RSC Drug Discovery

Edited by Dennis Allen Smith

Metabolism, Pharmacokinetics
and Toxicity of Functional Groups
Impact of Chemical Building Blocks on ADMET
Metabolism, Pharmacokinetics and Toxicity of Functional
Groups
Impact of Chemical Building Blocks on ADMET
RSC Drug Discovery Series

Editor-in-Chief
Professor David Thurston, London School of Pharmacy, UK

Series Editors:
Dr David Fox, Pfizer Global Research and Development, Sandwich, UK
Professor Salvatore Guccione, University of Catania, Italy
Professor Ana Martinez, Instituto de Quimica Medica-CSIC, Spain
Dr David Rotella, Wyeth Research, USA

Advisor to the Board:

Professor Robin Ganellin, University College London, UK

Titles in the Series:

1: Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact
of Chemical Building Blocks on ADMET

How to obtain future titles on publication:

A standing order plan is available for this series. A standing order will bring
delivery of each new volume immediately on publication.

For further information please contact:

Book Sales Department, Royal Society of Chemistry, Thomas Graham House,
Science Park, Milton Road, Cambridge, CB4 0WF, UK
Telephone: +44 (0)1223 420066, Fax: +44 (0)1223 420247, Email: books@
rsc.org
Visit our website at http://www.rsc.org/Shop/Books/
Metabolism, Pharmacokinetics
and Toxicity of Functional
Groups
Impact of Chemical Building Blocks on
ADMET

Edited by

Dennis A. Smith
Sandwich Laboratories, Pfizer Global Research and Development, Kent, UK
RSC Drug Discovery Series No. 1

ISBN: 978-1-84973-016-7
ISSN: 2041-3203

A catalogue record for this book is available from the British Library

r Royal Society of Chemistry 2010

All rights reserved

Apart from fair dealing for the purposes of research for non-commercial purposes or for
private study, criticism or review, as permitted under the Copyright, Designs and Patents
Act 1988 and the Copyright and Related Rights Regulations 2003, this publication may not
be reproduced, stored or transmitted, in any form or by any means, without the prior
permission in writing of The Royal Society of Chemistry or the copyright owner, or in the
case of reproduction in accordance with the terms of licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of the licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries concerning
reproduction outside the terms stated here should be sent to The Royal Society of
Chemistry at the address printed on this page.

The RSC is not responsible for individual opinions expressed in this work.

Published by The Royal Society of Chemistry,

Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, UK

Registered Charity Number 207890

For further information see our web site at www.rsc.org

Preface

When it was suggested to me that the area of Drug Metabolism and Phar-
macokinetics was lacking in works available for the Medicinal Chemist I was
somewhat surprised. On reflection a study of many of the volumes revealed
that although the scholarship was outstanding many had not been written to
provide information to the medicinal chemist as they encountered new com-
pound series.
So often the change of project and perhaps even the chance of chemical lead
series takes the chemist into an area that he has not encountered before. The
changes in absorption, distribution, metabolism and excretion properties
(ADME) can be profound. For instance a chemist working on aminergic
GPCR receptors will be used to compounds cleared by metabolism, with large
volumes of distribution and often good access to the CNS. Results from screens
often do not need to be re-intepreted to allow for protein binding effects.
Perhaps a switch to a non-aminergic GPCR will cause the chemist now to
work with acidic molecules. Now high intrinsic clearance can be disguised by
high protein binding, metabolic clearance is augmented by significant drug
transporter effects and some of the metabolic steps are reversible (e.g. acyl
glucuronides).
This volume attempts to fill this void. Metabolism, Pharmacokinetics and
Toxicity of Functional Groups tries to do what it says on the cover. Our
definition of the key functional groups seemed right at the outset. As we have
assembled the work it is pleasing to see the holistic nature of the concept. Many
of the chapters build off concepts described in others but each can be viewed as
a seperate entity and one we hope chemists find rewarding as they move into
new chemical areas or perhaps revisit others.
The aim of all the authors was to impart around 300 years of collective
knowledge and wisdom about the impact of ADME on Drug Discovery and

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

v
vi Preface
hopefully aid all those active in this noble profession. We like to think that
someone, somewhere, has used a concept in this volume that has led to a new
medicine, even better if it is medicines in the plural. If this turns out to be true
our time will have been well spent. All the authors are associated with Pfizer
and we would like to thank the company for encouraging the production of
this work.
Contents

Chapter 1 Drugs and their Structural Motifs 1

Alexander A. Alex and R. Ian Storer

1.1 Introduction 1
1.2 Launched Drugs 6
1.2.1 Target Space of Launched Drugs 7
1.2.2 Chemical Space of Launched Drugs 9
1.2.3 Molecular Properties of Launched Drugs 10
1.2.4 Polypharmacology 13
1.3 Drugs Bound to their Targets 15
1.3.1 Comparison of Binding Sites of Drugs Bound
in their Biological Targets 25
1.3.2 Phosphodiesterase 5 (PDE5) Drugs 25
1.3.3 Cyclooxygenase (Cox) Drugs 26
1.3.4 Classes of Drugs with High Structural
Similarity 28
1.4 Privileged Substructures in Drugs 28
1.4.1 Examples of Privileged Substructures 30
1.4.2 Benzodiazepines 30
1.4.3 Arylsulfonamides and Drugs Derived from
them 31
1.4.4 Chemokine Receptor 5 (CCR5) 37
1.4.5 Diaryl Heterocycles such as Cyclooxygenase
Inhibitors (COX-2) 39
1.4.6 Aminoheterocycles as Kinases Inhibitors 41
1.4.7 HMG-CoA Reductase Inhibitors 44
1.5 Discussion of Privileged Substructures and Chemical
Space 44

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

vii
viii Contents

1.6 Reasons for Compound Attrition 47

1.7 Summary and Outlook 51
1.8 Abbreviations 52
References 53

Chapter 2 ADMET for the Medicinal Chemist 61

K. Beaumont, S. M. Cole, K. Gibson and J. R. Gosset

2.1 Introduction 61
2.1.1 Physicochemical Principles for ADME 62
2.1.2 Physicochemistry Summary 65
2.2 Delivery of Drugs and Bioavailability 65
2.2.1 Oral Delivery 66
2.2.2 Intranasal Delivery 73
2.2.3 Inhaled Delivery 73
2.2.4 Sublingual Delivery 73
2.2.5 Rectal Delivery 74
2.2.6 Transdermal Delivery 74
2.2.7 Subcutaneous and Intramuscular
Administration 74
2.3 Tissue Distribution of Drugs 75
2.3.1 Distribution to the Central Nervous System 80
2.4 Clearance, Extraction, Metabolism and Excretion 82
2.4.1 Clearance 82
2.4.2 Clearance by the Liver 84
2.4.3 Metabolism 84
2.4.4 Biliary Elimination 91
2.4.5 Clearance by the Kidney 92
2.4.6 Clearance Summary 93
2.5 Toxicology related to ADME 94
References 94

Chapter 3 Carboxylic Acids and their Bioisosteres 99

Amit S. Kalgutkar and J. Scott Daniels

3.1 Introduction 99
3.2 Carboxylic Acid Containing Non-steroidal
Anti-inflammatory Drugs (NSAIDs) 103
3.2.1 Discovery of Aspirin 103
3.2.2 Mode of Inhibition of COX Activity by
NSAIDs 106
3.2.3 Molecular and Structural Basis for COX
Inhibition by NSAIDs 106
3.3 Carboxylic Acid Containing b-Lactam Antibiotics 107
3.3.1 Discovery of Penicillins 109
3.3.2 Mechanism of Action of b-Lactam Antibiotics 109
Contents ix

3.4 Carboxylic Acid Containing Statins 110

3.4.1 Discovery of the Statins 111
3.4.2 Molecular and Structural Basis for Inhibition
of HMG-CoA Reductase by Statins 113
3.5 Carboxylic Acid Containing Fibrates 113
3.6 From Terfenadine to Fexfofenadine–an Interesting
Case Study on the Utility of the Carboxylic Acid
Moiety in Drug Discovery 115
3.7 Bioisosteres of the Carboxylic Acid Moiety 116
3.7.1 Non-classical Bioisosteres of the Hydroxyl
Portion of the Carboxylic Acid Group 117
3.7.2 Non-classical Bioisosteres of the Entire COOH
Moiety 119
3.8 Absorption, Distribution, Metabolism and Excretion
(ADME) Profile of Carboxylic Acids 122
3.8.1 Oral Absorption 122
3.8.2 Distribution and Clearance 125
3.8.3 Metabolism of the Carboxylic Acid Moiety 126
3.8.4 P450 Isozymes Involved in the Oxidative
Metabolism of Carboxylic Acid Derivatives 140
3.8.5 Hepatobiliary Disposition of Carboxylic Acids 141
3.9 ADME Profile of Tetrazoles 143
3.9.1 Metabolism of the Tetrazole Motif 145
3.9.2 Role of Transporters in the Disposition
of Tetrazole-based Angiotensin II Receptor
Antagonists 146
3.10 ADME Profile of Thiazolidinedione Derivatives 146
3.10.1 Clearance and Oral Bioavailability 147
3.10.2 Metabolism of the Thiazolidinedione Ring
System 147
3.10.3 P450 isozymes Responsible for the
Metabolism of ‘glitazones’–DDI Potential 148
3.11 ADME Profile of Esters and Amides 149
3.12 Boronic Acid Derivatives 150
3.13 Concluding Remarks: Carboxylic Acid and
Drug Safety 151
References 154

Chapter 4 Primary, Secondary and Tertiary Amines and their Isosteres 168
D. K. Walker, R. M. Jones, A. N. R. Nedderman and
P. A. Wright

4.1 Introduction 168

4.1.1 Amines that Interact with Aminergic
Receptors 169
x Contents

4.1.2 Amines that Interact with Acetylcholine 170

4.1.3 Amines that Interact with Opioid Receptors 170
4.1.4 Amines that Interact with Ion Channels 171
4.1.5 Amine Antimalarial Drugs 171
4.1.6 Miscellaneous Amine Drugs 172
4.1.7 Amine Isosteres 172
4.2 Physicochemical Properties of Amines 173
4.2.1 Polarity of Amines 173
4.2.2 Basicity of Amines 174
4.3 Absorption Properties of Amine Containing
Drugs 176
4.3.1 Solubility and Absorption 176
4.3.2 Membrane Permeability and Absorption 177
4.3.3 Impact of P-glycoprotein on Absorption 180
4.4 Systemic Behaviour of Amine Containing Drugs 181
4.4.1 Tissue Affinity and its Impact on Distribution 181
4.4.2 Distribution and Duration 181
4.4.3 Additional Specific Interactions Enhancing
Tissue Affinity 183
4.4.4 Distribution Dependent on pH 183
4.4.5 Plasma Protein Binding 184
4.4.6 Brain Distribution 185
4.5 Clearance of Amine Containing Drugs 185
4.5.1 Metabolic Clearance 185
4.5.2 Phase 1 Metabolism 186
4.5.3 Phase 2 Metabolism 192
4.5.4 Non-metabolic Clearance 195
4.5.5 Renal Clearance 195
4.5.6 Biliary Clearance 195
4.6 Amines as Toxicophores and Toxicity of Amine
Containing Drugs 196
4.7 Zwitterions 199
4.8 Prodrugs of Amines to Change Physicochemical
Properties 200
4.8.1 Prodrugs to Enhance Absorption 201
4.8.2 Prodrugs to Achieve Tissue Specificity 202
4.8.3 Prodrugs Utilising Amine Functionality 203
References 204

Chapter 5 Sulfonamide as an Essential Functional Group in Drug

Design 210
Amit S. Kalgutkar Rhys Jones and Aarti Sawant

5.1 Introduction 210

5.1.1 Sulfanilamide Antibacterial Agents 212
Contents xi

5.1.2 Sulfonamide-based Anti-inflammatory

Agents 214
5.1.3 Sulfonamide-based Carbonic Anhydrase
Inhibitors 219
5.1.4 Sulfonylurea-based Hypoglycemic Agents 222
5.1.5 Miscellaneous Applications of the
Sulfonamide Group in Medicinal
Chemistry 224
5.2 Absorption, Distribution, Metabolism and
Excretion of Sulfonamides 230
5.2.1 Oral Absorption 230
5.2.2 Distribution 237
5.2.3 Metabolism 242
5.2.4 Renal Elimination 249
5.3 Adverse Drug Reactions (ADRs) with Sulfonamide
Drugs 251
5.3.1 Types of Hypersensitivity Reactions with
Sulfanilamide Antibacterials 251
5.3.2 Mechanism of Type 2 Hypersensitivity by
Sulfonamide Antibacterials 252
5.4 Bioactivation Pathways Involving the Sulfonamide
Motif 255
5.4.1 Bioactivation of Sulfonanilides 255
5.4.2 Intrinsic Electrophilicity of ‘Activated’
Sulfonamides 257
5.4.3 Bioactivation of Sulfonylureas 261
5.5 Conclusions 263
References 264

Chapter 6 Influence of Aromatic Rings on ADME Properties

of Drugs 275
Deepak Dalvie, Sajiv Nair, Ping Kang and Cho-Ming Loi

6.1 Introduction 275

6.2 Physicochemical Properties of Aromatic and
Substituted Aromatic Rings 276
6.2.1 Importance of Fluorine Substitution on
Phenyl Rings 283
6.3 Influence of Aromatic and Substituted Aromatic
Rings on ADME Properties of Compounds 285
6.3.1 Absorption 285
6.3.2 Distribution 290
6.3.3 Clearance 295
6.4 Toxicity 308
References 322
xii Contents

Chapter 7 Influence of Heteroaromatic Rings on ADME Properties

of Drugs 328
Deepak Dalvie, Ping Kang, Cho-Ming Loi, Lance Goulet
and Sajiv Nair

7.1 Introduction 328

7.2 Types of Heteroaromatic Rings and their
Physicochemical Properties 333
7.3 Influence of Heteroaromatic Rings on ADME
Properties of Compounds 338
7.3.1 Absorption 339
7.3.2 Distribution 344
7.3.3 Metabolism 348
7.3.4 Excretion 354
7.4 Influence of Heteroaromatic rings on Toxicity of
Compounds 357
7.5 Summary 364
References 365

Chapter 8 Peptidomimetics and Peptides as Drugs: Motifs Incorporated

to Enhance Drug Characteristics 370
Tracey Boyden, Mark Niosi and Alfin Vaz

8.1 Introduction 370

8.2 Peptidomimetics for Aspartic Acid Proteases 371
8.3 Anticancer Peptidomimetics 379
8.3.1 Summary 381
8.4 Peptide Drugs 382
8.4.1 Insulin and Insulin Analogs 382
8.4.2 Incetin Hormones 385
References 387

Chapter 9 Pharmacokinetics and Metabolism of Compounds that Mimic

Enzyme Transition States 390
Iain Gardner, Chris Barber, Martin Howard, Aarti Sawant
and Kenny Watson

9.1 Enzyme Transition States 390

9.2 Physicochemical Properties of Transition State
Analogues 393
9.3 ADME Properties of Transition State Analogue
Inhibitors against Different Enzyme Targets 396
9.3.1 Proteases 396
9.3.2 Neuraminidase TSAI 427
9.3.3 N-Ribosyltransferase TSAI 432
Contents xiii

9.3.4
Nucleoside Deaminase TSAI 435
9.3.5
Inosine 5-monophosphate Dehydrogenase
TSAI 440
9.3.6 Aspartate Carbamyl Transferase TSAI 441
9.3.7 Glycosidase Inhibitor TSAI 442
9.4 Conclusions 443
9.5 Abbreviations 444
References 445

Chapter 10 Alcohols and Phenols: Absorption, Distribution, Metabolism

and Excretion 460
Zhuang Miao and R. Scott Obach

10.1 Physicochemical Properties of Alcohols and

Phenols and their Prevalence in Drugs 460
10.2 Comparative Pharmacokinetics of Alcohols,
Phenols and their Counterparts Lacking the
Hydroxy Group 462
10.3 Biochemical Determinants of ADME
Characteristics of Drugs Possessing Hydroxyl
Groups 464
10.3.1 Plasma Protein Binding and Tissue
Distribution 466
10.3.2 Interactions of Hydroxyl Group
Containing Drugs with Drug Transporters
and Impact on Absorption, Distribution
and Excretion 467
10.3.3 Metabolism and Interaction with
Drug-metabolising Enzymes 468
10.3.4 Fluorine as an Isostere of Hydroxy Groups 480
10.4 Conclusions 482
References 482

Chapter 11 Future Targets and Chemistry and ADME Needs 486

Dennis A. Smith and David S. Millan

11.1 The Human Genome 486

11.2 Drug Targets within the Genome 487
11.3 The Genome Gap 488
11.3.1 New Mechanisms and the Druggable
Genome 490
11.4 The Need for New ADME tools 491
11.5 The Chemistry Gap 493
11.6 The Knowledge Gap in Drug Design 498
xiv Contents

11.7 Permeability of Membranes: A Pivotal

Role in Drug Disposition 498
11.8 Future CNS Targets and ADME Space 503
11.9 Penetration into the Cell—Intracellular Drug
Targets 506
11.10 Permeability and Large Molecules 507
11.11 Conclusion: Beyond PSA and ADME Space 508
References 509

Subject Index 512

CHAPTER 1

Drugs and their Structural

Motifs
ALEXANDER A. ALEX AND R. IAN STORER

Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent,

UK, CT13 9NJ

1.1 Introduction
The major focus of the research-based pharmaceutical industry is the discovery
of safe, efficacious, new chemical entities (NCEs) for therapeutic targets. The
pharmaceutical industry can look back at a history of successful innovations,
indicated by the fact that there are currently just over 1400 unique drugs on the
market.
The success of the industry can be measured in, for example, the increase in life
expectancy in men and women over the last four decades. For instance, a child born
in the United States in 2005 can expect to live nearly 78 years (77.9 years). The
increase in life expectancy represents a continuation of a long-running trend. Life
expectancy has increased from 75.8 years in 1995 and from 69.6 years in 1955.
(www.cdc.gov/nchs/pressroom/07newsreleases/lifeexpectancy.htm). Although there
are multiple factors which potentially contribute to the increase in life expectancy,
like for example diet and life style, the development and availability of new drugs
appear to have made a substantial contribution.
Equally impressive, the impact of the industry can also be highlighted by
the increase in five-year-survival rates for cancer when diagnosed 1975–1977

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

1
2 Chapter 1
compared to when diagnosed in 2000 (www.phrma.org/files/PhRMA%-
202009%20Profile%20FINAL.pdf). Between 1975 and 1979, the five-year sur-
vival rate for cancer was just 50%; by 2000, survival had risen to 67%. Survival is
increasing dramatically for many forms of cancer. The rate of five-year survival
went up 21% for breast cancer, 42% for prostate cancer, 28% for colon and
rectum cancer, and 25% for lung and bronchial cancer.
Drug discovery is a complex multivariate process, but the basic requirements
for orally administered NCEs include novelty and patentability, intrinsic
potency, oral bioavailability, no toxicological effects in humans, and a sig-
nificant advantage over existing accepted therapies (if applicable). A schematic
representation of the drug discovery process in the United States is shown in
Figure 1.1.
Although it is possible to predict, with varying accuracy, what a NCE will do
when orally administered to humans, the full potential of a NCE is not known
until it has been tested in clinical trials. Therefore, any investment made will not
yield any return until the NCE is on the market, which could be in the region of
ten years after patenting, and for the majority of compounds there will be no
return at all to offset the enormous costs of drug discovery and development.
Therefore, drug discovery is a high risk business with massive, long-term up-
front investments aiming at discovering the few blockbusters that are on the
market at any one time. In addition, the pharmaceutical industry is one of the
most research-intensive industries; in the United States, an average of 16% of
sales is spent on R&D, second only to the aerospace industry (www.nsf.gov/
statistics). The global pharmaceutical market is worth $553.4 billion in the top
ten markets alone (Table 1.1).
The top ten marketed drugs and their revenue between June 2007 and June
2008 are shown in Table 1.2; they account for a total of $67.4 billion,1 which is
only 12.2% of total sales in the top ten markets.
Among the top ten drugs, Pfizer’s Lipitor is by far the biggest seller, $5.5
billion ahead of a cohort of three drugs, Plavix, Nexium and Serentide with
sales of $8.3, $7.7 and $7.5 billion, respectively. The top ten therapies are shown

Figure 1.1 Schematic representation of the drug discovery process with typical time
frames and attrition rates from drug discovery to FDA approval (adapted
from www.phrma.org/files/PhRMA%202009%20Profile%20FINAL. pdf).
Drugs and their Structural Motifs 3

Table 1.1 Top ten pharmaceutical markets worldwide and their revenue for
the period June 2007 to June 2008.
Country Sales June 2007–June 2008 ($ billions) Share of global sales (%)

USA 288.9 40.7

Japan 63.5 8.9
France 42.5 6.0
Germany 40.4 5.7
Italy 25.1 3.5
UK 23.5 3.3
Spain 21.6 3.0
Canada 19.1 2.7
China 16.8 2.4
Brazil 11.9 1.7
Total 553.4 77.9

Table 1.2 Top ten marketed drugs worldwide for the period June 2007 to June
2008.
Sales June
2007–June
Name Compound Marketer Indication 2008

Lipitor Atorvastatin Pfizer Hypercholesterolemia 13.8

Plavix Clopidogrel Bristol-Myers Atherosclerotic 8.3
Squibb events
Nexium Esomeprazole AstraZeneca Acid reflux disease 7.7
symptoms
Serentide Fluticasone GlaxoSmithKline Asthma 7.5
and
salmeterol
Enbrel Etanercept Amgen Rheumatoid arthritis 5.6
Seroquel Quetiapine AstraZeneca Bipolar disorder, 5.1
schizophrenia
Zyprexa Olanzapine Eli Lilly & Co. Schizophrenia 5.1
Risperdal Risperidone Johnson & Schizophrenia 5.0
Johnson
Remicade Infliximab Centocor Crohn’s disease, 4.7
rheumatoid arthritis
Singulair Montelukast Merck & Co. Asthma, allergies 4.6
Top ten products 67.4

in Table 1.3 and account for 36.5% of global sales. The annual sales figures
indicate that oncologics are by far the biggest revenue stream for the phar-
maceutical industry, followed by lipid regulators. Interestingly, Pfizer’s Lipitor
alone accounts for almost half of lipid regulator sales.
Historically, big pharmaceutical companies delivered. The secret of their
success was simple: pharmaceutical companies brought a huge number of
innovative products to the market that genuinely helped sick people, and so
4 Chapter 1
Table 1.3 Top ten drug therapies and their annual global sales for the period
June 2007 to June 2008.
Therapy Sales June 2007–June 2008 Share of global sales(%)

Oncologics 45.8 6.4

Lipid regulators 34.2 4.8
Respiratory agents 30.7 4.3
Acid pump inhibitors 26.7 3.8
Antidiabetics 26.0 3.7
Antipsychotics 22.4 3.1
Angiotensin-II antagonists 21.6 3.0
Antiepileptics 16.5 2.3
Autoimmune agents 14.8 2.1
Total 259.1 36.5

60
New molecular entities
50 Biologic license application

40

30

20

10

0
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Figure 1.2 New molecular entities (NMEs) and biologic license applications
approved by the US FDA’s Center for Drug Evaluation and Research
between 1996 and 2008.

were readily prescribed, which generated solid sales.2 Even during times of
economic hardship, drugs continued to be an essential purchase. During this
flourishing period from the mid-1980s to the beginning of this decade, major
drug companies routinely generated double-digit growth in sales year after
year.
However, the pharmaceutical industry’s investment in R&D has also risen
steeply over the last 20 years, with R&D spending of $47.9 billion in 2007
compared with $26 billion in 2000 and $8.4 billion in 1990, and an average cost
of $1.3 billion for bringing a new drug to market—an increase of 65% since
2000 (www.phrma.org/files/PhRMA%202009%20Profile%20FINAL.pdf).3
Despite this increased investment in research and development, the number of
new molecular entities (NMEs) has not increased in line with rising investment;
in fact it declined between 1996 to 2008 from 54 to 21 (Figure 1.2).4–15
Drugs and their Structural Motifs 5
Figure 1.2 could suggest that there has been a significant decline in inno-
vation rates in the pharmaceutical industry over the last decade. The reasons
for this decline have been reviewed extensively16 and several causes have been
indicated as contributors to the R&D decline.17 Among these are for example
submaximal optimisation of resources and the inability to control costs as well
as negative impact of mergers and acquisitions, which have been grouped
together as factors internal to R&D. Alongside these, external reasons for the
decline include evolving healthcare, regulatory burden, lack of regulatory
harmonisation as well as changes in tolerance for risk.17,18 Looking back over
recent decades, total approvals by the US Food and Drug Administration
(FDA) reached a record high of 381 entities in the decade between 1995 and
2004 compared with the two previous decades (241 in the decade 1985–1994
and 190 in the decade 1975–1984). Thus, it would appear that a myopic focus
on near-term performance has given rise to a perception that bears very little
relationship to the actual innovation rates of the pharmaceutical industry in the
last decade.19
However, the issue of high attrition rates in drug discovery and development
still remains, without which the innovation rates would be even higher and,
potentially, would keep better track with the enormous increases in R&D
investment. Only about 11% of compounds entering clinical development ever
reach the market, being withdrawn for reasons such as efficacy (25%), tox-
icology (24%), clinical safety (12%), drug metabolism and pharmacokinetics
(DMPK, 8%), formulation (1%) and portfolio-related and other reasons
(30%).20 Therefore, out of the 70% of failures caused by specific effects, the
majority of 61% can be attributed to lack of efficacy, toxicology and clinical
safety, whereas DMPK (physicochemical properties, or drug likeness, of the
drug candidate itself) accounts for only 8% of attrition.21 However, the actual
proportion may be higher since some reported attrition, which was attributed
to lack of efficacy, might be due at least in part to poor DMPK.20 A similar
proportion of 7% was discussed as having inappropriate absorption, dis-
tribution, metabolism and excretion (ADME) properties among NCEs between
1964 and 1985.22 In addition, apparently only about 30% of marketed pre-
scription drugs produce revenues that match or exceed average R&D costs.22
The apparent decrease in productivity in the entire pharmaceutical industry
has put enormous financial pressures on individual companies and their share
price—one of the measures of confidence of investors in future profitability.23
Although the underlying reasons for this decline in productivity are complex,
many factors have been suggested, such as for example increasing clinical
development costs, FDA approval standards and political pressures on drug
pricing.20
One of the key reasons for the decline in productivity is without doubt the
high rate of attrition at all stages of the drug discovery process from failures in
the early pre-clinical stages to the very expensive late stage failures in the clinic
or even post-launch. Although exact figures on attrition in drug discovery are
difficult to derive due to the sparseness of publicly available data, it is clear that
success rates of discovery projects over the last decade, perhaps in part due to
6 Chapter 1
the very high attrition rates, have not been able to match expectations in terms
of productivity targets.
Therefore, attempts to reduce attrition early in the drug discovery process
have been a major focus over the last decade. During that time, the application
of guidelines linked to the concept of drug-likeness (in particular absorption)
such as the ‘rule of five’24 (see Section 2.1.1 for details) has gained wide
acceptance25 as an approach to reducing attrition in drugs.26,27 However,
despite this acceptance, an analysis of recent trends revealed that the physical
properties of molecules that are currently being synthesised in leading drug
discovery companies differ significantly from those of recently discovered oral
drugs and compounds in clinical development.26 This was particularly notable
for lipophilicity, where the consequences of a significant increase include a
greater likelihood of lack of selectivity and attrition in drug development.26
Physicochemical properties of molecules are completely under the control of
medicinal chemists and can be easily calculated for very large numbers, in some
cases for hundreds of thousands of designed structures prior to synthesis.
Close monitoring of physical properties during a drug discovery programme
and compound series selection based on orthogonal attrition risks as indicated
by compound properties and chemical scaffold may provide the medicinal
chemist with opportunities to significantly reduce attrition rates, which are
currently estimated at 93–96%.28
In this chapter, we focus on the relationship of molecular properties and
functional groups of compounds on their interactions with biological targets,
which can potentially impact on their pharmacological profile and their
potential attrition risks.29

1.2 Launched Drugs

The relationship between chemistry, biology and medicine has been a
remarkably productive one over the past century30 since Paul Ehrlich pioneered
the idea of systematically searching for drugs. By screening just over 600 syn-
thetic compounds, Ehrlich discovered arsphenamine (Salvarsan)31 in 1909
which, at the time, greatly improved the treatment of syphilis. Since then, there
have been a large number of very significant breakthroughs, for example
penicillin (1941), cortisone (1949), benzodiazepines (1960), beta blockers
(pronethalol, 1967), anti-histamines (cimetidine, 1977), ACE inhibitors (cap-
topril, 1981), insulin (1982), statins (lovastatin, 1987), HIV (zidovudine, AZT,
1987), COX-2 inhibitors (celecoxib, 1999) and kinase inhibitors (imatinib,
2001). Between 1983 and 2007, 907 different NCEs were approved as drugs.32
In their elegant analysis of drug targets, Overington et al. found the number
of unique launched drugs to be 1357,33 of which 1204 were considered to be
‘small-molecule’ drugs. Of those, 803 can be administered orally. The analysis
included data up to the end of 2005; the number of small molecule drugs has
since increased by 21 in 200611 and 19 in 200712, resulting in a total number of
launched small molecule drugs of 1244. Of the 1204 drugs used in the 2005
Drugs and their Structural Motifs 7
Table 1.4 Molecular targets of FDA approved drugs.
Number of
Class of approved drugs Species molecular targets

Targets of approved drugs Pathogen and 324

human
Human genome targets of approved Human 266
drugs
Targets of approved small-molecule Pathogen and 248
drugs human
Targets of approved small molecule Human 207
drugs
Targets of oral small molecule drugs Human 186

analysis, 1065 were assigned protein molecule targets believed to be responsible

for the efficacy of the drug.33 The data is summarised in Table 1.4.
In the following we discuss the target space and chemical space of drugs
separately, but it should be pointed out that these are not separate ‘spaces’ but
are interlinked through common, complementary properties. These are, for
example, the steric complementarity of a small molecule with a binding site—
not only in terms of shape but also in terms of electrostatic interactions and
physicochemical properties. This principle of complementarity of chemical and
biological space has been discussed extensively elsewhere34 and will not be
expanded upon as part of this chapter.

1.2.1 Target Space of Launched Drugs

The first analysis of the draft sequence of the human genome resulted in an
estimate of B31 000 protein-coding genes;35 the current estimate has dropped
to 22 287 genes.36 It is generally estimated that 3000 of these are druggable.37,38
The relationships of drugs and their targets has been studied extensively.33,39 In
this context, the term ‘chemogenomics’, described as ‘the discovery and
description of all possible drugs for all possible drug targets’, has been coined.40
Chemogenomics has been identified as a new approach that can guide drug
discovery based on integration of all information within a protein family, for
example sequence, structure–activity relationship (SAR) data and protein
structure. This allows very efficient cross-SAR analysis and exploration
between targets that share small molecule inhibitors, leading to identification of
new lead structures.41 Chemogenomic approaches to drug discovery effectively
explore the observation that similar receptors bind similar ligands and have
shifted traditional receptor-specific studies towards a more cross-receptor view
of pharmaceutical research.42 Chemogenomic approaches have been exempli-
fied recently for cardiovascular diseases43 as well as for kinases.44
The druggable genome has been initially quantified as 483 small molecule
drug targets,45,46 with a later figure suggesting that number to be between 600
and 1500.37,40 However, it has also been shown that out of these potential drug
8 Chapter 1
targets, only a total of 324 drug targets account for all classes of approved
therapeutic drugs. This number is reduced further to only 186 for targets of
approved oral small molecule drugs. The gene family distribution of current
drugs is shown in Figure 1.3.33
The concept of druggability, which has been used widely in recent years,
postulates that since the binding sites on biological molecules are com-
plementary with their ligands in terms of volume, topology and physico-
chemical properties, then only certain binding sites on putative drug targets will
be compatible with high-affinity binding to compounds with drug-like prop-
erties.47 The extension of this concept to the whole genome analysis led to the
identification of the druggable genome. This is the expressed proteome pre-
dicted to be amenable to modulation by compounds with drug-like proper-
ties.37 However, it needs to be noted that the meaning of the term druggability
has broadened beyond its generally accepted definition to signify very different
aspects along the discovery, development and clinical pipeline.48
A very useful categorisation of druggability has been published
by Sugiyama49 which differentiates between the druggable genome and
druggable proteins, and the druggability of compounds in terms of their

35

30

25

20

15

10

0
Rhodopsin-like
receptors

Ligand-gated ion

Penicillin-binding

Myeloperoxidas

Neurotransmitter
topoisomerase

Fibronectin type

Cytochrome

Others
Voltage-gated

symporter family
ion channels
Nuclear

Type II DNA

P450
GPCRs

channels

protein

e-like

III

Figure 1.3 The gene family share as a percentage of all FDA-approved drugs for the
top ten families. Beyond the ten most commonly drugged families, there
are a further 120 domain families or singletons for which only a few drugs
have been successfully launched. The data is based on 1357 dosed com-
ponents from 420 000 approved products, FDA, December 2005.
Drugs and their Structural Motifs 9
molecular properties. The usefulness of the concept of druggability from a
medicinal chemistry standpoint has been summarised by pointing out that the
rule of five (Ro5) and its extensions have generated awareness about the
importance of pharmacokinetic parameters for drug discovery and develop-
ment. In addition, the concept of druggability has led to the realisation that
there may be whole families of proteins for which it is either extremely chal-
lenging or impossible to design small molecules with acceptable oral
bioavailability.50
Another concept related to druggability is ligand efficiency,51,52 which gen-
erates a quantitative relationship between drugs and their biological targets,
and is defined as the binding energy per non-hydrogen atom in a particular
molecule. This concept can be very useful for lead selection by normalising
binding energy for molecular weight, but also for differentiation between gene
families that have a high or low probability of binding Ro5-compliant small
molecules based on an analysis of experimentally determined ligand-binding
energies for a particular target. These concepts of maximal affinity and ligand
efficiency have been developed further into a computational approach to pre-
dict druggability.53
In the past ten years or so, expectations in the pharmaceutical industry
have been raised as many companies have invested significantly in high-
throughput technologies that would make use of information derived from
the sequencing of the human genome.54 Therefore, it would seem that
companies are now well-placed to take advantage of the discovery of new
targets that have appeared in the post-genomic era. However, there appears to
be a reduced likelihood of delivering a preclinical drug development candi-
date against a new target, which could lead to a temptation to concentrate on
more established targets to reduce risk in current development portfolios.54
More recent in silico approaches such as high-throughput electronic biology
may help in identifying, for example, previously unknown complex relation-
ships between targets as well as compounds and targets in biological pathways
on a large scale in order to support many parallel work streams in a drug
discovery portfolio.55

1.2.2 Chemical Space of Launched Drugs

Chemical space, like target space and druggability, is another concept used
frequently in the literature.56,57 Defined as the number of synthesisable small
molecule compounds, it is believed to be in the order of 1060 individual
drug-like molecules. Analyses of chemical space have been published, for
example, for natural products58 and antibiotics.59 Recently, drug–target
interaction networks have been defined from the integration of chemical and
genomic spaces for four target classes—enzymes, ion channels, G-protein
coupled receptors (GPCRs) and nuclear receptors. The results indicate that
there is significant correlation between drug structure similarity, target
sequence similarity and the drug–target interaction network topology.60 Below
10 Chapter 1
we focus particularly on the molecular and pharmacological properties of
drugs.

1.2.3 Molecular Properties of Launched Drugs

The physicochemical properties of molecules are important factors for phar-
macological profiles and attrition risks. The distribution of the molecular
properties of small-molecule launched drugs has changed little in the past 25
years, despite changes in the types of clinical indications for which drugs have
been developed as well as in the range of targets.61
There have been a number of recent publications on the analysis of mole-
cular properties of oral drugs,62,63 as well as on how molecular pro-
perties influence oral drug-like behaviour.64 A number of researchers have
highlighted an upward trend in molecular weight (MW) and lipophilicity
between older and newer drugs,65,66 which has also been linked to attrition rates
in clinical trials.66
The first systematic analysis of molecular properties relating to compound
attrition resulted in the rule of five.24 This states that a molecule is less likely to
be absorbed if its molecular weight is above 500, its clogP is above 5, and the
number of hydrogen bond acceptors and donors is more than ten and five,
respectively. Although Ro5 parameters are interrelated, with cLogP being an
additive property dependent on fragment values and therefore directly related
to molecular weight, as well as hydrogen bond acceptor and donor function-
alities of fragments, its simplicity and ease of calculation even for very large
numbers of molecules has made it a very influential indicator for the likelihood
of compound absorption, which is also often referred to as drug-likeness.26
It has been shown that the distribution of molecular weight and lipophilicity
between marketed drugs and development phase I oral drugs is significantly
different, with marketed drugs on average having a lower molecular weight
compared with compounds in the discovery and development phases. It has
also been shown that molecular properties of compounds vary significantly
between gene families.63 A graphical representation of the average molecular
weight for oral drugs with respect to gene families is shown in Figure 1.4.63
Drugs in the protease and GPCR-peptidic families are characterised by
significantly higher average molecular weight, while those in the ion channel
family have lower average molecular weight. Drugs in the GPCR-lipid, GPCR-
peptidic and nuclear hormone receptor (NHR) families have significantly
higher cLogP. Drugs in the GPCR-peptidic and protease families have more
acceptors, while those in NHR families have fewer acceptors. In only four
families—CYP450, kinase, phosphodiesterase (PDE) and transporter—are the
mean values of all four properties statistically similar to those of all oral drugs.
Similar observations can be made while looking at the percentage of drugs in
each family passing all four (or three of four) original Lipinski rules. GPCR-
peptidic, GPCR-lipid and protease family targeted drugs have the lowest Ro5
compliance.63
Drugs and their Structural Motifs 11

700

600

500

400

300

200

100

0
P450

Transporter

GPCR aminergic

Ion channel

NHR

Other

PDE

Kinase

GPCR lipid

GPCR peptidic

Protease
Figure 1.4 Average molecular weight of oral drugs for gene families.

Although the data seems to indicate that overall molecular weight of dis-
covery and development compounds has increased over recent decades, it has
been suggested that the upward trend can be explained largely by variations in
the target portfolios of pharmaceutical companies. Most notably, a significant
decrease of biogenic amine GPCR drugs in the recent decades (43% to 28%)
and increases in protease and peptidic GPCR targeted drugs may explain much
of the overall molecular weight trend. Variation in properties over time for a
given family may result from varying pharmaceutical interest in its members
(e.g. serine proteases, metalloproteases, etc.).63
The central assumption in the applicability of standard rules for drug likeness
is that the target of interest requires molecular properties similar to those of the
average drug. Since bioavailability results from the interactions of drugs with
the same biological systems (i.e. membranes in the gastrointestinal tract as well
as metabolic enzymes like P450s and transporters), it is possible that well-
defined ranges of molecular properties can account for favourable interactions
with those systems. For certain proteomic families, application of standard
rules of drug likeness could bias the composition of corporate screening col-
lections away from the molecular properties needed for achieving high affinity
(e.g. for protein–protein interactions).67 The need to balance bioavailability
and affinity suggests that modified rules of drug likeness need to be adopted for
certain target classes.63
12 Chapter 1
The properties showing the clearest influence on the successful passage of a
candidate drug through the different stages of development are molecular
weight and lipophilicity. Statistical analysis shows that the mean molecular
weight of orally administered drugs in development decreases on passing
through each of the different clinical phases and gradually converges toward
the mean molecular weight of marketed oral drugs. It is also clear that the most
lipophilic compounds are being discontinued from development.66 This work
supports Lipinski’s findings that there are limiting factors to the molecular
weight and lipophilicity of a candidate drug that are reflected in the current
physicochemical property profiles of the marketed oral drug data set. This
study also suggests that these limiting values of physicochemical properties are
not historical artefacts but are under physiological control.66 In addition, an
analysis of the difference between drugs and their original lead compounds
shows that, for the majority of cases, only small structural changes are made in
the lead to drug process.68
Therefore, it can only be advantageous if a screening collection already
contains drug-like compounds with the right physical properties which carry a
lower risk of attrition during drug development. This has been developed into
the lead-like paradigm, which states that lead compounds need to be left-shifted
in terms of molecular weight and lipophilicity compared to drugs in order to
allow for the additional molecular weight and lipophilicity which has histori-
cally been added in the lead to drug process.69 Recent studies have shown that
molecular weight and log D are the most important factors in determining the
permeability of drug candidates.70 It has also been shown that the log D limits
are dependent on molecular weight, and rules have been defined for log D limits
required to achieve 450% chance of high permeability for a given MW band
(Table 1.5).
Although both molecular weight and log D have been linked to permeability,
log D may be the more important factor since it is an expression of multiple
molecular properties such as hydrogen bond donors and acceptors, lipophilicity
as well as dipole and polarisability, which are linked to physicochemical events
like permeability and binding.

