IUPAC.

435

Adapted from deﬁ nitions provided by the U.S. Ofﬁ ce of Dietary Supplements (ODS).

437

PRACTICE OF MEDICINAL CHEMISTRY 21

TABLE 2.1 (Continued)

22 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Acronym Designation

ADME Absorption, distribution, metabolism, and excretion

ADMET ADME and toxicity

CADD Computer-assisted drug design

CAS American Chemical Society’s Chemical Abstract Services

CCD Charge-coupled device (as in x-ray area detectors)

CoMFA Comparative molecular ﬁ eld analysis

CPT Camptothecin

CYP Cytochrome P450 metabolizing enzyme

439

FDA U.S. Food and Drug Administration

GI Gastrointestinal tract

HPLC High-performance liquid chromatography

HTS High-throughput screening

IND Investigational new drug

IP Intellectual property (e.g., trade secrets and patents)

IT Information technology

KDD Knowledge discovery in databases

440

LC Liquid chromatography

Log P Log value of a compound’s n-octonal/H

O partition coefﬁ cient

MDR Multidrug resistance

MS Mass spectrometry

NBE New biological entity,

441

taken to mean that such a compound has

potential diagnostic, therapeutic, or prophylactic value

NCE New chemical entity, taken to mean that such a compound has potential

diagnostic, therapeutic, or prophylactic value

NDA New drug application

NMR Nuclear magnetic resonance spectroscopy

PAC Paclitaxel

Pgp P-glycoprotein pump associated with the ABC class of membrane

transporter systems

PK Pharmacokinetic

QSAR Quantitative structure–activity relationships; note that the Q designation

also becomes applicable to all of the other SXR possibilities

SAbR Structure–absorption relationships

SAR Structure–activity relationships wherein activity is equated herein with

therapeutic efﬁ cacy-related elements from either an agonist or an

antagonist type of interaction

SDM Site-directed mutagenesis

SDR Structure–distribution relationships

SER Structure–excretion relationships

SMR Structure–metabolism relationships

SNP Single nucleotide polymorphism within a genome

SPR Surface plasmon resonance spectroscopy

STR Structure–toxicity relationships

STrR Structure–transporter relationships

SXR Generic representation for simultaneous SAR, SAbR, SDR, SER, SMR,

STR, and STrR

UHTS Ultrahigh-throughput screening

2D Two-dimensional structure representation

3D Three-dimensional structure representation

Phrases and acronyms commonly used in the U.S. drug regulatory process.

TABLE 2.2 Acronyms and Designations

an eye toward enhancing the efﬁ cacy of substances thought to be of therapeutic value.

the United States, medicinal chemistry became formalized as a graduate-level discipline

about 75 years ago within the academic framework of pharmacy education. From this set-

ting, overviews of medicinal chemistry’s subject matter have been offered to undergradu-

ate pharmacy students for many years.

20,21

Understanding structure–activity relationships

(SARs) at the level of inherent physical organic properties (i.e., lipophilic, electronic, and

steric parameters) coupled with consideration of molecular conformation soon became the

hallmark of medicinal chemistry research. Furthermore, it follows that because these fun-

damental principles could be useful during the design of new drugs, applications toward

drug design became the principal domain for a still young, basic science discipline. Per-

haps somewhat prematurely, medicinal chemistry’s drug design role became especially im-

portant within the private sector, where its practice quickly took root and grew rampantly

across the rich ﬁ elds being staked out within the acres of patents and intellectual property

that were of particular interest to the industry.

2.2.2 Early Developments

As a more comprehensive appreciation for the links between observed activity and pharma-

cological mechanisms began to develop about 50 years ago and then also proceeded to grow

rapidly in biochemical sophistication, medicinal chemistry, in turn, entered into what can

now be considered to be an adolescent phase. Conﬁ dently instilled with a new understand-

ing of what was happening at the biomolecular level, the ensuing period was characterized

by the high hope of being able to design new drugs independently in a rational (i.e., ab

initio) manner rather than by relying solely on nature’s templates and guidance for such.

Although this adolescent “heyday of rational drug design”

should certainly be credited

with having spurred signiﬁ cant advances in the methods that can be deployed for consider-

ing molecular conformation, the rate of actually delivering clinically useful therapeutic enti-

ties having new chemical structures within the private sector was not signiﬁ cantly improved

for most pharmacological targets unless the latter’s relevant biomolecules also happened

to lend themselves to rigorous analysis (e.g., obtaining an x-ray diffraction pattern for a

crystallized enzyme’s active site with or without a bound ligand). One of the major reasons

that rational drug design fell short of its promise was because without experimental data like

that afforded by x-ray views of a drug’s target site, medicinal chemistry’s hypothetical SAR

models often reﬂ ected speculative notions that were typically far easier to conceive than

were the actual syntheses of the molecular probes needed to assess a given model’s associ-

ated hypotheses. Thus, with only a small number of clinical success stories to relay, medici-

nal chemistry’s “preconceived notions about what a new drug ought to look like” began to

take on negative rather than positive connotations, particularly when being “hand-waved”

within the context of a private-sector drug discovery program (e.g., ref. 23). Furthermore,

from a practical point of view, the pharmaceutical industry, by and large, soon concluded

that it was more advantageous to employ synthetic organic chemists and have them learn

some pharmacology than to employ formally trained medicinal chemists and have them rec-

tify any shortcomings in synthetic chemistry that they might have due to their exposure to a

broader range of nonchemical subject matter during graduate school. Indeed, given the pro-

pensity for like-to-hire-like, the vast majority of today’s investigators who practice medici-

nal chemistry within big pharma, and probably most within the smaller-company segment

of the pharmaceutical industry as well, have academic backgrounds from organic chemis-

try rather than from formalized programs of medicinal chemistry. Although this particular

PRACTICE OF MEDICINAL CHEMISTRY 23

24 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

glimpse is important to appreciate as a historical note in that it provides useful insights

for the review’s later discourse about the formalized training of future medicinal chemists,

further references to medicinal chemistry throughout the remainder of this chapter intend to

imply medicinal chemistry’s practice as a discipline regardless of how a given investigator

may have become trained to do so. Importantly, no matter how its practitioners were being

derived, and even though it was still very much under-the-gun within the pharmaceutical

industry, medicinal chemistry did continue to thrive quite nicely during this period and cer-

tainly moved forward signiﬁ cantly as a recognized discipline within all sectors.

Arriving at the next historical segment, however, one ﬁ nds that medicinal chemistry’s in-

ability to accelerate the discovery of new chemical entities (NCEs) by using rational drug de-

sign became greatly exacerbated when the biotechnology rainfall began to hover over the ﬁ eld

of drug discovery just somewhat less than about 25 years ago.

With this development, not

only did the number of interesting biological targets begin to rise rapidly, but also, the ability

to assay many of these targets in a high-throughput manner suddenly prompted the screening

of huge numbers of compounds in very quick time frames. Ultimately, the need to satisfy

high-throughput screening’s (HTS’s)

25,26

immense appetite for compounds was addressed not

by either natural product or synthetic medicinal chemistry but by further developments within

what had quickly become a ﬂ ood-level

continuing downpour of biotechnology-related break-

throughs. Starting as gene cassette-directed peptide libraries and quickly moving into solid-

phase randomly generated peptide and nucleotide libraries,

28–32

this novel technology soon

spread across other disciplines, eventually spawning the new ﬁ eld of small-molecule combi-

natorial chemistry. Today, using equipment and platforms available from a variety of suppliers

(e.g., Table 2.3), huge libraries of compounds can readily be produced in either a random or a

directed manner.

33–40

Once coupled with HTS, these paired technologies have prompted what

has now come to be regarded as a new paradigm for the discovery of NCEs across the entire

big-pharma segment of the pharmaceutical enterprise.

41–43

Figures 2.1 and 2.2 provide a com-

parison of the old (classical) and new drug discovery paradigms, respectively.

As largely reported in 1999 in an article by Brown.

The number of reported libraries with or without disclosure of their biological

activities has grown from only a few that were produced primarily by the private

sector in the early 1990s to nearly 1000 in 1999, with nearly half of the latest

total now being contributed by the academic sector.

Over this same period, the

percentage of libraries directed toward unbiased discovery has gone from about

60% to 20%, while that for targeted/optimization of biased structural systems has

risen from about 40% to 80%.

Vendor Web Address

Advanced Chem Tech Peptide.com

Argonaut Technologies Argotech.com

Bohdan Automation Bohdan.com

Charybdis Technologies Charybtech.com

Chiron Technologies Chirontechnologies.com

Gilson Gilson.com

PE Biosystems Pebiosystems.com

Robbics Scientiﬁ c Robsci.com

Tecan Tecan-us.com

Zymark Zymark.com

TABLE 2.3 Suppliers of Combinatorial Chemistry Systems

and Compound Library Trends

Figure 2.1 Classical drug discovery and development paradigm. This model portrays interactions with

U.S. regulatory agencies and uses terms related to those interactions for steps 11 to 17. All of the other

terms typify generic phrases that have commonly been used by the international pharmaceutical com-

munity. Although some of the noted activities can be conducted in parallel or in an overlapping manner,

the stepwise sequential nature of this paradigm’s overall process is striking. Furthermore, whenever a

progressing compound fails to meet criteria set at an advanced step, the process returns or draws again

from step 3 for another reiteration. Numerous reiterations eventually identify a compound that is able to

traverse the entire process. A successful passage through the entire process to produce just one product

compound has been estimated to require about 15 years at a total cost of about $500 million. While

the largest share of these time and cost requirements occur during the later steps, the identiﬁ cation of a

promising preclinical development compound, step 7, can be estimated to take about four years from the

time of initiating a therapeutic concept. Step 1 is typically associated with some type of physiological

or pharmacological notion that intends to amplify or attenuate a speciﬁ c biological mechanism so as to

return some pathophysiology to an overall state of homeostasis. Step 2 typically involves one or two

biochemical-level assay(s) for the interaction of compounds intending to amplify or attenuate the con-

cept-related mechanism. As discussed in the text, steps 3 and 4 reﬂ ect key contributions from medicinal

chemistry and typically use all sources of available information to provide for compound efﬁ cacy hits

(e.g., everything from natural product surveys to rational approaches based on x-ray diffractions of the

biological target). Step 5 generally involves larger in vitro models (e.g., tissue level rather than biochemi-

cal level) for efﬁ cacy and efﬁ cacy-related selectivity. Step 6 generally involves in vivo testing and utilizes

a pharmacodynamic (observable pharmacologic effect) approach toward compound availability and du-

ration of action. Step 7 typically derives from a formal review conducted by an interdisciplinary team

upon examination of a formalized compilation of all data obtained to that point. Step 8 speciﬁ es parallel

activities that are typically initiated at this juncture by distinct disciplines within a given organization.

Step 9 begins more reﬁ ned pharmacokinetic evaluations by utilizing analytical methods for the drug

itself to address in vivo availability and duration of action. Step 10 represents short-term (e.g., two-week)

dose-ranging studies to identify toxic markers initially within one or more small animal populations.

Expanded toxicology studies typically progress while overlapping with steps 11 to 14. Steps 11 to 13 rep-

resent formalized reviews undertaken by both the sponsoring organization and the U.S. Food and Drug

Administration (FDA). Step 14 is typically a dose-ranging study conducted in healthy humans. Steps 15

and 16 reﬂ ect efﬁ cacy testing in sick patients, possible drug interactions, and so on. Step 17 again reﬂ ects

formalized reviews undertaken by both the sponsoring organization and the FDA. The FDA’s fast-track

review of this information is now being said to have been reduced to an average of about 18 months. It is

estimated that it costs a company about $150,000 for each day that a compound spends in development.

Finally, step 18 represents the delivery of an NCE to the marketplace. (From refs. 41 and 477 to 482.)

26 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

2.2.3 Present Status

Interestingly, the marriage of HTS with combinatorial chemistry has led to a situation

where identifying initial lead compounds is no longer considered to be a bottleneck for the

Figure 2.2 New drug discovery and development paradigm. This model portrays interactions with

U.S. regulatory agencies and uses terms related to those interactions for steps 11 to 17. All of the

other terms typify generic phrases that are commonly used by the international pharmaceutical com-

munity. The battery of proﬁ ling included in step 3 represents a striking contrast to the classical drug

discovery paradigm (Fig. 2.1). Furthermore, all of these screens are/will be of a high-throughput

nature such that huge numbers of compounds can be tested simultaneously in extremely short time

periods. As the predictive value of the resulting proﬁ les improve, selected compounds will have

higher and higher propensities to proceed successfully through steps 5, 6, 9, and 10 (Fig. 2.1) to the

point that these assays may become more of a conﬁ rmatory nature or may even be able to be omitted

completely (hence their dashed lines here).

483

The efﬁ ciency of traversing the various clinical test-

ing steps 14 to 16 successfully will also be improved, but their complete removal from the overall

process is highly unlikely. After the initial investments to upgrade step 3 and enough time has passed

to allow for the generation of knowledge from step 3’s raw data results (see text for discussion), the

overall time frame and cost for a single NCE to traverse the new paradigm should be considerably

improved from the estimates provided in Fig. 2.1. Step 1 is likely to be associated with some type of

genomic and/or proteomic derived notion that intends to amplify or attenuate a speciﬁ c biological

mechanism so as to return some pathophysiology to an overall state of homeostasis. Step 2 will be a

high-throughput assay derived from using molecular biology and bioengineering techniques. In step

3a, MC  medicinal chemistry. Because step 3a exploits actual pictures of what type of structural

arrangements are needed to interact with the biological targets, this approach toward identifying

new compound hits will continue to operate with high efﬁ ciency. However, because of this same

efﬁ ciency, the biological targets that lend themselves to such experimental depiction (by affording

crystals suitable for x-ray diffraction) are likely to be quickly depleted very early into the new mil-

lennium. Step 3b represents various combinatorial chemistry-derived libraries, natural product col-

lections, and elicited natural product libraries (see text for discussion of this topic). Steps 4 to 18 are

similar to the descriptions noted in Fig. 2.1. (From refs. 41 and 477 to 482.)

overall process of drug discovery and development. Indeed, many of the programs within

pharmaceutical companies are presently considered to be suffering from “compound over-

load,”

with far too many initial leads to follow up effectively. ADMET assessments are

now regarded as the new bottleneck, along with the traditionally sluggish clinical and regu-

latory steps. This situation, in turn, has prompted an emphasis to move ADMET-related

parameters into more of an HTS format undertaken at earlier decision points. Thus, even

though efﬁ cacy-related HTS and combinatorial chemistry reﬂ ect very signiﬁ cant incorpo-

rations of new methodologies, from a strategic point of view the most striking feature of the

new drug discovery and development paradigm shown in Fig. 2.2, compared to the clas-

sical approach depicted in Fig. 2.1, actually becomes the trend to place ADMET-related

assays closer to the beginning of the overall process by also deploying HTS methods.

Clearly, with the plethora of biologically based therapeutic concepts continuing to rise

even further and the identiﬁ cation of lead compounds now being much quicker because of

the HTS–combinatorial chemistry approach, more efﬁ cient handling of ADMET-related

concerns represents one of the most signiﬁ cant challenges now facing drug discovery and

development. Because of its importance, this challenge is likely to be resolved within the

near term of the new millennium. Medicinal chemistry’s critical role during this further

evolution of the present drug discovery paradigm is highlighted in subsequent sections.

2.2.4 Examples Involving Site-Directed Mutagenesis

While the drug discovery process has been inﬂ uenced by biotechnology in numerous ways

(e.g., Tables 2.4 and 2.5), one development deserves to be especially noted as this brief

account of medicinal chemistry’s history and present status is brought to a close. This

development is already having a major impact directly on the process of uncovering SARs

relevant to small-molecule drug design.

The method involves site-directed mutagenesis

(SDM). SDM applies to systematic point mutations directed toward speciﬁ c sites on genes

that translate to proteins associated with enzyme-active sites or drug receptor systems such

that the targeted changes can be used to study SARs while holding one or more active

site/receptor ligands constant during the assessment (Fig. 2.3). Numerous investigators are

now utilizing this reverse SAR approach to explore both enzyme and receptor ligand inter-

actions. Three examples are provided below wherein the site-directed mutagenesis studies

have been further coupled with one or more of the latest analytical chemistry techniques

such as microcalorimetry, as well as with sophisticated computational approaches. The

ﬁ rst pair of examples involve the active sites of some isolable enzymes; the third example

involves a nonisolable membrane-bound receptor complex.

Slama et al. have conducted studies involving a phosphatidylinositol-4-phosphate phos-

phatase (Scheme 2.1) designated as Sac1p.

With this enzyme’s peptide sequence in hand,

transformation of a bacterial host with an appropriate gene copy plasmid has allowed at

least one point-mutated protein to be examined per month in a functional biochemical

assay. While this particular example happens to represent a system that lends itself to over

expression coupled with a functional protein product that lends itself to ready isolation,

it clearly demonstrates that site-directed mutagenesis should no longer be considered to

be lengthy and tedious compared to classical SAR studies undertaken by rational syn-

thetic modiﬁ cations of an enzyme’s substrates. Slama et al.

are now studying poly(ADP-

ribose)glycohydrolase or PARG (Scheme 2.1) to more deﬁ nitively assign a tyrosine (i.e.,

796

Tyr) to a key role within this enzyme’s active site. Previously, this particular residue

has only been able to be implicated as being important by using classical inhibitor and

PRACTICE OF MEDICINAL CHEMISTRY 27

28 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

photoafﬁ nity labeling studies.

48,49

In this case, one of the PARG assays deploys micro-

calorimetry measurement of the binding energies for a designated series of ligands that is

then held constant as it is surveyed across the various mutant enzymes.

Similarly, Messer

and Peseckis et al. have collaborated in several site-directed mutagenesis studies involving

M1 muscarinic receptors.

51–53

These topographical mapping studies began as a follow-up

Activity Impact

Genomics and proteomics Plethora of new and better deﬁ ned mechanisms

to pursue as therapeutic targets.

High-throughput efﬁ cacy assays Screen huge numbers of therapeutic candidates

in short time frames using low compound

quantities.

High-throughput ADMET assays An evolving development: once validated

and coupled with efﬁ cacy assays, should

eventually allow for selection or drug design/

synthesis of clinical candidate compounds

rather than lead compounds that still require

considerable preclinical testing and additional

chemical tailoring.

Peptide and oligonucleotide compound libraries Provide huge numbers of compounds for

screening (spawned the ﬁ eld of combinatorial

chemistry as now applied to small organic

compounds); can be used as SAR probes and,

pending further developments in formulation

and delivery, may also become useful as drug

candidates.

Site-directed mutagenesis Allows “reverse” structure–activity relationship

explorations.

Transgenic species Novel in vivo models of pathophysiology that

allow pharmacological proof of principle

in animal models that mimic the human

situation; and animal models modiﬁ ed to have

human metabolism genes so as to provide

more accurate PK data and risk assessment.

442

Peptide version of pharmacological prototype Developed to the IND phase as an intravenous

agent can allow for clinical proof of principle

or concept.

Pharmacogenetics An evolving development: should soon

reﬁ ne clinical studies, market indications/

contraindications, and allow for subgrouping

of populations to optimize therapeutic

regimens; eventually, should allow for

classiﬁ cation of prophylactic treatment

subgroups.

TABLE 2.4 Impact of Biotechnology on Small-Molecule Drug Discovery and Development

This list is not intended to highlight the numerous activities associated with the development of speciﬁ c “biotech”

or large-molecule therapeutics (e.g., see Table 2.5). The arrangement of activities follows the order conveyed in

Figs. 2.1 and 2.2 rather than being alphabetical.

TABLE 2.5 Examples of Approved

Biotechnology-Based Drugs

Name

Company Clinical Use

Tissue plasminogen

activators (tPAs)

Activase

Retevase

Genentech

Boeringher-Mannheim;

Centocor

Dissolution of clots associated

with myocardial infarction,

pulmonary embolism, and

stroke

Clotting factors

BeneFIX Genetics Inst. Treatment of various

hemophilias

KoGENate Bayer

Recombinate Baxter/Genetics Inst.

Ceredase-

glucocerebrosidase

Cerezyme Genzyme Gaucher’s disease

DNAse

Pulmozyme Genentech Cystic ﬁ brosis

Insulins

Humalog Eli Lilly

Humulin

Eli Lilly

Novolin Novo Nordisk Insulin-sensitive diabetes

Novolin L Novo Nordisk

Novolin R Novo Nordisk

Erythropoiten-related

growth factors

Epogen

Procrit

Amgen

Ortho Biotech

Anemia associated with renal

failure, chemotherapy,

surgery, or loss of blood

Growth hormone

Biotropin

Genetropin

Humatrope

Neutropin*

Norditropin

Protropin

Saizen

Serostim**

Bio-Tech. Gen.

Pharmacia & Upjohn

Eli Lilly

Genentech

Novo Nordisk

Genentech

Serono Labs

Growth hormone deﬁ ciency

in children* and for use

in adults and in Turner’s

syndrome**; AIDS-related

wasting and for pediatric

HIV patients

Growth hormone-

releasing factor

Sermorelin Serono Labs Evaluation and use in pediatric

growth hormone deﬁ ciency

Platelet-derived

growth hormone

Regranex Ortho-McNeil Lower-extremity ulcers in

diabetic neuropathy

Fertility hormones

Gonal-F Serono Female infertility

(Continued)

PRACTICE OF MEDICINAL CHEMISTRY 29

30 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Name

Company Clinical Use

Interferon-α

Alferon N

Infergen*

Intron**

Roferon-A**

Interferon Sci.

Amgen

Schering-Plough

Hoffman-LaRoche

Genital warts*; chronic

hepatitis C infections**;

several antiviral and

anticancer indications

Interferon-β

Avonex

Betaseron

Biogen

Berlex

Multiple sclerosis

Interferon-γ

Actimmune Genentech Chronic graulomatous disease

Granulocyte-colony and

macrophage-stimulating

factor

Leukine

Neupogen

Immunex

Amgen

Treatment of blood disorders

during chemotherapy and

blood cell/marrow transplants

Interleukins

Neumega

Proleukin*

Genetics Inst.

Chiron

Thrombocytopenia subsequent

to chemotherapy*; certain

cancers

Monoclonal antibodies

CEA-scan Immunomedics Cancer diagnostic

Herceptin Genetech Breast cancer

MyoScint Centacor Cardia imaging agent

Neumega Genetics Inst. Prevention of chemotherapy-

induced thrombocytopenia

OncoScint Cytogen Cancer diagnostic

Orthoclone Ortho Biotech Transplant rejection

Prostascint Cytogen Cancer diagnostic

ReoPro Lilly Prevention of blood clots during

angioplasty

Rituxan Genetech; IDEC Non-Hodgkin’s lymphoma

Simulect Novartis Transplant rejection

Synagis Med Immune Respiratory syncytial virus

disease

Vezluma Boechringer-Ingelheim/NesRx Cancer diagnostic

Zenapax Hoffman-LaRoche; Protein

Design Labs

Transplant rejection

Antisense nucleotides

Vitravene ISIS/Ciba Retinitis

TABLE 2.5 (Continued)

Prior to mid-year 1999 and as largely reported in the three-part series published in 1999 by Hudson and Black.

In comparison, a more recent survey reveals that about 375 additional agents were already in, or were nearing

entry into, the clinical testing pipeline.

443

These compounds have also been referred to as new biological entities (NBEs) by analogy to new chemical enti-

ties (NCEs).

441

In general, NBEs use, recreate, or improve on proteins and other biomolecular polymers produced

in the body to counter disease.

443

Asterisks in this column relate to those in the “clinical use” column.

First recombinant-DNA-produced therapeutic protein to enter the market.

to better deﬁ ne the receptor interactions exhibited by CDD-0102 (Scheme 2.2), a selec-

tive M1 agonist that is undergoing preclinical development for the potential treatment of

Alzheimer’s disease.

54–56

Interestingly, these investigator’s more recent efforts have addi-

tionally turned toward addressing questions about the multivalent nature of M1 muscarinic

receptors.

57–60

Thus, in this case the site-directed mutagenesis studies are helping to re-

solve fundamental issues about the nature of these receptors, as well as helping to implicate

amino acid residues involved in the binding of agonists for the purpose of identifying drug

structural themes having enhanced selectivity for certain of the receptor subtypes. In all of

the M1-related studies, the site-directed mutagenesis results are being coupled to computa-

tional studies directed toward mapping the receptor’s key topological features.

2.2.5 Latest Trends

As has been noted by others, the deployment of these new approaches to studying SARs

represents an “exciting”

development that has clearly had an invigorating effect on the

practice of medicinal chemistry in many areas. Alternatively, that the SAR hallmark and

Series of Potential Receptor Ligands Explored

as the Experimental Variable

Classical SAR

Protocol

Wild Type

Receptor

Series of Point Mutated Receptors

Explored as the Experimental

Variable

SDM SAR Protocol

(Reverse SAR Protocol)

Selected Receptor

Ligand(s) Held Constant

Figure 2.3 Classical versus site-directed mutagenesis (SDM) exploration of structure–activity

relationships (SARs).

PRACTICE OF MEDICINAL CHEMISTRY 31

32 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

drug design intellectual domains of medicinal chemistry, along with chemically servicing

the experimental approaches toward identifying novel, lead structures, have all been over-

run by technologies initially derived from other disciplines, has also had somewhat of an

unnerving impact on medicinal chemistry. This is because it has previously been the nearly

exclusive deliverance of these roles from a chemical orientation that has served to distin-

guish medicinal chemistry as a distinct discipline. Thus, as a ﬁ eld, medicinal chemistry has

had to mature quickly from its troubled adolescence only to ﬁ nd itself in the middle of an

identity crisis. This crisis has been occurring among both its private-sector practitioners

Scheme 2.1 Cleavages effected by Sac1p and PARG. 1, Phosphatidylinositol-4-phosphate, wherein

R and R' are fatty acid side chains; 2, poly(ADP-ribose), wherein Ad is adenine and Glu is glucose. In

both cases, arrows indicate bonds undergoing enzymatic hydrolyses. (From refs. 46 to 49.)

Scheme 2.2 Suspected interactions of CDD-0102 with muscarinic M1 receptors. 3, 5-(3-Ethyl-

1,2,4-oxa-diazol-5-yl)-1,4,5,6-tetrahydropyrimidine (CDD-0102) shown as its protonated species;

Asn, asparagine 382; Asp, aspartic acid 105; Thr, threonine 192. Dashed lines signify suspected

hydrogen bonds. (From refs. 51 to 53).

and its academicians. Indeed, today’s trend within the public-sector funding arena, with a

major emphasis on genomics and proteomics, is causing some academic medicinal chemis-

try and chemistry investigators to turn their intellectual pursuits further and further toward

molecular biology.

Similarly, the undergraduate instruction of pharmacy students, which

has for so long represented medicinal chemistry’s academic bread and butter, has shifted

its emphasis away from the basic sciences toward more of a clinically oriented curriculum.

A ﬁ nal development that has contributed to medicinal chemistry’s identity crisis is the fact

that the new drug discovery paradigm supplants medicinal chemistry’s long-standing posi-

tion wherein its practitioners have typically been regarded as the primary inventors of the

composition of matter speciﬁ cations associated with NCE patent applications. This last

development, along with the overall IP arena of the future, is addressed in Section 2.9.

Although as alluded to earlier and as now practiced in tandem with SDM studies, clas-

sical medicinal chemistry rationale is still being relied on heavily to effectively identify

and ﬁ ne-tune lead compounds for systems whose biomolecules have lent themselves to

x-ray diffraction and/or NMR analyses (i.e., structure-based drug design),

64,65

medicinal

chemists of the new millennium must be prepared to face the possibility that the present

complement of these more amenable pharmacological targets is likely to become quickly

exploited, perhaps eventually even exhausted. Such a scenario, in turn, suggests that this

last stronghold for today’s practice of rational medicinal chemistry could also be lost as a

bastion against what could then potentially become an even more serious identity crisis in

the future.

Present trends thus leave us with the following paradigm for the immediate future of

small-molecule drug discovery (Fig. 2.2):

1. Genomics and proteomics will continue to uncover numerous new pharmacological

targets, to the extent that choosing the most appropriate and validating such targets

among the many therapeutic possibilities will also continue to rise as a growing chal-

lenge in itself.

2. Biotechnology will, in turn, continue to respond by generating quick ligand-

identiﬁ cation assays for all new targets chosen to be pursued: namely, by deploying

HTS protocols.

3. Targets that lend themselves to x-ray diffraction and structure-based drug design are

likely to be quickly exploited.

4. Ligands for HTS will be supplied by existing and new combinatorial-derived com-

pound libraries as well as from wild

and biotechnology-elicited natural sources.

5. Assessment of ADMET parameters, presently considered to be the bottleneck for the

overall drug discovery process (Table 2.6), will continue to move toward HTS modes

that can be placed at earlier and earlier positions within the decision trees utilized to

select lead compounds for further development as drugs.

It is important to emphasize at this juncture that in order to place conﬁ dence in the

predictive value of ADMET HTS surveys, these particular screens must become validated

relative to actual clinical-related outcomes. For the moment, however, this situation is

best likened to a deep, dark chasm that the rapidly evolving ADMET HTS surveys still

need to traverse if they are ultimately to become successful. Nevertheless, because of its

importance, it is proposed that not only will this type of validation be forthcoming within

the near term, but as the new millennium then continues into the more distant future,

PRACTICE OF MEDICINAL CHEMISTRY 33

Activity Estimated Time Frame Percentage Successfully Traversing Associated Criteria

Biological conception A plethora of genomic/proteomic characterizations presently

lies waiting to be exploited; this situation is expected to

prevail well into the new millennium.

The challenge lies in prioritizing which of the numerous

mechanisms might be best to pursue (see next entry).

Proof of therapeutic principle Ultimately requires reaching phase II clinical testing;

BIOTECH-derived humanized and/or transgenic disease

state models may be able to be substituted at an earlier point,

depending on the conﬁ dence associated with their validation.

Generally high, although there are some distinct therapeutic

categories that continue to have low success rates or lack

deﬁ nitive validation, such as the attempted treatments of

septic shock or the pursuit of endothelin modulators.

Identiﬁ cation of lead

compound based on efﬁ cacy

screen

Using HTS, thousands of compounds can be tested in a

matter of days or less (10 to 100 times more with UHTS);

companies are beginning to have more lead compounds than

they can move forward in any given program.

One compound out of 5000 from random libraries/one out of

10 from directed libraries; despite the low efﬁ ciency, this is

not regarded as a bottleneck because HTS can be done so

quickly; much higher percentages can be obtained during

ligand- and structure-based drug design, but synthesis is then

correspondingly slower.

Progression to preclinical

development compound

Approximately two years. About one out of 50 wherein all can be examined during the

time frame indicated.

Progression to clinical

development compound

Approximately two years. About one out of 10 wherein all can be examined during the

time frame indicated.

Phase I clinical study

Approximately one year.

About one out of two.

Phase II clinical study

Approximately two years.

About one out of two.

Phase III clinical study Approximately three years.

About one out of 1.5 and often with modiﬁ ed labeling details.

Product launch Approximately two years (NDA submission/approval). About one out of 1.5.

TABLE 2.6 Assessment of Drug Discovery and Development Bottlenecks

Presently regarded as the bottleneck for the overall process. These are the points where ADMET properties have historically been assessed. Approximately 40% are rejected owing to poor

pharmacokinetics and about another 20% because they show toxicity in animals. In the new drug discovery paradigm (Fig. 2.2), the ADMET assessments are being moved to an earlier point

in the overall process and are being conducted in an HTS mode. However, in most cases, validation of the new methods relative to clinical success still needs to be accomplished.

Although these studies may not be able to be accomplished any quicker, they may be able to be done more efﬁ ciently (e.g., smaller numbers and focused phenotypes within selected patient

populations) and with a greater success rate based on making the same improvements in the ADMET assessment area as noted in footnote a.

Timing includes generation and submission of formal reports.

About one of ﬁ ve compounds entering into clinical trials becomes approved. The overall process to obtain one marketed drug takes about 12 to 15 years at a cost of about $500 million.

Source: Refs. 444 to 449.

ADMET-related parameters derived from HTS will become even further manipulated for

their potential to allow for synergistic relationships within the overall course of a given

therapeutic or prophylactic treatment. These intriguing future possibilities for exploiting

ADMET-related parameters in a proactive and synergistic manner rather than in just a

negative ﬁ ltering mode, along with the likely move toward prophylactic medicines, are

discussed further in subsequent sections.

2.3 EVOLVING DRUG DISCOVERY AND DEVELOPMENT PROCESS

2.3.1 Working Deﬁ nition for Medicinal Chemistry

If the targets that are readily amenable to x-ray diffraction and structure-based drug design

do become exhausted with time, the only title role presently highlighted for medicinal

chemistry within the new paradigm of drug discovery will also disappear. To determine if

medicinal chemistry will still be operative under such a circumstance, let us ﬁ rst review

medicinal chemistry’s present deﬁ nition. Perhaps further attesting to medicinal chemistry’s

present identity crisis, however, is the fact that a purview of several of today’s major medic-

inal chemistry texts reveals that although one can ﬁ nd topic-related versions of such within

the context of various other discussions, even the textbooks seem reluctant to provide an

explicitly stated general deﬁ nition for medicinal chemistry (e.g., refs. 68 to 73). Turning,

instead, to this document’s earlier consideration of medicinal chemistry’s history, and for

the moment disengaging ourselves from any biases that might be interjected by overreact-

ing to the continuing ﬂ ood of biotechnology-related trends, it does become possible to

devise a general, working deﬁ nition for medicinal chemistry that can be used to address its

present-day identity crisis while also serving in a search for any key roles that medicinal

chemistry ought to be playing now and into the future of life science research.

As a working deﬁ nition for this review, let us simply state that medicinal chemistry uses

physical organic principles to understand the interaction of small molecular displays with

the biological realm. Physical organic principles encompass overall conformational con-

siderations, chemical properties, and molecular electrostatic potentials, as well as distinctly

localized stereochemical, hydrophilic, electronic, and steric parameters. Understanding

such interactions can provide fundamental, basic knowledge that is general as well as com-

pound-speciﬁ c in its applications directed toward either enhancing the overall proﬁ le of a

certain molecular display or designing an NCE (e.g., by effecting small molecule-driven

perturbations of discrete biological processes or of overall biological pathways for the

purpose of eliciting a speciﬁ ed therapeutic endpoint). Small molecular displays should

be thought of in terms of low-molecular-weight compounds (e.g., usually less than 1 kg)

that are typically of a xenobiotic origin and thus not in terms of biotechnology-derived

polymers. While the latter are being addressed aggressively by other ﬁ elds, it should ad-

ditionally be noted that a consideration of speciﬁ c details associated with the interaction

of small molecular portions of more complex biomolecular systems still falls within the

purview of medicinal chemistry’s stated focus. Alternatively, the biological realm should

be thought of very broadly, so as to encompass the complete span of new ADMET-related

systems as well as the traditional span of biological surfaces that might be exploited for

some type of efﬁ cacious interaction. The technologies that can be deployed as tools to

study these interactions at medicinal chemistry’s fundamental level of understanding are,

by intent, dissociated from medicinal chemistry’s deﬁ nition. Presently, such tools include

EVOLVING DRUG DISCOVERY AND DEVELOPMENT PROCESS 35

36 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

biotechnology-related methods such as site-directed mutagenesis, combinatorial chemistry

methods provided that the latter are coupled with knowledge-generating structural data-

bases (later discussion), and long-standing synthetic chemistry manipulations that can be

conducted in a systematic manner on either one of the interacting species in order to ex-

plore SARs. In the future, medicinal chemistry should be equally prepared to exploit what-

ever new tools and techniques that become available to allow it to proceed more efﬁ ciently

along the lines of its technology-independent deﬁ nition. Finally, it should be appreciated

that this deﬁ nition merges both the basic and applied natures of medicinal chemistry’s sci-

entiﬁ c activities into a key mix of endeavors for which a new research paradigm (Fig. 2.4)

has also been proposed recently as being a signiﬁ cant trend,

even if potentially “danger-

ous” in that it could compromise the longer-term pursuit of fundamental knowledge by

bringing applied science decision criteria into funding programs that have previously sup-

ported pure, basic science.

2.3.2 Immediate- and Long-Term Roles for Medicinal Chemistry

Importantly, no matter what pace targets amenable to x-ray may continue to allow struc-

ture-based drug design to be pursued in the new millennium, applying the aforementioned

deﬁ nition across the present drug discovery paradigm reveals that there is an even more

vital activity with which medicinal chemistry needs to become more involved. Indeed,

as HTS results are amassed into mountains not just for efﬁ cacy data but for each of the

ADMET parameters as well, it should ultimately become medicinal chemistry’s role, by

deﬁ nition, to attempt to understand and codify these awesome, crisscrossing ranges if such

data are to be merged and used either to select or design the most promising preclinical de-

velopment compounds. For example, while medicinal chemistry’s principles and logic may

not be needed to identify hits or leads from a single HTS efﬁ cacy parameter survey across a

library of potential ligands, or perhaps not even needed for two or three of such consecutive

surveys involving a few additional ADMET-related parameters, it is extremely doubtful

that the same series of compounds identiﬁ ed from within an initial library as a hit subset,

Figure 2.4 Nonlinear relationship of medicinal chemistry to basic and applied research. Surveys

suggest that chemical pharmaceutical companies spend about 9%, 37%, and 54% of their research

dollars on basic, applied, and developmental aspects of research, respectively.

484

(Adapted from a

ﬁ gure provided by Cotton

as part of his summary and commentary about a book entitled Pasteur’s

Quadrant.

)

or as further generated within a directed library based on the initial hit subset, will be able

to sustain themselves as the most preferred leads upon continued HTS parameter surveys

in the future when the latter become ramped up to their full potential. Furthermore, this

should still be the case even if the downstream selection criteria become more and more

relaxed through any of such progressions that ultimately strive to merge even a preliminary

HTS-derived ADMET portfolio in conjunction with one or more selective efﬁ cacy HTS

parameter surveys. In other words, identiﬁ cation of the optimal end product (i.e., the best

preclinical candidate compound) is unlikely to be derivable from an experimental process

that does not represent a “knowledge”

76,77

generating system that also allows for rational-

based assessments and adjustments, or even complete revamping, to be interspersed at

several points along the way. In the end, today’s move toward “focused libraries”

78,79

and

“smarter,” presorted relational databases may thus represent a lot more than just the often-

touted desire to “be more efﬁ cient.”

Indeed, this may be the only way for the new drug

discovery paradigm, now in its own adolescent phase, eventually to work as it continues

to mature and to take on more ADMET-related considerations in an HTS format. In this

regard it can now be emphasized that the common denominator required to correlate the

HTS data from one pharmacological setting to that of another ultimately resides in the

precise chemical structure language that medicinal chemistry has been evolving since its

emergence as a distinct ﬁ eld (i.e., SAR deﬁ ned in terms of physical organic properties dis-

played in three-dimensional space). This, in turn, suggests that within the immediate future

of the new millennium it should be medicinal chemistry that rises to become the central

interpreter and distinct facilitator that will eventually allow the entire new drug discovery

paradigm to become successful.

This central role for medicinal chemistry may become even more critical longer term

(i.e., for the next 50 to 75 years of the new millennium). Speculating that the new drug

discovery paradigm will indeed mature within the next 25 years into a synergistic merger

of efﬁ cacy and thorough ADMET HTS systems that allows for an effective multiparameter

survey to be conducted at the onset of the discovery process, the accompanying validated

predictive data that will have been generated over this initial period should be statistically

adequate to actually realize today’s dream of virtual or in silico screening

81–86

through vir-

tual compound and virtual informational libraries (not just for identifying potential efﬁ cacy

leads to synthesize, as is already being attempted, but across the entire preclinical portion

of the new paradigm wherein the best overall preclinical candidate compound is selected

with high precession for synthesis at the outset of a new therapeutic program) (Fig. 2.5).

However, this futuristic prediction again depends on the entire maturation process being

able to proceed in a knowledge-generating manner. Again central to the latter is medicinal

chemistry as the common denominator. For example, with time it can be expected that just

as various pharmacophores and toxicophores have already been identiﬁ ed for various por-

tions of the biological realm associated with efﬁ cacious or toxic endpoints, respectively,

speciﬁ c molecular properties and structural features will become associated with each of

the ADME behaviors. Indeed, work toward such characterizations is already progressing in

all of these areas (Table 2.7). Understanding the pharmacophores, metabophores,

toxico-

phores, and so on, in terms of subtle differences in molecular electrostatic potentials (from

which medicinal chemistry’s physical organic properties of interest are derived), as well

as in terms of simple chemical structural patterns, will eventually allow for identifying

optimal composites of all of these parameters across virtual compound libraries as long as

the latter databases have also been constructed in terms of both accurate three-dimensional

molecular electrostatic potentials and gross structural properties.

EVOLVING DRUG DISCOVERY AND DEVELOPMENT PROCESS 37

38 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

It should be clear that in order to play this key role effectively, the medicinal chem-

ist of the new millennium (Fig. 2.6) will have to remain well versed in physical organic

principles and conformational considerations while becoming even more adept at applying

them within the contexts of each of the ADMET areas, as has previously been done during

Figure 2.5 Future drug discovery and development paradigm. This model portrays interactions

with U.S. regulatory agencies and uses terms related to those interactions for steps 11 to 17. Future

implies about 50 to 75 years into the new millennium. The most striking feature of this paradigm

compared to Figs. 2.1 and 2.2 is the considerable number of decisions that will be made from virtual

constructs rather than from experimental results. Conﬁ dence in the virtual decisions will be directly

proportional to the level of knowledge that is learned from the huge amounts of drug screening data

being amassed during the ﬁ rst 25 to 50 years of the new millennium, coupled with the overall abil-

ity to predict clinical outcomes. Step 1 is likely to be associated with some type of genomic and/or

proteomic derived notion that intends to amplify or attenuate one or more speciﬁ c biological mecha-

nisms so as either to return some pathophysiology to an overall state of homeostasis or to modify

some system in a manner that prevents or provides prophylaxis toward an otherwise anticipated

pathophysiological development. A growing emphasis of treatments will be directed toward preven-

tion. Step 2 may be based on an actual x-ray diffraction version of the biological target or on a com-

putationally constructed version derived from similar known systems that have been catalogued for

such extrapolations. In either case, docking studies will be conducted in a virtual mode. Steps 3, 4,

and 5 will be conducted in a virtual mode. Steps 6 and 7 represent the ﬁ rst lab-based activities. After

submission of patents, it is proposed that in vivo testing involving steps 9 and 10 will be able to take

advantage of project-speciﬁ c and FDA-approved generic toxicity model transgenic species. Steps 11

to 18 are similar to those in Figs. 2.1 and 2.2, except that the likelihood for a compound to fall short

of the desired criteria will be signiﬁ cantly reduced. Subject inclusion/exclusion criteria will also be

much more reﬁ ned based on advances in the ﬁ eld of pharmacogenetics.

ADME Area Parameters

Absorption Physicochemical (e.g., log P

312,450

)

Functional groups (e.g., rule of ﬁ ve

311

)

Distribution Two-directional ﬂ ows across Caco-2 cell monolayers

451

Metabolism 2D structure-metabolism databases

452

(e.g., Metabol

Expert,

453

META,

454

Metabolite,

455

and Synopsis

Metabolism Database

456

) 3D models (e.g., CYP1A,

457

CYP2A,

458

CYP2B,

459

CYP2C,

560

CYP2D,

461,462

and

CYP3A

463

Excretion (half-lives, etc.) Structural patterns across species

464

Physicochemical (e.g., log P

465,466

)

TABLE 2.7 Examples of Efforts to Establish ADME-Related Structural Patterns

It is anticipated that the ADME informational area will soon become heavily inundated as pharmaceutical com-

panies begin to share their rapidly accumulating data through publications in the public domain (analogous to

what has happened historically for receptor/enzyme active site efﬁ cacy information but with the latter having

previously had to transpire in a much slower fashion without the arrival of today’s HTS methodologies).

Figure 2.6 Practice of medicinal chemistry (MC) in the new millennium. The most striking dif-

ferences from the long-standing practice of MC are (1) data reduction of huge amounts of rapidly

derived HTS biological results, (2) greater emphasis on multitechnique chemical structure consider-

ations, and most important, (3) the simultaneous attention given to all of the ADMET-related param-

eters along with efﬁ cacy and efﬁ cacy-related selectivity (E/S) during lead compound selection and

further design or enhancement, coupled with an expanding knowledge base that offers the possibility

for achieving synergistic beneﬁ ts by taking advantage of various combinations of multiagent, pro-

drug, soft drug, and/or multivalent drug strategies.

40 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

medicinal chemistry’s pursuit of distinct efﬁ cacy-related biochemical scenarios. Because

they will continue to represent part of the basic thrust of medicinal chemistry, the pursuit

of efﬁ cacy into the new millennium and the assessment of molecular conformation are

each considered further in the next two sections relative to medicinal chemistry’s working

deﬁ nition. Similarly, because the nature of the ADMET area’s maturation is regarded as

being pivotal toward shaping medicinal chemistry’s evolution in the new millennium, each

ADMET area is covered within a separate, ﬁ fth section. Although space limitations dictate

that these topics can be highlighted only in an abbreviated manner, a somewhat longer and

more technically oriented discourse is provided for the molecular conformation section

because it is envisioned that this area will constitute the alphabet for the universal language

that medicinal chemistry will help to elaborate in order to bridge and ultimately unite all

of the other areas.

2.4 PURSUING EFFICACY

From an experimental point of view, medicinal chemistry’s pursuit of the efﬁ cacy dis-

played by small molecules in a direct manner probably won’t change much in the near

term beyond what has already been proposed in terms of logical extensions from some

of the latest trends. For example, even the intriguing directions that are encompassed by

chemical genomic

strategies, where in some cases a small molecule may provide efﬁ cacy

in an indirect manner through its interactions with biochemically modiﬁ ed genomic and

proteomic signaling systems,

88–90

still reduce to being able to exploit the same fundamental

principles already cited within medicinal chemistry’s deﬁ nition so as to achieve the speciﬁ c

interaction with the modiﬁ ed biological surface, which then serves as a mediator toward

efﬁ cacy.

In terms of targets, bioinformatics will certainly associate pathophysiology and in-

dividual variation with useful genomic and proteomic information so as to maintain the

plethora of traditional and novel pharmacological targets well into the millennium.

91–98

Web-based public efforts

and commercial databases

100

coupled with desktop programs

101

are already being positioned to make such information readily available to everyone. The

ability of infectious microorganisms and viruses to evolve into resistant forms at a pace

at least equal to our ability to produce NCE chemotherapeutic agents can also be counted

on to continually provide new targets.

102,103

Alternatively, longer term into the new mil-

lennium, gene therapy will hopefully have eradicated many of today’s targets that derive

strictly from hereditary, gene-based abnormalities.

Somewhat along these same lines, the

ongoing characterization of single nucleotide polymorphisms (SNPs) and the further pur-

suit of individually tailored therapies, as presently being promoted by the new ﬁ eld of

pharmacogenetics, may also contribute toward some new targets.

104–111

Web-based public

and commercial databases in the pharmacogenetic

112

and speciﬁ c SNP areas,

113

along with

commercial sources for the latter’s experimental technologies,

114

are also making this type

of information generally available. In the near term, the ﬁ eld of pharmacogenetics is likely

to have its ﬁ rst major impacts on reﬁ ning clinical studies for late-stage preclinical can-

didates and on developing improved indications, contraindications, and dosage regimens

for marketed drugs and for compounds undergoing clinical study. Longer term, however,

pharmacogenetics should become instrumental in shaping the overall nature and subtleties

of the efﬁ cacy targets that are pursued rather than the numbers of new targets that relate

to novel mechanisms that might be deployed. For example, although one can speculate

that the present trend to pursue curative rather than palliative treatments will continue for

at least the short term, it is likely that in the longer term a growing emphasis will be

placed on preventative rather than on either palliative or curative treatments.

115,116

Toward

this end, pharmacogenetics should become of central importance, owing to its potential to

divide recipient populations into distinct treatment subgroups based on their predisposition

proﬁ les coupled with their general ADMET-related drug handling proﬁ les, wherein both

sets of criteria may eventually become accessible before or shortly after birth using gene-

based assays and to a lesser extent, administration of diagnostic probe molecules.

117,118

Depending on the variability of an individual’s environmental exposures, it can be imag-

ined that in the future, pharmacogenetic proﬁ ling will be done at routinely scheduled inter-

vals through the entire course of one’s life. Regardless of the number and nature of future

pharmacological targets, advances in biotechnology can be expected to continue to ﬂ ood

the overall life sciences arena and to get even better at deriving the required HTS assays.

Alternatively, HTS microengineering,

119

it would seem, may have to level off at about

9600-well (or well-less) tests per plate or, perhaps, move to other platforms involving chip

or bead technologies.

120

2.4.1 Gathering Positive, Neutral, and Negative SARs During HTS

As mentioned, HTS efﬁ cacy hits per se can certainly be pursued without the aid of me-

dicinal chemistry. Indeed, one can imagine that with one or more compound libraries

already in hand from an automated synthesis,

121

and the areas of robotics

122–124

and labo-

ratory information management or LIMS

125

also continuing to evolve rapidly, HTS in

the brute force mode may be able to proceed without any signiﬁ cant human intervention,

let alone without the need for an interdisciplinary group of investigators from a variety

of disciplines. However, as has been emphasized, if the new drug discovery paradigm

is ultimately to become successful, this type of screening will need to be accompanied

by structure-associated knowledge generation and assessment, with the latter being con-

ducted using the rationale and logic that can only be interjected by human intervention.

Furthermore, even though it could run the risk of placing medicinal chemistry into a fall-

guy position somewhat analogous to its earlier adolescent phase, it should also be noted

that according to the working deﬁ nition cited herein, as soon as any knowledge assess-

ments become at all sophisticated in terms of molecular structure and biological proper-

ties, they quickly fall right into the middle of the domain of medicinal chemistry and its

distinct area of expertise.

In this regard, it becomes important to brieﬂ y review some aspects of SARs that would

be worthwhile to include within the database assemblies that are currently being drawn up

to handle the mountains of data already arriving from today’s HTS programs. One can pre-

dict that once ADMET proﬁ ling by HTS is validated in the future, it will become extremely

valuable for a knowledge-generating paradigm to be able to discern not just the most active

compounds within an efﬁ cacy database and to be able to compare their structural patterns

to those in another database, but to also be able to ﬂ ag the regions on compounds that can

be altered with little effect on the desired biological activity as well as those areas that are

intolerant toward structural modiﬁ cation. The neutral areas, in particular, represent ideal

points for seamless merging of one set of a database’s hits with that of another regard-

less of the degree of pattern overlap, or for further chemical manipulation of a hit so as to

PURSUING EFFICACY 41

42 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

adjust it to the structural requirements deﬁ ned by another data set that may be so distant in

structural similarity space that attempted overlap or pattern recognition routines are oth-

erwise futile. The regions that are intolerant of modiﬁ cation represent areas to be avoided

during knowledge-based tailoring of an efﬁ cacy lead. Alternatively, the intolerant regions

represent areas that can be exploited when attempting to negate a particular action (e.g.,

metabolism or toxicity). An actual example of utilizing both neutral and negative SARs to

advantage is provided below to further illustrate how these types of data sets might also be

deployed simultaneously by future medicinal chemists (albeit with signiﬁ cantly stepped-

up complexity) as more and more parameters become added to the process of early lead

identiﬁ cation/optimization.

2.4.2 Example Involving Multidrug Resistance of Anticancer Agents

The investigations to be exempliﬁ ed in this case have been directed toward studying the

SAR associated with biological transporter systems with the hope of establishing a data-

base of transportophore relationships that might be generally applicable toward enhanc-

ing the selection and/or development of efﬁ cacy leads from any other type of data set.

To provide immediate relevance to this long-term project, it has initially focused on the

P-glycoprotein pump (Pgp)

126,127

that is associated, in part,

128,129

with the development

of multidrug resistance (MDR)

130,131

during cancer chemotherapy. Pgp is a 170-kD trans-

membrane glycoprotein that serves as an energy-dependent unidirectional efﬂ ux pump

having broad substrate speciﬁ city. In humans it is encoded by the MDR gene, MDR1,

whose classical phenotype is characterized by a reduced ability to accumulate drugs intra-

cellularly, and thus the deleterious impact of Pgp activity on cancer chemotherapy.

132–136

By way of practical example, the cytotoxicity of paclitaxel (PAC) is decreased by nearly

three orders of magnitude when breast cancer cell lines become subject to MDR, largely

via a Pgp mechanism.

137

In order to explore a series of probes that will systematically span

a speciﬁ ed range of physicochemical properties when coupled to the PAC framework, one

needs ﬁ rst to identify a region on PAC that is tolerant toward such modiﬁ cation in terms

of PAC’s inherent efﬁ cacy (i.e., where changes are known not to signiﬁ cantly alter PAC’s

cytotoxicity toward nonresistant breast cancer cells). Scheme 2.3 provides a summary of

the accumulated SAR data obtained from the PAC-related review literature,

138–142

wherein

it becomes clear that several positions along the northern edge represent neutral areas that

can lend themselves toward such an exploration.

Inhibitors of Pgp have already been identiﬁ ed by several different investigators, and

these types of compounds belong to a class of agents referred to as chemosensitizer drugs

for which there are a variety of mechanisms.

143,144

Although Pgp inhibitors can be coad-

ministered with a cytotoxic agent in order to negate MDR toward the latter when studied

in cell culture, to date these types of chemosensitizers have not fared well clinically.

145

One of the reasons that the inhibitors have not fared well is that they must compete with

the accompanying cytotoxic agent for access to the Pgp MDR receptors. Thus, it can

be imagined that if an SAR can be identiﬁ ed that is unfavorable for binding with Pgp

MDR receptors, and furthermore, that if such a negative transportophore could be incor-

porated onto the original cytotoxic agent in a neutral position, the cytotoxic agent might

itself avoid MDR or at the very least become better equipped to do so in the presence

of a coadministered Pgp MDR inhibitor (Fig. 2.7).

146

Toward this end, initial studies

are being directed toward exploring the possibility that it may be feasible to identify

negative SAR that is undesirable to the Pgp system within the speciﬁ c chemical context

of PAC by manipulating the latter at neutral positions that do not signiﬁ cantly affect

PAC’s inherent efﬁ cacy. To ascertain the generality toward potentially being able to place

such a negative transportophore onto other established chemotherapeutic agents and onto

lead compounds being contemplated for preclinical development, an identical series of

negative SAR probes is being examined within the context of a completely different mo-

lecular scaffold: that of the camptothecin (CPT) family of natural products for which

topotecan represents a clinically useful anticancer drug.

147

CPT, accompanied by a sum-

mary account of its SAR-related literature,

138,147–170

is depicted in Scheme 2.4, wherein

it becomes clear that the 7- and 9-positions represent the key neutral areas in CPT that

might be manipulated analogously to those in PAC. Since these two compounds have

very different molecular templates and owe their cytotoxicities to two distinctly different

mechanisms (i.e., PAC largely, but not exclusively

171–173

to overstabilization of microtu-

bules

174,175

and CPT largely, but not exclusively,

176

to “poisoning” of topoisomerase I

177

and because topotecan, a clinically deployed CPT analog (Scheme 2.4), is at the lower

end of the spectrum in terms of being subject to Pgp-related MDR (it loses about one or-

der of magnitude from its initial potency

137

), taken together these two molecules represent

an excellent pair to examine the generality of the transportophore-related SAR ﬁ ndings.

Other molecular scaffold systems and biological testing models can also be imagined so

as to extend such Pgp investigations into the areas of drug absorption, uptake into hepatic

tissue (drug metabolism), and passage across the blood–brain barrier (e.g., for either en-

hancing or attenuating drug penetration into the CNS). Additional transporters within the

ABC class can be explored systematically using a similar approach.

Although this particular example reﬂ ects a rational SAR strategy, the same types of

informational endpoints can certainly be achieved via a coupled HTS/combinatorial

Scheme 2.3 Overall SAR proﬁ le for paclitaxel-related compounds. This summary represents a

consolidation of SAR information contained in several review articles.

138–142

Note the tolerance for

structural modiﬁ cation along the northern hemisphere of the taxane ring system. Paclitaxel has R  ∅

and R′  CH

CO.

PURSUING EFFICACY 43

44 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Scheme 2.4 Overall SAR proﬁ le for camptothecin-related compounds. This summary represents a

consolidation of SAR information contained in several primary references.

138,147–170

Note the toler-

ance for structural modiﬁ cation along the northern hemisphere of the overall molecule, especially

when approaching the western edge. Camptothecin, the parent natural product, has R

 R

 H.

Topotecan, used clinically, has R

 OH and R

 CH

N(CH

)

Panel A

Panel B

Panel D

Panel C

Drug

Membrane Pgp

Transporter System

Cancer Cell

Chemosen-

sitizer

Chemosen-

sitizer

Figure 2.7 Diminishing the Pgp transporter’s role in multidrug resistance (MDR). Panel A: Extru-

sion of chemotherapeutic agents (drug) by cancer cells upon overexpression of the Pgp transporter

system is one mechanism associated with MDR. Panel B: Coadministration of a chemotherapeutic

agent and a chemosensitizing agent wherein the latter preferentially interacts with Pgp, thus dimin-

ishing the extrusion of the desired drug. The chemosensitizer may block the pump competitively

(wherein it becomes extruded) or noncompetitively. Panel C: Administration of a modiﬁ ed che-

motherapeutic agent (drug) that retains its desired efﬁ cacy-related properties but has a diminished

afﬁ nity for the Pgp system. Panel D: Coadministration of a modiﬁ ed chemotherapeutic agent and a

chemosensitizing agent such that the latter has a better chance of interacting with the Pgp system

relative to the desired chemotherapeutic agent.

chemistry approach, provided that the chemical structure components within the resulting

databases are initially constructed with architectures ﬂ exible enough to allow for such

queries. Likewise, while this example reﬂ ects a simple query between two different bio-

logical behaviors, the same types of queries can be conducted across multiple databases

for multiple parameters. That the next 25 years of medicinal chemistry will involve a

considerable amount of making sense out of such multiple parameter correlations based

on experimentally derived data is quite clear. That the next 50 to 75 years might then be

able to be fruitfully spent in more of a virtual correlations mode is certainly more specula-

tive but is, at least, probably reasonable provided that we can build our knowledge base

and fundamental understanding of how the various parameters, as assessed in isolation

according to the HTS scenarios described above, interact simultaneously within the entire

system.

2.4.3 Compound Libraries: Example of Working with Nature to Enhance

Molecular Diversity

Before closing this section, two additional trends need to be mentioned. The ﬁ rst involves

the likelihood that industry will even more heavily embrace site-directed mutagenesis as an

additional component of its efforts to identify lead compounds. Such modiﬁ cation can be

contemplated during the initial development of an HTS assay and then used to contribute

to the deﬁ nition of an overall pharmacophore as the latter is probed via various compound

testing paradigms. The importance of gaining a thorough appreciation for the overall phar-

macophore rather than for just identifying distinct lead structures is discussed in later sec-

tions. The second trend in the efﬁ cacy arena that is also likely to become very important in

the future can be illustrated by an example from a different research program. Instead of

focusing on ADMET-related parameters during rational drug design, this program involves

the chemical or library side of the new drug discovery paradigm. In particular, this exam-

ple seeks to enhance molecular diversity

178,179

along phytochemical structural themes that

have shown activity of either a toxic

180

or a promising nature during initial efﬁ cacy screen-

ing such that having related compound libraries would be highly desirable. As described

below, this particular example has a certain appeal in its ultimate practicality since it seeks

initially to produce directed molecular diversity within common plants indigenous to the

midwestern United States. This possibility is being explored by exposing plants simul-

taneously to both an elicitor (botany’s designation for an inducing agent) and selected

biochemical feedstocks. For example, the biochemical pathways leading to the anticancer

phytoalexins from soybeans shown in Scheme 2.5 may be able to be elicited by soybean

cyst nematode infections to produce a more diverse family of active principles when grown

in environments containing biochemically biased nutrients.

181,182

Toward this end, it has

been established that the statistical reproducibility of HPLC-derived phytochemical con-

stituent ﬁ ngerprints from soybean controls is adequate to discern real ﬂ uctuations in these

types of natural products.

183

Work is now progressing toward ascertaining the differences

that result upon exposures of soybeans to various stimuli and feedstocks. Interesting results

will be followed-up by studying the genetic control of the involved pathways. In this re-

gard it should be noted that an opposite approach that leads to similar combinatorial bio-

synthesis

184

endpoints is also being undertaken by various other groups, particularly with

an interest toward the production of proteins and peptide families from plant systems.

185

In those studies, directed modiﬁ cation of the genetic regions controlling one or more es-

tablished phytochemical pathways is ﬁ rst effected, and then these types of biotechnology

PURSUING EFFICACY 45

46 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

interventions are followed up by characterization of the altered biocombinatorial expres-

sion products.

2.5 ASSESSING AND HANDLING MOLECULAR CONFORMATION

2.5.1 Chemoinformatics

Given the exponential proliferation of technical data and our increasing ability to dissemi-

nate it rapidly through a vast maze of electronic networks, it is no wonder that new systems

capable of “managing and integrating information”

186

are regarded among “the most im-

portant of the emerging technologies for future growth and economic development across

the globe.”

186,187

That information technology (IT), in turn, is now receiving high prior-

ity in all sectors is quite clear,

188–195

particularly with regard to systems directed toward

integrating bioinformatic-related information as promoted via the World Wide Web.

196

Scheme 2.5 Elicitation of directed and novel natural product families from soybean. Panel A: Nor-

mal biochemical pathways leading to the ﬂ avones (5) to the key anticancer isoﬂ avones genistein and

diadzein, wherein phenylalanine is ﬁ rst converted to cinnamic acid, p-coumaroyl CoA, and ﬁ nally to

key intermediate 4, naringenin chalcone. Panel B: Feeding unnatural starting materials such as the

aryl-substituted phenylalanine and cinnamic acid derivatives shown as 6 and 7, respectively, under

circumstances where this pathway is also being elicited by external stimuli (e.g., soybean cyst nema-

tode infections), could be expected to produce new ﬂ avone derivatives (8), isoﬂ avonic derivatives (9),

or completely novel natural product families. This diagram of the phytochemical pathway leading to

ﬂ avones and isoﬂ avones represents a composite of several references.

485–487

An array of inexpensive

analogs related to 6 and 7 is available from commercial sources.

COSCoA

(OH)

(HO)

or or

Panel A

Genistein

Daidzein

6 7 8 9

Panel B

Comlpletely

Novel Natural

Product

Families

Medicinal chemistry’s contributions toward this sweeping assessment of the future im-

portance of IT reside primarily in the area of handling chemical structures and chemical

information, a specialized exercise complicated enough to merit its own designation as a

new ﬁ eld, that of chemoinformatics.

197–200

In this regard, the increasing use of databases

to link chemical structures with biological properties has already been alluded to in terms

of both real experimental data sets and virtual compilations. Although serious strides are

being taken in this area, however, there is a signiﬁ cant need for improvement in the han-

dling of chemical structures beyond what is suggested for the immediate future by what

now appears to be occurring within today’s database assemblies. For example, that “better

correlations are sometimes obtained by using two-dimensional displays of a database’s

chemical structures than by using three-dimensional displays” only testiﬁ es to the fact that

we are still not doing a very good job at developing the latter.

201

How medicinal chemistry

must step up and rise to the challenges already posed by this situation in order to fulﬁ ll

the key roles described for its near- and longer-term future is addressed in the next several

paragraphs of this section.

Assessment of molecular conformation, particularly with regard to database-housed

structures, represents a critical aspect of chemoinformatics. While new proteins of inter-

est can be addressed reasonably well by examining long-standing databases such as the

Protein Data Bank

202

and other Web-based resources

203

for either explicit or similar struc-

tural motifs and then deploying x-ray (pending a suitable crystal), NMR, and molecu-

lar dynamic/simulation computational studies

204–209

as appropriate, the handling of small

molecules and of highly ﬂ exible molecular systems in general remains controversial.

210

As alluded to above, the only clear consensus is that treatments of small molecules for

use within database collections “have, to date, been extremely inadequate.”

211

Certainly,

a variety of automated three-dimensional (3D) chemical structure drawing programs are

available that can start from simple 2D representations by using Dreiding molecular me-

chanics or other, user-friendly automated molecular mechanics-based algorithms, as well

as when data are expressed by a connection table or linear string.

212

Some programs are

able to derive 3D structure “from more than 20 different types of import formats.”

213

Fur-

thermore, several of these programs can be directly integrated with the latest versions of

more sophisticated quantum mechanics packages, such as Gaussian 98, MOPAC (with

MNDO/d), and extended Hückel.

200,212

Thus, electronic handling of chemical structures,

and to a certain extent comparing them, in 3D formats has already become reasonably well

worked out.

197–200,212–215

Table 2.8 provides a list of some of the 3D molecular modeling

products that have become available during the 1990s.

Nevertheless, a fundamental problem remains: how the 3D structure is derived ini-

tially in terms of its chemical correctness based on what assumptions might have been

made during the process. Further, there are still challenges associated with how readily

3D structure information can be linked with other, nonchemical types of informational

ﬁ elds. As has been pointed out by others, the reason that such mingling of data ﬁ elds

often does not afford good ﬁ ts is because “each was initially designed to optimize some

aspect of its own process and the data relationships and structures are not consistent.”

197

At this point, inexpensive Web-based tools that can integrate chemical structure data

with other types of information from a variety of sources, including genomic data, have

already begun to emerge.

216

This trend will continue to pick up into the new millen-

nium, and because of its overall importance for bioinformatics, user-friendly solutions

are likely to arrive early on.

ASSESSING AND HANDLING MOLECULAR CONFORMATION 47

Package Company Platform Description

Low-end sophistication

Nano Vision ACS Software Mac Simple, effective tool for viewing and rotating structures,

especially large molecules and proteins

Ball & Stick Cherwell Scientiﬁ c Mac Model building and visualization; analysis of bond distances,

angles

MOBY Springer-Verlag IBM (DOS) Model building and visualization; classical and quantum

mechanical computations; large molecules and proteins;

PDB ﬁ les

Nemesis Oxford Molecular IMB (Windows), Mac Quick model building and high-quality visualization;

geometry optimization (energy minimization)

CSC Chem. 3D/Chem 3D

Plus

Cambridge Scientiﬁ c Mac Easy-to-use building and visualization; geometry

optimization; integrated 2D program and word processing

Alchemy III Tripos Assoc. IBM (DOS, Windows), Mac Quick model building; energy minimization; basic

calculations; easy integration to high-end systems

PC Model Serena Software IBM (DOS), Mac Low cost with sophisticated calculations; platform ﬂ exibility

Midrange sophistication

CAChe Tektronic Mac Sophisticated computation tools; distributed processing

HyperChem Auto Desk IBM (Windows), Silicon

Graphics

Easy-to-use array of computation tools (classical and

semiempirical quantum mechanics)

Lab Vision Tripos Assoc. IBM (RISC-6000), Silicon

Graphics, DEC, VAX

Sophisticated but practical modeling for research

High-end sophistication

SYBL Tripos Assoc. IBM (RISC-6000), Silicon

Graphics, DEC VAX,

Sun 4, Convex

Integrated computation tools for sophisticated structure

determination and analysis; database management

CERIUS Molecular Simulations Silicon Graphics, IBM

(RISC-6000), Stardent

Titan

Suite of high-performance tools for building and simulating

properties

TABLE 2.8 Three-Dimensional Molecular Modeling Packages That Became Available During the 1990s

Source: Ref. 214.

2.5.2 Obtaining Chemically Correct 3D Structures

Unfortunately, the quick assignment of chemically correct 3D structures may not be read-

ily solvable. Recalling from the ﬁ rst sections of this chapter, medicinal chemistry has

been concerning itself with this task for quite some time. Medicinal chemistry’s interest in

chemical structure is further complicated, however, by the additional need to understand

how a given drug molecule’s conformational family behaves during its interactions with

each of the biological environments of interest. For example, as a drug embarks on its

“random walk”

217

through the biological realm (Fig. 2.8), the ensuing series of interactions

have unique effects on each other’s conformations

218,219

at each step of the journey and not

with just the step that ﬁ nally consummates the drug’s encounter and meaningful relation-

ship that is struck with its desired receptor/active site.

To track such behavior in a comprehensive manner, it becomes necessary to consider

a drug’s multiple conformational behaviors by engaging as many different types of con-

formational assessment technologies as possible, while initially taking an approach that

is unbiased by any knowledge that may be available from a speciﬁ c interacting environ-

ment. For example, the three common approaches depicted in Fig. 2.9 include (1) x-ray

(itself prone to bias from solid-state interactions within the crystal lattice); (2) solution

spectroscopic methods (i.e., NMR), which can often be done in both polar and nonpolar

media (this technique, however, being more limited by the amount of descriptive data that

it can generate); and (3) computational approaches that can be done with various levels

of solvation and heightened energy content (limited, however, by the assumptions and

approximations that need to be taken in order to simplify the mathematical rigor so as to

allow solutions to be derived in practical computational time periods). Analogous to the

simple, drawing program starting points, programs are also available for converting x-ray

ASSESSING AND HANDLING MOLECULAR CONFORMATION 49

Figure 2.8 Random walk taken by an oral drug on route to its point of efﬁ cacious contact within a hu-

man target cell. This continuum of interactions between a drug and various biological surfaces within the

human biological realm is typically divided into categories associated with ADMET and efﬁ cacy. Biolog-

ical milieu marked with an asterisk represent compartments having particularly high metabolic capabili-

ties. Blood is notably high in esterase capability. In the future, medicinal chemists will utilize knowledge

about ADMET-related SARs to more effectively identify the best drug leads and to further enhance the

therapeutic proﬁ les of selected compounds. (The phrase “random walk” is taken from ref. 217.)

50 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

and NMR data into 3D structures (e.g., ref. 220). While such a three-pronged approach

is not new,

221

it is emphasized herein because today’s medicinal chemistry literature sug-

gests that some investigators still fall into the single technique trap from which further

extrapolations of data are then sometimes made with great conviction. This may be be-

cause it is often difﬁ cult to obtain an acceptable crystal for x-ray analysis, have adequate

solubility for high-ﬁ eld conformational analysis by NMR, or perhaps, to become aligned

with appropriate computational expertise and computing power. At any rate, practical

advances in all three of these areas can be expected to alleviate such implementation-

related shortcomings so that the medicinal chemist of the near-term future will be more

readily able to consider structures from at least a three-pronged starting point either in-

dependently or through collaboration with other specialists and experts dedicated to each

of these areas. A real example that serves to further illustrate how entry structures might

be handled and matured using a computational approach is provided below. This prob-

lem pertains to the consideration of structures to be placed in a human drug metabolism

database.

221

2.5.3 Inﬂ uence of Biological Environments: Example Involving Drug Metabolism

This example is also cited in Chapter 9 of this text. Structures are initially considered as

closed-shell molecules in their electronic and vibrational ground states, with protonated

and unprotonated forms, as appropriate, also being entered. If a structure possesses tau-

tomeric options or if there is evidence for the involvement of internal hydrogen bonding,

the tautomeric forms and the hydrogen-bonded forms are considered additionally from

the onset. Determination of 3D structure is carried out in two steps. Preliminary geometry

optimization is affected by using a molecular mechanics method. For example, in this case

the gas-phase structure is determined by applying the MacroModel 6.5 modeling package

running on a Silicon Graphics Indigo 2 workstation with modiﬁ ed (and extended) AMBER

parameters also being applied from this package. Multiconformational assessment us-

ing systematic rotations about several predeﬁ ned chemical bonds with selected rotational

Figure 2.9 Techniques employed to assess conformational detail. X-ray diffraction requires a

suitable crystal, and its results are subject to solid-state interactions. Computational paradigms are

most accurate when done at the highest levels of calculation, but these types of calculations become

computer-time intensive. NMR requires that the molecule be soluble in the chosen solvent and that

adequate compound supplies be available. Mass spectrometry is also becoming an important tool for

larger molecules, although it provides smaller amounts of descriptive data. A composite of all ap-

proaches provides for the best possible assessment of molecular conformation.

angles is then conducted to deﬁ ne the low-energy conformers and conformationally ﬂ ex-

ible regions for each starting structure. In the second step, the initial family of entry struc-

tures are subjected to ab initio geometry optimizations, which in this case use a Gaussian

98 package running on a T90 machine housed in a state-level supercomputer center re-

source. Depending on the size of the molecule, 3-21G* or 6-31G* basis sets

223

are used for

conformational and tautomeric assessments. Density functional theory using the B3LYP

functional

224

is applied for the consideration of exchange correlation energy while keeping

the required computer time at reasonable levels. The highest-level structure determination

is performed at the B3LYP/6-31G* level. To ascertain the local energy minimum character

of an optimized structure, vibrational frequency analysis is carried out using the harmonic

oscillator approximation. Determination of vibrational frequencies also allows for obtain-

ing thermal corrections to the energy calculated at 0 K. Free energies are then calculated at

310 K (human body temperature). The latter values become particularly important for cases

where structural (conformational or tautomeric) equilibria occur.

From the calculated relative free energies, the gas-phase equilibrium constant and the

composition of the equilibrium mixture can be directly determined. Although these values

may not be relevant in an aqueous environment or in the blood compartment, the calculated

conformational distribution is relevant for nonpolar environments such as may be encoun-

tered when a drug passively traverses membranes or enters the cavity of a nonhydrated

receptor/enzyme active site just prior to binding. Repetition of this computational scheme

from biased starting structures based on actual knowledge of the interacting biological

systems or from x-ray or NMR studies (particularly when the latter have been conducted

in polar media), followed by studies of how the various sets of information become in-

terchanged and how they behave when further raised in energy, complete the chemical

conformational analysis for each structure being adopted into the human drug metabolism

database.

As mentioned earlier, after taking an unbiased structural starting point, medicinal chem-

istry needs especially to consider structures (and the energies thereof) by ascertaining

what their relevant conformations might be during interactions within various biological

milieus. It can be imagined that at least within the immediate future, a useful range of such

media to be considered will include aqueous solutions at acidic and neutral pH, namely at

ca. 2 (stomach) and 7.4 (physiological), respectively; one or more lipophilic settings, such

as might be encountered during passive transport through membranes; and ﬁ nally, speciﬁ c

biological receptors and/or enzyme active site settings that are of particular interest. Impor-

tantly, with time this list can then be expected to grow further so as to also include several

distinct environmental models deemed to be representative for interaction with various

transportophore relationships; several distinct environmental models deemed to be relevant

for interaction with speciﬁ c metabophore relationships such as within the active site of a

speciﬁ c cytochrome P450 metabolizing enzyme; and ﬁ nally, several distinct environmental

models deemed to be relevant for interaction with speciﬁ c toxicophore relationships. It

should also be appreciated that the interaction of even just one ligand within just one of the

various biological settings could still involve a wide range of conformational relationships

wherein the biological surface may also exist as an equilibrium mixture of various confor-

mational family members. If x-ray, NMR, and so on, can be deployed further to assess any

one or combination of these types of interactions, a composite approach that deploys as

many as possible of these techniques will again represent the most ideal way to approach

future conformational considerations within the variously biased settings. Advances to-

ward experimentally studying the nature of complexes where compounds are docked into

ASSESSING AND HANDLING MOLECULAR CONFORMATION 51

52 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

real and model biological environments are proceeding rapidly in all of these areas, with

11–13

and microcalorimetry

14,15

now adding themselves alongside x-ray (later discus-

sion) and NMR

225,226

as extremely useful experimental techniques for the study of such

SARs. Besides the experimental approaches, computational schemes will probably always

be deployed because they can provide the relative energies associated with all of the differ-

ent species. Furthermore, computational methods can be used to derive energy paths to get

from the ﬁ rst set of unbiased structures to a second set of environmentally accommodated

conformations in both aqueous media and at biological surfaces. Importantly, these paths

and their energy differences can then be compared along with a direct comparison of the

structures themselves, while attempting to uncover and deﬁ ne correlations between chemi-

cal structure and some other informational ﬁ eld within or between various databases.

Finally, it should be noted that by using computational paradigms, these same types of

comparisons (i.e., among and between distinct families of conformationally related mem-

bers) can also be made for additional sets of conformational family members that become

accessible at appropriately increased energy levels (i.e., at one or more 5 kcal/mol incre-

ments of energy) to thus address the beneﬁ cial losses of energy that might be obtained

during favorable binding with receptors or active sites.

227

These types of altered conforma-

tions can also become candidates for structural comparisons between databases. The latter

represents another important reﬁ nement that could become utilized as part of SAR queries

that will need to be undertaken across the new efﬁ cacy and ADMET-related parameters

of the future. With time, each structural family might ultimately be addressed by treating

the 3D displays in terms of coordinate point schemes or graph theory matrices.

228

This is

because these types of methods lend themselves to the latest thoughts pertaining to utiliz-

ing intentionally fuzzy coordinates

229,230

[e.g., x ± x', y ± y', and z ± z' (rather than just x, y,

and z plots)] for each atomic point within a molecular matrix wherein the speciﬁ ed varia-

tions might be intelligently derived from the composite of aforementioned computational

and experimental approaches. Alternatively, the fuzzy strategy might become better de-

ployed during the searching routines, or perhaps both knowledgeably fuzzy data entry and

knowledgeably fuzzy data searching engines handled, in turn, by fuzzy hardware,

231

will

ultimately best identify the correlations that are being sought in any given search paradigm

of the future. It should be noted, however, that for the fuzzy types of structural treatments,

queries will be most effective when the database has become large enough statistically

to rid itself of the additional noise that such fuzziness will initially create. An ongoing

example that serves to demonstrate the value of considering the dynamic energy relation-

ships associated with molecular trajectories as well as the more static conformational dis-

plays for a particular molecular interaction of interest is provided below.

2.5.4 Dynamic Energy Relationships: Example Involving a Small Ring System

As a different aspect of the aforementioned MDR-related anticancer chemotherapeutic pro-

gram, an effort has been directed toward replacing the complex scaffold of PAC (Scheme

2.3) with a very simple molecular format that still displays PAC’s key pharmacophoric

groups in the appropriate 3D orientations purported to be preferable for activity.

232–234

Toward this end, initial interest involved deﬁ ning the role of the β-acetoxyoxetane system,

particularly when the latter is adjacent to planar structural motifs. Since such systems are

rather unique among natural products

235

as well as across the synthetic literature,

236–238

ﬁ rst became necessary to study their formation within model systems relevant for this proj-

ect. 2-Phenylglycerol was synthesized

239

and deployed as a model to study the molecule’s

conformation by x-ray, NMR, and computational techniques as a prelude to affecting its

cyclization.

240

The energy differences that result as the molecule is reoriented so as to be

lined up for the cyclization were also calculated. Finally, once properly oriented in 3D

space, the energy required to actually traverse the S

2 reaction trajectory between the 1

and 3 positions was calculated (one of which positions utilizes its oxygen substituent for

the attack while the other relinquishes its oxygen as part of a leaving group). The synthesis

of 2-phenylglycerol and the pathway and energies associated with the intermediate species

and cyclization process to form the oxetane are summarized in Scheme 2.6.

Not surprisingly, given the strained-ring nature of this system, the energy needed to

effect the ring closure from the lowest of three closely related local minima conforma-

tions belonging to a family common to the independent x-ray and computationally de-

rived starting points was about 28 kcal/mol. What becomes interesting, however, is that

within this particular system, nearly half of this energy requirement results from the need

to disrupt hydrogen bonds in order to reorient the molecule initially into a conformation

ASSESSING AND HANDLING MOLECULAR CONFORMATION 53

Scheme 2.6 Synthesis of 2-phenylglycerol and investigation of its conversion to 3-hydroxy-3-

phenyloxetane. 2-Phenylglycerol (10) is depicted so as to convey the lowest-energy structure of the

three close local minima observed during ab initio calculations performed at the HF/6-31G* level.

TS1 and TS2 represent transition conformers obtained after the indicated bond rotations, while 11

represents the desired 3,3-disubstituted oxetane sytem. The respective relative energies in kcal/mole

for 10, TS1, and TS2, along with the product oxetane (11), are as follows: 0.00, 13.6, 12.4, and 28.2.

(From refs. 239, 240, and 488.)

54 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

appropriate for the reaction. Despite the resulting strain that becomes placed upon the

overall system’s bond angles, the actual movement of the relevant atoms along the reaction

trajectory (Scheme 2.6, dashed line) then accounts for only slightly more than one-half

of the total reaction energy. Therefore, from a synthetic point of view the results suggest

that it should be beneﬁ cial to employ a hydrogen bond acceptor solvent that has a high

boiling point, the ﬁ rst property assisting in disruption of the hydrogen bonds that need

to be broken for conformational reorientation and the second property allowing enough

thermal energy to be added conveniently to prompt progression across the reaction trajec-

tory. That such favorable conformational perturbations could indeed be achieved simply by

deploying these types of solvents was then conﬁ rmed by reexamination of the independent

results obtained from our initial NMR studies conducted in polar protic media. Alterna-

tively, from a medicinal chemistry point of view, these results serve as a reminder about

the long-standing arguments pertaining to the importance and energetics of drug desolva-

tion prior to receptor–active site interaction and, alternatively, the roles that stoichiometric

water molecules can play within such sites. That such concerns will be addressed in a

much more deliberate manner in the future by using multidisciplinary approaches similar

to the example cited herein seems clear. Indeed, a quick survey of the present medicinal

chemistry literature suggests that consideration of the dynamic nature of conformational

perturbations associated with efﬁ cacious events is already beginning to take hold.

241–246

is predicted, however, that it will be even more critical in the future to correlate SARs from

one database to another according to the dynamic energy differences between the various

molecule’s conformational family members when several ADMET-related interactions are

additionally factored into the overall behavior of a molecule being contemplated for further

development. In other words, simple comparisons of static structures, even when rigorously

assigned in 3D, will probably not be adequate to address a molecule’s behavior across all of

the efﬁ cacy and ADMET-related biological surfaces that become of interest as part of the

molecule’s optimization during future new drug design and development paradigms.

What this section points to is that, ultimately, structural databases of the future will

probably have several “tiers”

247

of organized chemical and conformational information

available which can be distinctly mined according to the speciﬁ ed needs of a directed

(biased) searching scheme while still being able to be completely mixed within an over-

all relational architecture such that undirected (unbiased) knowledge-generating mining

paradigms can also be undertaken.

248–253

Certainly, simple physicochemical data will need

to be included among the parameters for chemical structure storage. Similarly, searching

engines will need to allow for discrete substructure queries as well as for assessing overall

patterns of similarity and dissimilarity

254–261

across entire electronic surfaces.

2.5.5 Druglike Properties and Privileged Structures

It can be noted that it is probably already feasible to place most clinically used drugs into

a structural database that could at least begin to approach the low- to midtier levels of

sophistication because considerable portions of such data and detail are probably already

available in the literature for each drug even if they are spread across a variety of techni-

cal journals. On the other hand, it should also be clear that an alternative strategy will be

needed to handle the mountains of research compounds associated with just a single HTS

parameter survey. Unfortunately, it appears that some of the large compound surveys be-

ing conducted today do not even have systematically treated 2D structural representations.

Indeed, while the present status of handling chemical structure and data associated with

HTS is wisely being directed toward controlling the size of the haystack,

262

the dire status

of handling conformational detail is reﬂ ected by attempts that try to grossly distinguish be-

tween druglike and nondruglike molecules

263

in a 2D manner or, at best, to identify certain

“privileged structures”

264,265

while using 3D constructs derived from less than completely

rigorous experimental and computational assessments. Furthermore, in certain companies,

notions about druglike patterns (or actually the lack thereof) are already being set up as the

ﬁ rst screen or in silico ﬁ lter to be deployed against a given compound library’s members

while the latter are still on route to an HTS efﬁ cacy screen. Unfortunately, this scenario can

detract from the deﬁ nition of an initial efﬁ cacy pharmacophore along structural motifs that

might, alternatively, be able to take advantage of neutral areas by making straightforward

chemical modiﬁ cations that then serve to avoid the non-druglike features. At present, and

probably for much of the near term as well, strategies that use non-druglike parameters

to limit the number of compounds that can otherwise contribute toward the deﬁ nition of

a given efﬁ cacy-related structural space would appear to be premature. At the very least,

such strategies are counter to the need to continue to accumulate greater knowledge in the

overall ADMET arena, let alone in the speciﬁ c handling of 3D chemical structure at this

particular time. Finally, when it is appreciated that in most cases the connection of HTS

ADMET data with actual clinical outcomes still remains to be validated much more se-

curely, the strategy to deploy notions about non-druglike structural hurdles as decision steps

prior to efﬁ cacy screening becomes reminiscent of medicinal chemistry’s own adolescent

phase, wherein medicinal chemistry’s efforts to design drugs rationally without the beneﬁ t

of the additional knowledge afforded by an x-ray of the actual target site ultimately did not

enhance either the production of NCEs or the image of medicinal chemistry. A more ap-

propriate strategy toward addressing this area that is knowledge building and, instead, can

eventually expect to deploy the evolving ADMET druglike patterns in a proactive manner

is discussed further in Section 2.6. With regard to chemical structure, the present situation

thus indicates that we have a long way to go toward achieving the aforementioned tiers of

conformational treatments when dealing with large databases and applying them toward

the process of drug discovery. Nevertheless, because of the importance of chemoinformat-

ics toward understanding, fully appreciating, and ultimately, actually implementing bioin-

formatics along the practical avenues of new drug discovery, it can be imagined that future

structural ﬁ elds within databases, including those associated with HTS, may be handled

according to the following scenario, as summarized from the ongoing discussion in this

section and as also conveyed in Fig. 2.10.

2.5.6 Tiered Structural Information and Searching Paradigms

For optimal use in the future, it is suggested that several levels of sophistication will be

built into database architectures so that a simple 2D format can be input immediately.

Accompanying the simple 2D structure ﬁ eld would be a ﬁ eld for experimentally obtained or

calculated physicochemical properties (the latter data also to be upgraded as structures are

matured). While this simple starting point would lend itself to some types of rudimentary

structure-related searching paradigms, the same compound would then gradually progress

by further conformational study through a series of more sophisticated chemical struc-

ture displays. As mentioned earlier, x-ray, NMR, and computational approaches toward

considering molecular conformation will be deployed for real compounds given that it is

also likely that advances in all of these areas will allow them to be applied more readily in

each case. Obviously, virtual compound libraries and databases will have to rely solely on

ASSESSING AND HANDLING MOLECULAR CONFORMATION 55

56 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

computational approaches and on knowledgeable extrapolation from experimental data de-

rivable by analogy to structures within overlapping similarity space. Eventually, structures

would be manipulated to a top tier of chemical conformational information. This tier might

portray the population ratios within a conformational family for a given structure entry

expressed as both distinct member and averaged electrostatic surface potentials, wherein

the latter can be expanded further to display their atomic orientations by fuzzy graph theory

or fuzzy 3D coordinate systems. Thus, at this point one might speculate that an intelli-

gently fuzzy coordinate system could eventually represent the highest level of development

for the 3D quantitative SAR (3D-QSAR)

266,267

-based searching paradigms seemingly ris-

ing to the forefront of today’s trends in the form of comparative molecular ﬁ eld analyses

(CoMFA).

268,269

Furthermore, one can imagine that this tier might actually be developed

in triplicate for each compound: that is, one informational ﬁ eld for the environmentally

unbiased structural entries, another involving several subsets associated with known or

suspected interactions with the biological realm, and a third for tracking conformational

families when raised by about 5 and 10 kcal/mol in energy. Finally, as the new millennium

Figure 2.10 Handling chemical structures within databases of the future. This ﬁ gure depicts the

quick entry and gradual maturation of structures. Search engines, in turn, would also provide for a

variety of ﬂ exible paradigms involving physical properties with both full and partial (sub)structure

searching capabilities using pattern overlap/recognition, similarity–dissimilarity, CoMFA, and so

on. Structure entry would be initiated by a simple 2D depiction that is gradually matured in con-

formational sophistication via experimental and computational studies. Note that structures would

be evolved in both an unbiased format and in several environmentally biased formats. The highest

structural tier would represent tracking or searching the energies required for various conformational

movements that members would take when going from one family to another.

continues to churn its computational technologies forward, conformational and energetic

considerations pertaining to a compound’s movement between its various displays, similar

to that conveyed by the very simple example provided from the oxetane-related study, can

also be expected to be further reﬁ ned so as to allow future characterization and searching

of the dynamic chemical events that occur at the drug–biological interface (e.g., modes and

energies of docking trajectories and their associated molecular motions relative to both li-

gand and receptor/active site). That this top tier is extremely valuable for understanding the

interactions of interest to medicinal chemistry is apparent from the large amount of effort

already going on today in this area,

241–246

particularly when such studies are able to take

advantage of an x-ray-derived starting point.

By the same token, chemical structure search engines of the future will probably be set

up so that they can also be undertaken at several tiers of sophistication, the more sophis-

ticated requiring more expert-based enquiries and longer search times for the attempted

correlations to be assessed. A reasonable hierarchy for search capability relative to the

structural portion of any query might become (1) simple 2D structure with and without

physicochemical properties, (2) 3D structure at incremented levels of reﬁ nement, (3) 2D

and 3D substructures, (4) molecular similarity–dissimilarity indices, (5) fuzzy coordinate

matrices, (6) docked systems from either the drug’s or the receptor–active site’s view at

various levels of speciﬁ able precision, and ﬁ nally, in the more distant future (7) energy

paths for a drug’s movement across various biological milieu, including the trajectories and

molecular motions associated with drug–receptor/active site docking scenarios. Emphasiz-

ing informatics ﬂ exibility, this type of approach, where data entry can occur rapidly for

starting structure displays and then gradually be matured to more sophisticated displays as

conformational details are accurately accrued, coupled with the ability to query at different

levels of chemical complexity and visual displays

270

at any point during database matura-

tion, should allow for chemically creative database mining strategies to be effected in the

new millennium’s near term as well as into its more distant future.

2.6 ADMET CONSIDERATIONS

2.6.1 Assuring Absorption

In addition to conducting in vivo bioavailability studies on selected compounds at a later

stage of development, early in vitro assessments of structural information that might be use-

ful toward assuring absorption after a drug’s oral administration have now been going on

for several years.

271–273

Somewhat more recently, similar studies also began to be directed

toward assessing penetration across the blood–brain barrier (BBB).

274,275

Thus, determina-

tion of the pK

values for ionizable groups, determination of partition coefﬁ cients (e.g.,

using various types of log P calculations and measurements), and measurement of passage

across models of biological membranes (e.g., Caco cell lines) represent data that have now

been shifted toward HTS experimental and purely computational modes.

276–287

These types

of studies can be designated as AHTS (absorption high-throughput screening) (Table 2.2).

Since recent results suggest that the biological transporter systems are extremely important

factors in this area,

288–290

their study is also becoming part of AHTS (e.g., passage of drugs

across Caco cell layers from both directions

291,292

). This trend toward increasing sophisti-

cation within AHTS can be expected to continue. That genomics and proteomics will help

to identify and initially deﬁ ne absorption-related systems biochemically should be clear.

ADMET CONSIDERATIONS 57

58 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Alternatively, that beyond establishing the complementary AHTS systems’ biotechnology

might also be directed toward instilling passageways or speciﬁ c pores for drugs across the

human GI endothelial system

293

is certainly speculative even for the more distant future, as

are chemical

294–296

and nanotechnology

116,297–300

approaches to prompting or constructing

passageways, respectively. Similarly, that advances in formulation and alternate delivery

technologies

301–310

could eventually obviate the need for oral administration is also specu-

lative. Nevertheless, all of these possibilities need to be mentioned because, taken together,

they make the point that signiﬁ cant advances in any of the ADMET areas, regardless of

their technological source, have the potential to eliminate the need for assaying certain of

their presently related parameters, perhaps even returning the initial portion of the present

drug discovery paradigm (Fig. 2.2) back to where it originated (i.e., to being concerned

primarily with efﬁ cacy and selectivity during front-line testing) (Fig. 2.1).

As for deciphering selective efﬁ cacy-related SARs, medicinal chemistry’s role within

the more discernible future is likely to be directed toward making sense out of the AHTS

data mountains looming ahead using molecular structure information as the common code,

in this instance by relating the latter to structure–absorption relationships, or SAbRs. Such

efforts might eventually culminate in affording molecular blueprints for affecting absorp-

tion-related structural modiﬁ cations that are correlated with certain structural themes and

absorption characteristics for which efﬁ cacy hits may be able to be categorized using struc-

tural similarity–dissimilarity indices. Notable advances have already been made toward

deﬁ ning useful SAbRs in terms of database and virtual compound proﬁ ling (e.g., the rule of

ﬁ v e ).

311

The latter should be recognized as an important ﬁ rst step in this direction that can

be expected to continue in a more sophisticated manner in the future (e.g., along the lines

of 3D structural considerations relevant to the transporter systems, as well as more reﬁ ned

parameterization of physicochemical properties).

312

2.6.2 Directing Distribution

The types of studies mentioned above, along with a panel of assays speciﬁ c for certain

depot tissues such as red blood cells, plasma protein binding factors, and adipose tis-

sue,

313–316

will be mobilized toward directing distribution of a xenobiotic. Thus, as the

handling of chemical structure improves and more sophisticated correlations begin to un-

fold in the future, AHTS can be thought of as A/DHTS that provides both SAbR and SDR.

Simultaneous collection of such data will allow investigators to reﬂ ect upon drug absorp-

tion and distribution as a continuum of drug events that can effectively be incorporated

together at an earlier point of the overall lead decision process. Furthermore, in the case

of directing distribution it can be anticipated that genomics and proteomics will become

instrumental toward identifying numerous key factors that are overexpressed in various

patho-physiological states. For example, cancer cells are already known to overexpress a

variety of speciﬁ ed factors.

317–321

Ligands designed to interact with such factors residing

on cell surfaces can then be coupled with diagnostic and therapeutic agents so as to be

delivered at higher concentrations to these locales. For therapeutics, such strategies can be

thought of as placing both an address and a message within a molecular construct

322,323

that may involve an overlap of two small moleculerelated SAR patterns, or perhaps a

small molecule conjugated to a bioengineered biomolecule wherein the latter typically

serves as the address system. Indeed, the bioconjugate or immunoconjugate strategy has

been around for a while

324

and it appears to be beneﬁ ting from a renewed interest

325

that chemotherapeutic “smart bombs”

326

are now being added to our older arsenals of

single arrows and combinations of small-molecule “magic bullets.”

327

The earlier example

involving PAC and CPT (Schemes 2.3 and 2.4) can also be used to further emphasize

this theme wherein the chemical knowledge in the area of PAC and CPT protection and

coupling reactions can be used to construct compounds that would be directed toward

some of the factors that are overexpressed on certain human cancer cells so as to en-

hance selective toxicity,

328

particularly since there is some precedent in this case that this

might be feasible by combining two small molecules. One can imagine that as data are

amassed in the future for these types of factors, the most promising ones will be quickly

pursued according to both of the aforementioned scenarios, paired small-molecule SARs

and small molecule–bioconjugate pairs. Whether undertaken in a rational manner or via

the merger of two HTS-generated databases (i.e., one for an efﬁ cacious message and one

for determining a selective address), these types of pursuits fall into the general category

of tailoring a lead. Therefore, it can be expected that the expertise afforded by medicinal

chemistry will again be an integral component of such activities. Similarly, as suggested

by the earlier PAC/CPT MDR-related example, medicinal chemistry’s expertise will also

be vital toward exploiting the opposite cases, where it becomes desirable to avoid certain

systems (addresses) that become overexpressed as part of a given pathophysiology’s resis-

tance mechanisms or because messages delivered to such locales lead to toxicity within a

healthy compartment.

Before turning to those parameters that might be considered to be associated with end-

ing a drug’s random walk through the biological realm (e.g., metabolism and excretion),

it is necessary to discuss a practical limitation to where this overall discourse is leading.

Clearly, there will be ceilings for how many molecular adjustments can be stacked into a

single compound, no matter how knowledgeable we become about the various ADMET-

related structural parameters and how they might be merged so as best to take advantage

of molecular overlaps. This will be the case even when prodrug strategies are adopted

329

(Fig. 2.11) wherein certain addresses or messages that have been added to deal with one or

more aspects of ADMET, become programmatically jettisoned along the way while simul-

taneously activating the efﬁ cacy payload that is to be delivered only to the desired locale

as the ﬁ nal statement.Thus, this situation prompts the prediction that to interact optimally

with the entire gamut of efﬁ cacy and ADMET-related parameters during a given course of

drug therapy, the latter may need to be delivered not as a single agent but as a distinct set of

multiple agents wherein each individual component or player makes a speciﬁ ed contribu-

tion toward optimizing one or more of the efﬁ cacy and ADMET parameters relative to the

overall drug team’s therapeutic game plan.

2.6.3 Herbal Remedies: Example of Working with Nature to Discover

ADMET-Related Synergies

Today’s trend to self-administer herbal remedies and preventives, admittedly driven by

rampant consumerism in the United States rather than by solid science,

330,331

thus becomes

an important topic to be considered at this point in the review. In this regard, the reconnec-

tion to medicinal chemistry’s historical roots also becomes interesting to note. One of the

major, basic science questions about herbals (which do possess validated pharmacological

properties) is why their natural forms are sometimes superior to the more puriﬁ ed versions

of their active constituents, even when the latter are adjusted to reﬂ ect varying concentra-

tion ratios thought to coincide with their natural relative abundances. Given the notoriously

incomplete analytical characterizations of most herbal products, it should be apparent rela-

tive to the present discourse that numerous unidentiﬁ ed, nonefﬁ cacious, and otherwise

silent constituents within any given herb could have an interaction with one or more of the

ADMET CONSIDERATIONS 59

60 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

efﬁ caciously active constituents at any one or more levels of the latter’s ADMET steps.

When these interactions are favorable, the resulting overall pharmacological proﬁ le be-

comes altered in a seemingly synergistic manner that is obtainable from the more natural

forms of the mixture but lost upon puriﬁ cation to matrices containing only the actives.

332

Indeed, there is already some experimental precedent for this scenario relative to efﬁ ca-

ciously silent components improving the absorption,

333

enhancing the distribution,

334,335

and favorably altering the metabolism

336

of their active herbal counterparts, as well as more

classical synergies involving direct interactions that occur at the sites involved with efﬁ -

cacy.

337

Therapeutic enhancements derived directly from multiple interactions at efﬁ cacy

sites have been pursued for many years, with multivalent single drug entities reﬂ ecting the

latest trend in this direction.

338

What will be remarkable is that the new millennium will

Figure 2.11 Soft drug actions compared to standard drug and prodrug actions. Panel A depicts a

generalized version of a standard drug’s pattern of observed activities. Panel B depicts how a prodrug

approach can be used to modify the entry-side portion of a given drug’s overall proﬁ le of actions.

Panel C depicts how a soft drug approach can be used to modify the elimination-side portion of a

given drug’s overall proﬁ le of action. Both prodrugs and soft drugs can be used to decrease toxicity.

(From refs. 329, 352, and 353.)

continue to add the sophistication of the entire ADMET proﬁ le into such multi-action-di-

rected considerations.

339–343

Optimization of the overall pharmacological proﬁ le is precisely what is being striven for

when selecting and/or chemically tailoring an NCE lead according to either the old or new

paradigm of drug discovery. Restating, however, that it may be expecting too much even

upon extending the new paradigm into the future as a knowledge-generating process, to

obtain complete optimization within a single multiparameterized molecule, perhaps it will

be nature that will again lend its hand within the next millennium by revealing some of the

modes of ADMET synergy that have long been part of some herbal productions. At the very

least, medicinal chemistry should take care not to forget its roots in natural product chem-

istry as it marches forward with biotechnology just behind genomics and proteomics into

the new millennium. For example, efforts can be directed toward uncovering efﬁ cacy and

ADMET-related synergies that may be present among the constituents of herbs purported

to have anticancer or cancer-preventive properties by taking advantage of the common cell

culture panels already in place to assess anticancer activity, along with various transporter

system interactions via HTS format. However, because anticancer/cancer-preventive syn-

ergy could derive from favorable interactions across a wide variety of ADMET processes

relative to any combination of one or more efﬁ cacy-related endpoints, several mechanism-

based assays associated with several key possibilities for efﬁ cacy will also need to be

deployed as part of such a program. One can only imagine how sophisticated this type of

pursuit will become in the future when such highly interdisciplinary efﬁ cacy networks are

coupled to an even wider network of ADMET parameter experimental protocols.

A more classical approach toward the interactions of multicomponent systems would

be to utilize clinical investigations to study the interactions, either positive or negative,

that herbals may have with drugs when both are administered to humans. For example,

Bachmann and Reese et al.

344,345

have begun to study the interaction of selected herbals

with speciﬁ c markers for several drug metabolism pathways, while V. Mauro and L. Mauro

et al.,

346,347

among others, are studying the clinical pharmacokinetic consequences of se-

lected herb–drug administrations, such as ginkgo biloba with digoxin. Importantly, for all

of these herb-related studies it becomes imperative that extensive chemical constituent ﬁ n-

gerprinting is also undertaken so that the effects observed, particularly those suggestive of

synergy, can be correlated with overall chemical composition patterns and not just with the

distinct concentrations of preselected components already known to possess established

activity.

183

In contrast to both of the aforementioned types of studies that can be considered to rep-

resent systematic examinations of herbal-directed small libraries and speciﬁ ed herb–drug

clinical combinations, it becomes interesting to speculate how a truly random brute force

approach to identifying synergy might proceed not too far down the road into the new mil-

lennium (e.g., as an HTS survey of a huge random compound library in pursuit of optimal

pair or even triple compound teams rather than as the pursuit of a single blue-chip drug that

can do it all). In this regard, however, it must ﬁ rst be recognized that the present trend to test

mixtures of several compounds within a given well does not even begin to address synergy.

This is because based on considerable experience with various chemotherapeutic agents,

348

synergy is most likely to be observed at very select ratios within very distinct concentra-

tions of the players involved. In other words, looking at the simplest case of assessing the

potential synergy between just two molecules, A and B, requires testing A in the presence

of B across a range of molar ratios presented across a range of absolute concentrations.

This situation is depicted in Fig. 2.12.

349,350

ADMET CONSIDERATIONS 61

62 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

2.6.4 Brute Force HTS to Uncover Multicomponent Synergies

Pursuing the brute force approach from a purely mathematical viewpoint and in a mini-

mally elaborated pharmacological format, suppose that the possibility for A plus B synergy

relating to just a single efﬁ cacy or ADMET-related HTS parameter is examined across a

compound library having only 100 members wherein paired combinations are tested at just

three relative molar ratios (e.g., A/B at 0.5/1, 1/1, and 1/0.5) at only three total concentra-

tions of both members (e.g., 0.1, 1.0, and 10 µM). Then a total of 44,850 drug tests plus

numerous control runs would be required for an N  1 pass through the library.

351

Perhaps

because of these rather impressive numbers, brute force HTS in the new millennium will

undoubtedly relish such pursuits. Indeed, it strikes this author very surprisingly that noth-

ing along these lines seems yet to have appeared in the literature. At any rate, once the

HTS forces do become mobilized in this area, such testing could set up an interesting

“John Henry” competition with more directed investigations, such as those that have been

elaborated above that seek systematically to identify the speciﬁ c synergies seemingly pres-

ent within certain herbals. Ultimately, no matter how the identiﬁ cation of such favorable

drug–drug partnering possibilities are uncovered and are able to better deal with the vari-

ous ADMET parameters of tomorrow, as well as for the classical efﬁ cacy relationships of

today, they will certainly prove to be invaluable toward alleviating the situation of trying

to establish all of the most desired behaviors for a given therapeutic target within the con-

text of a single molecular framework. Furthermore, it can be anticipated that this type of

Synergy

Line of simple

additivity

Antagonism

50,T

Proportion of A

Figure 2.12 Drug interaction plot for two drugs, A and B. EC

50,T

is the total concentration of the

combined drugs which gives 50% of the maximum possible effect. The EC

50,T

is shown as a func-

tion of the fraction of drug A (drug B’s fraction is 1 minus the fraction shown). Rescaling of drug

concentrations to units of their EC

values allows simple additivity to be set at unity such that devia-

tions below or above this line indicate synergism or antagonism, respectively. The dots are actual

experimental results obtained for two anticancer agents, wherein the observed EC

50,T

values reﬂ ect

20 rays of ﬁ xed drug fractions as estimated from the data along that ray alone. The ﬁ tted curve was

generated by the global model for the entire data set and indicates the complicated nature of inter-

action relationships within even a well-controlled cell culture enviroment. That synergism can be

accompanied not only by simple additivity but also by ratio-dependent antagonistic relationships is

apparent. (From ref. 348).

information will become extremely useful when it becomes further elaborated by medicinal

chemistry into general structural motifs that have potential synergistic utilities and applica-

tions beyond what was initially uncovered by the speciﬁ c mixtures of deﬁ ned compounds.

2.6.5 Controlling Metabolism: Example Involving a Soft Drug Strategy

Although aspects of drug metabolism are covered more seriously in other chapters such

as Chapter 9, there is one general area pertaining to controlling metabolism that falls so

speciﬁ cally into medicinal chemistry’s domain of lead tailoring that it merits at least a

brief discussion herein. This topic involves exploiting what has come to be called

352,353

soft drug technology (Fig. 2.11), where a metabophore is placed within an established drug

or lead compound in order to program a speciﬁ ed course of metabolism for the resulting

combination. Although nature has provided numerous examples of soft drugs, esmolol

(Scheme 2.7) has come to be regarded as the prototypical soft drug that was obtained

via rational design.

354

In this case, a methyl propionate was appended to the classical

aryloxypropanolamine template associated with β-adrenergic receptor blockade (Scheme

2.7) in order to program the latter’s metabolism along the ubiquitous esterase pathways

such that the resulting β-blocker would possess an ultrashort duration of action.

355–357

Thus, a methyl 3-arylpropionate system (boldface atoms in Scheme 2.7) represents a use-

ful metabophore already having clinical proof of principle within the molecular context of

an aryloxypropanolamine template. This metabophore can be used to program human drug

metabolism by esterases. It can be noted that the rational design of esmolol simultaneously

drew upon several of medicinal chemistry’s basic science principles mentioned thus far: (1)

negative SAR, wherein it was determined that only lipophilic or, at most, moderately polar

groups could be deployed in the aryl portion of the aryloxypropanolamine pharmacophore

if activity was to be retained; (2) electronic physicochemical properties operative within

a biological matrix, wherein it was imagined that while an ester would be permissible in

the aryl portion (neutral SAR), a carboxylic acid moiety placed in the same aryl portion

would become too foreign to be recognized by β-adrenergic receptors upon ionization of

the carboxylic acid at physiological pH; (3) general structure–metabolism relationships

(SMR), wherein it was appreciated that an ester linkage might be relied on to program a

ADMET CONSIDERATIONS 63

ONH

12 13

Scheme 2.7 Esmolol as the prototypical soft drug. Compound 12 represents the classical aryloxy-

propanolamine pharmacophore associated with blockade of β-adrenergic receptors. Compound 13 is

esmolol, a soft drug version of 12 that has been programmed to have an ultrashort duration of action

due to hydrolysis of the methyl ester by the ubiquitous esteases. The methyl 3-arylpropionate (bold in

13) thus represents a useful metabophore

for the associated human esterases.

489

(From ref. 354.)

64 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

quick metabolism; (4) steric physicochemical properties, wherein it was imagined that the

metabolic hydrolysis rate could be quickened by extending the initial ester linkages away

from the bulky aryl group such that the methyl 3-arylpropionate metabophore was identi-

ﬁ ed; and (5) appreciation for the physiologic drug excretion structural relationship (SER),

where there is a general propensity to excrete low-molecular-weight acids. These are all

fundamental physical organic principles applied in a straightforward manner within very

speciﬁ c contexts of the biological realm. Thus, this example serves four purposes. The ﬁ rst

is to emphasize again that beyond activity hits per se, neutral and negative SARs should

also be tracked so as to be readily retrievable from the databases associated with a given

parameter survey of the future. The second is to emphasize again that medicinal chemistry

will need to become an active participant in the merging of various HTS parameter sur-

veys by using chemical structure as a common denominator, especially when such activi-

ties become considerably more complicated than in the esmolol case. Third, the esmolol

case demonstrates that even when problems can be reduced to what appears to be a rather

simple set of factors, it will still be medicinal chemistry’s unique desire to systematically

characterize the complete pattern of chemical structural relationships that is likely to be

called upon to ﬁ nalize what other disciplines might consider at that point to be rather sub-

tle and mundane details. In other words, who besides a medicinal chemist can be expected

to enthusiastically pursue methyl-, ethyl-, propyl-, and so on, relationships either syntheti-

cally or by tediously purveying huge databases of the future, just to look for those SXR

“Goldilocks” situations

354

that could become relevant toward addressing a problem within

another structural setting while attempting to merge the two data sets within a common

chemical context? The latter was precisely the case for the esmolol-related metabophore

upon comparison of methyl benzoate, methyl α-phenylacetate, methyl 3-phenylpropionate,

and methyl 4-phenylbutyrate, wherein the half-lives observed for these systems when in-

corporated into the molecular context of a β-blocker pharmacophore became about 40, 20,

10, and 60 min, respectively (Fig. 2.13). More current uses of the soft drug technology are

Duration

Action

(min)

(CH

)

Figure 2.13 Relationship between methylene-extended esters and duration of action within a series

of esmolol analogs. The “Goldilocks” nature of the ethylene extension relative to the desired 10-min

duration of action is apparent. (From refs. 354 to 357.)

also underway. For example, the esterase capability in newborns was recently compared to

that of adults, and it was found that esmolol’s half-life in cord blood (baby side) is about

twice as long as that in adult blood. Furthermore, individual variation is signiﬁ cantly more

pronounced within the newborns.

358,359

These ﬁ ndings, in turn, have prompted an explora-

tion of the generality of deploying the esmolol metabophore within the chemical contexts

of several other types of therapeutic agents: namely, those that are commonly used to treat

the neonatal population in critical care settings. Thus, this fourth and last aspect of the

esmolol example clearly demonstrates the potential impact that such classical medicinal

chemistry studies can have on the new ﬁ eld of pharmacogenetics, as the latter is surely

to become further evolved within the new millennium. Analogous to the importance of

merging SAbRs and SDRs with efﬁ cacy and selectivity-related SARs, SMRs and numer-

ous metabophore patterns can be expected to be gradually discerned and put to extensive

use by medicinal chemistry in the future either to enhance or to detract from a candidate

drug’s metabolism.

2.6.6 Optimizing Excretion

Clearly, SAbRs, SDRs, and SMRs can all be used to manipule and optimize the elimination

pattern of administered drugs, the esmolol example applying here as well. Analogous to

the distribution area, genomics and proteomics can soon be expected to delineate important

systems, such as specialized transporters within tissues like the kidney and liver, that are

especially responsible for the excretion of xenobiotic drugs and their metabolites. Medici-

nal chemistry’s involvement in uncovering SERs and deriving generally useful structural

patterns that might be used to tailor lead compounds or for merging of different types of

databases while attempting to select lead compounds again falls into the central theme for

medicinal chemistry’s future as being elaborated in this review. As for the other areas, more

speculative notions in this area can provide some interesting alternatives for this aspect of

ADMET. Although SERs are also likely to encompass various endogenous materials and

their catabolic fragments, one might still imagine that just like the futuristic examples sited

for absorption, in the more distant future, biotechnology, chemical, and nanotechnology

approaches might all be used successfully to engineer speciﬁ c drug excretion passages

through selected tissues.

2.6.7 Avoiding Toxicity

It may very well be that the most profound effect that genomics and proteomics are going

to have within the ADMET arena will ultimately pertain to avoiding toxicity. Indeed, that

toxicology has now become a protagonist through its participation in the design or early

selection of drug leads already represents a remarkable turnaround from its historical an-

tagonist role as a gatekeeper or policeman standing at an advanced stage of drug develop-

ment with an eye toward halting the progression of potentially toxic compounds on route to

the clinic.

360

Like the ﬁ eld of drug metabolism (Table 2.7), toxicology has been collecting

its data within databases for quite some time (Table 2.9). In fact, some of the structural pat-

terns that have come to be associated with distinct toxicities (toxicophores) are probably

on much ﬁ rmer ground than are the metabophore relationships. On the other hand, drug

metabolism derives from a ﬁ nite number of genetic constructs that translate into metabolic

activity (albeit notorious for their seeming molecular promiscuities) such that with enough

ADMET CONSIDERATIONS 65

66 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Database or Organization Description

Centers for Health Research [formerly,

Chemical Industry Institute of Toxicology

(CIIT)]

Industry consortium-sponsored collection/

dissemination of toxicology data; also

conducts research and training in toxicology

467

American Chemistry Council Long-Range

Research Initiative

Industry consortium-sponsored initiative to

advance knowledge about the health, safety,

and environmental effects of products and

processes

468

LHASA, Ltd. (UK-based, nonproﬁ t segment) Facilitates collaborations in which companies

share information to establish rules for

knowledge bases associated with toxicology

469

International Toxicology Information Center

(ITIC)

Pilot program to share data in order to

eventually be able to predict the toxicology of

small molecules, thus lessoning the expense

of in vitro and in vivo testing

469

U.S. Environmental Protection Agency (EPA)

High-Volume Chemical (HPV) Screening

Information Data Set (SIDS)

User-friendly version that will also be

submitted to the Organization for Economic

Cooperation and Development (OECD) and

its tie-in with IUCLID

a469

SNP Consortium: nonproﬁ t; makes its

information available to public

Addresses phenotypic aspects relative to

individual responses to xenobiotics (e.g.,

metabolic phenotype and toxicity)

Tox Express/Gene Express database offered by

Gene Logic (commercial

)

Offers a gene-expression approach toward

toxicity assessment

469

National Institute for Environmental Health

Science (NIEHS)

Compiling a database of results from

toxicogenomic studies in order to divide

chemicals into various classes of toxicity based

on which genes they stimulate or repress

470

International Program on Chemical Safety/

Organization for Economic Cooperation

IPCS/OECD

Risk assessment terminology standardization

and harmonization

471,472

MULTICASE (commercial) Prediction of carcinogenicity and other potential

toxicities

473

MDL Toxicity Database (commercial) Allows structure-based searches of more

than 145,000 (Jan. 2001) toxic chemical

substances, drugs, and drug-development

compounds

DEREK and STAR (LHASA, commercial

segments)

Prediction of toxicity

474

SciVision’s TOXSYS (commercial) General toxicity database to be developed in

collaboration with the U.S. FDA

475

Phase-1’s Molecular Toxicology Platform gene

expression microarays (commercial

)

Allows detection of gene expression changes in

many toxicologic pathways

476

TABLE 2.9 Toxicology Databases and Related Organizations

Includes cooperative efforts with the European Union and the European Chemicals Bureau (ECB) in using the

International Uniform Chemical Database (IUCLID) and its relationship to high-volume chemicals (HVPs).

This company’s product is representative of several such technologies that are also being made available by a

variety of other vendors.

Includes cooperative efforts with the U.S. Environmental Protection Agency (EPA) and the Information Division

at the National Institute for Occupational Safety and Health (NIOSH).

time the entire set of metabolic options should eventually become well characterized. Toxic

endpoints, alternatively, have no such limitation associated with their possible origins. In

other words, to show that a drug and its known or anticipated metabolites are completely

nontoxic is comparable to trying to prove the null hypothesis, even when a limited con-

centration range is speciﬁ ed so as to circumvent the situation that everything becomes

toxic someplace at high enough concentration. Nevertheless, genomics, proteomics, and

biotechnology do, indeed, appear to be producing some promising technologies that can be

directed toward this area. For example, array technologies are already becoming available

to assess the inﬂ uence of a drug on enormous numbers of genes and proteins in HTS fash-

ion.

361–368

Once enough standard data of this type are produced by taking known agents up

in dose until their toxicity becomes ﬁ ngerprinted via distinctive patterns of hot spots, array

patterns may be used to cross-check against the proﬁ les obtained in the same HTS mode

for new lead compounds. Given the quick rate that these important trends are likely to be

further developed and eventually validated within the new millennium, medicinal chemis-

try could certainly become overwhelmed trying to keep up with its complementary role to

identify the corresponding STR for each array hot spot.

In the case of toxicity, then, medicinal chemistry will probably need to approach STRs

in a different manner [e.g., initially from just the exogenous compound side of the equa-

tion for a given toxicity relative to the observed hot spot patterns (unless genomics, pro-

teomics, and biotechnology also quickly step in to deﬁ ne the biochemical nature of the

actual endogenous partners that are involved in a given toxic event)]. Taking a chemically

oriented starting point, however, should serve reasonably well for at least awhile into the

new millennium in that there will likely become a ﬁ nite number of chemical reactivity pat-

terns that can be associated with toxicity. Medicinal chemistry can be expected to elaborate

these reactivities into general STRs and then to use them toward deﬁ ning the liabilities in

new compounds. The notion that there should be a ﬁ nite number of structurally identiﬁ -

able toxiclike patterns is analogous to the notion that there should be a restricted number

of amenable druglike patterns that reside within structural databases having high degrees

of molecular diversity. Indeed, the case for toxicity is certainly on ﬁ rmer ground at this

particular point in time since there will likely be little added to the area of fundamental

chemical reactivity in the new millennium as opposed to proteomic’s anticipated revela-

tion of numerous new biochemistries that will, in turn, provide numerous new pharmaco-

logic targets wherein many can be expected to have their own distinct pharmacophore (and

potentially new druglike patterns). Finally, since the precise locales where the toxicity hot

spots may ultimately occur are endless, the latter will perhaps be better addressed by direct-

ing a second set of database queries toward the ADME proﬁ le and intracellular localization

patterns that a given drug may exhibit. In the end, after array technologies are producing

useful toxicology-related knowledge, the interplay of all of the ADME parameters with

STR should become just as important as they are for efﬁ cacy in terms of what type of toxic-

ity may ultimately be observed within the clinic.

2.6.8 Weighting Decision Criteria from Efﬁ cacy and ADMET SAR

Figures 2.14 to 2.17 convey how all of the efﬁ cacy and ADMET HTS proﬁ ling data may

eventually be simultaneously deployed toward the design of an optimal preclinical can-

didate compound. Based on the magnitudes of the various molecular similarities and

dissimilarities across each HTS parameter survey of the compound data set relative to

the locations of the generalized pharmacophores associated with various parameters, a

ADMET CONSIDERATIONS 67

68 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

balance will be sought between options involving any combination of single or coadmin-

istered multiple entities wherein each member has been further tailored according to an

unchanging multivalent prodrug and/or soft drug strategy once interactions within the

biological realm have been initiated. Figure 2.14 attempts to convey some of these possi-

bilities using a hypothetical set of structural space. Alternatively, serving as a real example

that couples several of the aforementioned anticancer studies, Fig. 2.15 captures some of

the results already obtained from the pursuit of transportophore SARs relative to the pacli-

taxel template by placing the results within the formalism of the present discussion. Figure

2.16 provides an overall step-by-step strategy for deploying the consideration of ADMET

parameters proactively to enhance the process of drug discovery rather than just to play a

negative role as a series of ﬁ lters toward either the entry or the continued progression of

a given compound through such a process. Concerns about today’s trend toward ﬁ ltering

the entry of compounds have been alluded to in Section 2.5. In particular, it should be

noted from Fig. 2.16 that it is a completely deﬁ ned efﬁ cacy-related pharmacophore that

serves as the central structural theme to drive the proactive ADMET strategy in that all

Figure 2.14 3D pattern recognition example pertaining to the simultaneous consideration of ef-

ﬁ cacy and ADMET-related pharmacophoric parameters during lead selection and drug design. In

this example, SXR space has been mapped according to experimental HTS results from a mod-

erately sized, directed compound library wherein an efﬁ cacious template was identiﬁ ed. A and C

represent requisite pharmacophoric features oriented in space by structural elements B. Lacking

distinctive functionality, B, in turn, tends to prolong the elimination half-life and eventually causes

toxic concentrations to be produced in compartments not associated with efﬁ cacy. D represents the

structural space that provides for a desirable absorption and initial distribution proﬁ le. E represents

a structural space that is subject to rapid metabolism. Thus, in this case a range of suitable func-

tionality deﬁ ned by A can be utilized in the northwest region of the efﬁ cacious pharmacophore

while selected functionalities or bioisosteres that reside within the structural space deﬁ ned by the

overlap of C and D should be utilized in the southeastern region. The latter strategy optimizes

absorption and initial distribution features while retaining efﬁ cacy such that more complicated

prodrug scenarios are not necessary. A soft drug version, however, should be contemplated so that

the eventual toxicity problems derived from the connecting chain can be circumvented, especially

since the SMR portion of the overall map indicates that there is an intrusion of B by structural fea-

tures E that prompt rapid metabolism and elimination. The latter can thus be readily exploited by

incorporating them into the overall molecular construct and then making adjustments or ﬁ ne-tuning

them to a desired metabolic rate by the incremental insertion of steric impediments near the point

of metabolic contact.

Aq. Solubility

(SAbR)

MDR

(Avoidance

STR)

Cancer Cell

Selectivity

(Address SDR)

Attachment To Efficacy

Tolerant Area on

Paclitaxel (Neutral SAR)

Figure 2.15 Design of optimized drug candidate based on simultaneous consideration of several

ﬁ elds of paclitaxel-related SARs. Studies indicate that there is a distinct region of structural space

that is overlapped simultaneously by SARs pertaining to enhanced aqueous solubility, avoidance of

multidrug resistance (MDR), and the propensity to associate selectively with cancer cells compared

to healthy cells. Coupling of this distinct structural space (depicted as the shaded region) onto an area

of paclitaxel that can accommodate structural modiﬁ cation without losing efﬁ cacy provides an opti-

mized drug candidate. The speciﬁ c details of the distinct structural space and the synthetic methods

that can be used to couple its useful molecular displays onto paclitaxel are the subject of pending

patent applications.

490,491

Figure 2.16 Drug design and development strategy. Note that this strategy emphasizes a complete

deﬁ nition of the efﬁ cacy-related pharmacophore as the central theme such that it can be merged with

simultaneously generated ADMET-related pharmacophores via a knowledge-driven proactive process

that will ultimately produce the optimized clinical candidate or combination of agents to be deployed

for therapy or prophylaxis. It should be noted that this knowledge-based decision tree contrasts some

of the futuristic schemes that have been suggested by others wherein the various ADMET issues

are simply used as consecutive or simultaneous ﬁ lters to eliminate compounds being selected from

huge compound databases. As elaborated within the text, it is this author’s opinion that by medicinal

chemistry input, HTS data of the future will be able to take drug discovery investigations to signiﬁ -

cantly greater heights of knowledge such that the latter can then be used for proactively assembling

the positive type of enhanced property molecular constructs mentioned above. Indeed, if this scenario

does not unfold, the overall new and future processes will forever be locked into a negative mode that

simply keeps eliminating compounds failing to meet certain criteria placed at each parameter.

Compound

Library

Define Pharmacophore

Plus Negative And

Neutral Structural

Space

Directed

Library

Define Associated

ADMET Structural

Space, e.g.

Metabophores, etc.

Tailor Compound To Desired

Profiles Including Use Of

Prodrug And/Or Soft Drug

Technologies Either Singly Or

As A Combination Of Agents

Use Knowledge-

Based Approach To

Merge Or Avoid

Overlapping

Structural Space e.g.

Alter Metabolism

70 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

other structural modiﬁ cations/biological enhancements are then conducted according to

knowledge-based scenarios generated via previous ADMET proﬁ ling experiences with

analogous structural space or by immediate HTS ADMET proﬁ ling of appropriately di-

rected libraries. Clearly, while nondruglike structural space may thus be used effectively

to contribute toward deﬁ ning the efﬁ cacy pharmacophore’s electronic surface potentials,

such structural components within a library should still be tagged with red ﬂ ags, indicating

that they are also destined to be altered or removed completely by tailoring of the spatial

overlaps according to the type of knowledge base that can be afforded by medicinal chem-

istry. That the various ADMET-related pharmacophores will eventually be evolved so as

to be able to be deployed more independently from a given efﬁ cacy-related pharmacophore

and in a completely in silico manner at any point along the drug design and development

ﬂ owchart will indeed occur as well. However, the latter will probably be realized only

after we are well into the future and probably only after we have accumulated consider-

able knowledge about how to manipulate their structural patterns for optimal overlap and/

or avoidance with a variety of model efﬁ cacy platforms wherein all structural space has

been accounted for by 3D electrostatic potential maps derived from rigorous experimental

and computational considerations of molecular conformation. Finally, drawing from the

overall strategy listed in Fig. 2.16, Fig. 2.17 illustrates the speciﬁ c interplay of all of the

aforementioned efﬁ cacy, selectivity, and ADMET considerations, along with some other

practical drug discovery considerations.

2.7 PROCESS CHEMISTRY CONSIDERATIONS

As shown in Fig. 2.17, practical issues pertaining to intellectual property (IP), such as

the structural novelty of biologically interesting compositions of matter, as well as to the

latter’s synthetic accessibility (relative to process chemistry and manufacturing costs) will

also continue to be factored into earlier decision points intending to select the optimal

preclinical candidate compounds of the future.

369–371

The changing landscape of IP-related

structural novelty is addressed in the ﬁ nal, summary section of this review. Three key issues

pertaining to the interplay of medicinal chemistry with process chemistry’s responses to

current trends are mentioned below.

2.7.1 Cost and Green Chemistry

First, the eventual production cost for a new therapeutic agent is much more important

today than it has been in the past. This is because pharmaceutical companies must now

garner their proﬁ ts from a marketplace that has become sensitized about the cost of ethical

pharmaceutical agents. The days of simply raising the price of such products in parallel to

increasing costs associated with discovering and developing them have been over for quite

some time.

372

In this regard, the cost-effectiveness of small-molecule drugs will probably

maintain an edge over biotechnology-derived therapeutic agents for at least the near-term

portion of the new millennium. The second point to be mentioned pertains to the impact

of the green chemistry movement.

373–376

This movement has prompted pharmaceutical

companies to ensure that their productions of drugs are friendly toward the environment

in terms of all materials and methods that may be deployed in the process. Finally, the

U.S. FDA’s initiative to have all stereoisomers present within a drug deﬁ ned both chemi-

cally and biologically has prompted industry’s pursuit of drugs that either do not contain

asymmetric centers or are enantiomerically pure.

377

This, in turn, has prompted the need

for better stereochemically controlled processes during production. Stereocontrol has al-

ways represented an extremely interesting area for synthetic chemistry exploration and

now for biotechnology-derived chemistry and reagent research as well (e.g., exploitation

of enzymes at the chemical manufacturing scale). Considerable progress is being made

toward developing such methods on many fronts, including enzymatic

378–380

and microar-

ray technologies.

381

Often, however, the new laboratory techniques do not readily lend

themselves to inexpensive scale-up/manufacturing type of green chemistry. Alternative ap-

proaches that seek to address this situation in a very practical manner can be exempliﬁ ed by

the following study that intends to exploit simple α-substituted benzylamine systems. The

latter are being explored as chiral auxiliary synthetic reagents for the delivery of a nitrogen

atom in a stereochemically biased fashion during the synthesis of end-product amines that

contain neighboring asymmetry (Scheme 2.8).

2.7.2 Deﬁ ning Stereochemistry: Example Involving Benzylamine Chiral Auxiliary

Synthetic Reagents

Nitrogen systems having α, β, or γ asymmetry represent an extremely common struc-

tural motif within drug molecules such that the proposed methods immediately become

PROCESS CHEMISTRY CONSIDERATIONS 71

Figure 2.17 Lead selection and drug design decision ﬂ owchart based on efﬁ cacy- and ADMET-

related pharmacophoric parameters. This ﬂ owchart has been set up to represent the case where

a single-molecule construct having the lowest level of complexity/sophistication is initially

sought. However, in the future it is also likely that well-established templates that optimize a

certain parameter will be able to be paired with the efﬁ cacy-related agent at an early point in

the overall design process. For example, a compound that inhibits a transporter system respon-

sible for a given lead’s poor passage through the GI endothelium might ideally be coadminis-

tered as a soft drug version. In this way, the partner compound would solve the oral absorption

problem and then be metabolized and eliminated quickly without doing much of anything else.

Structural manipulation of the efﬁ cacy construct could then be directed toward enhancing other

DMET-related proﬁ les.

72 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

of interest to medicinal chemistry. Given the ease of selectively debenzylating tertiary

amines, the methodology should be particularly well suited for the production of asym-

metric secondary amines. While a benzyl moiety has previously been deployed as either

a common N or O protecting group,

382,383

the utilization of benzylamine to deliver a

nitrogen while simultaneously controlling the degree of substitution is less common, even

though such systems can sometimes be quite effective in this regard, due to the phenyl

group’s inherent steric properties.

384

The use of chiral α-substituted benzylamine systems

has been even more limited, with such systems most typically being deployed as resolving

agent counterions for carboxylic acid partners.

385

The rarer deployment of such systems

in covalent relationships probably results from the prevailing notion that as the steric

environment about the N atom is increased, it eventually becomes more difﬁ cult to ef-

fect debenzylation via catalytic hydrogenolysis.

386

Thus, in order to ﬁ rst scope the overall

Scheme 2.8 Potential stereoselective synthesis of secondary amines using asymmetric benzyl-

amine-related systems as chiral auxiliary reagents. 14 is a racemic precusor where R has an asym-

metric center α, β, or γ to X; 15 is an optically pure chiral (*) auxiliary reagent wherein R' is a

substituent having selected physical properties such as speciﬁ ed elements of steric bulk; 16 and 17

are diastereomeric tertiary amine intermediates which are mixtures or separated single diastereomers,

respectively; and 18 is the desired optically pure secondary amine product. Step 1 represents a va-

riety of N-alkylation or reductive alkylation methods (e.g., X  halide, carbonyl, etc.), step 2 repre-

sents a fractional recrystallization or chromatographic separation if necessary, and step 3 represents

a catalytic hydrogenolysis reaction. It is reasonable to anticipate that during step 1, racemic 14 could

combine with optically pure reagent 15 to provide one or the other of the two possible diastereomers

directly (path b). Even when no asymmetric bias is observed during step 1 (path a), the two diastereo-

mers present as intermediate 16 will differ in physical properties such as solubility, chromatographic

behavior, boiling point, or melting point. Thus, the desired diastereomer, 17, may be able to be sepa-

rated conveniently under selected conditions involving recrystallization or chromatography during

the workup (step 2) of the ﬁ rst reaction. Alternatively, it may also be possible to effect an asymmetric

cleavage during step 3 if the chiral auxiliary in one or the other of the diastereomeric tertiary amines

can be removed preferentially by hydrogenolysis. Also note that step 3 can be delayed so as ﬁ rst to

effect other chemical modiﬁ cations associated with an overall synthesis while the amino functional-

ity is still 3. (From ref. 389.)

applicability of this type of chemistry, recent studies have systematically examined the

relationship between a nitrogen’s immediate steric environment and the propensity to-

ward debenzylation within relevant model systems. Surprisingly, these results indicate

that steric attenuation of debenzylation is not likely to be problematic.

387,388

Explora-

tions are now proceeding toward assessing the diastereomer bias that may be achievable

upon reaction of racemic electrophiles with various members of a readily obtainable fam-

ily of chiral auxiliary benzylamine synthetic reagents.

389

Scheme 2.9 depicts the initial

approach toward an area that targets enantiomerically pure aryloxypropanolamines as

an appropriately challenging chemical model that is very relevant to the production of

pharmaceuticals.

390,391

In addition to its relation to practical asymmetric pharmaceutical

process chemistry applications of the future, the benzylamine example has been cited

herein because it further illustrates the important role that physical organic chemistry

considerations play within the medicinal chemistry thought process, and vice versa. By

analogy to medicinal chemistry terms, the preliminary benzylamine studies have deter-

mined that there is an allowable region of potential structure stereochemical relationship

(SSR) space to explore because of the neutral effects that were observed relative to a co-

event that is required for the overall chemical process (i.e., subsequent debenzylation).

The neutral space was mapped out by using precisely deﬁ ned structure–debenzylation

relationship (SDebR) steric probes. The present benzylamine studies, in turn, are now

taking advantage of the neutral space to identify useful synthetic stereophores that can

elicit selected asymmetries analogous to the situation of pharmacophores that can elicit

selected efﬁ cacies. Finally, upon turning all of these analogies around, exactly the same

“library” of steric probes can also be used to actually explore the steric tolerance associ-

ated with metabolic N-dealkylation (i.e., wherein a cytochrome P450 biological surface

and its requisite cofactors then substitute for the inorganic catalytic surface and its hydro-

genated atmosphere that were present during the aforementioned hydrogenolysis studies).

Importantly, such metabolism data will have the potential to be knowledge-generating in

that they can then be used to assist in the prediction of the susceptibility toward metabolic

N-dealkylation not just by speciﬁ c structural pattern recognition, but also by a parameter-

ized and well-deﬁ ned physicochemical property that is inherently localized in the space

immediately residing about the N atom within any metabolic candidate. The value of

this example’s endpoint merits reemphasis. In the future, efﬁ cacy and ADMET parameters

will be deﬁ ned maximally in terms of their associated pharmacophore’s electrostatic po-

tential space and not just in terms of distinct compound hits or leads. Although compound

hits and leads can certainly reﬂ ect desirable structural prototypes for a given parameter,

Scheme 2.9 Attempted diastereoselective opening of a model epoxide using benzylamine-related

chiral auxiliary synthetic reagents. (From ref. 391.)

PROCESS CHEMISTRY CONSIDERATIONS 73

74 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

they can also become restrictive toward a broader conceptualization of the more diverse

structural space that can be invoked to better purview the full spectrum of efﬁ cacy and

ADMET overlaps, as the latter parameters are all being addressed simultaneously. Thus,

from a broader, knowledge-derived vantage point, the ideal candidate drugs (candidate

drug teams) of the future will ultimately be designed by considering all of the efﬁ cacy and

ADMET parameter pharmacophores simultaneously while not being restricted to any type

of predetermined structural template associated with a hit or lead structure that may have

been obtained from any one of them.

2.8 ANALYTICAL CHEMISTRY/X-RAY DIFFRACTION

As mentioned in the introduction, new developments in any of several key analytical tech-

niques can be expected to have a profound impact upon medicinal chemistry’s future. X-

ray diffraction is taken herein as just one example for these types of possibilities.

2.8.1 Latest Trends

At the forefront of current progress and trends in the ﬁ eld of x-ray technology are efforts

to properly derive and readily portray electrostatic potentials.

392

As alluded to earlier, the

initial recognition and driving forces associated with the interactions between a xenobi-

otic and the biological surfaces that it will encounter in vivo are due to a complementary

match between the topography of the electrostatic potential of the xenobiotic ligand and

that of the biological site (less anything energetically favorable that is given up when

the xenobiotic leaves its solvated environment). As discussed in Section 2.5, consider-

able effort has been expended toward calculating accurate depictions of the electrostatic

potentials of molecules theoretically. This has been most beneﬁ cial, however, for only

very small molecules because extended basis sets are required to obtain accurate results.

Calculations with smaller basis sets have been carried out for larger molecules, but as

already pointed out, the results are typically unreliable, owing to the nature of the math-

ematical approximations that will have necessarily been taken. Using larger basis sets for

small fragments and then using the fragments as building blocks for larger molecules is

also being done. The latter approach will probably become much more prevalent in the

new millennium.

Alternatively, it has been possible for several years to obtain experimental electrostatic

potential maps from the molecular charge distribution derived from x-ray diffraction data.

This approach has had limited appeal, however, because the experiments needed to pro-

vide the large amount of quality data that is used as the starting point have themselves

been extremely time-consuming, typically taking many weeks even for quite small mol-

ecules. Nevertheless, from the studies undertaken to date,

392

there is growing evidence

that results for small molecules can indeed be extrapolated to similar fragments in larger

molecules,

393

just like what is being suggested by the theoretical approaches that were

reiterated above. The practical implications of potentially using this approach to simul-

taneously address huge numbers of compounds when they reside along distinct structural

themes within compound libraries or databases (e.g., wherein a given scaffold then be-

comes the common fragment) is worth noting for both the x-ray and computational types

of approaches.

Today, the experimental approach afforded by x-ray diffraction has become much more

tractable. With the new generation of x-ray diffractometers using charge-coupled device

(CCD) area detectors, the necessary experimental data can be obtained in just a few days,

a duration comparable to that currently required for a routine x-ray structure determina-

tion.

394,395

Furthermore, with access to x-ray synchrotron beam lines, the time of the ex-

periment may be reduced to a few hours.

396

Shorter experiments, in turn, have allowed

development of cooling devices using liquid helium, thus giving access to lower tempera-

tures and improved data.

397

In addition, whereas with a serial diffractometer the length of

the experiment scales with the number of atoms in the molecule under study, the size of

the molecule is less important when collecting the data using a CCD detector. How large

a molecule is tractable in terms of converting these data into electrostatic potentials is still

unknown but is likely to be forthcoming within just the near-term future. In this regard,

one might speculate that this approach may even be able to handle larger molecules more

reliably than will the theoretical approaches elaborated in Section 2.5. From the present

situation it is already clear that it is now possible to map the topology of the electrostatic

potential for a typical small-molecule therapeutic agent within very reasonable time pe-

riods. This means that the electrostatic potentials may be able to be compared for series of

molecules having established biological activities so as to produce reﬁ ned SAR and to pro-

vide the most meaningful data possible relative to the x-ray contributed structures within

future databases. Not too much further into the new millennium, one might imagine that it

will also be feasible to use this approach toward mapping the complementary electrostatic

potential of receptor/enzyme active sites at highly improved resolutions,

397–401

especially

as promising results have already been obtained for some small proteins.

393

Indeed, certain

of the techniques described above for small molecules are already being applied to the

analysis of macromolecules.

401

2.8.2 Examples Involving Dopamine Receptors, c-AMP Phosphodiesterase

Enzymes, and the Dynamics of Protein Folding

Since applications of these latest trends in x-ray technology are themselves under investiga-

tion, studies like the ones being undertaken by Pinkerton et al.,

402

which intend to validate

the utility of deploying the CCD type of cutting-edge diffractometer approaches toward

the study of drug design, represent a critical step at this juncture. For these studies, clas-

sical SAR that is available within some established systems of therapeutic interest, such

as that for renal dopamine receptors

403

and for c-AMP phosphodiesterase active sites,

404

are being reexamined. Long-standing topographical models for these systems (Schemes

2.10 and 2.11, respectively) will certainly be interesting to reevaluate based on reﬁ ned

analyses of the most relevant structural probes that lend themselves to crystallization and

extremely accurate x-ray analysis. In this regard, the example depicted by Scheme 2.11 is

especially noteworthy because it clearly demonstrates the “broadest conceptualization of

a pharmacophore” theme that was emphasized as being extremely important at the close

of Section 2.8.1 (i.e., note that it is the electrostatic surface potential of a lead compound’s

imidazolone system that is further likened to the electronic topography of the cyclic phos-

phate’s trigonal bipyramid transition-state species traversed when cAMP is hydrolyzed by

phosphodiesterase).

Assuming that reﬁ ned x-ray techniques will become commonplace within the ﬁ rst 25

years of the new millennium and that they will be coupled with HTS approaches toward

crystallization and actual obtainment of diffraction data,

122,405,406

it becomes interesting

ANALYTICAL CHEMISTRY/X-RAY DIFFRACTION 75

76 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

to ponder what might be on the more distant horizon of x-ray-related technologies. Cer-

tainly, the ability to derive x-ray diffraction patterns from noncrystalline small and large

molecules would allow medicinal chemistry to embark on a rational approach toward drug

design immediately upon the obtainment of such “pictures” for every new pharmacologi-

cal target of the future. In silico screening of real and virtual libraries, having matured

considerably at that point as well, would also beneﬁ t enormously from such a development

and would be expected to be equally interactive with such data in terms of docking virtual

compounds into x-ray pictures of enzyme active sites and/or receptors followed by rank-

ing them as potential drug candidates.

407

Further out, but perhaps not too far past the next

75 years of imaginable future, it might be expected that x-ray-type data (even if no longer

strictly derived from x-ray’s present physics-related principles

408,409

) will be able to be

collected at fast enough real-time intervals such that it may become possible to observe

molecular interactions and motions within userfriendly videos after the data are processed

appropriately. One can imagine that videos of events such as the actual docking (binding or

afﬁ nity

410

) of a drug, the motions that transpire upon an agonist’s triggering of a receptor

(intrinsic activity

410

), or a substrate’s alterations upon action by an enzyme, could all be-

come commonplace in the somewhat more distant future. These displays would be similar

to the cartoon (i.e., in Section 2.1, “hand-waved”) versions that we presently generate using

computational or theoretical approaches coupled to molecular modeling packages

411–414

except that the entire process would be based on actual experimental data obtained in a

Scheme 2.10 Topographical model of the renal vascular dopamine receptor. A, C, Hp, and Hm

reside in a single plane and represent regions that interact with dopamine’s (19) amine, catechol ring,

p-hydroxy and m-hydroxy, respectively. Region B represents an auxiliary binding site suggested by

apomorphine’s SAR, whereas S1 and S2 represent steric limitations toward the binding of receptor

ligands. Note that while dopamine, 19, is nicely accommodated by the model, neither enantiomer of

the cyclopropyl analog, 20, can be accommodated because they will collide with either the receptor’s

planar ﬂ oor or S2 ceiling. The cyclopropyl analogs were found to be devoid of activity at either pe-

ripheral (renal) or central dopamine receptors.

492

(From ref. 403.)

real-time fashion, as the event actually occurred rather than by what we have only been

able to so far catch experimental glimpses of at stabilized junctures or what we are able to

imagine what such conversions may look like. For example, observing an actual trigonal

bipyramid transition state for the phosphorus atom within c-AMP during its hydrolysis by

a phosphodiesterase (Scheme 2.11) would certainly be a crowning analytical experimental

achievement of the future, especially since today’s sophisticated theoretical approaches

cannot even provide a good cartoon version for this process due to collapse of such a spe-

cies to lower energies during minimization. Similarly, observing the methylene portion of

cyclopropyl dopamine actually crashing into the proposed molecular ceiling on the renal

vasculature’s dopamine receptor (Scheme 2.10) would be equally impressive since this bit

of older SAR appears to have gradually become lost from the dopamine ﬁ eld, due to lack of

timely advances able to clarify its reality. Finally, observing that the conformationally re-

dundant nature of endothelin’s 3,11-disulﬁ de bond (Scheme 2.12) actually serves to initiate

its post-translational folding paradigm because that bond is easier (in terms of distance) to

form than the subsequent 1,15-disulﬁ de link

415

represents just one of many older hypoth-

eses that might be taken from the area of proteomics.

PDE

Arg-PDE

HNH

′

5′

HN NH

3′

5′

3′

PDE-Arg

PDE

21 22 23

Panel A

Panel B

Panel C

Scheme 2.11 Conformational and electrostatic potential topographies of c-AMP phosphodiester-

ase III (PDE III) active site ligands. In panel A, compound 21 represents a c-AMP substrate with

its adenine (Ad) and ribose moieties in an “anti” relationship. Interaction 22 depicts binding of the

phosphate portion using an arginine residue and a water molecule that was initially associated with

2ⴙ

in a stoichiometric relationship. Complex 23 depicts S

2 attack of phosphorus by H

O, with

formation of a trigonal bipyramid (TBP) transition state (TS). Compound 24 represents 5'-AMP as its

inverted product. The electronic charges indicated conserve the net charge overall and across the TS.

Panel B represents an overhead view of the atoms in the single plane of the TS, which forms the com-

mon base for the two pyramids of the TBP system. Also shown is the proposed electrostatic potential

map for the same atoms. Panel C shows a classical PDE III inhibitor ligand and the AM1 derived

in-plane molecular electrostatic potential map of its imidazolone ring. Because of the very notable

similarity between these two electrostatic potential maps, it has been proposed that these types of

compounds, as well as several other heterocycles that have electron-rich heteroatoms in analogous

locations, act as TS inhibitors of PDE-III.

404,493,494

(From ref. 404.)

ANALYTICAL CHEMISTRY/X-RAY DIFFRACTION 77

78 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

2.9 SUMMARY

2.9.1 General Points

Medicinal chemistry has been deﬁ ned as a pure science that explores fundamental physi-

cal organic principles to understand the interactions of small molecular displays with the

biological realm. Regardless of discipline or background, when investigators seek this level

of understanding, they are embarking upon basic medicinal chemistry research. Using a

variety of input data, medicinal chemistry’s applications, in turn, become that of designing

or selecting new drug candidates as well as providing molecular blueprints for improving

the therapeutic proﬁ les of existing drugs or of new pharmaceutical agent lead structures.

Historically, medicinal chemistry has also been heavily involved with generating input data

by designing and synthesizing probe molecules that can systematically test the roles being

played by the various physical organic principles during a given interaction. More recently,

medicinal chemistry has begun to utilize site-directed mutagenesis as an additional tool to

understand these same types of interactions. Future developments in store for drug discov-

ery are summarized in Table 2.10. The overall progression conveyed by Table 2.10 can be

≡

Scheme 2.12 Proposed folding scheme for the endothelins (ETs). Positions 1, 3, 11, and 15 are

cysteines. Note that formation of the 3,11-disulﬁ de linkage is proposed to occur ﬁ rst given their

closer proximity. Subsequent formation of the 1,15-disulﬁ de can then occur more readily to ﬁ nally

adopt the tightly folded form depicted as 25. At this juncture, the 3,11-disulﬁ de may then represent a

conformational redundancy that is not actually required in the ﬁ nal structure. This sequence of events

is equally applicable to big endothelin and to pre-proendothelin. Structure 25 represents a molecu-

lar dynamics-derived minimum energy conformation for ET-1 after starting from an entry structure

where residues 9 to 16 were initially constrained in an α-helix.

495

Note that the α-helix is retained

and that

His,

Asp, and

Trp form a close triad wherein

His and

Asp are probably internally

hydrogen bonded.

415

(From ref. 415.)

•

Genomics, and especially proteomics, will continue to unveil myriad new biomechanisms appli-

cable to modiﬁ cation for therapeutic gain and applicable to better understanding processes associ-

ated with ADMET.

•

Biomechanistic systems lending themselves to crystallization followed by x-ray diffraction will be

readily exploited by structure-based drug design.

•

BIOTECH will continue to evolve HTS/UHTS methodologies to assay therapeutic mechanisms

and to assess ADMET properties as a front-line initiative.

•

For cases where structure-based drug design is not possible, combinatorial libraries along with

both wild and genetically altered or elicited collections of natural materials will provide initial hits,

which will then be followed up by directed libraries and by ligand-based drug design.

•

Accumulating data in all areas will become well managed via a variety of databases, including,

in particular, an evolved level of 3D sophistication within chemoinformatics. The latter will ulti-

mately provide the common language for the integration of information across all databases.

•

As ADMET information accumulates from preclinical models, which will indeed become vali-

dated relative to the human case, distinct structural motifs will arise that can be used with statisti-

cally derived conﬁ dence limits in a predictive manner during drug design.

•

UHTS of man-made libraries, directed analyses of nature’s mixtures, and importantly, rational

exploitation of the newly elaborated and fully understood ADMET structural motifs will reveal

synergistic combinations that involve more than one compound.

•

For each case of new drug design, a summation of the accumulated knowledge from efﬁ cacy test-

ing, ADMET, and potential synergistic relationships will guide ﬁ nal tailoring of the clinical candi-

date drug or multidrug combination. Whereas a reductionist approach toward achieving the most

simple single-molecular system possible will remain appealing, any combination of one or more

agents wherein each member can independently use prodrug, soft drug, and multivalent strategies

will be considered within the context of deploying whatever is envisioned ultimately to work the

best for the clinical indication at hand.

•

As clinical successes unfold, a gradual move from experimental to virtual screening will occur.

Although this is already being done for initial lead ﬁ nding by docking druglike virtual compound

libraries into pockets deﬁ ned by x-ray, virtual methods should eventually be able to replace the

entire preclinical testing paradigm.

•

Paliative and curative targets will remain, the latter particularly for combating the ever-evolving

populations of microorganisms and viruses, and the former for retaining the quality of life as the

human lifespan continues to elongate. However, preventive and prophylactic treatment paradigms

will gradually take on more signiﬁ cance and will eventually reside at the highest priority of life

sciences research.

•

For all treatments, and especially for preventive and prophylactic paradigms, pharmacogenetics

will deﬁ ne population subgroups that will then receive treatment protocols optimized for their

individuality relative to the particular treatment that is being rendered. Testing to ascertain an

individual’s pharmacogenetic proﬁ le will begin at birth and continue at periodic intervals through-

out one’s life relative to the optimal deployment of preventive protocols. Such testing will also be

conducted immediately prior to any treatment of a detected pathophysiology.

•

In the more distant future, the combination of nanotechnologies with bioengineering and biotech-

nology will allow instillation of devices not only for immediate diagnoses of any deviations from

homeostasis, but also of programmable portals for user-friendly drug administration other than via

the oral route. Eventually, this same mix of technologies will also allow for instillation of program-

mable exits for the controlled excretion of drugs with or without the need for metabolism. Such

developments will, in turn, cause drug discovery to return primarily to the pursuit of only efﬁ cacy

and toxicity issues.

TABLE 2.10 Future Events Predicted for Drug Discovery and Development

(Continued)

80 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

Discovery libraries

•

Size is less important than diversity (with allowance for structurally redundant series, e.g., Me,

Et, etc.).

•

Diversity is also much more important than druglike properties (e.g., presence of non-druglike

members can be extremely useful toward initially probing overall structural space).

•

Alternatively, assay likable properties are mandatory [i.e., compound members must be able to be

delivered (e.g., solubilized, etc.) according to demands of a given assay].

Compound hits

•

Druglike properties are less important than efﬁ cacy tolerance [i.e., ﬂ exibility for altering structure

without altering efﬁ cacy (access to neutral regions)].

Initial lead structures

•

Individual structures are less important than a detailed description of the pharmacophore in terms

of electrostatic surface potentials plus knowledge about structural space that is neutral or intoler-

able toward modiﬁ cations relative to the measured biological parameter (e.g., efﬁ cacy).

•

Nondruglike features that may be present within certain members contributing to the overall phar-

macophore should be “red-ﬂ agged” but not necessarily ruled out as potential building blocks,

while the overall process of merging structural space across all parameters is continued (e.g., ef-

ﬁ cacy plus ADMET).

Final lead compounds

•

Optimal blend of efﬁ cacy and druglike properties (nondruglike features now completely removed

or adequately modiﬁ ed according to experimentally ascertained criteria that have been validated

for their correlation to the clinical response).

•

At least one neutral region or prodrug/soft drug option remains such that unanticipated hurdles

presenting themselves during further development might still be addressed by additional chemical

modiﬁ cation.

•

One or more backup compounds having distinctly different molecular scaffolds while still fulﬁ ll-

ing the overall ensemble of pharmacophore and druglike patterns.

TABLE 2.11 Molecular Attributes of Discovery Libraries, Compound Hits, Lead Structures,

and Final Compounds

Notice that while emphasis is placed on deﬁ ning a given pharmacophore to the maximum possible detail by

deemphasizing the use of druglike property parameters as an early ﬁ ltering mechanism, components of the

pharmacophore that are presumed to be undesirable should still be red-ﬂ agged as such. ADMET SAR can then

be superimposed within the distinct molecular contexts of each identiﬁ ed pharmacophore so as to be deployed

more efﬁ ciently as a ﬁ lter and, importantly, in a proactive manner while initial lead structures move toward ﬁ nal

lead compounds. This two-step approach will also allow for continued knowledge building within all of the key

parameters relative to various therapeutic areas.