Table 1.5 Permeability rules defining LogD limits

required to achieve 450% chance of
high permeability for a given molecular
weight band
Molecular weight AZLogD

o300 40.5
300–350 41.1
350–400 41.7
400–450 43.1
450–500 43.4
4500 44.5
Drugs and their Structural Motifs 13
Differentiation between drugs and non-drugs on the basis of molecular
properties has been reported in the literature, either based on statistical
approaches71 or analysis of structural features72 with an accuracy of between 71
and 92%, respectively. However, these approaches are mostly limited to
oral drugs and it has been shown that, for example, inhaled drugs reside in a
region of molecular property space which is very different from that of oral
drugs.73
In addition to predicting druggability for biological targets which can bind
small molecules, there has been considerable interest in targeting protein–
protein interactions with small molecules. Protein–protein interactions are
highly attractive for drug discovery because they are involved in a large number
of disease pathways where therapeutic intervention would bring widespread
benefit. Recent successes have challenged the widely held belief that these
targets are ‘undruggable’.74,75 Targeting interfaces between proteins has huge
therapeutic potential, but discovering small-molecule drugs that disrupt pro-
tein–protein interactions is an enormous challenge, which is being faced with
the support of for example bioinformatics approaches.76 This vast new field of
drug discovery has enormous potential but is outside the remit of this chapter;
for further reading we refer you to recent successes reported in the
literature.67,74,75

1.2.4 Polypharmacology
The understanding of pharmacological space is one of the fundamental aspects
of drug discovery, relating to off-target activity and in turn to compound
attrition. Pharmacological space has been mapped recently by Paolini et al.
through large-scale data integration of proprietary and published screening
data.39 They have assigned 2876 targets to protein sequences from 55 organ-
isms, with biologically active chemical tools for 1306 proteins. After removing
redundancies in the mammalian genes due to orthologs among species, 836
genes could be unambiguously identified in the human genome for which small-
molecule chemical tools with biological activity of IC50o10 mM have been
discovered. This number drops to 529 when a perhaps more realistic threshold
of 100 nM is applied (Table 1.6).
Of the pharmacological targets selected, 158 human proteins have been
identified as the primary modes-of-action for approved small-molecule drug
targets, with oral small-molecule drugs primarily targeting only 141 human
proteins.
A key question in global pharmacological space is how extensive is pro-
miscuity, which is defined as the specific binding of a chemical to more than one
target. Considering each pair of targets in turn, if two proteins both bind to the
same ligand, they can be considered as interacting in chemical space even if they
have no other interaction in physical space or similarity in sequence space. The
concept of ‘target-hopping’, where chemical matter for one target can be
considered as the basis for leads or tools for another target, has historically
14 Chapter 1
39
Table 1.6 Pharmacological target space.
Human targets with Human targets with
Gene family o10 mM binding affinity o100 nM binding affinity

Protein kinases 105 83

Peptide GPCRs 63 42
Transferases 49 24
Aminergic GPCRs 35 35
GPCRs (class A and 44 32
others)
Oxidoreductases 40 25
Metalloproteinases 44 35
Hydrolases 36 21
Ion channels (ligand-gated) 29 22
Nuclear hormone 24 19
receptors
Serine proteases 30 21
Ion channels (others) 18 11
Phosphodiesterases 19 18
Cysteine proteases 16 13
GPCRs (class C) 10 6
Kinases (others) 12 5
GPCRs (class B) 7 3
Aspartyl proteases 7 4
Miscellaneous 139 63
Enzymes (others) 109 47
Total 836 529

been an extremely fruitful method of drug discovery.77,78 Of all the 276 122
active compounds found in the database used by Paolini et al.,39 65% have
recorded activity for one target, whereas 35% are reported to hit more than one
target.
The observed polypharmacology interaction network for human proteins
was mapped to navigate polypharmacology relationships between targets. The
entire protein interaction network consists of 700 proteins connected by 12 119
interactions for all compounds below the affinity threshold of IC50 10 mM, and
with a difference in affinity of up to three orders of magnitude between two
targets. Promiscuity can be considered from the perspective of both the com-
pound and the pharmacological target, to measure compound selectivity and
target overlap.79,80 Table 1.7 shows the top ten promiscuous targets taken from
ref. 39.
Although different definitions of promiscuity result in different rankings, the
same target classes (aminergic GPCRs, cytochrome P450s and protein kinases)
appear at the top regardless of the method used in the analysis (Table 1.6).
Aminergic GPCRs and protein kinases exhibit the greatest intra- as well as
inter-gene family promiscuity. The data set used for this work is a sparse
matrix, since activity data for each compound is mostly limited to only a few
targets. There are indications that molecular properties and the potentially
Drugs and their Structural Motifs 15
39
Table 1.7 Top ten most promiscuous targets.
Number Protein Gene family

1 Cytochrome P450 1A2 Enzyme

2 5-hydroxytryptamine 2C receptor Aminergic GPCR
3 Cytochrome P450 3A4 Enzyme
4 D2 dopamine receptor Aminergic GPCR
5 SRC kinase Protein kinase
6 5-hydroxytryptamine 1A receptor (5HT1A) Aminergic GPCR
7 5-hydroxytryptamine 2A receptor (5HT2A) Aminergic GPCR
8 D4 dopamine receptor Aminergic GPCR
9 Alpha-1A adrenergic receptor Aminergic GPCR
10 5-hydroxytryptamine 7 receptor (5HT7) Aminergic GPCR

resulting promiscuity play an important role associated with in vivo tox-

icological outcomes.81 It appears that the statistical odds for toxicity are sig-
nificantly higher for compounds with a clogP 43 and a topological polar
surface area (TPSA) of less than 75Å2.81 This indicates that lipophilicity is
potentially linked to toxicity, which is in agreement with the perception that
lipophilic binding is non-specific, whereas polar binding is related to specificity
and therefore selectivity. This is in contrast to molecular weight, where low
molecular weight and complexity will increase promiscuity and lead to lower
selectivity.69
An alternative way to assess polypharmacology has been introduced by Fliri
et al. with the concept of ‘biological spectra’ based on the BioPrints data set.82
This enables the comparison of compound similarity at the biological level
rather than at the level of chemical structure similarity. The underlying idea is
that compounds with similar bio spectra are by definition similar, even though
their scaffold similarity might be very low. The relevance of this approach has
been demonstrated with examples that some pharmacology is associated with in
vivo clinical effects.83 In vitro–in vivo associations have been established from
experimental and predicted data between M1 activity and tachycardia, H1
activity and somnolence, as well as D1 activity and tremor. Further evidence is
provided by examples from corticosteriods, adrenergics, sedatives84 as well as
ligands of the dopamine receptors D2–D4.85

1.3 Drugs Bound to their Targets

In order to understand the interactions at a molecular level of drugs with their
targets, we have analysed examples of experimental protein–ligand structures
of some marketed drugs. Experimental structures for a large number of drugs
in their pharmaceutically relevant targets (see Table 1.8) are available in the
Protein Data Bank (PDB) (www.rcsb.org/pdb/home/home.do). These struc-
tures are a manifestation of the increasing value and application of structural
biology in drug discovery.86
 16
Table 1.8 Experimental protein–ligand structures for marketed drugs and their targets available in the Protein Data Bank
PDB
Drug name Structure Target Indication code

Oseltamivir (Tamiflu, prodrug) Neuraminidase Influenza 1l7f

O
O O
HN
O
H2N

Zanamivir (Relenza) HO OH Neuraminidase Influenza 3b7e

H
HO O OH
HN
O O
HN
NH
H2N

Imatinib (Gleevec) N Bcr-abl kinase Cancer 3gvu

HN N N

HN

O N
Chapter 1

N
Argatroban Thrombin Anti-coagulant 1dwc

HN

O
NH2 O S
NH
H 2N N

N O
O
Drugs and their Structural Motifs

OH

Gefitinib (Iressa) F EGFR kinase Cancer 2ity

O HN Cl
N O
N

O N

Erlotinib (Tarceva) O N EGFR kinase Cancer 1m17

O
O N
O
HN

Sildenafil (Viagra) O Phosphodiesterase 5 (PDE5) Erectile dysfunction 1udt

HN N
O O N
S
N N
N
O
17
 18

Table 1.8 (continued )

PDB
Drug name Structure Target Indication code

Vardenafil (Levitra) O Phosphodiesterase 5 (PDE5) Erectile dysfunction 3b2r

O O HN
N
S N
N N
N
O

Tadalafil (Cialis) O Phosphodiesterase 5 (PDE5) Erectile dysfunction 1udu

N

N N
H
O

O
O

Donepezil (Aricept) Acetylcholinesterase Alzheimers 1eve

O N

O
O
Chapter 1
Celecoxib (Celebrex) (structure of ana- Cyclooxygenase 2 Inflammation, pain, 6cox
logue SC-544) arthritis

F
N F
N F
H2N
S
O O

Diclofenac (Voltaren) Cl Cyclooxygenases Inflammation, pain, 1pxx

arthritis
NH
Drugs and their Structural Motifs

Cl O

OH

Ibuprofen Cyclooxygenases Inflammation, pain, 1eqg

O arthritis
OH

Flurbiprofen (Ansaid) Cyclooxygenases Inflammation, pain, 1eqh

OH arthritis
O

Aspirin (structure of acetyl-salicylic OH Cyclooxygenases Inflammation, pain 1pth

acid)
O

O
19

O
 20
Table 1.8 (continued )
PDB
Drug name Structure Target Indication code

Pindolol (structure of analogue b-adrenergic receptor Hypertension 2vt4

Cyanopindolol) OH
N
H
O

N
H

Aciclovir (Zovirax) O Herpes simplex type-1 thymi- Herpes 2ki5

N dine kinase
HN

N OH
H2N N
O

Ganciclovir O Herpes simplex type-1 thymi- Herpes 1ki2

N dine kinase
HN

H2N N N
O OH

OH
Penciclovir O Herpes simplex type-1 thymi- Herpes 1ki3
N dine kinase
HN

H2N N N
OH
Chapter 1

OH
Ritonavir HIV-1 protease HIV 1rl8

S N
N S

N
O O
O HN
HN OH
HN
O
Drugs and their Structural Motifs

Saquinavir O NH2 HIV-1 protease HIV 2nmw

H
N
N
O
HN O
OH

N
O

NH

Indinavir HIV-1 protease HIV 1hsg

O H
N

HN N
O HO
N
OH
21

N
 22
Table 1.8 (continued )
PDB
Drug name Structure Target Indication code

Nelfinavir (Viracept) HIV-1 protease HIV 1ohr

S
O
HO OH
N
H H
N H
N
H
O

Amprenavir (Agenerase) HIV-1 protease HIV 1t7j

H2N
OH
H
N N O
S
O O O O

Lopinavir HN HIV-1 protease HIV 1mui

O N
O

HN

HO
O
HN
O
Chapter 1
Atazanavir (Reyataz) HIV-1 protease HIV 2aqu
N

O OH O
H H
N N N O
O N N
H H
O O

Tipranavir (Aptivus) HIV-1 protease HIV 2o4l

Drugs and their Structural Motifs

F O
F
F N HO O
H
N
S
O O

Darunavir O O HIV-1 protease HIV 3d1z

S
N NH2

HO

NH
O H
O

O O
H
23
 24
Table 1.8 (continued )
PDB
Drug name Structure Target Indication code

Nevirapine (Viramune) O HIV reverse transcriptase HIV 1s1u

NH

N N N

Efavirenz (Sustiva) H HIV reverse transcriptase HIV 1fk9

N O

O
Cl
F
F F

Methotrexate O COOH Dihydrofolate reductase Cancer 4dfr

NH2 N
H
N
N N HOOC

H2N N N
Chapter 1
Drugs and their Structural Motifs 25
1.3.1 Comparison of Binding Sites of Drugs Bound in their
Biological Targets
One of the concepts discussed above is druggability, which relates to the like-
lihood of success for discovery of a drug based on the molecular properties of a
small molecule ligand and the binding site properties of the receptor. The
druggability concept is also based on the assumption that there are particular
features of a binding site which enable it to bind small molecules with sufficient
potency in order to meet requirements for a drug.
Rules have been derived in order to quantify druggability87 and make pre-
dictions based on analysis of the receptor structure as to whether a target is
likely to be druggable.53 Some of the parameters linked to druggability are:53

 the degree to which the binding site is buried inside the protein;
 the curvature of the binding pocket;
 the topology of multiple pockets and their relative positions in the binding
site;
 lipophilicity;
 polarity;
 the ability to form hydrogen bonds.

Below is the comparison of three marketed drugs, Tamiflu, Sildenafil and

Iressa, bound to their respective receptors, neuraminidase, phosphodiesterase 5
and EGFR kinase (Figure 1.5).
In the binding site of neuraminidase, the whole of the ligand oseltamivir can
be seen in the view from solvent into the binding site. In contrast, the binding
site of PDE5 is deeper and much narrower, and it also encloses most of the
ligand. The binding site of EGFR kinase binding Iressa forms a relatively
narrow cleft compared to the binding sites of PDE-5 and neuraminidase, and
the ligand structure is flat in comparison to Sildenafil and oseltamivir with
quinazoline as the central template. Following the concept of druggability, it
would be fair to say that for example PDE-5 and epidermal growth factor
receptor (EGFR) kinase are druggable targets, whereas it would appear to be
more difficult to develop a small molecule drug for neuraminidase. This
example also highlights the difficulty with predicting druggability, since osel-
tamivir (Tamiflu) is a marketed drug acting on neuraminidase, as is Relenza.
Therefore, druggability as a concept is perhaps not very useful as a filter but
rather for pointing out potential opportunities for discovery.

1.3.2 Phosphodiesterase 5 (PDE5) Drugs

Figure 1.6 indicates how three drugs marketed against erectile dysfunction
(Sildenafil, Vardenafil and Tadalafil) bind to their target phosphodiesterase 5.
Sildenafil and Vardenafil are structurally very similar—in fact they only differ
by two atoms and have a very similar binding mode—whereas Tadalafil has a
molecular structure which is very different from the other two drugs but still is
26 Chapter 1

Figure 1.5 Tamiflu bound to neuraminidase (PDB 1l7f, left), Sildenafil bound to
phosphodiesterase 5 (PDB 1udt, right) and Iressa bound to EGFR kinase
(PDB 2ity, bottom). Carbon atoms are yellow, light blue and purple,
respectively; oxygen atoms are red, nitrogen atoms are blue and sulfur
atoms are yellow. Surfaces are coloured by atom type: carbon ¼ green,
nitrogen ¼ blue, oxygen ¼ red, sulfur ¼ yellow.

able to bind to the same binding site, occupying significantly different areas in
the receptor (e.g. such as at the bottom and to the right of Figure 1.6).
Therefore, it appears that although sildenafil and vardenafil may contain
substructures which bind strongly to phosphodiesterase, they are two very
different interpretations of the phosphodiesterase pharmacophore which can
both show significant levels of activity in the receptor sufficient for the desired
pharmacological effect.
This ability to bind diverse substructures which are expressing a similar
pharmacophore does not seem to be specific to PDE5, but is observed perhaps
in an even more striking example of cyclooxygenase as highlighted below.

1.3.3 Cyclooxygenase (Cox) Drugs

Cyclooxygenase binds a variety of ligands, a number of them with acidic
groups, but also neutral compounds. The overlay of ibuprofen, flurbiprofen,
diclofenac and the celecoxib derivative SC-544 shown in Figure 1.7 highlights
the structural diversity of Cox inhibitors. Interestingly, the acid groups of
ibuprofen and diclofenac bind in very different areas of the binding pocket, so
that the compounds show very different pharmacophores.
Drugs and their Structural Motifs 27

Figure 1.6 Overlay of experimental binding orientations for three marketed drugs for
erectile dysfunction: sildenafil (light grey), vardenafil (dark grey) and
tadalafil (white). Compound structures are shown in Table 1.8.

Figure 1.7 Overlay of ibuprofen (white), flurbiprofen (black), diclofenac (medium grey),
celecoxib analogue SC544 (light grey) and salicylic acid (dark grey). The acid
groups of diclofenac (top right) and ibuprofen, flurbiprofen and salicylic acid
(bottom) are 9.1Å apart. Comound structures are shown in Table 1.8.
28 Chapter 1
Although a CF3 group is considered to be a bioisostere for an acid group, in
this case the CF3 group overlaps with the methyl groups of ibuprofen and
flurbiprofen, rather than the acid groups, and is therefore a bioisostere for a
lipophilic rather than a polar group. A more detailed analysis of the binding of
Cox inhibitors (particularly Cox-2 inhibitors) is presented in Chapter 5.1.2.

1.3.4 Classes of Drugs with High Structural Similarity

The two examples of the drug targets PDE5 and Cox discussed above raise a
number of questions relating to the ability of targets to accommodate and bind
a relatively diverse set of structural features in ligands. For example, ibuprofen
and flurbiprofen are structurally closely related, and were discovered during the
same screening programme in the 1960s when many analogues of phenylacetic
and phenoxyacetic acid were made in an attempt to identify potential herbi-
cides. Thus the similarity of ibuprofen and flurbiprofen can be explained by the
composition of the screening collection used rather than through a strong
receptor preference for a particular structure. Similarly, sildenafil (UK patent
1993, FDA approval 1998) and vardenafil (FDA approval 2003) are both
structural analogues of cyclic AMP, the natural substrate of PDE5.
In principle there are two compelling reasons why drugs have similar
structures. First, screening files contain structurally similar compounds due to
the fact that very often drug discovery programmes are driven by synthetic
accessibility rather than structural diversity, yielding similar chemical matter
through preparation of closely related analogues to study structure–activity
relationships. Secondly, many drugs are structurally related to the natural
substrate or inhibitor of the biological target receptor. Even though natural
substrates or inhibitors have been developed through millions of years of
evolution, their ability to bind to a certain receptor is not exclusive to a closely
related series of compounds, as has been shown above. This is the case parti-
cularly in the gene family of kinases, where a large number of scaffolds are able
to mimic adenosine triphosphate (ATP) and adenosine diphosphate (ADP),
which bind in the hinge region of the catalytic domain of kinases.

1.4 Privileged Substructures in Drugs

The concept that similar molecules act in a similar manner is a fundamental
principle of medicinal chemistry. The earliest reference in SciFinder to struc-
ture–activity relationship (SAR) was in 1899, describing the diuretic action in
relation to osmotic properties of sugars.88 The concept of SAR forms the basis
of analogue-based discovery89 and has been validated through decades of
empirical observations. However, as our ability to evaluate small molecules on
a relatively large number of targets in vitro expanded—a process called ‘pro-
filing’ or ‘secondary pharmacology’—there was a realisation that similar
molecules can have very different overall profiles, making the prediction of
secondary pharmacology very difficult. This observation of large changes in
Drugs and their Structural Motifs 29
pharmacology caused by rather small changes in molecular structure has also
been referred to as ‘activity cliffs’.90 Therefore, the similarity principle is a good
approximation for near-neighbours for a range of properties, but is not a
reliable concept even within the same chemical scaffold.
The concept of privileged structures corresponds to the smallest structural
subunit that has been encountered in several drugs or lead compounds which is
able to provide ligands for multiple families of drug targets. In the late 1970s,
Ariens et al. observed that many biogenic amine antagonists contain hydro-
phobic double ring systems as key structural elements.91 Other groups noted
that recurring molecular cores could exist across diverse drug targets.92 How-
ever, the term ‘privileged structure’ was first used by Evans et al. in reference to
their work with benzodiazepine ligands.93–95 Numerous other authors have
since identified other substructures that have also been observed in numerous
drug programmes. Notably work by Patchett and Nargund advanced the dis-
cussion around understanding the inherent reasoning behind some groups
appearing privileged by identifying properties in the substructures that facil-
itate their interactions with biomolecules, often via distinctly different inter-
actions to the respective endogenous ligands.96 These observations raised the
possibility that there existed preferred molecular scaffolds that have an inherent
tendency towards biological activity and that these groups could be modified to
provide ligands for a range of biological targets. In agreement with this
hypothesis, an analysis of all known drugs by Murko et al. in 1996 revealed that
the 5120 compounds in the Comprehensive Medicinal Chemistry (CMC)
database contain 1179 different frameworks; however 32 (3%) of those fra-
meworks accounted for 50% of all drugs.97 The analysis was purely based on
size and shape did not fully take into account atom types or bond order.
However, the existence of preferred molecular frameworks is still evident.
When atom type and bond order are included, a larger diversity of frameworks
results but again a large proportion of drug molecules (24%) are based on a
small number of molecular frameworks.92
Other groups have modified the definition of privileged structure to
encompass commonly occurring fragments within ligands that are promiscuous
within only a single target family.98 The motivation to identify such sub-
structures is derived from the need to avoid indiscriminate off-target activity.
More recently, however, this family specific concept has been re-categorised as
distinct from the original definition of a privileged structure and been termed
‘target family-directed masterkeys’ by Mueller.98
Recent discussions by Schnur et al. have reverted back to the original defi-
nition of privileged structures whereby the motif is observed across numerous
target families but by more distinctly defining differences between potentially
hazardous ‘frequent hit’ motifs that systematically appear in high throughput
screening (HTS) from those groups that occupy good ‘drug-like’ physico-
chemical space where the desired selectivity profile can be modulated via per-
ipheral structural modification.99,100
The concept of privileged structures has been applied to the planning of new
chemical libraries and has been associated with the application of
30 Chapter 1
98–99,101
computational methods and pharmacophore models. Furthermore,
fragmentation of bioactive molecules and drugs has permitted a more thorough
identification of relevant structural patterns that represent authentic biophores,
providing frameworks for the generation of new compound databases.
Numerous investigations have focussed on the identification of desirable pri-
vileged structural elements.
Two early studies examined methods to fragment compounds into core
structures with peripheral modification. Both studies identified substructures
with the intention of using the results to focus chemistry efforts for file
enrichment strategies, via chemically enabled libraries.97,102,103 In an alternative
approach, Fesik et al. investigated a new experimentally based method to
identified privileged structures by NMR-based screening of 10 000 selected
fragments across 11 diverse target proteins, and identified 12 privileged sub-
structures that appeared to bind to a range of targets with higher than average
probability.104 Naturally this work was only able to cover a fairly narrow
collection of core fragments and target proteins. In addition, no clear rationale
was obvious as to the reasoning for why the fragments bound to a particular
target. Siegel and Vieth studied a set of marketed drugs and pointed out that
there are a large number of drugs contained in other drugs either in their
entirety or as substructures.105 More recently, Sutherland et al. investigated the
relevance of chemical fragments as foundations for understanding target space
and activity prediction by decomposition of molecules into fragments and
comparing the similarity of those fragments and their relationship in target
space in an attempt to better understand cross-target activities.106 In addition,
the construction and use of small molecule libraries for fragment-based primary
screening based on privileged substructures has been discussed recently.107

1.4.1 Examples of Privileged Substructures

The concept of privileged substructures is highlighted below with examples of
marketed drugs for a variety of indications. Over the past two decades,
numerous structures and functionalities have been labelled as privi-
leged.92,98,100,101,108,109 For illustrative purposes, some key examples and
highlights from these analyses are discussed below covering classes of benzo-
diazepines, sulphonamide antibacterials, chemokine receptor 5 (CCR5)
antagonists, 3-hydroxy-3-methylglutaryl-coenzyme A CoA (HMG-CoA)
reductase inhibitors, cyclooxygenase inhibitors as well as kinase inhibitors.

1.4.2 Benzodiazepines
In the late 1980s, Evans and co-workers first defined the concept of privileged
structures.95 They worked on the development of novel non-peptidic chole-
cystokinin (CKK) receptor antagonists for the treatment of gastrointestinal
disorders (e.g. pancreatitis and gastroesophageal reflux) based on analogues of
the natural product asperlicin via structural modification of anxiolytic
Drugs and their Structural Motifs 31
110
benzodiazepine drugs such as diazepam. This work resulted in the devel-
opment of devazepide (MK-329) as the first non-peptidic benzodiazepine
antagonist, highly selective for cholecystokinin-1 (CCK-1) (IC50 0.8 nM).111–
113,93–95
Interestingly, this work involved key elements of structural planning
and molecular simplification from the natural product asperlicin. It was sug-
gested that both the benzodiazepine and tryptophan subunits of asperlicin are
key elements of the pharmacophore for molecular recognition be the CCK-1
receptors. Natural product guided development of CCK-1 antagonists is shown
in Figure 1.8.
Evans et al. recognised that this was not the first successful incorporation of
benzodiazepines into bioactive molecules. In fact the benzodiazepine motif
constitutes a broad class of neuroactive compounds acting as ligands to ion
channels and GPCRs. Notable examples of anxiolytic drugs of this class are
diazepam and lorazepam, which are ligands of central nervous system (CNS)
gabaergic receptors.114 In addition, there are the extensive numbers of CCK-1
diazepine containing ligands and numerous other applications as ligands to
other GPCRs such as k-opioid agonists like tifluadom115 for the treatment of
visceral pain, antithrombitic platelet activation factor (PAF) antagonists,116
analgesic and anti-inflammatory neurokinine (NK-1) receptor antagonists,117
and GPIIbIIIa receptor antagonists with antithrombitic profiles.118 In addition,
multiple classes of enzyme inhibitors have been developed that contain the
benzodiazepine unit; for example, HIV reverse transcriptase inhibitors such as
nevirapine,119 RAS-farnesyltransferase inhibitors for the treatment of cancer
(e.g. BMS-214662120). This diversity of bioactivity for benzodiazepines across a
broad range of classes of drug target led Evans to define this group as privileged.
A group of representative diazepine-containing drugs is shown in Figure 1.9.
Examining reasons for benzodiazepines to be privileged in this manner has
led to them being identified as b-turn peptidomimetics.121–123,92 The presence of
such structural motifs that are complementary to an array of primary and
secondary structural elements in proteins offers a potential explanation for the
promiscuous nature of the binding of many recurring scaffolds.

1.4.3 Arylsulfonamides and Drugs Derived from them

The following section describes the pivotal role of derivatives of sulphonamide
drugs in opening up new areas of pharmacology. These are also reviewed from
an ADMET perspective in Sections 5.1.1, 5.1.2 and 5.1.4 of Chapter 5.

1.4.3.1 Arylsulfonamides as Antibacterial Drugs

One of the best known classes of drugs are the antibacterial sulfonamides,
which were first patented in 1932,124 and were originally derived from azo dyes,
whose antibacterial properties were discovered in the early 20th century.125 The
first patent featured a red dye, sulfamidochrosoidine (Prontosil Rubrums)
(Figure 1.10), which made medical history by treating streptococcal infections,
32 Chapter 1

O
N
N

OH H H H
N

NH
N
O
O

Benzodiazepine

asperlicin
Tryptophan

CCK-1 IC50 = 1.4 μM

O
N

Cl N

diazepam

CCK-1 IC50 = >100 μM

O O
N

N
H
N HN

devazepide

CCK-1 IC50 = 0.8 nM

Figure 1.8 Natural product guided development of CCK-1 antagonists.

Drugs and their Structural Motifs 33

H O
N N

OH O
Cl N N HN

Cl F
S

Iorazepam tifluadom
BZD/Gabba k-opioid agonist

H O
N H O
N
CF3

N
N N N
N
O
CF3
NK1 antagonist nevirapine
RT inhibitor
HN
N O
N

O
N
N

Ar O
NC N CO2H
S GPIIbIIIa antagonist
O S
O
BMS-214662
RAS-farnesyl
transferase inhibitor

Figure 1.9 Representative diazepine-containing drugs.

including pneumonia and septicaemia, that were considered to be largely fatal.

For this extremely significant discovery, Gerhard Domagk, the director of the
IG Farbenindustrie laboratories at Elberfeld in Germany, was awarded
the Nobel Prize for Medicine in 1939. It was later discovered that it was indeed
the active metabolite sulfanilamide that was responsible for the antibacterial
activity. The first of the sulfanilamide analogues, sulfacefamide (Figure 1.11)
was marketed in 1938 and was used for many years to treat urinary infections.
Various derivatives of sulfanilamide were later marketed as drugs, their main
differentiation being around variation of the pKa of the sulfonamide to reduce
34 Chapter 1

NH2 NH2 NH2

N S H2N S
O O
NH2 N O O

Figure 1.10 Structures of sulfamidochrosoidine (Prontosil Rubrums, left), and its

active metabolite sulfanilamide.

N
HN HN
H2N S O H2N S N
O O
O O

HN

H2N S N
O
O

Figure 1.11 Structures of sulfacetamide (top left), sulfadiazine (top right) and
sulfadimidine.

the observed deposition of crystals in the kidneys. This problem of crystals

depositing in the kidneys was finally overcome by the introduction of sulfo-
namides that were more acidic and therefore more highly ionised in the urine.
The ideal sulfonamides required a pKa in the range of 6.5–7.5 in order to
balance the risk of kidney damage against rapid excretion. Two drugs, sulfa-
diazine and sulfadimidine (Figure 1.11), were in this range and both remain in
use for the treatment of meningitis.125
Since sulfonamides could be easily synthesised from commercially available
4-aminobenzenesulfonylchloride, many companies were researching in this area
resulting in the discovery of several classes of novel chemotherapeutic agents.
Although their side effects caused many problems in the clinic, they were
successfully exploited to provide oral antidiabetic drugs and valuable diuretics
as highlighted below.125

1.4.3.2 Arylsulfonamides as Antileprotic Drugs

A further discovery related to new indications for sulfonamides was made when
analogues of sulphanilamide were investigated for antibacterial properties. One
of the analogues, 4,4 0 -diaminodiphenylsulfone (Dapsone), was found to be 30
times more potent than sulphanilamide against streptococci; however, it was
Drugs and their Structural Motifs 35
126
also significantly more toxic. Therefore, several derivatives of Dapsone were
synthesised and analysed, and the close analogue glucosulfone was found to be
active against Mycobacterium tuberculosis. Since another mycobacterium, M.
leprae, is responsible for leprosy, both compounds were tested and found to be
the first effective treatments of the disease.127 Dapsone was also found to be
effective against leprosy and it is still in use as the standard antileprotic
sulfone.125

1.4.3.3 Arylsulfonamides as Diuretics

One of the effects of some sulfonamides was a reduction in carbon dioxide
binding power of the blood. This effect was discovered in 1940 and was
associated with the inhibition of the enzyme carbonic anhydrase,128 which
reversibly converts carbon dioxide into carbonate. However, only those sul-
fonamides in which both hydrogen atoms on the sulfonamide function were
unsubstituted were enzyme inhibitors. These included sulfanilamide and seven
sulfonamides that lacked antibacterial activity. A derivative of these, acet-
azolamide, which was far more potent than sulfanilamide, was subsequently
used as an orally active diuretic. Inhibition of carbonic anhydrase was also
turned to advantage in the treatment of glaucoma. By acting in the eye in the
same way as in the kidneys by reducing bicarbonate levels and the water
secreted with it, the build-up of pressure from excess fluid was overcome.
Acetazolamide still remains in use for this treatment today.125
A further potential use of sulfonamides was for the lowering of the con-
centrations of sodium and chloride in the body. It was believed that such a drug
would have the added bonus of being useful as an antihypertensive agent, since
clinicians in the 1950s were already beginning to believe that low salt diets were
effective in treating high blood pressure.125 The first carbonic anhydrase inhi-
bitor that increased chloride excretion was 4-sulfonamidebenzoic acid, receiv-
ing the approved name carzenide. This scaffold led to the discoveries of a
number of diuretics (Figure 1.12).
Chlorothiazide was one of the first of many thiazide diuretics. It effectively
made mercurial diuretics obsolete for the treatment of cardiac oedema asso-
ciated with congestive heart failure. Thiazide diuretics are still in use today for
the treatment of hypertension,125 and one of the analogues, bendrofluazide,
which had a longer duration of action, remains widely used in patients with
either mild heart failure or hypertension. The realisation that a second acidic
group in dichlorphenamide may be replaced with a carboxyl group, as long as
an appropriate substituent is present on the amino group, led to the intro-
duction of frusemide in 1962. It had a quicker onset of activity, which was more
intense and of shorter duration than that of other diuretics. Despite thiazides
being indicated for most patients requiring a diuretic, frusemide is widely
prescribed. Bumetanide is a more potent loop diuretic introduced ten years
after frusemide.129 Hoechst introduced the analogue piretanide when its patent
on frusemide expired.125
36 Chapter 1

HOOC Cl N

NH2 H2N NH
S S S

O O O O O O

H
F3C N
Cl NH

H2N NH
S S H2N
S COOH
O O O O
O O

N
HN
O
O
O
O
S COOH
S COOH O
O
NH2
NH2

Figure 1.12 Structures of diuretics: carzenide (top left), chlorothiazide (top right),
bendrofluazide (middle left), frusemide (middle right), bumetanide
(bottom left) and piretanide (bottom right).

1.4.3.4 Arylsulfonamides and Arylsulfonylureas as Antidiabetics

In the early 1940s, a clinical trial with an experimental sulfonamide, 2254 RP,
showed severe side effects which were linked to hypoglycaemia. This discovery
was confirmed in the mid-1950s with another experimental sulfonamide devel-
oped by C. H. Boehringer Company, which also induced hypoglycaemia.125 This
compound was introduced as an oral hypoglycaemic agent under the name
carbutamide. However, the compound showed unacceptable side effects in clin-
ical trials in the United States, even though the drug was already being used in
Europe. Meanwhile, Upjohn conducted a trial of Hoechst’s closely related
compound tolbutamide,130 which then received approval from the FDA for the
treatment of type 2 (non-insulin dependent) diabetes. Unlike carbutamide, it did
not have any antibacterial properties and therefore eliminated the risk of inducing
Drugs and their Structural Motifs 37
resistant bacteria. However, it had to be taken three times a day due to rapid
metabolism to the carboxylic acid. Soon after the introduction of tolbutamide,
Pfizer marketed chlorpropamide which did not have the metabolically sensitive
methyl group of tolbutamide and was about twice as potent.131 Other long-acting
sulfonylureas (structures are shown in Figure 1.13) have been marketed, including
highly potent agents such as glibenclamide (glyburide) and glipizide.125
All the compounds discussed above for various indications all contain the
common privileged substructure phenylsulfonamide. Although the reasons are
not completely understood, it can therefore be concluded that this substructure
appears to have physicochemical as well as binding properties which are
favourable in pharmacologically active compounds. One reason could be that
the sulfonamide group introduces polarity, which can place compounds in a
favourable region of chemical space with regards to drug properties like for
example for clearance. Therefore, despite the fact that phenylsulfonamide is a
substructure which is common in drugs, it is by no means specific to a particular
gene family. It can therefore be concluded that the reasons for its prevalence is
more related to ease of synthesis and the often favourable contribution to the
overall molecular properties such as polarity and hydrogen bonding potential;
these are related to specificity of binding and therefore increased selectivity,
which can so often determine the success or failure of a drug candidate. In
contrast, in the case of carbonic anhydrase inhibitors, the sulfonamide acts as a
zinc-binding group most likely in its deprotonated form, leading to a strong
electrostatic interaction which is probably, in a large part, responsible for the
good potency of the phenylsulfonamide substructure. It can only be speculated
as to which role the sulfonamide is playing in the other indications mentioned
above, since conclusive evidence as to its specific function in compounds acting
as antibacterials, diuretics or antidiabetics is still lacking. The role of the sul-
fonamide group in ADME is reviewed extensively in Chapter 5.

1.4.4 Chemokine Receptor 5 (CCR5)

The GPCR gene family is one of the most important drug target classes.
Therefore, an understanding of what features are contained in active GPCR
ligands and drugs has been of significant interest to the pharmaceutical industry
over recent decades. A recent analysis of over 17 000 GPCR ligands revealed
well-known motifs and also new substructural features such as the imidazole-
like substructure common for the histamine binding receptor ligands, as well as
the indole-like substructure which is common for serotonin receptor ligands.132
The chemokine receptor CCR5, a member of the family of GPCRs, is a target
for anti-HIV therapy, which is being targeted by the recently launched
antagonist maraviroc.133
All these CCR-5 antagonists have a common phenylpropylpiperidine sub-
structure (Figure 1.14), effectively a basic centre with a lipophilic group linked
by three carbon atoms. However, this substructure is also contained in a
number of other GPCR inhibitors, and is therefore not indicative of CCR-5 but
rather a large proportion of the GPCR gene family.
38 Chapter 1

N
N O
O O
O O
S N
S S
N N H
H H

NH2 NH2

O O
O O
O O
S N N
N H S H
N
H H

Cl

O
O
O
S N
N H
H

N
H
Cl O
O
O
OCH3
S N
N H
H

N N
H

Figure 1.13 Structures of antidiabetic sulfonamides and sulfonylureas: 2254 RP (top

left), carbutamide (top right), tolbutamide (second row left), chlorpro-
pamide (second row right), glibenclamide (second from bottom) and
glipizide (bottom).
Drugs and their Structural Motifs 39
N
N
N
N N
F
N N

O NH F

N OH

O
F F

N
F
N
N

O N
N O
F S O

N O
O
S
O NH2
S
O
O

Figure 1.14 Representative structures of published CCR-5 antagonists from Pfizer

(Maraviroc, Celsentrit, top left), Merck (top right), AstraZeneca (bot-
tom left) and GlaxoSmithKline (bottom right).

Therefore, privileged substructures can be indicative of a preference of a

receptor for a certain molecular structure but are not necessarily sufficient to
describe the minimum pharmacophoric requirements for activity at a particular
receptor. This is particularly the case for large gene families like GPCRs and
also kinases (discussed in more detail below). Both recognise binding motifs
which, although indicative of activity potentially against a whole gene family,
will not enable differentiation between its members for achieving selectivity—
particularly against closely related targets. Many GPCRs require a basic
function which is normally a primary, secondary or tertiary amine and the
ADME influences of these groupings are reviewed in Chapter 4.

1.4.5 Diaryl Heterocycles such as Cyclooxygenase Inhibitors

(COX-2)
A five-membered heterocycle with a conserved vicinal 1,2-diphenyl substitution
pattern has been observed in numerous medicinal chemistry programmes.
Probably the most notable application of this group is among the second
40 Chapter 1
generation non-steroidal anti-inflammatory (NSAID) agents that act via
selective COX-2 inhibition such as celecoxib134 and rofecoxib.135 However this
motif has been orthogonally optimised in other programmes to produce ligands
for multiple targets including P38 MAP kinase,136 adenosine A3 and phos-
phodiesterase-4 (PDE4).98 Besides anti-inflammatory targets, diarylheterocycle
derivatives have also been optimised as CB1 receptor agonists like Rimona-
bant98 and dopamine transporter inhibitors.137 By definition the broad spec-
trum application of this group in selective compounds for a range of target
families renders this a privileged subunit. A group of representative structures
of published diaryl heterocycle compounds is shown in Figure 1.15.
Although protein–ligand X-ray structures are known for a variety of these
compounds, a structural interpretation of the privileged status of the common
fragment motif is not clear.98 Interestingly, the nature of the five-membered
heterocycle is quite diverse and helps define different reactivity space. From a
basic level, this 1,2-diphenyl heterocycle offers a rigid, chemically enabled

F3C O H2N
O N Cl
N S
N
Me Cl

celecoxib refecoxib
COX-2 inhibitor COX-2 inhibitor N
IC50 40 nM IC50 20 nM CGS-2466
SO2NH2 SO2Me A3 antagonist, PDE4 inhibitor,
p38 inhibitor IC50 40 nM

AcS Cl

O
N
N Me
H
NH
N N N
N N
N N Cl
BuO Cl

rimonabant
SB-203580 glucagen antagonist
CB1 inverse agonist
p38 inhibitor
Ki 56 nM
IC50 600 nM
F OBu Cl

Figure 1.15 Representative structures of published diaryl heterocycle compounds.

Drugs and their Structural Motifs 41
template from which a phenyl ring can be positioned into pockets in an active
site of a protein. As phenyl rings are rigid lipophilic groups that are able to
occupy lipophilic sites or interact directly with other aromatic groups and side
chains, it can be argued that it is not surprising that this simple motif is able to
display activity for a range of proteins. Furthermore the option to add differ-
ential polarity and H-bonding capacity to both the core and periphery in a
chemically enabled manner provides optimal opportunity to build in target
selectivity. In accordance with this, multiple companies have opted to enrich
their compound collections with this template for high throughput screening,
thereby maximising probability of discovering new pharmacologies for this
group and cementing its place among identified privileged substructures.

1.4.6 Aminoheterocycles as Kinases Inhibitors

The purine scaffold is a key component of DNA and an important recognition
element in endogenous signalling molecules such as ATP and guanosine tri-
phosphate GTP.92 Consequently a large number of proteins have evolved to
recognise the purine structure and related mimics. Such molecules have been
identified as adenosine receptor ligands,138 gamma-aminobutyric (GABA)
receptor ligands,139 kinase enzyme inhibitors,109,140 antivirals such as Abacavir
(anti HIV) and acyclovir (herpes treatment).92 The structures of the last two are
shown in Figure 1.16.
Building on this theme, kinases are an example of a drug target family that is
host to a range of substructures which are known to be largely promiscuous
within the target family.100,109 The promiscuity for compounds within kinases
with broad selectivity over other targets can be reasoned based on the high
degree of similarity of kinase targets coupled with key structural knowledge of
the key recognition interactions. The majority of work towards the design of
kinase inhibitors has concentrated on inhibiting the ATP binding site. As the
family of over 500 distinct proteins have all evolved to share ATP as a common
natural ligand, it is unsurprising that there is broad conservation of size, shape

HN OH

N N
N OH
N

N H2N N N
H 2N N
O
Abacavir Acyclovir
OH

Figure 1.16 Representative structures of published purine-based antiviral com-

pounds: Abacavir (left) and Acyclovir (right).
42 Chapter 1

Kinase Hinge

HBA

HBD

Adenine Selectivity
Gate site
site Keeper

Activation
loop site

DFG-out site
Ribose
site
Asp
Phosphate
site

OH
R

Substrate binding site

Figure 1.17 ATP purine recognition by kinases.

and residue selection within this site across kinases. Furthermore, the ATP site
is well adapted to recognise the ATP purine ring head group via a series of
hydrogen bonding interactions.140 A schematic of ATP recognition by kinases
is shown in Figure 1.17.
Aranov et al. from Vertex discussed in detail the concept of kinase privileged
hinge-binding fragments and the notion of kinase targeted libraries whereby
libraries of known promiscuous kinase inhibitor scaffolds were made to spe-
cifically target SAR in kinase drug space.109 They investigated the idea of
kinase likeness in the context of kinase privileged fragments. Initially they
carried out an analogous method to that described by Murcko97,102 to define
the structures of kinase inhibitors in the context of their framework and
side chain atoms. The analysis was performed on 119 published kinase inhi-
bitors and revealed that the structural diversity at the level of rings and
linkers was relatively low. A combination of four rings and eight linkers was
found to describe 90% of the dataset.109 In particular, amino-substituted
Drugs and their Structural Motifs 43

N H N H
N N N N

Cl

O O F O O

Gefitinib Erlotinib

O O
N

N
HN
N H
N H N N
N
OH N
Cl N
KDR inhibitor TGFβ inhibitor

Figure 1.18 Representative structures of kinase inhibitor drugs.

heteroarylanilines have presented the majority of the kinase inhibitors under-

going clinical trials and most of the launched kinase drugs. Although the key
hydrogen bond elements that are crucial for recognition of the ATP purine
group are conserved for these bisarylamine groups, an analysis of X-ray
structures has revealed that the closely related groups do in fact bind in dif-
ferent orientations and locations of the ATP site. Representative structures of
kinase inhibitor drugs are shown in Figure 1.18.
Further to these observations, Aronov et al. proposed and internally vali-
dated a kinase-likeness rule termed the ‘2-0’ rule. This rule stated that a com-
pound is likely to have kinase activity if:

i) it contains two or more heteroaromatic nitrogens;

ii) it contains one or more heteroaromatic NH groups;
iii) it contains one or more anilines; and
iv) it contains one or more nitriles.
44 Chapter 1
When tested against the Vertex file, this rule was observed to accurately
describe between 80% and 100% of known kinase hinge-binding inhi-
bitors. When used prospectively, the authors suggested that a five-fold
enrichment in the discovery of new kinase inhibitors had been observed
and that this was likely due to the fact that the rule of thumb ensured that
the key hydrogen bonding recognition elements for hinge binding were
present.109

1.4.7 HMG-CoA Reductase Inhibitors

Statins are a class of hypolipidemic drugs used to control hypercholesterolemia
(elevated cholesterol levels) and to prevent cardiovascular disease. They act
through inhibition of HMG-CoA. The world’s best selling drug, Atorvastatin
(Lipitort), with worldwide sales of around $13.8 billion in 2008 (see Table 1.2),
belongs to this drug class.
Two distinct classes of HMG-CoA inhibitors appear to have been marketed
so far. One class, containing for example Atorvastatin, is made up of synthetic
statins. The other class, containing for example Pravastatin, which was initially
known as CS-514 and was originally identified in the bacterium Nocardia
autotrophica141, is made of natural products; another example is Mevastatin,
which was obtained from Penicillium citrinum.142,143 Pravastatin is also an
active metabolite of mevastatin.144 Both classes of compounds (illustrated in
Figure 1.19) can be regarded as containing privileged substructures, such as for
example the acid side chain in Cerivastatin, Rosuvastatin as well as Atorvas-
tatin, and the lactone side chain as well as the bicyclic core in the other class of
statins. However, this is debatable since the second class in particular is very
similar in terms of molecular structure (differing effectively by no more than
two atoms) and is based on natural products rather than having been synthe-
sised on the basis of receptor SAR and observed preference for a particular
substructure.
In addition, the two classes of statins are quite dissimilar in molecular
structure. Therefore, it appears that there is no particular single privileged
substructure responsible for activity against HMG-CoA in marketed statins,
but that there are at least two distinct chemical structures that may express a
very similar pharmacophore, which then accounts for the common activity
against the target.

1.5 Privileged Substructures and Chemical Space

A growing number of substructures have been classed as privileged over the
past two decades from a variety of analyses across industry and academia. It is
evident that rigid aromatic and polyaromatic rings are common features of
privileged structures. This makes sense when one considers the nature of the
majority of targeted binding pockets. The majority of these have been
Drugs and their Structural Motifs 45

F F

OH OH O OH OH O

O OH OH
O O N
S
N N N
Cerivastatin R

OH OH O
NH
N OH
O

Atorvastatin
F

HO O
HO O
O
O O
O
O
H O
CH3 H
CH3

HO
Pravastatin CH3 Lovastatin

HO O HO O

O O
O O

O O
H H
CH3 CH3

CH3 Simvastatin Mevastatin

Figure 1.19 Cerivastatin (Baycolt), Rosuvastatin (Crestort), Atorvastatin (Lipi-

tort), Pravastatin (Pravacolt), Lovastatin (Mevacort) and Simvastatin
(Zocort) and Mevastatin.
46 Chapter 1
hydrophobic pockets with p-stacking to phenylalanines and tyrosines being
commonly observed; hydrophobic and aromatic interactions between ligands
and protein targets play significant roles in the overall binding energy. Addi-
tionally the rigidity of many privileged groups enables presentation of the
peripheral functionality in an ordered fashion, incurring minimal entropic
penalty.100,145
It is worth considering why such groups have been discovered as privileged;
are they truly privileged? In particular, it is interesting to note that many leads
have been discovered from HTS strategies whereby a company’s compound file
is tested against a new target. The nature of the chemical matter in these col-
lections is composed of a combination of substrate from previous drug pro-
grammes and file enrichment library enabled compounds. As a consequence, hit
identification for a new programme potentially provides established groups a
disproportionate opportunity to be rediscovered as hits over defining oppor-
tunities for discovery of series that describe entirely new chemical space. When
this is considered in conjunction with the fact that templates that have a proven
safety track record are often viewed as appealing leads, it seems hardly sur-
prising that the same structural elements have been optimised for multiple drug
programmes.27
A recent analysis of 1386 marketed drugs revealed that 15% are contained
within other drugs and that 30% contain other drugs as substructure frag-
ments.145 A variety of recent papers have discussed these concepts of chemical
and biological space in the context of drug discovery. In particular discussions
consider future opportunities and methods to inspire new directions for drug
discover chemistry to find additional chemical matter to enable targets that
have been either unsuccessful or classed as undruggable due to the fact that
current chemical matter is unsuitable. Opportunities such as using natural
products as bio-active templates for further elaboration, broader fragment
screening strategies and further library manipulation of currently existing
templates via the development of more inventive library designs have all been
considered.58,145,146
Although privileged substructures are a useful concept, there are aspects of
this principle that are still ambiguous. It is not clear whether privileged sub-
structures are a result of nature’s preference for particular molecular entities, or
whether it is the chemist’s preference for particular synthetic routes, biased by
synthetic feasibility and precedence. Either way, we suggest that it is a useful
concept in the design of drugs which highlights the link between active com-
pounds and their SAR. Also, chemists generally think in terms of two-
dimensional Lewis structures, as used in this chapter, despite the fact that
receptors don’t recognise Lewis structures but rather electron densities around
ligand atoms. However, it is much more practical and efficient to view mole-
cules in the usual Lewis depiction rather than as for humans difficult to
interpret surfaces. Therefore, the existence of the concept of privileged sub-
structures could be more a result of how chemists are trained to recognise
molecules rather than the actual recognition processes that happen when small
molecules bind to a receptor.
Drugs and their Structural Motifs 47
1.6 Reasons for Compound Attrition
As pointed out earlier, only about 11% of compounds entering clinical devel-
opment ever reach the market; more than one third are withdrawn for reasons
like toxicology (24%) and clinical safety (12%), making toxicity-related factors
one of the major contributors to compound attrition.20 Smith and Schmid
reviewed drug withdrawals over recent decades and found that, in the cases
having the greatest impact, the reason for withdrawal was the interaction of a
drug with a single receptor, ion channel or enzyme.147 Once the mechanism has
been identified, screens can be established; however, when the mechanism is
more complex such as for example in organ toxicity, it is far more difficult to
establish those screens.147
One of the most significant developments in compound attrition in the last
ten years was the effect of QT prolongation on drug approvals. By 1998, QT
prolongation emerged as a major safety issue affecting many classes of drugs.
This was precipitated by the withdrawal of Serenading and Cisalpine from the
US market because of sudden deaths associated with QT interval prolongation.
The subsequent focus on QT led to the re-evaluation of many drugs on the
market and in development, and is likely to have contributed to the lower NCE
approval rates from 1998 onwards.16 QT prolongation can be related to inhi-
bition of the human ether-a-go-go related gene (hERG) potassium channel. The
function of this channel is to conduct the rapidly activating delayed rectifier
potassium current (IKr), which has a key role in the control of cardiac
rhythm.148 Examples of compounds removed from the market due to concerns
with this issue include the antihistamine Terfenadine (withdrawn February
1998) and 5HT4 partial agonist Cisapride (withdrawn July 2000) (Figure 1.20).
Both of these compounds are high affinity ligands for the hERG ion channel
with IC50 of 56 nM and 6 nM, respectively.
Both compounds have structural similarities, for example the piperidine ring
with a flexible hydrophobic substituent (phenylbutyl or phenyloxypropyl).
However, the substructure is not specific to hERG activity, since a number of
compounds with different substructures are also active. It appears to be more a
case of having the relevant pharmacophoric elements, which have been

MeO
OH
OH H
N N
NH2

O N O Cl
OMe

Figure 1.20 Terfenadine (left) and cisapride.

48 Chapter 1
identified as an amine as a basis centre and at least two lipophilic groups at
certain distances on either side of the amine.149,150
It is widely accepted that the withdrawal of compounds like Terfenadine and
Cisapride has had a considerable impact on the pharmaceutical industry,
particularly in terms of designing compounds that are less likely to have affinity
for the hERG channel. There are successful examples of avoiding hERG
assisted by molecular modelling, such as for example in the discovery of the
anti-HIV drug Maraviroc.133,149 More recently, evidence has been presented for
a strong link of molecular properties like molecular weight and logP with the
probability of polypharmacology or toxicity of compounds.81,151 Therefore, it
is not only the structure of compounds but also their physicochemical prop-
erties which can influence their toxicological profile, significantly adding to the
complexity of compound attrition.
Attrition rates have also been analysed in terms of different disease areas and
target class, finding that for example that the attrition rate of kinase inhibitors
in oncology is only 53% compared with the overall attrition rate of anti-cancer
drug candidates of 82%. This appears to indicate the benefits of developing
molecularly targeted therapeutics for cancer.152
Seemingly minor differences can be responsible for the ‘launched’ or ‘with-
drawn’ status for drugs, as in the example of Cerivastatin (Baycolt) and
Rosuvastatin (Crestort). Launched in 1997, Cerivastatin was voluntarily
withdrawn from all markets worldwide by Bayer in 2001 following reports of
side-effects of potentially fatal myopathy and rhabdomyolysis, in particular
when the drug was co-administered with gemfibrozil. The structures of Cer-
ivastatin and Rosuvastatin are shown in Figure 1.21.
At its peak, Baycol’s global sales in 2000 exceeded $586 million. Launched by
AstraZeneca in 2003, Rosuvastatin is now marketed in over 50 countries, with
global sales over $2 billion in 2006. Both compounds are HMG-CoA reductase
inhibitors and have been observed to have an antihypercholesterolemic and
antihyperlipidemic effect by depleting cells of mevalonic acid, a cholesterol
precursor. The event faced by Bayer in 2001 is similar to the global withdrawal

F F

OH OH O OH OH O

O OH OH
O O N
S
N N N

Figure 1.21 HMG-CoA reductase inhibitors on the market, withdrawn or dis-

continued in development: Cerivastatin (Baycolt, left) and Rosuvastatin
(Crestort, right).
Drugs and their Structural Motifs 49
of Vioxxt (Rofecoxib) by Merck and Co. in 2004 (global sales over $2.5 billion
in 2003).
Events like these have been categorised as a ‘Black Swan’.163 Coined by
Nassim Nicholas Taleb,153 the ‘Black Swan’ is an unpredictable event, which
has a massive impact from a business perspective. The term ‘Black Swan’ is a
metaphor for the first sighting of the black swan, which (a) invalidated the
assumption that all swans are white—based on millions of previous observa-
tions; (b) changed our perception of those birds; and (c) was retrospectively
‘assimilated’ as a highly predictable event. Thus, the accumulation of past data
cannot be used to predict the ‘unknown unknown’. Considered by philosophers
and military alike, three cognitive categories can be discussed: (a) the ‘known
knowns’, for which sets and models are available, they have been validated
externally and can be considered quite reliable; (b) the ‘known unknowns’, for
which sets and models can be documented and proved, but otherwise lack
external validation (i.e. true predictivity is of limited reliability); and finally, (c)
the ‘unknown unknowns’, for which we do not have sets and models, thus lack
any validation and predictivity.163
On the surface, the concept of the ‘unknown unknown’ does not appear to be
applicable for, for example, Cerivastatin which has good affinity for a known
target (HMG-CoA reductase), well-categorised pharmacokinetic profile (e.g.
60% orally bioavailable, 80% bound to albumin), and documented clinical
effect (it lowers cholesterolemia). An ‘unknown unknown’ element was intro-
duced by clinical practitioners who prescribed this lipid lowering drug in com-
bination with an older antilipidemic, gemfibrozil (marketed since 1982 as
Lipurt/Lopidt). Besides its serum lipid regulating effect, gemfibrozil also
blocks CYP 2C8, the cytochrome P450 isozyme primarily responsible for
metabolising Cerivastatin, leading to an unanticipated drug–drug interaction.163
A further example of the complexity of the reasons for attrition is the anti-
inflammatory drug ibuprofen. After being launched in 1969, it became widely
prescribed throughout the world in the wake of increasing concern about the
hazard of gastric bleeding caused by aspirin. Such was its relative safety
that in 1983 it became available in the UK as a non-prescription analgesic
on account of its having the lowest overall rate of reporting of suspected
adverse reactions among NSAIDs, some 20 million prescriptions having
been issued over the preceding 15 years. However, ibuprofen was only selected
to become a drug after its close analogue ibufenac, which only differs by a
methyl group (Figure 1.22), had to be withdrawn soon after being launched

O O

OH OH

Figure 1.22 Ibuprofen (left) and its close analogue ibufenac.

50 Chapter 1
31
in 1966. A change as small as one methyl group can, therefore, make the
difference between a safe drug and one which leads to unacceptable adverse
reactions.
Interestingly the ibuprofen analogue, flurbiprofen, which is also marketed as
an anti-inflammatory drug, also carries the motif of the alpha-substituted
carboxylic acid, as do several other marketed analogues such as naproxen,
ketoprofen and fenoprofen. However, a further analogue, benoxaprofen,
which also has the alpha-methylcarboxylic acid motif, had to be withdrawn
from the market after it was discovered that it had serious side effects in
patients relating to renal or hepatic failure. However, the adverse reactions
were attributed to a rather long half-life of the compound, particularly in
elderly patients.31 This highlights the complexity of the relationship between
compound structure and adverse reactions, which make it very difficult to
predict toxicity and side effects, even when the pharmacological or toxico-
logical profiles of close analogues are known.154 This may be in contrast to
attempts to predicting adverse reactions from in vitro pharmacological profiles
of structurally very similar compounds or close analogues containing identical
substructures.82,84,85
Balancing early market access to new drugs with the need for the assessment
of benefit and risk is an ongoing dilemma for the drug regulatory agencies.
This dilemma is not new, but has been made more prominent by recent high-
profile drug withdrawals and conflicting demands, including the need to
improve the efficiency of drug development on one hand and the need to avoid
exposing patients to unnecessary risks and possibly ineffective treatments on
the other.18
Drug target and candidate selection are two of the key decision points within
the drug discovery process for which all companies use certain selection criteria
to make decisions on which targets to accept into their discovery pipelines and
which compounds will pass into development.154 These steps do not only help
define the overall productivity of every company, but they are also decisions
taken without full predictive knowledge of the risks that lie ahead or how best
to manage them. In particular, the process of selecting new targets does not
normally involve full evaluation of the risks in the mechanism under investi-
gation, which may result in an inability to fully connect in vitro and animal
model results to the disease setting. The resulting poor progression statistics of
many compounds in the clinic is at least partially the result of a lack of
understanding of disease pathophysiology. Notably, the lack of efficacy is still a
major reason for failure in the clinic. Creating a more holistic understanding of
disease pathophysiology and an early confidence in the mechanism under
investigation could help facilitate the selection of not only the most appropriate
targets but also the best mechanisms for disease intervention and how to select
and optimise the best compounds.154
However, despite the continuing improvements in our understanding of the
origins of diseases and how to treat them, the hurdles to get drugs to market
may appear higher every year, perhaps accounting for the increasing rate in
attrition and the drop in productivity. Bringing a drug to the market today may
Drugs and their Structural Motifs 51
well be far more difficult than for example 50 or 100 years ago. As Walter
Sneader pointed out in his book Drug Discovery, A History:125 ‘It is fair to say
that if an attempt were to be made today to introduce either paracetamol or
aspirin into medicine, they might be denied a license’.
The knowledge about the pharmacology of drug candidates (e.g. through
broad-spectrum screening for example against a panel of common targets) may
also identify relationships with compound attrition and aid selection of clinical
candidates.82,84,85,155,156 Perhaps new knowledge-based approaches (e.g. sys-
tems toxicology) will be able to identify attrition risks more effectively and
reduce attrition rates by utilising toxicogenomics knowledge that combines
molecular expression data sets from transcriptomics, proteomics, metabo-
nomics and conventional toxicology with metabolic, toxicological pathway and
gene regulatory network information relevant to human disease.157,158

1.7 Summary and Outlook

We hope that we have shown that drug discovery is an extremely diffi-
cult endeavour, which is becoming ever increasingly complex due to the
changes in the regulatory requirements as well as the pressure on cost and
productivity.
Despite these pressures, the pharmaceutical industry keeps discovering new
drugs, albeit at what appears to be a slower rate. Whether the apparent
reduction in productivity is a temporary phase or will be a continuing trend is
not possible to determine at this time. However, it is almost inconceivable that
without a significant change in drug discovery paradigm and increase in pro-
ductivity (particularly drug output), the industry in its current state will con-
tinue to be as attractive to investors it appeared to be in the past. These changes
in paradigm will need to come from the business end of the industry, which is of
course research. The trend has been to suggest that the downward trajectory in
productivity and diminishing returns on investment are purely a ‘science pro-
blem’.159 In addition, pharmaceutical research has been considered a knowl-
edge-based and skill-based approach with research largely operating in
‘chemogenomics knowledge space’.160 Although scientific innovation is clearly
part of the solution, a recent analysis suggested that stronger management
attention to well-known value-creation levers such as for example cost, speed
and decision making, could increase productivity and return on investment
very significantly.159
Recently, an analysis of pharmaceutical industry performance metrics and
current portfolios has been used to project the future productivity of the
industry over the years 2007–2012 for the top 14 pharmaceutical companies.161
The results indicate, for example, that the collection of branded drugs each
of which is projected to achieve at least $500 million in annual sales, will
only rise slightly (1.1%) over the five years. This is considered to be largely a
consequence of loss of exclusivity for major products such as Pfizer’s
Lipitor (Atorvastatin), Wyeth’s Effexor (Venlafaxine), Johnson & Johnson’s
52 Chapter 1
Risperdal (Risperidone) and Eli Lilly’s Zyprexa (Olanzapine). However, bio-
logics sales are projected to grow significantly in the region of 13% over the
same time period. The companies with the largest percentage of biologics in
2007 were Amgen (96.7%, $13.6 billion), Roche (70.8%, $16.9 billion) and
Wyeth (39.6%, $5.3 billion). The forecasted growth is driven by products
such as Roche’s Averting (Bevacizumab) and Abbott’s Humira (Adalimumab),
new products anticipated to launch during the forecast period such as Amgen’s
Denosumab and Johnson & Johnson’s Golimumab as well as vaccines, in
particular GlaxoSmithKline’s Cervarix, Merck’s Gardasil and Wyeth’s
Prevnar. In addition, there is a projected overall 4.4% reduction in internally
developed products, with products from in-licensing and acquisition rising
correspondingly. Companies with the highest proportion of internally
developed product-derived revenue in 2007 include Novartis (93.4%), Eli Lilly
(84.2%) and GlaxoSmithKline (80.3%), whereas those with the lowest pro-
portion include Roche (15.1%), Schering-Plough (11.8%) and Sanofi-Aventis
(8.4%).
Although research in itself does not create revenue, it is here where future
value is created. Therefore, new thinking is perhaps required in research and a
better understanding of fundamental principles such as the influence of the
molecular properties of drugs on activity, ADME and toxicity. These principles
also include those of the rule of five as well as the concept of druggability,
ligand efficiency and privileged substructures.
We hope that we have highlighted the relevance and importance of privileged
substructures in drug design and discovery, using a number of examples
spanning a large area of pharmacological and target as well as chemical space.
New approaches to discovery such as indications discovery could also con-
tribute to maximising the value out of the chemical matter already available
that has already passed essential compound safety and attrition hurdles.162
However, as outlined above, future target space may well look very different
from current target space due to the exploration of new members of the pro-
teome discovered through sequencing of the human genome. The future of drug
discovery and the expansion of target as well as compound space are the subject
of the last chapter.

1.8 Abbreviations
ADMET absorption, distribution, metabolism, excretion, toxicity
ADP adenosine diphosphate
ATP adenosine triphosphate
DMPK drug metabolism and pharmacokinetics
CCK cholecystokinin
CMC Comprehensive Medicinal Chemistry
CNS central nervous system
COX cyclooxygenase inhibitor
EGFR epidermal growth factor receptor
Drugs and their Structural Motifs 53
FDA Food and Drug Administration
GABA gamma-aminobutyric acid
GPCR G-protein coupled receptor
GTP guanosine triphosphate
HERG human ether-a-go-go related gene
HIV human immunodeficiency virus
HMG-CoA 3-hydroxy-3-methyl-glutaryl-coenzyme A
HTS high throughput screening
MW molecular weight
NCE new chemical entity
NHR nuclear hormonal receptor
NME new molecular entity
NMR nuclear magnetic resonance
NSAID non-steroidal anti-inflammatory drug
PAF platelet activation factor
PDB Protein Data Bank
PDE phosphodiesterase
Ro5 rule of five
R&D research and development
SAR structure–activity relationship
TPSA topological polar surface area

References
1. S. J. Ainsworth, Chem. Eng. News, 2008, 86, 15–24.
2. N. N. Malik, Drug Discov. Today, 2008, 13, 909–912.
3. N. N. Malik, Expert Opin. Drug Discov., 2009, 4, 15–19.
4. B. Gaudilliere and P. Berna, Annu. Rep. Med. Chem., 2000, 35, 331–
356.
5. B. Gaudilliere, P. Bernardelli and P. Berna, Annu. Rep. Med. Chem., 2001,
36, 293–318.
6. P. Bernardelli, B. Gaudilliere and F. Vergne, Annu. Rep. Med. Chem.,
2002, 37, 257–277.
7. C. Boyer-Joubert, E. Lorthiois and F. Moreau, Annu. Rep. Med. Chem.,
2003, 38, 347–374.
8. S. Hegde and J. Carter, Annu. Rep. Med. Chem., 2004, 39, 337–368.
9. S. Hegde and M. Schmidt, Annu. Rep. Med. Chem., 2005, 40, 443–473.
10. S. Hegde and M. Schmidt, Annu. Rep. Med. Chem., 2006, 41, 439–477.
11. S. Hegde and M. Schmidt, Annu. Rep. Med. Chem., 2007, 42, 505–553.
12. S. Hegde and M. Schmidt, Annu. Rep. Med. Chem., 2008, 43, 455–497.
13. J. Owens, Nat. Rev. Drug Discov., 2007, 6, 99–101.
14. B. Hughes, Nat. Rev. Drug Discov., 2009, 8, 93–96.
15. B. Hughes, Nat. Rev. Drug Discov., 2008, 7, 107–109.
16. E. Schmidt and D. Smith, Drug Discov. Today, 2005, 10, 1031–1039.
17. R. R. Ruffolo, Expert Opin. Drug Discov., 2006, 1, 99–102.
54 Chapter 1
18. H.-G. Eichler, F. Pignatti, B. Flamion, H. Leufkens and A. Breckenridge,
Nat. Rev. Drug Discov., 2008, 7, 818–826.
19. E. F. Schmid and D. A. Smith, Annu. Rep. Med. Chem., 2005, 40,
431–441.
20. P. D. Leeson and A. M. Davis, J. Med. Chem., 2004, 47, 6338–6348.
21. H. Kubinyi, Nat. Rev. Drug Discov., 2003, 2, 665–668.
22. H. Grabowski, J. Vernon and J. A. DiMasi, Pharmacoeconomics, 2002,
20, 11–29.
23. J. A. Vernon, Expert Opin. Drug Discov., 2009, 4, 21–22.
24. C. A. Lipinski, F. Lombardo, D. B. W. and P. J. Feeney, Adv. Drug Del.
Rev., 1997, 23, 3–25.
25. C. A. Lipinski, Drug Discov. Today: Technologies, 2004, 1, 337–341.
26. P. D. Leeson and B. Springthorpe, Nat. Rev. Drug Discov., 2007, 6,
881–890.
27. P. D. Leeson, A. M. Davis and J. Steele, Drug Discov. Today: Technol-
ogies, 2004, 1, 189–195.
28. I. Kola and J. Landis, Nat. Rev. Drug Discov., 2004, 3, 711–715.
29. U. A. K. Betz, Drug Discov. Today, 2005, 10, 1057–1063.
30. I. Cavero and H. R. Kaplan, Expert Opin. Drug Discov., 2008, 3,
1145–1154.
31. W. Sneader, Drug Prototypes and their Exploitation, Wiley, London, 1996.
32. E. M. John, Annu. Rep. Med. Chem., 2008, 43, 525–544.
33. J. P. Overington, B. Al-Lazikani and A. L. Hopkins, Nat. Rev. Drug
Discov., 2006, 5, 993–996.
34. S. J. Haggarty, Curr. Opin. Chem. Biol, 2005, 9, 296–303.
35. IHGS Consortium, Nature, 2001, 409, 860–921.
36. IHGS Consortium, Nature, 2004, 431, 931–945.
37. A. L. Hopkins and C. R. Groom, Nat. Rev. Drug Discov., 2002, 1,
727–730.
38. A. P. Orth, S. Batalov, M. Perrone and S. K. Chanda, Expert Opin. Ther.
Targets, 2004, 8, 587–596.
39. G. V. Paolini, R. H. B. Shapland, W. P. van Hoorn, J. S. Mason and A. L.
Hopkins, Nat. Biotech., 2006, 24, 805–815.
40. P. R. Caron, M. D. Mullican, R. D. Mashal, K. P. Wilson, M. S. Su and
M. A. Murcko, Curr. Opin. Chem. Biol., 2001, 5, 464–470.
41. C. J. Harris and A. P. Stevens, Drug Discov. Today, 2006, 11, 880–888.
42. T. Klabunde, Brit. J. Pharmacol., 2007, 152, 5–7.
43. M. Casesa and J. Mestres, Drug Discov. Today, 2009, 14, 479–485.
44. P. Bamborough, D. Drewry, G. Harper, G. K. Smith and K. Schneider,
J. Med. Chem., 2008, 51, 7898–7914.
45. J. Drews, Nat. Biotech., 1996, 14, 1516–1518.
46. J. Drews and S. Ryser, Nat. Biotech., 1997, 15, 1318–1319.
47. A. L. Hopkins and C. R. Groom, Ernst Schering Res. Found. Workshop,
2003, 42, 11–17.
48. G. Vistoli, A. Pedretti and B. Testa, Drug Discov. Today, 2007, 13,
285–294.
Drugs and their Structural Motifs 55
49. Y. Sugiyama, Drug Discov. Today, 2005, 10, 1577–1579.
50. T. H. Keller, P. Arkadius and Y. Zheng, Curr. Opin. Chem. Biol., 2006,
10, 357–361.
51. I. D. Kuntz, K. Chen, K. A. Sharp and P. A. Kollman, Proc. Natl. Acad.
Sci. U.S.A., 1999, 96, 9997–10002.
52. A. L. Hopkins, C. R. Groom and A. Alex, Drug Discov. Today, 2004.
53. A. C. Cheng, R. G. Coleman, K. T. Smyth, Q. Cao, P. Soulard, D. R.
Caffrey, A. C. Salzberg and E. S. Huang, Nat. Biotechnol., 2007, 25,
71–75.
54. S. Carney, Drug Discov. Today, 2005, 10, 1012–1013.
55. W. T. Loging, L. Harland and B. Williams-Jones, Nat. Rev. Drug Discov.,
2007, 6, 220–230.
56. C. M. Dobson, Nature, 2004, 432, 824–828.
57. C. A. Lipinski and A. L. Hopkins, Nature, 2004, 432, 855–861.
58. J. Rosen, J. Gottfries, S. Muresan, A. Backlund and T. I. Oprea, J. Med.
Chem., 2009, 52, DOI: 10.1021/jm801514w.
59. M. Gualtieri, F. Banéres-Roquet, P. Villain-Guillot, M. Pugnière and
J.-P. Leonetti, Curr. Med. Chem., 2009, 16, 390–393.
60. Y. Yamanishi, M. Araki, A. Gutteridge, W. Honda and M. Kanehisa,
Bioinformatics, 2008, 24, i232–i240.
61. M. Vieth, M. G. Siegel, R. E. Higgs, I. A. Watson, D. H. Robertson,
K. A. Savin, G. L. Durst and P. A. Hipskind, J. Med. Chem., 2004, 47,
224–232.
62. M. Feher and J. M. Schmidt, J. Chem. Inf. Comput. Sci., 2003, 43,
218–227.
63. M. Vieth and J. J. Sutherland, J. Med. Chem., 2006, 49, 3451–
3453.
64. M. S. Lajiness, M. Vieth and J. Erickson, Curr. Opin. Drug Discov. Dev.,
2004, 7, 470–477.
65. C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeney, Adv. Drug
Del. Rev., 2001, 46, 3–26.
66. M. C. Wenlock, R. P. Austin, P. Barton, A. M. Davis and P. D. Leeson,
J. Med. Chem., 2003, 46, 1250–1256.
67. A. V. Veselovsky and A. I. Archakov, Curr. Comput. Aided Drug Des.,
2007, 3, 51–58.
68. T. I. Oprea, A. M. Davis, S. J. Teague and P. D. Leeson, J. Chem. Inf.
Comput. Sci., 2001, 41, 1308–1315.
69. M. M. Hann, A. R. Leach and G. Harper, J. Chem. Inf. Comput. Sci.,
2001, 41, 856–864.
70. M. J. Waring, Bioorg. Med. Chem. Lett., 2009, 19, 2844–2851.
71. M. C. Hutter, J. Chem. Inf. Model., 2006, 47, 186–194.
72. M. Wagener and V. J. van Geerestein, J. Chem. Inf. Comput. Sci., 2000,
40, 280–292.
73. T. J. Ritchie, C. N. Luscombe and S. J. F. Macdonald, J. Chem. Inf.
Model., 2009, 49, 1025–1032.
74. J. A. Wells and C. L. McClendon, Nature, 2007, 450, 1001–1009.
56 Chapter 1
75. J. C. Fuller, N. J. Burgoyne and R. M. Jackson, Drug Discov. Today,
2008, 14, 155–161.
76. N. J. Burgoyne and R. M. Jackson, Bioinformatics, 2006, 22, 1335–
1342.
77. C. G. Wermuth, J. Med. Chem., 2004, 47, 1303–1314.
78. S. Van Gestel and V. Schuermans, Drug Dev. Rev., 1986, 8, 1–13.
79. M. Vieth, R. E. Higgs, D. H. Robertson, M. Shapiro, E. A. Gragg and
H. Hemmerle, Biochim. Biophys. Acta, 2004, 1697, 243–257.
80. M. Vieth, J. J. Sutherland, D. H. Robertson and R. M. Campbell, Drug
Discov. Today, 2005, 10, 839–846.
81. J. D. Hughes, J. Blagg, D. A. Price, S. Bailey, G. A. DeCrescenzo, R. V.
Devraj, E. Ellsworth, Y. M. Fobian, M. E. Gibbs, R. W. Gilles,
N. Greene, E. Huang, T. Krieger-Burke, J. Loesel, T. Wager, L. Whiteley
and Y. Zhang, Bioorg. Med. Chem. Lett., 2008, 18, 4872–4875.
82. A. F. Fliri, W. T. Loging, P. F. Thadeio and R. A. Volkmann, Proc. Natl.
Acad. Sci. U.S.A., 2005, 102, 261–266.
83. C. M. Krejsa, D. Horvath, S. L. Rogalski, J. E. Penzotti, B. Mao,
F. Barbosa and J. C. Migeon, Curr. Opin. Drug Disc. Dev., 2003, 6,
470–480.
84. A. F. Fliri, W. T. Loging, P. F. Thadeio and R. A. Volkmann, Nat. Chem.
Biol., 2005, 1, 389–397.
85. A. F. Fliri, W. T. Loging, P. F. Thadeio and R. A. Volkmann, J. Med.
Chem., 2005, 48, 6918–6925.
86. M. Congreve, C. W. Murray and T. L. Blundell, Drug Discov. Today,
2005, 10, 895–907.
87. T. A. Halgren, J. Chem. Inf. Model., 2009, 49, 377–389.
88. E. Hedon and J. Arrons, C. R. Hebd. Seances Acad. Sci., 1899, 129,
778.
89. J. Fischer and C. R. Ganellin, Analogue-based Drug Discovery, Wiley-
VCH, Weinheim, 2006.
90. G. M. Maggiora, J. Chem. Inf. Model., 2006, 46, 1535–1535.
91. E. J. Ariens, A. J. Beld, J. F. Rodrigues de Miranda and A. M. Simonis, in
The Receptors: A Comprehensive Treatise, ed. R. D. O’Brien, Plenum
Press, New York, 1979, vol. 1, pp. 33–91.
92. R. W. DeSimone, K. S. Currie, S. A. Mitchell, J. W. Darrow and D. A.
Pippin, Comb. Chem. High Throughput Screening, 2004, 7, 473–493.
93. M. G. Bock, R. M. DiPardo, K. E. Rittle, B. E. Evans, R. M. Freidinger,
D. F. Veber, R. S. L. Chang, T. B. Chen, M. E. Keegan and V. J. Lotti, J.
Med. Chem., 1986, 29, 1941–1945.
94. B. E. Evans, M. G. Bock, K. E. Rittle, R. M. Di Pardo, W. L. Whitter, D.
F. Veber, P. S. Anderson and R. M. Freidinger, Proc. Natl. Acad. U.S.A.,
1986, 83, 4918–4922.
95. 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. Kinkel, J.
P. Speinger and J. Hirshfield, 1988, 31, 2235–2246.
Drugs and their Structural Motifs 57
96. A. A. Patchett and R. P. Nargund, Annu. Rep. Med. Chem., 2000, 35,
289–298.
97. G. W. Bemis and M. A. Murcko, J. Med. Chem., 1996, 39, 2887–2893.
98. G. Muller, Drug Discov. Today, 2003, 8, 681–691.
99. D. M. Schnur, Abstracts, 36th Middle Atlantic Regional Meeting of the
American Chemical Society, Princeton, NJ, June 8–11, 2003, 19.
100. D. M. Schnur, M. A. Hermsmeier and A. J. Tebben, J. Med. Chem., 2006,
49, 2000–2009.
101. T. Guo and D. W. Hobbs, Assay Drug Dev. Technol., 2003, 1,
579–592.
102. G. W. Bemis and M. A. Murcko, J. Med. Chem., 1999, 42, 5095–5099.
103. X. Q. Lewell, D. Judd, S. Watson and M. Hann, J. Chem. Inf. Comput.
Sci., 1998, 38, 511–522.
104. P. J. Hajduk, M. Bures, J. Praestgaard and S. W. Fesik, J. Med. Chem.,
2000, 43, 3443–3447.
105. M. G. Siegel and M. Vieth, Drug Discov. Today, 2006, 12, 71–79.
106. J. J. Sutherland, R. E. Higgs, I. Watson and M. Vieth, J. Med. Chem.,
2008, 51, 2689–2700.
107. E. Gianti and L. Sartori, J. Chem. Inf. Model., 2008, 48, 2129–2139.
108. C. D. Duarte, E. J. Barreiro and C. A. M. Fraga, Mini Rev. Med. Chem.,
2007, 7, 1108–1119.
109. A. M. Aronov, B. McClain, C. Stuver Moody and M. A. Murcko, J. Med.
Chem., 2008, 51, 1214–1222.
110. R. S. L. Chang, V. J. Lotti, R. L. Monaghan, J. Birnbaum, E. O. Stapley,
M. A. Goetz, G. Albers-Schonberg, A. A. Patchett and J. M. Liesch,
et al., Science, 1985, 230, 177–179.
111. M. G. Bock, R. M. DiPardo, B. E. Evans, K. E. Rittle, D. F. Veber, R. M.
Freidinger, R. S. L. Chang and V. J. Lotti, J. Med. Chem., 1988, 31,
176–181.
112. M. G. Bock, R. M. DiPardo, B. E. Evans, K. E. Rittle, W. L. Whitter,
V. M. Garsky, K. F. Gilbert, J. L. Leighton and K. L. Carson, et al.,
J. Med. Chem., 1993, 36, 4276–4292.
113. M. G. Bock, R. M. DiPardo, B. E. Evans, K. E. Rittle, W. L. Whitter,
D. F. Veber, R. M. Freidinger, R. S. L. Chang, T. B. Chen and V. J. Lotti,
J. Med. Chem., 1990, 33, 450–455.
114. L. H. Sternbach, J. Med. Chem., 1979, 22, 1–7.
115. D. Roemer, H. H. Buescher, R. C. Hill, R. Maurer, T. J. Petcher,
H. Zeugner, W. Benson, E. Finner, W. Milkowski and P. W. Thies,
Nature, 1982, 298, 759–760.
116. A. Walser, T. Flynn, C. Mason, H. Crowley, C. Maresca, B. Yaremko
and M. O’Donnell, J. Med. Chem., 1991, 34, 1209–1221.
117. D. R. Armour, N. M. Aston, K. M. L. Morriss, M. S. Congreve, A. B.
Hawcock, D. Marquess, J. E. Mordaunt, S. A. Richards and P. Ward,
Bioorg. Med. Chem. Lett., 1997, 7, 2037–2042.
118. R. S. McDowell, B. K. Blackburn, T. R. Gadek, L. R. McGee,
T. Rawson, M. E. Reynolds, K. D. Robarge, T. C. Somers, E. D.
58 Chapter 1
Thorsett, M. Tischler, R. R. Webb II and M. C. Venuti, J. Am. Chem.
Soc., 1994, 116, 5077–5083.
119. K. D. Hargrave, J. R. Proudfoot, K. G. Grozinger, E. Cullen, S. R.
Kapadia, U. R. Patel, V. U. Fuchs, S. C. Mauldin and J. Vitous, et al., J.
Med. Chem., 1991, 34, 2231–2241.
120. J. T. Hunt, C. Z. Ding, R. Batorsky, M. Bednarz, R. Bhide, Y. Cho, S.
Chong, S. Chao, J. Gullo-Brown, P. Guo, S. H. Kim, F. Y. F. Lee, K.
Leftheris, A. Miller, T. Mitt, M. Patel, B. A. Penhallow, C. Ricca, W. C.
Rose, R. Schmidt, W. A. Slusarchyk, G. Vite and V. Manne, J. Med.
Chem., 2000, 43, 3587–3595.
121. W. C. Ripka, G. V. De Lucca, A. C. Bach II, R. S. Pottorf and J. M.
Blaney, Tetrahedron, 1993, 49, 3609–3628.
122. W. C. Ripka, G. V. De Lucca, A. C. Bach II, R. S. Pottorf and J. M.
Blaney, Tetrahedron, 1993, 49, 3593–3608.
123. R. A. Fecik, K. E. Frank, E. J. Gentry, S. R. Menon, L. A. Mitscher and
H. Telikepalli, Med. Res. Rev., 1998, 18, 149–185.
124. IG Farbenindustrie, Ger. Pat., 607537, 1935.
125. W. Sneader, Drug Discovery: A History, Wiley, Chichester, 2005.
126. E. Fromm and J. Wittmann, Ber., 1909, 41, 2264–2273.
127. G. Wozel, Int. J. Dermatol., 1989, 28, 17–21.
128. T. Mann and D. Keilin, Nature, 1940, 146, 164–168.
129. P. W. Feit, J. Med. Chem., 1971, 14, 432–439.
130. G. Erhart, Naturwissenschaften, 1956, 43, 93.
131. F. J. Marshall and M. V. J. Sigal, J. Org. Chem., 1980, 23, 927–929.
132. E. van der Horst, Y. Okuno, A. Bender and A. P. Ijzerman, J. Chem. Inf.
Model., 2009, 49, 348–360.
133. D. A. Price, D. Armour, M. de Groot, D. Leishman, C. Napier, M.
Perros, B. L. Stammen and A. Wood, Bioorg. Med. Chem. Lett., 2006, 16,
4633–4637.
134. T. D. Penning, J. J. Talley, S. R. Bertenshaw, J. S. Carter, P. W. Collins,
S. Docter, M. J. Graneto, L. F. Lee, J. W. Malecha, J. M. Miyashiro, R.
S. Rogers, D. J. Rogier, S. S. Yu, G. D. Anderson, E. G. Burton, J. N.
Cogburn, S. A. Gregory, C. M. Koboldt, W. E. Perkins, K. Seibert, A. W.
Veenhuizen, Y. Y. Zhang and P. C. Isakson, J. Med. Chem., 1997, 40,
1347–1365.
135. P. Prasit, Z. Wang, C. Brideau, C. C. Chan, S. Charleson, W. Cromlish,
D. Ethier, J. F. Evans, A. W. Ford-Hutchinson, J. Y. Gauthier, R.
Gordon, J. Guay, M. Gresser, S. Kargman, B. Kennedy, Y. Leblanc, S.
Leger, J. Mancini, G. P. O’Neill, M. Ouellet, M. D. Percival, H. Perrier,
D. Riendeau, I. Rodger, P. Tagari, M. Therien, P. Vickers, E. Wong, L. J.
Xu, R. N. Young, R. Zamboni, S. Boyce, N. Rupniak, M. Forrest, D.
Visco and D. Patrick, Bioorg. Med. Chem. Lett., 1999, 9, 1773–1778.
136. A. Cuenda, J. Rouse, Y. N. Doza, R. Meier, P. Cohen, T. F. Gallagher, P.
R. Young and J. C. Lee, FEBS Lett., 1995, 364, 229–233.
137. K. H. Bleicher, H. -J. Böhm, K. Müller and A. I. Alanine, Nat. Rev. Drug
Discov., 2003, 2, 369–378.
Drugs and their Structural Motifs 59
138. S. J. F. Macdonald, M. D. Dowle, L. A. Harrison, P. Shah, M. R.
Johnson, G. G. Inglis, G. D. Clarke, R. A. Smith, D. Humphreys, C. R.
Molloy, A. Amour, M. Dixon, G. Murkitt, R. E. Godward, T. Padfield,
T. Skarzynski, O. M. Singh, K. A. Kumar, G. Fleetwood, S. T. Hodgson,
G. W. Hardy and H. Finch, Bioorg. Med. Chem. Lett., 2001, 11, 895–
898.
139. K. A. Jacobson, Drug Dev. Res., 2001, 52, 178–186.
140. M. W. Karaman, S. Herrgard, D. K. Treiber, P. Gallant, C. E. Atteridge,
B. T. Campbell, K. W. Chan, P. Ciceri, M. I. Davis, P. T. Edeen, R.
Faraoni, M. Floyd, J. P. Hunt, D. J. Lockhart, Z. V. Milanov, M. J.
Morrison, G. Pallares, H. K. Patel, S. Pritchard, L. M. Wodicka and P. P.
Zarrinkar, Nat. Biotech., 2008, 26, 127–132.
141. G. Yoshino, T. Kazumi, T. Kasama, I. Iwatani, M. Iwai, A. Inui,
M. Otsuki and S. Baba, Diabetes Res. Clin. Pract., 1986, 2, 179–181.
142. A. Endo, M. Kuroda and Y. Tsuyita, J. Antibiot., 1976, 29, 1346–
1348.
143. A. Endo, Y. Tsuyita, M. Kuroda and K. Tanzawa, Eur. J. Biochem., 1977,
77, 31–36.
144. N. Serizawa, K. Nakagawa and K. Hamano, J. Antibiot., 1983, 36,
604–607.
145. D. J. Triggle, Biochem. Pharmacol., 2009, 78, 217–223.
146. R. A. Wiley and D. H. Rich, Med. Res. Rev., 1993, 13, 327–384.
147. D. A. Smith and E. F. Schmid, Curr. Opin. Drug Discov. Dev., 2006, 9,
38–46.
148. M. T. Keating and M. C. Sanguinetti, Cell, 2001, 104, 569–580.
149. D. A. Price, D. Armour, M. de Groot, D. Leishman, C. Napier,
M. Perros, B. L. Stammen and A. Wood, Curr. Topics Med. Chem., 2008,
8, 1140–1151.
150. J. Scheiber, J. L. Jenkins, S. C. K. Sukuru, A. Bender, D. Mikhailov,
M. Milik, K. Azzaoui, S. Whitebread, J. Hamon, L. Urban, M. Glick and
J. W. Davies, J. Med. Chem., 2009, 52, 3103–3107.
151. J. Blagg, Annu. Rep. Med. Chem., 2006, 41, 353–368.
152. I. Walker and H. Newell, Nat. Rev. Drug Discov., 2008, 8, 15–16.
153. N. N. Taleb, The Black Swan: The Impact of the Highly Improbable,
Random House, New York, 2007.
154. A. Bakker, A. Caricasole, G. Gaviraghi, G. Pollio, G. Robertson, G. C.
Terstappen, M. Salerno and P. Tunici, ChemMedChem, 2009, 4, 1–12.
155. A. Bender, J. Scheiber, M. Glick, K. Azzaoui, J. Hamon, L. Urban,
S. Whitebread and J. L. Jenkins, ChemMedChem, 2007, 2, 861–873.
156. K. Azzaoui, J. Hamon, B. Faller, S. Whitebread, E. Jacoby, A. Bender,
J. L. Jenkins and L. Urban, ChemMedChem, 2007, 2, 874–880.
157. M. Waters, G. Boorman, P. Bushel, M. Cunningham, R. Irvin, A. Merrick,
K. Olden, R. Paules, J. Selkirk, S. Stasiewicz, B. Weis, B. van Houten,
N. Walker and R. Tennant, Environ. Health Perspect., 2003, 111,
811–824.
158. G. Brambilla and A. Martelli, Mutat. Res., 2008, 681, 209–229.
60 Chapter 1
159. E. David, T. Tramontin and R. Zemmel, Nat. Rev. Drug Discov., 2009, 8,
609–610.
160. A. L. Hopkins and A. Polinsky, Annu. Rep. Med. Chem., 2006, 41,
425–437.
161. M. Goodman, Nat. Rev. Drug Discov., 2008, 7, 795.
162. A. Hopkins, J. Lanfear, C. Lipinski and L. Beeley, Annu. Rep. Med.
Chem., 2005, 40, 349–358.
163. T. I. Oprea, Towards Drugs of the Future. Proceedings of the Solvay
Pharmaceuticals Symposium, C. G. Kruse, H. Timmerman, eds. IOS
Press, Amsterdam, 2008, 29–36.
CHAPTER 2

ADMET for the Medicinal

Chemist
K. BEAUMONT,a S. M. COLE,a K. GIBSONb AND J. R.
GOSSETa
a
Department of Pharmacokinetics, Dynamics and Metabolism; b Department
of Medicinal Chemistry, Pfizer Global Research and Development, Sandwich
Laboratories, Sandwich, Kent, CT13 9NJ, UK

2.1 Introduction
In order to become an effective therapy against a disease, a compound must
exhibit potency versus a particular therapeutic target, combined with a degree of
selectivity over other targets; this drives to an appropriate safety margin. In
addition, the vast majority of drugs are delivered at sites which are remote from
the site of action. Over millions of years of human evolution, the body has
developed a host of defence mechanisms designed to protect against exogenous
substances that may cause harm. These mechanisms now form barriers to the
passage of therapeutic drugs from their site of administration to their site of
action. The extent to which a drug can avoid these barriers will to a great extent
determine the therapeutic potential of that particular compound. Successful drugs
need to strike a balance between three major determinants of therapeutic potential:

1) Potency against a pharmacological target to drive efficacy

2) Selectivity over the large number of potential pharmacological/tox-
icological targets to drive safety

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

61
62 Chapter 2
3) Absorption, distribution, metabolism and excretion (ADME) to drive to
an acceptable dose and frequency of administration.

Consequently, the modulation of human ADME of a candidate drug series

has become one of the major goals for a medicinal chemist. In order to achieve
this goal, medicinal chemists need to understand the impact of structure/phy-
sicochemistry on ADME properties and how the human body deals with
chemicals. Furthering this understanding within the medicinal chemistry
community is the aim of this chapter.

2.1.1 Physicochemical Principles for ADME

The ultimate ownership of the physicochemical properties of candidate drugs
lies with the medicinal chemist. Since many of the physicochemical properties
that drive ADME properties are calculable using in silico methods, the med-
icinal chemist should understand the properties of the molecules they design
before they are synthesised. However, this simplistic view conveniently side-
steps the issue that the physicochemical requirements for potency versus a
pharmacological target are often at odds with those required for optimal
pharmacokinetics and a balance must be struck. To understand the trade-offs
required to achieve this balance, it is first appropriate to consider the physi-
cochemical requirements for pharmacological potency.
In crude terms, the pharmacological potency of a compound is determined
by the free energy generated on binding of the molecule at the active site. The
relationship between free energy of binding and Kd versus the target is repre-
sented by eqn (2.1):

DG ¼ RT ln Kd ð2:1Þ

In general, pharmacological potency is achieved by specific interactions

(these tend to be polar in nature) and non-specific lipophilic interactions of the
compound within the active site. Lipophilic groups add binding energy at rate
of 0.7 kcal1 mol1 for every methyl equivalent. Effectively, this suggests that
the levels of desired pharmacological potency can be achieved by adding car-
bon to an active scaffold. If this approach is widely adopted, compounds
become large and lipophilic as well as highly potent. Thus, compound series
optimised solely for potency will tend to high molecular weight and lipophili-
city, a situation that led to significant compound attrition due to poor phar-
macokinetics across the industry prior to the introduction of drug metabolism
into drug discovery.
As previously stated, the human body comprises a series of barriers to drug
movement between the sites of administration and action. These barriers have
been described previously1 and consist of aqueous and lipophilic environments,
as well as a large number of active processes (e.g. enzymic metabolism and
transporter-mediated excretion). To perform well in the body, a compound
must be able to exist in both aqueous and lipophilic environments. The
ADMET for the Medicinal Chemist 63
physicochemistry of a compound will, to a large extent, define how successful it
will be in overcoming these biological barriers.
The most important physicochemical properties from an ADME perspective
are lipophilicity, molecular weight and polar surface area. Lipophilicity is
measured as the partitioning between an aqueous and an organic (lipophilic)
environment. LogD is the partitioning between octanol and buffer at a parti-
cular pH to take into account ionisation—hence log D7.4. Lipophilicity mimics
the passage from aqueous solution through a membrane. When the lipophili-
city of a compound is too low, membrane permeation is less likely. When
lipophilicity is too high, aqueous solubility is compromised and, for example,
oral absorption is unlikely. Molecular weight is important as larger molecules
are likely to have greater difficulty permeating membranes due to steric inter-
actions with triglycerides making up a lipid bilayer. Finally, polar surface area
(PSA), an indicator of hydrogen bonding capacity, also plays a role since it will
define the number of associated water molecules in aqueous solution. For
passage across a membrane, these water molecules need to be removed—a
process that requires energy. Thus, the more hydrogen bonding potential a
molecule has, the more energy will be required to remove associated water
molecules and the less likely it will be to cross a membrane.
The importance of these parameters for oral absorption was recognised over
a decade ago.2 From an analysis of approximately 50 000 compounds, Lipinski
et al. proposed that poor oral absorption was most likely for a compound
when:

 log P was greater than five;

 molecular weight was greater than 500;
 there were more than five hydrogen bond donors; and
 the sum of nitrogen and oxygen atoms was greater than 10.

This is the ‘rule of five’.