•

Somewhere during gene therapy’s complete eradication of defective gene-based disease and bol-

stering of gene-linked defenses, along with the nanotechnology–bioengineering effort to recon-

struct humans relative to improving health, public policies and opinions dealing with what other

attributes might be manipulated to enhance the quality of life will need to be clariﬁ ed by signiﬁ cant

input from the nonbasic science disciplines. Thus, the social sciences, humanities, and philosophy

ﬁ elds, along with religion and the lay public at large, should look forward to providing what will

soon become desperately needed input into the continued directions that life sciences research is

likely to take well before the end of the next century, let alone before the conclusion of the present

millennium.

TABLE 2.10 (Continued)

thought of as being driven by an ever-increasing accumulation of knowledge within the

life science arena. Table 2.11 summarizes some key points relevant to chemical compound

categories associated with drug discovery.

2.9.2 Attributes of Drug Discovery Libraries, Compound Hits,

and Lead Compounds

Table 2.11 is important because it addresses a noteworthy concern, namely that today’s

trend to aggressively ﬁ lter non-druglike compounds out of the initial drug discovery HTS

process may work against the accumulation of knowledge that will be vital toward con-

tinually improving the overall process. In other words, although it can be argued that a

certain efﬁ ciency in the production of NCEs might already be obtainable at this juncture

by engaging in this type of negative strategy, an overemphasis upon this approach runs the

risk of having the drug discovery and development process becoming forever locked into

just this point of evolution. Alternatively, Table 2.11 conveys how some of today’s trends

that pertain to the molecular attributes of discovery libraries, compound hits, and lead com-

pounds, might best be deployed so as to allow the future developments conveyed in Table

2.10 to be derived from a truly solid base of accumulating knowledge.

At the front of the push to move forward remain the trends inspired by biotechnol-

ogy. HTS has already led to the situation where there are now mountains of in vitro data

available for input toward drug-related considerations. Within just the near term of the

new millennium, the entire gamut of ADMET parameters can be expected to join efﬁ cacy

surveys being conducted by HTS. Importantly, during this period the latter’s output will

have also become validated in terms of predicting clinical correlates. The common link

between these databases will be molecular structure as afforded by the probe compounds or

compound library members that become deployed during a given assay. Molecular struc-

ture can best be appreciated by the precise language that medicinal chemistry has been

learning since its formalization as a distinct discipline about 75 years ago. Thus, medicinal

chemistry is also obliged to step to the forefront and assist in understanding and translating

what the mountains of new data mean so that they might be optimally applied toward the

development of new therapeutic agents.

Recognizing that we have not been very effective to date, the appropriate handling of 3D

chemical structure within large databases represents a signiﬁ cant challenge that needs to be

resolved by a cooperative effort between medicinal chemists, computational chemists, and

both bioinformatic and chemoinformatic database experts as quickly as possible. That me-

dicinal chemists have a good appreciation for the biological nature of the data within one

mountain versus that of another is an equally challenging interdisciplinary problem that

will need to be addressed by cooperative efforts between medicinal chemists and investiga-

tors from all of the biochemical- and biological-related sciences. Resolving both of these

challenges will eventually allow the in vitro data sets to be intermeshed so as to provide

knowledge-generating assemblies that accurately predict the results that are eventually ob-

tained in vivo and, ultimately, within the clinic.

2.9.3 Formalized Instruction of Medicinal Chemistry

Faced with these immediate critical roles for medicinal chemistry within drug discovery

research, how should academia be preparing doctoral-level investigators to contribute as

medicinal chemists of the new millennium? First and foremost should be to retain medicinal

SUMMARY 81

82 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

chemistry’s emphasis on the physical organic principles that deﬁ ne chemical behavior in any

setting. This is fundamental to being a medicinal chemist. Such principles cannot be learned

well by relying only on textbook/e-instruction or even by predesigned laboratory outcome

exposures. Thus, a laboratory-based thesis project that involves physical organic principles

as the underlying variables of its scientiﬁ c enquiries seems mandated. While several types

of chemical problems might be envisioned to provide such a learning experience, the lab-

oratory practices of synthetic and physical organic chemistry represent extremely useful

tools to drive home permanently the principles associated with intra- and intermolecular

behavior and chemical reactivity. Similarly, with regard to synthesis/compound production,

it is also extremely important to learn ﬁ rst how to isolate and characterize pure materials.

Combinatorial mixtures and biochemical manipulations that rely on chemical kinetics and

the process of natural selection to dictate their concentrations can then be better appreci-

ated if approached at a later point in time. Finally, while a multistep synthesis of a complex

natural product can instill fundamental chemical principles, it may be more effective for

a budding medicinal chemist to prepare one or more series of probes wherein most of the

members in the series are novel in structure but are (seemingly) still readily obtainable via

reasonably close literature precedent according to short synthetic sequences (e.g., ﬁ ve or six

steps to each template that is to be further derivatized). This positions the student closer to

eventually appreciating structural trends and patterns that may reside within databases.

While this solid foundation is being derived from experimental lab work, graduate-level

exposures to various other ﬁ elds and aspects of life science research will, instead, have to

rely on available courses, seminars, or independent reading. Merging a student’s chemi-

cal learning base with a speciﬁ c biological area being targeted by the student’s molecu-

lar probes, however, ought to be feasible via actual experimentation without jeopardizing

either subject’s rigor. In the end, however, continuing postgraduate education is probably

the only way that an investigator intending to practice medicinal chemistry will even be

able to purview the explosion of information occurring in all of the areas relevant to assum-

ing the roles needed to resolve the aforementioned challenges of the new millennium. That

a practitioner may be able to have a head start along this learning path by initially pursuing

a formalized medicinal chemistry curriculum rather than an organic chemistry curriculum

has recently been suggested by others.

416

It should be emphasized that the broader exposure to the life sciences is a critical com-

ponent for a medicinal chemist’s continuing education not because medicinal chemists

should eventually attempt to pursue such endeavors independently but because these expo-

sures will allow them to interact more intelligently and meaningfully and collaborate with

dedicated experts in each of the numerous other ﬁ elds. Thus, the ability of medicinal chem-

ists to participate in interdisciplinary research while serving as scientiﬁ c “scholars” during

their attempts to integrate knowledge across broad sets of data and scientiﬁ c ﬁ elds

417

are

key operational behaviors that also need to be instilled early in the overall graduate-level

educational process. By their very nature, medicinal chemistry experiments often prompt

fundamental questions or hurdles that may be related to a variety of other disciplines. For

example, while in pursuit of dopamine receptor ligands having a cyclopropyl template

(Scheme 2.10), it became necessary to devise a new chemical method for effecting the

Curtius conversion under neutral conditions while preserving benzyl-protecting groups

that were located on the catechol moiety.

418

A similar chemical hurdle pertaining to the

formation of β-acetoxyoxetane systems relative to the paclitaxel studies was likewise en-

countered

239,240

as part of an example cited earlier. From the biological side, one of the ﬁ rst

questions that a medicinal chemist cannot help but ponder immediately upon entry into the

paclitaxel arena

175

is whether or not an endogenous material also exists, perhaps similar

to but distinct from the microtubule-associated proteins, which normally interacts with the

paclitaxel receptors purported to reside on intact microtubules in a stoichiometric man-

ner.

419

Similar biological questions pertain to the results from probing the SAR associated

with the Pgp MDR system. The latter appear to place the investigators in the middle of a

“one versus two distinct Pgp binding sites” controversy.

420

In this case, further experimen-

tal clariﬁ cation of this situation may eventually allow these medicinal chemists to pursue

site-directed selective Pgp ligands that are less prone to affect normal cells than cancer

cells, or to pursue bivalent superligands, well before the details of this controversy have

become fully deﬁ ned and resolved by genomic and proteomic approaches. Thus, the inter-

play of the subject matter from various biological disciplines during medicinal chemistry

research is as inherent in the broader medicinal chemistry intellectual process and notion

of scholarship as is the practical requisite for a solid-based knowledge of fundamental

physical organic principles.

In the future, increasing numbers of formalized short programs pertaining to a given

biological area are likely to be offered to practicing medicinal chemists at technical meet-

ings, academic centers, home cites, and via e-instruction.

421

Given the interdisciplinary

nature of the problems already at hand, along with the proposition that they will become

signiﬁ cantly more complex as we progress further into the new millennium, it is likely that

companies that encourage such interdisciplinary types of continuing education will also

eventually become the leaders that are able to most effectively implement the new para-

digm of drug discovery (i.e., not just toward generating more data faster while working on

smaller scale, but toward producing knowledge systems that actually lead to better NCEs

at a quicker pace while spending less money).

2.9.4 Intellectual Property Considerations

Before closing this chapter it also becomes appropriate to consider how all of the afore-

mentioned technical and operational possibilities could affect where medicinal chemistry

may be headed in terms of pharmaceutical IP.

422,423

Comments in this area will be directed

only toward small-molecule compounds and not toward biomolecules, despite the noted

turmoil that was initially created in the gene-related arena.

424,425

As indicated earlier, the

highly interdisciplinary nature of today’s life science research endeavors, coupled with the

new paradigm in drug discovery, indicates that the small-molecule composition of matter

arena is no longer the exclusive domain of medicinal chemistry. Nevertheless, even though

the appropriate list of inventors for any given case that has utilized HTS and combinato-

rial chemistry could become quite complex, with patience these situations should all be

reconcilable. Alternatively, there are some other issues that are also beginning to hit the IP

arena for which answers and appropriate operational models may not be as clear. Given

that the desirable goal of enhancing world trade has prompted the need to recognize (if not

to completely harmonize) patents on a global basis,

426,427

it is likely that the unique posi-

tion held by the United States with regard to acknowledging notebook entries as the earliest

dates of an invention’s conception will ultimately give way to the more practical European

process that simply acknowledges the ﬁ rst to ﬁ le. However, this move will further encour-

age the ﬁ ling of patents on technologies that are still very immature. For example, casting

this possibility within the trends elaborated throughout this review, companies will need

to resist the urge to ﬁ le on complete compound libraries and instead focus on claims that

protect a reasonable family of leads for which several members have indeed been identiﬁ ed

SUMMARY 83

84 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

as being meritorious by both efﬁ cacy and selectivity and at least preliminary ADMET HTS

(i.e., experimentally ascertained privileged structures). Unfortunately, an even worse sce-

nario has already begun, in that applications appear to be pending and arguments are being

directed toward the validity of patenting huge virtual libraries considered to be druglike in

their makeup. Emphasizing the notion that an actual reduction to practice is paramount for

a patent, this author presently stands in opposition to the attempts to garner protection of

virtual libraries. Along this same line, this author feels compelled to further note that the

current motion to assign CAS numbers to virtual compounds also represents a step in the

wrong direction. Finally, while patent protection of an existing scaffold that has experimen-

tally demonstrated its utility in one or more therapeutic areas is certainly meritorious, in

the future, companies will still need to refrain from overelaborating these same scaffolds in

an attempt to generate NCEs across several other therapeutic areas based solely on having

already secured IP protection within the context of compositions of matter. In other words,

force-ﬁ tting a given scaffold via its array of appendage options rather than by conducting

an HTS survey of other structural systems across the complete proﬁ le of selective efﬁ cacy

and ADMET parameters could easily be taken as a step backward in terms of both the

molecular diversity and therapeutic quality that is ultimately being delivered to the market-

place down the road of the new millennium. A broader discourse on business and scientiﬁ c

ethics at this juncture is beyond the scope of this review even though the rapid biotechnol-

ogy advances are certainly pressing the need for in-depth discussions in these areas and

their fundamental ties with philosophy and religion.

428

2.9.5 Knowledge Versus Diversity Paradox

In this same regard, however, a seeming paradox will be created by the insertion of knowl-

edge systems into the new drug discovery paradigm. Since medicinal chemistry will seek

to deﬁ ne SXR in terms of 3D electrostatic potentials that become predictive of preferred

ADMET and efﬁ cacy and selectivity behaviors so that their various assemblages can lead

to privileged drug structural motifs (or to ensembles of privileged structural motifs that

are deployed as drug teams), once this process begins to become effective, it will also play

against molecular diversity. Although this situation is not nearly as limiting as the situation

conveyed in the preceding paragraph and will always be subject to an expansion of diversity

based on the uniqueness of the efﬁ cacy pharmacophoric components, enhanced ADMET

knowledge in particular will indeed work in a direction away from overall diversity. Hope-

fully, however, the saving factor in this evolution will remain the pursuit of therapeutically

preferred arrangements and not the overutilization of a particular motif just because it has

been able to garner an exceptionally favorable ADMET proﬁ le, perhaps accompanied by

a strong patent position as well. Finally, while enhanced knowledge inherently leads to

more credible and useful predictions, it is the overextrapolation, extra weight, or zeal that

is sometimes placed on a given prediction versus other options, including that of having

no prediction, that can become problematic. Thus, even when all of the challenges cited

in this review appear to be resolved, the various disciplines caught up in drug discovery,

including that of medicinal chemistry, should all remain cognizant of the earlier days of

“preconceived notions” while also recalling the old adage that “a little bit of knowledge can

sometimes be dangerous” such that when the ideal drugs and drug ensembles of the near-

term future are constructed from experimental data, and those of the more distant future

from virtual data, the subsequent lab-based preclinical and clinical investigations will still

remain open to the possibility that at any point along the way, anything might still be able

to happen. Casting this last sentiment in a favorable direction, medicinal chemists of the

future, no matter how knowledgeably and guided by wisdom the overall process of drug

discovery may seem to have become, should always remain on the alert for serendipity.

Toward this end, the following three quotes, already revived by others relative to recent de-

velopments important to medicinal chemistry, have been strung together as an apt closing

for this chapter. Each is just as relevant today as it was when it was ﬁ rst pronounced.

We have scarcely as yet read more than the title page and preface of the great volume of na-

ture, and what we do know is nothing in comparison with that which may be yet unfolded and

applied.

—Joseph Henry, more than 100 years ago, as quoted recently by Jacobs

And if we are indeed to go forth and “…see further, then it will be by standing on the shoulders

of the giants…” as well as on the shoulders of the many others like us who have gone before

and who have thus brought us to where we are now.

—Isaac Newton, as quoted recently by Wedin

429

and as modiﬁ ed in tense so as to be used

within the present context

For “…we shall not cease from exploration. And the end of all of our exploring will be to

arrive where we started, and know the place for the ﬁ rst time.”

—T. S. Eliot, as recently quoted by the International Human Genome Sequencing

Consortium while concluding their report on the analysis of a substantially

complete version of the human genome sequence,

430

a historic accomplishment also

reported recently in a similar manner by a Celera Genomics–led consortium

431

ACKNOWLEDGMENTS

The secretarial assistance of Mrs. Pam Hennen is gratefully acknowledged.

REFERENCES AND NOTES

1. R. Breslow. Chem. Eng. News, Nov. 20, 5 (2000).

2. M. Jacobs. Chem. Eng. News, Dec. 6, 5 (1999).

3. M. Jacobs, A. Newman, J. Ryan (Eds.). The Pharmaceutical Century, American Chemical So-

ciety, Washington, DC (2000).

4. J. Steiner. Drug Discov. World 1(2), 61–66 (2000); 2(1), 55–61 (2000/1).

5. H. O. Alpar, K. R. Ward, E. D. Williamson. STP Pharm. Sci. 10(4), 269–278 (2000).

6. R. A. Hudson and C. D. Black. Biotechnology—The New Dimension in Pharmacy Practice: A

Working Pharmacist’s Guide, 2nd ed., ACPE Publications, Columbus, OH (1999).

7. C. M. Henry. Chem. Eng. News, July 3, 41–47 (2000).

8. U. Holzgrabe, I. Wawer, B. Diehl (Eds.). NMR Spectroscopy in Drug Development and Analy-

sis, Wiley-VCH, New York (1999).

9. W. Maas. Drug Discov., Aug./Sept., 87 (2000).

10. M. Mayer and B. Meyer. J. Med. Chem. 43, 2093–2099 (2000).

11. P. J. Hajduk, S. Boyd, D. Nettesheim, V. Nienaber, J. Severin, R. Smith, D. Davidson,

T. Rockway, S. W. Fesik. J. Med. Chem. 43, 3862–3866 (2000).

12. K. Tang, D.-J. Fu, D. Julien, A. Braun, C. R. Cantor, H. Koster. Proc. Natl. Acad. Sci. USA 96,

10016–10020 (1999).

REFERENCES AND NOTES 85

86 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

13. R. H. Griffey, S. A. Hofstadler, K. A. Sannes-Lowery, D. J. Ecker, S. T. Crooke. Proc. Natl.

Acad. Sci. USA 96, 10129–10133 (1999).

14. G. Karet. Drug Discov. Develop., Oct., 38–40 (2000).

15. M. L. C. Montanari, A. E. Beezer, C. A. Montanari, D. Pilo-Veloso. J. Med. Chem. 43,

3448–3452 (2000).

16. M. L. C. Montanari, A. E. Beezer, C. A. Montanari. Thermochim. Acta 328, 91–97 (1999).

17. L. Stolz. Annu. Rep. Med. Chem. 33, 293–299 (1998).

18. E. Danelian, A. Karlen, R. Karlsson, S. Winiwarter, A. Hansson, S. Lofas, H. Lennernas, M.

Hamalainen. J. Med. Chem. 43, 2083–2086 (2000).

19. A. Burger (Ed.). Medicinal Chemistry, 3rd ed., Part 1, Wiley-Interscience, New York (1970)

(ﬁ rst published in 1951).

20. C. O. Wilson, O. Gisvold, R. F. Doerge (Eds.). Textbook of Organic Medicinal and Pharmaceu-

tical Chemistry, 6th ed., Lippincott, Philadelphia (1971) (ﬁ rst published in 1949).

21. J. H. Block, E. B. Roche, T. O. Soine, C. O. Wilson (Eds.). Inorganic Medicinal and Pharma-

ceutical Chemistry, Lee & Febiger, Philadelphia (1974).

22. P. W. Erhardt. Drug metabolism data: past, present and future considerations, in Drug Metabo-

lism: Databases and High-Throughput Testing During Drug Design and Development, P. W.

Erhardt (Ed.), pp. 2–15, IUPAC/Blackwell, Geneva (1999).

23. P. W. Erhardt. Esmolol, in Chronicles of Drug Discovery, D. Lednicer (Ed.), Vol. 3, pp. 191–

206, American Chemical Society, Washington, DC (1993).

24. M. C. Venuti. Annu. Rep. Med. Chem. 25, 289–298 (1990).

25. K. R. Oldenburg. Annu. Rep. Med. Chem. 33, 301–311 (1998).

26. D. J. Ausman. Mod. Drug Discov., Jan., 18–23 (2001).

27. M. J. Felton. Mod. Drug Discov., Jan., 24–28 (2001).

28. P. J. Hylands and L. J. Nisbet. Annu. Rep. Med. Chem. 26, 259–269 (1991).

29. W. J. Dower and S. P. Fodor. Annu. Rep. Med. Chem. 26, 271–280 (1991).

30. W. H. Moos, G. D. Green, M. R. Pavia. Annu. Rep. Med. Chem. 28, 315–324 (1993).

31. M. A. Gallop, R. W. Barrett, W. J. Dower, S. Fodor, E. M. Gordon. J. Med. Chem. 37,

1234–1251 (1994).

32. E. M. Gordon, R. W. Barrett, W. J. Dower, S. Fodor, M. A. Gallop. J. Med. Chem. 37,

1385–1401 (1994).

33. N. K. Terrett. Combinatorial Chemistry, American Chemical Society/Oxford University Press,

New York (1998).

34. G. L. Trainor. Annu. Rep. Med. Chem. 34, 267–286 (1999).

35. J. N. Kyranos and J. C. Hogan, Jr. Modern Drug Discov., July/Aug., 73–82 (1999).

36. R. K. Brown. Mod. Drug Discov., July/Aug., 63–71 (1999).

37. S. Borman. Chem. Eng. News, May 15, 53–65 (2000).

38. E. LeProust, J. P. Pellois, P. Yu, H. Zhang, X. Gao. J. Comb. Chem. 2, 349–354 (2000).

39. D. Scharn, H. Wenschuh, U. Reineke, J. Schneider-Mergener, L. Germeroth. J. Comb. Chem. 2,

361–369 (2000).

40. R. E. Dolle. Mod. Drug Discov., Feb., 43–48 (2001) [the complete survey article, which appeared

in J. Comb. Chem. 2(5), 383–433 (2000) and includes 290 references, 10 tables, and 26 ﬁ gures,

is also available on the Web at http://pubs.acs.org/hotarticl/jcchff/2000/cc000055x_rev.html].

41. J. Boguslavsky. Drug Discov. Develop., Sept., 71 (1999).

42. G. Karet. Drug Discov. Develop., Mar., 28–32 (2000).

43. D. C. U’Prichard. Drug Discov. World, Fall, 12–22 (2000).

44. Personal communications while visiting several large pharmaceutical companies during 2000.

45. (a) C. D. Strader. J. Med. Chem. 39, 1 (1996); (b) P. S. Portoghese. Alfred Burger award in me-

dicinal chemistry. Molecular recognition at opiod receptors: from concepts to molecules, 219th

ACS Natl. Meet., Mar. 26–30, Abstract 302 (2000).

46. T. Maehama, G. S. Taylor, J. T. Slama, J. E. Dixon. Anal. Biochem. 279, 248–250 (2000).

47. J. T. Slama, S. Ramsinghani, D. W. Koh, J.-C. Ame, M. K. Jacobson. Identiﬁ cation of the active

site of poly(ADP-ribose)glycohydrolase using photoafﬁ nity labeling and site directed mutagen-

esis. 20th Midwest Enzyme Chem. Conf., Sept. 23, abstract 73 (2000).

48. J. T. Slama, N. Aboul-Ela, D. M. Goli, B. V. Cheesman, A. M. Simmons, M. K. Jacobson. J.

Med. Chem. 38, 389–393 (1995).

49. S. Ramsinghani, D. W. Koh, J.-C. Ame, M. Strohm, M. K. Jacobson, J. T. Slama. Biochemistry

37, 7801–7812 (1998).

50. H. Zeng, R. S. Miller, R. A. Flowers II, B. Gong. J. Am. Chem. Soc. 122, 2635–2644 (2000).

51. X.-P. Huang, F. E. Williams, S. M. Peseckis, W. S. Messer, Jr. J. Pharmacol. Exp. Ther. 286,

1129–1139 (1998).

52. X.-P. Huang, F. E. Williams, S. M. Peseckis, W. S. Messer, Jr. Am. Soc. Pharmacol. Exp. Ther.

56, 775–783 (1999).

53. X.-P. Huang, P. I. Nagy, F. E. Williams, S. M. Peseckis, W. S. Messer, Jr. Br. J. Pharmacol. 126,

735–745 (1999).

54. W. S. Messer, Jr., Y. F. Abuh, Y. Liu, S. Periyasamy, D. O. Ngur, M. A. N. Edgar, A. A. El-Assadi,

S. Sbeih, P. G. Dunbar, S. Roknich, T. Rho, Z. Fang, B. Ojo, H. Zhang, J. J. Huzl III, P. I. Nagy.

J. Med. Chem. 40, 1230–1246 (1997).

55. W. S. Messer, Jr., Y. F. Abuh, K. Ryan, M. A. Shepherd, M. Schroeder, S. Abunada, R. Sehgal,

A. A. El-Assadi. Drug Develop. Res. 40, 171–184 (1997).

56. W. S. Messer, Jr., W. G. Rajeswaran, Y. Cao, H.-J. Zhang, A. A. El-Assadi, C. Dockery, J. Liske,

J. O’Brien, F. E. Williams, X.-P. Huang, M. E. Wroblewski, P. I. Nagy, S. M. Peseckis. Pharm.

Acta Helv. 74, 135–140 (2000).

57. Y. Cao, M. E. Wroblewski, P. I. Nagy, W. S. Messer, Jr. Synthesis and biological evaluation

of bis[3-(3-alkoxy-1,2,5-thiadiazol-4-yl)-1,2,3,4-tetrahydropyrid-1-yl]alkane dihydrochlo-

rides as muscarinic agonists, 9th Int. Symp. Subtypes of Muscarinic Receptors, Nov., abstract

(2000).

58. W. S. Messer, Jr., W. G. Rajeswaran, Y. Cao, X.-P. Huang, M. E. Wroblewski, P. I. Nagy. Syn-

thesis and biological evaluation of bivalent xanomeline analogs as M1 muscarinic agonists, 9th

Int. Symp. Subtypes of Muscarinic Receptors, Nov., abstract (2000).

59. Y. Cao, X.-P. Huang, W. S. Messer, Jr. Soc. Neurosci. 26, abstract, p. 1912 (2000).

60. W. S. Messer, Jr., F. Liu, Y. Cao, F. E. Williams, X.-P. Huang. Soc. Neurosci. 26, abstract,

p. 1912 (2000).

61. P. I. Nagy. Recent Res. Develop. Phys. Chem. 3, 1–21 (1999).

62. P. S. Portoghese. Am. J. Pharm. Educ. 63, 342–347 (1999).

63. M. Jacobs. Chem. Eng. News, Oct. 30, 43–49 (2000).

64. H. Kubinyi. Curr. Opin. Drug Discov. Develop. 1(1), 4–15 (1998).

65. M. A. Murcko, P. R. Caron, P. S. Charifson. Annu. Rep. Med. Chem. 34, 297–306 (1999).

66. K. Rubenstein. Drug Market Develop., Report 1993, 1–200 (2000).

67. J. Josephson. Mod. Drug Discov., May, 45–50 (2000).

68. C. Hansch. Comprehensive Medicinal Chemistry, Wiley, New York (1990).

69. R. Silverman. The Organic Chemistry of Drug Design and Drug Action, Academic Press, New

York (1992).

70. M. E. Wolff (Ed.). Burger’s Medicinal Chemistry and Drug Discovery, 5th ed., Vol. 1, Wiley,

New York (1994).

REFERENCES AND NOTES 87

88 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

71. W. Foye, T. Lemke, D. Williams (Eds.). Principles of Medicinal Chemistry, 4th ed., Williams

& Wilkins, Baltimore (1995). During the last stages of this manuscript’s review, the author

was pleased to see that the latest edition (5th) of this text has explicitly incorporated an earlier

IUPAC deﬁ nition for medicinal chemistry.

72. A. Gringauz. Introduction to Medicinal Chemistry: How Drugs Act and Why, Wiley, New York

(1997).

73. H. J. Smith (Ed.). Smith and Williams’ Introduction to the Principles of Drug Design and Ac-

tion, 3rd ed., Harwood, Amsterdam (1998).

74. D. E. Stokes. Pasteur’s Quadrant: Basic Science and Technological Innovation, Brookings

Press, Washington, DC (1997).

75. F. A. Cotton. Chem. Eng. News, Dec. 4, 5 (2000).

76. T. Flores, D. Garcia, T. McKenzie, J. Pettigrew. In Innovations in Pharmaceutical Technology,

P. A. Barnacal (Ed.), pp. 18–22, Chancery Media., London (2000).

77. T. Studt. Drug Discov. Develop. Mar., 7 (2000).

78. E. Hodgkin and K. Andrews-Cramer. Modern Drug Discov., Jan./Feb., 55–60 (2000).

79. N. Sleep. Mod. Drug Discov., July/Aug., 37–42 (2000).

80. G. Karet. Drug Discov. Develop., Nov./Dec., 39–43 (1999).

81. J. Boguslavsky. Drug Discov. Develop., Mar., 67 (2000).

82. S. Young and J. Li. Drug Discov. Develop., Apr., 34–37 (2000).

83. M. B. Brennan. Chem. Eng. News, June 5, 63–73 (2000).

84. E. Estrada, E. Uriarte, A. Montero, M. Teijeira, L. Santana, E. De Clercq. J. Med. Chem. 43,

1975–1985 (2000).

85. H.-J. Boehm, M. Boehringer, D. Bur, H. Gmuender, W. Huber, W. Klaus, D. Kostrewa,

H. Kuehne, T. Luebbers, N. Meunier-Keller, F. Mueller. J. Med. Chem. 43, 2664–2674 (2000).

86. C. Bissantz, G. Folkers, D. Rognan. J. Med. Chem. 43, 4759–4767 (2000).

87. S. Borman. Chem. Eng. News, Dec. 18, 24–31 (2000).

88. S. Borman. Chem. Eng. News, Nov. 15, 55–57 (1993).

89. D. M. Spencer, T. J. Wandess, S. L. Schreiber, G. R. Crabtree. Science 262, 1019–1024 (1993).

90. A. C. Bishop, J. A. Ubersax, D. T. Petsch, D. P. Matheos, N. S. Gray, J. Blethrow, E. Shimizu,

J. Z. Tsien, P. G. Schultz, M. D. Rose, J. L. Wood, D. O. Morgan, K. M. Shokat. Nature 407,

395–401 (2000).

91. I. Wickelgren. Science 285, 998–1001 (1999).

92. R. Pettipher. In Innovations in Pharmaceutical Technology, P. A. Barnacal (Ed)., pp. 62–67,

Chancery Media, London (2000).

93. M. Swindells. In Innovations in Pharmaceutical Technology, P. A. Barnacal (Ed.), pp. 69–74,

Chancery Media, London (2000).

94. J. Boguslavsky. Drug Discov. Develop., June/July, 11–12 (2000).

95. S. Borman. Chem. Eng. News, Sept. 11, 6–7 (2000).

96. R. Willis and M. J. Felton. Modern Drug Discov., Jan., 31–36 (2001).

97. P. Elmer-DeWitt. Time 157(2), 57 (2001).

98. M. D. Lemonick. Time 157(2), 58–69 (2001).

99. National Human Genome Research Institute at www.nhgri.nih.gov; National Center for Bio-

technical Information in the NIH at www.ncbi.nlm.nih.gov; Human Genome Project via Gene

Bank at www.ncbi.nlm.nih.gov/genbank; and European Bioinformatics Institute at www.ebi.

ac.uk; as well as the home pages for the 16 institutions participating in the Human Genome

Sequencing Consortium.

100. AxCell Bioscience Corp. at www.axcellbio.com; BIND at http://bioinfo.mshri.on.ca; Celera

Genomics at www.celera.com; Curagen Corp. at www.curagen.com; Gene Logic at www.ge-

nelogic.com; Geneva Bioinformatics SA at www.genebio.com; Genome Database at www.gdb.

org; Genome Therapeutics Corp. at www.genomecorp.com; Hybrigenics SA at www.hybrigen-

ics.com; Incyte Genomics, Inc. at www.incyte.com; Lab On Web at www.labonweb.com; Life

Span BioSciences at www.lsbio.com; PRESAGE at http://csb.Stanford.edu; Protein Data Bank

at www.rcsb.org/pdb/; Proteome, Inc. at www.proteome.com; Structural Bioinformatics, Inc. at

www.strubix.com; and Swiss Institute of Bioinformatics at www.isbsib.ch/.

101. J. Buguslavksy. Drug Discov. Develop., Mar., 52–54 (2000).

102. C. S. W. Koehler. Mod. Drug Discov., Nov./Dec., 75–76 (2000).

103. J. M. Nash. Time 157(2), 90–93 (2001).

104. W. E. Evans and M. V. Relling. Science 286, 487–491 (1999).

105. D. S. Bailey and P. M. Dean. Annu. Rep. Med. Chem. 34, 339–348 (1999).

106. K. F. Lau and H. Sakul. Annu. Rep. Med. Chem. 35, 261–269 (2000).

107. R. Pettipher and R. Holford. Drug Discov. Develop., Jan./Feb., 53–54 (2000).

108. M. P. Murphy. Pharmacogenomics 1(2), 115–123 (2000).

109. M. P. Murphy. Drug Discov. World 1 (2), 23–32 (2000).

110. G. Karet, J. Boguslavsky, T. Studt. Drug Discov. Develop., Nov./Dec., 55–514 (2000).

111. J. Boguslavsky. Drug Discov. Develop., Nov./Dec., 521–524 (2000).

112. Curagen at www.curagen.com; Gene Logic at www.genelogic.com; Genome Therapeutics

Corp. at www.genomecorp.com; Immunogenetics Database at http://imgt.cnusc.fr:8104/;

Signalling Pathway Database at www.gvt.kyushu-u.ac.jpl/spad/.

113. The National Center for Biotechnical Information at www.ncbi.nlm.nih.gov/SNP/; the SNP

Consortium at http://snp.cshl.org/; as well as the home pages for the several organizations par-

ticipating in or supporting the SNP Consortium.

114. Affymetrix, Inc. at www.affymetrix.com; Applera Corp. at www.applera.com; Applied Bio-

systems at www.pebio.com/ab; Molecular Devices Corp. at www.moldev.com; Orchid BioSci-

ences Inc. at www.orchid.com; QIAGEN Genomics Inc. at www.qiagengenomics.com; Seque-

nom Inc. at www. sequenom.com; Third Wave Technologies at www.twt.com.

115. J. A. Halperin. The Standard (U.S. Pharmacopeia newsletter), Mar./Apr., 2 (2000).

116. C. M. Henry. Chem. Eng. News, Oct. 23, 85–100 (2000).

117. L. P. Rivory, H. Qin, S. J. Clarke, J. Eris, G. Duggin, E. Ray, R. J. Trent, J. F. Bishop. Eur. J.

Clin. Pharmacol. 56, 395–398 (2000).

118. O. Zelenkova, E. Hadasova, E. Ceskova, M. Vojtiskova, M. Hyksova. Human Psychopharma-

col. Clin. Exp. 15, 303–305 (2000).

119. L. Bellavance, J. Burbaum, D. Dunn. In Innovations in Pharmaceutical Technology, P. A.

Barnacal (Ed.), pp. 12–17, Chancery Media, London (2000).

120. J. Burbaum, C. Dilanni-Carrol, M. Traversari. Biotech. Int. II, 213–216 (1999).

121. J. Harness. In Innovations in Pharmaceutical Technology, P. A. Barnacal (Ed.), pp. 37–45,

Chancery Media, London (2000).

122. G. Karet. Drug Discov. Develop., June/July, 34–40 (2000).

123. J. Boguslavsky. Drug Discov. Develop., Aug./Sept., 51–54 (2000).

124. J. Boguslavsky. Drug Discov. Develop., Oct., 54–58 (2000).

125. J. Gibson and G. Karet. Drug Discov. Develop., June/July, 61–64 (2000).

126. J. A. Endicott and V. Ling. Annu. Rev. Biochem. 58, 137–171 (1989).

127. M. M. Gottesman and I. Pastan. Annu. Rev. Biochem. 62, 385–427 (1993).

REFERENCES AND NOTES 89

90 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

128. G. L. Scheffer, M. Kool, M. Heijn, M. deHaas, A. Pijnenborg, J. Wijnholds, A. van Helvoort, M.

C. de Jong, J. H. Hooijberg, C. Mol, M. van der Linden, J. de Vree, P. van der Valk, R. Elferink,

P. Borst, R. J. Scheper. Cancer Res. 60, 5269–5277 (2000).

129. J. D. Allen, R. F. Brinkhuis, L. van Deemter, J. Wijnholds, A. Schinkel. Cancer Res. 60,

5761–5766 (2000).

130. S. Gupta and T. Tsuruo (Eds.). Multidrug Resistance in Cancer Cells: Molecular, Biochemical,

Physiological and Biological Aspects, Wiley, Chichester, UK (1996).

131. J. A. Moscow, E. Schneider, S. P. Ivy, K. H. Cowan. Cancer Chemother. Biol. Response Modif.

17, 139–177 (1997).

132. M. M. Gottesman, I. Pastan, S. V. P. Ambudkar. Curr. Opin. Genet. Develop. 6, 610–617

(1996).

133. P. Gros, J. Croop, D. Housman. Cell 47, 371–380 (1986).

134. S. P. Cole, G. Bhardwaj, J. H. Gerlach, J. E. Mackie, C. E. Grant, K. C. Almquist, A. J. Stewart,

E. U. Kurz, A. M. Duncan, R. G. Deeley. Science 258, 1650–1654 (1992).

135. L. A. Doyle, W. Yang, L. V. Abruzzo, T. Krogmann, Y. Gao, A. K. Rishi, D. D. Ross. Proc. Natl.

Acad. Sci. USA 95, 15665–15670 (1998).

136. C. Ramachandran and S. J. Melnick. Mol. Diagn. 4, 81–94 (1999).

137. Values reported in the NIH/NCI database at http://dtp.nci.nih.gov/docs/cancer/searches/

standard_mechanism_list.html#topo1. Similar values are obtained in our labs using MCF-7

versus NCI/ADR-RES cell lines (unreported data).

138. M. E. Wall. In Chronicles of Drug Discovery, D. Lednicer (Ed.), pp. 327–348, American Chem-

ical Society, Washington, DC (1993).

139. M. Suffness. Annu. Rep. Med. Chem. 28, 305–314 (1993).

140. G. George, S. M. Ali, J. Zygmunt, L. R. Jayasinghe. Exp. Opin. Ther. Patents 4(2), 109–120

(1994).

141. K. C. Nicolaou, W.-M. Dai, R. K. Guy. Angew. Chem. Int. Ed. Engl. 33, 15–44 (1994).

142. G. I. Georg, T. Chen, I. Ojima, D. M. Vyas (Eds.) Taxane Anticancer Agents: Basic Science and

Current Status, American Chemical Society, Washington, DC (1995).

143. P. Sonneveld and E. Wiemer. Curr. Opin. Oncol. 9, 543–548 (1997).

144. V. Ling. Cancer Chemother. Pharmacol. 40, S3–S8 (1997).

145. G. A. Fisher and B. I. Sikic. Hematol. Oncol. Clin. North Am. 9, 363–382 (1995).

146. J. Sarver, P. Erhardt, W. Klis, J. Byers. Modiﬁ cation of anticancer drugs to avoid multiple drug

resistance, SGK 2000 Reaching for the Cure/Making a Difference Mission Conf., Sept. (2000).

147. W. D. Kingsbury, J. C. Boehm, D. R. Jakas, K. G. Holden, S. M. Hecht, G. Gallagher, M.

J. Caranfa, F. L. McCabe, L. F. Faucette, R. K. Johnson, R. P. Hertzberg. J. Med. Chem. 34,

98–107 (1991).

148. L. Liu. Adv. Pharmacol. 29B, 1–298 (1994).

149. B. Sinha. Drugs 49(1), 11–19 (1995).

150. P. Pantazis, B. Giovanella, M. Rothenberg (Eds.). Ann. N.Y. Acad. Sci. 803, 1–328 (1996).

151. C. Jaxel, K. Kohn, M. Wani, M. Wall, Y. Pommier. Cancer Res. 49, 1465–1469 (1989).

152. M. Wall, M. Wani, A. Nicholas, G. Manikumar, C. Tele, L. Moore, A. Truesdale, P. Leitner, J.

Besterman. J. Med. Chem. 36, 2689–2700 (1993).

153. Y. Pommier. Semin. Oncol. 23, 3–10 (1996).

154. M. Wall and M. Wani. Ann. N.Y. Acad. Sci. 803, 1–12 (1996).

155. D. Uehling, S. Nanthakumar, D. Croom, D. Emerson, P. Leitner, M. Luzzio, G. McIntyre, B.

Morton, S. Profeta, J. Sisco, D. Sternbach, W.-Q. Tong, A. Vuong, J. Besterman. J. Med. Chem.

38, 1106–1118 (1995).

156. S. Sawada, S. Matsuoka, K. Nokata, H. Nagata, T. Furuta, T. Yokokura, T. Miyasaka. Chem.

Pharm. Bull. 39, 3183–3188 (1991).

157. M. Sugimori, A. Ejima, S. Ohsuki, K. Uoto, I. Mitsui, K. Matsumoto, Y. Kawato, M. Yasuoka,

K. Sato, H. Tagawa, H. Terasawa. J. Med. Chem. 37, 3033–3039 (1994).

158. B. Giovanella, J. Stehlin, M. Wall, M. Wani, A. Nicholas, L. Liu, R. Silber, M. Potmesil. Sci-

ence 246, 1046–1048 (1989).

159. H. Hinz, N. Harris, E. Natelson, B. Giovanella. Cancer Res. 54, 3096–3100 (1994).

160. H. Wang, S. Liu, K. Hwang, G. Taylor, K.-H. Lee. Bioorg. Med. Chem. 2, 1397–1402 (1994).

161. M. Luzzio, J. Besterman, D. Emerson, M. Evans, K. Lackey, P. Leitner, G. McIntyre, B.

Morton, P. Myers, M. Peel, J. Sisco, D. Sternbach, W.-Q. Tong, A. Truesdale, D. Uehling, A.

Vuong, J. Yates. J. Med. Chem. 38, 395–401 (1995).

162. K. Lackey, J. Besterman, W. Fletcher, P. Leitner, B. Morton, D. Sternbach. J. Med. Chem. 38,

906–911 (1995).

163. K. Lackey, D. Sternbach, D. Croom, D. Emerson, M. Evans, P. Leitner, M. Luzzio, G. McIntyre,

A. Vuong, J. Yates, J. Besterman. J. Med. Chem. 39, 713–719 (1996).

164. A. Nicholas, M. Wani, G. Manikumar, M. Wall, K. Kohn, Y. Pommier. J. Med. Chem. 33,

972–978 (1990).

165. R. Crow and D. Crothers. J. Med. Chem. 35, 4160–4164 (1992).

166. C. Jaxel, K. W. Kohn, M. C. Wani, M. E. Wall, Y. Pommier. Cancer Res. 49, 1465–1469

(1989).

167. Y. Pommier, C. Jaxel, C. Heise. Structure–activity relationship of topoisomerase I inhibition by

camptothecin derivatives: evidence for the existence of a ternary complex, in DNA Topoisomer-

ases in Cancer, M. Potmesil and K. Kohn (Eds.), pp. 121–132, Oxford University Press, New

York (1991).

168. K. Lackey, J. Besterman, W. Fletcher, P. Leitner, B. Morton, D. Sternbach. J. Med. Chem. 38,

906–911 (1995).

169. M. Sugimori, A. Ejima, S. Ohsuki, K. Uoto, I. Mitsui, K. Matsumoto, Y. Kawato, M. Yasuoka,

K. Sato, H. Tagawa, H. Terasawa. J. Med. Chem. 37, 3033–3039 (1994).

170. S. Carrigan, P. Fox, M. Wall, M. Wani, J. Bowen. J. Comput.-Aided Mol. Des. 11, 71–78

(1997).

171. D. J. Rodi, R. W. Janes, H. J. Sanganee, R. A. Holton, B. A. Wallace, L. Makowski. J. Mol. Biol.

285, 197–203 (1999).

172. C.-P. Yang and S. B. Horwitz. Cancer Res. 60, 5171–5178 (2000).

173. N. Andre, D. Braguer, G. Brasseur, A. Goncalves, D. Lemesle-Meunier, S. Guise, M. A. Jordan,

D. Briand. Cancer Res. 60, 5349–5353 (2000).

174. P. B. Schiff, J. Fant, S. B. Horwitz. Nature 277, 665–667 (1979).

175. P. W. Erhardt. The Taxane J. 3, 36–42 (1997).

176. A. Nakashio, N. Fujita, S. Rokudai, S. Sato, T. Tsuruo. Cancer Res. 60, 5303–5309 (2000).

177. S. Froelich-Ammon and N. Osheroff. J. Biol. Chem. 270, 21429–21432 (1995).

178. W. Moos, G. D. Green, M. R. Pavia. Annu. Rep. Med. Chem. 28, 315–324 (1993).

179. D. C. Spellmeyer and P. D. J. Grootenhuis. Annu. Rep. Med. Chem. 34, 287–296 (1999).

180. A. Dorfman. Time 157(2), 97–99 (2001).

181. J.-S. Huang and K. Barker. Plant Physiol. 96, 1302–1307 (1991).

182. J. J. Cheong, R. Alba, F. Cote, J. Enkerli, M. G. Hahn. Plant Physiol. 103, 1173–1182 (1993).

183. J. Faghihi, X. Jiang, R. Vierling, S. Goldman, S. Sharfstein, J. Sarver, P. Erhardt. J. Chromatogr.

A915, 61–74 (2001).

184. S. Borman. Chem. Eng. News, Oct. 30, 35–36 (2000).

REFERENCES AND NOTES 91

92 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

185. V. Gruber. Annu. Rep. Med. Chem. 35, 357–364 (2000).

186. Interactive video conference: The role of higher education in economic development, Old

Dominion University Academic Television Service, Jan. 31, University of Toledo site (1997).

187. P. W. Erhardt. In Drug Metabolism Databases and High-Throughput Testing During Drug

Design and Development, P. W. Erhardt (Ed.), pp. viii–ix, IUPAC/Blackwell, Geneva (1999).

188. D. Ausman (Ed.). Data, data, everywhere. Molecular Connection (MDL’s Newsletter) 18(2),

3 (1999).

189. A. J. Recupero. Drug Discov. Develop., Nov./Dec., 59–62 (1999).

190. T. Agres. Drug Discov. Develop., Jan./Feb., 11–12 (2000).

191. T. Studt. Drug Discov. Develop., Jan./Feb., 68–69 (2000).

192. S. Zarrabian. Chem. Eng. News, Feb. 7, 5 (2000).

193. A.M. Thayer. Chem. Eng. News, Feb. 7, 19–32 (2000).

194. S. Tye and G. Karet. Drug Discov. Develop., Mar., 35 (2000).

195. W. Schulz. Chem. Eng. News, Feb. 5, 25 (2001).

196. DARPA at www.darpa.mil; FAO at www.fao.org/GENINFO/partner/defautl.htm; Gene Bank

at www.ncbi.nlm.nih.gov/entrez/nucleotide.html; DDBJ at www.ddbj.nig.ac.jp/; EMBL at

www.ebi.ac.uk/ebi_docs/embl_db/ebi/topembl.html; Human Genome Sequencing at www.

ncbi.nlm.nih.gov/genome/seq; KEEGG 12.0 at www.genome.ad. jp/kegg/; NASA at www.

nasa.gov; National Academy of Science STEP at www4.nas.edu/pd/step.nsf; National Center

for Biotechnical Information at www.ncbi.nlm.nih.gov; National Human Genome Research

Institute at www. nhgri.nih.gov/index.html; National Science Foundation Mathematics/

Physical Sciences at www.nsf.gov/home/mps /start.htm; Net Genics SYNERGY at www.

netgenics.com/SYNERGY.html; Pfam 4.3 at pfam.wustl.edu/; PROSITE at www3.oup.co.uk/

nar/Volume_27/Issue_01/html/gkc073_ gml.html#hdO; Sanger Center at www.sanger.ac.uk/;

SRS at www.embl–heidelberg–de/srs51; TIGER Database at www.tigr.org/tdb/tdb.html;

UNESCO at www.unesco.org/general/eng./ programmes/science/life/index.htm; University

of California–Berkeley at www.Berkeley.edu.

197. F. K. Brown. Annu. Rep. Med. Chem. 33, 375–384 (1998).

198. M. Hann and R. Green. Curr. Opin. Chem. Biol. 3, 379–383 (1999).

199. P. Ertl, W. Miltz, B. Rohde, P. Selzer. Drug Discov. World 1 (2), 45–50 (2000).

200. ChemFinder at www.chemﬁ nder.com; Chem Navigator at www.chemnavigator.com; Info

Chem GmbH at www.infochem.com; Institute for Scientiﬁ c Information at www.isinet.com;

MDL Information Systems at www.mdli.com; MSI at www.msi.com; Oxford Molecular at

www.oxmol.com; Pharmacopeia at www.pharmacopeia.com; QsarIS at www.scivision. com/

qsaris.html; Synopsys Scientiﬁ c Systems at www.synopsys.co.uk; Tripos at www.tripos.com.

201. Personal communications while attending several recent workshops directed toward expediting

drug development.

202. Protein Data Bank at www.rcsb.org/bdb/.

203. Biotech Validation Suite For Protein Structure at http://biotech.embl–heidelberg.de:8400/;

BLAST at www.ncbi.nlm.nih.gov/BLAST/; Build Proteins from Scratch at http://csb.stanford.

edu/levitt and www.chem.cornell.edu/department/Faculty/Scheraga/ Scheraga.html; CATH

at www.biochem.ucl.ac.uk/bsm/cath/; FASTA at www2.ebi.ac.uk/ fasta3/; MODELLER at

http://guitar.rockfeller.edu; PDB at www.pdb.bnl.gov/; Pedant at MIPS at http:pedant.mips.

biochem.mpg.del; PIR at www.nbrf.georgetown.edu/ pir/searchdb.html; Predict Protein at

www.embl–heidelber.de/predictprotein/ predictprotein.html; PSIPRED at http://insulin.

brunel.ac.uk/psipred; Rosetta at http://depts..Washington.edu/bakerpg; SAM–98 at www.cse.

ucsc.edu/research/compbio; SCOP at http://scop.mrc–lmb.cam.ac.uk/sopl; Structure at www.

ncbi.nlm.nih.gov/Structure/ RESEARCH/res.shtml; Swiss–Model at http://expasy.hcuge.

ch/swissmod/ SWISSMODEL.html; SWISS–PROT at http://expasy.hcuge.ch/sprot/sprottop.

html; Threader at http://insulin.brunel.ac.uk/threader/threader.html.

204. G. E. Schulz and R. H. Schirmer. In Principles of Protein Structure, C. R. Cantor (Ed.),

pp. 108–130, Springer-Verlag, New York (1999).

205. E. K. Wilson. Chem. Eng. News, Sept. 25, 41–44 (2000).

206. L. Holliman. Mod. Drug Discov., Nov./Dec., 41–46 (2000).

207. M. J. Felton. Mod. Drug Discov., Nov./Dec., 49–54 (2000).

208. M. Vieth and D. J. Cummins. J. Med. Chem. 43, 3020–3032 (2000).

209. M. A. Kastenholz, M. Pastor, G. Cruciani, E. Haaksma, T. Fox. J. Med. Chem. 43, 3033–3044

(2000).

210. G. Karet. Drug Discov. Develop., Jan., 28–32 (2001).

211. Personal communications while attending several recent scientiﬁ c conferences in the area of

medicinal chemistry.

212. ChemPen 3D at http://home.ici.net/~hfevans/chempen.htm; CS Chem Draw and CS Chem

Draw 3D at www.camsoft.com; Corina at www.mol–net.de.

213. A. Nezlin. Chem. News Commun. 9, 24–25 (1999).

214. M. Endres. Today’s Chem. at Work, Oct., 30–44 (1992).

215. G. Karet. Drug Discov. Develop., Aug./Sept., 42–48 (2000).

216. R. S. Rogers. Chem. Eng. News, Apr. 26, 17–18 (1999).

217. C. Hansch and T. Fujita. J. Am. Chem. Soc. 86, 1616–1626 (1964).

218. D. E. Koshland, Jr. and K. E. Neet. Annu. Rev. Biochem. 37, 359–410 (1968).

219. B. Belleau. Adv. Drug Res. 2, 89–126 (1965).

220. QUANTA (system for determining 3D protein structure from x-ray data) and FELIX (sys-

tem for determining 3D protein structure from NMR data) available from MSI-Pharmacopeia

(www.msi.com).

221. J. H. Krieger. Chem. Eng. News, Mar. 27, 26–41 (1995).

222. P. W. Erhardt. Epilogue. In Drug Metabolism Databases and High-Throughput Testing During

Drug Design and Development, P. W. Erhardt (Ed.), p. 320, IUPAC/Blackwell, Geneva (1999).

223. W. J. Hehre, L. Radom, P. vR. Schleyer, J. A. Pople. Ab Initio Molecular Orbital Theory, Wiley,

New York (1986).

224. A. D. Becke. J. Chem. Phys. 98, 5648–5652 (1993).

225. A. Medek, P. J. Hajduk, J. Mack, S. W. Fesik. J. Am. Chem. Soc. 122, 1241–1242 (2000).

226. M. Rouhi. Chem. Eng. News, Feb. 21, 30–31 (2000).

227. (a) B. C. Oostenbrink, J. W. Pitera, M. van Lipzig, J. H. Meerman, W. F. van Gunsteren. J.

Med. Chem. 43, 4594–4605 (2000); (b) N. Wu, Y. Mo, J. Gao, E. F. Pai. Proc. Natl. Acad. Sci.

USA 97, 2017–2022 (2000); (c) A. Warshel, M. Strajbl, J. Villa, J. Florian. Biochemistry 39,

14728–14738 (2000).

228. J. Devillers and A. T. Balaban (Eds.). Topological Indices and Related Descriptors in QSAR

and QSPR, Gordon and Breach, Amsterdam (1999).

229. C. Wrotnowski. Modern Drug Discov., Nov./Dec., 46–55 (1999).

230. C. M. Henry. Chem. Eng. News, Nov. 27, 22–26 (2000).

231. E. K. Wilson. Chem. Eng. News, Nov. 6, 35–39 (2000).

232. Z. Hu. Synthesis of C-13 side chain analogues related to the anticancer agent paclitaxel, M.S.

thesis, University of Toledo (1997).

233. Z. Hu and P. W. Erhardt. Org. Proc. Res. Develop. 1, 387–390 (1997).

234. Z. Hu, M. J. Hardie, P. Burckel, A. A. Pinkerton, P. W. Erhardt. J. Chem. Crystallog. 29(2),

185–191 (1999).

235. A. P. Marchand, Y. Wang, C.-T. Ren, V. Vidyasagar, D. Wang. Tetrahedron 52, 6063–6072

(1996).

REFERENCES AND NOTES 93

94 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

236. (a) M. Pomerantz and P. H. Hartman. Tetrahedron Lett. 991–993 (1968); (b) N. J. Turro and

P. A. Wriede. J. Am. Chem. Soc. 92, 320–329 (1970).

237. F. D. Lewis and N. J. Turro. J. Am. Chem. Soc. 92, 311–320 (1970).

238. L. E. Friedrich and P. Y.-S. Lam. J. Org. Chem. 46, 306–311 (1981).

239. W. A. Klis and P. W. Erhardt. Synth. Commun. 30, 4027–4038 (2000).

240. P. W. Erhardt, W. A. Klis, P. I. Nagy, K. Kirschbaum, N. Wu, A. Martin, A. A. Pinkerton. J.

Chem. Crystallogr. 30, 83–90 (2000).

241. H. A. Carlson, K. M. Masukawa, K. Rubins, F. D. Bushman, W. L. Jorgensen, R. D. Lins, J. M.

Briggs, J. A. McCammon. J. Med. Chem. 43, 2100–2114 (2000).

242. E. K. Bradley, P. Beroza, J. E. Penzotti, P. D. Grootenhuis, D. C. Spellmeyer, J. L. Miller. J.

Med. Chem. 43, 2770–2774 (2000).

243. M. Graffner-Nordberg, J. Marelius, S. Ohlsson, A. Persson, G. Swedberg, P. Andersson, S. E.

Andersson, J. Aqvist, A. Hallberg. J. Med. Chem. 43, 3852–3861 (2000).

244. R. Garcia-Nieto, I. Manzanares, C. Cuevas, F. Gago. J. Med. Chem. 43, 4367–4369 (2000).

245. A. Vedani, H. Briem, M. Dobler, H. Dollinger, D. R. McMasters. J. Med. Chem. 43, 4416–4427

(2000).

246. M. L. Lopez-Rodriguez, M. J. Morcillo, M. Fernandez, M. L. Rosada, L. Pardo, K. J. Schaper.

J. Med. Chem. 44, 198–207 (2001).

247. B. Ladd. Mod. Drug Discov., Jan./Feb., 46–52 (2000).

248. V. Venkatsubramanian. Computer-aided molecular design using neural networks and genetic

algorithms, in Genetic Algorithms in Molecular Modeling, J. Devillers (Ed.), Academic Press,

New York (1996).

249. A. Globus. Nanotechnology 10(3), 290–299 (1999).

250. T. Studt. Drug Discov. Develop., Aug./Sept., 30–36 (2000).

251. R. Resnick. Drug Discov. Develop., Oct., 51–52 (2000).

252. J. Jaen-Oltra, T. Salabert-Salvador, F. J. Garcia-March, F. Perez-Gimenez, F. Tohmas-Vert. J.

Med. Chem. 43, 1143–1148 (2000).

253. Bio Reason at www.bioreason.com; Columbus Molecular Software (Lead Scope) at www.

columbus- molecular.com; Daylight Chemical Information Systems at www.daylight.com;

IBM (Intelligent Miner) at www.ibm.com; Incyte (Life Tools and Life Prot) at www.incyte.

com; Molecular Applications (Gene Mine 3.5.1) at www.mag.com; Molecular Simulations

at www.msi.com; Oxford Molecular (DIVA 1.1) at www.oxmol.com; Pangea Systems (Gene

World 3.5) at www.pangeasystems.com; Pharsight at www.pharsight.com; SAS Institute

(Enterprise Miner) at www.sas.com; SGI (Mine Set 3.0) at www.sgi.com; Spotﬁ re (Spotﬁ re

Pro 4.0 and Leads Discover) at www.spotﬁ re.com; SPSS (Clementine) at www.spss.com;

Tripos at www.tripos. com.

254. M. A. Johnson, E. Gifford, C.-C. Tsai. In Concepts and Applications of Molecular Similarity,

M. A. Johnson (Ed.), pp. 289–320, Wiley, New York (1990).

255. E. Gifford, M. Johnson, C.-C. Tsai. J. Comput.-Aided Mol. Des. 5, 303–322 (1991).

256. E. M. Gifford, M. A. Johnson, D. G. Kaiser, C.-C. Tsai. J. Chem. Inf. Comp. Sci. 32, 591–599

(1992).

257. E. M. Gifford, M. A. Johnson, D. G. Kaiser, C.-C. Tsai. Xenobiotica 25, 125–146 (1995).

258. E. M. Gifford. Applications of molecular similarity methods to visualize xenobiotic metabo-

lism structure–reactivity relationships, Ph.D. thesis, University of Toledo (1996).

259. K. M. Andrews and R. D. Cramer. J. Med. Chem. 43, 1723–1740 (2000).

260. T. Borowski, M. Krol, E. Broclawik, T. C. Baranowski, L. Strekowski, M. J. Mokrosz. J. Med.

Chem. 43, 1901–1909 (2000).

261. E. A. Wintner and C. C. Moallemi. J. Med. Chem. 43, 1933–2006 (2000).

262. D. B. Boyd. Modern Drug Discov., Nov/Dec., 41–48 (1998).

263. Personal communications while visiting several large pharmaceutical companies and while at-

tending recent scientiﬁ c conferences in the area of medicinal chemistry.

264. B. E. Evans, K. E. Rittle, M. G. Bock, R. M. DiPardo, R. M. Freidinger, W. L. Whitter, G. F.

Lundell, D. F. Veber, P. S. Anderson, R. S. L. Chang, V. J. Lotti, D. J. Cerino, T. B. Chen, P. J.

Kling, K. A. Kunkel, J. P. Springer, J. Hirshﬁ eld. J. Med. Chem. 31, 2235–2246 (1988).

265. A. A. Patchett and R. P. Nargund. Annu. Rep. Med. Chem. 35, 289–298 (2000).

266. M. Pastor, G. Cruciani, I. McLay, S. Pickett, S. Clementi. J. Med. Chem. 43, 3233–3243

(2000).

267. C. Gnerre, M. Catto, F. Leonetti, P. Weber, P.-A. Carrupt, C. Altomare, A. Carotti, B. Testa. J.

Med. Chem. 43, 4747–4758 (2000).

268. R. E. Wilcox, W. H. Huang, M. Y. Brusniak, D. M. Wilcox, R. S. Pearlman, M. M. Teeter, C. J.

DuRand, B. L. Wiens, K. A. Neve. J. Med. Chem. 43, 3005–3019 (2000).

269. P. R. N. Jayatilleke, A. C. Nair, R. Zauhar, W. J. Welsh. J. Med. Chem. 43, 4446–4451 (2000).

270. R. Wedin. Modern Drug Discov., Sept./Oct., 39–47 (1999).

271. S. H. Yalkowski and S. C. Valvani. J. Pharm. Sci. 69, 912–922 (1980).

272. R. P. Mason, D. G. Rhodes, L. G. Herbette. J. Med. Chem. 34, 869–877 (1991).

273. L. G. Herbette, D. G. Rhodes, R. P. Mason. Drug Des. Deliv. 7, 75–118 (1991).

274. M. H. Abraham, S. Chada, R. Mitchell. J. Pharm. Sci. 83, 1257–1268 (1994).

275. S. C. Basak, B. D. Gute, L. R. Drewes. Pharm. Res. 13, 775–778 (1996).

276. M. Kansy, F. Senner, K. Gubernator. J. Med. Chem. 41, 1007–1010 (1998).

277. K. Palm, K. Luthman, A.-L. Ungell, G. Strandlund, F. Beigi, P. Lundahl, P. Artursson. J. Med.

Chem. 41, 5382–5392 (1998).

278. D. A. Smith and H. Van de Waterbeemd. Curr. Opin. Chem. Biol. 3, 373–378 (1999).

279. B. H. Stewart, Y. Wang, N. Surendran. Annu. Rep. Med. Chem. 35, 299–307 (2000).

280. P. Crivori, G. Cruciani, P.-A. Carrupt, B. Testa. J. Med. Chem. 43, 2204–2216 (2000).

281. F. Yoshida and J. G. Topliss. J. Med. Chem. 43, 2575–2585 (2000).

282. F. Lombardo, M. Y. Shalaeva, K. A. Tupper, F. Gao, M. H. Abraham. J. Med. Chem. 43,

2922–2928 (2000).

283. P. Ertl, B. Rohde, P. Selzer. J. Med. Chem. 43, 3714–3717 (2000).

284. W. J. Egan, K. M. Merz, Jr., J. J. Baldwin. J. Med. Chem. 43, 3867–3877 (2000).

285. A. Chait. Drug Discov. Develop., Apr., 63 (2000).

286. L. Pickering. Drug Discov. Develop., Jan., 34–38 (2001).

287. Absorption Systems’ BCS Biowaivers at www.absorption.com; ArQule’s Pilot at www.argule.

com; Camitro Corp’s ADME/Tox at www.camitro.com; Cerep SA’s BioPrints at www.cerep.

fr; Exon Hit Therapeutics SA’s Genetic Makeup/Drug Toxicity at www.exonhit.com; LION

Bioscience AG’s i-Biology at SRS Databases at www.lionbioscience.com; Quintiles’ Early

ADMET Screening at qkan.busdev@quintiles.com; Schrodinger, Inc’s QikProp at http://

www.schrodiner.com; Simulations Plus’s GastroPlus at www.simulationsplus.com; Trega

Biosciences’ In Vitro ADME at www.trega.com.

288. G. L. Amidon, P. I. Lee, E. M. Topp (Eds.). Transport Processes in Pharmaceutical Systems,

Marcel Dekker, New York (1999).

289. G. Zimmer (Ed.). Membrane Structure in Disease and Drug Therapy, Marcel Dekker, New

York (2000).

290. J. B. Dressman and H. Lennernas (Eds.). Oral Drug Absorption: Prediction and Assessment,

Marcel Dekker, New York (2000).

291. E. W. Taylor, J. A. Gibbons, R. A. Braeckman. Pharm. Res. 14, 572–577 (1997).

REFERENCES AND NOTES 95

96 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

292. J. A. Gibbons, E. W. Taylor, C. M. Dietz, Z.-P. Luo, H. Luo, R. A. Braeckman. In Drug Metabo-

lism: Databases and High-Throughput Testing During Drug Design and Development, P. W.

Erhardt (Ed.), pp. 71–78, IUPAC/Blackwell, Geneva (1999).

293. M. Rouhi. Chem. Eng. News, Jan. 29, 10 (2001).

294. V. Janout, C. DiGiorgio, S. L. Regen. J. Am. Chem. Soc. 122, 2671–2672 (2000).

295. P. A. Wender, D. J. Mitchell, K. Pattabiraman, E. T. Pelkey, L. Steinman, J. B. Rothbard. Proc.

Natl. Acad. Sci. USA 97, 13003–13008 (2000).

296. J. B. Rothbard, S. Garlington, Q. Lin, T. Kirschberg, E. Kreides, P. L. McGrove, P. A. Wendy,

P. A. Kharari. Nature Med. 6, 1253–1257 (2000).

297. M. Rouhi. Chem. Eng. News, Sept. 4, 43 (2000).

298. R. Dagani. Chem. Eng. News, Oct. 16, 27–32 (2000).

299. W. Schulz. Chem. Eng. News, Oct. 16, 39–42 (2000).

300. P. Zurer. Chem. Eng. News, Mar. 12, 12 (2001).

301. K. Park and R. J. Mrsny (Eds.). Controlled Drug Delivery, American Chemical Society,

Washington, DC (2000).

302. M. Mort. Med. Drug Discov. Apr., 30–34 (2000).

303. C. M. Henry. Chem. Eng. News, Sept. 18, 49–65 (2000).

304. S. Bracht. Innovations Pharm. Technol. 5, 92–98 (2000).

305. A. Levy. Innovations Pharm. Technol. 5, 100–109 (2000).

306. J. Southall and C. Ellis. Innovations Pharm. Technol. 5, 110–115 (2000).

307. C. Winnips and M. Keller. Innovations Pharmaceut. Tech. 6, 70–75 (2000).

308. C. Winnips. Drug Discov. World 2(1), 62–66 (2000/2001).

309. K. J. Watkins. Chem. Eng. News, Jan. 8, 11–15 (2000).

310. M. Jacoby. Chem. Eng. News, Feb. 5, 30–35 (2001).

311. C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney. Adv. Drug Deliv. Rev. 23, 3–25

(1997).

312. F. Darvas, K. Valko, G. Dorman, P. W. Erhardt, M. T. D. Cronin, I. Szabo. Biol. Techniques,

Accepted for publication.

313. A. Frostell-Karlsson, A. Remaeus, H. Roos, K. Andersson, P. Borg, M. Hamalainen, R. Karlsson.

J. Med. Chem. 43, 1986–1992 (2000).

314. N. Diaz, D. Suarez, T. L. Sordo, K. M. Merz, Jr. J. Med. Chem. 44, 250–260 (2001).

315. P. J. Hajduk, M. Bures, J. Praestgaard, S. W. Fesik. J. Med. Chem. 43, 3443–3447 (2000).

316. M. P. Czech. CIIT Activities (Chemical Industry Institute of Toxicology newsletter), May/June,

2–4 (2000).

317. M. A. Dechantsreiter, E. Planker, B. Matha, E. Lohof, G. Holzemann, A. Jonczyk, S. L.

Goodman, H. Kessler. J. Med. Chem. 42, 3033–3040 (1999).

318. R. Pasqualini, E. Koivunen, R. Kain, J. Lahdenranta, M. Sakamoto, A. Stryhn, R. A. Ashmun,

L. H. Shapiro, W. Arap, E. Ruoslahti. Cancer Res. 60, 722–727 (2000).

319. A. Eberhard, S. Kahlert, V. Goede, B. Hemmerlein, K. H. Plate, H. G. Augustin. Cancer Res.

60, 1388–1393 (2000).

320. L. Wyder, A. Vitaliti, H. Schneider, L. W. Hebbard, D. R. Moritz, M. Wittmer, M. Ajmo,

R. Klemenz. Cancer Res. 60, 4682–4688 (2000).

321. M. Matsuda, S.-I. Nishimura, F. Nakajima, T. Nishimura. J. Med. Chem. 44, 715–724 (2001).

322. T. G. Metzger, M. G. Paterlini, P. S. Portoghese, D. M. Ferguson. Neurochem. Res. 21,

1287–1294 (1996).

323. D. L. Larson, R. M. Jones, S. A. Hjorth, T. W. Schwartz, P. S. Portoghese. J. Med. Chem. 43,

1573–1576 (2000).

324. J. L. Marx. Science 216, 283–285 (1982).

325. P. A. Baeuerle and E. Wolf. Mod. Drug Discov. Apr., 37–42 (2000).

326. G. W. Geelhoed. S. Afr. J. Surg. 26, 1–3 (1988).

327. P. Ehrlich. Lancet, 445–451 (1913).

328. A. Albert. Selective Toxicity, Wiley, New York (1960).

329. (a) T. Higuchi and V. Stella (Eds.). Prodrugs as Novel Drug Delivery Systems, American Chem-

ical Society, Washington, DC (1975); (b) H. Bundgaard (Ed.). Design of Prodrugs, Elsevier,

Amsterdam (1985).

330. F. Wood. The Blade, Toledo, OH Jan. 25, 3A (2000).

331. K. Howe. San Francisco Chronicle, Jan. 21, 5 (2000).

332. J. A. Barnes. Inpharma 1185, May (1999).

333. W.-M. Keung, O. Lazo, L. Kunze, B. Vallee. Proc. Natl. Acad. Sci. USA 93, 4284–4289 (1996).

334. F. Stermitz, P. Lorenz, J. Tawara L. Zenewicz, K. Lewis. Proc. Natl. Acad. Sci. USA 97,

1433–1438 (2000).

335. N. R. Guz and F. R. Stermitz. J. Nat. Prod. 63, 1140–1145 (2000).

336. W. M. Keung, O. Lazo, L. Kunze, B. L. Vallee. Proc. Natl. Acad. Sci. USA 92, 8990–8993 (1995).

337. S. Verma, E. Salamone, B. Goldin. Biochem. Biophys. Res. Commun. 233, 692–696 (1997).

338. M. Mammen, S. K. Choi, G. M. Whitesides. Angew. Chem. Int. Ed. 2754–2794 (1998).

339. L. L. Kiessling, L. E. Strong, J. E. Gestwicki. Annu. Rep. Med. Chem. 35, 321–330 (2000).

340. H. Annoura, K. Nakanishi, T. Toba, N. Takemoto, S. Imajo, A. Miyajima, Y. Tamura-Horikawa,

S. Tamura. J. Med. Chem. 43, 3372–3376 (2000).

341. A. Gangjee, J. Yu, J. J. McGuire, V. Cody, N. Galitsky, R. L. Kisliuk, S. F. Queener. J. Med.

Chem. 43, 3837–3851 (2000).

342. S. Borman. Chem. Eng. News, Oct. 9, 48–53 (2000).

343. S. J. Sucheck, A. L. Wong, K. M. Koeller, D. D. Boehr, K. Draker, P. Sears, G. D. Wright, C.-H.

Wong. J. Am. Chem. Soc. 122, 5230–5231 (2000).

344. K. Bachmann and R. Belloto, Jr. Drugs and Ageing 15, 235–250 (1999).

345. K. Malmstrom, J. Schwartz, T. F. Reiss, T. J. Sullivan, J. H. Reese, L. Jauregui, K. Miller,

M. Scott, S. Shingo, I. Peszek, P. Larson, D. Ebel, T. L. Hunt, R. Huhn, K. Bachmann. Am. J.

Ther. 5, 189–195 (1998).

346. L. S. Mauro, D. A. Kuhl, J. R. Kirchhoff, V. F. Mauro, R. W. Hamilton. Am. J. Ther. 8, 21–25

(2001).

347. V. F. Mauro, L. S. Mauro, J. F. Kleshinski, P. W. Erhardt. Pharmacotherapy 21, 373, abstract

(2001).

348. Y. Brun, C. Wrzossek, J. Parsons, H. Slocum, D. White, W. Greco. Am. Assoc. Cancer Res.

Proc. 41, 102 (Abstract) (2000).

349. D. White and L. James. J. Clin. Epidemiol. 49, 419–429 (1996).

350. (a) W. Greco, G. Bravo, J. Parsons. Pharmacol. Rev. 47, 331–385 (1995); (b) H. M. Faessel,

H. K. Slocum, Y. M. Rustum, W. R. Greco. Biochem. Pharmacol. 57, 567–577 (1999).

351. Paired agent combinations  100!/(2! 98!)  4950 agent pairs. Test each pair at 3 concentra-

tions for each agent  9 combinations per pair. Number of pair tests  9(4950)  44,550. Num-

ber of single agent tests at 3 concentrations per agent  3(100)  300. Total tests for an N  1

pass through the library  44,550  300  44,850.

352. N. Bodor. In Strategy in Drug Research, J. A. K. Buisman (Ed.), pp. 137–164, Elsevier,

Amsterdam (1982).

353. N. Bodor and P. Buchwald. Med. Res. Rev. 20, 58–101 (2000).

354. P. Erhardt. Chron. Drug Discov. 3, 191–206 (1993).

REFERENCES AND NOTES 97

98 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

355. P. W. Erhardt, C. M. Woo, R. J. Gorczynski, W. G. Anderson. J. Med. Chem. 25, 1402–1407

(1982).

356. P. W. Erhardt, C. M. Woo, W. G. Anderson, R. J. Gorczynski. J. Med. Chem. 25, 1408–1412

(1982).

357. P. W. Erhardt, C. M. Woo, W. L. Matier, R. J. Gorczynski, W. G. Anderson. J. Med. Chem. 26,

1109–1112 (1983).

358. J. Liao, K. Golding, K. Dubischar, J. Sarver, P. Erhardt, H. Kuch, M. Aouthmany. Determination

of blood esterase capability in newborns, 27th Nat. ACS Med. Chem. Symp., June (2000).

359. J. Liao. Development of soft drugs for the neonatal population, M.S. thesis, University of

Toledo (2000).

360. P. Erhardt. Using drug metabolism databases as predictive tools during drug design and devel-

opment: past, present and future status, Soc. Tox. Canada 32nd Annu. Symp., Dec. (1999).

361. J. F. Sina. Annu. Rep. Med. Chem. 33, 283–291 (1998).

362. R. B. Conolly. CIIT Activities (Chemical Industry Institute of Toxicology newsletter) 19 (10),

6–8 (1999).

363. M. Jacobs. Chem. Eng. News, Mar. 13, 29 (2000).

364. J. Boguslavsky. Drug Discov. Develop. June/July, 43–46 (2000).

365. M. Eggers. Innovations Pharm. Technol. 6, 36–44 (2000).

366. J. Boguslavsky. Drug Discov. Develop., Oct., 30–36 (2000).

367. R. E. Neft, S. Farr. Drug Discov. World 1(2), 33–34 (2000).

368. For general toxicology: Cellomics, Inc. at www.cellomics.com; Celera, Inc. at http://www.

celera.com; Chemical Industry Institute of Toxicology at http://www.ciit.org; Inpharmatica at

www.inpharmatica. co.uk; National Center for Biotechnical Information at http://www.ncbi.

nlm.nih.gov/Web/Genebank/index.html and at http://www.ncbi.nlm.nih.gov/Entvez/Genonome/

org.html; Rosetta Inpharmatics at www.vii.com; SciVision (ToxSYS) at www.scivision. com.

For drug–drug interactions: Georgetown University at http://dml.Georgetown.edu/depts/

pharmacology/clinlist.html; FDA at http://www.fda.gov/cder/drug/advisory/stjwort.htm; Public

Citizen Health Research Group at http://www.citizen.org/hrg/newsletters/pillnews.htm.

369. R. Kumar. Innovations Pharm. Technol. 6, 120–123 (2000).

370. B. Lewis, M. Renn, K. Manning. Innovations Pharm. Technol. 6, 125–128 (2000).

371. M. McCoy. Chem. Eng. News, Mar. 12, 27–42 (2001).

372. Personal communications during the 175th Anniversary USP Convention, Mar. (1995).

373. P. Anastas and T. C. Williamson (Eds.). Green Chemistry: Frontiers in Benign Chemical Syn-

theses and Processes, Oxford University Press, New York (1998).

374. P. Anastas and J. Warner. Green Chemistry: Theory and Practice, Oxford University Press, New

York (1998).

375. P. T. Anastas, L. G. Heine, T. C. Williamson (Eds.). Green Chemical Syntheses and Process

(Part 1) and Green Engineering (Part 2), Oxford University Press, New York (2000).

376. P. Tundo and P. Anastas (Eds.). Green Chemistry: Challenging Perspectives, Oxford University

Press, New York (2000).

377. S. C. Stinson. Chem. Eng. News, Oct. 23, 55–78 (2000).

378. K. Auclair, A. Sutherland, J. Kennedy, D. J. Witter, J. P. Van den Heever, C. R. Hutchinson, J.

C. Vederas. J. Am. Chem. Soc. 122, 11519–11520 (2000).

379. K. Watanabe, H. Oikawa, K. Yagi, S. Ohashi, T. Mie, A. Ichihara, M. Honma. J. Biochem. 127,

467–473 (2000).

380. K. Watanabe, T. Mie, A. Ichihara, H. Oikawa, M. Honma. J. Biol. Chem. 275, 38393–38401

(2000).

381. S. Borman. Chem. Eng. News, Jan. 15, 9 (2001).

382. T. W. Greene and P. G. M. Wuts. Protective Groups in Organic Synthesis, 3rd ed., Wiley, New

York (1999).

383. L. J. Silverberg, S. Kelly, P. Vemishetti, D. H. Vipond, F. S. Gibson, B. Harrison, R. Spector, J.

L. Dillon. Org. Lett. 2, 3281–3283 (2000).

384. P. W. Erhardt. Synth. Commun. 13, 103–113 (1983).

385. S. Stinson. Chem. Eng. News, Dec. 4, 35–44 (2000).

386. (a) R. Baltzly and P. B. Russell. J. Am. Chem. Soc. 75, 5598–5602 (1953); (b) M. Freifelder

(Ed.) Catalytic Hydrogenation in Organic Synthesis Procedures and Commentary, Wiley, New

York (1978).

387. Y. Ni and P. Erhardt. Synthesis of steric probes to examine the chemical behavior of potential

benzylamine-related chiral auxiliary synthetic reagents, ACS Nat. Meet., Apr. (1997).

388. Y. Ni. Synthesis and N-debenzylation of steric probes to deﬁ ne the practical limit for employ-

ing potential benzylamine-related chiral auxiliary reagents, M.S. thesis, University of Toledo

(1996).

389. P. Erhardt. U.S. patent 5,977,409, 1–25 (1999).

390. C. Zhang and P. Erhardt. Benzylamine-related chiral synthetic reagents, Natl. ACS Med. Chem.

Symp., June (2000).

391. C. Zhang. Benzylamine-related chiral synthetic reagents: asymmetric opening of terminal

epoxides, M.S. thesis, University of Toledo (1999).

392. P. Coppens (Ed.). X-Ray Charge Densities and Chemical Bonding, Oxford University Press,

New York, (1997).

393. (a) V. S. Lamzin, R. J. Morris, Z. Dauter, K. S. Wilson, M. M. Teeter. J. Biol. Chem. 274,

20753–20755 (1999); (b) D. Housset, F. Benabicha, V. Pichon-Pesme, C. Jelsch, A. Maierhofer,

S. David, J. C. Fontecilla-Camps, C. Lecomte. ACTA Crystallogr. Sect. D Biol. Crystallogr.

56, 151–160 (2000); (c) C. Jelsch, M. M. Teeter, V. Lamzin, V. Pichon-Pesme, R. H. Blessing,

C. Lecomte. Proc. Natl. Acad. Sci. USA 97, 3171–3176 (2000); (d) M. V. Fernandez-Serra,

J. Junquera, C. Jelsch, D. Lecomte, E. Artacho. Solid State Commun. 116, 395–400 (2000).

394. A. Martin and A. A. Pinkerton. Acta Crystallogr. B54, 471–478 (1998).

395. A. A. Pinkerton. In Electron, Spin and Momentum Densities and Chemical Reactivity, P. G.

Mezey and B. Robertson (Eds.), pp. 213–223, Klewer, London (2000).

396. B. B. Iversen, F. K. Larsen, A. A. Pinkerton, A. Martin, A. Darovsky, P. A. Reynolds. Acta

Crystallogr. B55, 363–374 (1999).

397. K. Palczewski, T. Kumasaka, T. Hori, C. A. Behnke, H. Motoshima, B. A. Fox, I. Le Trong,

D. C. Teller, T. Okada, R. E. Stenkamp, M. Yamamoto, M. Miyano. Science 289, 739–745

(2000).

398. V. Comello. Drug Discov. Develop. Apr., 26–28 (2000).

399. (a) N. Ban, P. Nissen, J. Hansen, P. B. Moore, T. A. Steitz. Science 289, 905–920 (2000). (b) P.

Nessen, J. Hansen, N. Ban, P. B. Moore, T. A. Steitz. Science 289, 920–930 (2000).

400. F. Schluenzen, A. Tocilj, R. Zarivach, J. Harms, M. Gluehmann, D. Janell, A. Bashan, H.

Bartels, I. Agmon, F. Franceschi, A. Yonath. Cell 102, 615–623 (2000).

401. (a) B. T. Wimberly, D. E. Brodersen, W. M. Clemons, Jr., R. J. Morgan-Warren, A. P. Carter,

C. Vonrhein, T. Hartsch, V. Ramakrishnan. Nature 407, 327–339 (2000); (b) B. L. Hanson, A.

Martin, J. M. Harp, C. G. Bunick, D. A. Parrish, K. Kirschbaum, G. J. Bunick, A. A. Pinkerton.

J. Appl. Crystallogr. 32, 814–820 (1999); (c) B. L. Hanson, J. M. Harp, K. Kirschbaum, D. A.

Parrish, D. E. Timm, A. Howard, A. A. Pinkerton, G. J. Bunick. J. Cryst. Growth. In press.

402. M. J. Hardie, K. Kirschbaum, A. Martin, A. A. Pinkerton. J. Appl. Crystallogr. 31, 815–817

(1998).

403. P. W. Erhardt. J. Pharm. Sci. 69, 1059–1061 (1980).

404. P. W. Erhardt and Y.-L. Chou. Life Sci. 49, 553–568 (1991).

REFERENCES AND NOTES 99

100 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

405. S. Moran and L. Stewart. Drug Discov. World 2(1), 41–46 (2000).

406. Z. P. Wu. Accelerating lead and solid form selection in drug discovery and development using

non-ambient x-ray diffraction, New Chemical Technologies Satellite Program, 221st Nat. ACS

Meet., Mar. (2001).

407. D. A. Pearlman and P. S. Charifson. J. Med. Chem. 44, 502–511 (2001).

408. R. W. Schoenlein, S. Chattopadhyay, H. H. W. Chong, T. E. Glover, P. A. Heimann, C. V. Shank,

A. A. Zholents, M. S. Zolotorev. Science 287, 2237–2240 (2000).

409. E. Wilson. Chem. Eng. News, Mar. 27, 7, 8 (2000).

410. E. J. Ariens (Ed.). Molecular Pharmacology: The Mode of Action of Biologically Active Com-

pounds, Vol. I, Academic Press, New York (1964).

411. I. Schlichting, J. Berendzen, K. Chu, A. M. Stock, S. A. Maves, D. E. Benson, R. M. Sweet, D.

Ringe, G. A. Petsko, S. G. Sligar. Science 287, 1615–1622 (2000).

412. A. M. Rouhi. Chem. Eng. News, Mar. 13, 50–51 (2000).

413. E. K. Wilson. Chem. Eng. News, Apr. 24, 39–45 (2000).

414. U. Kher. How to design a molecule, Time, Jan. 15, 67 (2001).

415. P. W. Erhardt. In Endothelin, G. M. Rubanyi (Ed.), pp. 41–57, Oxford University Press, New

York (1992).

416. C. R. Ganellin. Chem. Int. 23(2), 43–45 (2001).

417. J. Brent. CLA scholars spin curiosity into discovery, Univ. Minn. CLA Today (1998–1999

Annual Report). Fall, 10–13 (1999).

418. P. W. Erhardt. J. Org. Chem. 44, 883 (1979).

419. J. Parness and S. B. Horwitz. J. Cell Biol. 91, 479–587 (1981).

420. A. B. Shapiro and V. Ling. Eur. J. Biochem. 250, 130–137 (1997).

421. For example, respectively: the long-standing Pharmacology for Medicinal Chemistry course

often offered by the ACS as a satellite program to their national scientiﬁ c meetings; the

Residential School on MC offered at Drew University (http://www.depts.drew.edu/resmed);

Drug Metabolism for Medicinal Chemists short course that I offer on-site; and a long-distance

Continuing Learning Program in MC offered by the University of Nottingham (http://www.

nottingham.ac.uk/pazjac/sbddpapertext. html).

422. E. G. Wright. Mod. Drug Discov., Oct., 69–70 (2000).

423. E. Ledbetter. Mod. Drug Discov., Apr. and May, 25–28 and 81–84 (2000).

424. M. Jacobs (Ed.). Challenging times for patent ofﬁ ce, Chem. Eng. News, Apr. 10, 39–44

(2000).

425. M. Jacobs (Ed.). PTO ﬁ nalizes guidelines for gene patents (http://www.uspto.gov/.) Chem.

Eng. News, Jan. 15, 33 (2001).

426. D. J. Hanson. Chem. Eng. News, Sept. 11, 19–20 (2000).

427. M. D. Kaminski. Mod. Drug Discov., Jan., 36–37 (2001).

428. W. Schulz. Chem. Eng. News, Nov. 27, 18–19 (2000).

429. R. Wedin. Chemistry (ACS newsletter), Spring, 1, 17–20 (2001).

430. International Human Genome Sequencing Consortium (representing more than 100 contrib-

uting authors). Initial sequencing and analysis of the human genome, Nature 409, 860–921

(2001).

431. Human Genome Research Team (over 200 participating authors represented largely from Celera

Genomics). The sequence of the human genome, Science 291, 1304–1351 (2001).

432. H. van de Waterbeemd, R. E. Carter, G. Grassy, H. Kubinyi, Y. C. Martin, M. S. Tute, P. Willett.

Annu. Rep. Med. Chem. 33, 397–409 (1998).

433. K. J. Watkins. Chem. Eng. News, Feb. 19, 29–45 (2001).

434. T. Agres. Drug Discov. Develop., Feb. 9, 10 (2001).

435. C.-G. Wermuth, R. Ganellin, P. Lindberg, L. A. Mitscher. Annu. Rep. Med. Chem. 33, 385–395

(1998).

436. M. Beggs. Drug Discov. World 2(1), 25–30 (2000).

437. http://dietary-supplements.info.nih.gov/.

438. S. Borman. Chem. Eng. News, July 31, 31–37 (2000).

439. K. A. Bachmann. Drug Metabolism: Databases and High-Throughput Testing During Drug

Design and Development, P. W. Erhardt (Ed.), pp. 79–121, IUPAC/Blackwell, Geneva (1999).

440. W. Klosgen and J. Zytko (Eds.). Handbook of Data Mining and Knowledge Discovery, Oxford

University Press, New York (2000).

441. A. M. Doherty. Annu. Rep. Med. Chem. 35, 331–356 (2000).

442. D. Hanson. Chem. Eng. News, June 19, 37 (2000).

443. M. Jacobs (Ed.). Chem. Eng. News, Apr. 3, 15 (2000).

444. G. Karet. Drug Discov. Develop., Nov./Dec., 71–74 (1999).

445. M. S. Lesney. Mod. Drug Discov., May, 31–32 (2000).

446. M. B. Brennan. Chem. Eng. News, June 5, 63–73 (2000).

447. F. Golden. Time 157(2), 74–76 (2001).

448. A. Thayer. Chem. Eng. News, Feb. 5, 18 (2001).

449. Clinical trial information: Royal Coll. Physicians of Edinburgh at www.rcpe.ac.uk/cochrane/

intro.html; Pharmaceutical Information Network at htt://pharminfo.com/conference/clintrial/

fda_1.html; PhRMA (Assoc. of 100 major pharmaceutical comp.) at www.pharma.org/charts/

approval.html; Univ. Pittsburgh IRB manual at www.ofreshs.upmc.edu/irb/preface.htm; Center

Watch (a comprehensive venue and listing for clinical trials) at www.centerwatch.com; FDA

Ctr. For Drug Evaluation and Res. (CDER) at www.fda.gov/cder; eOrange Book (Approved

drug products with therapeutic equivalence evaluations) at www.fda.gov/cder/ob.; FDA CDER

Handbook at www.fda.gov/cder/handbook/index.htm; FDA CDER Fact Book 1997 at www.fda.

gov/ cder/reports/cderfact.pdf; FDA Consumer Magazine at www.fda.gov/fdac/fdacindex.html;

and FDA Guidelines on content and format for INDs at www.fda.gov/cder/guidance/clin2.pdf.

450. M. Kansy, F. Senner, K. Gubernator. J. Med. Chem. 41, 1007–1010 (1998).

451. C. J. Cole and W. P. Pfund. Mod. Drug Discov., Oct., 73–74 (2000).

452. T. T. Wilbury. Drug Metabolism: Databases and High-Throughput Testing During Drug Design

and Development, P. W. Erhardt (Ed.), pp. 208–222, IUPAC/Blackwell, Geneva (1999).

453. F. Darvas, S. Marokhazi, P. Kormos, G. Kulkarni, H. Kalasz, A. Papp. In Drug Metabolism:

Databases and High-Throughput Testing During Drug Design and Development, P. W. Erhardt

(Ed.), pp. 237–270, IUPAC/Blackwell, Geneva (1999).

454. G. Klopman, M. Tu. In Drug Metabolism: Databases and High-Throughput Testing During

Drug Design and Development, P. W. Erhardt (Ed.), pp. 271–276 IUPAC/Blackwell, Geneva

(1999).

455. R.W. Snyder and G. Grethe. In Drug Metabolism: Databases and High-Throughput Testing

During Drug Design and Development, P. W. Erhardt (Ed.), pp 277–280, IUPAC/Blackwell,

Geneva (1999).

456. J. Hayward. In Drug Metabolism: Databases and High-Throughput Testing During Drug

Design and Development, P. W. Erhardt (Ed.), pp. 281–288, IUPAC/Blackwell, Geneva

(1999).

457. D. F. Lewis and B. G. Lake. Xenobiotica 26, 723–753 (1996).

458. D. F. Lewis and B. G. Lake. Xenobiotica 25, 585–598 (1995).

459. D. F. Lewis, C. Ioannides, D. V. Parke. Biochem. Pharm. 50, 619–625 (1995).

460. T. Walle. Methods Enzymol. 272, 145–151 (1996).

REFERENCES AND NOTES 101

102 MEDICINAL CHEMISTRY IN THE NEW MILLENNIUM

461. M. J. de Grout, G. J. Buloo, K. T. Hansen, N. P. Vermeulen. Drug Metab. Dispos. 23, 667–669

(1995).

462. M. J. de Groot, N. P. Vermeulen, J. D. Kramer, F. A. van Acker, G. M. Donne-Op den Kelder.

Chem. Res. Toxicol. 9, 1079–1091 (1996).

463. D. F. Lewis, P. J. Eddershaw, P. S. Goldfarb, M. H. Tarbit. Xenobiotica 26, 1067–1086 (1996).

464. K. W. DiBiasio. Univ. Micro. Int. (University of California–Davis), 1–5 (1989).

465. K. Bachmann, D. Pardoe, D. White. Environ. Health Perspect. 104, 400–407 (1996).

466. J. G. Sarver, D. White, P. Erhardt, K. Bachmann. Environ. Health Perspect. 105, 1204–1209

(1997).

467. B. J. Kuypu (Ed.). CIIT Activities (Chemical Industry Institute of Toxicology newsletter)

20(11–12), 1–2 (2000).

468. C. J. Henry and J. S. Bus. CIIT Activities (Chemical Industry Institute of Toxicology newsletter)

20(7), 1–5 (2000).

469. M. J. Felton. Mod. Drug Discov., Oct., 81–83 (2000).

470. C. Hogue. Chem. Eng. News, Mar. 19, 33–34 (2001).

471. P. Lewalle. Terminol. Standard. Harmon. 11(1–4), 1–28 (1999).

472. J. H. Duffus. Chem. Int. 23(2), 34–39 (2001).

473. G. Klopman. Quant. Struct.–Act. Relat. 11, 176–184 (1992).

474. D. Sanderson, C. Earnshaw, P. Judson. Hum. Exp. Toxicol. 10, 261–273 (1991).

475. S. C. Gad. SciVision Update (SciVision newsletter), Jan., 1–2 (2001).

476. R. E. Neft and S. Farr. Drug Discov. World, Feb., 33–34 (2000).

477. W. Cabri and R. Di Fabio (Eds.). From Bench to Market, Oxford University Press, New York

(2000).

478. K. Knapman. Modern Drug Discov., June, 75–77 (2000).

479. S. Hume. Innovations Pharm. Technol. 6, 108–111 (2000).

480. J. Court and N. Oughton. Innovations Pharm. Technol. 6, 113–117 (2000).

481. M.D. Lemonick. Time, Jan. 15, 58–67 (2001).

482. M. Zall. Mod. Drug Discov., Mar., 37–42 (2001).

483. M. B. Brennan. Chem. Eng. News, Feb. 14, 81–84 (2000).

484. M. S. Reisch. Chem. Eng. News, Feb. 12, 15–17 (2001).

485. T. Geissman and D. Crout (Eds.). Organic Chemistry of Secondary Plant Metabolism,

Freeman, San Francisco (1969).

486. R. Herbert (Ed.). The Biosynthesis of Secondary Metabolites, Chapman & Hall, New York

(1989).

487. T. Murashige and F. Skoog. Physiol. Plant 15, 473–497 (1962).

488. P. Erhardt and W. Klis. U.S. patent allowed.

489. P. Erhardt. U.S. Patent pending.

490. P. Erhardt, J. Sarver, W. Klis. U.S. patent pending.

491. P. Erhardt, W. Klis, J. Sarver. U.S. patent pending.

492. P. W. Erhardt, R. J. Gorczynski, W. G. Anderson. J. Med. Chem. 22, 907–911 (1979).

493. P. W. Erhardt, A. A. Hagedorn III, M. Sabio. Mol. Pharmacol. 33, 1–13 (1988).

494. P. W. Erhardt. Second generation phosphodiesterase inhibitors: Structure–activity relationships

and receptor models, in Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug

Action, J. Beavo and M. Housley (Eds.), Wiley, New York (1990).

495. J. C. Hemple, W. A. Ghoul, J.-M. Wurtz, A. T. Hagler. In Peptides: Chemistry, Structure and

Biology, J. E. River and G. R. Marshall (Eds.), pp. 279–284, ESCOM, Leiden, The Netherlands

(1990).

103

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

CONTEMPORARY DRUG DISCOVERY

LESTER A. MITSCHER AND APURBA DUTTA

Kansas University

Lawrence, Kansas

3.1 INTRODUCTION

It is unlikely to come as a startling revelation that the challenges faced by chemists in-

volved in drug seeking are signiﬁ cantly different from those faced by chemists occupied

in drug development even though they both work toward the same ultimate objective. The

reasons for this are, perhaps, not fully realized by those occupying either one of the two

camps. Primarily, medicinal chemists ﬁ nd their route to success, with all of its subthemes,

signiﬁ cantly constrained, but the ﬁ nal product is not, whereas process chemists ﬁ nd their

ﬁ nal product highly constrained, but the route to it is signiﬁ cantly variable. In this essay we

attempt to illustrate the challenges facing the medicinal chemist that lead to this aphorism.

At its core, every branch of chemistry involves achieving an understanding of the rela-

tionship between chemical structure and molecular properties. At its zenith, this involves

the ability to design a new substance intended to ﬁ ll a perceived need and, following prepa-

ration and evaluation, getting it right. To design and construct a molecule to possess, say, a

particular color is complex enough, but it is dramatically simpler than the problems involved

in designing and constructing a novel therapeutic agent. Many more characteristics must be

introduced into a molecule before it can become a drug, and many of those are often found

to be in opposition, so that one must make a series of appropriate structural compromises.

3.1.1 Getting Started

The drug-seeking chemist seeks a molecule with a sufﬁ cient number of favorable proper-

ties to justify the very major investment in time and treasure required to bring the drug

before the public and return a proﬁ t. The fundamental complexity of the process stems

from the Byzantine complexity of the human body and its processes. Whereas enormous

104 CONTEMPORARY DRUG DISCOVERY

strides have been made in unraveling these in the last half-century, our understanding of

these is still in a comparatively primitive state. The average time required between concep-

tion and introduction of a drug approaches 12 years and the cost in 2002 (factoring in the

failed initiatives) lies somewhere between $300 million and $800 million. Only then can it

begin to amortize the cost of its discovery and earn a return on investment. These breath-

catching costs stem in large part from the high failure rate and signiﬁ cant constraints put

upon the prices that society will accept. It also clearly explains why no ﬁ rm will make such

an investment unless a proprietary patent position can be achieved and the disease target is

sufﬁ ciently prevalent among persons capable of paying for their treatment.

At the outset of the project, the speciﬁ c identity of the compound sought is almost

always obscure. Indeed, ﬁ nding a suitable starting molecule is often the most challenging

feature of the search. As the project proceeds and the manifold barriers to success are pro-

gressively overcome by creative analoging, the ﬁ nal structure comes more and more into

focus and the molecular options become fewer. During the bulk of this time, the particular

chemistry involved in the route is not a major concern. It is necessary only that the route

be amenable to the reasonably efﬁ cient production of the necessary analogs. It is also im-

portant in the later stages that the chemistry be suitable for the production of the quantities

required for intensive animal studies. The cost of goods is not trivial but is not yet a major

consideration. The synthetic steps are usually somewhat optimized, but much room is pres-

ent for improvement and many of the steps will involve reagents and processes that cannot

safely be translated into large-scale production when this becomes a relevant consideration.

The contemporary position is that ease of economic synthesis is increasingly being consid-

ered at earlier stages, but this is not always possible. In any case, handing the process over

from discovery to development at the optimal stage requires judgement.

After the project is assigned to development, the structure of the molecule to be pro-

duced is quite clear and will not be changed casually. Hundreds to thousands of chemicals

will have been prepared, evaluated, and discarded by this time. The challenge now is to

develop a synthetic route to the surviving molecule that is economically attractive, scal-

able, uses safe chemicals and processes, does not require chromatography, produces benign

wastes, and in which each step is optimized. It is not surprising therefore that the discovery

route is rarely the development route and that success in both stages requires a high level

of expertise considering the constraints imposed.

3.2 CHARACTERISTICS OF A SUITABLE LEAD SUBSTANCE

There is a partially resolved debate among medicinal chemists as to which property is more

important at the outset of a drug-seeking campaign: potency and selectivity or suitable

druglike characteristics. Ultimately, as shown in Fig.3.1, both are required for success. The

∆

Potency /

Selectivity

Absorption,

Distribution,

Metabolism,

Excretion etc.

Drug

Figure 3.1 Successful drug seeking requires a satisfactory balance between efﬁ cacy and other

drug-like characteristics.

real question is which one to start with if one can have only one of the two. Most discovery

chemists now display a bias toward druglike characteristics in lead molecules, as many

more variables are involved and many more macromolecular ﬁ ts must be optimized in solv-

ing these than are involved with enhancing potency and selectivity.

3.2.1 Potency and Selectivity

A drug and its receptor mostly possess exquisitely demanding mutual molecular compat-

ibilities, leading to a tight host–guest complex possessing signaling properties for cells

that neither the drug nor the receptor alone possesses. Similarly, an inhibitor must bind to

an enzyme and inhibit its functioning. The difﬁ culty in accomplishing this stems from the

large and complex structure of the receptor or the enzyme, whose detailed composition and

topography are generally unknown at the outset of the project.

The original drug–receptor theory was that they were both rigid substances, so that the

drug needed to be crafted to ﬁ t the desired receptor and no other, just as a speciﬁ c key is

required to open a speciﬁ c lock. This simple picture is now recognized to be only one ex-

treme of a continuum. The opposite of this concept is a ﬂ exible drug and a ﬂ exible receptor.

In this view both interacting molecules must adjust in shape to accommodate each other.

This is analogous to a zipper. Following an initial docking interaction, the two components

interact progressively and ultimately form a new supermolecule with new properties. All

intermediate degrees of ﬂ exibility between these extremes are recognized (i.e., rigid drug

and ﬂ exible receptor, ﬂ exible drug and rigid receptor, etc.). An example can help clarify

these concepts for those who ﬁ nd these ideas to be new.

Methotrexate is an antitumor drug classiﬁ ed as an antimetabolite in that it inhibits the

formation of an essential component of DNA. DNA is the central biochemical from which

all other molecules ﬂ ow. It is made up of four essential monomers—adenine, cytosine,

guanine, and thymine—joined together in a linear manner in speciﬁ c sequences as de-

oxyribosyl phosphodiesters. Of these, thymine is unique to DNA, not being present in

RNA. Instead, it is biosynthesized by addition of a methylene group to deoxyuridine 5'-

monophosphate in a reaction catalyzed by thymidylate synthase, an enzyme using tetrahy-

drofolate as a source of the needed carbon. Cellular growth, repair, and reproduction are

impossible without this reaction. One way to prevent cancer cells from growing is to poison

them selectively by starving them of thymine by inhibiting this reaction. Methotrexate acts

as an anticancer agent in just this way as it competes with the cofactor for binding to the

enzyme. It was found following up a natural product lead and its molecular mode of action,

not known at the time, is now generally understood.

Dihydrofolate reductase alters its conformation slightly but importantly when it is

liganded with its cofactor or its inhibitors. This complicates the rational design of inhibitors,

for crafting a ﬁ t to the ground state that did not trigger this movement would produce

molecules that would probably be ineffective. Fortunately, none of this was known at the

time that methotrexate was developed.

Later, numerous x-ray studies with the normal substrate and its inhibitors, including

methotrexate, became available so it was possible not only to understand generally how this

inhibition works but to design novel agents using this information. From Scheme 3.1, which

illustrates schematically mutual interactions between the enzyme and methotrexate as well

as dihydrofolic acid, it may be seen that many distinct molecular interactions are involved.

Prime among these is the docking interaction between one of the glutamate carboxyls and a

conserved arginine of the enzyme. For the inhibitor to be potent and speciﬁ c, the remaining

interactions must be not only electronically but also sterically correct. It is important to note

CHARACTERISTICS OF A SUITABLE LEAD SUBSTANCE 105

106 CONTEMPORARY DRUG DISCOVERY

that the natural substrate, dihydrofolic acid, although structurally similar to methotrexate,

actually binds rather differently to the enzyme in that the pteridine rings are rotated with

respect to one another! This is not common in drug seeking and could have been a big

problem. More commonly, one starts with the structure of the normal substrate and designs

inhibitors following the belief that they will bind in the same manner. As is common in

enzyme–substrate interactions, the binding site for dihydrofolic acid and for methotrexate

is in a groove in the enzyme’s structure that is accessible only in speciﬁ c ways from the

exterior of the molecule. One might visualize it as a frankfurter in a bun. Most important,

methotrexate bound in the enzyme’s active site is not capable of providing the one-carbon

unit needed for thymine biosynthesis and also blocks dihydrofolic acid’s access.

Having a reliable picture of how a drug interacts with its target is a powerful aid to drug

seeking. This particular mode is known as structure-based drug design. Fortunately, it is

becoming more common. One notes the complexities of the task even with this information

in hand. Designing from ﬁ rst principles a successful alternative ligand for a pair that has

been optimized for each other evolutionarily is a tall order. The job is even more complex

when the alternative ligand binds to the same site but in a different way!

Because of the difﬁ culty of representing complex three-dimensional relationships in

two dimensions, for clarity of exposition the steric aspects of the relationship illustrated in

Scheme 3.1 have been simpliﬁ ed by drawing the molecules as though they were coplanar.

This is not actually the case. The central para-aminobenzoate moiety, for example, actually

projects upward toward the viewer’s eyes.

This example serves as one of many indicating that ﬁ nding potent and selective ligands

is not easy. It is, perhaps, comforting to note that it was originally accomplished empiri-

cally without knowing the molecular details presented in Scheme 3.1. It is discomforting

to note that a ligand that is successful in vitro all too often fails as a drug because other

factors prevent it from reaching its target. Getting a good ﬁ t is only part of the job. This

reemphasizes the thought conveyed by Fig. 3.1.

The parts of a drug such as methotrexate that are essential for its pharmacological action

are those in direct contact with its target macromolecule and are known as the pharmaco-

phore. The other functionalities of a pharmaceutical are less essential and are manipulated

Arg

Tr p

Asp

Thr

Leu

HN O

Asp

Ile

Tr p

Ile

Thr

Ile

Tr p

Ile

Scheme 3.1 Dihydrofolic acid (A) and methotrexate (B) in the active site of human dihy-

drofolate reductase. [From Klebe, G. (1994), J. Mol. Biol. 237: 212 – 235.]

in order to deal with pharmacokinetic deﬁ ciencies to be discussed later. These might be

termed auxophores.

The point of this exposition is that a successful inhibitor must possess suitable function-

ality and stereochemistry to ﬁ t tightly into the critical region of the receptor but not func-

tion in the same manner as the normal ligand. Whereas nature solves the problem through

evolutionary processes, a chemist must prepare such an agent through analoging, often

without knowing all of the signiﬁ cant details and in a much, much shorter time period.

The medicinal chemist must be familiar with the many ways in which successes have been

achieved in the past and recognize when these may be applied productively to a present

problem and be prepared to devise novel approaches as well.

3.2.2 Structure–Activity Relationships

Molecules are altered progressively, guided by biological data so that the desirability of

the substances is dramatically enhanced as one after another of the perceived deﬁ ciencies

is overcome. The structural and biological scorecard kept during this process is known as

a structure–activity relationship (SAR). The simplest of these involve lists of structures

and the comparative potency associated with each. From such lists the chemist can rapidly

sort out which molecular features are helpful and which are not. Multiple SARs involve

analogous data associated with serum protein binding, achievable blood levels following

oral administration, and so on.

3.2.3 Toxicity

The reactivity of the functional groups in the molecule should not be great. For example,

drugs containing α-haloketone moieties or other reactive electrophilic groups react rela-

tively indiscriminately with tissue macromolecules, producing covalently bound products.

At best these are sites of loss of molecular identity. Highly reactive molecules often fail to

reach their drug targets. At worst, they can result in drug allergy or overt toxicity, even mu-

tagenicity. These untoward effects are rarely tolerable. Stability of the agent is also related

to reactivity. The drug must be stable enough to allow for preparation of dosage forms that

will survive in active form to reach the patient and be administered.

The importance of lack of toxicity is obvious. Acute toxicity is the amount of a given

agent that will kill or injure following a single or a few doses. The bigger the separation

between the effective dose and the toxic dose (the therapeutic index), the safer the agent

will be to use. A minimum separation of 1000-fold or higher is a common target. Chronic

toxicity, on the other hand, refers to the dose that will produce deleterious effects when the

drug is administered for a signiﬁ cant number of doses. The chronically deleterious dose is

almost always signiﬁ cantly smaller than the acutely toxic dose.

Certain functional groups are recognized to produce toxicities all too often and are

therefore put into drug molecules only with great reservation. These groups are often called

toxicophores. Flat three-ring-containing aromatic moieties intercalate into DNA and inter-

fere with its function. Usually, this is undesirable. Aromatic nitro groups are reduced to a

variety of intermediately oxygenated species such as hydroxylamines and nitroxides that

are capable of self-condensation into toxic moieties. Similarly, aniline groups are capable

of oxidation into the same sort of undesirable metabolites. Thiols are capable of conden-

sation with sulfhydryl/disulﬁ de-containing tissue constituents, leading to toxic problems.

Many other toxicophores are well known to medicinal chemists and are avoided.

CHARACTERISTICS OF A SUITABLE LEAD SUBSTANCE 107

108 CONTEMPORARY DRUG DISCOVERY

Genetic toxicity is the most dangerous kind. DNA is a self-replicating polymer, and a

damaged DNA molecule can be copied into multiple altered copies, magnifying and pre-

serving the initial injury. Since damaged DNA, if not repaired promptly and accurately,

leads to mutations that can even lead to cancer, there is really no consensually safe dose of

a mutagen. Consequently, mutagenicity is a serious liability. Not all mutations are associ-

ated with cancer, but almost all cancers involve mutagenicity. If mutagenicity is relatively

molecule speciﬁ c, it can often be engineered out of a structure by appropriate analoging. If

it is series related, however, this is a showstopper and another lead series must be sought.

Known mutagenic moieties are therefore avoided at the outset.

3.2.4 Changing Appellation of the Best in Series: Analog Attrition

There are some key deﬁ nitions that are associated with molecules as the hunt progresses.

A hit is a substance that produces a signiﬁ cant response in a screen designed to reveal

promising substances. Its potency and selectivity might be quite low, but it should have a

promising structure for further elaboration. High-throughput screens enable one to search

rapidly through large collections of molecules (often called compound libraries). Usu-

ally, high-throughput screens are based on particular biological mechanisms, are cell-free,

sample sparing, and have relatively modest information content. A lead is a more important

molecule most often descended from a hit that produces promising activity in a whole cell,

intact organ system, or whole animal, conﬁ rming the relevance of the activity found in the

initial screen. Much more compound is needed to reveal this, the criteria for acceptance

are much higher, and the information content is much higher. The number of compounds

that advance from screening substances to hits to leads is an order of magnitude less at

each stage. A candidate drug is one of the few that has graduated from a lead substance

by demonstrating favorable properties in more than one relevant animal species suffering

from a model disease and so becomes a serious contender for clinical evaluation. A clinical

candidate is a molecule that is judged to be worth the major costs involved in evaluation

in humans, both healthy and diseased. A drug is a molecule that has passed successfully

through all of these stages and advances into human use. It is only at this last point that it

begins to return a proﬁ t to justify the expenditures already made.

Figure 3.2 illustrates schematically that the number of compounds one starts with ﬁ lters

down quickly into a very few survivors. The aim of the medicinal chemist is to make the

attrition signiﬁ cantly smaller.

3.3 SOME CRITERIA THAT A HIT MUST SATISFY TO BECOME A DRUG

In addition to satisfying requirements for suitable potency and selectivity, at least 20 addi-

tional characteristics are signiﬁ cant. A drug must possess all of these to a favorable degree,

but they are almost never equally important. Early identiﬁ cation of the most signiﬁ cant

variables hindering further advancement is of prime importance in drug seeking. One can

almost always enhance any one of these characteristics through analoging, but this often

results in a detriment to at least one of the other characteristics. Thus, one is almost always

forced to make compromise choices. For example, if one has outstanding potency, one

may be willing or forced to sacriﬁ ce some to enhance some other feature that must be im-

proved, such as solubility or absorption. Thus, the guiding principle is “good enough, soon

enough.” Perfection is rarely achieved. One interesting consequence of this multiplicity of

choices is that two different groups of investigators starting with the same lead molecule

will usually ﬁ nish with different ﬁ nal drugs.

The steps along the way are usually listed as though they are traversed linearly. All too

often, however, the medicinal chemist becomes stymied at a given point and has to back up

and bring a different substance forward to make additional progress. Whenever possible, to

save time, the problems are identiﬁ ed and attacked in parallel rather than sequentially.

Some of the properties that must be present or introduced into a drug substance are:

3.3.1 Level of Potency

The potency of the substance must be signiﬁ cant in order to identify early hits worth pur-

suing. Since there are comparatively few copies of a given receptor in a cell, the potency

needed for a hit is usually on the order of 100 µmol/mL or less. The potency of a candidate

drug is often in the range of 10 nmol/mL or less. This is somewhat related to speciﬁ city. It

is generally recognized that any effect of a drug other than that for which it is administered

represents a side effect and is therefore undesirable. The more potent a drug is, the more

Compound

Libraries

oooooooooooooooo

“Hits”

oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo

ooooo

oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo

Lead

ooo

Candidate

Drug

Clinical

Candidate

Drug

> 10

–10

10–10

< 10

Figure 3.2 The drug discovery pyramid.

Potency Efﬁ cacy Selectivity

Toxicity Absorption Distribution

Metabolism Excretion Acute toxicity

Chronic toxicity Mutagenicity Stability

Accessibility Cost Patentability

Clinical efﬁ cacy Water solubility Taste

Formulability Idiosyncratic problem(s)

SOME CRITERIA THAT A HIT MUST SATISFY TO BECOME A DRUG 109

110 CONTEMPORARY DRUG DISCOVERY

tightly and speciﬁ cally it ﬁ ts its receptor, and thus it is expected that it will not bind equally

well to extraneous receptors and so will have fewer side effects. As with most generaliza-

tions, this one has signiﬁ cant exceptions, but it nonetheless colors one’s attitude in drug

seeking. Thus, potency and selectivity are often considered as closely related.

3.3.2 Comparison of Potency and Efﬁ cacy

There is sometimes confusion about the difference between potency and efﬁ cacy. Potency

is the concentration of an agent that must be present to produce a biological result. Efﬁ cacy

is how efﬁ cient a given dose is in producing a given result. Two substances possessing equal

potency (in this case, producing their half-maximal effect at a similar dose) may produce

signiﬁ cantly different efﬁ cacies (produce much different total responses, regardless of the

dose). In Fig. 3.3, compounds A and B are similary potent and both are much more potent

than C. Compounds A and C are equally efﬁ cacious, but compound B is signiﬁ cantly less

so. Unfortunately, much more than potency, efﬁ cacy, and selectivity is required. Otherwise,

drug seeking would be a much simpler task than it is in reality.

3.3.3 Druglike Character

Pharmacodynamics is the study of the effect that a drug has on the body (e.g., lowering

blood pressure); pharmacokinetics is the study of the effect that the body has on the drug

(e.g., metabolism). Contemporarily, it is recognized that the pharmacokinetic (PK) proper-

ties of a substance are extremely important but are complex to produce at satisfactory levels

in a lead substance. Compound collections to be screened are most valuable, in this view, if

they contain many molecules that are likely to possess satisfactory pharmacokinetic prop-

erties at the outset, that is, are intrinsically druglike. These properties are often referred to

by the acronym ADME (absorption, distribution, metabolism, and excretion). Getting these

right in a lead series as soon as possible avoids wasting resources on molecules that have

little to no chance of becoming drugs.

3.3.4 Efﬁ cacy Following Oral Administration

One of the most desired characteristics of a drug is its ability to deliver therapeutically use-

ful concentrations in the body following oral administration. Although this property can of-

ten be manipulated favorably by suitable analoging, it is a complex undertaking. The drug

substance must ﬁ rst dissolve and escape digestion into polar small pieces. Next, it must

pass through the cells lining the gastrointestinal tract and pass into the blood. A number of

Percent of

maximal

response

Log of drug concentration

Figure 3.3 A comparison of potency and efﬁ cacy for drugs A, B, and C.

competing processes are involved, but most drugs do so by passive diffusion. Since the cell

membranes are lipoidal in nature, most drugs must be at least partially nonionized at the

cell surface for passive diffusion through this barrier. Since the small intestine has about

the surface of a tennis court, due to its highly convoluted nature, this is the body chamber

where the bulk of food and drug absorption takes place. Since the normal pH of the up-

per small intestine ranges from 4 to 7, nonpolar neutral compounds and compounds only

slightly ionized at this pH range (e.g., weak acids and unprotonated tertiary amines) are

absorbed well by passive diffusion. Thus, a drug’s pK

value should usually be less than 5,

so that a sufﬁ cient percentage is nonionized. In contrast, if it is too lipophilic, it will remain

in the lipid bilayer and not pass through to the other side. If it is too water soluble, it will

not penetrate at all (see Fig. 3.4). Small water-soluble compounds can enter cells through

active

transport

porin

passage

passive

uptake

lipophillic

passage

paracellular passage

Tight CNS

junction

Comparatively

loose paracellular

junction

inhibitor

drug

SCHEMATIC OF MULTIDRUG EXPORTER AND ITS INHIBITION

Figure 3.4 Drug uptake and exclusion at cells possessing a lipid bilayer. Passive diffusion

requires lack of ionization and dehydration. Small hydrophilic molecules can pass between

cells in the periphery but not in the CNS where the capillary cells are more tightly assopciated

with their neighbors. More hydrophilic drugs can pass through proteins with a water-lined in-

terior passage (porins) that span the membrane or by active transport using protein receptors.

The active transport is very speciﬁ c in structural requirements and can be in or out.

SOME CRITERIA THAT A HIT MUST SATISFY TO BECOME A DRUG 111

112 CONTEMPORARY DRUG DISCOVERY

porins (i.e., membrance-spanning proteins with a central water-lined well) or may even

pass between cells (paracellular passage), avoiding their membrance barriers altogether.

There are also protein-based selective uptake mechanisms in the membranes providing ac-

tive transport for some polar materials (e.g., amino acids, glucose), but this does not often

work for drugs. Very precise host–guest compatibilities must be present for active transport

to work. These active transporters can also work in the opposite direction to expel drugs

from cells. These processes prevent many drugs from reaching their cellular targets and

play a very signiﬁ cant role in resistance to antibiotic and anticancer chemotherapy.

To avoid excess polarity, molecules not only must not be too ionized but also must be

able, reversibly, to shed their adhering water. As a general matter, molecules should have

fewer than ﬁ ve hydrogen-bond donating moieties and fewer than 10 hydrogen-bond accept-

ing groups for this to be efﬁ cient. More than these and it will be too difﬁ cult to desolvate.

Peripheral cells are not tightly bound together, often leaving small spaces between cells

through which small hydrophilic molecules (such as ethyl alcohol) can pass. On the other

hand, the capillaries of the central nervous system are so tightly packed together that this

does not take place. Consequently, drugs intended to act in the central nervous system

(CNS) must be more lipophilic than drugs intended to act elsewhere in the body. Con-

versely, drugs to be excluded from the CNS should be more hydrophilic.

Also, the inﬂ uence of molecular weight on oral absorption is signiﬁ cant. Within reason-

able limits, molecules exceeding about 500 atomic mass units are often poorly absorbed in

the intestine. This is largely accounted for by the fact that passive diffusion becomes much

slower as molecular weight increases.

The overall shape of the substance is also of signiﬁ cance. Compact, rigid molecules

generally are absorbed more efﬁ ciently than ﬂ oppy molecules with signiﬁ cant cross-

sectional areas.

The combined inﬂ uence of all these factors lead to the conclusion that the physico-

chemical characteristics of a substance are in many ways more important than its speciﬁ c

chemical structure if it is to beneﬁ t from passive uptake. The medicinal chemist must learn

how to design molecules with favorable absorption characteristics.

This brief exposition avoids mention of ion channels that allow certain ions to penetrate

into and escape from cells against concentration gradients. This important topic will have

to be covered at another time in another place.

3.3.5 Lipinski Rules for Oral Absorption

The Lipinski rules were clearly enunciated and given scalar dimensions by Christopher

Lipinski and are widely known as the Lipinski rules of ﬁ ve. They were developed by com-

paring the properties of a large number of candidate drugs to their relative absorption fol-

lowing oral administration. Although there are many exceptions, particularly for drugs

taken up by active transport mechanisms for which the rules do not apply, the Lipinski

rules are a very useful. Their popularity stems from their simplicity, their accord with real-

ity, and the fact that they can readily be compreheneded by chemists and applied through

inspection of structural formulas in advance of synthesis. The rule of ﬁ ve is recognized as

being less useful for hits and leads than for the ﬁ nal drugs from which they were derived.

To improve the properties of these classes of agents it is usually necessary to introduce ad-

ditional functionality. Almost inevitably this results in molecular weight and grease creep.

If the molecular weight of a lead compound is too high and/or the molecule is quite lipo-

philic to start with, it is difﬁ cult to produce a ﬁ nal molecule with a good chance of being

orally active following further analoging. To accommodate this, chemists now often refer

to a rule of threes, instead, for hits and leads.

It is also well known that compact molecules are more efﬁ ciently absorbed than those

with a big cross section. Flexible molecules with many rotatable bonds are among the bad

actors. Daniel Veber has pointed out that molecules with no more than 10 rotatable bonds

are more likely to be orally active, and the Veber rule is often added to the Lipinski rules

in drug seeking.

3.3.6 Injectable Medications

When an illness is acute or a patient is unable to achieve beneﬁ t from oral absorption, the

drug is injected into the bloodstream or into tissues. For these purposes, signiﬁ cant water

solubility is required so that the volume administered does not itself create problems for

the patient. In this form of administration, the drug is rapidly available for the needs of the

various tissues, and the Lipinski rules are not relevant for these.

3.3.7 Distribution

Distribution refers to the ability of an agent to arrive at the source of pathology in the body

in concentrations sufﬁ cient to produce a favorable response. Many potentially intervening

molecular interactions play a role in this process, as illustrated in Fig. 3.5. Some of these

have just been discussed in dealing with oral uptake. However, there are many more not

yet mentioned. Ability to deliver an agent to a particular target organ by analoging is not

yet highly developed. The ﬁ rst organ that receives a drug-following uptake through the

gut wall and passage into the blood is the liver. The liver is a sort of metabolic factory that

transforms lipoidal materials into more polar molecules that are more readily transported

in the blood and excreted either back into the intestines or distributed throughout the body,

followed by passage into the urine via the kidneys. The liver in this sense is seen as a kind

of grease trap. These concepts are illustrated schematically in Fig. 3.5.

Dose

Solubility,

Taste,

Smell, etc.

Stability Permiability

Metabolism

Blood

Gut Wall

Blood

Excretion

Liver

Metabolism

Protein

Binding

Tissue

Distribution

Toxicity/

Side-Effects

Figure 3.5 Some of the successive hurdles that a lead must overcome before marketing

would be possible. In this diagram potency is taken for granted.

SOME CRITERIA THAT A HIT MUST SATISFY TO BECOME A DRUG 113

114 CONTEMPORARY DRUG DISCOVERY

3.3.8 Serum Protein Binding

After a drug enters the bloodstream, it is subject to a series of competing equilibriums. The

ﬁ rst of these involves binding to serum proteins. One of the major normal physiological

functions of serum protein binding is to carry fatty acids safely to the tissues. Otherwise,

these surface-active molecules would cause hemolysis of red blood corpuscles, destroying

their ability to function in their normal role in carrying gases to and from tissues. Acidic,

lipophilic molecules are particularly subject to this phenomenon, although several other

drug-binding systems are also found in the blood. The percentage of the drug that is bound

to serum proteins is inhibited from entering into the other equilibriums that otherwise

would come into play. The capacity of serum protein albumen to bind drugs is rather high

compared to the normal doses administered, so it can be a major factor in drug perfor-

mance. If the drug is highly protein bound and released slowly, it is not readily available

to other tissues. This does protect it from liver metabolism but also makes it difﬁ cult for it

to reach the target tissues where its receptors or enzyme target lies and so interferes with

its action. If it is minimally protein bound and readily released, it can freely reach tissue

enzymes and receptors, so serum protein binding is less consequential. For these reasons

the medicinal chemist avoids making lipophilic acids if possible.

3.3.9 Metabolism

Drug metabolism can take place in most tissues of the body, including the gut wall, but

the liver is the main site of metabolism. The liver is the ﬁ rst organ to receive blood-

carrying absorbed materials from the intestines. Metabolism is a complex process often

strongly affecting the performance of a drug. Metabolism can enhance, result in an indiffer-

ent effect, or be deleterious to drug action. In many cases, enzymes alter the structure of ad-

ministered drugs by converting them to more polar substances by hydrolysis or oxidation.

The modiﬁ ed molecule will distribute differently and have a different ﬁ t to its receptor(s)

than the molecule administered. Metabolism is often benign, but in some cases results in

toxic products and must be prevented or avoided. Medicinal chemists have become skilled

in designing molecules that are protected from metabolic change or in utilizing metabolism

to enhance the desirable properties of a drug. For example, changing a hydrogen atom to

a halogen on an aromatic ring is often effective in preventing oxidative hydroxylation at

that position. Many other metabolic reactions are recognized. Hydrolysis and oxidation are

common metabolic reactions. When metabolism fails to produce sufﬁ cient water solubility,

conjugation with sulfate or glucuronic acid (an acidic sugar) can be employed to enhance

solubility further. When these processes are complete, the drug escapes from the liver and

passes either back into the gut or into the general circulation. Many drug molecules, par-

ticularly those with sufﬁ cient intrinsic water solubility, are not metabolized at all in the

liver and pass rapidly into the general circulation without signiﬁ cant change. The medicinal

chemist must be able to control or modify metabolic processes.

3.3.10 Distribution

Most tissues are richly vascularized (e.g., lung and heart), whereas others are not (e.g., bone).

Thus, even under ideal conditions, organ distribution can be expected to be uneven and to

play a role in the action of a drug. The brain deserves special mention in this regard. Only

quite lipophilic materials can enter the brain by passive diffusion, and the range of molecules

that are taken in actively (e.g., glucose) is quite limited. This special characteristic is referred

to as the blood–brain barrier. Drugs intended for central nervous system delivery generally

must be more lipophilic than drugs designed to perform in other tissues. The lungs, heart,

liver, and kidneys receive comparatively signiﬁ cant concentrations of drug administered

early. Other organs get their drug concentrations in more dilute form and more slowly.

3.3.11 Excretion

Excretion, primarily in the urine, requires signiﬁ cant water solubility to be efﬁ cient. Another

important point to consider is that normal urine is a protein-free ﬁ ltrate. Thus, compounds

that are both extensively and tightly serum protein bound are excreted slowly, resulting in

long residence times in the body. The ability to adjust the excretion rate of a substance is

crucial to ensure that the disease being treated beneﬁ ts from a sufﬁ cient quantity of the

agent being present at the receptor for a sufﬁ cient period of time.

3.3.12 Patenting

For intellectual property rights and freedom to operate, they must be achievable. The drug

industry is a worldwide enterprise, and the costs of a new agent are so high that no drug will

be introduced if there are insufﬁ cient proprietary rights in all signiﬁ cant global markets to

give the investing group a reasonable chance for a satisfying return on their money. Since

a patent requires at a minimum both novelty and utility, the agent must be not only chemi-

cally unprecedented but also efﬁ cacious in an economically signiﬁ cant disease state.

3.3.13 Pharmaceutical Properties

The organoleptic properties of a drug must also be satisfactory. The agent must not have an

unattractive color, a disgusting taste, or a pungent odor. The agent must also be compatible

with the ingredients normally necessary for formulation in the form of tablets, capsules,

injectable solutions, and the like. The agent must not be hygroscopic. It also cannot possess

metastable isomeric crystal habits. This last factor can lead to subsequent alterations in be-

havior due to undesirable changes in crystal form and thus modiﬁ ed solubility properties.

In that case, the entire evaluation process must be redone at a huge expense.

3.3.14 Idiosyncratic Problems

There are other problems of a more idiosyncratic nature (i.e., associated with individual

projects and not occurring generally) with which a medicinal chemist must be prepared to

deal. In every project there are unanticipated difﬁ culties that develop between inception

and introduction. The medicinal chemist must be skillful and resourceful enough to pro-

vide acceptable solutions for these as the need arises.

3.3.15 Summary

Taken together, it can be seen that the synthetic chemist must be able to modulate structures

through thoughtful analoging so as to ﬁ nd ﬁ nal structures with an acceptable collection

of additional properties without losing the efﬁ cacy sought. The development chemist can

readily appreciate that a compound that survives all of this is not discarded lightly. Even if

SOME CRITERIA THAT A HIT MUST SATISFY TO BECOME A DRUG 115

116 CONTEMPORARY DRUG DISCOVERY

the synthesis in use up to this point is complex, dangerous, and costly, these represent chal-

lenges that, fortunately, the development chemist is skilled at resolving without requiring

that the clinical candidate be discarded.

3.4 EXAMPLE OF DRUG DEVELOPMENT THAT ILLUSTRATES

MANY OF THE AFOREMENTIONED CONSIDERATIONS

Unfortunately, space is not available to give speciﬁ c examples of all of these problems and

of the various solutions that have been devised. Nonetheless, it is likely that a general discus-

sion such as the preceding may not be entirely clear to those who are new to these consider-

ations. The methotrexate example given earlier deals mostly with host–guest relationships

rather than ADME and other considerations. Thus, the following is a brief recounting of

a different drug-seeking campaign from which the interested reader can see how many of

those inﬂ uences play out in actual practice. The practical result of the investigations re-

counted brieﬂ y in this section is that more than a dozen commercially successful compounds

have been introduced into human use, and a few of these are billion-dollar-a-year drugs.

3.4.1 Control of Blood Pressure with Drugs

Despite many decades of intensive research and many notable successes, cardiovascular

disease remains one of the major killers of humankind. Unfortunately, the symptoms of hy-

pertension are subtle, and many sufferers are unaware of their problem until signiﬁ cant and

irreversible pathology to vital organs has set in, leading to strokes, heart failure, and kidney

damage. There are three main mechanisms by which blood pressure becomes pathologi-

cally high: increased pumping of blood by the heart, decreasing elasticity of the arteries,

and decrease in kidney ﬁ ltration so that ﬂ uids in the body build up.

3.4.2 Historical Background

A century ago the major therapy available for treating high blood pressure comprised diuret-

ics and related materials (e.g., digitalis) that reduced blood volume. Added to this was nitro-

glycerine. Discovered accidentally as a by-product of explosives manufacture, nitroglycerine

alleviates some of the symptoms of angina pectoralis (chest pain associated with narrowing of

the arteries that supply the heart with its own blood supply) by relaxing certain blood vessels.

Nitroglycerine serves as a source of the gaseous neurohormone nitric oxide. Another useful

substance was quinidine, whose action is to treat cardiac arrhythmia (weak and improperly

synchronized heart beats) by altering nervous conductivity. Quinidine was discovered as the

result of observing a useful side effect in treating patients suffering from malaria.

All of these agents treat symptoms rather than the underlying pathology, so are not

curative. Even half a century ago we had hardly progressed much beyond these means, so

that the causes and treatment of cardiovascular disorders were still comparatively primi-

tive. Since that time a group of remarkably effective agents have been discovered that have

saved a multitude of lives and returned multibillions of dollars to shareholders.

The story below is offered to demonstrate how an incompletely met medical need can be

addressed, utilizing knowledge gained painfully from close observation of natural phenom-

ena. In the right hands, this can be translated through intelligence, diligence, creativeness,

and luck into a medically useful solution by analoging.

3.4.3 Finding a Starting Point: A Clue from Nature

To do something different than previously existed, how should one begin? In this case the

initial clue came from an understanding of a natural phenomenon. The deadly South Amer-

ican pit viper Bothrops jararaca has solved in an interesting way the problem that many

snakes have. That is, if they kill their prey successfully through the injection of venom, it

is important that the prey die quickly so that they cannot run off before dying, as the snake

cannot follow far and it does the pit viper little good to kill something that it cannot eat.

A profound drop in blood pressure following intoxication leads to fainting, if not outright

death. The venom that produces this result solves the snake’s dietary problem promptly

because it can now catch up with its immobilized victim and consume it.

A few insightful scientists understanding these considerations recognized the potential

value of snake venom to control malignant blood pressure if administered carefully to suf-

ferers. To understand how the venom works in reducing blood pressure and how it might be

used for deliberate human therapeutic work, it is important to understand a bit of physiol-

ogy. The entire process is wonderfully complex, and the nonspecialized reader can easily

lose the thread of the narrative if it is explicated in complete detail. Thus, the following is

an abridged version that covers only the main points.

3.4.4 Renin–Angiotensin–Aldosterone System

Maintenance of blood pressure and composition is vital for survival. The volume and initial

contents are provided by intake of water and nutrients into the gestrointestinal tract, from

which they pass into the blood. The heart keeps the blood moving. It supplies oxygen for

energy generation from burning (oxidizing) food by circulation to the lungs, from which at

the same time volatile wastes are removed. The oxygenated blood is returned to the heart

and pumped into the blood vessels. These must distend to hold this volume. They posses

the property of elastic rebound, which process squeezes the blood onward to the tissues.

Backﬂ ow is prevented by a one-way valve in the heart. The body clears its wastes in large

part by excretion in the urine. The kidneys play an essential role in this and function as a

pressure-driven ﬁ lter, leading to excretion of salt and water. Blood volume and content are

regulated by selective reuptake of necessary amounts of these things. Equilibrium between

intake and output is important for the system to work property. If the kidney ﬁ ltration

rate drops too low, the heart can beat more strongly, and if it is too high, it can beat less

powerfully. The blood vessels can alter their diameter to cooperate with this process such

that decreased diameter raises the blood pressure. These considerations are illustrated

schematically in Fig. 3.6.

To regulate these activities, a variety of hormones and enzymes play various roles. One

of these is the enzyme renin, produced and stored in kidney blood vessel wall, from which

it is released as needed when sodium ion concentration or blood pressure falls. A simpliﬁ ed

version of this complex sequence of events is given in Fig. 3.7.

Renin, an aspartoyl peptidase, catalyzes the cleavage of four amino acids from the acid

end of angiotensinogen by water, so producing angiotensin I. This 10-unit peptide is only

weakly active in conrolling blood pressure. A second enzyme, angiotensin-converting

enzyme (ACE), cuts dipeptide (two amino acids) units from the acid ends of a number of

peptides, of which angiotensin I is most relevant to our story. This produces the octapeptide

angiotensin II. This octapeptide hormone acts very powerfully to increase blood pressure

following binding to speciﬁ c receptors. Angiotensin II also triggers the formation of another

EXAMPLE OF DRUG DEVELOPMENT 117

118 CONTEMPORARY DRUG DISCOVERY

DRV

YIH

DRVY

Plasma Angiotensinogen

Renin,

Water

Angiotensin I

Angiotensin Converting Enzyme

(ACE), Water

Angiotensin II - A powerful vasoconstrictor hormone. It stimulates

release of aldosterone which helps retain water and salt in the blood.

Occupies specific vascular receptors.

Enzyme,

Water

Inactivation by hydrolysis.

Figure 3.7 A simpliﬁ ed diagram of the renin-angiotensin system for regulating blood pres-

sure. Renin is released by certain kidney cells when the blood pressure or salt concentration

fall too low. The circles represent individual amino acids and the letters are the standard

one-letter code for the amino acids. The fat arrows indicate where hydrolysis takes place.

Angiotensin-converting enzyme (ACE) is released from the lungs. The letters in the circles

are individual amino acids joined together by peptide bonds. Bradykinin is another peptide

hormone that plays a role. It combats the blood pressure elevating action of angiotensin II.

It is also hydrolyzed with the aid of ACE.

Gastrointestinal

System

Blood

Heart

Lungs

Kidneys

Urine

Blood

Vessels

Tissues

Figure 3.6 Interplay of several organs in the regulation of blood pressure. Blood volume

and content is regulated from intake from the GI system and excretion of ﬂ uid from the kid-

neys. The heart keeps it moving and the lungs add oxygen for the generation of energy. The

kidneys ﬁ lter the blood, removing soluble wastes and passing them into the urine. Water,

salt, etc., are reabsorbed to prevent excessive losses. The whole system is also regulated by

the brain in ways not illustrated here.

contributing hormone, aldosterone. The action of aldosterone is signiﬁ cant in reabsorb-

tion of sodium ions from kidney tubule ﬁ ltrates. This salt reabsorption concomitantly also

produces signiﬁ cant water reuptake, maintaining blood volume and consequently blood

pressure. Fine control of blood pressure in this system is achieved by yet another hormone,

bradykinin which antagonizes the action of angiotensin II and so is hypotensive. The action

of ACE is nearly equipotent in forming angiotensin II and destroying bradykinin, both ac-

tions resulting in blood pressure rise. The action of angiotensin II is essentially terminated

by cleavage into a smaller peptide, angiotensin III, by an aminopeptidase (an enzyme that

chews off an amino acid from the amino end). Other angiotensin peptides are known, but a

discussion of these would needlessly confuse the story.

It may seem strange and cumbersome to involve so many molecules in this process, but

it allows for ﬁ ne control. In sum, angiotensin II and aldosterone combine to increase blood

pressure. Bradykinin antagonizes the action of angiotensin II. Thus, it seemed reasonable

to presume that inhibiting one of these enzymes would lead to a decrease in blood pressure.

In case the reader is wondering why bradykinin was not chosen as a target instead, much

experience has shown that it is far easier to inhibit an enzyme than it is to enhance its action.

Thus, renin- and angiotensin-converting enzymes appeared to be more tractable targets.

3.4.5 Attempts to Inhibit Renin

Initially, a major effort was made in many laboratories to search for suitable inhibitors

of renin, the enzyme that starts the cascade. Despite ﬁ nding a number of inhibitors with

powerful in vitro potency and the discovery of much useful information about the various

ways to convert peptides into more druglike molecules, no orally active agent emerged

from this work. Among the difﬁ culties that could not be solved satisfactorily, the molecular

weight of strong inhibitors always turned out to be too high for satisfactory absorption.

Clearly, control of blood pressure through the renin–angiotensin–aldosterone cascade

would have to be found elsewhere.

3.4.6 Attempts to Inhibit Angiotensin-Converting Enzyme

Attention was turned to the next enzyme in the system, ACE. It was clear at the outset from

a study of snakes that this was a feasible means of reducing blood pressure, so proof of

principle was not a problem. By about 1965 it was found that the active principles in the

venom of the South American pit viper Bothrops jararaca were a mixture of closely related

peptides that inhibit angiotensin-converting enzyme. The venom is collected in the gums of

the snake and is injected into the victim through hollow teeth (fangs) when the snake strikes

because the gums are compressed as a consequence of the bite. This inhibition reduces

blood pressure and immobilizes the victim. The question, then, was whether medicinal

chemists and physicians could utilize this mechanism to control hypertension through the

crafty application of drugs based on the venon. Unfortunately, the active constituents in the

venom, being peptides, must be injected in order to work. The snake has no problem with

this. Physicians and patients, however, would. As noted, most hypertensive individuals do

not perceive unpleasant symptoms from their disease until their disease is far advanced.

Consequently, there would be little motivation to take a lifelong series of injections when

generally feeling rather ﬁ t. Under these circumstances, the treatment would be more appar-

ently troublesome than the disease. Consequently, successful drugs for this indication would

have to be taken orally and not have unpleasant side effects or they would not be used.

EXAMPLE OF DRUG DEVELOPMENT 119

120 CONTEMPORARY DRUG DISCOVERY

3.4.7 Peptides Make Poor Orally Active Drugs

Peptides make poor drugs for chronic oral administration, as most of them are digested be-

fore absorption from the gastrointestinal tract. An intense effort has been expended to over-

come this. In most cases where this has been successful, they have been molecules with

comparatively low molecular weight and whose peptide bonds have been altered to make

them resistant to digestion. These altered, unnatural analogs are known as peptidomimetics,

and the hypotensive agents inspired by snake venom peptides are one of the classical ex-

amples of how this can be done.

3.4.8 Analoging Studies of Pit Viper–Inspired Peptides

By 1971 several synthetic peptide-based analogs of the pit viper agents had been prepared

at Squibb (now Bristol–Myers Squibb) without ﬁ nding an orally active analog. The main

understanding that came out of this preliminary work was that the intensity of the effect

could be modulated by changing the speciﬁ c identity and sequence of amino acids in the

inhibitor and that proline was a good amino acid to have in the ﬁ rst position. Persistence

ultimately produced dramatic results when the peptide character of the inhibitors was suc-

cessfully reduced.

3.4.9 Peptidomimetics

The next important clue came from the literature. Converting enzyme was known to be a

zinc protease, and the general catalytic mechanism of such enzymes was known. Water

itself is much too weak to cleave peptides (or taking a shower would be a deadly act!); it

takes centuries to cleave a peptide bond unless the reaction is catalyzed. The zinc atom of

the enzyme bonds to a key water residue and greatly activates it so that peptide hydrolysis

is rapid, speciﬁ c, and effective. This is illustrated in Fig. 3.8.

NHR′

His

Glu

His

Glu

His

NHR′

–RCO

His

Glu

His

R NHR′

His

–R′NH

A. ACE or Carboxypeptidase

in the absence of substrate

B. Bonding to substrate

C. Tetrahedral intermediate

D. Hydrophillic collapse E. Release and return of the enzyme

to the hydrated ground state

Figure 3.8 Peptide bond hydrolysis by zinc-metalloproteases.

Work on an apparently analogous enzyme, carboxypeptidase A (CPA), also known to

be a zinc-based peptidase, had previously shown that comparatively simple compounds,

especially 2-benzylsuccinic acid, would inhibit the enzyme. In this simpliﬁ ed system

(illustrated in Scheme 3.2) a docking carboxyl group was present in the ﬁ rst position so

as to align the enzyme with its ligand, and a second carboxyl group was present down

the chain to interact with the zinc atom inactivating the enzyme. Binding the zinc atom

to the carboxyl group of the inhibitor produces a much weaker nucleophile and displaces

the water that would normally function to cleave the peptide bond. Interrupting this process

by providing an alternative ligand inhibits the enzyme. The benzyl moiety evidently ﬁ ts

into a lipophilic pocket and/or helps align the rest of the carbon chain. The success of this

inhibition served as a model for the ACE project.

3.4.10 Adaptation to Inhibition of ACE

In contrast to carboxypeptidase A, which cleaves a single amino acid at a time from the end

of a peptide chain, ACE cleaves a dipeptide. The linker separating the docking carboxyl and

the zinc-interacting carboxyl would therefore have to be longer. Thus, a series of dicarboxylic

acids were prepared with a variety of aliphatic spacers between the two carboxyl groups to

see if this idea would work with ACE. These were all found to be weak inhibitors of ACE

and were not very stable. Only much later did it become apparent that the carboxypeptidase

A model was wrong in some respects. In any case, to enhance potency, the distal carboxyl

group was exchanged for a sulfhydryl group, as this was proposed to be a stronger ligand

for the zinc atom than is a carboxylate. This bioisosteric exchange proved promising, even

though the potency was still unsatisfactory, so a further series of analogs was prepared in

which additional systematic structural variations were explored (Scheme 3.3).

This type of analoging of a lead substance to reveal the inﬂ uence of the various func-

tional groups and selecting the best embodiments is the soul of drug seeking and is called

establishing a structure–activity relationship. In this particular instance, it was quickly

found that the initial carboxyl group, preferably attached to a proline moiety, an alkyl

group with the appropriate stereochemistry adjacent to the amide (peptide bond), the space

between the docking carboxyl and the zinc ligand, and a thiol rather than a carboxyl che-

lating to the zinc were all important. In particular, it was noted that an analog ending in a

thiol at the left end was more than 1000 times more effective than one with a carboxyl in

the analogous position.

Base

Lipophilic

Residues

Zinc

Scheme 3.2 Putative binding mode of benzylsuccinic acid to carboxypeptidase A.

EXAMPLE OF DRUG DEVELOPMENT 121

122 CONTEMPORARY DRUG DISCOVERY

The best of the analogs prepared in this work, now known as captopril, is a very effec-

tive drug for reducing high blood pressure. Its interaction with the enzyme was schema-

tized at the time as shown in Scheme 3.4, in comparison with angiotensin I substrate. This

sort of scheme is commonly used in advance of knowledge of the molecular interactions

actually occuring. It is a convenient way to summarize the structure–activity ﬁ ndings but

must not be taken literally. Among other problems with a literal interpretation, the reader

will note that the chemical bond angles are all wrong in such a drawing. In fact, the present

picture has been established recently by x-ray determination of the enzyme and its various

inhibitors and differs in some signiﬁ cant details from the early cartoon.

The carboxyl and the sulfhydryl play their anticipated key role in successful inhibition.

The carboxyl associates with an amine in the enzyme, and this is the docking interaction.

Analog

Relative k

vs. ACE

Remark

1.0

Captopril (standard)

12,500

ACE = 1.7×10

–9

CPA = 6.2×10

–4

CPB = 2.5×10

–4

(5 orders of magnitude

more selective)

Docking carboxyl essential

“Greasy” methyl is helpful

12,000

Distance between amide carbonyl

and the zinc ligand is important

120

The pyrrolidine ring is helpful

120

The stereochemistry of the

“greasy” methyl is important

1100

A sulfhydryl group is more

useful than a carboxyl

Scheme 3.3 Structure – activity relationships (SAR) of captopril analogs.

The other interactions take place next and establish potency and selectivity. Of great inter-

est is the ﬁ nding that captopril’s very signiﬁ cant potency against ACE (K

⫽ 1.7 ⫻ 10

⫺9

is dramatically higher than its potency against carboxypeptidase A (K

⫽ 6.2 ⫻ 10

⫺4

or carboxypeptidase B (K

⫽ 2.5 ⫻ 10

⫺4

M). (Carboxypeptidase comes in more than one

isomeric form.)

The implication of the preference for proline at the ﬁ rst position is that the speciﬁ c

conformation imposed by this amino acid is particularly well suited for the needs of ACE.

One notes that the normal substrate does not begin with this amino acid at all, so the normal

substrate probably must adopt an unusual conformation when bound to ACE. With pro-

line as the lead amino acid, the inhibitors are putatively preorganized in the most suitable

conformation for inhibition of the enzyme. This idea was reinforced by the ﬁ nding that a

number of dipeptides terminating in proline inhibited the enzyme somewhat and that the

strongly inhibitory snake venom peptides frequently ended in proline.

Captopril was subsequently introduced into the clinic as a very useful and popular orally

active antihypertensive agent. This is the most gratifying proof of principle demonstrating

that a peptidomimetic inhibitor of the renin–angiotensin–aldosterone system would be an

excellent way to control hypertension. As a further beneﬁ t, ACE inhibitors can be com-

bined with diuretics to give even more powerful control of blood pressure.

This brief recounting leaves out a number of false starts, as the pathway to success is

usually obscure before the fact and often requires the synthesis of hundreds or thousands

of analogs before ﬁ nishing. Many false trails are followed for a time, and the effort usu-

ally requires much successive hypothesis formulation and synthesis and evaluation cycling

before a useful ﬁ nal result emerges.

3.4.11 Success Inspires Competition

Captopril’s acceptance inspired others to enter the ﬁ eld with competing molecules. Proof

of concept was now at hand and ACE had become a well-validated therapeutic target. To

be able to compete successfully with captopril, it was essential that new contenders possess

–

ACE

–

Captopril

Angiotensin I

–

Scheme 3.4 Early beliefs about the interactions between captopril and angiotensin I with angiotensin-

converting enzyme (ACE).

EXAMPLE OF DRUG DEVELOPMENT 123

124 CONTEMPORARY DRUG DISCOVERY

signiﬁ cant advantages. In this context, physicians soon noted that acceptability of captopril

was diminished by the incidence of rashes and a comparative loss of taste sensations. These

side effects were not severe but were disincentives to compliance, and a number of laborato-

ries attempted to ﬁ nd agents retaining the desirability of captopril as a hypotensive but with

side effects reduced or eliminated. By analogy to another drug that had been marketed for

another indication but that also had these untoward properties, a research group at Merck hy-

pothesized that the taste and rash were associated with the presence of the sulfhydryl group.

3.4.12 Taking a Different Approach

If this hypothesis were correct, it would be necessary to get rid of the sulfhydryl moiety.

After examination of a number of potential alternatives, it was therefore decided to go back

to the carboxyl group as a zinc ligand, but it was necessary to enhance potency by further

analoging, as this route had proven very much less successful in the precedent work at

Squibb. It is sometimes found in medicinal chemical studies that this can be accomplished

by introducing another molecular feature into a lead substance that can interact with a

feature in the enzyme that is not utilized by the initial lead. When this works, the added

binding site gives additional energy of association and speciﬁ city. It also gives much greater

inhibitory power. This must not make the molecular weight too high or introduce otherwise

unsatisfactory interactions.

There were other problems to address as well. A sulfur atom is roughly twice as large as

an oxygen atom (strictly speaking, it is more bioisoelectronic than bioisosteric), so the spacer

group would probably have to be made somewhat longer than had been done with captopril.

Adding an additional methylene unit to accomplish this was helpful, as was use of an alanine

linker arm, but both together were not enough (Scheme 3.5). Partly counterbalancing the

polarity of the NH unit by adding a methyl gave a rather signiﬁ cant enhancement of potency.

After signiﬁ cant empirical experimentation, projecting an arm ending in an aromatic ring in

the form of a homophenylalanine residue apparently associated with a greasy pocket in the

enzyme and enhanced bonding dramatically. Enalaprilat was found to be very potent, even

more potent than captopril itself in vitro. The putative interaction of enalaprilat with ACE is

analogous to that of captopril, as illustrated in Scheme 3.4, except that a third greasy pocket

would need to be added to the left in the diagram to accomodate this ﬁ nding.

Enalaprilat was not very effective when given orally, however, due to poor absorption.

This was judged to be a consequence of the possession of two carboxyl groups, making

it too ionizable for efﬁ cient oral activity. Thus, it was decided to modify the structure so

that the distal group was present as an ethyl ester. This solved the polarity problem, and

this comparatively well-absorbed analog was given the name enalapril. The ester function,

however, greatly decreased inhibition of the enzyme in vitro, but after oral absorption,

esterases cleaved enalapril to produce the much more active enalaprilat in the blood where

ACE is found. Thus, one had the best of both worlds. Enalaprilat inhibits ACE at 1.2 nM!

The ethyl ester of enalaprilat (enalapril) has found great acceptance in oral treatment of hu-

man hypertension. The alteration of a functional group so as to enhance absorption with the

intention that the blocking group be removed by enzymic action in the body to regenerate

the active drug is known as prodruging.

3.4.13 Analoging to Enhance Absorption

Another way to overcome the absorption problem, but without resort to prodruging, with its re-

quirement that the patient cooperate by activating the drug (not everyone is capable of doing this

efﬁ ciently), is to alter the isoelectric point of the drug so that its degree of ionization (acidity) is

reduced sufﬁ ciently that it passes through the gut wall at an acceptable level. Lisinopril (Scheme

3.6) embodies this concept. In this important drug, an amino group has been introduced into a

side chain of the molecule to balance the acidity of one of the two carboxyl groups.

One popular way to investigate drugs such as these is intellectually to split them into

component parts each of which can be synthesized readily and then assembled. In this man-

ner, enalaprilat and lisinopril could be envisioned as being a combination of three parts,

A–B–C. A thorough investigation would amount to analoging A, B, and C and then join-

ing them together. Hopefully, the best embodiment of A would add its virtues to the best

embodiments of B and C as well.

Lisinopril apparently emerged from such as exercise in which the central alanine of

enalaprilat (B in the example of preceding paragraph) was replaced by a series of amino

acid residues. It had been postulated that the function of the methyl group of alanine was to

dictate the most useful conformation, but it was not clear from this whether further beneﬁ ts

could be achieved by analoging at this position. From this group of analogs, lisinopril’s

relative potency leaps out (Scheme 3.6, item 8). Its speciﬁ c potency is equal to that of

enalaprilat. Even more important, lisinopril is orally effective at nearly the same dose as

its ethyl ester, so prodruging is not necessary to achieve control of high blood pressure.

Lisinopril has become an extremely important drug.

Relative I

18,333

4,083

2,000

Scheme 3.5 Synopsis of some of the structure – activity alterations leading to enalaprilat (bottom

structure).

EXAMPLE OF DRUG DEVELOPMENT 125

126 CONTEMPORARY DRUG DISCOVERY

Lysine is particularly convenient since it is a natural amino acid. To see whether it was

the best amine to use, a number of analogs were investigated systematically. It was indeed

the best at that position. Thus, its contribution to activity is not only a consequence of its

polarity, but it apparently ﬁ nds a compatible partner in the enzyme to which it binds par-

ticularly well. The data in Scheme 3.7 support this conclusion. It is also now known that

lisinopril beneﬁ ts from active uptake by a peptide transporter.

This abbreviated exposition of their development of ACE inhibitors illustrates the sys-

tematic work that goes into drug design and discovery. It shows why advances achieved as

a consequent of so much effort are not abandoned without great reluctance.

3.4.14 Clinical SAR

The great clinical acceptance of these three drugs led to intense competition by a variety of

ﬁ rms, and a substantial number of additional analogs are now available. As usual, as time elapses

and the patent literature becomes increasingly congested, the various analogs introduced bear

a less and less obvious molecular resemblance to captopril, where the story began. This is

Relative I

R =

NMe

NHAc

480

4.5

4.3

7.6

Scheme 3.7 Investigation of lisinopril analogs for comparative potency.