The important physicochemical parameters of lipophilicity, molecular weight
and polar surface area are, in fact, highly interrelated. Indeed, lipophilicity can
be viewed as a combination of the molecular weight (how many carbons in the
molecule) with polar surface area (how many heteroatoms). The Lipinski rules
are explainable by the fact that, as molecular weight approaches 500, the
probability increases that either lipophilicity or the polar surface area of a
molecule will be too high, driving to poor aqueous solubility or membrane
permeation respectively.
Other ADME processes are also highly associated with these physicochem-
ical properties. For example, the clearance of a molecule is defined by a com-
bination of structure and physicochemistry. Passive renal clearance is only
highly apparent when log D7.4 is below 0, whereas many metabolising enzymes
are membrane-bound and show a preference for lipophilic compounds. Plasma
protein binding (particularly to human serum albumen) is to a large extent
determined by lipophilicity. Brain penetration requires passage across a
membranous barrier that is a more significant hurdle than the gut wall
64 Chapter 2
membrane and is driven by the lipophilicity, molecular weight and polar sur-
face area of a molecule.
We have established that the ADME trinity of lipophilicity, polar surface
area and molecular weight are interrelated and define the ADME character-
istics of molecules. Smith and coworkers3,4 put forward the concept of ADME
space defined by the three axes of lipophilicity, molecular weight and polar
surface area. The challenge for the medicinal chemist is always to balance
ADME with pharmacological potency, a balance that is largely determined by
the physicochemical properties required to achieve binding to the active site of
the target.
The aminergic G-protein coupled receptors (GPCRs) are exemplified by
natural ligands (e.g. adrenaline and dopamine) that exhibit physicochemical
properties well within ADME space.3 Drugs emerging from these particular
targets (e.g. beta blockers and beta agonists) have tended to share the physi-
cochemical properties of the leads, driven by the size and shape of the active site
to which the modulators (and ligands) need to bind. Thus, drugs against the
aminergic GPCRs tend to fall well within ADME space, meaning that potency
and ADME properties can be rationalised in single molecules with drug-like
potential. In contrast, peptidic GPCR ligands bind to extracellular sites of their
target and tend to require high molecular weight to achieve adequate target
binding. Thus, ligands for these particular targets are not ideally placed to
promote drug-like ADME properties.
Targets that lie outside of ADME space cannot be ignored on the basis that
drugs with an appropriate ADME profile cannot be delivered. When working
under these circumstances, a degree of compromise needs to be considered.
This is exemplified by the HIV protease compounds.5 This topic is be explored
in Chapters 10 and 11.
Another potential method of making a compound series more physico-
chemically tractable from an ADME perspective is the ‘prodrug’ approach.
This has been extensively reviewed.4,6 A prodrug is used to mitigate a physi-
cochemical liability of a candidate through the addition of a pro-moiety that
incorporates the desired physicochemical profile to address the ADME issue
but can be removed once inside the body.
The most successful examples of prodrugs tend to be in the addition of polar
functionality to improve aqueous solubility (and thus oral absorption) and the
addition of lipophilicity to improve membrane permeation (and hence oral
absorption). These approaches are not without their difficulties. In general,
they can only be used successfully when the active drug is close to defined
ADME space. For example, the addition of a pro-moiety to improve mem-
brane permeation will have only limited impact on the degree of hydrogen
bonding4 (and hence polar surface area) and will derive most benefit from the
addition of lipophilicity to the active moiety. The improvement in oral
absorption will then depend upon the balance of lipophilicity and aqueous
solubility as well as the extent of release of active principle from the prodrug
once absorbed. The further away from ADME space the active principle is, the
more lipophilicity will need to be added and the more likely the prodrug
ADMET for the Medicinal Chemist 65
strategy will fail. Therefore, the prodrug strategy is not an easy win for the
medicinal chemist confronted with an ADME space issue.
It is clear that the pharmaceutical industry cannot ignore the challenge of
targets that require physicochemistry outside of defined ADME space. How-
ever, the medicinal chemistry challenge of approaches to mine drugs out of
these targets is an appreciable one. To enable these targets to yield acceptable
drugs, a degree of compromise on the quality of the pharmacokinetic profile
will need to be accepted. When working in these areas, the risks of slipping into
unacceptable ADME properties will be significant and predictability will be
lower. This suggests that larger numbers of compounds will need to be taken to
the clinic at risk to achieve the goal.
From a medicinal chemistry perspective, this does not signal a return to the
days of optimising chemical matter solely on the basis of potency. In recent
years the terms of ‘ligand efficiency’7 and ‘lipid efficiency’8 have entered the
literature. These calculations allow medicinal chemists to pose questions about
whether they are using molecular weight and lipophilicity to the greatest
advantage to improve potency. If used appropriately, we can be sure that each
atom and each degree of lipophilicity is being used optimally in compound
design.

2.1.2 Physicochemistry Summary

The therapeutic potential of a drug is defined in large part by its potency,
selectivity over other targets and ADME properties. Potency and ADME
properties are strongly defined by structure and physicochemistry. The extent
to which a compound overcomes the barriers to drug delivery in the body will
be defined by lipophilicity, molecular weight and polar surface area. The series
is considered chemically tractable when the physicochemistry required for
potency is within the ADME space defined by lipophilicity, molecular weight
and polar surface area. However, many new targets require physicochemistry
that is outside of ADME space, requiring alternative approaches.

2.2 Delivery of Drugs and Bioavailability

Typically, drugs are delivered at sites remote from the site of action. The major
drug delivery routes are exemplified in Table 2.1.
The movement of a drug away from the site of administration is termed the
absorption of that drug and the extent to which a drug overcomes the barriers
to its passage to its site of action is termed the bioavailability. The distinction
between the two terms is important as a drug can be completely absorbed but
exhibit no oral bioavailability due to post-absorption metabolism.
In our description of the pharmacokinetics of a drug, the site of action is
assumed to be the blood, which is also the site of measurement. For the intra-
venous route, all the dose is delivered directly into the blood. Consequently, the
66 Chapter 2
Table 2.1 Overview of the potential routes of administration of drugs and the
barriers to those routes.
Route Barriers Comment

Intranasal Mucocilliary clearance Limited to low doses

(solubility)
Buccal Swallowing Limited to low doses
(solubility)
Inhalation Deposition in mouth and Limited to low doses, patient
throat needs device training
Dermal Skin poorly permeable Limited to low doses, acci-
dental removal of dose by
washing etc.
Intravenous None Limited to in-clinic delivery
Oral Dissolution, gut perme- Can deliver range of doses
ability, first pass metabo- conveniently
lism in gut and liver
Subcutaneous and Blood flow to site and irrita- Low doses
intramuscular tion of tissue

bioavailability of an intravenous dose is always 100% and all other dose routes
must be measured relative to an intravenous dose.
Intravenous administration gives complete bioavailability and the potential
to rapidly achieve therapeutic concentrations. A good example of a drug that is
given by intravenous infusion is lidocaine, which is used to treat severe neu-
ropathic pain in the clinical setting.9
There are several important considerations for a drug to be administered by
the intravenous route. First, it must be sufficiently soluble in the proposed
formulation (mainly saline with limited potential for organic co-solvents) to
deliver the whole dose in an acceptable volume. Secondly, the compound must
have prolonged chemical stability in the formulation, especially as the dose
must be taken through sterilisation procedures. Overall, the requirement for
venous cannulation means that intravenous administration tends to be limited
to severe and life-threatening indications, and usually under qualified medical
supervision.

2.2.1 Oral Delivery

Most drugs are formulated for delivery by the oral route. For the patient this
route is the most convenient method of drug delivery. However, from a
pharmacokinetic point of view, it represents a significant challenge. There are a
number of barriers via the oral route that can limit oral bioavailability of a drug
(Figure 2.1).
Modulation of oral bioavailability is a key parameter for many drug dis-
covery programmes. High oral bioavailability is often required to limit inter-
patient variability and maintain an acceptable dose size.
ADMET for the Medicinal Chemist 67

Gut Lumen Gut Wall Portal

Vein

Efflux
Transporter Liver
Paracellular
Absorption
Dissolution
Transcellular
Absorption

Degradation Drug in
Solution
Influx
Transporter Metabolism
Blood
Enterocyte

Plasma
Site of Measurement

Gut Metabolism
Excretion to Faeces

Figure 2.1 Summary of the barriers to drug delivery following oral administration.

The human gastrointestinal tract has evolved to be the organ of food

digestion and nutrient absorption. In addition, it has in place a number of
mechanisms for preventing the entry of unwanted molecules (e.g. toxins). These
mechanisms often limit the oral bioavailability of drugs.
The physiology of the human gastrointestinal tract consists of three main
parts: the stomach, small intestine and large intestine. These are considered in
turn below.
The first organ that a drug will reach following oral administration is the
stomach. The major function of the stomach is the preparation of food for
further digestion. Gastric pH is acidic, ranging from 2 to 6, depending on the
presence of food.
The small intestine is the major absorbing organ. It is divided into three
parts: duodenum, jejunum and ileum. These regions display differences in their
absorptive and secretory capabilities. The duodenum is responsible for neu-
tralising the gastric acid to near physiological pH in preparation for absorption
across the gut wall.
The major adaptation of the small intestine is the membrane surface area
available for oral absorption. There is extensive inward folding on the luminal
68 Chapter 2
surface of the small intestine. On the surface of these folds are finger-like
projections called villi containing absorptive epithelial cells. These cells are
characterised by small projections on their luminal surface called microvilli,
which form the brush border membrane. The combination of luminal folds,
villi and microvilli increases the membrane surface area for absorption by
600-fold over that of the internal surface of a cylindrical surface alone.
The large intestine comprises three sections: cecum, colon and rectum. The
major role for the large intestine is reabsorption of water involved in digestion.
It is not a major absorptive organ as fluid for dissolution is limited and the
surface area for absorption is far less than for the small intestine.
To design compounds with acceptable oral bioavailability, it is important to
understand a drug’s journey from its site of administration into the systemic
circulation. There are a number of steps in this process.

2.2.1.1 Dissolution
To be absorbed a drug must be in solution in the gastrointestinal tract lumen.
The fluids in the lumen are predominantly aqueous in nature with a pH ranging
from acidic to physiological. Consequently, an important parameter for the
medicinal chemist to modulate within a chemical series is the aqueous
solubility.
Oral drugs tend to be administered in tablet or capsule formulation. The
formulation must first disintegrate before the active drug undergoes a dis-
solution process. These processes are governed in part by excipients in the
formulation and in part by the physical form and aqueous solubility of the
formulated drug.
Food may affect the absorption of drugs through the dissolution process.
The delay in gastric emptying with food after a meal may result in a longer time
to achieve maximal plasma concentrations, since the drug is held in the stomach
for longer than in the fasted state. The release of bile salts on ingestion of food
may result in enhanced dissolution of poorly soluble drugs. In addition, certain
compounds may form complexes with food resulting in a reduction in their
absorption.10

2.2.1.2 Permeation Across the Gastrointestinal Tract Epithelium

Once in aqueous solution, a drug must cross the gastrointestinal tract epithe-
lium to be absorbed. The gut wall epithelial cell is called the enterocyte. The
drug can cross the enterocyte layer by passive diffusion through the tight
junctions between the enterocytes (paracellular absorption) or by crossing the
enterocyte membrane (transcellular absorption). The route for passive
absorption of a drug depends upon the physicochemistry of the drug.
Paracellular absorption involves the passage of the drug through the aqueous
filled channels between the epithelial cells. Due to this size constraint, drugs
that are predominantly absorbed by the paracellular route are generally of low
ADMET for the Medicinal Chemist 69
molecular weight (o300) and relatively polar (log D7.4o0). An example of
such a drug is the beta blocker, atenolol. Paracellular absorption is limited to
the small intestine as, in the large intestine, aqueous pores are fewer in number
and smaller. In addition, there is an important species difference in paracellular
absorption with pore size being comparable in human and rodent, but sig-
nificantly larger in the dog.11 For this reason, the oral bioavailability of ate-
nolol is limited to approximately 50% in rat and human, but is complete in
dog.12,13
The majority of oral drugs cross the enterocyte by the transcellular route. As
outlined previously, the physicochemistry of the drug will determine the extent
of transcellular absorption. Lipophilicity, molecular size and hydrogen bonding
potential will in large part determine the ability of a drug to cross the gut wall
epithelial cell membrane (see above).
However, the vision of the enterocyte membrane as a simple lipid bilayer is
overly simplistic. There are a number of proteins expressed within the mem-
brane that have the potential to either facilitate (active absorption) or hinder
(active efflux) the passage of drug molecules across the membrane.

2.2.1.3 Active Absorption of Drugs

Given that the major role of the gastrointestinal tract is the digestion of food
and the absorption of nutrients, proteins have evolved to facilitate the active
uptake of nutrients with poor membrane permeation characteristics (e.g. di-
and tri-peptides). A small number of drugs take advantage of these uptake
transporters to facilitate their absorption into the body.
For example, the human peptide transporter 1 (hPEPT1) is a low affinity,
high capacity system expressed on the apical brush border membrane of
enterocytes.14,15 hPEPT1 has been shown to mediate the transport of di- and
tri-peptides into the systemic circulation from the gut lumen utilising a H1
gradient dependent transport system. This influx transporter has been impli-
cated in the absorption of a diverse range of drugs including b-lactam anti-
biotics such as cefadroxil (Figure 2.2).16
Targeting intestinal transporters by means of prodrugs has been a successful
strategy for improving oral absorption. Uptake via hPEPT1 is thought to be
the primary mechanism for valacyclovir absorption (Figure 2.3).17 Once
absorbed, valacyclovir is hydrolysed to its active form acyclovir. Consequently,

HO O

O
N
O Cefadroxil
OH
S N
H H
H2N

Figure 2.2 Structure of cefadroxil.

70 Chapter 2

O O
N N

H2N N N H2N N N
H O H O

O
O OH
Valacyclovir Acyclovir
NH2

Figure 2.3 Structure of valacyclovir.

Cl

Cl

O N N
N NH2
N S
O
O

Figure 2.4 Structure of UK-224 671.

valacyclovir pharmacokinetics are characterised by non-linear absorption as a

result of the saturation of this influx process at higher doses.18

2.2.1.4 Active Efflux of Drugs

In addition to uptake transporters, a number of efflux transporters are
expressed at the gut lumen facing (apical) membrane of the enterocyte to limit
the passage of molecules that could be potentially damaging to the body. They
can act as a significant barrier to the absorption of drugs.
The most well-studied drug efflux transporter is P-glycoprotein (P-gp). P-gp
is an ATP-binding cassette transporter highly expressed in many human tissues
including the enterocyte.19 It intercepts compounds as they pass through the
enterocyte membrane and effluxes them back into the gut lumen.20
The functional activity of P-gp is saturable. P-gp substrates that are poorly
membrane permeable and that are administered at non-saturating doses will
demonstrate limited oral absorption or non-linear absorption.
There are a number of examples where P-gp has been shown to affect the oral
absorption of its substrates. UK-224 671 is a potent selective NK2 receptor
antagonist (Figure 2.4) that has been shown to be a substrate for P-gp.21,22 In
P-gp knockout mice, the oral bioavailability of UK-224 671 was 22%, whereas
in the wild-type mice expressing P-gp, the value was less than 5%. The low oral
bioavailability translated to humans where UK-224 671 was shown to be
ADMET for the Medicinal Chemist 71

O N
O HN N
N
N

O S O
N

Figure 2.5 Structure of UK-343 664.

approximately 10% absorbed, presumably as a consequence of limited mem-

brane permeation and P-gp mediated efflux.
UK-343 664 (Figure 2.5) is a potent and selective phosphodiesterase 5
(PDE5) inhibitor.23 Over an 80-fold oral dose range (10–800 mg), it displayed a
1300-fold increase in Cmax and a 1900-fold increase in AUCt for UK-343 664.
This was accompanied by a reduction in the Tmax from 3.5 to 0.6 h. This non-
linear profile was ascribed to saturation of P-gp in the gut as the doses were
increased. At the lower doses, gut lumen concentrations were below the Km for
P-gp and consequently it was possible to delay the rate and impact of the extent
of absorption of UK-343 664. As the dose increased, the gut lumen con-
centrations increased significantly above the P-gp Km and this saturation led to
an increase in the rate and extent of absorption of the compound.
It is possible for co-administered P-gp substrates to promote a P-gp drug–
drug interaction by inhibition of this efflux transporter. For example, the oral
bioavailability of digoxin can be enhanced by co-administration of talinolol as
a result of P-gp inhibition.24

2.2.1.5 First-Pass Metabolism

Once a drug has crossed the enterocyte membrane, it has been absorbed into
the body. However, it has not yet reached its site of action and two significant
further metabolic barriers need to be overcome, both of which can contribute to
reduction in the overall bioavailability of the drug.

(a) Gut Wall Metabolism. The small intestine contains enzymes that can
metabolise drugs through both phase 1 (oxidative) and 2 (conjugation) reac-
tions.25,26 The involvement of the gut wall in the metabolism of drugs has
been extensively reviewed.27,28
A range of P450 enzymes have been identified in the human small intestine
including CYP3A4, 3A5, 2C9, 2C19, 2J2 and 2D6. CYP3A is the most highly
expressed CYP accounting for around 80% of the total CYP content.29 Even
72 Chapter 2
though the total mass of CYP3A in the intestine is only 1% of that in the liver,
clinical studies have demonstrated that intestinal metabolism can have a sig-
nificant effect on the overall first-pass metabolism of some drugs.30,31
The oral bioavailability of midazolam in humans has been estimated at
approximately 30%.32 This is significantly lower than would be expected from
the plasma clearance of midazolam (approximately 5 ml min1 kg1) with
respect to human hepatic blood flow, knowing that midazolam is completely
absorbed. When midazolam was administered to patients undergoing hepatic
transplantation,33 the gut wall first-pass extraction following intraduodenal
administration was calculated to be 0.43 (range 0.14 to 0.59). Thus, the oral
bioavailability of midazolam is a consequence of complete oral absorption and
loss of approximately 40% of the drug on first-pass through the gut wall, with
the remainder lost by hepatic first-pass extraction.
The gut wall first-pass extraction of drug molecules is thought to be enhanced
by the expression of CYP3A4 and P-gp together close to the apical membrane
of the enterocyte. It is thought that these two proteins can act as a concerted
barrier to the passage of their substrates across the enterocyte. The P-gp
intercepts its substrates on passage through the enterocyte and effluxes them
back into the gut lumen. Thus, the drug is cycled in and out of the enterocyte,
allowing CYP3A4 a number of opportunities for metabolism. In this way, it is
possible for a relatively low expression of enzyme to exert a significant gut wall
first-pass extraction.

(b) Hepatic. The entire blood supply of the upper gastrointestinal tract pas-
ses into the hepatic portal vein which flows directly to the liver. Conse-
quently, drug that has crossed the enterocyte and avoided gut wall
metabolism is also subject to first-pass extraction by the liver. Extraction of
drugs by the liver is discussed in detail later.

2.2.1.6 Oral Bioavailability

Oral bioavailability is the amount of drug that reaches the systemic circulation
following oral administration relative to the same intravenous dose of the same
compound. The oral bioavailability (F) of a drug is determined by the fraction of
the dose absorbed (Fa), the fraction escaping gut wall first-pass extraction (Fg)
and the fraction escaping hepatic first-pass extraction (Fh) as shown in eqn (2.2):

F ¼ Fa  Fg  Fh ð2:2Þ

Given that the physicochemical factors determining absorption and first-pass

extraction are often in opposition, the oral route provides significant challenges
to the medicinal chemist in attaining the goal of high oral bioavailability.
When it proves too difficult to balance the physicochemistry for potency with
that required for extensive oral bioavailability, it is possible to consider alter-
native delivery routes.
ADMET for the Medicinal Chemist 73
2.2.2 Intranasal Delivery
The nasal route has become a popular alternative to the oral route because it
provides an easy method of administration. The nasal cavity presents the main
barrier to drug absorption; low permeability of the nasal epithelium, rapid
mucociliary clearance and enzymatic degradation in the mucus layer are lim-
iting factors. The nasal cavity is easily accessible and is extensively vascu-
larised.34 Compounds that are administered by this route avoid hepatic first-
pass metabolism and so absorption will determine bioavailability.
Several peptides and proteins are given as nasal sprays.35 These tend to be
inactive when given orally, as they are quickly destroyed in the gastrointestinal
tract, but enough is absorbed from the nasal mucosa to provide a suitable
therapeutic effect.
Nasal administration is primarily suitable for highly potent, low dose and
aqueous soluble drugs since only a limited volume can be sprayed into the nasal
cavity.

2.2.3 Inhaled Delivery

The inhaled route has traditionally been used for the delivery of small mole-
cules to the lung (e.g. in the treatment of asthma). An example is the b2 agonist,
salmeterol. This asthma drug is administered by inhalation device at doses as
low as 50 mg for the treatment of the bronchoconstriction associated with an
asthma attack.36 Due to its rapid systemic clearance, plasma concentrations of
salmeterol are low, which reduces the potential for systemic side effects. Thus,
salmeterol can be considered lung focused.
More recently, the lung has been considered as a delivery route for large
macromolecules37 by virtue of its large surface area for absorption and the
highly vascularised nature of the tissue.

2.2.4 Sublingual Delivery

Delivery of drugs via the oral cavity is a useful route of administration. In the
oral cavity, the primary absorptive areas are the non-keratinised buccal and
sublingual mucosae. The buccal mucosa is the lining of the cheek and the lips,
while the sublingual mucosa is the membrane on the ventral surface of the
tongue and floor of the mouth. The oral mucosa is relatively permeable and has
an abundant blood supply, which drains from the mouth to the superior vena
cava.
Drugs absorbed from the mouth pass straight into the systemic circulation
without entering the hepatic portal system, thus escaping first-pass metabolism
and degradation by gastric acids.
Glyceryl trinitrate (for the treatment of acute angina) is given sublingually38
to avoid extensive first-pass metabolism following oral administration. Its
lipophilicity ensures rapid and effective absorption via the sublingual route.
Disadvantages of this route include potential irritation of the mucous
74 Chapter 2
membrane, as well as excessive salivation which promotes swallowing and
hence loss of dose.

2.2.5 Rectal Delivery

The rectal mucosa is highly vascularised with a rich blood and lymph supply.
However, absorption through the rectal mucosa is often unreliable and
incomplete. The route a drug takes depends on how it distributes within the
rectum and this is somewhat unpredictable. If a drug is absorbed from the
lower rectum via the inferior or middle haemorrhoidal veins, it can avoid first-
pass metabolism because these veins empty into the vena cava and bypass the
hepatic portal system. Substances that cross the upper rectal mucosa will be
carried by the superior haemorrhoidal vein to the hepatic portal circulation.
The rectal route can be useful to patients who have difficulty in swallowing or
have nausea or gastric pain. It may offer an advantage for drugs that are
destroyed by gastric acid or by microflora or enzymes in the intestine. In
addition, drugs may be given rectally when oral ingestion is prohibited because
the patient is unconscious. An example of this is the rectal administration of
diazepam employed in the acute treatment of epileptic seizures.39

2.2.6 Transdermal Delivery

Transdermal absorption of a drug through the skin to the systemic circulation
can occur via a transfollicular or transepidermal pathway. To be effective in
any transdermal delivery system, a drug must be absorbed in sufficient quan-
tities and the extent of absorption depends upon molecular weight and the lipid
and water solubility of the drug. Physiological factors including surface area of
application, skin condition and location can affect drug penetration.
Transdermal application is of utility for systemic delivery of small lipophilic
and potent molecules that require low input rates to achieve effective therapy.
An example of a drug that is administered transdermally is fentanyl for the
treatment of pain.40,41

2.2.7 Subcutaneous and Intramuscular Administration

Subcutaneous administration can be used in either acute or chronic therapies
and may be self-administered. It is often the route of choice for large molecules
such as insulin. Drug is either injected or delivered via a device placed in the
interstitial tissue beneath the dermis—most commonly in the upper arm, the
upper thigh, the lower part of the abdomen and the upper part of the back.
Intramuscular administration involves the injection of drug into the mus-
cular layer below the subcutaneous tissue. The most common sites of admin-
istration are the shoulder, the buttocks and the upper thigh muscles.
Absorption of most drugs administered via these routes is dependent on
blood flow to the site as the capillary wall between the injection site and the
ADMET for the Medicinal Chemist 75
blood offers little resistance to the passage of drugs. Absorption can be
manipulated by heating, vasodilation through massaging the injection site or
pharmacologically to increase the blood flow and thus absorption rate.
Absorption can occur through the blood or lymph, which is mainly determined
by the molecular weight of the drug. Lower molecular weight drugs are
absorbed into blood vessels whilst higher molecular weight molecules of greater
than 2000 daltons preferentially use the lymphatic system. Other properties that
influence drug absorption via these routes include those affecting dissolution
rate and lipophilicity.
An example of a drug administered either subcutaneously or intramuscularly
is interferon a2a for the treatment of hepatitis C or certain cancers.42

2.3 Tissue Distribution of Drugs

The pharmacokinetic approach that considers the blood as the site of action for
a drug is overly simplistic. Most drugs modulate targets that are in the tissues
supplied by the blood. As stated earlier, all of an intravenous dose is admi-
nistered into the blood. On administration, several distribution processes begin
to take place. Within the blood itself, the drug can bind to plasma proteins and
distribute into the cellular component of the blood (mainly erythrocytes).
Blood is rich in proteins which make up the plasma. The predominant
plasma proteins are albumin, alpha 1 acid glycoprotein and lipoproteins. Drugs
can bind to these proteins, with the extent of binding dependent on their
physicochemistry. Highly lipophilic compounds tend to show high binding to
albumin. In addition, acidic groups will also increase binding to albumin as
these interact with the basic lysine residues of the protein. Thus, lipophilic acids
tend to be most highly bound to plasma protein. Neutral compounds will bind
to albumin with the extent being solely determined by lipophilicity. Basic
compounds bind to alpha 1 acid glycoprotein by virtue of their positive charge
and to albumin by virtue of their lipophilicity. Plasma protein binding is an
important parameter to understand in drug disposition and further detailed
information is available in the literature.43–47
In addition to binding to plasma proteins, compounds can distribute into the
cellular component of blood. From a pharmacokinetic perspective, this is a
much neglected compartment since most bioanalytical methods measure
plasma concentrations and dispose of the cellular fraction.
Drug distribution into erythrocytes has been extensively reviewed.48 Drugs
can distribute into erythrocytes through non-specific interactions with mem-
branes and proteins within the erythrocyte such as haemoglobin. In addition,
certain drugs (e.g. acetazolomide) can bind specifically to carbonic anhydrase
and as such exhibit extensive erythrocyte distribution. Drugs that distribute
evenly between the plasma and cells of blood will exhibit a Kb/p of 1 and the
plasma pharmacokinetics will reflect those of blood. However, extensive dis-
tribution (Kb/p41) will lead to a disconnect between the plasma pharmaco-
kinetics and those determined in blood. Thus, it is important to understand the
76 Chapter 2
blood distribution behaviour of a compound. The Kb/p for acetazolomide is
2.9.48
Arterial blood flows to the tissues and into the microvasculature within the
tissues. The endothelial walls of the blood vessels within most tissues are
relatively leaky, which allows certain constituents of the blood to perfuse the
cells that make up the tissues. The aqueous portion of the blood readily escapes
the blood vessels, whereas the plasma proteins tend to be slower to diffuse due
to their molecular size. Erythrocytes do not generally escape the micro-
vasculature. Consequently, drugs that are bound to plasma proteins or exten-
sively distributed into erythrocytes tend to have more difficulty than unbound
drug in distributing into the tissues.
Physicochemical factors generally determine the extent of distribution of
drugs not bound to plasma proteins. Tissues are made up of cells, which in turn
are predominantly made up of cell membranes. Cell membranes are made up of
phospholipids, which are amphoteric in nature. The phospholipid membranes
arrange into bilayers with the negatively charged phosphate head groups on the
outside facing the extracellular water and the triglyceride tails on the inside,
producing a highly lipophilic environment. Lipophilic bases tend to permeate
these membranes well as the basic group interacts with the acid groups of
phospholipids on the membranes and the lipophilicity facilitates passage
through the hydrophobic lipid core. Lipophilic acids, as well as being highly
bound to albumin, also tend to be repelled by the negatively charged phos-
pholipid heads and consequently their tissue distribution is limited. The extent
of distribution of neutral molecules tends to be determined by their lipophili-
city, driven by interaction with the triglyceride tails of the cell membrane.49
Consequently, medicinal chemists can utilise physicochemical approaches to
calculate expected distribution of drugs in the body either prior to synthesis
(virtual molecules) or prior to in vivo administration.46,50–52 Such knowledge of
physicochemistry can also be used in compound design (e.g. to improve dis-
tribution to the required site or to generally improve the pharmacokinetics of a
drug). For example, the introduction of basic centres has been used successfully
to increase elimination half-life and thus the duration of effects of a number of
drugs relative to their neutral analogues.53–55
In terms of pharmacokinetics, the extent of distribution of a drug is repre-
sented by the volume of distribution. This parameter relates the total amount of
drug in the body at any given time to the plasma concentration at that time. It
can only be calculated following intravenous administration. In its most simple
form, it is obtained by extrapolating the terminal elimination phase back to
determine the concentration at time zero. This is the volume the total dose
would have to be instantaneously distributed into to achieve the time zero
plasma concentration, C0 (Figure 2.6).
Immediately following intravenous administration (A in Figure 2.6), all of
the drug will be in the blood as no tissue distribution has taken place. Since the
blood is the site of measurement, measured concentrations will be at their
highest. Several minutes after the dose (B in Figure 2.6), blood will have flowed
around the body many times, allowing distribution of the drug into the tissues
ADMET for the Medicinal Chemist 77

Figure 2.6 Plasma concentration versus time curve for an intravenous dose and
corresponding schematic of drug distribution at early times following
intravenous administration.

to take place. This will occur by the law of mass action down a concentration
gradient. At this point, the rate into the tissues will be greater than the rate of
return into the blood. When tissue distribution and blood clearance (see below)
is taken into consideration, the concentrations in the blood will fall rapidly over
this period.
At some point following intravenous administration (C in Figure 2.6), the
unbound concentrations in the tissue will be in equilibrium with the unbound
78 Chapter 2
concentrations in the blood. This is termed the steady state of distribution.
After this point, tissue distribution is complete and plasma concentrations
decline solely as a result of clearance from the blood. Extrapolation of this
phase back to the concentration axis will give the concentration of the drug at
time zero (Co) or the blood concentration if all the drug was distributed
instantaneously following dosing. This is related to the distribution volume
(Vd) and the dose by eqn (2.3):

Co ¼ Dose=Vd ð2:3Þ

The volume of distribution of a drug is not related to an actual volume, but

instead when compared with various physiological volumes, it provides an
indication of how well a drug distributes into tissues. Compounds that are
primarily restricted to the blood compartment will exhibit volumes of dis-
tribution that reflect blood volume (i.e. approximately 0.15 L kg1). Since
blood volume is fixed and blood is the compartment that is measured, it is not
possible for a compound to exhibit a volume of distribution value less than
blood volume. Compounds that distribute into extracellular water, but which
are not permeable enough to cross into the cell itself, will tend to exhibit
volumes of distribution equivalent to extracellular water (i.e. around
0.4 L kg1). Compounds that equally distribute into cells can exhibit volumes of
distribution equivalent to total body water (i.e. approximately 0.7 L kg1).
Finally, compounds that enter tissues and bind extensively will exhibit volumes
of distribution in excess of total body water (i.e. any value above 1 L kg1).
The relationship between volume of distribution and physicochemistry has
long been established.49 This is exemplified in Figure 2.7.
The determination of volume of distribution has traditionally always
required intravenous administration. However, it is now possible to use phy-
siologically based pharmacokinetic modelling (PBPK) to predict the distribu-
tional behaviour of compounds in silico. These models divide the body into a
number of compartments and determine the partitioning into each tissue either
by measurement or (more likely) by mechanistic equations that are based on
tissue physiology and the phospholipid composition of tissues. These can be
used to estimate distribution based on the physicochemistry of each com-
pound.56–58
In order to exert a pharmacological response the drug must bind to its target.
In the majority of cases, only unbound drug is available to interact with the
target (Figure 2.8).
As stated earlier, drug is absorbed (or administered) into the blood, where it
can bind to plasma proteins, distribute into red cells or begin to distribute into
tissues. An equilibrium will eventually be reached between binding in blood and
tissues. Assuming that the drug only distributes passively and that the laws of
mass action apply, then at equilibrium unbound drug concentrations should
be equivalent in all phases (i.e. the unbound concentration in the tissues should
be the same as the unbound concentration in the blood). Determination of
unbound concentrations in tissues is technically challenging. However,
ADMET for the Medicinal Chemist 79

Figure 2.7 Relationship between logD(7.4) and Vd for acids, bases and neutral
compounds.

Absorption

Free Drug Free Drug Free Drug

Receptor Tissue bound Plasma bound

bound drug drug drug

Clearance

Figure 2.8 The role of free drug.

unbound drug in blood is relatively easy to determine and is used as a surrogate

for target concentrations.
This concept is central to drug pharmacodynamics.49,59,60 Supporting evi-
dence for this hypothesis is shown in Figure 2.9. For a series of G-protein
80 Chapter 2

Figure 2.9 The relationship between unbound plasma concentration at efficacious

dose and receptor occupancy for a series of GPCR antagonists (ref. 49).

coupled receptor (GPCR) antagonists, the unbound concentrations at ther-

apeutic doses show a correlation with the concentration required for 75%
receptor occupancy at the target.
Unbound concentrations in blood will only reflect those in the tissue if there
is free movement of the drug between the blood and the tissue. In many tissues
this is the case, but there are a number of tissues that are behind significant
barriers to drug movement. The brain is the most important of these cases since
many targets for drug modulation are in the central nervous system (CNS).

2.3.1 Distribution to the Central Nervous System

The endothelial cells of the blood vessels supplying the brain have evolved to
form the blood–brain barrier, which can act to exclude the passage of drug
molecules. The adaptations include:

 tight junctions that are extremely resistant to the passage of drugs;

 membrane phospholipid composition that differs from other endothelial
cell membranes making them more rigid and resistant to passive
permeation;
 expression of efflux transporters (e.g. P-gp)61–63 and metabolizing enzymes.

These adaptations make the blood–brain barrier a significant hurdle for a

drug to cross in order to reach the site of action. Consequently, the physico-
chemical constraints for a CNS targeted drug are much more severe than for a
non-CNS drug (Figure 2.10).
As with all membrane permeation, lipophilicity, molecular size and hydrogen
bonding capacity are important determinants of brain penetration. A drug is
more likely to penetrate the CNS when polar surface area is below 90Å and
molecular weight (MW) is below 450, in contrast to non-CNS drugs where the
ADMET for the Medicinal Chemist 81

750

650

550

450
MW

350

250

150 non-CNS drugs

CNS drugs
50
0 40 80 120 160 200
PSA (polar surface area)

Figure 2.10 Physicochemical parameters for CNS and non-CNS drugs.

H
N
Cl
NH N
H
N N N
S S Cl S
O O Cl
O O
OMe

SB 271046 699929
Brain:Blood = 0.05:1 Brain:Blood = 3:1

Figure 2.11 Conformational constraint and removal of hydrogen bonding capacity

to promote CNS penetration.

physicochemical space is significantly larger. This reflects the relative difficulty

of permeation across the blood-brain barrier versus the enterocyte membrane.
Physicochemical strategies to improve blood–brain barrier penetration
have been suggested.64 The potent 5HT6 antagonist SB271046 exhibited poor
CNS penetration in rat (brain to blood ratio 0.05 to 1) by virtue of its poor
membrane permeation and P-gp substrate potential. A medicinal chemistry
strategy to remove the acidic NH, reduce hydrogen bonding potential and
restrict conformation led to the identification of 699929. This compound
exhibited greater CNS penetration in rat with a brain to plasma ratio of 3 to 1
(Figure 2.11)
Due to the presence of the blood–brain barrier, the unbound concentration
in the plasma may not reflect that at a CNS target. As in other tissues it is the
82 Chapter 2
unbound drug concentration in the brain that is important. Strategies that drive
compound into the brain using excessive lipophilicity are likely to fail, since
unbound concentrations are likely to be low.
A method has been developed in which brain binding in relation to
plasma protein binding is compared and used to judge whether a compound
will distribute well into the CNS. If this is the case, then unbound drug mea-
sured in plasma can be used as a surrogate for that in brain.65 This topic of
blood–brain penetration is also discussed in Section 11.8 when future targets
are considered

2.4 Clearance, Extraction, Metabolism and Excretion

The aim of many drug discovery projects is to provide a drug candidate with a
duration of action consistent with the required dose regimen. For many
approaches, the desired regimen is once per day. In order to produce a once
daily regimen, the drug must exhibit a significantly long elimination half-life
relative to the dose interval (every 24 h). The elimination half-life is determined
by the clearance and the volume of distribution. In addition, the importance of
clearance in defining dose is exemplified by eqn (2.4):

Css;avg  CL  t
Dose ¼ ð2:4Þ
F

where:

Css,avg is the average concentration at steady state (target concentration)

Cl is the plasma clearance of the molecule
t is the dose interval
F is the fraction bioavailability by the particular route.

Thus, clearance is an important parameter for the medicinal chemist to

understand as it will determine the dose necessary to achieve pharmacologi-
cally active concentrations. Indirectly, clearance will determine the frequency
of dosing and drive peak-to-trough ratio as a component of elimination
half-life.

2.4.1 Clearance
The clearance of a molecule is defined as the amount of blood (blood clea-
rance) or plasma (plasma clearance) cleared of drug per unit time and body
weight. The units of clearance are ml min1 kg1. Clearance is dependent
upon the ability of organs to metabolise or excrete the compound, the
plasma protein binding of the molecule and the blood flow to the clearing
organ. The importance of clearance as a parameter is that, for a given dose, it
defines the exposure (or AUC) to a given compound as determined by
ADMET for the Medicinal Chemist 83
eqn (2.5):

Dose
Clearance ¼ ð2:5Þ
AUC
Thus, a compound with a low clearance will always exhibit a high exposure
at a given dose, relative to a compound with a higher clearance.
The free drug hypothesis suggests that therapeutic efficacy is driven by the
unbound concentrations of the drug. Thus, unbound clearance is the plasma
clearance corrected for the fraction unbound in plasma.
In addition, the potential amount of blood cleared of drug in the absence of
flow limitations and plasma protein binding considerations is defined as the
unbound intrinsic clearance (Clintu) of that particular molecule. For metabo-
lised compounds, this is an important parameter as it is related to the Vmax
and Km of the enzyme for metabolism of any particular compound. This is the
parameter that medicinal chemists attempt to reduce when required to modify
the plasma clearance of any particular molecule. Clintu is related to the clear-
ance by the well stirred model as described by eqn (2.6):

fu  Clintu
Cl ¼ ð2:6Þ
fuðClintu þ QÞ

where fu is the fraction unbound in plasma and Q is the hepatic blood flow.

2.4.1.1 Organ Extraction

The major drug metabolising and excreting organs are the liver, the kidney and
the gut. The extent to which a compound can avoid metabolism and excretion
by these organs will define the clearance and hence the dose and duration of
action of that particular molecule. The fraction of a drug removed from the
blood on single perfusion through any particular organ is termed the extraction
of that compound.
For instance where extraction by an organ is 0.5 (50% of flow through it) the
concentration of a particular compound is halved by passage through that organ.
The clearance by that organ relates to the blood flow and the extraction (meta-
bolism etc.). This process is continuous for all further passages through the
extracting organ, such that plasma concentrations of the compound will fall in an
exponential manner. Consequently, if we know the blood flow to a single
extracting organ and the extraction ratio across that organ, we can calculate
amount of blood freed per unit time (i.e. the clearance of that drug) from eqn (2.7):

Clearance ¼ Extraction ratio  Organ blood flow ð2:7Þ

However, since many drugs are cleared by more than one organ and single
human organs cannot be studied in isolation, the clearance of a drug is always
calculated from the plasma concentration versus time curve using eqn (2.5).
84 Chapter 2
Clearance is most often calculated following intravenous administration
because the total dose of the drug is known to be delivered into the blood.
However, most drugs are delivered at sites remote from the blood, with the most
popular route being oral administration. As stated earlier, following oral admin-
istration the drug must overcome the barriers of dissolution, absorption, gut wall
metabolism and hepatic first-pass metabolism, dissolve in the gastrointestinal tract
lumen contents and cross the gastrointestinal tract enterocyte membrane.
Thus, at a number of points between the site of administration and the blood,
there is capacity for an oral drug to be lost. Consequently, the oral clearance
will often be higher than the intravenous clearance. The unifying factor is the
fraction bioavailable (F). Clearance is calculated using eqn (2.8).

DoseðoralÞ
ClearanceðoralÞ ¼ ð2:8Þ
AUCðoralÞ  F

Drugs can be cleared by a variety of organs in the human body. However, the
major clearing organs are the liver and the kidney.

2.4.2 Clearance by the Liver

The liver is at the centre of drug clearance by virtue of its drug metabolising
enzyme expression and its blood perfusion. The liver is perfused by blood from
two sources, the hepatic vein and the hepatic portal vein. Compounds being
presented to the liver through the hepatic vein are available for hepatic
extraction, which contributes to the systemic clearance of that drug. The
hepatic portal vein carries blood from the gut to the liver. Consequently, any
drug that has been absorbed from the gastrointestinal tract is available for
extraction by the liver, before it reaches the systemic circulation. The fraction of
a drug removed from the portal vein by the liver prior to accessing the systemic
blood is termed the hepatic first-pass extraction (see Section 2.2.1.5). This is a
significant factor contributing to the oral bioavailability of a drug. When
hepatic first-pass extraction is high, oral bioavailability will be low.
The hepatocyte is the liver cell responsible for the clearance of drugs
(Figure 2.12). Hepatocytes contain an arsenal of drug metabolising enzymes
and drug transport proteins that are capable of irreversibly removing com-
pounds from the circulating blood.

2.4.3 Metabolism
2.4.3.1 Cytochrome P450
The family of enzymes that metabolise the majority of drugs are the cytochrome
P450s (CYPs). These are a super-family of haem-containing enzymes that use
NADPH to catalyse the single electron oxidation of substrates. The overall
family consists of several hundred isoforms, with a variety of endogenous roles.
ADMET for the Medicinal Chemist 85

Efflux Hepatic uptake

Transporters Transporters
(eg MDR1, MRP) (e.g. OATPs, OCTs) Endplasmic
Reticulum
Sinusoidal (CYPs, FMOs
Membrane UDP glucuronyl
tranferases)

Mitochondria Sinusoidal
(MAO) Membrane Cytoplasm
Canalicular (Aldehyde
Bile
Membrane Oxidase,
Canaliculus
Sulfotransferases,
Amidases,
Esterases)

Figure 2.12 Representation of two hepatocytes showing the bile canaliculus.

However, there are only a small number of major human drug metabolising
CYPs, with the most important being CYP3A4, CYP2C9 and CYP2D6.
The CYP enzymes are expressed in many tissues, but they are found in
highest concentrations in the hepatocyte. They are membrane-bound enzymes
found within the cell on the endoplasmic reticulum. Human liver microsomes
are preparations of endoplasmic reticulum made by differential centrifugation
of human liver homogenates. On preparation, these intracellular membranes
form into spheres that express CYPs on their inner and outer surfaces.
The substrate structure metabolism relationships for these major drug meta-
bolising CYPs have been extensively reviewed.49,66 CYP2D6 will tend to meta-
bolise compounds with a metabolically vulnerable site a certain distance away
from a basic centre. Since CYP2D6 is polymorphically expressed across the
human population (6–8% of the Caucasian population do not express CYP2D6),
compounds that are predominantly metabolised by this isoform will exhibit
significant variability in pharmacokinetic profile. For this reason, CYP2D6
substrates tend to be avoided in compound selection. CYP2C9 will tend to
metabolise acids or compounds with a significant degree of hydrogen bonding.
CYP3A4 tends to be more promiscuous, driven by a large active site, and will
metabolise relatively large molecules with no major preference for ionisation.
These CYP isoforms share a common requirement for lipophilic compounds.
In general, CYP metabolism is positively correlated with log D.67
Therefore, across a series of compounds, the most lipophilic analogue will tend
to be more rapidly metabolised than its analogues with lower log D. As oxidative
enzymes, CYPs will tend to oxidise lipophilic compounds at sites of significant
electron density. This points the way to medicinal chemistry strategies to address
CYP metabolism. Clearly, reducing log D is likely to lead to a reduction in
86 Chapter 2
intrinsic clearance observed in human liver microsomes. However, as previously
established, lipophilicity tends to be correlated positively with potency against
the pharmacological target. Therefore, a balance needs to be struck. Perhaps a
more informed strategy to reduce CYP metabolism would be to identify the sites
of metabolic attack and reduce their susceptibility to metabolism by blocking
with metabolically inert groups or by reducing electron density at that position.
In this way, lipophilicity to drive potency can be maintained whilst reducing
metabolic clearance due to CYP metabolism.
It is easily overlooked in compound optimisation (based on human liver
microsomal intrinsic clearance) that the liver contains a vast armoury of other
drug metabolising enzymes. Thus, it is not unusual to find that a successful
campaign to lower the CYP mediated clearance within a series has not reduced
the in vivo clearance as much as expected due to metabolism by other hepatic
enzymes. This has been exemplified by Williams et al.68 A review of the top 200
prescribed drugs suggests that metabolism is responsible for the clearance of
approximately two-thirds. Of these, approximately two-thirds are pre-
dominantly CYP metabolised, suggesting that the emphasis on CYP-mediated
metabolism is well placed. However, there is approximately a further third that
are substrates for metabolic clearance mediated by enzymes other than CYPs.
Of these, the most prevalent are UDP-glucuronosyltranferases and esterases,
representing approximately 8% and 5% of the metabolised drugs, respectively.
More minor contributions come from flavin monooxygenases (FMOs),
monoamine oxidases (MAOs) and N-acetyltransferases.

2.4.3.2 Flavin Monooxygenase

Flavin monooxygenases (FMOs) are a family of membrane-bound endo-
plasmic reticulum enzymes consisting of six isoforms. FMO3 is the major
human adult FMO expressed in the liver. Consequently, it is present in human
liver microsomes and hepatocytes.
FMO3 utilises NADPH to produce the two electron oxidation of substrates
at heteroatoms such as nitrogen and sulphur69,70 to produce N- and S-oxides.
One of the endogenous substrates of human FMO3 is trimethylamine, which is
cleared by N-oxidation. Humans lacking FMO3 cannot metabolise trimethyla-
mine and exhibit the characteristic ‘fish odour’ syndrome, trimethylaminuria.
Drugs that are in part cleared by FMO3 are nicotine (Figure 2.13), clozapine
(N-oxidation), methimazole and cimetidine (S-oxidation).

+
FMO3
N O-
N
N
N
Nicotine Nicotine N-oxide

Figure 2.13 Metabolism of nicotine by FMO3.

ADMET for the Medicinal Chemist 87
2.4.3.3 Monoamine Oxidases
Monoamine oxidases (MAOs) use FAD to catalyse the oxidative deamination
of primary, secondary and tertiary amines. The reaction proceeds through
oxidation of the amine to form an imine, which is then hydrolysed to form the
aldehyde. The aldehyde is then further oxidised chemically or enzymatically to
produce the carboxylic acid.71–73
Two forms of MAO (MAO A and B) are involved in the metabolism of
endogenous catecholamine neurotransmitters such as dopamine and trypta-
mine. These enzymes are predominantly found in the outer mitochondrial
membrane in a variety of tissues, including the liver. MAO A has been shown to
be involved in the clearance of sumatriptan74 producing the only human phase I
metabolite, indole acetic acid (Figure 2.14).

2.4.3.4 Aldehyde Oxidase

Aldehyde oxidase is a molybdenum-containing liver cytosolic enzyme that can
be involved in both oxidative and reductive reactions. It is responsible for the
oxidation of a wide range of aldehydes and a number of nitrogen-containing
heterocycles. For example, aldehyde oxidase is the major enzyme involved in the
oxidation of the phthalazine containing compound, carbazeran (Figure 2.15).75

H O COOH
N H O
S MAO N
S
O N O
H N
H
Sumatriptan Sumatriptan
indole acetic acid

Figure 2.14 Metabolism of sumatriptan by MAO.

OH
O O
N N
N Aldehyde Oxidase N
O O
N N

H H
O N O N

O O

Figure 2.15 Metabolism of carbazeran by aldehyde oxidase.

88 Chapter 2
Carbazeran exhibits a species difference in pharmacokinetics that is driven by
differences in aldehyde oxidase activity.76 In the dog, carbazeran exhibits
greater than 60% oral bioavailability, whereas in humans oral bioavailability is
less than 5%.75 This species difference is driven by differences in hepatic first-
pass extraction mediated by aldehyde oxidase. Rat and dog livers show negli-
gible aldehyde oxidase activity, whereas human liver cytosol exhibits significant
activity. Guinea pig liver cytosol is the most appropriate pre-clinical model for
human aldehyde oxidase.
In addition, aldehyde oxidase is involved in the reductive metabolism of
ziprasidone,77 where it accounts for approximately two-thirds of the observed
metabolism (Figure 2.16).

2.4.3.5 Hydrolases
Hydrolases encompass a wide range of enzymes that use water to break-
down their substrates. In general, they catalyse the reaction exemplified by
Figure 2.17. Hydrolases comprise a huge family of enzymes (including estera-
ses, amidases and peptidases) that have been thoroughly reviewed in terms of
classification, mechanism of action and structure–activity relationship (SAR)
by Testa and Krämer.78
Probably the most important group of hydrolases from a drug metabolism
and prodrug point of view are the esterases including the cholinesterases

O O
HN HN

Cl Cl
Aldehyde Oxidase
Ziprasidone
N N

N N

N N
S
SH

Figure 2.16 Metabolism of ziprasidone by aldehyde oxidase.

O O
R H H R H
R X + O R OH + X

Figure 2.17 General reaction scheme for hydrolase enzymes.

ADMET for the Medicinal Chemist 89

OH O
O H H
+ O +
OH OH
OH

O O

Aspirin Salicylic Acid

Figure 2.18 Metabolism of aspirin by carboxylesterase.

(predominantly in plasma), arylesterases (predominantly in blood cells and

plasma) and carboxylesterases (predominantly in gut and liver).79–81
There are two major isoforms of human carboxylesterase. Human carbox-
ylesterase 1 (hCE-1) is highly expressed in the liver, whereas hCE-2 is most
highly expressed in the intestine. Both isoforms are not highly expressed in
human blood, whereas in preclinical species there is significant expression in the
blood. In addition, carboxylesterase-2 is not expressed in dog intestine. This
points to some of the difficulties encountered in preclinical species by programs
that target ester prodrugs.6
The carboxylesterases are serine hydrolases that catalyse the base-mediated
hydrolysis of esters using a three amino acid triad of serine, histidine and
glutamic acid. They are involved in the conversion of aspirin to acetyl salicylic
acid (Figure 2.18).

2.4.3.6 UDP-glucuronosyltranferases (UGTs)

The glucuronidation of drugs is an important clearance pathway for a number
of xenobiotics.68 The conjugation of a drug molecule with glucuronic acid
(most often at hydroxyl, carboxylic acid or nitrogen containing functions)
reduces lipophilicity and increases water solubility, thus rendering the molecule
more readily excretable in urine and bile.
The UGT enzymes catalysing glucuronidation are part of a larger family of
UDP-glycosyltranferase enzymes that conjugate lipophilic compounds with
glycosyl groups such as glucose, glucuronic acid and galactose.82 The UGT 1
and UGT 2 sub-families use UDP-glucuronic acid as the activated sugar donor,
whereas the UGT 8 sub-family uses UDP-galactose. Within the UGT 1 and 2
sub-families, there are a total of 18 enzymes, but only a small number catalyse
the majority of xenobiotic glucuronidation.68,83 For example, UGT2B7 meta-
bolises 35% of the glucuronidation substrates within the top 200 drugs, with
UGT1A4 responsible for 20% and UGT1A1 15%.
UDP-glucuronosyltransferase enzymes are expressed in many tissues within
the body, with major expression sites in the liver, kidney and intestine. The
UGT1 and UGT2 sub-families are expressed on the endoplasmic reticulum. It
is possible to demonstrate glucuronidation in human liver microsomes,
90 Chapter 2
provided UDG-glucuronic acid is added as a cofactor. Hepatocytes are also
potential in vitro assays for glucuronidation.

2.4.3.7 Sulfotransferases (SULTs)

In terms of drug metabolism, sulfotransferases use 3 0 -phosphoadenosine 5 0 -
phosphosulfate (PAPS) to transfer a sulfate group onto phenol/hydroxyl and
amine functions of their substrates.
The human sulfotransferase family of enzymes have a variety of roles from
xenobiotic metabolism to regulation of endogenous processes. There are two
main classes of SULTs: membrane-bound enzymes that in general are involved
in endogenous processes and the cytosolic SULTs that are responsible for drug
metabolism.84 The cytosolic SULTs have been extensively reviewed.85
There are three sub-families of cytosolic SULTs representing approximately
13 enzymes. The SULT1 sub-family is responsible for the majority of drug
sulfation, with SULT1A1 and SULT1A3 being the most active. SULT1A1
exhibits highest expression in human liver, whereas SULT1A3 is hardly
expressed in liver and shows significant expression in the intestine. Thus, these
isoforms are well placed to exert a first-pass sulfation on their substrates fol-
lowing oral administration. Examples of drugs that undergo significant sulfa-
tion are salbutamol, paracetamol and diflunisal.

2.4.3.8 N-Acetyltransferases (NATs)

N-Acetyltransferases are cytosolic enzymes that use acetyl-coenzyme A to
acetylate their substrates at amine functions. Arylamine N-acetyltransferases
have been extensively reviewed.86
There are two human NATs. NAT1 exhibits a wide range of tissue expres-
sion, including liver and lung. It tends to be mainly involved in endogenous
function, although it does have some drug substrates. NAT2 is the major drug
metabolising NAT, with expression in liver and intestine. NAT2 is poly-
morphically expressed with 50–55% of Caucasians and 10% of Asians classi-
fied as slow acetylators. This is important as the anti-tuberculosis drug
isoniazid is metabolised by NAT2 and shows higher circulating concentrations
in slow acetylators that makes them more prone to the peripheral neuropathy
side effect.
Recently, Rawal et al.87 documented their efforts to address NAT2 meta-
bolism in a research project. They found that UK-469 413 was metabolised by
NAT2, unusually at the piperazine nitrogen (Figure 2.19). As would be
expected for a NAT2 substrate, this compound was not turned over in rat and
human liver microsomes, but was cleared at greater than liver blood flow in the
rat. Further investigations showed that it was N-acetylated in rat and human
liver cytosol and hepatocytes. Incubations with singly expressed enzymes
showed that UK-469 413 was a specific substrate for NAT2. Analogues with
methyl groups adjacent to the piperazine nitrogen or bridging of the piperazine
ADMET for the Medicinal Chemist 91

N N

HN O HN
N NAT2
N + CoA SH
N + CoA S N
N N N N
NH N

O
UK-469,413 N-acetylated UK-469,413

Figure 2.19 Metabolism of UK-469 413 by N-acetyltransferase.

ring tended to block N-acetylation, suggesting that steric bulk is an appropriate

strategy to reduce NAT2 clearance.

2.4.4 Biliary Elimination

In addition to the metabolic capacity of the liver, hepatic clearance can occur
via biliary elimination. This is a three-stage process that largely involves active
transport of compounds from the blood into the bile. A review of the large
variety of drug-transporting proteins is beyond the scope of this chapter but
this been comprehensively reviewed elsewhere.20,88
The first stage in biliary clearance is the passage of the compound across the
sinusoidal membrane of the hepatocyte. For the majority of drugs, this process
occurs by passive diffusion, although these compounds tend to be substrates for
metabolic clearance that occurs within the hepatocyte. The products of this
metabolism can be excreted from the body in the bile. However, compounds
with inappropriate physicochemistry (high molecular weight, significant charge
and/or polarity) for rapid passage across the sinusoidal membrane can be
recognised in the blood and transported into the hepatocyte by drug transport
proteins (e.g. OATP1). By virtue of their physicochemistry, these types of
molecules are less prone to metabolic clearance, and so the second step in
hepatobiliary elimination is transfer of the molecule to the biliary canalicular
membrane (by diffusion or intracellular transfer proteins). Once at the cana-
licular membrane, these compounds can be recognised by efflux proteins (e.g.
P-gp and MRP2) and actively transported into the bile, to be excreted in the
faeces.
The study of biliary clearance in humans is difficult because the bile duct
empties into the gall bladder, which subsequently delivers bile into the faeces.
In the absence of bile duct surgery, it is impossible to determine the amount of a
compound excreted in the bile. Rather this information needs to be inferred
from the observed biliary clearance in animals and the excretion of unchanged
drug in the faeces. One such example is susalimod, which is completely elimi-
nated unchanged in bile in animals and shows a high excretion of unchanged
compound in the faeces of humans following intravenous administration.89
92 Chapter 2
2.4.5 Clearance by the Kidney
The main role of the kidney is to regulate water loss from the body and to
produce urine.
Blood flow to the Bowman’s capsule is filtered through the glomerulus into
the proximal convoluted tubule. Water and small molecules can pass freely
through the glomerulus whereas cells and large molecules such as proteins are
excluded. Thus, the fluid in the proximal convoluted tubule is simply a filtrate
of plasma. The blood flow to the glomerulus is approximately 1–2 ml min1
kg1, suggesting that around 30 litres of blood is filtered through the human
kidneys per day. The purpose of the rest of the nephron (the proximal and
distal convoluted tubules and the Loop of Henle) is to concentrate this fluid to
produce urine, ensuring water homeostasis and balance of salts. Thus, these
areas have a significant blood supply to enable water reabsorption by an
osmotic process.
The kidney is predominantly an organ of excretion, rather than metabolism.
Clearance of compounds by the kidney can be affected in three ways: glo-
merular filtration, renal reabsorption and renal secretion.90

2.4.5.1 Glomerular Filtration

In humans, blood flows to the glomerulus at a rate of 1 to 2 ml min1 kg1. This is
called the glomerular filtration rate (GFR). Any drug that is not bound to plasma
proteins in the blood will be filtered with the aqueous component. Once in the
filtrate, the fate of a compound will depend upon its ability to cross the mem-
branes of cells lining the nephron. Compounds that are significantly hydrophilic
will not be able to be reabsorbed passively across the cell membranes and
therefore will remain to be excreted in the urine. For such compounds, the renal
clearance (CLr) of the unbound drug will be the GFR as indicated by eqn (2.9):

Clr ¼ fu  GFR ð2:9Þ

Since these compounds tend to be largely unbound in plasma and not

metabolised to any great extent (because they do not cross membranes), they
are likely to exhibit a plasma clearance at GFR (i.e. 1–2 ml min1 kg1).
In common with many of the hydrophilic beta blockers, nadolol is
100% excreted unchanged in human urine with a plasma clearance of
1.6 ml min1 kg1 equivalent to GFR.91

2.4.5.2 Renal Reabsorption

The unbound fraction of lipophilic drugs will also be filtered through the glo-
merulus. However, the filtered component is likely to be reabsorbed across the
membranes of the cells lining the nephron. For these compounds, depending on
the lipophilicity, very little compound will remain in the urine and renal
ADMET for the Medicinal Chemist 93
clearance will be negligible (i.e. much less than GFR). However, these types of
compounds are more likely to exhibit clearance by (hepatic) metabolism.
A good example of a compound undergoing significant renal reabsorption is
fluconazole,92,93 which is 80% excreted unchanged in urine. The clearance of
fluconazole in humans is 0.23 ml min1 kg1, which is approximately a fifth of
GFR, suggesting that up to 80% is reabsorbed following glomerular filtration.

2.4.5.3 Renal Secretion

Some drugs are excreted in the urine at renal clearances that are greater than
GFR. Such compounds are substrates for transporter proteins that are
expressed in the cells lining the nephron (e.g. P-gp). The drugs can be actively
removed from the blood and secreted into the filtrate, and subsequently excreted
into the urine. For example, an intravenous dose of ranitidine is approximately
70% excreted unchanged in human urine. The renal clearance of ranitidine is
approximately 7 ml min1 kg1, which is significantly higher than GFR (i.e.
1.5 ml min1 kg1). The difference between the renal clearance and GFR is the
renal secretion clearance of the compound (i.e. 5.5 ml min1 kg1).94,95
Overall, renal clearance of xenobiotics can be both passive by filtration and
active via transport proteins. Renal clearance is inversely correlated with
lipophilicity, positively correlated with polarity and related to ionisation
state.96 Generally, compounds with log D values o0 will tend to be passively
renally excreted at GFR. However, as lipophilicity increases in a series, passive
renal excretion is reduced to significantly lower than GFR due to passive
reabsorption across the cells lining the nephron. In addition, drugs can be
actively reabsorbed from or secreted into the urine if they are substrates for
transporter proteins. In the case of active renal secretion, the renal clearance
values can be significantly in excess of GFR.

2.4.6 Clearance Summary

Throughout compound optimisation strategies, medicinal chemists most often
attempt to balance potency improvements with strategies to reduce clearance.
This is not surprising since clearance is a key parameter in modulating oral
bioavailability and elimination half-life and is a major determinant of daily dose.
All molecules within the body are subject to clearance, such that the human
body renews itself on a regular basis. The same is true of drug molecules. Indeed, a
compound that is not cleared would only require a single dose to be effective for a
lifetime. Therefore, it is important that drug molecules exhibit a degree of clear-
ance and it is only the rate of clearance that medicinal chemists seek to modulate.
The clearance of a potential drug molecule is dependant upon its physi-
cochemistry and its structure. Drugs can be cleared in many ways from
metabolism to excretion of unchanged drug in urine and faeces.
In general, hits emerging from high throughput screening (HTS) tend to exhibit
a significant degree of lipophilicity, as specific interactions with the target are not
94 Chapter 2
optimised and potency is driven by lipophilic interactions. Consequently, it is not
unusual for these types of molecules to be high affinity substrates for metabolising
enzymes such as CYPs. Medicinal chemistry strategies have emerged that allow
reduction in CYP-mediated metabolic clearance, whilst maintaining and
improving potency against the target. On many occasions these strategies have led
to clearance by other enzymes, such as FMO and UDP-glucuronyl transferases,
albeit at lower rates. In addition, metabolically inert molecules can also be sub-
strates for clearance by drug transporters in the liver and kidney.
The challenge of maintaining/improving potency whilst lowering clearance
will continue to be a major one for medicinal chemists. Thus, it is important that
chemists responsible for compound optimisation strategies continue to under-
stand the range of clearance routes arrayed against potential drug molecules.

2.5 Toxicology related to ADME

Metabolism is a key part of the detoxifying strategy employed by the body to
deal with xenobiotics such as drug molecules. However, there is growing evi-
dence that, in some cases, these strategies may be counterproductive and that
metabolism is capable of acting as a toxification process.
Some molecules undergo metabolism to yield chemically reactive products or
intermediates. Such molecules are at an increased risk of covalent binding to
proteins within the body97 and are also at an increased risk of rare but poten-
tially serious idiosyncratic adverse drug reactions (IADRs).98–100 These drug
reactions are frequently not picked up in pre-clinical toxicology and may not be
revealed until after launch when a drug is used by large numbers of people. In
addition, many of the structural fragments known to be activated by metabo-
lism to chemically reactive groups also cause signals in genetic toxicity testing
and adducts with DNA itself have been reported for some of these fragments.
The ‘reactive metabolite’ hypothesis states that the covalent binding of meta-
bolic products to biomolecules such as DNA and proteins is a necessary step for
these molecules to show genetic toxicity101 or other toxicological outcome.102
Furthermore, the drug–protein adducts may act as a ‘hapten’ triggering the
immune reaction which is a feature of many, but not all IADRs.103
In Section 2.1 we saw the key role that physicochemical properties play in the
metabolic fate of a drug. In the case of toxicology liabilities related to ADME,
there is a closer link with functional groups and structural motifs with a specific
chemical reactivity than with physicochemistry per se. However, the physico-
chemical properties of a drug will govern the rates of the biochemical reactions
which result in toxic products and the distribution of these metabolites.

References
1. E. H. Kerns and L. Di, Drug Discov. Today, 2003, 8, 316.
2. C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeney, Adv. Drug
Deliv. Rev, 1997, 23, 3.
ADMET for the Medicinal Chemist 95
3. K. Beaumont, E. Schmid and D. A. Smith, Bioorg. Med. Chem. Lett,
2005, 15, 3658.
4. D. A. Smith, Curr. Opin. Drug Discov. Dev, 2007, 10, 550.
5. G. C. Williams and P. J. Sinko, Adv. Drug Deliv. Rev, 1999, 39, 211.
6. K. Beaumont, R. Webster, I. Gardner and K. Dack, Curr. Drug Metab,
2003, 4, 461.
7. A. L. Hopkins, C. R. Groom and A. Alex, Drug Discov. Today, 2004, 9,
430.
8. P. D. Leeson and B. Springthorpe, Nat. Rev. Drug Discov, 2007, 6, 881.
9. F. M. Ferrante, J. Paggioli, S. Cherukuri and G. R. Arthur, Anesth.
Analg. (Baltimore), 1996, 82, 91.
10. B. N. Singh, Clin. Pharmacokinet, 1999, 37, 213.
11. Y.-L. He, S. Murby, G. Warhurst, L. Gifford, D. Walker, J. Ayrton,
R. Eastmond and M. Rowland, J. Pharm. Sci., 1998, 87, 626.
12. P. R. Reeves, D. J. Barnfield, S. Longshaw, D. A. D. McIntosh and M. J.
Winrow, Xenobiotica, 1978, 8, 305.
13. P. R. Reeves, J. McAinsh, D. A. D. McIntosh and M. J. Winrow,
Xenobiotica, 1978, 8, 313.
14. M. Brandsch, I. Knuetter and E. Bosse-Doenecke, J. Pharm. Pharmacol,
2008, 60, 543.
15. C. Hilgendorf, G. Ahlin, A. Seithel, P. Artursson, A.-L. Ungell and
J. Karlsson, Drug Metab. Dispos., 2007, 35, 1333.
16. B. Bretschneider, M. Brandsch and R. Neubert, Pharm. Res., 1999, 16, 55.
17. M. E. Ganapathy, W. Huang, H. Wang, V. Ganapathy and F. H.
Leibach, Biochem. Biophys. Res. Commun, 1998, 246, 470.
18. S. Weller, M. R. Blum, M. Doucette, T. Burnette, D. M. Cederberg, P. D.
Miranda and M. L. Smiley, Clin. Pharmacol. Ther, 1993, 54, 595.
19. F. Thiebaut, T. Tsuruo, H. Hamada, M. M. Gottesman, I. Pastan and
M. C. Willingham, Proc. Natl. Acad. Sci. U.S.A, 1987, 84, 7735.
20. A. Ayrton and P. Morgan, Xenobiotica, 2008, 38, 676.
21. K. Beaumont, A. Harper, D. A. Smith and S. Abel, Xenobiotica, 2000, 30,
627.
22. K. Beaumont, A. Harper, D. A. Smith and J. Bennett, Eur. J. Pharm. Sci,
2000, 12, 41.
23. S. Abel, K. C. Beaumont, C. L. Crespi, M. D. Eve, L. Fox, R. Hyland,
B. C. Jones, G. J. Muirhead, D. A. Smith, R. F. Venn and D. K. Walker,
Xenobiotica, 2001, 31, 665.
24. K. Westphal, A. Weinbrenner, T. Giessmann, M. Stuhr, G. Franke,
M. Zschiesche, R. Oertel, B. Terhaag, H. K. Kroemer and W. Siegmund,
Clin. Pharmacol. Ther, 2000, 68, 6.
25. T. Prueksaritanont, L. M. Gorham, J. H. Hochman, L. O. Tran and K. P.
Vyas, Drug Metab. Dispos, 1996, 24, 634.
26. J. H. Lin, M. Chiba and T. A. Baillie, Pharmacol. Rev., 1999, 51, 135.
27. K. Beaumont, in Methods and Principles in Medicinal Chemistry, ed.
H. van de Waterbeemd, H. Lennernäs and P. Artursson, 2003, vol. 18
(Drug Bioavailability), ch. 13, pp. 311–328.
96 Chapter 2
28. U. Fagerholm, J. Pharm. Pharmacol, 2007, 59, 1335.
29. M. F. Paine, H. L. Hart, S. S. Ludington, R. L. Haining, A. E. Rettie and
D. C. Zeldin, Drug Metab. Dispos, 2006, 34, 880.
30. J. Yang, T. Tucker Geoffrey and A. Rostami-Hodjegan, Clin Pharmacol
Ther, 2004, 76, 391.
31. M. F. Paine, M. Khalighi, E. M. Fisher, D. D. Shen, K. L. Kunze, C. L.
Marsh, J. D. Perkins and K. E. Thummel, J. Pharmacol. Exp. Ther, 1997,
283, 1552.
32. K. E. Thummel, S. D. O., M. F. Paine, D. D. Shen, K. L. Kunze and J. D.
Perkins, Clin. Pharmacol. Ther, 1996, 59, 491.
33. M. F. Paine, D. D. Shen, K. L. Kunze, J. D. Perkins, C. L. Marsh, J. P.
McVicar, D. M. Barr, B. S. Gillies and K. E. Thummel, Clin. Pharmacol.
Ther, 1996, 60, 14.
34. H. Kublik and M. T. Vidgren, Adv. Drug Deliv. Rev., 1998, 29, 157.
35. L. Illum, Trends Biotechnol., 1991, 9, 284.
36. M. Cazzola, R. Testi and M. G. Matera, Clin. Pharmacokinet, 2002, 41,
19.
37. J. S. Patton and P. R. Byron, Nat. Rev. Drug Discov, 2007, 6, 67.
38. P. W. Armstrong, J. A. Armstrong and G. S. Marks, Circulation, 1979,
59, 585.
39. C. T. Lombroso, Epilepsia, 1989, 30(Suppl 2), S11.
40. W. Jeal and P. Benfield, Drugs, 1997, 53, 109.
41. I. Power, Br. J. Anaesth, 2007, 98, 4.
42. R. J. Wills, Clin Pharmacokinet, 1990, 19, 390.
43. G. L. Trainor, Expert Opin. Drug Discov, 2007, 2, 51.
44. W. E. Lindup, Prog. Drug Metab, 1987, 10, 141.
45. P. T. Schoenemann, D. W. Yesair, J. J. Coffey and F. J. Bullock, Ann. N.
Y. Acad. Sci., 1973, 226, 162.
46. Z. Y. Wu, S. E. Cross and M. S. Roberts, J. Pharm. Sci., 1995, 84, 1020.
47. J. Krieglstein, Arzneim.-Forsch., 1973, 23, 1527.
48. P. H. Hinderling, Pharmacol. Rev., 1997, 49, 279.
49. D. A. Smith, B. C. Jones and D. K. Walker, Med. Res. Rev, 1996, 16, 243.
50. F. Lombardo, R. S. Obach, F. M. DiCapua, G. A. Bakken, J. Lu, D. M.
Potter, F. Gao, M. D. Miller and Y. Zhang, J. Med. Chem, 2006, 49, 2262.
51. F. Lombardo, R. S. Obach, M. Y. Shalaeva and F. Gao, J. Med. Chem,
2002, 45, 2867.
52. A. Petrauskas, P. Japertas, R. Didziapetris, K. Lanevskij and D. Bondarev,
Abstracts of Papers, 231st ACS National Meeting, Atlanta, GA, 26–30
March 2006, 2006, MEDI.
53. D. Alker, R. A. Burges, S. F. Campbell, A. J. Carter, P. E. Cross, D. G.
Gardiner, M. J. Humphrey and D. A. Stopher, J. Chem. Soc, Perkin
Trans. 2, 1992, 7, 1137.
54. D. A. Stopher, A. P. Beresford, P. V. Macrae and M. J. Humphrey,
J. Cardiovasc. Pharmacol., 1988, 12, S55.
55. M. S. Whitman and A. R. Tunkel, Infect. Control Hosp. Epidemiol, 1992,
13, 357.
ADMET for the Medicinal Chemist 97
56. P. Poulin and F. -P. Theil, J. Pharm. Sci., 2002, 91, 129.
57. G. E. Blakey, I. A. Nestorov, P. A. Arundel, L. J. Aarons and
M. Rowland, J. Pharmacokinet. Biopharm., 1997, 25, 277.
58. S. S. De Buck, V. K. Sinha, L. A. Fenu, M. J. Nijsen, C. E. Mackie and
R. A. H. J. Gilissen, Drug Metab. Dispos, 2007, 35, 1766.
59. J. P. Tillement, S. Urien, P. Chaumet-Riffaud, P. Riant, F. Bree, D. Morin,
E. Albengres and J. Barre, Fundam. Clin. Pharmacol, 1988, 2, 223.
60. L. A. Peletier, N. Benson and P. H. van der Graaf, J. Theor. Biol., 2009,
256, 253.
61. T. Obradovic, G. G. Dobson, T. Shingaki, T. Kungu and I. J. Hidalgo,
Pharm. Res, 2007, 24, 318.
62. P. H. Elsinga, N. H. Hendrikse, J. Bart, W. Vaalburg and A. Van Waarde,
Curr. Pharm. Des., 2004, 10, 1493.
63. N. Jung, C. Lehmann, A. Rubbert, M. Knispel, P. Hartmann, J. van
Lunzen, H.-J. Stellbrink, G. Faetkenheuer and D. Taubert, Drug Metab.
Dispos., 2008, 36, 1616.
64. M. Ahmed, M. A. Briggs, S. M. Bromidge, T. Buck, L. Campbell, N. J.
Deeks, A. Garner, L. Gordon, D. W. Hamprecht, V. Holland, C. N.
Johnson, A. D. Medhurst, D. J. Mitchell, S. F. Moss, J. Powles, J. T. Seal,
T. O. Stean, G. Stemp, M. Thompson, B. Trail, N. Upton, K. Winborn
and D. R. Witty, Bioorg. Med. Chem. Lett., 2005, 15, 4867.
65. J. C. Kalvass and T. S. Maurer, Biopharm. Drug Dispos., 2002, 23, 327.
66. D. A. Smith, M. J. Ackland and B. C. Jones, Drug Discov. Today, 1997, 2,
479.
67. H. van de Waterbeemd, D. A. Smith and B. C. Jones, J. Comput. Aided
Mol. Des., 2001, 15, 273.
68. J. A. Williams, R. Hyland, B. C. Jones, D. A. Smith, S. Hurst, T. C.
Goosen, V. Peterkin, J. R. Koup and S. E. Ball, Drug Metab. Disp., 2004,
32, 1201.
69. J. R. Cashman, Curr. Drug Metab., 2000, 1, 181.
70. J. R. Cashman and J. Zhang, Annual Rev. Pharmacol. Toxicol., 2006, 46,
65.
71. P. L. Dostert, B. M. Strolin and K. F. Tipton, Med. Res. Rev., 1989, 9, 45.
72. D. E. Edmondson, A. Mattevi, C. Binda, M. Li and F. Hubálek, Curr.
Med. Chem., 2004, 11, 1983.
73. K. F. Tipton, S. Boyce, S. J.O., G. P. Davey and J. Healy, Curr. Med.
Chem., 2004, 11, 1965.
74. C. M. Dixon, G. R. Park and M. H. Tarbit, Biochem. Pharmacol., 1994,
47, 1253.
75. B. Kaye, J. L. Offerman and J. L. Reid, Xenobiotica, 1984, 14, 935.
76. C. Beedham, S. E. Bruce, D. J. Critchley, T. Y. Al and D. J. Rance, Eur. J.
Drug Metab. Pharmacokinet., 1987, 12, 307.
77. C. Beedham, J. J. Miceli and R. S. Obach, J. Clin. Psychopharmacol.,
2003, 23, 229.
78. B. Testa and S. D. Krämer, Chem. Biodivers., 2007, 4, 2031.
79. F. M. Williams, Clin. Pharmacokinet., 1985, 10, 392.
98 Chapter 2
80. M. R. Redinbo and P. M. Potter, Drug Discov. Today, 2005, 10, 313.
81. T. Imai, Drug Metab. Pharmacokinet., 2006, 21, 173.
82. P. I. Mackenzie, K. W. Bock, B. Burchell, C. Guillemette, S. i. Ikushiro,
T. Iyanagi, J. O. Miners, I. S. Owens and D. W. Nebert, Pharmacogenet.
Genomics, 2005, 15, 677.
83. B. Burchell, D. J. Lockley, A. Staines, Y. Uesawa and M. W. H.
Coughtrie, Methods Enzymol., 2005, 400, 46.
84. N. Gamage, A. Barnett, N. Hempel, R. G. Duggleby, K. F. Windmill,
J. L. Martin and M. E. McManus, Toxicolog. Sci., 2006, 90, 5.
85. G. M. Pacifici and M. W. H. Coughtrie, Human Cytosolic Sulfo-
transferases, CRC Press, Boca Raton, FL, 2005.
86. E. Sim, K. Walters and S. Boukouvala, Drug Metab. Rev., 2008, 40, 479.
87. J. Rawal, R. Jones, A. Payne and I. Gardner, Xenobiotica, 2008, 38, 1219.
88. A. Ayrton and P. Morgan, Xenobiotica, 2001, 31, 469.
89. I. Pahlman, M. Edholm, S. Kankaanranta and M. L. Odell, Pharm.
Pharmacol. Commun., 1998, 4, 493.
90. U. Fagerholm, J. Pharm. Pharmacol., 2007, 59, 1463.
91. J. G. Riddell, D. W. Harron and R. G. Shanks, Clin. Pharmacokinet.,
1987, 12, 305.
92. K. W. Brammer, A. J. Coakley, S. G. Jezequel and M. H. Tarbit, Drug
Metab. Dispos., 1991, 19, 764.
93. S. G. Jezequel, J. Pharm. Pharmacol., 1994, 46, 196.
94. S. M. Grant, H. D. Langtry and R. N. Brogden, Drugs, 1989, 37, 801.
95. C. J. Roberts, Clin. Pharmacokinet., 1984, 9, 211.
96. M. V. S. Varma, B. Feng, R. S. Obach, M. D. Troutman, J. Chupka,
R. H. Miller and A. El-Kaman, J. Med. Chem., 2009, 52, 4844.
97. H. Takakusa, H. Masumoto, H. Yukinaga, C. Makino, S. Nakayama,
O. Okazaki and K. Sudo, Drug Metab. Dispos., 2008, 36, 1770.
98. B. Seguin and J. Uetrecht, Curr. Opin. Allergy Clin. Immunol., 2003, 3,
235.
99. N. Kaplowitz, Nat. Rev. Drug Discov., 2005, 4, 489.
100. J. L. Walgren, M. D. Mitchell and D. C. Thompson, Crit. Rev. Toxicol.,
2005, 35, 325.
101. E. C. Miller and J. A. Miller, Pharmacol. Rev., 1966, 18, 805.
102. J. R. Mitchell, D. J. Jollow, W. Z. Potter, D. C. Davis, J. R. Gillette and
B. B. Brodie, J. Pharmacol. Exp. Ther., 1973, 187, 185.
103. J. Uetrecht, Chem. Res. Toxicol., 2008, 21, 84.
CHAPTER 3

Carboxylic Acids and their

Bioisosteres
AMIT S. KALGUTKARa AND J. SCOTT DANIELSb
a
Pharmacokinetics, Dynamics and Metabolism Department, Pfizer Global
Research and Development, Eastern Point Road, Groton, Connecticut,
06340, USA; b Pharmacokinetics, Dynamics and Metabolism Department,
Pfizer Global Research and Development, 700 Chesterfield Village Parkway,
Chesterfield, Missouri, 63107, USA

3.1 Introduction
Exposing the importance of the carboxylic acid functional group is best achieved
by examining the number of endogenous processes and individual molecules
which rely on the intrinsic chemical nature (e.g. pKa and hydrogen bonding
characteristics) of this functional group. From amino acid conjugation (peptide
synthesis - proteins) and post-translational protein acylation, to triglycerides,
bile acids, prostanoids, messenger molecules and hormone catabolites, it is evident
that the carboxylic acid represents a key functional group contributing to the
biochemistry critical to mammalian physiology. Not surprisingly then, there exists
an extensive number of drugs possessing the carboxylic acid functional group.
The compounds represent a heterogeneous group comprising, among others, non-
steroidal anti-inflammatory drugs (NSAIDs), b-lactam antibiotics, statins,
fibrates, and even food additives such as preservatives and flavouring agents; these
compounds range from hydrophilic to lipophilic organic compounds.
Over 450 drugs containing a free carboxylic acid group are marketed in var-
ious countries worldwide (see Table 3.1 for select examples). In addition to the

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

99
 100

Table 3.1 Examples of marketed drugs containing carboxylic acid functionality.

Amoxicillin Ubenimex Diflunisal Ofloxacin Ramatroban Repaglinide
Cefadroxil Ketorolac Alprostadil Thymopentin Tiagabine Ganirelix
Cetraxate Acitretin Methotrexate Captopril Cefpodoxime Tirofiban
Sofalcone Argipidine Baclofen Imipemide Raltitrexed Imidapril
Chenodeoxycholic Mebrofenin Methyldopa Cilastatino Acitazanolast Cefprozil
acid
Pentagastrin Quinapril Bumetanide Indobufen Marbofloxacin Hirulog-1
Sincalide Lisinopril Metirosine Lomefloxacin Ecabet Valsartan
Cefoxitin Zaltoprofen Aminocaproic acid Desglugastrin Olopatadine Olmesartan
Flucloxacillin Melphalan Difenoxin Enrofloxacin Pidotimod Pazufloxacin
Sulphasalazine Azelaic acid Dinoprostone Norfloxacin Tosufloxacin Finofibric acid
Amphotericin B Proamipide Lysine Aceclofenac g-Linolenic acid Eptifibatide
Ciprofloxacin Tranilast Leucovorin Rufloxacin Sparfloxacin (S)-(+)-Ketoprofen
Aspoxicillin Cefdinir Iloprost Adapalene d-Aminolevulinic
acid
Eflornithine Butoctamide Perindopril Seratrodast Prulifloxacin Zanamivir
Enoxacin Aztreonam Flunoxaprofen Felbinac Ulifloxacin Bepotastine
Pefloxacin Acemetacin Cilazapril Erdosteine Orbifloxacin
Etodolac Oxaprozin Fosinoprilat Spirapril Eicosapentaenoic Pregabalin
acid
Diacerhein Pirprofen Nadifloxacin Ursodiol Dexibuprofen Telmisartan
Moexipril Alminoprofen Cefaclor Droxidopa Sarpogrelate Ziconotide
Lonidamine Piretanide Cefprozil Tamibarotene Fosfosal Triproamylin
Fleroxacin Norfloxacin Docosahexaenoic Actarit Proxigermanium Prezatide
acid
Bucillamine Cefotiam Somatostatin-14 Mofezolac Betamipron Nesiritide
Amlexanox 13-cis-Retinoic acid Levocabastine Ademetionine Telmesteine Recombinant human
parathyroid hor-
Chapter 3

mone (1-84)
Delapril Gemfibrozil Proglumide Nateglinide Pazufloxacin Glycopin
Vigabatrin Clinofibrate Romurtide Ceftibuten Zofenoprilat Alvimopan
Cefminox Carprofen Sermorelin Limaprost Salmeterol Glycyrrhizinic acid
Gabapentin Cefmenoxime Cefalexin Gadopentetate Eprosartan Alatrofloxacin
Epalrestat Triflusal Pranoprofen Unoprostone Febuxostat Alitretinoin
Cefixime Levodopa Frusemide Tranexamic acid Angiotensin II Bexarotene
Alacepril Carbidopa Panipenem Temocapril Fudosteine Etacrynic acid
Enalaprilat Acetylcysteine Daptomycin Ibuprofen alpha-Lipoic acid Thymalfasin
Acrivastine Artesunic acid Tosufloxacin Flurbiprofen Trovafloxacin Enfuvirtide
Sarafloxacin Sulindac Trandolapril Levofloxacin Folinic acid Vedaprofen
Tropesin Tolmetin Acipimox Cetirizine Cefoselis Bendamustine
Ramipril Aspirin Ciprofibrate Gadoterate Sitafloxacin Gemifloxacin
Cefodizime Indomethacin Pseudomonic acid Loracarbef all-trans-Retinoic 4-Aminosalicylic
acid acid
Carboxylic Acids and their Bioisosteres

Benazepril Mefenamic acid Leuprorelin Gatifloxacin Gadobenate Fexofenadine

Moxifloxacin Cefamandole Probenecid Rosoxacin Ampicillin Simvastatin
Teriparatide Amineptine Valproic acid Lidofenin Atorvastatin Ertapenem
Garenoxacin 5-Aminosalicylic Deferasirox Naproxen Mitiglinide Cefmetazole
acid
Niacin Chlorambucil Fenbufen Diclofenac Pitavastatin
Ceftobiprole Lumiracoxib Salicylic acid Meclofenamate Rosuvastatin
Levocetirizine Lodoxamide Ambrisentan Sulbactam Doripenem
Exenatide Penicillin V Bezafibrate Vancomycin Montelukast
Tolfenamic acid Ceftazidime Folic acid Fluvastatin Ertapenem
Cefuroxime Pyrantel Cinoxacin Fenoprofen Penicillin G
101
102 Chapter 3
biochemical processes that underscore the importance of this functional group, it
is equally important to point to the physiochemical properties of the carboxylate
ion which make it an attractive moiety for installation into drug candidates. For
example, ionisation of a carboxylic acid, at physiological pH (7.4), improves its
ability to hydrogen bond with neighbouring water molecules, and thus may
improve its overall water solubility. Likewise, a carboxylic acid group can also
serve to promote intermolecular hydrogen bonding at a particular pH, thus
resulting in an alteration in a physiochemical property (Figure 3.1). Continuing
with the theme of the importance of carboxylic acids to mammalian physiology,
it is the intramolecular hydrogen bonding type which results in the three-
dimensional structure of proteins.1
There may be multiple factors which vary the relative acidity (i.e. pKa)
of a carboxylic acid group, including neighbouring group and long-range
inductive effects. The simple straight- and branched-chain carboxylic acids
shown in Table 3.2 depict the attenuation of pKa observed with the intro-
duction of electron-withdrawing neighbouring groups and long-range sub-
stitutions;2 it is noteworthy that the relative acidities of carboxylic acids are also
related to their ability to stabilise the developing charge upon ionisation. The
hydrophilic nature of a particular drug may govern absorption, distribution
and elimination, and thus may bear on its overall in vivo disposition in a
mammal.
While the ionisation state of the carboxylic acid functional group is an
important determinant to a drug’s physiochemical properties, it is not necessarily
the governing characteristic contributing to the overall ionisation state of a

H-bond Zwitterion

H
O NH2 O NH3

O OH O OH

Figure 3.1 Example of intramolecular hydrogen bonding in the amino acid tyrosine
yielding a zwitterion intermediate.

Table 3.2 Inductive and neighbouring

group effects on the pKa of
the carboxylic acid group.
Acid pKa

HCOOH 3.77
(CH3)3CCOOH 5.05
ClCH2COOH 2.86
HOCH2COOH 3.83
Carboxylic Acids and their Bioisosteres 103

O O O O
F F
HO O

N N N N
NH2 NH2

gastric (≤ 3.5) intestinal (≤ 7)

Figure 3.2 Ionisation state of ciprofloxacin within the gastrointestinal tract.

particular drug in a particular in vivo compartment. For example, the fluor-

oquinoline antibiotic ciprofloxacin contains several ionisable groups, including a
carboxylic acid group and multiple amine functionalities. Because of the pH
gradient unique to the gastrointestinal tract, it is the piperazine moiety (i.e. sec-
ondary amine) which governs the charge state within the upper gastrointestinal
tract (gastric region). A subsequent elevation in pH in the proximal intestine
results in a zwitterionic state for ciprofloxacin; importantly it is in the upper
gastrointestinal tract where efficient ciprofloxacin absorption occurs (Figure 3.2).3
Mammalian systems are quite capable of processing and metabolising a
range of endogenous carboxylic acid containing compounds (e.g. triglycerides,
bile acids, prostanoids and hormone catabolites). These systems include but are
not limited to glucuronidation, amino acid and thioester (acyl CoA) conjuga-
tion as well as active uptake and efflux transport. It is not surprising then that
the physiological mechanisms predisposed to process endogenous substrates
are quite suited to metabolise drugs bearing this functional group.

3.2 Carboxylic Acid Containing Non-steroidal

Anti-inflammatory Drugs (NSAIDs)
NSAIDs are compounds with analgesic, antipyretic and, in higher doses, anti-
inflammatory effects. The term ‘non-steroidal’ is used to distinguish these
compounds from steroids (e.g. dexamethasone), which have a similar mode of
anti-inflammatory action. There are 420 different NSAIDs currently in
worldwide use. All classical NSAIDs introduced in the 1960s and 1970s contain
either a free carboxylic acid group or the acidic enol–carboxamide moiety (e.g.
piroxicam). As shown in Figure 3.3, carboxylic acid based NSAIDs can be sub-
categorised into: (a) salicylic acid, (b) anthranilic acid, (c) acetic acid, and (d)
propionic acid derivatives.

3.2.1 Discovery of Aspirin

The history of NSAIDs dates back many thousands of years to the early uses of
plant preparations that contained salicylate.4 A feasible commercial synthesis
of salicylic acid from phenol and carbon dioxide was formulated by Kolbe in
 104

Salicylic acid derivatives Anthranilic acid derivatives Acetic acid derivatives Propionic acid derivatives
O
R O OH O
O O H
CH3
H3CO OH O CH3
N Cl OH
OH CH3 H OH
N HO
N
H 3C
O
CH3 O
Salicylic acid: R = H O Cl Cl

Aspirin: R = COCH3 Mefenamic acid

Indomethacin Diclofenac
Ibuprofen Naproxen
Enol-carboxamide derivatives O

OH O OH
OH O N F O CH3
H
CH3 N OH
N N CH3
H N S OH
N H
S CH3 N O O
S CH3 CH3 O
O O
O O S

Piroxicam Meloxicam Sulindac Etodolac Ketoprofen

Figure 3.3 Structures of classical NSAIDs.