Analog X = Relative I

Gly

L-Ala

Alpha-MeAla

N-Me-L-Ala

Val

His

Phe

Lys

192

3.2

2083

Scheme 3.6 Amino acid substitutions leading to lisinopril analog 8.

clear from inspection of the structural formulas of these agents (Scheme 3.8). In time it was

found that the docking carboxyl-bearing ring could be other than proline (component C in the

ﬁ gure) but with the exception of temocapril, the bridging alanine unit (B) has generally been

preserved, with a few exception, as has been the homophenylalanine ester moiety (A).

Most of these analogs use the prodrug concept and compete for a more and more frag-

mented market. As often happens, the ﬁ rst drugs introduced continue to generate the big-

gest rewards to their ﬁ rms. First is best, and good enough, soon enough, are mottoes that

one hears frequently in discussions between medicinal chemists.

AcS

EtO

OMe

EtO

Quinapril (1989; Warner-Lambert)

Ramipril (1989; Hoechst)

Perindopril (1988; Servier)

Alacepril (1988; Dainippon)

Cilazepril

(1990; Hoffmann-LaRoche)

Benazepril (1990; CIBA-Geigy)

Delapril (1989; Takeda) Imidapril (1993; Tanabe)

Moexipril

(1995; Warner-Lambert)

Trandolapril (1993; Roussel Uclaf)

Spirapril (1995; Schering)

Temocapril (1994; Sankyo)

Fosinopril (1983; Tanabe)

Scheme 3.8 Clinical competitors for captopril, enalapril, and lisinopril.

EXAMPLE OF DRUG DEVELOPMENT 127

128 CONTEMPORARY DRUG DISCOVERY

3.4.15 More Recent Work

Later work demonstrated that successful manipulation of the renin–angiotensin–aldosterone

system could also be achieved by interfering with the productive association of angiotensin

II with its receptors. This stratagem allows the body to produce angiotensin II unimpeded

but interferes with binding of the hormone to its receptor and decreases blood pressure in

this way. Recounting this exciting story would also be proﬁ table but would take us away

from illustrating our main theme.

3.4.16 Résumé

This account beneﬁ ts from some retrospective wisdom in order to make a coherent story

from a complex sequence of events. More important, however, it serves not only to illus-

trate many of the principles set forth in the ﬁ rst part of this chapter but also as one of the

comparatively few examples of the rational conversion of a parenteral peptide lead into an

orally active peptidomimetic drug.

3.5 CONCLUSIONS

It should now be amply clear that drug design and optimization is a complex activity re-

quiring an understanding of cellular biochemistry, pathology, and synthetic chemistry. The

molecules that emerge from drug-seeking campaigns are often only a very small percent-

age of those prepared and tested. These molecules are not abandoned readily. They are,

however, all too often not yet ready for the market because a practical synthesis must yet

be found. This requires a high level of skill on the part of the developmental chemist. The

remainder of this book is devoted to this important topic.

ADDITIONAL READING

The reader who is interested in digging deeper into a few of the topics covered in this sur-

vey can proﬁ t from consulting the following articles prepared by authors actively engaged

in the work synopsized.

Cody, V., Galitsky, N., Rak, D., Luft, J. R., Pangborn, W., and Queener, S. F. (1999). Ligand-induced

conformational changes in the crystal structures of Pneumocystis carinii dihydrofolate reductase

complexes with folate and NADP, Biochemistry 38:4303–4312.

Lipinski, C. A., Lombardo, F., Dominy, B. W., Feeney, P. J. (2001). Experimental and computational

approaches to estimate solubility and permeability in drug discovery and development settings,

Adv. Drug Deliv. Rev 46:3–26.

Ondetti, M. A. (1981). Inhibitors of angiotensin-converting enzyme, in Biochemical Regulation of

Blood Pressure, R. L. Sofer, Ed., Wiley-Interscience, New York, pp. 165–204.

Patchett, A. A. (1993). Enalapril and lisinopril, in Chronicles of Drug Discovery, Vol. 3, D. Led-

nicer, Ed., ACS Professional Reference Books, American Chemical Society, Washington, DC,

pp. 125–162.

Veber, D. F., Johnson, S. R., Cheng, H.-Y., Smith, B. R., Ward, K. W., and Kopple, K. D. (2002).

Molecular properties that inﬂ uence the oral bioavailability of drug candidates, J. Med. Chem.

45:2615–2623.

129

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

COMBINATORIAL CHEMISTRY

IN THE DRUG DISCOVERY PROCESS

IAN HUGHES

GlaxoSmithKline Pharmaceuticals

Harlow, Essex, UK

4.1 INTRODUCTION

What is combinatorial chemistry, and what is its value to medicinal chemists engaged in

the drug discovery process? During the decades leading up to the 1990s, approaches to

synthetic chemistry remained virtually unchanged. Obviously, the tools available to the

chemist improved considerably, albeit incrementally, over the years: new, milder reaction

conditions, improvements to puriﬁ cation and analytical methods, and the ability to synthe-

size ever more complex molecules. However, during this period a single medicinal chemist

would be expected to make, on average, around 50 new molecules for screening per year

and would typically work on only one or two molecules at a time. In the early 1990s, the

medicinal chemistry community witnessed the start of a paradigm shift in the way they

would work in the future. A variety of new techniques that had been developing in the ﬁ eld

of peptide synthesis, with little impact on most synthetic chemists, were shown to be appli-

cable to small molecule synthesis. The following years witnessed a ﬂ urry of activity as the

new technologies that now constitute the ﬁ eld of combinatorial chemistry were developed.

Driving this surge of interest was the potential to increase the speed and efﬁ ciency of drug

discovery with the obvious economic advantages, as well as opportunities to commercial-

ize the technologies themselves.

Our goal in this chapter is to review key aspects of the “combinatorial revolution” and

their application to the drug discovery process. It would be foolish to attempt compre-

hensive coverage of such a broad ﬁ eld in a few pages, so the highlights selected are ac-

companied by a bibliography of reviews and books giving more in-depth coverage of the

subjects at the end of the chapter. As with any ﬁ eld of science, a terminology has developed

130 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

to communicate the new concepts, and a useful glossary has been published.

A similar

glossary of medicinal chemistry terms might also be useful for the reader.

The selection

of topics covered is based on the author’s experience and perspective and should in no

way be taken to imply inferiority of any approaches omitted. Again, to limit the material

covered, the primary focus is on the application of combinatorial techniques to small mol-

ecule synthesis, as opposed to that of biopolymers (oligopeptides, oligonucleosides, and

oligosaccharides).

4.1.1 The Birth of Combinatorial Chemistry

When and how did combinatorial chemistry begin? Lebl gives an interesting perspective

of some “classical” papers in the area.

Some of the key milestones are summarized in

Table 4.1 and discussed below. It is interesting to note that many of the main developments

were ﬁ rst applied to peptide chemistry with translation to small molecule synthesis often

Year Class

Contributors Development Ref.

1963 P Merriﬁ eld Solid-phase peptide synthesis 4

1970 SM Leznoff Early nonpeptide solid-phase synthesis 6

1984 P Geysen Multipins for parallel synthesis 7

1985 P Houghten Teabags 9

1988 P Furka Mix-and-split synthesis 10, 11

1991 P Fodor Light-directed spatially addressable

parallel synthesis

P Houghten Screening of mixtures 13

P Lam On-bead screening: one bead, one peptide 22

1992 P Houghten Positional scanning 14, 15

P Brenner, Lerner Oligonucleotide tags for encoding 23, 24

SM Ellman Solid-phase synthesis of

benzodiazepines

25, 26

1993 SM De Witt Diversomers; parallel solid-phase

synthesis on resin

P Ohlmeyer Haloaromatic tags for binary encoding 28

1994 SM Smith Indexed libraries 16

1995 P Deprez Orthogonal libraries 17

P Nicolaou Radio-frequency tags 32

P/SM Janda Liquid-phase combinatorial synthesis 36

1996 P Ni Secondary amine tags for encoding 29

P/SM Geysen Mass and isotopic encoding 30

SM Curran Fluorous tags for reagents and

substrates

39, 42

SM Cheng, Boger Mixtures by solution chemistry 44, 45

SM Kaldor Supported reagents 46, 47

1997 SM Lipinski Design: developability 51

1999 SM Ley Multistep solution synthesis with

supported reagents

49, 50

TABLE 4.1 Key Developments in Combinatorial Chemistry

P, peptides; SM, small molecules.

occurring a number of years subsequently. In this context, few would question the pioneer-

ing contribution of Merriﬁ eld in utilizing functionalized cross-linked polystyrene beads

as a solid support for the synthesis of peptides.

Key attributes of solid-phase chemistry

in terms of using forcing conditions (excess reagents), simpliﬁ ed reaction workup (ﬁ ltra-

tion), and ease of automation are now well known. In a similar fashion, Letsinger applied

the concept to solid-phase oligonucleotide synthesis.

Solid-phase approaches were also

applied to nonpeptide molecules during the 1970s,

but the potential utility of this pioneer-

ing work was not fully appreciated and did not ﬁ nd widespread acceptance.

Geysen’s use of an array of polyethylene rods (multipins) as supports for peptide syn-

thesis

introduced the concept of improving the speed and efﬁ ciency of synthesis by per-

forming many reactions in parallel. A similar process, using synthesis resin in the wells of

a microtiter plate (MTP), combined with automated liquid handling was reported several

years later.

In 1995, Houghten introduced numbered polypropylene mesh packets of resin (“tea

bags”) as an alternative means of achieving efﬁ ciency of synthetic effort, allowing many

different substrates to be subjected to the same reaction conditions in a single reaction

vessel.

Furka, who in 1988 reported the ﬁ rst split-and-mix synthesis (also referred to as

pool/split, portion/mix, and divide, couple, recombine), took the idea a stage further.

this process the resin is split into batches, each of which is reacted separately with a single

diversity reagent, then the batches are mixed together thoroughly. Assuming the use of n,

m, p diversity reagents in successive cycles of splitting, reacting, and mixing, the process

results in n ⫻ m ⫻ p products in only n ⫹ m ⫹ p synthetic steps. A subsequent publication

described the split-and-mix synthesis of a mixture of 180 pentapeptides in 15 coupling

cycles.

The potential for efﬁ ciency of synthetic operations was now enormous, and the

concept ﬁ red the imagination of chemists, as witnessed by the escalation of publications

over the following years.

Around the same time, Fodor’s group extended and miniaturized the multipin concept

by making use of photolithography techniques from the semiconductor industry.

In com-

bination with photolabile protecting groups, the synthesis of an array of 1024 peptides

grafted onto the surface of a glass microscope slide in an area of only 1.6 cm

was achieved.

Densities of up to 250,000 synthesis sites per square centimeter were considered realistic

goals. The approach above was somewhat specialized in its application and its enabling

technologies. In contrast, the ready availability of solid-phase beads and their ease of han-

dling in a conventional laboratory led to their continued widespread use.

4.1.2 Development of Screening Strategies for Libraries

It soon became apparent that the full latent power of combinatorial chemistry would beneﬁ t

the drug discovery community only if effective methods for screening the libraries were de-

veloped. In 1991, in the same issue of Nature, groups led by Houghten and Lam described

two distinct strategies, which were to form the basis of future approaches to screening

libraries. The ﬁ rst strategy, described by Houghten, comprised an iterative deconvolution

approach for ascertaining the active component(s) in a large library of 34 million hexapep-

tides.

In the ﬁ rst round of screening the library was divided into 324 pools, each having

a unique combination of residues in the ﬁ rst two positions but all possible variations in the

remaining four positions. Having ascertained the most potent residues in positions 1 and

2, a second set of libraries were prepared where these two residues were ﬁ xed and position

3 was held constant while positions 4 to 6 were varied. The active library thus deﬁ ned the

INTRODUCTION 131

132 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

optimal residue at position 3. The process was repeated until the preferred residue at each

position was discovered.

Houghten developed the approach further in a process known as positional scanning, in

which six hexapetide libraries were prepared, each composed of 18 mixtures with a single

position deﬁ ned as one of 18 natural amino acids.

14,15

Hence, by screening the individual

pools, the optimum monomer at each position could be determined. Several alternatives to

this technique were to be reported over the next few years: for example, Smith’s indexed

libraries

and the orthogonal combinatorial libraries of Deprez.

These all rely on synthe-

sizing subsets of a library in several different combinations in such a way that the active

subsets will share a single compound. These methods all utilize mixtures of compounds

released from the solid-phase supports into solution. Although the solutions thus obtained

are amenable to virtually all existing assay systems, it is always possible that optimal com-

pounds may be missed as a result of averaging of activities, or combinations of agonist and

antagonist molecules within a mixture. The statistical implications of pooling and deconvo-

lution strategies have been considered by several authors,

18–20

and tested experimentally.

The second key strategy, reported by Lam’s group,

capitalized on a consequence of the

mix-and-split process: namely, that each bead in the ﬁ nal mixture bears a single product,

since the bead has only been exposed to a single diversity reagent at each synthetic step. So,

instead of cleaving the products from the beads, a library of several million pentapetides

was screened by exposing the beads to a solution of acceptor molecules covalently tagged

with a reporter system (e.g., the enzyme alkaline phosphatase, or ﬂ uorescene). Beads bear-

ing products that bound strongly to the acceptor became stained and could be identiﬁ ed

visually, removed, washed free of acceptor, and microsequenced to determine the peptide

structure of interest.

This one-bead, one-peptide approach was subsequently made more powerful, and ex-

tended beyond the realms of peptide chemistry, by the concept of encoded libraries. The

key idea here is, at each synthetic step, to introduce a “readable” chemical moiety onto the

bead, such that this code or tag encodes the diversity reagents used at each step. Of course,

a key property of a coding system is that the tag molecules can be identiﬁ ed more easily

than the products they encode. For example, because of the small amount of material pres-

ent on a single bead, Brenner and Lerner used oligonucleotide sequences to encode peptide

libraries.

23,24

Ampliﬁ cation of the tag by the polymerase chain reaction (PCR) then allowed

the sequence to be read, and this could be related back to the peptide product structure.

4.1.3 From Peptides to Small Molecule Synthesis

The focus of the discussion so far has concerned peptide chemistry, and much of this was

probably considered an academic curiosity by many medicinal chemists. In 1992, however,

Ellman demonstrated the multistep solid-phase synthesis of a 1,4-benzodiazepine library,

highlighting the fact that such methods need not be restricted to oligomeric species. Subse-

quently, the chemistry was used in the synthesis of a library of 192 benzodiazepine deriva-

tives using Geysen’s multipin approach.

These publications established that solid-phase

chemistry and the associated technologies could be transferred to the small molecule me-

dicinal chemistry arena and inspired a rapid expansion of interest.

Shortly afterward, De Witt described a simple apparatus comprising a matrix of gas

dispersion tubes, each containing resin, for the simultaneous synthesis of an array of 40

products using solid-phase chemistry.

This Diversomer approach, coupled with simple

automation, was applied to the parallel synthesis of hydantoin and benzodiazepine arrays.

That this innovation had come from an industrial rather than an academic laboratory may

have been the signal to many industrial chemists that combinatorial chemistry tools had

genuine applicability.

Two fundamental philosophies were now established (split-and-mix and parallel synthe-

sis), and each continued to be developed independently. Further innovations were made in

techniques for encoding split-and-mix bead libraries. A binary encoding system, introduced

by Ohlmeyer et al.

in 1993 made use of combinations of chemically inert haloaromatic

tags to cap off 1% of free amino groups on the growing peptide prior to adding the encoded

amino acid. Following screening of the bead-bound peptides, photolytic cleavage of the

tags on the active beads allowed rapid assignment by electron-capture gas chromatogra-

phy. Alternative encoding strategies introduced subsequently include the use of secondary

amine tags,

mass and isotope tags readable by mass spectrometry,

and a tag-free system

where libraries are designed such that each component has a unique molecular weight and

is thus directly identiﬁ able by mass spectroscopy.

In 1995, Houghtens’ teabag idea was developed into a more robust process that could be

automated. Nicolaou reported

the use of porous microreactors that contained not only the

synthesis resin, but also a small SMART (single or multiple addressable radio-frequency

tag) semiconductor unit. Because the synthetic history of each microreactor could be re-

corded on the chip, they could be combined in pools for common reactions, then resorted

for subsequent steps by electronic scanning. Hence, the efﬁ ciencies of split-and-mix syn-

thesis were combined with the ability to synthesize useful quantities (ca. 10 to 20 mg) of

discrete compounds.

4.1.4 Beyond Solid-Phase Chemistry

Although solid-phase chemistry was the driving force behind the developments described

so far, there was an underlying penalty to be paid. Because the behavior of molecules

when tethered to a solid support is inﬂ uenced by the type of support and the linker used to

tether the molecule, it has been necessary to optimize reaction conditions on solid phase

for reactions that were already well understood in solution. The plethora of publications

in this area over recent years (see Table 4.7 for reviews) indicate that many reactions can

indeed be transferred to solid phase, sometimes advantageously when excess reagents can

be used to force reactions, but sometimes with limited scope compared to the solution

counterpart. In most cases a route development phase is essential prior to embarking on

the library synthesis if high-quality products are desired, because of the inability to per-

form puriﬁ cation during a solid-phase synthesis. This has necessitated the development

of new procedures for the analysis of bead-bound intermediates and products.

33–35

More-

over, a necessity of solid-phase chemistry is a convenient handle on the target molecule

through which it can be linked to the solid support. With peptides this was typically the

terminal carboxylic acid group. Much effort has been expended in identifying suitable

linker groups for the wide range of functional groups present in druglike molecules, as

well as traceless linkers for situations where no convenient functional group is available.

(See Table 4.7 for reviews.)

Notwithstanding these considerable efforts, it was clear that many medicinal chemistry

targets did not readily lend themselves to solid-phase synthesis, or that such routes were

precluded due to lengthy validation phases. To retain some advantageous features of solid-

phase chemistry, yet utilize a homogeneous reaction mixture, Janda’s group introduced

the concept of liquid-phase combinatorial synthesis in 1995.

Substrates were linked to

INTRODUCTION 133

134 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

polyethylene glycol monomethyl ether, a soluble linear polymer that could be precipitated

with ether to give a solid that was ﬁ ltered and washed free of excess reagents. Similarly,

methanol was used to precipitate substrates on non-cross-linked chloromethylated poly-

styrene.

37,38

A related strategy, introduced by Curran, was to capitalize on the insolubility

of ﬂ uorocarbon ﬂ uids in both aqueous and organic solvents. Products tagged with a per-

ﬂ uoro group were fully soluble during synthesis but could be separated from products by

a triphasic (aqueous, organic, ﬂ uorous) workup.

39–41

In a reverse strategy, reagents bearing

ﬂ uorous groups can be separated from nonﬂ uorous products.

42,43

Since 1996 there has been growing attention to truly solution-phase combinatorial

chemistry, and Boger

44,45

described the synthesis of mixtures by applying split-and-mix ap-

proaches to solution chemistry. Key to the success of this approach was a carefully designed

strategy for application of acid–base workup procedures to purify the libraries at interme-

diate stages. A more recent innovation of great impact is the use of solid-supported scav-

engers,

46,47

which are introduced at the end of a solution-phase synthesis step to sequester

unreacted building blocks or reagents. For example, aminomethyl polystyrene can be used

as a scavenger for acylating agents. In tandem, a renaissance in the use of solid-supported

reagents has essentially reversed the solid-phase synthesis paradigm; reaction workup is

still by simple ﬁ ltration, but now the product remains in solution while reagents and scav-

enged materials are removed on the solid support. In combination, solid-supported reagents

and scavengers are not only enjoying widespread application in high-throughput parallel

solution-phase library synthesis,

but are also proving invaluable in multistep drug

and

natural product

syntheses.

Library design (discussed in detail below) has always been an important prelude to

library synthesis to maximize interactions with target proteins and thus produce molecules

of increasing potency and selectivity toward their target. An oft-quoted publication by

Lipinski

in 1997 highlighted the additional need to constrain certain physical properties

within ranges deﬁ ned by his rule of ﬁ ve if oral activity is to be achieved.

Alongside the events noted above has been the development of commercial hardware

for implementing the ideas and concepts of combinatorial chemistry. Solid-phase synthe-

sizers, originally developed for peptide chemistry, were adapted for wider use. Heating and

cooling were added as well as the ability to handle slurries (e.g., resins) and the provision

of inert, moisture-free atmospheres. Simple block-based systems have also been devel-

oped for both solution and solid-phase chemistry and have become commonplace in most

laboratories.

Following synthesis, the downstream processes of workup, liquid handling, evapora-

tion, analysis, quantiﬁ cation, and puriﬁ cation have all been made more efﬁ cient by the ap-

plication of automated or parallel techniques. Interestingly, the most successful hardware

developments have come about through close collaborations between the users (industry)

and the vendors. Additionally, most combinatorial chemistry groups have implemented

software systems to manage the deluge of data generated during the various phases of

library production.

The last decade has seen major advances in the tools that medicinal chemists have at

their disposal to apply to the problems of drug discovery. As the ﬁ eld has developed, there

have been a number of noticeable trends, outlined in Table 4.2, in the drivers and applica-

tions of combinatorial methods. After a period of innovation during which not all dreams

became reality, we have now entered an era in which the application of appropriate meth-

ods can truly have an impact on productivity within the drug discovery process. It is aspects

of these applications that are discussed in more detail in the remainder of the chapter.

4.2 THE ROLE OF COMBINATORIAL CHEMISTRY IN DRUG DISCOVERY

Although the drug discovery process is highly complex, it may be considered to consist

of four key stages, the ﬁ rst being the discovery and deﬁ nition of a biological target. In

the majority of cases this is a protein with some known or unknown involvement in a dis-

ease process. A consequence of the success of the Human Genome Project in providing a

blueprint of the human genome, in combination with the complementary approach of pro-

teomics, is the unprecedented rate of discovery and understanding of new targets. This is

set to have a profound effect on drug discovery in the new millennium.

52,53

Once a relevant

target has been identiﬁ ed, assays are developed to identify molecules that bind to the target

and modify its behavior in the desired manner.

The process now enters the lead generation stage. Here, the goal is to identify a lead

compound whose properties (principally, but not exclusively its potency against the target)

make it a suitable candidate for more in-depth exploration. In most cases this involves

a high-throughput screening exercise, where large numbers of compounds are screened

against the target in the assay developed previously. Clearly, the greater the number of

appropriate compounds available for screening, the greater the chance of success. The pro-

duction of lead discovery libraries was an early driver for combinatorial chemistry and con-

tinues to be an important application. Libraries for this purpose are typically of signiﬁ cant

size, from a few thousand to tens or hundreds of thousands of compounds and may exist

in a number of formats that match the screening protocols in use.

For example, libraries

of discrete compounds would typically be of general utility and enter a standard screening

system along with a company’s historical compound collection. Soluble mixture libraries

might fulﬁ ll a similar role, but with the added necessity of deconvolution procedures to

pinpoint the active component. Libraries on solid supports require alternative screening

strategies,

which may be particularly suited to certain speciﬁ c target classes but are un-

likely to be as generally applicable as solution assays.

When knowledge of the three-dimensional structure of a protein is known, either through

x-ray diffraction studies or through homology modeling, de novo design of a library of

compounds is a possibility. Computer modeling of the interactions between candidate small

molecules and the known (or assumed) active site of the protein can provide a starting

From: To:

Quantity (numbers) Quality

Few large libraries Many smaller libraries

Innovation (and failure) Application (and focus)

Chemistry driven Biology (target) driven

Peptides Small molecules

Mixtures Singles

Solid-phase chemistry Any chemistry

Synthesizability Developability

Specialist tools General tools

Speculative Routine

Manual Automated

Fragmented Integrated

TABLE 4.2 Trends in Combinatorial Chemistry

THE ROLE OF COMBINATORIAL CHEMISTRY IN DRUG DISCOVERY 135

136 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

point for library design. Such an approach will often take place in parallel with high-

throughput screening exercises.

Once a suitable lead is identiﬁ ed, the compound enters the third stage of the process,

optimization. The goal here is to modify the lead structure such that a number of criteria

for progression into the ﬁ nal, development, stage are satisﬁ ed. Important criteria, the ac-

tual values of which will be set on a case-by-case basis, include high potency against the

target as well as selectivity against related targets. The development candidate will also

exhibit appropriate absorption, distribution, metabolism, and excretion (ADME) and toxic-

ity properties and will have a positive proﬁ le against P450 enzymes. Screening for these

properties at this stage, whether by in vitro, in vivo, or computational protocols is becom-

ing an essential ﬁ lter prior to entering the costly development phases.

56,57

The key role of combinatorial chemistry at this point is to provide sets of compounds

with which to derive an understanding of the contribution of structural elements of the

molecule toward its binding to the protein target, so that structure–activity relationships

(SAR) can be developed. Lead optimization libraries typically consist of hundreds or a

few thousand compounds at most. As the iterative cycles of library design, synthesis, and

screening progress and more is known about SAR, the library size might fall to a handful

of compounds focused on ﬁ ne-tuning some aspect of the overall proﬁ le.

Eventually, the optimized product will enter the ﬁ nal, although most lengthy and costly

stage, that of development into a marketable drug, including optimization of synthetic

routes for production-scale synthesis, stringent trials in patients to prove clinical efﬁ cacy

and safety, regulatory approval, and marketing; It is only once sales have begun that the

massive research and development costs can start to be recovered, and this is possible only

until the patent life of the compound expires. Any time savings in the overall process there-

fore represent opportunities for increased revenue, and combinatorial chemistry clearly has

an important role to play in this respect.

Some of the characteristic features of lead generation and lead optimization libraries are

summarized in Table 4.3. These classiﬁ cations are extremes, and there are frequently cross-

overs. For example, generic libraries are lead generation libraries designed with the intent

Lead Generation Libraries

Lead Optimization Libraries

Populate screening collection Modify leads to improve proﬁ le

Provide leads for range of targets Optimize leads for speciﬁ c target

Wide coverage of chemical space Focus on deﬁ ned area of chemical space

Larger arrays Smaller arrays

Small scale (few milligrams) Larger scale (tens or hundreds of milligrams)

High quality preferred High quality essential

General druglike properties Speciﬁ c potency, selectivity, ADME properties

Compounds made in one batch Compounds made in iterative cycles with

feedback from biology

Signiﬁ cant route development (often, new

chemistry)

Minimal route development time (often, known

chemistry)

Specialized groups Medicinal chemists

TABLE 4.3 Characteristic Features of Lead Generation and Lead Optimization Libraries

Also called prospecting or generic libraries.

Also called focused or directed libraries.

of interacting with some general class of target (e.g., protease, kinase, G-protein-coupled

receptor). Conversely, the products from highly focused optimization libraries may ulti-

mately join the screening collection for future lead generation operations.

4.3 DESIGNING COMBINATORIAL LIBRARIES

There have been several estimates of the number of druglike molecules that could theo-

retically be synthesized,

58,59

ranging from 10

to 10

200

. The impossibility of even con-

templating such a goal is evident when one considers the availability of materials (the

total mass of the observable universe is around 10

kg) or the time required (the Earth is

only 10

seconds old). The key aim of library design is to select from this overwhelming

number of possible compounds a subset that can be synthesized efﬁ ciently and which has a

high chance of either producing new leads or improving the properties of existing leads.

The wide range of design methodologies described in the literature bears testament to the

observation that there is no deﬁ nitive solution to this problem.

61–64

Some commercial soft-

ware systems for this purpose have been reviewed.

Although beyond the scope of the current discussion, some of the techniques described

below have been applied to selecting compounds (either individuals or library sets) for pur-

chase to augment existing screening collections. Also, when screening capacity is less than

the number of compounds available, these methods can be used to select representative or

biased subsets of the entire collection for screening.

In most cases, the starting point in library design is a virtual library, which deﬁ nes

all the conceivable compounds accessible by a particular synthetic route. Frequently, the

virtual library is deﬁ ned as a ﬁ xed template (scaffold, core structure) with a number of

variable (diversity) sites. The potential set of substituents (R-groups, fragments) at each

site is derived from diversity reagents (building blocks, monomer sets) deﬁ ned in scope by

their availability and suitability for the particular synthetic route. Alternatively, the virtual

library may be deﬁ ned as a series of transforms (reactions) applied to the monomer sets.

This transform approach is more intuitive to the medicinal chemist and can represent the

products of certain reactions, such as the Diels–Alder reaction, more satisfactorily than the

template–substituent approach.

However, the latter is more efﬁ cient computationally and

hence encountered most commonly.

4.3.1 Describing and Measuring Diversity

A major emphasis in the literature has been methodology for selecting the diversity reagents

that will be used to decorate the template. In this respect, an inescapable concept, with

many deﬁ nitions, is diversity.

67–69

A diverse subset of compounds is considered to be one

that represent the chemical space occupied by the complete virtual library. Deﬁ ning chemi-

cal space, and hence measuring diversity, requires descriptors that represent properties of

compounds in a computer-friendly format (numbers, matrices, or bit strings).

70,71

A large

number of descriptors have been reported (see Table 4.4 for examples)

and are frequently

used in combination, resulting in a multidimensional chemical space. Imaginative pictorial

representations such as ﬂ ower plots

are useful for visualizing many properties simultane-

ously. However, when data sets are large, the dimensionality is often reduced by techniques

such as principal component analysis

69,73

to simplify both computation and visualization

of the results.

DESIGNING COMBINATORIAL LIBRARIES 137

138 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

Because of the size of virtual libraries, the descriptors for each compound should be

easy and quick to compute since, by deﬁ nition, we cannot use measured properties for

molecules that do not exist. Ideally, the descriptors used should in some way represent po-

tential interactions between the molecules and their biological targets. The BCUT descrip-

tors of Pearlman,

74,75

as well as various fragment keys and ﬁ ngerprint descriptors,

fulﬁ ll

this role. Descriptors representing bulk molecular properties such as molecular weight,

lipophilicity, or number of rotatable bonds have greater value as ﬁ lters (see below) than for

diversity assessment.

Clearly, molecules and their sites of interaction with targets are not two-dimensional

entities, so intuitively three-dimensional descriptors should better represent their intermo-

lecular interactions, although this remains the subject of debate.

76–78

However, new issues,

such as how to align diverse sets of molecules for meaningful comparison and how many

of the multiple conformations of a ﬂ exible molecule to incorporate must be addressed

carefully and will signiﬁ cantly challenge computational resources. It is likely that two-

dimensional descriptors are adequate for lead generation libraries, while the speciﬁ city

needs of an optimization library are better catered for by three-dimensional descriptors.

An important class of three-dimensional descriptor is based on the types of pharmo-

cophoric groups in a molecule and the distances between them.

80–82

Pharmocophores are

features, such as acidic or basic centers, hydrogen bond (H-bond) donors and acceptors,

aromatic ring centroids, and hydrophobic groups, known to be important in drug–receptor

interactions. The relative positions of pharmocophores must, of course, be calculated for

each likely conformation of a ﬂ exible molecule.

Descriptor Comments

Molecular weight Whole molecule property; use as simple ﬁ lter

Hydrogen-bond donor, acceptor counts Whole molecule property; simple indication of

number of potential binding sites

Rotatable bonds Indication of molecular ﬂ exibility

c log P Calculated log P (measure of lipophilicity)

CMR Calculated molar refractivity (measure of size)

Structural keys Presence of small structural fragments

Two-dimensional ﬁ ngerprints Bit-string representation of atom paths through

molecule

Atom pairs Atom type and number of bonds separating pairs

of atoms

Topological torsion Atom types in four atom paths

Topological indices Single value representing molecular shape and

connectivity

BCUT descriptors Burden–CAS–University of Texas; use

atomic properties related to strength of

intermolecular interactions

Three-dimensional screens Encode distances and angles between features in

molecule

Pharmacophore keys Various approaches; commonly, triplets of

pharmacophoric groups deﬁ ned by their types

and separation

TABLE 4.4 Descriptors Used in Library Design

There has been considerable debate as to whether the less computationally intensive

evaluation of reagent diversity, rather than product diversity, is an acceptable approach, and

the outcome appears to be descriptor dependent.

83,84

Certainly, analysis of a few thousand

reagents is considerably faster than that of many millions of enumerated products and may

sufﬁ ce in some cases. However, product evaluation is likely to be essential when three-

dimensional descriptors are used.

A number of different algorithms have been developed and applied to the sampling of

the chemical space occupied by the virtual library, of which two major types, clustering

and bin-based partitioning, will be described. Clustering algorithms aim to group together

those compounds with some degree of similarity to each other, and will be illustrated by

two widely used methods.

Jarvis–Patrick clustering

groups molecules according to the number of nearest neigh-

bors they have in common. The method is rapid and suited to large libraries, but requires

careful setting of near-neighbor thresholds if clustering of diverse compounds or excessive

numbers of singletons (clusters of one) are to be avoided. Being a nonhierarchical method, a

ﬁ xed number of clusters are produced with no speciﬁ c relationships between the clusters.

In contrast, Ward’s agglomerative hierarchical clustering method

progressively com-

bines the most similar molecules into related groups, and the results can be viewed as

a dendrogram, similar to a family tree, showing the relationships between clusters and

allowing any number of clusters to be selected. Although slower, Ward’s method has been

shown to outperform Jarvis–Patrick and other clustering methods

and in our hands has

produced chemically intuitive results and proven very useful for the selection of reagent

sets. A limitation of clustering methods is that they give no information on the actual cover-

age of compound space by the library.

Bin-based partitioning methods, on the other hand, subdivide multidimensional chemi-

cal space along the axes into hypercubes or bins.

Compounds falling into the same bin

volumes are deemed to have similar chemical and, by assumption, biological properties.

Empty bins indicate areas of compound space not represented by the library under investi-

gation. Bin-based methods are useful for comparing libraries and for identifying “holes” in

a collection to be ﬁ lled by future libraries or compound acquisitions.

Whatever descriptors or algorithms are applied to the selection of a subset of com-

pounds from a virtual library, one additional criterion should be applied if the compounds

are to be synthesized efﬁ ciently by combinatorial methodologies, whether split-and-mix or

parallel array chemistry. It is not sufﬁ cient simply to select the most diverse subset of com-

pounds (called cherry picking); the compounds selected must be related through common

sets of building blocks. So, as an example, the problem is not to select the 96 most diverse

compounds but the most diverse 12 ⫻ 8 array of compounds. This problem is most easily

solved by selection of diverse reagent sets. However, it has been solved successfully at the

product level using a number of algorithms.

83,88–90

4.3.2 A More Focused Approach

The discussion so far has focused on designing libraries that give a good representation

of chemical space. We now consider what the bounds of this space are and two impor-

tant methodologies to limit the space that we desire to ﬁ ll. The ﬁ rst approach is to limit

chemical space to those property ranges that are likely to give druglike molecules; the

second is to restrict attention to structures likely to interact with speciﬁ c targets or target

classes. Like diversity, the term druglike has a range of deﬁ nitions, but essentially it is an

DESIGNING COMBINATORIAL LIBRARIES 139

140 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

attempt to confer those properties on a molecule that will increase its chances of being

orally absorbed, nontoxic, avoiding ﬁ rst-pass metabolism, and in an ideal world, passing

unhindered through costly development phases into clinical use.

The publication of Lipinski’s rule of ﬁ ve

in 1997 reawakened the awareness of some

chemists that the drive for large numbers of compounds was not the ultimate aim in drug

discovery. Lipinski analyzed a set of drugs known to be well absorbed and discovered that

the vast majority satisﬁ ed most of the following criteria: molecular weight (MW) below

500, calculated log P (c log P)

below 5, ﬁ ve or fewer H-bond donors and 10 or fewer

H-bond acceptors. Compounds failing more than one of these tests were predicted to be

poorly absorbed unless by some active transport mechanism.

Since this paper, there have been many other publications attempting to distinguish

drugs and nondrugs.

One very simple ﬁ lter to consider is the removal of compounds with

reactive functional groups (likely to be hydrolyzed or to react with biological nucleophiles)

or those known to contribute to toxicity.

Analysis of drug databases allows the frequency

of occurrence of molecular frameworks and side chains to be ranked and used as a guide

in library design.

93,94

Another approach is to use neural networks to perform the drug/non-

drug classiﬁ cation.

95,96

This technique has also been applied to the more speciﬁ c problem

of determining the features necessary for penetration of the blood–brain barrier, an essen-

tial requirement for central nervous system–active drugs.

The results of any of these studies should be considered guidelines rather than rules

since none of the methods are much more than 80% successful. Furthermore, the methods

are based on historic data, and it is unlikely that all future drugs will fall within their scope.

So although it makes sense that a large percentage of proposed library members should be

kept within the bounds of current wisdom, it would be foolish not to explore new areas to

some extent.

A further observation

is that the process of optimizing a lead is often associated with

an increase in molecular weight and lipophilicity. Hence, instead of designing lead gen-

eration libraries with druglike properties (typically, MW ⬎ 350, c log P ⬎ 3) which may

already be close to the limits allowed for these properties, attention should be focused on

leadlike property ranges (MW ⬍ 350, c log P ⬍ 3), leaving scope for further optimization.

Another factor to take into consideration is the demonstrated tendency of molecular com-

plexity to increase during the lead optimization process.

The second approach to chemical space limitation is in generic or focused library

design, where some degree of prior knowledge of structural requirements reduces the

chemical space of interest. A focused library is considered to be one directed toward a

speciﬁ c target receptor or enzyme. Its design may be de novo (based on knowledge of

receptor structure only), or based on a series of compounds with known SAR against

the speciﬁ c target (lead optimization library), or a combination of both. The former is

becoming increasingly applicable as the number of x-ray crystal structures and reliable

homology models of biomolecules grows. In structure-based design,

100–103

members of

a virtual library are screened in silico against their ability to bind to the active site of the

receptor, using programs such as DOCK,

104

GRID,

105

and LUDI.

106

An early example

was described by Kick et al.

107

in which the crystal structure of the pepstatin/cathepsin D

complex was used to reduce a virtual library of more than a billion statinelike compounds

to a set of 1000 for synthesis (see below). The performance of this set against a second li-

brary of 1000 optimized for diversity without consideration of the target structure showed

the former to give a signiﬁ cantly better hit rate against cathepsin D, especially in the high-

potency range.

The design of a generic library may use similar methods, but applied to a set of related

targets [e.g., aspartic proteases or G-protein-coupled receptors (GPCRs)]. Probably the

most common approach here is based on the concept of privileged structures,

108,109

deﬁ ned

as structural types that provide high-afﬁ nity ligands for more than one type of receptor or

enzyme. Libraries based around privileged structures are likely to give faster identiﬁ cation

of novel ligands for new members of known receptor or enzyme families. An example

given by Mason is to include the privileged structure as an additional “dummy” pharmaco-

phore in a four-point pharmacophore model.

A design criterion likely to receive increased attention is that of synthetic accessibility.

The RECAP technique

110

has been used to fragment biologically active molecules in a

retrosynthetic manner into building blocks rich in biologically recognized substructures.

However reagent sets may be chosen, their structural diversity will frequently also equate

to variations in their reactivity. A quantitative structure–property relationship (QSPR)

model has been applied to predict the reactivity of carboxylic acids toward the acylation of

an amine.

111

Classiﬁ cation,

112

modeling,

113

and simulation

114

of organic reactions will also

provide valuable predictive tools. At an experimental level, factorial design methods have

been applied to the optimization of reaction conditions for library synthesis.

115,116

4.4 TOOLS FOR SYNTHESIS OF COMBINATORIAL LIBRARIES

In the preceding section we focused on deﬁ ning the compounds to be synthesized in a

library and how the design criteria were matched to the desired function of the library. We

now consider the choice of tools available to facilitate library synthesis and how these, too,

must be matched to the type of library and the environment in which it is to be synthesized.

The equipment may achieve its goal of increased productivity per chemist in a number of

ways. At one end of the spectrum one ﬁ nds relatively simple, low-cost apparatus for manu-

ally performing operations in parallel. At the other extreme, sophisticated, fully automated

synthesizers offer round-the-clock unattended operation.

117

Several factors must be consid-

ered when selecting the most appropriate platform: suitability for solution or solid-phase

chemistry, throughput, modular or self-contained system, synthesis scale, footprint, ease

of use, cost, and chemical performance. Representative examples of currently available

hardware are assembled in Table 4.5. For a more detailed discussion, the reader is referred

to review articles in Table 4.7.

4.4.1 Nonautomated Tools

Perhaps the simplest devices for increasing synthetic throughput are those designed to

support, heat, and stir a number (6 to 24) of round-bottomed ﬂ asks, bottles, or test tubes

simultaneously. These low-cost devices occupy a small footprint and thus are suitable for

use alongside conventional equipment in a standard fumehood and are applicable to a range

of reaction scales from a few milligrams to several grams.

Manual systems of somewhat higher throughput are the reactor blocks based on the

ubiquitous microtiter plate format so well known in biological laboratories. Throughput

ranges from 96 down to 12 compounds per block, corresponding to 2 to 16 mL well vol-

umes. Blocks may be made of polypropylene, PTFE, or glass, they can be sealed, and

they often incorporate sintered frits to allow separation of soluble products from supported

reagents and sequestering agents. Agitation is usually by means of vibrating or oscillating

TOOLS FOR SYNTHESIS OF COMBINATORIAL LIBRARIES 141

142 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

Application Examples URLs

Stirring blocks (with

heating or cooling)

Carousel, greenhouse www.radleys.com

Reacto-stations www.stemcorp.com

KEM-Prep, reaction blocks www.jkem.com

Reactor blocks MiniBlock www.bohdan.com

Calypso system www.charybtech.com

FlexChem system www.robsci.com

Liquid handling Microlab, Star workstations www.hamilton.com

MiniPrep, Genesis www.tecan-us.com

Quad-Z 215 www.gilson.com

Mix-and-sort AccuTag-100 www.irori.com

SynPhase crowns and lanterns www.mimotopes.com

XYZ synthesizers ACT 496, Benchmark www.advancedchemtech.com

Myriad Discoverer, Myriad MCS www.mtmyriad.com

Sophas www.zinsser-analytic.com

Neptune www.bohdan.com

Automated synthesis workstations www.chemspeed.com

Fluidic synthesizers Nautilus www.argotech.com

Domino blocks www.torviq.com

Quest www.argotech.com

Other synthesizers Zymate www.zymark.com

Trident automated synthesizer www.argotech.com

Microwave synthesizers Smith synthesizer, Creator www.personalchemistry.com

Discover Systems www.cem.com

Ethos SYNTH www.milestonesci.com

Delivery of solids Redi, Lipos www.zinsser-analytic.com

Titan resin loader www.radleys.com

Resin and powder dispenser system www.anachem.co.uk

Reaction workup ALLEX www.mtmyriad.com

Trident sample processing station www.argotech.com

Lollipop www.radleys.com

Evaporation SpeedVac www.thermo.com

HT systems, megasystems www.genevac.com

Syncore PolyVap www.buchi.com

TurboVap www.zymark.com

IR-Dancer www.hettlab.ch

ALPHA-RVC, BETA-RVC www.matinchrist.de

Flash chromatography Flashmaster www.joneschrom.com

Quad3 www.biotage.com

CombiFlash, Optix 10 www.isco.com

Preparative HPLC Flex/Parallex www.biotage.com

FractionLynx autopuriﬁ cation

system

www.waters.com

1100 Series puriﬁ cation system www.chem.agilent.com

Discovery VP preparative system www.shimadzu.com

High-throughput puriﬁ er www.hii.hitachi.com

Weighing AWS www.bohdan.com

Calli, Moss www.zinsser-analytic.com

TABLE 4.5 Selection of Commercially Available Equipment

for High-Throughput Chemistry

shakers. In some cases, temperature control may be achieved by circulating cooled or

heated ﬂ uids through channels in the block assemblage. Multichannel handheld pipettes

facilitate reagent delivery, but this can be enhanced by the use of simple liquid-handling

robots. The Diversomer technology of Parke-Davis was interfaced to an XYZ robot for

liquid delivery.

118

4.4.2 Mix-and-Sort Systems

Interesting and widely used applications of the teabag approach to split-and-mix synthesis

have been developed by Irori and Mimotopes. Although applicable to small arrays, the

real power of these technologies lie in multistep syntheses of hundreds or thousands of

products. The Irori AccuTag-100 system comprises small mesh-sided microreactors that

the user loads with solid-phase resin and a radio-frequency (RF) tag.

119

A range of micro-

reactors can accommodate from approximately 30 mg of resin (which yields about 10 to

20 mg of product on cleavage) up to 10 times that quantity. Prior to the ﬁ rst synthetic step,

the microreactors are scanned and the unique code of the RF tag is assigned to one of the

library components. This code is used to sort the microreactors before each synthetic step

and enables a mix-and-sort strategy requiring exactly one microreactor for each planned

product. Chemistry is performed in conventional labware (ﬂ asks, bottles, etc.) and culmi-

nates in an archival process where the microreactors are sorted into a 96-well plate format

prior to cleavage, again directed by the RF tag codes. Scanning and sorting can be manual

or by using an automated tag reader/sorter. Hence the synthetically efﬁ cient mix-and-sort

process is combined with the production of useful quantities of discrete cleaved products.

The Mimotope SynPhase system circumvents the use of resin beads by using linker

groups specially grafted onto easily handled support structures, with loadings from 4 to

35 µmol and again capable of delivering milligram quantities of products. In addition to RF

tagging strategies similar to the Irori system, product identiﬁ cation can be achieved by the

use of color coding or by creating spatially addressed arrays in 96-well MTP format.

4.4.3 Automated Synthesizers

Automation of synthesis comes at a higher cost than the manual tools described above,

and the equipment typically ﬁ lls a standard fumehood and may even require a custom-built

installation. The advantages include the possibility of continuous, unattended operation,

reproducibility, and reduced exposure of operators to toxic chemicals. The ﬁ rst successful

automation was for solid-phase peptide chemistry, which is particularly well suited since

only two reactions (coupling and deprotection) are involved, both having been highly

optimized to give quantitative yields at room temperature. Moreover, a limited set of well-

characterized amino acid building blocks are used, and standardized deprotection, cou-

pling, and wash cycles are easily programmed.

Subsequently, modiﬁ ed peptide synthesizers were marketed for general solid-phase

organic chemistry where fewer consecutive steps, but much more variation in reaction

conditions and reagent, are the norm. The addition of slurry movement made split-and-mix

synthesis possible and heating–cooling capabilities, as well as improved provision of inert

atmospheres, extended the range of chemistry possible. Recently, the ﬁ rst fully automated

oligosaccharide synthesis was reported.

120

Several types of automated synthesizers have evolved. The most commonplace are those

based around the XYZ liquid handler. These are capable of delivering reagent solutions to

TOOLS FOR SYNTHESIS OF COMBINATORIAL LIBRARIES 143

144 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

a grid of reaction vessels or a reactor block, and can also be programmed to move solutions

from one vessel to another. They may have multiple probes and in most cases are sealed

by a septum on the reaction vessel, which is pierced to deliver reagents and maintains an

inert atmosphere. Alternatively, a twist cap mechanism may be used, which avoids the need

for small-diameter needles for reagent delivery, allowing the use of wider-bore positive-

displacement pipettes.

Using ﬂ uidic systems, where reagents and solvents ﬂ ow from reservoirs, through a per-

manent network of tubes and valves, to the reactors can largely eliminate moving parts.

Because such systems are entirely closed, control of atmosphere is straightforward, but the

pipework can be complex and may be prone to blockages if suspensions or poorly soluble

reagents are used. Other systems are available that use a multiarticulated robotic arm to

move vessels from one workstation to another, much as a human operator might.

A multistep synthesis need not be constrained to a single synthesizer. As an alternative

to truly parallel synthesis, in which every step is performed separately for every product,

a split–split strategy has been proposed

121

in which each diversity step is carried out only

once, and the products are split for subsequent diversity steps. Because the reaction scale

decreases at each stage, the use of a range of synthesizers is necessary to match both the

scale and number of vessels required for each step.

A recent development is the use of microwaves as a source of heat energy for accel-

erating chemical reactions.

122

Reactions that might ordinarily take many hours can be

completed in a few minutes. Hence, increased throughput is by way of time compression

rather than parallel operation. Early systems were based on modiﬁ ed domestic microwave

ovens and suffered from uneven and unpredictable heating. Recent systems have purpose-

designed single-vessel cavities where the power is controlled and delivered uniformly, and

allow real-time monitoring of both temperature and pressure, resulting in reproducible

results. Coupling such a reactor with simple robotics enables multiple reactions to be run

unattended.

Delivery of soluble reagents to any of these synthesis systems requires simple liquid-

handling robots, either stand alone or, more often, integrated with the synthesizer. Delivery

of solids, in particular synthesis resins, polymer-supported reagents, and scavengers is less

straightforward, although robotic dispensers have been developed. The use of isopycnic

suspensions in appropriate solvent mixes facilitates pipetting and is the technique used

in some automated synthesizers. The use of resins prepackaged in capsules

123

or in tablet

form

124

is attractive but suffers from the need to dissolve away the capsule or binder mate-

rial. This is inappropriate for the supported reagents and scavengers of solution chemistry,

where additional impurities would be introduced, but may be suitable for distribution of

resins for solid-phase chemistry. A recent publication introduced sintered resin blocks,

which are proposed as an alternative to microreactors or crowns but might also ﬁ nd use

in delivering reagents and scavengers.

125

Monolithic disks, which can be crafted in a va-

riety of shapes and sizes, offer an alternative means of conveniently handling solid sup-

ports.

126–128

Resin dispensers based on simultaneous volumetric measurement and delivery

of 96 aliquots of resin to a MTP format have been developed in our own laboratories and

commercialized by others.

4.4.4 Postsynthesis Processing

Even modest increases in synthesis throughput have a knock-on effect on downstream

sample processing. Routine events such as reaction workup, solvent evaporation, product

puriﬁ cation, quantiﬁ cation, and analysis all require high-throughput counterparts. As stated

previously, reaction workup of solid-phase syntheses involves simple ﬁ ltration. In a similar

way, supported reagents frequently facilitate the workup of solution-phase chemistry, and

scavenger resins remove excess reagents. Conventional liquid–liquid extraction procedures

have also been automated. The interface between the two liquid layers may be located by

dead reckoning of solvent volumes, or proprietary interface detection technologies can be

used. A simple but elegant concept is to freeze the aqueous layer and separate it from the

liquid organic layer.

High-throughput solvent evaporation is facilitated by vacuum centrifuges, which are

available in many sizes and can accommodate tens or hundreds of tubes or MTPs in a range

of formats. An alternative means to prevent bumping under vacuum is rapid agitation. Par-

allel evaporation can also be achieved by a suitably angled gas ﬂ ow over the solvent sur-

face, although some means of solvent capture (condensation) is also required.

A variety of tools are available for parallel puriﬁ cation. Solid-phase extraction (SPE),

in which product solutions are passed through plugs of sorbent (in ﬁ lter plates or syringe

barrels), is in widespread use.

129

In the simplest strategy, the sorbent retains the impurities

and the product is collected. Alternatively, the product is retained on the ﬁ rst pass and sub-

sequently eluted by a change in solvent. SPE is often performed in direct conjunction with

reactor blocks but has also been automated.

130

In many instances, true chromatographic puriﬁ cation methods are necessary, and when

gram quantities of material are involved, ﬂ ash chromatography systems are appropriate.

These allow automated parallel or serial chromatography of 10 or 12 samples and may

have ultraviolet (UV) detectors for recording or triggering fractions. In our hands these

systems have proven very useful for purifying building blocks and other intermediates.

Smaller-scale puriﬁ cations are most frequently effected by preparative high-performance

liquid chromatography (HPLC). A number of automated systems are commercially avail-

able, sometimes operating multiple columns in parallel to boost throughput. Depending

on the system, fractionation may be triggered by UV

131

or mass detection.

132,133

The latter

is appealing in that only a single fraction needs to be collected per product (i.e., that with

the correct MW). However, error recovery is impossible unless separate waste streams are

collected for each sample. In contrast, the more robust method of collecting all signiﬁ cant

peaks by UV requires downstream structure conﬁ rmation as well as effective fraction man-

agement.

134

Supercritical ﬂ uid chromatography (SFC) using pressurized carbon dioxide as

the major eluent offers the advantage of small fraction volumes since the eluent evaporates

instantly and avoids the costs and environmental issues of high solvent use.

135

If the products are collected in preweighed tubes or vials, the amount of product may

be quantiﬁ ed simply by weighing on an automated weighing station. This is not possible if

products are collected directly into MTPs. Alternative techniques such as evaporative light-

scattering detection (ELSD)

136,137

and chemiluminescent nitrogen detection (CLND)

138,139

are becoming popular, especially when used in conjunction with analytical HPLC or

HPLC-MS. The latter is the most widely used technique for structure conﬁ rmation, and

high throughput has been achieved by combining four or eight LC columns with a single

multiplexed mass spectrometer.

140

Multiple MS or HPLC-MS results can be evaluated rap-

idly using color-coded plate maps and other visualization tools.

141,142

Flow nuclear magnetic resonance (NMR) systems are making high-throughput NMR

analysis of products in MTP format routine

143

and deliver much more structural informa-

tion than MS techniques. However, interpretation of the many spectra produced is currently

a bottleneck that will only be resolved as spectrum prediction software improves. The use

TOOLS FOR SYNTHESIS OF COMBINATORIAL LIBRARIES 145

146 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

of internal standards such as 2,5-dimethylfuran

144

allows quantiﬁ cation estimates to be

made at the same time.

In summary, recent advances in automation and parallel processing of virtually all pro-

cedures in synthesis, puriﬁ cation, and analysis have rendered the high-throughput produc-

tion of high-quality compound libraries a reality. Of course, as each advance is made,

expectations rise, so there will doubtless be signiﬁ cant further developments in the future.

4.5 MANAGING THE COMBINATORIAL PROCESS

An inescapable consequence of the combinatorial chemistry revolution is the increased

data-handling burden compared to traditional methods. A single chemist might be expected

to design and synthesize thousands of discrete compounds in a year in contrast to 50 or

so previously. Each library of compounds must be designed, the building blocks acquired,

reaction quantities, molecular weights, and yields of products calculated, analytical results

interpreted, products registered into databases, and so on. Additionally, the range of equip-

ment, often robotic, in use has grown, each piece having its own operational and data

input–output requirements. Furthermore, depending on circumstances, a single chemist

might handle a library through all the stages of production, or increasingly commonly, the

work might be shared between a group of chemists working as a team who each need ac-

cess to data at different stages.

4.5.1 Speciﬁ cation of Combinatorial Libraries

Within the author’s laboratories, the impending complexities of data and process manage-

ment issues were recognized around 1996. Analysis of the repeated transfer or reentry of

data between a variety of electronic and hardcopy formats revealed signiﬁ cant potential

sources of error and tedium (time wasting). In a collaboration between medicinal chem-

ists, computational chemists, analysts, and the newly formed Cheminformatics group,

an integrated software application, RADICAL (Registration, Analysis and Design Inter-

face for Combinatorial and Array Libraries), was developed and deployed throughout the

organization.

The scope of RADICAL has been extended continually through incremental releases

in order to accommodate changes to the processes and equipment in use. A further advan-

tage of having a standard, but ﬂ exible, library management tool is that all data generated

during library production is automatically captured for archival. Of particular value is the

capture of synthetic routes used and their outcome, which offers great potential as a search-

able source of information for subsequent library route development. In our experience,

compliance with documentation and data logging is much higher if they form part of the

natural workﬂ ow rather than being additional tasks. Some key features of the RADICAL

environment are outlined in Table 4.6. Related systems, such as CICLOPS,

145

ADEPT,

and AIDD,

146

have been developed by other pharmaceutical companies; other packages,

such as CHEM-X

147

and Afferent,

148

are available commercially.

4.5.2 Controlling the Automated Workﬂ ow

One disadvantage of developing a single, albeit component-based application to support

high-throughput synthesis was that a new software release was required to incorporate

frequent changes necessitated by the purchase and modiﬁ cation of commercial hardware

such as synthesizers, analytical instruments, and balances. Also, the variety of equipment

in use required that data exchange be performed at the lowest common denominator of text

ﬁ les, resulting in the users having to exchange ﬁ les and become familiar with a number of

interfaces on the various pieces of hardware. A particular limitation was the lack of support

for tracking physical racks or plates of compounds around the laboratory. The decision was

therefore made to separate the deﬁ nition of chemistry speciﬁ cation and other structurally

related matters (RADICAL’s role) from the hardware, data, and process management-

dependent issues. The requirements for the latter role were to be achieved through code-

velopment of a separate but integrated software application, ACE (Automated Chemistry

Environment) with the Technology Partnership, Cambridge, UK.

The role of ACE is to deﬁ ne and manage the workﬂ ow that the compounds deﬁ ned in

RADICAL will follow during their synthesis, analysis, and puriﬁ cation in the laboratory.

The library deﬁ nition from RADICAL is mapped onto an ACE workﬂ ow which is stored

on the ACE manager (a server application with a database containing knowledge of all

compounds in process and their associated data), which is networked to the PCs controlling

the automated equipment around the laboratory. The operating software of each piece of

equipment is concealed from the user within an agent, which acts as an interface between

the ACE manager and the equipment. All agents have a similar appearance, which includes

a listing (queue) of all jobs currently available to run on that machine, so users only have

to be familiar with one interface type. Wherever possible, the agent automatically provides

the equipment with the necessary data and parameters to run the current job, then guides

the user through setup or loading procedures, runs the equipment, and returns any data col-

lected to the ACE manager. As a set of compounds (e.g., a rack of vials, a microtiter plate)

completes one process within its workﬂ ow (e.g., synthesis robot) the “job” will appear on

the queue of the agent of the next process (e.g., workup robot). All racks, plates, and vials

are bar-coded, which allows conﬁ rmation that the correct job is being processed and facili-

tates the identiﬁ cation of any sample within the laboratory.

ACE provides agents for both automated and manual steps, the latter often simply pre-

senting a means of indicating that the job is complete. Additionally, “decision” agents have

been developed to allow the user to inspect analytical data and indicate whether the com-

pound is of sufﬁ cient quality and/or quantity to progress. When the workﬂ ow is complete,

the appropriate data collected by ACE is transferred back to RADICAL for archival and

registration into the corporate database. The advantage of this approach is that whenever

•

Chemistry editor Deﬁ nition of synthetic route

Deﬁ ne steps, reactants, products, conditions.

•

Library design Deﬁ nition of building blocks

Select reagents from databases (commercial and

in-house), links to library design packages.

Acquire reagents (suppliers, in-house inventory).

Calculate reagent quantities.

•

Library results Enumerated products matrix

View and enter compound-speciﬁ c data.

Track status of compounds.

Link to registration and screening.

•

Plate manager Management of subsets of

library

Map products to plates.

Submit to quality control and report results

using color mapping.

TABLE 4.6 Key Components of RADICAL

MANAGING THE COMBINATORIAL PROCESS 147

148 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

a new piece of hardware is incorporated into the process, a relatively small change to the

software environment is required.

However good the tools provided to aid in the management of combinatorial chemistry,

whether for library design or production, the overall complexity of the process should not

be trivialized. When errors occur, recovery is much more complicated than in traditional

medicinal chemistry environments, if only because of the number of entities involved. We

implement regular appraisal of processes, seeking opportunities for further streamlining

and error reduction. Furthermore, it is important to provide adequate support for chemists,

in terms of training, assistance in error recovery, and capturing feedback regarding future

enhancements to the systems.

4.6 FROM SPECIALIST DISCIPLINE TO STANDARD TOOL

The following general remarks are based on published commentaries,

149–151

discussions

at conferences, and our own experiences. In the early 1990s, the potential of the newly

developing discipline of combinatorial chemistry to accelerate, and indeed revolutionize,

the drug discovery process was recognized throughout the pharmaceutical industry. Most

major companies established small groups of interested persons to investigate this potential

and bring the necessary tools in-house. The initial remit of such groups was typically to

review and develop new technologies and apply these to the synthesis of lead generation

libraries to increase corporate screening collections. This was driven in part by the newly

developed high-throughput screening capabilities,

152

with capacities to screen many more

compounds than were available at the time.

Within this fairly generic model, each company had its own approach. Differences

included the decision whether to use solid-phase or solution chemistry (or both) as their

major platform and whether to rely on commercially available technologies or develop

proprietary methods. A further choice was necessary between preparing discrete products

or the synthetically more attractive mixtures.

Additionally, an increasingly important part of the combinatorial groups’ brief was to

move the technologies out into the mainstream medicinal chemistry population for applica-

tion to lead optimization. In this respect a number of issues made the uptake less rapid than

anticipated. The scope of documented chemistry suitable for high-throughput synthesis

was restrictive, and in the case of solid-phase chemistry, the lengthy route development

times were much less acceptable for the optimization phase, where time and resources

were often at a premium. Many of the techniques developed, such as mix and split and en-

coding, offered the greatest efﬁ ciencies when applied to large libraries, whereas optimiza-

tion libraries were much smaller. The equipment associated with combinatorial chemistry

tended to be too bulky to accommodate readily in standard fumehoods and was operated

via computer interfaces that required a greater degree of familiarity than would be acquired

by occasional use outside the specialist groups.

Many of these hurdles have subsequently been overcome to a large extent, as a result of

changes to technology, processes, and mindsets. The recent explosion in the range of solid-

supported reagents and scavengers has had a dramatic inﬂ uence. The workup of many

types of solution-phase reactions can now be simpliﬁ ed by the application of these prin-

ciples. Interestingly, these ideas are also ﬁ nding imaginative applications in mainstream

multistep organic synthesis.

The restrictions in the scope of solid-phase chemistry are now far fewer than in the

pioneering days, when “any compound as long as it’s an amide” was the perceived dogma.

The number of publications in the ﬁ eld continues to grow, and considerable databases of

knowledge have been accumulated both publicly and within companies. Also, the inability

to purify compounds at intermediate stages of multistep solid-phase syntheses contrib-

uted to the need for long development phases in route optimization. With the advent of

high-throughput puriﬁ cation techniques, recovery of products from suboptimal synthetic

sequences became possible, and route development times could be compressed signiﬁ -

cantly. Automated puriﬁ cation of ﬁ nal products as well as intermediates obviously also

beneﬁ ts solution-phase protocols.

A further change has been the development of automated systems better suited to

the rigors of general organic chemistry. Careful control of reaction conditions, heating

and cooling, and maintenance of inert atmospheres have been signiﬁ cant improvements.

Appropriate system size has also been addressed; a number of systems designed for spe-

cialist high-throughput environments have undergone modiﬁ cation to provide related sys-

tems for the more modest needs of medicinal chemistry laboratories, and their control

systems are becoming more straightforward to use. A number of tools that rely less on

automation and more on parallel manual processing are very popular with chemists wish-

ing to perform 10 to 100 simultaneous reactions. The need for tools to support the process

downstream of synthesis has also been addressed. We have modiﬁ ed our own laboratories

to include automation workstations including synthesis, workup, evaporation, and puriﬁ ca-

tion capabilities.

However well suited the equipment, an essential component in successful implementa-

tion of high-throughput techniques in medicinal chemistry is training and support. In our

own laboratories, a short intensive training program was made available to 48 chemists

per year. This involved a simple one-step solution-phase synthesis of 192 compounds.

Over a period of two to three days, chemists were introduced to several pieces of robotic

and ancillary equipment as well as process philosophies, analytical tools, and the previ-

ously mentioned RADICAL application. As well as providing training, the compounds

produced have augmented the corporate compound collection and have produced several

interesting leads from high-throughput screening. In addition to this short familiariza-

tion exercise, a number of medicinal chemists have joined the high-throughput chem-

istry group for a more prolonged exposure to the technologies in a series of six-month

secondments.

4.7 APPLICATION OF COMBINATORIAL CHEMISTRY

IN DRUG DISCOVERY

Comprehensive surveys

153–156

indicate the large number of combinatorial libraries that

have been documented in the literature, many of which are accompanied by biological

data. Furthermore, publications describing the efforts of medicinal chemistry groups regu-

larly include reference to part of the work beneﬁ ting from rapid compound synthesis using

parallel or other combinatorial methods, and recent successful applications have been re-

viewed.

157

Four case histories are presented below to illustrate some of the ways in which

high-throughput methodologies are contributing to a more rapid identiﬁ cation and optimi-

zation of biologically active compounds.

APPLICATION OF COMBINATORIAL CHEMISTRY IN DRUG DISCOVERY 149

150 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

4.7.1 Case History 1

The ﬁ rst example illustrates the successful application of high-throughput screening of a

diverse compound collection followed by lead optimization using parallel synthesis tech-

niques.

158

The distribution of the 5-HT

receptor in the brain, in association with its high

afﬁ nity for a range of drugs used in psychiatry, made the search for selective antagonists

of this G-protein-coupled receptor an attractive proposition. The excellent afﬁ nity of bisa-

ryl sulfonamide (1; Scheme 4.1) for the 5-HT

receptor (pK

8.3) was discovered follow-

ing high-throughput screening of the SmithKline Beecham compound bank. Subsequent

studies demonstrated the compound to have no appreciable afﬁ nity for a range of over 50

receptors, enzymes, and ion channels and to be moderately brain penetrant (25%) but with

low oral availability (12%) as a result of rapid blood clearance. Using parallel solution

synthesis techniques (Scheme 4.1) a series of analogs were rapidly prepared to investi-

gate SAR around this lead compound, culminating in the identiﬁ cation of benzothiophene

derivative (2) with subnanomolar 5-HT

receptor afﬁ nity. The compound demonstrated

improved selectivity, reduced blood clearance, but was metabolically N-dealkylated to the

piperazine (3). Subsequent synthesis and testing of 3 showed that it maintained high af-

ﬁ nity (pK

8.9) and selectivity. Furthermore, pharmakokinetic studies in rats demonstrated

moderate brain penetrancy (10%), low blood clearance, and, importantly, excellent oral

bioavailability (⬎80%). The piperazine (3) has entered phase I clinical trials for the treat-

ment of cognitive disorders.

159

4.7.2 Case History 2

The second example illustrates the successful iterative deconvolution of a combinatorial

(mixture) library based around a compound discovered by database searching of a com-

pound collection.

160

Somatostatin is a widely distributed tetradecapeptide whose variety

OMe

O O

OMe

O O

2 R = Me

3 R = H

OMe

O O

(i)

(i) ArSO

Cl, acetone, rt, 18 h

Scheme 4.1

of actions are mediated through high-afﬁ nity membrane-associated receptors (GPCRs).

Five human somatostatin receptors (hSSTR1–5) have been cloned and characterized, for

which a number of moderately subtype-selective petide agonists have been identiﬁ ed. Us-

ing a model based on the conformation of cyclic hexapeptide agonist c(Pro-Tyr-D-Trp-

Lys-Thr-Phe), a three-dimensional similarity search of the Merck compound sample col-

lection identiﬁ ed 75 compounds for screening, of which 4 had high afﬁ nity (K

⫽ 200 nM)

in the mouse SSTR2 assay. A library was designed around this structure, synthesized by a

split-and-mix strategy according to Scheme 4.2 from 20 diamines, 20 amino acids, and 79

amines. Following the introduction of diamine and amino acid building blocks, portions of

resin were archived in readiness for the deconvolution process. When stereo- and regioi-

somers were taken into account, the library consisted of 131,670 entities in 79 pools of

1330 or 2660 members. The ﬁ rst deconvolution experiments focused on hSSTR2, against

which, reassuringly, the most potent mixture was the one containing compound 4. A sec-

ond, equipotent mixture was also deconvoluted and yielded the very potent (K

⫽ 0.04 nM)

and selective benzimidazolonylpiperidine (5). In a similar manner, subtype selective com-

pounds for hSSTR1, hSSTR4, and hSSTR5 were discovered. The experiments also fur-

nished useful structure–activity relationships that were subsequently exploited in further

library designs.

161

4.7.3 Case History 3

The next example demonstrates the application of combinatorial chemistry to the discov-

ery and lead optimization phases using both solid-phase

162

and solution-phase

163

chemis-

try. P-glycoprotein (Pgp) provides an active efﬂ ux mechanism for the clearance of toxic

substances. The overexpression of Pgp in cancer cells is responsible for the intrinsic or

acquired immunity of tumor cells to a wide range of chemotherapeutic agents, a phenom-

enon known as multidrug resistance (MDR). Unfortunately, toxicity issues have limited the

clinical use of compounds known to desensitize resistant tumor cells. Using solid-phase

-NH

FmocNH

-NH

Scheme 4.2

APPLICATION OF COMBINATORIAL CHEMISTRY IN DRUG DISCOVERY 151

152 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

synthesis, exempliﬁ ed in Scheme 4.3, a library of 500 imidazoles was prepared, with char-

acteristics typical of known Pgp modulators and substrates: namely, hydrophobicity and

multiple amine groups. The SAR discovered from screening this library, followed by some

preliminary solution-phase optimization, showed the cinnamic methyl ester (6), and sub-

sequently, the less metabolically liable ether (7) to have good activities (ED

⫽ 300 and

90 nM, respectively). While these modulators had excellent potency against a variety of re-

sistant cell lines, pharmacokinetic measurements indicated that the dimethylamino groups

were rapidly metabolized. Using solution-phase combinatorial chemistry (Scheme 4.4) a

range of alternative aminoaryl groups were investigated as 3- and 4-substituents on the

imidazole ring, of which the bis-isopropylamino compound (8) proved most potent and had

a good half-life in the dog (t

1/2

2.79 h). This compound has been progressed as a clinical

candidate and has shown virtually no adverse side effects in healthy male volunteers.

4.7.4 Case History 4

The ﬁ nal example was mentioned brieﬂ y in the library design section and demonstrates

the power of coupling combinatorial chemistry with structure-based design.

107

A second

library based on diversity considerations only was prepared as a control. The goal was to

develop small-molecule nonpeptidic inhibitors of cathepsin D, an aspartyl protease im-

plicated in tumor metastasis and Alzheimer’s disease, based on a stable mimetic of the

tetrahedral intermediate of peptide hydrolysis by this enzyme. The basic premise was to

construct molecules of type 9 using amines (R

) and acylating agents (R

, R

) as building

blocks to decorate the well-known (hydroxyethyl)amine isostere, according to the discon-

nection in Scheme 4.5. A search for suitable commercially available building blocks re-

vealed around 700 amines and 1900 acylating agents, which in combination would give

rise to more than a billion products. Two strategies were used to reduce this number to two

libraries of 1000 components each.

In the directed library design, the scaffold was modeled into the enzyme active site.

Each of the building blocks was added to the scaffold independently, and a full range

of conformations were assessed for intramolecular clashes with the scaffold and overlap

with cathepsin D, and the 50 best scoring building blocks at each position were retained.

Following removal of high-cost reagents, the remainder were clustered, and three sets of

10 building blocks were selected from unique clusters. In the diverse library design, the

original building block lists were simply clustered using the Jarvis–Patrick algorithm,

Wang

CHO

Wang

(a)R

-NH

, NH

OAc, R

COCOR

(b) 20% TFA, DCM

Scheme 4.3 Solid-phase synthesis.

ONos

POL

POL = Polystyrene resin with tetrahydropyran linker

Scheme 4.5

Scheme 4.4 Solution-phase synthesis.

(6) R = CH=CHCO

Me, R' = NMe

(7) R = CH=CHCH

OEt, R' = Me

(8) R = CH=CHCH

OEt, R' = i-PrNH

(a) R

NH, K

, DMSO, 90°C

(b) R

NH, K

, DMSO, 90°C

CHO, NH

OAc, 100°C

a, b, c

APPLICATION OF COMBINATORIAL CHEMISTRY IN DRUG DISCOVERY 153

154 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

and diverse sets of 10 were selected, with additional consideration of cost and balance of

functional groups in each set. Both libraries were prepared using parallel solid-phase syn-

thesis methods in a 96-well ﬁ lter apparatus. When screened against cathepsin D at 1 µM

concentration, 67 compounds from the directed library showed inhibitions greater than

50%, compared to 26 compounds from the diverse library. At 100 nM, the corresponding

ﬁ gures were 7 (directed, including 10 with K

⫽ 73 nM) and 1 (diverse). The authors con-

cluded that computational methods could be applied effectively to the reduction of large

virtual libraries to practical numbers for synthesis and evaluation. In the absence of target

information, however, diverse libraries remain an important strategy.

4.8 THE FUTURE OF COMBINATORIAL CHEMISTRY

The past decade has seen a number of developments, many of which have passed through

several generations of evolution. Some, at ﬁ rst exciting and promising a great deal, have

been replaced by alternatives, and overall strategies have been reviewed in light of the

available technologies. Doubtless, much of what has been described above will continue

to evolve with incremental improvements contributing to further ease of use, reliability,

reproducibility, elimination of process bottlenecks, faster cycle times, improved library

design processes (as our knowledge base expands), a greater range of suitable chemistry,

and so on.

164

The speciﬁ c roles of combinatorial chemistry may well evolve as information

derived from genomics leads to new strategies.

165

It is probable that many high-throughput

operations will move away from the conventional laboratory environment to a “drug dis-

covery factory” employing industrial-scale engineering.

117,166,167

Further lessons may be

learned from nature, with more libraries being designed around natural products,

168,169

by employing biocatalysts in the execution of their synthesis.

170

It is also likely that there

will be some surprises as new technologies are developed that shift the paradigm signiﬁ -

cantly. Two such ideas that are already emerging are outlined below.

4.8.1 Dynamic Combinatorial Libraries

A concept that has developed during recent years is that of dynamic combinatorial libraries

(DCLs).

171–173

A DCL is an equilibrating pool of reagents and products (e.g., amines,

aldehydes, imines) in which the products are not intended to be stable and isolable, but

rather, continuously dissociate and re-form in different combinations. The idea is that in

the presence of a receptor, the combinations that have the highest afﬁ nity will be formed

preferentially, and their concentration will increase relative to other combinations. Hence,

library generation and screening are combined in one process. Although the principle has

been elegantly demonstrated in a number of simple systems, there are obstacles to over-

come before this approach achieves sufﬁ cient generality to make a signiﬁ cant contribution

to drug discovery.

4.8.2 Miniaturization

Some degree of miniaturization of chemistry is already evident—the very idea of working

on a scale of 1 to 2 mL and routinely generating only a few milligrams of product would

have seemed unlikely a decade or two ago. High-throughput screening is already moving

to much smaller volumes; 384-, 1536- and even 3456- well plates occupy the same foot-

print as do traditional 96-well MTPs but use much smaller quantities of scarce biological

reagents.

174

Of course, there is also a concomitant reduction in the quantity of test com-

pound needed. Indeed, some ﬂ uorescence detection methods give measurable responses

from single molecular interactions.

152

The very basis of split-and-mix procedures relies on synthesis beads acting as tiny reac-

tors. However, even if beads are segregated for screening purposes, they are handled in bulk

quantities during synthesis using conventional labware. There are a number of physical

hurdles encountered in moving to signiﬁ cantly smaller synthesis scales. Increased surface

area relative to volume exacerbates moisture exclusion and evaporation problems, and cap-

illary action and surface tension become dominant forces. Hence, microscale synthesis

is unlikely to be performed in open wells, and a move toward enclosed ﬂ uidic systems is

probable.

Fabrication techniques such as photolithography, chemical etching, and laser micro-

forming have been applied to create networks of channels (typically, 50 to 300 µm) and

reactors in materials such as silicon, quartz, glass, and plastic, allowing the manufacture

of microﬂ uidic devices.

175

Applied initially to chromatographic and analytical devices,

176

these techniques have already been used to construct microﬂ uidic devices capable of

performing solid-phase or solution chemistry.

177,178

Advantages offered by miniaturized

systems include exquisite control of reaction conditions, low reagent and solvent

consumption, and potential integration of synthesis, analysis, and screening into a single

device.

4.9 CONCLUSIONS

Against an ever-changing scientiﬁ c, economic, and political backdrop, medicinal chem-

ists must continue to rise to the challenge of developing better, safer, more effective drugs

more quickly and more economically than ever before, using whatever technologies are

most appropriate. The concepts and technologies of combinatorial chemistry described

in this chapter will doubtless contribute greatly toward this goal, as they are incorpo-

rated into everyday working practices. Perhaps as important as providing new tools, the

development of combinatorial chemistry has opened chemists’ minds to the possibilities

of searching for further innovative means of attaining their goal. No single technology

is likely to be a universal panacea, however convincingly it may be marketed. Medicinal

chemists must continue to use judgment and scientiﬁ c reasoning to assess the possible

contributions new technologies might make, and how best to apply them. Similarly, manu-

facturers must continue to focus on the needs of the chemists. The ﬁ eld of combinato-

rial chemistry has undergone a rapid development in a relatively short period of time

and revolutionized the modus operandi of the medicinal chemist. In whatever directions

drug discovery may evolve in the future, we should anticipate combinatorial chemistry, in

one form or another, to make signiﬁ cant further impact through many new and exciting

developments.

In addition to the leading references provided throughout the chapter, a selection of

recent review articles and books are assembled in Table 4.7 to aid the interested reader in

exploring the subject further.

CONCLUSIONS 155

156 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

REFERENCES

1. Maclean, D.; Baldwin, J. J.; Ivanov, V. T.; Kato, Y.; Shaw, A.; Schneider, P.; Gordon, E. M.

Glossary of terms used in combinatorial chemistry. Pure Appl. Chem. 1999, 71, 2349–2365.

2. Wermuth, C. G.; Ganellin, C. R.; Lindberg, P.; Mitscher, L. A. Glossary of terms used in

medicinal chemistry. Pure Appl. Chem. 1998, 70, 1129–1143.

3. Lebl, M. Parallel personal comments on “classical” papers in combinatorial chemistry. J. Comb.

Chem. 1999, 1, 3–24.

4. Merriﬁ eld, R. B. Solid phase peptide synthesis. 1. The synthesis of a tetrapeptide. J. Am. Chem.

Soc. 1963, 85, 2149–2154.

5. Letsinger, R. L.; Mahadevan, V. Oligonucleotide synthesis on a polymer support. J. Am. Chem.

Soc. 1965, 87, 3526–3527.

6. Leznoff, C. C. The use of insoluble polymer supports in general organic synthesis. Acc. Chem.

Res. 1978, 11, 327–333.

7. Geysen, H. M.; Meleon, R. H.; Barteling, S. J. Use of peptide synthesis to probe viral antigens

for epitopes to a resolution of a single amino acid. Proc. Natl. Acad. Sci. U.S.A. 1984, 81,

3998–4002.

8. Schnorrenberg, G.; Gerhardt, H. Fully automatic simultaneous multiple peptide synthesis in

micromolar scale: rapid synthesis of series of peptides for screening in biological assays. Tet ra-

hedron 1989, 24, 7759–7764.

9. Houghten, R. A. General method for the rapid solid-phase synthesis of large numbers of pep-

tides: speciﬁ city of antigen–antibody interaction at the level of individual amino acids. Proc.

Natl. Acad. Sci. U.S.A. 1985, 82, 5131–5135.

10. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Cornucopia of peptides by synthesis. Abstr.

14th Int. Congr. Biochem. Prague, Czechoslovakia 1988, 47.

11. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. General method for rapid synthesis of multi-

component peptide mixtures. Int. J. Peptide Protein Res. 1991, 37, 487–493.

Subject Reviews Books

General 179–184

Glossaries, internet resources 1, 2, 185

Library design general, descriptors 61–63, 68, 69, 79, 82 186–188

Property-based, druglike 92, 189, 190

Reaction prediction 191

Solid-phase synthesis 192–197 198–201

Solid-phase supports 202, 203

Linkers for solid-phase chemistry 204–206

Encoding methods 207, 208

Liquid-phase synthesis 209, 210

Solution synthesis of libraries 211, 212

Supported reagents and scavengers 213

Solid-phase extraction 129

Separation strategies 214

Analytical methods 34, 35, 135, 215, 216 217

Automation 117, 218–221 222

Application to medicinal chemistry 153–155, 157, 223, 224

TABLE 4.7 Recent Reviews and Further Reading

12. Fodor, S. P. A.; Read, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Solas, D. Light-directed, spa-

tially addressable parallel chemical synthesis. Science 1991, 251, 767–773.

13. Houghten, R. A.; Pinilla, C.; Blondelle, S. E.; Appel, J. R.; Dooley, C. T.; Cuervo, J. H. Genera-

tion and use of synthetic peptide combinatorial libraries for basic research and drug discovery.

Nature 1991, 354, 84–86.

14. Pinilla, C.; Appel, J. R.; Blanc, P.; Houghten, R. A. Rapid identiﬁ cation of high afﬁ nity pep-

tide ligands using positional scanning synthetic peptide combinatorial libraries. BioTechniques

1992, 13, 901–905.

15. Dooley, C. T.; Houghten, R. A. The use of positional scanning synthetic peptide combina-

torial libraries for the rapid determination of opioid receptor ligands. Life Sci. 1993, 52,

1509–1517.

16. Smith, P. W.; Lai, J. Y. Q.; Whittington, A. R.; Cox, B.; Houston, J. G.; Stylli, C. H.; Banks,

M. N.; Tiller, P. R. Synthesis and biological evaluation of a library containing potentially 1600

amides/esters: a strategy for rapid compound generation and screening. Bioorg. Med. Chem.

Lett. 1994, 4, 2821–2824.

17. Deprez, B.; Williard, X.; Bourel, L.; Coste, H.; Hyaﬁ l, F.; Tartar, A. Orthogonal combinatorial

chemical libraries. J. Am. Chem. Soc. 1995, 117, 5405–5406.

18. Burgess, K.; Liaw, A. I.; Wang, N. Combinatorial technologies involving reiterative division/

coupling/recombination: statistical considerations. J. Med. Chem. 1994, 37, 2985–2987.

19. Topiol, S.; Davies, J.; Vijayakuma, S.; Wareing, J. R. Computer aided analysis of split and mix

combinatoral libraries. J. Comb. Chem. 2001, 3, 20–27.

20. Konings, D. A. M.; Wyatt, J. R.; Ecker, D. J.; Freier, S. M. Deconvolution of combinatorial

libraries for drug discovery: theoretical comparison of pooling strategies. J. Med. Chem. 1996,

39, 2710–2719.

21. Wilson-Lingardo, L.; Davis, P. W.; Ecker, D. J.; Hebert, N.; Acevedo, O.; Sprankle, K.;

Brennan, T.; Schwarcz, L.; Freier, S. M.; Wyatt, J. R. Deconvolution of combinatorial libraries

for drug discovery: experimental comparison of pooling strategies. J. Med. Chem. 1996, 39,

2720–2726.

22. Lam, K. S.; Salmon, S. E.; Hersh, E. M.; Hruby, V. J.; Kazmierski, W. M.; Knapp, R. J. A

new type of synthetic peptide library for identifying ligand-binding activity. Nature 1991, 354,

82–84.

23. Brenner, S.; Lerner, R. A. Encoded combinatorial chemistry. Proc. Natl. Acad. Sci. U.S.A. 1992,

89, 5381–5383.

24. Nielsen, J.; Brenner, S.; Janda, K. D. Synthetic methods for the implementation of encoded

combinatorial chemistry. J. Am. Chem. Soc. 1993, 115, 9812–9813.

25. Bunin, B. A.; Ellman, J. A. A general and expedient method for the solid-phase synthesis of

1,4-benzodiazepine derivatives. J. Am. Chem. Soc. 1992, 114, 10997–10998.

26. Bunin, B. A.; Plunkett, M. J.; Ellman, J. A. The combinatorial synthesis and chemical and

biological evaluation of a 1,4-benzodiazepine library. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,

4708–4712.

27. DeWitt, S. H.; Kiely, J. S.; Stankovic, C. J.; Schroeder, M. C.; Reynolds Cody, D. M.; Pavia,

M. R. “Diversomers”: an approach to nonpeptide, nonoligomeric chemical diversity. Proc.

Natl. Acad. Sci. U.S.A. 1993, 90, 6909–6913.

28. Ohlmeyer, M. H. J.; Swanson, R. N.; Dillard, L. W.; Reader, J. C.; Asouline, G.; Kobayashi,

R.; Wigler, M.; Still, W. C. Complex synthetic chemical libraries indexed with molecular tags.

Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 10922–10926.

29. Ni, Z.-J.; Maclean, D.; Holmes, C. P.; Murphy, M. M.; Ruhland, B.; Jacobs, J. W.; Gordon,

E. M.; Gallop, M. A. Versatile approach to encoding combinatorial organic syntheses using

chemically robust secondary amine tags. J. Med. Chem. 1996, 39, 1601–1608.

REFERENCES 157

158 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

30. Geysen, H. M.; Wagner, C. D.; Bodnar, W. M.; Markworth, C. J.; Parke, G. J.; Schoenen, F. J.;

Wagner, D. S.; Kinder, D. S. Isotope or mass encoding of combinatorial libraries. Chem. Biol.

1996, 3, 679–688.

31. Hughes, I. Design of self-coded combinatorial libraries to facilitate direct analysis of ligands by

mass spectrometry. J. Med. Chem. 1998, 41, 3804–3811.

32. Nicolaou, K. C.; Xiao, X.-Y.; Parandoosh, Z.; Senyei, A.; Nova, M. P. Radiofrequency encoded

combinatorial chemistry. Angew. Chem. Int. Ed. 1995, 34, 2289–2291.

33. Yan, B. Monitoring the progress and the yield of solid-phase organic reactions directly on resin

supports. Acc. Chem. Res. 1998, 31, 621–630.

34. Swali, V.; Langley, G. J.; Bradley, M. Mass spectrometric analysis in combinatorial chemistry.

Curr. Opin. Chem. Biol. 1999, 3, 337–341.

35. Shapiro, M. J.; Wareing, J. R. NMR methods in combinatorial chemistry. Curr. Opin. Chem.

Biol. 1998, 2, 372–375.

36. Han, H.; Wolfe, M. M.; Brenner, S.; Janda, K. D. Liquid-phase combinatorial synthesis. Proc.

Natl. Acad. Sci. U.S.A. 1995, 92, 6419–6423.

37. Chen, S.; Janda, K. D. synthesis of Prostaglandin E2 methyl ester on a soluble-polymer support

for the construction of prostanoid libraries. J. Am. Chem. Soc. 1997, 119, 8724–8725.

38. Chen, S.; Janda, K. D. Total synthesis of naturally occurring prostaglandin F2alpha on a non-

cross-linked polystyrene support. Tetrahedron Lett. 1998, 39, 3943–3946.

39. Struder, A.; Hadida, S.; Ferritto, R.; Kim, S.-Y.; Jeger, P.; Wipf, P.; Curran, D. P. Fluorous

synthesis: a ﬂ uorous-phase strategy for improving separation efﬁ ciency in organic synthesis.

Science 1997, 275, 823–826.

40. Struder, A.; Jeger, P.; Wipf, P.; Curran, D. P. Fluorous synthesis: ﬂ uorous protocols for the Ugi

and Biginelli multicomponent condensations. J. Org. Chem. 1997, 62, 2917–2924.

41. Struder, A.; Curran, D. P. A strategic alternative to solid-phase synthesis: preparation of a small

isoxazoline library by “ﬂ uorous synthesis”. Tetrahedron 1997, 53, 6681–6696.

42. Curran, D. P.; Hadida, S. Tris(2-(perﬂ uorohexyl)ethyl)tin hydride: a new ﬂ uorous reagent for

use in traditional organic synthesis and liquid phase combinatorial synthesis. J. Am. Chem. Soc.

1996, 118, 2531–2532.

43. Curran, D. P. Combinatorial organic synthesis and phase separation: back to the future. Chem-

Tracts Org. Chem. 1996, 9, 75–87.

44. Boger, D. L.; Tarby, C. M.; Myers, P. L.; Caporale, L. H. Generalized dipeptidomimetic tem-

plate: Solution phase parallel synthesis of combinatorial libraries. J. Am. Chem. Soc. 1996, 118,

2109–2110.

45. Cheng, S.; Comer, D. D.; Williams, J. P.; Myers, P. L.; Boger, D. L. Novel solution phase strat-

egy for the synthesis of chemical libraries containing small organic molecules. J. Am. Chem.

Soc. 1996, 118, 2567–2573.

46. Kaldor, S. W.; Siegel, M. G.; Fritz, J. E.; Dressman, B. A.; Hahn, P. J. Use of solid supported

nucleophiles and electrophiles for the puriﬁ cation of non-peptide small molecule libraries.

Tetrahedron Lett. 1996, 37, 7193–7196.

47. Kaldor, S. W.; Fritz, J. E.; Tang, J.; McKinney, E. R. Discovery of antirhinoviral leads by

screening a combinatorial library of ureas prepared using covalent scavengers. Bioorg. Med.

Chem. Lett. 1996, 6, 3041–3044.

48. Caldarelli, M.; Habermann, J.; Ley, S. V. Synthesis of an array of potential matrix metallopro-

teinase inhibitors using a sequence of polymer-supported reagents. Bioorg. Med. Chem. Lett.

1999, 9, 2049–2052.

49. Baxendale, I. R.; Ley, S. V. Polymer-supported reagents for multi-step organic synthesis:

application to the synthesis of Sildenaﬁ l. Bioorg. Med. Chem. Lett. 2000, 10, 1983–1986.

50. Ley, S. V.; Schucht, O.; Thomas, A. W.; Murray, P. J. Synthesis of the alkaloids (⫹/⫺)-oxo-

maritidine and (⫹/⫺)-epimaritidine using an orchestrated multi-step sequence of polymer sup-

ported reagents. J. Chem. Soc., Perkin Trans. 1 1999, 1251–1252.

51. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Experimental and computational

approaches to estimate solubility and permeability in drug discovery and development settings.

Adv. Drug Del. Rev. 1997, 23, 3–25.

52. Ohlstein, E. H.; Ruffolo, R. R.; Elliott, J. D. Drug discovery in the next millennium. Annu. Rev.

Pharmacol. Toxicol. 2000, 40, 177–191.

53. Drews, J. Drug discovery: a historical perspective. Science 2000, 287, 1960–1969.

54. Venton, D. L.; Woodbury, C. P. Screening combinatorial libraries. Chemom. Intell. Lab. Syst.

1999, 48, 131–150.

55. Lam, K. S.; Lebl, M.; Krchnak, V. The “one-bead–one-compound” combinatorial library

method. Chem. Rev. 1997, 97, 411–448.

56. Eddershaw, P. J.; Beresford, A. P.; Bayliss, M. K. ADME/PK as part of a rational approach to

drug discovery. Drug Discov. Today 2000, 5, 409–414.

57. Brennan, M. B. Drug discovery. Filtering out failured early in the game. Chem. Eng. News

2000, 78, 63–73.

58. Czarnik, A. W. Why combinatorial chemistry, why now (and why you should care). Chem-

Tracts Org. Chem. 1995, 8, 13–18.

59. Martin, Y. C. Challenges and prospects for comutational aids to molecular diversity. Perspect.

Drug Discov. Des. 1997, 7/8, 159–172.

60. Walters, W. P.; Stahl, M. T.; Murcko, M. A. Virtual screening: an overview. Drug Discov. Today

1998, 3, 160–178.

61. Drewry, D. H.; Young, S. S. Approaches to the design of combinatorial libraries. Chemom.

Intell. Lab. Syst. 1999, 48, 1–20.

62. Bajorath, J. Selected concepts and investigations in compound classiﬁ cation, molecular

descriptor analysis, and virtual screening. J. Chem. Inf. Comput. Sci. 2001, 41, 233–245.

63. Leach, A. R.; Hann, M. M. The in silico world of virtual libraries. Drug Discov. Today 2000, 5,

326–336.

64. Gorse, D.; Lahana, R. Functional diversity of compound libraries. Curr. Opin. Chem. Biol.

2000, 4, 287–294.

65. Warr, W. A. Commercial software systems for diversity analysis. Perspect. Drug Discov. Des.

1997, 7/8, 115–130.

66. Leach, A. R.; Bradshaw, J.; Green, D. V. S.; Hann, M. M. Implementation of a system for re-

agent selection and library enumeration, proﬁ ling, and design. J. Chem. Inf. Comput. Sci. 1999,

39, 1161–1172.

67. Martin, Y. C. Diverse viewpoints on computational aspects of molecular diversity. J. Comb.

Chem. 2001, 3, 1–20.

68. Mason, J. S.; Hermsmeier, M. A. Diversity assessment. Curr. Opin. Chem. Biol. 1999, 3,

342–349.

69. Agraﬁ otis, D. K.; Myslik, J. C.; Salemme, F. R. Advances in diversity proﬁ ling and combinato-

rial series design. Mol. Diversity 1999, 4, 1–22.

70. Livingstone, D. J. The characterization of chemical structures using molecular properties: a

survey. J. Chem. Inf. Comput. Sci. 2000, 40, 195–209.

71. Katritzky, A. R.; Maran, U.; Lobanov, V. S.; Karelson, M. Structurally diverse quantitative

structure–property relationship correlations of technologically relevant physical properties. J.

Chem. Inf. Comput. Sci. 2000, 40, 1–18.

72. Brown, R. D. Descriptors for diversity analysis. Perspect. Drug Discov. Des. 1997, 7/8, 31–49.

REFERENCES 159

160 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

73. Martin, E. J.; Blaney, J. M.; Siani, M. A.; Spellmeyer, D. C.; Wong, A. K.; Moos, W. H. Measur-

ing diversity: experimental design of combinatorial libraries for drug discovery. J. Med. Chem.

1995, 38, 1431–1436.

74. Pearlman, R. S.; Smith, K. M. Novel software tools for chemical diversity. Perspect. Drug

Discov. Des. 1998, 9/10/11, 339–353.

75. Pearlman, R. S.; Smith, K. M. Software for chemical diversity in the context of accelerated drug

discovery. Drugs Future 1998, 23, 885–895.

76. Brown, R. D.; Martin, Y. C. Use of structure–activity data to compare structure-based clustering

methods and descriptors for use in compound selection. J. Chem. Inf. Comput. Sci. 1996, 36,

572–584.

77. Patterson, D. E.; Cramer, R. D.; Ferguson, A. M.; Clark, R. D.; Weinberger, L. E. Neighborhood

behavior: a useful concept for validation of “molecular diversity” descriptors. J. Med. Chem.

1996, 39, 3049–3059.

78. Matter, H.; Potter, T. Comparing 3D pharmacophore triplets and 2D ﬁ ngerprints for selecting

diverse compound subsets. J. Chem. Inf. Comput. Sci. 1999, 39, 1211–1225.

79. Van Drie, J. H.; Lajiness, M. S. Approaches to virtual library design. Drug Discov. Today 1998,

3, 274–283.

80. Pickett, S. D.; Mason, J. S.; McLay, I. M. Diversity proﬁ ling and design using 3D pharmacoph-

ores: pharmacophore-derived queries (PDQ). J. Chem. Inf. Comput. Sci. 1996, 36, 1214–1223.

81. Mason, J. S.; Morize, I.; Menard, P. R.; Cheney, D. L.; Hulme, C.; Labaudiniere, R. F. New 4-

point pharmacophore method for molecular similarity and diversity applications: overview of

the method and applications, including a novel approach to the design of combinatorial libraries

containing privileged substructures. J. Med. Chem. 1999, 42, 3251–3264.

82. Beno, B. R.; Mason, J. S. The design of combinatorial libraries using properties and 3D phar-

macophore ﬁ ngerprints. Drug Discov. Today 2001, 6, 251–258.

83. Jamois, E. A.; Hassan, M.; Waldman, M. Evaluation of reagent-based and product-based strate-

gies in the design of combinatorial library subsets. J. Chem. Inf. Comput. Sci. 2000, 40, 63–70.

84. Gillet, V. J.; Nicolotti, O. Evaluation of reactant-based and product-based approaches to the

design of combinatorial libraries. Perspect. Drug Discov. Des. 2000, 20, 265–287.

85. Jarvis, R. A.; Patrick, E. A. Clustering using a similarity measure based on shared near neigh-

bors. IEEE Trans. Comput. 1973, C22, 1025–1034.

86. Ward, J. H. Heirarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 1963,

58, 236–244.

87. Mason, J. S.; Pickett, S. D. Partition-based selection. Perspect. Drug Discov. Des. 1997, 7兾8,

85–114.

88. Gillet, V. J.; Willet, P.; Bradshaw, J. The effectiveness of reactant pools for generating structur-

ally-diverse combinatorial libraries. J. Chem. Inf. Comput. Sci. 1997, 37, 731–740.

89. Stanton, R. V.; Mount, J.; Miller, J. L. Combinatorial library design: maximizing model-ﬁ tting

compounds within matrix synthesis constraints. J. Chem. Inf. Comput. Sci. 2000, 40, 701–705.

90. Bravi, G.; Green, D. V. S.; Hann, M. M.; Leach, A. R. PLUMS: a program for the rapid optimi-

zation of focused libraries. J. Chem. Inf. Comput. Sci. 2000, 40, 1441–1448.

91. Leo, A. J. Calculating log P

oct

from structures. Chem. Rev. 1993, 93, 1281–1306.

92. Walters, W. P.; Ajay; Murcko, M. A. Recognizing molecules with drug-like properties. Curr.

Opin. Chem. Biol. 1999, 3, 384–387.

93. Bemis, G. W.; Murcko, M. A. The properties of known drugs. 1. Molecular frameworks. J. Med.

Chem. 1996, 39, 2887–2893.

94. Bemis, G. W.; Murcko, M. A. Properties of known drugs. 2. Side chains. J. Med. Chem. 1999,

42, 5095–5099.

95. Ajay; Walters, W. P.; Murcko, M. A. Can we learn to distinguish between “drug-like” and “non-

drug-like” molecules? J. Med. Chem. 1998, 41, 3314–3324.

96. Sadowski, J.; Kubinyi, H. A scoring scheme for discriminating between drugs and nondrugs. J.

Med. Chem. 1998, 41, 3325–3329.

97. Ajay; Bemis, G. W.; Murcko, M. A. Designing libraries with CNS activity. J. Med. Chem. 1999,

42, 4942–4951.

98. Teague, S. J.; Davis, A. M.; Leeson, P. D.; Oprea, T. The design of leadlike combinatorial librar-

ies. Angew. Chem. Int. Ed. 1999, 38, 3743–3748.

99. Hann, M. M.; Leach, A. R.; Harper, G. Molecular complexity and its impact on the probability

of ﬁ nding leads for drug discovery. J. Chem. Inf. Comput. Sci. 2001, 41, 856–864.

100. Kuntz, I. D. Structure-based strategies for drug-design and discovery. Science 1992, 257,

1078–1082.

101. Li, J.; Murray, C. W.; Waszkowycz, B.; Young, S. C. Targeted molecular diversity in drug dis-

covery: integration of structure-based design and combinatorial chemistry. Drug Discov. Today

1998, 3, 105–112.

102. Kirkpatrick, D. L.; Watson, S.; Ulhaq, S. Structure-based drug design: combinatorial chemistry

and molecular modeling. Comb. Chem. High Throughput Screen. 1999, 2, 211–221.

103. Bohm, H.-J.; Stahl, M. Structure-based library design: molecular modelling merges with com-

binatorial chemistry. Curr. Opin. Chem. Biol. 2000, 4, 283–286.

104. DesJarlais, R. L.; Sheridan, R. P.; Seibel, G. L.; Dixon, J. S.; Kuntz, I. D. Using shape comple-

mentarity as an initial screen in designing ligands for a receptor binding site of known three-

dimensional structure. J. Med. Chem. 1988, 31, 722–729.

105. Goodford, P. J. A computational procedure for determining energetically favorable binding

sites on biologically important macromolecules. J. Med. Chem. 1985, 28, 849–857.

106. Bohm, H.-J. The computer program LUDI: a new method for the de novo design of enzyme

inhibitors. J. Comput.- Aided Mol. Des. 1992, 6, 61–78.

107. Kick, E. K.; Roe, D. C.; Skillman, A. G.; Lui, G.; Ewing, T. J. A.; Sun, Y.; Kuntz, I. D.; Ellman,

J. A. Structure-based design and combinatorial chemistry yield low nanomolar inhibitors of

cathepsin D. Chem. Biol. 1997, 4, 297–307.

108. Evans, B. E.; Rittle, K. E.; Bock, M. G.; DiPardo, R. M.; Freidinger, R. M.; Whitter, W. L.;

Lundell, G. F.; Veber, D. F.; Anderson, P. S.; Chang, R. S. L.; Lotti, V. J.; Cerino, D. J.; Chen,

T. B.; Kling, P. J.; Kunkel, K. A.; Springer, J. P.; Hirshﬁ eld, J. Methods for drug discovery:

development of potent, selective, orally effective cholecystokinin antagonists. J. Med. Chem.

1988, 31, 2235–2246.

109. Patchett, A. A.; Nargund, R. P. Privileged structures: an update. Annu. Rep. Med. Chem. 2000,

35, 289–298.

110. Lewell, X. Q.; Judd, D. B.; Watson, S. P.; Hann, M. M. RECAP-retrosynthetic combinato-

rial analysis procedure: a powerful new technique for identifying privileged molecular frag-

ments with useful applications in combinatorial chemistry. J. Chem. Inf. Comput. Sci. 1998,

38, 511–522.

111. Braban, M.; Pop, I.; Willard, X.; Horvath, D. Reactivity prediction models applied to the selec-

tion of novel candidate building blocks for high-throughput organic synthesis of combinatorial

libraries. J. Chem. Inf. Comput. Sci. 1999, 39, 1119–1127.

112. Satoh, H.; Sacher, O.; Nakata, T.; Chen, L.; Gasteiger, J.; Funatsu, K. Classiﬁ cation of organic

reactions: similarity of reactions based on changes in the electronic features of oxygen atoms at

the reaction sites. J. Chem. Inf. Comput. Sci. 1998, 38, 210–219.

113. Gasteiger, J.; Hondelmann, U.; Rose, P.; Witzenbichler, W. Computer-assisted prediction of the

degradation of chemicals: hydrolysis of amides and benzoylphenylureas. J. Chem. Soc., Perkin

Trans. 1, 1995, 193–204.

REFERENCES 161

162 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

114. Hollering, R.; Gasteiger, J.; Steinhauer, L.; Schulz, K.-P.; Herwig, A. Simulation of organic

reactions: from the degradation of chemicals to combinatorial synthesis. J. Chem. Inf. Comput.

Sci. 2000, 40, 482–494.

115. Kuo, P. Y.; Du, H.; Corkan, L. A.; Yang, K.; Lindsey, J. S. A planning module for perform-

ing grid search, factorial design, and related combinatorial studies on an automated chemistry

workstation. Chemom. Intell. Lab. Syst. 1999, 48, 219–234.

116. Pilipauskas, D. R. Can the time from synthesis design to validated chemistry be shortened?

Med. Res. Rev. 1999, 19, 463–474.

117. Hird, N. W. Automated synthesis: new tools for the organic chemist. Drug Discov. Today 1999,

4, 265–274.

118. DeWitt, S. H.; Czarnik, A. W. Combinatorial organic synthesis using Parke-Davis’s Diversomer

method. Acc. Chem. Res. 1996, 29, 114–122.

119. Xiao, X.-Y.; Li, R.; Zhuang, H.; Ewing, B.; Karunaratne, K.; Lillig, J.; Brown, R.; Nicolaou,

K. C. Solid-phase combinatorial synthesis using MicroKan reactors, rf tagging, and directed

sorting. Biotech. Eng. 2000, 71, 44–50.

120. Plante, O. J.; Palmacci, E. R.; Seeberger, P. H. Automated solid-phase synthesis of oligosac-

charides. Science 2001, 291, 1523–1527.

121. Brooking, P.; Doran, A.; Grimsey, P.; Hird, N. W.; MacLachlan, W. S.; Vimal, M. Split-split: a

multiple synthesiser approach to efﬁ cient automated parallel synthesis. Tetrahedron Lett. 1999,

40, 1405–1408.

122. Larhed, M.; Hallberg, A. Microwave-assisted high-speed chemistry: a new technique in drug

discovery. Drug Discov. Today 2001, 6, 406–416.

123. Mahar, L. Resin capsules for solid-phase organic synthesis. Spec. Chem. 1998, 15–15.

124. Hird, N. W.; Hughes, I.; Re, V.; North, N. C. Formulation. 1999, WO 99/11676.

125. Atrash, B.; Bradley, M.; Kobylecki, R.; Cowell, D.; Reader, J. Revolutionizing resin handling

for combinatorial synthesis. Angew. Chem. Int. Ed. 2001, 40, 938–941.

126. Hird, N.; Hughes, I.; Hunter, D.; Morrison, M. G. J. T.; Sherrington, D. C.; Stevenson, L.

Polymer disks: an alternative support format for solid phase synthesis. Tetrahedron 1999, 55,

9575–9584.

127. Tripp, J. A.; Stein, J. A.; Svec, F.; Frechet, J. M. J. “Reactive ﬁ ltration”: use of functionalized po-

rous polymer monoliths as scavengers in solution-phase synthesis. Org. Lett. 2000, 2, 195–198.

128. Vaino, A. R.; Janda, K. D. Euclidean shape-encoded combinatorial chemical libraries. Proc.

Natl. Acad. Sci. U.S.A. 2000, 97, 7692–7696.

129. Nilsson, U. J. Solid-phase extraction for combinatorial libraries. J. Chromatogr. A 2000, 885,

305–319.

130. Siegel, M. G.; Hahn, P. J.; Dressman, B. A.; Fritz, J. E.; Grunwell, J. R.; Kaldor, S. W. Rapid

puriﬁ cation of small molecule libraries by ion exchange chromatography. Tetrahedron Lett.

1997, 38, 3357–3360.

131. Schultz, L.; Garr, C. D.; Cameron, L. M.; Bukowski, J. High throughput puriﬁ cation of combi-

natorial libraries. Bioorg. Med. Chem. Lett. 1998, 8, 2409–2414.

132. Kiplinger, J. P.; Cole, R. O.; Robinson, S.; Roskamp, E. J.; Ware, R. S.; O’Connell, H. J.;

Brailsford, A.; Batt, J. Structure-controlled automated puriﬁ cation of parallel synthesis prod-

ucts in drug discovery. Rapid Commun. Mass Spectrom. 1998, 12, 662.

133. Diggelmann, M.; Sporri, H.; Gassmann, E. Preparative LC/MS technology: a key component of

the existing high speed synthesis at Syngenta. Chimia 2001, 55, 23–25.

134. Hughes, I. Separating the wheat from the chaff: high throughput puriﬁ cation of chemical librar-

ies. J. Assoc. Lab. Autom. 2000, 5, 69–71.

135. Ripka, W. C.; Barker, G.; Krakover, J. High-throughput puriﬁ cation of compound libraries.

Drug Discov. Today 2001, 6, 471–477.

136. Hsu, B. H.; Orton, E.; Tang, S.-Y.; Carlton, R. A. Application of evaporative light scattering

detection to the characterization of combinatorial and parallel synthesis libraries for pharma-

ceutical drug discovery. J. Chromatogr. A 1999, 725, 103–112.

137. Fang, L.; Wan, M.; Pennacchio, M.; Pan, J. Evaluation of evaporative light-scattering detec-

tor for combinatorial libray quantitation by reversed phase HPLC. J. Comb. Chem. 2000, 2,

254–257.

138. Fitch, W. L.; Szardenings, A. K.; Fujinari, E. M. Chemiluminescent nitrogen detection for

HPLC: an important new tool in organic analytical chemistry. Tetrahedron Lett. 1997, 38,

1689–1692.

139. Shah, N.; Gao, M.; Tsutsui, K.; Lu, A.; Davis, J.; Scheuerman, R.; Fitch, W. L. A novel ap-

proach to high-throughput quality control of parallel synthesis libraries. J. Comb. Chem. 2000,

2, 453–460.

140. Biasi, V.; Haskins, N.; Organ, A.; Bateman, R.; Giles, K.; Jarvis, S. High throughput liquid

chromatography/mass spectrometric analyses using a novel multiplexed electrospray interface.

Rapid Commun. Mass Spectron. 1999, 13, 1165–1168.

141. Gorlach, E.; Richmond, R.; Lewis, I. High-throughput ﬂ ow injection analysis mass spectros-

copy with networked delivery of color-rendered results. 2. Three-dimensional spectral mapping

of 96-well combinatorial chemistry racks. Anal. Chem. 1998, 70, 3227–3234.

142. Tong, H.; Bell, D.; Tabei, K.; Siegel, M. M. Automated data massaging, interpretation, and

e-mailing modules for high throughput open access mass spectrometry. J. Am. Soc. Mass Spec-

tron. 1999, 10, 1174–1187.

143. Keifer, P. A.; Smallcombe, S. H.; Williams, E. H.; Salomon, K. E.; Mendez, G.; Belletire, J.

L.; Moore, C. D. Direct-injection NMR (DI-NMR): a ﬂ ow NMR technique for the analysis of

combinatorial chemistry libraries. J. Comb. Chem. 2000, 2, 151–171.

144. Gerritz, S. W.; Seﬂ er, A. M. 2,5-Dimethylfuran (DMFu): an internal standard for the “traceless”

quantitation of unknown samples via 1H NMR. J. Comb. Chem. 2000, 2, 39–41.

145. Gobbi, A.; Poppinger, D.; Rohde, B. Developing an in-house system to support combinatorial

chemistry. Perspect. Drug Discov. Des. 1997, 7兾8, 131–158.

146. Manly, C. J. Managing laboratory automation: integration and informatics in drug discovery. J.

Autom. Methods Manage. Chem. 2000, 22, 169–170.

147. Davies, K.; White, C. Managing combinatorial chemistry information. In Molecular Diversity

in Drug Design, Dean, P. M.; Lewis, A. L. (Eds), Kluwer Academic Publishers, Dordrecht, The

Netherlands, 1999, pp. 175–196.

148. Goodman, B. A. Managing the workﬂ ow of a high-throughput organic synthesis laboratory:

a marriage of automation and information management technologies. J. Assoc. Lab. Autom.

1999, 4, 48–52.

149. Merritt, A. T. Uptake of new technology in lead optimization for drug discovery. Drug Discov.

Today 1998, 3, 505–510.

150. Labaudiniere, R. F. RPR’s approach to high-speed parallel synthesis for lead generation. Drug

Discov. Today 1998, 3, 511–515.

151. Coates, W. J.; Hunter, D. J.; MacLachlan, W. S. Successful implementation of automation in

medicinal chemistry. Drug Discov. Today 2000, 5, 521–527.

152. Herzberg, R. P.; Pope, A. J. High-throughput screening: new technology for the 21st century.

Curr. Opin. Chem. Biol. 2000, 4, 445–451.

153. Dolle, R. E. Comprehensive survey of combinatorial library synthesis: 1999. J. Comb. Chem.

2000, 2, 383–433.

REFERENCES 163

164 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

154. Dolle, R. E.; Nelson, K. H. Comprehensive survey of combinatorial library synthesis: 1998. J.

Comb. Chem. 1999, 1, 235–282.

155. Dolle, R. E. Comprehensive survey of chemical libraries yielding enzyme inhibitors, receptor

agonists and antagonists, and other biologically active agents: 1992 through 1997. Mol. Diver-

sity 1998, 3, 199–233.

156. Dolle, R. E. Comprehensive survey of combinatorial libraries with undisclosed biological activ-

ity: 1992-1997. Mol. Diversity 1998, 4, 233–256.

157. Golebiowski, A.; Klopfenstein, S. R.; Portlock, D. E. Lead compounds discovered from librar-

ies. Curr. Opin. Chem. Biol. 2001, 5, 273–284.

158. Bromidge, S. M.; Brown, A. M.; Clarke, S. E.; Dodgson, K.; Gager, T.; Grassam, H. L.;

Jeffrey, P. M.; Joiner, G. F.; King, F. D.; Middlemiss, D. N.; Moss, S. F.; Newman, H.; Riley, G.;

Routledge, C.; Wyman, P. 5-Chloro-N-(4-methoxy-3piperazin-1-ylphenyl)-3-methyl-2-benzo-

thiophenesulfonamide (SB-271046): a potent, selective, and orally bioavailable 5-HT6 receptor

antagonist. J. Med. Chem. 1999, 42, 202–205.

159. Bromidge, S. M.; Clarke, S. E.; Gager, T.; Grifﬁ th, K.; Jeffrey, P.; Jennings, A. J.; Joiner, G.

F.; King, F. D.; Lovell, P. J.; Moss, S. F.; Newman, H.; Riley, G.; Rogers, D.; Routledge, C.;

Seraﬁ nowska, H.; Smith, D. R. Phenyl benzenesulfonamides are novel and selective 5-HT6

antagonists: identiﬁ cation of N-(2,5-dibromo-3-ﬂ uorophenyl)-4-methoxy-3-piperazin-1-ylben-

zenesulfonamide (SB-357134). Bioorg. Med. Chem. Lett. 2001, 11, 55–58.

160. Berk, S. C.; Rohrer, S. P.; Degrado, S. J.; Birzin, E. T.; Mosley, R. T.; Hutchins, S. M.; Pasternak,

A.; Schaeffer, J. M.; Underwood, D. J.; Chapman, K. T. A combinatorial approach toward

the discovery of non-peptide, subtype-selective somatostatin receptor ligands. J. Comb. Chem.

1999, 1, 388–396.

161. Rohrer, S. P.; Birzin, E. T.; Mosley, R. T.; Berk, S. C.; Hutchins, S. M.; Shen, D.-M.; Xiong, Y.;

Hayes, E. C.; Parmar, R. M.; Foor, F.; Mitra, S. W.; Degrado, S. J.; Shu, M.; Klopp, J. M.; Cai,

S.-J.; Blake, A.; Chan, W. W. S.; Pasternak, A.; Yang, L.; Patchett, A. A.; Smith, R. G.; Chapman,

K. T.; Schaeffer, J. M. Rapid identiﬁ cation of subtype-selective agonists of the somatostatin

receptor through combinatorial chemistry. Science 1998, 282, 737–740.

162. Sarshar, S.; Zhang, C.; Moran, E. J.; Krane, S.; Rodarte, J. C.; Benbatoul, K. D.; Dixon, R.;

Mjalli, A. M. M. 2,4,5-Trisubstituted imidazoles: novel nontoxic modulators of P-glycoprotein

mediated multidrug resistance. Part 1. Bioorg. Med. Chem. Lett. 2000, 10, 2599–2601.

163. Zhang, C.; Sarshar, S.; Moran, E. J.; Krane, S.; Rodarte, J. C.; Benbatoul, K. D.; Dixon, R.;

Mjalli, A. M. M. 2,4,5-Trisubstituted imidazoles: novel nontoxic modulators of P-glycoprotein

mediated multidrug resistance. Part 2. Bioorg. Med. Chem. Lett. 2000, 10, 2603–2605.

164. Borman, S. Combinatorial chemistry: redeﬁ ning the scientiﬁ c method. Chem. Eng. News 2000,

78, 53–65.

165. Thorpe, D. S. Forecasting roles of combinatorial chemistry in the age of genomically derived

drug discovery targets. Comb. Chem. High Throughput Screen. 2000, 3, 421–436.

166. Harrison, W. Changes in scale in automated pharmaceutical research. Drug Discov. Today

1998, 3, 343–349.

167. Archer, R. The drug discovery factory: an inevitable evolutionary consequence of high-through-

put parallel processing. Nat. Biotechnol. 1999, 17, 834–834.

168. Wessjohann, L. A. Synthesis of natural-product-based compound libraries. Curr. Opin. Chem.

Biol. 2000, 4, 303–309.

169. Nicolaou, K. C.; Pfefferkorn, J. A.; Roecker, A. J.; Cao, G.-Q.; Barluenga, S.; Mitchell, H. J.

Natural product-like combinatorial libraries based on privileged structures. 1. General prin-

ciples and solid-phase synthesis of benzopyrans. J. Am. Chem. Soc. 2000, 122, 9939–9953.

170. Krstenansky, J. L.; Khmelnitsky, Y. Biocatalytic combinatorial synthesis. Bioorg. Med. Chem.

1999, 7, 2157–2162.

171. Lehn, J.-M. Dynamic combinatorial chemistry and virtual combinatorial libraries. Chem. Eur.

J. 1999, 5, 2455–2463.

172. Cousins, G. R. L.; Poulsen, S.-A.; Sanders, J. K. M. Molecular evolution: dynamic combinato-

rial libraries, autocatalytic networks and the quest for molecular function. Curr. Opin. Chem.

Biol. 2000, 4, 270–279.

173. Huc, I.; Nguyen, R. Dynamic combinatorial chemistry. Comb. Chem. High Throughput Screen.

2001, 4, 53–74.

174. Wolke, J.; Ullmann, D. Miniaturized HTS technologies: uHTS. Drug Disc. Today 2001, 6,

637–646.

175. Barrow, D.; Cefai, J.; Taylor, S. Shrinking to ﬁ t. Chem. Ind. 1999, 591–594.

176. Cowen, S. Chip service. Chem. Ind. 1999, 584–586.

177. DeWitt, S. H. Microreactors for chemical synthesis. Curr. Opin. Chem. Biol. 1999, 3, 350–356.

178. Haswell, S. J.; Middleton, R. J.; O’Sullivan, B.; Skelton, V.; Watts, P.; Styring, P. The applica-

tion of micro reactors to synthetic chemistry. Chem. Commun. 2001, 391–398.

179. Terrett, N. K. Combinatorial Chemistry, Oxford University Press, Oxford, 1998.

180. Miertus, S.; Fassina, G. (Eds), Combinatorial Chemistry and Technology: Principles, Methods

and Applications, Marcel Dekker, New York, 1999.

181. Jung, G. (Ed.), Combinatorial Chemistry: Synthesis, Analysis, Screening, Wiley-VCH,

Weinheim, Germany, 1999.

182. Fenniri, H. (Ed.), Combinatorial Chemistry, Oxford University Press, Oxford, 2000.

183. Seneci, P. Solid Phase Synthesis and Combinatorial Technologies, Wiley-Interscience, New

York, 2000.

184. Bannwarth, W.;Felder, E. (Eds.), Combinatorial Chemistry: A Practical Approach, Wiley-

VCH, Weinheim, 2000.

185. Nakayama, G. R. Combinatorial chemistry web alert. Curr. Opin. Chem. Biol. 2001, 5,

239–240.

186. Gordon, E. M.; Kerwin, J. F. (Eds.), Combinatorial Chemistry and Molecular Diversity in Drug

Discovery, Wiley, New York, 1998.

187. Dean, P. M.; Lewis, R. A. (Eds.), Molecular Diversity in Drug Design, Kluwer Academic Pub-

lishers, Dordrecht, The Netherlands, 1999.

188. Ghose, A. K.; Viswanadhan, V. N. (Eds.), Combinatorial Library Design and Evaluation. Prin-

ciples, Software Tools, and Applications in Drug Discovery, Marcel Dekker, New York, 2001.

189. Clark, D. E.; Pickett, S. D. Computational methods for the prediction of “drug-likeness.”

Drug Discov. Today 2000, 5, 49–58.

190. Van de Waterbeemd, H.; Smith, D. A.; Beaumont, K.; Walker, D. K. Property-based design:

optimization of drug absorption and pharmacokinetics. J. Med. Chem. 2001, 44, 1313–1333.

191. Gasteiger, J.; Pfortner, M.; Sitzmann, M.; Hollering, R.; Sacher, O.; Kostka, T.; Karg, N. Com-

puter-assisted synthesis and reaction planning in combinatorial chemistry. Perspect. Drug Dis-

cov. Des. 2000, 20, 245–264.

192. Hermkens, P. H. H.; Ottenheijm, H. C. J.; Rees, D. Solid-phase organic reactions: a review of

the recent literature. Tetrahedron 1996, 52, 4527–4554.

193. Hermkens, P. H. H.; Ottenheijm, H. C. J.; Rees, D. Solid-phase organic reactions II: a review of

the literature Nov. 95–Nov. 96. Tetrahedron 1997, 53, 5643–5678.

194. Booth, S.; Hermkens, P. H. H.; Ottenheijm, H. C. J.; Rees, D. C. Solid-phase organic reactions

III: a review of the literature Nov. 96–Dec. 97. Tetrahedron 1998, 54, 15385–15443.

195. Lorsbach, B. A.; Kurth, M. J. Carbon-carbon bond forming solid-phase reactions. Chem. Rev.

1999, 99, 1549–1581.

REFERENCES 165

166 COMBINATORIAL CHEMISTRY IN THE DRUG DISCOVERY PROCESS

196. Franzen, R. G. Recent advances in the preparation of heterocycles on solid support: a review of

the literature. J. Comb. Chem. 2000, 2, 195–214.

197. Sammelson, R. E.; Kurth, M. J. Carbon-carbon bond-forming solid-phase reactions. Part II.

Chem. Rev. 2001, 101, 137–202.

198. Obrecht, D.;Villalgordo, J. M. (Eds.), Solid-Supported Combinatorial and Parallel Synthesis of

Small-Molecular-Weight Compound Libraries, Pergamon Press, Oxford, 1998.

199. Bunin, B. A. The Combinatorial Index, Academic Press, San Diego, CA, 1998.

200. Dorwald, F. Z. Organic Synthesis on Solid Phase: Supports, Linkers, Reactions, Wiley-VCH,

Weinheim, Germany, 2000.

201. Burgess, K. (Ed.), Solid-Phase Organic Synthesis, Wiley-Interscience, New York, 2000.

202. Hudson, D. Matrix assisted synthetic transformations: a mosaic of diverse contributions. 1. The

pattern emerges. J. Comb. Chem. 1999, 1, 333–360.

203. Hudson, D. Matrix assisted synthetic transformations: a mosaic of diverse contributions. 2. The

pattern is completed. J. Comb. Chem. 1999, 1, 403–457.

204. James, I. W. Linkers for solid phase synthesis. Tetrahedron 1999, 55, 4855–4946.

205. Guillier, F.; Orain, D.; Bradley, M. Linkers and cleavage strategies in solid-phase organic syn-

thesis and combinatorial chemistry. Chem. Rev. 2000, 100, 2091–2157.

206. Comely, A. C.; Gibson, S. E. Tracelessness unmasked: a general linker nomenclature. Angew.

Chem. Int. Ed. 2001, 40, 1012–1032.

207. Barnes, C.; Balasubramanian, S. Recent developments in the encoding and deconvolution of

combinatorial libraries. Curr. Opin. Chem. Biol. 2000, 4, 346–350.

208. Afﬂ eck, R. L. Solutions for library encoding to create collections of discrete compounds. Curr.

Opin. Chem. Biol. 2001, 5, 257–263.

209. Gravert, D. J.; Janda, K. D. Organic synthesis on soluble polymer supports: liquid-phase meth-

odologies. Chem. Rev. 1997, 97, 489–509.

210. Toy, P. H.; Janda, K. D. Soluble polymer-supported organic synthesis. Acc. Chem. Res. 2000,

33, 546–554.

211. An, H.; Cook, P. D. Methodologies for generating solution-phase combinatorial libraries. Chem.

Rev. 2000, 100, 3311–3340.

212. Baldino, C. M. Perspective articles on the utility and application of solution-phase combinato-

rial chemistry. J. Comb. Chem. 2000, 2, 89–103.

213. Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach, A. G.; Longbottom, D. A.;

Nesi, M.; Scott, J. S.; Storer, R. I.; Taylor, S. J. Multi-step organic synthesis using solid-sup-

ported reagents and scavengers: a new paradigm in chemical library generation. J. Chem. Soc.,

Perkin Trans. 1 2000, 3815–4196.

214. Curran, D. P. Strategy-level separations in organic synthesis: from planning to practice. Angew.

Chem. Int. Ed. 1998, 37, 1175–1196.

215. Hughes, I.; Hunter, D. Techniques for analysis and puriﬁ cation in high-throughput chemistry.

Curr. Opin. Chem. Biol. 2001, 5, 243–247.

216. Kassel, D. B. Combinatorial chemistry and mass spectrometry in the 21st century drug discov-

ery laboratory. Chem. Rev. 2001, 101, 255–267.

217. Swartz, M. (Ed.), Analytical Techniques in Combinatorial Chemistry, Marcel Dekker, New

York, 2000.

218. Hardin, J. H.; Smietana, F. R. Automating combinatorial chemistry: a primer on benchtop ro-

botic systems. Mol. Diversity 1995, 1, 270–274.

219. Cargill, J. F.; Lebl, M. New methods in combinatorial chemistry: robotics and parallel synthe-

sis. Curr. Opin. Chem. Biol. 1997, 1, 67–71.

220. Antonenko, V. V. Automation in combinatorial chemistry. In Combinatorial Chemistry and

Technology, Miertus, S.; Fassina, G. (Eds.), Marcel Dekker, New York, 1999, pp. 205–232.

221. Bondy, S. S. The role of automation in drug discovery. Curr. Opin. Drug Discov. Dev. 1998, 1,

116–119.

222. Hoyle, W. (Ed.), Automated Synthetic Methods for Speciality Chemicals, Royal Society of

Chemistry, Cambridge, 2000.

223. Floyd, C. D.; Leblanc, C.; Whittaker, M. Combinatorial chemistry as a tool for drug discovery.

Prog. Med. Chem. 1999, 36, 91–168.

224. Houghten, R. A.; Pinilla, C.; Appel, J. R.; Blondelle, S. E.; Dooley, C. T.; Eichler, J.; Nefzi,

A.; Ostresh, J. M. Mixture-based synthetic combinatorial libraries. J. Med. Chem. 1999, 42,

3743–3778.

REFERENCES 167

169

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

PARALLEL SOLUTION-PHASE

SYNTHESIS

NORTON P. PEET AND HWA -OK KIM

CreaGen Biosciences, Inc.

Woburn, Massachusetts

5.1 INTRODUCTION

The concept of performing tasks in parallel rather than in series is simple. Parallel processing

was the central concept of industrialization as it dawned in this country, and assembly lines

made it possible to produce commercial goods in multiples rather than one at a time. The

concept as applied to organic and medicinal chemistry debuted in the 1990s and is termed

parallel solution-phase synthesis (PSPS). In this chapter we trace some personal history

with this very useful technique and highlight important contributions that have been made

by others to the methodology of PSPS.

It is very clear that PSPS is an extremely versatile tool for the simultaneous construc-

tion of related molecules. A variety of equipment has been employed for performing this

commonsense chemistry, ranging from nonautomated procedures with multiple reaction

vessels to fully automated programmable medicinal chemistry synthesis instruments. In

between these extremes is semiautomated equipment such as reaction blocks (96-well and

smaller) and partially programmable devices. The latter are perhaps the most useful to the

medicinal chemist who is intent on building relatively small, focused libraries. It is not in

the scope of this review to describe the equipment that has been developed and used for

PSPS, but rather, to present some of the chemistry that has been enabled with these tools.

5.2 AHEAD OF OUR TIME

Several reports from our laboratories from the 1980s demonstrate that we were performing

parallel chemistries in solution phase. The enactment of these chemistries was clearly more

170 PARALLEL SOLUTION-PHASE SYNTHESIS

behavioral than technological. Perhaps it was this early mindset, however, in our labora-

tories and those of others, that led to the widespread use of automated and semiautomated

techniques for doing parallel chemistries in solution in the next decade.

Our early PSPS experiments were earmarked by simple shortcuts that we employed.

These did not include special equipment or reagent-choosing software, but they did involve

the parallel use of standard glassware and notebook entries in which all experimental and

analytical data for a series of compounds were reported in tabular form in one entry rather

than as single entries.

To investigate specially substituted 1,2,4-triazolo[4,3-b]pyridazines with potential bron-

chodilator activity, we treated 3,4,5-trichloropyridazine (1) with secondary amines to pre-

pare requisite starting materials.

The reactive intermediate (1) underwent a regiospeciﬁ c

monodisplacement reaction at the 3-position to give compounds (2) with two equivalents

of amine and double displacement with excess amine to provide the 3,5-disubstituted prod-

ucts (3), as shown in Scheme 5.1.

Purity ranged from 78 to 100% for these conversions and was determined by

H nuclear

magnetic resonance (NMR) spectroscopy prior to puriﬁ cation. Since the positions for

disubstitution reactions of 1 with other nucleophiles produced products with substituents on

adjacent positions, the structure of 3 (NR

 1-pyrrolidinyl) was unequivocally determined

by hydrogenolysis of the chloro group and demonstrating by

H NMR spectroscopy that

the pyridazine protons were meta-coupled.

Another early example of PSPS from our laboratories was the preparation of a series

of 4-(1,3,4-oxadiazol-2-yl)-N,N-dialkylbenzenesulfonamides (7). This was an example of

a 2  2  4 cross PSPS. Carboxybenzenesulfonyl chlorides (4) were treated with two dif-

ferent secondary amines to produce benzenesulfonamides (5), which were then esteriﬁ ed

to give the corresponding ethyl esters (6). Treatment of the esters with hydrazine hydrate

gave the hydrazides (7), which were converted to oxadizoles (8) upon cyclization with four

different orthesters,

as shown in Scheme 5.2.

Another early example of PSPS from our laboratories involved the synthesis of several

key intermediates, which were new reactive intermediates for the preparation of tricyclic

systems, using a novel double displacement reaction.

The reactive intermediates (10) were

prepared by treating benzoyl chloride (8) with S-methyl cyclic thioureas (9). Subsequent

double displacement reactions with hydrazines gave the imidazo[2,1-b]quinozolin-5-(3H)-

ones (11) as shown in Scheme 5.3.

This novel cyclization was an example of a 3  2  5 cross PSPS. In addition, it was

shown that other aroyl chlorides would substitute for the benzoyl chlorides in this scheme

to produce a variety of tricyclic systems. Selected compounds that were prepared in this

PSPS protocol were evaluated as musculatropic bronchodilator agents in an artiﬁ cially

insufﬂ ated guinea pig animal model.

EtOH

excess HNR

2 equiv. HNR

Scheme 5.1

We also prepared compounds for respiratory research that were very effective in a rat

passive cutaneous anaphylaxis (PCA) model. These compounds were active as mediator

release inhibitors in this model,

and one of these agents (MDL 427)

was studied exten-

sively in phase I human clinical trials.

Again, 2-nitrobenzoyl chlorides (12) were starting materials of choice for the preparation

of a 6  4 cross of quinazolinones bearing a tetrazole substituted at the 3-position, all of

OEt

NHNH

meta or para position

1. R

NH, H

2. HCl

ClO

, R

= Me or Et

EtOH, HCl

NNH

C(OEt)

Scheme 5.2

SMe

NNR

diglyme or neat

SMe

HNNR

= F, NO

= H, Cl, CH

( )

n = 1, 2

( )

pyridine

= 4-morpholinyl, 1-piperidinyl,

1-homopiperidinyl, dimethylamino,

benzylamino

Scheme 5.3

AHEAD OF OUR TIME 171

172 PARALLEL SOLUTION-PHASE SYNTHESIS

which were active as antiallergic agents. As shown in Scheme 5.4, acid chlorides (12) were

treated with 5-aminotetrazole to produce the tetrazolyl amide (13). Catalytic reduction of

the sodium salts of 13 in aqueous media cleanly gave the corresponding anthranilamides

(14), which were cyclized with orthoesters to produce the 3-(1H-tetrazol-5yl)-4(3H)-

quinazolinones (15). Compounds 15 were converted to sodium salts to provide water-

soluble agents, which greatly facilitated their administration to animals for biological

evaluation.

We were also interested in adenosine receptor ligands for the treatment of respiratory

diseases and other disorders. The preparation of 8-substituted xanthines allowed us to use

PSPS for the preparation of a group of designed adenosine antagonists.

Key intermediate

was coupled with 16 different carboxylic acids, using the mixed anhydride method, to

provide the pyrimidinones (17) in which the amino group at the 5-position was regioselec-

tively acylated. To protect the chiral integrity of the acyl substituent during construction

of the xanthine nucleus, we prepared imino ethers (18) using Meerwein’s reagent under

relatively mild conditions. Imino ethers (18) could then be cyclized, again under the same

relatively mild conditions, to the (very) selective adenosine A

receptor antagonists (19).

This route, which is shown in Scheme 5.5, provided a novel and very convenient route to

xanthines bearing chiral substituents at the 8-position. Interestingly, the facile closure of 18

to 19 constitutes a 5-endo-trig anti-Baldwin ring closure.

8,9

Alternatively, when racemiza-

tion of the acyl group was not an issue, we were able to effect the closure of intermediates

17 directly to 19 with ethanolic potassium hydroxide.

5.3 RECENT REPORTS OF PARALLEL SOLUTION-PHASE SYNTHESIS

One of the companies that spearheaded PSPS and built a successful business around an auto-

mated platform for performing PSPS is ArQule. Several reports from their laboratories cite

= 3-CH

, 4-CH

, 5-CH

3-OCH

, 5-Cl,

3,4-OCH

1. H

, Pd/C

1N NaOH

2. 1N HCl

5-aminotetrazole

= H, CH

, CH

, C

C(OEt)

Scheme 5.4

successful examples of PSPS in building diverse chemical libraries. These libraries include

triazine-based compounds containing both carbohydrates and peptides,

chalcone-based

screening libraries,

spiro[pyrrolidine-2,3'-oxindole] libraries,

and several others.

13,14

representative and typical instructive example of convergent automated PSPS is shown in

Scheme 5.6.

Chalcones were used as versatile, key intermediates for the construction

Pr R

OEt

R = (R)-1-(phenylmethyl)ethyl, (S)-1-(phenylmethyl)ethyl, 1-(phenylmethyl)ethyl,

(R)-1-phenylpropyl, (S)-1-phenylpropyl, 1-phenylpropyl,

(R)-1-phenylethyl, (S)-1-phenylethyl, 1-phenylethyl, 1-(phenylmethyl)propyl,

1-(phenylmethyl)butyl, 2-indanyl, 1,2,3,4-tetrahydro-2-naphthyl,

1-(hydroxymethyl)-2-phenylethyl,

trans-

2-phenylcyclopentyl

RCOOH

mixed anhydride coupling

OBF

benzene

reflux

1. silica gel chromatography

2. benzene, reflux

Scheme 5.5

RECENT REPORTS OF PARALLEL SOLUTION-PHASE SYNTHESIS 173

EtOH

HN N

NHN

EtOH

NOH

EtOH

Scheme 5.6

174 PARALLEL SOLUTION-PHASE SYNTHESIS

of several heterocycles. Treatment of 1280 chalcones with hydroxylamines gave the 1280

isoxazolines (21). Additional diversity was added by cyclization of compounds (20) with

substituted phenylhydrazines to produce 7680 pyrazolines 22. Similarly, addition of substi-

tuted aminobenzamides to enones (20) gave a library of 7680 fused pyrimidines (23).

Many laboratories are presently using PSPS as an integral part of their discovery pro-

grams. In some research organizations, PSPS is part of discovery chemistry groups. In

other organizations, PSPS is a part of collaborative units, such as high-throughput or high-

speed chemistry groups, SWAT teams, or combinatorial chemistry groups. A few universi-

ties are now engaging and teaching students to use PSPS. There are numerous examples

of successful PSPS campaigns, both published and unpublished. In this review we do not

attempt to report these successes comprehensively, but instead, highlight a few representa-

tive recent examples from both our laboratories and other laboratories.

A series of antimalarial rhodacyanine dyes has recently been reported

employing a

3  3  3 cross PSPS as displayed in Scheme 5.7. Methylthioiminium salts (24) were con-

densed with rhodamines (25) to give vinylogous amides (26). Tosylated salts (27) were

prepared and coupled with iminium salts (28) to afford the ﬁ nal library of compounds (29).

A representative compound with high antimalarial activity is MKH-57.

A hit compound arising from an anti-infective screen of 1500 higher plants

provided

the rationale for building the library shown in Scheme 5.8. Dimethoxybenzoyl chlorides

N+

SMe

N+

SMe

TsO

N+

MKH-57: high antimalarial activity

TsO

CN, rt

TsOMe, DMF

120

N, CH

Scheme 5.7

(30) were treated with triphenylphosphonium ylides (31) to provide, after acylation and

intramolecular Wittig cyclization, benzo-fused heterocycles (33). Subsequent treatment of

33 with pyridinium bromide hydrobromide produced the corresponding didemethylated

compounds (34). The focused array of compounds represented by general structure 34 was

evaluated in a panel of antifungal assays.

A small library of ﬂ avones was prepared

using PSPS in a 4  9 cross sequence as

shown in Scheme 5.9 and evaluated as high-afﬁ nity benzodiazepine receptor ligands. This

PPh

TEA, PhMe,

heat

pyridinium bromide

hydrobromide

X = CH

or CH; NH or N

Y= O, S, NH

CO/HO groups: 1,2; 1,3; 2,3; 2,4

R = H, F, Cl, CN,

-Pr, others

Scheme 5.8

pyridine

pyridine,

KOH

, HOAc

= H, F, Cl, Br

= 2-F, 3-F, 4-F, 3-Cl, 3-Br, 4-Br, 3-OMe, 4-OMe, 4-NO

Scheme 5.9

RECENT REPORTS OF PARALLEL SOLUTION-PHASE SYNTHESIS 175

176 PARALLEL SOLUTION-PHASE SYNTHESIS

library was prepared from acetophenones (35) and benzoyl chlorides (36), which produced

a series of phenyl esters (37). Esters 37 were converted to β-diketones (38) with pyridine

and potassium hydroxide. Subsequent cyclization of 38 under acidic conditions gave 36

ﬂ avones of general structure 39. Interestingly, to facilitate a screening protocol, nine mix-

tures of four compounds were also produced, using the same synthetic protocol but em-

ploying a mixture of all four acetophenones with each acid chloride.

We have employed PSPS to prepare a substantial set of what we refer to as ade-

nine scaffold-derived cyclin-dependent kinase (cdk2 and cdk4) inhibitors, as shown in

Scheme 5.10. These compounds were evaluated, and many shown to be active in in vitro

tumor cell proliferation models for breast, colon, and prostate cancer.

After deﬁ ning

optimal substituents for the 2- and 9-positions of the adenine scaffolds, we used PSPS to

optimize the substituent at the 6N-position. We knew from modeling and x-ray crystal-

lographic studies

that substituents at this position protruded out of the enzyme active

site and into a solvent-accessible site, and thus referred to it as the ADME handle. We

could, therefore, adjust this substituent to affect ADME parameters such as solubility and

metabolism without greatly affecting enzyme inhibitory activity.

Our two-step one-pot PSPS products initiated from 2,6-dichloro-9-cyclopentylpurine

(41), which was easily accessible from dichloropurine (40) by treatment with cyclopenta-

nol under Mitsunobu conditions. Treatment of 41, as shown in Scheme 5.10, with a variety

of 4-amino-1-benzylpiperidines provided the 6,9-disubstituted purines (42) in situ. Sub-

sequent treatment of compounds 42 with trans-1,4-diaminocyclohexane, after removal of

solvent and heating to 150 C, gave the target compounds 43 in quite good overall yields.

In other recent studies we have developed process chemistry and performed proof-

of-principle experiments on key intermediates for the preparation of PSPS libraries. Thus,

PPh

, DEAD, THF

C to rt

68%

neat, 150

50-85%

EtOH, reflux

>90%

R = Ph; MDL108,522 (H717)

Scheme 5.10

under Swern conditions, we have optimized the conversion of amino alcohols (45), pre-

pared from 3,6-dichloropyridazine (44) by displacement with simple amino alcohols, to

imidazo[1,2-b]pyridazines (46) and minimized the production of ketones (47).

Imidazo-

pyridazines (46) are useful starting points for PSPS, for the introduction of a wide variety

of substituents at the 6-position, after nucleophilic displacement using a palladium cross-

coupling procedure (Scheme 5.11).

We have also extended the Iqbal multicomponent procedure

21–23

for preparing

2-[(acetylamino)methyl]-1,3-dicarbonyl compounds to the preparation of heterocycles.

Thus, compounds 49 were prepared from ethyl acetoacetate (48) by treatment with various

substituted benzaldehydes, acetyl chloride, and acetonitrile. As proof-of- principle for the

conversion of 49 to heterocycles, which can be accomplished using PSPS, we treated 49

(R  H) with phenylhydrazine to produce pyrazolone (50), as shown in Scheme 5.12.

There are only a limited number of multicomponent condensations, which are general

reactions that reliably produce the desired target compounds and are amenable to parallel

synthesis. Two other multicomponent condensations that fall into this category are the

Ugi

and the Passerini

reactions. The Passerini reaction involves the treatment of an

isocyanide with a carboxylic acid and an aldehyde or a ketone. The Ugi reaction uses

Cl Cl

EtOH

reflux

(COCl)

DMSO, TEA

, -78

= H, CH

, COPh

= H, CH

, Ph, CH

OPh, others

Scheme 5.11

C OEt

O O

O NHAc

EtOOC

AcCl

4-R-PhCHO

AcHN

R = Cl, CO

Me, NO

, Me, OMe

NNHPh

R = H

Scheme 5.12

RECENT REPORTS OF PARALLEL SOLUTION-PHASE SYNTHESIS 177

178 PARALLEL SOLUTION-PHASE SYNTHESIS

all of these components plus ammonia or an amine. Another multicomponent reaction is

the Biginelli condensation, whose reactants are aldehydes, ketones with activated methy-

lene groups, and ureas or thioureas.

27–29

A recent example of microwave-assisted Biginelli

condensations, using a variety of catalysts to optimize yields, has been reported.

Microwave-assisted parallel synthesis (MAPS) is a combination of new technologies,

which will undoubtedly be used for many future applications.

Several excellent review articles on PSPS have appeared. One of these articles focuses

on PSPS and solid-phase strategies that use natural product templates, and covers PSPS

strategies for libraries based on distamycin, ﬂ avonoids, kramerixin (benzofurans), map-

picine, curacin, stipiamide, sarcodictine, and taxoids.

Several additional articles are cited

in Section 5.4.

Many additional examples of PSPS, which have been applied to the synthesis of me-

dium-sized target libraries with biological activity, are shown in Table 5.1. It is very clear

that PSPS facilitated the time-efﬁ cient preparation of these focused libraries and has de-

creased the cycle time for hit-to-lead and lead-to-high value lead iteration.

5.4 SOLID SUPPORTED REAGENTS, SCAVENGERS, AND CATALYSTS

A critically important supporting technology for parallel solution-phase synthesis is that

of functionalized polymers. There are three classes of functionalized polymers: solid-

supported reagents (SSRs), solid-supported scavengers (SSSs), and solid-supported cata-

lysts (SSCs). It is these reagents that allow solution-phase synthesis to produce reaction

solutions and products that are ultimately quite clean, simply by allowing excess reagent

or catalyst to be removed by ﬁ ltration. Similarly, excess reagents that are not resin-bound

can be trapped by designed SSSs and removed by ﬁ ltration, as can by-products of the reac-

tion or unreacted starting materials. Combinations of SSRs, SSSs, and SSCs can be used

for PSPS steps. Indeed, the use of these reagents is what allows the execution of multistep

solution-phase synthetic procedures.

In our laboratory we designed a resin-based reagent for diazo transfer,

related to an

earlier reagent that was reported by the Rebek group.

Our reagent is polystyrene-sup-

ported benzenesulfonyl azide (52), which is thermally stable and not friction sensitive.

Resin-bound reagent 52 was prepared by treatment of commercially available polymer-

supported benzenesulfonyl chloride with sodium azide in aqueous dimethylformamide

at room temperature. Transfer of the diazo group from 52 to a variety of active methy-

lene compounds (53) was demonstrated to produce diazo compounds (54), as shown in

Scheme 5.13.

A recent example of PSPS which employs both resin-bound reagents and scavengers is

shown in Scheme 5.14 and describes the syntheses of trisubstituted benzimidazolones.

Thus, 2-ﬂ uoronitrobenzenes (55) were treated with primary amines to produce anilines

(56) via Meisenheimer displacement. Excess 2-ﬂ uoronitrobenzenes (55) were removed us-

ing the polyamine scavenger A. Treatment of 2-nitroanilines (56) with Raney nickel gave

phenylenediamines (56), which were cyclized to benzimidazolones (58) with 1,1'-carbon-

yldiimidazole (CDI). Alkylation of benzimidazolones (58) was achieved by treatment with

resin-bound base A and an alkyl bromide, followed by removal of the excess alkyl bromide

with resin-bound thiourea scavenger B. This overall procedure allowed the production of

an array of compounds using puriﬁ cation strategies that were differentiated completely

from conventional techniques.

Entry Library Structure Biological Activity Comments Ref.

1 Cyclic (depsi)peptides HUN-7293 analogs Inhibition of cell adhesion

molecule expression

VCAM-1, ICAM-1, and E-selectin are adhesion

molecules that can induce inﬂ ammation.

2 Spiperone analogs

N N

5-HT

/D2 ligands Improved selectivity for 5-HT

with respect to

spiperone was achieved.

3 Acylguanidines

R N N

O NH

Sodium channel blockers Selected compounds were evaluated in the

audiogenic mouse antiseizure model.

4 N-(1-Phenylethyl)-5-

phenylimidazole-2-

amines



ATPase inhibitors

Inhibitors have potential utility for congestive

heart failure (CHF).

5 5-Carboxamido-1-benzyl-

(3-dimethylamino-

propyloxy-1H-pyrazoles

Activators of soluble guanylate

cyclase

Compounds also inhibit platelet aggregation

and display oral bioavailability in this

lipophilicity-restrained library.

6 2,4,6-Trisubstituted

quinazolines

Cdk4/cyclin D1 and cdk2/

cyclin E inhibitors

X-ray structure of inhibitor bound to cdk2 was

obtained.

TABLE 5.1 Parallel Solution-Phase Synthesis Libraries

(Continued)

179

Entry Library Structure Biological Activity Comments Ref.

7 3-Substituted indoles

OCH

Neurokinin-1 (NK-1)

antagonists

Subnanomolar afﬁ nity, orally bioavailable

compound was obtained.

8 Proline derivatives

(

)

FKBP12 inhibitors Compounds are nanomolar inhibitors of

peptidyl-prolyl isomerase (PPIase or

rotamase).

9 N-Aryl-N-oxalylanthranilic

acids

HOOC

NHR

Protein tyrosine phosphatase

1B (PPT1B) inhibitors

Selectivity is observed over T-cell PTPase

(TCPTP).

10 CC-1065 analogs

NHBoc

Sequence selective alkylation

of duplex DNA

CC-1065 and the related duocarmycins are

potent antitumor antibiotics.

TABLE 5.1 (Continued)

180

11 Tethered dimers

-(CH

)

NAD synthetase inhibitors

Compounds also inhibited growth of gram ()

bacteria.

12 p-Acylthiocinnamides

Antagonists of leukocyte

function-associated

antigen-1/intracellular

adhesion molecules-1 (LFA/

ICAM-1) interaction

Compounds have potential utility for

inﬂ ammatory diseases, autoimmune disorders,

tumor metastasis, allograft rejection, and

reperfusion injury.

13 3-Imidazolylcarbolines

Somatostatin antagonists Nonpeptidic antagonists are sst

selective with

potential to treat acromegaly, neuroendocrine

tumors and gastrointestinal disorders.

14 β-C-Mannosides

Selectin inhibitors The tetrasaccharide sialyl Lewis

is the

carbohydrate epitope recognized by selectins

E, P, and L.

15 Pyrrolidines

COOHH

O N

Inﬂ uenza neuraminidase (NA)

inhibitors

NA is necessary for virus replication and

infectivity; x-ray structure determined for

inhibitor A-192558.

(Continued)

181

Entry Library Structure Biological Activity Comments Ref.

16 N-(Pyrroldinylmethyl)

hydroxamic acids

Transition-state inhibitors of

fucosyltransferases

Potential therapeutic areas include inﬂ ammatory

diseases and cancer metastasis.

17 Mercaptomethyl ketones

Mechanism-based inhibitors of

cysteine proteinases

A seven-step synthesis of substrate-based

ketone inhibitors was developed; calpains

are implicated in neurodegenerative diseases

and osteoporosis; caspases are involved with

programmed cell death.

18 Kerolides

macrocycle

These semisynthetic

erythromycins have potential

for treating macrolide MLS

resistance

Puriﬁ cation of library was accomplished using

ion exchange.

19 Tetrahydrofurans

Lead discovery library One-step construction of diverse heterocycles

via [3  2] cycloaddition reaction.

20 Oxazoles

Lead discovery library One-step construction of heterocycles using

p-toluenesulfonylmethylisocyanide (TosMIC).

21 1,2-Phenethyldiamines

NHR

Lead generation library A wide range of biological activities has been

reported for this class of compound.

TABLE 5.1 (Continued)

182

22 Imidazolinones

None reported Structures produced resemble the natural

product creatine.

23 Spiropyrrolopyrroles

Neurokinin (NK) receptor

antagonists

Low nanomolar afﬁ nities for the NK-1 receptor

were achieved.

24 3,4-Dihydroquinoxalin-

2-ones

Aldose reductase inhibitor,

antagonism of AMPA and

angiotensin II receptors

Combination of solution- and solid-phase

synthesis was used.

25 Purines

OCH

Potential inhibitors of

ATP-dependent proteins

In addition, Suzuki coupling of aryl boronic

acids produced N

-aryl derivatives.

26 3-Arylindoles

Potential inhibitors of

topoisomerase, related to

BE10988

Suzuki coupling strategy was used. 57

183

184 PARALLEL SOLUTION-PHASE SYNTHESIS

Another recent report, which used polymer-bound reagents, bases, and scavengers in

concert with PSPS, described the preparation of peptidomimetic pyrazinone antithrombot-

ics (Scheme 5.15).

These pyrazinones were shown to be inhibitors of tissue factor VIIa

complex. Treatment of pyrazinone (60) with HOBt (61) in the presence of resin-bound

carbodiimide (reagent A) gave the activated pyrazinone (62), which then was coupled with

primary amines to yield amides (63). The reaction mixture was treated with scavenger A

NaN

O/DMF

52: PS-SO

O O

PS-SO

(52)

N, CH

, rt

O O

a: R

= CH

O, R

= OCH

b: R

= CH

O, R

= C

c: R

= R

= OCH

d: R

= CH

O, R

= OC(CH

)

e: R

= R

= OC(CH

)

and others

54a-e

PS-SO

53a-e

Scheme 5.13

NHR

t-Bu

NEt

= H, CO

= alkyl, arylakyl

= alkyl, phenacyl, arylmethyl, others

1. R

, DMF, rt

2. scavenger A

, Raney Ni

MeOH

base A

Br, DMF

scavenger B

Scavenger A base A

Scavenger B

Scheme 5.14

to sequester unreacted 60 and scavenger B to capture excess HOBt (61). The sequestering

agents in this and related protocols allowed the preparation of hundreds of pure compounds

without chromatographic puriﬁ cation.

In Table 5.2 are compiled representative resin-bound agents that have been employed in

PSPS formats. From inspection of the classiﬁ cation column it is evident that a wide vari-

ety of processes can be mediated with these resin-bound agents. Included are resin-bound

reagents such as a wide variety of oxidizing and reducing agents; nucleophiles of various

kinds for displacement reactions; coupling reagents such as Wittig, and peptide coupling

reagents; hydrazine delivery reagents, using a catch and release method; and many others.

Deprotection, dehydration, isomerization, and catalysis are additional processes that are

mediated by these resin-bound agents.

Resin-bound acids and bases are employed for initiating reactions, absorbing bases or

acids that are generated by PSPS processes, or for scavenging excess or unreacted start-

ing materials or by-products that are produced. A variety of other scavengers has been

developed, such as immobilized isocyanates for removal of reactive amines in case of acid

sensitivity and immobilized thiols for the removal of ketones.

Interestingly, PSPS reagents or scavengers can be tagged for removal by resin-bound

agents. Entries 20 and 21 in Table 5.2 exemplify these processes. Thus, a soluble carbodi-

imide reagent can be tagged with a tertiary amine, which can be removed by an ion-exchange

resin (e.g., a resin-bound sulfonic acid). Similarly, tetraﬂ uorophthalic anhydride, which is

used to scavenge excess secondary amines, produces a tagged species after scavenge oc-

curs. The benzoic acid that is produced by opening the anhydride with the secondary amine

is a tagged scavenger, which can be removed with a resin-bound primary amine.

COOH

reagent A

DMF/CH

OBt

HOBt

NHR

DMF, NMM

60 + 61 +

scavenger A

scavenger B

Pure 63

Reagent A

Scavenger A

Scavenger B

Scheme 5.15

SOLID SUPPORTED REAGENTS, SCAVENGERS, AND CATALYSTS 185

Entry Compound Class Prepared Resin-Bound Agent Used Classiﬁ cation of Agent Comments Refs.

1 Carbamates

Reagent (activation) Catch and release reagent. 62

2 Secondary amines

CrO

Reagent (oxidation) Tandem three-phase reaction. 63

3 Secondary amines

Reagent (reduction) Tandem three-phase reaction. 63

4 Viscinol diols

[4  2] Cycloaddition removal

9-Anthrylmethyl ester tag employed

for capture.

5 2-Pyrazolines

NCO

Scavenger (amine) Excess intermediate pyrazoline with

free NH was scavenged.

6 α-Ketothiazole

-2

Reagent (reduction) Excess Dess–Martin reagent is reduced

by thiosulfate resin.

()-Plicamine

OAc

Reagent (oxidation) Natural product plus its enantiomer

were prepared using thirteen resin-

bound agents, including the solid

supported iodonium diacetate.

TABLE 5.2 Resin-Bound Agents Used for PSPS

186

187

8 Carponone

(THF)

Ph Ph

Catalyst (isomerization) Reagent is an immobilized version of

Felkin’s iridium catalyst.

9 (R)-Salmeterol

PPh

Reagent (activation) Used with CBr

to convert alcohol to

alkyl bromide.

10 Bicyclo[2.2.2]octanes

Scavenger (ketone) Used in conjunction with resin-bound

diisopropylethylamine.

11 1,2,3-Thiadiazoles

NHNH

Reagent (hydrazine delivery) Catch and release reagent. 71

12 Oleﬁ ns

Reagent (coupling) Wittig reaction. 72–74

13 4-Phenyl-2H-phthalazin-

1-one

NMe

NEt

-t-

Reagent (deprotonation) PBEMP is one of several bases that

can be used for N-alkylation of

weakly acidic N-heterocycles.

14 α-Bromoketones; vicinal

dibromides

Reagent (halogenation) PM5VTHT. 76

15 α-Azidoalcohols; alkyl

azides

Reagent (nucleophile) Basic ion-exchange resins can be

loaded with both inorganic and

organic anions.

77, 78

(Continued)

188

Entry Compound Class Prepared Resin-Bound Agent Used Classiﬁ cation of Agent Comments Refs.

16 Alkyl and aryl nitriles

PPh

/CCl

Reagent (dehydration) Both carboxamides and aryl oximes

can be converted to nitriles with this

reagent.

17 α,β-Unsaturated esters

and nitriles

NMe

X P(OR)

Reagent (Horner-Emmons) This ion-pair reagent was the ﬁ rst

reported polymer-supported reagent.

18 α,β-Unsaturated esters

O P

OEt

COOEt

Reagent (Horner–Emmons) Barton’s base used to generate anion. 81

19 Bromohydrins and

iodohydrins

PPh

X= Br, I

Reagent (Mitsunobu) Polymer-bound triarylphosphine used

with DEAD.

82, 83

20 Aryl ethers

O N

OCH

Reagent (Mitsunobu) Polymer-bound DEAD used with PPh

. 84–87

21 Derivatives of secondary

amines

Reagent (tagging) This tagged carboxy amide is removed

with

88, 89

TABLE 5.2 (Continued)

22 Aldehydes and ketones

"tagged" by

C N

Reagent (tagging) Tagged carbodiimide reagent is

removed by ion exchange resin.

23 Perhydro-3-oxo-1,4-diaz-

epinium derivatives

Wang aldehyde HL resin Scavenger (amine) Microwave irradiation was employed

to increase scavenger reactivity.

24 Flavanoid cycloadducts

[4  2] Cycloaddition removal

Compounds prepared are related to

prenylﬂ avanoid Diels–Alder natural

products.

25 Amines (α-substituted)

Capture (amine) Capture of the product amines, derived

from alcoholysis of corresponding

sulfonamides, was microwave-as-

sisted.

26 Terminal acetylenes

EtO

: ROMP

Reagent (Horner–Wadsworth–

Emmons oleﬁ nation)

This ROMP gel-supported ethyl

1-diazo-2-oxopropylphosphonate is

one example of many ROMP agents.

189

190 PARALLEL SOLUTION-PHASE SYNTHESIS

Another approach to PSPS, which is closely related to the tagging methodology, is the

ﬂ uorous synthesis technique that was introduced in 1997.

This technique adds another

tool to the armamentarium of scavenging and capturing techniques represented by attach-

ment to polymer support and linkage to ionizable functionality. The ﬂ uorous synthesis

technique, simply described, is a third liquid phase (e.g., perﬂ uorohexanes), which is im-

miscible with both water and common organic solvents.

96–98

Typical organic compounds have virtual no solubility in the ﬂ uorous phase. They can,

however, be rendered soluble in the ﬂ uorous phase by attachment of a ﬂ uorous tag. In a

number of ﬂ uorous synthetic techniques, the tris(perﬂ uorohexylethyl)silyl group has been

employed as a convenient ﬂ uorous phase tag.

99,100

The synthesis of ﬂ uorous amine scaven-

ger (66) has been described, as shown in Scheme 5.16. The tris(perﬂ uorohexylethyl)silyl

group was doubly appended to N,N'-diallyltriﬂ uoroacetamide (64), and the resulting new

triﬂ uoroacetamide (65) was converted to an amine scavenger (66) by treatment with lithium

aluminum hydride. Compound 66 was shown to be a good scavenger for arylisocyanates,

for example, to produce ureas (67), which had been used in excess as reagents in the PSPS

synthesis of a library of ureas.

101

Fluorous tethered reagents have also been developed. A Staudinger protocol using

a ﬂ uorous tethered triarylphosphine (69) has been developed. In a one-pot, two-step

procedure, 4-azidobenzoic acid (68) was converted quantitatively to 4-aminobenzoic acid

(70) in a 4-hour time frame, as shown in Scheme 5.17. Puriﬁ cation was accomplished by

FluoroFlash

™

chromatography.

102

This Staudinger protocol was then applied successfully

to a structurally diverse group of azides. The use of ﬂ uorous tethered reagents with

FluoroFlash

™

chromatography has been reviewed.

103,104

O CF

)

SiH

PtCl

, 80

C, 12 h

33–37%

[(C

)

SiCH

]

N CF

[(C

)

SiCH

]

[(C

)

SiCH

]

N NHAr

LAH, ether

97%

ArNCO

Scheme 5.16

COOH

PPh

COOH

1) T

HF, rt, 1h

2) H

O, 60

C, 3h

3) FluoroFlash

quantitative yield

68 69 70

Scheme 5.17

Entry 26 describes one ROMP (ring-opening metathesis polymerization) reagent that

has been used for the preparation of terminal acetylenes from aldehydes.

ROMP re-

agents, which are prepared from monomeric scavengers or reagents by a metathesis polym-

erization process, include an N-hydroxysuccinimide coupling agent,

105

a primary alcohol

for scavenging a variety of electrophiles,

106

sulfonyl chlorides for scavenging a variety of

nucleophiles,

107

and bis-acid chlorides as nucleophile scavengers.

108

In addition, ROMP

oligomers have been used as solid supports for multistep reaction sequences

109

and also as

capture–release agents for the synthesis of amines and alkyl hydrazines.

110

An older topic,

which is not covered in this review, is the use of organic soluble supports and reagents. This

topic has been reviewed previously.

111,112

A number of extensive reviews have appeared that cover the use of resin-bound agents

in PSPS. A comprehensive review by Kirschning et al.

113

covers all classiﬁ cations of im-

mobilized agents, regardless of their function. This reference serves as the premier lead-

ing reference for the reader who wants a cursory view of any subgroup of agent and who

requires additional references to consult for more detail. In addition, other more special-

ized review articles deal with the topic of immobilized reagents,

114–116

catalysts,

117–119

and

scavenging agents.

120

5.5 THE FUTURE

It has become increasingly clear that the production of large random chemical libraries us-

ing combinatorial chemistry techniques has not been productive for the efﬁ cient generation

of hit compounds. Problems inherent in this approach include (1) lack of chemical diversity

of the compound sets; (2) the relatively high cost of screening protocols for meaningful

biological assays; and (3) the inattention to important compound properties such as solubil-

ity, cLogP, polar surface area, and number of rotatable bonds.

Thus, the continuing and increasing need is established for the production of focused,

carefully designed libraries with a high likelihood of generating high-quality hit or lead

compounds. These relatively small libraries, numbering in the hundreds rather than the

thousands, will contain leadlike

121–123

or druglike

124,125

molecules.

Semiautomated PSPS methods will be increasingly useful tools to generate these re-

quired focused libraries. To minimize the need for puriﬁ cation, which always is an issue

with solution-phase chemistry protocols, resin-bound reagents and scavengers will also

become increasingly useful (and less expensive).

Thus, in the future we will see a deﬁ nite shift to the production of small, thought-driven

target libraries for hit generation. Many of these will initiate from natural product scaffolds,

which will introduce new chemical diversity and inherent biological activities. Iterations of

small, focused libraries will produce higher-value compounds in a faster time frame than

was achieved from large random libraries.

REFERENCES

N. P. Peet, Synthesis of 7-(1-pyrrolidinyl)-1,2,4-triazolyl[4,3-b]pyridazines, J. Heterocycl.

Chem., 21, 1389–1392 (1984).

N. P. Peet and S. Sunder, Factors which inﬂ uence the formation of oxadiazoles from anthranil-

hydrazides and other benzoylhydrazines, J. Heterocycl. Chem., 21, 1807–1816 (1984).

REFERENCES 191

192 PARALLEL SOLUTION-PHASE SYNTHESIS

N. P. Peet, J. Malecha, M. E. LeTourneau, and S. Sunder, Preparation of imidazo[2,1-

b]quinazolin-5(3H)-ones and related tricyclic systems using a novel, double displacement reac-

tion, J. Heterocycl. Chem., 26, 257–264 (1989).

N. P. Peet, L. E. Baugh, S. Sunder, and J. E. Lewis, Synthesis and antiallergic activity of some

quinolinones and imidazoquinolinones, J. Med. Chem., 28, 298–302 (1985).

N. P. Peet, L. E. Baugh, S. Sunder, J. E. Lewis, E. H. Matthews, E. L. Olberding, and D. N.

Shah, 3-(1H-Tetrazol-5-yl-4(3H)-quinazolinone sodium salt (MDL 427): a new antiallergic

agent, J. Med. Chem., 29, 2403–2409 (1986).

N. P. Peet, N. L. Lentz, M. W. Dudley, A. M. L. Ogden, D. R. McCarty, and M. M. Racke,

Xanthines with C

chiral substituents as potent and selective adenosine A

antagonists, J. Med.

Chem., 36, 4015–4020 (1993).

J. W. Daly, W. L. Padgett, M. T. Shamim, P. Butts-Lamb, and J. Waters, 1,3-Dialkyl-8(p-

sulfophenyl)xanthines: potent, water-soluble antagonists for A

- and A

-adenosine receptors, J.

Med. Chem., 28, 487–492 (1985).

J. E. Baldwin, Rules for ring closure, J. Chem. Soc., Chem. Commun., 734–736 (1976).

J. E. Baldwin, J. Cutting, W. Dupont, L. Kruse, L. Silberman, and R. C. Thomas, 5-Endo-trigonal

reactions: a disfavoured ring closure, J. Chem. Soc., Chem. Commun., 736–738 (1976).

G. R. Gustafson, C. M. Baldino, M. E. O’Donnell, A. Sheldon, R. J. Tarsa, C. J. Verni, and D. L.

Coffen, Incorporation of carbohydrates and peptides into large triazine-based screening librar-

ies using automated parallel synthesis, Tetrahedron, 54, 4051–4065 (1998).

D. G. Powers, D. S. Casebier, D. Fokas, W. J. Ryan, J. K. Troth, and D. L. Coffen, Automated

parallel synthesis of chalcone-based screening libraries, Tetrahedron, 54, 4085–4096 (1998).

D. Fokas, W. J. Ryan, D. S. Casebier, and D. L. Coffen, Solution phase synthesis of a

spiro[pyrrolidine-2,3'-oxindole] library via a three component 1,3-dipolar cycloaddition reac-

tion, Tetrahedron, 39, 2235–2238 (1998).

C. M. Baldino, Perspective articles on the utility and application of solution-phase combinato-

rial chemistry, J. Comb. Chem., 2, 89–103 (2000).

C. M. Baldino, D. S. Casebier, J. Caserta, G. Slobodkin, C. Tu, and D. L. Coffen, Convergent

parallel synthesis, Synlett, 488 (1997).

K. Takasu, H. Terauchi, H. Inoue, H.-S. Kim, Y. Wataya, and M. Ihara, Parallel synthesis of

antimalarial rhodacyanine dyes by the combination of three components in one pot, J. Comb.

Chem., 5, 211–214 (2003).

R. A. Fecik, K. E. Frank, E. J. Gentry, L. A. Mitscher, and M. Shibata, Use of combinatorial and

multiple parallel synthesis methodologies for the development of anti-infective natural prod-

ucts, Pure Appl. Chem., 71, 559–564 (1999).

M. Marder, H. Viola, J. A. Bacigaluppo, M. I. Colombo, C. Wasowski, C. Wolfman, J. H.

Medina, E. A. Ruveda, and A. C. Paladini, Detection of benzodiazepine receptor ligands in

small libraries of ﬂ avone derivatives synthesized by solution phase combinatorial chemistry,

Biochem. Biophys. Res. Commun., 249, 481–485 (1998).

P. W. Shum, N. P. Peet, P. M. Weintraub, T. B. Le, Z. Zhao, F. Barbone, B. Cashman, J. Tsay,

S. Dwyer, P. C. Loos, E. A. Powers, K. Kropp, P. S. Wright, A. Bitonti, J. Dumont, and

D. R. Borcherding, The design and synthesis of purine inhibitors of CDK2: part 3, Nucleosides

Nucleotides, 20, 1067–1078 (2001).

M. K. Dreyer, D. R. Borcherding, J. A. Dumont, N. P. Peet, J. Tsay, P. S. Wright, A. J. Bitonti,

J. Shen, and S.-H. Kim, Crystal structure of human cyclin-dependent kinase 2 in complex with

the adenine-derived inhibitor H717, J. Med. Chem., 44, 524–530 (2001).

P. Raboisson, B. Mekonnen, and N. P. Peet, Tetrahedron Lett., 44, 2919–2921 (2003).

M. M. Reddy, B. Bhatia and J. Iqbal, A one pot synthesis of β-acetamido ketones and furans by

cobalt(II) catalyzed coupling of 1,3- or 1,4-dicarbonyl compounds and aldehydes: a remarkable

role of dioxygen, Tetrahedron Lett., 36, 4877–4880 (1995).

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

M. Mukhopadhyay, B. Bhatia, and J. Iqbal, Cobalt catalyzed multiple component condensation

route to β-acetamido carbonyl compound libraries, Tetrahedron Lett., 38, 1083–1086 (1997).

B. Bhatia, M. M. Reddy, and J. Iqbal, Cobalt-catalyzed three-component coupling involving

ketones or ketoesters, aldehydes and acetonitrile: a novel one-pot synthesis of β-acetamido

ketones, J. Chem. Soc., Chem. Commun., 6, 713–714 (1994).

S. M. Antonelli, M. Tandon, and N. P. Peet, Mechanistic and preparative studies of a multi-

component condensation reaction, Org. Lett., manuscript submitted for publication.

I. Ugi, Neuere Methoden der präparativen organischen Chemie IV. Mit sekundär-reaktionen

Gekoppelte α-Additionen von Immonium-Ionen und Anionen an Isonitrile, Angew. Chem. Int.

Ed. Engl., 1, 8–21 (1962).

M. Passerini, Passerini reaction, Gazz. Chim. Ital., 51, 126–181 (1921).

P. Biginelli, Aldehyde–urea derivatives of aceto- and oxaloacetic acids, Gazz. Chim. Ital., 23,

360–413 (1893).

C. O. Kappe, 100 Years of the Biginelli dihydropyrimidine synthesis, Tetrahedron, 49, 6937–

6963 (1993).

C. O. Kappe, Recent advances in the Biginelli dihydropyrimidine synthesis: new tricks from an

old dog, Acc. Chem. Res., 33, 879–888 (2000).

A Stadler and C. O. Kappe, Automated library generation using sequential microwave-assisted

chemistry: application toward the biginelli multicomponent condensation, J. Comb. Chem., 3,

624–630 (2001).

D. G. Hall, S. Manku, and F. Wang, Solution- and solid-phase strategies for the design, synthe-

sis and screening of libraries based on natural product templates: a comprehensive survey, J.

Comb. Chem., 3, 125–150 (2001).

Y. Chen, M. Bilban, C. A. Foster, and D. L. Boger, Solution-phase parallel synthesis of a phar-

macophore library of HUN-7293 analogues: a general chemical mutagenesis approach to deﬁ n-

ing structure–function properties of naturally occurring cyclic (depsi)peptides, J. Am. Chem.

Soc., 124, 5431–5440 (2002).

H. C. Hansen, R. Olsson, G. Croston, and C.-M. Andersson, Multistep solution-phase parallel

synthesis of spiperone analogues, Bioorg. Med. Chem. Lett., 10, 2435–2439 (2000).

S. Padmanabhan, R. C. Lavin, P. M. Thakker, J. Guo, L. Zhang, D. Moore, M. E. Perlman, C.

Kirk, D. Daly, K. J. Burke-Howie, T. Wolcott, S. Chari, D. Berlove, J. B. Fischer, W. F. Holt,

G. J. Durant, and R. N. McBurney, Solution-phase, parallel synthesis and pharmacological

evaluation of acylguanidine derivatives as potential sodium channel blockers, Bioorg. Med.

Chem. Lett., 11, 3151–3155 (2001).

B. E. Blass, C. T. Huang, R. M. Kawamoto, M. Li, S. Liu, D. E. Portlock, W. M. Rennells,

and M. Simmons, Parallel synthesis and evaluation of N-(1-phenylethyl)-5-phenylimidazole-2-

amines as Na



ATPase Inhibitors, Bioorg. Med. Chem. Lett., 10, 1543–1545 (2000).

D. L. Selwood, D. G. Brummell, R. C. Glen, M. C. Goggin, K. Reynolds, M. A. Tatlock, and

G. Wishart, Solution-phase parallel synthesis of 5-carboxamido 1-benzyl-3-(3-dimethylamino-

propyloxy)-1H-pyrazoles as activators of soluble sodium guanylate cyclase with improved oral

bioavailability, Bioorg. Med. Chem. Lett., 11, 1089–1092 (2001).

T. M. Sielecki, T. L. Johnson, J. Liu, J. K. Muckelbauer, R. H. Grafstrom, S. Cox, J. Boylan,

C. R. Burton, H. Chen, A. Smallwood, C.-H. Chang, M. Boisclair, P. A. Benﬁ eld, G. L. Trainor,

and S. P. Seitz, Quinazolines as cyclin dependent kinase inhibitors, Bioorg. Med. Chem. Lett.,

11, 1157–1160 (2001).

J. E. Fritz, P. A. Hipskind, Karen L. Lobb, J. A. Nixon, P. G. Threlkeld, B. D. Gitter, C. L.

McMillian, and S. W. Kaldor, Expedited discovery of second generation NK-1 antagonists: iden-

tiﬁ cation of a nonbasic aryloxy substituent, Bioorg. Med. Chem. Lett., 11, 1643–1646 (2001).

C. Choi, J.-H. Li, M. Vaal, C. Thomas, D. Limburg, Y.-Q. Wu, Y. Chen, R. Soni, C. Scott,

D. T. Ross, H. Guo, P. Howorth, H. Valentine, S. Liang, D. Spicer, M. Fuller, J. Steiner, and

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

REFERENCES 193

194 PARALLEL SOLUTION-PHASE SYNTHESIS

G. S. Hamilton, Use of parallel-synthesis combinatorial libraries for rapid identiﬁ cation of po-

tent FKBP12 inhibitors, Bioorg. Med. Chem. Lett., 12, 1421–1428 (2002).

Z. Xin, T. K. Oost, C. Abad-Zapatero, P. J. Hajduk, Z. Pei, B. G. Szczepankiewicz, C. W.

Hutchins, S. J. Ballaron, M. A. Stashko, T. Lubben, J. M. Trevilly, M. R. Jirousek, and G. Liu,

Potent, selective inhibitors of protein tyrosine phosphatase 1B, Bioorg. Med. Chem. Lett., 13,

1887–1890 (2003).

D. L. Boger, H. W. Schmitt, B. G. Fink, and M. P. Hedrick, Parallel synthesis and evaluation

of 132 ()-1,2,9,9a-tetrahydrocyclopropa[c]benz[e]indol-4-one (CBI) analogues of CC-1065

and the duocarmycins deﬁ ning the contribution of the DNA-binding domain, J. Org. Chem., 66,

6654–6661 (2001).

S. E. Velu, W. A. Cristofoli, G. J. Garcia, C. G. Brouillette, M. C. Pierson, C.-H. Luan, L. J.

DeLucas, and W. J. Brouillette, Tethered dimers as NAD synthetase inhibitors with antibacterial

activity, J. Med. Chem., 46, 3371–3381 (2003).

G. Liu, J. T. Link, Z. Pei, E. B. Reilly, S. Leitza, B. Nguyen, K. C. Marsh, G. F. Okasinski, T. W.

von Geldern, M. Ormes, K. Fowler, and M. Gallatin, Discovery of novel p-arylthio cinnamides

as antagonists of leukocyte function-associated antigen-1/intracellular adhesion molecule-1 in-

teraction. 1: identiﬁ cation of an additional binding pocket based on an anilino diaryl sulﬁ de

lead, J. Med. Chem., 43, 4025–4040 (2000).

L. Poitout, P. Roubert, M.-O. Contour-Galcéra, C. Moinet, J. Lannoy, J. Pommier, P. Plas,

D. Bigg, and C. Thurieau, Identiﬁ cation of potent non-peptide somatostatin antagonists with

sst

selectivity, J. Med. Chem., 44, 2990–3000 (2001).

N. Kaila, L. Chen, B. E. Thomas, D. Tsao, S. Tam, P. W. Bedard, R. T. Camphausen, J. C. Alvarez,

and G. Ullas, β-C-Mannosides as selectin inhibitors, J. Med. Chem., 45, 1563–1566 (2002).

G. T. Wang, Y. Chen, S. Wang, R. Gentles, T. Sowin, W. Kati, S. Muchmore, V. Giranda, K.

Stewart, H. Sham, D. Kempf, and W. G. Laver, Design, synthesis and structural analysis of

inﬂ uenza neuraminidase inhibitors containing pyrrolidine cores, J. Med. Chem., 44, 1192–1201

(2001).

M. Takayanagi, T. Flessner, and C.-H. Wong, A strategy for the solution-phase synthesis of N-

(pyrrolidinylmethyl)hydroxamic acids, J. Org. Chem., 65, 3811–3815 (2000).

A. Lee and J. A. Ellman, Parallel solution-phase synthesis of mechanism-based cysteine prote-

ase inhibitors, Org. Lett., 3, 3707–3709 (2001).

A. Denis and C. Renou, Novel N-demethylation of ketolide: application to the solution phase

parallel synthesis of N-desosaminyl-substituted ketolides using ion exchange resins, Tetrahe-

dron Lett., 43, 4171–4174 (2002).

M. Cavicchioli, X. Marat, N. Monteiro, B. Hartmann, and G. Balme, Solution-phase parallel

tetrahydrofuran synthesis with propargyl alcohols and benzylidene-(or alkylidene-)malonates,

Tetrahedron Lett., 43, 2609–2611 (2002).

B. A. Kulkarni and A. Ganesan, Solution-phase parallel oxazole synthesis with TosMIC, Tetra-

hedron Lett., 40, 5637–5638 (1999).

C. T. Lowden and J. S. Mendoza, Solution phase parallel synthesis of 1,2-phenethyldiamines,

Tetrahedron Lett., 43, 979–982 (2002).

L. Varga, T. Nagy, I. Kovesdi, J. Benet-Buchholz, G. Dorman, L. Urge, and F. Darvas, Solution-

phase parallel synthesis of 4,6-diarylpyrimidine-2-ylamines and 2-amino-5,5-disubstituted-3,5-

dihydroimidazol-4-ones via a rearrangement, Tetrahedron, 59, 655–662 (2003).

K. H. Bleicher, Y. Wuthrich, G. Adam, T. Hoffmann, and A. J. Sleight, Parallel solution- and

solid-phase synthesis of spiropyrrolopyrazoles as novel neurokinin receptor ligands, Bioorg.

Med. Chem. Lett., 12, 3073–3076 (2002).

E. Laborde, B. T. Peterson, and L. Robinson, Traceless, self-cleaving solid- and solution-phase

parallel synthesis of 3,4,7-trisubstituted 3,4-dihydroquinoxalin-2-ones, J. Comb. Chem., 3,

572–577 (2001).

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

G. M. Green, N. P. Peet, and W. A. Metz, Polystyrene-supported benzenesulfonyl azide: a diazo

transfer reagent that is both efﬁ cient and safe, J. Org. Chem., 66, 2509–2511 (2001).

S. Ding, N. S. Gray, Q. Ding, and P. G. Shultz, Expanding the diversity of purine libraries, Tet-

rahedron Lett., 42, 8751–8755 (2001).

N. Zou, J.-F. Liu, and B. Jiang, Solution-phase synthesis of a thiazoyl-substituted indolyl library

via Suzuki cross-coupling, J. Comb. Chem., 5, 754–755 (2003).

W. R. Roush, D. Feitler, and J. Rebek, Polymer-bound tosyl azide, Tetrahedron Lett., 1391–

1392 (1974).

B. Raju, N. Nguyen, and G. W. Holland, Solution-phase parallel synthesis of substituted benz-

imidazoles, J. Comb. Chem., 4, 320–328 (2002).

J. J. Parlow, B. L. Case, T. A. Dice, R. L. Fenton, M. J. Hayes, D. E. Jones, W. L. Neumann,

R. S. Wood, R. M. Lachance, T. J. Girard, N. S. Nicholson, M. Clare, R. A. Stegeman, A. M.

Stevens, W. C. Stallings, R. G. Kurumbail, and M. S. South, Design, parallel synthesis, and

crystal structures of pyrazinone antithrombotics as selective inhibitors of the tissue factor VIIa

complex, J. Med. Chem., 46, 4050–4062 (2003).

H. Sumiyoshi, T. Shimizu, M. Katoh, Y. Baba, and M. Sodeoka, Solution-phase parallel syn-

thesis of carbamates using polymer-bound N-hydroxysuccinimide, Org. Lett., 4, 3923–3926

(2002).

J. C. Pelletier, A. Khan and Z. Tang, A tandem three-phase reaction for preparing secondary

amines with minimal side products, Org. Lett., 4, 4611–4613 (2002).

X. Wang, J. J. Parlow and J. A. Porco, Parallel Synthesis and Puriﬁ cation Using Anthracene-

Tagged Substrates, Org. Lett., 2, 3509–3512 (2000).

U. Bauer, B. J. Egner, I. Nilsson, and M. Berghult, Parallel solution phase synthesis of N-

substituted 2-pyrazoline libraries, Tetrahedron Lett., 41, 2713–2717 (2000).

M. S. South, T. A. Dice, T. J. Girard, R. M. Lachance, A. M. Stevens, R. A. Stegeman, W. C.

Stallings, R. G. Kurumbail, and J. J. Parlow, Polymer-assisted solution-phase (PASP) parallel

synthesis of an α-ketothiazole library as tissue factor VIIa inhibitors, Bioorg. Med. Chem. Lett.,

13, 2363–2367 (2003).

I. R. Baxendale, S. V. Ley, and C. Piutti, Total synthesis of the amaryllidaceae alkaloid ()-

plicamine and its unnatural enantiomer by using solid-supported reagents and scavengers in a

multistep sequence of reactions, Angew. Chem. Int. Ed. Engl., 41, 2194–2197 (2002).

I. R. Baxendale, A. Lee, and S. V. Ley, A concise synthesis of the natural product carpanone

using solid-supported reagents and scavengers, Synlett, 1482–1484 (2001).

R. N. Bream, S. V. Ley, and P. A. Procopiou, Synthesis of the β

-agonist (R)-salmeterol using a

sequence of supported reagents and scavenging agents, Org. Lett., 4, 3793–3796 (2002).

S. V. Ley and A. Massi, Polymer supported reagents in synthesis: preparation of

bicyclo[2.2.2]octane derivatives via tandem michael addition reactions and subsequent combi-

natorial decoration, J. Comb. Chem., 2, 104–107 (2000).

Y. Hu, S. Baudart, and J. A. Porco, Parallel synthesis of 1,2,3-thiadiazoles employing a “catch

and release” strategy, J. Org. Chem., 64, 1049–1051 (1999).

M. Bernard, W. T. Ford, and E. C. Nelson, Synthesis of ethyl retinoate with polymer-supported

wittig reagents, J. Org. Chem., 48, 3164–3168 (1983).

S. D. Clarke, C. R. Harrison, and P. Hodge, Phase transfer catalyzed polymer-supported Wittig

reactions, Tetrahedron Lett., 21, 1375–1378 (1980).

J. Castells, J. Font and A. Virgili, Reaction of dialdehydes with conventional and polymer-

supported Wittig reagents. J. Chem. Soc., Perkin Trans. 1, 1–6 (1979).

S. Shinkai, H. Tsuji, H. Yoichiro, and O. Manabe, Polymer-bound dimethylaminopyridine as a

catalyst for facile ester synthesis, Bull. Chem. Soc. Jpn., 54, 631–632 (1981).

A. Badadjamian and A. Kessat, Bromination of various unsaturated ketones and oleﬁ ns with

poly(4-methyl-5-vinylthiazolium) hydrotribromide, Synth. Commun., 25, 2203–2209 (1995).

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

REFERENCES 195

196 PARALLEL SOLUTION-PHASE SYNTHESIS

A. Hassner and M. Stern, Synthesis of alkyl azides with a polymeric reagent, Angew. Chem. Int.

Ed. Engl., 25, 478–479 (1986).

M. Lakshman, D. V. Narkarni, and R. E. Lehr, Regioselective ring opening of polycyclic aromatic

hydrocarbon epoxides by polymer-supported azide anion, J. Org. Chem., 55, 4892–4897 (1990).

C. R. Harrison, P. Hodge, and W. J. Rodgers, Conversion of carboxamides and oximes to nitriles

or imidoyl chloride using a polymer-supported phosphine and carbon tetrachloride, Synthesis,

41–43 (1977).

G. Cainelli, M. Contento, F. Manescalchi, and R. Regnoli, Polymer-supported phosphonates:

oleﬁ ns from aldehydes, ketones, and dioxolanes by means of polymer-supported phosphonates,

J. Chem. Soc., Perkin Trans., 1, 2516–2519 (1980).

A. G. M. Barrett, S. M. Cramp, R. S. Roberts, and F. J. Zeeri, Horner–Emmons synthesis with

minimal puriﬁ cation using ROMPGEL: a novel high-loading matrix for supported reagents,

Org. Lett., 1, 579–582 (1999).

A. R. Tunoori, D. Dutta, and G. I. Georg, Polymer-bound triphenylphosphine as traceless re-

agent for Mitsunobu reactions in combinatorial chemistry: synthesis of aryl ethers from phenols

and alcohols, Tetrahedron Lett., 39, 8751–8754 (1998).

J. C. Pelletier and S. Kincaid, Mitsunobu reaction modiﬁ cations allowing product isolation

without chromatography: application to a small parallel library, Tetrahedron Lett., 41, 797–800

(2000).

L. D. Arnold, H. I. Assil, I. Hanaa, and J. C. Vederas, Polymer-supported alkyl azodicarboxyl-

ates for Mitsunobu reactions, J. Am. Chem. Soc., 111, 3973–3976 (1989).

M. C. Desai and L. M. S. Stramiello, Polymer-bound EDC (P-EDC): a convenient reagent for

formation of an amide bond, Tetrahedron Lett., 34, 7685–7688 (1993).

M. Adamczyk, J. R. Fishpaugh, and P. G. Mattingly, An easy preparation of hapten active esters

via solid supported EDAC, Tetrahedron Lett., 36, 8345–8346 (1995).

M. Adamczyk and J. R. Fishpaugh, A solid supported synthesis of thiol esters, Tetrahedron

Lett., 37, 4305–4308 (1996).

J. J. Parlow and D. L. Flynn, Solution-phase parallel synthesis of a benzoxazinone library using

complementary molecular reactivity and molecular recognition (CMR/R) puriﬁ cation technol-

ogy, Tetrahedron, 54, 4013–4031 (1998).

J. J. Parlow, W. Naing, M. S. South, and D. L. Flynn, In situ chemical tagging: tetraﬂ uoroph-

thalic anhydride as a “sequestration enabling reagent” (SER) in the puriﬁ cation of solution-

phase combinatorial libraries, Tetrahedron Lett., 38, 7959–7962 (1997).

D. L. Flynn, J. Z. Crich, R. V. Devraj, S. L. Hockerman, J. J. Parlow, M. S. South, and S.

Woodard, Chemical library puriﬁ cation strategies based on principles of complementary mo-

lecular reactivity and molecular recognition, J. Am. Chem. Soc., 119, 4874–4881 (1997).

I. Masip, C. Ferrandiz-Huertas, C. Garcia-Martinez, J. A. Ferragut, A. Ferrer-Montiel, and A.

Messeguer, Synthesis of a library of 3-oxopiperazinium and perhydro-3-oxo-1,4-diazepinium

derivatives and identiﬁ cation of bioactive compounds, J. Comb. Chem., 6, 135–141 (2004).

X. Lei and J. A. Porco, Synthesis of a polymer-supported anthracene and its application as a

dienophile scavenger, Org. Lett., 6, 795–798 (2004).

T. Mukade, D. R. Dragoli, and J. A. Ellman, Parallel solution-phase asymmetric synthesis of

α-branched amines, J. Comb. Chem., 5, 590–596 (2003).

A. G. M. Barrett, B. T. Hopkins, A. C. Love, and L. Tedeschi, Parallel synthesis of termi-

nal alkynes using a ROMPgel-supported ethyl 1-diazo-2-oxopropylphosphonate, Org. Lett., 6,

835–837 (2004).

A. Studer, S. Hadida, R. Ferritto, S.-Y. Kim., P. Jeger, P. Wipf, and D. P. Curran, Fluorous

synthesis: a ﬂ uorous-phase strategy for improving separation efﬁ ciency in organic synthesis,

Science, 275, 823–826 (1997).

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

I. T. Horvath and J. Rabai, Facile catalyst separation without water: ﬂ uorous biphase hydrofor-

mylation of oleﬁ ns, Science, 266, 72–75 (1994).

B. Cornils, Fluorous biphase systems: the new phase-separation and immobilization technique,

Angew. Chem. Int. Ed. Engl., 36, 2057–2059 (1997).

I. T. Horvath, Fluorous biphase chemistry, Acc. Chem. Res., 31, 641–650 (1998).

A. Studer, P. Jeger, P. Wipf, and D. P. Curran, Fluorous synthesis: ﬂ uorous protocols for the Ugi

and Biginelli multicomponent condensations, J. Org. Chem., 62, 2917–2924 (1997).

A. Studer and D. P. Curran, A strategic alternative to solid phase synthesis: preparation of a

small isoxazoline library by “Fluorous Synthesis,” Tetrahedron, 53, 6681–6696 (1997).

B. Linclau, A. K. Sing, and D. P. Curran, Organic-ﬂ uorous phase switches: a ﬂ uorous amine scav-

enger for puriﬁ cation in solution phase parallel synthesis, J. Org. Chem., 64, 2835–2842 (1997).

C. W. Lindsley, Z. Zhao, R. C. Newton, W. H. Leister, and K. A. Strauss, A general Staudinger

protocol for solution-phase parallel synthesis, Tetrahedron Lett., 43, 4467–4470 (2002).

D. P. Curran, Fluorous reverse phase silica gel: a new tool for preparative separations in syn-

thetic organic and organoﬂ uorine chemistry, Synlett, 9, 1488–1496 (2001).

Z. Lou, J. Williams, R. W. Read, and D. P. Curran, Fluorous Boc (FBoc) carbamates: new amine

protecting groups for use in ﬂ uorous synthesis, J. Org. Chem., 66, 4261–4266 (2001).

A. G. M. Barrett, S. M. Cramp, R. S. Roberts, and F. J. Zecri, A ROMPGel supported N-hydroxy-

succinimide: a host of acylations with minimal puriﬁ cation, Org. Lett., 2, 261–264 (2000).

J. D. Moore, A. M. Harned, J. Henle, D. L. Flynn, and P. R. Hanson, Scavenger-ROMP-ﬁ lter:

a facile strategy for soluble scavenging via norbornenyl tagging of electrophilic reagents, Org.

Lett., 4, 1847–1849 (2002).

J. D. Moore, R. H. Herpel, J. R. Lichtsinn, D. L. Flynn, and P. R. Hanson, ROMP-generated

oligomeric sulfonyl chlorides as versatile soluble scavenging agents, Org. Lett., 5, 105–107

(2003).

J. D. Moore, R. J. Byrne, P. Vedantham, D. L. Flynn, and P. R. Hanson, High-load, ROMP-

generated oligomeric bis-acid chlorides: design of soluble and insoluble nucleophile scaven-

gers, Org. Lett., 5, 4241–4244 (2003).

A. M. Harned, S. Mukherjee, D. L. Flynn, and P. R. Hanson, Ring-opening metathesis phase-

trafﬁ cking (ROMPpt) synthesis: multistep synthesis on soluble ROM supports, Org. Lett., 5,

15–18 (2003).

S. Murkherjee, K. W. C. Poon, D. L. Flynn, and P. R. Hanson, Capture-ROMP-release: applica-

tion to the synthesis of amines and alkyl hydrazines, Tetrahedron Lett., 44, 7187–7190 (2003).

D. J. Gravert and K. D. Janda, Organic synthesis on soluble polymer supports: liquid-phase

methodologies, Chem. Rev., 97, 489–509 (1997).

P. H. Toy and K. D. Janda, Soluble polymer-supported organic synthesis, Acc. Chem. Res., 33,

546–554 (2000).

A. Kirschning, H. Monenschein, and R. Wittenberg, Functionalized polymers: Emerging versa-

tile tools for solution-phase chemistry and automated parallel synthesis, Angew. Chem. Int. Ed.

Engl., 40, 650–679 (2001).

D. H. Drewry, D. M. Coe, and S. Poon, Solid-supported reagents in organic synthesis, Med. Res.

Rev., 19, 97–148 (1999).

S. J. Shuttleworth, P. M. Allin, and K. Sharma, Functionalized polymers: recent developments

and new applications in synthetic organic chemistry, Synthesis, 1217–1239 (1997).

S. W. Kaldor and M. G. Siegel, Combinatorial chemistry using polymer-supported reagents,

Curr. Opin. Chem. Biol., 1, 101–106 (1997).

B. Jandeleit, D. J. Schaefer, T. S. Powers, H. W. Turner, and W. H. Weinberg, Combinatorial

materials science and catalysis, Angew. Chem. Int. Ed. Engl., 38, 2494–2532 (1999).

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

REFERENCES 197

198 PARALLEL SOLUTION-PHASE SYNTHESIS

E. Lindner, T. Schneller, F. Auer, and H. A. Mayer, Chemistry in interphases: a new approach to

organometallic syntheses and catalysis, Angew. Chem. Int. Ed. Engl., 38, 2154–2174 (1999).

J. H. Clark and D. J. Macquarries, Environmentally friendly catalytic methods, Chem. Soc. Rev.,

25, 303–310 (1996).

J. J. Parlow, R. V. Devraj, and M. S. South, Solution-phase chemical library synthesis using

polymer-assisted puriﬁ cation techniques, Curr. Opin. Chem. Biol., 3, 320–336 (1999).

S. J. Teague, A. M. Davis, P. D. Leeson, and T. I. Oprea, The design of leadlike combinatorial

libraries, Angew. Chem. Int. Ed. Engl., 24, 3743–3748 (1985).

T. I. Oprea, A. M. Davis, S. J. Teague, and P. D. Leeson, Is there a difference between leads and

drugs? A historical perspective, J. Chem. Inf. Comput. Sci., 41, 1308–1315 (2001).

T. I. Oprea, I. Zamora, and A.-L. Ungell, Pharmacokinetically based mapping device for chemi-

cal space navigation, J. Comb. Chem., 4, 258–266 (2002).

J. Sadowski and H. Kubinyi, A scoring scheme for discriminating between drugs and nondrugs,

J. Med. Chem., 41, 3325–3329 (1998).

Ajay, W. P. Walters, and M. A. Murcko, Can we learn to distinguish between “drug-like” and

“nondrug-like” molecules?, J. Med. Chem., 41, 3314–3324 (1998).

118.

119.

120.

121.

122.

123.

124.

125.

199

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

TIMING OF ANALOG RESEARCH

IN MEDICINAL CHEMISTRY

JÁNOS FISCHER AND ANIKÓ GERE

Gedeon Richter Ltd.

Budapest, Hungary

6.1 INTRODUCTION

The term analog is used according to the IUPAC recommendations,

where an incremental

innovation

differentiates the drug from the original one. This overview focuses on ana-

logs, but bioisosteres

are also mentioned if appropriate. Since not all analogs ﬁ t into a

general formula, we depict all the formulas individually. We would like to demonstrate that

analog research can be successful both in the early phase, when no product has yet been

introduced to the market, and also after a drug has been launched successfully. The analogs

afford in most cases an incremental and in some cases an essential innovation. We propose

a classiﬁ cation of drug discoveries according to their timing: early phase analogs and drug

analogs. With the help of some signiﬁ cant examples from the past few decades, we show

how analogs from both approaches can contribute to drug discovery.

6.2 EARLY PHASE ANALOGS

Early phase analogs are structurally similar drugs discovered before the original drug is

launched. As a result of early phase parallel research, the discovery dates of such deriva-

tives are often very close to each other.

6.2.1 ACE Inhibitors

The ﬁ rst successful angiotensin-converting enzyme (ACE) inhibitor was captopril.

The

pioneer discovery by Ondetti and Cushman was achieved by replacing the carboxyl group

200 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

of the analogous carboxyalkanoyl-L-proline by an SH group. To get more active and mer-

capto-free analogs, Patchett started from an homologous carboxyalkanoyl-L-proline, and

among N-carboxyalkyl dipeptides, enalapril

and lisinopril proved to be long-acting ACE

inhibitors. In a parallel research activity many analogous ACE inhibitors were discovered

before enalapril appeared on the market to create a class of drugs (Table 6.1), which can be

used in the treatment of hypertension and congestive heart disease.

6.2.2 AT

Antagonists

In 1982, hypotensive imidazole-5-acetic acid derivatives were published,

which antago-

nized angiotensin II evoked vasoconstriction. The clinical breakthrough came with losar-

tan

and its analogs, which except for eprosartan and telmisartan, have a biphenyltetrazole

moiety (Table 6.2). The angiotensin II antagonists are competitive with the ACE inhibitors,

and further clinical trials will decide which class is more effective.

6.2.3 Proton Pump Inhibitors

Omeprazole,

which is the ﬁ rst successfully introduced H

⫹

-ATPase inhibitor, was fol-

lowed by novel analogs (Table 6.3). They are irreversible blockers of the proton pump that

is responsible for acid secretion by the gastric parietal cells. According to the mechanism

of action,

omeprazole itself is inactive, but it is transformed into a sulfenamide, which is

the active inhibitor in vivo. A comparison of omeprazole, lansoprazole, and pantoprazole

demonstrated differences in pharmacokinetics and drug interaction proﬁ le.

6.2.4 Insulin Sensitizers: Glitazones

The ﬁ rst member of the thiazolidine-2,4-diones, ciglitazone,

reduced plasma glucose

after oral administration in several insulin-resistant animal models, but a more potent com-

pound was needed. The clinical breakthrough was troglitazone,

which was introduced

in 1997. In contrast to the three classes of drugs noted above, where the ﬁ rst member of

the class remained successful even after the introduction of their analogs, the case of the

“glitazones” shows an opposite situation. Troglitazone was withdrawn

from the market

COOH

COOEt

COOH

captopril enalapril lisinopril

Name Originator Basic Patent Launch

Captopril Squibb (Bristol-Myers Squibb) 1976 1980

Enalapril Merck & Co. 1978 1984

Lisinopril Merck & Co. 1978 1987

TABLE 6.1 ACE Inhibitors

COOH

HOOC

losartan eprosartan valsartan

COOH

candesartan irbesartan telmisartan

Name Originator Basic Patent Launch

Losartan DuPont 1986 1994

Eprosatran SmithKline Beecham

(GlaxoSmithKline)

1989 1997

Valsartan Ciba-Geigy (Novartis) 1990 1996

Candesartan Takeda 1990 1999

Irbesartan Sanoﬁ 1990 1997

Telmisartan Boehringer-Ingelheim 1991 1999

TABLE 6.2 Angiotensin II Antagonists

omeprazole pantoprazole lansoprazole

rabeprazole

Name Originator Basic Patent Launch

Omeprazole Hässle (AstraZeneca) 1978 1988

Pantoprazole Byk-Gulden 1983 1994

Lansoprazole Takeda 1984 1991

Rabeprazole Eisai 1986 1997

TABLE 6.3 Omeprazole and Its Analogs

202 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

in 2000 because of liver toxicity in humans (Table 6.4). The clinical application of rosigli-

tazone and pioglitazone is carried out according to the U.S. Food and Drug Administration

(FDA) labeling, including the need for liver enzyme monitoring before the start of therapy

and periodically during the treatment. The glitazones exert their insulin sensitizer activity

via stimulation of peroxisome proliferator activated receptor gamma subtype (PPAR-γ).

6.2.5 HMG-CoA Reductase Inhibitors

Mevastatin (compactin), a fungal metabolite

and a potent inhibitor of hydroxymethyl-

glutaryl (HMG) CoA reductase, initiated a series of statins for treatment of lipoprotein

disorders (Table 6.5). The clinical breakthrough was lovastatin,

followed by simvastatin

and pravastatin. Their therapeutic ﬁ eld is the treatment of hypercholesterolemia. As a result

of intense activity in the design of synthetic analogs of the foregoing statins, new analogs

were obtained where the decalin moiety was replaced by different heterocyclic rings bear-

ing almost the same substituents, such as 3,5-dihydroxyheptanoic acid derivative, 4-ﬂ uo-

rophenyl, and isopropylsubstituents. The ﬁ rst member of these heterocyclic statins was

ﬂ uvastatine sodium, followed by atorvastatin, which lower both cholesterol and triglyc-

eride levels. The lactone forms are prodrugs, which are metabolized to the corresponding

active hydroxy acid form.

6.2.6 Antimigraine Drugs

The ﬁ rst representative of drug with 5-HT

1B/1D

agonist mechanism was sumatriptan,

which proved to be a useful drug for the treatment of acute attacks of migraine. Among

the analogs 5-HT

selectivity and pharmacokinetic parameters play an important role

(Table 6.6). Rizatriptan showed both an increased oral bioavailability and more rapid

absorption to oral sumatriptan.

6.3 DRUG ANALOGS

A drug analog is a structurally similar drug which was discovered later (much later) than

the launch of the original drug. There are some lonely drugs, such as aspirin, levodopa,

troglitazone rosiglitazone pioglitazone

Name Originator Basic Patent Launch

Troglitazone Sankyo 1983 1997

Pioglitazone Takeda 1985 1999

Rosiglitazone SmithKline Beecham

(GlaxoSmithKline)

1987 1999

TABLE 6.4 Glitazones

methyldopa, metformin, PAS, and colchicine, without analogs, but they represent only a

minority group of the drugs. In most cases several examples of successful analog-research

based on pioneer drugs can be observed.

6.3.1 Metoclopramide Analogs

The history of these drugs amounts to about four decades (Table 6.7). Metoclopramide was

discovered in 1961. It is a centrally acting antiemetic agent,

but its mechanism of action

has not been fully elucidated. Antidopaminergic properties at both D

and D

receptor

subtypes play an important role in its activity,

but its extrapyramidal side effects are also

correlated with this mechanism. Twenty years later analogous cisapride was dicovered,

which does not exhibit potent dopamine receptor antagonist activity. Its main therapeutic

COONa

lovastatin simvastatin pravastatin

COONa

COO

Ca 1/2

fluvastatin atorvastatin

Name Originator Basic Patent Launch

Lovastatin Merck & Co. 1979 1987

Simvastatin Merck & Co. 1980 1989

Pravastatin Sankyo 1980 1989

Fluvastatin Sandoz (Novartis) 1982 1994

Atorvastatin Pﬁ zer 1986 1997

TABLE 6.5 Statins

DRUG ANALOGS 203

204 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

use is the treatment of gastroesophageal reﬂ ux disease and it was one of the most suc-

cessful drugs of the last decade. Agonistic action at 5-HT

receptors, and thereby facilita-

tion of cholinergic excitatory neurotransmission, has been suggested as the mechanism

by which these agents enhance gastric motility.

Current data suggest that concomitant

administration of cisapride and certain azole-derivatives (e.g., ketoconazole) can result in

OMe

metoclopramide cisapride mosapride

Name Originator Basic Patent Launch

Metoclopramide Société d’Études Scientiﬁ ques et

Industrielles de l’Île-de-France

1961 1964

Cisapride Janssen 1981 1988

Mosapride Dainippon 1986 1998

TABLE 6.7 Gastroprokinetic Drugs

O O

sumatriptan naratriptan zolmitriptan

O O

eletriptan rizatriptan frovatriptan

Name Originator Basic Patent Launch

Sumatriptan Glaxo (GlaxoSmithKline) 1982 1991

Naratriptan Glaxo (GlaxoSmithKline) 1987 1997

Zolmitriptan Wellcome (GlaxoSmithKline) 1990 1997

Eletriptan Pﬁ zer 1990 1999

Rizatriptan Merck & Co. 1991 1998

Frovatriptan SmithKline Beecham

(GlaxoSmithKline)

1991 2000

TABLE 6.6 Antimigraine Drugs

prolongation of the QT interval. The marketing of cisapride was terminated in the United

States in 2000,

but further research is going on in this ﬁ eld. Mosapride was discovered

in 1986. The mode of action of mosapride on gastrointestinal motor activity was clearly

different from that of cisapride, which stimulates motor activity in all sites of the gastro-

intestinal tract.

Further clinical trials are needed to evaluate the drug interaction proﬁ le

of mosapride.

6.3.2 Azatadine Analogs

It was a popular view a generation ago that a nonsedating H

-receptor antagonist is unob-

tainable.

There was a lack of validated methods to predict the sedative liability of a new

antihistamine. Terfenadine served as a clinical breakthrough, which proved to be such an

agent in 1978.

This initiated researchers at Schering-Plough to carry out drug analog

research, whose lead molecules were terfenadine and azatadine.

A selected battery of

central nervous system (CNS) tests in guinea pigs and mice using terfenadine and azata-

dine as reference drugs helped them to screen analogs. The carbamate analog of azatidine

removed its central nervous system (CNS) activity while retaining much of its antihista-

mine potency. Further optimizing the carbamate azatadine afforded the 8-chloro derivative

loratadine,

with a longer duration of action (Table 6.8). Its active metabolite, deslorata-

dine was launched in 2001.

6.3.3 Miconazole Analogs

Miconazole is used primarily as a topical antifungal agent. Ketoconazole and ﬂ uconazole

can also be given orally. The terminal half-life of ﬂ uconazole is approximately three times

higher than that of ketoconazole. Oral ﬂ uconazole with a single dose produces clinical cure

of uncomplicated vulvovaginal candidiasis. All of these agents are fungistatic by inhibiting

COOEt

azatadine loratadin desloratadine

Name Originator Basic Patent Launch

Azatadine Schering-Plough 1963 1977

Loratadine Schering-Plough 1980 1988

Desloratadine Sepracor/Schering-Plough 1984 2001

TABLE 6.8 Azatadine Analogs

DRUG ANALOGS 205

206 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

the biosynthesis of ergosterol. Clinical failure of antifungal therapy due to resistance to

existing agents is spreading rapidly and is often multifactorial

(Table 6.9).

6.3.4 Nifedipine Analogs

The calcium channel blocking mechanism was discovered by Fleckenstein

in 1967.

Nifedipine,

the ﬁ rst member of this class, has a short duration of action. Many structural

miconazole sulconazole oxiconazole

ketoconazole fenticonazole fluconazole

itraconazole sertraconazole

Name Originator Basic Patent Launch

Miconazole Janssen 1968 1971

Sulconazole Syntex 1974 1985

Oxiconazole Siegfried 1975 1983

Ketoconazole Janssen 1977 1981

Fenticonazole Recordati 1978 1987

Fluconazole Pﬁ zer 1981 1988

Itraconazole Janssen 1983 1988

Sertraconazole Ferrer 1984 1992

TABLE 6.9 Miconazole Analogs

analogs were developed with a better pharmacokinetic proﬁ le (gradual onset, long dura-

tion of action). These drugs are used for the treatment of mild and moderate hypertension

(Table 6.10).

6.3.5 Propranolol Analogs

Propranolol is the ﬁ rst nonselective β-adrenergic blocking agent with no intrinsic sympa-

thomimetic activity. It is used for the treatment of arrhythmias, angina pectoris, and hyper-

tension. Because of its ability to block β-receptors in bronchial smooth muscle, the drug

is generally not used in people with bronchial asthma. As a consequence, there has been

a search for β-adrenergic blocking agents that are cardioselective. Practolol, discovered

in 1966, was the ﬁ rst such agent, but it was withdrawn because of its toxic side effects.

Table 6.11 summarizes the β

-selective (cardioselective) antagonists, which contributed an

essential improvement to the therapy.

6.3.6 Clodronate Analogs

Bisphosphonates are powerful bone resorption inhibitors that have been found to be clini-

cally useful in the treatment of osteoporosis. Bisphosphonates are generally very poorly

absorbed when given orally, but once absorbed they are taken up preferentially in bones.

COOMe

MeOOC

COOEtMeOOC

nifedipine felodipine lercanidipine

COOEtMeOOC

COOEtEtOOC

amlodipine lacidipine

Name Originator Basic Patent Launch

Nifedipine Bayer 1967 1975

Felodipine Hässle (AstraZeneca) 1978 1988

Lercanidipine Recordati 1984 1997

Lacidipine GlaxoSmithKline 1985 1991

Amlodipine Pﬁ zer 1986 1990

TABLE 6.10 Calcium Channel Blockers

DRUG ANALOGS 207

208 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

Clodronate disodium and etidronate sodium were used for the treatment of Paget’s dis-

ease.

The second-generation products (alendronate and pamidronate) are much more

effective. A third-generation agent, the risedronate, seems to have fewer esophageal side

effects

(Table 6.12).

6.4 SUMMARY

This short overview of analog research in medicinal chemistry has deﬁ ned two main di-

rections according to the timing: early phase analogs and drug analogs. A breakthrough

discovery in medicinal chemistry initiates parallel early phase research activities in several

research centers of the world. The marketing of new drugs, however, takes on average 10

propranolol atenolol metoprolol

bopindolol betaxolol esmolol

bisoprolol carvedilol

Name Originator Basic Patent Launch

Propranolol ICI (AstraZeneca) 1962 1964

Atenolol ICI (AstraZeneca) 1969 1975

Metoprolol Hässle (AstraZeneca) 1970 1975

Bopindolol Sandoz (Novartis) 1975 1985

Betaxolol Synthélabo 1975 1983

Esmolol American Hospital Supply (DuPont) 1980 1987

Bisoprolol Merck (Damstadt) 1976 1986

Carvedilol Boehringer-Mannheim (Roche) 1978 1991

TABLE 6.11 b

-Adrenergic Blocking Agents

to 15 years, and during this long period several similar lead compounds are identiﬁ ed to

give at the end several early phase analogs on the market. In the case of drug analogs the

situation is different. A drug is the end product of a long optimizing process in research

and development, nevertheless, during their clinical trials, side effects, drug interactions,

and other weak points can be observed, which stimulates researchers to make drug analogs

which also provide remarkable achievements, as shown with the examples above.

These two approaches overlap in some cases. If a parallel research activity is starting in

a period when the clinical results of a drug candidate are published (phase III), the products

of the analog research will be regarded as a drug analog because of the long development

process of today. There are no general rules as to whether it is preferable to undertake early

phase or a drug analog research. It depends on the marketing conditions, the company strat-

egy, the medicinal chemistry possibilities, and last but not least, on the inventive capacities

of the people involved.

ONa

clodronate disodium etidronate disodium pamidronate disodium alendronate sodium

ONa

tiludronate disodium risedronate sodium zoledronate disodium ibandronate sodium

ONa

ONaO

incadronate sodium

Name Originator Basic Patent Launch

Clodronate Procter & Gamble 1963 1986

Etidronate Procter & Gamble 1966 1977

Pamidronate Gador and Henkel 1971 1987

Alendronate Gentili/Merck 1982 1993

Tiludronate Sanoﬁ -Synthélabo 1982 1993

Risedronate Procter & Gamble 1984 1998

Zoledronate Novartis 1986 2000

Ibandronate Roche 1986 1996

Incadronate Yamanouchi 1988 1997

TABLE 6.12 Clodronate Analogs

SUMMARY 209

210 TIMING OF ANALOG RESEARCH IN MEDICINAL CHEMISTRY

ACKNOWLEDGMENTS

The authors thank Professor C. R. Ganellin for his helpful comments and S.Lévai for tech-

nical assistance.

REFERENCES AND NOTES

IUPAC Recommendations: glossary of terms used in medicinal chemistry, Pure Appl. Chem. 70

(1998) 1129–1143.

Riefberg V., Pinkus G., In Vivo (1996) 18–23.

Wermuth C. G. (Ed.), The Practice of Medicinal Chemistry, Academic Press, San Diego, CA

(1996) p. 207.

Ondetti M. A., Rubin B., Cushman D. W., Science 196 (1977) 441; U.S. patent 4,046,889

(1976).

Patchett P. A., Harris E., Tristram E. W., Wyvratt M. J., Wu M. T., Taut D., Peterson E. R.,

Ikeler T. J., ten Broeke J., Payne L. G., Ondeyka D. L., Thorsett E. D., Greenlee W. J., Lohr

N. S., Hoffsommer R. D., Joshua H., Ruyle W. V., Rothrock J. W., Aster S. D., Maycock A. L.,

Robinson F. M., Hirschmann R., Sweet C. S., Ulm E. H., Gross D. M., Vassil T. C., Stone C. A.,

Nature 288 (1980) 280; European patent 12,401 (1978).

Furakawa Y., Kishimoto S., Nishikawa K. (Takeda Chemical Industries Ltd.), U.S. patents

4,340,598 and 4,355,040 (1982).

Carini D. J., Duncia J. V., Aldrich P. E., Chiu A. T., Johnson A. L., Pierce M. E., Price W. A.,

Santella J. B. III, Wells G. J., Wexler R. R., Wong P. C., Yoo S.-E., Timmermans P. B. M., J. Med.

Chem. 34 (1991) 2525–2547; European patent 253,310 (1986).

Sjöstrand S. E., Junggren U. K. (Hässle Läkemedel AB), European patent 5129 (1978).

Lindberg P., Nordberg P., Alminger T., Bränström A., Wallmark B., J. Med. Chem. 29 (1986)

1327–1329.

Zech K., Steinijans V. W., Huber R., Radtke H. W., Int. J. Clin. Pharmacol. Ther. 34 (Suppl. 1)

(1996) 3–6.

Sohda T., Mizuno K., Imamiya E., Suguyama Y., Fujita T., Kawamatsu Y., Chem. Pharm. Bull.

30 (1982) 3580–3600.

Yoshioka T., Fujita T., Kanai T., Aizawa Y., Hasegawa K., Horikoshi H., J. Med. Chem. 32 (1989)

421–428.

Warner-Lambert decided to discontinue marketing troglitazone (Rezulin) for the treatment of

type II diabetes for safety and efﬁ cacy reasons (Warner-Lambert news release in Prous Science

Daily Essentials, March 22, 2000).

Lehmann J. M., Moore L. B., Smith-Oliver T. A., Wilkison W. O., Willson T. M., Kliewer S. A.,

J. Biol. Chem. 270 (1995) 12953–12956.

Endo A., Kuroda M., Tsujita Y., J. Antibiot. 29 (1976) 1346.

Alberts A. W., Chen J., Kuron G., Hunt V., Huff J., Hoffman C., Rothrock J., Lopez M., Joshua

H., Harris E., Patchett A., Monaghan R., Currie S., Stapley E., Albers-Schonberg G., Hensens O.,

Hirschﬁ eld J., Hoogsteen K., Liesch J., Springer J., Proc. Natl. Acad. Sci. USA 77 (1980) 3957.

Todd P. A., Goa K. L., Drugs 40 (1990) 583–607.

Heuring R. E., Peroutka S. J., J. Neurosci. 7 (1987) 894–903.

Sciberras D. G., Polvino W. J., Gertz B. J., Cheng H., Stepanavage M., Wittreich J., Olah T.,

Edwards M., Mant T., Br. J. Clin. Pharmacol. 43 (1997) 49–54.

French patent 1 313 758 (Societé d’Études Scientiﬁ ques et Industrielles de l’Île-de-France, 1961).

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

MacDonald T. M., Eur. J. Clin. Pharmacol. 40 (1991) 225–230.

Craig D. A., Clarke D. E., J. Pharmacol. Exp. Ther. 252 (1990) 1378–1386.

Janssen decided to stop marketing cisapride (Propulsid) in the United States as of July 14, 2000

(Janssen news release in Prous Science Daily Essentials, March 28, 2000).

Yoshida N., Ito T., Karasawa T., Itoh Z., J. Pharmacol. Exp. Ther. 257 (1991) 2572–2578.

Barnett A., Green M. J., in Lednice D. (Ed.), Chronicles of Drug Discovery, Vol. 3, ACS Profes-

sional Reference Book, Washington, DC 1993, p. 83.

Clarke C. H., Nicholson A. N., Br. J. Clin. Pharmacol. 6 (1978) 31–35.

U.S. patent 3,301,863 (Schering Corp., 1963).

Villani F. J., Wefer E. A., Mann T. A., Peer L., Levy A. S., J. Heterocycl. Chem. 9 (1972)

1203–1207.

Watkins W. J., Renau T. E., Annu. Rep. Med. Chem. 35 (2000) 157–166.

Fleckenstein A., Arzneim.-Forsch. (Drug Res.) 22 (1967) 22–33.

U.S. patent 3,485,847 (Bayer, 1967).

Harting J., Becker K. H., Bergmann R., Bourgois R., Enenkel H. J., Fuchs A., Jonas R., Lettenbau

K., Minck K. O., Schelling P., Schulze E., Arzneim.-Forsch. (Drug Res.) 36 (1986) 200–208.

Meunier C., Chapuy M. C., Courpron P., Vignon E., Edouard C., Bernard J., Rev. Rheum. 42

(1975) 699–705.

Lanza F. L., Hunt R. H., Thomson A. B. R., Provenza J. M., Blank M. A., Gastroenterology, 119

(2000) 631–638.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

REFERENCES AND NOTES 211

213

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

POSSIBLE ALTERNATIVES

TO HIGH-THROUGHPUT SCREENING

CAMILLE G. WERMUTH

Prestwick Chemical, Inc.

Illkirch, France

7.1 INTRODUCTION

A program for therapeutic discovery typically starts with the search for hits: molecules that

exhibit a certain afﬁ nity for a target. Identifying hits for a new target usually involves the

screening of a wide range of structurally diverse small molecules in an in vitro bioassay.

Alternatively, small molecules can be screened for their potential for modulation of a bio-

logical process that is considered to be critical in a disease or in which the target is thought

to play a major role. Thanks to miniaturization and robotics, the number of compounds that

can be screened has greatly increased, and several thousand compounds can be screened in

a single day. Once a hit has been identiﬁ ed, its activity must be conﬁ rmed and validated.

The following are typical hit validation criteria: (1) the activity of the hit must be reproduc-

ible; (2) the hit must not display activity against many diverse targets; (3) the analogs of the

hit must display SAR; and (4) the hit must not contain chemically reactive functions.

Only

then does it become a lead substance, commonly referred to as the lead.

The pharmaceutical industry has come to rely heavily on high-throughput screening

(HTS), and virtually every pharmaceutical company in the world has established HTS as

an integral part of its discovery process. High-throughput screening technology can cur-

rently be regarded as being the preferred method for identiﬁ cation of new hits. There are

many accounts in the literature of successes based on this approach, including the discovery

of insulin mimetics

or of the new opioid ORL1 receptor agonists.

This strategy for drug

discovery has several limitations, however, that are in the main due to the nature of the

chemical libraries that are input to the robots. These libraries are usually huge and typically

contain from 100,000 to 1,000,000 compounds. They are most often assembled by parallel

214 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

or combinatorial chemistry. Some of the limitations exhibited by these libraries include

inadequate diversity (which leads to a decrease in the chances of success), rather moderate

or even low hit rates, and ﬁ nally, the low biopharmaceutical quality of the hits (“A hit is not

a drug!”). Present trends in HTS are focused on reducing the size of the libraries, on making

them more effective, and on excluding ADME (absorption, distribution, metabolism, and

elimination)-, and toxicological-inadequate candidates at as early a stage as possible. Com-

putational chemists can help to reduce the number of compounds that are selected for HTS.

HTS technology has overshadowed all other drug discovery paradigms, and its almost

exclusive use as a method for discovering drugs can lead to rather disappointing results.

The aim of the present chapter is to draw the attention of scientists involved in drug dis-

covery to the existence of some valuable alternative approaches. The ﬁ rst of these is analog

design, a second discusses the usefulness of physiopathological hypotheses, and a third

deals with contributions from clinical observations. Finally, in the last part, opportunities

to derive new drugs from old drugs that are provided by the selective optimization of side

activities (SOSA) approach are considered.

7.2 ANALOG DESIGN

7.2.1 Deﬁ nitions

The term analogy, which is derived from the Latin and Greek analogia, has been used

in natural sciences since 1791 to describe structural and functional similarity.

When ex-

tended to apply to drugs, this deﬁ nition implies that the analog of an existing drug molecule

shares chemical and therapeutic similarities with the original compound. In formal terms,

this deﬁ nition means that three categories of drug analogs may be anticipated: (1) analogs

that exhibit chemical and pharmacological similarity, (2) analogs that exhibit only chemi-

cal similarity, and (3) analogs that display similar pharmacological properties but exhibit

totally different chemical structures.

Since they exhibit both chemical and pharmacological similarities, analogs in the ﬁ rst

category, may be regarded as direct analogs. These correspond to the class of drugs that are

often referred to as me-too compounds. They are usually improved versions that exhibit phar-

macological, pharmacodynamic, or biopharmaceutical advantages over a pioneer drug. Ex-

amples are the ACE inhibitors derived from captopril, the histamine H

antagonists derived

from cimetidine, and the HMG-CoA reductase inhibitors derived from mevinolin. There are

production and marketing reasons behind such analogs which can be justiﬁ ed in much the

same way as any other industrial products, such as laptop computers or automobiles.

The second class of analogs is made of molecules that have chemical resemblances

and for which the term structural analogs is suggested. This class contains compounds

that were originally prepared as close and patentable analogs of a novel lead, but in which

biological assays have revealed totally unexpected pharmacological properties. A historical

example of the emergence of a new activity is provided by the discovery of the antidepres-

sant properties of imipramine, which was originally designed as an analog of the potent

neuroleptic drug chlorpromazine. Another example which illustrates that chemical simi-

larity does not necessarily mean biological similarity is to be found in steroid hormones:

although testosterone and progesterone are chemically very similar, they have totally dif-

ferent biological functions. The observation of an “emergent” activity may be purely fortu-

itous or may be the result of a deliberate and systematic investigation.

In the third class of analogous compounds, no chemical similarity is observed, but they,

compounds share common biological properties. The term functional analogs is suggested

for such compounds. Examples include the neuroleptics chlorpromazine and haloperidol

and the tranquillizers diazepam and zopiclone. Despite the fact that they have totally dif-

ferent chemical structures, they show similar afﬁ nities for dopamine and benzodiazepine

receptors, respectively. Currently, the design of such drugs has been facilitated, thanks to

virtual screening of large libraries of diverse structures.

7.2.2 Pharmacophore-Based Analog Design: Scaffold Hopping

or Scaffold Morphing

One option for designing functional analogs that are of interest involves searching large

virtual compound libraries for structures that are isofunctional but which are based on a

different scaffold. This approach is referred to as scaffold hopping. The objective is to es-

cape from a patented chemical class by identifying molecules in which the central scaffold

is different but in which the essential function-determining points are retained to form the

basis of a relevant pharmacophore.

In an exploratory test

a program called CATS (chemically advanced template search)

was applied to the prediction of novel cardiac T-type Ca

2⫹

channel blocking agents by

exploring the Roche in-house compound depository. Mibefradil (1), a known T-channel

blocking agent (IC

⫽ 1.7 µM) (Scheme 7.1), was used as the seed structure for CATS. The

12 highest-ranking molecules were tested for their ability to inhibit cellular Ca

2⫹

inﬂ ux us-

ing a cell culture assay. Nine compounds (75%) showed signiﬁ cant activity (IC

⬍ 10 µM),

of which one compound (2, clopimozid) had an IC

⬍ 1 µM. The IC

values for the next

best structures (not given in the article) were 1.7, 2.2, 3.2, and 3.5 µM.

These hits have structural scaffolds that show signiﬁ cance differences from the query

structure 1. Essential function-determining points are, however, retained, and these form the

basis for a relevant pharmacophore pattern. A similar approach, described as scaffold mor-

phing, involves a series of directed chemical transformations of the initial structure, the aim

of which is to generate new chemotypes with enhanced properties such as potency, selectiv-

ity, safety, and novelty.

An early claim of scaffold morphing was made in 1994 by Buehl-

mayer et al.

The objective was the design of the Novartis drug valsartan (Diovan)

using the

DuPont angiotensin II receptor antagonist (3, losartan, Scheme 7.2) as the starting model.

This last example can be regarded as being a particularly reﬁ ned case of bioisostery

in which the n-butyl chains overlap, as does the carbonyl with the imino dipole, and the

lipophilic chlorine with the isopropyl group. In valsartan (4) the metabolic oxidation of

the hydroxymethylene function has already been anticipated. Rational morphing strategy

Scaffold hopping

1 2

Scheme 7.1 Query structure 1 (mibefradil) and a high-ranking isofunctional structure 2 (clopimo-

zid) derived from 1 by CATS.

ANALOG DESIGN 215

216 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

is closely associated with virtual screening.

Other detailed descriptions of modern com-

puter-based hit location and optimization technologies are described in the excellent book

by Kubinyi and Müller.

7.2.3 Natural Compounds as Models

There are many historical examples recorded of drugs that are the result of modiﬁ cation,

usually a simpliﬁ cation, of a natural model: morphinanes, benzomorphanes, and phenyl-

piperidines derived from morphine; hundreds of steroid analogs derived from the endog-

enous hormones, and procaine and various other local anesthetics derived from cocaine,

for example. More recently, a second generation of statins characteristically contains only

two chiral centers instead of the eight in the natural lead lovastatin (5) or in its close analog

pravastatin (6). A typical example is found in rosuvastatin (7) (Scheme 7.3).

7.2.4 Emergence of New Activities

Imipramine It can also happen that a totally new property which is absent in the original

molecule appears unexpectedly during pharmacological or clinical studies on a me-too

Scheme 7.2 The transition from the DuPont angiotensin II receptor antagonist losartan (3) to the Novartis

analog valsartan (4) represents an example of computer-guided scaffold morphing. (From ref. 7.)

3 4

Scheme 7.3 The more recent statins contain only two chiral centers instead of the eight in natural

lovastatin.

OCH

O O

5 lovastatin (mevinolin)

6 pravastatin 7 rosuvastatin

compound. The emergence of such a new activity means that the therapeutic copy in turn

becomes a new lead structure. This was the case for imipramine (9) (Scheme 7.4), which

was initially synthesized as an analog of chlorpromazine (8) and presented to clinical

investigators to study its antipsychotic proﬁ le.

11,12

During its clinical evaluation this

substance exhibited much greater activity against depressive states than against psychoses.

Since 1954, imipramine has opened up genuine therapeutic avenues for the pharmacological

treatment of depression.

Sildenaﬁ l A more recent example is provided by the discovery that sildenaﬁ l (11,

Scheme 7.5; Viagra), a phosphodiesterase type 5 (PDE5) inhibitor, can be used as an ef-

ﬁ cacious, orally active agent for the treatment of male erectile dysfunction.

13,14

On its way

to becoming Viagra, the compound UK-92,480, which was prepared in 1989 by Pﬁ zer

scientists in Sandwich, England, went from being at ﬁ rst a drug for hypertension to a drug

for angina. Then it changed again when a 10-day tolerance study in Wales discovered its

unusual side effect: penile erection.

Sildenaﬁ l was originally provided for clinical use as

a hypotensive and cardiotonic substance; its usefulness in male erectile dysfunction clearly

resulted from the clinical observations.

7.3 PHYSIOPATHOLOGICAL HYPOTHESES

7.3.1 Discovery of Levodopa

The amino acid L-dihydroxyphenylalanine (levodopa) was not used in medicine until the

role of dopamine as a neurohormone was discovered by Carlsson et al.

Work carried out

from the 1930s through to the 1950s identiﬁ ed the biosynthetic pathway (in chromafﬁ n

tissue and in adrenergic neurons) that linked dietary amino acids with catecholamines.

17,18

Scheme 7.4 Structures of imipramine (9) and chlorimipramine (10) compared to that of

chlorpromazine (8).

8 chlorpromazine

9 imipramine (X = H)

10 chlorimipramine (X = Cl)

Scheme 7.5 Structure of the phosphodiesterase type 5 (PDE5) inhibitor sildenaﬁ l (11).

PHYSIOPATHOLOGICAL HYPOTHESES 217

218 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

A multistage biosynthetic pathway was proposed for the synthesis of adrenaline from

L-tyrosine (Scheme 7.6). In this pathway, L-tyrosine is converted to levodopa, which

undergoes decarboxylation to form dopamine. Dopamine is the immediate precursor of

noradrenaline, which is converted to adrenaline.

An enzyme known as levodopa decarboxylase was discovered in 1939 which degrades

any levodopa present in mammalian tissues and which thus hindered its detection. Le-

vodopa can, however, be administered to correct dopamine deﬁ ciency in Parkinson’s dis-

ease. It behaves like a pro-drug, undergoing metabolic conversion to dopamine once it has

entered the brain.

The introduction of racemic dihydroxyphenylalanine into therapeutic use occurred at

the University of Vienna after Hornykiewicz

had gathered evidence that pointed to there

being a depletion of dopamine reserves in the brains of patients with Parkinson’s disease.

Since dopamine was too polar to cross the blood–brain barrier, Hornykiewicz attempted

to alleviate the disease by administering 50 to 150 mg of dihydroxyphenylalanine intrave-

nously to 20 patients and used this metabolic precursor of dopamine, since the neurohor-

mone itself could not cross over into the brain from the general circulation.

His results

seemed favorable, as were those reported around the same time by Barbeau

at the Univer-

sity of Montreal. Barbeau administered dihydroxyphenylalanine by mouth to six patients.

These ﬁ ndings were, however, disputed by others, and it was not until 1967 that treatment

Scheme 7.6 Dopamine biosynthesis and metabolism.

OCH

COMT

OCH

COMT

OCH3

COMT

L-dopa

tyrosine

hydroxylase

3-O-methyl-L-dopa

L-tyrosine

3-methoxytyramine

3-methoxy-4-dihydroxy-

phenylacetic acid

dopamine

3,4-dihydroxyphenyl-

acetic acid (DOPAC)

aromatic acid

decarboxylase

monoamine-

oxidase

aldehyde-

deshydrogenase

protocols were perfected by Cotzias

and his colleagues at the Medical Research Centre

of Brookhaven National Laboratory. They demonstrated that oral doses of up to 16 g each

day consistently improved the general clinical condition of more than 50% of patients. This

improvement lasted only while treatment continued. Because of the expense involved, ra-

cemic dihydroxyphenylalanine had been used in the early trials. Since that time, levodopa

(the optically active isomer that is the metabolic precursor of dopamine) has become the

universal treatment for Parkinson’s disease. Several hundred thousand patients have ben-

eﬁ ted from this treatment. Of the DOPA administered by the oral route, however, 95%

undergoes decarboxylation in the peripheral circulation before crossing the blood–brain

barrier. To prevent the peripheral DOPA from undergoing this unwanted premature degra-

dation, a peripheral inhibitor of DOPA–decarboxylase is usually added to the treatment. An

additional improvement to the treatment involves the simultaneous addition of a catechol

O-methyltransferase (COMT) inhibitor such as tolcapone or entacapone.

7.3.2 H

-Receptor Antagonists

Research into the development of speciﬁ c antagonists for the H

histamine receptor in the

treatment of gastric ulcers has also been carried out using a rational physiopathological

process.

23,24

Starting from the observation that the antihistaminic compounds known at

the time (H

-receptor antagonists) were not capable of acting as antagonists for the gastric

secretion provoked by histamine, Black and his collaborators envisaged that an unknown

subclass of the histamine receptor (the future H

receptor) existed. From 1964 on, they

initiated a program of systematic research for speciﬁ c antagonists for this receptor.

The starting point was guanylhistamine (12, Scheme 7.7), which exhibits weak antago-

nistic properties against the gastric secretion that histamine induced. Lengthening the side

chain of this compound clearly increased H

antagonistic activity, but a residual agonist

PHYSIOPATHOLOGICAL HYPOTHESES 219

Scheme 7.7 Structures of some key compounds in the development of H

-receptor antagonists.

12 N

-guanylhistamine

13 burimamide

14 metiamide 15 cimetidine

16 ranitidine 17 famotidine

18 roxatidine

220 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

effect remained. By replacing the strongly basic guanidino function by a neutral thiourea,

burimamide (13), was obtained. Although very active, this compound was rejected due to

its low oral bioavailability. The addition of a methyl group in position 4 of the imidazolic

ring, followed by the introduction of an electron-withdrawing sulfur atom in the side chain,

ﬁ nally led to a compound that was both very active and less highly ionized, properties

that improved its absorption by the oral route. The derivative thus obtained, metiamide

(14), was excellent and, moreover, 10 times more potent than burimamide. Because of its

thiourea grouping, however, metiamide exhibited undesirable side effects (agranulocytosis,

nephrotoxicity) that would limit its clinical use. Replacement of the thiourea by an isosteric

group that had the same pK

(N-cyanoguanidine) ﬁ nally led to cimetidine (15), which be-

came a medicine of choice for the treatment of gastric ulcers. It later appeared that the

imidazolic ring that was present in histamine and in all H

antagonists discussed hitherto

was not essential for the H

antagonistic activity. Thus, ranitidine (16), which possesses a

furan ring, has proved to be even more active than cimetidine. The same proved to be true

for famotidine (17) and roxatidine (18).

7.3.3 Rimonabant and Obesity

Among drug-taking communities it was well known that smoking of marijuana was fol-

lowed by an increase in appetite. One consequence of this was the widespread use of

smoked marijuana as a treatment for HIV-related anorexia and weight loss. This activity

also suggested that orally administrated synthetic tetrahydrocannabinol (THC, dronabinol,

Marinol), the main psychoactive ingredient in marijuana, could be used as an appetite

stimulant. The treatment has proved, in fact, to be efﬁ cacious.

Numerous pharmacologi-

cal studies conﬁ rmed that the CNS receptor involved was the cannabinoid CB1 receptor.

Exogenously administered cannabinoid receptor agonists such as ∆

-tetrahydrocannabinol

stimulate food consumption in animals as well as in humans. Endogenous cannabinoid

receptor agonists are present in the brain, and the level of these agonists in the brain in-

creases with greater demand for food by rodents. Conversely, administration of CB1 recep-

tor antagonists was hypothesized as being a means of reducing food intake and represented

a treatment for obesity. This hypothesis prompted the scientists of Sanoﬁ Aventis to start a

screening program for CB1 receptor antagonists. This search culminated in the discovery

of compound SR 141716A (rimonabant; 19, Scheme 7.8). This selective CB1 receptor

antagonist is in phase III clinical trials for the treatment of obesity and has been found to

decrease appetite and body weight in humans.

Scheme 7.8 Structures of the cannabinoid receptor CB1 antagonists rimonabant (19) and SLV 319 (20).

,HCl

O O

19 rimonabant (SR141716A) 20 SLV 319

Cannabinoid CB1 receptor antagonists are currently the subject of intensive research,

due to their highly promising therapeutic prospects. Novel chemical entities that have

CB1 antagonistic properties have recently been disclosed by several pharmaceutical com-

panies and some academic research groups. Some of these entities are close structural

analogs of the lead compound rimonabant. A considerable number of these CB1 antagonists

are bioisosteres that derived from rimonabant by the replacement of the pyrazole moiety

with an alternative heterocycle. As well as these achiral compounds, Solvay Pharmaceuticals

have disclosed a novel class of chiral pyrazolines [such as compound SLV 319 (20), (Scheme

7.8)] which are potent and CB1/CB2 subtype-selective cannabinoid receptor antagonists.

7.4 CONTRIBUTIONS FROM CLINICAL INVESTIGATIONS

The clinical observation of entirely unexpected side effects constitutes a nearly inexhaust-

ible source of avenues to follow in the search for lead compounds. Indeed, in addition to the

desired therapeutic action, most drugs possess side effects that are either accepted from the

beginning as a necessary evil or are recognized only after some years of utilization. When

side effects themselves are of medical interest, the disassociation of the primary effect from

the side-effect activities may become an objective: Enhance the activity that was originally

considered as secondary and minimize or eliminate the activity that was initially dominant.

Promethazine, for example, an antihistaminic derivative of phenothiazine, has important

undesirable sedative effects. To their credit, clinicians such as Laborit

have promoted

the utilization of this side effect and have directed research toward better proﬁ led analogs.

The emergence of chlorpromazine, the prototype for a new therapeutic series, the neuro-

leptics (the existence of which was previously unsuspected and which have revolutionized

psychiatric practice) was the result of this impulse.

28,29

Countless other examples can be

found in the literature, such as the hypoglycemic effect of some antibacterial sulfamides,

the uricosuric effect of the coronaro-dilating drug benziodarone, the antidepressant effect

of isoniazide, an antituberculosis drug, and the hypotensive effect of β-blocking agents.

In some cases a new clinical activity observed in an existing drug is sufﬁ ciently potent

and of sufﬁ cient interest to justify the immediate use of the drug in the new indication, as

will be illustrated hereafter. Amiodarone (21, Scheme 7.9), for example, was introduced

as a coronary dilator for angina. Concern about corneal deposits, discoloration of skin that

was exposed to sunlight and thyroid disorders led to the drug being withdrawn in 1967. In

1974, however, it was discovered that amiodarone was highly effective in the treatment of

a rare type of arrhythmia known as the Wolff–Parkinson–White syndrome. Accordingly,

amiodarone was reintroduced speciﬁ cally for this purpose.

CONTRIBUTIONS FROM CLINICAL INVESTIGATIONS 221

Scheme 7.9 Structures of the benzofuranic arones.

21 amiodarone 22 benziodarone 23 benzbromarone

222 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

Benziodarone (22), used initially in Europe as a coronary dilator, later proved to

be a useful uricosuric agent. It is at present withdrawn from the market as a result of

there having been several cases of jaundice associated with its use.

The corresponding

brominated analog, benzbromarone (23), was marketed speciﬁ cally for its uricosuric

properties.

Thalidomide (24), was initially launched as a sedative/hypnotic drug (Scheme 7.10), but

withdrawn because of its extreme teratogenicity. Under restricted conditions (no adminis-

tration during pregnancy or to any woman of childbearing age), it has found a new use as an

immunomodulator. In particular, it appears to be efﬁ cacious in the treatment of erythema

nodosum leprosum, a possible complication in the chemotherapy of leprosy.

In 2001, the antimalarial drug quinacrine (25) and the antipsychotic drug chlorproma-

zine (8, Scheme 7.11) were shown to inhibit prion infections in cells. Prusiner and co-

workers

identiﬁ ed the drugs independently and found that they inhibit conversion of nor-

mal prion protein into infectious prions and clear prions from infected cells. Both drugs can

cross over from the bloodstream into the brain, where prion diseases are localized. In many

therapeutic families a new generation of compounds is born from the previous generation.

In the past this occurred for the sulfonamides, penicillins, steroids, prostaglandins, and

tricyclic psychotropics families, and one can draw true genealogical trees that represent the

progeny of these discoveries. More recent examples are to be found in ﬁ elds that include

statins, ACE inhibitors, and in the family of histaminergic H

antagonists.

Research programs based on the exploitation of side effects are of great interest in

the discovery of new avenues insofar as they depend on information about activities that

have been observed directly in humans and not in animals. On the other hand, they enable

new therapeutic activities to be detected even when no pharmacological models exist in

animals.

Scheme 7.10 Structure of thalidomide. The marketed compound is the racemate.

24 thalidomide

Scheme 7.11 Old drugs, new use: the antimalarial drug quinacrine (25) and the antipsychotic drug

chlorpromazine (8) can inhibit prion infections.

MeO

25 quinacrine 8 chlorpromazine

7.5 NEW LEADS FROM OLD DRUGS: THE SOSA APPROACH

The SOSA (selective optimization of side activities) approach represents an original alter-

native to HTS.

33–35

It involves two steps:

1. Screening of newly identiﬁ ed pharmacological targets using a limited set (approxi-

mately 1000 compounds) of well-known drug molecules for which bioavailability

and toxicity studies have already been performed and which have proven to be useful

in human therapy.

2. Once a hit has been obtained with a given drug molecule, the task is then to prepare

analogs of this molecule to transform the observed side activity into the main effect

and to reduce signiﬁ cantly or eliminate the initial pharmacological activity.

7.5.1 Rationale

The rationale behind the SOSA approach is based on the fact that in addition to their main

activities, almost all drugs used in human therapy exhibit one or more side effects. In other

words, as well as being capable of exerting a strong interaction with the main target, they

also exert weaker interactions with other biological targets. Most of these targets bear no

relation to the primary therapeutic activity of the compound. The objective is then to move

toward a reversal of these afﬁ nities so that the side effect identiﬁ ed becomes the main ef-

fect, and vice versa.

A chemical library that is available for the SOSA approach is the Prestwick Chemi-

cal Library (Prestwick Chemical, Inc., 1825 K Street NW, Suite 1475, Washington, DC

20006-1202; www.prestwickchemical.com). It contains 1120 biologically active com-

pounds that exhibit a high degree of chemical and pharmacological diversity as well as

known bioavailability and safety in humans. Over 85% of the compounds are well-es-

tablished drugs, with 15% of them being bioactive alkaloids. For scientists who have an

interest in drug similarities, this library most certainly fulﬁ lls their requirements in the

quest for druglike leads!

7.5.2 Examples

Selective Ligands for the Endotheline ET

Receptors The development of these ligands

by scientists from Bristol-Myers Squibb provide us with an illustration of the SOSA

approach.

36,37

Starting with an in-house library, the antibacterial compound sulfathiazole

(26, Scheme 7.12) was an initial, but weak, hit (IC

⫽ 69 µM). Testing of related

sulfonamides identiﬁ ed the more potent sulﬁ soxazole (27) (IC

⫽ 0.78 µM). Systematic

variation ﬁ nally led to the potent and selective ligand BMS-182874. In vivo, this compound

was orally active and produces a long-lasting hypotensive effect.

Further optimizations that were guided by pharmacokinetic considerations led the BMS

scientists to replace the naphtalene ring with a diphenyl system.

Among the compounds

prepared 29 (BMS-193884, ET

⫽ 1.4 nM; ET

⫽ 18,700 nM) showed promising he-

modynamic effects in a phase II clinical trial on congestive heart failure. More recent stud-

ies have led to an extremely potent antagonist (30; BMS-207940 ET

⫽ 10 pM), which

exhibits an 80,000-fold selectivity ratio for ET

versus ET

. The bioavailability of 30 is

100% in rats and it exhibits oral activity even at a 3 µM/kg dosage.

NEW LEADS FROM OLD DRUGS: THE SOSA APPROACH 223

224 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

Cholinergic Agonists In our second example, the starting lead was the antidepressant

minaprine (31, Scheme 7.13). In addition to properties of reinforcing serotonergic

and dopaminergic transmission, this aminopyridazine possesses a weak afﬁ nity for

muscarinic M

receptors (K

⫽ 17 µM). Simple chemical variation was used to eliminate

dopaminergic and serotonergic activities and to boost the cholinergic activity to nanomolar

concentrations.

40–42

Acetylcholinesterase Inhibitors Starting from the same minaprine lead, since this

molecule is recognized by the acetylcholine receptors, it was conceivable that it might

Scheme 7.12 A successful SOSA approach was used in the identiﬁ cation of the antibacterial sul-

fonamide sulfathiazole (26) as a ligand for the endothelin ETA receptor and in its optimization to pro-

vide the selective and potent compounds BMS-182874 (28), BMS-193884 (29), and BMS-207940

(30). (From ref. 38 and 39.)

O O

26 sulfathiazole 27 sulfisoxazole

28 BMS-182874

= 0.15 µMET

= 0.78 µMET

= 69 µM

29 BMS-193884

Ki = 1.4 nM ET

Ki = 0.010 nM

30 BMS-207940

Scheme 7.13 Progressive transformation from minaprine to a potent and selective partial musca-

rinic M1 agonist. (From refs. 40 to 42.)

N N

31 minaprine

=17,000 nM

= 50 nM

= 550 nM

= 3 nM

also be recognized by the acetylcholine enzyme. It turned out that minaprine had only a

very weak afﬁ nity for acetylcholinesterase (600 µM on electric eel enzyme). Relatively

simple modiﬁ cations, however (creation of a lipophilic cationic head, increasing the length

of the side chain, and bridging the phenyl and the pyridazinyl rings) allowed us to obtain

nanomolar afﬁ nities (Scheme 7.14).

43,44

CRF Antagonists Another interesting switch involved the progressive transformation

from desmethylminaprine (40) to the bioisosteric thiadiazole (41, Scheme 7.15) and then

to bioisosteric thiazoles. Trisubstitution of the phenyl ring and replacement of the aliphatic

morpholine by a pyridine led to compound 42, which exhibited some afﬁ nity for the

receptor of the 41-amino acid neuropeptide, corticotrophin releasing factor (CRF). Further

optimization led to nanomolar CRF antagonists such as 43.

45,46

Scheme 7.14 IC

values for acetylcholinesterase inhibition (electric eel enzyme). (From refs. 43 and 44.)

36 desmethyl-deoxy-

minaprine 13 µM

37 side chain variant

1.9 µM

31 minaprine

600 µM

38 N-benzyl-piperidino

side chain

0.12 µM

39 bridged analogue

0.010 µM

NEW LEADS FROM OLD DRUGS: THE SOSA APPROACH 225

Scheme 7.15 Switch from the antidepressant molecule minaprine to the potent CRF receptor an-

tagonist.

(From refs. 45 and 46.)

N N

MeO

226 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

HIV Protease Inhibitors Yielded by The Anticoagulant Phenprocoumon At Upjohn,

phenprocoumon (44, Scheme 7.16), used therapeutically as an anticoagulant, was

independently discovered to be a moderately active HIV protease inhibitor (K

⫽ 1 µM).

Optimization produced the bis-aralkyl-substituted 4-hydroxypyrone (45; PNU-96 988,

⫽ 38 nM)

, and ﬁ nally, the picomolar inhibitor tipranavir (46; R,R-diastereomer:

⫽ 8 pM). Surprisingly, other diastereomers of tipranavir show a very low stereospeciﬁ city

of drug action; they are also very potent HIV protease inhibitors (R,S-diastereomer:

⫽ 18 pM; S,R-diastereomer: K

⫽ 32 pM; S,S-diastereomer: K

⫽ 220 pM).

From the Antibiotic Erythromycin to LHRH Antagonists

Screening of the Abbott

chemical repository identiﬁ ed erythromycin A derivatives that bound to rat LHRH receptor

with submicromolar afﬁ nity and which exhibited LHRH antagonistic properties. One

of the most potent antagonists was an anilinoethyl cyclic carbamate (47, Scheme 7.17).

Initial SAR studies, guided by overlaying compound 47 with the potent cyclic decapeptide

(48), led to the para-chlorophenyl analog 49, which exhibited a 20-fold improvement in

potency relative to compound 47. Optimization efforts based mainly on replacement of the

cladinose moiety in position 3 led to compound 50, which has 1 to 2 nM afﬁ nity for both rat

and human LHRH receptors and is a potent in vitro inhibitor of LH release (pA

⫽ 8.76). In

vivo, compound 50 was found to produce a dose-dependent suppression of LH in castrated

male rats via both intravenous end oral dosing.

7.5.3 Discussion

The SOSA approach appears to be an efﬁ cient strategy for the discovery of drugs, particu-

larly since it is based on the screening of drug molecules and thus yields druglike hits auto-

matically. It can represent an attractive alternative before a costly HTS campaign is started.

Once the initial screening has provided a hit, this will be used as the starting point for a

drug discovery program. Using both traditional medicinal chemistry and parallel synthesis,

the initial side activity is transformed into the main activity, and conversely, the initial main

activity is greatly reduced or eliminated. There is a high probability that this strategy will

lead to safe, bioavailable, original, and patentable analogs.

Scheme 7.16 The anticoagulant phenprocoumon yields the HIV-protease inhibitor tipranavir.

O O

(R)

46 tipranavir Ki = 8 pM

44 phenprocoumon Ki = 1 µM

45 PNU-96-988 Ki = 38 nM

Safety and Bioavailability Over a period of years of practicing SOSA approaches, it has

been observed that when analog synthesis is performed with a drug molecule as the lead

substance, there is a notably increased probability of obtaining safe new chemical entities.

In addition, most of these satisfy Lipinski’s,

Veber’s,

Bergström’s,

and Wenlock’s

observations in terms of solubility, oral bioavailability, and drug-likeness.

Patentability When a well-known drug hits with a new target, there is a possibility that

several hundred or even thousands of analogs of the original drug molecule have already

been synthesized by the original inventors and their competitors. These molecules are

usually protected by patents or already belong in the public domain. At ﬁ rst glance it would

appear that there is a high probability of interference. In fact, during optimization of a

therapeutic proﬁ le different than that of the original inventors, medicinal chemists rapidly

prepare analogs whose chemical structures are notably different from that of the original

hit. As an example, a medicinal chemist interested in phosphodiesterases (PDEs) and using

diazepam as a lead will inevitably synthesize compounds that are in terms of structure

outside the scope of the original patents, precisely because they exhibit predominantly

PDE-inhibiting properties and show practically no further afﬁ nity for the benzodiazepine

receptor (Scheme 7.18).

Originality The screening of a library of several hundred therapeutically diverse drug

molecules sometimes ends up with very surprising results. A nice example of unexpected

ﬁ ndings resulting from systematic screening is to be found in the tetracyclic compound

O O

OMe

O O

OMe

(pKi =7.0)

(pKi = 8.4)

(pKi = 9.17)

Scheme 7.17 From derivatives of the antibiotic erythromycin A to LHRH antagonists.

NEW LEADS FROM OLD DRUGS: THE SOSA APPROACH 227

228 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

54 (BMS-192548) extracted from Aspergillus niger WB2346 (Scheme 7.19). For any

medicinal chemist or pharmacologist, the similarity of this compound to the antibiotic

tetracycline is striking. No one, however, would forecast a priori that BMS-192548 would

exhibit CNS activities. The compound turns out in fact to be a ligand for neuropeptide Y

receptor preparations.

Orphan Diseases As mentioned above, a peculiarity of this type of library that

differentiates it from others is that it is made up of compounds that have already been given

safely to humans. Thus, if a compound were to hit with sufﬁ cient potency on an orphan

target, there is a signiﬁ cant chance that it could be tested rapidly on patients for proof of

principle. This possibility represents another advantage of the SOSA approach.

7.6 CONCLUSIONS

The relative lack of pharmaceutical creativity over recent years is at least in part attribut-

able to a misleading faith that identiﬁ cation of a new hit or lead molecule is the most impor-

tant step in a drug discovery program. As mentioned earlier, “hits are not drugs.” Problems

in later development, such as pharmacokinetic and toxicological proﬁ ling (ADME, bio-

availability, genotoxicity, etc.) or pharmaceutical formulation (salts, polymorphism, water

solubility, etc.) are much too often regarded as being straightforward ancillary activities. It

can currently be assumed that most pharmaceutical companies possess dozens of hits and

that their main concern is to transform them into valuable drug candidates.

Scheme 7.18 Starting from the tranquillizer diazepam (51), the improvement in the PDE (4) inhibi-

tory activity leads to structures such as 52 that are sufﬁ ciently original to render them patentable.

51 diazepam 52 CI-1044

Scheme 7.19 Unexpected CNS activity of the tetracycline analog 54 (BMS-192548). (From

ref. 55.)

53: tetracycline

54: BMS-192548

The four alternatives to high-throughput screening that have been proposed here (i.e.,

analog design, research based on physiopathological hypotheses, exploitation of clinical

observation, and the identiﬁ cation of new leads from old drugs) share a precious common

advantage: They are all derived from a knowledge of properties that are inherent in existing

drug molecules of proven activity and/or usefulness in humans. Many of the drawbacks

that are encountered in the development of HTS-derived molecules are therefore bypassed

and fewer safety problems should arise. For all of these reasons, it is recommended that

pharmaceutical company management adopt these alternatives as fruitful and complemen-

tary approaches to HTS.

REFERENCES

Ramesha, C. S. How many leads from HTS?—Comment. Drug Discov. Today 2000, 5, 43–44.

Liu, K.; Xu, L.; Szalkowski, D.; Li, Z.; Ding, V. Discovery of a potent, highly selective, and

orally efﬁ cacious small-molecule activator of the insulin receptor. J. Med. Chem. 2000, 43,

3487–3494.

Roever, S.; Adam, G.; Cesura, A. M.; Galley, G.; Jenck, F. E. A. High-afﬁ nity, non-peptide ago-

nists for the ORL1 (orphanin FQ/nociceptin) receptor. J. Med. Chem. 2000, 43, 1329–1338.

Rey, A. Dictionnaire historique de la langue française, Dictionnaires Le Robert, Paris, 1992.

Schneider, G.; Giller, T.; Neidhart, W.; Schmid, G. “Scaffold-hopping” by topological phar-

macophore search: a contribution to virtual screening. Angew. Chem. Int. Ed. 1999, 38, 2894–

2896.

Balakin, K. V.; Tkachenko, S. E.; Okun, I.; Skorenko, A. V.; Ivanenkov, Y. A.; Savchuk, N. P.;

Ivashchenko, A. A.; Nikolsky, Y. Bioisosteric morphing in primary hit optimization. Chimi. Oggi

(Chem. Today) 2004, 15–18.

Buehlmayer, P.; Furet, P.; Cricione, L.; de Gasparo, M.; Whitebread, S.; Schmidlin, T.; Lattman,

R.; Wood, J. Valsartan, a potent, orally active angiotensin II antagonist developed from the struc-

turally new amino acid series. Bioorg. Med. Chem. Lett. 1994, 4, 29–34.

Croom, K. F.; Keating, G. M. Valsartan: a review of its use in patients with heart failure and/or

left ventricular systolic dysfunction after myocardial infarction. Am. J. Cardiovasc. Drugs 2004,

4, 395–404.

Bajorath, J. Integration of virtual and high-throughput screening. Nature Rev. Drug Discov.

2002, 1, 882–894.

Kubinyi, H.; Müller, G. Chemogenomics and drug discovery. In Methods and Principles in

Medicinal Chemistry, Mannhold, R., Kubinyi, H., Folkers, G. (Eds), Wiley-VCH, Weinheim,

Germany, 2004, p. 463.

Kuhn, R. Über die Behandlung depressiver Zuständen mit einem Iminodibenzylderivat (G22355).

Schweiz. Med. Wochenschr. 1957, 87, 1135–1140.

Kuhn, R. The treatment of depressive states with G22355 (imipramine hydrochloride). Am. J.

Psychiatry 1958, 115.

Terret, N. K.; Bell, A. S.; Brown, D.; Ellis, P. Sildenaﬁ l (Viagra), a potent and selective inhibitor

of type 5 cGMP phosphodiesterase with utility for the treatment of male erectile dysfunction.

Bioorg. Med. Chem. Let. 1996, 6, 1819–1824.

Boolell, M.; Allen, M. J.; Ballard, S. A.; Gepi-Attee, S.; Muirhead, G. J.; Naylor, A. M.; Osterloh,

I.; Gingell, C. Sildenaﬁ l: an orally activetype 5 cyclic GMP-speciﬁ c phosphodiesterase inhibitor

for the treatment of penile erectile dysfunction. Int. J. Impot. Res. 1996, 8, 47–52.

Kling, J. From hypertension to angina to Viagra. Mod. Drug Discov. 1998, 31–38.

10.

11.

12.

13.

14.

15.

REFERENCES 229

230 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

Carlsson, A.; Lindquist, M.; Magnusson, T.; Waldeck, P. On the presence of 3-hydroxytyramine

in the brain. Science 1958, 127, 471.

Blaschko, H. Metabolism and storage of biogenic amines. Experientia 1957, 13, 9–12.

Blaschko, H. The development of current concepts of catecholamine formation. Pharmacol. Rev.

1959, 11, 307–316.

Ehringer, H.; Hornykiewicz, O. Verteilung von Noradrenalin und Dopamin (3-hydroxytyramin)

im Gehirn des Menschen und ihr Verhalten bei Erkrankungen des extrapyramidalen Systems.

Wien. Klin. Wochenschr. 1960, 38, 1236–1239.

Birkmayer, W.; Hornykiewicz, O. The

L-dihydroxyphenylalanine (L-dopa) effect in Parkinson’s

syndrome in man: on the pathogenesis and treatment of Parkinson akinesis. Arch. Psychiatr.

Nervenkr. Z. Gesamte Neurol. Psychiatr. 1962, 203, 560–574.

Barbeau, A. The use of L-dopa in Parkinson’s disease: a 20 year follow-up. Trends Pharmacol.

Sci. 1981, 2, 297–299.

Cotzias, G. C.; Van Woert, M. H.; Schiffer, L. M. Aromatic amino acids and modiﬁ cation of

parkinsonism. New Engl. J. Med. 1967, 276, 374–379.

Black, J. W.; Duncan, W. A. M.; Durant, J. C.; Ganellin, C. R.; Parsons, M. E. Deﬁ nition and

antagonism of histamine H2-receptors. Nature 1972, 236, 385–390.

Ganellin, C. R. Cimetidine. In Chronicles of Drug Discovery, Wiley, New York, 1982,

pp. 1–38.

Abrams, D. I. Medical marijuana: tribulations and trials. J. Psychoactive Drugs 1998, 30,

163–169.

Lange, J. H.; Kruse, C. G. Recent advances in CB1 cannabinoid receptor antagonists. Curr. Opin.

Drug Discov. Dev. 2004, 7, 498–506.

Laborit, H.; Huguenard, P.; Alluaume, R. Un nouveau stabilisateur végétatif, le 4560 R.P. Presse

Méd. 1952, 60, 206–208.

Thuilier, J. Les dix ans qui ont changé la folie, Robert Laffont, Paris, 1981 pp. 253–257.

Maxwell, R. A.; Eckhardt, S. B. Drug Discovery: A Casebook and Analysis, Humana Press,

Clifton, NJ, 1990.

Sneader, W. Drug Prototypes and Their Exploitation, Wiley, Chichester, West Sussey, England,

1996, p. 242.

Iyer, C. G. S.; Languillon, J.; Ramanujam, K. WHO coordinated short-term double-blind trial

with thalidomide in the treatment of acute lepra reactions in male lepromatus patients. Bull.

W.H.O. 1971, 45, 719–732.

Korth, C.; May, B. C. H.; Cohen, F. E.; Prusiner, S. B. Acridine and phenothiazine derivatives as

pharmacotherapeutics for prion disease. Proc. Natl. Acad. Sci. USA 2001, 98, 9836–9841.

Wermuth, C. G. Search for new lead compounds: the example of the chemical and pharmacologi-

cal dissection of aminopyridazines. J. Heterocycl. Chem. 1998, 35, 1091–1100.

Wermuth, C. G.; Clarence-Smith, K. “Drug-like” leads: bigger is not always better. Pharm. News

2000, 7, 53–57.

Wermuth, C. G. The “SOSA” approach: an alternative to high-throughput screening. Med. Chem.

Res. 2001, 10, 431–439.

Riechers, H.; Albrecht, H.-P.; Amberg, W.; Baumann, E.; Böhm, H.-J.; Klinge, D.; Kling, A.;

Muller, S.; Raschack, M.; Unger, L.; Walker, N.; Wernet, W. Discovery and optimization of a

novel class of orally active non-peptidic endothelin-A receptor antagonists. J. Med. Chem. 1996,

39, 2123–2128.

Poroikov, V.; Akimov, D.; Shabelnikova, E.; Filimonov, D. Top 200 medicines: can new actions

be discovered through computer-aided prediction? SAR QSAR Environ. Res. 2001, 12, 327–344.

Stein, P. D.; Hunt, J. T.; Floyd, D. M.; Moreland, S.; Dickinson, K. E. J.; Mitchell, C.; Liu, E.

C.-K.; Webb, M. L.; Murugesan, N.; Dickey, J.; McMullen, D.; Zhang, R.; Lee, V. G.; Serdino,

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

R.; Delaney, C.; Schaeffer, T. R.; Kozlowski, M. The discovery of sulfonamide endothelin an-

tagonists and the development of the orally active ET

antagonist 5-(dimethylamino)-N- (3,4-

dimethyl-5-isoxazoly1)-l-naphthalenesulfonamide. J. Med. Chem. 1994, 37, 329–331.

Murugesan, N.; Gu, Z.; Spergel, S.; Young, M.; Chen, P.; Mathur, A.; Leith, L.; Hermsmeier,

M.; Liu, E. C.-K.; Zhang, R.; Bird, E.; Waldron, T.; Marino, A.; Koplowitz, B.; Humphreys,

W. G.; Chong, S.; Morrison, R. A.; Webb, M. L.; Moreland, S.; Trippodo, N.; Barrish, J. C.

Biphenylsulfonamide endothelin receptor antagonists. 4. discovery of N-[[(4,5-dimethyl-3-iso

xazolyl)amino]sulfonyl]-4-(2-oxazolyl)[1,1'-biphenyl]-2-yl]methyl]-N,3,3-trimethylbutanamide

(BMS-207940, a highly potent and orally active ET

selective antagonist. J. Med. Chem. 2003,

46, 125–137.

Wermuth, C. G.; Schlewer, G.; Bourguignon, J.-J.; Maghioros, G.; Bouchet, M.-J.; Moire, C.;

Kan, J.-P.; Worms, P.; Bizière, K. 3-Aminopyridazine derivatives with atypical antidepressant,

serotonergic and dopaminergic activities. J. Med. Chem. 1989, 32, 528–537.

Wermuth, C. G. Aminopyridazines: an alternative route to potent muscarinic agonists with no

cholinergic syndrome. Il Farmaco 1993, 48, 253–274.

Wermuth, C. G.; Bourguignon, J.-J.; Hoffmann, R.; Boigegrain, R.; Brodin, R.; Kan, J.-P.; Sou-

brié, P. SR 46559A and related aminopyridazines are potent muscarinic agonist with no cholin-

ergic syndrome. Bior. Med. Chem. Lett. 1992, 2, 833–836.

Contreras, J.-M.; Rival, Y. M.; Chayer, S.; Bourguignon, J. J.; Wermuth, C. G. Aminopyridazines

as acetylcholinesterase inhibitors. J. Med. Chem. 1999, 42, 730–741.

Contreras, J.-M.; Parrot, I.; Sippl, W.; Rival, R. M.; Wermuth, C. G. Design, Synthesis and struc-

ture-activity relationships of a series of 3-[2-(1-benzylpiperidin-4-yl)ethylamino]pyridazine de-

rivatives as acetylcholinesterase inhibitors. J. Med. Chem. 2001, 44, 2707–2718.

Gully, D.; Roger, P.; Valette, G.; Wermuth, C. G.; Courtemanche, G.; Gauthier, C. Dérivés alkyl-

amino ramiﬁ és du thiazole, leurs procédés de préparation et les compositions pharmaceutiques

qui les contiennent; Elf-Sanoﬁ : French Demande 9207736; June 24, 1992.

Gully, D.; Roger, P.; Wermuth, C. G. 4-Phenyl-aminothiazole derivatives: method for preparing same

and pharmaceutical compositions containing said derivatives; Sanoﬁ : World patent, Jan. 9, 1997.

Thaisrivongs, S.; Tomich, P. K.; Watenpaugh, K. D.; Chong, K. T.; Howe, W. J.; Yang, C. P.;

Strohbach, J. W.; Turner, S. R.; McGrath, J. P.; Bohanon, M. J.; Lynn, J. C.; Mulichak, A. M.;

Spinelli, P. A.; Hinshaw, R. R.; Pagano, P. J.; Moon, J. B.; Ruwart, M. J.; Wikinson, K. F.; Rush,

B. D.; Zipp, G. L.; Dalga, R. J.; Schwende, F. J.; Howard, G. M.; Padbury, G. E.; Toth, L. N.;

Zhao, Z.; Koeplinger, K. A.; Kakuk, T. J.; Cole, S. L.; Zaya, R. M.; Piper, R. C.; Jeffrey, P. Struc-

ture-based design of HIV protease inhibitors: 4-hydroxycoumarins and 4-hydroxy-2-pyrones as

non-peptidic inhibitors. J. Med. Chem. 1994, 37, 3200–3204.

Turner, S. R.; Strohbach, J. W.; Tommasi, R. A.; Aristoff, P. A.; Johnson, P. D.; Skulnick, H. I.;

Dolak, L. A.; Sees, T. E. P.; Tomich, P. K.; Bohanon, M. J.; Horng, M. M.; Lynn, J. C.; Chong, K.

T.; Hinshaw, R. R.; Watenpaugh, K. D.; Janakiraman, M. N.; Thaisrivongs, S. Tipranavir (PNU-

140690): a potent, orally bioavailable nonpeptidic HIV protease inhibitor of the 5,6-dihydro-4-

hydroxy-2-pyrone sulfonamide class. J. Med. Chem. 1998, 41, 3467–3476.

Randolph, J. T.; Waid, P.; Nichols, C.; Sauer, D.; Haviv, F.; Diaz, G.; Bammert, G.; Besecke, L.

M.; Segreti, J. A.; Mohning, K. M.; Bush, E. N.; Wegner, C. D.; Greer, J. Nonpeptide luteinizing

hormone-releasing hormone antagonists derived from erythromycin A: design, synthesis, and

biological activity of cladinose replacement analogues. J. Med. Chem. 2004, 47, 1085–1097.

Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Experimental and computational ap-

proches to estimate solubility and permeability in drug discovery and development settings. Adv.

Drug. Deliv. Rev. 2001, 46, 3–26.

Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D. Molecular

properties that inﬂ uence the oral bioavailability of drug candidates. J. Med. Chem. 2002, 45,

2615–2623.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

REFERENCES 231

232 POSSIBLE ALTERNATIVES TO HIGH-THROUGHPUT SCREENING

Bergström, C. A. S.; Strafford, M.; Lazorova, L.; Avdeef, A.; Luthman, K.; Artursson, P. Absorp-

tion classiﬁ cation of oral drugs based on molecular surface properties. J. Med. Chem. 2003, 46,

558–570.

Wenlock, M. C.; Austin, R. P.; Barton, P.; Davis, A. M.; Leeson, P. D. A comparison of physi-

cochemical property proﬁ les of development and marketed oral drugs. J. Med. Chem. 2003, 46,

1250–1256.

Burnouf, C.; Auclair, E.; Avenel, N.; Bertin, B.; Bigot, C.; Calvet, A.; Chan, K.; Durand, C.;

Fasquelle, V.; Féru, F.; Gilbertsen, R.; Jacobelli, H.; Kebsi, A.; Lallier, E.; Maignel, J.; Martin,

B.; Milano, S.; Ouagued, M.; Pascal, Y.; Pruniaux, M.-P.; Puaud, J.; Rocher, M.-N.; Terasse,

C.; Wrigglesworth, R.; Doherty, A. M. Synthesis, structure-activity relationships, and pharma-

cological proﬁ le of 9-amino-4-oxo-1-phenyl-3,4,6,7-tetrahydro[1,4]diazepino[6,7,1-hi]indoles:

discovery of potent, selective phosphodiesterase type 4 inhibitors. J. Med. Chem. 2000, 43,

4850–4867.

Shu, Y. Z.; Cutrone, J. Q.; Klohr, S. E.; Huang, S. BMS-192548, a tetracyclic binding inhibitor of

neuropeptide Y receptors, from Aspergillus niger WB2346. II. Physico-chemical properties and

structural characterization. J. Antibiot. 1995, 48, 1060–1065.

52.

53.

54.

55.

233

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

PROTEOMICS AND DRUG DISCOVERY

SUSAN DANA JONES

BioProcess Technology Consultants, Inc.

Acton, Massachusetts

PETER G. WARREN

Independent Biotechnology Consultant

Lexington, Massachusetts

8.1 INTRODUCTION

Drug discovery is a process that is composed of many essential steps. A key component

of drug discovery is the identiﬁ cation of a therapeutically relevant molecular target whose

activation or inhibition by a pharmaceutical agent will have an impact on the state or pro-

gression of a disease. For example, inhibition of the proton pump in the gastric mucosa is

correlated directly with reduced stomach acidity and therefore reduced pain in the stomach

and esophagus caused by extended low-pH conditions. Direct or indirect inhibition of this

proton pump is the target of a number of successful drugs developed by companies such

as Astra Zeneca, Abbott, Altana Pharma, and others. The challenge for new drug discovery

efforts is to identify a novel molecular target for therapeutic intervention. In recent years,

genomics and proteomics have been hailed as technologies whose application to drug dis-

covery will revolutionize and enable novel therapeutic concepts through identiﬁ cation of

previously unknown disease-associated targets. Herein, the tenets of these technologies

are reviewed along with the ways in which data generated through genomics, proteomics,

and other related approaches are currently used to improve the success and efﬁ ciency of

drug discovery. More advanced applications of these technologies are on the horizon for

drug discovery, and these are reviewed here as well. Throughout, the primary focus is on

proteomics.

234 PROTEOMICS AND DRUG DISCOVERY

8.2 DRUG DISCOVERY PROCESS

8.2.1 Process Overview

Target Identiﬁ cation and Validation The process of drug discovery starts long before

the screening of compound libraries for molecules that bind to and affect the action of a

protein associated with disease. Drug discovery starts with the identiﬁ cation of the disease-

associated protein or molecular target. In the past, identiﬁ cation of molecular targets was a

painstakingly slow process, generally carried out by investigators who could spend entire

careers studying one pathway or one protein that is involved in some aspect of human

development, metabolism, or disease. In the course of deﬁ ning the role of a given protein

in the cellular life cycle, the direct relevance of this protein to a disease state would often

become clear. By removing or altering the target’s function from the cell, the target would

then be validated as a suitable target for drug screening. In recent times, methods to identify

panels of targets that are associated with particular pathways or disease states have emerged,

and the acceleration of target identiﬁ cation has led to a bottleneck in target validation.

Methods described in this chapter are being deployed to reduce the bottleneck and increase

the availability of new, useful, disease-associated molecular targets for development of new

classes of effective therapeutic compounds.

Screening for Hits Once a disease-associated molecular target has been identiﬁ ed and

validated in disease models, screening for a selective and potent inhibitor (or activator)

of the target is the next step. Libraries of compounds that are either synthetic chemicals,

peptides, natural or engineered proteins, or antibodies are exposed to the target in a manner

that will detect and isolate those members of the library that interact with and, preferably,

have an effect on the target. The compounds selected are called hits. Initially, screening

can be performed by searching for compounds that bind to the target, but binding is not

sufﬁ cient for therapeutic activity. More recent screening procedures include an activity-

based readout as part of the initial screening assay. For example, if the goal is to inhibit a

protein that is involved in activating the expression of a particular gene or set of genes, the

assay can include a readout to determine if the expression of the gene is reduced by the

compound. Such assays can be cell-based, but more often they are enzymatic assays that

can be performed in a high-throughput manner.

Lead Optimization Once the initial screening is performed, a large collection of hits is

obtained. These hits are then evaluated and the best ones identiﬁ ed in a process known

as lead optimization. During this stage, scientists determine which, if any, of the hits

selected has the appropriate properties to justify continued expenditure of resources on

the development of the compound as a clinical candidate. These properties include ease

of synthesis; adherence to the Lipinsky rules, which describe chemical characteristics

that are predictive of biodistribution and in vivo activity; speciﬁ city for the target; and

efﬁ cacy in the disease. The hits discovered during the screening process are therefore

characterized by a variety of biochemical, biophysical, and biological methods to narrow

the set down to a handful of compounds. Assays that measure each compound’s activity

directly are a useful ﬁ rst step at determining potency and can enable the most effective

compounds to be identiﬁ ed. The resulting smaller set of hits, or leads, is then tested in

more rigorous models of disease, including either cell-based or animal models of the

disease.

Pharmacology and Toxicology Once one or more potential lead compounds emerge from

the lead optimization efforts, the compound must be evaluated in multiple disease models,

and if possible compared to existing therapies for the same disease. Further, properties of the

compound are studied when delivered into a living organism. The classic set of properties

that must be appropriate include Absorption (through the intended route of administration),

distribution (what organs does it end up in?), metabolism (what are the by-products of cellular

metabolism of the compound, and what potential effects do these compounds have on the

organism and disease?), and excretion (how is it processed and eliminated from the body?).

Collectively known as ADME, this set of properties is essential to evaluate in multiple

species, in addition to measuring efﬁ cacy in sophisticated disease models. Design of ADME

and efﬁ cacy experiments is a crucial activity in drug discovery and development.

All drugs that are intended for human clinical applications must be tested for toxicity

using the same material that will be used in humans, manufactured by the same process.

Toxicology testing is a highly regulated process that is governed by regulatory authori-

ties such as the U.S. Food and Drug Administration (FDA). Developing a manufacturing

process for a drug (which is outside the scope of this chapter) and executing the necessary

toxicology program are very expensive and are therefore usually performed on only one

lead compound identiﬁ ed during lead optimization and subsequent efﬁ cacy testing. Only

if the compound fails at this stage will another compound from the same screening (if

possible) be advanced into preclinical development. Different companies, and in fact coun-

tries, have different standards for which products move into human clinical testing, but in

all cases the product’s safety must be adequately demonstrated before regulatory approval

can be obtained for advancing the product into humans.

Clinical Trials Human clinical testing follows an established process in most regulated

countries worldwide. If possible, the initial administration of a compound to a human subject

is performed on people who do not have the disease indication for which the product is

designed. The compound is administered at a dose that is signiﬁ cantly lower than the intended

therapeutic dose, and the trial subjects are monitored for any signs of toxicity. Most often,

these adverse events or side effects consist of headache, fever, nausea, or other discomforts.

If the trial subjects receiving the initial low dose have no or minimal side effects, the next

cohort of subjects is treated with a higher dose. Using this dose escalation method, the aim

is to determine the maximum tolerated dose (MTD) or to determine the safety of a dose

level that is known to generate therapeutic beneﬁ t. Depending on the disease indication, the

tolerance for adverse reactions is different. For example, almost all chemotherapeutic agents

for cancer are signiﬁ cantly toxic, but many cancer patients are willing to suffer these effects

in order to have a chance to overcome their disease and continue to live without cancer. Due

to their known toxicity, oncology drugs are therefore rarely tested in normal volunteers,

but in fact are often tested initially in patients who have failed all other chemotherapy or

other treatment regimens and have no other option for survival. Chronic, nonfatal diseases,

however, are not usually treated with agents that have severe side effects because the beneﬁ t

of the drug does not outweigh the risks and discomfort of the side effects.

Once the MTD has been determined, drugs usually move into phase II testing, in which

the drug’s efﬁ cacy is determined in a small select group of patients who have the disease

for which the drug was developed. These trials can also include a range of doses and dos-

ing regimens (modes of delivery, frequency, etc.), in order to measure the clinical efﬁ cacy

of different doses. Designing phase II trials and choosing endpoints or objectives for the

therapy that will accurately reﬂ ect the compound’s efﬁ cacy is a very demanding process,

DRUG DISCOVERY PROCESS 235

236 PROTEOMICS AND DRUG DISCOVERY

and many compounds can fail at this stage. Only 60% of drugs that enter phase II success-

fully complete this stage and move on to the pivotal phase III trial.

Phase III is considered pivotal because the drug is administered to a much larger group

of patients and is evaluated for efﬁ cacy with greater rigor than in phase II. These trials

can cost millions of dollars and therefore are attempted only when the phase II results are

highly convincing. There are regulatory hurdles that must be crossed to enter phase III.

The drug product that is used in phase III must be manufactured exactly as the ﬁ rst com-

mercial product will be made. The facility or reactor that is used, the process, the analytical

methods, and the formulation and vialing are all identical to the intended ﬁ nal product that

will be made and launched upon product approval. Following the successful completion of

phase III clinical testing, the owner of the compound then ﬁ les an application with regula-

tory authorities in various countries, such as the U.S. FDA, for permission to sell the com-

pound for the intended therapeutic indication. The FDA or other agency must then grant a

license to market the compound, and the new drug is launched onto the market.

8.2.2 Motivation for Improvement

The drug discovery and development process described above follows a logical and linear

path from target identiﬁ cation through completion of clinical trial and submission of an

application to the regulatory authorities seeking approval to market a drug. However, this

process is lengthy, cumbersome, and most important, generates compounds that are more

likely than not to fail during clinical testing. The major hurdles in drug discovery and

development today are:

• Time. From the initial discovery and validation of a molecular target to ﬁ nal market-

ing of an effective drug can take 10 to 15 years.

• Efﬁ ciency. a signiﬁ cant portion of the high cost of drug development is due to the very

high attrition rate of compounds that enter the clinic; only 12% succeed.

• Expense. Drug development can cost up to $800 million per successful compound.

Therefore, improved clinical outcomes of therapeutic candidates at all stages of develop-

ment would contribute signiﬁ cantly to reducing development costs and improving the num-

ber of candidates that succeed in obtaining approval for commercialization. New methods

that enable better target identiﬁ cation and validation, a deeper understanding of total systems

biology and the implications for any speciﬁ c drug or target in development, and methods to

understand the nature of molecular targets to enable more effective screening and lead identi-

ﬁ cation would be extremely valuable in meeting the needs for cost-effective development of

new and useful therapeutics for human disease. The emerging technologies described in this

chapter aim to address the many aspects of drug development and to enable more effective

understanding of targets and compounds through high-throughput high-content analysis.

8.3 HIGH-THROUGHPUT SCREENING APPROACHES

TO DRUG DISCOVERY

High-throughput screening (HTS) is a term that became common in drug discovery within

the past decade. The term refers to the ability to search through large compound libraries

to identify those compounds that have a desired binding or activity property. Together

with combinatorial chemistry, in which large chemical or biological libraries could be de-

signed and built, HTS was envisioned as the solution to a shrinking pool of new therapeutic

compounds in development. Unfortunately, the actual mechanics of HTS did not enable

screening of large libraries in a reasonable amount of time, and therefore although more

compounds were being screened, fewer hits were identiﬁ ed. For chemical libraries, the

trend in the recent past has been to design smaller libraries focused on particular families

of molecular targets.

During the screening process, the library is exposed to a protein target or, in some cases,

a cell, and those hits that bind and/or act on the target are identiﬁ ed. Most screens are

searching for an inhibitor of the target, and using this approach many inhibitors can readily

be discovered. The problem that is faced in drug discovery is not the output of the screening

procedures, but the failure of many of the compounds in subsequent efﬁ cacy studies in cells

or animals. Failure can occur for many reasons, but often the cause of compound failure in

a disease model is that the compound is only screened against a single target and in fact has

cross-reactivity to other proteins that may be necessary for a completely different function

in the organism. Understanding the exact structure and function of the target and of those

molecules that are most closely related to it is the ﬁ rst step in increasing the identiﬁ cation

of hits whose preclinical and clinical performance is more likely to be successful. The

emerging technologies described in the rest of this chapter are those technologies that are

designed to meet this need.

8.4 EMERGING TECHNOLOGIES AND APPROACHES: SCALE AND SPEED

We now turn our attention to newer, high-throughput technologies that show considerable

promise for improving and accelerating drug discovery and development. The biggest

shifts we consider here are in scale and focus: from primarily low-scale biological, bio-

chemical, and pharmacological methods to high-throughput high-content system-wide

data-driven approaches. These approaches can generate data on the scale of an entire bio-

logical system (cell, tissue, or organism). We begin to see the emergence of genomics,

proteomics, metabolomics, and so on, leading in the direction of an integrated systems

biology perspective. With this major shift to larger scale comes a resulting major increase

in speed. Much more signiﬁ cant biological information can be gathered and processed in

a much shorter time.

We look next brieﬂ y at genomics, then focus primarily on proteomics as a key set of

technologies that can help achieve this broadening of scope. We then look at the implica-

tions of this wider focus for improving the drug discovery process.

8.5 GENOMICS

The genome of an organism is the complete genetic makeup, the entire DNA complement,

of that organism. Genomics, then, is the study of entire genomes. The intention of execut-

ing the sequencing and analysis of the entire human genome was to enable more rapid and

effective identiﬁ cation of disease-associated genes and thereby provide drug companies

with prevalidated targets. The key areas of genomics study are the development and ap-

plication of tools for the prediction and detection of (1) genes, (2) sequence similarity, (3)

GENOMICS 237

238 PROTEOMICS AND DRUG DISCOVERY

motif/domain similarity, and (4) gene expression variations (measurement of mRNA levels

through microarray analysis: coexpression, regulation, etc.).

Genomics has indeed aided the drug discovery process signiﬁ cantly. Gene expression

analysis in particular has been used extensively with some success in the search for new drug

targets and for biomarkers of disease and therapeutic activity. However, in many respects,

genomics has raised more questions than it has answered. This is due to a fundamental

limitation: mRNA expression levels do not reliably predict protein expression levels.

It has been shown experimentally that gene expression results can be quite different

from protein expression results, even when assayed on the same sample under identical

conditions [Ideker et al., 2001]. There are a number of possible reasons for these discrepan-

cies. The main differentiators are as follows:

1. Half-life. RNA and proteins may have very different half-lives in the cell. RNA may

get degraded before translation [e.g., by RNA interference (RNAi)].

2. Post translational modiﬁ cations (PTMs). No information on PTMs is available from

transcripts. These are modiﬁ cations that, by deﬁ nition, occur after the transcript has

been translated into a protein.

3. Localization of protein. The site of activity of a protein cannot be predicted reliably

from transcripts.

4. Protein interactions. There is no information from transcripts on protein interactions

with nucleotides, phospholipids, ligands, or other proteins.

Since all four of these are central to a great many biological functions, these limitations

illuminate the need to study the function, structure, and interactions of the proteins them-

selves. It is the proteins in a living organism that are the molecular targets of most drugs on

the market and in development. The study of individual or small numbers of proteins has

been performed in academic laboratories worldwide for some time, but there is a growing

need, both for basic science and in drug discovery, to study the function, structure, and

interactions of proteins at a system-wide scale: This is the emerging ﬁ eld of proteomics.

In addition, it is expected that proteomics will become more tightly linked with genomics,

metabolomics, and so on, leading toward an integrated systems biology approach.

8.6 PROTEOMICS

In its widest sense, proteomics is the systematic study of the proteome: all the proteins that are

encoded in the genome of an organism. Normally, this means studying the set of proteins in a

cell or tissue. The term refers both to the study itself and to the set of techniques and methods

that enable this study. Due in part to the success of high-scale gene expression studies, it is

being recognized that studying proteins on the proteomic scale yields insights that are just not

possible when studying proteins in isolation. A good example of such an insight is when a

protein–protein interaction study results in the mapping and functional elucidation of an entire

metabolic or signaling pathway, with all its constituent proteins and their interconnections.

The types of information that proteomics studies aim to produce include:

1. Protein structure and function

2. Protein expression levels

3. Posttranslational modiﬁ cations

4. Subcellular localization

5. Protein–protein interactions

6. Protein–nucleotide interactions (e.g., to identify and study transcription regulating

proteins, known as transcription factors)

7. Protein–lipid interactions (e.g., for membrane-associated proteins, very important in

signal transduction pathways)

In this section we look at some of the key technologies and methods used in proteomics

studies to achieve these goals.

8.6.1 Functional Areas of Proteomics

Proteomics activities normally fall in one of several functional categories. We outline these

ﬁ rst, then look at each in detail. Note that for each area there are low-scale methods that

are typically used in individual protein studies, and high-throughput methods that are more

applicable for proteome-scale investigations. We outline these functional areas ﬁ rst, then

look in detail at the methods used in each area.

1. Fractionation and puriﬁ cation (separation, isolation). Normally, protein studies be-

gin with a cell lysate, the result of tissue fractionation, or a biological ﬂ uid (e.g.,

blood, urine). The proteins of interest must be isolated and separated from the mix-

ture, which may contain many other proteins as well as metabolites, lipids, and so

on.

2. Identiﬁ cation (primary sequence). The proteins that have been separated out must

now be identiﬁ ed. This involves determining the primary amino acid sequence of the

proteins or matching peptide information to known proteins (ﬁ ngerprinting).

3. Quantitation. Frequently, it is important to know how much of a protein has been

detected. In particular, the relative quantities of various proteins in a cell or tissue

yield important information.

4. Characterization. Now the proteins can be analyzed and studied. Key areas of study

include:

a. Sequence homologies

b. Posttranslational modiﬁ cations (PTMs)

c. Functional analysis, including interactions, pathways, and networks

d. Structural analysis and structure–function relationships

8.6.2 Fractionation and Puriﬁ cation

There are many ingenious ways that have been developed to separate and purify proteins.

Low-scale techniques include gel- and column-based fractionation methods, as well as

differential centrifugation or centrifugation through density gradients. In all cases, the pro-

teins are separated into bulk fractions or bands based on a single global property such as

size or charge. More advanced methodologies allow the fractionation and puriﬁ cation of an

individual protein from a protein mixture, through methods such as immunoprecipitation in

PROTEOMICS 239

240 PROTEOMICS AND DRUG DISCOVERY

which an antibody to a particular protein is used to separate that protein from all others, or

afﬁ nity chromatography, in which one or a set of related proteins are captured on an immo-

bilized ligand. These methods, although practical from an analytical point of view, do little

to assist in understanding complex biological pathways and essential protein interactions

with other proteins or other components of the cell.

The two most common separation methods for high-throughput applications described

next.

Two-Dimensional Polyacrylamide Gel Electrophoresis. (2D-PAGE) Standard (one-

dimensional) gel electrophoresis separates proteins based on mass alone. The sample is

loaded onto a gel slab and an electric current is passed through the gel. The current pulls the

proteins down the gel at a rate proportional to their mass, resulting in a series of bands that

represent proteins of different mass. However, separation by mass alone is often not sufﬁ cient.

2D-PAGE separates along the ﬁ rst dimension of the gel by isoelectric point, or pI (the pH at

which the protein’s net charge is zero), then along the second dimension by mass, as above.

The resulting two-dimensional arrangement gives better separation and results in a pattern

of spots, each representing a protein (or proteins) of speciﬁ c mass and pI. It is an effective

and widely used technique. However, there are some inherent problems with the technology:

2D-PAGE results do not replicate well; it does not detect proteins at low concentrations; it

does not work well for hydrophobic (e.g., membrane-associated) proteins; and it is possible

for several proteins to inhabit one spot on the gel, as they may have similar mass and pI.

High-Performance Liquid Chromatography (HPLC) HPLC uses a tube (known as a

column) ﬁ lled with some form of treated packing material (matrix or beads) that separates

proteins by selectively capturing and eluting them. A detector at the end of the column

identiﬁ es the proteins as they elute. There are several types of HPLC, using different

techniques to separate proteins based on the characteristic desired. These include gel

ﬁ ltration, ion exchange, afﬁ nity, and hydrophobic interaction. Each of these techniques

captures the protein of interest and holds it in the column while the rest of the sample

solution washes through. Then a different solution, which dissociates the protein of interest

from the column matrix, is applied and washed through the column. This elutes the protein,

carrying it past the detector, which typically registers it as a characteristic peak on a time

plot, allowing it to be identiﬁ ed. Many types of detectors are used with HPLC, including

refractive index, ultraviolet, ﬂ uorescent, radiochemical, electrochemical, near-infrared,

mass spectrometry, nuclear magnetic resonance, and light scattering. HPLC has some

advantages over 2D-PAGE in high-throughput environments: It is much faster, the results

are more repeatable, and it is more sensitive to low-concentration proteins. Also, it can be

used to separate and identify proteins by many more attributes than just mass and pI. For

example, proteins can be separated by ionization, by speciﬁ c afﬁ nity to an antibody or

ligand, or by hydrophobicity.

8.6.3 Identiﬁ cation

Once the proteins are separated, they must be identiﬁ ed. A standard method is to use the

output of 2D-PAGE directly: using visualization tools, the gel can be “read” in such a way

as to determine the mass and pI of the dark protein spots. If the protein of interest is not

novel, its mass and pI can then be looked up in a database and compared to known proteins

to ﬁ nd a match. This is often sufﬁ cient, but there are problems associated with this method.

For example, posttranslational modiﬁ cations can change a protein’s mass, thus shifting the

gel spot such that it may be mistaken for a different protein. Also, 2D-PAGE by itself can-

not identify an unknown protein. For this, additional steps are needed: The protein’s amino

acid sequence must be determined.

Edman Sequencing There are two main methods for identifying amino acid sequence.

Edman sequencing is a well-established method that degrades the protein one residue at a

time and determines each residue. It is effective, especially for de novo sequencing of an

unknown protein. However, it is relatively slow. Mass spectrometry (MS) is becoming the

preferred method because it can be applied in high-throughput environments.

Mass Spectrometry Analysis Mass spectrometry has really revolutionized the study of

proteins, due largely to its ability to identify proteins quickly and accurately and provide

sequence information. It has become an important and ubiquitous tool in proteomic studies.

There are many variants of the technology, but the basic elements are as follows, from input

to output:

1. Sample inlet and ionization chamber. MS works by separating peptide ions accord-

ing to mass and charge. (Most whole proteins are too large to be analyzed intact by

MS.) Therefore, sample molecules must ﬁ rst be broken up and charged by ionization.

The two most common forms of ionization are electrospray ionization (ESI) (used

with a liquid sample such as HPLC output), and matrix-assisted laser desorbtion

ionization (MALDI), which is used to ionize proteins by laser bombardment of solid-

state or viscous sample preparations on a plate. In either case, the ionized fragments

are then injected into the mass analyzer section.

2. Mass analyzer. This section separates the ions by mass-to-charge ratio (m/z). Two of

the main methods to accomplish this are as follows:

a. Electromagnetic ﬁ eld separation. Apply an electromagnetic ﬁ eld in such a way as

to let through only those of a certain charge. A quadropole analyzer is an example

of this type. Other types let all ions through, but use a magnetic ﬁ eld to deﬂ ect

them according to their m/z ratio.

b. Time-of-ﬂ ight separation. In this case, the analyzer is essentially a long chamber.

Lighter ions traverse the chamber faster than do heavier ones. A time-of-ﬂ ight

(TOF) analyzer exempliﬁ es this type.

3. Detector. The detector registers the arrival of an ion and sends the information to a

recorder to be processed and graphed for analysis.

4. Recorder. The recorder processes the detector’s output, generating intensity versus

m/z ratio results in text and graphical form.

Tandem mass spectrometry (MS/MS) is an important type of MS. It uses two mass

analyzer sections, separated by a collision cell that takes peptide ions separated by the ﬁ rst

MS, breaks them down into smaller ions of all possible lengths from one amino acid up to

the full peptide length (these are called b and y ions), then uses the second MS section to

separate these so that the detector identiﬁ es the b and y ions of each length. The action of

this middle chamber is called collision-induced dissociation (CID).

To provide peptides for MS identiﬁ cation, separated proteins can be obtained by excis-

ing gel spots and digesting with trypsin to give peptides. However, in high-throughput

PROTEOMICS 241

242 PROTEOMICS AND DRUG DISCOVERY

situations, HPLC is often used to fractionate the proteins, with the output of the column

directly feeding the MS ionizer input. The separated peptides from the column are then

ionized by ESI and run through MS/MS to be identiﬁ ed in a continuous ﬂ ow. The output of

the MS is a spectrum of mass/charge (m/z) ratios, showing intensity peaks corresponding

to all the detected b and y peptide fragments. Due to the CID that occurs between the two

MS stages, the fragments correspond to cleavages at every bond between the amino acids.

By analyzing the differences between one fragment’s m/z and the next, the sequence can

be deduced: each difference is the m/z of one amino acid in the sequence. However, there

can be ambiguity, since more than one amino acid may share the same m/z. So the resulting

sequence data for several peptides in the protein are used to ﬁ nd matches against a database

of existing sequences, and the best match identiﬁ es the protein.

Peptide mass ﬁ ngerprinting is a simpler and often sufﬁ ciently effective method of pro-

tein identiﬁ cation. MALDI/TOF is typically used for this. The protein is digested into

peptides; these are ionized and fed into a single MS stage (without further fragmentation);

and the resulting m/z spectrum identiﬁ es the m/z value of each peptide. This is the ﬁ nger-

print: this set of m/z values are then used to search a database of known peptide m/z values,

and the best match across all peptides identiﬁ es the protein.

8.6.4 Quantitation

To fully understand the relevance of a given protein to a disease state, quantitation of the

protein under different conditions is essential. For an isolated protein, many methods are

used. Ultraviolet (UV) absorption at a wavelength of 280 is a very rapid, simple, and ef-

fective method of measuring a protein concentration in solution, but this is accurate only

if the extinction coefﬁ cient for this actual protein is known. More often, one does not have

a single protein or does not know the extinction coefﬁ cient, as determining this is an ex-

pensive and time-consuming process that is usually completed only when the protein has

already been shown to be important or is being used as a biopharmaceutical.

Gel-based methods described above for separation and identiﬁ cation of proteins can

also be used to quantitate numerous proteins that are within a given sample, provided that

they separate well from each other on the gel. To use gels for quantitation requires staining

the gel with agents that recognize and bind to the proteins in some predictable manner and

requires that several lanes be devoted to standards of known quantity for comparison. The

accuracy and sensitivity of gel-based methods depends on the ability of the agent to bind

indiscriminately to proteins of all amino acid compositions, and to do so efﬁ ciently within

the gel matrix. Coomassie blue staining is often used and can be quantitative. Silver stain-

ing is more sensitive and will detect minor contaminant or other bands if they are present,

but it is less quantitative. Recent advances in developing new reagents have made various

ﬂ uorescent or light-emitting dyes available. Companies such as Amersham, Invitrogen, and

others all market products that are described as accurate and sensitive at monitoring exact

protein levels in a gel-or solution-based assay.

Mass spectrometry is generally not considered quantitative but is used to measure the

components that are present in a mixture or in a protein. The relative peak heights of an

MS measurement are not at all correlated with levels of the protein or peptides in solution.

However, MS can be used for relative quantitation of a protein under two different condi-

tions. Here, the protein from one sample is labeled with a stable isotope, changing its mass

so that MS can detect and quantitate its presence relative to the unlabeled protein.

8.6.5 Characterization

Sometimes, identiﬁ cation and quantitation are end goals in themselves. Perhaps we run an

experiment to determine how much of protein X exists in tissue sample Y. Often, however,

these are preliminary steps, and the real goal is to characterize the protein(s) we ﬁ nd.

Before proceeding, however, a few words about protein sequence, structure, and function

are necessary to set the stage. A cardinal rule of proteins is that their structure determines

their function. The ﬁ nal three-dimensional conﬁ guration of a protein, especially any active

site or binding sites it may have, determines the protein’s biological or chemical function.

There are four categories of protein structure:

1. Primary structure refers to the sequence of amino acids that form the protein chain.

The sequence is critical; it determines the three-dimensional conﬁ guration, or fold,

of the protein.

2. Secondary structure indicates the three-dimensional structure of subsections of the

protein chain, which fall into a handful of characteristic conﬁ gurations, such as alpha

helices and beta sheets. These are the ﬁ rst structures to form as the protein is being

created.

3. Tertiary structure occurs next, when the alpha helices, beta sheets, and other second-

ary structures assemble into the globular protein structure.

4. Quaternary structure results if the individual protein chains further assemble into

dimers, trimers, and so on.

The protein’s structure may not be complete at the tertiary or quaternary stage, however.

Often, one or more posttranslational modiﬁ cations (PTMs) change the structure further.

PTMs are important players that inﬂ uence protein stability, function, localization, and

interactions with other proteins, lipids, ligands, and so on. Some of the most important

PTMs are described in Table 8.1.

PROTEOMICS 243

PTM Description

Phosphorylation Addition or removal of a phosphate group. Quickly modiﬁ es the three-

dimensional structure of the protein or of its active site. Acts like a

switch: for example, activating/deactivating an enzyme.

Proteolysis Cleavage of proteins or peptides into smaller pieces. This cleaves off a

precursor segment to produce the mature protein. Also used to destroy

proteins marked for destruction (see Ubiquitination).

Glycosylation Addition of carbohydrate chains to proteins. This occurs while the

protein is still unfolded and emerging from the ribosome. This affects

many functions, including targeting a protein to different tissues.

Cystein oxidation Formation of disulﬁ de bonds. Stabilizes the protein; ensures correct

folding; enables certain protein–protein binding.

Ubiquitination Attachment of a small ubiquitin protein to a target protein, thus marking

the target for destruction.

TABLE 8.1 Signiﬁ cant PTMs

244 PROTEOMICS AND DRUG DISCOVERY

There are two more terms related to structure and function that should be deﬁ ned, since

they are a core part of protein characterization.

1. Domain. A protein domain is a region of the protein that is responsible for a par-

ticular function. An example is the SH2 domain, which speciﬁ cally binds to another

protein’s phosphorylated tyrosine area.

2. Motif. A motif is the amino acid sequence for a functional domain. It is conserved

across a set of related proteins that share the function.

Earlier in this section we summarized four major areas of protein characterization. With

this structure–function background, we can now elaborate on each.

Sequence Homologies Let us suppose that we have separated and identiﬁ ed a new human

protein whose function is unknown. Once the protein’s amino acid sequence is determined,

it is possible to search among other known protein sequences in the sequence databases

for homologs (sequences that share an evolutionary relationship). (In Section 8.8 and the

Appendix, sequence analysis and domain/motif databases and software, such as BLAST,

are discussed in more detail.) In sequence analysis, we look for signiﬁ cant similarities

between regions of our new protein and those of other proteins. A region that appears with a

high degree of similarity across multiple proteins identiﬁ es a motif. Given this motif, we can

then check databases to determine if the motif has been identiﬁ ed as a functional domain.

By using this kind of knowledge of the homologous proteins, we can do the following with

our new protein:

• Identify its probable function

• Assign it to a known protein family

• Determine many of its structural features (e.g., α-helixes, β-sheets).

Posttranslational Modiﬁ cations Characterizing proteins with PTMs essentially means

determining the presence or absence of the particular PTM. Several ingenious ways have

been devised to accomplish this. We discuss three ways of using afﬁ nity puriﬁ cation

followed by MS analysis to identify and quantify proteins in the presence of PTMs. In

each case, the ﬁ rst ligand of the pair is the capture ligand: It is immobilized by covalent

bonding to a matrix in order to bind with speciﬁ city to the substance of interest, which is

the second of the pair.

1. Lectin/carbohydrate, for detecting glycosylation. Lectin binds the sugar part of a gly-

coprotein directly, making lectin an effective means of directly separating out glyco-

sylated proteins. In this case, the separation is fully complete before the MS analysis.

2. Antibody/epitope, for detecting any PTM. Here, an immobilized antibody is used to

bind with high speciﬁ city to an epitope-tagged protein of interest. Then the protein is

analyzed with MS to detect whether PTMs have occurred, since all PTMs alter either

the mass or the charge of a protein, or both. For instance, a sample might have a mix

of proteins, including some unknown ratio of phosphorylated to unphosphorylated

protein X. If all of protein X is epitope tagged, the antibody can separate it all out

and send it to MS. MS can then quantify how much of X is phosphorylated and how

much is not based on the m/z difference.

3. Receptor/ligand tag, for detecting any PTM. An example of this is the use of a biotin-

tagged protein of interest, which is captured by avidin (which has a high afﬁ nity for

biotin) immobilized on the matrix. When through, only the bound biotin-tagged pro-

teins remain. To detect PTMs on the protein, MS analysis is required to differentially

quantitate the tagged protein with and without the PTM, as in the antibody/epitope

discussion above.

Functional Analysis, Including Interactions, Pathways, and Networks The basis of

functional analysis is the study of protein interactions: interactions with ligands, nucleotides,

small molecules, lipids, and especially, with other proteins. All are important, both for

basic science and in drug discovery. (A case study of proteome-wide protein–ligand study

leading to a successful drug discovery appears in Section 8.9.7.). In the present section we

concentrate on protein–protein interactions (PPIs).

Protein–protein interactions provide substantial insight into the functions of a protein.

Proteins frequently bind together into complexes, which carry out certain functions only

when combined in this way. Examples of such complexes are commonly found in signal

transduction pathways and gene transcription complexes. Therefore, knowing the other

proteins with which a new protein associates can help assign a putative function to the new

protein and to identify a possible role in a network or pathway, such as:

• Signal transduction

• Cell cycle regulation

• Transcription control

• Immunology

• Metabolism

Much study has gone into associating certain diseases with abnormal behavior of particular

pathways. Therefore, if a new protein can be identiﬁ ed with a pathway, it could be impli-

cated in a disease (“guilt by association”) and could then be considered a possible new

drug target.

High-throughput methods have recently been applied to the study of protein–protein

interactions at the proteome-wide scale. Some problems remain to be addressed, but this

area holds considerable promise. There are three main techniques: yeast two-hybrid,

phage display, and protein chips. A fourth, TAP, is normally a low-scale method, but it

has recently been used successfully in a proteome-wide study (Gavin et al., 2002). The

details:

1. Yeast two-hybrid (Y2H) uses two hybrids expressed in yeast to detect binding. Each

is encoded by its own plasmid inserted into the yeast cell. The two hybrids are:

a. Bait: a DNA binding domain fused to a protein of interest, which is to be tested

for binding against a large number of….

b. Prey: transcription activation domains fused to transcribed and translated open

reading frames (ORFs). ORFs are putative genes, each one read off the genome

from start codon to stop codon.

This tests one prey ORF at a time per cell culture, and is repeated for each ORF per

culture against the one bait. Once expressed, the two hybrids will bond if the bait and

PROTEOMICS 245

246 PROTEOMICS AND DRUG DISCOVERY

prey proteins interact. This brings the activation domain close to the DNA binding

domain, triggering the activation of RNA polymerase and thus transcribing a reporter

gene such as His3. If the reporter protein is detected (say, by growth/no growth), a

positive interaction is reported (Ito et al., 2001; Uetz et al., 2000).

2. In phage display, a fusion gene sequence is inserted into a bacteriophage and al-

lowed to express. This contains a sequence that codes a peptide for display, and this

is usually fused to the gene for the phage surface tip protein, pIII. When expressed

and translated, the phage displays the peptide on its surface. This peptide is then used

as a recognition site to attract the binding sites of a number of molecules that being

tested for interactions with the protein that would normally contain this displayed

peptide (Danner et al., 2001).

3. Protein chips are microarrays, but instead of using DNA on the spots, proteins (or

protein-binding molecules such as antibodies) are applied to the spots. This technol-

ogy is dealt with separately in the next section.

4. Tandem afﬁ nity puriﬁ cation (TAP) uses a single-compound epitope tag of calmodu-

lin binding peptide with protein A. This tag is used through two afﬁ nity separations

in sequence. The technique separates complexes of associated proteins of interest

from background proteins. At the end, the interacting proteins in the resulting com-

plex can then be detected and analyzed (Gavin et al., 2002).

Structural Analysis and Structure–Function Relationships A thorough understanding

of a protein’s structure is one of the key goals of protein study. It is of fundamental

importance in rational, or structure-based, drug discovery. As noted, amino acid sequence

dictates structure, and structure dictates function. This three-way relationship cannot be

separated. Having looked at sequence and function, let us now turn to structural analysis.

There are two broad areas for determining protein structure. The ﬁ rst involves laboratory

techniques, and the second uses computational techniques.

Laboratory Structure Determination Laboratory determination of protein structure is

still considered the gold standard. For drug discovery in particular, current computational

approaches are not capable of generating a precise enough model of the protein to enable

optimized structure-based drug design. There are several laboratory approaches for

accurately determining protein structure. The two key methods are x-ray crystallography,

which uses x-ray diffraction to resolve the structure of crystallized proteins, and nuclear

magnetic resonance imaging (NMR), which uses nuclear spin transitions to resolve the

structure of proteins in solution. Both are capable of providing images down to a very

ﬁ ne resolution, and each has advantages and disadvantages. With x-ray crystallography,

proteins are often crystallized with a ligand bound in the active site. This helps determine

the physical conﬁ guration and energetics of the binding pocket, which are critical factors

in structure-based drug design.

Laboratory methods are not without their drawbacks. For instance, a protein is not a

static structure; there can be signiﬁ cant motion in some of its parts. (This is particularly

true for an allosteric protein, one that changes shape based on whether or not a ligand is

bound to its active site) So x-ray crystallography can only give a snapshot of a protein and

may miss other important conﬁ gurations. But this is the current state of the art, and despite

the drawbacks it is still possible to obtain a tremendous amount of highly useful structural

information with these methods.

Computational Structure Determination Although not as accurate as the laboratory

techniques, computational approaches provide signiﬁ cant value. There are three main

classes of computational approaches that attempt to predict protein structure based on

amino acid sequence:

1. Homology modeling. This approach starts with a template protein of known struc-

ture whose sequence is homologous with the new protein of unknown structure. A

model is built by “threading” the novel protein’s sequence along the scaffold of the

known protein. The threading is fairly complex: The growing model must be checked

repeatedly for energy minimization as the sequences are aligned, residues are substi-

tuted, and loops are constructed. The conﬁ guration with globally minimized energy

is assumed to be the most likely model. This method has too many uncertainties to

generate high conﬁ dence, but it can help provide valuable probable structural infor-

mation (Kopp et al., 2004).

2. Ab initio fold prediction. The goal of solving a protein’s three-dimensional structure

computationally from sequence data alone has been a holy grail for some time. This

is partly due to the difﬁ culties of solving structures using laboratory techniques.

Considerable research has gone into this area. The annual competitions, compar-

ing fold predictions with known (but held secret) crystallographic results, show

that ab initio prediction algorithms are continuing to improve. The algorithms are

too complicated to go into here. However, there is one, known as Rosetta (see the

Appendix), which has scored much better than competitors in the competitions and

has come close with several new proteins (Bonneau et al., 2001; Rohl et al., 2002).

However, the current view is that even the best ab initio folding algorithms do not

yet generate a good enough model to be used with high conﬁ dence in structure-

based drug discovery.