Chapter 3
Carboxylic Acids and their Bioisosteres 105

O
OH OH O
H3C O O
CO2 OH (CH3CO)2O
OH
pressure

Phenol Salicylic acid Acetylsalicylic acid

(Aspirin)

Scheme 3.1 Synthesis of aspirin, the first carboxylic acid containing NSAID.

1874 (Scheme 3.1), which led to the introduction of sodium salicylate in the
treatment for chronic rheumatoid arthritis and gout.
The search for a superior salicylic acid derivative was initiated at Bayer in
1895. Chemist Felix Hoffman, to whom the task was presented, also had a
personal reason for this endeavour; his father had been taking salicylic acid for
many years to treat arthritis and encountered emesis as a major side effect.
Hoffman found a way of acetylating the phenol group of salicylic acid to form
acetylsalicylic acid (aspirin) (see Scheme 3.1).4 After initial laboratory tests,
Hoffman’s father was given the drug; it was pronounced effective and later
confirmed as such in ‘impartial’ clinical trials. The drug was introduced in 1899
with a report suggesting that aspirin was a convenient way of delivering sal-
icylic acid to the body.5 That aspirin is a mere prodrug for salicylic acid has
been debated since its discovery, but as discussed below, aspirin clearly has
potent actions of its own that are not shared by salicylic acid.
It was not until 1971 that Vane and co-workers proposed that the anti-
inflammatory and analgesic properties of NSAIDs are due to inhibition of
prostaglandin biosynthesis, which is catalysed by the enzyme prostaglandin
endoperoxide synthase or cyclooxygenase (COX).6–8 COX catalyses the for-
mation of prostaglandins and thromboxane from the fatty acid substrate ara-
chidonic acid (itself derived from the cellular phospholipid bilayer by the action
of phospholipase A2).9
COX activity originates from two distinct and independently regulated
enzymes, termed COX-1 and COX-2.10,11 COX-1 is the constitutive isoform and
is mainly responsible for the synthesis of cytoprotective prostaglandins in the
gastrointestinal tract. COX-2 is inducible and short-lived; its expression is sti-
mulated in response to pro-inflammatory mediators and the isozyme plays a
major role in prostaglandin biosynthesis in inflammatory cells (monocytes/
macrophages).12 Classical NSAIDs act as non-selective inhibitors of COX-1 and
COX-2 isozymes.13 Inhibition of COX-1 is thought to be responsible for the
gastrointestinal liabilities associated with most NSAIDs, while inhibition of the
inducible COX-2 isozyme is thought to be responsible for the anti-inflammatory
effects.14 The hypothesis led to substantial research efforts towards the discovery
of selective COX-2 inhibitors and has resulted in the introduction of celecoxib,
valdecoxib and rofecoxib into the market as the next generation of NSAIDs
with reduced gastrointestinal liabilities.15
106 Chapter 3

k1 k2
E + I (E.I) (E.I)*
k-1

Figure 3.4 Kinetics of COX inhibition by NSAIDs.

3.2.2 Mode of Inhibition of COX Activity by NSAIDs

The interaction of NSAIDs with COX follows a two-step kinetic sequence
(Figure 3.4) as originally proposed by Rome and Lands in the 1970s.16 The first
step involves the formation of a rapidly reversible (E  I) complex between COX
and NSAIDs, leading to competitive inhibition. The second step is the time-
dependent conversion of the initial (E  I) complex to one, [E  I]*, in which the
inhibitor is bound more tightly. Formation of the [E  I]* complex occurs in
seconds to minutes and is thought to reflect the induction of a subtle protein
conformational change.
Based on their mode of inhibition, NSAIDs can be sub-categorised into: (a)
competitive inhibitors, (b) slow, tight-binding inhibitors and (c) covalent
modifiers.16 Competitive inhibitors comprises of purely reversible COX inhi-
bitors. Most compounds in this category compete reversibly with the fatty acid
substrate, arachidonic acid, for binding at the COX active site. Examples of
NSAIDs that fall into this category include ibuprofen, piroxicam, naproxen
and mefenamic acid.
Slow, tight-binding inhibitors exhibit more complex kinetics than simple,
competitive inhibitors. Like competitive inhibitors, they too form the initial,
reversible (E I) complex, but this complex is converted to a more stable (E I)*
complex in a time-dependent fashion at an enzyme–inhibitor ratio of 1 : 1.16,17
In vitro, dissociation of the inhibitor from the (E I)* complex occurs very
slowly. Examples of NSAIDs in this category include indomethacin, meclofe-
namic acid and diclofenac. Aspirin is the only NSAID that covalently modifies
COX-1 and COX-2.
The mechanism of COX inactivation involves initial, reversible binding at
the active site, followed by irreversible acetylation of an active site Ser530
residue in the two isozymes (Figure 3.5).18–20 Ser530 is not important for COX
activity; mutagenesis of this residue to alanine does not affect catalysis or
arachidonate binding, suggesting that covalent modification of this residue by
aspirin inhibits COX activity by simply blocking arachidonic acid binding to
the COX active site.

3.2.3 Molecular and Structural Basis for COX Inhibition by

NSAIDs
Crystal structures of complexes of sheep COX-1, mouse COX-2 and human
COX-2 with NSAIDs have been solved.21–24 Despite their structural diversity
and differences in modes of inhibition, all NSAIDs bind in the substrate
Carboxylic Acids and their Bioisosteres 107

O
O
HO Ser530
O O
H3C O O HN
H3C O Ser530
OH
HN

Acetylsalicylic acid Acetylated COX-1/-2

(Aspirin) OH O

OH

Figure 3.5 Covalent modification of COX-1 and COX-2 by aspirin.

access channel with their carboxylic acid moiety ion paired to an active site
Arg120 residue. The Arg120 residue also ion pairs with the carboxylate of ara-
chidonic acid. Site-directed mutagenesis of the arginine residue in COX-1 to
glutamine or glutamate renders the protein resistant to inhibition by carbo-
xylic acid containing NSAIDs.25,26 Crystallisation of COX-1 acetylated by
bromoacetylsalicyclic acid not only confirms Ser530 acetylation but also reveals
a salicylate ion-paired to Arg120.22 Arg120 is part of a hydrogen bonding
network with Glu524 and Tyr355 which stabilises substrate–inhibitor interac-
tions and closes off the upper part of the COX active site from the spacious
opening at the base of the channel. Disruption of this hydrogen bonding
network opens the constriction and enables substrate–inhibitor binding and
release to occur.
It is important to note that selective COX-2 inhibitors are actually compe-
titive inhibitors of both COX-1 and COX-2, but exhibit selectivity for COX-2
in the time-dependent step by binding tightly at the active site and causing a
conformational change in the isozyme structure (see Chapter 5 for a detailed
description of sulfonamide-based selective COX-2 inhibitors).

3.3 Carboxylic Acid Containing b-Lactam Antibiotics

b-Lactam antibiotics are a broad class of antibacterial agents that include
penicillin derivatives, cephalosporins and cephamycins.27–29 Common struc-
tural features in these compounds is the presence of the b-lactam nucleus
and a carboxylic acid moiety (Figure 3.6). The differences in these various
derivatives (other than chemical structure) are related to absorption pro-
perties, resistance to penicillinases and specificity for organisms for which they
are most effective.27–29 Penicillin antibiotics are historically significant
because they were the first drugs that were effective against many pre-
viously serious diseases such as tuberculosis, syphilis, and staphylococcus
infections.
 108

H H H
R N S H R2 H
R1 N S
O N
O O N
O R3
COOH
COOH

General scaffold of penicillins

General scaffold of cephalosporins/cephamycins

CH2 OCH2
Penicillin V: R =
Cephalosporins: R2 = H; Cephamycins: R2 = OCH3
Penicillin G (Bicillin): R =

Cl
H2C NH2 R2 = H R3 = Cl
H2 C Cefaclor: R1 =
Cloxacillin (Cloxapen): R =
Oxacillin (Bactocill): R =
CH3 CH3 O
N O N O
Cefoxotin: R1 = CH2 R2 = OCH3 R3 = O NH2
S
CH
CH NH2
NH2
Amoxicillin: R = Ampicillin: R =
HO

Figure 3.6 b-Lactam antibiotics.

Chapter 3
Carboxylic Acids and their Bioisosteres 109
3.3.1 Discovery of Penicillins
The discovery of penicillin in 1928 is attributed to Scottish scientist and Nobel
laureate, Sir Alexander Fleming, who noticed a green mould growing in a
culture of Staphylococcus aureus; where the two had converged, the bacteria
were lysed.30–32 This led to the discovery of penicillin, which was produced by
the mould. This serendipitous observation began the modern era of antibiotic
discovery. In the early 1940s, the chemical structure of penicillin was deter-
mined by Dorothy Crowfoot Hodgkin, while a team of Oxford scientists led by
Sir Howard Florey and Ernst Boris Chain demonstrated the in vivo bacterial
action of penicillin and also discovered a method of producing the drug in
adequate quantities to treat humans.30–32 Florey and Chain shared the 1945
Nobel Prize in Medicine with Fleming for their work. Strides in the fermen-
tation technology arena and the discovery of moulds containing the highest
quality of penicillin allowed the mass production of the drug; approximately
2.3 million doses were prepared in time for the Allied invasion of Normandy in
the spring of 1944.

3.3.2 Mechanism of Action of b-Lactam Antibiotics

Unlike sulfonamide antibacterial agents that exhibit bacteriostatic activity,
penicillins and cephalosporins/cephamycins are bacteriocidal, i.e. they destroy
existing bacteria.33,34 The biosynthesis of bacterial cell wall peptidoglycan is
catalysed and controlled in its final stages by a class of transpeptidase enzymes,
which act on the D-alanyl-D-alanine peptide appendages of glucosyl poly-
saccharides. The enzymes, after removing the COOH-terminal D-alanine, use
the new carbonyl to form a peptide bond with an amino acceptor group on a
neighbouring polysaccharide peptide. This transpeptidation produces a cross-
linked cell wall network. Wall biosynthesis is inhibited by penicillins and
cephalosporins.34 The b-lactam is able, because of a structural resemblance to
the D-alanyl-D-alanine segment, to compete with the catalytic process and form
a transient penicilloyl–enzyme complex (Figure 3.7); the biochemical phe-
nomenon has been proven via solution of the crystal structures of cephalos-
porin bound to a bifunctional serine-type D-alanyl-D-alanine carboxypeptidase/
transpeptidase.35–37 Penicillins function as affinity labels of the peptidoglycan
transpeptidase by irreversibly acylating a catalytically active serine residue (see
Figure 3.7); covalent binding at the active site prevents the substrate from
binding.38–42 The beauty of the penicillins (and cephalosporins) is that the b-
lactam ring is not exceedingly reactive; consequently, few non-specific acylation
reactions occur with these molecules in mammals.
Although penicillins are ‘wonder drugs’ in their activity against a variety of
bacteria, and are still used widely today, many strains of bacteria have become
resistant to their effects. This has been attributed to the excretion of the enzyme
b-lactamase in resistant bacteria, which catalyses the hydrolysis of the b-lac-
tams covalently attached to the target transpeptidase enzyme.43 In the 1970s,
several naturally occurring b-lactams, which lacked the general penicillin or
110 Chapter 3

O
O H
H3C N
H COOH
NH
O NH COOH O NH S CH3
H3C
peptidoglycan peptidoglycan
Acyl D -alanyl-D-alanine
motif in transpeptidase

O O
O
HO HN R
H
R N S HN O S
HN
O N Peptidoglycan transpeptidase O HN
O
COO COO
β-lactam antibiotic Acylated enzyme

Figure 3.7 Mechanism of pharmacological action of b-lactam antibiotics.

NH2 O O
NH2 H H H
H H H O OH N S S
N S
N O N N
O N O O O
HO O COO K COOH
COOH
COOH
Ampicillin Sulbactam
Amoxicillin Clavulanate potassium

Augmentin Unasyn

Figure 3.8 Penicillin–b-lactamase inhibitor combination to combat bacterial resistance.

cephalosporin structure were isolated from various organisms and were found
to be potent mechanism-based inactivators of b-lactamases (Figure 3.8).44,45
These compounds are used in combination with penicillins to destroy penicillin-
resistant strains of bacteria. For example, the combination of amoxicillin and
clavulanate (a b-lactamase inactivator) is sold as Augmentin, and ampicillin
plus sulbactam is sold as Unasyn (see Figure 3.8).46–48 The b-lactamase inhi-
bitors have no antibiotic activity, but they protect the penicillin from
destruction so that it can interfere with cell wall biosynthesis.

3.4 Carboxylic Acid Containing Statins

The discovery of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA)
reductase inhibitors—called statins—was a breakthrough in the prevention of
hypercholesterolemia (high cholesterol) and related diseases. Hypercholester-
olemia is considered to be one of the major risk factors for atherosclerosis,
Carboxylic Acids and their Bioisosteres 111
which often leads to cardiovascular, cerebrovascular and peripheral vascular
diseases.49,50 HMG-CoA reductase is the rate-limiting enzyme of the mevalo-
nate pathway, the metabolic pathway that produces cholesterol and other
isoprenoids.51,52 Inhibition of this enzyme in the liver by statins results in
decreased cholesterol synthesis and an increased synthesis of low-density
lipoprotein (LDL) receptors, which leads to the increased LDL clearance from
the bloodstream, ultimately reducing the risk of atherosclerosis and diseases
caused by it.51–54

3.4.1 Discovery of the Statins

Over hundred years ago, a German pathologist named Rudolph Virchow dis-
covered the presence of cholesterol in the arterial walls of humans that died from
occlusive vascular diseases such as myocardial infarction.54 In the 1950s, the
Framingham heart study led by Thomas Royle Dawber revealed the correlation
between high blood cholesterol levels and coronary heart diseases;55 this led
scientists to explore novel ways of lowering cholesterol levels without significant
changes in diet and lifestyle. Because the primary goal was to inhibit cholesterol
biosynthesis in the body, HMG-CoA reductase became a natural target. In the
1970s, Akira Endo and Masao Kuroda initiated research into inhibitors of
HMG-CoA reductase.56 The Japanese team reasoned that certain microorgan-
isms may produce inhibitors of this enzyme as a defence mechanism against
other organisms, as mevalonate is a precursor of many substances required by
organisms for cell wall maintenance. During the course of these studies, the team
isolated mevastatin (Figure 3.9), a potent HMG-CoA inhibitor from a fer-
mentation broth of Penicillium citrinum.56,57 Likewise, in 1978, Alfred Alberts
and co-workers at Merck discovered a new natural product-based HMG-CoA
reductase inhibitor in a fermentation broth of Aspergillus terreus, which later
became known as lovastatin (Figure 3.9)—the first commercially marketed
statin.58 Commercially available statins are categorised into two groups: fer-
mentation-derived and synthetic. Fermentation-derived statins include lovasta-
tin, simvastatin, and pravastatin, whereas, synthetic statins include atorvastatin,
fluvastatin, cerivastatin, pitavastatin and rosuvastatin (see Figure 3.9).
In addition, statins have sometimes been grouped according to their structure
into type 1 and type 2 statins.59 Type 1 statins (e.g. lovastatin, pravastatin and
simvastatin) possess a substituted decaline-ring structure which resembles
mevastatin. Type 2 statins (e.g. atorvastatin, cerivastatin, fluvastatin and rosu-
vastatin) have larger hydrophobic groups linked to the carboxylic acid side chain.
One of the main differences between the type 1 and type 2 statins is the
replacement of the butyryl group in type 1 statins with a fluorophenyl group in
type 2 statins (see Figure 3.9). The fluorophenyl group is responsible for
additional polar interactions that causes tighter binding of the statin to the
HMG-CoA reductase enzyme.59 As discussed later in great detail, the type 1
statins simvastatin and lovastatin are commercially available in the inactive
lactone forms; they undergo metabolism to their active hydroxy acid forms
 112

Derived from fermentation (Type 1)

HO
HO O HO COOH
HO O HO O COOH
HO
COOH OH
O OH O
O O O
O OH O
O O
O
O O H
O O H
H O H
H H

HO

Pravastatin
Mevastatin Lovastatin (Lactone-form) Lovastatin (hydroxyacid-form) Simvastatin (Lactone-form) Simvastatin free hydroxy acid

Synthetic origin (Type 2) F

F

N OH OH OH
OH OH
COOH
COOH COOH
H N
N OH
N OH H3CO H3C
OH N N
O N O S
COOH CH3 F
F O
Atorvastatin Fluvastatin Cerivastatin Rosuvastatin

Figure 3.9 Categorisation of statins based on origin and chemical structure.

Chapter 3
Carboxylic Acids and their Bioisosteres 113
in vivo. In contrast, type 2 statins are commercially available in their active
hydroxy acid forms.

3.4.2 Molecular and Structural Basis for Inhibition of HMG-

CoA Reductase by Statins
Biochemical studies have shown that statins bind reversibly to the HMG-CoA
reductase enzyme with nanomolar range affinity. The affinity of the natural
substrate, 3-hydroxy-3-methylglutaryl-CoA towards conversion to malonyl-
CoA (Figure 3.10) is in the micromolar range.60,61
The essential structural components of all statins are a dihydroxyheptanoic
acid unit and hydrophobic ring system architecture. The statin pharmacophore
is a modified hydroxyglutaric acid component, which is structurally analogous
to the endogenous substrate and product transition state intermediate (see
Figure 3.10). Because HMG-CoA reductase reveals a stereoselective bias in
statin binding, all statins require a 3R,5R stereochemistry in the dihydrox-
yheptanoic acid unit for inhibition. Co-crystallisation of HMG-CoA reductase
with statins reveals that the carboxylic acid inhibitors exploit a shallow
hydrophobic groove to accommodate their hydrophobic domains.59,62–64
The specificity and the tight binding of statins is due to orientation and
bonding interactions that form between the statin and the HMG-CoA reduc-
tase. Polar interactions are formed between the HMG-like moiety in the statin
and residues that are located in the cis loop of the enzyme. The terminal car-
boxylic acid group in statins forms a salt bridge with a positively charged lysine
(Lys735) residue in the active site. In addition, Lys691 participates in a hydrogen
bonding network with Glu559, Asp767 and the hydroxyl group of the hydro-
xyglutartic acid component in statins.59

3.5 Carboxylic Acid Containing Fibrates

Fibrates are a class of lipophilic carboxylic acid derivatives which are used in
accessory therapy in many forms of hypercholesterolemia, usually in combi-
nation with statins.65 Although less effective in lowering LDL than statins,

HO
HO HO COO
COO COO
H3C H3 C OH
O OH
H
S-CoA S-CoA
HMG-CoA Ring
Mevaldyl-CoA
transition state intermediate Statin pharmacophore

Figure 3.10 Molecular/structural basis for inhibition of HMG-CoA reductase by

statins—the statin pharmacophore.
114 Chapter 3
fibrates improve high-density lipoprotein (HDL) by B20–30% and triglyceride
levels by B40%.65,66
Although used clinically since the 1930s, the mechanism of action of fibrates
remained unelucidated until it was discovered in the 1990s that fibrates activate
PPAR (peroxisome proliferator-activated receptors), especially PPARa in
muscle, liver and other tissues.67,68 The PPARs comprise a family of ligand-
activated transcription factors that play a key role in lipid homeostasis via
modulation of carbohydrate, fat metabolism and adipose tissue differentiation.
There are three members of the family: PPARa, PPARd(or b) and PPARg.69 In
humans, PPARa activation results in increased clearance of triglyceride-rich
very low-density lipoprotein and upregulation of ApoA1, the principal lipo-
protein component of HDL.70 As a consequence, the fibrates lower triglyceride
and raise HDL levels. Fibrates prescribed commonly include fenofibrate,
gemfibrozil, clofibrate, ciprofibrate and bezafibrate (Figure 3.11). Fenofibrate
and clofibrate are sold as the corresponding ester derivatives; the active
metabolites, responsible for the pharmacological activity, are fenofibric acid
and clofibric acid, respectively.
Though effective for dyslipidemia, fibrates are weak PPARa agonists (EC50
B30–50 mM) and their subtype selectivity is poor. Much research effort has
been invested within the pharmaceutical industry to improve upon the potency
of human PPARa agonism and several groups have reported the discovery of
potent PPARa agonists (EC50’s in the low nanomolar range) and their effects in
animal models of dyslipidemia.71–74 The archetypal small molecule PPAR
agonists are structurally divided into three parts, for example, (1) a carboxylic
acid head piece, (2) a linker part, and (3) a hydrophobic tail part. Examples of
some novel, selective PPARa agonists are shown in Figure 3.12 Increase in the

O O
O O
O OH
O O
O O
O O
O OH

Cl Cl
Cl Cl
Fenofibrate Fenofibric acid Clofibrate Clofibric acid
O
O
OH
CH3
O O
O
HN
OH O
OH
O Cl
CH3
Cl
Cl

Bezafibrate Gemfibrozil Ciprofibrate

Figure 3.11 Fibrates in the treatment of dyslipidemia.

Carboxylic Acids and their Bioisosteres 115
CF3 O
HOOC Cl S
N CF3
H
O O O N
HOOC O

H O
N
O HOOC
N N N
HOOC O H
OCH3 CF3

Figure 3.12 Novel PPARa agonists for the potential treatment of dyslipidemia.

O O O

OH R R
O N
H
O R = alkyl, aryl

OH O
O O O O
S
N R N
H H
R = alkyl, NH2, NH-R1

Figure 3.13 Non-classical bioisosteres of the hydroxyl group in a carboxylic acid

moiety.

amphipathic (lipophilic) character generally results in improved PPAR agon-

ism against the respective subtypes.

3.6 From Terfenadine to Fexfofenadine–an Interesting

Case Study on the Utility of the Carboxylic Acid
Moiety in Drug Discovery
The selective H1 antihistamine terfenadine (Seldanes) (Figure 3.13) was the
first non-sedating antihistamine to be introduced for the treatment of allergic
rhinitis. It exhibited little or no incidence of central nervous system (CNS)
sedative effects associated with the first-generation antihistamines.75,76 Since its
introduction in the 1980s, terfenadine ranked as one of the most widely pre-
scribed drugs in the United States. However, a published report of ventricular
arrhythmia associated with terfenadine use first appeared in 1990, wherein a
patient developed Torsades de Pointes while on the recommended daily doses of
terfenadine concomitantly with cefaclor, ketoconazole and progesterone.77 The
risk of cardiac arrhythmia caused by QT interval prolongation was traced back
to excessively high serum concentrations of terfenadine, which occurred due to
116 Chapter 3

P4503A4 P4503A4
OH OH OH
HO HO HO
N N N

Terfenadine 3.1 3.2

OH O OH
(Fexofenadine)

Scheme 3.2 Discovery of non-cardiotoxic and non-sedating H1 antagonist fexofenadine.

ketoconazole-mediated inhibition of its principal metabolic elimination path-

way.78–80
Terfenadine is extensively metabolised by cytochrome P450 (P450) 3A4 via
initial hydroxylation on its t-butyl group to the primary alcohol metabolite 3.1
followed by its oxidation to the corresponding carboxylic acid metabolite 3.281
(Scheme 3.2). Under normal recommended dosages, terfenadine is not asso-
ciated with cardiotoxicity because very low, free systemic concentrations of
terfenadine are achieved in vivo, a consequence of its extensive and rapid first
pass metabolism.82 However, toxic effects on the heart’s rhythm and electrical
conduction such as ventricular tachycardia and Torsades de Pointes are dis-
cerned upon concomitant administration of terfenadine with P4503A4
inhibitors.
In early 1997, given the increased number of cases of pharmacokinetic
interactions between terfenadine and P450 3A4 inhibitors,83,84 the US
Food and Drug Administration (FDA) recommended that terfenadine be
removed from the market and that physicians consider alternative medications
for their patients. Terfenadine was formally removed from the US market in
late 1997.
Of much interest against this backdrop were the findings that the carboxylic
acid metabolite 3.2 retained all the primary pharmacology (non-sedating H1
antagonism) associated with terfenadine and that in vivo metabolite 3.2 appears
to exert most, if not all, of the pharmacological actions associated with the
administration of the parent compound.82 Importantly, 3.2 was devoid of the
cardiotoxic potential associated with the parent compound in the clinic; sub-
sequently, 3.2 (later named as fexofenadine) replaced terfenadine on the
market.85

3.7 Bioisosteres of the Carboxylic Acid Moiety

Non-classical bioisosteres for the carboxylic acid moiety consists of replace-
ments which involve (a) only the hydroxyl portion or (b) both the hydroxyl and
carbonyl fragments of the functional group. The determination of suitable
replacements for the carboxylic acid group is often based on the ability of the
bioisostere to possess similar acidity and to exhibit similar physiochemical
properties.
Carboxylic Acids and their Bioisosteres 117
3.7.1 Non-classical Bioisosteres of the Hydroxyl Portion of the
Carboxylic Acid Group
The types of non-classical bioisosteres typically used as hydroxyl replacements
are similar to the non-classical bioisosteres of the phenolic hydroxyl group.86
Of the prototypic fragments shown in Figure 3.13, replacement of the hydroxyl
group in the COOH motif with a phenylsulfonamide results in the formation of
a acylsulfonamide, which possesses a pKa comparable to the carboxylic acid
moiety (see Chapter 5 for a detailed description of this bioisostere).
Because the molecular size of these individual fragments is larger than that of
the hydroxyl group,86 these non-classical bioisosteres are unlikely to be suitable
in cases where pharmacological activity is adversely effected by an increased
molecular size in the vicinity of the carboxylic acid group. These non-classical
bioisosteres tend to be most effective in those instances where the role of the
hydroxyl group in the carboxylic acid group is to act as either a hydrogen bond
acceptor or donor. An interesting example where non-classical bioisosteres of
the hydroxyl group significantly impacted primary pharmacology is evident
with recent studies on the conversion of certain non-selective NSAIDs into
non-ulcerogenic, selective COX-2 inhibitors.87–89

3.7.1.1 Neutral Derivatives of Non-selective Carboxylic Acid

Containing NSAIDs as Selective COX-2 Inhibitors
Derivatisation of the carboxylic acid moiety in the selective COX-1 inhibitors
from the arylacetic and anthranilic acid class of compounds—exemplified by
indomethacin and meclofenamic acid, respectively—affords potent and selec-
tive COX-2 inhibitors (Figure 3.14).87–89 Within the indomethacin series,
esters (compounds 3.3 and 3.4) and primary and secondary amides (compounds
3.5 and 3.6) are superior to tertiary amides (compound 3.7) as selective inhi-
bitors. Furthermore, increase in the lipophilicity of the amide and/or
ester substituent also increased COX-2 inhibitory potency and selectivity
(see Figure 3.14—compounds 3.3 and 3.5 and compounds 3.4 and 3.6). Ela-
boration of structure–activity relationship (SAR) efforts also led to the dis-
covery of the corresponding reverse esters (e.g. compound 3.8) and reverse
amides (e.g. compound 3.9) of indomethacin as selective COX-2 inhibitors.90
Inhibition kinetics reveal that the neutral indomethacin derivatives behave as
slow, tight-binding inhibitors of COX-2 and that selectivity is a function of the
time-dependent step, as is the case with the diarylheterocycle-based COX-2
inhibitors celecoxib, valdecoxib and rofecoxib. Studies with site-directed COX-
2 mutants, however, indicated that the molecular basis for COX-2 selectivity of
indomethacin derivatives differs from the parent NSAIDs and from diarylhe-
terocycles. For example, the Arg120 residue, which is a critical determinant of
fatty acid substrate arachidonate as well as inhibitor binding, is not important
for inhibition by indomethacin amides and esters. COX-2 selectivity was shown
to arise from novel interactions at the opening and at the apex of the arachi-
donic acid-binding site.87,89,90 In in vivo animal models of acute inflammation,
 O
O
118
O
X
H3CO
R
OH H3CO CH3
H3CO
CH3 N
CH3 N
N
O Cl
O Cl
O Cl
Indomethacin 3.3: R = OCH3 3.4: R = NH2 3.5: R = O 3.6: R = NH
IC50 (COX-2) ~ 0.01 μM IC50 (COX-2) ~ 0.25 μM IC50 (COX-2) ~ 0.70 μM IC50 (COX-2) ~ 0.05 μM IC50 (COX-2) ~ 0.06 μM
IC50 (COX-1) ~ 0.006 μM IC50 (COX-1) ~ 33 μM IC50 (COX-1) ~ 25 μM IC50 (COX-1) > 66 μM IC50 (COX-1) > 66 μM

Cl

X
N O
O R
CH3
H3CO H
O N
H3CO
CH3
CH3 N
N CH3
O Cl Cl
O Cl
3.7 3.8: R = O 3.9: R = NH Meclofenamic acid: R = OH 3.10: R = NH(CH2)2OC6H5
IC50 (COX-2) > 33 μM IC50 (COX-2) ~ 0.05 μM IC50 (COX-2) ~ 0.05 μM IC50 (COX-2) ~ 0.05 μM IC50 (COX-2) ~ 0.15 μM
IC50 (COX-1) > 66 μM IC50 (COX-1) > 66 μM IC50 (COX-1) > 66 μM IC50 (COX-1) ~ 0.04 μM IC50 (COX-1) ~ 66 μM
Chapter 3

Figure 3.14 Neutralisation of the carboxylic acid group in non-selective COX inhibitors indomethacin and meclofenamic acid yields selective
COX-2 inhibitors.
Carboxylic Acids and their Bioisosteres 119
candidate compounds were also shown to possess oral anti-inflammatory
activity without the ulcerogenic effects associated with the parent NSAIDs.87 In
a manner similar to indomethacin, some secondary amide derivatives of the
fenamic acid NSAID, meclofenamic acid, also demonstrated potent and
selective COX-2 inhibition.89 The 2-phenoxyethylamide derivative 3.10 was the
most selective inhibitor, with a COX-2 selectivity ratio of B440 (Figure 3.14).
Unlike indomethacin SAR, only the amide derivatives of meclofenamic acid
demonstrated COX-2 selectivity. The esters were either inactive or non-selective
COX inhibitors. The reason(s) for this discrepancy is unclear. Finally, it is
interesting to note that simple derivatisation involving amidation and/or
esterification is not a universal strategy to convert all traditional carboxylic acid
NSAIDs into selective COX-2 inhibitors. For instance, amidation or ester-
ification of naproxen, sulfindac and/or ibuprofen yields inactive compounds
(A. S. Kalgutkar, unpublished observations).

3.7.2 Non-classical Bioisosteres of the Entire COOH Moiety

Non-classical bioisosteres as replacement of the entire carboxylic acid group
are also widely known; in particular, sulfonamides, tetrazolyl and thiazolidi-
nedione derivatives as non-classical bioisosteric replacements have yielded
many commercially successful medicines. Of late, the boronic acid group has
also emerged as a carboxylic acid surrogate and has resulted in the discovery of
the peptidyl proteasome inhibitor, bortezomib. Discussion on the sulfonamide
group as a non-classical bioisostere of the carboxylic acid moiety is provided in
Chapter 5.

3.7.2.1 5-Substituted-1H-tetrazoles as Carboxylic Acid

Bioisosteres
5-Substituted-1H-tetrazoles are excellent non-classical bioisosteres of car-
boxylic acid derivatives.91–94 Because of the presence of a free N–H bond, the
tetrazole moiety can exist in a B1:1 ratio of the 1H- and 2H-tautomeric forms
(Figure 3.15). The free N–H bond in tetrazole analogs also imparts acidity due
to its ability to stabilise negative charge via electron delocalisation. Both ali-
phatic and aromatic tetrazoles possess pKa values (B4.5–4.9) which are
comparable to carboxylic acids.95–97 Like their carboxylic acid counterparts,

4 4
O 3 3
N N
5 N 5 N
R R R
N NH
OH N1 2 1 N 2
H
(1H )-tetrazole (2H )-tetrazole

Figure 3.15 5-Substituted-1H-tetrazoles as non-classical bioisosteres of carboxylic

acids.
120 Chapter 3
tetrazoles are ionised at physiological pH; however, it is important to note that
tetrazoles in the anionic form are B10-fold more lipophilic than the corre-
sponding carboxylate derivatives.98

3.7.2.2 Tetrazole-based Angiotensin II Receptor Antagonists:

the Discovery of Losartan and Related Analogs
The renin–angiotensin system (RAS) plays a key role in blood pressure reg-
ulation and sodium balance.99 RAS can be interrupted at various levels; it can
be blocked either by inhibition of rennin or angiotensin-converting enzyme, or
via direct antagonism of the G protein-coupled angiotensin II receptors.
The angiotensin II type I (AT1) receptor is the best elucidated in terms of its
biochemistry and signalling functions. The receptor is activated by the vaso-
constricting octapeptide angiotensin II, which in turn results in an increase in
cytosolic Ca21 concentrations (through activation of phospholipase C).
Amongst other actions, blockade of the AT1 receptor directly causes vasodi-
lation, reduces secretion of vasopressin, reduces production and secretion of
aldosterone; the combined effect of which is reduction of blood pressure.
In the 1990s, numerous pharmaceutical companies were engaged in research
directed at discovering novel AT1 receptor antagonists for the treatment of
hypertension. Although potent peptide-based AT1 receptor antagonists (e.g.
saralasin) have been used as pharmacological tools for many years, the ther-
apeutic utility has been limited primarily due to poor oral bioavailability and
significant agonist activity at the receptor.100,101
Figure 3.16 depicts key medicinal chemistry milestones that led to increa-
sing more potent and, eventually, orally bioavailable non-peptide AT1 receptor
antagonists. Benzimidazole derivatives 3.11 and 3.12 were initially disclosed by
Takeda as non-peptide hits that demonstrated weak but selective AT1 receptor
antagonist properties.102,103 Another attractive feature was the lack of func-
tional agonist activity at the receptor.102,103 Because the C-terminus carboxylic
acid group in angiotensin II is essential for binding to the receptor,104
researchers at Du Pont decided to incorporate a second carboxylic acid
moiety on the N-benzyl group in 3.11 and 3.12, which led to 3.13, a vastly
improved AT1 receptor antagonist.105 Further improvements in potency were
gained by ‘enlarging’ the molecular size and lipophilicity of 3.13; although
analogs such as 3.14 and 3.15, which incorporated an additional phenyl ring via
amide, ether, amine and/or thiol linkage exhibited significant gains in
antagonist potency, the compounds were devoid of oral bioavailability in
preclinical species.105,106
A breakthrough was achieved when the biphenyl carboxylic acid derivative
EXP7711 (compound 3.16) was synthesised and shown to be an orally active
AT1 receptor antagonist.106 While orally active, 3.16 was slightly less potent
than 3.14 in inhibiting angiotensin II binding.107 In the hope of further
improving oral activity and potency of the biphenyls, a number of carboxylic
acid bioisosteres were evaluated.108 The tetrazole bioisostere proved to be the
 Cl Cl Cl Cl
N N N N
COOH COOH COOCH3 OH
N N N N

COOH
H
N
R COOH X
3.11: R = Cl; IC50 ~ 40 μM 3.13 (IC50 ~ 1.2 μM) O COOH
3.12: R = NO2; IC50 ~ 13 μM 3.14 (IC50 ~ 0.12 μM) 3.15 (X = O, CH2, NH, S)
Carboxylic Acids and their Bioisosteres

Cl Cl Cl
N N N
CYP Metabolism OH OH
N COOH N
N

N N
N HOOC
N
N NH N NH 3.16 (EXP7711) IC50 ~ 0.23 μM

3.17 (DuP 753, Losartan) Orally active analog

3.18 (EXP3174)
IC50 ~ 0.001 μM IC50 ~ 0.005 μM

Figure 3.16 Historical overview of the discovery of the tetrazole-based angiotensin II receptor antagonist, losartan.
121
122 Chapter 3
key in the discovery of the first non-peptide AT1 receptor antagonist losartan
(DuP 753) (3.17) with good oral bioavailability, vastly improved antagonist
potency (IC50 B5.5 nM) and a long duration of action.109–111 It is noteworthy
to point out that the carboxylic acid metabolite of losartan, i.e. EXP3174
(compound 3.18), is an active metabolite with greater AT1 receptor antagonism
than the parent compound.112,113
The value of the tetrazole motif in the discovery of novel, selective and orally-
active AT1 receptor antagonists is clearly evident from the fact that five out of
the six drugs in this class that are currently marketed for the treatment of
hypertension contain the tetrazole group. The list includes losartan, irbesartan,
olmesartan medoxomil (active metabolite: olmesartan), valsartan and cande-
sartan cilexetil (active metabolite: candesartan); telmosartan contains a car-
boxylic acid group instead of the tetrazole motif (Figure 3.17). It is important
to note that site-directed mutagenesis studies have provided evidence that the
tetrazole moiety in the non-peptide antagonists interacts with a protonated
lysine and histidine at the recognition site of the AT1 receptor in a manner
similar to the interaction of the carboxy terminus of the natural ligand
angiotensin with the receptor.114,115

3.8 Absorption, Distribution, Metabolism and

Excretion (ADME) Profile of Carboxylic Acids
Structural diversity within the carboxylic acid based drugs results in subtle
differences in physiochemical properties [e.g. molecular weight (MW), clogP,
log D] which then influences aqueous and lipid solubility, and subsequently the
pharmacokinetic disposition profile.

3.8.1 Oral Absorption

With the exception of some b-lactam-based antibiotics, most small molecule
carboxylic acid based drugs are administered by the oral route and hence
absorption into the target tissue (e.g. liver in the case of statins and systemic
circulation in the case of NSAIDs) is essential for their pharmacological action.
As such, several marketed carboxylic acid drugs obey Lipinski’s ‘rule of five’
(see Section 2.1.1) for good oral absorption; Table 3.3 depicts the aqueous
solubility, Caco-2 cell permeability and human oral bioavailability character-
istics of a series of low MW NSAIDs.116
The presence of the free carboxylic acid group in drugs also provides a
convenient handle for preparation of salts to improve aqueous solubility (and
therefore oral absorption and bioavailability). In other cases, the carboxylic
acid moiety can be neutralised to corresponding ester prodrug derivatives to
improve oral absorption profile by virtue of improved membrane permeability
especially in the case of polar carboxylate analogs (e.g. certain b-lactam
antibiotics).
 O
Cl OH O
Cl OH
N N N
N O N
OH O
N O N COOH
N N COOH N
P450 O

N N N
N N
N N N
N N
N NH N NH N NH
N NH N NH
Carboxylic Acids and their Bioisosteres

Losartan EXP3174 Irbesartan Olmesartan medoxomil Olmesartan

O N H3 C
N

N COOH N O O N N
N
O HOOC
O N N
O
H3 C
O
O
N N N
N N N
N NH N NH N NH HOOC

Valsartan Candesartan cilexetil Candesartan Telmisartan

Figure 3.17 Structures of the marketed angiotensin II receptor antagonists for treatment of hypertension.
123
124 Chapter 3
Table 3.3 Structure, dose, Caco-2 permeability and solubility characteristics
of NSAIDs.
High Papp (Caco-2)a Equilibrium solubilityb
dose Oral F
Compound MW (mg) (%) AB BA pH 1.2 pH 7.4

Diclofenac 295 50 54 20.2 21.3 1.0 15 900

Indomethacin 357 50 98 10.4 24.5 1.0 1300
Ibuprofen 206 800 480 9.60 19.2 60 2300
Sulindac 356 200 88 6.30 12.2 7.0 1300
Aspirin 180 975 68 25.5 19.1 6200 6400
Ketorolac 255 20 100 4.30 18.6 110 1300
Naproxen 230 500 99 12.3 20.0 5.0 2500
Diflunisal 250 500 90 12.5 17.0 3.0 2500
Salicylic acid 138 750 100 17.6 20.5 180 2400
Papp106 cm sec1.
a

mg mL1.
b

Drug O O R Esterase Drug O OH Non-enzymatic Drug OH

O O RCOOH O CH2O O

Scheme 3.3 Tripartate prodrug concept.

3.8.1.1 Carboxylic Acid Prodrugs as a Tactic to Improve Oral

Absorption
The carboxylic acid group can be coupled with various alcohol derivatives to
afford neutral ester analogs. Because, esterases are ubiquitous in mammals and
can hydrolyse structurally diverse substrates, metabolic regeneration of the
parent carboxylic acid drug is often a facile process. It is possible to prepare
ester derivatives with virtually any degree of hydrophilicity or lipophilicity;
furthermore, electronic and/or steric factors (on the carboxylic acid or alcohol
moieties) can be manipulated to control the rates of enzymatic hydrolysis and
therefore ester stability.
In some cases, prodrugs can be ineffective because the ester bond is too labile
or too stable. A remedy towards this situation is the design of tripartate (self-
immolative) prodrugs where the carrier is linked to the drug via a linker group
(Scheme 3.3).117,118 This feature allows for different kinds of functionalities to
be incorporated for varying stabilities; it also displaces the drug farther from
the hydrolysis site, which decreases steric interference by the carrier. The drug–
linker connection, however, must be designed so that it cleaves spontaneously
(i.e. is self-immolative) after the carrier has detached. A practical approach to
accomplish this is via the double prodrug as shown in Scheme 3.3.
The tripartate prodrug strategy has been employed to improve the oral
absorption of ampicillin (Scheme 3.4). The absorption of ampicillin in healthy
Carboxylic Acids and their Bioisosteres 125
NH2
H
N S

O N
O
COOH
Ampicillin
NH2
H
R R R N
Esterases O OH OH S
O O R2
R=
O O O O N
R1 O R1
R2COOH O
3.19

Bacampicillin: R1 = CH3; R2 = OC2H5

Pivampicillin: R1 = H; R2 = t-Bu

Scheme 3.4 Double ester prodrugs of ampicillin: discovery of bacampicillin and

pivampicillin.

humans was shown to be dose-dependent with decreased absorption at higher

doses. In healthy humans, the absorption decreased from 72% to 45% when
the oral dose increased from 500 to 3000 mg.119 The dose-dependent intestinal
absorption is in accordance with the carrier-mediated absorption process of
the b-lactam derivative observed in animal studies, although the decrease in the
absorption at very high doses (43000 mg) could also be partially a result of
the incomplete dissolution of ampicillin in the intestinal tract.119–122 Although
simple alkyl and/or aryl esters of ampicillin are rapidly hydrolysed in rodents,
they are resistant to esterase-mediated hydrolysis in humans presumably due to
steric hindrance of the ester carbonyl by the fused b-lactam ring system.
A solution to the dilemma was the construction of a double ester, an acy-
loxymethyl ester such as bacampicillin or pivampicillin (see Scheme 3.4) which
extends the terminal ester carbonyl away from the fused b-lactam ring system
and eliminates the inherent steric hindrance with the human esterases.118,123,124
Hydrolysis of the terminal ester (or carbonate in the case of bacampicillin)
affords the unstable hydroxymethyl ester 3.19 which spontaneously decom-
poses to ampicillin. Unlike ampicillin, bacampicillin is almost completely
absorbed, and ampicillin is liberated into the systemic circulation in o15
minutes.125,126 An additional example of the effectiveness of the tripartate
prodrug strategy is evident with the acyloxy prodrug of the angiotensin II
receptor antagonist candesartan (see Figure 3.17).

3.8.2 Distribution and Clearance

Because of extensive binding to albumin in plasma, many carboxylic acid drugs
(especially NSAIDs) possess low tissue affinity, resulting in a small volume of
distribution at steady state (Vdss), approaching plasma or blood volume (0.1 to
0.2 L kg1). The pharmacokinetic parameters [plasma clearance (CLp), Vdss,
half-life (T1/2)], plasma free fraction (fu) and physiochemical parameters of
126 Chapter 3
Table 3.4 Physiochemical and pharmacokinetic attributes of carboxylic acid-
based drugs.
CLp Vdss
Drug MW cLogP Log D7.4 fu (mL/min/kg) (L/kg) T1/2 (hr)

Diclofenac 296 4.73 0.95 0.005 3.5 0.22 1.4

Ibuprofen 206 3.68 0.80 0.006 0.82 0.15 1.6
Ketoprofen 254 2.76 0.25 0.008 1.6 0.13 2.1
Ketorolac 255 1.62 0.95 0.0068 0.35 0.11 5.1
Cerivastatin 459 3.68 1.5–1.7 0.01 2.9 0.33 1.8
Fluvastatin 411 4.04 1.0–1.2 0.0079 16 0.42 0.70
Pravastatin 424 2.04 0.75 0.5 14 0.46 0.78
Atorvastatin 558 4.46 1.11 0.01 8.9 5.4 7.8
Rosuvastatin 481 1.89 0.25 0.12 14 1.20 20
Penicillin G 334 1.75 2.06 0.40 6.9 0.24 0.70
Sulbactam 233 0.31 5.11 0.62 5.1 0.32 1.1

structurally diverse carboxylic acid containing drugs in humans are shown in

Table 3.4.127
Within this context, it is noteworthy to comment on statin lipophilicity in
relationship to their localised pharmacological action in the liver. The more
lipophilic statins tend to achieve higher levels of exposure in non-hepatic tis-
sues, while the hydrophilic statins tend to be more hepatoselective. The dif-
ference in selectivity is because lipophilic statins passively and non-selectively
diffuse into both hepatocyte and non-hepatocyte, while the hydrophilic statins
rely largely on active transport into hepatocyte to exert their effects. High
hepatoselectivity is thought to translate into reduced risks of adverse effects
including myopathy and rhabdomyolysis. Of the marketed statins, cerivastatin
was the most lipophilic (log D7.4) and also had the largest percentage of serious
adverse effects due to its ability to inhibit vascular smooth muscle proliferation;
as a result, it was voluntarily removed from the market by the manufacturer.