3. Secondary structure prediction. An example will illustrate the use of computa-

tional secondary structure determination and the practical importance of the re-

sults. Suppose that you have the sequence of a novel protein of unknown structure

and function. You note that several regions of the protein are a mixture of hydro-

phobic and hydrophilic residues. Looking at the sequence alone may not tell you

whether these regions are likely to result in overall hydrophobic or hydrophilic

areas, or neither. However, you next determine that the sequences of these regions

are characteristic of alpha helices (using available structure prediction software

tools such as PredictProtein or ProtScale; see the Appendix). Once these second-

ary structures are folded correctly, it turns out that they present all their alternating

hydrophobic residues outward and therefore create an overall hydrophobic region.

These structures now tell you that these regions are likely to belong to a mem-

brane-associated protein. If you ﬁ nd further that you have seven such alpha heli-

ces, each about 20 amino acids long and separated by beta turns, you may deduce

important functional information from this: You have almost certainly found a new

G-protein-coupled receptor (GPCR). GPCRs are by far the most common target

for drugs (Drews et al.; 2000), so with further functional analysis, you may have

discovered a new drug target.

It must be stressed that computational approaches are not foolproof. They are somewhat

probabilistic, indicating for instance that a particular string of amino acids is likely to fold

into an alpha helix or a beta sheet. They are usually based on a good deal of experience, but

PROTEOMICS 247

248 PROTEOMICS AND DRUG DISCOVERY

there are always exceptions to the derived rules. So laboratory techniques should always

be used to verify the computational ﬁ ndings. As noted, an experimentally derived three-

dimensional structure (usually via x-ray crystallography) is considered the most authorita-

tive given the current state of the art.

8.7 PROTEIN CHIP TECHNOLOGY

Protein microarrays are a new technology for highly parallel high-throughput sensitive

analysis of proteins across different conditions and time frames. It is the ﬁ rst technology

that is showing the potential for simultaneous very high scale (up to full proteome) analysis

of protein functions, interactions, and comparative expression. Applications include pro-

tein identiﬁ cation, quantitation, and functional analysis, all of which can be applied to drug

discovery, biomarker discovery, and pathway building. The protein microarray concept

grew out of the DNA microarray, which has proven such a valuable tool in genomics stud-

ies, such as comparative gene expression studies. Protein array technology is considerably

more challenging than its DNA precursor, for reasons we elaborate on below. Therefore, it

not only shows great promise but also has signiﬁ cant hurdles to surmount. However, it is

such a potentially important tool for proteomics and for drug discovery that it warrants its

own section.

We review the issues that the technology aims to address: the current state of the tech-

nology, the underlying principles, different implementations, and the advantages and draw-

backs of the technology as it stands today.

There are two types of protein arrays:

1. Functional protein microarrays (also known as interaction arrays). These are arrays

of proteins or peptides immobilized to spots on the array substrate. The arrays are

then exposed to solutions of substances of interest (usually, label tagged), such as

other proteins, ligands, small molecules, nucleotides, or phospholipids to see if they

bind to the arrayed proteins. Protein spots where binding has occurred are generally

detected by the tag (usually ﬂ uorescent). The arrayed proteins may also be enzymes,

for studying highly parallel enzyme–substrate interactions.

2. Protein capture microarrays. These are usually arrays of ﬁ xed capture agents such as

antibodies or other compounds: for example, aptamers, which, like antibodies, can

bind with speciﬁ city to a protein target. The arrayed capture agents then capture and

bind proteins of interest that are applied to the ﬁ nished array. As with the functional

protein arrays, proteins of interest to be applied to the array are normally tagged, and

detected if bound. Alternatively, a sandwich method can be used [basically, a high-

scale enzyme-linked immunosorbent assay (ELISA)], where a second antibody is

label tagged and applied after the proteins of interest have been applied to the array

of immobilized ﬁ rst antibodies. Where the second antibodies bind to the proteins

captured by the ﬁ rst antibodies, the labels are detected. Capture arrays are especially

useful for diagnostics and comparative expression studies.

8.7.1 Issues Addressed

The chief aim of protein microarray technology is to enable highly parallel, simultaneous,

high-scale, high-throughput screening and analysis of proteins: identiﬁ cation, quantitation,

interactions, and function. The technology also enables comparative protein studies under

varying conditions and as a function of time. Further, the technology provides high selec-

tivity among the arrayed proteins, and high sensitivity to proteins at very small expres-

sion levels. These capabilities address limitations in other technologies (e.g., 2D-PAGE,

immunoassays, mass spectrometry, ICAT, Y2H): in particular, lack of wide dynamic range

and sensitivity to low expression levels and lack of simultaneous high-scale parallelism

(Hanash, 2003; Lopez and Pluskal, 2003; Phizicky et al., 2003; Taussig et al., 2002). In

addition, protein microarrays can be used with hydrophobic (i.e., membrane-associated)

proteins (Lopez and Pluskal, 2003). Most other existing technologies have difﬁ culty ana-

lyzing hydrophobic proteins. Capture arrays can in addition give direct protein expression-

level information and can be used to detect different posttranslational modiﬁ cations of a

protein.

8.7.2 Current State of the Technology

Protein microarray technology is not yet considered a mature technology. According to

Gavin MacBeath at Harvard (who is himself a core researcher in protein microarrays),

most protein microarray work that has been reported to date has been proof-of-concept:

“We’re at the point now where people need to start applying this technology to real ques-

tions and use it to learn something about biology” (Gershon, 2003). A seminal 2001 study

by Zhu et al. (2001) falls into that category, as does their previous study on protein kinases

(Zhu et al., 2000) and MacBeath and Schreiber’s (2000) own important study.

However, signs of maturation are beginning to appear. Some papers are showing inter-

esting practical applications with potential clinical signiﬁ cance, and commercial develop-

ments are proliferating. One recent paper by W. Robinson et al. detailed a study using

arrayed antigens to help guide tolerance-inducing therapy against autoimmune diseases,

in this case a mouse model for multiple sclerosis (Robinson et al., 2003). The therapy was

a “cocktail” vaccine of DNA, each component of which was aimed at one of numerous

antibody responses that were detected by the array assay. The authors were able to show a

signiﬁ cant (45%) reduction in autoimmune response, thus showing promise for application

in humans with MS.

Commercially, there are a number of suppliers with chips or assay services on the mar-

ket (or announced). Of particular interest in the functional array area is Michael Snyder’s

new company, ProtoMetrix (recently purchased by Invitrogen), which aims to commercial-

ize and sell a nearly full-proteome yeast proteome chip reported on in a 2001 study (Zhu

et al., 2001). Another interesting venture is Sweden’s Biacore, which markets a variety of

protein arrays based on a different implementation, microﬂ uidics, and uses laser-based

surface plasmon resonance to detect interactions without labels. Most of their stock sensor

chips are protein capture chips, but they also make a custom chip that has powerful func-

tional protein array capabilities.

There is a general lack of industry acceptance of the technology so far. It is not yet

widely perceived as a trustworthy mainstream tool for drug discovery. For now, many pro-

tein chip companies are going with a service model, where they will do the assays for the

drug companies and biotechs and return the data (Uehling, 2003). However, there is a gen-

eral sense that this is probably a true breakthrough technology that when somewhat more

mature, could prove to be of major importance in drug discovery, biomarker discovery, and

pathway building. We will now look at the details of the two major types: functional arrays

and capture arrays.

PROTEIN CHIP TECHNOLOGY 249

250 PROTEOMICS AND DRUG DISCOVERY

Functional Protein Arrays

Underlying Principles Functional protein microarrays are based on essentially the same

principle as DNA microarrays: Bind a large number of whole or partial proteins (rather

than DNA strands) to a surface in an array (typically, high-density) of spots, wells, and so

on, and then probe the entire array with solutions containing one or more substances of

interest. The results are measured and analyzed for identiﬁ cation and strength of binding.

Heng Zhu of Michael Snyder’s lab has provided a good summary: “In principle, the

biochemical activities of proteins can be systematically probed by producing proteins in a

high-throughput fashion and analyzing the functions of hundreds or thousands of protein

samples in parallel using protein microarrays” (Zhu et al., 2001). Although there are a

number of different ways of implementing these (see below), the basic principle remains the

same: Probe an array of immobilized proteins or peptides with solutions of other proteins,

small molecules, ligands, lipids, and so on, and screen for the resulting interactions.

Implementations A number of ways of implementing functional protein microarrays have

been devised. First, a large number of proteins must be created, ampliﬁ ed, and puriﬁ ed.

All protein chip implementations share this problem: how to successfully create the large

number of proteins to be arrayed. In some high-scale cases, a recombinant technique

similar to that described in Zhu et al. (2001) is used (Lopez and Pluskal, 2003). This

process involves the following: Create constructs of open reading frames (ORFs), usually

along with a fusion tag; amplify these with PCR and clone into expression vectors (e.g.,

plasmid); insert the vector into cells for expression (e.g., yeast, bacteria, or other); then

extract and purify the resulting proteins. Another interesting method used in generating

high-scale peptide microarrays involves direct synthesis of the peptides in situ directly on

the chip using digital photolithography to achieve programmable light activation of small

local chemical reactions on each spot (Pellois et al., 2002).

Next comes the implementation options of the arrays themselves. The proteins are im-

mobilized onto a surface, often using the same type of robotic application systems as that

used to spot DNA microarrays. The array surface can be glass (usually coated or deriva-

tized), silicon, nitrocellulose, or PVDF membranes, MS plates, microtiter or other well

technology, or even microbeads or other particles (Taussig et al., 2002). Zhu et al. (2001)

used nickel-coated glass slides, while MacBeath and Schreiber (2001) derivatized their

glass slides chemically with a cross-linking agent that reacted with primary amines on the

proteins they applied. An earlier Zhu et al. study (2000) of yeast protein kinases using pro-

tein chips used arrays of microwells in silicone elastomer sheets which sat atop glass slides.

Solutions containing the substances of interest (proteins, ligands, lipids, small molecules,

etc.) are then applied to the chips and incubated to achieve binding.

The chips are then scanned to detect binding patterns and strengths. This detection is

done in several ways. Typically, the probe molecules are labeled with ﬂ uorescent or radio-

active labels so that when they bind, their signal can be read by a scanner. (An interesting

non labeled technique is the laser-based surface plasmon resonance used by Biacore, dis-

cussed previously.) The scanner reads and records the intensity of the signal at each spot on

the array, and these data are then analyzed in software.

Advantages Functional protein microarray technology holds considerable promise for

high-scale highly parallel high-throughput assays for investigating a number of things:

protein interactions with proteins, phospholipids, ligands, small molecules, nucleotides,

and so on. This could become a powerful tool for drug discovery. Among its advantages

over other technologies:

• Sensitivity over a wider dynamic range: in particular, for proteins at very low expres-

sion levels.

• Ideal for proteome-wide protein expression studies.

• Ability to screen large numbers of interactions at one time: high parallelism, high

throughput.

• Able to accommodate hydrophobic (e.g. membrane-associated) proteins.

• Potential for enabling simultaneous small molecule screening over a wide range

of human proteins. This could identify both on-target efﬁ cacy and off-target cross-

reactivity (indicating negative side effects or toxicity).

• Can make use of some existing DNA array scanning technology and software.

Disadvantages

• Generating proteins from ORFs has disadvantages. Although it is an efﬁ cient way to

generate a large library of expressed products, there is no guarantee that the products

are in fact valid proteins, since only one of six reading frames is valid for any one

gene. Also, the ORF-based approach usually generates only one isomer, and may

miss other splice variants.

• In some high-scale cases, especially when using bacteria to express the proteins, the

protein extraction process may denature the proteins, so they may not be properly

folded and thus be nonfunctional. [However, even screening with denatured pro-

teins has been shown to be useful for certain applications (Lopez and Pluskal, 2003;

Taussig et al., 2002).]

• Problems can arise in using fusion tags for protein puriﬁ cation step and/or for cova-

lently bonding the tagged proteins to the slide or other substrate. The tags may inhibit

proper folding of the protein and may interfere with PTMs.

• Difﬁ culty in generating thousands of veriﬁ ably valid, functional, full-length proteins.

The high-throughput approach may miss some slow-developing proteins. Active sites

may be blocked or facing down.

Protein Capture Arrays

Underlying Principles Capture arrays use the same basic principles as do established

low-throughput immunoassay technologies. It is essentially afﬁ nity capture, like that used

in fractionation and puriﬁ cation. A capture agent such as an antibody or aptamer with high

speciﬁ city for a particular protein is immobilized on the array and then used to attract and

bind that protein.

Implementations For protein capture arrays, the capture agents must ﬁ rst be developed,

and in suitable quantity. Several types are used for arrays. Two of the most commonly

used are antibodies and aptamers. Antibodies are glycoproteins which are normally used

by the immunological system to selectively detect recognition sites on foreign invader

proteins. Their natural selectivity, or afﬁ nity for a speciﬁ c protein makes them a good

PROTEIN CHIP TECHNOLOGY 251

252 PROTEOMICS AND DRUG DISCOVERY

choice as capture agents. Aptamers are synthetic ligands composed of DNA or RNA that

are designed to bind with speciﬁ city to a particular protein. They are short oligonucleotides

that fold into predictable three-dimensional conﬁ gurations which complement the binding

sites of the target proteins.

Antibodies can be made by conventional immunization techniques (monoclonal or poly-

clonal). Another common technique is to create an antibody library using phage display,

then selecting from the library and using recombinant expression, generally using E. coli

(Taussig et al., 2002). Aptamers can be created as large libraries of oligonucleotides, then

selected using the SELEX procedure, which is amenable to automation for high-throughput

applications (Mayer et al., 2003; Taussig et al., 2003). In both cases, speciﬁ c antibodies or

aptamers are selected using high-throughput capture methods that apply the protein of in-

terest, then select out only those capture agents that have bound to that protein. These are

then further puriﬁ ed and ampliﬁ ed.

The implementation options of the arrays themselves are similar to those used for

functional protein arrays. Being proteins themselves, antibodies are amenable to many

of the same surfaces and immobilization procedures. Elaborate methods using covalent

cross-linking or tag afﬁ nity (e.g., biotin/streptavidin) are sometimes used to try to orient

the immobilized proteins such that their binding sites face away from the surface and are

therefore more available for analyze binding. Nucleic acid aptamers are simpler to deal

with. They are less susceptible to orientation issues, and they can use the same surface

technology (usually coated or derivatized glass slides) that is used in the more mature DNA

microarrays.

Detection of binding on capture arrays uses many of the same techniques as are used

for functional protein arrays. Fluorescent labeling of the analyze protein is most common.

However, in sandwich ELISA assays, the secondary antibody carries the label tag. For

instance, for two-color comparative protein expression studies, the secondary antibodies

are normally tagged with Cy3 and Cy5 labels. These provide red and green colors indicat-

ing the relative binding intensity of normal vs. disease analyses, for instance. Label-free

methods can also be used with capture arrays, such as the laser-based surface plasmon

resonance method discussed previously.

Advantages

• High sensitivity: protein capture arrays have demonstrated successful detection as

low as 10 fM (Bock et al., 2004).

• High selectivity and afﬁ nity if used in sandwich ELISA mode.

• Potential to enable high-scale protein expression proﬁ ling (e.g., diseased vs. normal

tissues) to help determine proteins that over- or underexpress in the disease state.

This can be a powerful drug target identiﬁ cation tool as well as a biomarker discovery

tool.

• Potentially effective as diagnostic chips, to detect disease by known biomarkers in a

clinical setting.

• Multiple biomarkers can be identiﬁ ed and assayed in parallel. Some diseases (e.g.,

cancers) are characterized by protein signatures, that is, changes in expression of

multiple proteins vs. normal tissue expression levels.

• Can make use of some existing DNA array technology, as with functional protein

arrays.

Disadvantages

• Cannot be used to discover unknown proteins. By deﬁ nition, you must know what

proteins you are trying to capture, then design a capture agent for those proteins.

• Scale is currently limited: It is very challenging to create thousands of unique capture

agent types. Most current capture array implementations are limited to tens or at most

hundreds of unique spots.

• It can be difﬁ cult to immobilize antibodies to the array while maintaining their bio-

logical activity and structure, as with functional protein arrays.

• Antibodies do not always bind with high selectivity and speciﬁ city, particularly if non

target proteins are in high abundance. (But sandwich ELISA methods can improve

speciﬁ city signiﬁ cantly, as noted above.)

Summary Protein microarrays have generated considerable excitement due to the potential

use of simultaneous high-scale high-throughput assays in quantifying and characterizing

all manner of protein interactions and for proteome-wide protein expression proﬁ ling. Of

particular interest is their potential value in the drug discovery and development process, in

biomarker discovery, and as clinical diagnostic tools. However, the technology is still quite

immature and has many hurdles to surmount before it is accepted by industry as a valid,

reliable tool. Protein chip companies may need to adopt a services model at ﬁ rst, until

they can repeatedly and reliably generate useful high-quality, data for the pharmaceutical

and biotech industries on a contractual basis. This should help the technology gain wider

acceptance.

8.8 PROTEOMICS DATA ANALYSIS: COMPUTATIONAL BIOLOGY

AND BIOINFORMATICS

The key to the success and utility of genomics and proteomics is the ability to deal with the

voluminous quantities of data they generate. Perhaps the most major difference between the

new high-scale approaches and standard biological methods is the shift from hypothesis-

driven to data-driven exploration. Hypothesis-driven methodology will always be a core

part of basic science and drug discovery. But high-scale approaches are complementing this

with data-driven methodology. The activities relating to managing and analyzing these data

can collectively be called informatics. We now look at the informatics side of proteomics.

Several functions are needed to deal with these data. They can be grouped as follows:

1. Data generation. Laboratory instruments used in proteomics have become extremely

sophisticated computer-controlled machines, including mass spectrometers, 2D-PAGE ma-

chines, robots for creating microarrays, and microarray scanners. These all require special-

ized software to control and monitor the process, generate and sometimes preprocess the

data, and generate reports on the results.

2. Storage, retrieval, and tracking. Data must be retrieved from laboratory machines and

stored in an appropriate format and conﬁ guration to promote access and mining for rela-

tionships among the data.

a. Storage. Large quantities of storage space must be allocated. Since multiple storage

servers are often required, it can be helpful to install a storage access network (SAN)

PROTEOMICS DATA ANALYSIS 253

254 PROTEOMICS AND DRUG DISCOVERY

to help with the speed and efﬁ ciency of data access. The SAN manages a cluster of

storage servers and takes over the task of deciding on which machine(s) to allocate

storage for a particular dataset. All of this must be well architected and managed, and

regular backups are imperative.

b. Database management. Considerable effort goes into the design, management, and

maintenance of databases for biological data. It helps for database managers in this

area to have a solid grounding in biology and biochemistry. This promotes database

designs that make it easier for biologists to ﬁ nd biologically meaningful associations

in their data mining. The databases should, if at all possible, “speak the same lan-

guage.” That is, for any one protein or gene, the same key (accession number) should

be acceptable across all databases. (Among public databases, for instance, this is not

always the case. Signiﬁ cant efforts are under way to try to coordinate the databases

to ﬁ x this problem.)

c. LIMS. Laboratory information management systems is a specialized area that com-

bines data storage and retrieval, database management, and data tracking for data-

intensive laboratory environments. These systems combine biological databases

with the means to organize and track the sources of these data and coordinate the

source information with the experimental data. Thus, all data resulting from a par-

ticular tissue sample on a particular date can be effectively tracked. LIMSs are

used primarily to do sample tracking and to aid in project management. In a drug

discovery environment, they can be especially useful in generating, maintaining,

and archiving the kinds of information required in ﬁ ling for FDA approval for a

drug candidate.

3. Computation infrastructure. High-scale genomics and proteomics methods often re-

quire massive computing power. In structural proteomics, for example, computational pro-

tein folding is an example of a particular compute-intensive area: given a primary amino

acid sequence of a protein, there are an enormous number of possible three-dimensional

conﬁ gurations. Protein folding algorithms must explore all the feasible folds, generate a

free-energy function for each, and try to ﬁ nd the conﬁ guration with the global minimum

free energy. This is the kind of process that can take a supercomputer to accomplish in

any reasonable time, and supercomputers have been used in these problems. However,

supercomputer use can be quite expensive and may be out of reach for some organiza-

tions. In recent years, cluster computing has come into its own for these and many other

proteomics computation requirements. Clusters of simple relatively inexpensive PC-type

hardware, often running Linux, can do distributed processing on complex problems after

they are appropriately broken down into smaller, parallel chunks that can be assigned to

the multiple individual processors. Cluster computing has proven extremely effective and

is considerably less expensive than the use of supercomputers. However, the latter still have

their place, especially when symmetric multiprocessing, using one large memory space, is

required.

4. Data checking and manipulation. These address the crucial area of data integrity and

validity. Before the data can be used in biological analysis, manipulation and checking

must usually be done to minimize sources of error and to help ensure the biological sig-

niﬁ cance of the data. Microarray gene expression data provide a good example. Normally,

the laboratory instruments generating the data come with some software to preprocess

the raw data they generate. This may involve normalization and error checking. However,

additional manipulation and checking is usually needed. For example, data from several

repeat experimental runs must be cross-checked and correlated; they may need to be nor-

malized across all runs so that the data are comparable. Data are subjected to statistical

analysis to:

a. Perform normalization.

b. Check for accuracy and precision.

c. Check for sufﬁ cient signal-to-noise ratio.

d. Calculate signiﬁ cance values.

5. Analysis: software and databases. Computational biologists and bioinformaticians cre-

ate and use sophisticated software data analysis and visualization tools. In addition to the

software tools, and tightly linked with them, well-curated databases have become ubiq-

uitous and essential in genomics and proteomics analysis. Refer to the Appendix for a

representative sample of some important publicly available databases and software tools.

In proteomics, areas addressed by software and databases include:

a. Protein identiﬁ cation.

• Analysis and visualization of 2D-PAGE data

• Analysis and visualization of mass spectrometry peptide spectra

• Detection of primary amino acid sequence from MS/MS spectra or Edman

sequencing

• Peptide ﬁ ngerprinting from MS output for protein identiﬁ cation

• Protein identiﬁ cation from 2D-PAGE spot localization or MS/MS peptide

sequencing

b. Protein quantitation

c. Characterization

• Sequence homologies, mapping to protein families

• Posttranslational modiﬁ cations

• Functional analysis (construction and visualization of signaling pathways, and

mapping and visualizing protein–protein interactions)

• Structural analysis(two- and three-dimensional)

• Structure–function correlation

The effectiveness of all this elaborate data generation, storage, handling, computing,

and analysis is, of course, dependent on one key prerequisite: proper experimental design!

Just as in standard biological methodology, it is imperative to consider carefully the in-

tended outcome of the experiment and then to design it in such a way as to maximize the

potential for that outcome and minimize the chances of spurious conclusions. One of the

dangers of data-driven high-scale approaches is that the data can appear so authoritative,

so deﬁ nitive, that new practitioners may be tempted to accept them at face value and to

assume their validity. Recent history has given ample evidence of this danger. Microarray

gene expression studies, especially in the early days of this technology, roughly the late

1990s, provide a good example. In that period, many exciting results were generated, and

papers published, based on single experimental runs with microarrays. Several of these

results could not be replicated later and therefore had to be considered suspect. It turns

out that there is often highly signiﬁ cant variance in the data from one experimental run

to another, and that the results of a single experiment therefore proved inadequate, and

in some cases, erroneous. This variance can be due to different tissue sample processing,

PROTEOMICS DATA ANALYSIS 255

256 PROTEOMICS AND DRUG DISCOVERY

different microarray processing, variations in quality and quantity of cDNA probes on the

microarray spots, varying background noise, and so on. Therefore, it is now generally ac-

cepted procedure to perform at least three separate microarray experiments with the same

probe sets. The resulting data can then be cross-correlated, normalized, and checked for

errors. Outliers on one array may be discarded. Therefore, the importance of proper experi-

mental design cannot be overstressed.

8.9 PROTEOMICS AND DRUG DISCOVERY

Earlier in this chapter, the current practices in drug discovery were outlined. Recall the is-

sues with the current drug discovery process:

• Very long time span between initial conception and ﬁ nal marketing of a drug: 10 to

15 years is not uncommon.

• Very high attrition rate: about 5000 compounds in for every one drug out.

• Exceedingly costly process: a major contributor to high costs when a drug candidate

fails late in the process (e.g., in phase III trials).

Therefore, it is desired to improve the process. For instance:

• More targets identiﬁ ed and validated, and faster

• Fail problematic drug candidates earlier and more cheaply

• Enable targeted therapeutics

Now that some background has been presented, we can consider how proteomics can be

applied to help address these issues. We look at proteomics applications in the drug discov-

ery process, both those in current use as well as potential. A number of new possibilities are

opening up, due to the increases of scale and speed that result from these new technologies.

For instance, there are new capabilities for:

• Systematic high-scale searches for new drug targets

• Identiﬁ cation of thousands of previously uncharacterized proteins through high-

throughput expression studies and functional proteomics

• Biomarker detection using high-scale differential protein expression studies of patho-

logical vs. normal samples

In this section we outline some of the key ways these capabilities can assist in each

stage of the drug discovery process. Reference is made to the methods and technologies

discussed earlier in the chapter. Some illustrative examples are provided along the way.

8.9.1 Target Identiﬁ cation

The initial stage of the drug discovery process is the identiﬁ cation of a molecular target,

almost always a protein, which can be shown to be associated with disease. To accelerate

this, more proteins need to be identiﬁ ed and characterized. Although the set of human

genes is by now quite well understood, this is not as true for all the proteins in the human

proteome. Because of splice variants and posttranslational modiﬁ cations, the number of

proteins far exceeds the number of genes. As of today, a great many human proteins have

yet to be fully identiﬁ ed and characterized. The new high-throughput technologies of pro-

teomics are helping to close this knowledge gap in a number of ways:

1. Identify large numbers of novel proteins in the search for new drug targets. High-

throughput separation, puriﬁ cation, and identiﬁ cation methods are being used for this. 2D-

PAGE is often used for separation, even though it has some drawbacks. HPLC is a common

separation and puriﬁ cation technique, and is particularly adapted for high-throughput use

when coupled directly to tandem MS for identiﬁ cation. These procedures produce peptide

sequences that must then be compared to possible homologs in other organisms (or paralogs

in humans) using informatics. Edman sequencing may be needed for sequence veriﬁ cation.

2. Perform initial functional characterization of novel proteins. Initially, assign them to

functional areas that might implicate some of them in disease pathology:

a. Metabolic pathways

b. Transcriptional regulatory networks

c. Cell cycle regulatory networks

d. Signal transduction pathways

Computational sequence homology analysis can help putatively assign a novel protein to a

known pathway or network that its homolog participates in. Sequence analysis also helps iden-

tify known motifs and functional domains that the new protein shares with known proteins.

This information can provisionally assign membership in a protein family to the new protein.

3. Screen a compound with known therapeutic effect against a large number of human

proteins to identify the exact target. It is not uncommon to have a compound with known

but perhaps limited or suboptimal therapeutic effect against a disease, yet not know the mo-

lecular target for this compound. A proteome-scale search can pinpoint a possible target,

once binding has been identiﬁ ed. The case study of the development of LAF389, presented

at the end of this section, describes a successful example of this in detail.

4. Perform high-throughput differential protein expression proﬁ ling, comparing diseased

with normal tissue samples, to identify the biomarker proteins that are possible contribu-

tors to the disease by their over- or underexpression. This is one of the most promising

areas of application of proteomics in target identiﬁ cation. Several proteomics technologies

can be used for expression proﬁ ling. Two-dimensional PAGE has been used, although com-

paring two different gels (one for normal, one for disease state) can be problematic, due to

the replicability issue discussed earlier. A variant of this called two-dimensional differen-

tial in-gel electrophoresis (2D-DIGE) overcomes this drawback. 2D-DIGE has been used

successfully to identify biomarkers that could be potential drug targets (Lee et al., 2003).

As noted previously, protein chips have perhaps the greatest potential for high-throughput

simultaneous differential protein expression proﬁ ling. Several successful attempts to use

protein arrays for target identiﬁ cation have been reported (Simon et al., 2003). However,

the technology is still relatively immature and is therefore further out on the horizon as far

as general industry acceptance for target identiﬁ cation.

Example A recent study showed how an integrated approach involving proteomics,

bioinformatics, and molecular imaging was used to identify and characterize disease-tissue-

speciﬁ c signature proteins displayed by endothelial cells in the organism (Oh et al., 2004).

PROTEOMICS AND DRUG DISCOVERY 257

258 PROTEOMICS AND DRUG DISCOVERY

Working with blood vessels in normal lungs and in lung tumors in rats, the researchers

used several high-throughput afﬁ nity-based separation procedures, followed by MS and

database analysis to identify and map the proteins displayed by endothelial cells that line the

blood vessels. With this approach they identiﬁ ed proteins that are displayed only on solid

tumor blood vessel walls. They then demonstrated that radioisotope-labeled antibodies can

recognize these tumor-speciﬁ c proteins, allowing them to be imaged. The radioisotope

labeling itself also resulted in signiﬁ cant remission of solid lung tumors, demonstrating

tissue-targeted therapeutic potential. Although individual endothelial proteins had been

identiﬁ ed and targeted in previous studies, this is the ﬁ rst time a proteomics approach had

been used to move toward a complete tissue-speciﬁ c mapping of the proteins displayed

on the blood-exposed surfaces of blood vessels. This demonstrates a new approach for

identifying potential novel targets for therapy.

8.9.2 Target Validation

At this point in the drug discovery process, one or more potentially disease-related pro-

tein targets have been identiﬁ ed. For now, let us assume one. The next step is to validate

the target. Primarily, this means that the target’s relevance to disease pathology must be

determined unambiguously. This involves more detailed functional characterization, more

evidence for the pathway or network assignment, and modulating the protein’s activity to

determine its relationship to the disease phenotype. Some determination of tractability may

be done in this stage as well. Proteomics can assist in several ways:

1. Determine in what tissues and cell components it appears and in which developmen-

tal stages. High-throughput techniques such as protein chips and 2D-DIGE can be used

for proteome-scale expression studies comparing different tissue types and developmental

stages. These add evidence that the putative target is found in a disease-related tissue and

at the expected developmental stage. Interaction studies can help determine subcellular

localization by showing binding to proteins or phospholipids in the cell that have known

location. Further, sequence analysis can identify known location-specifying signal peptides

on the protein. Finally, posttranslational modiﬁ cation analysis can identify certain PTMs

that determine the destination of the protein.

2. Understand when and for how long the target gets expressed, and when degraded.

High-throughput protein expression studies can be done in multiple runs over time, and

then compared. In this case, expression patterns are assayed and then compared across

multiple time points rather than normal vs. disease states. For additional evidence it would

be very informative to do a time-based expression study ﬁ rst of normal tissue, then another

one of diseased tissue, and compare the behavior over time of normal vs. disease tissues.

This would provide a multidimensional target validation.

3. Verify the target protein’s speciﬁ c role within the protein family and the pathway

or network identiﬁ ed in the previous stage. Initial putative functional assignments for

the new protein were made in the target identiﬁ cation stage. In target validation, high-

throughput protein–protein interaction studies can be used to strengthen the evidence for

the protein family, network, or pathway involvement. Protein–phospholipid interaction

assays can determine whether the new protein is membrane associated. Technologies

such as protein chips are beginning to be used for these interaction studies and have great

potential in this area (Walgren et al., 2004; Zhu et al., 2001). Other techniques such as

Y2H, phage display, and tandem afﬁ nity puriﬁ cation have also been used with success

in this area.

Posttranslational modiﬁ cations to the new protein can also be identiﬁ ed by methods

discussed earlier. Knowledge of the binding partners and posttranslational modiﬁ cations

of a new protein goes a long way to help characterize it functionally, to solidify its assign-

ment to a pathway, and so on. Since the association of a disease and a particular function or

pathway is often already known, a solid assignment of the new protein to such a function

or pathway implicates the protein in the disease. This adds evidence toward validating the

target.

4. Determine the effect of inhibiting the putative target.

a. Does target inhibition disrupt a disease-related pathway?

b. Does target inhibition slow or stop uncontrolled growth? Have other effects?

c. Does this effect validate the target as unambiguously related to the disease

pathology?

Several methods can be used to answer these questions. Gene knockout studies in mice

have been an effective tool for some time, but knockouts cannot be done with all putative

target proteins. It is in any case a slow, laborious process to breed the knockout strain cor-

rectly and reliably and may not result in viable mice to study.

One recent alternative method has garnered a great deal of attention and is currently

achieving rapid adoption in the industry due to its relative ease, rapidity, effectiveness, and

lower cost. It is called RNA interference (RNAi) (Hannon, 2002). This technique has roots

in both genomics and proteomics. Small interfering RNAs (siRNA) are synthetic 19- to 23-

nucleotide RNA strands that fold in hairpin conﬁ guration. They elicit strong and speciﬁ c

suppression of gene expression; this is called gene knockdown or gene silencing. RNAi

works by triggering the degradation of the mRNA transcript for the target’s gene before

the protein can be formed. RNAi is done in vitro in the lab to verify disease relevance of a

putative target in several possible ways:

• If the target is believed to cause pathology by overexpression, investigators can knock

down the gene for the target and observe whether the disease pathology is reduced.

• If underexpression is assumed to contribute to disease, the gene can be knocked down

in healthy tissue samples to see whether this elicits the same disease phenotype.

• The pathway in which the target is believed to participate can itself be validated as

disease-related by performing knockdowns of some or all of the genes in the path-

way. Then the putative target gene can be knocked down, and the effect of this on the

pathway’s function can be observed. This both validates disease relevance and veri-

ﬁ es the functional assignment of the target to the pathway.

Note that RNAi is also being investigated for its therapeutic potential. Although there are

major ADME issues to be addressed, RNAi molecules, with their high speciﬁ city and ef-

ﬁ cacy in gene suppression, may themselves hold great promise as drug candidates.

8.9.3 Screening for Hits

Now we have one or more possible targets. For each target we need to screen many com-

pounds to look for drug candidates that show activity against the target (i.e., hits). This

PROTEOMICS AND DRUG DISCOVERY 259

260 PROTEOMICS AND DRUG DISCOVERY

can be like searching for a needle in a haystack, so any techniques that can help accelerate

and focus this search are of great value. The following techniques are fairly new but are in

current use.

1. Develop sets of compounds to screen for activity against the target. The ﬁ rst order of

business is to construct focused sets of compounds to screen against the target. Structural

proteomics and combinatorial chemistry can play major roles at this stage. As mentioned

before, an x-ray crystallographic structure for the protein provides the golden standard for

three-dimensional structure. Given such structural information, it is possible to develop

much more focused compound sets for screening libraries than would otherwise be possible.

Combinatorial chemistry can then be used to design the libraries of such compounds. These

are numerous small modiﬁ cations of a basic small molecule or side group that is likely to ﬁ t

the known binding pocket of the target protein based on the structural information.

Computational chemistry is also adding to this effort. It is being used to generate large

virtual compound libraries as a part of structure-based drug design (see below). Like the

combinatorial compound libraries, these are sets of compounds that are likely to ﬁ t well with

the target’s binding site. However, these compounds are all in silico, in the computer only.

2. Screen for compounds that affect the target. (In most cases, inhibition is aimed for. Most

drug targets are enzymes, and most of these are overactive, either by being overabundant or

by being stuck in an active state. Therefore, most drug candidates attempt to inhibit or shut

off the enzyme by binding to its active site.)

High-throughput laboratory screening can now proceed. This technique uses 96- or 384-

well plates to combine the target protein with each of the screening compounds, one per

well. In this way, the compounds are individually tested for activity against the target. Once

activity is detected, even if it is only moderate, that compound is designated as a hit. Chem-

ists can later attempt to increase its activity by modifying the compound and retesting.

Virtual screening is another, newer technique that is showing some promise. Using the

virtual compound libraries, virtual screening uses elaborate computational chemistry tech-

niques to determine in silico the ﬁ t between each virtual compound model and the binding

site of the protein model derived from the x-ray crystallographic structure. This involves

computing the chemical afﬁ nity, the steric ﬁ t, and the energetics of the compound in the

binding pocket. When certain thresholds are reached, a hit is declared, and reserved for

further optimization (Sneider et al., 2002).

3. Structure-based drug design. As mentioned above, virtual screening entails the com-

putational identiﬁ cation of a drug candidate from the ground up. This is called structure-

based or rational drug design. Structural proteomics provides the core information needed

to achieve this. There are two main approaches used in structure-based drug design (Wolf

et al., 2003):

a. Building up an optimized ligand from a known inhibitor molecule. If structural in-

formation is available for a known inhibitor ligand, this can be modeled and used as

a starting point. Placed computationally into the target protein’s binding site, it can

be manipulated on the computer by chemical changes or the addition, moving, or

subtraction of chemical groups or even atoms until its ﬁ t is considered strong. Some

docking programs ﬁ rst decompose the known ligand into fragments before the user

docks them appropriately in the binding pocket and begins optimizing.

b. De novo ligand design and/or docking into the binding pocket model. If no known

inhibitor exists, or there is no structural information on one, a ligand can be built up

from scratch. A base fragment is initially placed in the binding site, then additional

fragments or atoms are added according to sets of rules derived from many known

protein–ligand structures.

8.9.4 Lead Optimization

At this point, one or more promising drug candidate hits have been identiﬁ ed and are ready

to be promoted to lead status. This entails verifying or optimizing the previously mentioned

qualities necessary for the drug candidate to be pursued: ease of synthesis, adherence to

the Lipinsky rules, target speciﬁ city, and efﬁ cacy against the disease. Following are two

lead optimization endeavors in which proteomics and related computational techniques are

making signiﬁ cant inroads.

1. Optimize the hits rationally (i.e., not by trial and error). Continuing the theme of ratio-

nal drug design, let us suppose we now have a model of a lead compound that was gener-

ated by one of the methods outlined above. Structural proteomics can continue to be of

use in providing precise docking and binding information for the optimization of the lead.

Computational chemistry is used to calculate the compound’s adherence to the Lipinsky

rules. Then structural proteomics and computational chemistry can team up to evaluate

quantitative structure–activity relationships (QSARs) of the lead candidate. QSAR is the

attempt to correlate the structural descriptors of a compound with its activity. The analysis

is done computationally, and can help predict not only the strength of the desired activity

of the compound, but also activities that might tend toward toxicity or impaired ADME

characteristics. The generation and interpretation of QSAR models is complex and can be

greatly aided by the application of cheminformatics and data mining techniques (Weaver,

2004). The application of this sort of structural proteomics screening and optimization has

been applied successfully in developing HIV protease inhibitors, antimicrobial drugs, and

inﬂ uenza virus neuraminidase inhibitors (Walgren et al., 2004).

Of course, such computational results must be veriﬁ ed in the laboratory. The in vitro

analog of QSAR analysis is accomplished by the use of activity-based probes. These are

chemically reactive compounds that can be used to identify and quantify a speciﬁ c biologi-

cal activity of a protein based on its active component (Walgren et al., 2004). Both QSAR

analysis and in vitro testing with activity-based probes help to verify target speciﬁ city and

strength of activity.

2. Identify a closely related animal model for initial in vivo testing. Eventually, it is highly

desirable to identify and use a valid animal model for in vivo testing of the lead compound

to help verify efﬁ cacy against the disease in a living organism. This can be a very difﬁ cult

endeavor, but it can be helped by the application of sequence analysis and comparative

genomics. Assuming that we now have a validated target protein, it is possible to take the

amino acid or nucleotide sequence of the protein and search for orthologs in animals typi-

cally used for models. Hopefully, an ortholog with high sequence similarity can be found,

and that animal would then be a candidate for in vivo testing of the lead candidate. Much

effort goes into developing a breed of laboratory animal that exhibits the disease phenotype

and in which the disease is demonstrably similar to that in humans. If such a model has

been developed for this disease, the association between the homologous protein and the

disease phenotype can be established. For instance, if that protein is overexpressed in the

disease, RNAi can be used to silence the gene for the protein, and the effect on the course

PROTEOMICS AND DRUG DISCOVERY 261

262 PROTEOMICS AND DRUG DISCOVERY

of the disease is observed. Once this association has been established, the animal is now

ready for in vivo testing with the lead candidate.

8.9.5 Pharmacology and ADME-Tox

Solid leads have been identiﬁ ed and optimized by this point, and shown to possess the

right druglike qualities and efﬁ cacy in in vitro studies. However, the large gap between the

lab assay and the living being must now be narrowed. Further tests with animal models,

plus early human trials, have been and will continue to be the mainstay of this assessment.

However, proteomics can play a major role by enabling ADME-tox analyses and screen-

ings earlier in the drug discovery process.

1. Eliminate early on those candidate compounds that have poor “druggability” charac-

teristics (i.e., poor ADME or toxicity proﬁ le). In traditional drug discovery and develop-

ment, ADME and toxicity assessments were normally done at this stage, with in vitro and,

especially, in vivo studies. In fact, this assessment continues on into clinical trials, and it

has sometimes been true that ADME-tox properties were unable to be fully understood un-

til late in clinical trials, after tens or hundreds of millions of dollars had been spent. Clearly,

failure of a drug candidate at such a late stage is ﬁ nancially very onerous. Therefore, there

has been intense interest in the notion of “fail fast”: Try to identify and eliminate as early as

possible any compound which has ADME-tox properties that are likely to make it a poor or

unacceptable drug. Accordingly, drug companies are working to move ADME-tox analysis

into increasingly earlier stages of the discovery process.

Many of the analytical methods discussed above can aid in this. Structural proteomics-

based QSAR and its relative, quantitative structure–toxicity relationship (QSTR), help to

identify qualities such as poor absorption or distribution, suboptimal metabolism (e.g., too

many metabolic products), too-fast or too-slow elimination, and toxicity. These analyses

can be performed earlier, in the lead optimization stage, so that by the time the lead candi-

date enters clinical trials there is a better understanding of its pharmacology and toxicol-

ogy. It is even possible to perform preliminary ADME-tox screenings earlier than that, such

as in the target validation or screening stages.

2. Verify speciﬁ city; screen lead drug candidates against nontarget proteins in the pro-

teome to ensure minimal cross-reactivity. Proteomics can be used in another way to build

conﬁ dence in low side effects and toxicity. High-throughput protein interaction studies can

determine whether the lead candidate shows activity against other human proteins that are

not the intended target. As the technology matures, functional protein microarrays should

prove to be an excellent method for this type of assessment. A chip can be arrayed with

a proteome-scale collection of proteins and then incubated with the lead candidate com-

pound, then monitored for off-target protein interactions. In the near term, 96- or 384-well

arrays used in high-throughput mode are more feasible for this type of assay, although

somewhat lower scale.

There is another interesting use of this approach. This kind of high-throughput assay can

also identify potentially beneﬁ cial off-target interactions. For instance, imatinib mesylate

(Gleevec) has been developed as a very successful therapeutic for chronic myeloid leu-

kemia. It works by speciﬁ cally inhibiting BCR-ABL tyrosine kinase. Recently, it has

been shown that Gleevec is also active against other tyrosine kinases and therefore may

prove to be useful as a drug for other diseases where other tyrosine kinases are overactive

(Nadal et al., 2004). That kind of beneﬁ cial off-target activity of a lead candidate can

be identiﬁ ed much earlier, using proteomics approaches such as functional protein chips

arrayed, for instance, with a variety of kinases in the case of a kinase-inhibitor such as

Gleevec. The drug development process could then split into several parallel tracks, each

targeting a different disease and a different kinase.

3. Enable tissue-speciﬁ c delivery of a drug candidate within the body. This is an area of

intense interest to the industry. Very promising drug leads have had to be abandoned due

to the inability to get them delivered to the target tissue. It is particularly difﬁ cult to get

drugs delivered with speciﬁ city to solid tumor cancer cells. However, new approaches are

showing potential to help.

Example In the target identiﬁ cation section, we gave the example of proteomic

identiﬁ cation of tissue-speciﬁ c blood-vessel signature display proteins as potential targets

(Oh et al., 2004). The methods used in that study also uncovered a means for enabling

tissue-speciﬁ c drug delivery. The study demonstrated the use of proteomic techniques to

identify and map display proteins that can help direct drugs to their intended speciﬁ c tissue

target of action. Enabling tissue-speciﬁ c drug targeting in this way can also help reduce

side effects by directing drugs away from other areas of the body.

8.9.6 Clinical Trials: Biomarkers and Pharmacogenomics

Even in the clinical trials phase of drug development, proteomics can play several pivotal

roles: the identiﬁ cation of biomarkers for diagnostics, for safety, and for measuring thera-

peutic efﬁ cacy. One signiﬁ cant area enabled by biomarkers is pharmacogenomics.

Biomarkers Biomarkers are biological indicators found in tissue, serum, urine, and so

on, used to determine the following:

• Normal versus disease states

• Pathogenic processes

• Positive and negative responses to therapeutic intervention

In clinical trials, they can play several crucial roles: as diagnostics, to indicate the efﬁ cacy

of a drug, to help guide dose selection, and to predict or indicate toxicity. Historically,

biomarkers have typically been single indicators. However, it has been increasingly recog-

nized that many diseases, such as many forms of cancer, are complex and heterogeneous. In

such cases, single-analyte biomarkers simply cannot capture enough information to serve

as valid indicators of disease, drug efﬁ cacy, and so on. For these cases the much broader-

scope approaches of genomics and proteomics are being used increasingly to discover and

provide multianalyte biomarker proﬁ les, or panels.

In genomics, this strategy has already proven useful in several cases. For instance,

Genomic Health of Redwood City, California, has launched its Oncotype DX diagnostic,

which measures expression levels of a panel of 16 genes to predict the course of breast

cancer. In this case, that means predicting which tumors are likely to metastasize.

Another company, Agendia of Amsterdam, is also marketing a breast cancer proﬁ ling test,

Mammaprint. This uses a panel of 70 genes found to correlate very highly with future

metastasis (Garber, 2004).

PROTEOMICS AND DRUG DISCOVERY 263

264 PROTEOMICS AND DRUG DISCOVERY

These diagnostic panels were developed using a technique that is common to both

genomic and proteomic biomarker proﬁ le discovery. The technique uses the analysis of

large-scale expression patterns to uncover statistically strong associations between over-

or underexpression of sets of genes or proteins and the target biological metric, such as

disease vs. normal states. This is made possible by a combination of high-throughput tech-

nologies such as microarrays along with advanced computational and statistical analysis

tools that can mine these large datasets for such associations and determine their statistical

signiﬁ cance. One interesting point about this approach is that it is not absolutely necessary

to know the function of all the genes or proteins in the panel. Some may even be novel

genes or proteins whose functions have never been elucidated. Still, if their expression dif-

ferences can be strongly and repeatably correlated with the desired biological metric, they

can become validated biomarkers.

Genomic biomarker proﬁ les can be very useful, but they also have some inherent draw-

backs: They are usually inadequate for determining drug efﬁ cacy and toxicity, two key

biological metrics to be measured in clinical trials. This is because it is the proteins that

are ultimately involved in every disease process, and almost all drugs have proteins as their

target. Also, toxicity is usually due to a drug’s undesirable metabolic changes or its bind-

ing to off-target proteins that negatively alters normal functioning. Therefore, proteomics

approaches can play a major part in discovering and developing biomarker proﬁ les that

measure drug efﬁ cacy and toxicity.

Several proteomics approaches are showing utility in single- and multianalyte biomarker

panel discovery. In many cases, they can be used with easily collected biological ﬂ uids such

as blood or urine. Many of the technologies discussed in Section 8.6 are applicable. 2D-PAGE

has been used successfully for identifying differential protein expression, although it has the

aforementioned limitations, such as replicability issues and the requirement for relatively

high analyte volumes. Protein chips, although not yet a fully mature technology, are a natu-

ral ﬁ t for highly parallel protein expression proﬁ ling leading to biomarker panels (Rocken

et al., 2004). For example, an ELISA-based capture chip was used successfully to identify

a panel of ﬁ ve circulating angiogenic factor protein levels as biomarkers for gynecological

tumors (Huang et al., 2004). Liquid chromatography and mass spectrometry are being used

in several ways for biomarker discovery. One is to analyze the low-molecular-weight portion

of the proteome to generate new classes of disease biomarker proﬁ les. These have shown

signiﬁ cant success in early disease detection (Johann et al., 2004). In another approach, a

version of MALDI called surface-enhanced laser desorbtion ionization (SELDI) has been

used with MS to identify a number of panels of multiple protein biomarkers, each with an

expression proﬁ le speciﬁ c to one of several different tumor types, including breast cancer

(Li et al., 2002; Shin et al., 2002). One very promising area is the harnessing of known im-

mune response to some early-stage cancers to discover and validate tumor-speciﬁ c antigens

in serum and in tumor lysates as biomarkers for cancer. This has been demonstrated effec-

tively using 2D-PAGE and MS to identify biomarkers for breast cancer, and validated with

Western immunoblot analysis (Shin et al., 2002). Another area where proteomics techniques

are clearly indicated is the identiﬁ cation of posttranslationally modiﬁ ed forms of proteins. In

some cases it is the modiﬁ cation, especially cleavage or glycosylation, that identiﬁ es a pro-

tein as a biomarker. For example, several techniques, including 1- and 2-D PAGE followed

by MS, as well as protein chips, have been used to detect new PTM-speciﬁ c biomarkers for

breast cancer (Shin et al., 2002). These techniques show signiﬁ cant promise in developing

biomarkers for diagnostics, for understanding pathogenic process, and for measuring drug

efﬁ cacy and toxicity in clinical trials and beyond.

Pharmacogenomics Often, signiﬁ cant differences in drug efﬁ cacy and/or side effects

appear in clinical trials among the test subjects. There has been considerable recent interest

in ﬁ nding correlations between these differences and gene variations or expression patterns

of genes or proteins among the test subject population. If such a correlation is found, the

variation or expression pattern can become biomarkers that indicate higher likelihood of either

efﬁ cacy or unwanted side effects (or even toxicity). This area is called pharmacogenomics.

In November 2003, the FDA ruled for the ﬁ rst time that pharmacogenomic data can be

included in drug-approval applications. This is widely believed to be the opening of the

door to greater FDA support for targeted medicine, that is, therapeutics that are marketed to

the subset of the population that is likely to beneﬁ t from a particular drug, or that is unlikely

to experience serious negative side effects. Biomarkers are the key enabler of this kind of

medicine; they are essential in identifying the right patients for targeted therapeutics.

In the genomics world, genotyping is used effectively in determining correlations

between a single nucleotide polymorphism (SNP, an inherited gene variation) and the like-

lihood of a person contracting a particular disease. Thus, SNP genotyping is a statisti-

cally predictive tool. However, SNP variants are not considered the best biomarkers for

either the presence of a disease or as a measure of therapeutic activity against the disease.

There are too many other ways for a gene to get damaged than just a SNP: environmental

inﬂ uences are a major factor. In addition, the SNP gene can be inactivated anyway (e.g.,

through methylation, which silences genes). Here again, gene and protein expression anal-

ysis can detect the activation or inactivation of a gene(s), regardless of the cause. Most of

the biomarker discovery techniques mentioned above can be enlisted to identify single- or

multianalyte biomarkers for pharmacogenomics. To date, however, the most fully devel-

oped biomarkers that are in clinical use are single gene traits that inﬂ uence drug response

in humans.

Example One of the most dramatic examples of pharmacogenomics and targeted therapy

is the HER2/Herceptin story. It was discovered that the gene for HER2, a growth factor

receptor protein, is overexpressed in 25 to 30% of breast cancers. In those cases, tumor

aggressiveness is greatly increased. It was further discovered that a humanized anti-HER2

antibody, Herceptin (trastuzumab), signiﬁ cantly increases effectiveness of chemotherapy

against HER2-overexpressing metastatic breast cancer in a combined regimen (Slamon

et al., 2001). This combined therapy has signiﬁ cantly improved the outlook in those 25

to 30% of patients for later stages of breast cancer that had previously been much more

resistant to treatment. In the Slamon et al. study, the HER2 overexpression biomarker was

measured by immunohistochemical analysis. Since that study, a number of other means,

both genomic and proteomic, for detecting and measuring this important biomarker have

been studied. The proteomic quantitation of circulating HER2 receptor protein in serum

has been shown to hold promise in predicting disease outcome and response to therapy

(Ross et al., 2003).

8.9.7 Case Study

We have seen how proteomics approaches can inﬂ uence drug discovery signiﬁ cantly

at every stage of the process. The following is a speciﬁ c case study showing how pro-

teomics can be used successfully in the target identiﬁ cation and validation stages. It is

about the development of LAF389, an antitumor drug candidate developed at Novartis

(Towbin et al., 2003). It depicts the successful application of 2D-PAGE scaled up as a

PROTEOMICS AND DRUG DISCOVERY 265

266 PROTEOMICS AND DRUG DISCOVERY

proteome-wide high-throughput method for protein proﬁ ling. The example demonstrates

the effective use of proteomic techniques for target identiﬁ cation and validation.

The scientists had a compound, bengamide, which was known to exhibit a signiﬁ cant

antitumor effect. However, the mechanism of action of action of bengamide was unknown,

and its target was yet to be identiﬁ ed. The main use of a proteomic approach was to ap-

ply high-scale 2D gel electrophoresis to narrow down the possible targets for bengamides

without bias, then pursue the candidates with more detailed assays.

The scientists used 2D gel electrophoresis at high scale to generate 1500 protein spots

before and after treatment with bengamide E, with no presumption of a speciﬁ c target.

Any protein spots that showed reproducible differences following bengamide treat-

ment were then considered candidate bengamide targets and were analyzed further with

MALDI-MS to identify them. The spots showed differences in charge, indicating a pos-

sible PTM effect.

Isoforms of the protein called 14-3-3 stood out in this set, due to their known role in

modulation of signaling, cell cycle control, transcriptional control, and apoptosis, suggest-

ing a role in cancer if impaired. So the scientists tested induction of 14-3-3 proteins with

LAF389, a bengamide analog. But which isoform was being induced? The scientists used

an analytical blotting method to separate 14-3-3 isoforms, then did Western blotting us-

ing antibodies against 14-3-3 proteins. IEF immunoblotting was then used with isoform-

speciﬁ c polyclonal antibodies to identify 14-3-3γ as the induced isoform.

Two different MALDI-MS analyses were used to detect the cause of the induced pI

difference. These determined that an N-terminal 14-3-3γ peptide sequence was acetylated

in the absence of LAF389, but retained the unprocessed N-terminal Met (oxidized) in the

LAF389-induced samples. This was veriﬁ ed by using a monoclonal antibody to detect un-

processed 14-3-3γ. These results suggested that LAF389 might be inhibiting MetAp, which

normally acetylates 14-3-3γ.

An enzymatic assay veriﬁ ed that LAF389 does inhibit MetAp (1 and 2). This established

MetAps as direct targets of LAF389. (Their later in vivo studies showed that LAF389

inhibited MetAp2 only.)

Finally, the scientists validated the target using the following two methods:

1. They used siRNA to knock down MetAp2 level by 75%, after which unprocessed

14-3-3γ was detected, verifying the role of MetAp2 n processing 14-3-3γ.

2. Using x-ray crystallography, they determined the crystal structure of MetAp2 bound

with LAF153, another bengamide analog. The structure shows that LAF153 binds in

the same manner expected of a polypeptide substrate.

Note that a proteomic approach like this was essential for target identiﬁ cation. Gene

expression studies could not have detected this mechanism (and did not, as the scientists

report). This is because 14-3-3γ is expressed equally in cultures with and without LAF389;

the effect of LAF389 (via inhibition of MetAp) is strictly on a posttranslational modiﬁ ca-

tion (i.e., the acetylation of 14-3-3γ).

8.10 CONCLUSIONS

In this chapter we have described several recently developed enabling technologies that

are being applied to drug discovery and development. There has been much interest in

increasing the rate and productivity of drug discovery, as the blockbuster drugs that have

sustained the industry recently are rapidly coming off-patent and being replaced by cheaper

generics. The use of technologies based on genomics, proteomics, metabolomics, and so

on, has led drug discovery and development efforts away from a single target mentality

and in the direction of an integrated systems biology perspective. With this major shift

to larger scale comes a resulting major increase in both speed and quality of leads. The

emerging technologies described in this chapter aim to address the many aspects of drug

development and to enable more effective understanding of targets and compounds through

high-throughput high-content analysis. These emerging high-throughput technologies are

beginning to show considerable promise for improving and accelerating drug discovery

and development and helping to ﬁ ll the pipelines of pharmaceutical companies.

Improved clinical outcomes of therapeutic candidates at all stages of development

would contribute signiﬁ cantly to reducing development costs and improving the number

of candidates that succeed in obtaining approval for commercialization. By using the tech-

nologies described herein, companies will ﬁ nd that the quality of their drug leads should

be higher and they should be less likely to fail in the clinic. New methods that enable better

target identiﬁ cation and validation, a deeper understanding of total systems biology and

the implications for any speciﬁ c drug or target in development, and methods to understand

the nature of molecular targets to enable more effective screening and lead identiﬁ cation

will be extremely valuable in meeting the needs for cost-effective development of new and

useful therapeutics for human disease.

ACKNOWLEDGMENTS

The authors wish to thank Ioannis Moutsatsos, Director of Protein Informatics at Wyeth,

whose exceptional graduate course in proteomics at Brandeis University has been an in-

valuable source of knowledge and inspiration in writing this chapter.

REFERENCES

Angenendt, P., et al., (2004), Cell-free protein expression and functional assay in nanowell chip

format, Anal. Chem. 76:1844–1849.

Bock, C., et al., (2004), Photoaptamer arrays applied to multiplexed proteomic analysis, Proteomics

4:609–618.

Bonneau, R., et al. (2001), Rosetta in CASP4: progress in ab initio protein structure prediction,

Proteins, Suppl. 5:119–126.

Danner, S., et al. (2001), T7 phage display: a novel genetic selection system for cloning RNA-binding

proteins from cDNA libraries, Proc. Natt. Aced. Sci. USA 98 (23):12954–12959; Nov. 6.

Drews, J. (2000), Drug discovery: a historical perspective, Science 287:1960–1964.

Garber, K. (2004), Gene expression tests foretell breast cancer’s future, Science 303:1754–1755,

Mar. 19.

Gavin, A. C., et al. (2002), Functional organization of the yeast proteome by systematic analysis of

protein complexes, 1, Nature 415(6868):141–147., Jan. 10.

Gershon, D. (2003), Proteomics technologies: probing the proteome, Nature 424:581–587, July 31.

Hanash, S. (2003), Disease proteomics, Nature, 422, March 13.

Hannon, G. (2002), RNA interference, Nature, 418, July.

REFERENCES 267

268 PROTEOMICS AND DRUG DISCOVERY

Huang, R. et al. (2004), Enhanced protein proﬁ ling arrays with ELISA-based ampliﬁ cation for high-

throughput molecular changes of tumor patients’ plasma, Clin. Cancer Res. 10:598–609.

Ideker, T., et al. (2001), Integrated genomic and proteomic analysis of a systematically perturbed

metabolic network, Science 292:929–934.

Ito, T., et al. (2001), A comprehensive two-hybrid analysis to explore the yeast protein interactome,

Proc. Natl. Acad. Sci. USA 98(8):4569–4574, Apr. 10.

Johann, D. J. et al. (2004), Clinical proteomics and biomarker discovery, Ann. N.Y. Acad. Sci.

1022:295–305, June.

Kopp, J. et al. (2004), Automated protein structure homology modeling: a progress report, Pharma-

cogenomics 5(4):405–416, June.

Lee, J., et al. (2003), Differential protein analysis of spasomolytic polypeptide expressing metaplasia

using laser capture microdissection and two-dimensional difference gel electrophoresis, Appl.

Immunohistochem. Mol. Morphol. 11(2):188–193, June.

Li, J., et al. (2002), Proteomics and bioinformatics approaches for identiﬁ cation of serum biomarkers

to detect breast cancer, Clin. Chem. 48:1296–1314.

Lopez, M., and Pluskal, M. (2003), Protein micro- and macroarrays: digitizing the proteome, J.

Chromatog. B, 787:19–27.

Luthra, R., et al. (2004), TaqMan RT-PCR assay coupled with capillary electrophoresis for quantiﬁ -

cation and identiﬁ cation of bcr-abl transcript type, Mod. Pathol. 17(1):96–103, Jan.

MacBeath, G., and Schreiber, S. (2000), Printing proteins as microarrays for high-throughput func-

tion determination, Science, 289; Sept. 8.

MacNeil, J. (2004), Bring on the protein biochips, Genome Technol. 43, May.

Mayer, G., et al. (2003), Aptamers: multifunctional tools for target validation and drug discovery,

DrugPlus Int., Nov–Dec.

Nadal, E., et al. (2004), Imatinib mesylate (Gleevec/Glivec) a molecular-targeted therapy for chronic

myeloid leukaemia and other malignancies, Int. J. Clin. Pract. 58(5):511–516, May.

Oh, P., et al. (2004), Subtractive proteomic mapping of the endothelial surface in lung and solid

tumours for tissue-speciﬁ c therapy, Nature 429(6992):629–635, June 10.

Pellois, J., et al. (2002), Individually addressable parallel peptide synthesis on microchips, Nature

Biotechnol. 20(9):922–926, Sept.

Phizicky, E., et al. (2003), Protein analysis on a proteomic scale, Nature 422:208–215, Mar. 13.

Robinson, W., et al. (2003), Protein microarrays guide tolerizing DNA vaccine treatment of autoim-

mune encephalomyelitis, Nature Biotechnol. Sept.

Rocken, C., et al. (2004), Proteomics in pathology, research and practice, Pathol. Res. Pract.

200(2):69–82.

Rohl, C. A., et al. (2002), De novo determination of protein backbone structure from residual dipolar

couplings using Rosetta, J. Am. Chem. Soc. 124(11):2723–2729, Mar. 20.

Ross, J. S., et al. (2003), Targeted therapy for cancer: the HER-2/neu and Herceptin story, Clin. Lead-

ership Manage. Rev. 17(6):333–340, Nov–Dec.

Shin, B. K., et al. (2002), Proteomics approaches to uncover the repertoire of circulating biomarkers

for breast cancer, J. Mammary Gland Biol. Neoplasia 7(4), Oct.

Simon, R., et al. (2003), Tissue microarrays in cancer diagnosis, Expert Rev. Mol. Diagn. 3(4):

421–430, July.

Slamon, D. J., et al. (2001), Use of chemotherapy plus a monoclonal antibody against HER2 for

metastatic breast cancer that overexpresses HER2, N. Engl. J. Med. 344:783–792.

Sneider, G., et al. (2002), Virtual screening and fast automated docking methods, Drug Discov.

Today 7:64–70.

Taussig, M., et al. (2002), Protein arrays resource page, ESF Programme on Integrated Approaches

for Functional Genomics, 2002.

Taussig, M., et al. (2003), Progress in antibody arrays, Targets 2(4), Aug.

Towbin, H., et al. (2003), Proteomics-based target identiﬁ cation: bengamides as a new class of

methionine aminopeptidase inhibitors, J. Biol. Chem. 278(52):52964–52971, Dec. 26.

Uehling, M., (2003), Fishing chips, Bio-IT World, Sept.

Uetz, P., et al. (2000), A comprehensive analysis of protein–protein interactions in Saccharomyces

cerevisiae, Nature 403(6770):623–627, Feb. 10.

Walgren, J., et al. (2004), Application of proteomic technologies in the drug development process,

Toxicol. Lett. 149(1–3):377–385, Apr. 1.

Weaver, D. C. (2004), Applying data mining techniques to library design, lead generation and lead

optimization, Curr. Opin. Chem. Biol. 8(3):264–270, June.

Wolf, K., et al. (2003), Information-based methods in the development of antiparasitic drugs, Para-

sitol. Res. 90:S91–S96.

Zhu, H., et al. (2000), Analysis of yeast protein kinases using protein chips, Nature Genet. 26, Nov.

Zhu, H., et al. (2001), Global analysis of protein activities using proteome chips, Science 293,

Sept. 14.

APPENDIX: PUBLIC-DOMAIN SOFTWARE TOOLS AND DATABASES

Nucleotides, Genomics, Documentation, General

Databases Purpose

NCBI Entrez Gateway page for all the NCBI databases: nucleotide and protein

DBs, PubMed, Genome, Structure, OMIM, etc.

http://www.ncbi.nlm.nih.gov/Entrez/

GeneCard Compendium of information related to genes gathered automatically;

useful; good collection of links in one place, but slow

http://bioinfo.weizmann.ac.il/cards/index.html

UCSC Genome Browser Graphical genome browser; highly conﬁ gurable

http://genome.ucsc.edu/

APPENDIX: PUBLIC-DOMAIN SOFTWARE TOOLS AND DATABASES 269

Tools Purpose

BLAST Sequence alignment and homology studies (amino acids and

DNA formats)

http://www.ncbi.nlm.nih.gov/BLAST/

WebLogo Consensus sequence logo generator

http://weblogo.berkeley.edu/

GenScan, WebGene Gene prediction (including splice sites, CpG islands, repeat

elements, polyA sites, promoter regions)

http://genes.mit.edu/GENSCAN.html

RepeatMasker Annotation and masking of interspersed repeats and low-

complexity areas (used in conjunction with GenScan)

http://repeatmasker.genome.washington.edu/cgi-bin/

RepeatMasker

(Continued)

270 PROTEOMICS AND DRUG DISCOVERY

Proteins, Proteomics

Databases Purpose

Map Viewer Graphical exploration of sequence data with respect to

chromosome loci

http://www.ncbi.nlm.nih.gov/mapview/

LocusLink Chromosome and gene location information

http://www.ncbi.nlm.nih.gov/LocusLink/

KEGG Kyoto Encyclopedia of Genes & Genomes: interaction networks

(e.g., metabolic and regulatory pathways)

http://www.genome.ad.jp/kegg/

DAVID DB for Annotation, Visualization, and Integrated Discovery:

GoCharts (functional), KeggCharts (metabolic), DomainCharts

http://apps1.niaid.nih.gov/david/

Homologene Homologs in model organisms

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=homologene

Entrez SNP SNP analysis

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=snp

Entrez Taxonomy Taxonomic information for phylogenetic work

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Taxonomy

PubMed Published information on genes and proteins

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed

OMIM Information and references (good summaries)

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM

Tool Purpose

Clustal Software for creating multiple sequence alignment (MSA). The

source code for CLUSTAL X and several executable versions for

different machines are freely available by anonymous ftp to

ftp igbmc.u-strasbg.fr. Hypertext documentation can be viewed at:

http://www-igbmc.u-strasbg.fr/BioInfo/clustalx/

The NCBI Vibrant Toolkit is available by anonymous ftp from:

http://ncbi.nlm.nih.gov

BioEdit MSA editing and manipulating

http://www.mbio.ncsu.edu/BioEdit/bioedit.html

MEGA Molecular evolutionary genetics analysis: phylogenetic trees from

MSAs

http://www.megasoftware.net/

DeepView (spdbv) Swiss-PDB viewer: for protein 3D structure analysis, homology

modeling, and visualization

http://www.expasy.org/spdbv/

PeptIdent Identiﬁ cation of proteins using experimental pI, Mw, and peptide

mass ﬁ ngerprinting data; compares to SwissProt database to

ﬁ nd closest matches

http://us.expasy.org/tools/peptident.html

Mascot Uses mass spectrometry data to identify proteins from primary

sequence databases

http://www.matrixscience.com/

PredictProtein Predicts secondary protein structure, given an amino acid sequence

http://www.embl-heidelberg.de/predictprotein/

APPENDIX: PUBLIC-DOMAIN SOFTWARE TOOLS AND DATABASES 271

Tool Purpose

ProtScale Protein proﬁ ling tools using a number of scales: hydrophobicity,

polarity, α-helix, β-sheet, etc.

http://www.expasy.org/cgi-bin/protscale.pl

Rosetta Ab initio 3D protein fold prediction using David Baker’s Rosetta

algorithm. Rosetta server available at:

http://www.bioinfo.rpi.edu/~bystrc/hmmstr/server.php

Database Purpose

Expasy–SwissProt Curated, highly annotated protein database

http://us.expasy.org/

http://us.expasy.org/sprot/

InterPro, PFAM Protein families, domains; graphical domain structure alignments

http://www.ebi.ac.uk/interpro/

http://www.sanger.ac.uk/Software/Pfam/index.shtml

InterDom Database of interacting domains: database of putative interacting protein

domains from both experimental and computational sources

http://interdom.lit.org.sg/

ProSite Documentation; consensus patterns

http://us.expasy.org/prosite/

CDD Conserved domain database: more comparative structure and domain

information; includes MSAs

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=cdd

PDB, PDBsum Protein database; primarily for 3D structures; crystallographic

http://www.rcsb.org/pdb/index.html

http://www.biochem.ucl.ac.uk/bsm/pdbsum/index.html

DIP Database of interacting proteins

http://dip.doe-mbi.ucla.edu/

Reactome Curated database of core human biology reactions and pathways

http://www.reactome.org/

BioCarta Categorization of pathways of all kinds: metabolic, cell signaling, cell

cycle regulation, apoptosis, etc

http://www.biocarta.com/genes/index.asp

Swiss-2DPage Two-dimensional polyacrylamide gel electrophoresis database: identiﬁ -

cation of proteins by mass/charge spots in 2D gel database

http://au.expasy.org/ch2d/

SCOP Structural classiﬁ cation of proteins (e.g., by secondary, tertiary, etc.)

http://scop.mrc-lmb.cam.ac.uk/scop/index.html

273

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

USING DRUG METABOLISM

DATABASES DURING DRUG DESIGN

AND DEVELOPMENT

PAUL W. ERHARDT

The University of Toledo College of Pharmacy

Toledo, Ohio

9.1 INTRODUCTION

The type of drug typically pursued during pharmaceutical research and development is

one that can be used with a wide margin of safety via oral administration to humans. In

this chapter we consider drug metabolism within that context. Thus, it should be appreci-

ated immediately that beyond a drug’s speciﬁ c structural motifs that dictate its distinct

interactions with selected sets of human metabolizing enzymes, other structural features

that affect a drug’s absorption, distribution, and elimination also represent key factors that

become equally important toward inﬂ uencing the overall course of its metabolism. For ex-

ample, a drug that is metabolized by enzyme A to a greater extent than by enzyme B during

a pair of in vitro assays may still be signiﬁ cantly metabolized by B within an in vivo setting

if its exposure to B is substantially greater than its exposure to A. The degree of exposure

to an enzyme will be determined by the prevalence of that enzyme in various compart-

ments coupled with the drug’s distribution into those compartments. Figure 9.1 displays the

interplay of these considerations by tracing the path that a central nervous system (CNS)

drug will traverse on route to its site of efﬁ cacious action after oral administration. Steps

highlighted by an asterisk represent compartments that are known to have high levels of

metabolic activity. Tables 9.1 and 9.2 indicate how the spectrum, as well as the overall

level, of metabolic activity varies across several different tissues and across several clas-

sic human phenotypes, respectively.

In the end, the overall course of a drug’s metabolism

will depend upon the drug’s initial and continued distribution to the sets of metabolizing

enzymes variously displayed within the metabolically active compartments combined with

274 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

Oral Cavity/Mucosa* Intestinal Cavity/Gut Bacterial Flora* and Mucosa*

(Stomach pH 1-3 Duodenum pH 5–7 Jejunum Ileum

pH 7–8) Hepatoportal Circulation (Blood*) Liver* Venous

Return Heart Bronchopulmonary Circulation/Lungs*

Heart Systemic Circulation (Toward Major Organs, Periphery and CNS)

Vascular Endothelium (Blood-Brain Barrier) CNS Interstial

Fluid (Cerebrospinal Fluid) CNS Target Cells/Cell Membrane

Intracellular Fluid/Organelles and Target Receptor

Figure 9.1 Path taken by a CNS drug to its site of action after oral administration. Compartments

having high metabolic activity are marked by an asterisk. Although the intestinal mucosa and liver are

regarded as the principal organs associated with the ﬁ rst-pass effect, the lungs also display consider-

able xenobiotic metabolizing capability and can be considered to represent the key pseudo-ﬁ rst-pass

organs

for all routes of administration other than oral or intraperitoneal. The asterisk placed on the

blood compartment reﬂ ects high levels of speciﬁ c esterase and amidase activities rather than a high

level of overall metabolic capability.

TABLE 9.1 Metabolic and Excretion Capabilities Displayed by Selected Tissues

Tissue Metabolic/Excretion Activities

Gastrointestinal mucosa Rich in CYP 3A, but essentially all CYPs are represented, gradu-

ally peaking in the duodenum region and then falling off toward

the ileum; signiﬁ cant glucuronide and sulfate conjugation

pathways; signiﬁ cant monamine oxidase activity.

Blood Rich in various esterase and amidase enzymes.

Liver Rich in essentially all CYPs and conjugation pathways. Excrete

compounds into bile having MW ⱖ 500 g, particularly when

highly polar, conjugated molecules.

Lungs Represented by most of the CYPs and conjugation pathways

with certain levels being comparable or even higher than levels

found in the liver (e.g., CYP 2A13 as recently implicated in the

carginogenicity of nicotine).

Vasculature Neither the BBB, the placenta, or the mammary gland barriers

exhibit signiﬁ cantly enhanced overall xenobiotic metabolizing

capabilities.

Kidneys Rich in the various protease enzymes; excrete compounds having

MW ⱕ 500 g; highly polar water-soluble compounds are not

reabsorbed from the urine.

CYP, cytochrome P450 enzyme

; MW, molecular weight; BBB, blood–brain barrier.

the duration of the drug’s stay and its substrate suitability for those enzymes, all as a func-

tion of drug concentration and as the summation of competing metabolic processes from

one compartment to another over time.

Given the complexity of the aforementioned scenario, two approaches can be contem-

plated toward predicting the metabolism of new drugs being considered for use in humans.

The ﬁ rst involves dividing the in vivo setting into simpler parameters that are amenable to

either quick experimental or theoretical assessment, followed by reassembling the individ-

ual sets of data obtained for a new compound in a manner that accounts for the complexity

of the intact human. The second approach involves the query of established databases for

structures or structural elements that are similar to the new drug wherein the database’s in-

formation has been assembled from drug metabolism studies already undertaken within the

in vivo setting. Both of these approaches are being deployed by the pharmaceutical indus-

try at very early points in the overall process of drug discovery, the ﬁ rst by adopting high-

throughput screening (HTS) methods to assess at least certain of the simpliﬁ ed parameters,

and the second by using commercially available drug metabolism databases.

The status of

the latter approach is reviewed speciﬁ cally in the remainder of this chapter.

9.2 HISTORICAL PERSPECTIVE

Drug metabolism efforts within the pharmaceutical industry have necessarily tended to

focus on the speciﬁ c issues associated with the development of lead compounds selected

for their potential as therapeutic agents and not for their suitability to serve as molecular

TABLE 9.2 Metabolic Capabilities Displayed by Selected Human Phenotypes

Phenotype Characteristic Metabolic Activities

Debrisoquine

Extensive versus poor metabolizers. Inherited defect in CYP 2D6 expression:

5 to 10% Caucasian, 2% Oriental, and 1% Arabic populations.

Phenytoin

Normal versus poor metabolizers. Inherited defect in CYP 2C18: 15 to 20%

Oriental and 2 to 6% Caucasian populations.

Aldehyde oxidase

Lack of indicated capability; signiﬁ cant polymorphism among the Oriental

population (ca. 50%).

Acetylation Fast and slow acetylators. Individual variations in Nat2: 30 to 40%

Caucasian, 80 to 90% Oriental, and 100% Eskimo populations are fast.

Neonatal Overall, CYP isoform activities typically lower than children to adult ranges

except for the 3A subfamily; blood esterase activity appears to be about

50% at birth

; immature UGT pathways fairly common leading to

decreased clearance of bilirubin or gray baby syndrome.

Elderly Most studies directed toward enzymatic activity per se suggest that levels re-

main comparable to younger populations; alternatively, decreased hepatic

blood ﬂ ow can signiﬁ cantly lead to a decrease in the overall metabolism

and excretion of xenobiotics.

CYP, cytochrome P450 enzymes

; NAT2, N-acetyl transferase isoform 2

; UGT, uridinediphosphoglucuronosyl

transferase.

This phenotype extends to numerous other drugs.

This phenotype is largely characterized by an altered pathway for ethanol metabolism.

HISTORICAL PERSPECTIVE

275

276 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

or mechanistic probes to answer metabolic questions. As a result, the drug metabolism

literature has to a large extent become anecdotal, replete with case-by-case examples

but having less information pertaining to systematic investigations directed toward the

rational development of broader principles useful for predicting the biotransformations of

new, biologically interesting compounds in general.

6,7

Nevertheless, as a result of many

early workshops, reviews, and textbook treatments (e.g., refs. 8, 9 to 11, and 12 to 14,

respectively) as well as to numerous updates in the form of periodic reviews and specialist

reports dedicated to the topic of drug metabolism (e.g., refs. 15 to 17), generalizations

about the likely metabolic reactions of certain chemical features have gradually come to

be accepted with a high degree of conﬁ dence, the propensity for ester hydrolysis, O- and

N-dealkylations, and for aromatic hydroxylation being exemplary.

As drug metabolism data have continued to accumulate, it has become convenient to

store and sort this information using computerized approaches (e.g., refs. 19 to 26). In this

regard, two expert systems emerged as the ﬁ rst leaders in this area. Metabol Expert

27–29

and META

30–33

have been constructed by compiling literature intentionally limited to

well-established sources such as textbooks and reviews so as to assure the quality of the

information contained within the database. By using computer-driven queries across these

databases, one can identify sites on a new molecule where metabolic biotransformations

are likely to occur. Unfortunately, these early databases have indiscriminately combined

metabolism data available from studies that employed a variety of mammalian species, and

consequently, their programs tend to predict all of the metabolic possibilities for an exog-

enous material when the latter is placed in a theoretical “average” mammal.

Recognizing

the need to prioritize the often numerous metabolic possibilities, a priority number based

on a scale of 1 (fast biotransformation) to 9 (slow biotransformation) also accompanies

each prediction from the META database.

Subsequent to the development of the early expert systems, two databases that repre-

sent collections of drug metabolism data have also become available. Metabolite

34,35

is a

broad collection of metabolism data that are being accumulated without bias as to literature

source. Thus, the strength of this database may eventually lie in its extensive quantity of

data rather than in the quality of each of its data entries. Alternatively, the Accelrys Me-

tabolism Database

36,37

represents a computerized version of data taken from the well-re-

spected Biotransformations series edited by Hawkins.

Substructure queries across these

databases can be used to predict biotransformations for new compounds by ﬁ nding data for

actual compounds within the database that are closest in structure to each query. Finally,

METEOR

38,39

is a recently released expert system that, like META, has sought to provide a

ranking for the predicted metabolic possibilities arising from a new structural query. These

ﬁ ve databases are summarized in Table 9.3.

9.3 PRESENT STATUS

A survey on the use of drug metabolism databases within the industry has recently been

assembled as part of a book devoted to this overall topic.

While the survey’s responses

largely reﬂ ect experiences that relate to using some of the older databases during the later

stages of the drug discovery process, the survey’s consensus points become informative

toward appreciating the pros and cons of today’s attempts to develop and deploy such

databases during the early stages of drug discovery. The survey’s consensus points are

highlighted in Table 9.4.

TABLE 9.3 Commercially Available Drug Metabolism Databases

Database Vendor

Approx. Number

Search Paradigms

Recommended

Platform

Compatible

Software

Approx. Cost

(thousands of dollars)Cmpd. Biotran.

MetabolExpert CompuDrug, Inc.

South San Francisco, CA

415-271-8800

www.compudrug.com

— — Chemical structure;

transformation;

test system

PC PALLAS 10

META Multicase, Inc.

Beachwood, OH

216-831-3740

www.multicase.com

— — Chemical structure VAX; Open VMS;

Windows-based

—20

Metabolite MDL Information Systems, Inc.

San Leandro, CA

510-895-1313

www.mdli.com

30,000 100,000 Chemical structure;

substructure;

similar structure;

transformation;

data searcher

PC; MAC 40

Metabolism

Database

Accelrys, Ltd.

Leeds, UK

44 113 224 9788

www.accelrys.com

3,000 25,000 Chemical structure;

transformation;

test system; author;

journal name;

keywords

SGI; Windows

NT; Sun Solaris;

Open VMS

ISIS; REACCS;

ACCORD

Database

Explorer

METEOR LHASA

University of Leeds

Leeds, UK

44 113 233 6531

www.chem.leeds.ac.uk/luk

Entries reﬂ ect status at end of 2001.

277

278 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

In addition to the points listed in Table 9.4, two major areas of concern were identiﬁ ed.

Both concerns need to be addressed if we are to more advantageously exploit computer-

ized approaches toward the adoption of metabolism considerations into the early stages

of drug discovery. The two areas involve initial biological data entry/sorting, followed by

chemical structural entry/inquiry. In terms of biologically-related issues, concerns about

indiscriminately combining metabolic data from different species have already been alluded

to. Although there is some analogy to the drug action literature where simple allometric

scaling and physiologically based pharmacokinetic (PBPK) modeling methods can be em-

ployed to adjust or better quantitate appropriate dosing of the efﬁ cacious levels for a drug

to be used in different species,

the potential impact that species differences have upon the

metabolism pathways are still of concern because the latter can also deviate signiﬁ cantly in

a qualitative fashion. For example, consistently distinct differences in phase II metabolism

(conjugation pathways) between various species are well established.

41,42

Alternatively,

differences in phase I metabolism seem to vary in an inconsistent manner. In most cases

where systematic studies have been conducted across different species, qualitative as well

as quantitative differences have been observed between the metabolic proﬁ les obtained for

a given set of selected xenobiotics, even when the species have been restricted to various

mammals.

Similarly, as mentioned in the introduction, the potential for obtaining signiﬁ -

cantly divergent results within the same species when comparing in vitro versus in vivo data

should be reiterated. Thus, the metabolic proﬁ le for a drug that is distributed away from the

liver after intravenous administration is likely to be quite different from that which would

be suggested by its in vitro study using liver microsomal fractions. As above, in most cases

where systematic studies have been conducted across different tissues within the same spe-

cies, differences have been observed between the metabolic proﬁ les obtained for any given

series of selected xenobiotics

(Table 9.1).

Interestingly, the factors pertaining to distribution and pharmacokinetic half-life may

actually be larger concerns while attempting to cross-relate data obtained from in vitro

versus in vivo studies in the same species than while cross-relating data from in vivo stud-

ies obtained between two different species. For example, analyses among our collaborators

have shown that there is a statistically signiﬁ cant correlation between the half-lives of a

wide range of drugs when determined in two very different species, rat data versus human

data.

45,46

The ﬁ nal concern in the biological area has also become encompassed by the new

ﬁ eld of pharmacogenetics. It involves the growing appreciation that there can be signiﬁ cant

differences in drug metabolism due simply to subtle individual differences in enzymatic

TABLE 9.4 Consensus Points from Industry Survey About Using Drug Metabolism Databases

•

Predicting all metabolic possibilities within a theoretical “average” mammal creates a dense forest

of information from which it becomes difﬁ cult to discern the speciﬁ c pathways that might be as-

sociated with the human response.

•

In most cases, many of the suggestions were already suspected from a simple visual inspection of

the query molecules, whereas in other cases, biotransformations were missed despite the existence

of speciﬁ c literature precedent.

•

The need to convey statistically derived metabolic probabilities rather than a list of possibilities

is critical.

Source: Ref. 6.

phenotype. Here, the metabolism of the same drug studied in the same species using the

same method (in vitro or in vivo) can still vary from individual to individual based on clas-

sic differences in population phenotypes

(Table 9.2), on disease-related differences in

phenotype,

48,49

or due to the impact on phenotype resulting from a person’s total metabolic

history of previous and ongoing exposures to xenobiotics.

It should be clear at this point

that many of the issues that have complicated earlier systematic investigations within the

ﬁ eld of drug metabolism remain and serve to exacerbate the additional challenges associ-

ated with today’s construction and use of much larger, generalized databases.

It can be imagined that at least some of these issues might be addressed as a series of

initial sorting steps when data are being entered. Although this strategy would be cum-

bersome initially, it would result in a series of biologically intelligent and perhaps more

manageable databases. Alternatively, if enough searchable terms pertaining to the concern

areas were entered into a single relational database along with the actual metabolism data,

appropriately factored searching paradigms, perhaps coupled with rational PBPK consid-

erations throughout, could seemingly be devised to help surmount these concerns on the

query end. Ultimately, whether these biological “apples and oranges” are separated as the

data are entered, searched, or subsequently rationalized, the desired search query will even-

tually need to be linked with chemical structure. The latter constitutes a different aspect of

constructing databases which has its own set of challenges. An initial consideration of key

chemical issues follows.

In terms of chemical-related issues, there are two areas that come to immediate at-

tention. The ﬁ rst involves a fundamental question that is associated with the pursuit of

structure metabolism relationships (SMRs) and their use in the effective deployment of

metabolism databases. Namely, what is the proper role to expect the entire structure to be

playing versus the discrete roles of its displayed organic functionality? After all, it is the

latter that actually undergo metabolism, and just as a pharmacophore deﬁ nes the pattern

of structural elements requisite for a drug’s interaction with a biologic receptor or enzyme

active site, it will be the particular array of functional groups and their immediate molecu-

lar environments which dictate what happens once a drug is present at the compartment

where a speciﬁ c metabolic conversion is to take place. Numerous secondary issues also

stem from this question, but all of these lead to a common concern about how much detail

needs to be included for a given chemical structure data entry or query. While the level

of detail that can be applied toward molecular description/searching spans a considerable

range, a second chemical issue additionally presents itself at this juncture: namely, the ac-

curacy of the initial chemical depiction. For example, one extreme would be a thorough,

three-dimensional electronic surface map across the entire structure as obtained from x-ray

analysis and/or rigorous computational treatments, additionally coupled with experimental

physicochemical information or descriptors to also address the distribution issues. The

other extreme would be a simple two-dimensional ﬁ gure, perhaps energy minimized with

only the aid of an automated drawing program. This particular chemical quandary is ad-

dressed further in Section 9.4. For the present, practicality has dictated that the most simple

structural data entry possible be used due to the sheer amount of valuable anecdotal infor-

mation that is already available. However, even in this case it may be expecting too much

from a given database to be able to accurately predict the entire metabolic outcome for a

new compound using a single structural query. Rather, it may be more appropriate to think

in terms of a multistep approach that might ﬁ rst include analyses across a parent database

in terms of various metabolic functional groups so as to initially produce a series of speciﬁ c

SMR maps which characterize key metabophore

2,7

elements in terms of the probabilities

PRESENT STATUS 279

280 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

that they might undergo a particular metabolic reaction based on such occurrences relative

to other possibilities. The metabophores may then need to be better reﬁ ned by inputting

considerably more precise structural detail. Such a strategy would certainly be in line with

today’s trend toward smarter libraries as well as with the trends being utilized to integrate

various [other bioinformatic and chemoinformatic] systems.

Employing this set of spe-

cies speciﬁ c atlases, each having its own list of reﬁ ned metabophore/relative metabolic

probability maps, in conjunction with a new drug structure query then becomes a third step

in an overall searching paradigm that will probably also need to be continuously guided

in a rational manner by appropriately considering distribution and pharmacokinetic issues

throughout.

9.4 FUTURE PROSPECTS

From the foregoing discussions, two critical hurdles loom in front of using xenobiobitic

metabolism databases to more effectively predict human drug metabolism in the future.

The ﬁ rst involves the need for an improved treatment of three-dimensional structure within

chemical-related databases in general, and the second involves the need to be able to better

correlate the various types of metabolism-related data to the human clinical experience.

The chemical structure hurdles are addressed ﬁ rst.

Handling chemical structures and chemical information within the setting of large

databases represents a specialized exercise complicated enough to merit its own designation

as a new ﬁ eld, that of chemoinformatics.

52–55

As mentioned, there is a signiﬁ cant need for

improvement in the handling of chemical structures beyond what appears to be occurring

within today’s database assemblies. For example, that “better correlations are sometimes

obtained by using 2D displays of a database’s chemical structures than by using 3D displays”

only testiﬁ es to the fact that we are still not doing a very good job at developing the later.

In general, the handling of small molecules and of highly ﬂ exible molecular systems

controversial, with the only clear consensus being that treatments of small molecules for

use within database collections “have, to date, been extremely inadequate.”

Certainly, a

variety of automated, three-dimensional chemical structure drawing programs are available

that can start from simple two-dimensional representations by using Dreiding molecular

mechanics or other user-friendly automated molecular mechanics-based algorithms, as

well as by using data expressed by a connection table or linear string.

Some programs

are able to derive three-dimensional structure “from more than 20 different types of import

formats.”

Furthermore, several of these programs can be directly integrated with the latest

versions of more sophisticated quantum mechanics packages such as Gaussian 98 MOPAC

(with MNDO/d) and extended Hückel.

55,59

Thus, electronic handling of chemical structures

and to a certain extent comparing them in three-dimensional formats has already become

reasonably well worked out.

52–55,59–62

Table 9.5 is a list of some of the three-dimensional

molecular modeling products that have become available during the 1990s.

The fundamental problem that remains, however, is how the three-dimensional structure

is derived initially in terms of its chemical correctness, the latter being dependent on what

assumptions might have been made during the process of energy minimization. This situ-

ation is further complicated by the additional need to understand how a given drug mole-

cule’s conformational family behaves during its interactions with each of the biological en-

vironments of interest: those associated with all of the compartments traversed in Fig. 9.1,

along with those associated with each of the speciﬁ c metabolizing enzymes that the drug

TABLE 9.5 Three-Dimensional Molecular Modeling Packages that Became Available During the 1990s

Package Company Platform Description

Low-end sophistication

Nano Vision ACS Software MAC Simple, effective tool for viewing and rotating

structures, especially large molecules and proteins

Ball & Stick Cherwell Scientiﬁ c MAC Model building and visualization; analysis of bond

distances, angles

MOBY Springer-Verlag IBM (DOS) Model building and visualization; classical and

quantum mechanical computations; large

molecules and proteins; PDB ﬁ les

Nemesis Oxford Molecular IMB (Wind) MAC Quick model building and high-quality visualization;

geometric optimization (energy minimization)

CSC Chem. 3D/Chem 3D Plus Cambridge Scientiﬁ c MAC Easy-to-use building and visualization; geometric

optimization; integrated 2D program and word

processing

Alchemy III Tripos Assoc. IBM (DOS, Wind) MAC Quick model building; energy minimization; basic

calculations; easy integration to high-end systems

PC Model Serena Software IBM (DOS) MAC Low cost with sophisticated calculations; platform

ﬂ exibility

Mid range sophistication

CAChe Tektronic MAC Sophisted computation tools; distributed processing

HyperChem Auto Desk IBM (Wind) Silicone Graphics Easy-to-use array of computation tools (classical and

semiempirical quantum mechanics)

Lab Vision Tripos Assoc. IBM (RISC-6000) Silicone Graphics

DEC, VAX

Sophisticated but practical modeling for research

High-end sophistication

SYBL Tripos Assoc. IBM (RISC-6000) Silicone Graphics

DEC VAX, Sun 4, Convex

Integrated computation tools for sophisticated

structural determination and analysis; database

management

CERIUS Molecular Simulations Silicone Graphics IBM (RISC-6000)

Stardent Titan

Suite of high-performance tools for building and

simulating properties

Source: Ref. 61.

281

282 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

will eventually encounter. To track these conformational behaviors in a comprehensive

manner, it becomes necessary to consider a drug’s multiple conformational possibilities by

engaging as many different types of conformational assessment technologies as possible

while initially taking an approach that is unbiased by any molded relationship dictated by a

speciﬁ c interacting environment. For example, three common methodological approaches

include (1) x-ray; (2) solution spectroscopic methods such as nuclear magnetic resonance

(NMR), which can often be done in both polar and nonpolar media; and, (3) computational

approaches, which can be done with various levels of solvent and heightened energy con-

tent but are limited by the assumptions and approximations that need to be taken in order

to simplify the mathematical rigor so as to allow computational solutions to be derived in

practical time periods. Analogous to the simple, drawing program starting points, programs

are available for conversion of x-ray and NMR data into three-dimensional structures.

An example of how three-dimensional conformation might be addressed is provided

by the following description of an ongoing project in our labs that pertains to construction

of a human drug metabolism database. Structures are initially considered as closed-shell

molecules in their electronic and vibrational ground states with protonated and unproton-

ated forms, as appropriate, also being entered. If a structure possesses tautomeric options,

or if there is evidence for the involvement of internal hydrogen bonding, the tautomeric

forms and the hydrogen-bonded forms are additionally considered. Determination of

three-dimensional structure is carried out in two steps. Preliminary geometry optimization

is affected by using a molecular mechanics method, in our case the gas-phase structure be-

ing determined by applying the MacroModel 6.5 modeling package running on a Silicone

Graphics Indigo 2 workstation with modiﬁ ed (and extended) AMBER parameters. Multi-

conformational assessment using systematic rotations about several predeﬁ ned chemical

bonds with selected rotational angles is then conducted to deﬁ ne the low-energy conform-

ers and conformationally ﬂ exible regions for each starting structure. In the second step,

the initial family of entry structures are subjected to ab initio geometry optimizations,

which in our case use the Gaussian 98 package running at the T90 machine in the Ohio

Supercomputer Center resource. Depending on the size of the molecule, 3-21G* or 6-32G*

basis sets

are used for conformational and tautomeric assessments. Density functional

theory using the B3LYP functional

is applied for the consideration of exchange correla-

tion energy while keeping the required computer time at reasonable levels. The highest-

level structure determination is performed at the B3LYP/6-31G* level. To ascertain the

local energy minimum character of an optimized structure, vibrational frequency analysis

is carried out using the harmonic oscillator approximation. Determination of vibrational

frequencies also allows for obtaining thermal corrections to the energy calculated at 0 K.

Free energies are then calculated at 310 K (human body temperature). From the relative

free energies calculated, the gas-phase equilibrium constant and the composition of the

equilibrium mixture can be determined. Although these values may not be relevant in polar

media such as an aqueous environment or the blood compartment, the calculated con-

formational distribution is relevant for nonpolar environments that may be encountered

when a drug passively traverses membranes or begins to enter the cavity of a nonhydrated

receptor/enzyme active site just prior to binding. Repetition of this computational scheme

from biased starting structures based on actual knowledge about the interacting biological

systems or from x-ray or NMR studies (particularly when the latter have been conducted

in polar media), followed by studies of how the various sets of information become inter-

changed and how they additionally behave when further raised in energy, complete the

chemical conformational analyses that are being done for each structure being adopted into

our human drug metabolism database.

As mentioned earlier, however, after taking an unbiased structural starting point, struc-

tures also need to be considered by ascertaining what their relevant conformations might

be during interactions within various biological milieus. It can be imagined that at least

within the immediate future, a useful range of such environments to be considered will

include aqueous solutions of acidic and neutral pH: namely, at about 2 (stomach) and 7.4

(physiological), respectively; one or more lipophilic settings, such as might be encountered

during passive transport through membranes; and ﬁ nally, speciﬁ c biological receptors and/

or enzyme active site settings that are of particular interest. Importantly, with time this list

can then be expected to grow further so as to include several distinct environmental mod-

els deemed to be representative for interaction with various transportophore relationships;

several distinct environmental models deemed to be relevant for interaction with speciﬁ c

metabophore relationships such as within the active site of a speciﬁ c cytochrome P450

metabolizing enzyme; and ﬁ nally, several distinct environmental models deemed to be rel-

evant for interaction with speciﬁ c toxicophore relationships. If x-ray, NMR, and so on, can

be further deployed to assess any one or combination of these types of interactions, a com-

posite approach that deploys as many as possible of these techniques will again represent

the most ideal way to approach future conformational considerations within the variously

biased settings. Advances toward experimentally studying the nature of complexes where

compounds are docked into real and model biological environments are proceeding rapidly

in all of these areas. In addition to the experimental approaches, computational schemes

will probably always be deployed because they can provide the relative energies associated

with all of the various species. Furthermore, computational methods can be used to derive

energy paths to get from the ﬁ rst set of unbiased structures to a second set of environ-

mentally accommodated conformations in both aqueous media and at biological surfaces.

Importantly, these paths and their energy differences can then be compared within database

settings along with the direct comparison of the structures themselves, while attempting to

uncover and deﬁ ne correlations between chemical structure and some other informational

ﬁ eld.

Finally, it should be noted that by using computational paradigms, these same types

of comparisons (i.e., among and between distinct families of conformationally related

members) can also be done for additional sets of conformations that become accessible

at increased energy levels (i.e., at one or more 5 kcal/mol increments of energy) so as to

simulate the beneﬁ cial losses of energy that might be obtained during favorable binding

with receptors or active sites.

These types of altered conformations can also become

candidates for structural comparisons between databases. The latter represents another im-

portant reﬁ nement that could become utilized as part of SAR queries that will need to be

undertaken in the future. With time, each structural family might be addressed by treating

the three-dimensional displays in terms of coordinate point schemes or graph theory matri-

ces.

This is because these older methods lend themselves to the latest thoughts pertaining

to utilizing intentionally fuzzy coordinates

68,69

(e.g., x ± x', y ± y', and z ± z' for each atomic

point within a molecular matrix wherein the speciﬁ ed variations can be derived intelligently

from the composite of aforementioned computational and experimental approaches). Al-

ternatively, the fuzzy strategy might become better deployed during the searching routines,

or perhaps both knowledgeably fuzzy data entry and knowledgeably fuzzy data searching

engines handled, in turn, by fuzzy hardware

will ultimately best identify the correlations

that are being sought in any given search paradigm of the future. It should be noted, how-

ever, that for the fuzzy types of structural treatments, queries will be most effective when

the database has become large enough to rid itself statistically of the additional noise that

such fuzziness will initially create.

FUTURE PROSPECTS 283

284 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

Regardless of how they are evolved exactly, what this section points to is that ultimately,

the chemical structural databases of the future will probably have several tiers

of organized

chemical and conformational information available which can be mined distinctly according

to the speciﬁ ed needs of a directed searching scheme while still being able to be mixed com-

pletely within an overall relational architecture such that undirected knowledge-generating

mining paradigms can also be undertaken.

72–77

Certainly, simple physicochemical data will

need to be included among the parameters for chemical structure storage. Similarly, search-

ing engines will need to allow for discrete substructure queries as well as for assessing over-

all patterns of similarity and dissimilarity

78–85

across entire electronic surfaces.

It can be noted that it is probably already feasible to place most of the clinically used

drugs into a structural database that could at least begin to approach the low to midtier

levels of sophistication because considerable portions of such data and detail are prob-

ably already available within the literature, even if it is spread across a variety of technical

journals for each drug. On the other hand, it should also be clear that an alternative strategy

will be needed to handle the mountains of research compounds associated with a single

HTS survey. Taken together, the present discussions suggest that we have a long way to go

toward achieving the aforementioned tiers of conformational treatments when dealing with

large databases and applying them to the process of drug discovery. Nevertheless, because

of the importance of chemoinformatics toward understanding, fully appreciating, and ul-

timately implementing bioinformatics along the practical avenues of new drug discovery,

it can be imagined that future structural ﬁ elds within databases, including those associated

with drug metabolism, may be handled according to the following scenario, as summarized

from the ongoing discussion in this section and as also conveyed in Fig. 9.2.

For optimal use in the future, it is suggested that several levels of sophistication will be

built into database architectures so that a simple two-dimensional format can be input im-

mediately. Accompanying the simple two-dimensional structure ﬁ eld would be a ﬁ eld for

experimentally obtained or calculated physicochemical properties. Although this simple

starting point would lend itself to some types of rudimentary structure-related searching

paradigms, the same compound would then gradually progress by further conformational

study through a series of more sophisticated chemical structure displays. As mentioned ear-

lier, x-ray, NMR, and computational approaches toward considering conformation will be

deployed for real compounds, whereas virtual compound libraries and databases will rely

on computational approaches or on knowledgeable extrapolation from experimental data

derivable by analogy to structures within overlapping similarity space. Eventually, struc-

tures would be manipulated to a top tier of structural information. This tier might portray

the population ratios within a conformational family for a given structure entry expressed

as both distinct member and averaged electrostatic surface potentials wherein the latter

can be further expanded so as to display their atomic orientations by fuzzy graph theory

for fuzzy three-dimensional coordinate systems. Thus, at this point it might be speculated

that an intelligently fuzzy coordinate system could eventually represent the highest level

of development for tomorrow’s three-dimensional quantitative SAR

86,87

-based searching

paradigms. Furthermore, it can be imagined that this top tier might actually be developed

in triplicate for each compound: that is, one informational ﬁ eld for the environmentally

unbiased structural entries, another involving several subsets associated with known or

suspected interactions with the biological realm, and a third for tracking conformational

families when raised by about 5 to 10 kcal/mol in energy. Finally, conformational and en-

ergetic considerations pertaining to a compound’s movement between its various displays

can also be expected to be further reﬁ ned so as ultimately to allow future characterization

and searching of the dynamic chemical events that occur at the drug–biological interface

(e.g., modes and energies of docking trajectories and their associated molecular motions

relative to both ligand and receptor/active site). This top tier is extremely valuable for fully

understanding the interactions of interest to drug metabolism, a situation made apparent by

the large amount of effort already going on today in this area.

88–93

Similarly, chemical structure search engines of the future will probably be set up so that

they can be undertaken at several tiers of sophistication, the more sophisticated requiring

more expert-based inquiries and longer search times for the correlations to be assessed. A

reasonable hierarchy for search capability relative to the structural portion of any query

might become (1) simple two-dimensional structure with and without physicochemical

properties; (2) three-dimensional structure at incremented levels of reﬁ nement; (3) two-

dimensional and three-dimensional substructures; (4) molecular similarity–dissimilarity

indices; (5) fuzzy coordinate matrices; (6) docked systems from either the drug’s or the re-

ceptor or active site’s view at various levels of speciﬁ able precision; and ﬁ nally, in the more

distant future, (7) energy paths for a drug’s movement across various biological milieu,

2D Input of Structure

and Physicochemical

Properties

Environmentally

Unbiased

3D Structures

· X-Ray

· Computational

Conformational

Family Members

Depicted Individually

and as Averaged

Graph Theory or

Fuzzy 3D

Coordinate System

Composite

Members and

Averaged Composite

at Incremental

Increases of Energy

Environmentally

Biased 3D Structures

· X-Ray

· NMR

· Computational

(Docked Molecules)

Efficacy

Surfaces

Various

ADMET-Related

Surface Models

Specified Conformations

Depicted as Graph

Theory or Fuzzy 3D

Coordinate Systems

(critical interactions

likely to have low

tolerance for variability)

Track Energies For

Movements Between

Various Conformational

Family Members

Figure 9.2 Handling chemical structures within databases of the future. This ﬁ gure depicts the

quick entry and gradual maturation of structures. Structure entry would be initiated by a simple two-

dimensional depiction that is gradually matured in conformational sophistication through experi-

mental and computational studies. Note that structures would be evolved in both an unbiased and in

several environmentally biased formats. The highest structural tier represents tracking and searching

the energies required for various conformational movements that members would take when going

from one family to another. Search engines, in turn, would also provide for a variety of ﬂ exible query

paradigms involving physical properties with both full and partial (sub)structure searching capabili-

ties using pattern overlap/recognition, similarity–dissimilarity, CoMFA, and so on.

FUTURE PROSPECTS 285

286 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

including the trajectories and molecular motions associated with drug-receptor/active site

docking scenarios. Emphasizing informatics ﬂ exibility, this type of approach, where data

entry can occur rapidly for starting structure displays and then be gradually matured to

more sophisticated displays as conformational details are accurately accrued, coupled with

the ability to query at different levels of chemical complexity and visual displays

at any

point during database maturation, should allow for chemically creative database mining

strategies to be effected in the near term as well as into the more distant future.

The second major hurdle toward more effectively deploying drug metabolism databases

involves the critical need to better deﬁ ne the correlation of preclinical data with what can

then be expected to occur in humans. In other words, a ready mechanism or protocol needs

to be available that can be used to provide for validation of a given HTS model relative to

the model’s contribution toward projecting the eventual clinically observed composite of

variously parameterized HTS events. In this regard, it can be emphasized that the amount

of metabolism data produced at the clinical level is actually quite small compared to the

rather large amount of metabolism data available from preclinical studies. This situation

has prompted an effort by our laboratories to establish a speciﬁ c human drug metabolism

database.

This relational database coupled with substructure searching capability should

be derived solely from human clinical results that are continually being contributed by

all interested practitioners. In turn, the database should be available on the WWW via a

nonproﬁ t mechanism. Thus, the operation and utility of this metabolism database might be

imagined to be somewhat similar to that of the Cambridge x-ray collection, the protein da-

tabank, or to some of the newer gene-related informational resources that have been made

available over the WWW on a nonproﬁ t basis.

The sheer size of such a common database can overcome the anecdotal nature of the nu-

merous smaller collections presently being held individually by the big pharma members

of the pharmaceutical enterprise. Importantly, the database’s growing size will eventually

allow it to be utilized to develop more accurate and meaningful human SMRs. Selected

aspects of the overall SMRs, in turn, can still be applied by individuals in a proprietary

fashion to better predict the metabolic fate of their own, speciﬁ c structural motifs. Simi-

larly, a speciﬁ c human metabolism database would support rather than compete with the

ongoing activities of the existing metabolism database vendors. The latter have already

collected data from numerous species and various testing paradigms, all of which will still

be very much required as critical road maps during new drug development for quite some

time. Finally, and perhaps most important, the assembly of this type of database may be the

only way to assess and validate the actual utility of the ongoing explosion of biochemical

and in vitro metabolism data and HTS techniques presently being directed toward resolv-

ing metabolism issues at the earliest possible stages of drug discovery. The beneﬁ ts of such

a database are summarized in Table 9.6.

TABLE 9.6 Utility of a Human Xenobiotic Metabolism Database

•

Will be available on the WWW via a nonproﬁ t format

•

Will allow explicit structure searching of standards selected to validate proprietary drug

metabolism screens

•

Will allow substructure searching to identify analogous metabolic occurrences within humans

relative to proprietary compounds undergoing drug development

•

Will have large number of biotransformation entries so that statistically derived probability

assessments can be made about all metabolic possibilities

The informational ﬁ elds being constructed within this database are shown in Fig. 9.3,

which also depicts the database’s overall architecture. The treatment of three-dimensional

chemical structures has already been outlined in Fig. 9.2, which depicts the progression for

maturing structural entries according to the prior discussion.

9.5 SUMMARY

The gradual accumulation of drug metabolism studies has afforded a vast ﬁ eld of data

which offers the potential to be used for predicting the metabolic outcomes of new drug

candidates. Five major expeditions have ventured into this ﬁ eld to provide maps of vary-

ing detail which are available commercially as expert systems or databases with chemical

structure searching capabilities. An analysis of several case studies indicates that attempts

to predict metabolic outcomes for new structures by using these types of databases have

been only marginally successful. The reasons for this shortcoming include both biologi-

cal and chemical factors. Some of the biological issues include species and phenotypic

variation, as well as how to properly interrelate PBPK types of parameterization and HTS

screening results so as to better reﬂ ect the intact human situation. The major chemical is-

sues include a fundamental question about how much of a structure should be included

within a metabophore during SMR assessments, and the long-standing issue of how to

accurately derive three-dimensional structure. The discussions within this chapter have

considered these issues and have suggested some new approaches that could enhance the

utility of future chemical structure–biological information databases in general. In par-

ticular, the suggestions may also be useful for the speciﬁ c assembly of drug metabolism

databases that might partner in a synergistic manner with HTS metabolism data acquisition

methodologies.

For the biological issues it is imperative that a human drug metabolism database be

made available as a standard so that all types of preclinical sets of data from whatever type

Metabolic Probes

Clinical

Data

Chemical

Structure

Data

Phenotype

Data

Global Network

of Investigators

Existing

Literature

Data

Three-Dimensional

X-Ray; NMR;

Computational

Physicochemical

Properties

Big Pharma

Gov. Regulatory

Agencies, e.g. FDA

Existing

Literature

Data

Gene/Expression

Based Markers

Directed And

Relational

Database

Mining And

Knowledge

Discovery

Paradigms

Figure 9.3 Informational ﬁ elds to be included in the human drug metabolism database. This data-

base should be available on the WWW via a nonproﬁ t format. A ready mechanism should be made

available to continually receive human (clinical) metabolism data from any source. However, actual

data entry will have to be monitored for quality control via the database’s maintenance organization.

SUMMARY 287

288 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

of theoretical model, in vivo or in vitro experimental model, or parameterized HTS assay

can then be monitored for their predictabilities in statistical terms if not actually validated

within a more traditional format for their direct utilities. Similarly, in terms of chemical

issues, it has been suggested that although two-dimensional representations may constitute

a practical starting point for the input of structures, it is imperative that methods be evolved

to mature these displays into accurate representation of the relevant three-dimensional

conformational families. Speciﬁ c approaches toward the construction of a commonly held

human drug metabolism database and for the handling of three-dimensional chemical

structure have been elaborated herein. Both approaches are presently being explored within

our laboratories.

REFERENCES AND NOTES

Tables 9.1 and 9.2 represent a compilation of information taken largely from the drug metabo-

lism sections of three long-standing textbooks in the areas of medicinal chemistry, pharmacol-

ogy, and toxicology, respectively: (a) D. Williams. Drug metabolism, in Medicinal Chemistry,

4th ed., W. Foye, T. Lemke, and D. Williams (Eds.), Williams & Wilkins, Baltimore, pp. 83–140

(1995). (b) L. Z. Benet (Ed.). General principles, Sec. I (5 chapters), in Goodman & Gilman’s

The Pharamacological Basis of Therapeutics, 9th ed., J. G. Hardman, L. E. Lombard, P. B.

Molino, R. W. Radon and A. G. Gilman (Eds.), McGraw-Hill, New York, pp. 1–101 (1996).

contributing authors. Disposition of toxicant, Unit 2 (3 chapters), in Casarett & Doull’s Toxicol-

ogy: The Basic Science of Poisons, 5th ed., C. D. Klaassen, M. O. Amdur, and J. Doull (Eds.),

McGraw-Hill, New York, pp. 89–198 (1996).

This phrase is being considered for adoption by the International Union of Pure and Applied

Chemistry’s nomenclature working party on drug metabolism terms chaired by P. W. Erhardt.

T. Su, Z. Bao, Q-Y. Zhang, T. Smith, J-Y. Hong, and X. Ding. Human cytochrome P450

CYP2A13: predominant expression in the respiratory tract and its high efﬁ ciency metabolic

activation of a tobacco-speciﬁ c carcinogen, 4-(methylnitrosamino)-1-)3-pyridyl)-1-butanone.

Cancer Res., 60, 5074–5079 (2000).

Nomenclature for selected metabolizing enzymes. (a) Cytochrome P450s (CYPs): family, a sub-

set of the CYP superfamily sharing at last 40% amino acid sequence homology denoted by an

Arabic numeral after the uppercase letters CYP, (e.g., CYP3); subfamily, a subset of the CYP

superfamily sharing at lest 55% homology denoted by an Arabic uppercase letter after the family

designation (e.g., CYP3A); isoform, an individual form of CYP denoted by adding a second Ara-

bic numeral after the subfamily designation (e.g., CYP3A4). (b) N-Acetyl transferases (NATs)

are divided into two major families (i. e., 1 and 2), for which there are also many individually

displayed polymorphic versions. (c) Uridinedisphosphoglucuronosyl transferases (UGSs) are di-

vided into two major families (i. e., 1 and 2), based on having at least 50% amino acid sequence

homology followed by additional subfamily designations such as A and B, all of which also have

many individually displayed additional polymorphic versions (e.g., UGT1A1).

Recent results from this laboratory: e.g., J. Liao. Development of Soft Drugs for the neonatal

population, M. S. thesis, Univ. Toledo (2000).

Drug Metabolism: Databases and High Throughput Testing During Drug Design and Develop-

ment, IUPAC, Chapel Hill, NC (1999).

P. Erhardt. Drug metabolism data: past and present status, Med. Chem. Res., 8, 400–421 (1998).

National Academy of Sciences (U.S.) Committee. Problems of drug safety, Workshops (1966–

1968).

E. Boyland and J. Booth. The metabolic fate and excretion of drugs, Annu. Rev. Pharmacol., 2,

129–142 (1962).

R. T. Williams. Detoxication mechanisms in man, Clin. Pharm. Ther., 4, 234–254 (1963).

J. R. Gillette. Metabolism of drugs and other foreign compounds by enzymatic mechanisms,

Prog. Drug Res., 6, 11–73 (1963).

R. T. Williams, Detoxication Mechanisms: The Metabolism and Detoxication of Drugs, Toxic

Substances and Other Organic Compounds, Wiley, New York (1959).

T. B. Binns. Absorption and Distribution of Drugs, Williams & Wilkins, Baltimore (1964).

B. N. LaDu, H. G. Mandel, and E. L. Way. Fundamentals of Drug Metabolism and Drug Disposi-

tion, Williams & Wilkins, Baltimore (1972).

J. W. Bridges and L. F. Chasseaud. Progress in Drug Metabolism, Vol. 1, Wiley, New York

(1976).

The Chemical Society/The Royal Society of Chemistry (London), Foreign Compound Metabo-

lism in Mammals, Vols. 4–6, Specialist Periodical Reports (1977, 1979, and 1981).

D. R. Hawkins. Biotransformations: A Survey of the Biotransformations of Drugs and Chemicals

in Animals, Vols. 1–8, Royal Society of Chemistry, London (1989–1995).

R. V. Smith, P. W. Erhardt, and S. W. Leslie. Microsomal O-demethylation, N-demethylation

and aromatic hydroxylation in the presence of bisulﬁ te and dithiothreitol, Res. Commun. Chem.

Pathol. Pharmacol., 12, 181–184 (1975).

D. E. Atkinson. Cellular Energy Metabolism and Its Regulation, Academic Press, New York

(1977).

M. L. Spann, K. C. Chu, W. T. Wipke, and G. Ouchi. Use of computerized methods to predict

metabolic pathways and metabolites, J. Environ. Pathol. Toxicol., 2, 123–130 (1978).

T. H. Varkony, D. H. Smith, and C. Djerassi. Computer-assisted structure manipulation: studies

in the biosynthesis of natural products, Tetrahedron, 34, 841–852 (1978).

G. I. Ouchi. Computer-assisted prediction of plausible metabolites of xenobiotic compounds,

Univ. Micro. Int., Univ. California, Santa Cruz, CA (1978).

L. Goldberg. Structure–Activity Correlation as a Predictive Tool in Toxicology, Hemisphere,

New York (1983).

J. F. Tinker and H. Gelernter. Computer simulation of metabolic transformation, J. Comput.

Chem., 7, 657–665 (1986).

M. L. Mavrovounitis and G. Stephanopoulous. Computer-aided synthesis of biochemical path-

ways, Biotechnol. Bioeng., 36, 1119–1132 (1990).

S. Barcza, L. A. Kelly, and C. D. Lenz. Computerized retrieval of information on biosynthesis

and metabolic pathways, J. Chem. Inf. Comput. Sci., 30, 243–251 (1990).

F. Darvas. Metabol expert: an expert system for predicting the metabolism of substances, in

QSAR in Environmental Toxicology II, K. L. Kaiser (Ed.) Reidel Co., Dordrecht, The Nether-

lands (1987), pp. 71–81.

F. Darvas. Predicting metabolic pathways by logic programming, J. Mol. Graph., 6, 80–86

(1988).

Commercially available from ComGenex, Inc.

G. Klopman, M. Dimayuga, and J. Talafous. META. 1. A program for the evaluation of metabolic

transformation of chemicals, J. Chem. Inf. Comput. Sci., 34, 1320–1325 (1994).

J. Talafous, L. M. Sayre, J. J. Mieyal, and G. Klopman. META. 2. A dictionary model of mam-

malian xenobiotic metabolism, J. Chem. Inf. Comput. Sci., 34, 1326–1333 (1994).

G. Klopman and M. Tu. Meta. 3. A genetic algorithm for metabolic transform priorities optimiza-

tion, J. Chem. Inf. Comput. Sci., 37, 329–339 (1997).

Commercially available from Multicase, Inc.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

REFERENCES AND NOTES 289

290 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

E. M. Gifford, M. A., Johnson, and C.-C. Tsai. Using the xenobiotic metabolism database to rank

metabolic reaction occurrences, Mol. Connect. MDL Newsl. Commun. Customers, 13, 12–13

(1994).

The metabolite database is available from MDL Information Systems, Inc.

J. Hayward. Synopsys metabolism database, in Drug Metabolism: Databases and High Through-

put Testing During Drug Design and Development, P. Erhardt (Ed.), IUPAC, Chapel Hill, NC,

pp. 281–288 (1999).

Commercially available from Accelrys, Ltd.

N. Greene. Knowledge based expert systems for toxicity and metabolism prediction, in Drug

Metabolism: Databases and High Throughput Testing During Drug Design and Development,

P. Erhardt (Ed.), IUPAC, Chapel Hill, NC, pp. 289–296 (1999).

Commercially available from Lhasa, UK.

K. A. Bachmann and R. Ghosh. The use of in vitro methods to predict in vivo pharmacokinetics

and drug interactions, Curr. Drug Metabol., 2, 299–314 (2001).

J. N. Smith. Comparative Biochemistry of Detoxiﬁ cation, in Comparative Biochemistry,

M. Florkin and H. S. Mason (Eds.), Academic Press, New York, pp. 403–457 (1964).

G. Mulder. Conjugation Reactions in Drug Metabolism, Taylor & Francis, London (1990).

K. W. DiBiasio. In vitro testicular xenobiotic metabolism of model substrates in rats, mice, mon-

keys, and humans, Univ. Micro. Int., Univ. California. Davis, CA (1989).

J. Baron, J. M. Voigt, T. B. Whitter, T. T. Kawabata, S. A. Knapp, F. P. Guengerich, and W. B.

Jakoby. Identiﬁ cation of intratissue sites for xenobiotic activation and detoxication, Adv. Exp.

Med. Biol., 197, 119–144 (1986).

K. Bachmann, D. Pardoe, and D. White. Scaling basic toxicokinetic parameters from rat to man,

Environ. Health Perspect., 104, 400–407 (1996).

J. Sarver, D. White, P. Erhardt, and K. Bachmann. Estimating xenobiotic half-lives in humans

from rat data: inﬂ uence of log P, Environ. Health Perspect., 105, 1204–1209 (1997).

S. Hayashi, J. Watanabe, and K. Kawajiri. Genetic polymorphisms in the 5'-ﬂ anking region

change transcriptional regulation of the human cytochrome P450IIE 1 gene, J. Biochem. (Tokyo),

110, 559–565 (1991).

H. Sugimura, N. H. Caporaso, G. L. Shaw, R. V. Modali, F. J. Gonzalez, R. N. Hoover, J. H.

Resau, B. F. Trump, A. Weston, and C. C. Harris. Human debrisoquine hydroxylase gene poly-

morphisms in cancer patients and controls, Carcinogenesis, 11, 1527–1530 (1990).

S. Kato, M. Onda, N. Matsukura, A. Tokunaga, T. Tajiri, D. Y. Kim, H. Tsuruta, N. Matsuda,

K. Yamashita, and P. G. Shields. Cytochrome P4502E1 (CYP2E1) genetic polymorphism in a

case–control study of gastric cancer and liver diseases, Pharmacogenetics, 5, S141–S144 (1995).

K. Tominaga, Y. Koyama, M. Sasagawa, M. Hiroki, and M. Nagai. A case–control study of stom-

ach cancer and its genesis in relation to alcohol consumption, smoking and familial cancer his-

tory, Jpn. J. Cancer Res., 82, 974–979 (1991).

S. Borman. Combinatorial chemistry. C & E News, Feb. 24, 43–62 (1997).

F. K. Brown. Chemoinformatics: what is it and how does it impact drug discovery, Annu. Rep.

Med. Chem., 33, 375–384 (1998).

M. Hann and R. Green. Chemoinformatics: a new name for an old problem, Curr. Opin. Chem.

Biol., 3, 379–383 (1999).

P. Ertl, W. Miltz, B. Rohde, and P. Selzer. Web-based chemoinformatics for bench chemists,

Drug Discov. World, 1(2), 45–50 (2000).

ChemFinder at www.chemﬁ nder.com; Chem Navigator at www.chemnavigator.com; Info Chem

GmbH at www.infochem.com; Institute for Scientiﬁ c Info. at www.isinet.com; MDL Informa-

tion Systems at www.mdli.com; MSI at www.msi.com; Oxford Molecular at www.oxmol.com;

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Pharmacopeia at www.pharmacopeia.com; QsarIS at www.scivision. com/qsaris.html; Synopsys

Scientiﬁ c Systems at www.synopsys.co.uk; Tripos at www.tripos.com.

Personal communications while attending several recent workshops directed toward expediting

drug development.

G. Karet. Multiplicity, Drug Discov. Dev., Jan., 28–32, (2001).

Personal communications while attending several recent scientiﬁ c conferences in the area of

medicinal chemistry.

ChemPen 3D at http://home.ici.net/~hfevans/chempen.htm; CS Chem Draw and CS Chem Draw

3D at www.camsoft.com; Corina at www.mol-net.de.

A. Nezlin. Chem Draw Ultra 5.0: an indispensable everyday tool for an industrial organic chem-

ist, Chem. News. Commun.. 9, 24–25 (1999).

M. Endres. High-resolution graphics combined with computational methodology make 3D molec-

ular modeling exciting, colorful and downright sexy, Today’s Chem. at Work, Oct., 30–44 (1992).

G. Karet. Software promotes targeted design. Drug Discov. Dev., Aug./Sept., 42–48 (2000).

QUANTA (system for determining three-dimensional protein structure from x-ray data) and FE-

LIX (system for determining three-dimensional protein structure from NMR data) available from

MSI-Pharmacopeia (www.msi.com).

W. J. Hehre, L. Radom, P. v. R. Schleyer, and J. A. Pople. Ab Initio Molecular Orbital Theory,

Wiley, New York (1986).

A. D. Becke. Density-functional thermochemistry. III. The role of exact exchange, J. Chem.

Phys., 98, 5648–5652 (1993).

(a) B. C. Oostenbrink, J. W. Pitera, M. van Lipzig, J. H. Meerman, and W. F. van Gunsteren.

Simulations of the estrogen receptor ligand-binding domain: afﬁ nity of natural ligands and xe-

noestrogens, J. Med. Chem., 43, 4594–4605 (2000). (b) N. Wu, Y. Mo, J. Gao, and E. F. Pai.

Electrostatic stress in catalysis: structure and mechanism of the enzyme orotidine monophos-

phate decarboxylase, Proc. Natl. Acad. Sci. USA, 97, 2017–2022 (2000). (c) A. Warshel, M.

Strajbl, J. Villa, and J. Florian. Remarkable rate enhancement of orotidine 5'-monophosphate

decarboxylase is due to transition-state stabilization rather than to ground-state destabilization,

Biochemistry, 39, 14728–14738 (2000).

J. Devillers and A. T. Balaban (Eds.). Topological Indices and Related Descriptors in QSAR and

QSPR, Gordon and Breach Science, Amsterdam (1999).

C. Wrotnowski. Counting on computational intelligence, Mod. Drug Discov., Nov./Dec., 46–55

(1999).

C. M. Henry. Protein structure by mass spec, Chem. Eng. News, Nov. 27, 22–26 (2000).

E. K. Wilson. Quantum computers, Chem. Eng. News, Nov. 6, 35–39 (2000).

B. Ladd. Intuitive data analysis: the next generation Rx. Mod. Drug Discov., Jan./Feb., 46–52

(2000).

V. Venkatsubramanian. Computer-aided molecular design using neural networks and genetic al-

gorithms, in Genetic Algorithms in Molecular Modeling, J. Devillers (Ed.), Academic Press,

New York (1996).

A. Globus. Automatic molecular design using evolutionary techniques, Nanotechnology, 10(3),

290–299 (1999).

T. Studt. Data mining tools keep you ahead of the ﬂ ood, Drug Discov. Dev., Aug./Sept., 30–36

(2000).

R. Resnick. Simpliﬁ ed datamining, Drug Discov. Dev., Oct., 51–52 (2000).

J. Jaen-Oltra, T. Salabert-Salvador, F. J. Garcia-March, F. Perez-Gimenez, and F. Tohmas-Vert.

Artiﬁ cial neural network applied to prediction of ﬂ uoroquinolone antibacterial activity by topo-

logical methods, J. Med. Chem., 43, 1143–1148 (2000).

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

REFERENCES AND NOTES 291

292 USING DRUG METABOLISM DATABASES DURING DRUG DESIGN AND DEVELOPMENT

Bio Reason at www.bioreason.com; Columbus Molecular Software (Lead Scope) at www.co-

lumbus-molecular.com; Daylight Chemical Information Systems at www.daylight.com; IBM

(Intelligent Miner) at www.ibm.com; Incyte (Life Tools and Life Prot) at www.incyte.com;

Molecular Applications (Gene Mine 3.5.1) at www.mag.com; Molecular Simulations at www.

msi.com; Oxford Molecular (DIVA 1.1) at www.oxmol.com; Pangea Systems (Gene World 3.5)

at www.pangeasystems.com; Pharsight at www.pharsight.com; SAS Institute (Enterprise Miner)

at www.sas.com; SGI (Mine Set 3.0) at www.sgi.com; Spotﬁ re (Spotﬁ re Pro 4.0 and Leads Dis-

cover) at www.spotﬁ re.com; SPSS (Clementine) at www.spss.com; Tripos at www.tripos.com.

M. A. Johnson, E. Gifford, and C.-C. Tsai. In Concepts and Applications of Molecular Similarity,

M. A. Johnson (Ed.), Wiley, New York, pp. 289–320 (1990).

E. Gifford, M. Johnson, and C.-C. Tsai. A graph-theoretic approach to modeling metabolic path-

ways. J. Comput.-Aided Mol. Des., 5, 303–322 (1991).

E. M. Gifford, M. A. Johnson, D. G. Kaiser, and C.-C. Tsai. Visualizing relative occurrences in

metabolic transformations of xenobiotics using structure–activity maps, J. Chem. Inf. Comput.

Sci., 32, 591–599 (1992).

E. M. Gifford, M. A. Johnson, D. G. Kaiser and C.-C. Tsai. Modeling the Relative metabolic

occurrence of alkyl-nitrogen bond cleavage using structure–reactivity maps, Xenobiotica, 25,

125–146 (1995).

E. M. Gifford. Applications of molecular similarity methods to visualize xenobiotic metabolism

structure–reactivity relationships, Ph.D. thesis, Univ. Toledo (1996).

K. M. Andrews and R. D. Cramer. Toward general methods of targeted library design: topomer shape

similarity searching with diverse structures as queries, J. Med. Chem., 43, 1723–1740 (2000).

T. Borowski, M. Krol, E. Broclawik, T. C. Baranowski, L. Strekowski, and M. J. Mokrosz. Ap-

plication of similarity matrices and genetic neural networks in quantitative structure–activity

relationships of 2- or 4-(4-methylpiperazino) pyrimidines: 5-HT

receptor antagonists, J. Med.

Chem., 43, 1901–1909 (2000).

E. A. Wintner and C. C. Moallemi. Quantized surface complementarity diversity (QSCD): a model

based On small molecule–target complementarity, J. Med. Chem., 43, 1933–2006 (2000).

M. Pastor, G. Cruciani, I. McLay, S. Pickett, and S. Clementi. Grid-independent descriptors

(GRIND): a novel class of alignment-independent three-dimensional molecular descriptors, J.

Med. Chem., 43, 3233–3243 (2000).

C. Gnerre, M. Catto, F. Leonetti, P. Weber, P.-A. Carrupt, C. Altomare, A. Carotti, and B. Testa.

Inhibition of monoamine oxidases by functionalized coumarin derivatives: biological activities,

QSARS, and 3D-QSARS, J. Med. Chem., 43, 4747–4758 (2000).

H. A. Carlson, K. M. Masukawa, K. Rubins, F. D. Bushman, W. L. Jorgensen, R. D. Lins, J. M.

Briggs, and J. A. McCammon. Developing a dynamic pharmacophore model for HIV-1 integrase,

J. Med. Chem., 43, 2100–2114 (2000).

E. K. Bradley, P. Beroza, J. E. Penzotti, P. D. Grootenhuis, D. C. Spellmeyer, and J.L. Miller.

A rapid computational method for lead evolution: description and application to α

-adrenergic

antagonists, J. Med. Chem., 43, 2770–2774 (2000).

M. Graffner-Nordberg, J. Marelius, S. Ohlsson, A. Persson, G. Swedberg, P. Andersson, S. E.

Andersson, J. Aqvist, and A. Hallberg. Computational predictions of binding afﬁ nities to di-

hydrofolate reductase: synthesis and biological evaluation of methotrexate analogues, J. Med.

Chem., 43, 3852–3861 (2000).

R. Garcia-Nieto, I. Manzanares, C. Cuevas, and F. Gago. Increased DNA binding speciﬁ city for

antitumor ecteinascidin 743 through protein–DNA interactions? J. Med. Chem., 43, 4367–4369

(2000).

A. Vedani, H. Briem, M. Dobler, H. Dollinger, and D. R. McMasters. Multiple-conformation and

protonation-state representation in 4D-QSAR: the neurokinin-1 receptor system, J. Med. Chem.,

43, 4416–4427 (2000).

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

M. L. Lopez-Rodriguez, M. J. Morcillo, M. Fernandez, M. L. Rosada, L. Pardo, and K.-J. Schaper.

Synthesis and structure–activity relationships of a new model of arylpiperazines. 6. Study of the

5-HT

/α

-adrenergic receptor afﬁ nity by classical Hansch analysis, artiﬁ cial neural networks,

and computational simulation of ligand recognition, J. Med. Chem., 44, 198–207 (2001).

R. Wedin. Visual data mining speeds drug discovery, Mod. Drug Discov., Sept./Oct., 39–47

(1999).

P. Erhardt. Epilogue, in Drug Metabolism: Databases and High-Throughput Testing During

Drug Design and Development, P. Erhardt (Ed.), IUPAC, Chapel Hill, NC, p. 320 (1999).

93.

94.

95.

REFERENCES AND NOTES 293

295

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

DISCOVERY OF THE ANTIULCER

DRUG TAGAMET

C. ROBIN GANELLIN

University College London

London, UK

10.1 HISTORICAL BACKGROUND

10.1.1 Prologue

Tagamet has the nonproprietary name (INN) cimetidine. At the time that it was discovered,

cimetidine represented a novel type of drug action that revolutionized the treatment of

peptic ulcer disease. It was the product of a rational approach to drug design, involving

a very close collaboration between pharmacologists and medicinal chemists. The phar-

macology involved a detailed analysis that ran counter to the general wisdom at the time,

and the chemistry was unusual in that there was no known lead molecule with the required

properties to build upon (neither natural product nor synthetic chemical compound). The

chemistry was also unusual in that it relied on a detailed application of physical–organic

chemistry to structure–activity analysis. The subject has become a classic textbook ex-

ample. To understand and appreciate why this discovery was unusual it is necessary to

consider the level of knowledge (or lack of it) prevailing at the time (in 1964–1972) the

research was carried out. Much of the time was spent in the usual fog that accompanies

research into the unknown where there is no precedent but only analogy on which to base

ideas. Now it all seems obvious and very logical, but at the time it was bedevilled with

controversy and uncertainty. Indeed, the work was conducted at the SmithKline & French

Research Institute in Welwyn Garden City, UK, and the lack of early progress led the top

American management in Philadelphia to order the closure of the research program after

the ﬁ rst three years. Fortunately, the UK scientists did not accede to this order.

296 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

10.1.2 Pharmacological Receptors

Natural transmitter substances such as the biogenic amines have provided a rich source

for the development of new drugs. They act at speciﬁ c sites on cells in the body that we

call receptors, and in so doing they stimulate the cells to produce a speciﬁ c response.

Usually, this involves a sequence of biochemical events that modulate the cell’s function.

The biogenic amines therefore act as chemical messengers to affect what the body does.

The remarkable thing is that the biogenic amines have a chemical individuality that is

speciﬁ cally recognized at their own receptor sites; they do not act at other receptor sites

because the sites discriminate among different messengers. This makes for a highly selec-

tive process. It follows that the design of compounds to act in competition with the bio-

genic amines for occupation of these sites is a potential way of attaining speciﬁ city of drug

action. It also appears that the receptor sites for a particular amine are not homogeneous.

Drugs have been discovered that differentiate among sites for the same amine, suggesting

different populations of amine receptors. This allows further scope for introducing selec-

tivity of drug action.

Now, in the twenty-ﬁ rst century we know that these receptor sites are protein molecules

on the cell membrane that are either ion channels or are coupled to enzymes involved

in phosphorylation which amplify the original stimulus. In the ﬁ rst half of the twentieth

century there was very little proof of the nature of the amine receptors. Indeed, for many

years they were the ﬁ gment of pharmacologists' imagination. Their presence was only

inferred from the nature of dose–response relationships obtained by relating the concentra-

tion of biogenic amine applied and its effect on the function of a tissue or organ. By 1950,

acetylcholine was considered to act at two types of receptor named after natural products

(nicotine and muscarine), adrenaline and noradrenaline had just been suggested to act at α

and β receptors, and histamine had recently been proposed to possibly act at a second type

of receptor. The pharmacological characterization of histamine receptors had been estab-

lished by the seminal work of Heinz Schild, who had used histamine and antihistamines as

the means to provide the general basis for classifying competitive reversible antagonism in

the decade 1945–1955 (recall the now well-known Schild plot). Subsequently (1966), he

classiﬁ ed the actions of histamine that were blocked by the antihistamine drugs as being

mediated by histamine type 1 (H

) receptors.

10.1.3 Peptic Ulcer Disease

Peptic ulceration is the most common disease of the gastrointestinal tract. It produces

considerable illness and pain, and it used to result in great economic loss to the patients

and their communities; it could even be fatal. It comprises duodenal and gastric ulcers

and affects large numbers of people who are otherwise relatively ﬁ t. Duodenal and gas-

tric ulcers are localized erosions of the mucous membrane of the duodenum or stomach,

respectively, which expose the underlying layers of the gut wall to the acid secretions of

the stomach and to the proteolytic enzyme pepsin. In the 1960s the cause of acute peptic

ulcer was not properly understood, but for many years the main medical treatment had

been aimed at reducing acid production, based on the hope that neutralizing gastric acid

would reduce its irritating effects and also reduce the efﬁ cacy of pepsin, and so allow

ulcers to heal.

Secretion of gastric acid by the parietal cells of the stomach is initiated by the thought,

sight, smell, or taste of food and is mediated by the autonomic nervous, system via the

vagus nerves, which provide parasympathetic innervation to the stomach and small intes-

tine; the neurotransmitter released by stimulation of the vagus is acetylcholine. Branches

of the vagus, innervating the antral region of the stomach, stimulate the release of the

peptide hormone gastrin from special gastrin-producing G cells. The presence of food in

the stomach further stimulates release of gastrin, which passes into the bloodstream and is

carried to the parietal cells, where it acts to stimulate them to secrete hydrochloric acid. In

addition to acetylcholine and gastrin, a third chemical secretagog, histamine, was known

to be involved (Scheme 10.1).

The relationship between the three secretagogues acetylcholine, gastrin, and histamine

has been a source of considerable controversy among physiologists for many years. When

it was found (in the late 1940s and early 1950s) that antihistamine drugs did not reduce

acid secretion, the role of histamine as a secretagog was placed in serious question. By

1964, when gastrin had been isolated (Gregory and Tracy, 1961) and sequenced (Gregory

et al., 1964) at Liverpool University, most gastric physiologists were convinced that gastrin

played the key role in the physiological control of gastric acid and considered histamine to

be unimportant (Johnson, 1971).

For many years the main medical treatment for peptic ulcers relied on the use of ant-

acids to neutralize the gastric acid, but when taken in sufﬁ cient quantities they may cause

unpleasant side effects. Anticholinergic drugs (to block the acetylcholine transmission) can

decrease gastric acid secretion, but their use in the treatment of peptic ulceration was lim-

ited by side effects such as dryness of the mouth, urinary retention, and blurred vision. The

alternative to drug treatment is surgery. This aims to cut out part of the acid secretory and

gastrin-producing regions of the stomach (e.g., partial gastrectomy) or to selectively cut the

branches of the vagal nerve (e.g., selective vagotomy) that supply the acid-secretory region.

This was a difﬁ cult and sometimes dangerous operation.

p-Glu-Gly-Pro-Trp-Leu-

[Glu]

-Ala-Tyr-Gly-

Trp-Met-Asp-Phe-NH

Gastrin

Parietal cell

HCl

Histamine

COCH

NMe

Acetylcholine

Scheme 10.1 Three chemical messengers stimulate the production of hydrochloric acid from

the gastric parietal cell. The formulas show acetylcholine cation, histamine monocation (the most

prevalent species at physiological pH 7.4), and human gastrin l.

HISTORICAL BACKGROUND 297

298 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

10.1.4 Search for New Antiulcer Drugs

With the background to treatment described above it is not surprising that most major phar-

maceutical companies set up research programs aimed at discovering antiulcer agents. For

many years the main approach was to induce ulcer formation in rats, or to stimulate pro-

duction of gastric acid in rats, and to screen compounds for their ability to protect against

ulcer formation or to inhibit acid production. The record of success was very poor. In the

1960s a more scientiﬁ c and rational approach was beginning to take shape. With increasing

understanding of the physiology of gastric acid secretion, several companies established

research programs to discover speciﬁ c inhibitors of the action of the then known chemical

messengers. In the UK, ICI Pharmaceuticals initiated a search for a gastrin antagonist, and

Pﬁ zer and SmithKline & French independently sought a histamine antagonist.

The two secretagogs, gastrin and histamine, therefore presented alternative targets for

inhibiting acid production. If one of them has a controlling inﬂ uence on acid secretion,

blocking it might succeed in controlling acidity, but since blocking one site still leaves two

others to act, it is by no means certain that such an approach will be successful. Furthermore,

as with acetylcholine, gastrin and histamine have other actions, and an agent that blocks their

effects on acid stimulation may also block other sites, leading to unacceptable side effects

(as with the anticholinergics). Thus, it can be appreciated that such an approach to drug

discovery is highly speculative, and one certainly has no reason to presume that it will be

successful in providing a new therapy. On the other hand, it does have a scientiﬁ c rationale

founded on working hypotheses that are capable of investigation and being put to the test. An

important part of this analysis is provided by the pharmacologist’s view of drug receptors.

10.2 SEARCH FOR AN H

-RECEPTOR HISTAMINE ANTAGONIST

10.2.1 Histamine Receptors

Histamine was discovered at the beginning of the twentieth century, and in 1910 Dale and

Laidlaw published their seminal work on the actions of histamine on smooth muscle and

blood pressure. Subsequent studies led to a view that histamine was a principal mediator

of inﬂ ammation and shock; eventually, Bovet (1950) and colleagues discovered antihista-

mines, and in the 1940s these new drugs were introduced for the treatment of allergic con-

ditions such as urticaria and hay fever. Pharmacological studies with antihistamines such

as mepyramine (1, Scheme 10.2) led to the view that they were reversible competitive an-

tagonists and that antagonism was surmountable. Schild identiﬁ ed pharmacological criteria

so that the antihistamines could be viewed as acting in competition with histamine for oc-

cupation of its speciﬁ c receptors sites. By 1948, studies on vascular tissues had shown that

the antihistamines could only reduce the intensity of the action of large doses of histamine

but did not totally abolish its vasodilator effects, and this led Folkow et al. (1948) to suggest

that there may be two types of histamine receptor, only one of which could be blocked by

antihistamines such as diphenhydramine (2, Scheme 10.2). The suggestion appeared to lay

dormant in the literature for many years.

The antihistamines were used by pharmacologists to explore the various actions of his-

tamine in different tissues. Several actions of histamine had been noted that could not be

speciﬁ cally antagonized by these drugs: for example, stimulation of gastric acid secretion

in the rat, cat, or dog; simulation of isolated atria of the guinea pig; inhibition of rat uterus

contractions. Some pointers to the differentiation of histamine receptors had been obtained

by considering the selectivity of action of agonists on these different tissue systems, and the

results led Ash and Schild to propose in 1966 that the actions of histamine blocked by the

antihistamine drugs characterized one type of histamine receptor, which they named the H

receptor. They suggested that other actions of histamine not speciﬁ cally antagonized were

probably mediated by other histamine receptors, but the characterization of these receptors

awaited the discovery of speciﬁ c antagonists; in the meantime they were to be regarded as

non-H

receptors.

10.2.2 Biological Approach to a Histamine Antagonist at Non-H

Receptors

The inability of the antihistamine drugs (H

-receptor antagonists) to inhibit histamine-

stimulated gastric acid secretion had been known for many years and there have been a

few published reports of concerted efforts to discover a speciﬁ c antagonist to this action

of histamine. In collaboration with the Eli Lilly Company in the United States, Grossman

et al. (1952) reported on an extensive study of compounds, chemically related to histamine,

that were examined for their action on acid secretion and also tested as possible inhibitors

of histamine stimulation, but Grossman did not uncover a histamine antagonist.

A similar analysis by Sir James Black led him to establish at the SmithKline & French

research Laboratories in Welwyn Garden City, in 1964, the test procedures needed to detect

antagonists of these other effects of histamine. It was hoped that the work would lead

to a new type of pharmacological agent with possible clinical utility: namely, a possible

means for selective pharmacological control of gastric acid secretion with the potential for

treating peptic ulcer disease. Compounds were tested for their ability to inhibit histamine-

stimulated gastric acid secretion in anaesthetized rats using a reﬁ ned Ghosh and Schild

(1958) preparation. Other researchers (Ash and Schild, 1966; Grossman et al., 1952; Lin

et al., 1962; Van den Brink, 1969) also examined close analogs of histamine for possible

antagonism of histamine-stimulated gastric acid secretion. None of these studies estab-

lished a histamine antagonist.

Since other types of inhibitors of gastric secretion could also act in this test, compounds

found to be active were also tested on isolated tissue systems to provide additional criteria

for speciﬁ c antagonism to histamine. Two in vitro test systems involving two different ani-

mal species were set up: histamine-induced stimulation of guinea-pig right atrium (which

continues to beat spontaneously in vitro because it contains the pacemaker and histamine in-

creases the rate of beating) and inhibition by histamine of evoked contractions of rat uterus.

It is worth noting that the atmosphere prevailing in gastroenterological science at that

time was strongly against the search for a histamine antagonist as a means of control-

ling gastric acid secretion. In the 1960s many researchers turned their attention to seeking

SEARCH FOR AN H

-RECEPTOR HISTAMINE ANTAGONIST 299

Scheme 10.2

NCH

N(CH

)

(1) Mepyramine

CHOCH

N(CH

)

(2) Diphenhydramine

300 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

speciﬁ c inhibitors of gastrin-induced acid secretion. The import of histamine was difﬁ cult

to prove and there was a widely held view that histamine had no place in the physiological

maintenance of gastric acid secretion. The unsuccessful effort by researchers at Eli Lilly in

the 1950s to ﬁ nd an antagonist of histamine-stimulated acid secretion added further to the

general feeling that the approach was “played out.”

10.2.3 Chemical Approach to an Antagonist: Generating a Lead

Once the problem of obtaining a competitive antagonist was posed in biological terms it

was necessary to consider how to approach it chemically. How could one obtain such a

compound? Where should one start, given no obvious lead compound? Nothing was known

chemically about the physiological site of action of histamine. Returning to ﬁ rst principles,

the structure of histamine was used as a chemical starting point. The view was taken that

since the search was for a molecule that would compete with histamine for its receptor site,

such a molecule would have to be recognized by the receptor and then bind more strongly

than histamine, but not trigger the usual response. It therefore seemed worthwhile to retain

in potential antagonist structures some chemical features of histamine to aid receptor rec-

ognition, and to include chemical groups that might assist the binding.

In this type of research one has to work by analogy. Considering other areas where other

antagonists had been developed by chemical means to assist binding (e.g., antimetabolites,

enzyme inhibitors, and other receptor systems, such as anticholinergic and antiadrenergic

drugs). The structure of histamine was therefore modiﬁ ed to alter its chemical properties

deliberately, while retaining some deﬁ nite aspect of its structure or chemistry. Some ex-

amples have been discussed elsewhere by Ganellin et al. (1976).

Many compounds were made, based on the structure of histamine. In the ﬁ rst four years

some 200 compounds were synthesized and tested, without providing a blocking drug.

The problem for the chemist is that there are too many possible compounds for synthesis.

Even small modiﬁ cations of the natural stimulant histamine introduce many variables. Two

hundred compounds may sound to be a rather low number by today’s standards of high-

throughput screening and parallel synthesis. It must be remembered, however, that most

of the compounds were selected to explore a possible mode of interaction. They were not

selected because there were suitable intermediates available commercially (i.e., ease of

synthesis was not allowed to dictate the medicinal chemistry), although when all other

things were equal, chemical accessibility was, of course, the determinant. Furthermore,

imidazole chemistry can be demanding and compounds had to be puriﬁ ed by ion-exchange

chromatography in aqueous solution.

During this period there developed considerable uncertainty about whether the receptors

for histamine in gastric acid secretion might not be accessible. Also during this period, all

seven isomers of monomethyl histamine had been synthesized and tested to explore where

methyl groups could be accommodated in the histamine molecule without loss of afﬁ nity.

There was considerable relief when it was found that 4-methylhistamine was selective for

stimulating acid secretion, in comparison with its effect in stimulating the ileum. Con-

versely, 2-methylhistamine showed some selectivity for the ileum (in comparison with his-

tamine). Here was evidence for the existence of at least two types of histamine receptors.

Toward the end of this time many doubts were expressed about whether it really would

prove possible to block the action of histamine on gastric acid secretion, and indeed, there

was considerable pressure within the company to abandon the project (see above). The

scientists involved in the project were, however, ﬁ rmly resolved to continue, and the test

system was reﬁ ned. It is very important to conduct research in such a way as to learn from

negative results. Even a list of inactive compounds is informative if they have been selected

for particular reasons. Having tested many compounds with lipophilic substituents without

seeing antagonism, the pharmacologists reexamined some of the early hydrophilic com-

pounds. One of these polar hydrophilic compounds showed some blocking activity. It was

very weak but it provided the vital lead. It was missed originally because this compound

also acted as a stimulant; in fact, it is a partial agonist. The compound is a histamine de-

rivative in which a guanidine group replaces the amino group in the side chain: namely,

-guanylhistamine (3) (Table 10.1).

10.2.4 Lead Optimization

The lead compound (3) was very weakly active, but within a few days an analogous com-

pound was retested and found to be more active: (S)-[2-(imidazol-4-yl)ethyl]isothiourea

(4) (Table 10.1). Still, a much more active compound was required.

An immediate question to be answered was whether activity was due to the presence of

the guanidine or isothiourea groups (amidines) per se or to the structural resemblance to

histamine. Structure–activity studies suggested that for these structures the imidazole ring

was important; antagonism did not appear to be a property of amidines in general. It was

also necessary to identify the particular chemical properties that conferred antagonist activ-

ity in order to make analogs of increased potency.

The amidine groups are strong bases and are protonated and positively charged at physi-

ological pH. Thus, the molecules resemble histamine monocation but also differ in several

ways; the amidinium group is planar (whereas the ammonium group of histamine is tetra-

hedral) and the positive charge is distributed over three heteroatoms. It was noted that the

distance between ring and terminal nitrogen is potentially greater than in histamine, and

(CH

)

(CH

)

Structure

Substituent Activity

X  NH 

X  S 

X  NH 

X  S

Z  SMe

Z  Me

Z  SMe 

Z  Me 

TABLE 10.1 Structure and Antagonist Activities of Some Simple

Imidazolylalkylisothioureas, Imidazoylalkylguanidines, and

Imidazoylalkylcarboxamidines

Tested for inhibition of histamine-stimulated gastric acid secretion in the lumen-perfused anaesthetized rat.

Results represented semiquantitively as ±, detectable; , ID

 500 µmol/kg; , ID

⬃ 200 µmol/kg; ,

 100 to 50 µmol/kg. ID

is the intravenous dose that reduces a near-maximal secretion to 50%.

SEARCH FOR AN H

-RECEPTOR HISTAMINE ANTAGONIST 301

302 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

that there are several nitrogen sites for potential interactions instead of one, thereby afford-

ing more opportunities for intermolecular hydrogen bonding.

It was envisaged that amidines might act as antagonists through additional binding being

contributed by the amidine group. A type of bidentate hydrogen bonding between ion pairs

can occur with an amidinium cation and an oxyacid anion (e.g., carboxylate, phosphate, or

sulfate), which might be part of the receptor protein. Structural variables therefore identiﬁ ed

for study initially were the amidine groups, amidino N-substituents (to search for hydrophobic

interactions as a contribution to afﬁ nity), side-chain length, and alternatives to imidazole.

Many analogs of compounds 3 and 4 were made, but most turned out to be less active;

at that time, unsubstituted imidazole appeared to be the best ring; alkyl substitution on the

amidine N gave inconsistent results, and lengthening the side chain gave another break-

through but threw up an apparent contradiction. For the guanidine structure, increasing the

chain length led to a compound (5) showing an increase in antagonist activity. However, for

the isothiourea, the reverse result was obtained; that is, increasing the chain length gave a

compound (6) of reduced antagonist activity (Ganellin, 1981).

Thus, although the ﬁ rst two compounds discovered [guanidine (3) and isothiourea (4)] ap-

peared to be closely related in structure in simply being isosteric nitrogen and sulfur analogs,

the results with the homologs suggested that the situation was more complex. In an attempt

to rationalize these differences, various related amidines were examined (Scheme 10.3). It

was found that the reversed isothioureas (7, 9) (side chain on N instead of S) resembled the

guanidines in chain-length requirements, as did carboxamidines (e.g., 8 and 10).

The apparent nonadditivity between structural change and biological effect posed a typi-

cal problem familiar to all practicing medicinal chemists: With so many structural variables

Scheme 10.3 Imidazolylalkylisothioureas, imidazolylalkylcarboxamidines, and imidazolylalkyl-

guanidines synthesized and tested as potential antagonists of histamine-stimulated gastric acid secre-

tion. The structures are shown as side-chain cations. (a) Isothioureas; (b) Reversed isothioureas; (c)

carboxamidines; (d) guanidines.

to study (e.g., ring, side-chain length, amidine system, amidine substituents), there are

many millions of structures incorporating different combinations of these variables, and

one cannot make and test them all. What, then, should govern the selection?

An essential feature of the discipline in medicinal chemistry is to ﬁ nd logical reasons

for deﬁ ning the boundary conditions for the selection of structures for synthesis. In the case

under study there was a continuous search for useful physicochemical models for studying

the chemistry of these compounds, and the inconsistencies in the structure–activity pattern

were used to challenge the model or to reexamine the meaning of the biological test results.

This dialogue, a search for self-consistency between the chemistry and biology, is vital to

new drug research where no precedent exists.

To explore structure–activity relationships further, it became desirable to increase the

side-chain length still more, but problems of chemical synthesis were experienced and new

synthetic routes were required. Exploration of amidines and substituents continued but

progress became very slow. The problem was that the compounds had mixed activities. In

the main they acted as both agonists and antagonists; that is, they appeared to be partial

agonists. This meant that although the compounds antagonized the action of histamine,

they were not sufﬁ ciently effective inhibitors of gastric acid secretion because of interfer-

ence through their inherent stimulatory activity. This appeared to impose a limitation on the

potential of this type of structure for providing antagonists.

10.2.5 Validating the Research Program

Thus, a critical stage was reached in the need for selectivity: to achieve a separation be-

tween agonist and antagonist activities. It seemed that these compounds might act as ago-

nists by mimicking histamine chemically, since like histamine, they have an imidazole

ring, and being basic amidines, the side chain at physiological pH is protonated and carries

a positive charge. It also seemed likely that these features would permit receptor recogni-

tion and provide binding for a competitive antagonist. This posed a considerable dilemma

because the chemical groups that appeared to be required for antagonist (blocking) activity

were the same groups that seemed to confer the agonist (stimulant) effect.

To separate these activities, the strongly basic guanidine group was replaced by non-

basic groups that although polar, would not be charged. Such an approach furnished ana-

logs that indeed were not active as agonists; however, the ﬁ rst examples were also not

active as antagonists. Eventually, one example, the thiourea derivative (11, SK&F 91581)

(Scheme 10.4), that did not act as a partial agonist exhibited weak activity as an antagonist.

Thioureas are essentially neutral in water because of the electron-withdrawing thiocarbonyl

group. Conjugation forces the nitrogen atoms into a planar form and limits the availability

of the nitrogen lone electron pairs, as in amides.

SEARCH FOR AN H

-RECEPTOR HISTAMINE ANTAGONIST 303

Scheme 10.4 Imidazolylalkylthioureas.

304

X CH

NHCNHMe

-Receptor Activities

In Vitro In Vivo

Atrium

(95% limits)

10

6

Uterus

(95% limits)

10

6

Acid Secretion

(µmol/kg)

Compound Structure

Number Trivial Name R

Burimamide (thiourea) H CH

S 7.8 (6.4–8.6) 6.6 (4.9–8.3) 6.1

Thiaburimamide H S S 3.2 (2.5–4.5) 3.2 (2.5–4.5) 5

Metiamide (thiourea) Me S S 0.92 (0.74–1.15) 0.75 (0.40–1.36) 1.6

Urea isostere Me S O 22 (8.9–65) 7.1 (1.6–30) 27

Guanidine isostere Me S



16 (8.1–32) 5.5 (2.8–13) 12

Nitroguanidine (isostere) Me S N–NO

1.4 (0.79–2.8) 1.4 (0.72–3.2) 2.1

Cimetidine (cyanoguanidine) Me S N–CN 0.79 (0.68–0.92) 0.81 (0.54–1.2) 1.4

TABLE 10.2 Structures and H

-Receptor Histamine Antagonist Activities of Burimamide, Metiamide, Cimetidine, and Isosteres

Activities determined against histamine simulation of guinea-pig right atrium in vitro. The dissociation constant (K

) was calculated from the equation K

 B /(x1), where x is

the respective ratio of concentrations of histamine needed to produce half-maximal responses in the presence and absence of different concentrations (B) of antagonist.

Activities determined against histamine inhibition of electrically evoked concentrations of rat uterus in vitro.

Activities as antagonists of histamine-simulated gastric acid secretion in the anaesthetized rat as indicated in footnote a of Table 10.1.

Problems of synthesizing higher homologous amines were solved by this time. The amine

with the four-carbon atom chain length was synthesized and further exploration revealed that

with this type of structure, extension of the alkylene side chain resulted in a marked increase

in antagonist potency. It was not until the side chain had been lengthened that the signiﬁ -

cance of the result with SK&F 91581 (11) became clear and the desired aim was achieved,

that is, a pure competitive antagonist without agonist effects. This compound (12, SK&F

91863) paved the way for an exploration of alkyl groups seeking a potential hydrophobic

interaction and led to the N-methyl analog, which was given the name burimamide (13).

Burimamide (13) was an extremely important compound. It was highly selective, showed

no agonist activity, and antagonized the action of histamine in a competitive manner on the

two in vitro non-H

systems, guinea pig atrium and rat uterus (Table 10.2). It fulﬁ lled the

criteria required for characterizing the existence of another set of histamine receptors, the

receptors. Thus, it allowed these tissue systems to be deﬁ ned as H

-receptor systems,

so burimamide was deﬁ ned as an H

-receptor antagonist. This discovery was announced

in Nature in 1972; the work had taken six years (Black et al., 1972). It provided the ﬁ rst

deﬁ nition of histamine H

receptors.

Burimamide also antagonized the action of histamine as a stimulant of gastric acid

secretion in the rat, cat, and dog, and it was the ﬁ rst H

-receptor antagonist to be investi-

gated in humans. Given intravenously, it blocked the action of histamine as a stimulant of

gastric acid secretion in humans, thereby conﬁ rming that burimamide behaves in humans

as it does in animals. Thus, burimamide validated the research program. However, its one

drawback was that it was not sufﬁ ciently active to be given orally. Thus, although buri-

mamide was selective enough to deﬁ ne H

receptors, it was not active enough to permit

proper drug development.

10.3 DEVELOPMENT OF A CLINICAL CANDIDATE DRUG

10.3.1 Dynamic Structure–Activity Analysis

Various ways to alter the structure of burimamide were examined in an attempt to increase

potency. One approach that proved successful resulted from two lines of exploration that

merged. Attempts were being made to overcome the problem of synthesizing the side chains

by inserting a thioether link. Meanwhile, a study was being made of the pK

characteristics

of burimamide since it was realized that burimamide in aqueous solution is a mixture of

many chemical species in equilibrium. At physiological pH there are three main forms of

the imidazole ring, three planar conﬁ gurations of the thioureido group (a fourth is theoreti-

cally possible but is disfavored by internal steric hindrance), and various trans and gauche

rotamer combinations of the side chain CH

–CH

bonds (Scheme 10.5). This means that at a

given instant only a small proportion of the drug molecules would be in a particular form.

The existence of a mixture of species leads one to question which may be biologically

active and whether altering drug structure to favor a particular species would alter drug

potency. This is a process of dynamic structure–activity analysis (DSAA) (Ganellin, 1981).

There are substantial energy barriers to interconversion between the species of burimam-

ide, so it is quite likely that a drug molecule presenting itself to the receptor in a form

unfavorable for drug–receptor interaction might diffuse away again before having time to

rearrange into a more favorable form. The relative population of favorable forms might

therefore determine the amount of drug required for a given effect.

DEVELOPMENT OF A CLINICAL CANDIDATE DRUG 305

306 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

The various species of burimamide do not inconvert instantaneously, but whereas the

rotamers of the side chain and thioureido groups are interconverted simply by rotation of a

C–C or C–N bond, interconversion of the ring forms probably involves a water-mediated

proton transfer. It was argued that if a molecule presents itself to the receptor with the ring

in an unfavorable form, it might not readjust unless there were suitably oriented water mol-

ecules (or other hydrogen donor–acceptors) present.

10.3.2 Imidazole Tautomerism and Sulfur Methylene Isosterism

The arguments above led to a study of imidazole tautomerism and the population of imid-

azole species. At physiological pH the main species are (Scheme 10.5) the cation (13c) and

= (CH

)

NHCNHMe

(a)

(13c)

(13a)

(13b)

(b)

(c)

(13d) (Z, E)

(13e) (Z, Z)

(13g) (E, Z)

(13f) (E, E)

(CH

)

Scheme 10.5 Burimamide species equilibria in solution: (a) imidazole ring (ionization and

tautomerism); (b) alkane chain (C—C bond rotation gives trans and gauche conformers); (c) thiourea

group (conﬁ gurational isomerism).

two uncharged tautomers (13a and 13b), and their populations were estimated qualitatively

from the electronic inﬂ uence of the side chain using pK

data and the Hammett equation

(Charton, 1965b):

where ρ is the Hammett reaction constant and σ

the Hammett substituent constant. For

burimamide, the ring pK

(7.25 at 37C) is greater than that of unsubstituted imidazole

(6.80), indicating that the side chain is mildly electron releasing. In contrast, for histamine

the ammonium ethyl side chain was seen to be electron withdrawing since it lowered the

of the imidazole ring (pK

 5.90). Thus, although both histamine and burimamide are

monosubstituted imidazoles, the structural similarity is misleading in that the electronic

properties of the respective imidazole rings are different.

If the active form of burimamide were tautomer 13a, the form most preferred for hista-

mine, increasing its relative population might increase activity; for example, incorporating

an electronegative atom into the antagonist side chain should convert it into an electron-

withdrawing group and favor species 13a. This would not be the only requirement for

activity, and it would be necessary to minimize disturbance to other biologically impor-

tant molecular properties such as stereochemistry and lipid–water interactions. For reason

of synthesis, the ﬁ rst substitution to be made was the replacement of a methylene group

(CH

) by the isosteric thioether linkage (S) at the carbon atom next but one to the

ring, to afford thiaburimamide (14) (Table 10.2), which was found to be more active as an

antagonist.

It was argued that further stabilization of tautomer 13a might be obtained by incorporat-

ing an electron-releasing substituent in the vacant 4(5) position of the imidazole ring. A

methyl group was selected since it was thought that it should not interfere with receptor

interaction, 4-methylhistamine having been shown to be an effective H

-receptor agonist.

This approach was successful, and introduction of a methyl group into the ring of the

antagonist furnished a more potent drug, which was named metiamide (15) (Black et al.,

1974) (Table 10.2). Metiamide represented a major improvement, being 10 times more po-

tent than burimamide in vitro and a potent inhibitor of stimulated acid secretion in humans.

It was investigated in patients and shown to produce a signiﬁ cant increase in the healing

rate of duodenal ulcers and marked symptomatic relief. However, of 700 patients treated,

there were a few cases of granulocytopenia (causing a reduction in the number of circulat-

ing white cells in the blood and leaving patients open to infection). Although reversible,

this severely limited the amount of clinical work, so that another compound was required

for clinical development.

10.3.3 Isosteres of Thiourea and the Discovery of Cimetidine

There now arose a truly critical issue for drug development. The question had to be faced

whether the granulocytopenia was due to the pharmacological properties of an H

-receptor

histamine antagonist. Remember, this was the ﬁ rst time that such a pharmacological agent

had been tested in the clinic. If H

receptors were involved in the generation of white blood

cells, this would clearly limit the usefulness of an H

-antagonist drug. Alternatively, the

problem might be due to a toxic effect of the drug structure, in which case it would be a

chemical problem, and therefore, in principle, it should be capable of a solution. One possi-

bility explored was that the granulocytopenia associated with metiamide was caused by the

a(R)

 pK

a(H)

 ρσ

DEVELOPMENT OF A CLINICAL CANDIDATE DRUG 307

308 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

thiourea group in the molecule. Exploration had continued with other possible structures,

in particular with alternatives to the thiourea group. Isosteric replacement of the thiourea

sulfur atom (S) of metiamide by carbonyl oxygen (O) gave the urea analog (16), but

this was much less active. Isosteric replacement by imino nitrogen (N) afforded the gua-

nidine (17), which interestingly, though charged, was not a partial agonist but a fairly active

antagonist. However, in vitro, the urea and guanidine isosteres were both approximately 20

times less potent than metiamide; therefore, other ways were investigated for removing the

positive charge on the guanidine derivative.

Guanidine basicity is markedly reduced by electron-withdrawing substituents, and

Charton (1965a) had demonstrated a high correlation between the inductive substituent

constant σ

and pK

for a series of monosubstituted guanidines. The cyano and nitro groups

are sufﬁ ciently electron withdrawing to reduce the pK

by over 14 units, to values  0.

The nitroguanidine (18) and cyanoguanidine (19) analogs of metiamide were synthesized

and found to be active antagonists comparable with metiamide (Durant et al. 1977). Of

these two compounds, the cyanoguanidine (19) is slightly more potent and was selected

for development (Brimblecombe et al., 1975, 1978), being given the nonproprietary name

cimetidine.

Cyanoguanidines exist predominantly in the cyanoimino form, and in cimetidine the cy-

anoimino group (NCN) replaces the thione (S) sulfur atom of metiamide. The similar

behavior of cimetidine and metiamide as histamine H

-receptor antagonists and the close

similarity in physicochemical characteristics of thiourea and cyanoguanidine permit the

description of thiourea and cyanoguanidine groups in the present context as bioisosteres

(Durant et al., 1977). Nitroguanidine may also be considered to be a bioisostere of thiourea

in this series of structures.

10.3.4 Cimetidine: A Breakthrough in the Treatment of Peptic Ulcer Disease

Cimetidine was shown to be a speciﬁ c competitive antagonist of histamine at H

receptors

in vitro and to be effective in vivo at inhibiting histamine-stimulated gastric acid secretion

in the rat, cat, and dog. The ID

values determined in the rat, cat, and dog were not sig-

niﬁ cantly different from each other. Cimetidine was also shown to be active when admin-

istered orally in the dog.

Cimetidine was also found to be an effective inhibitor of pentagastrin-stimulated acid

secretion. (Pentagastrin is a synthetic biologically active analog of gastrin which contains

the terminal four amino acid residues of gastrin: N-t-BOC-β-Ala-Trp-Met-Asp-Phe-NH

The ID

values indicated that the potency of cimetidine against pentagastrin-stimulated

secretion is very similar to its potency against histamine-stimulated secretion. This ﬁ nd-

ing demonstrated that gastrin and histamine are somehow linked in the gastric secretory

process, and the results ﬁ rmly established that histamine has a physiological role in gastric

acid secretion.

Cimetidine given orally at 1 to 1.2 g/day was shown to relieve symptoms and promote

healing of lesions in a majority of patients with peptic ulcer disease. Fortunately, cimeti-

dine did not cause granulocytopenia, thereby demonstrating the lack of any link with H

receptor antagonism. Cimetidine was marketed ﬁ rst in the United Kingdom in November

1976 and was marketed in the United States in November 1977; by 1979 it was sold in over

100 countries under the trademark Tagamet, representing all the major markets with one

important exception—it was not granted approval for use in Japan until 1982. Cimetidine

changed the medical management of peptic ulcer disease and became a very successful

product (Freston, 1982; Winship, 1978). In 1983 its annual worldwide sales reached nearly

$1 billion, and in several countries it was the leading prescription product in sales; indeed,

it was the ﬁ rst of the modern “blockbuster” medicines.

10.4 SUMMARY AND FURTHER OBSERVATIONS

In this discovery, the medicinal chemistry was very dependent on applications of physi-

cal–organic chemistry principles, at a time when this was rather unusual. In particular,

critically important contributions were made by considering pK

values and applications

of the Hammett equation. The design process also involved several examples of the use

of isosteric replacement. Retrospective structure–activity analysis has demonstrated the

existence of correlations with log P, dipole moments, and conformational analysis. The

clear message is that in medicinal chemistry it is necessary to think as an organic chemist

but not only about synthesis.

The long-term nature of pharmaceutical research and development and the need for

tenacity to continue in the face of considerable difﬁ culty and disappointment is well

illustrated by this case history of drug discovery (see Fig. 10.1). The research project was

initiated in 1964, and it took six years (to 1970) to obtain burimamide, which was used to

characterize pharmacologically histamine H

receptors, to verify the basic concept, and to

be studied in human volunteers. The next drug, metiamide, a more potent and orally active

compound, was investigated clinically, but its use was severely restricted by a toxic effect.

Cimetidine followed from metiamide: it was ﬁ rst synthesized in 1972 and made generally

available in 1977. An overall time of 13 years had elapsed!

It is self-evident that any account of drug discovery must be incomplete, and certainly

only a small proportion of the total studies made have been described in this chapter. Many

avenues examined during structure–activity analysis turned out to be ineffective, and since

in the main, these are not mentioned here, the net effect may be to make the work appear to

be more rational and more perceptive than is warranted. To limit the scale of the problem

for lead generation and optimization, the early work concentrated on imidazole derivatives.

SUMMARY AND FURTHER OBSERVATIONS 309

Program starts

First lead antagonist

Burimamide

Metiamide

1964 66 68 70 72 74 76 78

U.S.A.

U.K.

Into volunteers

Cimetidine

Figure 10.1 Thirteen years to discover and develop cimetidine and make it generally available for

therapeutic use.

310 DISCOVERY OF THE ANTIULCER DRUG TAGAMET

However, it was soon demonstrated that other heterocycles such as pyridine and thiazole

could be used effectively in place of imidazole. Later, researchers at Glaxo demonstrated

that it was not even necessary to have a nitrogen heterocycle, and they synthesized a furan

derivative, ranitidine (20), which became the second histamine H

-receptor antagonist to

be introduced clinically (under the trademark Zantac) for treatment of peptic ulcer disease

(Bradshaw et al., 1982) and eventually became the second modern blockbuster medicine

(Scheme 10.6). This was followed by three other H

-receptor antagonist products: nizati-

dine from Eli Lilly (Lin et al., 1983), famotidine from Yamanouchi (Takeda et al., 1982),

and roxatidine from Teikoku.

Post Scriptum The treatment of diseases is developing continuously. The last 40 years

have seen notable changes in gastroenterology. Until Tagamet became available, the

medicinal management of peptic ulcer disease and related problems of hypersecretion of

acid such as dyspepsia were not really effective. Surgery had become the mainstay for

treatment. Tagamet changed this situation dramatically and offered physicians a medicine

that acted pharmacologically to “turn off the tap” of acid secretion. Blocking histamine H

receptors gave physiologists a tool for studying the physiology of gastric acid secretion.

This led to the development of omeprazole, the ﬁ rst of the H



ATPase inhibitors, which

represented a new class of antiulcer product, the proton-pump inhibitors. This still left

a mystery problem: Why did patients relapse after treatment had permitted their ulcers

to heal? Then came another breakthrough: the discovery that ulcers were caused by an

infection of Helicobacter pylori. So treatment has developed into a regimen that aims ﬁ rst

to reduce acidity and so relieve patients of pain and then to give a combination of antibiotic

and antibacterial agents to eradicate the infection.

At the time that it was discovered, burimamide appeared to be a selective drug and was

used to characterize H

receptors. There were, however, some anomalous pharmacological

effects with its use; we now know that, ironically, it is actually much more potent as an

antagonist at histamine H

receptors than at H

receptors. Thus, from the early struggle in

the 1960s to prove the existence of more than one type of histamine receptor, we have now

had the contribution of molecular biological techniques that inform us that there are at least

four classes of histamine receptor, designated as H

, H

, and H

. It remains to be seen

whether new drug therapies will be developed as a consequence of these later discoveries.

REFERENCES

Ash, A. S. F., and H. O. Schild (1966), Br. J. Pharmacol. Chemother. 27, 427–439.

Black, J. W., W. A. M. Duncan, G. J. Durant, C. R. Ganellin, and M. E. Parsons (1972), Nature

236, 385–390.

Black, J. W., G. J. Durant, J. C. Emmett, and C. R. Ganellin (1974), Nature 248, 65–67.

(20) Ranitidine

CHNO

SCH

NHCNHCH

(CH

)

NCH

Scheme 10.6

Bovet, D. (1950), Ann. N. Y. Acad. Sci. 50, Art. 9, 1089–1126.

Bradshaw, J., M. E. Butcher, J. W. Clitherow, M. D. Dowle, R. Hayes, D. B. Judd, J. M.

McKinnon, and B. J. Price (1982), in A. M. Creighton and S. Turner, eds., Chemical Regula-

tion of Biological Mechanisms, Special Publication 42, Royal Society of Chemistry, London,

pp. 45–57.

Brimblecombe, R. W., W. A. M. Duncan, G. J. Durant, J. C. Emmett, C. R. Ganellin, and M. E.

Parsons (1975), J. Int. Med. Res. 3, 86–92.

Brimblecombe, R. W., W. A. M. Duncan, G. J. Durant, J. C. Emmett, C. R. Ganellin, G. B. Leslie,

and M. E. Parsons (1978), Gastroenterology 74, 339–347.

Charton, M. (1965a), J. Org. Chem. 30, 3346–3350.

Charton, M. (1965b), J. Org. Chem. 30, 969–973.

Dale, H. H., and P. P. Laidlaw (1910), J. Physiol. (London), 41, 318–344.

Durant, G. J., J. C. Emmett, C. R. Ganellin, P. D. Miles, H. D. Prain, M. E. Parsons, and G. R.

White (1977), J. Med. Chem. 20, 901–906.

Folkow, B., K. Haeger, and G. Kahlson (1948), Acta Physiol. Scand. 15, 264–278.

Freston, J. W. (1982), Ann. Intern. Med. 97, 573–580.

Ganellin, C. R. (1981), J. Med. Chem. 24, 913–920.

Ganellin, C. R., G. J. Durant, and J. C. Emmett (1976), Fed. Proc. Fed. Am. Soc. Exp. Biol. 35,

1924–1930.

Ghosh, M. M., and H. O. Schild (1958), Br. J. Pharmacol. Chemother. 13, 54–61.

Gregory, H., R. C. Sheppard, D. S. Jones, P. M. Hardy and G. W. Kenner (1964), Nature (London)

204, 931–933.

Gregory, R. A., and H. J. Tracy (1961), J. Physiol. (London) 156, 523–543.

Grossman, M. I., C. Robertson, and C. E. Rosiere (1952), J. Pharmacol. Exp. Ther. 104,

277–283.

Johnson, L. R. (1971), Gastroenterology 61, 106–118.

Lin, T. M., R. S. Alphin, F. G. Henderson, D. N. Benslay, and K. K. Chen (1962), Ann. N. Y. Acad.

Sci. 99, 30–44.

Lin, T. M., D. C. Evans, M. W. Warrick, R. P. Pioch, and R. R. Ruffalo (1983), Gastroenterology

84, 1231.

Takeda, M., T. Tagaki, Y. Yashima, and H. Maeno (1982), Arzneim. Forsch. 32(ii), 734–737.

Van den Brink, F. G. (1969), in Histamine and Antihistamines, Molecular Pharmacology,

Structure–Activity Relations, Gastric Acid Secretion, Drukkerij Gebr. Janssen, Nijmegen, The

Netherlands, p. 179.

Winship, D. H. (1978), Gastroenterology 74, 402–406.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

REFERENCES 311

313

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

DISCOVERY OF POTENT NONPEPTIDE

VASOPRESSIN RECEPTOR ANTAGONISTS

BRUCE E. MARYANOFF

Johnson & Johnson Pharmaceutical Research and Development

Spring House, Pennsylvania

11.1 INTRODUCTION

The practice of drug discovery has undergone a paradigm shift over the past 30 years

or so.

Around 1975, there were fairly limited options available to the “drug hunter”

keen on nabbing the ultimate prey: a new marketed product. Drug discovery at that time

was usually centered on the synthesis of sizable samples (2 to 5 g) of new chemical

compounds and broad pharmacological screening in animal models that represented a

disease state. For the most part, researchers did not conduct in vitro testing with enzymes

and receptors. If one were seeking a novel, patentable, biologically active compound in

the “old days,” one would prepare interesting compounds with structures like those of

known drugs, bioactive natural products, or endogenous mediators (e.g., norepinephrine

and acetylcholine), but with interesting chemical twists and turns, so to speak. Drug

discovery scientists have now become very partial to the pursuit of discrete molecular

targets such as receptors, enzymes, and ion channels, with an emphasis on ﬁ nding com-

pounds that act directly, with good selectivity. In addition, the abundance of detailed

structures for pharmaceutically interesting macromolecules has led to very “structure-

based” approaches to drug discovery. It would be wonderful if all of the technological

advances, such as genomics/proteomics, molecular biology, vast chemical libraries, high-

throughput screening, protein structures, and computational methods, could be applied to

a worthy research goal and actually deliver a clinical candidate that would sail all the way

to the marketplace. Unfortunately, the experience of the last 10 to 15 years has revealed a

disquieting paradox. Despite major investments of capital and human resources in high-

tech drug discovery, the output of pharmaceutical products has actually diminished.

314 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

Although we have numerous exciting modern tools and approaches at our disposal, we

probably should not lose sight of the old-fashioned low-tech approaches. It is my sen-

sibility to embrace all possible avenues of attack in trying to ﬁ nd drug candidates with

staying power, candidates that possess the attributes to enter human clinical trials and to

progress to the market.

My discovery of TOPAMAX topiramate, which is marketed worldwide for the treat-

ment of epilepsy and migraine, occurred under the old paradigm.

A synthetic inter-

mediate from an enzyme inhibitor project, McN-4853, was screened in a panel of

pharmacological assays, one of which was the maximal electroshock seizure test in

mice. The interesting anticonvulsant activity that we found was conﬁ rmed in an NIH

laboratory (National Institute of Neurological Diseases and Stroke), and a high level

of enthusiasm was generated. Topiramate was placed on the development pathway and

ultimately led to a commercial product with annual sales in excess of $1 billion. It is

doubtful that this valuable molecule would ever have surfaced through a process of

direct intent.

Since 1990, we have been pursuing serine protease inhibitors by the process of structure-

based drug discovery. Nearly eight years of direct intent transpired before we delivered a

compound into development that advanced to human clinical trials.

Our ﬁ rst drug candi-

date from this project was the tryptase inhibitor RWJ-56423, for the treatment of asthma

and allergic rhinitis.

1,4

Subsequently, on related enzyme targets, we delivered a ﬁ rst-in-

class, dual inhibitor of cathepsin G and chymase, RWJ-355871, for the treatment of airway

inﬂ ammatory diseases.

1,5

This novel protease inhibitor emanated from the auspicious con-

ﬂ uence of high-throughput screening and structure-based drug design. It can be viewed as

encouraging that these high-tech approaches yielded such favorable outcomes, but the time

frame was protracted and a marketed product is not even on the horizon.

In another fundamental design project, we sought compounds to block certain integ-

rins, which are cell-surface adhesion molecules. We wanted to mimic key aspects of the

KQAGD sequence in the ﬁ brinogen γ-chain to generate oral ﬁ brinogen receptor (glycopro-

tein IIb/IIIa) antagonists for treating thrombotic disorders. By applying high-throughput

solid-phase parallel synthesis in the lead optimization stage, we were able to identify elar-

oﬁ ban (RWJ-53308) and propel it into human clinical trials in a reasonably short time

frame.

Although our oral integrin antagonist elaroﬁ ban progressed successfully through

phase IIa, the therapeutic modality of oral GPIIb/IIIa antagonists was unfortunately drawn

into serious question by troublesome results in the clinical trials of certain oral agents from

other companies.

As a consequence, this promising antithrombotic therapy virtually dis-

integrated overnight.

Another opportunity to score a drug product entailed a low-tech rather old-fashioned

approach: the chemical modiﬁ cation of competitor structures to attain the desired biologi-

cal activity and suitable patentability. This “follow-on” regime may sound less risky in

that a validated starting point exists and the characteristics of the clinical candidate can

be envisioned more easily. However, the risk is just moved downstream in the pathway

in the sense that the clinical candidate must meet many additional requirements, unlike

an agent that is ﬁ rst-in-class. It is necessary to impart signiﬁ cant improvements over key

existing compounds from competitors, and in many cases, the requirements can be a

very tall order, indeed. Since it may not be readily apparent how to engineer the desired

molecular attributes, the acquisitive researchers must resort to Edisonian methods. We

came to apply this modus operandi to the discovery of potent, nonpeptide vasopressin

receptor antagonists.

11.2 GENESIS OF THE VASOPRESSIN RECEPTOR ANTAGONIST PROJECT

In 1997, the drug discovery team that I was leading in concert with biologist Dr. Patricia

Andrade-Gordon was working on a number of interesting molecular targets. However,

these targets were largely viewed as high risk in that their therapeutic potential had not

been deﬁ ned by existing compounds with that mechanism of action in a clinical environ-

ment. We wanted to balance the project portfolio of the Vascular Research Team by initiat-

ing work on some follow-on, or fast-second, targets. After reviewing various possibilities

with key team members, we whittled the original list of 10 targets down to two, one of

which was vasopressin receptor antagonists. The idea of entering this new area did not

capture our imagination immediately. Rather, it was our attendance at the annual meeting

of the American Heart Association (AHA) in November of that year that provided the mo-

tivation. After studying a poster presented by a scientist from Lederle Laboratories (now

part of Wyeth Pharmaceuticals), we became more intrigued with the potential for quickly

delivering a useful clinical entity. However, it was a luncheon engagement at the AHA

meeting that brought us to a crisp decision.

Our team was involved in a multiyear collaboration with COR Therapeutics, Inc.

(now part of Millennium Pharmaceuticals) in the pursuit of thrombin receptor (PAR-1)

antagonists, a cutting-edge class of antithrombotic agents.

We happened to encounter

Dr. Charles Homcy, the senior vice president of R&D at COR, at the AHA meeting and

went to lunch with him in Orlando’s Peabody Hotel. Since Charles had been at Lederle,

we mentioned the exciting poster on vasopressin receptor antagonists, whereupon he

indicated that the project was initiated by him and driven forward during his tenure.

Charles’s enthusiasm was so infectious that Patricia and I contracted the disease. Simul-

taneously, she and I felt a surge of resolution, and we were pleased to pay the entire bill.

Our determination to enter this ﬁ eld was communicated to key team members on our

return to Spring House.

A target proposal was written and submitted to Dr. Per Peterson, our senior vice presi-

dent of drug discovery, for review and approval. Interestingly, another team, one based in

Raritan, New Jersey, was also planning to embark on this target as a fast-second project.

The Endocrine Therapeutics Team viewed vasopressin from the perspective of its being

a hypophyseal hormone. How would this conﬂ ict be settled? Per’s philosophy was clear

on this issue: We compete vigorously with other drug companies but not within our own.

Therefore, our two teams combined forces and drew up a plan for a cross-site collaboration

early in 1998. Most of the biology and pharmacology would be conducted in Raritan, and

the chemistry would be split between the two sites, with different novel structures being

designed and synthesized at each site. This collaboration has been exceptionally fruitful,

with ﬁ ve compounds entering preclinical development, one of which progressed success-

fully through phase IIa clinical trials. In this chapter I intend to address the chemical series

that were pursued in my group at Spring House.

11.3 VASOPRESSIN, ITS RECEPTORS, AND DISEASE

Arginine vasopressin (AVP) is a nonapeptide hormone that is mainly secreted from the

posterior pituitary gland and that exerts multiple biological actions, as a hormone and a

neurotransmitter.

AVP is synthesized in magnocellular neurons within the hypothalamus

and released into the circulation from the posterior pituitary. It can have multiple peripheral

VASOPRESSIN, ITS RECEPTORS, AND DISEASE 315

316 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

actions, including contraction of uterine, bladder, and vascular smooth muscle, stimulation

of hepatic glycogenolysis, induction of platelet aggregation, release of corticotropin from

the anterior pituitary, and the stimulation of renal water reabsorption. As a neurotransmitter

within the central nervous system (CNS), AVP can affect various fundamental behaviors,

the stress response, and the memory process. These diverse actions of AVP are mediated

through speciﬁ c G-protein-coupled receptors (GPCRs), which have been classiﬁ ed into

three subtypes: V

, V

, and V

(Table 11.1). The V

and V

receptors signal through

phosphatidylinositol hydrolysis to mobilize intracellular calcium (Ca

2

), whereas V

re-

ceptors signal through cyclic adenosine monophosphate (cAMP). The V

receptors me-

diate the contraction of smooth muscle, hepatic glycogenolysis, and the CNS effects of

vasopressin; the V

receptors stimulate corticotropin release from the pituitary gland; and

the V

receptors, which are found only in the kidney, are responsible for the antidiuretic

actions of AVP. Within the renal collecting ducts, AVP regulates the water-selective channel

aquaporin-2, inducing the translocation of aquaporin-2 to the plasma membrane to increase

water permeability.

From a cardiorenal standpoint, we were especially interested in the V

and V

receptor

subtypes. Vasopressin-induced antidiuresis maintains normal plasma osmolarity and blood

volume via renal epithelial V

receptors, while vasoconstriction is mediated by V

recep-

tors. In pathological states, plasma AVP levels may be unusually high at a given plasma os-

molarity, resulting in water retention and hyponatremia (low plasma Na



concentrations),

an electrolyte abnormality that occurs in 1% of hospitalized patients.

Hyponatremia is

associated with edema in conditions such as hepatic disease/cirrhosis, congestive heart

failure (CHF), and renal failure. Presently, no drug is available that will selectively cause

water excretion, or aquaresis.

There is substantial evidence to support a pathophysiological role of AVP in CHF.

a rat model of CHF, resulting from ischemic cardiomyopathy, a peptide-based V

receptor

antagonist increased cardiac output, decreased peripheral resistance, and increased urine

output four- to ten-fold.

The ﬁ rst nonpeptide V

-receptor antagonist, OPC-31260, induced

marked aquaresis in conscious dogs with CHF. A combined V

receptor antagonist

could have particular beneﬁ t in heart failure via hemodynamic and renal mechanisms.

pivotal clinical trials, three nonpeptide vasopressin receptor antagonists improved the ﬂ uid

status, osmotic balance, and hemodynamics of patients with CHF; however, long-term,

adequately populated studies are needed to prove the value of such drugs relative to clinical

outcomes and quality of life.

12d

Subtype Location Physiological Actions Signal Transduction

Smooth muscle

Liver

CNS

Stimulates smooth muscle contraction

Stimulates hepatic glycogenolysis

Memory; behavior

coupled:

IP formation

2

mobilization

Anterior pituitary Stimulates ACTH release from

pituitary

coupled:

IP formation

2

mobilization

Kidney Stimulates renal reabsorption of water G

coupled:

adenylate cyclase

TABLE 11.1 Vasopressin Receptor Subtypes and Biological Actions

IP, inositol phosphate.

Nonosmotic release of AVP is well documented in liver cirrhosis,

which poses a seri-

ous unmet medical need. The failure to suppress AVP activity after central volume expan-

sion may be one of the early mechanisms responsible for water–electrolyte imbalance in

liver cirrhosis in children.

Administration of a V

-receptor antagonist in rats with ex-

perimental cirrhosis has produced signiﬁ cant improvement.

Vasopressin is implicated in

experimental brain edema, and a V

-receptor mechanism has been postulated.

In rats, a

antagonist increased plasma osmolarity and reversed the increase in brain water caused

by bilateral occlusion of the carotid arteries; it also prevented cerebral edema and increases

in brain sodium while normalizing diuresis induced by subarachnoid hemorrhage.

Researchers at several pharmaceutical companies, recognizing the potential therapeutic

utility of vasopressin receptor antagonists, have been inspired to identify clinical candi-

dates.

In early studies, marked species differences were found in the effects of peptide-

based vasopressin antagonists.

Although SKF-101926 was a V

-receptor antagonist in

rats, dogs, and monkeys in vivo, and human renal tissue in vitro, it showed agonist activity

in human clinical trials. The variability between rats, dogs, monkeys, and humans suggested

critical structural differences between vasopressin receptor proteins. A similar observa-

tion of interspecies divergence was made with backup compound SKF-105494. During the

1980s, SmithKline & French scientists sought a pure water diuretic, partly in support of

their diuretic franchise involving the treatment of hypertension, and thus advanced these

interesting peptide molecules. However, the problematic species dependency hampered

their development plans and suitably alerted subsequent practitioners.

A major step forward was made with the discovery of the ﬁ rst nonpeptide vasopressin an-

tagonists, OPC-21268 (V

-selective)

and OPC-31260 (V

-selective).

Prior to this mile-

stone, potent vasopressin antagonists were peptide analogs of AVP, with the expected liabili-

ties of poor oral bioavailability and short duration of action.

OPC-31260 (1; mozavaptan)

is an orally active V

-selective antagonist in rats and humans,

and numerous structure–ac-

tivity studies have been performed around this chemical class (Scheme 11.1).

23b,26

This work

at Otsuka had a great impact on the ﬁ eld because it established a key structural motif, or

pharmacophore, for nonpeptide vasopressin receptor antagonists: the 1-(4-benzamidoben-

zoyl)benzazepine template. Further strides on this central theme were made in the late

1990s by researchers at Lederle/Wyeth

and Yamanouchi.

At this point in time, various

benzazepine-type series have been disclosed in the scientiﬁ c literature,

including several

from Johnson & Johnson.

Some compounds of particular interest are the V

-selective

antagonists tolvaptan (2; OPC-41061)

26,31

and lixivaptan (3; VPA-985),

27,32

and the V

antagonist conivaptan (4; YM-087),

28a,33

all of which have advanced to phase II or phase III

clinical trials (Scheme 11.1). Other important contributions have emanated from research-

ers at Sanoﬁ –Synthelabo, who have reported V

-, V

-, and V

-receptor antagonists of

clinical interest with very different structural motifs: indoline prolinamides, such as V

selective SR-49059 (relcovaptan) and V

-selective SSR-149415, and spiroindolinones,

such as V

-selective SR-121463A.

Our drug discovery effort, which has focused on the

(benzamidobenzoyl)benzazepine class, is described in the remainder of this chapter.

11.4 THE GAME PLAN

When initiating a new drug discovery project, it helps to have a well-deﬁ ned goal, a vision

for success, some novel chemical ideas, and a workable plan for moving compounds for-

ward. In the absence of a sound plan with clear-cut decision-making steps (the critical

THE GAME PLAN 317

318 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

path), there is no reason at all to design and synthesize target molecules. Importantly, the

early bioassays in the critical path need to be established beforehand, and suitable pharma-

cology models need to be available. At the outset we were going to rely on human receptor

binding assays coupled with human receptor cell-based assays.

The 418-amino acid human V

receptor has a 72% sequence identity with the rat V

receptor, 36/37% identity with the human/rat V

receptors, and 45% identity with the hu-

man oxytocin receptor; the cloned 371-amino acid human V

receptor has 50% identity

with the rat V

receptor and 45% identity with the human oxytocin receptor.

A crucial

starting point for our project was establishing and validating binding assays based on cloned

and expressed human V

and V

receptors. We used recombinant human V

or V

receptor

preparations derived from the cell membranes of transfected HEK-293 cells, with test com-

pounds being evaluated for their ability to displace [

H]AVP. Cellular functional assays

were also put in place to ascertain whether a compound that binds is a receptor agonist or

antagonist. The inhibition of receptor activation caused by AVP was quantiﬁ ed in HEK-293

cells expressing either human V

or V

receptors, and changes in intracellular concentra-

tions of either Ca

2

(for V

) or cAMP (for V

) were measured. For follow-up work, we had

cellular functional assays based on rat V

and V

receptors to address any issues that might

arise in the rat pharmacology work. Also, a human oxytocin receptor binding assay was

available to rule out that undesirable action. Compounds of interest for in vivo evaluation

would be tested for their pharmacokinetic properties, such as metabolic stability to rat and

human liver microsomes, oral bioavailability and duration, and binding to plasma proteins.

Scheme 11.1

Obviously, if a compound with good in vitro activity were metabolically unstable or poorly

absorbed orally, it would not proceed to the next stage.

Our early in vivo assessment would entail hypertension (V

action) and aquaresis (V

ac-

tion) experiments in rats. A V

receptor antagonist should reverse the hypertension induced

by AVP (saline solution infused at 30 ng/kg per minute intravenously) in pentobarbital-

anesthetized rats. The degree of reduction in mean arterial blood pressure from the AVP-in-

duced level of 50 to 60 mmHg would be recorded, and ED

values would be calculated

from the linear portion of the dose–response curve. The acute diuretic effect of a V

-recep-

tor antagonist would be determined in hydrated, conscious rats given single oral doses of

test compounds. Spontaneously voided urine would be collected over 4 hours, and the urine

volume, osmolality, and electrolyte concentrations would be measured. Advanced leads

would be tested acutely for aquaresis in dogs and monkeys, then multidose studies would

be performed in rats and monkeys. Thus, a workable game plan was established.

11.5 NOVEL CHEMOTYPES: VARIATIONS ON A THEME

My experience at the Orlando AHA meeting in 1997 brought the (benzamidobenzoyl)be

nzazepine class of vasopressin antagonists into the forefront. Lixivaptan (3) had entered

human clinical trials

20a,32

and analog CL-385004 (5) also looked like an interesting com-

pound (Scheme 11.2).

27c

Such structures appeared to offer a suitable platform for branch-

ing off to a novel chemotype. We conceived of compounds with general formula 6,

30c,d

and

an examination of the patent literature in 1998 indicated that no compounds of this variety

had been disclosed. Since these are fairly sizable molecules that could present challenges

for oral drug development due to unfavorable physical properties, such as high molecular

weight (450 to 600 Da), high hydrophobicity (log P  4), and limited aqueous solubility

(1 mg/mL),

we thought that the presence of one, or even two, basic amine centers in tar-

gets 6 could be a positive inﬂ uence. It was hoped that these novel chemical entities would

give rise to receptor antagonists with good in vitro and in vivo potency and favorable oral

bioavailability. Even though we became highly enamored with this chemical approach,

we decided to diversify by pursuing parallel paths to enhance the probability of success.

Such a parallel-processing strategy can be applied in the early stage of a new project, but

ultimately needs to be narrowed down by the assessment of critical data. Thus, we also

conceived of novel indole-fused compounds with generic structure 7, which are somewhat

akin to 1 and 4,

30a

and novel bridged bicyclic derivatives with general structure 8,

30b

which

are related to targets 6 (Scheme 11.2).

11.5.1 Azepinoindoles

An assortment of indolobenzodiazepine derivatives, 7a–m, were prepared, mainly accord-

ing to the route in Scheme 11.3, and tested (Table 11.2).

30a

Intermediate azepino[4,3,2-

cd]indole (11) was obtained in seven steps from 4-nitroindole (10),

which was prepared

from 2-methyl-3-nitroaniline (9) in two steps.

Acylation of 11 with a 4-nitroaroyl chlo-

ride (RNO

PhCOCl) provided 12, which was reduced with zinc and ammonium chloride

to provide aniline 13. Acylation of 13 with aroyl chlorides (R

PhCOCl) provided target

compounds 14 (encompassing 7a–g). In general, this series of compounds (Table 11.2) did

not bind effectively to the V

receptor, although three compounds exhibited modest V

binding of 30 to 60% at 1 µM: 7b, 58%; 7d, 33%; 7h, 49%. Mozavaptan congener 7a was

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 319

320 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

Scheme 11.2 (S,S,S) or (S) enantiomer shown.

6 (X = CH

, O, S, NR)

(CH

)

8 (n = 1, 2)

Scheme 11.3 Synthesis of 14. Reagents and conditions: a, 4-NO

ArCOCl, Et

N, CH

, 0C; b,

Zn dust, NH

Cl, MeOH, reﬂ ux; c, Ar'COCl, Et

N, CH

, 0C.

2 steps

7 steps

inactive in V

binding, and lixivaptan congener 7b was moderately potent (V

 90 nM)

but much weaker than lixivaptan (V

 2.3 nM). However, it was the 2-phenylbenzoyl

group, which is contained in conivaptan (4), that made a big difference. For example, 7c–e

had V

values in the range 5 to 10 nM, and substitution of the indole 2-position with

a sulfonic acid group surprisingly did not diminish the binding potency (i.e., 7l and 7m).

Compounds 7c and 7d exhibited reasonable functional antagonism, as their V

values of

17 nM compared favorably with that of lixivaptan (VPA-985; V

 23 nM). Compound

7c was evaluated orally in the rat aquaresis assay at a dose of 10 mg/kg, but it lacked activ-

ity. Besides oral bioavailability being a problem with this series, the compounds tended to

have low aqueous solubility, which is why we tried the sulfonic acid group.

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 321

Compound

Binding

(nM)

Funct.

(µM)

H 2-Me H H H IA

H 2-Me 5-F H H 90

H 2-Ph H H H 9.0 0.070

H 2-(4-MePh) H H H 5.0 0.070

o-Cl 2-Ph H H H 7.0

m-Cl 2-Ph H H H 20 0.33

H 2-Me 3-F H H 83

H 2-(4-MePh) H Me H 10

H 2-(4-MePh) H C(O)Me H 80

H 2-(4-MePh) H Pr H IA

H 2-Ph H H Cl 20

H 2-Ph H H SO

H 6.0

o-Cl 2-Ph H H SO

H 6.0

VPA-985

2.3 0.023

TABLE 11.2 Indoloazepine Derivatives with Their Vasopressin Receptor Effects

Puriﬁ ed by reverse-phase semiprep. HPLC; 95% pure by reverse-phase HPLC/MS (215/254 nm); character-

ized by 300-MHz

H NMR and MS.

Position of substitution is ortho or meta to the carboxyl.

Binding to human V

receptors (N  3 to 6). IA  inactive, i.e., 30% inhibition of radioligand binding at

100 nM. All analogs had 100-fold selectivity for V

over V

Inhibition of functional human V

-receptor activity (N  3 to 5).

Reference standard. V

binding K

 44 nM.

322 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

11.5.2 Bridged Bicyclic Derivatives

Bridged bicyclic benzodiazepines, represented by 8a–l, were synthesized (Scheme 11.4)

and evaluated (Table 11.3).

30b,40

Fortunately, the chemistry for obtaining the key bridged

bicyclic amino acids in enantiomerically enriched form had already been described.

Het-

ero Diels–Alder reaction of cyclopentadiene (n  1) or 1,3 cyclohexadiene (n  2) with

activated imine 15 (from ethyl glyoxalate and α-methylbenzylamine) yielded exo-substi-

tuted cycloadducts 16 with high (n  1) to moderate (n  2) diastereoselectivity. The (S,S,S

or S)-cycloadduct was produced predominantly from the (R)-()-amine, and vice versa.

The diastereomers were separated by ﬂ ash-column chromatography, and the one isomer

(shown) was hydrogenated to reduce the alkene and deprotect the nitrogen to furnish bi-

cyclic amino esters 17 in good yields (95% e.e.). 2-Nitrobenzamides 18 were reduced

and cyclized to intermediate benzodiazepinediones (not shown), which were reduced with

LiAlH

to give the key benzodiazepine core structure, 19 (50 to 65% overall yield based

on 17).

Importantly, these transformations occurred without any loss of stereochemical

Scheme 11.4 Synthesis of 8. Reagents and conditions: a, EtO

CCHO, toluene, reﬂ ux (H

O);

b, CF

H (n  1) or CF

H/BF

O (n  2), CH

, 20C; c, H

, Pd/C, EtOH; d, Et

, 0C; e, Fe, AcOH, reﬂ ux; f, LiAlH

, THF, 0C; g, Zn dust, NH

Cl, MeOH, reﬂ ux; h, ArCOCl,

N, CH

, 0C.

Me Ph

CHCO

(CH

)

(CH

)

(CH

)

ab c

n = 1 or 2

C(O)Cl

(CH

)

e, f

C(O)Cl

(CH

)

(CH

)

d g, h

(CH

)

17 +

integrity at the originally established stereogenic center. Acylation of 19 with 4-nitroben-

zoyl chlorides (20) yielded precursors 21, which were reduced and acylated to provide the

desired target molecules, 8.

This series of compounds (Table 11.3) afforded very potent V

antagonists, possessing

single-digit nanomolar K

values in receptor binding (i.e., 8a, 8c, 8d, 8g, 8h, and 8i). Sev-

eral compounds with the (S) conﬁ guration were dual V

antagonists in the binding as-

says. Compound 8g is a notable example, with V

and V

values of 3.2 and 1.2 nM. This

compound and 8c also showed well in both functional assays: V

and V

values of 9 and

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 323

(CH

)

Compound

n Conﬁ g.

Binding

(nM)

Binding

(nM)

Funct.

(µM)

Funct.

(µM)

H Cl 2-Ph 1

7.0 1.8 0.13 0.009

H Cl 2-Ph 1

7.0 0.03

Cl H 2-(4-MePh) 1

7.0 2.3 0.023 0.013

Cl Cl 2-(4-MePh) 1

10 1.4

Cl H 2-F 1

20 12 2.6 0.180

Cl H 2-Cl 1

1.5 4.8 0.13 0.032

Cl H 2-CF

1.2 3.2 0.04 0.009

H H 2-Ph 2 R/S ⬃30 2.8

H Cl 2-Ph 2 R/S 30%

1.8

H Cl 3-F-5-Me 2

15%

9.0

Cl H 2-Cl 2

10 6.0 0.62 0.52

Cl H 2-(4-MePh) 2

17 33

VPA-985

44 2.3 6.0 0.023

OPC-31260

70 13 0.42 0.04

TABLE 11.3 Bridged Bicyclic Derivatives with Their Vasopressin Receptor Effects

Puriﬁ ed by chromatography and crystallized as HCl salts; 95% pure by reverse-phase HPLC/MS (215/254 nm);

characterized by 300-MHz

H NMR and MS. Elemental analyses for compounds tested in vivo. Where relevant,

99% e.e. by chiral HPLC.

Absolute conﬁ guration for the stereocenter with the asterisk. For n



1, (S,S,S) or (R,R,R) enantiomers; for



2, (S) and/or (R) enantiomers.

Binding to human V

and V

receptors (N  3 to 6).

Inhibition of functional human V

or V

activity (N  3 to 5).

Percent inhibition at 100 nM.

Reference standard.

324 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

40 nM for 8g and 13 and 23 nM for 8c. The (R)-enantiomer (8b) is a potent V

-selective an-

tagonist with V

values of 7 nM in the binding assay and 30 nM in the functional assay.

We tested potent derivatives 8a (10-fold V

-selective), 8b (V

-selective), and 8c (bal-

anced V

) in vivo for their aquaretic activity in rats (Table 11.4). Compounds 8a and 8b

exhibited reasonable activity at a low oral dose of 0.3 mg/kg, whereas 8c was considerably

less potent (3 mg/kg). Compounds 8a–c had modest oral bioavailability in rats (F  14%,

10%, and 7%, respectively), with 8c being the lowest; the oral half-lives were useful, rang-

ing from 2.5 to 5 h. The weaker potency of 8c may be related to its unimpressive oral bio-

availability. We studied the ability of the VPA-985, 8a, 8c, and 8g to reverse AVP-induced

hypertension in rats (V

related). VPA-985 had an ED

value of 1120 ± 520 µg/kg (mean

± SE, N  5), whereas 8a, 8c, and 8g showed greater potency with ED

values of 220

± 43 µg/kg (N  6), 170 ± 120 µg/kg (N  8), and 380 ± 290 µg/kg (N  5), respectively,

which relates to their V

afﬁ nities. The potent compounds in this series tended to have low

aqueous solubility and some signiﬁ cant inhibition of cytochrome P450 enzymes, which are

involved in the oxidative metabolism of drug molecules (e.g., inhibition of Cyp 3A4).

11.5.3 Thiazino-, Oxazino-, and Pyrazinobenzodiazepines

Our best prospects for clinical candidates were realized by investigating compounds in

series 6.

30c,d

Initially, we carried out some work with 6 (X  CH

) and its pyrrolidine

congener, but the patent landscape surrounding this type of compound became unattractive

as a result of information disclosed by Ohtake and co-workers.

29f,43

Thus, the compounds

with X  S, O, and NR became much more attractive. Most of the target compounds were

obtained by assembling the requisite tricyclic intermediates 25, as shown in Scheme 11.5,

and then converting them by the standard protocol mentioned earlier to target molecules,

which were subjected to biological assays (Tables 11.5 to 11.7).

30c,d,44

In certain cases the

racemates of the tricyclic intermediates were resolved into individual enantiomers, which

were converted to single-enantiomer targets.

30c,d,44

Herein, we discuss the following rep-

resentative target molecules: thiazino compounds (26–36), oxazino compounds (37–50),

Oral Dose

(mg/kg) N

Urine Vol.

(mL)

Urine Osmol.

(mOs/kg)

Oral Dose

(mg/kg) N

Urine Vol.

(mL)

Urine Osmol.

(mOs/kg)

8a 8b

Vehicle 10 1.1 ± 0.2 663 ± 62 Vehicle 10 0.6 ± 0.1 892 ± 70

0.3 9 2.5 ± 0.3 444 ± 79 0.3 9 1.4 ± 0.1 408 ± 37

1 10 6.0 ± 0.7 299 ± 34 1 10 3.4 ± 0.4 303 ± 29

3 10 15.1 ± 1.4 180 ± 9 3 10 9.4 ± 1.1 238 ± 19

10 10 29.9 ± 2.0 138 ± 11 10 10 19.5 ± 1.4 172 ± 9

OPC-31260

Vehicle 10 2.3 ± 0.3 441 ± 30 Vehicle 16 2.1 ± 0.4 608 ± 50

1 10 2.9 ± 0.2 387 ± 15 1 8 2.3 ± 0.5 530 ± 42

3 10 5.8 ± 0.5 226 ± 9 3 10 5.4 ± 0.7 372 ± 39

10 10 17.8 ± 1.3 159 ± 9 10 9 12.2 ± 1.0 238 ± 18

30 8 21.5 ± 3.0 165 ± 17

Mean ± SE.

TABLE 11.4 In Vivo Diuretic Effects of 8a–c in Rats

and pyrazino compounds (51–64). A synthetic example involving 54 serves to illustrate

the procedure for introducing a 2,2,2-triﬂ uoroethyl group and for executing the end game

(Scheme 11.6).

The thiazino series, 6 (X  S), was explored ﬁ rst with numerous racemic targets being

prepared from 25a, some of which are presented in Table 11.5 (26–37). The direct analog

of VPA-985, 26, had much weaker binding to the V

and V

receptors than the reference,

Scheme 11.5 Synthesis and resolution of tricyclic intermediates 25.

resolution

LiAlH

THF, 65˚C

(see ref 44)

C(O)Cl

N, CH

(

)-(+)-25

68-78%

53-89%

70-99%

AcOH

a: X = S

b: X = O

c: X = NMe

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 325

Scheme 11.6 Synthesis of target 54. ACE, α-chloroethoxycarbonyl; DCE, 1,2-dichloroethane; Tf,

triﬂ uoromethanesulfonyl (triﬂ yl).

1. ACE-Cl, DCE

reflux, 2 h

2. MeOH, reflux

3. CF

OTf

1. Zn dust

Cl, MeOH

C(O)Cl

N, CH

C(O)Cl

25c +

326 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

although the V

value of 60 nM is still quite respectable. The reduced V

afﬁ nity of 26

is consistent with its potency in the V

functional assay (K

 300 nM). Introduction of an

o-phenyl group, as in 27, enhanced the V

receptor afﬁ nity (K

 37 nM), which was re-

ﬂ ected in the V

functional assay (K

 70 nM). Compound 27 showed good V

selectivity,

especially as depicted from the functional assays, with a V

ratio of 200. Thus, we

adopted the o-phenyl substituent for other analogs. Des-halogen parent 28 had similar V

afﬁ nity (K

 5.0 nM) and selectivity (V

 10). Its two enantiomers, (S)-()-28 and

(R)-()-28, were examined and each had signiﬁ cant V

receptor afﬁ nity (K

 3.2 and

25 nM, respectively). The mere eightfold difference in V

afﬁ nity for the two enantiomers

26–36

Compound

Binding

(nM)

Binding

(nM)

Funct.

(µM)

Funct.

(µM)

H Cl Me 5-F IA 60 14 0.3

H Cl Ph 5-F IA 37 14 0.07

H H Ph H 49 5.0 8.0 0.02

(ⴙ)-28

d,e

H H Ph H 84 3.2 1.4 0.015

(ⴚ)-28

e,f

H H Ph H 290 25 1.3 0.04

H Cl Ph H 490 11

(ⴙ)-29

d,e

H Cl Ph H 250 3.7 14 0.017

H Cl Ph 4-F 410 3.7

H F Ph H IA 10

H Me Ph H IA 10

H OMe Ph H IA 15

9-Cl H Ph H IA 10

8-Me H Ph H 18 8.0 14 0.023

8-F H Ph H IA 9.0

VPA-985

44 2.3 6.0 0.023

Puriﬁ ed by reverse-phase semi-prep. HPLC and isolated as triﬂ uoroacetate salts unless noted otherwise; 95%

pure by reverse-phase HPLC/MS (215/254 nm); characterized by ESI-MS, with selected compounds analyzed by

300-MHz

H NMR. Compounds are racemates unless noted otherwise.

Binding to human V

and V

receptors (N  3 to 6). IA



inactive (30% inhibition at 100nM).

Inhibition of functional human V

or V

activity (N  3 to 5).

(S)-() enantiomer.

HCl salt.

(R)-() enantiomer.

Reference standard.

TABLE 11.5 Vasopressin V

and V

Binding and Functional Data

for Thiazinobenzodiazepines

of 28 suggests that the geometry around this portion of the ligand is not particularly critical

for V

receptor binding interactions.

The excellent V

afﬁ nity of (S)-()-28, coupled with the 26-fold selectivity for V

over

, was encouraging. Subsequent variation of R

on the 4-aminobenzamide ring provided

potent, reasonably V

-selective analogs (cf. 28 with 29 and 30–33). Overall, notable V

afﬁ nities were realized for (S)-()-29 (K

 3.7 nM), (S)-()-28 (K

 3.2 nM), and 30

 3.7 nM). The results for 34–36 indicate that R

substitution is fairly well tolerated, al-

though this modiﬁ cation can strengthen V

afﬁ nity, as observed for 35 (V

 18 nM).

To follow up on these results, we studied several compounds in cell-based functional

assays. Compounds ()-29, (S)-()-28, (R)-()-28, 35, and VPA-985 potently antago-

nized the effects of AVP on human V

receptors (K

 0.015 to 0.04 µM), whereas they

had just weak potency against human V

receptors (K

 1.3 to 15 µM). Enantiomer (S)-

()-28 gave V

and V

values of 1400 and 15 nM, for about 90-fold V

selectivity, and

(S)-()-29 gave V

and V

values of 14,000 and 17 nM, for an impressive functional

selectivity of about 820-fold. The potency of each enantiomer of 28 in the V

-receptor

functional assay (K

 15 to 40 nM) is reasonably consistent with the binding data.

For the oxazino series, 6 (X  O), we mostly studied the (S)-() enantiomers because

they have better V

receptor afﬁ nity (Table 11.6; 37–50). In comparing 40 and 41, each

enantiomer has a signiﬁ cant V

afﬁ nity (K

 0.9 and 13 nM, respectively), but there is

a 15-fold preference for 40. Interestingly, 40 also has a high afﬁ nity for the V

receptor

 24 nM), which accounts for a 25-fold V

binding selectivity. Direct VPA-985 analog

37 exhibited excellent V

receptor afﬁ nity as well as modest V

afﬁ nity, in contrast to

the related thiazinobenzodiazepine (26). In general, the o-phenyl group resulted in good-

to-excellent V

afﬁ nity. Indeed, several oxazino analogs had single-digit nanomolar V

receptor binding with good selectivity versus V

, such as 39, 40, 42–44, 46, 47, and 49.

The most potent compounds possessed the 2-phenyl (40, V

 0.9 nM), 2-phenyl-4-ﬂ u-

oro (46, V

 1.9 nM), and 2-phenyl-4-hydroxy (49, V

 1.4 nM) groups. In the V

functional assay, 37, 38, 40, 42–47, and 49 were very potent in antagonizing the effects of

AVP ( K

 0.002 to 0.02 µM), whereas they were weak in the V

functional assay (K

 0.7

to 11 µM). In the cell-based assays, 40 exhibited K

values of 420 and 3 nM for V

and

, reﬂ ecting 140-fold V

selectivity. The (R)-() enantiomer 41 showed less functional

potency (K

 15,000 and 170 nM) relative to 40, in contrast to a comparison of (S)-()-28

and (R)-()-28.

Oral administration of (S)-()-28 to rats elicited a dose-dependent aquaretic effect. At

a dose of 10 mg/kg p.o., urine output (N  8) was increased 300% over untreated con-

trols (N  8), with a reduction in urine osmolality of 70%. Oral administration of 40 to

rats produced a dose-dependent aquaretic effect with remarkable potency. An oral dose

of just 1 mg/kg caused a 700% increase of urine output (N  10) over untreated controls

(N  18), with a 60% reduction of urine osmolality. For comparison, at oral doses of 10 and

1 mg/kg, OPC-31260 (1) is reported to modify urine output/osmolality by 500%/–65%

and 0%/–10%, respectively;

23b

VPA-985 (3) is reported to modify urine output/osmolality

by 450%/70% and 200%/50%, respectively.

27a

Compound 40 showed excellent pharmacokinetics in rats and dogs. In rats, the oral

bioavailability was very favorable (F  68%), with an oral t

1/2

value of 3.7 h. The physical

properties of 40 (MW  538 Da) appeared to be better than those for reference compound

lixivaptan (3; MW  474 Da): for 40, log P  3.63; log D (pH3)  2.01 (pK

 4.75/12.7);

for lixivaptan: calculated log P  6.1 (calculated pK

 11.9).

The log P value for 40 of

3.63 is within a range (essentially 0 to 4) that is favorable for membrane transport.

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 327

The structure of 40, as a tosylate salt, was determined by x-ray diffraction. The tricycle

adopts a chairlike conformation for the seven-membered ring, with the pendant amide car-

bonyl (C-13) in an axial orientation. The carbonyl oxygen of this amide, O-2, is anti to the

fused benzene ring (Scheme 11.7). This arrangement is analogous to that observed for re-

lated N-acyltetrahydrobenzazepines in solution.

We carried out a Monte Carlo conforma-

tional search on protonated 40 with the OPLS–AA force ﬁ eld and GB/SA water model.

30c

The global energy minimum contains a trans-fused oxazinobenzodiazepine with a chair-

like seven-membered ring, an axial amide, and an amide carbonyl anti to the fused benzene

ring (Scheme 11.8). The tricyclic nucleus in this structure is closely superimposable on

Scheme 11.7 Solid-state structure of 40

TsOH [(S)-() isomer], showing the cationic subunit with

its atom-numbering scheme (standard atom color code; H, cyan).

Scheme 11.8 Structure of the global minimum-energy conformation of protonated 40 (standard

atom color code).

328

DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

that of the x-ray structure. The next-higher-energy structure is 1.7 kcal/mol (7.1 kJ/mol)

less stable than the global minimum. It has a cis-fused tricycle, resulting from inversion

of the ammonium center (NH



), a chairlike seven-membered ring, an axial amide, and an

amide carbonyl anti to the fused benzene.

30c

This structural information could be helpful

in developing a pharmacophore model for V

-receptor binding, and some modeling studies

in this regard have been reported.

From our studies with the thiazino- and oxazinobenzazepine series, we identiﬁ ed potent

nonpeptide vasopressin receptor antagonists, some of which have low-nanomolar V

receptor afﬁ nity and at least 20-fold selectivity for V

over V

receptors. In the thiazino

class, (S)-()-28 has excellent V

afﬁ nity (K

 3.2 nM), moderate binding selectivity (V

 26), good functional selectivity (V

 93), and oral efﬁ cacy as an aquaretic agent

in rats. Also, (S)-()-29 has excellent V

afﬁ nity (K

 3.7 nM) and notable V

selectivity

(binding V

 68; functional V

 820). In the oxazino class, (S)-() enantiomer

40 has excellent V

-receptor afﬁ nity (K

 0.9 nM), moderate binding selectivity (V

 27), good functional selectivity (V

 140), and impressive oral potency as an

aquaretic agent in rats. Although some compounds had reasonably high afﬁ nity for the hu-

man V

receptor, such as 35, 40, and, 49, signiﬁ cant activity was not manifested in the V

functional assay (Tables 11.5 and 11.6); this behavior was also observed with lixivaptan

(3; VPA-985). In general, the thiazino- and oxazinobenzodiazepine classes were not con-

ducive to providing potent functional V

-receptor antagonists. On the basis of an extensive

array of preclinical data, oxazinobenzodiazepine 40 (JNJ-17048434; RWJ-351647) was

advanced into human clinical studies. On oral administration to humans this V

-selective

antagonist was found to be effective as an aquaretic agent, with remarkable potency. Thus,

40 appears to have potential for the treatment of edematous conditions in patients.

Finally, the pyrazino series, 6 (X  NR), was examined with an eye toward identifying

a suitable backup to 40.

30d

In this case we hoped that the presence of two basic amine cen-

ters might offer advantages relative to the physical properties or, at least, offer somewhat

different and useful physical properties. The pyrazinobenzodiazepines were evaluated for

binding to human V

and V

receptors and for function activity on cells with human V

receptors. The results for representative compounds, 51–64, are given in Table 11.7. For 51

and its two enantiomers, there was very little difference in afﬁ nity, which is consistent with

the results for the corresponding sulfur and oxygen systems (vide supra), although the lack

of a distinction in this case is more emphatic. (R)-()-51 has a V

value of 12 nM, which

is 13-fold less potent than that for oxygen analog 40 (K

 0.9 nM).

The potency of (R)-

()-51 in the V

functional assay is also less than that of 40, by a factor of about 9, but it is

in the same range as VPA-985 (3). In general, the pyrazino series exhibited very weak V

afﬁ nity, except for 56. The nitrogen substituent, R

, plays an important role in determining

afﬁ nity. When R

is Me (51) or H (52), the V

value is 16 or 21 nM; however, when

is i-Pr (53) or CF

(54) the V

afﬁ nity is very weak. Another aspect relates to the R

substituent on the 4-aminobenzoyl unit, compare the V

afﬁ nity of 51 with that of 55–60.

The o-chloro (51) and o-methyl (56) groups gave reasonably potent V

afﬁ nity, whereas

the o-methoxy (57) and o-triﬂ uoromethyl (58) groups did not. As for R

, a 4-methyl group

(63) was quite good, but a 4-methoxy group (64) was not. Substantial potency in the V

functional assay (K

 250 nM) was obtained for 51, (R)-()-51, (S)-()-51, 56, 62, and

63. Compound 61, a direct analog of VPA-985 (3), showed 20-fold lower V

afﬁ nity; how-

ever, the afﬁ nity was improved to some degree (factor of 3) by using a 2-phenyl group for

(i.e., 62). (R)-()-51 turned out to be very selective for V

-receptor action versus V

, as

it had a V

binding K

 2900 nM and a V

functional K

 28,600 nM.

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 329

330 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

Additional studies were conducted with (R)-()-51. Its functional responses, in

terms of K

values, with human and rat V

receptors were 47 ± 5/28,600 nM and

16 ± 1/3900 nM, respectively. There was no inhibition of AVP-induced activation of hu-

man V

or oxytocin receptors (K

 17,000 nM).

In rats, (R)-()-51 had excellent oral

bioavailability (F  81%) with an oral plasma half-life of 3.8 h; its oral bioavailability was

also noteworthy in beagle dogs (F  49%) and cynomolgus monkeys (F  62%).

37–50

Compound

Binding

(nM)

Binding

(nM)

Funct.

(µM)

Funct.

(µM)

Cl Me 5-F 100 2.8 2.0 0.012

Cl Ph 5-F ca. 300

11 6.4 0.012

H Ph H ca. 30

3.7 0.016

g,h

Cl Ph H 24 ± 7 0.9 ± 0 0.42 ± 0.05 0.003 ± 0

h,i

Cl Ph H 640 ± 150 13 ± 1

 15

0.17 ± 0.02

Cl 4-MeOPh H IA 2.8 1.9 0.003

Cl 3-MeOPh H IA 3.7 3.4 0.002

Cl 4-OH-Ph H IA 4.2 4.2 0.011

Cl 3-OH-Ph H IA 5.0 11 0.02

Cl Ph 4-F IA 1.9 0.7 0.002

Cl Ph 4-OMe IA 4.2 1.3 0.005

Cl Ph 5-OMe IA 17

Cl Ph 4-OH ca. 30 1.4 1.6 0.006

Cl Ph 5-OH IA 13 1.9 0.07

VPA-985

44 2.3 6.0 0.023

Same as for Table 11.5, except compounds are (S)-() enantiomers (as depicted in the formula) unless noted

otherwise.

Binding to human V

and V

receptors (N  3 to 6). IA



inactive (30% inhibition at 100nM).

Inhibition of functional human V

or V

activity (N  3 to 5).

62% inhibition at 1000 nM.

Racemic mixture.

69% inhibition at 100 nM.

ClN

•HCl•1.3H

O (analyzed correctly for C/H/N/H

O); 98.7% enantiomeric purity by chiral HPLC.

data (V

) for racemic mixture (40  41): 75/2 nM; 1.0/0.01 µM.

HCl salt; (R)-() enantiomer related to 40.

Reference standard.

TABLE 11.6 Vasopressin V

and V

Binding and Functional Data for

Oxazinobenzodiazepines

Oral administration of (R)-()-51 to hydrated conscious rats produced a dose-

dependent aquaretic effect (Table 11.8). For example, at a oral dose of 10 mg/kg it caused

a 1030% increase of urine output over untreated controls, with a 23% reduction of urine

osmolality. For comparison, at an oral dose of 10 mg/kg, OPC-31260 is reported to alter

NOVEL CHEMOTYPES: VARIATIONS ON A THEME 331

51–64

Compound

Binding

(nM)

Binding

(nM)

Funct.

(µM)

o-Cl 2-Ph Me IA 16 0.099

(ⴙ)-51

e,f

o-Cl 2-Ph Me 2900 12 ± 2 0.047 ± 0.005

(ⴚ)-51

o-Cl 2-Ph Me IA 20 0.017

o-Cl 2-Ph H IA 21 0.129

o-Cl 2-Ph i-Pr IA IA

o-Cl 2-Ph CH

IA IA

H 2-Ph Me IA IA

o-Me 2-Ph Me ca. 30 ca. 50 0.077

o-OMe 2-Ph Me IA IA

o-CF

2-Ph Me IA IA

m-Me 2-Ph Me IA IA

m-OMe 2-Ph Me IA ca. 50 0.32

o-Cl 2-Me,5-F Me IA ca. 50

o-Cl 2-Ph,5-F Me IA 17 0.094

o-Cl 2-(4-MeC

) Me IA 15 0.062

o-Cl 2-(4-MeOC

)Me IA IA

24 0.9 0.004

VPA-985

43 2.3 0.023

Target compounds were puriﬁ ed by reverse-phase semi-prep. HPLC and isolated as bis-triﬂ uoroacetate salts,

unless noted otherwise; 95% pure by reverse-phase HPLC/MS (215/254 nm); characterized by ESI-MS, with

selected compounds analyzed by 300-MHz

H NMR. Compounds are racemates unless noted otherwise.

The position of substitution is ortho or meta to the carboxy group.

Binding to human V

and V

receptors (N  2 to 6). IA  inactive (i.e., 30% inhibition of radioligand binding

at 100 nM).

Inhibition of functional human V

activity (N  2 to 5).

(R)-() enantiomer; 99% enantiomeric purity by chiral HPLC.

Di-HCl salt.

98.3% enantiomeric purity by chiral HPLC.

Reference standard.

TABLE 11.7 Vasopressin V

and V

Binding and Functional Data for

Pyrazinobenzodiazepines

332 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

urine output/osmolality by 500%/65%,

23b

and lixivaptan is reported to alter urine

output/osmolality by 450%/70%.

27a

(R)-()-51 was also a very efﬁ cacious aquaretic

agent in beagle dogs: at an oral dose of 1 mg/kg, the values for urine output/osmolality

were 1016%/81%.

Our investigation of the pyrazinobenzodiazepine series was fruitful in that it furnished

(R)-()-51 (JNJ-16240198; RWJ-659834), a highly selective V

-receptor antagonist that

inhibits vasopressin-induced renal water resorption without inducing electrolyte loss in rats

and dogs. On the basis of an assortment of preclinical data, (R)-()-51 was advanced into

development as a backup to oxazinobenzodiazepine 40. Vasopressin V

-selective antago-

nists of this type could be useful for several clinical indications for which marketed diuret-

ics are used, such as congestive heart failure, hypertension, and edema, with the added

beneﬁ t of aquaresis without electrolyte loss. Particular areas for therapeutic application

would be hyponatremia and liver cirrhosis.

11.6 EPILOGUE

Our drug discovery efforts on vasopressin receptor antagonists were initiated in late

1997. To propel this project, we assembled a cross-site collaboration between my

research team in Spring House, Pennsylvania, and a research team in Raritan, New

Jersey. Together, a great drug discovery plan of action was formulated and executed.

Thus, this bilateral relationship turned out to be highly productive, perhaps even be-

yond our original dreams, with ﬁ ve compounds having entered preclinical development.

One of these, 40, progressed successfully through phase IIa clinical trials and another

(not discussed here) advanced to human clinical studies as a mixed V

-receptor

antagonist.

30f,g

In this chapter I adopted a personal perspective and concentrated on the

medicinal chemistry work that was performed by my group in Spring House.

30a–d,44

a similar vein, the excellent medicinal chemistry work of the scientists in Raritan was

also very fruitful.

30e–g

In fact, the combined operations of our two chemistry groups

served to greatly expand the range of structural diversity under exploration and proved

to be very complementary as well. When a chemotype from one side of the Delaware

River entered preclinical development and then fell out (for unforeseen reasons that will

go unmentioned), a chemotype from the other side ﬁ lled the vacuum. This alternation

Oral Dose

(mg/kg)

Urine Volume Urine Osmolality

mL % of Vehicle mOs/kg % of Vehicle

Vehicle 2.3 ± 0.2 692 ± 189

0.3 3.3 ± 0.4 143 368 ± 37

1 4.5 ± 0.8 196 334 ± 34

3 9.4 ± 1.2

409 239 ± 23

10 23.7 ± 1.9

1030 156 ± 12

(R)-()-51 was administered orally to conscious hydrated male rats at the dose speciﬁ ed. Each value represents

the mean ± SE (N  7 to 8). Values representing the percent of vehicle (vehicle  100%) are also provided.

P  0.05 versus vehicle values.

TABLE 11.8 Effect of (R)-(ⴙ)-51 on Urine Volume and Osmolality in Rats

of contributed clinical candidates from each group in the partnership was a powerful

formula for success. In retrospect, the results from our vasopressin-receptor antagonist

research exceeded our original expectations. We applied direct intent, and it seems that

clinical candidates kept coming our way. Point and click—well, perhaps not that easy.

At the end of the day, it has been a very rewarding experience to see our well-consid-

ered research plan come together in an excellent collaborative effort, to “deliver the

goods.”

ACKNOWLEDGMENTS

I am indebted to numerous colleagues at Johnson & Johnson PRD who have played im-

portant roles in our vasopressin antagonist studies; their names appear in the articles cited

in footnote 30. Relative to the work presented here, I want to recognize especially the out-

standing scientiﬁ c leadership afforded by Drs. Patricia Andrade-Gordon, Keith Demarest,

Joseph Gunnet, Dennis Hlasta, and William Hoekstra. Dr. Gunnet was particularly instru-

mental in driving the biology of this project forward. I also thank Jay Matthews and Alexey

Dyatkin for their excellent medicinal chemistry efforts, and Lawrence de Garavilla for his

pharmacology contributions. Jay Matthews’ chemical leadership was critical to the success

of our vasopressin V

antagonist series.

REFERENCES AND NOTES

Maryanoff, B. E. J. Med. Chem. 2004, 47, 769–787.

Thayer, A. M. Chem. Eng. News 2003, 81, 33–37 (July 28). Also see Service, R. F. Science 2004,

303, 1796–1799.

(a) For information on TOPAMAX topiramate, refer to: www.topamax.com. (b) Also see

Maryanoff, B. E.; Nortey, S. O.; Gardocki, J. F.; Shank, R. P.; Dodgson, S. P. J. Med. Chem.

1987, 3, 880–887. Shank, R. P.; Gardocki, J. F.; Vaught, J. L.; Davis, C. B.; Schupsky, J. J.;

Raffa, R. B.; Dodgson, S. J.; Nortey, S. O.; Maryanoff, B. E. Epilepsia 1994, 35, 450–460.

Maryanoff, B. E.; Costanzo, M. J.; Nortey, S. O.; Greco, M. N.; Shank, R. P.; Schupsky, J. J.;

Ortegon, M. P.; Vaught, J. L. J. Med. Chem. 1998, 41, 1315–1343. Shank, R. P.; Gardocki,

J. F.; Streeter, A. J.; Maryanoff, B. E. Epilepsia 2000, 41(Suppl. 1), S3–S9. Garnett, W. R.

Epilepsia 2000, 41 (Suppl. 1), S61–S65. Ben-Menachem, E. Rev. Contemp. Pharmacother.

1999, 10, 163–184. Tomson, T. Curr. Sci. 2002, 82, 698–706. Faught, E. Drugs Today

1999, 35, 49–57. Michelucci, R.; Passarelli, D.; Riguzzi, P.; Volpi, L.; Tassinari, C. A.

CNS Drug Rev. 1998, 4, 165–186. Brandes, J. L.; Saper, J. R.; Diamond, M.; Couch, J. R.;

Lewis, D. W.; Schmitt, J.; Neto, W.; Schwabe, S.; Jacobs, D. J. Am. Med. Assoc. 2004, 291,

965–973.

Costanzo, M. J.; Yabut, S. C.; Almond, H. R., Jr.; Andrade-Gordon, P.; Corcoran, T. W.;

de Garavilla, L.; Kauffman, J. A.; Abraham, W. M.; Recacha, R.; Chattopadhyay, D.; Maryanoff,

B. E. J. Med. Chem. 2003, 46, 3865–3876.

Greco, M. N.; Hawkins, M. J.; Powell, E. T.; Almond, H. R., Jr.; Corcoran, T. W.; de Garavilla, L.;

Kauffman, J. A.; Recacha, R.; Chattopadhyay, D.; Andrade-Gordon, P.; Maryanoff, B. E. J. Am.

Chem. Soc. 2002, 124, 3810–3811. Greco, M. N.; Almond, H. R., Jr.; de Garavilla, L.; Hawkins,

M. J.; Maryanoff, B. E.; Qian, Y.; Walker, D. G.; Cesco-Cancian, S.; Nilsen, C. N.; Patel, M. N.;

Humora, M. J. PCT Intl. Patent Appl. WO 0335654, 2003. A Novel, Potent Dual Inhibitor of the

Leukocyte Proteases Cathepsin G and Chymase: molecular mechanisms and anti-inﬂ ammatory

REFERENCES AND NOTES 333

334 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

activity in vivo. de Garavilla, Lawrence; Greco, Michael N.; Sukumar, Narayanasami; Chen,

Zhi-Wei; Pineda, Agustin O.; Mathews, F. Scott; Di Cera, Enrico; Giardino, Edward C.; Wells,

Grace I.; Haertlein, Barbara J.; Kauffman, Jack A.; Corcoran, Thomas W.; Derian, Claudia K.;

Eckardt, Annette J.; Damiano, Bruce P.; Andrade-Gordon, Patricia; Maryanoff, Bruce E. Drug

Discovery, School of Medicine, Johnson & Johnson Pharmaceutical Research and Development,

Spring House, PA, USA. Journal of Biological Chemistry (2005), 280(18), 18001-18007. Pub-

lisher: American Society for Biochemistry and Molecular Biology, CODEN: JBCHA3 ISSN:

0021-9258. Journal written in English. CAN 143:19271 AN 2005:379217 CAPLUS (Copyright

Hoekstra, W. J.; Maryanoff, B. E.; Andrade-Gordon, P.; Cohen, J. H.; Costanzo, M. J.; Damiano,

B. P.; Haertlein, B.; Harris, B. D.; Kauffman, J. A.; Keane, P. M.; McComsey, D. F.; Villani,

F. J., Jr.; Yabut, S. B. Bioorg. Med. Chem. Lett. 1996, 6, 2371–2376. Hoekstra, W. J.; Poulter,

B. L. Curr. Med. Chem. 1998, 5, 195–204. Hoekstra, W. J.; Maryanoff, B. E.; Damiano, B. P.;

Andrade-Gordon, P.; Cohen, J. H.; Costanzo, M. J.; Haertlein, B. J.; Hecker, L. R.; Hulshizer,

B. L.; Kauffman, J. A.; Keane, P.; McComsey, D. F.; Mitchell, J. A.; Scott, L.; Shah, R. D.;

Yabut, S. C. J. Med. Chem. 1999, 42, 5254–5265. Damiano, B. P.; Mitchell, J. A.; Giardino, E.;

Corcoran, T.; Haertlein, B. J.; de Garavilla, L.; Kauffman, J. A.; Hoekstra, W. J.; Maryanoff,

B. E.; Andrade-Gordon, P. Thromb. Res., 2001, 104, 113–126.

Quinn, M. J.; Byzova, T. V.; Qin, J.; Topol, E. J.; Plow, E. F. Arterioscler. Thromb. Vasc. Biol.

2003, 23, 945–952. Salam, A. M.; Al Suwaidi, J. Expert Opin. Invest. Drugs 2002, 11, 1645–

1658. Quinn, M. J.; Plow, E. F.; Topol, E. J. Circulation 2002, 106, 379–385. Chew, D. P.; Bhatt,

D. L.; Topol, E. J. Am. J. Cardiovasc. Drugs 2001, 1, 421–428. Chew, D. P.; Bhatt, D. L.; Sapp,

S.; Topol, E. J. Circulation 2001, 103, 201–206.

Derian, C. K.; Maryanoff, B. E.; Zhang, H.-C.; Andrade-Gordon, P. Drug Dev. Res. 2003, 59,

355–366. Maryanoff, B. E.; Zhang, H.-C.; Andrade-Gordon, P.; Derian, C. K. Curr. Med. Chem.–

Cardiovasc. Hematol. Agents 2003, 1, 13–36. Derian, C. K.; Maryanoff, B. E.; Zhang, H.-C.;

Andrade-Gordon, P. Expert Opin. Invest. Drugs 2003, 12, 209–221.

Reeves, W. B.; Andreoli, T. E. In Williams Textbook of Endocrinology, 8th ed., Wilson, J. D.;

Foster D. W. (Eds.), W.B. Saunders, Philadelphia, 1992, pp. 311–356.

Nielson, S.; Chou C. L.; Marples, D.; Christensen, E. I.; Kishore, B. K.; Knepper, M. A. Proc.

Natl. Acad. Sci. USA 1995, 332, 1540–1545.

Anderson, R. J.; Chung, H.-M.; Kluge, R.; Schrier, R. W. Ann. Intern. Med. 1985, 102, 164–168.

Chung, H. M.; Kluge, R.; Schrier, R. W.; Anderson, R. J. Arch. Intern. Med. 1986, 146, 333–336.

Sane, T.; Rantakari, K.; Poranen, A.; Taehtelae, R.; Vaelimaeki, M.; Pelkonen, R. J. Clin. Endo-

crinol. Metab. 1994, 79, 1395–1398.

(a) Kortas, C.; Bichet, D. G.; Rouleau, J. L.; Schreier, R. W. J. Cardiovasc. Pharmacol. 1986, 8

(suppl. 7), S107–S110. (b) Johnston, C. I.; Arnolda, L. F.; Tsunoda, K.; Phillips, P. A.; Hodsman,

G. P. Can. J. Physiol. 1987, 65, 1706–1711. (c) Van Zwieten, P. A. Prog. Pharmacol. Clin.

Pharmacol. 1990, 7, 49–66. (d) Thibonnier, M. Curr. Opin. Pharmacol. 2003, 3, 683–687. (e)

Thibonnier, M.; Coles, P.; Thibonnier, A.; Shoham, M. Annu. Rev. Pharmacol. Toxicol. 2001, 41,

175–202.

Mulinari, R. A.; Gavras, I.; Wang, Y. X.; Franko, R.; Gavras, H. Circulation 1990, 81, 308–311.

Naitoh, M.; Suzuki, H.; Murakamj, M.; Matsumoto, A.; Arakawa, K.; Ichihara, A.; Nakamoto,

H.; Oka, K.; Yamamura, Y.; Saruta, T. Am. J. Physiol. 1994, 267, H2245–H2254. Lee, C. R.;

Watkins, M. L.; Patterson, J. H.; Gattis, W.; O’Connor, C. M.; Gheorghiade, M.; Adams, K. F.,

Jr. Am. Heart J. 2003, 146, 9–18.

Martin, P-Y.; Schrier, R. W. Kidney Int. 1997, 59, S43–S49.

Winnicki, C.; Warnawin, K.; Januszewicz, P.; Socha, J. Pol. J. Pediatr. Gastroenterol. Nutr. 1994,

19, 425–430.

Tsuboi, Y.; Ishikawa, S.-E.; Fujisawa, G.; Okada, K.; Saito, T. Kidney Int. 1994, 46, 237–244.

10.

11.

12.

13.

14.

15.

16.

17.

Laszlo, F. A.; Csati, S.; Balaspiri, L. Acta Endocrinol. 1984, 106, 56–60. Reeder, R. F.; Nattie,

E. E.; North, W. G. J. Neurosurg. 1986, 64, 941–950. Rosenberg, G. A.; Kyner, W. T.; Estrada,

E.; Patlak, C. S. Ann. NY Acad. Sci. 1986, 481, 383–389.

Tang, A. H.; Ho, P. M. Life Sci. 1988, 43, 399–403. Laszlo, F. A.; Varga, C.; Balaspiri, L. Ann.

N.Y. Acad. Sci. 1993, 689, 627–629.

(a) Thibonnier, M. Expert Opin. Invest. Drugs 1998, 7, 729–740. (b) Trybulski, E. J. Annu. Rep.

Med. Chem. 2001, 36, 159–168. (c) Albright, J. D.; Chan, P. S. Curr. Pharm. Des. 1997, 3,

615–632. (d) Paranjape, S. B.; Thibonnier, M. Expert Opin. Invest. Drugs 2001, 10, 825–834.

Ruffolo, R. R.; Brooks, D. P.; Huffman, W. F.; Poste, G. Drug News Perspect. 1991, 4, 217–222.

Kinter, L. B.; Caltabiano, S.; Huffman, W. F. Biochem. Pharmacol. 1993, 45, 1731–1737.

(a) Yamamura, Y.; Ogawa, H.; Chihara, T.; Kondo, K.; Onogawa, T.; Nakamura, S.; Mori, T.;

Tominaga, M.; Yabuuchi, Y. Science 1991, 252, 572–574. (b) Ogawa, H.; Yamamura, Y.; Miya-

moto, H.; Kondo, K.; Yamashita, H.; Nakaya, K.; Chihara, T.; Mori, T.; Tominaga, M.; Yabuuchi,

Y. J. Med. Chem. 1993, 36, 2011–2017.

(a) Yamamura, Y.; Ogawa, H.; Yamashita, H.; Chihara, T.; Miyamoto, H.; Nakamura, S.; Onogawa,

T.; Yamashita, H.; Hosokawa, T.; Mori, T.; Tominaga, M.; Yabuuchi, Y. Br. J. Pharmacol. 1992,

105, 787–791. (b) Ogawa, H.; Yamashita, H.; Kondo, K.; Yamamura, Y.; Miyamoto, H.; Kan,

K.; Kitano, K.; Tanaka, M.; Nakaya, K.; Nakamura, S.; Mori, T.; Onogawa, T.; Tominaga, M.;

Yabuuchi, Y. J. Med. Chem. 1996, 39, 3547–3555.

Manning, M.; Sawyer, W. H. J. Receptor Res. 1993, 13, 195–214. Laszlo, F. A.; Laszlo, F., Jr.; De

Weid, D. Pharmacol. Rev. 1991, 43, 73–108. Manning, M.; Chan, W. Y.; Sawyer, W. H. Regul.

Peptides 1993, 45, 279–283.

Fujisawa, G.; Ishikawa, S.; Tsuboi, Y.; Okada, K.; Saito, T. Kidney Int. 1993, 44, 19–23. Ohnishi,

A.; Orita, Y.; Takagi, N.; Fujita, T.; Toyoki, T.; Ihara, Y.; Yamamura, Y.; Inoue, T.; Tanaka, T. J.

Pharmacol. Exp. Ther. 1995, 272, 546–551.

Kondo, K.; Ogawa, H.; Yamashita, H.; Miyamoto, H.; Tanaka, M.; Nakaya, K.; Kitano, K.;

Yamamura, Y.; Nakamura, S.; Onogawa, T.; Mori, T.; Tominaga, M. Bioorg. Med. Chem. 1999,

7, 1743–1757.

(a) Albright, J. D.; Reich, M. F.; Delos Santos, E. G.; Dusza, J. P., Sum, F.-W.; Venkatesan, A. M.;

Coupet, J.; Chan, P. S.; Ru, X.; Mazandarani, H., Bailey, T. J. Med. Chem. 1998, 41, 2442–2444.

(b) Aranapakam, V.; Albright, J. D.; Grosu, G. T.; Chan, P. S.; Coupet, J.; Saunders, T.; Ru, X.;

Mazandarani, H. Bioorg. Med. Chem. Lett. 1999, 9, 1733–1736. (c) Aranapakam, V.; Albright,

J. D.; Grosu, G. T.; Delos Santos, E. G.; Chan, P. S.; Coupet, J.; Ru, X.; Saunders, T.; Mazanda-

rani, H. Bioorg. Med. Chem. Lett. 1999, 9, 1737–1740. (d) Albright, J. D.; Delos Santos, E. F.;

Dusza, J. P.; Chan, P. S.; Coupet, J., Ru, X.; Mazandarani, H. Bioorg. Med. Chem. Lett. 2000, 10,

695–698. (e) Ashwell, M. A.; Bagli, J. F.; Caggiano, T. J.; Chan, P. S.; Molinari, A. J.; Palka, C.;

Park, C. H.; Rogers, J. F.; Sherman, M.; Trybulski, E. J.; Williams, D. K. Bioorg. Med. Chem.

Lett. 2000, 10, 783–786.

(a) Norman, P.; Leeson, P. A.; Rabasseda, X.; Castaner, J.; Castaner, R. M. Drugs Future 2000,

25, 1121–1130. Matsuhisa, A.; Taniguchi, N.; Koshio, H.; Yatsu, T.; Tanaka, A. Chem. Pharm.

Bull. 2000, 48, 21–31. (b) Shimada, Y.; Taniguchi, N.; Matsuhisa, A.; Sakamoto, K.; Yatsu, T.;

Tanaka, A. Chem. Pharm. Bull. 2000, 48, 1644–1651. Matsuhisa, A.; Kikuchi, K.; Sakamoto, K.;

Yatsu, T.; Tanaka, A. Chem. Pharm. Bull. 1999, 47, 329–339.

(a) Cho, H.; Murakami, K.; Nakanishi, H.; Fujisawa, A.; Isoshima, H.; Niwa, M.; Hayakawa, K.;

Hase, Y.; Uchida, I.; Watanabe, H.; Wakitani, K.; Aisaka, K. J. Med. Chem. 2004, 47, 101–109.

(b) Shimada, Y.; Taniguchi, N.; Matsuhisa, A.; Yatsu, T.; Tahara, A.; Tanaka, A. Chem. Pharm.

Bull. 2003, 51, 1075–1080. (c) Sum, F.-W.; Dusza, J.; Delos Santos, E.; Grosu, G.; Reich, M.;

Du, X.; Albright, J. D.; Chan, P.; Coupet, J.; Ru, X.; Mazandarani, H.; Saunders, T. Bioorg. Med.

Chem. Lett. 2003, 13, 2195–2198. (d) Tsukada, J.; Tahara, A.; Tomura, Y.; Wada, K.; Kusayama,

T.; Ishii, N.; Aoki, M.; Yatsu, T.; Uchida, W.; Taniguchi, N.; Tanaka, A. Eur. J. Pharmacol. 2002,

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

REFERENCES AND NOTES 335

336 DISCOVERY OF POTENT NONPEPTIDE VASOPRESSIN RECEPTOR ANTAGONISTS

446, 129–138. (e) Kakefuda, A.; Suzuki, T.; Tobe, T.; Tsukada, J.; Tahara, A.; Sakamoto, S.;

Tsukamoto, S. J. Med. Chem. 2002, 45, 2589–2598. (f) Ohtake, Y.; Naito, A.; Hasegawa, H.;

Kawano, K.; Morizono, D.; Taniguchi, M.; Tanaka, Y.; Matsukawa, H.; Naito, K.; Oguma, T.;

Ezure, Y.; Tsuriya, Y. Bioorg. Med. Chem. 1999, 7, 1247–1254.

(a) Matthews, J. M.; Greco, M. N.; Hecker, L. R.; Hoekstra, W. J.; Andrade-Gordon, P.;

de Garavilla, L.; Demarest, K. T.; Ericson, E.; Gunnet, J. W.; Hageman, W.; Look, R.; Moore,

J. B.; Maryanoff, B. E. Bioorg. Med. Chem. Lett. 2003, 13, 753–756. (b) Dyatkin, A. B.;

Hoekstra, W. J.; Hlasta, D. J.; Andrade-Gordon, P.; de Garavilla, L.; Demarest, K. T.; Gunnet, J.

W.; Hageman, W.; Look, R.; Maryanoff, B. E. Bioorg. Med. Chem. Lett. 2002, 12, 3081–3084.

[Corrigendum: Bioorg. Med. Chem. Lett. 2004, 14, 3363] (c) Matthews, J. M.; Hoekstra, W. J.;

Dyatkin, A. B.; Hecker, L. R.; Hlasta, D. J.; Poulter, B. L.; Andrade-Gordon, P.; de Garavilla, L.;

Demarest, K. T.; Ericson, E.; Gunnet, J. W.; Hageman, W.; Look, R.; Moore, J. B.; Reynolds,

C. H.; Maryanoff, B. E. Bioorg. Med. Chem. Lett. 2004, 14, 2747–2752. (d) Matthews, J. M.;

Hlasta, D. J.; Andrade-Gordon, P.; Demarest, K. T.; Ericson, E.; Gunnet, J. W.; Hageman, W.;

Look, R.; Moore, J. B.; Maryanoff, B. E. Lett. Drug Design Discov. 2005, 2, 219–223. (e) Ur-

banski, M. J.; Chen, R. H.; Demarest, K. T.; Gunnet, J.; Look, R.; Ericson, E.; Murray, W. V.;

Rybczynski, P. J.; Zhang, X. Bioorg. Med. Chem. Lett. 2003, 13, 4031–4034. (f) Xiang, M. A.;

Chen, R. H.; Demarest, K. T.; Gunnet, J.; Look, R.; Hageman, W.; Murray, W. V.; Combs, D. W.;

Patel, M. Bioorg. Med. Chem. Lett. 2004, 14, 2987–2989. (g) Xiang, M. A.; Chen, R. H.; Demar-

est, K. T.; Gunnet, J.; Look, R.; Hageman, W.; Murray, W. V.; Combs, D. W.; Rybczynski, P. J.;

Patel, M. Bioorg. Med. Chem. Lett. 2004, 14, 3143–3146. (h) RWJ-351647 and RWJ-659834 are

the subject of an issued U.S. patent (Mar. 30, 2004): Dyatkin, A. B.; Hoekstra, W. J.; Maryanoff,

B. E.; Matthews, J. M. U.S. patent 6,713,475 (2004) [WO 0043398; EP 1147115].

Sorbera, L. A.; Castaner, J.; Bayes, M.; Silvestre, J. Drugs Future 2002, 27, 350–357.

Gheorghiade, M.; Niazi, I.; Ouyang, J.; Czerwiec, F.; Kambayashi, J.; Zampino, M.; Orlandi, C.

Circulation 2003, 107, 2690–2696.

Martinez-Castelao, A. Curr. Opin. Invest. Drugs 2001, 2, 525–530. Chan, P. S.; Coupet, J.; Park,

H. C.; Lai, F.; Hartupee, D.; Cervoni, P.; Dusza, J. P.; Albright, J. D.; Ru, X.; Mazandarani, H.;

Tanikella, T.; Shepherd, C.; Ochalski, L.; Bailey, T.; Lock, T. Yeung W.; Ning, X.; Taylor, J. R.;

Spinelli, W. Adv. Exp. Med. Biol. 1998, 449, 439–443.

Martinez-Castelao, A. Curr. Opin. Invest. Drugs 2002, 3, 89–95. Tahara, A.; Tomura, Y.; Wada,

K.; Kusayama, T.; Tsukada, J.; Takanashi, M.; Yatsu, T.; Uchida, W.; Tanaka, A. J. Pharmacol.

Exp. Ther. 1997, 282, 301–308. Gerbes, A. L.; Guelberg, V.; Gines, P.; Decaux, G.; Gross, P.;

Gandjini, H.; Djian, J. Gastroenterology 2003, 124, 933–939.

Martinez-Castelao, A. Curr. Opin. Invest. Drugs 2001, 2, 1423–1427. Serradeil-Le Gal,

C. Cardiovasc. Drug Rev. 2001, 19, 201–214. Jimenez, W.; Serradeil-Le Gal, C.; Ros, J.; Cano,

C.; Cejudo, P.; Morales-Ruiz, M.; Arroyo, V.; Pascal, M.; Rivera, F.; Maffrand, J.-P.; Rodes,

J. J. Pharmacol. Exp. Ther. 2000, 295, 83–90. Serradeil-Le Gal, C.; Wagnon, J.; Garcia, C.;

Lacour, C.; Guiraudou, P.; Christophe, B.; Villanova, G.; Nisato, D.; Maffrand, J. P.; Le Fur,

G.; Guillon, G.; Cantau, B.; Barberis, C.; Trueba, M.; Ala, Y.; Jard, S. J. Clin. Invest. 1993,

92, 224–231. Serradeil-Le Gal, C.; Lacour, C.; Valette, G.; Garcia, G.; Foulon, L.; Galindo,

G.; Bankir, L.; Pouzet, B.; Guillon, G.; Barberis, C.; Chicot, D.; Jard, S.; Vilain, P.; Garcia, C.;

Marty, E.; Raufaste, D.; Brossard, G.; Nisato, D.; Maffrand, J. P.; Le Fur, G. J. Clin. Invest.

1996, 98, 2729–2738. Griebel, G.; Simiand, J.; Serradeil-Le Gal, C.; Wagnon, J.; Pascal, M.;

Scatton, B.; Maffrand, J.-P.; Soubrie, P. Proc. Natl. Acad. Sci. USA 2002, 99, 6370–6375.

Gard, P. Curr. Opin. Oncol. Endocrinol. Metabol. Invest. Drugs 2000, 2, 265–272.

(a) Barberis, C.; Morin, D.; Durroux, T.; Mouillac, B.; Guillon, G.; Seyer, R.; Hibert, M.;

Tribollet, E.; Manning, M. Drug News Perspect. 1999, 12, 279–292. (b) Barberis, C., Seibold,

A., Ishido, M., Rosenthal, W., Birnbaumer, M. Regul. Peptides 1993, 45, 61–69. (c) Thibonnier,

M.; Auzan, C.; Madhun, Z; Wilkins, P.; Berti-Mattera, L.; Clauser, E. J. Biol. Chem. 1994, 269,

3304–3310. (d) Hirasawa, A.; Shibata, K.; Kotosai, K.; Tsujimoto, G. Biochem. Biophys. Res.

30.

31.

32.

33.

34.

35.

Commun. 1994, 203, 72–79. (e) Morel, A.; Lolait, S. J.; Brownstein, M. J. Regul. Peptides 1993,

45, 53–59. (f) Morel, A.; O’Carroll, A. M.; Brownstein, M. J.; Lolait, S. J. Nature 1992, 356,

523–526. (g) Lolait, S. J.; O’Carroll, A. M.; McBride, O. W.; Konig, M.; Morel, A.; Brownstein,

M. J. Nature 1992, 357, 336–339.

Tahara, A.; Saito, M.; Sugimoto, T.; Tomura, Y.; Wada, K.; Kusayama, T.; Tsukada, J.; Ishii, N.;

Yatsu, T.; Uchida, W.; Tanaka, A. Br. J. Pharmacol. 1998, 125, 1463–1470.

Oprea, T. I. J. Comput.-Aided Mol. Design 2000, 14, 251–264. Walters, W. Patrick; A.; Murcko,

M. A. Curr. Opin. Chem. Biol. 1999, 3, 384–387. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.;

Feeney, P. J. Adv. Drug Deliv. Rev. 1997, 23, 3–25.

Hester, J. B., Jr. J. Org. Chem. 1964, 29, 1158–1160. Hester, J. B., Jr. J. Org. Chem. 1967, 32,

4095–4098. Nagasaka, T.; Ohki, S. Chem. Pharm. Bull. 1977, 25, 3023–3033.

Bergman, J.; Sand, P. Org. Synth. 1987, 65, 146–149. Quick, J.; Saha, B. Tetrahedron Lett. 1994,

35, 8553–8556.

There are some errors involving the interchange of compound numbers 4 and 5 in ref 30b, which

are reported in the noted Corrigendum.

Abraham, H.; Stella, L. Tetrahedron 1992, 48, 9707–9718. Bertilsson, S. K.; Ekegren, J. K.;

Modin, S. A.; Andersson, P. G. Tetrahedron 2001, 57, 6399–6406.

Carabateas, P. M. U.S. patent 3,860,600 (1975); Chem. Abstr. 1975, 83, 58892.

Ohtake, Y.; Naito, A.; Naito, K.; Matsukawa, H.; Saito, Y.; Toyofuku, H. PCT Int. Appl. 1998,

1–156 (WO 9843976); Chem. Abstr. 1998, 129, 290149.

(a) Dyatkin, A. B.; Freedman, T. B.; Cao, X.; Dukor, R. K.; Maryanoff, B. E.; Maryanoff, C. A.;

Matthews, J. M.; Shah, R. D.; Naﬁ e, L. A. Chirality 2002, 14, 215–219. (b) Matthews, J. M.;

Dyatkin, A. B.; Evangelisto, M.; Gauthier, D. A.; Hecker, L. R.; Hoekstra, W. J.; Poulter, B. L.;

Maryanoff, B. E. Tetrahedron: Asymmetry 2004, 15, 1259–1267.

SciFinder, Advanced chemistry development software, Solaris v4.67.

Hassner, A.; Amit, B.; Marks, V.; Gottlieb, H. E. J. Org. Chem. 2003, 68, 6853–6858.

Gieldon, A.; Kazmierkiewicz, R.; Slusarz, R.; Ciarkowski, J. J. Comput.-Aided Mol. Design

2001, 15, 1085–1104.

(R)-()-51 has the same spatial arrangement at the stereogenic center as does 40; however, the

stereochemical descriptor for the absolute conﬁ guration of ()-51 [“R”] differs from that of 40

[“S”] because of the Cahn–Ingold–Prelog priority rules: Cahn, R. S.; Ingold, C.; Prelog, V. An-

gew. Chem. Int. Ed. 1966, 5, 385–415.

Inhibition of AVP activation of human V

or oxytocin receptors in transfected HEK-293 cells.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

REFERENCES AND NOTES 337

339

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

DISCOVERY AND DEVELOPMENT

OF THE ULTRASHORT-ACTING

ANALGESIC REMIFENTANIL

PAUL L. FELDMAN

GlaxoSmithKline Research and Development

Research Triangle Park, North Carolina

12.1 INTRODUCTION

To the layman general anesthesia is archetypically viewed as a patient gently falling asleep

following the doctor or nurse placing a mask over the patient’s nose and mouth. The patient

loses consciousness and the surgeons begin their work. To the practitioner general anesthe-

sia is a state induced by a combination of agents that induce hypnosis, the sleep component,

along with amnesia, analgesia, and muscle relaxation (Fig. 12.1). The combinations of

agents that produce these effects facilitate surgical or procedural intervention at minimal

risk to the patient while providing for optimal recovery.

Hypnotics, such as propofol, induce sleep and a lack of awareness. Analgesics such as

fentanyl and remifentanil synergize with hypnotics to provide profound hypnosis as well as

to blunt the severe pain a patient may realize as a result of the surgical procedure. Amnes-

tics such as midazolam are anxiolytics and also provide for the absence of recall of surgical

events. It is most often the experience of patients undergoing a surgical event that they can-

not recall anything that took place after being placed under anesthesia. Muscle relaxants

such as cis-atracurium and succinylcholine disrupt the nerve to muscle signal transmission

and are used clinically as adjuncts to anesthesia to facilitate intubations and relax muscles.

The muscle relaxation, or paralysis, caused by these agents allow a surgeon to more easily

manipulate a patient undergoing a surgical procedure. It is the skillful combination of these

agents administered to patients by anesthesiologists that provides for humane and tolerable

surgical interventions.

340

DISCOVERY AND DEVELOPMENT OF THE ULTRASHORT-ACTING ANALGESIC REMIFENTANIL

Anesthesiologists require agents that are extremely safe and predictable because they

service a large and varied patient population. They must deliver agents that provide the ben-

eﬁ ts of anesthetic care while ensuring that the anesthetic agents themselves will manifest

minimal risk. Typically, the anesthesiologist titrates the anesthetic drugs throughout a surgi-

cal procedure depending on the manipulation the surgeon is performing on the patient. For

example, the anesthesiologist may need to deepen the level of hypnotic and analgesic when

a patient is undergoing a particularly painful part of the surgical procedure. Once that part

of the procedure is completed the anesthesiologist may lighten the anesthesia so that at the

end of the procedure they can provide for a rapid recovery of the patient. If a deep anesthetic

state is maintained throughout a surgical event, it may compromise a rapid recovery and

leave lingering effects of the anesthesia long after the patient has left the operating room.

One of the limitations of many anesthetic drugs is that they accumulate during a proce-

dure and this prolongs a patient’s recovery from the surgery and anesthesia. A goal of the

pharmaceutical industry has been to provide new anesthetic agents that have an ultrashort

duration of action so that there is a rapid and predictable recovery of patients. To date, there

are no clinically used ultrashort-acting hypnotics or amnestics.

The most popular hypnotics

and amnestics are short-acting, but if given over a prolonged period of time, they will accu-

mulate and prolong patient recoveries. Succinlycholine is an ultrashort-acting muscle relax-

ant that is used routinely given its favorable pharmacokinetic properties; however, it suffers

some serious pharmacodynamic limitations that preclude it being the ideal ultrashort-acting

muscle relaxant.

The introduction of remifentanil, Ultiva, provides anesthesiologists with

the analgesic component of anesthetic care that delivers a strong analgesic with a rapid

onset of action, ultrashort and predictable duration of action, and no cumulative effects.

Hopefully, over the next several years, safe, effective, and ultrashort-acting hypnotics, mus-

cle relaxants, and amnestics will be discovered and progressed to provide anesthesiologists

with the other agents they need to complement remifentanil in the anesthesia quadrad.

12.2 DISCOVERY OF REMIFENTANIL

Based on the premise that the clinically used analgesic opiate anesthetic agents were highly

efﬁ cacious yet suffered from pharmacokinetic liabilities, we began a research program

Anesthesia

Care

Hypnosis: sleep,

absence of awareness

(propofol)

Analgesia: freedom of pain

or response to surgical stimuli

(remifentanil)

Muscle Relaxation: facilitate intubation

and surgical access/manipulation

(cis-atracurium)

Amnesia: absence of recall

(midazolam)

Figure 12.1 Anesthesia quadrad.

to discover an analgesic opiate with an ultrashort duration of action for use as an adjunct

to anesthesia.

The popular opiate analgesics used as adjuncts to anesthesia are fentanyl,

sufentanil, and alfentanil (Scheme 12.1). All of the analgesics are potent µ-opioid ago-

nists that were discovered by Paul Janssen and co-workers at Janssen Pharmaceutica in the

1960s and 1970s. Our goal was to discover a drug with µ-opioid agonist properties, such

as fentanyl and its congeners, but which would undergo extremely rapid metabolism and

elimination. The rapid metabolism of the drug would be independent of liver and kidney

function and the rate of metabolism would not change with prolonged or high-dose admin-

istration. The importance of discovering a drug that undergoes metabolism independent of

liver or kidney function is directly related to the predictability of drug half-life. Anesthetics

that rely on the liver or kidney for clearance may have varied durations of action across

patient populations because the functions of these organs may differ depending on the state

of health of a patient. We sought to discover a drug whose metabolism would be absolutely

predictable regardless of the health status of the patient who received the agent.

The design of potent µ-opioid analgesics that are ultrashort acting was based on the

hypothesis that an ester incorporated into a fentanyl motif would retain the potent µ-opioid

agonist properties. In addition, the ester functional group would rapidly be metabolized

to an acid by ubiquitous blood and tissue esterases and would render the new molecule

inactive at the µ-opioid receptor. Inspecting the structure of fentanyl and its analogs, we

noted that the most potent µ-opioid agonists in this structural class possessed a two-carbon

chain attached to a lipophilic moiety pendant to the piperidine nitrogen. We reasoned that

the lipophilic group, a phenyl group in fentanyl, docked into a lipophilic portion of the

µ-opioid receptor. Replacement of the aryl group with an ester function, also a lipophilic

group, would provide for potent µ agonists. However, once the ester group was hydrolized

to an acid by blood or tissue esterases, the polar and charged acid function would lose sig-

niﬁ cant afﬁ nity for the µ receptor due to unfavorable interactions. One attractive feature to

this design strategy is the use of ubiquitous esterases to metabolize and inactivate the anal-

gesic versus reliance on the liver or kidney to inactivate or clear the drug. Given the broad

distribution of esterases in humans, it was hypothesized that the proposed mechanism of

metabolism would be operative in all patient populations.

The design we proposed had precedent in the medicinal chemistry literature

(Scheme 12.2). In the early 1980s Erhardt and co-workers published their successful

Scheme 12.1 Known opiate analgesics.

Fentanyl

Sufentanil

Alfentanil

NCO

Carfentanil

DISCOVERY OF REMIFENTANIL 341

342

DISCOVERY AND DEVELOPMENT OF THE ULTRASHORT-ACTING ANALGESIC REMIFENTANIL

approach to generating ultra short-acting beta blockers.

They replaced one of the aryl

rings of propanolol with a ethyl propionate moiety. Esmolol was found to be a potent and

effective ultrashort-acting beta blocker that was commercialized and used in the hospital

setting. Just after we had begun our work, two additional reports were published describing

similar approaches to generate ultrashort-acting antiarrythmics and opioid analgesics.

7,8

Unfortunately, the approach published using the fentanyl scaffold to generate ultrashort-

acting analgesics did not succeed. A possible explanation for the failure of this approach is

that the esters were hydrolized too slowly given the sterically hindered nature of the ester

moiety. In the successful approaches using this strategy, the esters are unhindered and thus

readily accepted as substrates for esterase enzymes. This general approach to discover-

ing drugs, active agents that are predictably metabolized to inactive metabolites, has been

coined soft drugs and has been applied to several different therapeutic areas.

The testing scheme we used to evaluate our opioid analgesics is shown in Fig. 12.2.

Initially, molecules were tested for their µ-opioid afﬁ nity and efﬁ cacy in the well-charac-

terized and standard in vitro guinea pig ileum assay. Compounds that were shown to be

potent µ-opioid agonists were then tested in the rat tail withdrawal reﬂ ex model. In this in

vivo model compounds were tested for their analgesic efﬁ cacy as well as their duration of

action. Ideally, we wished to discover compounds as potent as fentanyl and its congeners

both in vitro and in vivo, but with durations of action in the rat signiﬁ cantly less than

fentanyl. For those compounds with the desired proﬁ le in the primary assays, the in vitro

Scheme 12.2 Medicinal chemistry approach.

Esterases

Highly Potent Inactive

Figure 12.2 Testing scheme.

Synthesis

In Vitro

Guinea Pig Ileum

In Vivo

Rat Withdrawal Reflex Model

In Vitro

Human Whole Blood Hydrolysis Assay

half-lives in human whole blood were determined to assess the ability of the compounds to

be metabolized by esterases in human blood.

Our initial attempts to generate potent µ-opioid agonists were disappointing. When the

phenethyl group of fentanyl was replaced by acetate, propionate, and butyrate esters, the

compounds were approximately 1000-fold less active than fentanyl in the guinea pig ileum

assay. Although these compounds were weak µ-opioid agonists, we were encouraged to

pursue this approach given that these esters possessed the desired duration of action in the

rat. Compared to fentanyl, these compounds were approximately four to six times shorter

acting. Therefore, to discover the drug candidate with the desired proﬁ le, we needed to

retain the structural features of these initial compounds, which gave analgesics with an

ultrashort duration of action, but we needed to increase their µ-opioid agonist potency.

Janssen and co-workers had shown that the hydrogen at the 4-position of the piperidine

portion of fentanyl could be substituted with a methyl ester (cf. carfentanil) or methoxy

methyl group (cf. sufentanil and alfentanil) to obtain very potent µ-opioid agonists. Indeed,

we found that replacement of the hydrogen with a methyl ester had a profound affect on

the µ-opioid agonist properties in our series of molecules. Replacement of the phenethyl

group of carfentanil with a methyl propionate provided an extremely potent µ-opioid ago-

nist with an ultrashort duration of action. This compound, remifentanil, was approximately

twofold less potent than fentanyl in vitro (EC

3.5 nM versus 1.8 nM in the guinea pig

ileum assay), equipotent to fentanyl in the rat tail withdrawal reﬂ ex model (ED

4.5 µg/kg

for remifentanil versus 4.6 µg/kg for fentanyl), and was four times shorter acting (15-min

versus 60-min duration) in rat tail withdrawal reﬂ ex model (Table 12.1).

We hypothesized that the ultrashort duration of action that remifentanil possessed in the

rat is due to its hydrolysis to an inactive acid metabolite. Indeed, in subsequent biometabo-

lism studies we demonstrated that remifentanil is very rapidly and nearly quantitatively

metabolized to its acid metabolite in the rat. In addition to the ultrashort duration of action

in the rat, we also discovered that remifentanil is rapidly hydrolized to its acid metabolite

TABLE 12.1 Structure–Activity Relationships for Ultrashort-Acting Analgesics

R ⫽ HR ⫽ CO

1.27 ± 0.57 ⫻ 10

⫺5

5.39 ± 1.02 ⫻ 10

⫺6

1.66 ± 0.59 ⫻ 10

⫺6

M (3.2/10 ⫺ 15)

3.55 ± 0.23 ⫻ 10

⫺9

M (0.0044/15)

3.60 ± 0.30 ⫻ 10

⫺6

M 1.50 ± 0.98 ⫻ 10

⫺8

5.44 ± 1.51 ⫻ 10

⫺6

M 1.02 ± 0.63 ⫻ 10

⫺8

Fentanyl

1.76 ± 0.36 ⫻ 10

⫺9

M (0.0046/60)

values for inhibition of electrically evoked contraction in guinea pig ileum.

values (mg/kg) for analgesic effects in rat tail withdrawal assay/duration of in vivo effect in minutes.

DISCOVERY OF REMIFENTANIL

343

344

DISCOVERY AND DEVELOPMENT OF THE ULTRASHORT-ACTING ANALGESIC REMIFENTANIL

in vitro in human whole blood. The acid metabolite was synthesized and tested for its

µ-opioid in vitro and in vivo and efﬁ cacy. Gratifyingly, the acid metabolite of remifentanil

was shown to be more than 500-fold less active than remifentanil in the in vitro guinea pig

ileum assay and greater than 350-fold less potent in the rat tail withdrawal reﬂ ex model.

From these data we surmised that the acid has a signiﬁ cantly lower afﬁ nity for the µ-opioid

receptor due to the polar and ionized functional group being unable to bind effectively into

a lipophilic pocket in the µ-opioid receptor.

It was essential to demonstrate in animals that remifentanil has a signiﬁ cantly shorter

duration of action than fentanyl and its congeners after a single dose. In addition, it was

just as signiﬁ cant to demonstrate that upon prolonged infusion of remifentanil the recov-

ery of animals would be just as rapid as it was when given a single dose. To test whether

remifentanil would accumulate upon prolonged administration, rats were infused with high

doses of remifentail for 1 hour and the time it took to recover was measured. Indeed, even

after the 1-hour infusion the animals recovered from the analgesic affects of remifentanil

in the same time frame as if given a single bolus dose. Also, an experiment was conducted

where rats were given multiple successive bolus doses of remifentanil to see if the recov-

ery from its effects would be prolonged when treated successively with the drug. Again,

no prolongation of duration was noted, suggesting that the mechanism for metabolism of

remifentanil was not saturated and accumulation of the drug was not signiﬁ cant. All of

these preclinical data suggested that we had discovered a potent µ-opioid agonist that had

an ultrashort duration of action based on a liver- and kidney-independent metabolic path-

way to a signiﬁ cantly less active metabolite.

12.3 CHEMICAL DEVELOPMENT OF REMIFENTANIL

The synthetic route we employed to generate remifentanil and its analogs is patterned after

the route developed by Janssen and co-workers (Scheme 12.3).

The route began with the

Strecker reaction performed on N-benzyl-4-piperidinone to generate 2. Conversion of the

nitrile to the amide proved to be capricious, due to the reversion of the amino nitrile to N-

benzyl-4-piperidinone and decomposition depending on acid strength, reaction time, and

temperature. Carefully controlled hydrolysis of the nitrile to amide could be accomplished

with 85% sulfuric acid. Conversion of the primary amide to the acid was accomplished

under harsh basic conditions. Following conversion of the potassium salt to the sodium

salt the methyl ester was made by alkylating the sodium carboxylate with methyl iodide.

In addition to the desired product, this reaction also produced very polar material believed

to be the quaternary ammonium salt. This undesired by-product could easily be separated

from the desired ester by extractive workup; however, the ester was isolated as a viscous

oil. Acylation of the aniline nitrogen was accomplished by heating 5 in neat propionic

anhydride at 167⬚C followed by crystallization of the product as its oxalic acid salt. The

benzyl group was removed using high-pressure hydrogenation, and the secondary amine

was immediately subjected to a Michael reaction with methyl acrylate to form the free base

of remifentanil. The white crystalline hydrochloride salt of remifentanil was formed with

HCl in MeOH and ether.

Although the medicinal chemistry route to remifentanil is relatively short, there were a

number of problems that prevented the use of this route on a large scale (Scheme 12.4). In

the ﬁ rst stage the use of KCN in acetic acid generates HCN, which is a safety issue when

this reaction is performed on a large scale. In the hydrolysis of the nitrile to the primary

Scheme 12.3 Medicinal chemistry route to remifentanil.

KCN, PhNH

HOAc, H

20–25

°C, 2h

(71%)

CNHN

85% H

45 °C, 24h

(66%)

CONH

NaHN

1. 4M KOH

ethylene glycol

150

°C, 8h

2. conc. HCl

3. 10M NaOH

(93%)

(MeO)

DMF

°C, 0.75h

(72%)

MeHN

1. (EtCO)

167

°C, 3h

2. (HO

MeOH

(80%)

MeN

(50 psi)

10% Pd-C

MeOH/HOAc

20–25

°C, 24h

(99%)

MeN

1. CH

CHCO

MeOH, 20–25 °C, 16h

2. HCl, MeOH/Et

(89%)

MeN

23 4

Remifentanil-HCl

HCl

345

Scheme 12.4 Manufacturing considerations regarding the medicinal chemistry route.

Generation of HCN

CNHN

Neutralize

large volumes

of H

CONH

NaHN

Corrosive

materials used

at high temperatures

Quaternization

of piperidine

nitrogen and

noncrystalline

intermediate

MeHN

High temperature

acylation reaction

MeN

High pressure

hydrogenation

and unstable

intermediate

MeN

23 4

Remifentanil-HCl

HCl

346

amide the product was isolated as an oil after neutralizing large volumes of sulfuric acid.

This method of isolation proved to be impractical on a large scale. Subsequent hydrolysis

of the amide to acid 4 using KOH in ethylene glycol at 150⬚C needed to be avoided given

the corrosive nature of these reagents at high temperature, which would be difﬁ cult to oper-

ate on a large scale. Conversion of the acid to the methyl ester also needed to be modiﬁ ed

to avoid both isolation of the noncrystalline ester 5 and quaternization of the piperidine ni-

trogen. Milder conditions were also sought for the acylation of the ester 5 to give 6 as well

as the removal of the benzyl group to give 7. Avoiding the high-pressure hydrogenation to

remove the benzyl group would signiﬁ cantly simplify the manufacture of remifentanil. All

of these issues needed to be addressed to accomplish a cost-effective, simpliﬁ ed, robust,

and safe route to remifentanil.

The ultimate manufacturing route to remifentanil is depicted (in Scheme 12.5). Except

for stage 4, the general synthetic route to remifentanil is similar to the medicinal chemistry

route. However, a number of signiﬁ cant improvements were made which enhanced the

safety, robustness, and scaling of the reactions. In the ﬁ rst stage (the Strecker reaction), by

using acetone cyanohydrin, the cyanide is slowly released in a controlled fashion and thus

avoids the expulsion of gaseous HCN from the reaction. To avoid a complex workup of a

toxic mixture, a solvent from which the product would crystallize directly was sought. Use

of the organic reagent cyanide (acetone cyanohydrin) avoided the presence of metal salts,

and by using industrial mineral spirits the crystallization occurred directly, thus driving

the equilibrium in favor of the desired product (2), which crystallizes from the reaction,

Scheme 12.5 Remifentanil manufacturing route.

(CH

)

C(OH)CN

PhNH

95% EtOH, 5% MeOH

40–50

°C

(86%)

CNHN

1. 90% H

35 °C, 20–30h

2. H

(89%)

CONH

1. conc. HCl

reflux, 6h

2. H

(81%)

1. (EtCO)

O, EtOAc,

reflux, 16h

2. 4, TEA, 25

°C to reflux

3. MeOH, 70

°C to reflux

4. oxalic acid

5. MeOH recrystallization

(78%)

MeN

1. K

, H

O, EtOAc

2. 1M H

10% Pd-C, H

(1 atm)

3. pH 8.5 (NH

OH)

CHCO

, 20 °C, 18h

4. isopropyl acetate, HCl

5. MeOH, isopropyl acetate

recrystalization

(56%)

MeN

Remifentanil-HCl

HCl

(HO

CHEMICAL DEVELOPMENT OF REMIFENTANIL 347

348

DISCOVERY AND DEVELOPMENT OF THE ULTRASHORT-ACTING ANALGESIC REMIFENTANIL

mixture in very high purity. Hydrolysis of the nitrile to the primary amide (3) was ac-

complished with sulfuric acid, as had been done in the medicinal chemistry route, and the

product was crystallized from the reaction mixture by the careful addition of water to the

reaction once the hydrolysis had been completed. Isolation of the bis-sulfate salt simpliﬁ ed

this reaction workup considerably versus the tedious neutralization of sulfuric acid that

had been standard in the medicinal chemistry process. Hydrolysis of the amide to the acid

was accomplished using concentrated HCl versus the strongly basic conditions used in the

medicinal chemistry scheme. Following completion of the hydrolysis, the crude product is

ﬁ ltered, then slurried with water, and the monohydrochloride salt was isolated as a crystal-

line solid.

The most signiﬁ cant and elegant change to the medicinal chemistry route is the one-

stage conversion of acid 4 to ester 6

(Scheme 12.6). In this step the acid is esteriﬁ ed and

the aniline nitrogen is acylated under mild conditions. By using this procedure we have

obviated the need to isolate the oily ester 5 and avoid using the harsh conditions to acylate

5 to generate 6. The reaction is conducted by initially reﬂ uxing ethyl acetate with propionic

anhydride to remove any alcohols that may be contaminating the ethyl acetate solvent.

Once this part of the process is completed, the reaction is cooled to room temperature, and

4, along with triethyl amine, is added and the reaction is heated to reﬂ ux for 1 hour. Follow-

ing completion of the 1-hour reﬂ ux, the reaction is cooled to 70⬚C, methanol is added, and

then the reaction is again reﬂ uxed for 2 hours. Upon completion of the reaction the product

is isolated as its crystalline oxalate salt (6). This salt is recrystallized from methanol to

yield 6 in an overall yield of 78% from 4.

It is hypothesized that the initial step in the production of 6 is formation of the mixed

anhydride of the 4-anilidopiperidine acid and propionic acid. Ring closure onto the anilino

nitrogen generates intermediate 8, which can either open to form acid 9 or lose water to

generate iminium ion (10). Acid 9 can also be converted to 10 via formation of another

Scheme 12.6 Mechanism proposal for amide–ester (6) formation.

HHN

MeN

OEt

106

mixed anhydride, followed by ring closure. The electrophilic carbonyl of 10 is susceptible

to nucleophilic attack and in the presence of methanol will generate the desired product 6.

The mechanism of this reaction has precedent in the literature. Formation of α-amido

esters or amides from α-amido acids can be accomplished by treating the acids with an an-

hydride to generate azlactones, which are then opened by alcohols or amines with catalytic

acid to generate the desired products.

In the synthesis of 6 the α-amino nitrogen of 4 is a

secondary amine, so that when the azlactone-like intermediate (10) is formed it is activated

toward nucleophilic attack without the need for an acid catalyst. The generality of this reac-

tion has been demonstrated by the use of various anhydrides and alcohol nucleophiles.

The completion of the synthesis is accomplished by hydrogenolysis of the benzyl group

and subsequent Michael addition of methyl acrylate onto the piperidine nitrogen to yield

remifentanil. This same sequence was used in the medicinal chemistry route, however, in

the process route different conditions were used to avoid the use of high-pressure hydroge-

nation and the isolation of the oily intermediate secondary piperidine. After the free base

of 6 is made using K

in water–methanol, the piperidine is extracted into dilute phos-

phoric acid and hydrogenated at 1 atm over Pd–C. The pH of the solution containing the

secondary piperidine is adjusted to 8.5 using ammonium hydroxide, and methyl acrylate is

added to the aqueous solution to effect the Michael addition. The pH of the solution is ex-

tremely important because if it is too acidic, the nitrogen is protonated and the Michael ad-

dition does not take place. Alternatively, if the solution is too basic, the unhindered methyl

ester of remifentanil is rapidly hydrolized. The isolation of remifentanil is accomplished by

extracting the free base into isopropyl acetate, acidifying with HCl, and then crystallizing

remifentanil–HCl from a mixture of MeOH and isopropyl acetate.

The manufacturing process to generate remifentanil is ﬁ ve stages, all isolated inter-

mediates are solids, and the synthesis proceeds in approximately 25% overall yield. The

process used to generate remifentanil provides a safe, robust, inexpensive, and elegant

manufacturing route.

12.4 HUMAN CLINICAL TRIALS WITH REMIFENTANIL

The early clinical development of remifentanil demonstrated that it produced a dose-

dependent increase in analgesia, it had a rapid onset of action, and it had a signiﬁ cantly

more rapid offset of activity versus alfentanil (5.4 versus 54 minutes).

As predicted by the

preclinical work, the rapid inactivation of remifentanil in humans was shown to be primar-

ily the result of hydrolysis of the unhindered ester by nonspeciﬁ c plasma and tissue ester-

ases to the signiﬁ cantly less potent carboxylic acid. As a consequence of this mechanism of

inactivation it was not surprising, but gratifying, to demonstrate that the pharmacokinetics

of remifentanil in patients who had impaired hepatic or renal functions were unaltered

compared to healthy volunteers.

Remfentanil is indicated for use as an analgesic during induction and maintenance of

general anesthesia and for postoperative analgesia under close supervision in postanes-

thesia or intensive care unit setting. Based on the data gathered from human trials, the

recommended dose guidelines for remifentanil’s use in general anesthesia and as an an-

algesic in the immediate postoperative period were established. Induction of anesthesia is

typically done with an infusion of 0.5 to 1 µg/kg per minute of remifentanil along with a

hypnotic agent. Maintenance of anesthesia can be conducted with an infusion of remifent-

anil along with volatile agents, such as N

O or isoﬂ urane, or an intravenous hypnotic, such

HUMAN CLINICAL TRIALS WITH REMIFENTANIL 349

350

DISCOVERY AND DEVELOPMENT OF THE ULTRASHORT-ACTING ANALGESIC REMIFENTANIL

as propofol. Following completion of the surgical procedure, remifentanil can be infused at

a rate of 0.025 to 2 µg/kg per minute in the immediate postoperative arena to continue an-

algesic relief. Given the ultrashort duration of action of remifentanil, careful management

of a patient’s pain in the postoperative period must be considered. This can be achieved by

slowly reducing the infusion of remifentanil in the postoperative setting or administering

a long-acting analgesic just prior to cessation of the surgical procedure. Therefore, as the

analgesic effects of remifentanil rapidly dissipate, the patient’s pain will continue to be

managed with the longer-acting analgesic.

Remifentanil is also approved for use during monitored anesthesia care. Monitored an-

esthesia care is typically done for patients who are undergoing outpatient procedures. These

procedures include very short surgical procedures, such as a breast biopsy, or short surgical

procedures that utilize a nerve block. In a monitored anesthesia setting, remifentanil may

be administered either by using a single dose or by the continuous infusion method. In

either method remifentanil may be administered in combination with the anxiolytic mid-

azolam followed by a local or regional anesthetic block.

In humans, remifentanil demonstrates the typical pharmacology of a µ-opioid agonist.

In addition to its beneﬁ cial efﬁ cacy as an analgesic and hypnotic, remifentanil may cause

respiratory depression, bradycardia, hypotension, and skeletal muscle rigidity during pro-

cedures in which it is used for induction and/or maintenance of general anesthesia. Due to

the pharmacokinetic proﬁ le of remifentanil, the duration of these adverse side effects can

be controlled by discontinuing or decreasing the rate of the remifentanil infusion. In addi-

tion, the incidence and magnitude of these effects are dependent on the type and dose of

other anesthetic agents coadministered with remifentanil.

Remifentanil is unique among the clinically used opioid anesthetic agents as a result

of its pharmacokinetic proﬁ le. Its rapid and predictable pharmacokinetic proﬁ le provides

anesthesiologists with an analgesic that is optimal for different patient types and surgical

procedures ranging from short procedures performed in an outpatient setting to protracted

surgical cases that require rapid intraoperative titration and/or rapid recovery. The discov-

ery and development of remifentanil is a result of careful consideration of the needs of the

anesthesia community, the thoughtful design and execution of the chemistry and pharmo-

cology, the creative efforts in development to provide drug products suitable and safe for

clinical use, and the diligent efforts during the clinical phase to provide a product that is

differentiated and realizes its potential.

ACKNOWLEDGMENTS

Many talented drug discoverers are involved in any program that progresses successfully

from the lab to the market. The discovery and development of remifentanil is no exception

to this rule. I want to thank the many talented people who worked on remifentanil during all

phases of its progression. There are too many people who played signiﬁ cant roles to men-

tion all by name; however, Dr. Michael James, who served as the biology program leader

for the discovery of remifentanil, is gratefully acknowledged for his seminal contribution.

REFERENCES

For leading references and a general review of research on anesthetic agents, see Rees, D. C.;

Hill, D. R. Annu. Rep. Med. Chem. 1996, 31, 41.

Recent disclosures have described classes of ultrashort-acting benzodiazepines that may ﬁ nd

use as ultrashort-acting amnestics. (a) Stafford, J. A.; Pacofsky, G. J.; Cox, R. F.; Cowan, J. R.;

Dorsey, G. F., Jr.; Gonzales, S. S.; Jung, D. K.; Koszalka, G. W.; McIntyre, M. S.; Tidwell, J. H.;

Wiard, R. P.; Feldman, P. L. Bioorg. Med. Chem. Lett. 2002, 12, 3215. (b) Pacofsky, G. J.; Staf-

ford, J. A.; Cox, R. R.; Cowan, J. R.; Dorsey, G. F., Jr.; Gonzales, S. S.; Kaldor, I.; Koszalka,

G. W.; Lovell, G. G.; McIntyre, M. S.; Tidwell, J. H.; Todd, D.; Whitesell, G.; Wiard, R. P.;

Feldman, P. L. Bioorg. Med. Chem. Lett. 2002, 12, 3219.

Recent reports have described a new ultrashort-acting nondepolarizing neuromuscular blocking

agent (muscle relaxant) that is undergoing human clinical trials. (a) Boros, E. E.; Bigham,

E. C.; Mook, R. A., Jr.; Patel, S. S.; Savarese, J. J.; Ray, J. A.; Thompson, J. B.; Hashim, M. A.;

Wisowaty, J. C.; Feldman, P. L.; Samano, V. J. Med. Chem. 1999, 42, 206. (b) Samano, V.; Ray,

J. A.; Thompson, J. B.; Mook, R. A., Jr.; Jung, D. K.; Koble, C. S.; Martin, M. T.; Bigham, E. C.;

Regitz, C. S.; Feldman, P. L.; Boros, E. E. Org. Lett. 1999, 1, 1993. (c) Belmont, M. R.; Lien,

C. A.; Savarese, J. J.; Patel, S.; Fischer, G.; Mook, R. A., Jr. Br. J. Anaesth. 1999, 82, A419.

(d) Boros, E. E.; Samano, V.; Ray, J. A.; Thompson, J. B.; Jung, D. K.; Kaldor, I.; Koble, C. J.;

Martin, M. T.; Styles, V. L.; Mook, R. A., Jr.; Feldman, P. L.; Savarese, J. J.; Belmont, M. R.;

Bigham, E. C.; Boswell, S. E.; Hashim, M. A.; Patel, S. S.; Wisouaty, J. C.; Bowers, G. D.;

Moseley, C. L.; Walsh, J. S.; Reese, M. J.; Rutowske, R. D.; Seﬂ er, A. M.; Sptizer, T. D. J. Med.

Chem. 2003, 46, 2502.

A summary of the clinical proﬁ le of remifentanil is described in Patel, S. S.; Spencer, C. M.

Drugs, 1996, 52, 417.

The primary references for the medicinal chemistry and pharmacology leading to the discovery

of remifentanil are: (a) Feldman, P. L.; James, M. K.; Brackeen, M. F.; Bilotta, J. M.; Schuster,

S. V.; Lahey, A. P.; Lutz, M. W.; Johnson, M. R.; Leighton, H. J. J. Med. Chem. 1991, 34, 2202.

(b) James, M. K.; Feldman, P. L.; Schuster, S. V.; Bilotta, J. M.; Brackeen, M. F.; Leighton, H. J.

J. Pharmacol. Exp. Ther. 1991, 259, 712.

Erhardt, P. W.; Woo, C. M.; Anderson, W. G.; Gorczynski, R. J. J. Med. Chem. 1982, 25, 1408.

Stout, D. M.; Black, L. A.; Barcelon-Yang, C.; Matier, W. L.; Brown, B. S.; Quon, C. Y.; Stampﬁ li,

H. F. J. Med. Chem. 1989, 32, 1910.

Colapret, J. A.; Diamantidis, G.; Spencer, H. K.; Spaulding, T. C.; Rudo, F. G. J. Med. Chem.

1989, 32, 663.

Lutz, M. W.; Morgan, P. H.; James, M. K.; Feldman, P. L.; Brackeen, M. F.; Lahey, A. P.; James,

S. V.; Bilotta, J. M.; Pressley, J. C. J. Pharmacol. Exp. Ther. 1994, 271, 795.

A review of therapeutic agents that were discovered using the soft drug approach can be found in

Bodor, N.; Buchwald, P. Med. Res. Rev. 2000, 20, 58.

Van Daele, P. G. H.; DeBruyn, M. F. L.; Boey, J. M.; Sanczuk, S.; Agten, J. T. M.; Janssen, P. A. J.

Arzneim.-Forsch. 1976, 26, 1521.

Coleman, M. J.; Goodyear, M. D.; Latham, D. W. S.; Whitehead, A. J. Synlett 1999, 12, 1923.

Carter, H. B. Org. React. 1946, 3, 198.

10.

11.

12.

13.

REFERENCES 351

353

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

DISCOVERY AND DEVELOPMENT

OF NEVIRAPINE

KARL GROZINGER, JOHN PROUDFOOT, AND KARL HARGRAVE

Boehringer-Ingelheim Pharmaceuticals

Ridgeﬁ eld, Connecticut

13.1 INTRODUCTION

In October 1991 the Center for Disease Control in the United States declared the acquired

immune deﬁ ciency syndrome (AIDS) to be an epidemic in that country. From 151 cases

and 128 deaths reported in 1981, deaths from AIDS in the United States passed the 100,000

mark by the end of 1990. At the end of 2002, the World Health Organization estimated that

there were 42 million people living with HIV/AIDS worldwide and that about 3 million

people had died from AIDS in 2002 alone.

A massive scientiﬁ c effort aimed at under-

standing the etioliogy of this debilitating and deadly disease led to the recognition that a

retrovirus, human immune deﬁ ciency virus type 1 (HIV-1), was the causative agent.

The

virus invades and destroys CD4

⫹

T cells of the immune system, eventually compromising

the immune system to such an extent that normally innocuous pathogens become life-

threatening. The ﬁ rst drug approved for the treatment of AIDS, zidovudine (AZT), reached

the market in 1987 and was followed by several others. In this chapter we describe the

discovery and development of one of these drugs, nevirapine (Scheme 13.1).

The HIV life cycle, depicted schematically in Fig. 13.1, shows the targets for potential

therapeutic intervention. Virus replication requires entry into the cell, incorporation of the

virus genetic material into the host genome, and subsequent transcription and translation to

generate new virus particles. Transformation of the viral genetic material, single-stranded

RNA, into double-stranded proviral DNA is accomplished by the enzyme HIV-1 reverse

transcriptase (HIV-1 RT). The proviral DNA is subsequently integrated into the host ge-

nome by HIV-1 integrase. It remains dormant until activated, when transcription leads

to the production of viral RNA, some of which serves as genetic material for viral prog-

354 DISCOVERY AND DEVELOPMENT OF NEVIRAPINE

eny and some of which acts as a template for viral polyprotein production. A third virus

enzyme, HIV-1 protease, is essential for processing this material into individual proteins

that along with the viral RNA assemble to produce a new generation of infectious virus par-

ticles. The budding or escape of these new virus particles is lethal to the host cell. Therapies

based on vaccines or inhibition of the integrase enzyme have so far not progressed beyond

clinical evaluation. The ﬁ rst therapeutic agent for prevention of virus entry into the cell

was approved in 2003. Inhibition of the RT and protease enzymes has proven to be most

successful approach to effective therapies, and nearly all the approved AIDS drugs target

one of these two enzymes.

As our interest in developing a new drug for the treatment of AIDS took root, one

agent under evaluation in the clinic, AZT,

showed efﬁ cacy in slowing the progression of

the disease. AZT targets the HIV-1 RT enzyme by mimicking a natural substrate of this

enzyme, thymidine. Because it mimics a nucleoside naturally present in the cell, AZT can

also interfere with the activity of endogenous human enzymes. It was thought that some

of the side effects seen with AZT treatment were due to this lack of selectivity, and it was

clear that there were opportunities for additional, more selective inhibitors of HIV-1 RT.

HIV-1 RT is a heterodimeric enzyme with subunits of molecular weight 66 kDa (p66) and

51 kDa (p51). The enzyme possesses three distinct catalytic activities. It can function as an

RNA-dependent DNA polymerase that transforms the single-stranded RNA viral genome

RNA

Viral Protein

Synthesis

Viral RNA

Viral DNA

Protease

Integrase

Transcription

Reverse Transcriptase

Figure 13.1 HIV life cycle.

Nevirapine

Scheme 13.1

to an RNA–DNA hybrid. An RNAse H activity cleaves the original RNA strand, and a

DNA-dependent DNA polymerase activity synthesizes a complementary DNA providing

double-stranded proviral DNA. Since RT enzymatic activity is unique to retroviruses and

no corresponding activity is found in the normal human cell, this enzyme is a particularly

attractive target for drug discovery and holds the promise that a selective inhibitor of RT

could be a particularly safe and effective agent. Our goal, therefore, was to generate a drug

structure that did not resemble the natural nucleoside substrates of the enzyme.

13.2 LEAD DISCOVERY AND OPTIMIZATION

The decision to search for nonnucleoside inhibitors of HIV-1 RT required that we initiate

a screening program to identify a structural lead. This was based not only on the prefer-

ence to develop a novel structural class but also on the pragmatic basis that no suitable

lead structures were known. The screening capabilities at that time (⬍100 samples per

week) would now be considered (very) low-throughput screening compared to the high-

throughput screening (HTS) or ultrahigh-throughput screening (UHTS) available with

today’s robotic systems, which can screen thousands of compounds per hour. Thus, it was

fortuitous that after screening only approximately 600 random compounds from the com-

pany sample collection that a pyrido[2,3-b][1,4]benzodiazepinone (1) was found that was

weakly active (6 µM) in the HIV-1 RT enzyme assay.

The tricyclic backbone in this struc-

ture is a common feature in a large series of compounds in our company collection that

center around the M

-selective antimuscarinic agent pirenzepine (Scheme 13.2).

Although

there were many pirenzepine-analog pyrido[2,3-b][1,4]benzodiazepinones (1) and iso-

meric pyrido[2,3-b][1,5]benzodiazepinones (2) available for testing, there were very few

Pirenzepine

5a X = CH

5b X = N

Ring numbering for the

dipyridodiazepinone

7a R = Et

7b R = cyclopropyl

(Nevirapine)

Scheme 13.2

LEAD DISCOVERY AND OPTIMIZATION 355

356 DISCOVERY AND DEVELOPMENT OF NEVIRAPINE

dipyrido[3,2-b:2',3'-e][1,4]diazepinones (3) and dibenzo[b,e][1,4]diazepinones (4). A fo-

cused screening of the available analogs from these four series of related compounds sug-

gested that both the pyrido[2,3-b][1,5]benzodiazepinones (2) and dipyrido[3,2-b:2',3'-e]

[1,4]diazepinones (3) showed the most promise as lead structures, and at this point there

was no biological or physicochemical basis on which to choose one series over the other.

Accordingly, the synthetically more accessible pyrido[2,3-b][1,5]benzodiazepinone (2)

series was initially selected as the primary series for further studies, with the derivative

6,11-dihydro-11-ethyl-6-methyl-5H-pyrido[2,3-b][1,5]benzodiazepinone (5a) chosen as a

lead structure. This compound was active in the HIV-1 RT enzyme screening assay with an

value of 350 nM.

The lead optimization process requires that a set of criteria be established that are

deemed essential for the selection of compounds (for preclinical and clinical development)

with the characteristics necessary to achieve the therapeutic objective, desired route of ad-

ministration, and so on. For this program the HIV-1 RT enzymatic assay was the primary

screen. Active compounds (IC

⬍ 1 µM) in this assay were then tested for their ability to

block HIV-1 proliferation in cell culture, and a direct correlation was quickly found be-

tween the enzymatic and cellular potencies. Speciﬁ city was determined by testing selected

compounds against reverse transcriptase enzymes from HIV-2, SIV (simian immunodeﬁ -

ciency virus), and feline and murine leukemia viruses. In addition, the compounds were

tested for speciﬁ city against human DNA polymerases α, β, γ, and δ, and against calf thy-

mus DNA polymerase.

Additional criteria, not uncommon to most other drug candidates,

required that the compound(s) selected be nontoxic at the intended doses, that they have

reasonable aqueous solubility and be orally bioavailable, and that they be sufﬁ ciently stable

to metabolism to exhibit a reasonable half-life in vivo. Finally, the compound(s) selected

needed to have the ability to pass the blood–brain barrier. This was required by the fact that

the HIV-1 virus has the ability to infect the central nervous system, sometimes leading to

the potentially fatal AIDS dementia complex.

Although, as stated above, most of the initial synthetic efforts focused on the pyrido[2,3-

b][1,5]benzodiazepinone (2) series, a smaller synthetic effort was directed to the dipyr-

ido[3,2-b:2',3'-e][1,4]diazepinones (3), with the corresponding methyl–ethyl analog 5b

(IC

⫽ 125 nM) used as the lead for that series. It quickly became apparent that, in general,

structure–activity relationships developed for 2 also applied to 3. Also, the dipyrido com-

pounds are generally as potent as or more potent than the corresponding monopyrido 2, and

these SAR ﬁ ndings were conﬁ rmed as additional compounds were synthesized and tested.

Relatively early in the project, in vivo metabolism studies showed that N,N'-dealkyl-

ation occurred fairly rapidly in the compounds with no aromatic ring substitution, although

the rate of N,N'-dealkylation was slower in the dipyrido compound (5b) than was the corre-

sponding pyridobenzo substance 5a. Subsequent conﬁ rmation with other analogs demon-

strated that in general the rate of metabolism of the dipyrido compounds in different animal

species was signiﬁ cantly slower than the corresponding pyridobenzo analogs. In addition

to this important ﬁ nding, the dipyrido compounds were also found to have greater aqueous

solubility and were less cytotoxic than the corresponding pyridobenzo analogs. Accord-

ingly, the focus of the synthetic program shifted exclusively to the dipyrido series.

Since a primary goal was to develop an orally available drug, emphasis was placed on

measuring the rates of metabolism for selected analogs. It was known from the structure-

activity relationships that were being developed for this series that N(5)-dealkylation and

N(5),N(11)-dedialkylation resulted in substantial reductions in potency. On the other hand,

by attaching a methyl group at the 4-position (see 6) on the pyridyl ring rather than the

amide nitrogen (position-5, 6), three beneﬁ cial results were obtained. First, N(5)-dealkyl-

ation from the amide nitrogen was no longer possible. Second, rather than observing an

increased rate of N-dealkylation at position 11, N-dealkylation at that position actually

decreased signiﬁ cantly, with the primary metabolism (hydroxylation) taking place at the

4-methyl group. The rate of this oxidative metabolism was much slower than the N,N-

dealkylation observed with 5b. A third important ﬁ nding was an increased potency of the

4-methyl derivative compared to the corresponding 5-methyl compound (5b). Thus, by one

simple change in the structure, the rate of metabolism was lowered signiﬁ cantly and the

potency was increased appreciably.

Lead optimization of the series based on the dipyridodiazepinone structure (3) provided

structure–activity relationships (SAR) that supported the selection of nevirapine (7b) as a

preclinical candidate. These SAR studies indicated that a methyl group at position 4 re-

sulted in compounds with the highest potency and best overall characteristics. Derivatives

with small alkyl or acyl substituents at the lactam nitrogen (position 5) were also potent but

suffered from relatively rapid metabolism as noted above. Moreover, small alkyl groups

(ethyl and cyclopropyl) at the N-11-position were required for optimum enzyme inhibition

since both smaller functional groups (methyl and hydrogen) and larger alkyl groups (in-

cluding propyl and i-propyl) conferred signiﬁ cantly lower potency. Aromatic ring substitu-

tion other than at position 4 generally lowered potency and reduced aqueous solubility. The

most potent compounds had a 4-methyl substituent provided that there was no substitution

(i.e., R

⫽ H) at the adjacent 5-position.

On the basis of the results above, the N-11 ethyl (7a) and cyclopropyl (7b) derivatives

were selected as preclinical candidates. These two compounds were evaluated in a wide

range of tests and generally exhibited similar biological and physicochemical character-

istics, although the ethyl analog was more potent than the cyclopropyl analog against the

HIV-1 RT enzyme, and it was equipotent in its ability to block virus proliferation in the

cellular assay. The determining factor in the selection of the cyclopropyl analog (7b, ne-

virapine) over the ethyl analog for clinical evaluation was the signiﬁ cantly higher oral

bioavailability of nevirapine in monkeys and rats. This enabled the attainment of much

higher blood levels at a given oral dose. High oral bioavailability (93%) and a relatively

long half-life (25 hours) were subsequently measured in human clinical trials.

13.3 CHEMICAL DEVELOPMENT AND PROCESS RESEARCH

The initial synthesis

of nevirapine (7b) developed by the medicinal chemistry group em-

ployed 2-chloro-3-nitro-4-methylpyridine (8) as a key intermediate (Scheme 13.3). Cata-

lytic reduction to 2-chloro-3-amino-4-pyridine (9) and condensation with 2-chloronicoti-

noylchloride (10) gave the 2,2'-dihaloamide (11). Treatment of 11 with four equivalents of

cyclopropylamine (12) in xylene at 120 to 140⬚C under autogenous pressure produced the

2'-alkylamino adduct (13). Subsequent ring closure with sodium hydride in pyridine at 80

to 100⬚C gave nevirapine.

A signiﬁ cant concern during early drug development was the limited availability of in-

termediate 8, which could be obtained commercially only in small quantities. It is derived,

via 16, by nitration of 14 (Scheme 13.3), and a yield of 82% has been reported for this re-

action.

However, the method is not scalable, due to a rapid uncontrollable exotherm when

the reaction is conducted on a large scale. In addition, the nitration method gives a mix-

ture of regioisomers 15 and 16, and separation of these isomers proved to be industrially

CHEMICAL DEVELOPMENT AND PROCESS RESEARCH 357

358 DISCOVERY AND DEVELOPMENT OF NEVIRAPINE

impractical. Based on the information obtained from these experiences, the use of 8 as a

precursor was abandoned and chemical and process development activities were directed

toward development of an alternative approach to 3-amino-2-chloro-4-methylpyridine (9)

that could be scaled up to produce pilot plant quantities.

Although 2,3,4-trisubstituted pyridines are rather difﬁ cult to access, 2,3,4,6-tetrasubsti-

tuted pyridines, particularly those with substituent patterns similar to 17 (Scheme 13.3),

are relatively easy to produce on a large scale,

and in fact, 17 is commercially available

in multiton quantities. As part of the SAR efforts around the nevirapine structure, we had

experience in transforming this material (17) into dichloroaminopyridine (20). Treatment

with phosphorous oxychloride gives 18, which is followed by acid hydrolysis of the 3-cyano

substituent and conversion to the amine under Hofmann rearrangement conditions. Efforts

to selectively remove the 6-chloro substituent from intermediate 18, 19, or 20 were unsuc-

cessful. However, removal of both chlorine atoms by catalytic dechlorination followed by

selective rechlorination in the 2-position gave the required product (9).

Although this

NCl

NNH

ClCl

CONH

ClCl

5%Rh/C

EtOH

toluene

xylene

110

83%

NaH

diglyme

10 11

17 X = OH

18 X = Cl

98%

65%

67%

POCl

85%

84%

91%

92%

93%

Scheme 13.3

synthetic approach lacks atom economy with respect to the removal and addition of chlo-

rine atoms, it provided the opportunity to meet the short-term active pharmaceutical ingre-

dient (API) supply needs and established a synthetic strategy upon which further develop-

ment activities could be founded. It should also be noted that all process steps from the

commercially available raw material 2,6-dihydroxy-4-methyl-3-cyanopyridine (17) avoid

the use of halogenated organic solvents, making it environmentally attractive.

Intermediate 20 can also be transformed into nevirapine via 22, 23, and 24 in an al-

ternative synthetic sequence that eliminates the dechlorination and rechlorination process

steps (Scheme 13.4).

In this approach, 24 is dechlorinated with palladium on carbon and

hydrogen to give nevirapine in high yield.

Although this option presented the opportunity

to eliminate one process step, all the intermediates from this method differ in chemical

composition from the original process (Scheme 13.3). For this reason, reevaluation of the

API impurity proﬁ le, toxicology, and other pharmacological and regulatory issues would

be required. Because this option was identiﬁ ed late in the chemical development process,

it was decided that the potential process beneﬁ ts were more than offset by the additional

time and effort required to requalify this process, and this option for API production was

abandoned.

The nevirapine process scheme used during chemical development provided the basis

on which to begin process development studies, with the objective of deﬁ ning reaction con-

ditions that would allow this process to be carried out on a routine commercial basis. In this

process, the basic elements of the molecule are introduced at the step involving the con-

densation of 9 and 10. Using the U.S. Food and Drug Administration (FDA) guidelines

for deﬁ ning the starting point in the synthesis for regulatory purposes, 3-amino-2-chloro-

4-methylpyridine (9) was considered a raw material in the synthesis. This provided the

opportunity to implement further process improvements in the preparation of this molecule

after the product launch, with limited regulatory impact.

The use of cyclopropylamine (12) presented a signiﬁ cant process optimization opportu-

nity. In the initial chemical development pilot studies, the conversion of 11 to 13 required

4 molar equivalents of cyclopropylamine in the reaction medium. Although 12 is a simple

building block, it is rather expensive on a per kilogram basis. In this reaction, 1 mol of 12

is used to absorb the by-product HCl from the reaction. Calcium oxide was found to be a

ClCl

70%

84%

90%

72%

Scheme 13.4

CHEMICAL DEVELOPMENT AND PROCESS RESEARCH 359

360 DISCOVERY AND DEVELOPMENT OF NEVIRAPINE

much more cost-effective neutralizing agent. However, even with calcium oxide present, a

2.5 M excess of cyclopropylamine is required to carry the reaction to completion. Efforts

to combine this reaction with the subsequent cyclization step were successful, and 13 was

treated as a nonisolated intermediate in process development pilot runs as well as on a

commercial scale.