3.8.3 Metabolism of the Carboxylic Acid Moiety

Not surprisingly, the metabolism of carboxylate containing drugs mirrors that
of many endogenous processes. The opening discussion is a characterisation of
the principal routes of carboxylate biotransformation, and importantly, the
core set of pathways responsible for eliciting the pharmacokinetics, pharma-
cology and toxicology commonly associated with carboxylic acid containing
drugs.

3.8.3.1 Glucuronidation
From a quantitative perspective, glucuronidation is the most important route
of carboylic acid biotransformation yielding the corresponding b-1-O-acyl
glucuronides (also referred to as acyl glucuronides) (Scheme 3.5); these are
Carboxylic Acids and their Bioisosteres 127

1-O-β-glucuronide 3-O-β-glucuronide
HO2C
O HO2C HO2C OH
UGT O O HO
HO HO R O O
R OH HO O R R O OH OH
UDPGA OH OH O
Carboxylic acid O O NH2-protein
HX-protein
HO2C
X = S, O or NH OH protein
HO
HO2C R O N
O
O OH
HO protein
HO OH + R X O
OH
Schiff base
Amadori
Haptenization rearrangment

HO2C
OH protein
HO
Immune-mediated IADRs R O NH
O
O

Scheme 3.5 Mechanism of covalent adduction of acyl glucuronide metabolites to

proteins: plausible contributor to idiosyncratic drug toxicity of some
carboxylic acid drugs.

more polar than the parent acids due to the hydrophilic nature of the linked
glucuronic acid moiety. The biotransformation is catalysed by the family of
uridine 5 0 -diphospho-glucuronosyl transferases (UDP-glucuronosyl trans-
ferases, UGT, EC 2.4.1.17) that require uridine-diphosphate glucuronic acid
(UDPGA) as a co-factor. Depending on the structural features, molecular
weight and recognition pattern for active uptake and/or efflux, acyl glucur-
onides can be eliminated via renal or biliary excretion. Following biliary
excretion, acyl glucuronides may also be hydrolysed to regenerate the parent
carboxylic acid (aglycone) that can be reabsorbed from the gut into the portal
circulation via a process referred to as enterohepatic recirculation.128,129 Acyl
glucuronide hydrolysis is usually catalysed by b-glucuronidase enzymes,
although non-specific esterases can also participate in this process.130 Particu-
larly noteworthy are the examples in which rearranged isomers of some acyl
glucuronides (vide infra) display resistance to glucuronidase-mediated hydro-
lysis (e.g. diflunisal–acyl glucuronide) and thus present variations in enter-
ohepatic recirculation.131 Provided the acyl glucuronide is released into the
systemic circulation, hydrolysis can also occur in plasma. Numerous carboxylic
acid containing drugs including members from the NSAID, statin and fibrate
classes of compounds are subject to some degree of acyl glucuronidation as a
component of their elimination mechanism.132

3.8.3.2 Role of Acyl Glucuronide Metabolites in Drug Toxicity

Because acyl glucuronides are ester derivatives, they are intrinsically electro-
philic. The notion that acyl glucuronides could react with biological nucleo-
philes on proteins originated from observations that glucuronides of bilirubin
and a number of other carboxylic acid containing drugs (NSAIDs and fibrates)
128 Chapter 3
were able to form covalent adducts with human serum albumin in vitro.133–139
The in vitro observations have also been extended to the in vivo situation with
NSAIDs such as zomepirac and tolmetin.136,138,139 For instance, an in vivo
study in human volunteers revealed a linear correlation between the area under
the curve of zomepirac acyl glucuronide (but not the parent zomepirac itself) in
plasma and the amount of zomepirac covalently bound to plasma proteins.135
The mechanism of covalent adduction of acyl glucuronides to proteins can
proceed via two different pathways (Scheme 3.5).140–144 The first is a transa-
cylation mechanism where a nucleophilic amino acid on a protein macro-
molecule attacks the carbonyl group of the primary acyl glucuronide, leading to
the formation of an acylated protein and free glucuronic acid. The second
mechanism involves condensation between the aldehyde group of a rearranged
acyl glucuronide and a lysine residue or an amine group of the N-terminus,
leading to the formation of a glycated protein. The formation of the iminium
species is reversible but may be followed by an Amadori rearrangement of the
imino sugar to the more stable 1-amino-2-keto product.
A structural relationship between acyl glucuronide degradation to the Schiff
base and covalent binding has been established utilising carboxylic acid deri-
vatives with varying degrees of substitution on the carbon a to the carbonyl
group in the parent compounds (i.e. acetic, propionic and benzoic acid deri-
vatives).140–144 The results of these studies suggest that a higher degree of alkyl
substitution at the a-carbon leads to lower reactivity with biological nucleo-
philes giving rise to a general rank order of reactivity (acetic acid4propionic
acid 4 benzoic acid). These observations imply that inherent electronic and
steric properties must modulate the rate of acyl glucuronide rearrangement.
Of much interest within this context are the findings that acyl glucuronides of
acetic acid-based NSAIDs including ibufenac, tolmetin and zomepirac—all of
which have been withdrawn due to cases of idiosyncratic hepato- and/or renal
toxicity—exhibit the highest level of glucuronide rearrangement and covalent
binding, whereas mono-a-substituted acetic acids (2-substituted propionic
acids) such as ibuprofen and naproxen exhibit intermediate level of acyl glu-
curonide rearrangement and covalent binding.
The pair of NSAIDs, ibufenac and ibuprofen (Figure 3.18), provides one of
the most dramatic examples of structure–toxicity relationships in drug dis-
covery. While ibuprofen is one of the safest over-the-counter anti-inflammatory
agent on the market, its close-in analogue ibufenac was withdrawn due to

CH3
O O

OH OH

Ibufenac Ibuprofen

Figure 3.18 Structure–toxicity relationships for acyl glucuronidation—the magic

methyl: ibufenac (hepatotoxin) vs ibuprofen (non-hepatotoxin).
Carboxylic Acids and their Bioisosteres 129
severe hepatotoxicity. The daily doses of both NSAIDs are comparable (400–
800 mg) and the only structural difference between the two drugs is the presence
of the a-methyl substituent in ibuprofen. Both NSAIDs are subject to extensive
acyl glucuronidation in animals and humans;145,146 in the case of ibuprofen, it
has been shown that the presence of the extra a-methyl substituent slows acyl
glucuronide rearrangement to the electrophilic carbonyl intermediate capable
of covalently modifying critical proteins potentially leading to toxicity.146

3.8.3.3 Inhibition of UGT and Transport Proteins by Acyl

Glucuronides
A less common consequence, but noteworthy nonetheless, is the potential for
inhibition of active transport proteins by acyl glucuronide metabolites. This
occurrence was most recently reported in an account of an apparent renal
transporter-mediated drug–drug interaction (DDI) induced by the fibrate,
gemcabene.147,148 What was originally described as a synergstic lowering of
blood pressure during the concomitant administration of the angiotensin-
converting enzyme inhibitor, quinapril, and the fibrate gemcabene, was later
discovered to be the gemcabene-induced increase in the serum concentrations
of the active metabolite, quinaprilat, via inhibition of its renal excretion by
gemcabene and its acylglucuronide metabolite (Figure 3.19). Employing a rat in
vivo model and human transporters in vitro, it was demonstrated that the
acylglucuronide of gemcabene was a moderate inhibitor of human OAT3
(IC50 ¼ 197 mM; rOat3 IC50 ¼ 133 mM), the transporter responsible for the renal
uptake of quinaprilat.
Similarly, a hepatotoxicity characterised by transient hyperbilirubinemia was
observed in rats receiving oral administration of the drug candidate,
MLN8054149 (Figure 3.20). Particularly noteworthy was the efficiency at which
the acyl glucuronide metabolite disrupted bilirubin homeostasis via the inhi-
bition of OATP (IC50 B0.5 mM) mediated uptake of bilirubin as well as MRP2
(IC50 B3 mM) and MRP3 (IC50 B7 mM) mediated excretion of conjugated

OH
O O
OH OH
O HN
O O
N

O
Gemcabene (CI-1027) RO O

Quinapril: R=CH3CH2
Quinaprilat: R=H

Figure 3.19 Clinical DDIs mediated by acyl glucuronide metabolites.

130 Chapter 3

O
R
O

HN

N N

Cl
F N

HO2C
MLN8054: R = H O
MLN8054-glucuronide: R = HO
HO
OH

Figure 3.20 Inhibition of hepatobiliary transport of bilirubin by the acyl glucuronide

of MLN8054.

bilirubin.149 In addition to active transport inhibition, subsequent UGT inhi-

bition data (IC50 B200 mM) would also implicate the disruption of UGT1A1-
mediated bilirubin conjugation by the acyl glucuronide of MLN8054,
intensifying the disruption in bilirubin homeostasis. Importantly, the efficient
production of the acylglucuronide and hepatobiliary concentrations was only
observed in rat. Thus, the clinical manifestation of hyperbilirubinemia is not
expected.

3.8.3.4 Amino Acid Conjugation

Amino acid conjugation of carboxylic acids is an alternative to acyl glucur-
onidation and is considered to be a detoxication reaction leading to the for-
mation of amide metabolites. The conjugation of benzoic acid with glycine to
form hippuric acid (Scheme 3.6) was discovered in 1842, making it the first
biotransformation reaction described in the literature.150 The specific amino
acid involved in conjugation usually depends on the bioavailability of that
amino acid from endogenous and dietary sources. Glycine conjugates are
commonly observed as metabolites of carboxylic acids in mammals; glycine
conjugation in mammals follows the order herbivores4omnivor-
es4carnivores. Conjugation with L-glutamine is most common in primate drug
metabolism. It does not occur to any significant extent in non-primates. In
mammals, taurine is an alternate amino acid acceptor to glycine, although
arginine, asparagine, histidine, lysine, glutamate, aspartate, alanine and serine
conjugates also have been detected as carboxylic acid metabolites to some
degree or other in mammals.151 In addition, several dipeptides including glycyl-
glycine, glycyltaurine and glycyclvaline are known to participate in this con-
jugation pathway.151
 NH2
O OH N
CoA-S-COCH3 CH3COO NH2CH2COOH CoA-SH
O OH O O
Carboxylic Acids and their Bioisosteres

H H N O
N N P P N H
S O O O O N N
OH OH OH
O O H3C CH3
O O O
ATP AMP + PPi OH
P OH
Benzoic acid Acyl CoA thioester Hippuric acid
HO

Scheme 3.6 Mechanism of amino acid conjugation of carboxylic acids: conversion of benzoic acid to hippuric acid, the first biotransformation
reaction described in the scientific literature.
131
132 Chapter 3
As is usually the case with enzyme-catalysed reactions, the ability of
carboxylic acids to undergo amino acid conjugation depends on steric
hindrance around the carboxylic acid group and upon substituents on the
aromatic ring or aliphatic side chain. For instance, in rats, ferrets and
monkeys, the major pathway of phenylacetic acid biotransformation is
amino acid conjugation.152–155 However, due to steric hindrance, dipheny-
lacetic acid cannot be conjugated with an amino acid, so the major pathway
of diphenylacetic acid biotransformation in the same three species is acyl
glucuronidation.152,154,156

3.8.3.5 Mechanism of Amino Acid Conjugation

The mechanism of amino acid conjugation, as illustrated in the conversion of
benzoic acid to hippuric acid, is shown in Scheme 3.6. The carboxylic acid
moiety in xenobiotics and/or drugs is converted to the corresponding coenzyme
A thioester derivative by mitochondrial acyl CoA synthetases (long-chain fatty
acid-CoA ligases) and requires ATP. Conversion to the CoA thioester produces
a more hydrolytically stable product that can be transported in the cell readily
but is still quite reactive toward the appropriate amine nucleophiles. The
appropriate cytosolic and/or mitochondrial amino acid N-acyltransferase then
catalyses the condensation of the amino acid and the coenzyme A thioester to
give the amino acid conjugate. This step is analogous to amide formation
during the acetylation of aromatic amines by N-acetyltransferase. Two different
types of N-acyltransferases have been purified from mammalian hepatic
mitochondria. One prefers benzoyl-CoA as substrate, whereas the other prefers
arylacetyl-CoA.

3.8.3.6 Role of Acyl CoA Metabolites in Covalent Modification

of Proteins
The intermediate acyl CoA metabolites of carboxylic acids are thioester deri-
vatives, which possess sufficient electrophilicity towards nucleophilic reactions
with amino acid residue(s) on proteins as well as with the endogenous anti-
oxidant glutathione (GSH).157–160 For instance, the hypolipidemic drug nafe-
nopin (see Scheme 3.7) was able to transacylate liver proteins following in vitro
incubations with liver homogenates, resulting in amide and thioester linkages
with protein lysine and cysteine amino acid residues, respectively, and that the
AUC of nafenopin protein acylation correlated linearly with the AUC of
nafenopin-CoA formation.161 Conjugation of CoA thioesters with GSH has
also been discerned.161
As has been previously demonstrated for acyl glucuronides, the substi-
tution pattern around the acyl CoA metabolites greatly influences chemical
reactivity; increasing substitution at the a-carbon generally correlates with a
decrease in reactivity with nucleophiles.162 Alongside acyl glucuronides,
covalent adduction of acyl CoA metabolites of NSAIDs such as zomepirac
 H
O N COOH
H3 C CH3
S
O NH
O COOH
O
NH2
GSH
H

CoA-SH

H3 C CH3
Carboxylic Acids and their Bioisosteres

Acyl-glutathione conjugate
O COOH
H

Acyl CoA NH2

N H3 C CH3
Synthetase O OH O O
H N Nu
H H O Protein
O N N P P N
S O O O O N O
OH OH
H3 C CH3 O O H3 C CH3 CoA-SH
O O
Nafenopin OH
P OH
Nafenopin acyl CoA thioester H
HO Protein-NuH

Scheme 3.7 Acyl CoA thioester metabolites of carboxylic acid derivatives as electrophiles.
133
134 Chapter 3
and tolmetin with proteins and GSH has been proposed as a mechanism
for the idiosyncratic immune-mediated toxicity associated with these
drugs.157–160
A remarkable feature in the metabolism of NSAIDs such as ibuprofen
(see Figure 3.3) is the unidirectional chiral inversion from the pharmaco-
logically inactive (R)- to the active (S)-enantiomer. Such inversion has been
documented in several in vivo studies with 2-arylpropionic acid-based drugs
and xenobiotics.163 The mechanism of enantioselective inversion is believed
to involve the initial enantioselective formation of the acyl CoA thioester
followed by epimerisation by 2-arylpropionyl-CoA epimerase (this involves
the intermediacy of a symmetrical conjugated enolate anion), followed by
hydrolysis to regenerate the free acids. For each 2-arylpropionic acid drug
studied, almost no acyl CoA formation is observed for the S-enantiomers,
while the respective acyl CoA thioester derivatives are readily detected for
most R-enantiomers. The enantioselective covalent binding of the acyl
CoA thioester of R-2-phenylpropionic acid to hepatic tissue has been also
demonstrated.164

3.8.3.7 b-Oxidation of Carboxylic Acids

b-Oxidation is the process by which fatty acids are broken down in mito-
chondria and/or in peroxisomes by stepwise oxidation of the carbon chain (two
carbons for each cycle) to generate acetyl-CoA, the entry molecule for the
Krebs cycle.165 Mechanistically, b-oxidation comprises of an initial CoASH-
dependent activation of the carboxylate moiety to afford the CoA thioester
intermediate 3.20 (Scheme 3.8). Acyl-CoA-dehydrogenase mediated dehy-
drogenation (at the C2 and C3 carbons in 3.20) yields olefin 3.31, which
undergoes a stereospecific hydration at the double bond to afford the corre-
sponding L-b-hydroxyacyl CoA intermediate 3.22. Oxidation of the alcohol
group in 3.22 by NAD1 generates the b-ketoacyl CoA derivative 3.23, which
undergoes cleavage at the a,b-bond by the thiol group of another molecule of
CoA in a reaction catalysed by b-ketothiolase. This biochemical reaction results
in the formation of one molecule of acetyl-CoA and one molecule of the acyl
CoA derivative 3.24. Hydrolysis of thioester bond in 3.24 yields the carboxylic
acid metabolite 3.25, which is two carbons shorter.165

3.8.3.8 b-Oxidation of Statins

The dihydroxyheptanoic or heptanoic acid side chain in statins is particularly
prone to b-oxidation. Pentanoic acid derivatives of simvastatin and lovastatin,
corresponding to the loss of a two-carbon unit from the dihydroxy hepatanoic
acid side chain, have been reported to occur exclusively in rodents following the
administration of the lactone form of the statin derivatives.166,167 Carboxylic
acid metabolites shortened by two or four carbon units, resulting in pentanoate
or propionate derivatives, respectively, have been observed in vivo for the
 O O O NADH
3 FAD FADH2 OH O NAD
Acyl CoA CoA
+ H2O +H
CoA
OH Synthetase 2
S S CoA
Acyl CoA - H2O S
3.20 Dehydrogenase 3.21 3.22
Carboxylic Acids and their Bioisosteres

O O O O
+ CoA-SH,Thiolase H2O
CoA CoA
S S OH
O
3.23 3.24 3.25
CoA
H3C S
Acetyl-CoA

Scheme 3.8 b-Oxidation of fatty acids.

135
 F F
136
HO
COOH
COOH
OH
O
O

O O
H H N H N
H N OH N
OH
O O COOH
COOH
Lovastatin free hydroxy acid Atorvastatin

N
F COOH
F

OH OH N OH
OH
COOH
COOH
COOH
H3CO OH
H3CO F

N
N
F N
Cerivastatin Fluvastatin
COOH

Figure 3.21 b-Oxidation of statins.

Chapter 3
Carboxylic Acids and their Bioisosteres 137
dihydroxyheptanoic acid derivatives atorvastatin and pravastatin primarily in
rodents and minimally in humans (Figure 3.21).168–170 Analogous metabolites
have also been described for cerivastatin and fluvastatin, both of which contain
the dihydroxyheptanoic acid moiety (see Figure 3.21).171,172 Statins that form
propionic acid metabolites (loss of four carbon unit) are believed to undergo
two cycles of b-oxidation.172
Furthermore, since all statins possess a D-b-hydroxy configuration, an epi-
merisation to the L-configuration is needed for the b-oxidation cycle to occur.
The mechanism(s) for the formation of unsubstituted pentanoic acid metabo-
lites of statins proceeds via the initial b-oxidation cycle, which yields a D-b-
hydroxypentanoic acid derivative, followed by fatty acid biosynthetic processes
involving dehydration of the remaining D-hydroxyl group, followed by
hydrogenation to form the unsubstituted pentanoic acid metabolites as shown
in Scheme 3.9 for simvastatin free acid.167,173

3.8.3.9 b-Oxidation of Valproic Acid to Reactive Metabolites

Valproic acid (Scheme 3.10) is an anticonvulsant agent first introduced in
France in 1967 for the treatment of epilepsy.174 Although, valproic acid has
been shown to be effective against a broad spectrum of seizure types, its
usage has been associated with a rare but serious effect involving irrever-
sible liver failure (usually characterised by hepatic steatosis with or without
necrosis).175–178 The biochemical mechanisms that underlie valproic acid
hepatotoxicity are not clearly understood, although a number of hypotheses
have been advanced, including a role for toxic valproate metabolites. The
involvement of toxic and potentially reactive metabolites was first suggested
by Gerber et al. based on structural analogy with the known hepatotoxin
4-pentenoic acid, which is associated with mitochondrial damage and impair-
ment of fatty acid oxidation.179 The line of reasoning was developed further
by Zimmerman and Ishak, who proposed that the terminal olefin metabolite of
valproic acid [i.e. D4-valproic acid (3.26), Scheme 3.10], might be the respon-
sible hepatotoxin.180 Interestingly, 3.26 was first detected as a minor meta-
bolite in the plasma of epileptic children receiving valproic acid; much higher
levels were detected in the serum of paediatric patients (the group most
susceptible to valproic acid induced liver injury) than in either youths or
adults.181
In the case of 4-pentenoic acid, b-oxidation is believed to lead to 3-oxo-4-
pentenoyl-CoA, a reactive, electrophilic species that is proposed to alkylate 3-
ketoacyl-CoA-thiolase, the terminal enzyme for b-oxidation, resulting in the
mechanism-based inactivation of this enzyme complex.182 Studies on the
metabolism of D4-valproic acid (3.26) in perfused rat liver or in primates indeed
revealed products of b-oxidation as illustrated in Scheme 3.10.183–185 Following
conversion of 3.26 to its CoA derivative, sequential steps of b-oxidation lead to
2(E)- D2,4-valproic acid (3.27), 3-hydroxy-D4-valproic acid (3.28) and 3-oxo-D4-
valproic acid (3.29) as the corresponding CoA derivatives. The diene 3.27 and
 138

HO
COOH
COOH COOH
OH COOH
O HO
O COOH COOH COOH
O CoASH, ATP OH + H2O OH NAD CoASH NADPH
OH
H OH
- H2O - H2 O

Unsubstituted
Simvastatin free hydroxy acid pentanoic acid

Scheme 3.9 Proposed mechanism for b-oxidation of statins leading to unsubstituted pentanoic acid metabolites: simvastatin as an example.
Chapter 3
 O OH O OH O S
CoA O S O S O S
CoA CoA CoA

Valproic acid 3.26 3.27 OH 3.28 O 3.29

GSH
Carboxylic Acids and their Bioisosteres

O S O OH
CoA H2 O
GS GS

F F F
HO CoA-SH S S
P450 HO
CoA CoA
O O O O
3.30 3.31 3.32 3.33

Scheme 3.10 b-Oxidation of valproic to reactive, electrophilic intermediates: effect of 2-fluorine substitution.
139
140 Chapter 3
the allylic alcohol 3.28 intermediates (free acid forms) have been identified as
metabolites in perfused rat liver and in primates in vivo. The 3-oxo-D4-valproic
acid (3.29) is believed to be the reactive, electrophilic species that binds cova-
lently to the ketoacylthiolase protein resulting in its inactivation, while adducts
derived from the reaction of GSH and N-acetylcysteine with the diene 3.27 in
preclinical species and humans suggest a role for this reactive metabolite in the
hepatotoxic event.186–189
Of much interest in this aspect is the finding that substitution of the methine
hydrogen atoms on the C2 position in valproic acid with a fluorine atom yields
a non-heptatotoxic compound 3.30 that retains anticonvulsant activity of the
parent drug in mice.190 The fluorine atom prevents oxidation of the 4-ene-2-
fluoro valproic acid CoA intermediate 3.32 to the diene 3.33,191 although
subsequent experimental work argued that failure to form the acyl CoA
intermediate 3.32 prevents 4-ene-2-fluoro valproate from undergoing b-oxida-
tion (see Scheme 3.10).192

3.8.4 P450 Isozymes Involved in the Oxidative Metabolism of

Carboxylic Acid Derivatives
P4502C9 exhibits selectivity for the oxidation of relatively small and struc-
turally diverse lipophilic carboxylic acid derivatives such as NSAIDs, fibrates
and even some statins.193–198 The structural and molecular basis for P4502C9
selectivity for carboxylic acids is evident in the P4502C9 structure co-
complexed with the NSAID flurbiprofen, which is known to undergo 4 0 -
hydroxylation by the isozyme.199 The co-crystal structure highlights the
importance of several amino acid residues in the binding site that are likely to
be important for binding lipophilic carboxylic acids. In particular, the Arg108,
Asn289 and Asp293 residues were flagged as potential substrate recognition
moieties; data which are in agreement with site-directed mutagenesis studies on
P4502C9 catalysed oxidations.200–202 Some of these residues have also been
noted as important from in the various in silico approaches,203–206 highlighting
the utility of the pharmacophore-based models to predict P450 oxidation in
general.
The gene encoding for P4502C9 carries numerous inherited polymorphisms.
Those coding for R144C (*2) and I359L (*3) amino acid substitutions have
both significant functional effects and appreciable high population fre-
quencies.207,208 Consequently, drugs that are metabolised by P4502C9 are
prone to considerable interindividual variability in pharmacokinetics. For
example, mean CLp in homozygous carriers of the *3 allele were below 25% of
that of the wild-type for several P4502C9 substrates including warfarin, tol-
butamide, glipizide and fluvastatin.207,208 It is of interest to note that P4502C9
is not the exclusive P450 isoform responsible for the oxidative metabolism of
carboxylic acids. Highly lipophilic carboxylic acids such as the statins and/or
some PPAR-a agonists also tend to be undergo oxidative metabolism by
P4503A4 (in addition to P4502C9).209,210
Carboxylic Acids and their Bioisosteres 141
3.8.5 Hepatobiliary Disposition of Carboxylic Acids
Over the past 15 years, a number of important human drug transporters have
been identified that are expressed at the apical or basolateral side of the epi-
thelial cells in various tissues. Most drug transport proteins, which catalyse
cellular uptake and efflux belong to two super-families namely the SLC (solute-
linked carrier) and the ABC (ATP-binding cassette) transporters, respec-
tively.211,212 The combination of organic anion transporting polypeptides
(OATPs) and multidrug resistance-associated protein 2 (MRP2), which repre-
sents two classes of transporters from the SLC and ABC super-family,
respectively, play an important role in the hepatobiliary transport of organic
anions including carboxylic acid derivatives at the sinusoidal and canalicular
membranes. In the human liver, OATP1B1 (also known as OATP2 or
OATPC), OATP1B3 (OATP8) and OATP2B1 (OATPB) are predominant
transporters responsible for the hepatic uptake of a variety of organic anionic
compounds.213 Once taken up into hepatocytes, anionic compounds and/or
metabolites derived from phase II glucuronidation can undergo MRP2-
mediated biliary excretion.214 Besides MRP2, multidrug resistance 1 (MDR1,
P-glycoprotein) protein and breast cancer resistance protein (BCRP), which are
located on the bile canalicular membrane of the liver, can also be involved in
the active efflux of organic anions into bile.215 Uptake via OATPs followed by
excretion via MRP or other efflux transport proteins from the ABC family
constitutes vectorial transport for the hepatobiliary excretion of several car-
boxylic acid based drugs.
An example of this phenomenon is evident with the selective histamine
H1-receptor antagonist and carboxylic acid derivative fexofenadine (see
Scheme 3.2). Hepatic metabolism is of minimal importance in the elimination
of fexofenadine in rodents and human; this lipophilic carboxylic acid is pre-
dominantly eliminated via biliary excretion in the unchanged form.85 Biliary
excretion of fexofenadine in humans is mediated by MDR1.216 In addition,
fexofenadine also functions as a substrate for hepatic uptake by human and rat
OATPs.216,217 In the case of fexofenadine, inhibition/stimulation of these
uptake/efflux processes is known to lead to significant DDIs in humans.218,219
The role of transporters in the disposition of a carboxylic acid containing
phosphodiesterase-4 inhibitor CP-671 305 (Figure 3.22) has also been examined
in great detail.220 Like fexofenadine, CP-671 305 is resistant to metabolism by
either phase I or phase II drug metabolising enzymes in liver microsomes and
hepatocytes from preclinical species and human; these findings are supported
by the lack of detectable metabolites in pooled plasma, urine and/or bile from
rats, dogs and monkeys following CP-671 305 administration.221
Preliminary investigations into the clearance mechanism in rats revealed that
the compound undergoes substantial biliary excretion in the unchanged
form.221 In bile duct exteriorised rats, a 7.4-fold decrease in the half-life of CP-
671 305 was observed implicating enterohepatic biliary circulation of the parent
drug. A statistically significant difference in CP-671 305 pharmacokinetics was
also discernible in cyclosporin A- or rifampicin-pretreated rats as reflected from
142 Chapter 3

O F

N CH3
H
N O O COOH

O
O

CP-671,305

Figure 3.22 Chemical structure of a novel PDE4 inhibitor and carboxylate analog
which undergoes active hepatobiliary transport.

a four-fold decrease in clearance and a four-fold increase in the area under the
curve (AUC).220 Given the ability of cyclosporin A and rifampicin to inhibit
multiple drug transporters, the interactions of CP-671 305 with the major
human hepatic drug transporters, MDR1, MRP2, BCRP and OATPs, were
evaluated in vitro.220 CP-671 305 was identified as a substrate of MRP2 and
BCRP, but not MDR1. CP-671 305 was a high affinity substrate of human
OATP2B1, but not a substrate for human OATP1B1 or OATP1B3. Exam-
ination of the hepatobiliary transport of CP-671 305 in sandwich-cultured
hepatocytes indicated active uptake into hepatocytes followed by efflux into
bile canaliculi, consistent with the results from in vitro transporter studies. The
role of rat Mrp2 in the biliary excretion was also examined in TR (Mrp2-
deficient) rats, and the observations that CP-671 305 pharmacokinetics were
largely unaltered in TR rats were consistent with the finding that compro-
mised biliary clearance of CP-671 305 was compensated by increased urinary
clearance.220 As such, these in vitro and in vivo studies, which suggest an
important role for transport proteins in the hepatobiliary disposition of CP-671
305 in rat and human, could be valuable in the design of clinical DDI studies.

3.8.5.1 Hepatobiliary Transport of Statins

Statins are avidly taken into hepatocytes by active uptake transporters, among
which OATP1B1 appears to be the most important.213,222,223 Other hepatic
uptake transporters that can transport statins are OATP1B3, OATP2B1,
OATP1A2 and sodium-dependent taurocholate co-transporting polypeptide
(NTCP). The carrier-mediated hepatic uptake process not only represents the
first step of hepatic drug elimination, but is also an active drug delivery system
for many statins to the liver as a target organ.224 As all statins are eliminated
mainly by the liver, their active hepatic uptake, metabolism by P450 isozymes
and biliary excretion via MDR1, MRP2, BCRP and bile salt export pump
(BSEP) can regulate their total clearance.213,223,225
Carboxylic Acids and their Bioisosteres 143
To treat patients with dyslipidemias resistant to diet or single agent phar-
macotherapy, combination therapies of statins with other drugs are widely
used, which can result in DDIs. An increase in the plasma concentration of
many of the currently used statins can cause severe side effects such as muscle
toxicity or even rhabdomyolysis. Cerivastatin was withdrawn from the market
after combinations of the statin derivative with the fibrate gemfibrozil226,227 or
the immunosuppressant cyclosporin A228 led to marked increases in its systemic
exposure resulting in severe muscle toxicity.
The observed DDIs are thought to arise from the inhibition of P4502C8-
mediated cerivastatin oxidative metabolism by gemfibrozil229,230 and/or an
inhibition of the OATP1B1-mediated hepatic statin uptake into the liver by
gemfibrozil and its glucuronide conjugate.230,231 In addition, daily use of
gemfibrozil also increases the AUC of simvastatin and lovastatin free acid
derivatives and pravastatin by 2–3 fold.232,233 It is interesting to note that
neither fenofibrate nor bezafibrate increase the AUC of simvastatin, lovastatin
or pravastatin, which indicates that DDI with statins is not a group effect of the
fibric acid derivatives.234,235 Apart from DDI with ceravastatin, cyclosporin A,
which inhibits numerous membrane transporters including OATPs and MDR1,
also increases the AUCs of simvastatin, lovastatin and pravastatin about
10-fold.236–238 Cases of rhabdomyolosis have occurred during concomitant use
of cyclosporin A and different statins.239–241 Polymorphisms of SLCO1B1
(encoding OATP1B1) and ABCC2 (encoding MRP2) can cause considerable
interindividual variability in plasma concentrations of statins.242–247 Common
variants in SLCO1B1 have also been associated with an increased risk of
myopathy with certain statins such as simvastatin.246,247

3.9 ADME Profile of Tetrazoles

Structural diversity within the tetrazole-based AT1 receptor antagonists results
in subtle differences in physiochemical properties,248 which then influences: (a)
binding affinity to the AT1 receptor, (b) aqueous and lipid solubility, and (c)
pharmacokinetic profile (e.g. absorption, distribution, clearance and routes of
elimination). Table 3.5 lists key physiochemical attributes of representative
members of the tetrazole-based AT1 receptor antagonists.
The acidic nature of the tetrazole moiety can provide a convenient means for
preparation of salt forms to improve aqueous solubility (e.g. losartan is

Table 3.5 Some key physiochemical properties of tetrazole-based AT1

receptor antagonists.
AT1 receptor blocker MW cLogP Log D7.4 fu (% bound) Vdss (L/kg)

Losartan 422 4.1 1.7 98.6–98.8 0.49

Irbesartan 428 6.0 1.8 90 0.75–1.3
Candesartan 440 5.4 0.79 499 0.12
Valsartan 435 4.9 1.3 94–97 0.24
144 Chapter 3
commercially sold as the potassium salt). Candesartan and olmesartan, which
contain a carboxylic acid group (in addition to the presence of the tetrazole
motif), are administered as the corresponding carboxylic acid ester derivatives
to increase oral absorption (see Figure 3.17). Candesartan cilexetil (prodrug
form) is rapidly and completely activated via ester hydrolysis during absorption
from the gastrointestinal tract to active candesartan. Even upon administration
in the prodrug form, the absolute oral bioavailability of candesartan is low
(B15%), a likely characteristic of its high polarity despite the ‘prodrug’ handle.
Consistent with this hypothesis, food with a high fat content does not have any
effect on candesartan absorption or bioavailability.
Because of the bioisosteric relationship, the tissue distribution pattern is
similar for drugs that contain the either the carboxylate or tetrazole func-
tionalities. For example, tetrazoles, like their carboxylic acid counterparts, are
heavily bound to plasma proteins (mainly to albumin; a small proportion of
binding to a1-acid glycoprotein has also been noted) (see Table 3.5). Overall,
the extensive plasma protein binding is also reflected in a low Vdss for the
tetrazole AT1 receptor antagonists in humans.249–253 Furthermore, as seen in
Table 3.6, subtle differences in log D also impacts CLp and T1/2 of the tetrazole-
based AT1 receptor antagonists in humans.249
The elimination pathways of tetrazole-based AT1 receptor antagonists
involve phase I/II metabolism and/or non-metabolic (biliary and urinary)
excretion, a phenomenon that has some commonality with the excretion pat-
tern discerned with carboxylic acid containing drugs.254–258 For example, mass
balance studies on irbesartan in humans reveal that B9% of the orally admi-
nistered dose is metabolised via oxidative and conjugation pathways catalysed
by P450 and UGT isozymes; B80% of the administered dose is excreted in the
faeces via the bile and the remainder of the dose appears in the urine.254 In the
case of losartan, oxidative metabolism of its primary alcohol motif by P450
results in the formation of the active carboxylic acid metabolite EXP3174.
After oral administration to humans, B14% of the losartan dose is converted
to EXP3174; faecal and renal elimination account for the remainder of the
losartan dose. EXP3174 is B10- to 40-fold more potent than the parent
compound and it appears that most of the in vivo pharmacological activity of
losartan is derived from the metabolite.

Table 3.6 Human pharmacokinetics of tetrazole-based AT1 receptor

antagonists.
AT1 receptor blocker Dose (mg) CLpa T1/2 (h) Tmax (h) F (%)

Losartan 50–100 8.6 1.5–2.0 1.0 33

Irbesartan 150–300 2.2 13 0.3 60–82
Candesartanb 4–32 0.4 6.0–13 2.0–5.0 15
Valsartan 80–320 0.5 6.0–10 2.0–4.0 19 (10–35)
mL min1 kg1.
a
b
Administered as the ester prodrug (candesartan cilexetil).
Carboxylic Acids and their Bioisosteres 145

Scheme 3.11 N-Glucuronidation of tetrazoles.

3.9.1 Metabolism of the Tetrazole Motif

In contrast to carboxylic acids, tetrazoles are resistant to metabolic pathways
involving b-oxidation and amino acid conjugation. However, b-N-glucur-
onidation has been shown to be an important clearance pathway of tetrazole-
containing compounds in a similar as that discerned with the carboxylic acid
moiety. The biotransformation reaction results in the formation of O-b-glu-
curonides. Tetrazole glucuronidation can occur on the N-1 or the N-2 nitrogen
as shown in Scheme 3.11. Nohara and co-workers were the first to identify a
tetrazole-N-1-glucuronide (3.35) in urine of animals dosed with 6-ethyl-3-
(1H-tetrazol-5-yl)chromone (3.34) (Scheme 3.11).259 Glucuronide 3.35 was
identified as the exclusive isomer by chemical synthesis and NMR studies.
Recent studies with biphenyltetrazole derivatives, however, have indicated the
N-2-glucuronide conjugate to be the preferred metabolite over the N-1-glu-
curonide based on NMR and X-ray crystal structure characterisation.260,261
In fact, several N2-tetrazole glucuronide conjugates have been reported as
metabolites of tetrazole-based angiotensin II receptor antagonists losartan,
irbesartan, candesartan and zolarsartan in animals and/or humans.262–266
The optimal pH value for the transformation of tetrazoles to their respective
N2-glucuronide conjugates correlates very well with the reported pKa of tetra-
zole (pKa ¼ 4.9).267 Recent studies have also shown that UGT1A3 is highly
selective towards tetrazole-N2 glucuronidation in losartan, candesartan and
zolarsartan.268

3.9.1.1 Role of P4502C9 in the Oxidative Metabolism of

Tetrazole Derivatives
Considering the bioisosteric relationship between the carboxylic acid group and
the tetrazolyl moiety, it is not surprising that the polymorphic P4502C9 is also
146 Chapter 3
involved in the oxidative metabolism of many of the tetrazole-based AT1
receptor antagonists. This attribute often results in pharmacokinetic interac-
tions with other P4502C9 substrates or inhibitors in a manner analogous to the
situation with carboxylate-containing drugs.249,269 The substituent attached to
the biphenyl tetrazolyl scaffold in these compounds is usually the site of oxi-
dation. As described earlier, the conversion of losartan to its pharmacologically
active metabolite EXP3174 is principally mediated by the action of P4502C9
(with some contribution from P4503A4).112,270,271 Variability in losartan
metabolism to EXP3174 has been discerned in individuals with different
P4502C9 genotypes.272,273 P4502C9 also catalyses the oxidative metabolism of
irbesartan, candesartan and valsartan to some degree; the metabolism usually
results in weakly active or inactive compounds.274–276 Telmisartan and olme-
sartan are generally resistant to oxidation by P450 isoforms; telmisartan is
partially metabolised by glucuronidation, and olmesartan is excreted unchan-
ged.277,278 The ability of tetrazole-based AT1 receptor antagonist to inhibit
P4502C9 in human liver microsomes has also been examined; all compounds
were shown to possess weak inhibitory activity (IC50 430 mM) against
P4502C9 catalysed warfarin hydroxylation.279

3.9.2 Role of Transporters in the Disposition of Tetrazole-based

Angiotensin II Receptor Antagonists
The hepatobiliary and urinary excretion of tetrazole-based AT1 receptor
antagonists can be subject to active transport. The findings that olmesartan is
excreted in both bile and urine in the unchanged form has led to studies aimed
at characterising the role of active transporters in olmesartan disposition. On
the basis of in vitro studies as well as in vivo studies in Eisai hyperbilirubinemic
rats, a role for MRP-2 in olmesartan biliary has been established.280–282
Olmesartan was also shown to function as a OATP1B1, OATP1B3, organic
anion transporter (OAT) 1 and OAT3 substrate.281,282 Likewise, Yamashiro
et al. have shown the role of OATP1B1, OATP1B3 and MRP-2 in the hepa-
tobiliary transport of valsartan.283 Consistent with these findings, interindividual
variability in valsartan pharmacokinetics in human subjects with OATP1B1*1b
alleles has been noted.284 Finally, the finding that OATP1B3 is also involved in
the hepatic uptake of the carboxylic acid based angiotensin antagonist telmi-
sartan285 serves to illustrate the bioisosteric relationship between the carboxy-
late and the tetrazole group in terms of affinity towards active transport.

3.10 ADME Profile of Thiazolidinedione Derivatives

The thiazolidinedione (TZD) class of compounds—collectively referred to as
the ‘glitazones’—are PPAR g agonists used for the treatment of type 2 diabetes.
The thiazolidinedione group shares a non-classical biosisosteric relationship
with the carboxylic acid group owing to the acidic nature of the imide fragment
Carboxylic Acids and their Bioisosteres 147

Figure 3.23 Structures of thiazolidinedione-based anti-diabetic drugs.

in the 5-membered thiazolidinedione ring system. The first commercialised

TZD drug, troglitazone (Figure 3.23) was withdrawn from the US market after
numerous reported cases of severe liver failures leading to liver transplantation
or death; after being on the market for 17 months, the FDA received 560
reports of hepatotoxicity and 24 cases of acute liver failure.286–288 In contrast,
the related TZD analogs rosiglitazone and pioglitazone (Figure 3.23) are
devoid of the hepatotoxicity associated with troglitazone.

3.10.1 Clearance and Oral Bioavailability

Because they are non-classical bioisosteres of the carboxylic acid functionality,
their pharmacokinetic parameters are fairly similar to those discerned with
carboxylate-based drugs. For example, the human CLp, Vdss, plasma fu and
elimination T1/2 of rosiglitazone is 0.65 mL min1 kg1, 0.20 L kg1, 0.002 and
3.9 h, respectively.127 The oral bioavailability of rosiglitazone is nearly 100%
and that of pioglitazone 480%.289,290

3.10.2 Metabolism of the Thiazolidinedione Ring System

The thiazolidinedione ring is susceptible towards oxidative ring scission. The
observation was first noted during bioactivation studies on troglitazone.291,292
In vitro incubations of troglitazone in NADPH- and GSH-supplemented
human liver microsomes led to the characterisation of several GSH conjugates.
Based on these structures of these conjugates, the two proposed pathways for
the bioactivation of troglitazone include metabolism of the chromane ring to
the quinone or ortho-quinonemethide (Scheme 3.12, panels A and B) and
oxidative cleavage of thiazolidenedione ring (Scheme 3.12, panel C). While
rosiglitazone and pioglitazone do not contain the chromane ring system found
in troglitazone, they do contain the thiazolidinedione scaffold. And consistent
with the findings with troglitazone, both rosiglitazone and pioglitazone have
been shown to undergo thiazolidinedione ring scission mediated by P450
enzyme(s) in human microsomes resulting in reactive metabolites trapped by
GSH.293
The mechanisms of troglitazone-induced hepatotoxicity remain unclear at
the present time and seem to be multi-factorial. Besides differences in meta-
bolism (absence of quinonoid formation in rosiglitazone and pioglitazone), a
comparison of the effects of TZD drugs on toxicologically relevant gene
expression in primary culture hepatocytes using microarray analysis showed
148 Chapter 3
CH3 CH3
CH3 CH3
H3C O H3C O
GSH
P450 O HO
A CH2 H2C
SG
quinonemethide
S
CH3 O CH3 OH CH3
CH3 CH3 H3C O O
H3C O NH H3C O
O O P450 OH
O O
HO B O
CH3 CH3 CH3

Troglitazone
OH SG
O S S
P450 GSH, H2O
S N C O NH2
C O
NH O O
O CO2
GSH, [O]

O OH
S
H
N SG

O O

Scheme 3.12 Bioactivation of the thiazolidinedione ring system in ‘glitazones’ to

electrophilic intermediates.

that substantially higher numbers of genes were affected by troglitazone

treatment when compared to rosiglitazone and pioglitazone.294 Masubuchi
et al.295 have also investigated the effects of troglitazone, rosiglitazone, and
pioglitazone on mitochondrial function. Troglitazone, but not rosiglitazone or
pioglitazone, was shown to induce decreases in mitochondrial membrane
potential and mitochondrial Ca21 accumulation consistent with the induction
of mitochondrial permeability transition.