Particular attention was also paid during development activities to the speciﬁ c reaction

conditions employed in the ﬁ nal cyclization step. The medicinal chemistry procedure as

well as the initial chemical development efforts used either pyridine or diglyme as solvent

and sodium hydride as the base. Sodium hydride in 2.8 M excess is required to carry the cy-

clization reaction to completion. The ﬁ rst mole of sodium hydride is consumed with the de-

protonation of the more acidic amide proton. In the event that a base of insufﬁ cient strength

is used to remove the amino proton of 13, ring closure to the oxazolo[5,4]pyridine (28,

Scheme 13.5) from the displacement of the chlorine atom by the amide carbonyl oxygen

occurs. The same undesired product was observed when sodium carbonate or other base

systems were used. No industrially practical substitute for sodium hydride as a reagent

base was identiﬁ ed. It was recognized from solvent screening studies that the reaction

pathway for the ring closure was also very solvent dependent. When dimethylformamide

(DMF) was used as the solvent, oxazole is the exclusive reaction product.

One problem with the use of sodium hydride as a reagent in pilot and commercial opera-

tions relates to the storage and handling requirements for this material. Sodium hydride is

received from the supplier in 10-kg bags as a 60% amalgam in mineral oil to stabilize the

reagent. In the nevirapine process, the mineral oil tends to agglomerate with the product

upon precipitation from the reaction mixture. An intermediate puriﬁ cation step was de-

veloped using DMF as a crystallization medium. The crude product was dissolved in hot

DMF followed by charcoal treatment to remove the residual mineral oil. The charcoal was

then removed by ﬁ ltration followed by crystallization of the product after addition of water.

The method outlined in the box in Scheme 13.3 was used to produce the nevirapine-API

requirements through phase III clinical trials, commercial launch, and production.

13.4 MECHANISM OF ACTION

As nevirapine progressed through development into the clinic, the details of its mode of ac-

tion were gradually uncovered. Enzyme kinetics experiments revealed that nevirapine and

related molecules were noncompetitive inhibitors of HIV-1 RT

and bind at a site distinct

from the catalytic, or active, site. Nevirapine was also found to display a remarkable selec-

tivity proﬁ le. It did not inhibit human DNA polymerases, which gave hope for an improved

safety proﬁ le over nucleoside inhibitors in the clinic. Very surprisingly, nevirapine was

Scheme 13.5

also inactive against the reverse transcriptase (HIV-2 RT) from the closely related virus,

HIV-2. Initially, the origin of the selectivity toward HIV-2 RT was difﬁ cult to understand,

since nucleoside analogs typically display comparable potency against both HIV-1 RT and

HIV-2 RT. These data, in combination with the noncompetitive enzyme kinetics, prompted

experiments designed to identify the location of the binding site for nevirapine on RT.

Ultimately, a combination of biophysical

and crystallographic

studies located the drug-

binding site in the larger p66 subunit of the enzyme adjacent to tyrosine residues at posi-

tions 181 and 188. These amino acids are close in sequence to aspartic acids 185 and 186,

which constitute part of the enzyme active site. The corresponding residues in HIV-2 re-

verse transcriptase are isoleucine 181 and leucine 188. The data in Table 13.1 illustrate the

importance of the nature of the amino acid residues at these two positions for the binding

of nevirapine to RT. Replacement of even one of the tyrosine residues in the native HIV-1

sequence, for example with leucine as shown in Table 13.1, renders the mutant enzyme

insensitive to nevirapine.

Concurrent in vitro experiments in which virus was grown in

the presence of increasing concentrations of the drug yielded virus that was resistant to ne-

virapine. Analysis of the RT enzyme from the resistant virus showed that it had a mutation

encoding for the replacement of tyrosine 181 with a cysteine residue.

This single amino

acid change decreased the sensitivity of the enzyme to the drug by about 100-fold and

alerted us to the possibility of virus resistance to the drug in a clinical setting.

13.5 CLINICAL STUDIES

As nevirapine progressed into the clinic in 1991, the initial expectations were that it could

be used successfully as a monotherapeutic agent. Pharmacokinetic studies showed that the

drug was well absorbed, with high bioavailability and a long half-life allowing twice-daily

dosing.

19,20

Exposure form the 200- and 400-mg/day dosing regimens gave drug plasma

levels many times above the in vitro IC

value for inhibition of virus replication, and

at these doses the drug caused an initial rapid decrease in viral load in treated patients.

However, over a period of weeks to months there was a rebound in virus back to original

levels.

The rebound is due to the emergence of resistant virus

and is due mainly to the

virus containing the mutant Y181C RT enzyme discussed above.

The focus changed to determining the utility of the drug in combination with other

anti-AIDS agents. In vitro studies had provided evidence that combinations of anti-HIV

drugs could be more effective than single agents.

Evidence supporting combination ther-

apy in the clinic was also emerging; for example, AZT in combination with didanosine or

zalcitabine was more effective than AZT alone.

Clinical studies showed that nevirapine

in combination with a variety of other nucleoside and protease inhibitors is effective in

TABLE 13.1 Nature of Amino Acid Residues at Positions 181 to 188

181 182 183 184 185 186 187 188 Nevirapine IC

(nM)

HIV-1 sequence Y Q Y M D D L Y 40

HIV-2 sequence I Q Y M D D I L

⬎10,000

HIV-1 (Y188L) Y Q Y M D D L L

⬎10,000

HIV-1 (Y181C) C Q Y M D D L Y 3000

CLINICAL STUDIES 361

362 DISCOVERY AND DEVELOPMENT OF NEVIRAPINE

suppressing virus replication for prolonged periods of time.

Nevirapine was approved

by the FDA in 1996 and was the ﬁ rst member of the nonnucleoside reverse transcriptase

inhibitor (NNRTI) class to reach the market.

ACKNOWLEDGMENTS

The successful discovery and development campaign described above required the cre-

ative input and hard work of many individual scientists whose scientiﬁ c contributions are

captured in the references cited. We would also like to particularly acknowledge the lead-

ership and commitment of Jay Merluzzi, Alan Rosenthal, Julian Adams, Robert Eckner,

and Peter Grob.

REFERENCES

http://www.who.int/hiv/en/.

Curran, J. W.; Morgan, W. M.; Hardy, A. M.; Jaffe, H. W.; Darrow, W. W.; Dowdle, W. R. The

epidemiology of AIDS: current status and future prospects. Science 1985, 229, 1352.

Mitsuya, H.; Weinhold, K. J.; Furman, P. A.; St. Clair, M. H.; Nusinof Lehrman, S.; Gallo, R.

C.; Bolognesi, D.; Barry, D. W.; Broder, S. 3

'-Azido-3'-deoxythymidine: an antiviral agent that

inhibits the infectivity and cytopathic effect of human T cell lymphotropic virus type III/lymph-

adenopathy-associated virus in vitro. Proc. Natl. Acad. Sci. USA, 1985, 82, 7096.

Hargrave, K. D.; Proudfoot, J. R.; Grozinger, K. G.; Cullen, E.; Kapadia, S. R.; Patel, U. R.;

Fuchs, V. U.; Mauldin, S. C.; Vitous, J.; Behnke, M. L.; Klunder, J. M.; Pal, K.; Skiles, J. W.;

McNeil, D. W.; Rose, J. M.; Chow, G. C.; Skoog, M. T.; Wu, J. C.; Schmidt, G.; Engel, W. W.;

Eberlein, W. G.; Saboe, T. D.; Campbell, S. J.; Rosenthal, A. S.; Adams, J. Novel non-nucleoside

inhibitors of HIV-1 reverse transcriptase. 1. Tricyclic pyridobenzo- and dipyridodiazepinones.

J. Med. Chem., 1991, 34, 2231–2241.

Carmine, A. A.; Brogden, R. N. Drugs, 1985, 30, 85.

Merluzzi, V. J.; Hargrave, K. D.; Labadia, M.; Grozinger, K.; Skoog, M.; Wu, J. C.; Shih, C.

K.; Eckner, K.; Hattox, S.; Adams, J.; Rosenthal, A. S.; Faanes, R.; Eckner, R. J.; Koup, R. A.;

Sullivan, J. L. Inhibition of HIV-1 replication by a nonnucleoside reverse transcriptase inhibitor.

Science, 1990, 250, 1411–1413.

(a) Lamson, M. J.; Sabo, J. P.; MacGregor, T. R.; Pav, J. W.; Rowland, L.; Hawi, A.; Cappola,

M.; Robinson, P. Single dose pharmacokinetics and bioavailability of nevirapine in healthy

volunteers. Biopharm. Drug Dispos., 1999, 20, 285–291. (b) Cheeseman, S. H.; Hattox, S. E.;

McLaughlin, M. M.; Koup, R. A.; Andrews, C; Bova, C. A.; Pav, J. W.; Roy, T.; Sullivan, J. L.;

Keirns, J. J. Pharmacokinetics of nevirapine: initial single-rising-dose study in humans. Antimi-

crob. Agents Chemother., 1993, 37, 178–182.

Burton, A. G.; Halls, P. J.; Katritzky, A. R. Tetrahedron Lett., 1971, 24, 20211.

Bobbitt, M.; Scala, D. A. J. Org. Chem., 1960, 25, 560.

Grozinger, K. G.; Hargrave, K. D.; Adams, J. U.S. patent 5,668,287, 1997.

Grozinger, K. G.; Fuchs, V.; Hargrave, K. D.; Mauldin, S.; Vitous, J.; Campbell, S.; Adams, J.

Synthesis of nevirapine and its major metabolite. J. Heterocycl. Chem., 1995, 32, 259–263.

Grozinger, K. G.; Hargrave, K. D.; Adams, J. U.S. patent 5,571,912, 1996.

Guidance for Industry BACPAC I: Intermediates in Drug Substance Synthesis Bulk Active Post-

approval Changes: Chemistry, Manufacturing, and Controls Documentation, U.S. Department

10.

11.

12.

13.

of Health and Human Services Food and Drug Administration Center for Drug Evaluation and

Research (CDER), Center for Veterinary Medicine (CVM), February 2001.

(a) Wu, J. C.; Warren, T. C.; Adams, J.; Proudfoot, J.; Skiles, J.; Raghavan, P.; Perry, C.; Farina,

P. F.; Grob, P. M. A novel dipyridodiazepine inhibitor of HIV-1 reverse transcriptase acts through

a nonsubstrate binding site. Biochemistry, 1991, 30, 2022–2026. (b) Spence, R. A.; Kati, W. M.;

Anderson, K. S.; Johnson, K. A. Mechanism of inhibition of HIV-1 reverse transcriptase by non-

nucleoside inhibitors. Science, 1995, 267, 988–993.

Cohen, K. A.; Hopkins, J.; Ingraham, R. H.; Pargellis, C.; Wu, J. C.; Palladino, D. E. H.; Kincade,

P.; Warren, T. C.; Rogers, S.; Adams, J.; Farina, P. F.; Grob, P. M. Characterization of the binding

site for nevirapine (BI-RG-587) a nonnucleoside inhibitor of human immunodeﬁ ciency virus

type-1 reverse transcriptase. J. Biol. Chem., 1991, 266, 14670–14674.

Kohlstaedt, L. A.; Wang, J.; Friedman, J. M.; Rice, P. A.; Steitz, T. A. Crystal structure at 3.5A

resolution of HIV-1 reverse transcriptase complexed with an inhibitor. Science, 1992, 256, 1783–

1790.

Shih, C. K.; Rose, J. M.; Hansen, G. L.; Wu, J. C.; Bacolla, A.; Grifﬁ n, J. A. Chimeric human

immunodeﬁ ciency virus type 1/type 2 reverse transcriptases display reversed sensitivity to non-

nucleoside analog inhibitors. Proc. Natl. Acad. Sci. USA, 1991, 88, 9878–9882.

Richman, D.; Shih, C. K.; Lowy, I.; Rose, J.; Prodanovich, P.; Goff, S.; Grifﬁ n, J. Human immu-

nodeﬁ ciency virus type 1 mutants resistant to nonnucleoside inhibitors of reverse transcriptase

arise in tissue culture. Proc. Natl. Acad. Sci. USA, 1991, 88, 11241–11245.

Lamson, M. J.; Sabo, J. P.; MacGregor, T. R.; Pav, J. W.; Rowland, L.; Hawi, A.; Cappola, M.;

Robinson, P. Single dose pharmacokinetics and bioavailability of nevirapine in healthy volun-

teers. Biopharm. Drug Dispos., 1999, 20, 285–291.

Pollard, R. B.; Robinson, P.; Dransﬁ eld, K. Safety proﬁ le of nevirapine, a nonnucleoside reverse

transcriptase inhibitor for the treatment of human immunodeﬁ ciency virus infection. Clin. Ther.,

1998, 20(6), 1071–1092.

De Jong, M. D.; Vella, S.; Carr, A.; Boucher, C. A. B.; Imrie, A.; French, M.; Hoy, J.; Sorice, S.;

Pauluzzi, S.; et al. High-dose nevirapine in previously untreated human immunodeﬁ ciency virus

type 1-infected persons does not result in sustained suppression of viral replication. J. Infect.

Dis., 1997, 175, 966–970.

Richman, D. D.; Havlir, D.; Corbeil, J.; Looney, D.; Ignacio, C.; Spector, S. A.; Sullivan, J.;

Cheeseman, S.; Barringer, K.; et al. Nevirapine resistance mutations of human immunodeﬁ -

ciency virus type 1 selected during therapy. J. Virol., 1994, 68, 1660–1666.

Harmenberg, J.; Aakesson-Johansson, A.; Vrang, L.; Cox, S. Synergistic inhibition of human

immunodeﬁ ciency virus replication in vitro by combinations of 3

'-azido-3'-deoxythymidine and

'-ﬂ uoro-3'-deoxythymidine. AIDS Res. Hum. Retroviruses, 1990, 6, 1197–1202.

Jablonowski, H. Studies of zidovudine in combination with didanosine and zalcitabine. J. Acquir.

Immune Deﬁ c., Syndr. Hum. Retroviruses, 1995, 10(Suppl. 1), S52–S56.

(a) Reliquet, V.; Ferre, V.; Hascoet, C.; Besnier, J. M.; Bellein, V.; Arvieux, C.; Molina, J. M.;

Breux, J. P.; Zucman, D.; Rozenbaum, W.; Allavena, C.; Rafﬁ , F. Stavudine, didanosine and

nevirapine in antiretroviral-naive HIV-1-infected patients. Antiviral Ther., 1999, 4, 83–84. (b)

D’Aquila, R. T.; Hughes, M. D.; Johnson, V. A.; Fischl, M. A.; Sommadossi, J.-P.; Liou, S.-H.;

Timpone, J.; Myers, M.; Basgoz, N.; et al. Nevirapine, zidovudine, and didanosine compared

with zidovudine and didanosine in patients with HIV-1 infection: a randomized, double-blind,

placebo-controlled trial. Ann. Intern. Med., 1996, 124, 1019–1030.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

REFERENCES 363

365

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

APPLICATIONS OF NUCLEAR

IMAGING IN DRUG DISCOVERY

AND DEVELOPMENT

JOHN W. BABICH

Molecular Insight Pharmaceuticals, Inc.

Cambridge, Massachusetts

WILLIAM C. ECKELMAN

Molecular Tracer, LLC

Bethesda, Maryland

14.1 INTRODUCTION

14.1.1 Process and Challenges of Drug Development

The process of preclinical drug discovery has changed fundamentally over the past de-

cade, providing pharmaceutical companies with a large number of novel targets as well

as novel tools designed to accelerate the discovery process. Many of the new molecular

target structures, however, have not been thoroughly validated, which increases the risk of

unforeseeable side effects or lack of efﬁ cacy. In addition, novel drug discovery technolo-

gies, including high-throughput screening involving libraries of thousands of compounds,

have increased the number of molecules being considered for lead optimization with high-

speed chemistry and ultimately for in vivo investigation in a wide array of suitable animal

systems. So far, these developments have added additional costs to the drug development

process without necessarily increasing its efﬁ cacy or improving its cost-effectiveness

(Lehmann Brothers and McKinsey, 2001).

The process of early clinical drug development, in contrast, has changed little over the past

20 years. Chances for a drug candidate that has emerged successfully from preclinical stud-

ies to advance from a phase I trial to an approved product average around 22%, and even for

366 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

compounds in phase III, the failure rate is still 57% (Lawrence, 2004). A signiﬁ cant propor-

tion of these failures are associated with inappropriate drug metabolism and pharmacokinetics

of candidate molecules. At the same time, costs associated with successful drug development

are now estimated to exceed $800 million, with the bulk of these costs accruing in advanced

clinical trials (Dickson and Gagnon, 2004). This situation deﬁ nes the urgent need for drug

development companies to recognize the success and failure potential early in the R&D pro-

cess and has spurred interest for reliable cost-effective methods to generate pharmacokinetic,

toxicology, proof of concept, and efﬁ cacy data with high predictability in preclinical animal

studies and early clinical trials that provide a good rationale for go/no-go decisions.

14.1.2 Role and Contribution of Positron Emission Tomography

Positron emission tomography (PET) and single photon emission tomogrpahy (SPECT) are

compelling quantitative imaging techniques to monitor and assess intra- and extracellular

events such as biochemical pathways, molecular interactions, drug pharmacokinetics, and

pharmacodynamics in the living organism. Endogenous molecules, receptor and enzyme

ligands, drugs, or biomolecules are labeled with radionuclides and injected mostly intrave-

nously as molecular probes. In the living organism they participate in biochemical processes

while emitting gamma rays, which permit visualization of their location and concentration.

Only recently, PET and SPECT scanners were reﬁ ned and redesigned such that they

also became capable of providing high-resolution images of small animals, such as mice

and rats (Cherry et al., 1997; Cherry, 2001). This innovation has catapulted the use of PET

and SPECT from patient management or clinical trials to the preclinical stages of drug

development. It has also largely promoted its use to validate molecular target structures for

drug discovery in vivo, investigate and redesign experimental drugs, study the distribution

of a drug, monitor its interaction with its target across organs and tissue in the body, and

link that information to its biological function and clinical efﬁ cacy in the same organism.

Since PET and SPECT capture and visualize in vivo processes at the molecular level

and since new drug candidates increasingly target speciﬁ c molecular structures and events,

it may not be surprising to note that both disciplines, drug discovery and PET/SPECT,

become increasingly intertwined. PET and SPECT identify and validate in vivo molecular

structures that are proposed but not yet proven to present good drug development targets,

while on the other hand, pharmacogenomics and functional proteomics increasingly pro-

vide new probes for PET imaging (Eckelman, 2003).

14.2 PRINCIPLES AND EVOLUTION OF TECHNOLOGY

14.2.1 Introduction to PET Principles

PET does not provide a direct chemical analysis of reaction products; instead, the labeled

molecules function as tracers to depict individual steps in a biochemical cascade. The

labeled molecule emits positrons, which travel only a very short distance and then com-

bine with an electron. Annihilation occurs; the masses of positron and electron convert

into their energy equivalent through emission of two 511-keV photons that are about 180⬚

apart. Those emitted photons are detected as a coincidence event once they strike oppos-

ing detectors simultaneously. Between 6 and 70 million detection pairs are recorded from

many different angles and allow for reconstruction of an image. The kinetic information is

used to calculate the concentration of binding sites over time. Tissue concentration of the

probe in conjunction with the time course of the plasma concentration permit calculating

the transport and reaction process the probe undergoes in the living organism. As a result,

one obtains an image related to the rate of the process under investigation.

The radiolabeled probes are made in miniaturized self-shielded low-energy cyclotrons,

which bombard stable nucleon with protons or deuterons to generate proton-rich nuclei.

As a result of the proton–neutron imbalance, the nucleus is stabilized by conversion of

a proton to a neutron. This conversion occurs by electron capture whereby the charge of

a proton is “neutralized” by an orbital electron, or by emission of a positive electron or

positron whereby a proton gives up its charge resulting in conversion to a neutron. Those

minicyclotrons are combined in automated units that provide, usually under the control

of an integrated computer, solvents, reagent additions, solution transfer, column separa-

tion, and other steps to produce labeled imaging probes, creating mobile and versatile PET

radiopharmacies.

Positron-emitting radionuclides are available for the key elements found in almost all

biomolecules and drugs, such as carbon (

C), nitrogen (

N), and oxygen (

O). In addi-

tion, ﬂ uorine, although not normally a constituent of biomolecules, can act as a bioisostere

of hydrogen or the hydroxyl group and is commonly encountered in a variety of drugs.

Simple molecules such as

, or H

F are most commonly used as precursors

to the radiolabeled tracer. They substitute for a stable carbon, nitrogen, or ﬂ uorine to yield

labeled molecules or drugs that are chemically and biologically indistinguishable from the

stable or nonradioactive counterpart. With these building blocks a rich array of physiologi-

cally active molecules and drugs can be produced. Usually, compounds are administered in

picomole or nanomole amounts to provide dynamic and real-time information on in vivo

behavior of radiolabeled molecules. The direct in vivo imaging of molecular interaction

of biological processes provides insights into physiological and disease processes at the

molecular level that had not previously been possible and have tremendous value to accel-

erate, advance, and improve drug discovery and drug development processes.

14.2.2 Suitable Targets

The majority, if not all, novel compounds in drug development target in a highly speciﬁ c

and selective fashion cellular or nuclear receptors, cellular enzymes, or transporters or

interfere with protein–protein interaction. The tracer technology delivers high sensitivity

for imaging those low-density molecular systems in vivo, even though at a cost of lower

spatial resolution than for MRI or CT. The high speciﬁ c activity of the radioligands facili-

tates detection of low-density sites. Radiolabeled ligands or antibodies are instrumental to

map receptor expression across anatomical structures such as different regions of the brain,

and to understand in vivo binding kinetics and associated biological functions of drugs

and antibodies binding to cell-surface receptors. Radiolabeled ligands, for example, permit

quantiﬁ cation of receptor occupancy achieved by doses of a therapeutic agent targeting a

speciﬁ c receptor or receptor subtype.

14.2.3 Suitable Animal Models

Advances in PET imaging technology led in the late 1990s to the evolution of compact

low-cost dedicated small-animal PET scanners with high spatial resolution capability

that extended the use of PET as a research tool in clinical applications to preclinical drug

PRINCIPLES AND EVOLUTION OF TECHNOLOGY 367

368 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

discovery research by making laboratory animals such as mice and rats, which are so im-

portant in the early stages of drug discovery, accessible to in vivo imaging technology

(Cherry et al., 1997; Cherry, 2001; Myers and Hume, 2002) and deliver dynamic, real-time

information. Small animal imaging technology speciﬁ cally offers high volumetric resolu-

tion at an axial ﬁ eld of view of 1.8 cm. Three-dimensional collections of emission data in

combination with Bayesian reconstruction techniques generate high-deﬁ nition images of

the entire mouse body (Qi et al., 1998). These advancements have greatly accelerated PET

studies in a variety of animals, including those that are good model systems of human dis-

eases or carry deﬁ ned genetic alterations, such as gene knock-out animals.

Knock-mice have more recently been instrumental in investigating receptor-subtype

selectivity for muscarinic receptor ligands toward each of the four receptor subtypes M1,

M2, M3, and M4 in mice where either one of these subtypes was absent (Jagoda et al.,

2003). This particular study identiﬁ ed the M2 subtype receptor as the main target of the ra-

dioligand tested in distinct regions of the brain. More recently, protein–protein interactions

subject to a range of physiological and pathophysiological conditions have been investigated

in vivo using PET. The yeast/two-hybrid technology was key in allowing scientists to ﬁ nd

interaction partners for any given protein and further map down to the single-amino-acid

level the requirements for functional protein–protein interactions. With the help of PET,

these protein–protein interactions can now be examined and quantiﬁ ed not just in vitro un-

der highly artiﬁ cial nonphysiologic conditions but in the living animal. This is achieved by

linking the two-hybrid construct to a reporter gene whose expression in turn is visualized by

radiolabeled ligands (Luker et al., 2003), allowing researchers to monitor intrinsic binding

speciﬁ cities of proteins and gene regulation during development, disease progression, and

under the inﬂ uence of pharmacologic agents. Given the still prevailing shortcomings in the

efﬁ cacy of gene transfer, the strength of this approach lies in the preclinical animal studies.

14.3 ROLE IN DRUG DISCOVERY

PET is primarily a functional imaging technology designed for rapid and repeated non-

invasive in vivo assessment of molecular and ultimately biological processes. Molecular

imaging examines disease biology as the disease evolves and proceeds through its natural

history. This capability makes PET an ideal tool to accompany all stages of drug devel-

opment: from the identiﬁ cation of molecular structures involved in the genesis and evo-

lution of the disease, through the design of drugs to effectively modify those molecular

structures and their biological or pathological function in vivo, and ultimately to monitor

therapeutic efﬁ cacy and response in the patient. Products of modern drug discovery are

monoclonal antibodies, vaccines, therapeutic peptides or proteins, and enzyme inhibitors.

Usually, there are several iterations in the chemical design of these compounds before a

lead with the desired properties emerges. At this stage of the drug development process,

radionuclide-based methods are helpful to assess binding characteristics of antibodies or

ligands or to measure new transport or enzyme functions in vivo.

14.3.1 Target Validation and Drug Design

Target Identiﬁ cation: Linking Markers to Disease The most signiﬁ cant contribution of

PET in ﬁ nding or conﬁ rming molecular structures as targets for therapeutic interventions

has occurred over the past decade in neurodegenerative and psychiatric disorders which,

by their very nature, are not easily accessible for investigation in animal models. PET

visualizes brain neurochemistry in the living brain by using radiotracers that monitor

neurotransmitter synthesis, metabolism, enzymes, transporters, and receptors. From those

studies, insights into the biological role of these proteins and their involvement in psychiatric

or neurodegenerative disorders, including schizophrenia, anxiety disorders, depression, or

various forms of dementia, have emerged that in turn identify molecular target structures

suitable for pharmacological modulation (Smith et al., 2003). PET has, for example, been

instrumental in identifying altered serotonin pathway activities in the genesis and evolution

of several eating disorders such as anorexia nervosa and bulimia nervosa, providing a ﬁ rst

rational for therapeutic interventions strategies (Barbarich et al., 2003).

Similarly, PET studies with radiolabeled ligand of peripheral-type benzodiazepine

receptors, which were believed to be expressed exclusively by peripheral blood cells

and linked to peripheral inﬂ ammatory responses led to the discovery of benzodiazepine

receptor expression associated with neuroinﬂ ammatory processes in the brain. Speciﬁ cally,

PET identiﬁ ed benzodiazepine receptors in activated microglia cells as they accompany

neuroinﬂ ammatory diseases such as multiple sclerosis or neurodegerative diseases such

as Alzheimer’s. Now the expression of this peripheral benzodiazepine binding sites on

microglia is viewed as a key indicator for their transition from a normal resting state to

the activated state (Cagnin et al., 2001). Those receptors have subsequently been shown to

display all the functional characteristics associated with the initiation of an inﬂ ammatory

response, such as release of nitrite oxide and tumor necrosis factor-α (Wilms et al., 2003).

Taken together, these PET observations led to the identiﬁ cation of a molecular target struc-

ture and therapeutic rational to attenuate the neuroinﬂ ammatory process associated with

Alzheimer’s disease or multiple sclerosis.

In Vivo Analysis of Receptors, Enzymes, and Signal Transduction in Normal and

Diseased Cells and Animals The therapeutic value of a novel drug that acts through

binding a distinct molecular structure in a speciﬁ c organ or cell population is guided by

speciﬁ city, saturation kinetics, and selectivity. Each of these parameters can be assessed

in vivo using PET. 16α[

F]ﬂ uoro-17β-estradiol, a selective estrogen receptor ligand, has,

for example, been used to quantify the distribution of the tracer in discrete cerebral areas

and to quantify in vivo estrogen receptor binding parameters in the brain (Dehdashti et al.,

1999). Cell–cell interactions also govern host–pathogen interactions and can be visualized

to monitor pathogen tropism, pathogen life cycle, signal transduction, host response, and

cell trafﬁ cking in living animals (Piwnica-Worms et al., 2004). Radiolabeled rolipram, a

drug that inhibits phosphodiesterase 4, a key intracellular signal transduction molecule that

is implied in signaling pathways used by inﬂ ammatory signals as well as neurotransmitters,

has been used to monitor PDE4 activity in vivo in distinct regions of the brain (DaSilva

et al., 2002). Similarly, phosphoinositide turnover accompanies a wide variety of intracellular

signal transduction processes, including those initiated by cytokines or neurotransmitters.

In the brain its turnover is closely linked to synaptic functionality and to the production of

second messengers. Radiolabeled diacylglycerol as a marker for phosphoinositide turnover

is suitable to assess postsynaptic biological responses in healthy subjects and under various

disease conditions (Imahori et al., 2002).

Intracellular signaling is also visualized and quantiﬁ ed in vivo by examining gene reg-

ulation with the help of gene reporter constructs. The regulatory element of interest is

combined with reporter genes, such as the secreted alkaline phosphatase, that are easily

visualized by PET (Haberkorn et al., 2004). Hereby, gene expression throughout the body

ROLE IN DRUG DISCOVERY 369

370 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

is quantiﬁ ed in the presence or absence of a pharmacological agent (Iyer et al., 2001).

For example, human T-cells were engineered to express a reporter gene whose product is

visualized by PET and whose promoter is regulated by an intracellular transcription factor.

The T-cell receptor-dependent nuclear factor of activated T-cells (NFAT) have been used

for molecular imaging of T-cell activation in vivo (Ponomarev et al., 2001). This tool has

the potential to provide in vivo insights into the course of normal and pathologic immune

responses as well as temporal dynamics and immune regulation at different stages of dis-

ease and following therapy, such as adoptive immunotherapy for cancer, vaccination, or

immunosuppressive drugs.

In Vivo Characterization of Novel Targets: Deﬁ ning In Vivo Gene Function, Protein–

Protein Interaction, Signaling Pathways, and Gene Regulation Protein–protein

interaction or receptor occupancy to complex cellular and biological responses can only

be observed in the living being using PET. The dopamine D

receptor, for example, is

implied in some of the behavioral and psychological symptoms of dementia (BPSD),

including aggressiveness, wandering, and sleep disturbance. PET using a speciﬁ c probe

for the dopamine D

receptor in patients treated with risperidone, an approved drug for

Alzheimer disease, was instrumental in documenting the fact that risperidone enhanced

the binding potential of the dopamine D

receptor without interacting with it directly, thus

elucidating a molecular mechanism of action that was linked to some of its clinical effects

(Meguro et al., 2004).

Similarly, radiolabeled enzyme inhibitors or substrate are used to quantify in vivo

enzyme activity. Failure of cells to repair endogenous DNA continuously has, for example,

been identiﬁ ed as an initial step in the evolution of many human malignancies and has also

been related to therapy response and outcome. One of the key enzymes ensuring continu-

ous and faithful DNA repair is alkylguanine-DNA alkyltransferase (AGT), which is often

faulty in breast cancer. A radiolabeled substrate for AGT has been developed that monitors

in vivo: for example, in a rat model of breast cancer, the activity of this enzyme in vivo

(Zheng et al., 2003). This sets the stage to investigate, for example, whether novel antican-

cer compounds may be modulated in their efﬁ cacy by the presence or functional activity of

AGT, which may be distributed heterogeneously throughout the tumor.

PET advances the study of gene regulation, and signal transduction cascades from

extracellular or in vitro experimental systems to the in vivo situation. It monitors temporal

and spatial dynamics of expression and function of speciﬁ c genes and intrinsic binding

speciﬁ cities of proteins that govern gene regulation during disease progression or under

the inﬂ uence of a pharmacological agent. The study of gene regulation of human breast

cancer cells in vivo, for example, is suitable to gain more insights into the molecular events

accompanying disease progression (Berger and Gambhir, 2000). With the help of small

animal imaging, speciﬁ c binding of the tumor suppressor gene p53 to its viral target, the

large T-antigen, is detected and quantiﬁ ed in living mice (Luker et al., 2002). Transgenes

brought into living animals are instrumental in investigating the transcriptional activation

of endogenous genes by PET. For example, a reporter gene is under the transcriptional

control of regulatory elements speciﬁ c for the p53 tumor suppressor gene functions as an

in vivo readout system to measure the p53-governed signal transduction cascade and acti-

vation of p53-dependent genes in murine xenograft models of human tumors (Doubrovin

et al., 2001). These examples illustrate that PET advances functional proteomics from the

extracellular and the in vitro scenario to the in vivo situation and aids drug development

through direct interrogation of molecular targets within intact animals.

14.3.2 Preclinical Studies

Small animal PET scans transform established animal disease models to functionally vali-

dated visual models for drug development where the biological role of a given molecular

target structure is observed directly and the effect of a compound targeting that structure

can be visualized and assessed simultaneously in a functional readout system.

Organ Distribution, Pharmacokinetics, and Toxicity of Leads or Candidates PET

quantiﬁ es, for example, uptake and distribution of novel neuropharmacological agents

in experimental animals and links their binding to their respective target structure to

physiological and biochemical processes, such as glucose metabolism or blood ﬂ ow, which

can then be used to extrapolate those data to correlate physiological and pharmacological

effects to optimize drug design and ultimately, clinical treatment (Halldin et al., 2001).

Salazar and Fischman (1999) evaluated BMS 181101, a drug with agonist and antagonist

activity at various sites in the serotonin system. The

C-labeled form of the drug was used

to show that the residence time in the brain was short and, as a result, speciﬁ c binding could

not be determined by external imaging. These studies on receptor occupancy showed that

the drug may have a narrow therapeutic index and may not be suitable for once- or twice-

a-day dosage.

Monoamine oxidases (MAOs) A and B are ﬂ avoproteins residing in the outer mitochon-

drial membrane. They participate in the processing of various neurotransmitters; MAO-A

predominantly oxidizes serotonin, norepinephrine, and dopamine. Selective MAO-A inhib-

itors are used to treat patients with Alzheimer’s disease, Parkinson’s disease, or depression

and other psychiatric disorders. A radiolabeled reversible and speciﬁ c inhibitor of MAO-A

was instrumental in quantifying and evaluating its uptake into the brain of baboons, its tis-

sue distribution across the brain, and its binding kinetic with MAO-A (Bottlaender et al.,

2003), creating a tool to evaluate in vivo pharmacokinetics, substrate-binding afﬁ nity, and

functional efﬁ cacy of new MAO-A inhibitors.

Acetylcholinesterase (AChE) is an important target of several marketed drugs treating

neurodegenerative diseases. Novel radiotracers, such as [

C]physostigmine, depict regional

distribution of AChE activity, as it is known from histological studies on postmortem brains

and allow quantifying the occupancy of binding sites on AChE by AChE inhibitors. This

facilitates pharmakokinetic studies across brain regions at the molecular level. As a comple-

mentary approach, radiolabeled acetylcholine analog substrates measure AChE activity and

quantify the efﬁ cacy of AChE inhibitors. For example, in patients with Alzheimer’s disease,

endogenous AChE activity is known to be reduced. Here, the ability of current drugs such

as donepezil and rivastigmine to inhibit the activity of AChE allows therapeutic monitoring

of these drugs and in the design of improved versions of current drugs by using PET imag-

ing as an indicator for target inhibition in vivo (Shinotoh et al., 2004).

Achieving the right bioavailability and organ distribution of novel compounds which

have passed all the in vitro tests and possessed the desired mechanism of action in cell

cultures can be a challenging endeavor. For example, promising cancer compounds with

a dual-action mechanism that inhibit both topoisomerase I and II may suffer from poor

extravascular distribution, which limits their therapeutic efﬁ cacy even for compounds with

high in vitro cytotoxic proﬁ le. PET has been instrumental in studying biodistribution of

lead candidates in mice bearing human tumor xenografts and selecting candidate with the

best biodistribution, clearance, metabolic stability, and ultimately, efﬁ cacy proﬁ le against

human tumors in a mouse model (Osman et al., 2001).

ROLE IN DRUG DISCOVERY 371

372 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

In Vivo Mechanism of Action of Leads or Candidates in Cells and Animals Small

animal PET characterizes exact drug–target interactions and elucidates the mechanism

of action of pharmacologic agents that could not otherwise be observed. A dual PET

strategy, for example, has been instrumental in deﬁ ning the effect of methamphetamine

and scopolamine in the brain. Researchers investigated simultaneously the binding of

these agents to the dopamine D

receptor in the striatum of conscious monkeys as well

their ability to functionally activate signaling cascades known to be associated with the D

receptor (i.e., the D

receptor-coupled cAMP messenger determined by phosphodiesterase

IV activity as a readout system) (Tsukada et al., 2001). It was shown that neither compound

interfered with the binding capacity of the D

receptor, but both compounds enhanced

PDE4 activity. Dual PET analysis thus examined simultaneously receptor occupancy

and second messenger systems and thereby delivered in vivo functional insights on the

molecular mode of action of motion-sickness drugs.

Radiolabeled ligands targeting the imidazoline receptor (Hudson et al., 2003) not only

facilitated visualization of their uptake kinetics and biodistribution but also linked their

interaction with their respective receptor to a biochemical signaling cascade and ultimately

to a biological response. PET shows that receptor activation is associated with ameliora-

tion of depression in an established rat model of this disease. This experimental setup now

sets the stage to investigate simultaneously in vivo bioavailability, receptor binding, and

saturation as well as functional efﬁ cacy (i.e., behavioral changes of the rat model) of novel

compounds targeting these receptors.

Similarly, radiolabeled inhibitors of the 11β-hydroxylase, an enzyme involved in the

biosynthesis of cortisol and aldosterone and also known to function as anesthetic drug

through the GABAergic system, are useful not only for in vivo imaging of the adrenocortex

but also to investigate the function, dynamics, and kinetics of narcotic drugs with PET that

are designed to work by targeting the GABAA receptors (Mitterhauser et al., 2003).

Preclinical Proof of the Principle of Novel Treatment Paradigms Applying PET to small

animals not only visualizes molecular and biochemical processes in the in vivo situation

compared to extracellular or in vitro studies but also provides a unique opportunity to

monitor cell behavior and processes that were not accessible for direct investigation and

hence proof of principle before. This includes cell trafﬁ cking in vivo as it occurs during

tumor metastasis or during immune responses. PET detects, in vivo and in real time, tumor

cell migrating in mouse models, visualizes directly their organ preference, and investigates

how the metastatic potential can be modulated by pharmacological agents that alter the cell-

surface proﬁ le of the tumor cells, such as their expression pattern of adhesion molecules

(Koike et al., 1995).

Adoptive immune therapy using bioengineered T-cells to target tumor cells speciﬁ cally

and systemically is an emerging treatment paradigm in oncology. These T-cells can be

equipped with a marker gene that permits their tracking in vivo without compromising

their ability to identify and kill distant metastatic tumor cells, as shown recently in a murine

model of a human tumor. This technique for imaging the migration of ex vivo-transduced

antigen-speciﬁ c T-cells in vivo is informative, nontoxic, and potentially applicable to hu-

mans (Koehne et al., 2003). As PET permits monitoring and quantifying of in vivo expres-

sion of transgenes, it assess the tissue speciﬁ city, quality, and sustainability of foreign

gene expression in the context of preclinical gene therapy studies involving, for example,

adenovirus-mediated gene transfer techniques (Mayer-Kuckuk et al., 2003; Groot-Wassink

et al., 2004). Further, xenograft models of human tumors have been used to investigate

in vivo the feasibility of suicide-gene transfer for the treatment of prostate cancer (Pantuck

et al., 2002).

14.3.3 Clinical Studies

Patient Selection Through Noninvasive Identiﬁ cation of Molecular Targets With the

arrival of novel drugs designed to interfere with speciﬁ c molecular processes, the need to

monitor in vivo the proposed mechanism of action at the molecular level and to identify

prospectively patients who express the molecular target structure in sufﬁ cient amounts or

do not express target structures that may predispose them for side effects has emerged. PET

offers the unique opportunity to monitor patients in vivo in a noninvasive fashion for both

(Fishman et al., 1997a,b; Solomon et al., 2003). Good examples are experimental drugs

such as Erbitux, which targets the epidermal growth factor receptor, a cell-surface receptor

tyrosine kinase whose expression and activity is linked to the genesis of several cancers.

Efﬁ cacy of this drug in the individual patient is probably correlated with the expression

level of the EGF receptor as well as its functional activity. A PET biomarker recognizing

speciﬁ cally the functional active EGF-R tyrosine kinase was developed that also acts as an

inhibitor of this tumor-associated target (Ben-David et al., 2003) and should permit both

identiﬁ cation of patients suitable for treatment and the monitoring of therapeutic efﬁ cacy.

Similarly, very recently the ﬁ rst antiangiogenesis drug, Avastin, was approved for can-

cer treatment. Avastin targets the vascular–endothelial growth factor (VEGF) receptor on

endothelial cells and is designed to shut off tumor vascularization and thereby to kill the

cancer. It would be helpful both to monitor treatment efﬁ cacy and to identify patients who

are likely to beneﬁ t from this treatment by visualizing the expression of VEGF. A radio-

labeled probe to this end has been developed and has been shown in small animal models

of human tumors to recognize VEGF with high afﬁ nity and high speciﬁ city in both the

primary tumor and all metastatic sites, preparing the way for similar studies in patients

receiving this experimental drug (Collingridge et al., 2002).

Therapeutic inhibitors of intracellular signaling cascades are thought to constitute more

speciﬁ c and less toxic anticancer compounds. However, appropriate tools to assess in vivo

precisely the molecular mechanism by which a novel compound is believed to act have

been lacking, and PET may well ﬁ ll this gap: for example, an experimental compound,

17-allylaminogeldanamycin (17-AAG), designed to inhibit the heat-shock protein Hsp90.

17-AAG causes degradation of the oncogene Her2 as well as other proteins and blocks

tumor growth in preclinical animal experiments. Her2 is implemented in the pathogenesis

of breast cancer and other solid tumors, and 17-AAG is the ﬁ rst experimental drug tar-

geting this pathway to enter clinical trials. PET permits in vivo imaging of Her2 and the

kinetics of its degradation under the inﬂ uence of 17-AAG in animal tumors (Smith-Jones

et al., 2004). This approach allows noninvasive imaging of the pharmacodynamics of a

targeted drug and will facilitate the rational design of combination therapy based on tar-

get inhibition. Early identiﬁ cation of patients who are because of their pharmacogenomic

proﬁ le likely to be more susceptible to certain side effects of novel drugs is of pivotal

importance to smoothen the drug development process. PET has been instrumental in

identifying patients with a distinct pattern of histamine H

-receptor expression in the brain

who are most likely to suffer more severely from subjective sleepiness and objective seda-

tion under the inﬂ uence of histamine H

-receptor antagonists (Tashiro and Yanai, 2003).

Histamine H

-receptor occupancy, which is then used to assess the risk for sedation in the

individual patient. This experimental setup will also be useful in evaluating the overall

ROLE IN DRUG DISCOVERY 373

374 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

potential of novel antihistamines to elicit this side effect during preclinical and early stage

clinical studies.

Similarly, PET is helpful in determining whether apparent unresponsiveness to treat-

ment is caused by failure to deliver the drug to the site of disease or by other means of

pharmacodynamic resistance. For example,

C-labeled phenytion ([

C]DPH) was instru-

mental in quantifying the regional brain concentration of this drug in patients with medi-

cally resistant epilepsy. It was shown that the epileptogenic focus accumulated amounts of

the radiolabeled drug similar to those in normal brain areas in the same patient, pointing

to a pharmacodynamic rather than a pharmacokinetic cause for the resistance observed

(Baron et al., 1983). More recently, others have provided PET data to suggest that the corti-

cal GABA-A receptor plasticity differs in a normal brain area, and epileptogenic focus may

play a role in causing drug-resistant epilepsy (Marrosu et al., 2003).

In Vivo Biodistribution Radiolabeled drugs of interest (DOIs) enable the investigator to

determine the biodistribution or pharmacokinetics and target tissue in normal organs as a

function of time and dosage. This is even more important, as innovative drugs often times

target very speciﬁ cally distinct molecular processes and biochemical pathways that play

out differently in different organs. Understanding the in vivo distribution and observing in

vivo drug effects across organs and tissue can be crucial in deﬁ ning the appropriate dose,

understanding the risks of potential side effects, and obtaining some insight as to how the

disease itself, potential co-morbidities, or even unrelated disease states may interfere with

the kinetic, distribution, and site-speciﬁ c effects of a compound. Pathophysiological changes

in tissues, for example, may alter the delivery of drugs and make extrapolations from normal

tissue data unreliable. Radiolabeled DOI permit comparative pharmacokinetic analysis

under normal and pathological conditions: for example, to investigate tissue distribution

of antibiotics and determine actual concentrations at the sites of infection (Fischman et al.,

1997a–c). Inappropriate drug metabolism and pharmacokinetics is estimated to cause 40 %

of clinical drug development failures (Lappin and Garner, 2003). Microdosing combines

PET with accelerator mass spectrometry to investigate human metabolism by providing

both pharmacodynamic and pharmakokinetic information of minute amounts of drugs. It

permits safe human studies of experimental drugs in the very early development process,

thus reducing sunk costs associated with later-stage clinical trail failure as well as limiting

the need for animal experiments (Bergstrom et al., 2003).

PET has been instrumental in deﬁ ning the biodistribution and concentration range of

antibiotics such as trovoﬂ oxacin in human volunteers. It was determined that concentra-

tions sufﬁ cient to kill most members of enterobacteriacaeae and anaerobes were achieved

in all organs, including the brain (Babich et al., 1996; Fischman et al., 1997c, 1998). PET

also assists in optimizing formula or administration route of existing drugs. Two stud-

ies were carried out to evaluate drug distribution in the gut and in the lung. Producing

a true tracer situation is important to these studies. In the analysis of modiﬁ ed-release

formulations, drugs such as diltiazem are formulated with small amounts of

152

Sm sa-

marium oxide. The tablets can then be activated by neutron activation to produce radioac-

tive

153

Sm and can be followed in vivo using planar imaging. This phase I study allowed

quantitative distribution of the tablet in the gastrointestinal tract (Maziere et al., 1992)

In another example, the lung distribution of the corticosteroid triamcinolone acetonide

was studied using the dispenser, Azmacort, a pressurized aerosol metered-dose inhaler

formulation. The steroid was radiolabeled with

C and introduced into the dispenser. The

time–activity curve obtained using PET showed a signiﬁ cant increase in the steroid in the

lung and a signiﬁ cant decrease in the mouth. These data were submitted to the U.S. Food

and Drug Administration as supportive evidence of the Azmacort inhaler’s superiority

(Berridge and Heald, 1999). The pharmacokinetic proﬁ les of BCNU, an anticancer com-

pound, was compared after intravenous administration and intraarterial administration

using [

C]BCNU in patients with recurrent gliomas. Using intraarterial administration,

BCNU levels in tumors average 50-fold greater than comparable intravenous administra-

tion. The authors suggest that within this small group of patients (N ⫽ 10), the degree of

metabolic trapping of BCNU in tumors correlated with the clinical response to this agent

(Tyler et al., 1986).

It is also possible to determine the extent of interaction of competing drugs on the phar-

macokinetics and tissue concentration of the DOI. In this case, the DOI is radiolabeled and

the organ distribution and kinetics are determined before and after the administration of

a nonradioactive drug with a particular pharmacological proﬁ le. In this manner it may be

possible to determine the receptor afﬁ nity of a novel neuroleptic (DOI) in the presence of

a well-characterized drug having similar pharmacology.

Pharmacodynamic Evaluation of Biological Parameters to Select Dose and Assess

Therapy One of the most important and difﬁ cult steps in the drug development process

is deﬁ ning the dose–response relationship. Much can be learned about a candidate drug by

measuring its effect on physiological parameters and biochemical processes. For example,

several ﬂ uoroquinolone antibiotics have been evaluated as to their ability to alter cerebral

blood ﬂ ow, glucose metabolism, and oxygen consumption (Bednarczyk et al., 1990; Green

et al., 1991). Other studies aim at establishing in vivo a relationship between plasma drug

concentration and drug–target interactions in distinct organs such as the brain. If the new

drug acts by binding speciﬁ cally to a cell-surface or nuclear receptor, PET determines

receptor binding parameters (afﬁ nity and capacity) in vivo either preclinically in animals or

in early-stage trials in humans in binding-site experiments, whereby high-speciﬁ c-activity

radiolabeled drug is administered with increasing amounts of unlabeled drug (Morris

et al., 1996). These data should be useful in the determination of dosing, particularly for

neuroleptic drugs. If maximal therapeutic effect can be demonstrated to occur at a known

level of receptor occupancy and that degree of occupancy can be related to a given dose

level, increasing doses will probably result in increased incidence of side effects without

further therapeutic beneﬁ t. Similarly, Aprepitant, a highly selective substance P antagonist,

is an experimental drug designed to ameliorate chemotherapy-induced nausea and emesis.

Its efﬁ cacy is dictated by its ability to bind to the NK(1) receptor in the brain. PET has been

instrumental in a phase I study to assess NK(1) brain occupancy as a function of aprepitant

dose in healthy volunteers; these data now allow to predict NK(1) occupancy from plasma

drug concentrations and can be used to guide dose selection for clinical trials of NK(1)

receptor antagonists in central therapeutic indications (Bergstrom et al., 2004).

The putative antipsychotic drug M100907 was studied indirectly using

C-labeled spi-

perone, which binds to both the dopamine D

receptor and the 5-HT

serotonin receptor

(Offord et al., 1999). The therapeutic index of M100907 was deﬁ ned in phase I single- and

multiple-dose tolerability studies. PET was then used to conﬁ rm the mechanism of action

of M100907 in humans. The resulting data also helped investigators to deﬁ ne an appro-

priate dose range and regimen for subsequent clinical studies in schizophrenia patients.

(Offord et al., 1999).

C-labeled raclopride, a dopamine receptor antagonist, has been

used to measure increases in the neurotransmitter dopamine indirectly as a function of the

pharmacologic action of an amphetamine-like potential drug candidate under investigation

ROLE IN DRUG DISCOVERY 375

376 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

for schizophrenia, building on the insight that elevated synaptic dopamine concentrations

are associated with this disease (Breier et al., 1997).

PET has long enjoyed a pivotal position in the evaluation of tumor response to chemo-

therapy and hence patient management. This includes, foremost, the use of FDG to quantify

tumor metabolism as an indicator of tumor response to treatment with chemotherapeutic

agents (Eckelman et al., 2000; Lammertsma et al., 2001). In addition, biochemical indica-

tors of cellular proliferation, such as the incorporation of thymidine in actively dividing

cells, can be exploited to investigate in vivo in humans the effect of investigational cancer

drugs designed to stop the proliferation of tumor cells and to correlate the biological read-

out with in vivo plasma concentration and biodistribution data (Wells et al., 2003). Simi-

larly, anticancer drugs have been proposed to exert their effect by causing programmed cell

death of tumor cells. The molecular marker of apoptosis Annexin V has been instrumental

in quantifying in cancer patients undergoing chemotherapy the extent of tumor cell apopto-

sis. The investigators were further able, despite the small number of study patients, to relate

overall survival and progression free survival to the uptake of the apoptosis tracer, suggest-

ing that the radiolabeled apoptosis marker Annexin may be used as a surrogate marker to

assess therapeutic efﬁ cacy both for approved drugs in the clinical management of cancer

patients and in early clinical trials to monitor the efﬁ cacy of novel investigational cancer

compounds (Belhocine et al., 2002). In this context it is of note that cancer drug develop-

ment programs are speciﬁ cally burdened by long observation times required before clinical

efﬁ cacy is satisfactorily assessed and higher failure rate in the more expensive late-stage

clinical studies.

Similarly, the experimental cancer drug Comprestatin A4 phosphate is designed to dis-

rupt tumor blood supply by binding to endothelial cell tubulin, thereby causing morpho-

logical changes of endothelials cells that lead to their disruption. This mechanism of action

has been conﬁ rmed in animal studies. PET was instrumental in visualizing this effect in

patients as part of a phase I clinical trial and to assess cooperation and synergetic effects of

Comprestatin A4 phosphate when combined with cisplatin and to aid dose selection for a

subsequent phase II trial (West and Price, 2004). In a phase II study with the experimental

drug razoxane, also designed to interfere with tumor vascularization, PET-guided visual-

ization of the vascular physiology using [

O]H

and [

O]C as tracers monitored over time

the effect of razoxane on renal tumor perfusion compared to normal tissue perfusion. PET

provided valuable information on the in vivo biology of angiogenesis in patients and was

instrumental in assessing the effects of antiangiogenic therapy (Anderson et al., 2003).

14.4 SUMMARY AND OUTLOOK

The challenges for the pharmaceutical industry in the drug discovery and development

process range from the evaluation of potential new drug candidates, the determination

of drug pharmacokinetics and pharmacodynamics, the measurement of receptor occu-

pancy as a determinant of drug efﬁ cacy, and the pharmacological characterization of

mechanisms of action. Among the tools with signiﬁ cant potential to reduce overall costs

and improve the reliable identiﬁ cation of promising new compounds is PET. Often, the

mechanism of action of new compounds is well deﬁ ned based on an array of extracellular

and in vitro studies. Early recognition that the proposed mechanism of action also holds

true in an in vivo scenario such as a wild-type animal or an animal model representing the

target disease is of signiﬁ cant value for the drug-development process. Although PET may

appear to be expensive, by providing the basis for early go/no-go decisions, it can actu-

ally be cost-effective. With the help of PET, the drug candidate is being tested in an intact

species. Imaging is especially important in paradigms where animals are studied before

and after treatment with the drug candidate. Because paired statistics can be carried out in

the same animal, fewer animals are needed. Therefore, in vivo imaging has an advantage

over the autoradiographic approaches that are carried out in vivo but require the sacriﬁ ce

of the animal after each study.

Clinical-stage development programs often could be accelerated if reliable and predic-

tive surrogate markers were available that could be monitored noninvasively. This applies,

for example, to cancer trials, where PET could establish almost instantaneous whether a

novel compound is capable of affecting tumor cell proliferation. For example, FDG has

been instrumental in quantifying tumor response to chemotherapy (Lammertsma, 2001).

Fast, efﬁ cient, and cost-effective use of PET in the drug discovery and development pro-

cess relies on the availability of well-characterized binding sites and suitable radioli-

gands. If these are not readily obtainable, well-established radiopharmaceuticals such

as O-15 water for blood ﬂ ow and F-18 FDG for glucose metabolism or well-established

radiolabeled ligands that can measure the effect of drugs indirectly may provide valuable

alternatives.

Before PET can be ﬁ rmly integrated in the clinical evaluation of experimental drugs in

multicenter international trials, standards deﬁ ning the visual and analytical methods to be

used for quantiﬁ cation and reporting of PET data need to be established. Both the European

Organization for Research and Treatment of Cancer and the National Cancer Institute in

the United States have made recommendations along these lines for the use of PET in

assessing tumor response to chemotherapy (Eckelman et al., 2000). In phase I and II hu-

man studies, classic PK measurements can then be coupled with imaging measurements to

deﬁ ne an optimal dosing schedule, help formulate the design of phase III studies, and thus

contribute to improving their success rates. It is to be expected that pharmacogenomics

will identify new surrogate markers for therapy monitoring which may represent potential

new tracers for imaging suitable to accelerate and improve the clinical drug development

process and ultimately to guide treatment decisions.

REFERENCES

Anderson, H., Yap, J. T., Wells, P., Miller, M. P., Propper, D., Price, P., Harris, A. L. Measurement of

renal tumour and normal tissue perfusion using positron emission tomography in a phase II clini-

cal trail of razoxane. Br. J. Cancer, July 21, 2003, 89(2):262–267.

Babich, J. W., Rubin, R. H., Graham, W. A., Wilkinson, R. A., Vincent, J., Fishman, A. J. 18F-

labeling and biodistribution of the novel ﬂ uoro-quinolone antimicrobial agent, trovaﬂ oxacin (CP

99,219). Nucl. Med. Biol., Nov. 1996, 23(8):995–998.

Barbarich, N. C. Kaye, W. H., Jimerson, D. Neurotransmitter and imaging studies in anorexia ner-

vosa: new targets for treatment. Curr. Drug Target CNS Neurol. Disord. Feb. 2003, 2(1):61–72.

Baron, J. C., Roeda, D., Munari, C., Crouzel, C., Chodkiewicz, J. P., Comar, D. Brain regional

pharmacokinetics of

C-labeled diphenylhydantoin: positron emission tomography in humans.

Neurology (Cleveland), 1983, 33:580–585.

Bednarczyk, E. M., Green, J. A., Nelson, A. D., et al. Comparison of the effect of temaﬂ oxacin,

ciproﬂ oxacin, or placebo on cerebral blood ﬂ ow, glucose and oxygen metabolism in healthy sub-

jects by means of positron emission tomography. Clin. Pharmacol. Ther., 1990, 50:165–171.

REFERENCES 377

378 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

Belhocine, T., Steinmetz, N., Hustinx, R., Bartsch, P., Jerusalem, G., Seidel, L., Rigo, P., Green, A.

Increased uptake of the apoptosis-imaging agent (99m)Tc recombinant human Annexin V in hu-

man tumors after one course of chemotherapy as a predictor of tumor response and patient prog-

nosis. Clin. Cancer Res., Sept. 2002, 8(9):2766–2774.

Bench, C. J., Lammertsma, A. A. Dolan, R. J. Grasby, P. M., Warrington, S. J., Gunn, K.,

et al., Dose dependent occupancy of central dopamine D2 receptors by the novel neuroleptic CP-

88,059-01: a study using positron emission tomography and

C-raclopride. Psychopharmacol-

ogy, 1993, 112(2–3):308–314.

Ben-David, I., Rozen, Y., Ortu, G., Mishani, E. Radiosynthesis of ML03, a novel positron emission

tomography biomarker for targeting epidermal growth factor receptor via the labelling synthon:

[

C]acryloyl chloride. Appl. Radiat. Isot., Feb. 2003, 58(2):209–217.

Berger, F., Gambhir, S. S. Recent advances in imaging endogenous or transferred gene expression

utilizing redionuclide technologies in living subjects: applications to breast cancer. Breast Cancer

Res., 2001, 3(1):28–35.

Bergstrom, M., Grahnen, A., Langstrom, B. Positron emission tomography microdosing: a new

concept with application in tracer and early clinical drug development. Eur. J. Clin. Pharmacol.,

Sept. 2003, 59(5–6):357–366.

Bergstrom, M., Hargreaves, R. J., Burns, H. D., Goldberg, M. R., Sciberras, D., Reines, S. A., Petty,

K. J., Orgen, M., Antoni, G., Langstrom, B., Eskola, O., Schenin, M., Solin, O., Majumdar, A. K.,

Constanzer, M. L., Battisti, W. P., Bradstreet, T. E., Gargano, C., Hietala, J. Human positron emis-

sion tomography studies of brain neurokinin 1 receptor occupancy by aprepitant. Biol. Psychiatry,

May 15, 2004, 55(10):1007–1012.

Berridge, M. S., Heald, D. L. In vivo characterization of inhaled pharmaceuticals using quantitative

positron emission tomography. J. Clin. Pharmacol., 1999, Suppl., 25S–29S.

Bottlaender, M., Dolle, F., Guenther, I., Roumenov, D., Fuseau, C., Bramoulle, Y., Curet, O., Jegham,

J., Pinquier, J. L., George, P., Valette, H. Mapping the cerebral monoamine oxidase type A: posi-

tron emission tomography characterization of the reversible selective inhibitor [

C]beﬂ oxatone.

J. Pharmacol. Exp. Ther., May 2003, 305(2):467–473.

Breier, A., Su, T.-P., Saunders, R., et al. Schizophrenia is associated with elevated amphetamine

induced synaptic dopamine concentrations: evidence from a novel positron emission tomography

method. Proc. Natl. Acad. Sci. USA, 1997, 94, 2569–2574.

Cagnin, A., Brooks, D. J. Kennedy, A. M., Gunn, R. N., Myers, R., Turkheimer, F. E., Jones, T.,

Banati, R. B. In vivo measurement of activated microglia in dementia. Lancet, Aug. 11 2001,

358(9280):461–467.

Cherry, S. R. Fundamentals of positron emission tomography and applications in preclinical drug

development. J. Clin. Pharmacol., May 2001, 41(5):482–491.

Cherry, S. R., Shao, Y., Silverman, R. W., Meadors, K., Siegel, S., Chatziioannou, A., Yound, J. W.,

Jones, W. F., Moyers, J. C., Newport, D., Boutefnouchel, A., Farquhar, T. H. Andreaco, M., Pulus,

M. J., Binkley, D. M., Nutt, R., Phelps, M. E. Micro PET: a high resolution PET scanner for imag-

ing small animals. IEEE Trans. Nucl. Sci., 1997, 44:1161–1166.

Collingridge, D. R., Carroll, V. A., Glaser, M., Aboagye, E. O., Osman, S., Hutchinson, O. C.,

Barthel, H., Luthra, S. K., Brady, F., Bicknell, R., Price, P., Harris, A. L. The development of

[(124)I]iodinated-VG76e: a novel tracer for imaging vasucular endothelial growth factor in vivo

using positron emission tomography. Cancer Res., Oct. 15, 2002, 62(20):5912–5919.

DaSilva, J. N., Lourenco, C. M., Meyer, J. H., Hussey, D., Potter, W. Z., Houle, S. Imaging cAMP-

speciﬁ c phosphodiesterase-4 in human brain with R-[

C]rolipram and positron emission tomog-

raphy. Eur. J. Nucl. Med. Mol. Imaging, Dec. 2002, 29(12):1680–1683.

Dehdashti, F., Flanagam, F. L., Mortimer, J. E., Katzenellenbogen, J. A., Welch, M. J., Siegel, B. A.

Positron emission tomographic assessment of “metabolic ﬂ are” to predict response of metastatic

breast cancer to antiestrogen therapy. Eur. J. Nucl. Med., Jan. 1999, 26(1):51–56.

Dickson, M., Gagnon, J. P. Key factors in the rising cost of new drug discovery and development.

Nature Drug Discov., May 2004.

Doubrovin, M., Ponomarev, V., Beresten, T., Balatoni, J., Bornmann, W., Finn, R., Humm, J., Larson,

S., Sadelain, M., Blasberg, R., Gelovani, Tjuvajev, J. Imaging transcriptional regulation of p53-

dependent genes with positron emission tomography in vivo. Proc. Natl. Acad. Sci. USA., July

31, 2001, 98(16):9300–9305.

Eckelman, W. C. Labeled agents as tracers and carriers: theory and practice. In Feinendegen, L. E.,

Shreeve, W. W., Eckelman, W. C., Bahk, Y. W., Wagner, H. N., Jr; (Eds.), Molecular Nuclear

Medicine: The Challenges of Genomics and Proteomics to Clinical Practice. Springer-Verlag,

New York, 2003, pp. 85–98.

Eckleman, W. C. Tatum, J. L., Kurdziel, K. A., et al. Quantitative analysis of tumor biochemistry

using PET and SPECT(1). Nucl. Med. Biol., 2000, 27, 633–691.

Fischman, A. J., Babich, J. W., Alpert, N. M., Vincent, J., Wilkinson, R. A., Callahan, R. J., Correia,

J. A., Rubin, R. H. Pharmacokinetics of

F-labeled trovaﬂ oxacin in normal and Escherichia coli–

infected rats and rabbits studied with positron emission tomography. Clin. Microbiol. Infect., June

1997a, 3(3):379.

Fischman, A. J., Babich, J. W., Alpert, N. M., Vincent, J., Wilkinson, R. A., Callahan, R. J., Correia,

J. A., Rubin, R. H. Pharmacokinetics of

F-labeled trovaﬂ oxacin in normal and Escherichia coli–

infected rats and rabbits studied with positron emission tomography. Clin. Microbiol. Infect., Feb.

1997b, 3(1):63–72.

Fischman, A. J., Alpert, N. M., Babich, J. W., Rubin, R. H. The role of positron emission tomography

in pharmacokinetic analysis. Drug Metab. Rev., Nov 1997c, 29(4):923–956.

Fischman, A. J., Babich, J. W., Bonab, A. A., Alpert, N. M., Vincent, J., Callahan, R. J., Correia, J. A.,

Rubin, R. H. Pharmacokinetics of [

F]trovaﬂ oxacin in healthy human subjects studied with posi-

tron emission tomography. Antimicrob. Agents Chemother., Aug 1998, 42(8):2048–2054.

Green, J. A., Bednarczyk, E. M., Miraldi, F., et al. Reduction in cerebral glucose metabolism caused

by quinoline administration in healthy subjects. Nucl. Med. Biol. Int. J. Appl. Radiat. Isot. P. B,

1991, 19:126–129.

Groot-Wassink, T., Aboagye, E. O., Wang, Y., Lemoine, N. R., Reader, A. J., Vassaux, G. Quantitative

imaging of Na/I symporter transgene expression using positron emission tomography in the living

animal. Mol. Ther, Mar. 2004, 9(3):436–442.

Haberkorn, U., Altmann, A., mier, W., Eisenhut, M. Impact of functional genomics and proteomics

on radionuclide imaging. Semin. Nucl. Med., Jan. 204, 34(1):4–22.

Halldin, C., Gulyas, B., Farde, L. PET studies with carbon-11 radioligands in neuropsychopharmaco-

logical drug development. Curr. Pharm. Des., Dec. 2001, 7(18):1907–1929.

Herbst, R. S., Mullani, N. A., Davis, D. W., Hess, K. R., McConkey, D. J., Charnsangavej, C.,

O’Reilly, M. S., Kim, H. W., Baker, C., Roach, J., Ellis, L. M., Rashid, A., Pluda, J., Bucana, C.,

Madden, T. L., Tran, H. T., Abbruzzese, J. L. Development of biologic markers of response and

assessment of antiangiogenic activity in a clinical trial of human recombinant endostatin. J. Clin.

Oncol., Sept. 15, 2002, 20(18):3804–3814.

Hudson, A. L., Tyacke, R. J., Lalies, M. D., Davies, N., Finn, D. P., Marti, O., Robinson, E.,

Husbands, S., Minchin, M. C., Kimura, A., Nutt, D. J. Novel ligands for the investigation of imid-

azoline receptors and their binding proteins. Ann. N. Y. Acad., Dec. 2003, 1009:302–308.

Imahori, Y., Fujji, R., Tujino, H., Kimura, M., Mineura, K. Postitron emission tomography: measure-

ment of the activity of second messenger systems. Methods, July 2002, 27(3):251–262.

Iyer, M., Wu, L., Carey, M., Wang, Y., Smallwood, A., Gambhir, S. S. Two-step transcriptional am-

pliﬁ cation as a method for imaging reporter gene expression using weak promoters. Proc. Natl.

Acad. Sci. USA, Dec. 4 2001, 98(25):14595–14600.

Jagoda, E. M., Kiesewetter, D. O., Shimoji, K., Ravasi, L., Yamada, M., Gomeza, J., Wess, J.,

Eckelman, W. C. Regional brain uptake of the muscarinic ligand, [

F]FP-TZTP, is greatly

REFERENCES 379

380 APPLICATIONS OF NUCLEAR IMAGING IN DRUG DISCOVERY AND DEVELOPMENT

decreased in M2 receptor knockout mice but not in M1, M3 and M4 receptor knockout mice.

Neuropharmacology, Apr. 2003, 44(5):653–661.

Koehne, G., Doubrovin, M., Doubrovina, E., Zanzonico, P., Gallardo, H. F., Ivanova, A., Balatoni,

J., Teruya-Feldstein, J., Heller, G., May, C., Ponomarev, V., Ruan, S., Finn, R., Blasberg, R. G.

Bornmann, W., Rivierre, I., Sadelain, M., O’Reilly, R. J., Larson, S. M., Tjuvajev, J. G. Serial

in vivo imaging of the targeted migration of human HSV-TK-transduced antigen-speciﬁ c lympho-

cytes Nat. Biotechnol., Apr. 2003, 21(4):405–413.

Koike, C., Oku, N., Watanabe, M., Tsukada, H., Kakiuchi, T., Irimura, T., Okada, S. Real-time PET

analysis of metastatic tumor cell trafﬁ cking in vivo and its relation to adhesion properties. Bio-

chim. Biophys. Acta, Sept. 13, 1995, 1238(2):99–106.

Lammertsma, A. A. Measurement of tumor response using [

F]-2-ﬂ uoro-2-deoxy-D-glucose and

positron-emission tomography, J. Clin. Pharmacol., 2001, Suppl., 104S–106S.

Lappin, G., Garner, R. C. Big physics, small doses: the use of AMS and PET in human microdosing

of development drugs. Nat. Rev. Drug Discov., Mar. 2003, 2(3):233–240.

Lawrence, S. Priming the pipeline, Acumen Vol. II, p. 20, 2004.

Lehmann Brothers and McKinsey & Co., The Fruits of Genomics, Jan. 2001.

Luker, G. D., Sharma, V., Pica, C. M., Dahlheimer, J. L., Li, W., Ochesky, J., Ryan, C. E., Piwnica-

Worms, H., Piwnica-Worms, D. Noninvasive imaging of protein–protein interactions in living

animals. Proc. Natl. Acad. Sci. USA, May 14, 2002, 99(10):6961–6966.

Luker, G. D., Sharma, V., Piwnica-Worms, D. Visualizing protein–interactions in living animals.

Methods. Jan 2003, 29(1):110–122.

Marrosu, F., Serra, A., Maleci, A., Puligheddu, M., Biggio, G., Piga, M. Correlation between

GABA(A) receptor density and vagus nerve stimulation in individuals with drug-resistant partial

epilepsy. Epilepsy Res., June–July, 2003, 55(1–2):59–70.

Mayer-Kuckuk, P., Doubrovin, M., Gusani, N. J., Gade, T., Balatoni, J., Akhurst, T., Finn, R., Fong,

Y., Koutcher, J. A., Larson, S., Blasberg, R., Tjuvajev, J. G., Bertino, J. R., Banerjee, D. Imaging

of dihydrofolate reductase fusion gene expression in xenografts of human liver metastases of

colorectal cancer in living rats. Eur. J. Nucl. Med. Mol. Imaging, Sept., 2003, 30(9):1281–1291.

Maziere, B., Coenen, H. H., Halldin, C., Nagren, K., Pike V. W. PET radioligands for dopamine

receptors and re-uptake sites: chemistry and biochemistry. Int. J. Radiat. Appl. Instrum. P. B Nucl.

Med. Biol., 1992, 19(4):497–512.

Meguro, K., Meguro, M., Tanaka, Y., Akanuma, K., Yamaguchi, K., Itoh, M. Risperidone is effective

for wandering and disturbed sleep/wake patterns in Alzheimer’s disease. J. Geriatr. Psychiatry

Neurol., June 2004, 17(2):61–67.

Mitterhauser, M., Wadsak, W., Wabnegger L., Sieghart, W., Viernstein, H., Kletter, K., Dudczak, R.

In vivo and in vitro evaluation of [

F]FETO with respect to the adrenocortical and GABAergic

system in rats. Eur. J. Nucl. Med. Mol. Imaging, Oct. 2003, 30(10)1398–1401.

Morris, E. D., Babich, J. W., Alpert, N. M., Bonab, A. A., Livni, E., Weise, S., et al. Quantiﬁ cation

of dopamine transporter density in monkeys by dynamic PET imaging of multiple injections of

11C-CFT. Synapse, 1996, 24(3):262–272.

Myers, R., Hume, S. Small animal PET. Eur. Neuropsychopharmacol., Dec. 2002, 12(6):

545–555.

Offord, S. J., Wong, D. F., Nyberg, S. The role of positron emission tomography in the drug develop-

ment of M100907, a putative antipsychotic with a novel mechanism of action. J. Clin. Pharma-

col., 1999, Suppl., 17S–24S.

Osman, S., Rowlinson-Busza, G., Luthra, S. K., Aboagye, E. O., Brown, G. D., Brady, F.,

Myers, R., Gamage, S. A., Denny, W. A., Baguley, B. C., Price, P. M. Comparative biodistribution

and metabolism of carbon-11-labeled N-[2-(dimethylamino)ethyl]acridine-4-carboxamide and

DNA-intercalating analogues. Cancer Res. Apr. 1, 2001, 61(7):2935–2944.

Pantuck, A. J., Berger, F., Zisman, A., Nguyen, D., Tso, C. L., Matherly, J., Gambhir, S. S.,

Belldegrun, A. S. CL1-SR39: a noninvasive molecular imaging model of prostate cancer suicide

gene therapy using positron emission tomography. J. Urol., Sept. 2002, 168(3):1193–1198.

Piwnica-Worms, D., Schuster, D. P., Garbow, J. R., Molecular imaging of host–pathogen interactions

in intact small animals. Cell Microbiol., Apr. 2004, 6(4):319–331.

Ponomarev, V., Doubrovin, M., Lyddane, C., Beresten, T., Balatoni, J., Bornman, W., Finn, R.,

Akhurst, T., Larson, S., Blasberg, R., Sadelain, M., Tjuvajev, J. G. Imaging TCR-dependent

NFAT-mediated T-cell activation with positron emission tomography in vivo. Neoplasia, Nov.–

Dec. 2001, 3(6):480–488.

Qi, J., Leahy, R. M., Cherry, S. R., Chatziioannou, A., Farquhar, T. H. High-resolution 3D Bayesian

image reconstruction using the microPET small-animal scanner. Phys. Med. Biol., Apr. 1998,

43(4):1001–1013.

Salazar, D. E., Fischman, A. J. Central nervous system pharmacokinetics of psychiatric drugs. J. Clin.

Pharmacol., Aug. 1999, Suppl., 10S–12S.

Shinotoh, H., Fukushi, K., Nagatsuka, S., Irie, T. Acetylcholinesterase imaging: its use in therapy

evaluation and drug design. Curr. Pharm. Des., 2004, 10(13):1505–1517.

Smith, G. S., Koppel, J., Goldberg, S. Applications of neuroreceptor imaging to psychiatry research.

Psychopharmacol. Bull., Autumn, 2003, 37(4):26–65.

Smith-Jones, P. M., Solit, D. B., Akhurst, T., Afroze, F., Rosen, N., Larson, S. M. Imaging the phar-

macodynamics of HER2 degradation in response to Hsp90 inhibitors. Nat. Biotechnol., May 9,

2004.

Solomon, B., McArthur, G., Cullinane, C., Zalcberg, J., Hicks, R. Applications of positron emis-

sion tomography in the development of molecular targeted cancer therapeutics. BioDrugs, 2003,

17(5):339–354.

Tashiro, M., Yanai, K. A potential of positron emission tomography in the drug development of non-

sedative antihistamines. Nippon Yakurigaku Zasshi, Nov. 2003, 122, Suppl., 78P–80P.

Tsukada, H., Harada, N., Ohba, H., Nishiyama, S., Kakiuchi, T. Facilitation of dopaminergic neural

transmission doest not affect [(11)C]SCH23390 binding to the striatal D(1) dopamine receptors,

but the facilitation enhances phosphodiesterase type-IV activity through D(1) receptors: PET

studies in the conscious monkey brain. Synapse, Dec. 15, 2001, 42(4):258–265.

Tyler, J. L., Yamamoto, Y. L., Diksic, M., Theron, J., Villemure, J. G., Worthington, C., et al. Pharma-

cokinetics of superselective intra-arterial and intravenous [

C]BCNU evaluated by PET. J. Nucl.

Med., 1986, 27:775–780.

Wells, P., Aboagye, E., Gunn, R. N., Osman, S., Boddy, A. V., Taylor, G. A., Raﬁ , I., Hughes, A. N.,

Calvert, A. H., Price, P. M., Newell, D. R. 2-[

C]thymidine positron emission tomography as an

indicator of thymidylate synthase inhibition in patients treated with AG337. J. Natl. Cancer Inst.,

May 7, 2003, 95(9):675–682.

West, C. M., Price, P. Combretastatin A4 phosphate. Anticancer Drugs, Mar., 2004, 15(3):

179–187.

Wilms, H., Claasen, J., Rohl, C., Sievers, J., Deuschl, G., Lucius, R. Involvement of benzodiazepine

receptors in neuroinﬂ ammatory and neurodegenerative diseases: evidence from activated microg-

lial cell in vitro. Neurobiol. Dis., Dec. 2003, 14(3):417–424.

Zheng, Q. H., Liu, X., Fei, X., Wang, J. Q., Ohannesian, D. W., Erickson, L. C., Stone, K. L., Hutchins,

G. D. Synthesis and preliminary biological evaluation of radiolabeled O6-benzylguanine deriva-

tives, new potential PET imaging agents for the DNA repair protein O6-alkylguanine-DNA alkyl-

transferase in breast cancer. Nucl. Med. Biol., May 2003, 30(4):405–415.

REFERENCES 381

383

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

POLYMERIC SEQUESTRANTS AS

NONABSORBED HUMAN THERAPEUTICS

PRADEEP K. DHAL, CHAD C. HUVAL, AND S. RANDALL HOLMES-FARLEY

Genzyme Corporation

Waltham, Massachusetts

15.1 INTRODUCTION

The use of functional polymers in medicine has seen considerable growth during the past

two decades.

Polymers as biomaterials have found applications in such areas as artiﬁ cial

organs, tissue engineering, components of medical devices, and dentistry. A growing as-

pect of the ﬁ eld is the recognition of polymers as useful therapeutic agents: polymers that

either exhibit pharmacological properties themselves or that can be utilized as carriers for

selective and sustained delivery vehicles for small molecule or macromolecular (e.g., pro-

teins, genetic materials) pharmaceutical agents.

Over the past several years there has been a growing scientiﬁ c interest in the use poly-

mers as delivery agents for drug molecules. The goal of such research is to deliver drugs at

a sustained rate, deliver drugs targeted at speciﬁ c sites (to minimize toxicity and enhance

selectivity for, e.g., certain antitumor agents), and to deliver macromolecular prodrugs with

polymers acting as carrier molecules.

2–4

More recently, utility of polymers as nonviral

vectors for the delivery of genetic materials for gene therapy has also been evaluated.

Signiﬁ cant advancements in the area of polymeric drug delivery systems (including com-

mercial products) have taken place in recent years. Several research papers and review

articles pertaining to the use of biomedical polymers for drug delivery have been published

over the last 20 years.

6,7

Although polymers are used extensively as drug delivery agents, intrinsically bioactive

polymers (polymers as active pharmaceutical ingredients) are a relatively recent develop-

ment.

Due in part to their high molecular weight, polymers would appear to offer sev-

eral advantages over low-molecular-weight agents as potential therapeutic agents. These

384 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

beneﬁ ts may include lower toxicity, greater speciﬁ city of action, and enhanced activity

due to multiple interactions (polyvalency). Nevertheless, the concept of polymeric drugs

has not received widespread acceptance. Medicinal chemists have long considered syn-

thetic polymers as an uninteresting class of compounds to be evaluated as potential phar-

macophores. Some of the underlying concerns include the issue of undeﬁ ned molecular

weight, polydispersity, and compositional heterogeneity (copolymers in particular), since

such properties could complicate drug development process. The high-molecular-weight

characteristics of polymers would impede their systemic absorption through oral route as

well as complicate subsequent ADME (absorption, distribution, metabolism, and elimina-

tion) proﬁ le. However, as we describe in this chapter, these potential shortcomings associ-

ated with pharmacological characteristics of polymers can be carefully exploited to design

and develop therapeutic agents for disease conditions where low-molecular-weight drugs

have either failed or produced inadequate therapeutic beneﬁ ts. Although polymers do not

exhibit properties that ﬁ t most druglike deﬁ nitions and frequently violate Lipinski’s rules,

a reevaluation of the attributes of polymers highlights a number of beneﬁ ts of polymeric

drugs that are not achieved with traditional small molecule drugs.

The fact that high-molecular-weight polymers are not generally absorbed from the

gastrointestinal tract may be of particular advantage where it is desirable to prohibit a

drug molecule from systemic exposure. For example, by combining this characteristic

with the ability of polymers to selectively bind molecular and macromolecular species

in gastrointestinal ﬂ uids, it has become possible to develop a new class of therapeutic

agents that can selectively bind and remove detrimental species from the gastrointestinal

(GI) tract. As with any pharmaceutical, sequestration of each target molecule or pathogen

requires a unique strategy, and this strategy depends not only on the chemical nature of

the target, but on the location, concentration, and quantity of the target that needs to be

removed.

Since several books and review articles have been published covering the area of poly-

meric drug delivery systems,

2–7

this chapter is limited to case studies where polymers act

as active drugs. We focus on some of the most recent efforts in this regard and concentrate

on the development of polymeric drugs for the sequestration of low-molecular-weight spe-

cies such as bile acids, phosphate, and iron as well as polyvalent interactions to bind toxins,

viruses, and bacteria

15.2 POLYMERS AS SPECIFIC MOLECULAR SEQUESTRANTS

A number of potentially detrimental substances are present in the gastrointestinal (GI) tract.

These molecules and pathogens are implicated for a number of disease conditions. These

species can be either exogenic (i.e., they enter the body with food, drinks, and/or from

the environment) or they may be produced endogenously as a result of body metabolism.

Effective removal of these species in a selective manner offers a promising prophylactic

and therapeutic approach for the treatment of a number of disease states. Biocompatible

polymeric sequestrants represent an ideal class of agents for this purpose. Their nonabsorp-

tion through the intestinal wall should offer minimal toxicity, while the incorporation of

appropriate functional groups and manipulation of polymer physicochemical characteris-

tics provides opportunities to tailormake polymers with high selectivity and capacity. In

the following sections we describe speciﬁ c examples of polymeric drugs that have been

developed during the last few years.

15.3 SEQUESTRATION OF INORGANIC IONS IN THE GI TRACT

Electrolytes play a key role in regulating the distribution of different inorganic anions and cat-

ions between the intracellular and extracellular ﬂ uid compartments (homeostasis) and are vi-

tal in maintaining myocardial and neurological functions, ﬂ uid balance, oxygen delivery, and

acid–base balance. Electrolyte imbalance, caused by either excessive ingestion or impaired

elimination of an electrolyte from the body, has important physiological effects. Although

certain nonrenal tissues (e.g., muscle and liver) contribute to maintaining electrolyte balance,

the kidney plays the predominant role in maintaining electrolyte balance.

Therefore, renal

impairment is the most common reason for electrolyte imbalance and can lead to critical

health conditions. Conceptually, nonabsorbed polymeric sequestrants could be used to bind

these excessive ions (implicated in various pathologic conditions) selectively in the GI tract.

15.4 POLYMERIC POTASSIUM SEQUESTRANTS: A NONABSORBED

THERAPY FOR HYPERKALEMIA

Since the transmembrane potential is a major regulator of cell function, the ratio of po-

tassium between intra- and extracellular ﬂ uids is critically important to all living cells.

Hyperkalemia (elevated level of serum potassium, usually greater than 5.0 mEq/L) can

result from burn and crush muscle injuries, acidosis, or through the use of antihypertension

drugs (angiotensin-converting enzyme inhibitors). A rise in serum potassium can manifest

moderate to serious health problems, including paresthesias, areﬂ exia, respiratory failure,

and bradycardia.

Since the kidney is responsible for elimination of most excess potas-

sium, patients with impaired renal function are incapable of maintaining potassium homeo-

stasis. Traditional approaches to treat hyperkalemia include use of insulin, glucose, sodium

bicarbonate, and calcium chloride. However, these treatments have their own shortcom-

ings. For example, excess calcium could lead to hypercalcemia, which in turn leads to

myocardial infarction, kidney stones, and a variety of other conditions.

Use of an insoluble polyanionic resin to sequester excess potassium in the GI tract with

elimination in the feces could enable patients to excrete potassium despite impaired kid-

neys. There are a number of low-molecular-weight ligands that complex potassium ions.

Utilizing this principle, a cation-exchange resin based on sodium polystyrene sulfonate

(1) (Scheme 15.1) was developed to sequester potassium ion in the GI tract. This polymer,

385

Scheme 15.1

−

POLYMERIC POTASSIUM SEQUESTRANTS: A NONABSORBED THERAPY FOR HYPERKALEMIA

386 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

marketed under the brand name Kayexalate

and approved in the United States for the

treatment of hyperkalemia in 1975, is administered orally or rectally. A potential problem

with its use is induction of hypernatremia (elevated serum sodium), since the material ex-

changes 1.0 eQ potassium for 1.5 eQ sodium as well as incidences of intestinal necrosis.

These adverse effects resulting from Kayexalate treatment may provide an opportunity to

discover new generations of potassium sequestering polymers.

15.5 POLYMERIC DRUGS FOR CHRONIC RENAL FAILURE

Management and control of elevated-level serum phosphosphorus (hyperphosphatemia) is a

severe issue for patients suffering from chronic and end-stage renal failure.

16,17

The conse-

quences of inadequate control of serum phosphate level leads to various pathologies of clinical

signiﬁ cance. They include soft tissue calciﬁ cation (leading to cardiac calciﬁ cation and car-

diac-related complication), renal osteomalacia leading to reduced bone density, and second-

ary hyperparathyroidism. These conditions make hyperphosphataemia a major risk (including

mortality) among patients suffering from end stage renal disease, such as dialysis patients.

The major route for the excretion of phosphate from healthy human body is through the

kidney. Therefore, patients with an impaired renal system exhibit systemic accumulation of

phosphate due to an ingestion–excretion imbalance. Phosphate binder therapy has been the

mainstay for the treatment of hyperphosphatemia. The traditional phosphate binders have

been calcium- and aluminum-based agents that remove phosphate through the formation of

insoluble calcium or aluminum phosphate in the GI tract.

18,19

Unfortunately, since aluminum

and calcium salts have the propensity for systemic absorption through the linings of the GI

tract, they can lead to undesirable toxic and metabolic side effects (e.g., neurological disorders,

cardiac calciﬁ cations) in patients with impaired renal function. As a result, the side effects as-

sociated with these inorganic salt–based phosphate sequestrants limit their long-term use.

Design and development of nonabsorbed cationic polymers as sequestrants for phos-

phate ions offer an ideal approach to treat hyperphosphatemia in renal failure patients. The

known binding of phosphate by polycationic species has long been a well-studied phe-

nomenon. A number of studies have been carried out over the last two decades in design-

ing novel compounds such as macrocyclic oligomeric amines and oligomeric guanidinium

compounds (e.g., 2 and 3, Scheme 15.2) as receptors to study molecular recognition events

Scheme 15.2

NH O HN

HN HN

involving phosphate, pyrophosphate, and phosphonate anions as guests.

20,21

Electrostatic

interaction is the primary force for this complexation process. Hydrogen bonding is con-

sidered to contribute to further binding strength. Utilization of this principle of physical

organic chemistry has led to the discovery of a series of nonabsorbed polymeric amine

compounds and polymeric guanidinium compounds that show afﬁ nity toward dietary phos-

phate. Being nonabsorbed, these polymeric sequestrants would be conﬁ ned to the GI tract

and thus would act as effective therapeutic agents to treat hyperphosphatemia that are free

from the side effects arising from the use of calcium and related metal salts as phosphate

binders.

Hider and Rodriguez have discovered a series of cross-linked polymer resins containing

guanidine groups as speciﬁ c sequestrants for phosphate anions.

22,23

Binding of phosphate

ions to guanidinium groups of arginine residues of proteins involving two electrostatic

bonds and two stereochemically favorable hydrogen bonds is well known in biological

systems.

On the basis of biological principles, they developed the synthetic procedure

to prepare insoluble polymer resins containing guanidinium groups. The chemical syn-

thetic procedure to prepare these polymeric guanidinium salts is presented in Scheme 15.3.

In vitro phosphate binding studies involving these polymeric guanidinium salts shows that

the polymers bind phosphate more selectively in the presence of other biologically impor-

tant anions, such as salts of bile acids, chloride, bicarbonate, and so on.

In our laboratories a series of amine containing polymers were evaluated as nonab-

sorbed phosphate sequestrants. Using the knowledge of phosphate binding properties of

macrocyclic polyamines and related hosts (as molecular receptors for anions), a series of

functional polymers bearing pendant and backbone amino groups was evaluated as phos-

phate sequestrants. These amine-functionalized polymer hydrogels were prepared either by

cross-linking of amine polymers or by the cross-linking polymerization of amine-contain-

ing vinyl monomers.

25,26

Polymers bearing primary, secondary, and tertiary amine groups

as well as quaternary ammonium groups were evaluated for this purpose. The general

method for syntheses of polymeric amine gels by cross-linking of soluble polymers is

shown in Scheme 15.4. The structural repeat units of a selection of these polymers are il-

lustrated in Scheme 15.5.

Like their low-molecular-weight phosphate receptor counterparts, the binding strengths

and capacities of these polymeric phosphate sequestrants have been found to depend on

a number of parameters, including the pH of the medium. Since more ammonium groups

are present at lower pH, the polymers exhibit higher binding capacity at lower pH. Due to

higher concentrations of amine groups along polymer chains and the divalent (and possibly

CNCN

HCl

/Pd(0)

Scheme 15.3 Polymer-bound guanidinium salts obtained by chemical modiﬁ cation of

poly(acrylonitrile) resins.

POLYMERIC DRUGS FOR CHRONIC RENAL FAILURE 387

388 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

trivalent) nature of phosphate anions, polymeric amines probably also exhibit stronger af-

ﬁ nity toward phosphate anions than small molecule receptors, due to a polydentate chelate

effect. This phenomenon has not, however, been examined in detail. Furthermore, since

the degree of protonation of polymeric amines is dependent on polymer architectures

(due to charge–charge repulsion), incorporation of appropriate spacing between the amine

groups in the polymer chain and the backbone is key to the phosphate binding capacities

of these polymers. Finally, polymers containing primary amine groups were found to be

better phosphate sequestrants than polymers bearing secondary and tertiary amines, while

polymers containing quaternary amines were found to exhibit the least phosphate binding

property.

27,28

A systematic structure–activity relationship (SAR) study was carried out by using dif-

ferent polymeric amine, different cross-linking agents, and the extent of cross-linking

of these polymer gels. The study revealed that cross-linked polyallylamine gels are the

most potent phosphate-binding polymers and possess properties suitable for pharmaceu-

tical applications. This class of polymers exhibited maximum phosphate binding in the

pH range encountered in the ssmall intestine, whereas the polymer gel was found to be

nontoxic and is essentially nonabsorbed. From the systematic SAR study we identiﬁ ed

Scheme 15.4 Synthesis of polymeric hydrogels bearing amino groups by cross-linking of soluble

polymeric amines.

Scheme 15.5 Structural repeat units of representative polymers used as precursors for sequestrant

synthesis.

(CH

)

epichlorohydrin crosslinked polyallylamine (4, Scheme 15.6) as the lead candidate for sub-

sequent preclinical and clinical development. This polymer was found to exhibit maximum

capacity for phosphate. It is believed that the polymer binds phosphate anion through elec-

trostatic and possibly hydrogen-bonding interactions (see Scheme 15.7).

Thus, epichlorohydrin cross-linked polyallylamine hydrogel underwent clinical develop-

ment as the ﬁ rst metal-free phosphate sequestrant for the treatment of hyperphosphatemia.

The compound was approved in the United States by the Food and Drug Administration

in 1998 under the generic name sevelamer hydrochloride, and it has been marketed since

under the brand name Renagel by Genzyme Corporation. Subsequently, it was approved

in the Europe, Japan, and a number of other countries. Since its approval, Renagel has

demonstrated effective, long-term control of serum phosphate levels and has shown several

advantages over and above traditional agents for the management of hyperphosphatemia

in renal failure patients.

29,30

The ability for improved control of serum phosphate without

increasing the exposure to toxic metal ions such as aluminum and eliminating the intake of

additional calcium offers a number of clinical advantages. For example, without increasing

the calcium load or promoting calciﬁ cation, Renagel may help prevent cardiac complica-

tions in end-stage renal disease patients. Furthermore, Renagel has been found to reduce

serum parathyroid hormone and total and low-density lipoprotein (LDL) cholesterol in

hemodialysis patients. Since cardiovascular events are the most common causes of mortal-

ity of dialysis patients, Renagel thus offers a very promising treatment in the management

of renal failure.

It represents one of the ﬁ rst tailormade polymeric drugs that exhibits

its prophylactic and therapeutic properties through selective sequestration and removal of

unwanted dietary components in the GI tract without presenting any systemic side effects.

These documented beneﬁ ts of Renagel provide an opportunity to affect patient survival and

morbidity as well as to reduce health care expenses.

15.6 POLYMERIC IRON SEQUESTRANTS FOR THE TREATMENT OF IRON

OVERLOAD DISORDERS

Although iron is essential for the proper functioning of all living cells, it is toxic if present

in excess.

In the presence of molecular oxygen, excess iron can produce oxygen-derived

Scheme 15.6

POLYMERIC DRUGS FOR CHRONIC RENAL FAILURE 389

390 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

free radicals such as hydroxyl radical. These reactive free radicals interact with many

biological molecules, resulting in peroxidative tissue or organ damage. These adverse

events form the basis for a number of pathological conditions. Under normal conditions,

iron metabolism is highly conserved, with the majority of the iron being recycled within

the body. There is no normal iron loss mechanism; iron is not normally present in urine,

feces, or bile. Its removal from the body occurs only through bleeding or normal sloughing

of epithelial cells. Certain genetic disorders can, however, lead to increased absorption of

dietary iron (as in the case of hemochromatosis) or transfusion-induced iron over load (as

in the case of β-thalassaemia and sickle cell anemia).

33,34

Although excess iron can be removed by venesection (especially in hemochromatosis

patients), removal of iron using an iron chelator is the only effective way to relieve iron

overload in patients with β-thalasemia or sickle cell anemia.

The current standard of care

is desferrioxamine (5, Scheme 15.8), the only approved iron chelator for this condition in

the United States. This drug presents several shortcomings. It exhibits a narrow therapeutic

window and is orally inactive. As a result, it requires administration by parenteral infusions

Scheme 15.7 Binding interaction between polymer-bound ammonium groups and phosphate

anions.

OH + HCl

−

for 8 to 12 hours per day.

Thus, there is a need for the discovery and development of

orally active iron chelating agents for the treatment of iron overload.

Development of nonabsorbed polymeric ligands to sequester and remove dietary iron

selectively from the GI tract appears to be an attractive method for the treatment of iron over-

load. To be clinically useful, polymeric iron chelators need to possess several important fea-

tures. First, the polymeric ligand must possess high afﬁ nity, capacity, and selectivity toward

iron. Furthermore, the chelator should be biocompatible and should not be absorbed from

the GI tract. In general, design principles of polymeric chelators are based on the knowledge

of low-molecular-weight iron chelators. For clinical applications, the properties of chelators

in terms of metal ion selectivity and ligand–metal complex stability are important. Since iron

exists in two oxidation states [ferrous (⫹2) and ferric (⫹3)], chelators can be designed for

sequestering both forms of iron. Soft donor atoms (e.g., nitrogen-containing ligands such as

bipyridine and phenanthroline) can be employed to sequester Fe(II). Although these ligands

are selective for Fe(II), they also possess afﬁ nity for other biologically important divalent

metal ions, such as Zn(II) and Cu(II). On the other hand, oxyanions such as hydroxamates

and catecholates are selective toward Fe(III). These oxyanion ligands in general show higher

selectivity toward trivalent metal ions than toward divalent metal ions. Natural iron chela-

tors such as the siderophores desferrioxamine (5) and enterobactin (6, Scheme 15.9) contain

hydroxamate and catechol groups, respectively, and are selective toward Fe(III).

Scheme 15.8

N-OH

391

POLYMERIC IRON SEQUESTRANTS FOR THE TREATMENT OF IRON OVERLOAD DISORDERS

Scheme 15.9

392 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

By careful consideration of the foregoing facts, polymeric hydrogels containing hy-

droxamic acid and catechol moieties (7 and 8, Scheme 15.10) as well as cross-linked poly-

meric amines were prepared and were evaluated for their iron-binding properties.

Under

in vitro conditions, all of these polymers sequester iron at higher pH. At lower pH, the

polymers containing hydroxamic acids maintained their iron-binding properties, whereas

other polymers substantially lost their iron-binding properties. In vivo studies using rodents

have shown that treatment of the animals with the polyhydroxamic acid polymer lead to a

decrease in absorption of dietary iron.

However, no human clinical data is available on

the iron chelating polymers.

15.7 SEQUESTRATION OF BILE ACIDS: POLYMERS

AS CHOLESTEROL-LOWERING AGENTS

Increased plasma total cholesterol and low-density lipoprotein cholesterol (LDLc) are es-

tablished risk factors for antherosclerosis, which is the underlying cause of coronary heart

disease and most strokes.

The reduction of elevated LDLc is one of the most common

therapeutic treatments for cardiovascular disease. The majority of the people at risk require

only a modest (20 to 30%) reduction in LDLc level to minimize the risk of this serious

disease.

HMG-CoA reductase inhibitors (more commonly known as statins) are the most

widely used drugs for reducing blood levels of LDLc and have been shown to reduce the

risk of coronary events and strokes signiﬁ cantly. These ﬁ ndings have led to recent guide-

lines for expanding the use of cholesterol-lowering drug therapies to more patients.

In the

United States alone there are an estimated 36 million people who are in need of cholesterol-

lowering drug therapy.

Despite the spectacular success of statins in treating cardiovascular diseases, there is

still a need for new therapies to reduce blood LDLc. Statins are, for example, not indicated

for pregnant women, for pediatric use, or for patients with liver disease. Moreover, some

patients do not achieve the LDLc goal with statin therapy alone. There are also long-term

potential safety issues associated with statins, such as liver dysfunction and musculoskel-

etal symptoms. This is clearly evident from the recent withdrawal of a statin, cerevastatin

(Baycol), from the market.

The molecular mechanism underlying cholesterol metabolism was investigated system-

atically by Brown and Goldstein.

According to this metabolic pathway, cholesterol is syn-

thesized in the liver by the enzyme HMG-CoA reductase. Subsequently, it is transformed

Scheme 15.10

C-NHOH

into a bile acid in the liver and secreted to the gallbladder. The statins inhibit the function

of HMG-CoA reductase, which is the rate-limiting enzyme in cholesterol biosynthesis. The

presence of bile acid in the cholesterol metabolism pathway suggests that another approach

to reduce plasma LDLc is through effective removal of the bile acid from the bile pool,

resulting in upregulation of bile acid biosynthesis. Since the body attempts to maintain a

steady state of bile acid pool, the process of sequestration and removal of bile acid leads to

a corresponding drop in plasma cholesterol levels.

Bile acid sequestrants (BASs) are cross-linked polymeric cationic gels that bind anionic

bile acids in the GI tract and result in elimination of bile acid from the body.

The use

of these polymeric gels for sequestering bile acids is an established approach for treat-

ing elevated cholesterol.

Being systemically nonabsorbed, these polymeric cholesterol-

lowering drugs are free from the systemic side effects that are associated with statins.

Moreover, the BAS have over 30 years of clinical experience with a good safety record. Un-

til recently, two cationic polymers, cholestyramine (9) and colestipol (10, Scheme 15.11),

have been the only approved bile acid sequestrants on the market. Despite the appeal of

their safety proﬁ les, these two ﬁ rst-generation bile acid sequestrants have, despite their

high in vitro capacity, low clinical potency. This has led to reduced patient compliance and

hence limited use. For example, the doses required for a 20% cholesterol reduction with

cholestyramine and colestipol are typically 16 to 24 g/day.

This low clinical efﬁ cacy of

BAS has been ascribed to competition for bile acids with the active bile acid transporter

system of the GI tract.

It appears that for a cationic polymer to be a potent BAS, it must

have high binding capacity, strong binding strength, and selectivity toward bile acids in

the presence of competing desorbing forces of the GI tract. Thus, a potent BAS needs to

Scheme 15.11

NCH

SEQUESTRATION OF BILE ACIDS: POLYMERS AS CHOLESTEROL-LOWERING AGENTS 393

394 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

exhibit slow off-rates of bound bile acids from the polymer resin to effectively overcome

the active transport of bile acids from the GI tract.

The critical design rules for discover-

ing more potent BAS need to take into consideration the physicochemical features of bile

acids. A typical bile acid (11, Scheme 15.12) possesses an anionic group and a hydrophobic

core. Therefore, while electrostatic interaction is the primary force for bile acid binding

by cationic polymers, a second attractive force that needs to be considered is the hydro-

phobic interaction between the sequestrant and bile acid. Furthermore, favorable swelling

characteristics of these cationic hydrogels in physiological environments are required for

attaining high capacity. Thus, a balanced combination of hydrophilicity (high capacity)

and hydrophobicity (to slow down the rate of desorption), along with optimum density of

cationic groups, would constitute key features of the most potent BAS.

Consideration of these desired features has led to the discovery of a number of bile acid

sequestrants over the last decade from our laboratories as well as from other groups.

52–56

Chemical structures of some of these representative bile acid sequestrants are summarized

in Scheme 15.13. The common features of these hydrogels are the presence of amine and

ammonium groups. Additional structural features of these polymers include the presence

of hydrophobic chains. A variety of polymer backbones, including vinyl and allyl amine

polymers, (meth)acrylates, (meth)acrylamide, styrene, carbohydrates, polyethers, and other

condensation polymers, have been considered.

57–60

Although a large number of polymers

have been synthesized and tested in preliminary in vitro and in vivo for their bile acid se-

questration properties, very few polymers have proceeded to preclinical development and

clinical trials. Some of these promising BASs that entered the clinic include DMP-504 (12),

colestimide (13), SK&F 97426-A (14), and colesevelam hydrochloride (15, Scheme 15.14).

Scheme 15.13 Structural repeat units of representative ammonium salts as bile acid sequestrants.

–

+N(CH

)

Scheme 15.12

CONH

–

(CH

)

–

11 a:

= 1, X = COOH; b:

= 2, X = SO

0.98

0.02

+ Cl

Scheme 15.14

395

396 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

Structural features of these polymers comprise an optimum combination of charge density,

hydrophobic tails, and water swelling properties. From this new generation of bile acid se-

questrants that entered clinical trials, only two have been approved for marketing. Colestim-

ide has been approved for marketing in Japan and is sold under the trade name Cholebine;

colesevelam hydrochloride has been approved for marketing in the United States and is

being sold under the trade name WelChol.

61,62

Both polymers exhibit lower rates of side

effects and are better tolerated than previously marketed BASs. These BASs can be used as

monotherapy or in combination therapy by coadministration with statins.

15.8 SEQUESTRATION OF PATHOGENS: POLYMERIC

ANTI-INFECTIVE AGENTS

Polyvalent interactions are deﬁ ned as simultaneous binding interactions between multiple

ligands on one molecular entity and multiple receptors on another (cells, viruses, proteins,

etc.). These phenomena are prevalent in biological systems. Polyvalent interactions can

be collectively much stronger than the corresponding monovalent interactions.

These

interactions are responsible for the early stages of a large variety of important biological

processes, such as cell-surface or receptor–ligand recognition events. These processes can

provide the basis for mechanisms of both agonizing and antagonizing biological interac-

tions that are fundamentally different from those encountered in monovalent systems. A

schematic illustration of the principle of polyvalency is presented in Scheme 15.15. The

monovalent receptor

monovalent ligand monovalent complex

polyvalent receptor

polyvalent ligand

polyvalent complex

Scheme 15.15

underlying theory of polyvalency and its importance in biological processes have been

reviewed in detail.

The polyvalent interaction in biological systems has been considered as a new paradigm

for the design and discovery of a new generation of therapeutic agents.

Since polyva-

lency can enhance the binding strength signiﬁ cantly, the approach is particularly appealing

when the interaction between a monovalent ligand and a polyvalent receptor is weak.. For

example, a polyvalent agent carrying two or more linked ligands can bind at two or more

receptor sites on a target pathogen, resulting in enhanced binding strength. This would

lead to enhanced inhibition and/or sequestration of the target agent, such as a pathogen.

Polymeric systems provide an attractive platform to expand this approach for discovering

novel drug molecules wherein a collection of similar or different ligands can be linked

together covalently to a single polymer chain. The ﬂ exibility to choose different polymer

architectures (e.g., block, alternate, random, comb, graft, branched, dendritic), monomer

types, and the ease of polymer synthesis and modiﬁ cation enables one to synthesize a wide

variety of well-deﬁ ned polyvalent species with desired features such as controlled ligand

density, optimum binding strength and capacity, hydrophilicity and hydrophobicity, and the

incorporation of ancillary groups to enhance the recognition and binding of the target of in-

terest. Interplay of these features in polyvalent ligands can, in principle, lead to long-lasting

and more potent therapeutic agents. This new concept of ligand–substrate interaction has

led in recent years to the discovery of a variety of polymeric drugs that have been found to

sequester or inhibit toxins, viruses, and bacteria. Some of these polyvalent polymeric drugs

have advanced into human clinical trials, which provide further support to the validity of

this concept of drug discovery and development.

15.9 SEQUESTRATION OF TOXINS

Toxins are produced by pathogenic microorganisms in the host body. There are two kinds

of toxins: exotoxins and endotoxins.

67,68

Exotoxins are generally proteinous materials

released by pathogenic microorganisms; endotoxins are lipopolysaccharides and consist of

polysaccharide segments and glycolipid segments that constitute the outer cell membranes

of all gram-negative bacteria.

Upon secretion from microorganisms, these toxins can travel within the body of the

host organism and can cause damage in regions of the body farther away from the site of

infection. The pathogenic effects of these toxins—hemolysis, destruction of leucocytes,

paralysis, diarrhea, and septic shock—could be fatal.

Some of the pathogens that produce

life-threatening toxins in the GI tract are food poisoning organisms such as Staphylococ-

cus aureus, Clostridium perfringens, and Bacillus cereus and intestinal pathogens such as

Vibrio cholerae, Escherichia coli, and Salmonella enteritidis. Furthermore, anthrax toxin

is produced by Bacillus anthracis. In almost all cases, the causative agents responsible for

major symptoms of diseases associated with these pathogens are the exotoxins produced

by the organisms.

70,71

Clostridium Difﬁ cile Toxin Clostridium difﬁ cile is responsible for large numbers of

episodes of diarrhea that arise as a result of antibiotic treatment.

The outbreak of this

disease has been attributed to the disruption of normal colonic ﬂ ora by antibiotic treatment.

As a result, C. difﬁ cile invades and colonizes in the gut. Therefore, C. difﬁ cile infection is

prevalent in hospital settings. This bacterium releases two high-molecular-weight proteins

SEQUESTRATION OF TOXINS 397

398 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

(toxins A and B), which are implicated primarily in causing diarrhea.

73,74

The traditional

way to treat C. difﬁ cile–associated diarrhea has been the use of one of two antibiotics:

metronidazole or vancomycin. Although the use of antibiotics is effective in eliminating

C. difﬁ cile infection initially, repeated treatment with antibiotics increases the chance of

further infection by C. difﬁ cile, due to sterilization of the gut. Furthermore, there has been

increasing body of documentation pointing to growing resistance of C. difﬁ cile to antibiotic

treatment.

Since the disease symptoms are attributed to the toxins produced by this bacterium,

selective sequestration, neutralization, and elimination of these toxins by a polymeric se-

questering agent appears to offer an attractive and safe alternative to antibiotic therapy.

This method of treatment would treat the infection without disrupting the reestablishment

of normal bacterial growth in the gut.

Since C. difﬁ cile toxins are proteins, they possess multiple binding sites that are de-

rived from amino acid side chains. These side chain functional groups of the toxin can

effectively interact with multifunctional polymers carrying complementary binding sites

through simultaneous polyvalent interactions. Some of the early studies using polymeric

ligands to sequester C. difﬁ cile toxins include the use of anion-exchange resins such as

cholestryramine and colestipol (the bile acid sequestrants described above). However, the

effectiveness of these polymeric ion exchangers as sequestrants to bind and remove C. dif-

ﬁ cile toxins was found to be quite modest.

Systematic investigation in our laboratories has led to the discovery of some novel poly-

meric multivalent ligands as sequestrants that effectively bind and neutralize C. difﬁ cile–

derived toxins. Among the various polymers evaluated, a series of high-molecular-weight

water-soluble anionic polymers (16–18, Scheme 15.16) were found to be particularly ef-

fective in sequestering and neutralizing C. difﬁ cile toxins.

77,78

In general, these polymers

contain sulfonic acid groups and have high molecular weight. Careful in vitro studies using

pulsed ultraﬁ ltration binding experiments and ﬂ uorescence polarization spectroscopy re-

vealed that the binding constants for complexation between one of the polymers and toxins

A and B were 133 nM and 8.7 µM, respectively.

The ability to bind both the toxins is not

a general feature of these anionic polymers. For example, while sodium salt poly(styrene

sulfonic acid) effectively binds the polymers, poly(sodium 2-acrlamido-2-methyl-1-pro-

panesulfonate), a high-molecular-weight polyanion of similar charge density does not

bind either of the toxins to any measurable extent. Furthermore, the binding strength is

dependent on molecular weight, with lower-molecular-weight polymers exhibiting very

low binding strength. These results thus suggest that sequestration of C. difﬁ cile toxins

by polyanions is not purely electrostatic in origin. On the other hand, it appears that the

Scheme 15.16

SOO

-O

sodium salt of poly(styrene sulfonic acid) and related anionic polymers interact with toxins

A and B through multiple weak interactions that ampliﬁ es to a very high binding strength

as a result of polyvalent interaction. From ﬂ uorescence polarization data, it was estimated

that 1 molecule of toxin A interacts with about 800 monomer units on the polymer. Thus,

a single polymer chain of molecular mass 300 kDa would wrap around a toxin molecule

about three or four times. This suggests effective large polyvalent interactions between the

polymer and protein surface.

The in vitro binding activity of these high-molecular-weight

polyanions correlated well with the in vivo biological activity. The polymer GT prevented

the mortality of 80% of hamsters with severe C. difﬁ cile colitis.

These polymers are non-

antimicrobial and hence do not interfere with the activities of standard antibiotics. There-

fore, they are likely to overcome the problems associated with antibiotic resistance.

High-molecular-weight sodium salt of a poly(styrene sulfonic acid) derivative was se-

lected as the lead candidate for clinical development as a non-antibiotic-based polymer

therapy for treating C. difﬁ cile infection. This compound, Tolevamer, has been successful

in human clinical trials in both phases I and II. This polymer is currently undergoing phase

III human clinical trials.

In addition to the polyanions cited above, another class of polymers bearing pendant

oligosaccharide groups has been investigated as possible sequestrants for C. difﬁ cile tox-

ins. Some of these compounds have progressed to human clinical trial.

81,82

The underlying

principle behind these polymeric sugar agents for the treatment of C. difﬁ cile infection is

that toxin A has shown lectinlike activity, which allows it to bind to an oligosaccharide re-

ceptor on epithelial cells. Furthermore, toxin B has been found to bind erythrocytes. These

observations suggest that the cell invasion and binding of C. difﬁ cile may be mediated by

cell-surface carbohydrate receptors. Therefore, polymers bearing pendant sugar residue

may compete with human cells toward C. difﬁ cile toxin, After identifying oligosaccharide

sequences, appropriate oligosaccharide molecules that are speciﬁ c for both toxins were

conjugated to different polymer backbone (19, Scheme 15.17). Lengths of tethering arms

linking the polymer backbones with oligosaccharide moieties were optimized appropri-

ately to maximize the binding strength of ligand–polymer (polymer–toxin) interaction.

The oligosaccharide sequences that were found to improve toxin binding include maltose,

cellobiose, isomaltotriose, and chitobiose. Polymeric carriers examined include substituted

polystyrene and other oleﬁ n backbones. These polymeric toxin binders (under the generic

POLYMERIC ANTIMICROBIAL AGENTS 399

Scheme 15.17

400 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

name of Synsorb) were found to be effective in neutralizing toxins and in controlling diar-

rhea in animal models. One of these polyvalent polymeric carbohydrate agents had also

progressed to human clinical trials. However, due to lack of therapeutic efﬁ cacy, it was

withdrawn after a phase II clinical trial.

15.10 POLYMERIC ANTIMICROBIAL AGENTS

The emergence of microbial pathogens that are resistant to multiple classes of available

antimicrobial agents is becoming a global public health concern of signiﬁ cant magnitude.

These multidrug resistant bacteria are prevalent in both hospital and community environ-

ments.

The majority of these strains have been found to carry multiple drug resistance

factors. The only effective treatment for multiply resistant bacterial infections that is cur-

rently available is vancomycin. However, resistance to vancomycin is also a growing con-

cern. For example, methicillin-resistant Staphylococcus aureus (MRSA) strains, vancomy-

cin-resistant enterococci, and amikacin and β-lactam-resistant Kleisiella pneumoniae are

some of the bacterial species that are resistant to vancomycin.

84,85

This rapid emergence of

multidrug-resistant bacterial strains poses a potential threat to human life. As a result, there

is an urgent need to discover and develop novel antibacterial agents that can circumvent the

challenges posed by multidrug-resistant microorganisms.

Polyvalent ligands as antibacterial agents have been thought to exhibit potential advan-

tages over monomeric antimicrobial agents. Cluster effects from polyvalent ligands would

lead to ampliﬁ cation of weak nonbonding interaction between the bacterial surface recep-

tors and the polymeric ligands. Aggregation and precipitation of bacteria by polyvalent

ligands is potentially another favorable feature that could be achieved by using polyvalent

antimicrobial agents. Finally, polyvalent ligands could enhance the rate of lysis of the bac-

terial cell membrane and wall more rapidly, due to multipoint attachment.

Most bacterial infections are initiated by adhesion of microorganisms to the muco-

sal surfaces of hosts, mediated in part by bacterial protein adhesins.

These adhesins

interact with carbohydrate determinants of host cell glycolipids or glycoproteins. The

mechanism for bacterial infection through this pathway suggests that development of ap-

propriate polyvalent sugar derivatives could competitively block the attachment of micro-

bial adhesin to host mucosal surface, resulting in protection against infection. This concept

has been explored through the synthesis of a number of polymers bearing acid-functional-

ized glycoside moieties. Oleﬁ nic monomers containing glycoside moieties and acid func-

tional groups such as O-sulfo and O-carboxymethyl groups were prepared and converted to

various copolymers. These polymers were found to be effective in vitro against a number

of bacterial targets.

A second approach to design polyvalent ligands as antimicrobial agents based on cat-

ionic polymers has also been explored systematically in our laboratories. The mechanism

of the antimicrobial action of these polymers has been attributed to their ability to enhance

the rate of cell lysis. These polymers were designed as mimetics of certain cationic amphi-

philic peptides containing multiple arginine and lysine residues. These cationic peptides,

known as antimicrobial peptides, cause cell lysis through interaction of the positive charges

of the peptides with negative phosphate head groups of cell membrane phospholipids

A series of amphiphilic cationic polymers were prepared bearing amine and quaternary

ammonium groups as well as hydrophobic tails (see Scheme 15.18). These polymers were

found to exhibit antimicrobial activity against a number of microbes.

Until the advent of

protease inhibitor, C. parvum was a primary target for drug discovery to treat GI tract infec-

tions in persons with HIV infection.

Some of these polymers were found to be superior

to the commonly prescribed antibiotic, such as paromomycin.

15.11 CONCLUSIONS AND OUTLOOK

Although polymeric materials have been developed to produce useful biomaterials and

drug delivery systems, until recently intrinsically bioactive polymers as therapeutic agents

have remained largely unappreciated. In the present chapter we have attempted to illustrate

the potential of polymers in the discovery and development of novel therapeutic agents for

treating a number of human diseases. By careful consideration of both disease targets and

their mechanisms of action, a number of functional polymers have been discovered that

exhibit promising pharmacological properties. These polymeric drugs capitalize on the

unique physicochemical properties of polymer, and on many occasions, these polymers

exhibit therapeutic properties that cannot be achieved by traditional small molecule drugs.

The usefulness of these bioactive polymers for the treatment of different diseases that have

been marketed or are being developed is quite impressive. Polymeric sequestrants that

are conﬁ ned to the GI tract and carry out their disease-modifying activities are important

examples that attest to the validity of the concept of polymeric drugs. Recognition of the

phenomenon of polyvalency as a tool in drug discovery strategy has been carefully utilized

to discover polymeric drugs that provide a new paradigm for developing novel therapeu-

tics. Although a number of these polymers have not yet met the goal of exhibiting in vivo

biological activity, the development of toxin sequestrants suggests that it is not simply

an academic curiosity. Finally, further understanding of cellular and molecular basis of

various disease targets with the recent advancements of genomics and proteomics research

would further enable pharmaceutical researchers in designing novel polymeric drugs with

desired immunological and other related pharmacological properties for systemic applica-

tions of polymers against other disease targets. Once these design criteria are identiﬁ ed and

ﬁ ne tuned, potentially more selective polymer therapeutics will be discovered for medical

needs unmet or met inadequately.

Scheme 15.18

CONHR

–

CONHR'

CONHR''

–

CONCLUSIONS AND OUTLOOK 401

402 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

REFERENCES

R. Duncan, Nature Rev. Drug Discov., 2, 347 (2003).

R. M. Ottenbrite, in R. L. Dunn and R. M. Ottenbrite (Eds.), Polymeric Drugs and Drug Delivery

Systems, ACS Symposium Series 469, American Chemical Society, Washington, DC, 1991, 3 pp.

K. Ulbrich, M. Pechar, J. Strohalm, and V. Subr, Macromol. Symp., 118, 577 (1997).

R. Duncan and J. Kopecek, Adv. Polym. Sci., 57, 51 (1984).

M. C. Garnett, Crit. Rev. Ther. Drug Carrier Syst., 16, 147 (1999).

R. Langer, Nature, 392, 5 (1998).

M. Ahlers, W. Muller, A. Reichert, H. Ringsdorf, and J. Venzmer, Angew. Chem. Int. Ed. Engl.,

29, 1269 (1990).

R. Duncan, in K. Park and R. J. Mrsny (Eds.), Controlled Drug Delivery, ACS Symposium

Series 752, American Chemical Society, Washington, DC 2000, 350 pp.

S. R. Holmes-Farley, Polym. Mater. Sci. Eng., 80, 246 (1999).

L. R. Johnson, Physiology of the Gastrointestinal Tract, Raven Press, New York, 1994.

R. M. Rosa, M. E. Williams, and F. H. Epstein, in D. W. Seldin and G. Giebisch (Eds.), The Kid-

ney: Physiology and Pathophysiology, Raven Press, New York, 1992, 2165 pp.

G. O. Perez, J. R. Oster, R. Pelleya, P. V. Catalis, and D. C. Kem, Nephron, 36, 270 (1984).

T. J. Marrone and K. M. Terz, J. Am. Chem. Soc., 114, 7542 (1992).

G. W. Gardiner, Can. J. Gastroenterol., 11, 573 (1997).

B. B. Gerstman, R. Kirkman, and R. Platt, Am. J. Kidney Dis., 20, 159 (1992).

J. Delmez and E. Slatopolsky, Am. J. Kidney Dis., 4, 303 (1992).

J. W. Coburn and E. Slatopolsky, in B. M. Brenner and F. C. Rector, Jr. (Eds.), The Kidney, W.B.

Saunders, Philadelphia, 1991, 2036 pp.

C. Hsu, Am. J. Kidney Dis., 29, 641 (1997).

S. Ribiro, A. Ramos, and A. Brandao, Nephrol. Dial. Transplant, 13, 2037 (1998).

F. P. Schmidtchen and M. Berger, Chem. Rev., 97, 1609 (1997).

M. W. Hosseim, and J. M. Lehn, Helv. Chim. Acta, 70, 1312 (1987).

R. C. Hider and A. Canas-Rodriguez (to British Technology Group Ltd.), U.S. patent 5,698,190

(December 1997).

R. C. Hider and A. Canas-Rodriguez (to British Technology Group Ltd.), U.S. patent 1,132,706

(March 2000).

B. Dietrich, D. L. Fyles, T. M. Fyles, and J. M. Lehn, Helv. Chim. Acta., 62, 2763 (1979).

S. R. Holmes-Farley, W. H. Mandeville, and G. M. Whitesides (to Genzyme Corporation), U.S.

patent 5,496,545 (January 1996).

S. R. Holmes-Farley, W. H. Mandeville, and G. M. Whitesides (to Genzyme Corporation), U.S.

patent 5,667,775 (September 1997).

S. R. Holmes-Farley and W. H. Mandeville (to Genzyme Corporation), U.S. patent 6,726,905

(April 2004).

S. R. Holmes-Farley, W. H. Mandeville, J. Ward, and K. L. Miller, J. Macromol. Sci. Pure. Appl.

Chem., A36, 1085 (1999).

E. A. Slatopolosky, S. K. Burke, and M. A. Dillion, Kidney Int., 55, 299 (1999).

N. Amin, Nephrol. Dial. Transplant, 17, 340 (2002).

A. J. Bleyer, S. K. Burke, and M. A. Dillon, Am. J. Kidney Dis., 33, 694 (1999).

R. R. Crichton, in Inorganic Biochemistry of Iron Metabolism, Ellis Horwood, New York, 1991.

G. M. Brittenham, in R. Hoffman, E. Benz, S. Shattil, B. Furie, and H. Cohen (Eds.), Hematology:

Basic Principles and Practice, Churchill Livingstone, New York, 1991, 327 pp.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

D. J. Weatherall and J. B. Clegg, in The Thalassaemia Syndromes, 3rd ed., Blackwell Scientiﬁ c,

Oxford, 1981.

Z. D. Liu and R. C. Hider, Med. Res. Rev., 22, 26 (2002).

J. B. Porter, M. C. Jawson, E. R. Huehn, C. A. East, and J. W. P. Hazell, Br. J. Haematol., 73,

403 (1989).

K. N. Raymond, G. Muller, and B. F. Matzanke, Top. Curr. Chem., 58, 49 (1984).

W. H. Mandeville and S. R. Holmes-Farley (to GelTex Pharmaceuticals, Inc.), U.S. patent

5,487,888 (January 1996).

P. K. Dhal, P. Cannon, S. R. Holmes-Farley, W. H. Mandeville, S. R. Polomoscanik, and T. X.

Neenan, Biomacromolecules, 6, 2946 (2005).

S. M. Grundy, Circulation, 97, 1436 (1998).

S. M. Grundy, Endocrinol. Metab. Clin. North Am., 27, 655 (1998).

Executive summary of the third report of the National Cholesterol Education (NCEP) Expert

Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults, J. Am. Med.

Assoc., 285, 2486 (2001).

D. J. Rader, Nature Med., 7, 1282 (2001).

M. S. Brown and J. L. Goldstein, Proc. Natl. Acad. Sci. USA, 96, 11041 (1999).

S. M. Grundy, in R. Fears (Ed.), Pharmacological Control of Hyperlipidaemia, Prous Science

Publishers, Barcelona, Spain, 1986, 3 pp.

E. R. Stedronsky, Biochim. Biophys. Acta, 1210, 255 (1994).

W. H. Mandeville and D. I. Goldberg, Curr. Pharm. Design, 3, 15 (1997).

G. M. Benson, C. Haynes, S. Blanchard, and D. Ellis, J. Pharm. Sci., 82, 80 (1993).

M. Ast And W. H. Frishman, J. Clin. Pharmacol., 30, 99 (1990).

W. H. Mandeville, W. Braunlin, P. Dhal, A. Guo, C. Huval, K. Miller, J. Petersen, S. Polo-

moscanik, D. Rosenbaum, R. Sacchiero, J. Ward, and S. R. Holmes-Farley, Mater. Res. Soc.

Symp. Proc., 550, 3 (1999).

P. Zarras and O. Vogl, Prog. Polym. Sci., 24, 485 (1999).

S. R. Holmes-Farley and W. H. Mandeville (to GelTex Pharmaceuticals, Inc.), U.S. patent

5,607,669 (March 1997).

C. C. Huval, S. R. Holmes-Farley, J. S. Petersen, and P. K. Dhal (to Genzyme Corporation), U.S.

patent 6,248,318 (1997).

S. R. Holmes-Farley, P. K. Dhal, and J. S. Petersen (to Genzyme Corporation), U.S. patent

6,203,785 (1998).

S. R. Holmes-Farley, W. H. Mandeville, S. K. Burke, and D.I. Goldberg (to Genzyme Corpora-

tion), U.S. patent 6,423,754 (2002).

G. M. Benson, D. R. Alston, D. M. B. Hickey, A. A. Jaxa-Chamiec, C. M. Whittaker, C. Haynes,

A. Glen, S. Blanchard, S. R. Cresswell, and K. E. Suckling, J. Pharm. Sci., 86, 76 (1997).

C. C. Huval, M. J. Bailey, S. R. Holmes-Farley, W. H. Mandeville, K. Miller-Gilmore, R. J.

Sacchiero, and P. K. Dhal, J. Macromol. Sci. Pure Appl. Chem., A38, 1559 (2001).

C. C. Huval, S. R. Holmes-Farley, W. H. Mandeville, R. J. Sacchiero, and P. K. Dhal, Eur. Polym.

J., 40, 693 (2004).

L. Zhang, V. Janout, J. L. Renner, M. Uragami, and S. L. Regen, Bioconjug. Chem., 11, 397 (2000).

N. S. Cameron, A. Eisenberg, and G. R. Brown, Biomacromolecules, 3, 116 (2002).

B. Gaudilliere and P. Berna, Annu. Rep. Med. Chem., 35, 331 (2000).

B. Gaudilliere, P. Bernardellli, and P. Berna, Annu. Rep. Med. Chem., 36, 293 (2001).

W. H. Mandeville and C. Arbeeny, Idrugs, 2, 237 (1999).

S.-K. Choi, Synthetic Multivalent Molecules: Concepts and Biomedical Applications, Wiley-

Interscience, New York, 2004.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

REFERENCES 403

404 POLYMERIC SEQUESTRANTS AS NONABSORBED HUMAN THERAPEUTICS

M. Mammen, S.-K. Choi, and G. M. Whitesides, Angew. Chem. Int. Ed. Engl., 37, 2754 (1998).

S. Borman, Chem. Eng. News, 78, 48 (2000).

B. Henderson, M. Wilson, and B. Wren, Trends Microbiol., 5, 454 (1997).

M. Rauchhaus, A. J. Coats, and S. D. Anker, Lancet, 356, 930 (2000).

B. Beutler and A. Poltorak, Crit. Care Med., 29, 52 (2001).

Y. Gasche, D. Pittet, and P. Sutter, in W. J. Sibbald and J. L. Vincent (Eds.), Clinical Trials for

Treatment of Sepsis, Springer-Verleg, Berlin, 1995, 35 pp.

T. C. Dixon, M. Meselson, J. Guillemin, and P. C. Hanna, N. Engl. J. Med., 341, 815 (1999).

J. G. Bartlett, Clin. Infec. Dis., 18, (Suppl. 4), S265 (1994).

C. P. Kelly and T. LaMont, Annu. Rev. Med., 19, 375 (1998).

R. C. Spencer, J. Antimicrob. Chemother., 41 (Suppl. C), 5 (1998).

S. Lebel, S. Bouttier, and T. Lambert, FEMS Microbiol. Lett., 238, 93 (2004).

N. S. Taylor and J. G. Bartlett, J. Infect. Dis., 141, 92 (1980).

C. Bacon-Kurtz and R. Fitzpartick (to GelTex Pharmaceuticals, Inc.), U.S. patent 6,290,946

(September 2001).

C. Bacon-Kurtz and R. Fitzpartick (to GelTex Pharmaceuticals, Inc.), U.S. patent 6,270,755

(August 2001).

W. Braunlin, Q. Xu, R. Fitzpatrick, C. B. Kurtz, R. Burrier, and J. D. Klinger, Biophysics, 87,

534 (2004).

C. B. Kurtz et al., Antimicrob. Agents Chemother., 45, 2340 (2001).

L. D. Heerze and G. D. Armstrong (to Synsorb Biotech, Inc.), U.S. patent 6,013,635 (January

2000).

U. J. Nilsson, L. D. Heerze, Y.-C. Liu, G. D. Armstrong, M. M. Palcic, O. Hindsgaul, Bionconj.

Chem., 8, 466 (1997).

H. Pearson, Nature, 418, 469 (2002).

C. Walsh, Nature, 406, 775 (2000).

L. P. Kotra, D. Golemi, S. Vakulenko, and S. Mosashery, Chem. Ind., 10, 341 (2000).

G. Reid and J. D. Sobel, Rev. Infect. Dis., 9, 470 (1987).

W. H. Mandeville and V. R. Garigapati (to Genzyme Corporation), U.S. patent 5,821,312

(October 1998).

W. L. Maloy and U. P. Kari, Biopolymers, 37, 105 (1995).

W. H. Mandeville, T. X. Neenan, and S. R. Holmes-Farley (to Genzyme Corporation), U.S. pat-

ent 6,034,129 (March 1998).

J. M. Matukaitis, J. Community Health Nurs., 14, 135 (1997).

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

405

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

BOTANICAL IMMUNOMODULATORS

AND CHEMOPROTECTANTS

IN CANCER THERAPY

BHUSHAN PATWARDHAN AND MANISH GAUTAM

University of Pune

Pune, India

SHAM DIWANAY

Abasaheb Garware College

Pune, India

16.1 INTRODUCTION

Botanicals being chemically complex and diverse may provide combinations of syner-

gistic moieties useful in cancer therapy. With suitable examples, in this chapter we high-

light the importance of traditional medicine in natural product drug discovery related to

cancer. Most cancer chemotherapeutic agents are associated with toxicity toward normal

cells and tissues that share many characteristics with tumor cells, particularly high cell

turnover. Optimal dosing of cancer chemotherapeutic agents is often limited because of

severe nonmyelosuppressive and myelosuppressive toxicities. Therefore, it is a continuing

challenge to design therapy that is effective and also efﬁ ciently targeted to tumor cells.

Cytoprotective agents are expected to control or prevent these toxicities and include use of

synthetic and natural products. Speciﬁ c chemoprotectants are emerging for cisplatin and

anthracyclin antibiotics. None of the available agents satisfy criteria for an ideal chemopro-

tection. This has stimulated research for discovering natural resources with immunomodu-

latory and cytoprotective activities. Various botanicals and ethnopharmacological agents

used in traditional medicine have been investigated by various workers for their chemopro-

tective, immunomodulating, adaptogenic, and antitumor activities and have revealed prom-

ises toward developing into a potential drug for cancer treatment per se or as adjuvant.

406 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

Cancer chemotherapy is a major treatment modality used for the control of advanced

stages of malignancies and also as a prophylactic against possible metastases in combi-

nation with the radiotherapy.

The chemotherapeutic agents available today have mainly

immunosuppressent activity. Most of them are cytotoxic and exert a variety of side effects.

The metabolism and clinical safety of these agents has not been clearly established. This

has given rise to stimulation in research for locating natural resources showing immu-

nological activity. In past, the variety of naturally occurring agents, including living and

attenuated microorganisms, autologous and heterologous proteins, and injections of animal

organ preparations, were used to restore and repair defense mechanism. A number of plant

extracts have been shown to be immunomodulators.

Chemical agents preventing site-

speciﬁ c toxicity of cytotoxic drugs are in current practice in cancer chemotherapy. These

agents exhibiting chemoprotective and immunomodulatory activities are reviewed here.

16.2 IMMUNOMODULATION

The control of disease by immunologic means has two objectives: the development of

immunity and the avoidance of undesired immune reactions. Immunostimulation in a

drug-induced immunosuppression model and immunosuppression in an experimental

hyperreactivity model by the same preparation can be said to be true immunomodulation.

Immunomodulators are biological response modiﬁ ers (BRMs), used to treat cancer, which

exert their antitumor effects by improving host defense mechanisms against the tumor.

They have a direct antiproliferative effect on tumor cells and also enhance the ability of the

host to tolerate damage by toxic chemicals that may be used to destroy the cancer. Modu-

lation of immunity was previously attempted with glucocorticoids and cytotoxic drugs

such as cyclophosphamide. It is now recognized that immunomodulatory therapy could

provide an alternative to conventional chemotherapy for a variety of diseased conditions,

especially when the host’s defense mechanisms have to be activated under the conditions

of impaired immune responsiveness or when a selective immunosuppression has to be in-

duced in situations such as inﬂ ammatory diseases, autoimmune disorders, and organ/bone

marrow transplantation.

All three classes of immunomodulators—biologicals, chemical,

and cytokines—will continue to play a major role in advancing and improving the qual-

ity of treatment of several human as well as animal diseases.

There is need for further

research to better understand the biochemical mechanisms involved in immunoregulation

to maximize the beneﬁ ts of chemical immunomodulators as single agents or adjuvants in

cancer therapy.

16.3 ETHNOPHARMACOLOGY AND BOTANICAL IMMUNOMODULATORS

There are two major ways of bioprospecting natural products for investigation. The ﬁ rst

is the classical method, which relies on phytochemical factors, serendipity, and random

screening approaches. The second method uses traditional knowledge and practices as the

drug discovery engine. Known as the ethnopharmacology approach, this method is time-

and cost-effective and may lead to better success than routine random screening. Various

ethnopharmacological agents are under investigation as immunomodulators. Traditional

Chinese medicine, Japanese Kampo, Indian Ayurveda, and such are becoming important

bioprospecting tools. Ayurveda gives a separate class of immunomodulating botanicals

named Rasayanas. Ayurveda, one of the most ancient and yet living traditions practiced

widely in India, Sri Lanka, and other countries, has a sound philosophical and experien-

tial basis. India has about 45,000 plant species; medicinal properties have been assigned

to many to several thousands. Ayurveda has detailed descriptions of over 700 herbs and

400,000-registered Ayurvedic practitioners routinely prescribe them, particularly for treat-

ment of chronic disease conditions. A considerable research on pharmacognosy, chemistry,

pharmacology, and clinical therapeutics has been carried out and the Ayurvedic database

has detailed descriptions of over 700 medicinal plants.

6,7

Rasayanas are nontoxic herbal

preparations or individual herbs used to rejuvenate or attain the complete potential of a

healthy or diseased person in order to prevent diseases and degenerative changes that lead

to disease. Pharmacodynamic studies on Rasayana botanicals have suggested many possi-

ble mechanisms, such as nonspeciﬁ c and speciﬁ c immunostimulation, free-radical quench-

ing, cellular detoxiﬁ cation, cell proliferation, and cell repair. Ayurveda (with particular

reference to botanicals) may play an important role in modern health care, particularly

where satisfactory treatment is not available. There is a need to evaluate the potential of

Ayurvedic remedies as adjuvant to counteract side effects of modern therapy and compare

the cost-effectiveness of certain therapies vis-à-vis modern therapeutic schedules.

16.4 ADAPTOGENS OR ADJUSTIVE MEDICINE

Most of the synthetic chemotherapeutic agents available today are immunosuppressants,

are cytotoxic, and exert a variety of side effects. N. V. Lazarev, who developed the concept

of a state of nonspeciﬁ cally increased resistance of an organism (SNIR), laid down the

theoretical basis for separation of a new group of medicinal substances. The medicinal

substances causing SNIR were named adaptogens.

Generally, adaptogens are those drugs

that enable one to withstand the stress and strain of life, impart immunity to give protection

against diseases, postpone aging, and improve vigor, vitality, and longevity. The concept is

also referred to as adjustive medicine. The concept of adjustive remedies has been difﬁ cult

to prove experimentally. A bifunctional information exchange network between the ner-

vous and immune systems is established by speciﬁ c receptors for humoral substances on

cells of nervous and immune systems. In particular, neuroregulators (neurotransmitters and

neuromodulators) can modulate speciﬁ c immune system function(s), and immunoregula-

tors (immunomodulators) can modulate speciﬁ c nervous system function(s). Acute and

chronic inﬂ ammatory processes, malignancy, and immunological reactions stimulate the

synthesis and release of immunomodulators in various cell systems. These immunomodu-

lators have pivotal roles in the coordination of the host defense mechanisms and repair and

induce a series of endocrine, metabolic, and neurologic responses.

However, with recent

insight into the neuroendocrine immune system regulation, such adjustive effects on the

homeostatic system of the body seem very likely.

16.4.1 Botanicals with Adaptogenic Activity

Mistletoe Lectin Deﬁ ned nontoxic doses of the galactoside-speciﬁ c mistletoe lectin

(mistletoe lectin-I, a constituent of clinically approved plant extract) have immunomodu-

latory potencies. The obvious ability of certain lectins to activate nonspeciﬁ c mechanisms

supports the assumption that lectin–carbohydrate interactions may induce clinically

beneﬁ cial immunomodulation. Randomized multicenter trials are being performed to

ADAPTOGENS OR ADJUSTIVE MEDICINE 407

408 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

evaluate the ability of complementary mistletoe lectin-I treatment to reduce the rate of

tumor recurrences and metastases, to improve overall survival and the quality of life, and

to exert immunoprotection in cancer patients under tumor-destructive therapy.

Panax Ginseng Panax ginseng (Araliaceae), a Korean and Chinese medicine employed

for its putative medicinal properties in South Asia, stimulated basal natural killer (NK) cell

activity following subchronic exposure and helped stimulate the recovery of NK function

in cyclophosphamide-immunosuppressed mice but did not further stimulate NK activity

in poly I:C-treated mice; T- and B-cell responses were not affected. P. ginseng provided a

degree of protection against infection with Listeria monocytogenes but did not inhibit the

growth of transplanted syngenetic tumor cells. Increased resistance to L. monocytogenes

was not detected in challenged mice previously given immunosuppressive doses of

cyclophosphamide. These data suggest that P. ginseng have some immunomodulatory

properties, associated primarily with NK cell activity.

Ginseng alone, or in combination

with vitamins and minerals, is promoted primarily as general tonics which increase

nonspeciﬁ c resistance and sometimes even as an aphrodisiac. On prolonged use, ginseng

shows a few adverse effects, most notable of which is the Ginseng abuse syndrome.

Achyranthes Bidentata Achyranthes bidentata polysaccharide (ABP) root extract (25 to

100 mg/kg, day ⫺1 to 7) could inhibit tumor growth (S-180) by 31 to 40%. A combination

of cyclophosphamide and ABP increased the rate of tumor growth inhibition by 58%. ABP

could potentiate LAK cell activity and increase the Con A–induced production of tumor

necrosis factor (TNF-β) from murine spleenocytes. The S-180 cell membrane content of

sialic acid was increased, and phospholipid decreased after ABP had acted on cells for

24 hours. Data suggest that the antitumor mechanism of ABP may be related to potentiation

of host immunosurveillance mechanism and the changes in cell membrane features.

Viscum Album and Echinacea Purpurea Extracts of Viscum album (Plenosol) and

Echinacea purpurea (Echinacin) are used clinically for their nonspeciﬁ c action on cell-

mediated immunity. These two were shown to possess a stimulating effect on the production

of lymphokines by lymphocytes and in the transformation test. A toxic effect on cells was

produced only with very high, clinically irrelevant concentrations. Clinical application

of these extracts can produce a stimulation of cell-mediated immunity (one therapeutic

administration followed by a free interval of one week) or can have a depressive action

(daily administration of higher doses). These observations were conﬁ rmed by lymphokine

production and assay, 3H for at least three months, thymidine incorporation, and a skin test

with recall antigens.

16.4.2 Rasayana Botanicals as Adaptogens

Tinospora Cordifolia Treatment with aqueous, alcohol, acetone, and petroleum ether

extracts of stem of T. cordifolia resulted in signiﬁ cant improvement in mice swimming

time and body weights, and petroleum ether extract showed signiﬁ cant protective effect

against cyclophosphamide-induced immunosuppression. Prevention of cyclophosphamide-

induced anemia was also reported.

Withania Somnifera Pretreatment with W. somnifera, T. cordifolia, and Asparagus rac-

emosus induced a signiﬁ cant leucocytosis in cyclophosphamide-induced myelosuppresed

animals. In terms of phagocytosis and intracellular killing, PMN functions, were stimu-

lated, and reticuloendothelial system functions were greatly activated in treated animals.

The phagocytic functions of peritoneal and alveolar macrophages were also stimulated

with T. cordifolia, A. racemosus, and Emblica ofﬁ cinalis. T. cordifolia, A. racemosus, E.

ofﬁ cinalis, Terminalia chebula, Bacopa monira, and W. somnifera improved the carbon

clearance, indicating stimulation of the reticuloendothelial system. A signiﬁ cant increase

in the proliferative fraction in the bone marrow was observed in mice treated with T. cor-

difolia, as revealed in ﬂ ow cytometry analysis.

A comparative pharmacological inves-

tigation of W. somnifera (Ashwagandha) and ginseng showed a signiﬁ cant difference in

antistress activity. Ginseng exhibited higher antistress activity; however, gastric ulcers due

to swimming stress were notably less in ashwagandha. The anabolic study revealed that

the ashwagandha-treated group had a greater gain in body weight than did the ginseng

group.

16.5 CHEMOPROTECTION

Modern cancer therapy produces substantial acute and chronic toxicity, which impairs

quality of life and limits the effectiveness of treatment. Recent clinical and laboratory data

suggest that repair of treatment-related injury is a multiphase and continuous process pro-

viding multiple opportunities for pharmacologic intervention. A host of agents (toxicity

antagonists) are under development that modulate normal tissue response or interfere with

mechanisms of toxicity. Although signiﬁ cant challenges remain, the routine application of

such agents promises to reduce treatment-related morbidity substantially and potentially

to allow treatment intensiﬁ cation in high-risk disease.

The concept of site-speciﬁ c inac-

tivation of cytotoxic anticancer agents has been explored with numerous modalities. The

goal of such chemoprotection is to improve the therapeutic ratio of an agent by selectively

reducing its toxicity in non-tumor-bearing tissue, which is target for dose-limiting toxicity.

Furthermore, a chemoprotectant cannot add new toxicities that might otherwise limit the

administration of maximally tolerated doses of chemotherapeutic agent.

20,21

16.5.1 Drug Targets and Current Trends

Chemoprotection and cytoprotection are studied under preventive oncology and are inter-

changeable terms in cancer chemotherapy. Preventive oncology applies pharmacological

agents to reverse, retard, or halt progression of neoplastic cells to invasive malignancy.

Cancer chemoprevention is one of the newer approaches in the management of cancer.

Epidemiological observations, preclinical animal pharmacology, knockout models, can-

cer cell lines, and clinical trials have shown the efﬁ cacy of this approach. Many drug

targets are under clinical development; prostaglandin pathway, estrogen receptor mod-

ulation, gluthathione peroxidase inhibition, and immunomodulation appear promising.

Celecoxib, tamoxifen, retinoids, rexinoids, selenium, tocopherols, and mofarotene are

some of the promising leads and are in clinical development.

New opportunities in

clinical chemoprevention research include investigating chemopreventive effects of phy-

tochemicals.

Safer immunomodulating agents suitable for long-term therapy remain an

unmet therapeutic need. Figure 16.1 gives a schematic overview of cytoprotection and

immunomodulation.

CHEMOPROTECTION 409

410 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

16.5.2 Chemoprotectants for Antimetabolites

Methotrexate The best-studied chemoprotectant regimen involves rescue from high-

dose methotrexate bone marrow toxicity using reduced folate derivative leucovorin (folinic

acid). This agent is able to substitute for the endogenous reduced folate cofactor, which

methotrexate diminishes. Leucovorin can thereby replenish intracellular reduced folate

pools and prevent methotrexate-induced cytotoxicity via blockade of thymidine synthesis.

However, leucovorin rescue is not truly site-speciﬁ c and will principally depend on

differences in tumor and normal cell growth rates. Schedule-dependent chemoprotection

is described for asparaginase and methotrexate. Preclinical studies show that asparaginase

administered before methotrexate can attenuate its toxicity by inhibiting cellular protein

synthesis. Reduced intracellular polyglutamation of methotrexate by asparaginase

reduces drug retention and thereby causes more rapid methotrexate efﬂ ux, with reduced

toxicity.

25,26

Fluorouracil (5-FU) Chemoprotectants for 5-FU have been directed primarily at the

formation of 5-fulorouridylate triphosphate (5-FUTP), which is closely associated with

normal tissue toxicity in the gut and bone marrow. The best-studied clinical interaction

involves the modulation of 5-FU myelotoxicity with concomitant allopurinol. Selective

chemoprotection involves inhibition of orotidine monophosphate decarboxylase by

allopurinol metabolite oxipurinol. This enzymatic blockade leads to a buildup of orotidine

and orotic acid, which ultimately blocks pathways leading to FUTP formation and halts

RNA inhibition by 5-FU. Selectivity for allopurinol depends on different primary pathways

or ratios of 5-FU activation in normal cells and tumor cells.

The RNA base uridine

has also been reported to block 5-FUlethal toxicity in mice and humans. Modulation of

Figure 16.1 Chemoprotection and immunomodulation.

NEEDS IMMUNOCORRECTION

IMMUNOMODULATORS

Radiation

Surgery

Alkylating agents

Anti-

metabolites

Dose limiting side

effects

Immunosuppression

and myelotoxicities

to normal cells

Cytotoxic

antibiotics

Plant

derivatives

Hormones

and steroids

ANTICANCER EFFECT

CHEMOPROTECTION

RADIOPROTECTION

5-FU associated gastrointestinal toxicity leading to protection was afforded by uridine

diphosphoglucose (URDP), a precursor of uridine. It was found that URDP reduces the

toxic side effects and increases the therapeutic index. The incorporation of 5-FU metabolites

into RNA is blocked primarily as a result of increased intracellular uridine pools. Uridine

rescue from 5-FU toxicity depends on prolonged exposure to uridine rather than attaining

high peak plasma levels of uridine.

28–30

Arotinoid Ro 40-8757 The arotinoid Ro 40-8757 (mofarotene) exhibits a high degree

of activity against established chemically induced mammary tumors in rats. Treatment of

animals with high doses of arotinoid leads to reductions in tumor numbers, with many

animals becoming free of palpable tumors. The toxicities associated with these therapeutic

effects are relatively mild compared to those of all-trans retinoic acid or 13-cis retinoic

acids given at doses with little or no antitumor efﬁ cacy. However, long-term treatment

with Ro 40-8757 results in new growth of tumors. In a rat mammary tumor model, chronic

administration of cyclophosphamide (5 days/week at 10 mg/kg) plus daily administration

of arotinoid at a relatively low dose (75 mg/kg per day) indicated an additive antitumor

effect. However, the therapeutic effects were synergistic because all of the animals treated

with cyclophosphamide as a single agent died after 6 weeks of treatment, whereas all

of the animals given the combination survived the full 10 weeks of the experiment. The

results of detailed studies on the hemopoietic progenitor cells in mice treated with this

combination demonstrated that the protective effect of Ro 40-8757 occurred at the level

of the bone marrow progenitors.

Combination therapy of arotinoid Ro 40-8757 and

5-FU of established chemically induced mammary tumors in rats signiﬁ cantly enhanced

the reduction in tumor burden and tumor number. Ro 40-8757 did not have an effect on

tumor burden. This protective effect of arotinoid makes it a useful potential partner for

combination therapy with 5-FU.

Nucleophilic Sulfur Thiols as Alkylating Agent Chemoprotectants Administration of

supplemental thiols or other compounds with an available reduced sulfur atom against

DNA-alkylating or DNA-binding drugs forms a biochemical rationale of chemoprotection.

Commercially available thiols, or compounds containing a free sulfhydryl residue, include

the amino acid cysteine, its N-acetylated derivative mucomyst, and the sodium sulfonate

salt of 2-mercaptoethane sulfonate mesna. The speciﬁ city for chemoprotection by sulfur

mucleophiles lies in their physical and pharmacokinetic properties. For example, sodium

thiosulfate is a small molecule that is highly charged and hydrophilic at physiologic pH.

Therefore, it distributes in an active state in the bloodstream but achieves poor uptake into

lipophilic tissue compartments, including the central nervous system and the bone marrow.

Sodium thiosulfate concentrates in the renal tubules during its rapid urinary elimination,

indicating that thiosulfate will directly (and nonspeciﬁ cally) inactivate any electrophilic

(alkylating) species in the urine or in the bloodstream. For this reason, sodium thiosulfate is

useful only as a local chemoprotectant for alkylating agents such as mechlorethamine. The

poor distribution of sodium thiosulfate into bone marrow further lessens its chemoprotectant

utility for agents such as mechlorethamine.

16.5.3 Thiol-Based Chemoprotectants for Cisplatin and Oxazophosphorine-Based

Alkylating Agents

Reduction of associated nephrotoxicity and bone marrow suppression was studied by

allowing dose escalation and designing pharmacokinetically based dosing schedules.

CHEMOPROTECTION 411

412 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

However, new dose-limiting toxicities consisting of peripheral neuropathy and ototoxicity

have emerged, which continue to restrict the potential use of high-dose cisplatin therapy.

Chemoprotective agents, including sodium thiosulfate, WR2721, and diethyldithiocarba-

mate (DDTC) are being examined extensively as “rescue agents’’ for either regional or

systemic administration of cisplatin. Sodium thiosulfate (STS) is given intravenously con-

currently or following intraperitoneal (i.p.) cisplatin administration. This can produce a

12-fold greater exposure to cisplatin in the i.p. space and has been associated with tumor

regression. However, this approach is limited to locoregional drug treatment since active

form of sulfur nucleophile is distributed in the bloodstream and could complex with cis-

platin to abrogate its intended systemic therapeutic effects.

Many studies suggest local

application of STS to prevent cisplatin toxicity and maintain the systemic antitumoral

effectiveness of cisplatin. Cocheal admistration resulted in complete protection against

CDDP-induced hearing loss, with no change in compound action potential (CAP).

How-

ever, alleviation of cisplatin-induced side effects by administration of sodium thiosulfate

was not observed in a mice xenograft tumor model, suggesting that cisplatin–STS treat-

ment would provide no beneﬁ t in patients treated with cisplatin.

Several other thiol-

based compounds have shown activity in preventing cisplatin-induced toxicities. These

include the experimental aminothiol WR-2721, the disulﬁ de metal chelator diethyldithio-

carbamate (DDTC), mesna, N-acetylcysteine (NAC), and thiourea.

The oxazophosphorine-based antitumor agents include cyclophosphamide and ifos-

famide. These alkylating agents require metabolic activation by hepatic microsomal (P450)

enzymes to active species, including the bis functional alkylator phosphoramide mustard

and the protein-reactive aldehyde acrolein. A difference between cyclophosphamide and its

chloroethyl isomer ifosfamide involves primarily a slower rate of metabolic activation of

ifosfamide to 4-hydroxyifosfamide, the precursor of the active species of the drug. Large

amounts of toxin acrolein are produced and accumulate in the urinary bladder. In the ab-

sence of a chemoprotectant, active doses of ifosfamide can produce dose-limiting urotoxic-

ity manifested by hemorrhagic cystitis, bladder ﬁ brosis, and heightened long-term risk of

bladder cancer. Similar toxicities are noted with chronic or acute high-dose cyclophospha-

mide dosing. Thus, both ifosfamide and high-dose cyclophosphamide have limited clinical

efﬁ ciency in the absence of the effective chemoprotectant of the urinary bladder.

Mesna Mesna is the sodium salt of 2-mercaptoethanesulfonic acid, a selective urinary

tract protectant for oxazophosphorine-type alkylating agents. Mesna prevents bladder

damage from major toxic metabolites of ifosfamide and cyclophosphamide. Mesna can

bind speciﬁ cally to cisplatin or alkylating agent–generated free radicals or alkylating agent

metabolites to reduce the incidence of cisplatin-associated neurotoxicity and nephrotoxicity

or alkylating agents associated with myelosuppression and urothelial toxicity.

Mesna

does not block the antitumor activity of oxazophosphorines or of other classes of antitumor

agents. Mesna is superior to previous urinary prophylaxis regimens among a large series

of SH-based uroprotectants with the least toxicity of any of the agents tested.

Mesna also

attenuates the lethal effect and hematological toxicity of vespeside and taxol but does not

reduce speciﬁ c activities in mice with transplanted tumors.

Amifostine Amifostine is organic thiophosphate compound able to protect normal tissues

selectively against cytotoxic agents in cellular and animal models without protecting tumor

tissues. Amifostine is a prodrug that is dephosphorylated into its active metabolite, a free

thiol derivative, by the membrane alkaline phosphatase of target tissue. This unique

metabolism supports its cellular selectivity and its preferential uptake by normal tissues.

Preclinical animal studies have demonstrated that administration of amistoﬁ ne protects

against irradiation and a variety of chemotherapy-related toxicities, including cisplatin-

induced nephrotoxicity, neurotoxicity, and cyclophosphamide- and bleomycin-induced

pulmonary toxicity and cardiotoxicity induced by doxorubicin and related chemotherapeutic

agents. In nonrandomized and randomized trials in malignant melanoma, colorectal

cancer, head and neck cancer, non-small cell lung cancer, and epithelial ovarian carcinoma,

amifostine signiﬁ cantly reduced the hematological and nonhematological toxicity of DNA-

damaging agents such as alkylators, platinum compounds, and mitomycin C. In more

recent studies, amistoﬁ ne also protected patients from side effects produced by taxanes or

topoisomerase I inhibitors.

Currently, there is no evidence that amifostine compromises

the antineoplastic effect of the drugs studied. Moreover, amifostine appears to produce

growth factor–like properties, resulting in growth-promoting effects on primitive blood

progenitor cells ex vivo. In a randomized phase III study conducted in patients with ovarian

carcinoma receiving a combination of cisplatin and cyclophosphamide, a signiﬁ cant

decrease in hematological, renal, and neurologic toxicity was observed in amifostine-

treated patients compared with the control group.

The protective effect of amifostine has

been demonstrated for cisplatin-induced toxicity in lung and ovarian cancer, with particular

regard to nephrotoxicity, neurotoxicity, and neutropenia. No protective effect has been seen

for tumor cells, owing to a selective action of amifostine on healthy tissues. A frequent side

effect of amifostine is a transient decrease in blood pressure; it is usually asymptomatic

if an easily handled premedication is given. Cytoprotection by amifostine is also well

known for alkalyting drugs and radiation therapy, whereas it is still the object of study for

new drugs, especially taxanes.

Amifostine also protects bone marrow from cumulative

toxicity arising from chronic exposure to therapeutic agents such as alkylating agents.

Well-controlled clinical trials have shown that amifostine can ameliorate cumulative bone

marrow toxicity and the acute and chronic neutropenic and/or thrombocytopenic effects

of cyclophosphamide. In a pivotal phase III study of cisplatin–cyclophosphamide with or

without amifostine, amifostine reduced course-by-course cumulative bone marrow damage

compared with the course-by-course cumulative myelosuppression experienced by those

treated with cisplatin or cyclophosphamide alone.

Disulﬁ ram Disulﬁ ram (tetraethylthioperoxidicarbonic diamide), an aldehyde dehydroge-

nase inhibitor, prevented cyclophosphamide-induced bladder damage in a dose-dependent

manner when administered simultaneously with cyclophosphamide [100 to 400 mg/kg,

intraperitoneally (i.p.)] but failed to diminish the acute toxicity, leucotoxicity, and immuno-

toxicity of cyclophosphamide. Diethyldithiocarbamate (DDTC), a metabolite of disulﬁ ram,

did not interfere with cyclophosphamide antitumor activity when administered 3 hours

after cyclophosphamide. The protective effect of disulﬁ ram on the bladder was critically

dependent on administration timing. Disulﬁ ram slightly potentiated the antitumor activity

of cyclophosphamide against Sarcoma-180 or EL-4 leukemia in vivo when administered

simultaneously with cyclophosphamide.

Disulﬁ ram is an effective protective agent against bladder damage caused by ifosfamide

treatment.

Disulﬁ ram (DSF), in combination with ifosfamide (IFX), prevented IFX-

induced bladder damage but failed to diminish the acute lethal toxicity or leukocytotoxic-

ity of IFX. Diethyldithiocarbamate (DDTC) prevented IFX-induced bladder damage when

administered simultaneously with IFX or 1 to 5 hours afterward. The antitumor activity

of IFX in ddY-mice inoculated with Sarcoma-180 or in C57BL/6J mice inoculated with

CHEMOPROTECTION 413

414 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

EL-4 leukemia was not impaired when it was given simultaneously with DSF or 3 hours

before DDTC. Thus, neither DSF nor DDTC impaired the antitumor effect of IFX, and

both diminished its adverse effects. The bladder protection of DSF and DDTC appeared

to be resulting from adduct formation with acrolein and not from inhibition of the meta-

bolic activation of IFX.

Disulﬁ ram (DSF) blocks the urotoxicity of cyclophosphamide

in mice and increases the oncolytic effect of cyclophosphamide in the L1210 murine leu-

kemia. However, mice treated with cyclophosphamide and DSF appeared to have longer-

lasting neutropenia than did animals treated with cyclophosphamide alone. Bone marrow

granulocyte/macrophage progenitor cells (GM-CFCs) were relatively well preserved and

the recovery of the GM-CFCs was not prolonged by DSF, indicating that the acute cyto-

toxic effect of cyclophosphamide on the granulocyte/macrophage progenitor cells is not

enhanced by DSF.

16.5.4 Chemoprotectants for Anthracyclines

Anthracyclins such as doxorubicin and daunomycin comprise one of the most important

classes of DNA-binding antitumor agents. Short- and long-term cardiotoxicity can occur

at lower doses, and the use of these drugs is limited by a characteristic clinical cardiomy-

opathy that develops in approximately 5 to 15% of patients after cumulative doxorubicin

doses greater than 450 mg/m

. Children and adolescents appear to be particularly sensitive

to the cardiotoxic effects of doxorubicin. Although cumulative doxorubicin doses are usu-

ally limited to ⱕ450 mg/m

, up to 70% of long-term survivors of childhood cancer have

evidence of cardiac dysfunction, including overt congestive heart failure.

Anthracyclines

complexed with metals can sustain lipid peroxidation, and scavengers of reduced oxy-

gen free radicals, including superoxide dismutase, catalase, and mannitol, do not block its

action.

A variety of putative free-radical scavengers have been shown to protect against

doxorubicin cardiotoxicity in experimental animals. These include the lipid-soluble anti-

oxidant vitamin E

and N-acetylcysteine (NAC). Despite positive preclinical results, NAC

was not effective in cancer patients given high cumulative doxorubicin doses.

ICRF-187 The piperazine derivative of ethylenediaminetetraacetate razoxane (ICRF-

187) is a prodrug that is converted intracellularly to an iron-chelating agent that removes

iron from doxorubicin-iron complexes in vitro. Dexrazoxane is a cardioprotective

antioxidant that is used clinically to reduce the cardiotoxicity of the chemotherapeutic

drugs doxorubicin, paclitaxel, and other anthracyclins.

Although the cardioprotective

effect of dexrazoxane in cancer patients undergoing chemotherapy with anthracyclines

is well documented, the potential of this drug to modulate topoisomerase II activity and

cellular iron metabolism may hold the key for future applications of dexrazoxane in cancer

therapy, immunology, or infectious diseases.

16.5.5 Botanical Immunomodulators as Chemoprotectants

Withania Somnifera Withania somnifera is an ofﬁ cial drug mentioned in the Indian

Herbal Pharmacopoeia

and Ayurvedic Pharmacopoeia.

Studies indicate that W. som-

nifera (ashwagandha) (WS) possesses anti-inﬂ ammatory, antitumor, antistress, antioxidant,

immunomodulatory, hemopoietic, and rejuvenating properties. The chemistry of WS has

been studied extensively, and over 35 chemical constituents have been identiﬁ ed, extracted,

and isolated. The biologically active chemical constituents are alkaloids (isopelletierine,

anaferine), steroidal lactones (withanolides, withaferins), saponins containing an addition-

al acyl group (sitoindoside VII and VIII), and withanolides with a glucose at carbon 27

(sitoindoside IX and X).

The suppressive effect of cyclophosphamide-induced toxicity by WS extracts was ob-

served in mice. Administration of WS extracts signiﬁ cantly reduced leucopenia induced by

cyclophosphamide treatment, resulting in an increase in bone marrow cellularity. Adminis-

tration of W. somnifera extract for 5 days along with cyclophosphamide (CTX) (1.5 mmol/

kg body weight, i.p.) reduced the CTX-induced urotoxicity.

Treatment of W. somnifera

resulted in the enhancement of interferon gamma (IFN-gamma), interleukin-2 (IL-2), and

granulocyte macrophage colony stimulating factor (GM-CSF), which were lowered by cy-

clophosphamide administration. Pharmacodyanmic studies reveal that the major activity of

W. somnifera may be due to the enhancement of cytokine production and stem cell prolif-

eration and its differentiation.

These studies indicate that W. somnifera could reduce the

cyclophosphamide toxicity and its usefulness in cancer chemotherapy.

The major activity of WS may be due to stimulation of stem cell proliferation, indicat-

ing the fact that WS could reduce cyclophosphamide toxicity and its usefulness in cancer

chemotherapy.

WS was also shown to prevent lipid peroxidation (LPO) in stress-induced

animals, indicating its adjuvant as well as chemoprotectant activity.

Glycowithanolides,

consisting of equimolar concentrations of sitoindosides VII to X and withaferin A, isolated

from the roots of WS were evaluated for protection in iron-induced hepatoxicity in rats.

Ten days of oral administration of these active principles, in graded doses (10, 20, and 50

mg/kg), resulted in attenuation of hepatic lipid peroxidation (LPO) and the serum enzymes

alanine aminotransferase, aspartate aminotransferase, and lactate dehydrogenase during

iron-induced hepatoxicity.

Antistress activity observed with W. somnifera will be an ad-

ditional beneﬁ t, along with chemoprotectant activity.

Tinospora Cordifolia Tinospora cordifolia is widely used in ayurvedic medicines and is

known for its immunomodulatory, antihepatotoxic, antistress, and antioxidant properties.

It has been used in combination with other plant products to prepare a number of ayurvedic

preparations. The chemistry has been studied extensively, and its chemical constituents

can be broadly divided into alkaloids, diterpenoids, steroids, ﬂ avanoids, and lignans.

Reviews have appeared on quaternary alkaloids and biotherapeutic diterpene glucosides

of Tinospora species. Much of the work has been carried out on berberine, jatrorrhizine,

tinosporaside, and columbin. Extracts of T. cordifolia (TC) have been shown to inhibit

lipid peroxidation and superoxide and hydroxyl radicals in vitro. The extract was also

found to reduce the toxic side effects of cyclophosphamide (25 mg/kg, 10 days) in the mice

hematological system by free-radical formation as seen from total white cell count, bone

marrow cellularity, and α-esterase-positive cells.

The active principles of TC were found

to possess anticomplementary and immunomodulatory activities.

TC is reported for its

various immunopharmacological activities (e.g., inhibition of C3-convertase of the classical

complement pathway). Humoral and cell-mediated immunity were reported for cardioside,

cardifolioside A, and cardiol and their activation was more pronounced with increasing

incubation time.

Extracts of T. cordifolia has been shown to inhibit lipid peroxidation and

superoxide and hydroxyl radicals in vitro. The extract was also found to reduce the toxic

side effects of cyclophosphamide (25 mg/kg, 10 days) in the mice hematological system

by free-radical formation as seen from total white cell count, bone marrow cellularity, and

α-esterase-positive cells.

CHEMOPROTECTION 415

416 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

Tinospora Bakis Dose-dependent cytoprotection by Tinospora bakis, a plant from the

Senegalese pharmacopoeia, was observed in an in vitro model. A lyophilized aqueous

extract of plant roots decreased intracellular enzyme release (LDH and ASAT) from CCl

intoxicated hepatocytes isolated from rats. The cytoprotective effect was more effective for

long-course treatment.

Asparagus Racemosus Chemically, Asparagus racemosus (AR) contains steroidal

saponins, known as shatavarins, isoﬂ avaones, and isoﬂ avones, including 8-methoxy-5,

6,4'-trihydroxyisoﬂ avone 7-O-β-D-glucopyranoside, asparagamine, a polycyclic alkaloid,

racemosol, a cyclic hydrocarbon (9,10-dihydrophenanthrene), polysaccharides, and

mucilage. AR has been shown to stimulate macrophages and inﬂ uence long-term adaptation

favorably. Possible links between immunomodulatory and neuropharmacological activity

have been suggested. Extracts of A. racemosus were evaluated for its neuroendocrine

immune modulating effect. It prevents stress-induced increase in plasma cortisol along

with activation of peritoneal macrophages and inhibition of gastric vascular damage.

A comparative study between A. racemosus, T. cordifolia, glucan, and lithium carbonate

against the myelosuppressive effects of single does (200 mg/kg, subcutaneously) and

multiple doses (three doses, 30 mg/kg, i.p.) of cyclophosphamide in mice revealed that all

four drugs prevented, to varying degrees, leucopenia produced by cyclophosphamide.

Treatment of A. racemosus signiﬁ cantly inhibited ochratoxin A–induced suppression of

chemotactic activity and production of IL-1 and TNF-α by macrophages.

Our studies on these plants (i.e., WS, TC, and AR) revealed that they have signiﬁ cant

immunomodulatory activity. This activity can be useful in a variety of conditions, such as

myeloprotection in cancer chemotherapy and immunoprotection during infection. We ob-

served that treatment of ascitic sarcoma–bearing mice with formulation of total extracts of

WS and TC and alkaloid-free polar extract of WS resulted in protection toward cyclophos-

phamide-induced myelo- and immunosuppression. In another situation, these plants were

evaluated for their immunoadjuvant potential in a pertussis model, where aqueous extracts

of these plants reduced the dose of vaccine required to confer protection against pertussis

intracerebral challenge, increasing survival percentage and signiﬁ cantly increasing in per-

tussis antibody titers. These observations are of major importance in the immunochemical

industry and in vaccination strategies.

66–68

Crocetin A natural carotenoid, crocetin, at a dose of 50 mg/kg, modulated the release

of chloroaceteldehyde, a urotoxic metabolite of cyclophosphamide in the urine of

mice given a combined treatment. Crocetin at the same dose signiﬁ cantly elevated

glutathione-S-transferase enzyme activity in both the bladder and liver of mice treated

with cyclophosphamide. In Sarcoma-180 tumor–bearing mice, crocetin has the ability to

protect against cyclophosphamide-induced bladder toxicity without altering its antitumor

activity.

Crocetin also inhibited benzo[a]pyrene–induced genotoxicity and neoplastic

transformation in C3H10T1/2 cells. Crocetin was found to increase the activity of GST

and decreases the formation of a benzo[a]pyrene–DNA adduct.

UL-409 Oral administration of UL-409, a herbal formulation, at a dose of 600 mg/kg

signiﬁ cantly prevented the occurrence of cold-resistant stress-induced ulcerations in Wistar

rats, alcohol- and aspirin-induced gastric ulceration, and cysteamine- and histamine-

induced duodenal ulcers in rats and guinea pigs, respectively. The volume and acidity of

gastric juice in pylorius-ligated rats was also reduced by UL-409. It also signiﬁ cantly, and

dose dependently, promoted gastric mucus secretion in normal as well as in stress-, drug-,

and alcohol-induced ulceration in animals.

Mikania Cordata Induction of phase 2 enzymes is an effective and sufﬁ cient strategy

for achieving protection against the toxic and neoplastic effects of many carcinogens.

Literature reports suggest that the chemopreventive action of Mikania chordata is based on

its effect on phase 2 enzymes. M. cordata oral administration resulted in increased activity

of microsomal uridine diphosphoglucose dehydrogenase and reduced nicotinamide adenine

dinucleotide (phosphate): quinine reductase and cytosolic glutathione S-transferases, with

a concomitant elevation in the contents of reduced glutathione. M. chordata was also

found to increase the total protein mass, fractional rate of protein synthesis, ribosomal

capacity and efﬁ ciency (rate/ribosome), and high turnover rate of protein (protein/DNA)

on pretreatment in CCl

-treated hepatic tissue. This indicated tissue repair leading to a

functional improvement in CCl

disorganized hepatocytes.

Oral administration of a

methanolic fraction of M. cordata (Burm., B. L. Robinson) signiﬁ cantly prevented the

occurrence of water-immersion stress-induced gastric ulcers in a dose-responsive manner.

The extract also dose-dependently inhibited gastric ulcers induced by ethanol, aspirin,

and phenylbutazone. The volume, acidity, and peptic activity of the gastric juice in

pylorus-ligated rates were not altered upon administration of the extract but signiﬁ cantly

and dose-dependently promoted gastric mucus secretion in normal as well as stress- and

ethanol-induced ulcerated animals. It was claimed that the activity observed might be due to

the modulation of defensive factors through an improvement in gastric cytoprotection.

Indigenous Herbal Drug Formulations Brahma Rasayana and Ashwagandha Rasayana

were found to protect mice from cyclophosphamide-induced (50 mg/kg daily for 14 days)

myelosuppression and subsequent leucopenia.

Treatment with A. racemosus, T. cordifolia,

W. somnifera, and Picrorhiza kurrooa signiﬁ cantly inhibited carcinogen ochratoxin A

(OTA)–induced suppression of chemotactic activity and production of interleukin-1

(IL-1) and tumor necrosis factor–alpha (TNF-α) by macrophages. Immu-21, a polyherbal

formulation that contains extracts of Ocimum sanctum, W. somnifera, E. ofﬁ cinalis, and

T. cordifolia, at 100 mg/kg daily over 7 days and 30 mg/kg daily over 14 days prevented

cyclophosphamide-induced genotoxicity in mice.

16.6 RADIOPROTECTION

Radiotherapeutics is still one of the major treatment modalities practiced for control of

localized solid tumors. The major goal of therapy is the achievement of total tumor control

with limited toxicity and complications. Radiation doses that can be delivered without

causing severe damage to surrounding normal tissues can be insufﬁ cient to eradicate a

tumor. New strategies for the prevention of radiation injuries are currently being explored

with the ultimate aim of developing globally radioprotective nontoxic pharmacologics.

These include the development of agents as radiosensitizers, apoptosis inducers in tumor

cells, and radioprotectants, which can increase the radiosensitivity of such resistant tumors,

reduce the radiation dose required, and protect the normal tissue morbidity associated with

tumoricidal radiation doses. Radioprotective agents, although widely studied in the past

four decades and including several thousand agents, have not reached the level of provid-

ing the agent that conforms to criteria required of an optimal radioprotective, including

RADIOPROTECTION 417

418 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

effectiveness, speciﬁ city, availability, toxicity, and tolerance. The prophylactic treatments

under review encompass diverse pharmacological classes as novel immunomodulators, nu-

tritional antioxidants, and cytokines.

With the advancement in understanding of tumor cell biology, many molecular mecha-

nisms or targets have been identiﬁ ed for modulation of radiation response. Some important

ones are COX-II, MMPs, TRAIL (TNF-α-related apoptosis-inducing ligand)/Apo2L, and

epidermal growth factor receptor (EGFR).

77–79

Substances reducing the radiation-induced

toxicity by modulating the biological response to radiation injury may represent an alterna-

tive concept in radioprotection, and hence there is interest in and the need for new compounds

that can protect tissues from radiation injury. Natural compounds may have an advantage,

being more structurally diverse and safer, hence more acceptable for human application.

16.6.1 Radioprotectants from Botanicals

Withaferin A is a steroidal lactone from W. somnifera inhibited growth of Ehrlich ascites

carcinoma in Swiss mice with increased survival and life span. Antitumor and radiosensi-

tizing effects were observed with a combination treatment of abdominal gamma irradiation

with withaferin A, resulting in increased tumor cure and tumor-free survival.

Withaferin

A showed growth inhibitory effect in vitro on both Chinese hamster V79 and HeLa cells.

It reduced the survival of V79 cells in a dose-dependent manner.

Combination treatment

of an alcoholic extract of Withania somnifera (500 mg/kg, i.p., 10 days) with one local

exposure to gamma radiation (10 Gy) followed by hyperthermia (43⬚C for 30 minutes) sig-

niﬁ cantly increased the tumor cure (Sarcoma-180 grown on the dorsum of adult BALB/c

mouse), growth delay, and animal survival. This combination also depleted the tumor GSH

level signiﬁ cantly and synergistically. Thus, in addition to having a tumor inhibitory effect,

ashwagandha acts as a radiosensitizer, and heat enhances these effects. The severe deple-

tion in the tumor GSH content by the combination treatment must have enhanced the tumor

response, as the inherent protection of the thiol will be highly reduced.

The presence of

a wide variety of effects, such as hypotensive, antispasmodic, antitumor, antiarthritic, anti-

pyretic, analgesic, anti-inﬂ ammatory, and hepatoprotective activities and antistress proper-

ties, show that W. somnifera may be acting by nonspeciﬁ cally increasing the resistance of

the animals to various stressful conditions.

16.6.2 Botanical Immunomodulators as Antitumor Agents

Plant products have contributed several novel compounds that possess promising antitumor

activity. Crude extract of W. somnifera root has a strong tumoricidal and tumor growth

inhibitory activity. Withaferin A, an alkaloid isolated from the leaves, has been reported to

show marked tumor inhibitory activity in vitro against cells derived from human carcinoma

of nasopharynx and experimental mouse tumors. A single i.p. dose of withaferin A injec-

tion 24 or 48 hours after Ehrlich’s ascites tumor transplantation produced an immediate

growth reduction in 3 to 80% of mice, followed by complete disappearance of tumor cells

in the peritoneal cavity of the surviving mice with no signs of tumor development. How-

ever, withaferin A is toxic in mice, with an LD

of 54 mg/kg body weight after an i.p. injec-

tion. An effective dose of crude extract was much higher (a cumulative dose of more than

10 g—750 mg/kg daily for 15 days) with less toxicity than reported doses of puriﬁ ed witha-

ferin A and withanolide D, which exhibited toxic effects. Crude extract included a range of

chemicals (e.g., a few ﬂ avanoids, several alkaloids, and other withanolides) in addition to

withaferin A. In the ayurvedic system of treatment, dry powders or crude extract is used,

and hence the effects observed may not be attributed to a single component. The rationale

for this type of treatment is that the toxicity of an active component may be counteracted

by another component, which may not have the desired therapeutic property.

Resinous material from a methanol extract and orange-colored oil from a petroleum

ether extract of Semecarpus anacardium Linn. F. have been found to possess antitumor

properties in a P388 lymphocytic leukemia model.

Acetylated oil of Semicarpus ana-

cardium, which itself does not possess antitumor activity against experimental transplant-

able tumors, enhances the antitumor effect of anticancer drugs such as mitomycin-C,

6-mercaptopurine, and methotrexate when used in combination against P388 and S180

(ascites) tumor systems.

Extracts of T. cordifolia have been shown to inhibit lipid peroxidation and superoxide

and hydroxyl radicals in vitro. The extracts were also found to reduce the toxic side ef-

fects of cyclophosphamide administration in mice. Moreover, their administration partially

reduced elevated lipid peroxides in serum and liver as well as alkaline phosphatase and

glutamine pyruvate transaminase thus indicating the value of Tinospora extracts in reduc-

ing the chemotoxicity induced by free radical–forming chemicals.

Crude saponins obtained from shoots of Asparagus racemosus [asparagus crude

saponins (ACSs,] were found to have antitumor activity. They inhibited the growth of

human leukemia HL-60 cells in culture and macromolecular synthesis in a dose- and

time-dependent manner. The ACS in the range 75 to 100 µg/mL was cytostatic; at concen-

trations greater than 200 µ/mL it was cytocidal to HL-60 cells. ACSs at 6 and 50 µg/mL

inhibited the synthesis of DNA, RNA, and protein in HL-60 cells by 41, 5, and 4% or

by 84, 68, and 59%, respectively. The inhibitory effect of ACSs on DNA synthesis was

irreversible.

16.7 CONCLUSIONS

Many chemical agents are used as chemoprotectants for conventional cancer chemotherapy

and/or radiation therapy. However, their effect is locoregional and is dependent on dose

and time of administration, in contrast to anticancer drugs. The limitations and inconve-

nience of their use has stimulated research to discover natural resources with immunologi-

cal activity. Various botanicals and ethnopharmacological agents of traditional medicine

are under investigation for chemoprotective and immunomodulatory activities. The results

are encouraging. Chemoproﬁ ling of these botanicals have been reported, but most activity

reports are for crude semiprocessed extracts or fractions. Several reports suggest cyto-

static and cytotoxic properties, along with enhanced immune function in extract and or

polyherbal formulations. Nevertheless, any single component isolated from an extract or

formulation may not retain all three desired properties. There have been attempts to isolate

and characterize active moieties, with limited success. Many authors have hypothesized the

presence of synergism and buffering in extracts; however, systematic scientiﬁ c investiga-

tions on pharmacodyanmics, kinetics, dosing, and interactions need to be undertaken to

study these principles. Furthermore, studies are required to better understand the molecular

and biochemical mechanisms involved in immunoregulation and its role in cytoprotection

and radioprotection. Such efforts might lead to effective integration of botanical medicine

in cancer therapy. This review will be useful in bioprospecting exercises toward the devel-

opment of newer, safer, and effective agents for therapeutic management of cancer.

CONCLUSIONS 419

420 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

REFERENCES

Devi, P. U., Sharada, A. C., Solomon, E. F., and Kamath, M. S., In vivo growth inhibitory effect

of Withania somnifera (ashwagandha) on a transplantable mouse tumor, Sarcoma 180, Indian J.

Exp. Biol., 30; 169–172, 1992.

Patwardhan, B., Kalbag, D., Patki, P. S., and Nagasampagi, B. A., Search of immunoregulatory

agents: a review, Indian Drugs, 28(2); 56–63, 1990.

Upadhyay, S. N., Plant products as immune response modulators, Proceedings of the Interna-

tional Ayurveda Conference’ 97, Sanjay Gandhi Post Graduate Institute of Medical Sciences,

Lucknow, Indian 10, pp. 1997.

Chirigos, M. A., Immunomodulators: current and future development and applications, Thymus,

19(1); S7–S20, 1992.

Turowski, R. C., and Triozzi, P. L., Application of chemical immunomodulators to the treatment

of cancer and AIDS, Cancer Invest., 12(6); 620–643, 1994.

Patwardhan, B., Ayurveda: the designer medicine—review of ethnopharmacology and bio-

prospecting research, Indian Drugs, 37; 213–227, 2000.

Patwardhan, B., Vaidya, A., and Chorghade, M., Ayurveda and natural products drug discovery,

Curr. Sci., 86; 789–799, 2004.

Rege, N. N. Adaptogenic properties of six rasayana herbs in ayurvedic medicine, Phytother. Res.,

13; 275–291,1999.

Brekhman, I. L., and Dardimov, I. V., New substances of plant origin which increase nonspeciﬁ c

resistance, Annu. Rev. Pharmacol., 21; 419–426, 1969.

Plata-Salaman, C. R., Immunomodulators and feeding regulation: a humoral link between the

immune and nervous systems, Brain Behav. Immunol., 3(3); 193–213, 1989.

Beuth, J., Clinical relevance of immunoactive mistletoe lectin-I, Anticancer Drugs, 8(1);

S53–S55, 1997.

Kim, J. Y., Germolee, D. R., and Luster, M. I., Panax ginseng as a potential immunomodulator:

studies in mice, Immunopharmacol. Immunotoxicol., 12(2); 257–276, 1990.

Siegel, R. K., Ginseng abuse syndrome: problems with the panacea, J. Am. Med. Assoc., 241(15);

1614–1615, 1979.

Yu, S., and Zhang, Y., Effect of Achyranthes bidenta polysaccharide (ABP) on antitumor activity

and immune function of S-180 bearing mice, Chung Hua Chung Liu Tsa Chih (China), 17(4);

275–278, 1995.

Coeugniet, E. G., and Elek, E., Immunomodulation with Viscum album and Echinacea purpurea

extracts, Onkologie, 10(3); 27–33, 1987.

Patil, M, Patki, P., Kamath, H. V., and Patwardhan, B., Antistress activity of Tinospora cordifolia

(Wild) Miers, Indian Drugs, 34(4); 211–215, 1997.

Dahanukar, S. A., and Thatte, U. M., Current status of ayurveda in phytomedicine, Phytomedi-

cine, 4(4); 359–368, 1997.

Grandhi, A., Mujumdar, A. M., and Patwardhan, B., A comparative pharmacological investiga-

tion of ashwagandha and ginseng, J. Ethnopharmacol., 44; 131–135, 1994.

Trotti, A., Toxicity antagonists in cancer therapy, Curr. Opin. Oncol., 9(6); 569–578, 1997.

Dorr, R. T., Chemoprotectants for cancer chemotherapy, Semi. Oncol., 18(1); 48–58, 1991.

Matthew, L., and Craig, L., Chemoprotectants: a review of their clinical pharmacology and thera-

peutic efﬁ cacy, Drugs, 57; 293–308, 1999.

Krzystyniak K. L. Current strategies for anticancer chemoprevention and chemoprotection, Acta

Pol. Pharm., 59(6); 473–478, 2002.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Krishnan, K., Campbell, S., Abdel-Rahman, F., Whaley, S., and Stone, W. L., Cancer chemopre-

vention drug targets, Curr. Drug Targets, 4(1); 45–54, 2003.

Kucuk, O., New opportunities in chemoprevention research, Cancer Invest., 20(2); 237–245, 2002.

Bertino, J. R., Rescue techniques in cancer chemotherapy: use of leucovorin and other rescue

agents after methotrexate treatment, Semin. Oncol., 4; 203–216, 1977.

Jolivet, J., Cole, D. E., and Holcenberg, J. S., Prevention of methotrexate cytotoxicity by aspara-

ginase inhibition of methotrexate polyglutamate formation, Cancer Res., 45; 217–220, 1985.

Howell, S. B., Wung, W., and Taetle, R., Modulation of 5-ﬂ uorouracil toxicity by allopurinol in

man, Cancer, 48; 1281–1289, 1981.

Klubes, P., and Cerna, I., Use of uridine rescue to enhance the antitumor selectivity of 5-ﬂ uoro-

uracil, Cancer Res., 43; 3182–3186, 1983.

Stolﬁ , R. L., Martin, D. S., and Sawyer, R. C., Modulation of 5-ﬂ uorouracil-induced toxicity

in mice with interferon or with the interferon inducer, polyinosinic–polycytidylic acid, Cancer

Res., 43; 561–566, 1983.

Van Groeningen, C. J., Petes, G. J., and Leyva, A., Reversal of 5-ﬂ uorouracil-induced myelosup-

pression by prolonged administration of high-dose uridine, J. Natl. Cancer Inst., 81; 157–162,

1989.

Eliason, J. F., Inoue, T, Kubota, A., Teelmann, K., Horh, I., and Hartmann, D., The antitumor

arotinoid Ro 40-8757 protects bone marrow from the toxic effects of cyclophosphamide, Int. J.

Cancer, 55; 492–497, 1993.

Eliason, J. F., Inoue, T., Kubota, A., Hurh, I., and Hartmann, D., The anti-tumor arotinoid Ro 40-

8757 protects bone marrow from the toxic effects of 5-ﬂ uorouracil, Int. J. Cancer, 57; 192–197,

1994.

Shea., M., Koziol, J. A., and Howell, S. B., Kinetics of sodium thiosulfate, a cisplatin neutralizer,

Clin. Pharmacol. Ther., 35; 419–425, 1984.

Pfeiﬂ e, C. E., Howell, S. B., Felthouse, R. D., et al., High-dose cisplatin with sodium thiosulfate

protection, J. Clin. Oncol., 3; 237–244, 1985.

Wang, J., Lloyd Faulconbridge, R. V., Fetoni, A., Guitton, M, J., Pujol, R., and Puel, J. L., Local

application of sodium thiosulfate prevents cisplatin-induced hearing loss in the guinea pig,

Neuropharmacology, 45(3); 380–393, 2003.

Masaki, I., Chikako, S., Hiromu, S., and Osamu, T., Neutralizing effect of sodium thiosulfate on

antitumor efﬁ cacy of cisplatin for human carcinoma xenografts in nude mice, Gynecol. Oncol.,

40; 34–37, 1991.

Lewis, C. M., Chemoprotectants: a review of their clinical pharmacology and therapeutic

efﬁ cacy, Drugs, 57(3); 293–308,1999.

Dorr, R. T., Daniel, D., and Hoff, V., Cancer Chemotherapy Handbook, 2nd ed., Appleton &

Lange, East Norwalk, CT, 1994, pp. 685–688.

Hensley, M. L., Schuchter, L. M., Lindley, C., Meropol, N. J., Cohen, G. I., and Broder, J., J.

Clin. Oncol., 17(10); 3333–3355, 1999.

Santini, V., Amifostine: chemotherapeutic and radiotherapeutic protective effects, Expert Opin.

Pharmacother., 2(3); 479–489, 2001.

Lenoble, M., Amifostine: current and future applications in cytoprotection, Bull. Cancer, 83(9);

773–787, 1996.

Castiglione, F., Dalla, M., and Porcile, G., Protection of normal tissues from radiation and cyto-

toxic therapy: the development of amifostine, Tumori, 85(2); 85–91,1999.

Alberts, D. S., Protection by amifostine of cyclophosphamide-induced myelosuppression, Semin.

Oncol., 26(2); 37–40, 1999.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

REFERENCES 421

422 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

Ishikawa, M., Aoki, T., Yomogida, S., Takayanagi, Y., and Sasaki, K., Disulﬁ ram as protective

agent against cyclophosphamide-induced urotoxicity without compromising antitumor activity

in mice, Pharmacol. Toxicol., 74(4–5); 255–261,1994.

Ishikawa, M., and Takayanagi, Y., Inhibition of ifosfamide-induced urotoxicity by disulﬁ ram in

mice, Jpn. J. Pharmacol., 49(1); 147–150, 1989.

Ishikawa, M.,Takayanagi, Y., and Sasaki, K., Drug interaction effects on antitumor drugs: VIII.

Prevention of ifosfamide-induced urotoxicity by disulﬁ ram and its effect on antitumor activity

and acute toxicity of alkylating agents in mice, Pharmacol. Toxicol., 68(1); 21–25, 1991.

Gamelli, R. L., Ershler, W. B., Kacker, M. P., and Foster, P. S., The effect of disulﬁ ram on

cyclophosphamide-mediated myeloid toxicity, Cancer Chemother. Pharmacol., 16(2); 153–155,

1986.

Wexler, L. H., Andrich, M. P., and Venzon, D., Randomized trial of the cardioprotective agent

ICRF-187 in pediatric sarcoma patients treated with doxorubicin, J. Clin. Oncol., 14(2); 362–372,

1996.

Gianni, I., Vigano, L., and Lanzi, C., Role of daunosamine and hydroxyacetyl side chain in

reaction with iron and lipid peroxidation by anthracyclins, J. Natl. Cancer Inst., 80; 1104–1111,

1988.

Myers, C. E., McGuire, W. P., and Liss, R. H., Adriamycin: the role of lipid peroxidation in car-

diac toxicity and tumor response, Science, 197; 165–167, 1977.

Weiss, G., Loyevsky, M., and Gordeuk, V.R., Dexrazoxane (ICRF-187), Gen. Pharmacol., 32(1);

155–158, 1999.

Della Torre P., Imondi A, R., Bernardi, C., Podesta, A., Moneta, D., Riﬂ ettuto, M., and Mazue,

G., Cardioprotection by dexrazoxane in rats treated with doxorubicin and paclitaxel, Cancer

Chemother. Pharmacol., 44(2); 138–142, 1999.

Indian Herbal Pharmacopoeia, joint publication of Indian Drugs Manufacturers’ Association

and Regional Research Laboratory, Jammu-Tawi India, 1998, pp. 165–173.

Mishra, L. C., Singh, B. B., and Dagenais, S., Scientiﬁ c basis for the therapeutic use of Withania

somnifera (ashwagandha): a review, Altern. Med. Rev., 5(4); 334–346, 2000.

Davis, L., and Kuttan, G., Effect of Withania somnifera on cytokine production in normal and

cyclophosphamide treated mice, Immunopharmcol. Immunotoxicol., 21(4); 695–703, 1999.

Davis, L., and Kuttan, G., Suppressive effect of cyclophosphamide-induced toxicity by Withania

somnifera extract in mice, J. Ethnopharmcol., 62(3); 209–214, 1998.

Dhuley, J. N., Effect of ashwagandha on lipid peroxidation in stress-induced animals. J. Ethno-

pharmacol., 60(2); 173–178, 1998.

Bhattacharya, A., Ramanathan, M., Ghosal, S., and Bhattacharya, S. K., Anti-stress activity of

sitoindosides VII and VIII, new acylsterylglucosides from Withania somnifera, Phytother. Res.,

1; 32–37, 1987.

Mathew, S., and Kuttan, G., Antioxidant activity of Tinospora cordifolia and its usefulness in the

amelioration of cyclophosphamide activity, J. Exp. Clin. Cancer Res., 16(4); 407–411,1997.

Kapil, A., and Sharma, S., Immunopotentiating compounds from Tinospora cordifolia, J. Ethno-

pharmacol., 58(2); 89–95, 1997.

Patil, M., Patki, P., Kamath, H. V., and Patwardhan, B., Antistress activity of Tinospora cordifolia

(Wild) Miers, Indian Drugs, 34(4), 211, 1997.

Diallo-Sall, A., Niang-Ndiaye, M., Ndiaye, A. K., and Dieng, C. F., Hepato-protective effect of a

plant from the Senegalese pharmacopoeia: Tinospora bakis (Menispermaceae) using an in vitro

model, Dakar Med., 42(1); 15–18, 1997.

Dahanukar, S., and Thatte, U., Rasayana concept of ayurveda myth or reality: an experimental

study, Indian Pract., 245–252, 1988.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Thatte, U. M., and Dahanukar, S. A., Comparative study of immunomodulating activity of

Indian medicinal plants, lithium carbonate and glucan, Methods Find. Exp. Clin. Pharmacol.,

10; 639–644, 1988.

Dhuley, J. N., Effect of some Indian herbs on macrophage functions in ochratoxin A treated mice,

J. Ethnopharmacol., 58(1); 15–20, 1997.

Gautam, M., Diwanay, S. S., Gairola, S., Shinde, Y. S., Patki, P. S., and Patwardhan, B., Immu-

noadjuvant potential of Asparagus racemosus in experimental system, J. Ethnopharmacol., 91;

251–255, 2004.

Gautam, M., Diwanay, S. S., Gairola, S., Shinde, Y. S., Jadhav, S. S., and Patwardhan, B.,

Immune response modulation to DPT vaccine by aqueous extract of Withania somnifera in ex-

perimental system, Int. Immunopharmacol., 4, 841–849, 2004.

Diwanay, S., Chitre, D., and Patwardhan, B., Immunoprotection by botanical drugs in cancer

chemotherapy, J. Ethnopharmacol., 90, 49–55; 2004.

Chang, W. C., Lin, Y. L., Lee, M. J., Shiow, S. J., and Wang, C. J., Inhibitory effect of crocetin

on benzo[a]pyrene genotoxicity and neoplastic transformation in C3H10T1/2 cells, Anticancer

Res., 16(6B); 3603–3608, 1996.

Nair, S. C., Panikar, K. R., and Parthod, R. K., Protective effect of crocetin on bladder toxicity

induced by cyclophosphamide, Cancer Biother., 8(4); 339–343, 1993.

Mitra, S. K., Gopumadhavan, S., Hemavathi, T. S., Muralidhar, T. S., and Venkataranganna,

M. V., Protective effect of UL-409, a herbal formulation against physical and chemical factor in-

duced gastric and duodenal ulcers in experimental animals, J. Ethnopharmacol., 52(3); 165–169,

1996.

Bishayee, A., and Chatterjee, M., Anticarcinogenic biological response of Mikania cordata:

reﬂ ections in hepatic biotransformation systems, Cancer Lett., 30; 81(2); 193–200, 1994.

Mandal, P. K., Bishayee, A., and Chatterjee, M., Stimulation of hepatic protein synthesis in

response to Mikania cordata root extract in carbon tetrachloride–induced hepatotoxicity in mice,

Ital. J. Biochem., 41(6); 345–351,1992.

Praveenkumar, V., Kuttan, R., and Kuttan, G., Chemoprotective action of rasayanas against

cyclophosphamide toxicity, Tumori, 80; 306–308, 1994.

Jena G. B., Nemmani, K. V., Kaul, C. L., and Ramarao, P., Protective effect of a polyherbal for-

mulation (Immu-21) against cyclophosphamide-induced mutagenicity in mice, Phytother. Res.,

17(4); 306–310, 2003.

Weiss, J. F., and Landauer, M. R., Protection against ionizing radiation by antioxidant nutrients

and phytochemicals, Toxicology, 189(1–2); 1–20, 2003.

Lammering, G., Anti-epidermal growth factor receptor strategies to enhance radiation action,

Curr. Med. Chem. Anti-cancer Agents, 3(5); 327–333, 2003.

Marini, P., and Belka, C., New strategies for combined treatment with ionizing radiation, Curr.

Med. Chem. Anti-cancer Agents, 3(5); 334–342, 2003.

Petersen, C., Baumann, M., and Petersen, S., New targets for the modulation of radiation

response: selective inhibition of the enzyme cyclooxygenase 2, Curr. Med. Chem. Anti-cancer

Agents, 3(5); 354–359, 2003.

Devi, P. U., Withania somnifera Dunal (ashwagandha): potential plant source of a promising drug

for cancer chemotherapy and radiosensitization, Indian J. Exp. Biol., 34(10); 927–932, 1996.

Sharada, A. C., Soloman, F. E., Devi, P. U., Udupa, N., and Srinivasan, K. K., Antitumor and

radiosensitizing effects of withaferin A on mouse Ehrlich ascites carcinoma in vivo, Acta Oncol.,

35(1); 95–100, 1996.

Devi, P. U., Sharada, A. C., and Soloman, F. E., Antitumor and radiosensitizing effects of Witha-

nia somnifera (ashwagandha) on a transplantable mouse tumor, Sarcoma-180, Indian J. Exp.

Biol., 31(7); 607–611, 1993.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

REFERENCES 423

424 BOTANICAL IMMUNOMODULATORS AND CHEMOPROTECTANTS IN CANCER THERAPY

Asthana, R., and Raina, M. K., Pharmacology of Withania somnifera (Linn) Dunal: a review,

Indian Drugs, 26(5); 199–205, 1989.

Indap, M. A., Ambaye, R. Y., and Gokhale, S. V., Anti-tumor and pharmacological effects of the

oil from Semicarpus anacardium Linn. F., Indian J. Physiol. Pharmacol., 27(2); 83–91, 1983.

Indap, M. A., Ambaye, R. Y., and Gokhale, S. V., Potentiation of activity of anti-cancer drugs by

acetylated oil of Semicarpus anacardium Linn. F. in experimental tumors, Indian Drugs, 23(8);

447–451, 1986.

Mathew, S., and Kuttan, G., Antioxidant activity of Tinospora cordifolia and its usefulness in the

amelioration of cyclophosphamide activity, J. Exp. Clin. Cancer Res., 16(4); 407–411, 1997.

Shao, Y., Chin, C. K., Ho, C. T., Ma, W., Garrison, S. A., and Huang, M. T., Anti-tumor activity

of the crude saponins obtained from asparagus, Cancer Lett., 104(1); 31–36, 1996.

83.

84.

85.

86.

87.

425

Drug Discovery and Development, Volume 1: Drug Discovery, Edited by Mukund S. Chorghade

INDEX

Abbott Laboratories, 3, 5–6, 13–15, 233

Ab initio:

calculations, 20–21, 23

fold prediction, 247

geometry optimization, 51

implications of, 271

quantum mechanics, 50

Abnormalities, hereditary, 40

Absorption:

ADMET, see Absorption, distribution,

metabolism, elimination, and toxicity

(ADMET) studies

antimigraine drugs, 202

bisphosphonates, 207

enhancement strategies, 124–126

signiﬁ cance of, 108–109, 111

Absorption, distribution, metabolism, elimination,

and toxicity (ADMET) studies:

absorption assurance, 57–58

biomedical polymers, 384

characterized, 110, 136, 214, 235

deﬁ ned, 2

directing distribution, 58–59

drug discovery and development process, 18, 27,

33, 61, 79–81

drug handling proﬁ les, 41

efﬁ cacy, 73–74, 84

elimination, optimization of, 65

future directions for, 79

handle, 176

herbal remedies, 59–62

-HTS systems, 33, 35–37

metabolism controls, 63–65

molecular conformation, 55

parameter pharmacophores, 74

proﬁ ling, 18, 39, 41, 69–70, 84

proteomics, 259

SARs, 49, 67–70, 80

structural motifs/patterns, 39, 67, 79, 84

synergy, 61–62, 79

toxicity, avoidance strategies, 65–67, 262

Absorption high-throughput screening (AHTS)

systems, 57–58

Academic laboratories, 133

Accelerator mass spectrometry, 374

Accelrys, Ltd., 276

Acceptors, H-bond, 140

Accessibility, signiﬁ cance of, 109

Accession number, 254

ACE (α-chloroethyloxycarbonyl), 325

Acetophenones, 176

Acetylation, 5, 275

Acetylcholine, 296–297, 313

Acetylcholinesterase (AchE):

inhibitors, 224–225

deﬁ ned, 371

Acetylenes, 189

Achyranthes bidentata, 408

426 INDEX

Acid(s), 3, 111, 185. See also speciﬁ c types of acids

Acid-base, implications of, 134, 385

Acquired immunodeﬁ ciency syndrome (AIDS), 29

Acrolein, 412

Acromegaly, 181

Acronyms, 19, 22

ACS Software, 48, 281

Actimmune, 30

Activase, 29

Active pharmaceutical ingredient (API), 359

Active site docking, 57

Active transport, 111–112, 139, 394

Activity-based probes, 261

Acute inﬂ ammatory process, 407

Acute peptic ulcer, 296

Acute toxicity, 107, 109, 413

Acylation/acylating agents, 152, 175

Acyl groups, 172, 415

Acylguanidines, 179

Adaptogens:

botanicals, 407–408

deﬁ ned, 407

Rasayana botanicals, 408–409

Adenine, 105

Adenosine, 172

Adenovirus, 372

ADEPT, 146

A/DHTS, 58

Adipose tissue, 58

Adjustive medicine, 407–409

ADME, see Absorption, distribution, metabolism,

and excretion (ADME)

Administrative costs, 3

Adoptive immunotherapy, 370, 372–373

Adrenaline, 218, 296

Advanced Chem Tech, 24

Adverse effects, see speciﬁ c drugs

Afferent, 146

Afﬁ nity, 76, 150

Agendia, 263

Agglomerative hierarchical clustering method, 139

Aggressiveness, 370

Agranulocytosis, 220

AIDD, 146

Airway inﬂ ammatory diseases, 314

Alacepril, 127

Alanine, 124, 127

Alanine aminotransferase, 415

Albumen, 114

Alchemy III, 48, 281

Alcohol(s), 177, 186–187, 189, 191, 348–349

Alcoholysis, 189

Aldehyde dehydrogenase inhibitors, 413

Aldehydes, 178, 188, 275

Aldose reductase inhibitor, 183

Aldosterone, 119, 372

Alendronate sodium, 208–209

Alfentanil, 341, 343

Alferon N, 30

Aliphatic morpholine, 225

Alkaline phosphatase, 132, 369, 412, 419

Alkaloids, 223, 415, 418

Alkylating agents, 411

Alkylation, 178, 180

Alkylators, 413

Alkyl bromide, 178, 186

Alkyl groups, 357

Alkylguanine-DNA alkyltransferase (AGT), 370

Alkyl hydrazines, 191

Alkyl nitriles, 187

Allergies/allergic reaction, 111

Allograft rejection, 181

Allosteric protein, 246

Allyl amine polymers, 394

17-Allylaminogeldanamycin, 373

α-Azidoalcohols, 187

α-Bromoketones, 187

α-haloketone, 107

α-Ketothiazole, 186

Altana Pharma, 233

Aluminum salts, 386

Alveolar macrophages, 409

Alzheimer’s disease, 10, 31, 369–369, 152, 371

AMBER parameters, 20, 50, 282

American Chemistry Council Long-Range Research

Initiative, 66

American Heart Association (AHA), 315

American Hospital Supply, 208

Amersham, 242

Amgen, 29–30

Amide nitrogen, 357

Amidines, 301–303

Amidinium cation, 302

Amifostine, 412–413

Amikacim-resistant Klebsiella pneumoniae, 400

Amine(s):

acylation of, 141

characterized, 151–152, 189

end-product, 71

parallel solution-phase synthesis, 170, 178, 184

polymers, 387–388

quaternary, 388

secondary, 185–186, 188

synthesis of, 191

tags, 133

tertiary, 111, 185

Amino acids, 31, 117–118, 123, 125, 132–133, 151,

242, 247, 318, 361, 398

Amino alcohols, 177

Aminobenzamides, 174

4-Amino-1-benzylpiperidines, 176

Aminobenzamides, 174

Amino groups, 125

Aminomethyl styrene, 134

INDEX 427

Aminopyridazine, 224

5-Aminotetrazole, 172

Aminothiol, 412

Amiodarone, 221

Amlodipine, 207

Ammonia, 178

Ammonium groups, 387, 390, 400

Ammonium salts, 344, 394

Amnesia, 340

Amnestics, 340

5'-AMP, 77

AMPA, 183

Anaferine, 415

Analgesics:

characterized, 339–340

remifentanil, 340–350

ultrashort, 339–350

Analog(s):

attrition, 108

deﬁ ned, 199

design of, see Analog design

direct, 214

drug, 202–208

early phase, 199–202

production, 104, 107

research studies, 199–209

Analog design:

emergence of new activities, 216–217

natural compound models, 216

pharmacophore-based, 215

terminology, 214–215

Analoging, see Analog(s)

pit viper venom illustration, 120

signiﬁ cance of, 116

Analytical chemistry:

characterized, 18–19, 27

examples of, 75–78

trends in, 74–75

Andrade-Gordon, Patricia, Dr., 315

Anemia, 29, 408

Anesthesia, 339–340. See also Remifentanil

Angicoagulants, 226

Angina/angina pectoralis, 116, 207, 217,

221

Angioplasty, 30

Angiotensin-converting enzyme (ACE):

characterized, 117–118, 120

inhibitors, 119, 121–123, 126, 199–200, 214,

222

peptide bond hydrolysis, 120

Angiotensin II receptors:

antagonist losartan, 215–216

characterized, 117–119, 128, 183

Anhydride, 185, 349

Anhydroerythromycin, 3, 6

4-Anilidopiperidine acid, 348

Aniline groups, 107, 178

Animal models/studies:

amistoﬁ ne, 413

analgesic opiates, 342–344

antiulcer drugs, 298–300, 305

beneﬁ ts of, 222, 234, 261–262

clarithromycin, 6

Clostridium difﬁ cile toxins, 400

discovery route, 104

drug-target interaction, 372

erythromycin, 6–7

insulin-resistant, 200

in vivo diuretic effects, 324

nuclear imaging, 370

parallel solution-phase synthesis, 170

vasopressin receptor agonists, 316–319, 321, 327,

330–332

Anion-exchange resins, 398

Anionic polymers, 398–399

Annexin V, 376

Anorexia, HIV-related, 220

Anthracyclines, 414

Anthrax toxin, 397

9-Anthrylmethyl ester, 186

Antiadrenergic drugs, 300

Antiallergic agents, 172

Antiarrythmics, 342

Antibacterial agents, 223, 310, 400

Antibiotics:

antitumor, 180

azithromycin, 6, 9–12

characterized, 226, 375, 399

clarithromycin, 6–8, 10, 12

erythromycin, 3–8, 10, 12, 182, 226

resistance to, 112

tetracycline, 228

treatment, 397–398, 401

ulcer treatment, 310

Antibodies, 253, 368

Antibody/epitope, posttranslational modiﬁ cations,

244–245

Anticancer agents:

camptothecin (CPT), 43–44, 59

characterized, 13

diadzein, 46

genistein, 46

methotrexate, 105–106, 116, 410

multidrug resistance (MDR), 42–45, 59

paclitaxel (PAC), 42–43, 52, 59, 68–69, 82–83, 414

phytoalexins, 45–46

research study, ADMET SAR, 68

resistance to, 112

topotecan, 43–44

Anticholinergic drugs, 297, 300

Anticonvulsants, 314

Antidepressants, 14, 224

Antidiuresis, 316

Antidopaminergic properties, 203

428 INDEX

Antiemetic agents, 203

Antifungal:

assays, 175

therapy, 206

Antigens, tumor-speciﬁ c, 264

Antihepatotoxic agents, 415

Antihistamines, 205, 219, 221, 296–298, 374

Antihypertensives, terazosin, 13, 15

Anti-infective agents, 396–397

Anti-inﬂ ammatories, 414

Antimalarial drugs, 222

Antimetabolites:

arotinoid Ro 40-8757, 411

characterized, 300

ﬂ uorouracil (5-FU), 410–411

methotrexate, 410

nucleophilic sulfur thios, 411

Antimicrobial agents, 261, 400–401

Antimigraine drugs, 202

Antineoplastic effects, 413

Antioxidants, 414–415, 418

Antipsychotics, 217, 222

Antisense nucleotides, 30

Antistress agents, 414–415

Antithrombotic agents, 314–315

Antitrust litigation, 15

Antituberculosis drugs, 221

Antitumor:

agents, 105, 265, 405, 412, 414

antibiotics, 180

Antiulcer drugs, search for, 298. See also Tagamet

Anxiety disorders, 369

API (active pharmaceutical ingredient), 2, 13

Apomorphine, 76

Apoptosis, 417

Appetite stimulants, 220

Applied sciences, 17–18, 36

Aprepitant, 375

Aptamers, 252

Aquaporin-2, 316

Aralkyl-substituted 4-hydroxypyrone, 226

Areﬂ exia, 385

Arginine, 105

Arginine vasopression (AVP), 315–316

Argonaut Technologies, 24

Aromatic nitro groups, 107

Aromatic rings, 114, 124

ArQule, 172–173

Array technologies, 67

Arrhythmias, 207, 221

Artherosclerosis, 392

Aryl:

boronic acids, 183

ethers, 188

nitriles, 187

oximes, 187

3-Arylindoles, 183

Arylisocyanates, 190

Aryloxypropanolamine template, classical, 63

ASAT, 416

Ashwagandha Rasayana, 417

Asp, 78

Asparagine, 32

Asparagus crude saponins (ACSs), 419

Asparagus racemosus, 408–409, 416–417, 419

Aspartate aminotransferase, 415

Aspartic:

acids, 32, 361

proteases, 141

Aspergillus niger WB2346, 228

Aspirin, 202, 417

Asthma, 207, 314

AstraZeneca, 207–208, 233

antagonists, 200

Atenolol, 208

Atorvastatin, 202–203

ATPase inhibitors, 179, 310

Attrition rate, 256

Auto Desk, 48, 281

Autoimmune diseases/disorders, 181, 249

Automated Chemistry Environment (ACE), 147–148

Automated synthesizers, combinatorial library

synthesis, 141, 143–144

Automated tag reader/sorter, 143

Autonomic nervous system, 296

Auxophores, deﬁ ned, 107

Avidin, 245

Avonex, 30

Ayurveda/Ayurveda remedies, 406–407

Azatadine, 205

Azepinoindoles, 319, 321–322

Azides, parallel solution-phase synthesis, 190

4-Azidobenzoic acid, 190

Azithromycin, 6, 9–12

Azmacort, 374

Azole derivatives, 204

Bacillus:

anthracis, 397

cereus, 397

Backup compounds, 80

Bacteria, see speciﬁ c types of bacteria

gram, 180

gram-negative, 397

inhibition of, 397

multidrug-resistant, 400

Bacterial:

infections, 400

protein adhesins, 400

Bait, posttranslational modiﬁ cation, 245

Ball & Stick, 48, 281

Barton’s base, 187

Bases, resin-bound, 185. See also Acid-bases

Basis set, 20

INDEX 429

Baxter/Genetics Inst., 29

Baycol, 392

Bayer, 29, 207

B cells, 408

BCR-ABL tyrosine kinase, 262

BCUT descriptors, 138

Beckman rearrangement process, 9

Behavioral and psychological symptoms of dementia

(BPSD), 370

Benazepril, 127

BeneFIX, 29

Bengamide, 266

Benign prostatic hyperplasia (BPH), 13, 15

Benzamidobenzoyl, 319

Benzbromarone, 222

Benzenesulfonamides, 170

Benzenesulfonyl azide, 178

Benzimidazolones, 178

Benziodarone, 221–222

Benzodiazepine(s):

characterized, 319, 324–332

libraries, 132

receptors, 175–176, 215, 227, 369

Benzodiazepinone, 355–356

Benzofurans, 178

Benzoic acid, 185

Benzomorphanes, 216

Benzothiophene, 150

Benzoyl:

chlorides, 170, 176

group, 347, 349

Benzylamine chiral auxiliary synthetic reagents,

71–74

Benzyloxycarbonyl groups (Z-groups),

erythromycin, 7

Benzyl-protecting groups, 82

Benzylsuccinic acid, 121

BE 10988, 183

Berlex, 30

β-acetoxyoxetane system, 52, 82

β-adrenergic blockers, 207–208

β-adrenergic receptor blockade, 63

β-blocking agents, 221

β-C-Mannosides, 181

β-diketones, 176

-selective antagonists, 207

Betaseron, 30

β-thalassaemia, 390

Betaxolol, 208

Biacore, 249–250

Bicarbonate, 387

Bicyclo[2,2,2] octanes, 187

Biginelli condensation, microwave-assisted, 178

Bile acids:

characterized, 387, 392–393

sequestrants (BASs), 393–396

Binary encoding system, 133

Binding, 76

Bins, combinatorial library design, 139

Bioactive polymers, 401

Bioassays, 213

b ions, 241–242

Bioavailability, 179–180, 202, 220, 223, 226–227,

317, 330, 356, 372

BioCarta, 271

Biocatalysts, 154

Biochemical:

assays, 25, 27

manipulation, 81

pathways, 374

sciences, 81

Biochemistries, 67

Biocombinatorial expression, 46

Biodegradation, soft drugs, 21

BioEdit, 270

Bioengineering techniques, 26, 79, 379

Biogen, 30

Biogenic amines, 296

Bioinformatics, 20, 40, 46–47, 81, 253–257, 284

Bioisoelectronics, 124

Bioisosteric exchange, 121

Bioisostery, 20, 215, 221

Biological assays, 191, 214

Biological environments, impact of, 50–52

Biological libraries, 237

Biological-related sciences, 81

Biological target, 135

Biomarker(s):

discovery techniques, 265

signiﬁ cance of, 252, 256, 263–265

Biomechanistic systems, 79

Biomedical polymers, 383–384

Biometabolism studies, 343

Biomolecules, 58, 367

Biopolymers, 130

Biosynthesis, 105, 372

Biosynthetic pathways, 217–218

BIOTECH, 79

Bio-Tech Gen., 29

Biotechnology:

advances in, 28, 35–36, 41, 65, 67, 79, 81, 84

-derived therapeutic agents, 70

Biotin/streptavidin, 252

Biotropin, 29

Bipyridine, 391

Bis-acid chlorides, 191

Bis-aryl sulfonamide, 150

Bis-isopropylamino compound, 152

Bisoprolol, 208

Bisphosphonates, 207

Black, James, Sir, 299

Bladder toxicity, 416

BLAST, 244, 269

Bleomycin, 413

430 INDEX

Blood:

clots, prevention of, 30

disorders, 30

-level problems, 8

molecular conformation studies, 49

Blood-brain barrier (BBB), 43, 57, 115, 139, 218–219,

222, 356

Blood cell/marrow transplants, 30

Blood pressure regulation, drug development

example:

absorption, enhancement of, 124–126

analoging studies, pit viper-inspired peptides,

120

angiotensin-converting enzyme, inhibition of, 119,

121–123, 126

clinical SAR, 126–127

competition, 123–124, 126–127

historical background, 116

peptides, 120

peptidomimetics, 120–121, 128

renin-angiotensin-aldosterone system, 117–119,

123, 128

renin inhibition, 119

snake venom illustration, 117, 119–120

Blue-chip drug, 61

BMS:

181101, 371

182874, 223–224

192548, 228

193884, 224

207940, 224

Boechringer-Ingelheim/Nes Rx, 30, 201

Boehringer-Mannheim, 30, 208

Bohdan Automation, 24

Boiling point, 72

Bone density, 386

Bone marrow:

chemoprotectants for, 413

GM-CFCs, 414

suppression, 411

transplant, 30

Bopindolol, 208

Botanical(s):

antitumor agents, 418–419

characterized, 405

immunomodulators, 405–407

radioprotectants from, 418

Botany, 45

Bothrops jararaca, 117, 119

Bottlenecks, sources of, 26–27, 34, 154

Bradycardia, 385

Bradykinin, 118–119

Brahma Rasayana, 417

Brain:

adrenocortex, 372

blood-brain barrier, 43, 57, 115, 139, 218–219, 222,

356

blood pressure regulation, 118

drug distribution, 114

neurochemistry, 369

positron emission tomography, 372

Breast cancer, 30, 42, 176, 263–265, 370

Bridged bicycle derivatives, 322–324

Brimadmide, 219–220

Bristol-Myers Squibb, 120, 200, 223

Bromide hydrobromide, 175

Bromohydrins, 188

Bronchodilators, 170

B3LYP, 20, 51

B3LYP/6–31G*, 282

Bulkyl aryl group, 64

Burimamide, 304–305, 307, 309–310

By-products, 3, 178, 185

CAChe, 48, 281

Caco cell lines, 57

Calcium:

channel blockers, 206–207, 215

characterized, 215, 387

intracellular, 316

oxide, 359–360

salts, 386

Calpains, 182

Cambridge Scientiﬁ c, 48, 281

Camptothecin (CPT), 43–44, 59

Cancer, see speciﬁ c types of cancer

cells, characteristics of, 105, 370

chemotherapy, see Chemotherapeutic agents

diagnostic agents, 30

metastasis, 181, 408

statin treatment, 10

Candesartan, 201

Candidate drugs, 74, 108

Candidiasis, 205

Cannabinoid CB1 receptors, 220–221

Captopril:

analogs, 122–123

characterized, 200, 214

competition for, 127

side effects, 124

Capture-release agents, 191

Carbamate(s):

azatadine, 205

characterized, 186

Carbodiimide reagents, 184–185, 188

Carbohydrates, 173, 244, 394, 407

Carbon, 105–106, 367, 409

Carbon dioxide, 145

Carbonyl, 215

1,1'-Carbonyldiimidazole (CDI), 178

Carboxamides, 187

Carboxamidines, 302

Carboxyalkanoyl-

L-proline, 200

Carboxy amide, 188

INDEX 431

5-Carboxamido-1-benzyl-(3-dimethylamino-

propyloxy-1H-pyrazoles, 179

Carboxybenzenesulfonyl chlorides, 170

Carboxylate, 121

Carboxyl groups, 121–122, 124–125

Carboxylic acids, 72, 133, 141, 172, 349

Carboxypeptidase, 120–121, 123

Carcinogenicity, 66

Cardiac arrhythmia, 116

Cardia imaging agent, 30

Cardioprotective antioxidants, 414

Cardioside/Cardifolioside A/Carodiol, 415

Cardiotonic substances, 217

Cardiotoxicity, 413–414

Cardiovascular disease/disorders, 116, 392

Carfentanil, 341, 343

Carotenoids, 416

Carponone, 186

Carvedilol, 208

CAS Registry, 10, 84

Catalase, 414

Catalysis, 185

Catalysts, solid-supported (SSCs), 178

Catalytic hydrogenolysis, 72

Catch and release method, 185–187

Catecholamines, 217

Catecholates, 391

Catechols, 392

Cathepsin D, 140, 152

Cathepsin G, 314

Cationic polymers, 393–394

CATS (chemically advanced template search), 215

CB1 receptor antagonists, 220

CC-1065/CC-1065 analogs, 180

CDD (conserved domain database), 271

CDD-0102, 31–32

cDNA, 256

CEA-scan, 30

Celecoxib, 409

Cell(s):

adhesion molecules, 179

cell-cell interactions, 369

cell-surface adhesion, 314

culture assay, 215

cycle regulatory networks, 245, 257

proliferation, 407

repair, 407

trafﬁ cking, 369

Cellobiose, 399

Cellular biochemistry, 120

Cellular functional assays, 318

Centacor, 30

Center for Disease Control, 353

Centers for Health Research, 66

Centocor, 29

Central nervous system (CNS):

active drug requirements, 140

azatadine analogs, 205

cannabinoid CB1 receptor, 220

characterized, 273–274, 316, 411

drug penetration into, 43

drugs for, 115

efﬁ cacy studies, 111–112

tetracycline and, 228

Ceredase-glucocerebrosidase, 29

Cerevastatin, 392

Cerezyme, 29

CERIUS, 48, 281

Chalcones, 173–174

Characterization, 239

Charybdis Technologies, 24

Chelators, metal, 390–392, 412

Chemical:

approaches, absorption process, 65

etching, 155

genomic strategies, 40

hurdles, 82

kinetics, 82

libraries, 237, 313

modiﬁ cations, 80

reactivity, toxicology and, 67

structure, search engines, 57. See also

Structure-activity relationship (SAR)

Chemical Industry Institute of Toxicology (CIIT),

Chemiluminescent nitrogen detection (CLND), 145

Chemoinformatics, 20, 46–49, 79, 81, 146, 280, 284

Chemoprotectants/chemoprotectants:

for antimetabolites, 410– 411

for anthracyclines, 414

botanical immunomodulators, 414–417

characterized, 405, 419

drug targets, 409

goal of, 409

immunomodulation and, 409–410

thiol-based, 411–414

Chemosensitizer drugs, 42, 44

Chemotherapeutic agents, see also speciﬁ c

medications

characterized, 9, 235, 405–406

NCE, 40

P-glycoprotein pump (Pgp), 42, 44, 83, 151–152

smart bombs, 58

topotecan, 43

Chemotherapy:

anemia and, 29

beneﬁ ts of, generally, 377

blood disorder treatments, 30

for leprosy, 222

Chemotoxicity, 419

CHEM-X, 146

Cherry picking, 139

Cherwell Scientiﬁ c, 48, 281

Children, see Pediatrics

432 INDEX

Chiral:

auxiliary synthetic reagents, 71–74

pyrazolines, 221

Chiron Technologies, 24, 30

Chitobiose, 399

Chloride, 387

Chlorimipramine, 217

Chlorine, 215, 358

8-Chloro derivatives, 205

Chloromethylated polystyrene, 134

Chlorpromazine, 214, 221, 215, 217, 222

Cholebine, 396

Cholesterol:

levels, 202, 392–396

-lowering agents, 10, 392–396

metabolism pathway, 393–396

Cholestryramine, 3393, 98

Cholinergic:

agonists, 224

excitatory neurotransmission, 204

Chromafﬁ n tissue, 217

Chromatography:

applications, generally, 72

electron-capture gas, 133

ﬂ ash, 142, 145

ﬂ ash-column, 322

ﬂ uoroFlash chromatography, 190

high-performance liquid (HPLC), 45, 143, 145,

240–242, 257

ion-exchange, 300

liquid, 264

supercritical ﬂ uid (SFC), 245

Chronic granulomatous disease, 30

Chronic hepatitis C infections, 30

Chronic inﬂ ammatory process, 407

Chronic myeloid leukemia (CML), 262

Chronic renal failure and, 386–389

Chronic toxicity, 107, 109

Chymase, 314

Ciba-Geigy, 201

CICLOPS, 146

Ciglitazone, 200

Cilazepril, 127

Cimetidine, see Tagamet

characterized, 214, 219–220, 304

discovery of, 307–308

peptic ulcer disease treatment, 308–309

Cinnamic methyl ester, 152

Cirrhosis, 316

Cisapride, 203–205

Cisplatin, 411–412

CL-385004, 319

Clarithromycin, 6–8, 10, 12

Classical inhibitor studies, 27–28

Classical synergies, 60

Cleavage:

asymmetric, 72

characterized, 264

enalaprilat, 124

mass spectrometry analysis, 242

peptide bonds, 120–121

photolytic, 133

in screening combinatorial libraries, 132

Clinical candidate, deﬁ ned, 108

Clinical candidate drug:

cimetidine, 307–309

dynamic structure-activity analysis, 305–306

imidazole tautomerism, 306-

sulfur methylene isosterism, 306–307

thiourea, isoteres of, 307–308

Clinical efﬁ cacy, 109

Clinical investigations, 221–223

Clinical proof of principle, 28

Clinical SAR, 126–127

Clinical testing, 26

Clinical trials:

amifostine, 413

biomarker proﬁ les, 264

characterized, 2, 235–236, 262, 398–400

discovery process, 235–236

remifentanil, 349–350

Clodronate disodium, 208–209

cLogP, 191

Clopimozid, 215

Closed-shell molecules, 50

Clostridium:

difﬁ cil, 397–400

parvum, 401

perfringens, 397

Clots/clotting factors, 29

Clustal, 270

Cluster computing, 254

Clustering methods, combinatorial library design,

139

Cocaine, 216

Cocktail vaccine, 249

Colchicine, 203

Colesevelam hydrochloride, 394

Colestimide, 396

Colestipol, 393, 398

Colitis, 399

Collision-induced dissociation (CID), 241–242

Colon cancer, 176

Colorectal cancer, 413

Combinatorial biosynthesis, 45

Combinatorial chemistry:

case histories/illustrations, 149–154

combinatorial libraries, design of, 135–141

drug discovery, role in, 135–137, 149–154

future directions for, 154–155

historical perspectives, 130–131

information resources, 156

key developments in, 130

lead optimization, 136, 151, 267

INDEX 433

library screening strategies, 131–132

library synthesis, 141–146

management strategies, 146–148

overview of, 24, 26–28, 36, 39, 83, 129–130, 155,

214, 260

small molecule synthesis, 132–133, 135

solid-phase chemistry, 133–135

as standard tool, 148–149

trends in, 135

Combinatorial libraries:

characterized, 20, 26, 79

design of, 134–141

encoded, 132

indexed, 132

optimization, 136–137, 148

orthogonal, 132

screening strategies, 131–132

synthesis tools, 141–146

virtual, 137–138, 140

Combinatorial mixtures, 81

Combinatorial revolution, 129

Commercialization, impact of, 2

Compactin, 202

Comparative molecular ﬁ eld analysis (CoMFA), 20,

Competition, signiﬁ cance of, 123–124

Complementary medicine, 407–408

Co-morbidity, 374

Compound(s):

action potential (CAP), 412

attrition rate, 236

hits, attributes of, 80–81

libraries, see Compound libraries

overload, 27

proﬁ ling, virtual, 58

surveys, 54

Compound libraries:

characterized, 24, 33, 37, 39, 45–46, 69, 80–81,

108

virtual, 55–56, 79, 284

Comprestatin A4 phosphate, 376

CompuDrug, Inc., 276

Computational biology, 253–256

Computational chemistry, 260–261

Computational chemists, role of, 81, 214

Computational methods, 313

Computational structure determination, protein

structure analysis, 247

Computational studies, molecular conformation

assessment, 47, 50

Computational technologies, molecular

conformation studies, 53, 55–57

Computed tomography (CT), 367

Computer hardware, solid-phase synthesis, 134

Computer modeling, combinatorial chemistry, 135

Computer software programs:

Automated Chemistry Environment (ACE), 147

CAS Registry, 10

combinatorial library design, 137

combinatorial library speciﬁ cation, 146

public domain software tools and databases,

269–271

RADICAL, 146–147, 149

solid-phase synthesis, 134

spectrum prediction, 145–146

structure prediction, 247

Condensation:

implications of, 107

multicomponent, 177

polymers, 394

Conformational assessment, 282

Congestive heart failure (CHF), 179, 223, 316, 332

Conivaptan, 317

Contemporary drug discovery:

barriers to success, 104

cost of, 104

example of, 116–128

hit required criteria, 108–116

suitable lead substance, characteristics of,

104–108

Continuing education programs, 82–83

Controlling the size of the haystack, 55

Coomassie blue staining, 242

Copolymers, 400

Copper, 391

Core structure, combinatorial library design, 137

Corneal deposits, 221

Coronary dilators, 221–222

Coronary heart disease, 392

COR Therapeutics, Inc., 315

Corticotrophin releasing factor (CRF), 225, 316

Cortisol levels, 372, 416

Cost, drug development process, 109

Counterions, 72

Coupling:

characterized, 143, 187

combinatorial chemistry, 152

cycles, 131

COX-2 inhibitors, 10, 13–15

c(Pro-Tyr-D-Trp-Lys-Thre-Phe), 151

Creatine levels, 182

CRF antagonists, 225

Critical path, implications of, 317–318

Crocetin, 416

Cross-reactivity, 262

Crystal lattice, 49

Crystallization, 75, 79, 344, 347, 360

CSC Chem 3D/Chem 3D Plus, 48, 281

C3-convertase, 415

Curacin, 178

Curative treatment, 79

Curtius conversion, 82

Cyanide reagent, 347

Cyanoguanidines, 304, 308

434 INDEX

Cyanohydrin, 347

Cycle time, 154

Cyclic adenosine monophosphate (cAMP):

characterized, 75–78, 316, 372

phosphodiesterase enzymes, 75–78

Cyclic (depsi)peptides, 179

Cyclic phosphates, 75

Cycline-dependent kinase (cdk), 176, 179

Cyclin E inhibitors, 179

Cyclization:

intramolecular Wittig, 175

parallel solution-phase synthesis, 170, 172,

174–175

Cycloaddition reaction, 182, 188

Cyclopentadiene, 322

Cyclophosphamide, 412–413, 416

Cyclopropyl:

analogs, 76

characterized, 357

dopamine, 77

template, 82

Cyclopropylamine, 357, 359–360

CYP, 274–275

Cystein oxidation, 243

Cysteine proteinases, 182

Cystic ﬁ brosis, 29

Cytochrome P450, 51, 73, 283

Cytogen, 30

Cytokines, 369, 415, 418

Cytoprotection, 419

Cytosine, 105

Cytotoxicity, paclitaxel (PAC), 42

Dainippon, 204

Damstadt, 208

Data, generally:

-driven exploration, 253

fuzzy, 52, 56, 283

genomic, 47

handling, combinatorial library management, 146

integrity, 254

mining, 54, 254, 284

pk, 28

proﬁ ling, 41, 59–61, 67, 70, 79, 139

validity of, 234, 254, 256–259, 266

Database(s):

Ayurvedic, 407

combinatorial library design, 140

efﬁ cacy, 41

HTS-generated, 59

knowledge-generating structural, 36

management, 254

pharmacogenetic, 40

public domain, 269–271

2D structure-metabolism, 39

toxicology, 66

virtual, 55–56

DAVID, 270

DEAD, 188

Dealkylation, 356–357

Debenzylation, 72–73

Debrisoquine, 275

Decalin, 202

Decarboxylation, 218–219

DEC (1,2,-dichloroethane), 325

Dechlorination, catalytic, 358–359

Deconvolution, 131–132, 135, 150–151

Deep View, 270

Dehydration, 111, 185, 187

Delapril, 127

Dementia, 369–370

Density functional theory, 51

Deoxyuridine 5'-monophosphate, 105

Depression, 369, 372

Deprotection, 143, 185

Deprotonation, 187

DEREK, 66

Desferrioxamine, 390

Desloratadine, 205

Desmethyl-deoxy-minaprine, 225

Desmethylminaprine, 228

Dess-Martin reagent, 186

Detector, in mass spectrometry analysis, 241

Detoxiﬁ cation, cellular, 407

Deuterium labeling study, erythromycin, 3–4

Development costs, 136

Dexrazoxane, 414

Diabetes:

insulin-sensitive, 29

statin treatment, 10

Diabetic neuropathy, 29

Diacylglycerol, radiolabeled, 369

Diadzein, 46

Diagnoses, future events prediction process, 79

Dialysis, 386

Diamines, 151

Diastereoselectivity, 322

Diasteromers, 72

Diazepam, 215, 227

Diazepinones, 355

Diazo compounds, 178

Dichloroaminopyridine, 358

2,6-Dichloro-9-cyclopentylpurine, 176

Dichloropurine, 176

3,6-Dichloropyridazine, 177

Didanosine, 361

Didemethylated compounds, 175

Diels-Alder reaction, 137

Dietary phosphate, 387

Dietary supplements, see Nutraceuticals

Diethyldithiocarbamate (DDTC), 412–413

Diffraction studies, 75, 135

Diffractometer, 75

Diffusion, 111–112, 114

INDEX 435

Digitalis, 116

Diglyme, 360

Dihydrofolate reductase, 105–106

Dihydrofolic acid, 105–106

3,5-Dihydroxyheptanoic acid derivative, 202

Dihydroxyphenylalanine, 218

3,4-Dihydroquinoxalin-2-ones, 183

Diisopropylethylamine, 187

Dimers, 180

Dimethoxybenzoyl chlorides, 174

Dimethylformamide (DMF), 178, 360

2,5-Dimethylfuran, 146

Diovan, 215

DIP, 271

Dipeptides, 117, 121

Diphenhydramine, 298

Dipyrido compounds, 356

Dipyridodiazepinone structure, 357

Directing distribution, 58–59

Discovery chemists, 105

Discovery process:

case study, 265–266

clinical trials, 235–236, 263–265

efﬁ ciency of drug, 236

emerging technologies, 236–237

ﬁ nancial considerations, 25, 104, 236, 256

genomics, 237–238

high-throughput screening (HTS) approaches,

236–237

hits, screening for, 234, 259–261

hurdles in, 236

lead optimization, 234, 261–262

libraries, 80

pharmacology, 235, 262–263

proteomics, 238–248

pyramid, 109

target identiﬁ cation and validation, 234, 256–259,

266

time frame for, 236, 256

toxicology, 235

Discrete compounds, 135

Disease, see speciﬁ c diseases

biomarker proﬁ le, 264

model, 236–237

vs. normal state, 263–264

pathology, 259

Displacement reactions, 185

Distamycin, 178

Distribution, 18, 39, 109, 113

Disulﬁ ram (DSF), 413–414

Diuresis, 317

Diuretics, 116, 123, 324

Diversity, combinatorial library design,

137–139

Diversomer technology, 132, 143

Divide, 131

DMP-504 colestimide, 394

DNA:

alkylation, 411

array technology, 252

characteristics of, 105

duplex, 180

microarrays, 250

polymerases, 355–356, 360

synthesis, 419

toxicity and, 107–108

DNAse, 29

DOCK, 140

Docking interaction, 76, 79, 105, 121–122, 127, 283

Domain, deﬁ ned, 244

Donepezil, 371

Donors:

H-bond, 140

soft, 391

DOPA, 219

Dopamine:

biosynthesis and metabolism, 218

characterized, 218–219, 371

receptor antagonist, 203, 375

receptor ligands, 82

receptors, 75–78, 215, 370, 372

Dosage:

dose-dependent suppression, 226

dose-ranging studies, 25

dose-response curve, 319

dose-response relationship, 296, 375

serum protein binding, 114

toxicity and, 107

Double displacement reaction, 170

Doxorubicin, 414

Dreiding molecular mechanics, 47, 280

Dronabinol, 220

Drug(s):

administration, user-friendly, 79

deﬁ ned, 108

delivery system, 263

design components, 71, 260–261

efﬁ cacy, 264–265

interaction proﬁ le/interactions, 61–62, 200, 205

metabolism, see Drug metabolism

structural themes, 31

Drug-binding systems, 114

Drug-biological interface, 57

Drug-drug partnering, 62

Druglike/druglikeness:

deﬁ ned, 20, 139–140

implications of, 227

nondruglike distinguished from, 55

Drug metabolism:

biological environments and, 50–52

databases, see Drug metabolism databases

high-throughput screening studies, 284, 286, 288

implications of, 51–52, 61, 275

nuclear magnetic resonance studies, 282–284

436 INDEX

Drug metabolism: (Continued)

systematic studies, 278

x-ray studies, 282–284

Drug metabolism databases:

characterized, 273, 287–288

future directions for, 280, 282–287

historical perspective, 275–276

industry survey about, 278

informational ﬁ elds, 287 (ﬁ g.)

hurdles for, 286

metabolic and excretion capabilities, 274

metabolic capabilities displayed by human

phenotype, 275

present status of, 276, 278–280

xenobiotics, 286

Drug-receptor/active site docking, 57, 286

Drug-receptor interactions, 19

Drug-receptor theory, 105

Drugs of interest (DOIs), 374

Drug-target interaction, 372, 375

Duocarmycins, 180

Duodenal ulcers, 296, 307, 416

DuPont, 201, 208, 216

Duration of action, signiﬁ cance of, 13

Dynamic combinatorial libraries (DCLs), 154

Dynamic energy relationships, 52–54

Dynamic structure-activity analysis, 305–306

Dyspepsia, 310

Early phase analogs:

ACE inhibitors, 199–200

antimigrane drugs, 202

antagonists, 200

HMG-CoA reductase inhibitors, 202

insulin sensitizers, glitazones, 200, 202

proton pump inhibitors, 200

Echinacea purpurea, 408

Economic considerations, 13, 25, 104, 236, 256

Edema, 316–317, 332

Edisonian methods, 314

Edman sequencing, 241, 257

Efﬁ cacy:

deﬁ ned, 110

in drug development process, 109–110, 264–265

following oral administration, 110 –112

potency compared with, 110

pursuit of. 40–46

structural manipulation and, 71

suitable lead substances, 104–105

test/testing, 25, 79, 264

EGF-R tyrosine kinase, 373

Ehrlich ascites carcinoma, 418

Elderly metabolic capabilities, 275

Electrolyte balance, 385

Electromagnetic ﬁ eld separation, in mass

spectrometry analysis, 241

Electron-capture gas chromatography, 133

Electronic physicochemical properties, 63

Electronic topography, 75

Electrospray ionization (ESI), 241–242

Electrostatic:

interaction, 387, 398

potential, 20, 37, 70, 73, 77, 84

Eletriptan, 204

EL-4 leukemia, 413–414

Eli Lilly Company, 29, 299–300

Elimination structural relationship (ESR),

Emblica ofﬁ cinalis, 409, 417

Emergent activity, deﬁ ned, 214

Emerging technologies, 237

Enalapril, 200

Enalaprilat, 124–125, 127

Enantiomers, 13, 327

Endotheline ET

receptors, 223–224

Endothelins (ETs), 77–78

Endotoxins, 397

5-Endo-trig anti-Baldwin ring closure, 172

End-stage renal failure, 386

Entacapone, 219

Enterobactin, 391

Entrez:

SNP, 270

Taxonomy, 270

Enzymatic assays, 234

Enzyme(s), see speciﬁ c enzymes

exploitation of, 71

inhibitors, 300

kinetics, 361

-substrate interactions, 106

Epichlorohydrin cross-linked polyallylamine,

389

Epogen, 29

Epoxide, 73

Eprosatran, 201

Equilibrium, signiﬁ cance of, 51, 282

Erbitux, 373

Ergosterol, 206

Erythema nodosum leprosum, 222

Erythrocytes, 399

Erythromycin, 3–8, 10, 12, 182, 226

Erythropoiten-related growth factors, 29

Escherichia coli (E. coli), 252, 397

Esmolol, 63–65, 208, 342

Esterases, 65, 124, 341–342

Esters, 63–64, 124, 176

Estradiol, 369

Estrogen receptor modulation, 409

Ethanol, 417

Ether extracts, 408

Ethics, 84

Ethnopharmacology approach, 406–407

Ethyl acetate, 348

Ethyl 1-diazo-2-oxopropylphosphonate, 189

INDEX 437

Ethylenediaminetetraacetate razoxane (ICRF-187),

414

Ethylene glycol, 347

Ethyl ester, 124

5-(3-Ethyl-1,2,4-oxa-diazol-5-yl)-1,4,5,6-

tetrahydropyrimidine), see CDD-0102

Etidronate sodium, 208–209

European Chemicals Bureau (ECB), 66

European Organization for Research and Treatment,

377

Evaporation, 142, 144–145, 155

Evaporative light-scattering detection (ELSD), 145

Excess reagents, 131

Excretion, 18, 39, 65, 79, 109, 115, 118

Exotoxins, 397

Expasy-SwissProt, 271

Experientia, 5

Extant impurities, 3

Extended Hückel calculations, 20, 47

Fabrication techniques, combinatorial chemistry, 155

Factorial design methods, 141

Fail fast drug candidates, 262

Famotidine, 219–220

Fast-second projects, 315

Fatty acids, 114

FDA, see U.S. Food and Drug Administration (FDA)

Fe(II), 391

Feedstock, biochemical, 42

Feline leukemia virus, 356

Felkin’s iridium catalyst, 186

Felodipine, 207

Fentanyl, 341–343

Fenticonazole, 206

Ferrer, 206

Fibrinogen receptor, 314

1536-well microplates, 21

Filtration, 131, 178, 360

Final lead compounds, 80

Fingerprinting, toxicology studies, 67

First-pass effect, 274

FKBP12 inhibitors, 180

Flash chromatography, 142, 145, 322

Flavanoids, 178, 188, 418

Flavones, 46, 175–176

Flavoproteins, 371

Fluconazole, 205–206

Fluidic synthesizers, 142

Fluorescence detection methods, 132, 155

Fluorescence polarization spectroscopy, 398–399

Fluorescent labeling, 252

Fluorocarbon ﬂ uid, insolubility of, 134

FluoroFlash chromatography, 190

2-Fluoronitrobenzenes, 178

4-Fluorophenyl, 202

Fluoroquinolone antibiotics, 375

Fluorouracil (5-FU), 410–411

Fluorous synthesis, 190

Fluvastatine, 202–203

Focused libraries, 37, 140

Folk medicine, 19

Food:

absorption, 111

poisoning, 397

Formulability, 109

Fosinopril, 127

Fractionation, 239, 251

Free energies, 51, 282

Free radicals:

characterized, 390, 412, 415, 419

quenching, 407

Front-line testing, 58, 79

Frovartriptan, 204

Fucosyltransferases, transition-state inhibitors,

181

Functional analogs, 215

Functional analysis, 245

Functional groups, 107, 139

Functional protein arrays:

advantages of, 250–251

disadvantages, 251

implementations of, 250, 252

underlying principles, 250

Functional protein microarrays, 262–263. See also

Functional protein arrays

Funding sources, public-sector, 33

Fungistatic agents, 205

Furan, 13, 220

Fusion tags, 250

Fuzzy data, 52, 56, 283

GABA receptors, 372

Gador and Henkel, 209

Gallbladder, 393

Gas-phase equilibrium, 51, 282

Gastric:

acid, 296–300, 303, 308–310

parietal cells, 200

ulcers, 220, 296, 409, 416–417

Gastrin, 297–298, 300

Gastroenterology, 310

Gastroesophageal reﬂ ux disease, 204

Gastrointestinal (GI):

disorders, 181

motility assays, 6

tract, 110, 118, 205, 384–386, 391, 393, 397, 401

Gastroprokinetic drugs, 204

Gaucher’s disease, 29

Gaussian 98:

applications, 20, 47, 51

MOPAC, 280

package, 282

Gene-based disease, 80

GeneCard, 269

438 INDEX

Gene expression, 237–238

Gene Logic, 66

Genentech, 29–30

Gene regulation, 370

Generic companies/drug manufacturers, 3, 15

Generic library, 140–141

Gene therapy, 80

Genetic disorders, 390

Genetics Inst., 29–30

Genetic toxicity, 38, 108

Genetropin, 29

Genistein, 46

Genital warts, 30

Genomic data, 47

Genomic Health, Oncotype DX diagnostic, 263

Genomics, 18, 20, 26, 28, 33–34, 38–39, 58, 67, 79,

83, 237–238, 313, 401

Genotoxicity, 228, 416

GenScan, 269

Gentili/Merck, 209

Genzyme Corporation, 29, 389

Gilson, 24

Ginseng abuse syndrome, 408

Glass/glassware, 155, 170

GlaxoSmithKline, 201–202, 204, 207, 310

Gleevec, 262–263

Glitazones, 200, 202

Glucose, 115, 371

Glucuronic acid, 114

Glutamate carboxyls, 105

Glutamine pyruvate transminase, 419

Glutathione peroxidase inhibition, 409

Glutathione S-transferases (GST), 416–417

Glycoproteins, 244, 314

Glycosides, 400

Glycosylation, 243, 264

Gonal-F serono, 29

Gram bacteria, growth inhibition, 180

Gram-negative bacteria, 397

Granulocyte macrophage colony stimulating factor

(GM-CSF):

characterized, 30, 415

progenitor cells, 414

Granulocytopenia, 307–308

Graph theory, 283

Green chemistry, 70–71

GRID, 140

Growth hormones, 29

GT, 399

Guanidine, 302, 308

Guanidine isostere, 304

Guanidinium salts, 386–387

Guanidino, 220

Guanine, 105

Guanylate cyclase, 179

Guanylhistamine, 219

Guilt by association, 245

H, 170

Half-life, signiﬁ cance of, 13, 238

Haloaromatic tags, 133

Halogen, 114

Halogenation, 187

Haloperidol, 215

Hammett equation, 309

Hässle, 207–208

Hay fever, 298

Head and neck cancer, 413

Health policy, 80

Heart:

arrhythmia, 116

cardiovascular disease/disordres, 116, 392

congestive heart failure, 179, 223, 316, 332

coronary artery disease, 392

drug distribution in, 114–115

failure, 116

ischemic cardiomyopathy, 316

HEK-293 cells, 318

Helicobacter pylori, 310

Hematological toxicity, 413

Hemochromatosis, 390

Hemodialysis, 389

Hemodynamics, 316

Hemolysis, 114, 397

Hemophilias, 29

Hemopoietic agents, 414

Hepatic disease, 316

Hepatoxicity, 415

Herbal medicines/remedies, 21, 59–62

Herceptin, 30, 265

Heteroatoms, 77

Heterocycles, one-step construction of, 182

Heterocyclic statins, 202

Hexapeptides, 131–132

High blood pressure, see Hypertension

High-performance liquid chromatography (HPLC):

applications, 240–242, 257

combinatorial library synthesis, 145

phytoalexins studies, 45

preparative, 143

High-throughput, see High-throughput screening

(HTS)

assays, 28

chemistry, commercial equipment, 142

protein interaction studies, 262

synthesis, combinatorial libraries, 145–146

technologies, 265

High-throughput screening (HTS):

absorption, see Absorption high-throughput

screening (AHTS)

ADMET proﬁ ling, 69–70

alternatives to, 213–229

applications, 33, 108, 148, 155, 234, 236–237,

245–246, 248, 257, 260, 275, 300, 313–314,

355

INDEX 439

assay development, 41, 45

-combinatorial chemistry approach, 26–27

deﬁ ned, 20, 24

discovery process, 213

drug discovery and development, 79, 81

drug metabolism studies, 284, 286–288

efﬁ cacy screen, 55, 84

efﬁ cacy surveys, 27, 33, 35–37, 54–55, 64, 67, 81

future directions for, 79

intellectual property (IP), 83

microengineering, 41

multicomponent synergies, detection of, 62–63

process, 41–42, 81, 245–246

HIS, 78

Histamine:

biological approach to histamine antagonist,

299–300

characterized, 296–297, 301–302

receptors, 219, 298–299

Histaminergic H

antagonists, 214, 222

Hit(s):

biopharmaceutical quality of, 214

deﬁ ned, 108

drug development criteria, 108–116

generation of, 191

screening for, 234, 259–261

structure, 73–74

validation criteria, 213

HIV-1 RT (reverse transcriptase), 353–355, 357, 360

HIV-2:

RT, 361

SIV, 356

HL-60 cells, 419

HMG-CoA reductase inhibitors, 202, 214, 392–393

HOBt, 184–185

Hoffman-LaRoche, 30

Homcy, Charles, Dr., 315

Homeostasis, 18, 38, 385

Homologene, 270

Homologous proteins, 244, 261

Homology modeling, 135, 140, 247

Homophenylalanine esters, 127

Hormones:

estrogen, 409

fertility, 29

growth, 29

neurohormones, 218

steroidal, 214

Horner-Emmons reagent, parallel solution-phase

synthesis, 187

Horner-Wadsworth-Emmons oleﬁ nation, 189

Host defense mechanisms, 407

Host-guest:

compatibilities, 112

complex, 105

relationships, 116

Hot spots, 67

5-HT

, 204

5-HT

1B/1D

agonist, 202

3-(1H-tetrazol-5yl)-4(3H)-quinazolinones, 172

HTS efﬁ cacy process, 41–42

HTS/UHTS methodologies, 79

Humalog, 29

Human studies:

clinical trials, 349–350, 398–400

dihydrofolate reductase, 105–106

drug metabolism database, 51, 275

esterases, 63

genome, 21, 135

GI endothelial system, 58, 71

Human Genome Project, 135

Human immune deﬁ ciency virus type 1 (HIV-1),

353–354, 356

Human somatostain receptors (hSSTR1–5),151

Humatrope, 29

Humulin, 29

HUN-7293 analogs, 179

Hydantoin, 132

Hydrazides, 170

Hydrazine, 170, 185, 187

Hydrogen:

bond/bonding, 4–5, 20, 50, 54, 78, 112, 138, 140,

282, 302, 387

characterized, 114

groups, 357

Hydrogenation, 344

Hydrogenolysis, 7, 72–73, 349

Hydrolysis, 77, 114, 118, 120, 152, 344, 347–349

Hydrophilic drugs, 111

Hydrophilicity, 394, 397

Hydrophobicity, 20, 152, 394, 397

Hydroxamates, 391

Hydroxamic acid, 392

Hydroxy acid, 202

Hydroxy(ethyl)amine isotere, 152

Hydroxyl:

groups, 5–6

radicals, 390, 415

Hydroxylamines, 107, 174

Hydroxylase, 372

Hydroxylation, 43

Hydroxymethylglutaryl (HMG), CoA reductase, 202

HyperChem, 48, 281

Hypercholesterolemia, 202

Hyperkalemia, 385–386

Hyperparathyroidism, 386

Hyperphosphataemia, 386–387

Hypertension:

characterized, 116, 207, 217, 332

statin treatment, 10

terazosin, 13, 15

Hypnosis, 340

Hypnotics, 222, 339–340, 349

Hypoglycemic effect, 221

440 INDEX

Hyponatremia, 316

Hypotensives, 217

Hypothalamus, 315

Hypothesis-driven methodology, 253

Ibandronate sodium, 209

ICI Pharmaceuticals, 208, 298

ICRF-187, 414

IDEC, 30

Identiﬁ cation, in proteomics, 239

Idiosyncratic problems, in drug development

process, 109

Ifosfamide (IFX), 412–413

Imidapril, 127

Imidazole(s), 152, 303, 306, 309–310

Imidazole-5-acetic acid derivatives, 200

Imidazolic ring, 220

Imidazoline receptor, 372

Imidazolinones, 182

Imidazolone system, 75, 77

Imidazoylalkylcarboxamidines, 301

Imidazoylalkylguanidines, 301

Imidazolylalkylisothioureas, 301

Imidazolylalkythioureas, 303

3-Imidazolylcarbolines, 181

Imidazopyridazines, 176

Imidazo[2,1-b]quinozolin-5-(3H), 170

Iminium salts, 174

Imino ethers, 172

Imipramine, 214, 217

Immune systems, 407

Immunex, 30

Immunity, cell-mediated, 408

Immunization techniques, 252

Immunodulatory agents, 415

Immunology, 245, 414

Immunomedics, 30

Immunomodulators, 222, 406–407, 409, 414, 416,

418

Immunoprecipitation, 239240

Immunoprotection, 416

Immunostimulation, 407

Immunosuppression, 370, 408

Immunotherapy, 370

Immunotoxicity, 413

Indian Ayurveda, 406–407

Indian Herbal Pharmacopoeia, 414

Indigenous herbal drug formulations, 417

Indoloazepine, 321

Indolobenzodiasepine derivatives, 319

IND phase, 28

Infection, 40, 46. See also speciﬁ c types of infections

Infectious disease, 414

Infergen, 30

Inﬂ ammatory diseases, 181

Inﬂ uenza neuramidinase (NA) inhibitors, 181

Inﬂ uenza virus neuraminidase inhibitors, 261

Informatics, 253, 257. See also Bioinformatics

Information technology (IT), 46–47

Inhibitor(s), see speciﬁ c types of inhibitors

A-192558, 181

characterized, 105–106, 111, 260

Injectable medications, 113

Insulin:

characterized, 29

mimetics, 213

sensitizers, 200, 202

Integrins, 314

Intellectual property (IP), 18, 83–84, 115

Interdisciplinary research, 82–83

InterDom, 271

Interferon-∀, 30

Interferon-∃, 30

Interferon gamma (INF-(γ)), 30, 415

Interferon Sci, 30

Interleukins:

characterized, 30

Interleukin-1 (IL-1), 416–417

Interleukin-2 (IL-2), 415

International pharmaceutical community, 26

International Program on Chemical Safety/

Organization for Economic Cooperation

(IPCS/OECD), 66

International Toxicology Information Center (ITIC),

International Uniform Chemical Database

(IUCLID), 66

InterPro, 271

Intracellular localization, 67

Intravenous administration, 6, 13

Intrinsic activity, 76

Intron, 30

Inventorship, 19

Invitrogen, 242

Iodohydrins, 188

Iodonium diacetate, 186

Ion-exchange:

characterized, 182, 185, 187–188

chromatography, 300

Ionization/ionization chamber, 111–112, 125, 241

Ioselectric point, 240

Iqbal multicomponent procedure, 177

Irbesartan, 201

Iron:

cellular, 414

chelating polymers, 392

overload disorders, 389–392

Irori AccuTag-100 system, 143

Irradiation, 188, 413

Ischemic cardiomyopathy, 316

ISIS/Ciba, 30

Isocyanates, immobilized, 185

Isoﬂ avones, 46, 416

Isoﬂ urance, 349

INDEX 441

Isoleucine, 361

Isomaltotriose, 399

Isomerization, 185–186

Isoniazide, 221

Isopelletierine, 415

Isopropyl acetate, 349

Isopropyl groups, 215

Isopropylsubstituents, 202

Isopycnic suspensions, 144

Isosteres, 20, 304

Isothiourea, 302

Isoxazolines, 174

Itraconazole, 206

IUPAC, 199

Janssen, Paul, 204, 206, 341

Janssen Pharmaceutica, 341

Japanese Kampo, 406

Jarvis-Patrick clustering, 139, 152

Jaundice, 222

Johnson & Johnson, 317

Kayexalate, 386

KEGG (Kyoto Encyclopedia of Genes & Genomes),

270

Kerolides, 182

Ketoconazole, 204–206

Ketones:

characterized, 177, 185, 187–188

inhibitors, 182

Kidney(s):

damage, 116

drug distribution, 115

ﬁ ltration, 117–119

functions of, 113

stones, 385

transporters in, 65

Kinases, 27, 137, 176, 179, 262, 361, 373

Klebsiella pneumoniae, 400

Knowledge-generating mining paradigms, 54,

284

Knowledge systems, 84–85

KoGENate, 29

KQAGD sequence, 314

Kramerixin, 178

Laboratory, see speciﬁ c corporations

conventional environment, 154

determination, in protein structure analysis,

246–247

information management systems (LIMS), 254

Laborit, 221

Lacidipine, 207

Lactate dehydrogenase, 415

Lactone, 202

LAF389 case study, 257, 265–266

Lansoprazole, 200–201

Large-scale stereoselective synthetic methodologies,

Laser microforming, 155

LC-MS, 19

LC-MS/MS, 19

LC-NMR, 19

LDH, 416

Lead (n):

compounds, 81, 191, 275

structures, 73–74, 80

Lead:

decision process, 58

deﬁ ned, 108, 213

discovery libraries, 135, 182

generation libraries, 136, 138

generation stage, 135

optimization, 136–137, 140, 148, 150–151, 234,

261–262, 267, 314

selection, 71

Lectin/carbohydrate, posttranslational modiﬁ cations,

244

Lectins, 407

Lederle Laboratories, 315

Leprosy, 222

Lercanidipine, 207

Leucine, 361

Leucocytes, destruction of, 397

Leucocytosis, 408

Leucopenia, 415–417

Leucotoxicity, 413

Leukemia, 419. See also speciﬁ c types of leukemia

Leukine, 30

Leukocyte function-associated

antigen-1/intracellular adhesion molecules-1

(LFA/ICAM-1), 181

Leukocytotoxicity, 413

Levodopa, 202, discovery of, 217–219

LHASA, Ltd., 66, 277

LHRH antagonists, 226–227

Library information management (LIMS),

Life sciences research, 79–83

Ligand(s):

-based drug design, 21, 79

characterized, 58

endothiline ET

receptors, 223

-metal complex, 391

polymeric, 390, 400

potent and selective, 106

protein interactions and, 245

radiolabeled, 367, 369

Lipid peroxidation (LPO), 414–415, 419

Lipinski rules of ﬁ ve, 112–113, 134, 140, 234, 261,

384

Lipophilicity, 138

Lipopolyacchardies, 397

Lipoprotein disorders, 202

442 INDEX

Liquid:

chromatography, 264

handling, 131, 142–143

-liquid extraction, 145

-phase combinatorial synthesis, 133–134

Lisinopril, 125–127, 200

Listeria monocytogenes 408

Lithium aluminum hydride, 190

Liver:

bile acid in, 393

cirrhosis, 316–317

drug distribution process, 113, 115

dysfunction, 392

enzyme monitoring, 202

jaundice, 222

metabolism, 114

transporters in, 65

Lixivaptan, 317, 319

LocusLink, 270

Losartan, 200–201

Lovastatin, 202–203, 216

Low-density lipoprotein cholesterol (LDLc), 389,

392

Low-energy molecular conformation, 50–51

Low-molecular-weight, 64

LUDI, 140

Lung cancer, 413

Lungs, drug distribution, 114–115, 118

Lymphokines, 408

Lysine, 126

Macrocylic:

oligomeric amines, 386

polyamines, 387

Macrolactone, 4

Macrolide, 4, 8–9

MacroModel 6.5 modeling package, 50, 282

Macromolecules, 105–106

Macular degeneration, 10

Magnetic resonance imaging (MRI), 367

Malaria, 116

Male erectile dysfunction, 217

Maltose, 399

Mammaprint, 263

Mmanagement strategies, in combinatorial chemistry:

automated workﬂ ow, 146–148

speciﬁ cations, 146

Man-made libraries, 79

Mannitol, 414

Manufacturing costs, 3. See also Economic

considerations

Mappicine, 178

Map Viewer, 270

Marijuana, 220

Marinol, 220

Marketing, 3, 136

Mascot, 270

Mass analyzer, in mass spectrometry analysis, 241

Mass spectrometry (MS):

applications, generally, 19, 241, 264

combinatorial library screening, 133

combinatorial library synthesis, 133, 145

molecular conformation studies, 50, 52

Mass-to-charge ratio (m/z), 241–242, 244

Matrix-assisted laser desorbtion ionization

(MALDI), 241, 264

Matrix-assisted laser desorbtion ionization-mass

spectrometry (MALDI-MS), 266

Matrix-assisted laser desorption ionization time-of-

ﬂ ight (MALDI/TOF) analysis, 242

Maximum tolerated dose (MTD), 235

MDL:

427, 170–171

Information Systems, Inc., 276

Toxicity Database, 66

Mechloroethamine, 411

Media:

nuclear magnetic resonance (NMR), 49, 51

parallel solution-phase synthesis, 172

Medical Research Centre of Brookhaven National

Laboratory, 219

Medicinal chemistry:

acronyms, 19, 22

ADMET studies, 18, 27, 33, 36, 57–70, 84

analog research, 199–209

analytical techniques, 18–19, 27, 74–78

applications, generally, 130, 303, 343–347

chemoinformatics, 18

contemporary, 103–128

current status of, 26–27

deﬁ ned, 35–36

drug design role, 23

drug discovery and development process, 25,

35–41, 79–80, 84–85, 104, 109

efﬁ cacy, pursuit of, 40–46, 84

evolution of, 17–18

as formalized discipline, 18–19, 23, 81–83

future directions for, 19, 78–81

historical perspectives, 23–26, 103–104

intellectual property (IP), 18, 83–84

knowledge vs. diversity paradox, 84–85

metabolism, 341–342

molecular conformation, 18, 23, 46–57

nutraceuticals, 18

overview of, 17–19

process chemistry, 70–74

roles of, immediate- and long-term, 36–40, 81–83

site-directed mutagenesis examples, 27–28, 31

synthetic chemistry, 18

terminology, 19–21

trends in, 31–35

Medicinal chemists:

continuing education, 82–83

curriculum requirements, 82

INDEX 443

drug development process, 109, 112–115, 134

formal training of, 23–24

future directions for, 84–85

graduate-level programs, 23–24

roles of, generally, 81–82, 103

Medicinal plants, 407

MEGA (molecular evolutionary genetics analysis), 270

Meisenheimer displacement, 178

Melanoma, malignant, 413

Melting point, 72

Mepyramine, 298

2-Mercaptoethane sulfonate mesna, 411

2-Mercaptoethanesulfonic acid, 412

Mercaptomethyl ketones, 182

Merck & Co., 124, 151, 200, 203–204, 208–209

Merck Index, 10

Mesna, 412

Metabol Expert, 39, 276–277

Metabolic pathways, 257

Metabolic reaction, 275

Metabolism:

in ADME, 18, 39

drug, 109, 114

protein interactions, 245

Metabolism Database, 277

Metabolite, 39, 276

Metabophores, 21, 37, 63–65, 279

Metal:

chelators, 390–392, 412

ions, 391

salts, 387

Metastases, 152, 181, 372, 408

Metathesis polymerization, 191

METEOR, 277

Metformin, 203

(Meth)acrylamide, 394

(Meth)acrylates, 394

Methanol, 348–349

Methanolysis, erythromycin studies, 7–8

Methicillin-resistant Staphylococcus aureus

(MRSA), 400

Methotrexate, 105–106, 116, 410

3-Methoxytyramine, 218

3-Methoxy-4-hydroxyphenyl-acetic acid, 218

Methyl:

α-phenylacetate, 64

benzoate, 64

ester, 349

groups, 124, 356–357

propionate, 63, 343

Methylation, 7, 265

Methyldopa, 203

Methylene groups, 105, 124

Methyl 4-phenylbutyrate, 64

Methylhistamine, 300, 307

Methylthioiminium salts, 174

Methyl 3-arylpropionate system, 63

Methyl 3-phenylpropionate, 63

Metiamide, 219–220, 307–309

Metoclopramide, 204

Metoprolol, 208

Metronidazole, 398

Mevastain, 202

Mevinolin, 214

Mibefradil, 215

Michael addition, 349

Miconazole, 206

Microarray(s):

analysis, 238

DNA, 250

gene expression, 254–255

protein, 248, 250–253, 262–263

technologies, 71

Microcalorimetry, 19, 27–28, 52

Microdosing, 374

Microengineering, 41

Microﬂ uidic devices, 155

Microreactors, 133, 143

Microtiter plates (MTPs), 131, 143–145

Microwave-assisted parallel synthesis (MAPS), 178

Microwave synthesizers, 142

Midazolam, 350

Migraines, 202, 314

Mikania cordata, 417

Millennium Pharmaceuticals, 315

Mimotope SynPhase system, 143

Minaprine, 224–225

Miniaturization, 154–155

Minimum laboratory concentrations (MICs), 7–8

Mistletoe lectin, 407–408

Mitomycin C, 413

Mitsunobu reagent, 188

Mix-and-sort synthesis, 142–143

Mix-and-sort systems, combinatorial libraries, 143

Mix-and-split process, 132

MKH-57, 174

MLS

resistance, 182

MNDO calculations, 21

MOBY, 48, 281

Moexipril, 127

Mofarotene, 409, 411

Molar ratio, 61

Molecular, generally:

conformation, see Molecular conformation

diversity, 45–46, 80, 84–85

imaging, 257

mechanics, 21

modeling, see Molecular modeling

orbital calculations, 20–21

oxygen, 389

scaffold systems, 43, 80, 84, 137, 152

sequestrants, polymeric 384

weight, signiﬁ cance of, 119, 124, 133, 138–139

Molecular biology, 26, 33, 310, 313

444 INDEX

Molecular Biology, 39

Molecular conformation:

ACE inhibition, 123

assessment and handling of, 46–57

implications of, 18, 23

similarity-dissimilarity indices, 57

Molecular diversity, drug efﬁ cacy research, 45–46

Molecular modeling:

applications, 4, 76

3D, 47–48, 280–281

Molecular Simulations, 48, 281

Molecular Toxicology Platform (Phase-1), 66

Monoamine-oxidase aldehyde-dehydrogenase, 218

Monoamine oxidases (MAOs), 371

Monoclonal antibodies, 30, 252, 367

MOPAC, 47

Morphinanes, 216

Morphine, 216

Morpholine, 225

Mosapride, 204–205

Motif, deﬁ ned, 244

Mouse studies, 249, 356, 370, 409–418

Mozavaptan, 317, 319

mRNA expression, 238

Mucomyst, 411

Multiagent, 39

MULTICASE, 66

Multicase, Inc., 276

Multicomponent:

reactions, 177–178

synergies, 62–63

Multiconformational assessment, 50

Multidrug exporter, 111

Multidrug resistance (MDR), 42–45, 59, 69, 83, 151

Multipins, 131–132

Multiple interactions, 60

Multiple sclerosis, 10, 30, 369

Multiple sequence alignment (MSA), 270

Multistep drug synthesis, 134, 144

Multivalent:

drug strategies, 39

ligands, 21

strategies, 79

Murine leukemia virus, 356

Murine xenograft models, 370

Muscarine, 296

Muscarinic M1 agonist, 224

Muscarinic receptors, 296, 368

Muscle relaxation, 340

Mutagenesis, site-directed, see Site-directed

mutagenesis (SDM)

Mutagenicity, 108–109

Mutations, 108

Myeloprotection, 416

Myelosuppression, 412–413

Myocardial function, 29, 385

MyoScint, 30

NAD synthetase inhibitors, 180

Nanomolar afﬁ nities, 225

Nanotechnology, 58, 65, 79

Nano Vision, 48, 281

Naratriptan, 204

Naringenin chalcone, 46

National Institute for Environmental Health Science

(NIEHS), 66

National Institute for Occupational Safety and

Health (NIOSH), Information Division, 66

National Institute of Neurological Diseases and

Strokes, 314

National Institutes of Health (NIH), 314

Natural iron chelators, 391

Natural killer (NK) cells, 408

Natural product(s):

characterized, 39, 191

libraries, 26

parallel solid-phase synthesis (PSPS), 186

scaffolds, 191

synthesis, 134

Nature, 131

NCBI Entrez, 269

Negative SAR, 41–42, 63–64

Nemesis, 48, 281

Neonates:

esterase capability, 65

neonatal metabolic capabilities, 275

Neoplastic transformation, 416

Nephrotoxicity, 220, 411–413

Nervous systems, 407. See also speciﬁ c nervous

systems

Neumega, 30

Neupogen, 30

Neural networks, 139

Neurodegenerative diseases/disorders, 182, 368–369

Neuroendocrine:

immune system, 407

tumors, 181

Neurohormones, 218

Neurokinin -1 (NK-1), 180

Neurokinin (NK) receptor antagonists, 183

Neuroleptics, 214–215, 221, 375

Neurological function, 385

Neurologic toxicity, 413

Neuromodulators, 407

Neuroregulators, 407

Neurotoxicity, 412–413

Neurotransmitters, 315, 369, 407

Neutral SAR, 41–42, 63–64

Neutropenia, 413

Neutropin, 29

Nevirapine

characterized, 354–355

chemical development and process research,

357–360

chemical structure, 354

INDEX 445

clinical studies, 361–362

lead discovery, 355–357

lead optimization, 356–357

mechanism of action, 360–361

medicinal chemistry approach, 360

New activity, emergence of, see Emergent activity

New chemical entities (NCEs):

deﬁ ned, 24

drug discovery and development process, 25–26,

61, 81, 84

patents, 33

production of, 55

New drug discovery paradigm, 41, 45, 58

Nicotinamide adenine dinucleotide, 417

Nicotine, 296

Nicotinic receptors, 296

Nifedipine, 206

96-well, generally:

arrays, 262

ﬁ lter apparatus, 154

microplates, 20, 143, 169, 260

MTPs, 155

Nitric oxide, 116

Nitriles, 187

Nitrite oxide, 369

2-Nitroanilines, 178

Nitrobenzamides, 322

2-Nitrobenzoyl chloride, 171–172

Nitrogen/nitrogen group, 7, 71, 310, 349, 367

Nitroglycerine, 116

Nitroguanidines, 304, 308

Nitroxides, 107

Nizatidine, 310

Non-Hodgkin’s lymphoma, 30

Nonnucleoside reverse transcriptase inhibitor

(NNRTI), 362

Non-small cell lung cancer, 413

Noradrenaline, 218, 296

Norditropin, 29

Norepinephrine, 313, 371

Notebook entries, 170

Novartis, 30, 201, 203, 208–209, 215–216, 265

Novelty, analog design, 215

Novolin/Novolin L/Novolin R, 29

Novo Nordisk, 29

Nuclear imaging, see Nuclear magnetic resonance

(NMR) applications

clinical studies, 373–376

drug development process and challenges of,

365–366

future directions for, 376–377

positron emission tomography (PET), 366–368

role in drug discovery, 368–376

technology, principles and evolution of, 366–368

Nuclear magnetic resonance (NMR) applications:

characterized, 2, 19, 246

combinatorial library synthesis, 145

drug metabolism studies, 282–284

erythromycin, 5, 10

molecular conformation assessment, 47, 49–55

spectroscopy, 170

Nucleophiles, 140, 185, 187, 191

Nucleophilic attack, 34

Nucleotide libraries, 24

Nutraceuticals, 18, 21

Observational studies, COX-2 inhibitors, 13

Ochratoxin-A, 416

Ocimum sanctum, 417

Octapeptide(s), angiotensin II, 117–119

Oleﬁ nic monomers, 400

Oleﬁ ns, 186, 399

Oligomers, ROMP, 191

Oligonucleosides, 130

Oligonucleotide(s):

characterized, 252

compound libraries, 28

solid-phase synthesis, 131

Oligopeptides, 130

Oligosaccharide(s):

characterized, 130

receptors, 399

synthesis, 143

Omeprazole, 200–201

3-O-methyl-L-dopa, 218

OMIM, 270

OncoScint, 30

One-bead, one-peptide approach, 132

OPC-41061, 317

OPC-31260, 316–317

OPC-21268, 317

Open reading frames (ORFs), 245–246, 250–251

Opioid(s):

analgesics, 341–344, 350

anesthetic agents, 350

ORL1 receptor agonists, 213

Optimization library, 138. See also Lead,

optimization

Oral absorption, 112–113, 124

Oral administration:

ADMET studies, 57–58

DOPA, 219

implications of, 107, 110–112

Lipinski rules of ﬁ ve, 112–113

Oral bioavailability, 227

Oral uptake, 113

Organ:

damage, 390

distribution, 114–115

Organic:

chemistry, 23, 81–82, 149

compounds, 28, 190

synthesis, multistep, 148

Orphan diseases, 228

446 INDEX

Ortho Biotech, 29–30

Orthoclone, 30

Orthoesters, 172

Ortho-McNeil, 29

Osmolarity, 317

Osmotic balance, 316

Osteoporosis, 10, 182, 207

Ovarian carcinoma, 413

Overexpression, 259, 261, 265

Oxadizoles, 170

Oxalic acid salt, 344

Oxazinobenzodiazepines, 324–332

Oxazoles, 182

Oxetane, molecular conformation study, 53, 57

Oxford Molecular, 48, 281

Oxiconazole, 206

Oxidation, 107, 114, 186, 243

Oxidative hydroxylation, 114

4-(1,3,4-Oxadiazol-2-yl)-N,N-

dialkylbenzenesulfonamides, 170

Oxyacid anions, 302

Oxyanions, 391

Oxygen:

characterized, 118, 124, 367

free radicals, 414

Oxytocin receptor, 318

Paclitaxel (PAC), 42–43, 52, 59, 68–69, 82–83, 414

Paget’s disease, 208

Palladium cross-coupling, 176

Palliative treatment, 41, 79

Pamidronate disodium, 208–209

Panax ginseng, 408

Pantoprazole, 200–201

Para-aminobenzoate, 106

Parallel array chemistry, 139

Parallel chemistry, 213–214

Parallel solution-phase synthesis (PSPS):

catalysts, 178, 184–185, 190–191

future directions for, 191

historical perspectives, 169–172

implications of, 154

recent reports, 172–178

scavengers, 178, 184–185, 190–191

semiautomated, 191

solid reagents, 178, 184–185, 190–191

Parallel synthesis, generally, 133–134, 150, 300, 314

Paralysis, 397

Parasympathetic nervous system, 297

Paresthesias, 385

Parkinson’s disease, 218–219, 371

Paromomycin, 401

PAR-1, 315

Partition coefﬁ cients, 57

Partitioning, bin-based, 139

PAS, 203

Passerini reaction, 177

Passive cutaneous anaphylaxis (PCA), 171

Passive diffusion, 111–112, 114

Patentability, 109, 227. See also Patents

Patent Cooperation Treaty, 9

Patents:

ACE inhibitors, 200

-adrenergic blocking agents, 208

angiotensin II antagonists, 201

antimigraine drugs, 204

application, submission of, 38

azatadine analogs, 205

azithromycin, 9–10

calcium channel blockers, 207

claims, 83–84

clodronate analogs, 209

gastroprokinetic drugs, 204

glitazones, 202

miconazole analogs, 206

omeprazole and analogs, 201

prazosin, 13

statins, 203

suitable lead substance, 104

trends in, 19

Pathology, 128

Pathophysiology, 18, 28, 38, 79

PBEMP, 187

PC Model, 48, 281

PDB 271

PDBsum, 271

PDE III inhibitors, 77

PDE4 activity, 369, 372

PE Biosystems, 24

Pediatrics:

growth hormone deﬁ ciency, 29

HIV, 29

Penicillins, 222

Pentapeptides, synthesis of, 131

Pepsin, 296

Pepstatin/cathepsin D complex, 140

Peptic ulcer disease, 296–297, 299

Peptide(s):

antimicrobial, 400

bonds, 118, 120

characterized, 173

chemistry, 132, 134

chronic oral administration and, 120

coupling, 185

drug development example, 120

hydrolysis, 152

libraries, 24, 28

mass ﬁ ngerprinting, 242

parenteral lead, 128

phytoalexin studies, 45

pit viper-inspired, 120

sequencing, 255

synthesis, 131, 250

therapeutic, 368

INDEX 447

PeptIdent, 270

Peptidomimetic pyrazinone antithrombotics, 184

Peptidomimetics, 120–121, 128

Peptidyl-prolyl isomerase (PPIase), 180

Perhydro-3-oxo-1,4-diazepinium derivatives, 188

Perindopril, 127

Peritoneal macrophages, 409

Peroxidative tissue, 390

Peroxisome proliferator-activated receptor gamma

(PPAR-(γ)), 202

Pertussis, 416

Peterson, Per, Dr., 315

PFAM, 271

Pﬁ zer (Groton, CT):

azithromycin, 9–10

drug discovery/development, 203–204, 206–207,

217

partnership with Pliva, 15

prazosin, 13

pH:

chemoprotectant agents and, 411

efﬁ cacy studies, 111

molecular conformation assessment, 51

phosphate binding and, 387–388

Phage display, posttranslational modiﬁ cations, 246

Phagocytosis, 409

Pharmaceutical industry, 23–24

Pharmaceutical properties, 115

Pharmacia & Upjohn, 29

Pharmacodynamics, 110, 374, 419

Pharmacogenetics, 21, 28, 40–41, 65, 79, 278

Pharmacogenomic proﬁ le, 373

Pharmacogenomics, 265

Pharmacokinetic(s):

antimigraine drugs, 202

deﬁ ciencies, 107

deﬁ ned, 110

endotheline ET

receptors, 223

modeling methods, 278–279, 287

nevirapine studies, 361

nifedipine analogs, 207

proton pump inhibitors, 200

remifentanil, 349–350

signiﬁ cance of, 25, 150, 419

Pharmacological:

proﬁ le, 59–61

proof of principle, 28

prototype, 28

receptors, 296

Pharmacology, 235, 262–263

Pharmacophore(s):

combinatorial library design, 138

deﬁ ned, 106

efﬁ cacy-related, 68–69

functions of, 21, 37, 55, 63–64, 67, 80, 138, 317

model, 141

Phase-1, 66

Phase II/Phase III trials, 235–236

Phenanthroline, 391

1,2-Phenethyl diamines, 182

Phenothiazine, 221

Phenprocoumon, 226

Phenprocoumon K1, 226

Phenyl:

esters, 176

groups, 72

Phenylalanine, 46

Phenylbutazone, 417

2-Phenylglycerol, 52–53

4-Phenyl-2H-phthalazin-1-one, 187

Phenylhydrazines, 174, 177

Phenylpiperidines, 216

Phenytoin, 275

Phosphate:

anions, 389–390

binder therapy, 386

binding, 387–388

Phosphatidylinositol-4-phosphate phosphatase, 27,

Phosphodiesterases, 75–78, 227

Phosphodiesterase type 5 (PDE5) inhibitor, 217

Phospholipids, 400

Phosphorylation, 243

Photoafﬁ nity labeling studies, 28

Photolabile protecting groups, 131

Photolithography, 131, 155, 250

Physical organic chemistry, 82, 309, 387

Physical organic principles, 81–83

Physicians’ Desk Reference, 10, 12

Physicochemical properties, 57–58, 284

Physiopathological hypotheses:

-receptor antagonists, 219–220

levodopa, 217–219

rimonabant, 220–221

Phytoalexins, 45–46

Phytochemical(s):

implications of, 409

pathways, 45–46

structures, 45

Picrorhiza kurrooa, 417

Pioglitazone, 202

Piperazine, 150

Piperidine, 343, 349

Pipetting, 144

Pirenzepine, 355

value, 57, 111

PK data, 28

Plant(s):

extracts, 406

molecular diversity enhancement study, 45

Plasma protein binding factors, 58

Plastic, 155

Platelet aggregation, 179

Platelet-derived growth hormone, 29

448 INDEX

Platinum compounds, 413

Plicamine, 186

PM5VTHT, 187

PNU-96–988, 226

Polar surface area, 191

Poly(ADP-ribose), 32

Poly(ADP-ribose)glycohydrolase (PARG), 27–28

Polyanions, 398–399

Polyclonal antibodies, 252

Polydentate chelate effect, 388

Polyethers, 394

Polyethylene glycol monomethyl ether, 134

Polyethylene rods, 131

Polyhydroxamic acid polymers, 392

Polymer/polymeric:

amines, 392

anti-infective agents, 396–397

antimicrobial agents, 400–401

architecture of, 397

characterized, 383–384, 401

chronic renal failure and, 386–389

cholesterol-lowering agents, 392–396

functional, 383, 401

future research directions, 401

hyperkalemia, 385–386

inorganic ions in GI tract, 385

iron overload disorders, 389–392

physicochemical properties of, 401

as speciﬁ c molecular sequestrants, 384

toxin sequestration, 397–400

Polymerase chain reaction (PCR), 132, 250

Polymorphism, 228

Polypropylene mesh resin packets, 131

Polystyrene beads, cross-linked, 131

Polystyrenes, 398–399

Polyvalent interactions/polyvalency, 384, 396–397,

400–401

Pooling, 131–133

Pool/split, 131

Portion/mix, 131

Positional scanning, 132

Positive SAR, 41–42

Positron emission tomography (PET):

applications, generally, 365, 366–368

clinical studies, 373–374

preclinical studies, 371–373

principles of, 366–367

Positron-emitting radionucleides, 367

Postsynthesis processing, combinatorial libraries,

144–146

Posttranslational modiﬁ cations (PTMs), 238–239,

241, 243–245, 257, 259, 264

Potassium:

characterized, 385–386

hydroxide, 176

polymers, 385–386

salt, 344

Potency:

analog design, 215

combinatorial libraries, 134

deﬁ ned, 110

hypertension drugs and, 125–127

level of, 109

signiﬁ cance of, 104–108

Practolol, 207

Pravastatin, 202–203, 216

Prazosin, 13–15

Preclinical:

development compound, 25

testing paradigm, 79

PredictProtein, 247, 270

Pregnancy, 222

Prenylﬂ avanoid Diels-Alder natural products, 188

Prescriptions, physician authority, 13

Prestwick Chemical Library, 223

Preventive paradigms, 79

Prey, posttranslational modiﬁ cations, 245–246

Primary structure, deﬁ ned, 243

Principle component analysis (PCA), 137

Prion infections, 222

Private-sector drug discovery program, 23

Privileged structures, 21, 141

Process chemistry:

characterized, 18

cost and green chemistry, 70–71

stereochemistry deﬁ ned, 71–74

Procrit, 29

Procter & Gamble, 209

Prodrug strategies, 21, 39, 59–60, 68, 79–80, 124,

202, 218

Product(s):

diversity, 139

evaluation, 139

puriﬁ cation, 144–145

Production costs, 70–71

Production-scale synthesis, combinatorial chemistry,

136

Proﬁ ling data, 41, 59–61, 67, 70, 79, 136

Proﬁ t incentives, 3

Progenitor cells, 413

Proleukin, 30

Proline:

ACE inhibition, 123

derivatives, 180

Promethazine, 221

Proof-of-concept, 249

Proof-of-principle experiments, 176–177

Prophylactic treatment paradigms, 79

Propionic anhydrides, 344, 348

Propofol, 339, 350

Propranol, 207–208

Proprietary rights, 115

ProSite, 271

Prostaglandins, 222, 409

INDEX 449

Prostascint, 30

Prostate cancer, 176, 373

Protease:

characterized, 137, 141

inhibitors, 361, 401

Protection-deprotection scheme, 7

Protein(s):

ATP-dependent, 183

expression proﬁ ling, 252, 264

fold/folding, 75–78, 254, 271

identiﬁ cation, 242, 255

interactions, 238, 245

localization of, 238

microarrays, 248, 250–251, 253, 262–263

production of, 45

quantitation, software and databases, 255

receptors, 111

signatures, 252, 257

structure, see Protein structure

synthesis, 417

Protein capture arrays:

advantages of, 252

disadvantages of, 253

implementations of, 251–252

underlying principles, 251

Protein capture chips, 249

Protein chip technology:

current state of, 249–253, 264

functional protein microarrays, 248,

250–251

posttranslational modiﬁ cations, 246

protein capture microarrays, 248, 251–253

Protein Data Bank, 47

Protein Design Labs, 30

Protein-ligand structures, 261

Protein-lipid interactions, 238

Protein-nucleotide interaction, 239

Protein-phospholipid interaction assays, 258

Protein-protein interactions (PPI)s, 239, 245, 255,

368

Protein structure:

analysis, 246–248

computational determination, 247

laboratory determination, 246

signiﬁ cance of, 243, 313

Protein tyrosine phosphatase 1B (PPT1 B) inhibitors,

180

Proteolysis, 243

Proteome-wide protein-ligand study, 245

Proteomic data analysis:

computational infrastructure, 254

database management, 254–255

data checking and manipulation, 254–255

data generation, 253

LIMS, 254

software, 255

storage, 253–254

Proteomics:

characterized, 18, 21, 26, 28, 33–34, 38–39, 58, 67,

77, 79, 83, 261, 313, 401

data analysis, 253–256

fractionation, 239–240

functional areas of, 239, 249

identiﬁ cation, 240–242, 248

protein chip technology, 248–253

puriﬁ cation, 239–240

quantitation, 242, 248, 265

research goals, 238–239

ProtoMetrix, 249

Proton pump:

functions of, 200

inhibitors, 200, 310

Prototypes, 28

Protropin, 29

ProtScale, 247, 271

Psychiatric disorders, 367

Public policy, 80

PubMed, 270

Pulmonary:

embolism, 29

toxicity, 413

Pulmozyme, 29

Puriﬁ cation:

applicaitons, generally, 182, 239, 251

high-throughput techniques, 149

of nevirapine, 360

parallel solution-phase synthesis (PSPS), 185, 191

process characterized, 60

tandem afﬁ nity, 259

Purines, 183

Pyrazinobenzodiazepines, 324–332

Pyrazinones, 184

Pyrazoles, 221

Pyrazolines, 174, 177, 186, 221

Pyridazine, 170

Pyridazinyl rings, 225

Pyridines, 175–176, 310, 357–358

Pyridinium bromide hydrobromide, 175

Pyridobenzyo substance, 356

Pyrimidines, 174

Pyrimidinones, 172

Pyrrolidines, 181

Pyrrolidinyl, 170

Quadropole analyzer, 341

Quality of life, 79–80, 408

Quantitation, 239

Quantitative structure-activity relationships

(QSARs), 261, 284

Quantitative structure-property relationship (QSPR),

141

Quantitative structure-toxicity relationship (QSTR),

262

Quantum mechanics, 21, 47

450 INDEX

Quartz, 155

Quaternary amines, 388

Quaternary structure, deﬁ ned, 243

Quinacrine, 222

Quinapril, 127

Quinazolinones, 171

Quinidine, 116

Rabeprazole, 201

Racemic, 72–73

Racemic dihydroxyphenylalanine, 219

Racemization, 172

Raclopride, 375

Radiation, chemoprotectants, see Radioprotection

RADICAL (Registration, Analysis and Design

Interface for Combinatorial Libraries), 146–147,

149

Radio-frequency (RF) tag, 143

Radiolabeled probes, 367

Radioprotection:

antitumor agents, 418–419

from botanicals, 418

components of, 417–418

Ramipril, 127

Random walk, 49, 59

Raney nickel, 178

Ranitidine, 219–220

Rasayana:

botanicals, 408–409

characterized, 407

Rational drug design, 23–24

Rationally design molecules, 39

Rat studies, 226, 330–332, 342–343

Razoxane, 376

Reaction workup, 142, 144–145

Reactor blocks, 142

Reagent(s):

diversity, 139

research, 71

screening combinatorial libraries, 131–132

solid-supported (SSRs), 178

RECAP technique, 141

Receptor/ligand tag, posttranslational modiﬁ cations,

245

Rechlorination, 358–359

Recognition site, 246

Recombinate, 29

Recombine, 131

Recordati, 206–207

Recorder, in mass spectrometry analysis, 241

Recrystallization, 72

Rectome, 271

Red blood cells, 58

Reduction, parallel solid-phase synthesis (PSPS),

186

Reductive alkylation methods, 72

Regiospeciﬁ c monodisplacement reaction, 170

Regranex, 29

Relational databases, 37

Relcovaptan, 317

Remifentanil:

chemical development of, 344–349

discovery of, 340–344

human clinical trials, 349–350

manufacturing route, 347–348

medicinal chemistry route, 341–342, 348–349

Renagel, 389

Renal:

dopamine receptors, 75–78

failure, 29, 316

impairment, 385. See also Chronic renal failure

toxicity, 413

Renin, inhibition of, 119

Renin-angiotensin-aldosterone system, 117–119, 123,

128

ReoPro, 30

RepeatMasker, 269

Reperfusion injury, 181

Research and development (R&D) costs, 3

Resin-bound reagents, 178

Resistant virus, 361

Respiratory failure, 385

Respiratory syncytial virus disease, 30

Retesting, 260

Retevase, 29

Retinitis, 30

Retinoids, 409

Retroviruses, 355

Rexinoids, 409

Rhinitis, allergic, 314

Rhodacyanine dyes, antimalarial, 174

Rhodamines, 174

Rimonabant, obesity and, 220–221

Risedronate sodium, 208–209

Risk assessment, 28, 66

Risperidone, 370

Rituxan, 30

Rivastigmine, 371

Rizatriptan, 202, 204

RNA:

functions of, 105

interference (RNAi), 259, 261

polymerase, 246

synthesis, 419

Robbics Scientiﬁ c, 24

Robots/robotics, 144, 146–147, 149, 213, 253, 355

Robustness, 133

Roche, 208–209, 215

Rodent studies, 220

Roferon-A, 30

Rolipram, 369

ROMP (ring-opening metathesis polymerization)

agents, 189, 191

Rosetta, 247, 271

Rosiglitazone, 202

Rosuvastatin, 216

INDEX 451

Rotamase, 180

Rotatable bonds, 138, 191

Roxatidine, 219–220

Rule of ﬁ ve, 58, 134, 140

Rule of threes, 113

RWJ-56423, 314

RWJ-53308, 314

RWJ-355871, 314

Sac 1 p, 27, 32

Safety:

analog design, 215

bioavailability and, 226–227

Saizen, 29

Salmeterol, 186

Salmonella enteritidis, 397

Salt(s):

ammonium, 344, 394

calcium, 386

concentration, blood pressure regulation, 117–119

characterized, 228

guanidinium, 386–387

iminium, 174

metal, 387

oxalic acid, 344

potassium, 344

sodium, 344, 398–399

tosylated, 174, 328

Sample inlet, in mass spectrometry analysis, 241

Sample production, 2

Sandoz, 10, 203, 208

Sandwich ELISA, 252–253

Sankyo, 202–203

Sanoﬁ Aventis, 220

Sanoﬁ -Synthélabo, 201, 209, 317

Saponins, 415–416, 419

Sarcodictine, 178

Sarcoma-180, 413

Saturation kinetics, 369

Scaffold hopping/morphing, 215–216

Scale, in emergent technologies, 237

Scale-up issues, 2–3

Scavengers, solid-supported (SSSs), 178

Schering-Plough, 30, 205

Schild plot, 296, 298–299

Schizophrenia, 369, 376

Schrödinger equation, 21

Scientiﬁ c/primary news journals, 19

SciFinder, 10–11

SciVision, 66

Screening libraries, 260

Secondary amines, 72, 388

Secondary antibodies, 252

Secondary structure:

deﬁ ned, 243

structure prediction, 247

Secretagogs, 297–298

Sedatives, 205, 222

Selectin inhibitors, 181

Selective uptake mechanisms, 112

Selectivity, signiﬁ cance of, 104–107, 109, 134, 136,

215, 369, 383

Selenium, 409

SELEX procedures, 252

Semecarpus anacardium, 419

Semiempirical calculations, 21

Sepracor, 205

Septic shock, 397

Sequence:

analysis, 257

homologs, 244

homology, 255

Serena Software, 48, 281

Serine protease inhibitors, 314

Sermorelin, 29

Serono Labs, 29

Serostim, 29

Serotonin, 369, 371

Sertraconazole, 206

Serum protein binding, 114

Sexual dysfunction, 217

Shatavarins, 416

Sickle cell anemia, 390

Side effects, signiﬁ cance of, 3, 6, 124, 221, 235, 265,

350, 397

Siderophores desferrioxamine, 391

Siegfried, 206

Signal transduction, 245, 257, 369–370

Sildenaﬁ l, 217

Silico screening, 37

Silicon, 155

Silicon Graphics Indigo 2 workstation, 50

Simian immunodeﬁ ciency virus (SIV), 356

Similarity-dissimilarity indices, 57–58

Simulect, 30

Simvastatin, 202–203

Single-molecular systems, 79

Single nucleotide polymorphisms (SNPs), 21, 40, 265

Single photon emission tomography (SPECT), 366

Site-directed mutagenesis (SDM):

characterized, 27–31, 36

compound libraries and, 45

future directions for, 78

historical perspectives, 39

Sitoindoside VII/VIII, 415

6–31G*, 20, 51

6–32G*, 282

SKF 97426-A, 394

SKF-101926, 317

SKF-105494, 317

Skin discoloration, 221

Sleep disturbance, 370

Slurries, solid-phase synthesis, 134

SLV 319, 220–221

Small interfering RNAs (siRNA), 259, 266

Small intestine, 111

452 INDEX

Small-molecular combinatorial chemistry, 24

Small-molecule:

compounds, 83

drug discovery and development, 27–28, 33–35

synthesis, 132–133, 135

Small ring system, 52–54

SMART (single or multiple addressable

radio-frequency tag) semiconductor, 133

SmithKline & French Research Institute, 295, 305,

317, 394

SmithKline Beecham, 150, 201–202, 204

Snake venom, 117, 119–120, 123

SNP Consortium, 66

Société d’Études Scientiﬁ ques et Industrielles de

l’Île-de-France, 204

Sodium:

channel blockers, 179

hydride, 357, 360

polystyrene sulfonate, 385–386

salt, 344, 398–399

sulfonate, 411

thiosulfate (STS), 411–412

Soft donor atoms, 391

Soft drugs, 21, 39, 60, 63–65, 71, 79–80, 342

Soft tissue calciﬁ cation, 386

Solid-phase:

beads, 131–132

chemistry, 131–135, 149

extraction (SPE), 145

peptide chemistry, 143–144

synthesis, 145, 152

Solids, delivery of, 142

Solid-state interactions, 49

Solid-supported catalysts (SSCs), 178

Solid-supported reagents (SSRs), 178

Solid-supported scavengers (SSSs), 178

Solid tumor cancer cells, 263

Solubility, 108, 191, 227

Soluble mixture libraries, 135

Solution chemistry, 144

Solution-phase, see Parallel solution-phase

synthesis

combinatorial chemistry, 134, 152

synthesis, 153

Solution synthesis, 183

Solvay Pharmaceuticals, 221

Somatostatin:

antagonists, 181

characterized, 150–151

SOSA (selective optimization of side activities)

approach:

characterized, 214, 223, 226–228

examples of, 223–226

rationale for, 223

Sour Pliva (Zagreb):

azithromycin, 9–10

Pﬁ zer partnership, 15

Soybean cyst nematode infection, 45–46

Speciﬁ cation, combinatorial library design, 138

Speciﬁ city, 369

Spectroscopy applications, 39. See also Mass

spectrometry

Speed, in emergent technologies, 237

Spiperone analogs, 179

Spirapril, 127

Spiroindolinones, 317

Spiropyrrolopyrroles, 183

Spir[pyrrolidine-2,3N-oxindole] libraries, 173

Split-and-mix:

philosophy, 133–134, 139

synthesis, 131, 155

Split-split synthesis, 144

Springer-Verlag, 48, 281

Squibb, 120, 124, 200

SR 141716A, 220

SR-121463, 317

SSR-4-149415, 317

Stability, drug development process, 109

Staphylococcus aureus, 397, 400

STAR, 66

State of nonspeciﬁ cally increased resistance of an

organism (SNIR), 407

Statins, 10, 202, 222, 392, 396

Statistical analysis, 255

Staudinger protocol, 190

Stem cell proliferation, 415

Stereochemistry, 10, 71–74, 107, 121, 307

Stereocontrol, 71

Steric environment, 72

Steric physicochemical properties, 64

Steroid/steroidal:

analogs, 216

characterized, 222

hormones, 214

lactones, 415

saponins, 416

Stimulant effect, 303

Stipiamide, 178

Stirring blocks, 142

Stoichiometry, 83

Storage access network (SAN), 253–254

Strecker reaction, 344

Strokes, 29, 116, 392

Structural-absorption relationships (SAbRs), 58,

Structural analogs, 214

Structural database, 54

Structural proteomics, 261

Structure metabolism relationships (SMRs),

279

Structure stereochemical relationship (SSR), 73

Structure-activity analysis, 295, 301, 305–306, 317.

See also Dynamic structure-activity analysis

(DSAA)

INDEX 453

Structure-activity relationships (SARs):

captopril analogs, 122–124

checking and manipulation, 254–255

classical studies, 27, 31, 75

combinatorial chemistry, 136, 151–152

combinatorial library design, 140

directing distribution and, 58

efﬁ cacy-related, 41–42, 58

establishment of, 121

future research directions, 39

during HTS, 41–42, 81

models of, 23

molecular conformation studies, 52

multidrug resistance (MDR) and, 42, 83

negative, 41–42, 63–64

neutral, 41–42, 63–64

nevirapine, 356–358

paclitaxel (PAC), 42–43, 69

polymeric amines, 388–389

positive, 41–42

reverse, 28

selectivity-related, 65

signiﬁ cance of, 23, 80, 309

site-directed mutagenesis (SDM), 31

small-molecule drug design, 27, 58–59

suitable lead substance, 107

toxicology, 67–68

ultrashort-acting analgesics, 343–344

Sructure-based drug design, 21, 33, 35, 106, 314

Structure-debenzylation relationship (SDebR), 73

Structure-distribution relationship (SDR), 58–59, 65

Structure-elimination relationships (SERs), 65

Structure-function relationships, 246–248

Structure-metabolism relationship (SMR), 63–65,

68, 286

Structure-toxicity relationship (STR), 67

Styrene, 394

3-Substituted indoles, 180

Succinlycholine, 340

Sufentanil, 341, 343

Suitable lead substance, characteristics of:

analog attrition, 108

overview, 104–105

potency, 104–107

selectivity, 104–107

structure-activity relationships (SARs), 107

toxicity, 107–108

Sulconazole, 206

Sulfamides, 221

Sulfate, 114

Sulfathiazole, 223–224

Sulfenamide, 200

Sulfhydryl groups, 121–122, 124

Sulfhydryl/sulﬁ de-containing tissue, 107

Sulﬁ soxazole, 223–224

Sulfonamides, 189, 222

Sulfonic acid groups, 398

Sulfonyl chlorides, 191

Sulfur:

characterized, 124

electron-withdrawing, 220

methylene isosterism, 306–307

Sumatriptan, 202, 204

Supercritical ﬂ uid chromatography (SFC), 245

Supermolecule, formation process, 105

Superoxide dismustase, 414

Superoxide radicals, 415

Suppliers, of combinatorial systems and compound

library trends, 24

Surface-enhanced laser desorbtion ionization

(SELDI), 264

Surface plasmon resonance, 19

Surgery, anesthesia, 339–350

Suzuki coupling, 54

Swern conditions, 177

Swiss-2DPage, 271

SXR, 64, 84

SYBL, 48, 281

Synagis, 30

Synergism, 18, 60–63, 79

Synopsis Metabolism Database, 39

Synsorb, 400

Syntex, 206

Synthélabo, 208

Synthesizers, automated, 141, 143–144

Synthetic chemistry, 18, 36, 71, 128

Synthetic chemists, functions of, 129

Synthetic organic chemistry, 82

Synthetic organic chemists, functions of, 23

Synthetic polymers, 384

Synthetic stereophores, 73

Systematic studes, drug metabolism, 278

Systems biology perspectives, 237

Tag afﬁ nity, 252

Tagamet:

characterized, 295, 309

development of clinical candidate drug, 305–309

historical background, 295–298

search for H

-receptor histamine agonists,

298–305

Tagging reagent, 188

Taisho Pharmaceuticals (Japan), erythromycin

studies, 5

Takeda, 201–202

Tamoxifen, 409

Tandem afﬁ nity puriﬁ cation (TAP), 245–246

Tandem mass spectrometry (MS/MS), 241–242, 255,

257

Target(s):

identiﬁ cation, see Target identiﬁ cation

site, types of, 23–24

speciﬁ city, 261

x-ray diffraction, 25

454 INDEX

Target identiﬁ cation:

signiﬁ cance of, 1–2, 252

validation of, 234, 256–259, 266

Taste, in drug development process, 109

Taxanes, 43, 413

Taxoids, 178

Taxol, 412

T cells:

bioengineered, 372

CD4+, 353

Tea bags, 131

Tecan, 24

Tektronic, 48, 281

Telmisartan, 201

Temocapril, 127

Terazosin, 13–15

Terfenadine, 205

Tertiary:

amines, 72, 388

structure, deﬁ ned, 243

Testing process, overview of, 2–3. See also speciﬁ c

research studies

Tetracycline, 228

Tetraﬂ uorophthalic anhydride, 185

Tetrahydrocannabinol (THC), 220

Tetrahydrofolate, 105

Tetrahydrofuran (THF), 13, 182

Tetrazolyl amide, 172

Teva Pharmaceuticals, 10

Thalidomide, 222

Therapeutic activity, 265

Thiadiazoles, 127, 225

Thiazinobenzodiazepines, 324–332

Thiazole, 310

Thiazolidine-2,4-diones, 200

Thiocarbonyl, 303

Thiols, 107, 185, 411

Thiorea, 178

Thiosulfate, 186, 411

Thiourea, 220, 303, 307–308, 412

6–31G*, 20, 51

6–32G*, 282

Threading, 247

3D chemical structures, 47, 49–50, 57, 81

Three-dimensional descriptors, 138

3D quantitative SAR (3D-QSAR), 20, 56

384-well:

arrays, 262

microplates, 20–21, 260

Three-phase reaction, 186

Threonine, 32

Thrombin receptors agonists, 315

Thrombocytopenia, 30, 413

Thymidine, 408

Thymidylate synthase, 105

Thymine, 105–106

Thyroid disorders, 221

Tiered structures, 55–57

Tiers, in chemical structure databases, 284–285

Tiludronate disodium, 209

Time-of-ﬂ ight (TOF) separation, 241

Tinospora:

bakis, 416

cordifolia, 408, 415, 417, 419

Tipranavir, 226

Tissue factor VIIa complex, 184

Tissue plasminogen activators (tPAs), 29

Tissue-speciﬁ c delivery, 263

Tocopherols, 409

Tolcapone, 219

Tolevamer, 399

Tolvaptan, 317

TOPAMAX topiramate, 314

Topiramate, 314

Topographical mapping studies, 28, 31

Topoisomerase, characterized, 183

Topoisomerase I

177

, 43

Topoisomerase I/Topoisomerase II, 371, 413–414

Topotecan, 43– 44

Tosylated salts, 174, 328

Tox Express/Gene Express database, 66

Toxicity:

assessments, 262

avoidance strategies, 65–67

cisplatin-induced, 412

proﬁ le, 262

signiﬁ cance of, 18, 59–60, 109, 383–384

studies, 223, 228, 264

suitable lead substance, 107–108

Toxic markers, 25

Toxicology:

databases, 66

implications of, 235

Toxicophores, 21, 37, 65, 107

Toxins, polymeric sequestration, 397–400

TOXSYS (SciVision), 66

Traditional Chinese medicine, 406

Trancription factors, 239

Trandolapril, 127

Tranquilizers, 215

Transcriptional regulatory networks, 257

Transcription control, 245

Trans-1,4-diaminocyclohexane, 176

Transgenic species, 28

Transplantation:

blood/cell marrow, 30

rejection, 30

Transporter system, 71

Transportophores:

characterized, 21, 42–43

relationships, 51

Trastuzumab, 265

Triarylphosphine, 188, 190

Triazine-based compounds, 173

INDEX 455

1,2,4-Triazolo[4,3-b]pyridazines, 170

3,4,5-Trichloropyridazine, 170

Tricyclic psychotropics, 222

Triﬂ uoroacetamide, 190

2,2,2-Triﬂ uoroethyl group, 325

Triﬂ uoromethanesulfonyl (Tf), 325

Triglyceride levels, 202

Trigonal bipyramid (TBP) transition state (TS), 75,

Triphenylphosphonium ylides, 175

Tripos Assoc., 48, 281

Tris(perﬂ ourohexylethyl)silyl group, 190

2,4,6-Trisubstituted quinazolines, 179

Troglitazone, 200, 202

Trp, 78

Tumor(s):

cells, 151, 370

-destructive therapy, 370, 408

metabolism, 376

metastases, 152, 181, 372

vascularization, 373

xenografts, 371

Tumor necrosis factor (TNF), 408, 416–417

Tumor necrosis factor-∀ (TNF-∀), 369

Turner’s syndrome, 29

3–21G*, 20, 51, 282

2D chemical structures, 47, 54–55, 57

Two-dimensional descriptors, 138

Two-dimensional differential in-gel electrophoresis

(2D-DIGE), 257

Two-dimensional polyacrylamide gel electrophoresis

(2D-PAGE), 240–241, 253, 255, 257, 264–265

Tyrosine/tyrosine kinases, 27, 361

Ubiquitination, 243

UCSC Genome Browser, 269

Ugi reaction, 177–178

UK-92,480, 217

Ulcers/ulceration, 29, 220, 416– 417

UL-409, 416–417

Ultiva, 340

Ultrahigh-throughput screening (UHTS), 21, 79,

355

Ultraviolet (UV):

absorption, 242

detectors, 145

Underexpression, 259

U.S. Environmental Protection Agency (EPA),

High-Volume Chemical (HPV) Screening

Information Data Sets (SIDS), 66

U.S. Food and Drug Administration (FDA):

approval, 254, 265, 362

guidelines, 359

labeling, 202

regulation by, 25, 235–236

stereoisomers, 70–71

toxicity, 66

U.S. Patent and Trademark Ofﬁ ce, 9

Unsaturated esters, 187

Upjohn, 226

Urea:

isostere, 304

production of, 190

Uricosuric agents, 221–222

Uridinediphosphoglucose dehydrogenase, 417

Urinary tract protectants, 412

Urine, 113, 115

Urothelial toxicity, 412

Urotoxicity, 412, 415

Urticaria, 298

Vaccination/vaccines, 249, 368, 370

Validity of data, 234, 254, 256–259, 266

Valsartan, 201, 215–216

Vancomycin:

characterized, 398

resistant enterococci, 400

Vascular-endothelial growth factor (VEGF) receptor,

373

Vasoconstriction, 316

Vasopressin receptor antagonists:

azepinoindoles, 319, 321–322

background of, 313–315

bridged bicyclic derivatives, 322–324

characterized, 315–317

oxazinobenzodiazepines, 324–332

project genesis, 315

pyrazinobenzodiazepines, 324–332

study design, 317–319, 332–333

thiazinobenzodiazepines, 324–332

Veber, Daniel, 113

Veber rule, 113

Vespeside, 412

Vezluma, 30

Viagra, 217

Vibrational frequencies, molecular conformation,

Vibrio cholera, 397

Vicinal dibromides, 187

Vinyl amine polymers, 394

Virtual compound libraries/virtual library, 79, 84,

137, 284

Virtual screening, 37, 260

Virus(es):

evolution of, 40

infectivity, 181

inhibition of, 397

replication, 181, 361–362

Viscinol diols, 186

Viscum album, 408

Vitamin E, 414

Vitravene, 30

VPA-985, 317, 325, 327

selectivity, 326–327, 329, 332

456 INDEX

Wandering behavior, 370

Wash cycles, 143

Wasting, AIDS-related, 29

Watanabe, Yoshiaki, Dr., 5

Water:

retention, 316

solubility, 109, 111–115, 228

WebGene, 269

WebLogo, 269

Web sites, as information resource, 19, 24

Weighting decision criteria, 67–70, 143

Weight loss, 220

WelChol, 396

Wellcome, 204

Western immunoblot analysis, 264

Winn, Marty, Dr., 13

Withaferin A, 418–419

Withaferins, 415

Withaferin somnifera, 408–409, 417–418

Withanolides, 415, 418

Wittig cyclization/reaction, 175, 185, 187

Wolff-Parkinson-White syndrome, 221

Workﬂ ow, combinatorial library management, 146–147

Workstations, automation, 149

World Health Organization, 353

WR2721, 412

Wyeth Pharmaceuticals, 315

Xanthine, 172

Xenobiotic(s):

characterized, 21, 35, 278

directing distribution, 58–59

elimination of, 65

metabolism database, 286

Xenograft models, 372

X-ray(s):

ACE inhibition study, 122

combinatorial library design, 140

crystallographic studies, 21, 176, 246, 260

diffraction, 18, 23, 25–26, 33, 35, 38, 76, 79, 135,

328

drug metabolism studies, 282–284

inhibitor functions studies, 105

molecular conformation assessment, 47, 49–51,

53, 57

Xylene, 357

XYZ synthesizers, 142–144

Yamanouchi, 209

Yeast/two-hybrid (Y2H) studies, 245, 368

y ions, 241–242

YM-087, 317

Zalcitabine, 361

Zantac, 310

Zenapax, 30

Zhu, Heng, 250

Zidovudine, 353–354, 361

Zinc, 120, 391

Zinc-metalloproteases, 120

Zoledronate disodium, 209

Zolmitriptan, 204

Zopiclone, 215

Zymark, 24