3.10.3 P450 isozymes Responsible for the Metabolism of

‘glitazones’–DDI Potential
P4502C8 and P4503A4 are the major P450 isozymes, which catalyse the oxi-
dative biotransformation of troglitazone and pioglitazone, whereas rosiglita-
zone is metabolised by P4502C9 and P4502C8.296–298 The major oxidative
pathways of ‘glitazones’, however, do not involve thiazolidinedione ring
scission. For instance, N-demethylation and pyridine ring hydroxylation are
the principal metabolic pathways of rosiglitazone. For both rosiglitazone
and pioglitazone, the most relevant clinical pharmacokinetic interactions
have been described in healthy volunteers with rifampicin (rifampin), which
results in a significant decrease of AUC (54–65% for rosiglitazone; 54% for
pioglitazone), and with gemfibrozil, which results in a significant increase of
AUC (130% for rosiglitazone; 220–240% for pioglitazone).299–302 As noted
earlier with carboxylic acids, DDIs between ‘glitazones’ and gemfibrozil stems
from inhibition of the P4502C8 enzyme by gemfibrozil and its glucuronide
conjugate.303
Carboxylic Acids and their Bioisosteres 149
3.11 ADME Profile of Esters and Amides
Unlike carboxylic acids, esters and amides are neutral and more lipophilic in
nature, which results in vastly different pharmacokinetic parameters for com-
pounds containing these functional groups. Consequently, it can be debated
whether esters/amides are true bioisosteres of the carboxylic acid group. Apart
from the role of esters as carboxylic acid prodrugs, the utility of this functional
group as a carboxylic acid replacement is limited. This is primarily due to the
facile in vivo hydrolysis of the ester functional group by esterases in the gut,
liver and plasma, resulting in less than optimal pharmacokinetic parameters.
For example, the half-life for deacetylation of the acetoxy group present in
aspirin is B2 h in human plasma, and reduces to 1–3 minutes for carboxylic
acid ester derivatives of aspirin.304
Our recent study on the disposition of the non-selective COX inhibitor
indomethacin and its corresponding COX-2-selective neutral amides deriva-
tives 3.36–3.38 (Figure 3.24) in the rat provides a comparison of the ADME
profile of the two functional groups.305,306 Unlike the parent carboxylic acid
derivative, which is resistant to metabolic turnover in liver microsomes from rat
and human, neutral amide derivatives 3.36–3.38 were considerably less stable in
this biological matrix and underwent extensive oxidative metabolism by P450, a

O-demethylation
(P4502C9) Glucuronidation
O

OH
H3CO
O-demethylation
CH3 (P4503A4) O
N

N R
O Cl H3CO H
CH3
Indomethacin
N

O Cl
O-demethylation Oxidation
(P4503A4) (P4503A4)
O
3.37: R = F
N
H3CO H
CH3 3.38: R = F
N N

O Cl

3.36

Figure 3.24 Differences in the metabolic fate of carboxylic acid derivative indo-
methacin and its neutral amide derivatives.
150 Chapter 3
Table 3.7 Physiochemical and pharmacokinetic parameter comparison
between indomethacin and some of its neutral amide derivatives.
T1/2 (min)a
Fraction unbound CLp (mL/ Vdss
Compound MW cLogP (rat plasma) Rat Human min/kg) (L/kg) F (%)
Indomethacin 357 4.18 0.03 490 490 0.51 0.19 98
Amide 3.36 460 5.50 0.0045 1.0 1.5 155 5.3 o1
Amide 3.37 450 5.85 0.0040 8.5 25 26 2.3 20
Amide 3.38 451 5.05 0.0070 8.0 23 39 2.0 38
a
Liver microsomes.

HO
B OH
H2N OH
NH
O O H2N B H2N N
OH
N COOH O B
HO OH

Thrombin γ-Glutamyl transpeptidase DPP4

O
OH H3CO
R B
OH
NH2 O OH
dative H H3CO
bond HO S OH N N B
B N OH OCH3 OH
H B
O OH O
N CH3O OH

Arginase Proteasome α-Tubulin

Figure 3.25 Boronic acid-containing pharmacologic active compounds.

feature that is linked to the increase in molecular weight and lipophilic char-
acter (Table 3.7). As such, the increase in molecular weight and lipophilicity
also results in a decrease in plasma free fraction (see Table 3.7). A comparison
of the pharmacokinetic attributes of indomethacin and neutral amides 3.36–
3.38 is depicted in Table 3.7 and reveals a dramatic difference in CLp and Vdss
between the parent carboxylic acid and its amide derivatives; the neutral amide
analogs are cleared at a more rapid rate than the free carboxylic acid derivative,
a phenomenon that ultimately results in a lower oral bioavailability for the
amides. Unlike indomethacin, which is susceptible to glucuronidation and
P4502C9-catalysed O-demethylation, indomethacin amides 3.36–3.38 under-
went oxidative O-demethylation and oxidation on the amide substituent, which
was almost exclusively mediated by P4503A4 (Figure 3.24).305,306

3.12 Boronic Acid Derivatives

The boronic acid is a functional group with dual character. In addition to
delivering the electronic and hydrogen-bonding requirements of carboxylate
isostere (Figure 3.25, a-tubulin inhibitor), the boronate is also effective as a
Carboxylic Acids and their Bioisosteres 151

O OH O
H Human H
N N B Microsomes N N OH
N OH N
H H
O or r CYP O
N N

bortezomib deboronated
metabolites

Scheme 3.13 Biotransformation of the boronic acid group in bortezomib.

pharmacophore in itself. The vacant p-orbital of boron is amenable to dative

bonding with the nucleophilic active site residues of proteolytic enzymes
such as those bearing serine (e.g. thrombin, arginase, DPP4) and threonine
(e.g. proteasome, g-glutamyl transpeptidase).307–309 This dative bonding
capability of boron results in enzyme inhibition kinetics that are nearly indis-
tinguishable from irreversible inhibitors, a hallmark trait of a transition state
analog (Figure 3.25, inset).
Although 20 years of drug discovery had elapsed in the arena of boronate-
mediated inhibition of proteolysis, little had been reported regarding the bio-
transformation pathways which characterise boronate metabolism and dis-
position in humans. More recently, the discovery and clinical development of
the peptidyl boronate proteasome inhibitor, bortezomib, has resulted in the
characterisation of the hepatic metabolism of this functional group. And while
the metabolism of bortezomib did not proceed via prototypic carboxylic acid
biotransformation pathways (e.g. glucuronidation), the boronate was shown to
be equally labile to oxidation. Consistent with the in vitro metabolism appraisal
in human liver microsomes, the principal route of biotransformation of bor-
tezomib in patients was deboronation, the result of which was formation of a
pair of carbinolamide metabolites (Scheme 3.13).310 Multiple P450 enzymes
including P450 1A2, 2C9, 2C19, 2D6 and 3A4 catalysed deboronation in
bortezomib.310 A subsequent investigation demonstrated that the deboronation
reaction in human liver microsomes involved multiple oxidants, including both
reactive oxygen species (e.g. O2d) and specific activated-enzyme oxidants (e.g.
peroxo-iron), both generated during the CYP catalytic cycle.311

3.13 Concluding Remarks: Carboxylic Acid and Drug

Safety
From a drug discovery perspective, the presence of the carboxylic acid func-
tionality in drug candidates provides an interesting topic for debate. Carboxylic
acid containing drugs (e.g. NSAIDs, b-lactam antibacterials and statins) have
revolutionised drug discovery in the 20th century, yet a number of carboxylic
acid containing drugs have also been withdrawn from the market due to rare
but serious adverse reactions. Of 29 drugs withdrawn from the market in the
152 Chapter 3
UK, Spain or USA between 1974 and 1993, nine were carboxylic acid con-
taining drugs, making this compound class the most frequently involved in
drug discontinuations in this period.312 In a recent review by Fung et al., it was
found that of the 121 prescription drugs withdrawn worldwide between 1960
and 1999, 17 were carboxylic acid containing drugs.313
Many of the carboxylic acid containing drugs that have been associated with
toxicity (idiosyncratic or otherwise) belong to NSAID class. As NSAIDs are
some of the most used prescription and over-the-counter drugs, the number of
patients exposed to these drugs on a daily basis may be part of the explanation
why so many adverse drug reactions, including idiosyncratic ones, have been
observed with this therapeutic class. Some of the more prominent cases of
NSAIDs that have been withdrawn post-marketing include ibufenac and
benoxaprofen (Figure 3.26); this was due to incidents of overt liver toxicity, and
zomepirac, which caused anaphylactic shocks. Additional examples of car-
boxylic acid drugs that have been withdrawn after introduction to the market,
or have otherwise been associated with liver toxicity, are shown in Figure 3.26.
In many instances, bioactivation of the carboxylic acids to electrophilic esters
(e.g. acylglucuronides, acyl CoA thioesters) is thought to represent the rate-
limiting step in the aetiology of idiosyncratic adverse drug reactions (IADRs).
When such metabolites react with critical proteins, cellular functionality may
be disturbed or an immune response may be induced, eliciting adverse effects
that in serious cases can be fatal. For example, reported adverse effects to
carboxylic acid containing drugs span from mild elevation in serum liver
enzymes or jaundice over skin rash and eczema to fatal anaphylactic shock.
Although it is now generally accepted that there is a link between the for-
mation of chemically reactive metabolites and a number of IADRs, the
mechanisms by which this occurs are generally not well understood. Further-
more, several carboxylic acid-based drugs (e.g. diclofenac, suprofen, zome-
pirac) also contain additional structural alerts/toxicophores (thiophene,
aniline), susceptible to P450 catalysed bioactivation.314–316 As a consequence, it
is difficult to draw conclusions on the overall contribution of carboxylic acid
bioactivation versus P450 catalysed reactive metabolite formation towards
IADR occurrence.
From an industrial perspective, there is no clear rationale for avoiding the
carboxylic acid moiety in drug design especially since mammals are exposed to
many carboxylic acid based compounds from dietary sources every day and
there are a number of safe carboxylic acid drugs on the market. Research over
the last decades has, however, revealed the bioactivation potential of the car-
boxylic acid moiety to protein-reactive metabolites, and structure–toxicity
relationships between acyl glucuronide reactivity and IADRs have been fairly
compelling in some instances (e.g. ibufenac versus ibuprofen); consequently, it
is sensible to evaluate the ability of all carboxylate drugs to form electrophilic
acyl glucuronides (especially the propensity to the glucuronide to rearrange).317
Since several factors play a role in determining which metabolites are formed
and to what extent, and since little is known about the mechanisms that govern
 NSAIDs
O

CH3 COOH
O Cl
NH2 N
O Cl
COOH
COOH N
N
H
Br Cl
Cl O COOH
Benoxaprofen Bromfenac Clometacin Diclofenac
Hepatotoxicity, cholestasis, Hepatotoxicity, hepatic necrosis Hepatitis, renal injury Hepatitis, jaundice, skin rash,
skin rash, phototoxicity anaphylactic shock
H3 C
O COOH
O CH3
COOH N
COOH
COOH
Carboxylic Acids and their Bioisosteres

Cl
S CH3
O
Zomepirac
Fenbufen Ibufenac Suprofen
Hepatotoxicity, skin rash Hepatotoxicity, hepatic necrosis Anaphylactic shock, renal failure,
jaundice Renal injury hepatitis

O
F COOH COOH
O

H N N N COOH
Cl
F
H 2N
HN H Cl
S O
F
COOH Valproic acid
Amineptine Trovafloxacin Tienilic acid
(Antidepressant) (Antibiotic) (Diuretic) (Anticonvulsant)
Hepatotoxicity Hepatotoxicity Hepatotoxicity Hepatotoxicity, blood dyscrasia
jaundice, thrombocytopenia
153

Figure 3.26 Carboxylic acid based drugs withdrawn from the market or associated with significant incidences of idiosyncratic drug toxicity.
154 Chapter 3
toxicity caused by chemically reactive metabolites, it is difficult to conclude
whether specific carboxylic acid metabolites will ultimately cause toxicity.
From a safety risk mitigation perspective, additional considerations such as
the daily dose of the drug candidate may be a pivotal factor mitigating the risks
of IADRs. Examples of low dose drugs (o50 mg day1) that cause IADRs are
rare (whether or not these agents are prone to bioactivation).317,318 Atorvas-
tatin serves as the ideal example of this phenomenon; despite biotransforma-
tion to acyl glucuronide metabolites; there have been no instance of IADRs
with this blockbuster drug, a feature that can be linked with its low daily dose.
Likewise, it is important to note the differences in daily doses of troglitazone
(200–400 mg day1) when compared with the structurally related thiazolidine-
dione derivatives rosiglitazone and pioglitazone (10–40 mg day1). This feature
may offset the bioactivation liability associated with the thiazolidinedione ring
system in general resulting in an improved safety profile for the successor
agents relative to troglitazone, which has been withdrawn due numerous cases
of fatal hepatotoxicity.

References
1. J. S. Nowick, Org. Biomol. Chem., 2006, 4, 3869.
2. J. March, in Advanced Organic Chemistry: Reactions, Mechanisms, and
Structure, ed. J. March, Wiley Interscience, New York, 1985, p. 218–236.
3. M. E. Cavet, M. West and N. L. Simmons, Br. J. Pharmacol., 1997, 121,
1567.
4. J. R. Vane, R. J. Flower and R. M. Botting, Stroke, 1990, 21(Suppl IV),
IV–12.
5. H. Dreser, Pflugers Arch., 1899, 76, 306.
6. J. R. Vane, Nature New Biol., 1971, 231, 232.
7. R. J. Flower, Pharmacol. Rev., 1974, 26, 33.
8. S. H. Ferreira, S. Moncada and J. R. Vane, Nature, 1971, 231, 237.
9. C. J. Hawkey, Lancet, 1999, 353, 307.
10. D. L. DeWitt and W. L. Smith, Proc. Natl. Acad. Sci. U. S. A., 1988, 85,
1412.
11. T. Hla and K. Neilson, Proc. Natl. Acad. Sci. U. S. A., 1992, 89, 7384.
12. J. R. Vane, J. A. Mitchell, I. Appleton, A. Tomlinson, D. Bishop-Bailey,
J. Croxtall and D. A. Willoughby, Proc. Natl. Acad. Sci. U. S. A., 1994,
91, 2046.
13. C. J. Smith, Y. Zhang, C. M. Kobolt, J. Muhammad, B. S. Zweifel,
A. Shaffer, J. J. Talley, J. L. Masferrer, K. Seibert and P. C. Isakson,
Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 13313.
14. T. D. Warner, F. Giuliano, I. Vojnovic, A. Bukasa, J. A. Mitchell and
J. R. Vane, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 7563.
15. L. J. Marnett, Annu. Rev. Pharmacol. Toxicol., 2009, 49, 265.
16. L. H. Rome and W. E. M. Lands, Proc. Natl. Acad. Sci. U. S. A., 1975, 72,
4863.
Carboxylic Acids and their Bioisosteres 155
17. R. A. Copeland, J. M. Williams, J. Giannaras, S. Nurnberg, M.
Covington, D. Pinto, S. Pick and J. M. Trzaskos, Proc. Natl. Acad. Sci.
U. S. A., 1994, 91, 11202.
18. F. J. Van Der Ouderaa, M. Buytenhek, D. H. Nugteren and D. A. Van
Dorp, Eur. J. Biochem., 1980, 109, 1.
19. G. J. Roth, E. T. Machuga and J. Ozols, Biochemistry, 1983, 22, 4672.
20. D. L. DeWitt and W. L. Smith, Proc. Natl. Acad. Sci. U. S. A., 1988, 85,
1412.
21. D. Picot, P. J. Loll and R. M. Garavito, Nature, 1994, 367, 243.
22. P. J. Loll, D. Picot and R. M. Garavito, Nature Struct. Biol., 1995, 2, 637.
23. P. J. Loll, D. Picot, O. Ekabo and R. M. Garavito, Biochemistry, 1996, 35,
7330.
24. R. G. Kurumbail, A. M. Stevens, J. K. Gierse, J. J. McDonald, R. A.
Stegeman, J. Y. Pak, D. Gildehaus, J. M. Miyashiro, T. D. Penning,
K. Seibert, P. C. Isakson and W. C. Stallings, Nature, 1996, 384, 644.
25. D. K. Bhattacharyya, M. Lecomte, C. J. Rieke, R. M. Garavito and W. L.
Smith, J. Biol. Chem., 1996, 271, 2179.
26. J. A. Mancini, D. Riendeau, J.-P. Falgueyret, P. J. Vickers and G. P.
O’Neill, J. Biol. Chem., 1995, 270, 29372.
27. H. L. Hirsh, Med. Ann. Dist. Columbia, 1948, 17, 7.
28. A. Dalhoff, Infection, 1979, 7, 294.
29. E. P. Abraham, Drugs, 1987, 34(Suppl. 2), 1.
30. W. K. Joklik, FASEB J., 1996, 10, 525.
31. T. N. Raju, Lancet, 1999, 353, 936.
32. B. L. Lignon, Semin. Pediatr. Infect. Dis., 2004, 15, 52.
33. R. Bentley, J. Ind. Microbiol. Biotechnol., 2009, 36, 775.
34. R. F. Pratt, Cell. Mol. Life Sci., 2008, 65, 2138.
35. J. A. Kelly, J. R. Knox, P. C. Moews, G. J. Hite, J. B. Bartolone, H. Zhao,
B. Boris, J. M. Frère and J. M. Ghuysen, J. Biol. Chem., 1985, 260, 6449.
36. J. A. Kelly, J. R. Knox and H. Zhao, J. Mol. Graph., 1989, 7, 87.
37. J. A. Kelly, J. R. Knox, H. Zhao, J. M. Frère and J. M. Ghuysen, J. Mol.
Biol., 1989, 209, 281.
38. A. D. Russell, Prog. Med. Chem., 1969, 6, 135.
39. N. Georgopapadakou, S. Hammerström and J. L. Strominger, Proc. Natl.
Acad. Sci. U. S. A., 1977, 74, 1009.
40. A. Marquet, J. M. Frère, J. M. Ghuysen and A. Loffet, Biochem. J., 1979,
177, 909.
41. R. R. Yocum, D. J. Waxman, J. R. Rasmussen and J. L. Strominger,
Proc. Natl. Acad. Sci. U. S. A., 1979, 76, 2730.
42. R. R. Yocum, J. R. Rasmussen and J. L. Strominger, J. Biol. Chem., 1980,
255, 3977.
43. S. O. Meroueh, G. Minasov, W. Lee, B. K. Shoichet and S. Mobasherry,
J. Am. Chem. Soc., 2003, 125, 9612.
44. H. Aoki and M. Okuhara, Annu. Rev. Microbiol., 1980, 34, 159–181.
45. J. K. Noguchi and M. A. Gill, Clin. Pharm., 1988, 7, 37.
46. B. R. Smith and J. L. LeFrock, Drug Intell. Clin. Pharm., 1985, 19, 415.
156 Chapter 3
47. G. E. Stein and M. J. Gurwith, Clin. Pharm., 1984, 3, 591.
48. D. M. Campoli-Richards and R. N. Brogden, Drugs, 1987, 33, 577.
49. J. Stamer, Arch. Surg., 1978, 113, 21.
50. M. S. Brown and J. L. Goldstein, Proc. Natl. Acad. Sci. U. S. A., 1974, 71,
788.
51. M. S. Brown, T. F. Deuel, S. K. Basu and J. L. Goldstein, J. Biol. Chem.,
1978, 253, 1121.
52. E. Z. Dajani, T. G. Shahwan and N. E. Dajani, J. Assoc. Acad. Minor
Phys., 2002, 13, 27.
53. A. M. Gotto Jr., Clin. Cardiol., 2003, 26(Suppl. 1), 121.
54. J. A. Tolbert, Nature Rev. Drug. Discov., 2003, 2, 517.
55. M. J. Pencina, R. B. D’Agostino Sr., M. G. Larson, J. M. Massaro and R.
S. Vasan, Circulation, 2009, 119, 3078.
56. A. Endo, M. Kuroda and Y. Tsujita, J. Antibiot., 1976, 29, 1346.
57. A. Endo, Y. Tsujita, M. Kuroda and K. Tanzawa, Eur. J. Biochem., 1977,
77, 31.
58. A. W. Alberts, J. Chen, G. Kuron, V. Hunt, J. Huff, C. Hoffman, J.
Rothrock, M. Lopez, H. Joshua, E. Harris, A. Patchett, R. Monaghan, S.
Currie, E. Stapley, G. Albert-Schonberg, O. Hensens, J. Hirschfield, K.
Hoogsteen, J. Liesch and J. Springer, Proc. Natl. Acad. Sci. U. S. A., 1980,
77, 3957.
59. E. S. Istvan and J. Deisenhofer, Science, 2001, 292, 1160.
60. K. Tanzawa and A. Endo, Eur. J. Biochem., 1979, 98, 195.
61. C. E. Nakamura and R. H. Abeles, Biochemistry, 1985, 24, 1364.
62. E. S. Istvan, Am. Heart J., 2002, 144(Suppl. 6), S27.
63. E. S. Istvan and J. Deisenhofer, Biochim. Biophys. Acta, 2000, 1529, 9.
64. E. S. Istvan, M. Palnitkar, S. K. Buchanan and J. Deisenhofer, EMBO J.,
2000, 19, 819.
65. D. L. Sprecher, Am. J. Cardiol., 2000, 86(Suppl), 46L.
66. Rader, Am. J. Cardiol., 2003, 91(Suppl), 18E.
67. I. Issemann and S. Green, Nature, 1990, 347, 645.
68. S. A. Kliewer, S. S. Sundseth, S. A. Jones, P. J. Brown, G. B. Wisely, C. S.
Koble, P. Devchand, W. Wahli, T. M. Willson, J. M. Lenhard and J. M.
Lehmann, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 4318.
69. B. Desvergne and W. Wahli, Endocr. Rev., 1999, 20, 649.
70. C. Duval, M. Muller and S. Kersten, Biochim. Biophys. Acta, 2007, 1771, 961.
71. Y. Xu, D. Mayhugh, A. Saeed, X. Wang, R. C. Thompson, S. J.
Dominianni, R. F. Kauffman, J. Singh, J. S. Bean, W. R. Bensch, R. J.
Barr, J. Osborne, C. Montrose-Rafizadeh, R. W. Zink, N. P. Yumibe,
N. Huang, D. Luffer-Atlas, D. Rungta, D. E. Maise and N. B. Mantlo,
J. Med. Chem., 2003, 46, 5121.
72. M. L. Sierra, V. Beneton, A. B. Boullay, T. Boyer, A. G. Brewster,
F. Donche, M. C. Forest, M. H. Fouchet, F. J. Gellibert, D. A. Grillot,
M. H. Lambert, A. Laroze, C. Le Grumelec, J. M. Linget, V. G.
Montana, V. L. Nguyen, E. Nicodeme, V. Patel, A. Penfornis, O. Pineau,
Carboxylic Acids and their Bioisosteres 157
D. Pohin, F. Potvain, C. B. Ruault, M. Saunders, J. Toum, H. E. Xu,
R. X. Xu and P. M. Pianetti, J. Med. Chem., 2007, 50, 685.
73. M. Nomura, T. Tanase, T. Ide, M. Tsunoda, M. Suzuki, H. Uchiki,
K. Murakami and H. Miyachi, J. Med. Chem., 2003, 46, 3581.
74. R. C. Desai, E. Metzger, C. Santini, P. T. Meinke, J. V. Heck, J. P. Berger,
K. L. MacNaul, T. Q. Cai, S. D. Wright, A. Agarwal, D. E. Moller and
S. P. Sahoo, Bioorg. Med. Chem. Lett., 2006, 16, 1673.
75. J. L. R. Barlow, R. E. Beitman and T. H. Tsai, Arzneim.-Forsch., 1982, 32,
1215.
76. D. McTavish, K. L. Goa and M. Ferrill, Drugs, 1990, 39, 552.
77. B. P. Monahan, C. L. Ferguson, E. S. Killeavy, B. K. Lloyd, J. Troy and
L. R. Cantilena Jr., JAMA, 1990, 264, 2788.
78. K. T. Kivistö, P. J. Neuvonen and U. Klotz, Clin. Pharmacokinet., 1994,
27, 1.
79. M. Jurima-Romet, K. Crawford, T. Cyr and T. Inaba, Drug Metab.
Dispos., 1994, 22, 849.
80. L. L. von Moltke, D. J. Greenblatt, S. X. Duan, J. S. Harmatz and R. I.
Shader, J. Clin. Pharmacol., 1994, 34, 1222.
81. K. H. Ling, G. A. Leeson, S. D. Burmaster, R. H. Hook, M. K. Reith and
L. K. Cheng, Drug Metab. Dispos., 1995, 23, 631.
82. M. G. Eller, B. J. Walker, P. A. Westmark, S. J. Ruberg, K. K. Antony,
B. E. McNutt and R. A. Okerholm, J. Clin. Pharmacol., 1992, 32, 267.
83. R. E. Benton, P. K. Honig, K. Zamani, L. R. Cantilena and R. L.
Woosley, Clin. Pharmacol. Ther., 1996, 59, 383.
84. P. K. Honig, D. C. Wortham, R. Hull, K. Zamani, J. E. Smith and L. R.
Cantilena, J. Clin. Pharmacol., 1993, 33, 1201.
85. A. Markham and A. J. Wagstaff, Drugs, 1998, 55, 269.
86. G. A. Patani and E. J. LaVoie, Chem. Rev., 1996, 96, 3147.
87. A. S. Kalgutkar, B. C. Crews, S. W. Rowlinson, A. B. Marnett, K. R.
Kozak, R. P. Remmel and L. J. Marnett, Proc. Natl. Acad. Sci. U. S. A.,
2000, 97, 925.
88. A. S. Kalgutkar, A. B. Marnett, B. C. Crews, R. P. Remmel and L. J.
Marnett, J. Med. Chem., 2000, 43, 2860.
89. A. S. Kalgutkar, S. W. Rowlinson, B. C. Crews and L. J. Marnett, Bioorg.
Med. Chem. Lett., 2002, 12, 521.
90. A. S. Kalgutkar, B. C. Crews, S. Saleh, D. Prudhomme and L. J. Marnett,
Bioorg. Med. Chem., 2005, 13, 6810.
91. R. N. Butler, Adv. Het. Chem., 1977, 21, 323.
92. C. W. Thornber, Chem. Soc. Rev., 1979, 8, 563.
93. H. Singh, A. S. Chawla, V. K. Kapoor, D. Paul and R. K. Malhotra,
Prog. Med. Chem., 1980, 17, 151.
94. A. Burger, Prog. Drug Res., 1991, 37, 287.
95. A. J. Albert, J. Chem. Soc. B, 1966, 427.
96. J. M. McManus and R. M. Herbst, J. Org. Chem., 1959, 24, 1643.
97. V. A. Ostrovskii and A. O. Koren, Heterocycles, 2000, 53, 1421.
158 Chapter 3
98. C. Hansch and L. Leo, in Exploring QSAR. Fundamentals and Applica-
tions in Chemistry and Biology, ed. C. Hansch and L. Leo, American
Chemical Society, Washington DC, 1995, ch. 13.
99. R. T. Eberhardt, R. M. Kevak, P. M. Kang and W. H. Frishman, J. Clin.
Pharmacol., 1993, 33, 1023.
100. D. H. Streeten, G. H. Anderson, J. M. Freiberg and T. G. Dalakos,
N. Engl. J. Med., 1975, 292, 657.
101. G. H. Anderson Jr., D. H. Streeten and T. G. Dalakos, Circ. Res., 1977,
40, 243.
102. A. T. Chiu, J. V. Duncia, D. E. McCall, P. C. Wong, W. A. Price Jr., M. J.
Thoolen, D. J. Carini, A. L. Johnson and P. B. Timmermans, J. Phar-
macol. Exp. Ther., 1989, 250, 867.
103. P. C. Wong, A. T. Chiu, W. A. Price, M. J. Thoolen, D. J. Carini, A. L.
Johnson, R. I. Taber and P. B. Timmermans, J. Pharmacol. Exp. Ther.,
1988, 247, 1.
104. K. Hsieh and G. R. Marshall, J. Med. Chem., 1986, 29, 1968.
105. P. C. Wong, W. A. Price Jr., A. T. Chiu, M. J. Thoolen, J. V. Duncia,
A. L. Johnson and P. B. Timmermans, Hypertension, 1989, 13, 489.
106. R. R. Wexler, D. J. Carini, J. V. Duncia, A. L. Johnson, G. J. Wells, A. T.
Chiu, P. C. Wong and P. B. Timmermans, Am. J. Hypertension, 1992, 5,
209S.
107. P. B. Timmermans, D. J. Carini and A. T. Chiu, Blood Vessels, 1990, 27,
295.
108. D. J. Carini, J. V. Duncia, P. E. Aldrich, A. T. Chiu, A. L. Johnson, M. E.
Pierce, W. A. Price, J. B. Santella, G. J. Well and R. R. Wexler, et al.,
J. Med. Chem., 1991, 34, 2525.
109. A. T. Chiu, D. E. McCall, W. A. Price Jr., P. C. Wong, D. J. Carini, J. V.
Duncia, R. R. Wexler, S. E. Yoo, A. L. Johnson and P. B. Timmermans,
Am. J. Hypertension, 1991, 4, 282S.
110. P. C. Wong, W. A. Price Jr., A. T. Chiu, J. V. Duncia, D. J. Carini, R. R.
Wexler, A. L. Johnson and P. B. Timmermans, Am. J. Hypertension,
1991, 4, 288S.
111. J. V. Duncia, D. J. Carini, A. T. Chiu, A. L. Johnson, W. A. Price, P. C.
Wong, R. R. Wexler and P. B. Timmermans, Med. Res. Rev., 1992, 12,
149.
112. R. A. Stearns, P. K. Chakravarty, R. Chen and S.-H. L. Chiu, Drug
Metab. Dispos., 1995, 23, 207.
113. M. W. Lo, M. R. Goldberg, J. B. McCrea, H. Lu, C. I. Furtek and T. D.
Bjornsson, Clin. Pharmacol. Ther., 1995, 58, 641.
114. K. Noda, Y. Saad, A. Kinoshita, T. P. Boyle, R. M. Graham, A. Husain
and S. S. Karnik, J. Biol. Chem., 1995, 270, 2284.
115. R. R. Wexler, W. J. Greenlee, J. D. Irvin, M. R. Goldberg, K. Prendergast,
R. D. Smith and P. B. M.W. M. Timmermans, J. Med. Chem., 1996, 39,
625.
116. M. Yazdanian, K. Briggs, C. Jankovsky and A. Hawi, Pharmaceutical
Res., 2004, 21, 293.
Carboxylic Acids and their Bioisosteres 159
117. P. L. Carl, P. K. Chakravarty and J. A. Katzenellenbogen, J. Med. Chem.,
1981, 24, 479.
118. W. Daehne, E. Frederiksen, E. Gundersen, F. Lund, P. Morch, H. J.
Petersen, K. Roholt, L. Tybring and W. O. Godtfredsen, J. Med. Chem.,
1970, 13, 607.
119. G. Paintaud, G. Alván, M. L. Dahl, A. Grahnén, J. Sjövall and J. O.
Svensson, Eur. J. Clin. Pharmacol., 1992, 43, 283.
120. F. Margarit, J. Moreno-Dalmau, R. Obach, C. Peraire and J. M.
Pla-Delfina, Eur. J. Drug Metab. Pharmacokinet., 1991, 3, 102.
121. Y. H. Zhao, J. Le, M. H. Abraham, A. Hersey, P. J. Eddershaw, C. N.
Luscombe, D. Butina, G. Beck, B. Sherborne, I. Cooper and J. A. Platts,
J. Pharm. Sci., 2001, 90, 749.
122. H. Lennernäs, L. Knutson, T. Knutson, A. Hussain, L. Lesko, T.
Salmonson and G. L. Amidon, Eur. J. Pharm. Sci., 2002, 15, 271.
123. R. T. Scheife and H. C. Neu, Pharmacotherapy, 1982, 2, 313.
124. M. Ehrnebo, S. O. Nilsson and L. O. Boréus, J. Pharmacokinet. Bio-
pharm., 1979, 7, 429.
125. N. O. Bodin, B. Ekström, U. Forsgren, L. P. Jalar, L. Magni, C. H.
Ramsay and B. Sjöberg, Antimicrob. Agents Chemother., 1975, 8, 518.
126. H. C. Neu, Rev. Infect. Dis., 1981, 3, 110.
127. R. S. Obach, F. Lombardo and N. J. Waters, Drug Metab. Dispos., 2008,
36, 1385.
128. G. M. Pollack and K. L. Brouwer, J. Pharmacokinet. Biopharm., 1991, 19,
189.
129. S. M. Pond and T. N. Tozer, Clin. Pharmacokinet., 1984, 9, 1.
130. B. C. Sallustio, L. Sabordo, A. M. Evans and R. L. Nation, Curr. Drug
Metab., 2000, 1, 163.
131. R. G. Dickinson and A. R. King, Biochem. Pharmacol., 1993, 46, 1175.
132. C. D. King, G. R. Rios, M. D. Green and T. R. Tephly, Curr. Drug
Metab., 2000, 1, 143.
133. A. F. McDonagh, L. A. Palma, J. J. Lauff and T. W. Wu, J. Clin. Invest.,
1984, 74, 763.
134. R. B. van Breemen and C. Fenselau, Drug Metab. Dispos., 1985, 13, 318.
135. R. Drew and K. Knights, Agents Actions Suppl., 1985, 17, 127.
136. P. C. Smith, A. F. McDonagh and L. Z. Benet, J. Clin. Invest., 1986, 77,
934.
137. M. Stogniew and C. Fenselau, Drug Metab. Dispos., 1982, 10, 609.
138. M. L. Hyneck, P. C. Smith, A. Munafo, A. F. McDonagh and L. Z.
Benet, Clin. Pharmacol. Ther., 1988, 44, 107.
139. A. Munafo, M. L. Hyneck and L. Z. Benet, Pharmacology, 1993, 47, 309.
140. L. Z. Benet, H. Spahn-Langguth and S. Iwakawa, Life Sci., 1993, 53,
L141.
141. A. Ding, J. C. Ojingwa, A. F. McDonagh, A. L. Burlingame and L. Z.
Benet, Proc. Natl. Acad. Sci. U.S.A., 1993, 90, 3797.
142. S. Bolze, N. Bromet, C. Gay-Feutry, F. Massiere, R. Boulieu and
T. Hulot, Drug Metab. Dispos., 2002, 30, 404.
160 Chapter 3
143. J. Wang, M. Davis, F. Li, F. Azam, J. Scatina and R. Talaat, Chem. Res.
Toxicol., 2004, 17, 1206.
144. G. S. Walker, J. Atherton, J. Bauman, C. Kohl, W. Lam, M. Reily,
Z. Lou and A. Mutlib, Chem. Res. Toxicol., 2007, 20, 876.
145. M. Castillo and P. C. Smith, J. Chromatogr., 1993, 614, 109.
146. M. Castillo and P. C. Smith, Drug Metab. Dispos., 1995, 23, 566.
147. H. Yuan, B. Feng, Y. Yu, J. Chupka, J. Y. Zheng, T. G. Heath and B. R.
Bond, J. Pharmacol. Exp. Ther., 2009, 330, 191.
148. C. Xia, L.-S. Gan, V. Kadambi, Y. Li, N. Liu, V. Uttamsingh, R. Gallegos,
M. Milton, J.-T. Wu, S. Prakash, C. Alden, F. Lee and S. Balani, Toxicol,
Pathol., 2009, 37, 123.
149. M. G. Manfredi, J. A. Ecsedy, K. A. Meetze, S. K. Balani, O. Burenkova,
W. Chen, K. M. Galvin, K. M. Hoar, J. J. Huck, P. J. LeRoy, E. T. Ray,
T. B. Sells, B. Stringer, S. G. Stroud, T. J. Vos, G. S. Weatherhead, D. R.
Wysong, M. Zhang, J. B. Bolen and C. F. Claiborne, Proc. Natl. Acad.
Sci. U. S. A., 2007, 104, 4106.
150. J. Liberg, Ann. Phys. Chem., 1829, 17, 389.
151. A. J. Hunt and J. Caldwell, in Conjugation Reactions in Drug Metabolism,
ed. G. J. Mulder, Taylor & Francis, London, 1990, pp. 273–305.
152. P. A. F. Dixon, J. Caldwell and R. L. Smith, Xenobiotica, 1977, 7, 727.
153. A. R. Jones, Xenobiotica, 1982, 12, 387.
154. P. A. F. Dixon, J. Caldwell, C. J. Woods and R. L. Smith, Biochem. Soc.
Trans., 1976, 4, 143–145.
155. M. O. James, R. L. Smith, R. T. Williams and M. Reidenberg, Proc.
Royal Soc. London, Series B: Biol. Sci., 1972, 182, 25.
156. P. A. F. Dixon, J. Caldwell and R. L. Smith, West African J. Biol. Appl.
Chem., 1977, 20, 21.
157. K. M. Knights, M. J. Sykes and J. O. Miners, Expert Opin. Drug Metab.
Toxicol., 2007, 3, 159.
158. M. P. Grillo and L. Z. Benet, Drug Metab. Dispos., 2002, 30, 55.
159. J. Olsen, C. Li, C. Skonberg, I. Bjornsdottir, U. Sidenius, L. Z. Benet and
S. H. Hansen, Drug Metab. Dispos., 2007, 35, 758.
160. M. P. Grillo and F. Hua, Drug Metab. Dispos., 2003, 31, 1429.
161. B. C. Sallustio, S. Nunthasomboon, C. J. Drogemuller and K. M.
Knights, Toxicol. Appl. Pharmacol., 2000, 163, 176.
162. U. Sidenius, C. Skonberg, J. Olsen and S. H. Hansen, Chem. Res. Toxi-
col., 2004, 17, 75.
163. J. M. Mayer, V. M. Roy-De, C. Audergon, B. Testa and J. C. Etter, Int. J.
Tiss. React., 1994, XVI, 59.
164. C. Li, L. Z. Benet and M. P. Grillo, Chem. Res. Toxicol., 2002, 15, 1480.
165. D. L. Nelson and M. M. Cox, in Lehninger Principles of Biochemistry, ed.
A. L. Lehninger, Worth Publishers, New York, 2000, pp. 598–625.
166. S. Vickers, C. A. Duncan, I. -W. Chen, A. Rosegay and D. E. Duggan,
Drug Metab. Dispos., 1990, 18, 138.
167. R. A. Halpin, E. H. Ulm, A. E. Till, P. H. Kari, K. P. Vyas, D. B.
Hunninghake and D. E. Duggan, Drug Metab. Dispos., 1993, 21, 1003.
Carboxylic Acids and their Bioisosteres 161
168. D. W. Everett, T. J. Chando, G. C. Didonato, S. M. Singhvi, H. Y. Pan
and S. H. Weinstein, Drug Metab. Dispos., 1991, 19, 740.
169. A. E. Black, M. E. Sinz, R. N. Hayes and T. F. Woolf, Drug Metab.
Dispos., 1998, 26, 755.
170. A. E. Black, R. N. Hayes, B. D. Roth, P. Woo and T. F. Woolf, Drug
Metab. Dispos., 1999, 27, 916.
171. J. G. Dain, E. Fu, J. Gorski, J. Nicoletti and T. J. Scallen, Drug Metab.
Dispos., 1993, 21, 567.
172. M. Boberg, R. Angerbauer, W. K. Kanhai, W. Karl, A. Kern, M. Radtke
and W. Steinke, Drug Metab. Dispos., 1998, 26, 640.
173. T. Prueksaritanont, B. Ma, X. Fang, R. Subramanian, J. Yu and J. H.
Lin, Drug Metab. Dispos., 2001, 29, 1251.
174. P. M. Haddad, A. Das, M. Ashfaq and A. Wieck, Expert Opin. Drug
Metab. Toxicol., 2009, 5, 539.
175. E. S. Zafrani and P. Berthelot, Hepatology, 1982, 2, 591.
176. H. Nau and W. Loscher, Epilepsia, 1984, 25(Suppl. 1), S14.
177. S. Russell, Curr. Opin. Pediatr., 2007, 19, 206.
178. S. A. Koenig, D. Buesing, E. Longin, R. Oehring, P. Häussermann,
G. Kluger, F. Lindmayer, R. Hanusch, I. Degen, H. Kuhn, K. Samii,
A. Jungck, R. Brückner, R. Seitz, W. Boxtermann, Y. Weber, R. Knapp,
H. H. Richard, B. Weidner, J. M. Kasper, C. A. Haensch, S. Fitzek,
M. Hartmann, P. Borusiak, A. Müller-Deile, V. Degenhardt, G. C.
Korenke, T. Hoppen, U. Specht and T. Gerstner, Epilepsia, 2006, 47,
2027.
179. N. Gerber, R. G. Dickinson, R. C. Harland and R. K. Lynn, J. Pediatr.
(St. Louis), 1979, 95, 142.
180. H. J. Zimmerman and K. G. Ishak, Hepatology, 1982, 2, 591.
181. T. A. Baillie, Chem. Res. Toxicol., 1988, 1, 195.
182. H. Schulz, Biochemistry, 1983, 22, 1827.
183. A. W. Rettenmeier, K. S. Prickett, W. P. Gordon, S. M. Bjorge, S.-L.
Chang, R. H. Levy and T. A. Baillie, Drug Metab. Dispos., 1985, 13, 81.
184. A. W. Rettenmeier, W. P. Gordon, K. S. Prickett, R. H. Levy, J. S.
Lockard, K. E. Thummel and T. A. Baillie, Drug Metab. Dispos., 1986,
14, 443.
185. A. W. Rettenmeier, W. P. Gordon, K. S. Prickett, R. H. Levy and T. A.
Baillie, Drug Metab. Dispos., 1986, 14, 454.
186. K. Kassahun, K. Farrell and F. Abbott, Drug Metab. Dispos., 1991, 19,
525.
187. K. Kassahun, P. Hu, M. P. Grillo, M. R. Davis, L. Jin and T. A. Baillie,
Chem. Biol. Interact., 1994, 90, 253.
188. W. Tang and F. S. Abbott, Chem. Res. Toxicol., 1996, 9, 517.
189. W. Tang and F. S. Abbott, J. Mass Spectrom., 1996, 31, 926.
190. W. Tang, J. Palaty and F. S. Abbott, J. Pharmacol. Exp. Ther., 1997, 282,
1163.
191. W. Tang, A. G. Borel, T. Fujiyama and F. S. Abbott, Chem. Res. Tox-
icol., 1995, 8, 671.
162 Chapter 3
192. M. P. Grillo, G. Chiellini, M. Tonelli and L. Z. Benet, Drug Metab.
Dispos., 2001, 29, 1210.
193. J. O. Minors and D. J. Birkett, Br. J. Clin. Pharmacol., 1998, 45.
194. C. R. Lee, J. A. Goldstein and J. A. Pieper, Pharmacogenetics, 2002, 12,
251.
195. M. A. Hamman, G. A. Thompson and S. D. Hall, Biochem. Pharmacol.,
1997, 54, 33.
196. A. Mancy, M. Antignac, C. Minoletti, S. Dijols, V. Mouries, N. T.
Duong, P. Battioni, P. M. Dansette and D. Mansuy, Biochemistry, 1999,
38, 14264.
197. T. S. Tracy, C. Marra, S. A. Wrighton, F. J. Gonzalez and K. P.
Korzekwa, Biochem. Pharmacol., 1996, 52, 1305.
198. W. Tassaneeyakul, D. J. Birkett, M. C. Pass and J. O. Miners, Br. J. Clin.
Pharmacol., 1996, 42, 774.
199. M. R. Wester, J. K. Yano, G. A. Schoch, C. Yang, K. J. Griffin, C. D.
Stout and E. F. Johnson, J. Biol. Chem., 2004, 279, 35630.
200. C. M. Mosher, M. A. Hummel, T. S. Tracy and A. E. Rettie, Biochem-
istry, 2008, 47, 11725.
201. J. U. Flanagan, L. A. McLaughlin, M. J. Paine, M. J. Sutcliffe, G. C.
Roberts and C. R. Wolf, Biochem. J., 2003, 370, 921.
202. A. Melet, N. Assrir, P. Jean, M. Pilar Lopez-Garcia, C. Marques-Soares,
M. Jaouen, P. M. Dansette, M. A. Sari and D. Mansuy, Arch. Biochem.
Biophys., 2003, 409, 80.
203. M. J. Sykes, R. A. McKinnon and J. O. Minors, J. Med. Chem., 2008, 51,
780.
204. S. Rao, R. Aoyama, M. Schrag, W. F. Trager, A. Rettie and J. P. Jones,
J. Med. Chem., 2000, 43, 2789.
205. M. J. De Groot, A. A. Alex and B. C. Jones, J. Med. Chem., 2002, 45,
1983.
206. I. Zamora, L. Afzelius and G. Cruciani, J. Med. Chem., 2003, 46, 2313.
207. J. Kirchheiner, I. Roots, M. Goldammer, B. Rosenkranz and J. Brock-
möller, Clin. Pharmacokinet., 2005, 44, 1209.
208. J. Kirchheiner and J. Brockmöller, Clin. Pharmacol. Ther., 2005, 77, 1.
209. H. Lennernas, Clin. Pharmacokinet., 2003, 42, 1141.
210. P. J. Neuvonen, J. T. Backman and M. Niemi, Clin. Pharmacokinet.,
2008, 47, 463.
211. Y. Shitara, T. Horie and Y. Sugiyama, Eur. J. Pharm. Sci., 2006, 27, 425.
212. R. B. Kim, Mol. Pharm., 2006, 3, 26.
213. Y. Shitara, T. Horie and Y. Sugiyama, Eur. J. Pharm. Sci., 2006, 27,
501.
214. T. Mikkaichi, T. Suzuki, M. Tanemoto, S. Ito and T. Abe, Drug Metab.
Pharmacokinet., 2004, 19, 171.
215. C. Funk, Expert Opin. Drug Metab. Toxicol., 2008, 4, 363.
216. M. Cvetkovic, B. Leake, M. F. Fromm, G. R. Wilkinson and R. B. Kim,
Drug Metab. Dispos., 1999, 27, 866.
Carboxylic Acids and their Bioisosteres 163
217. M. Shimizu, K. Fuse, K. Okudaira, R. Nishigaki, K. Maeda, H. Kusuhara
and Y. Sugiyama, Drug Metab. Dispos., 2005, 33, 1477.
218. M. A. Hamman, M. A. Bruce, B. D. Haehner-Daniels and S. D. Hall,
Clin. Pharmacol. Ther., 2001, 69, 114.
219. G. K. Dresser, R. B. Kim and D. G. Bailey, Clin. Pharmacol. Ther., 2005,
77, 170.
220. A. S. Kalgutkar, B. Feng, H. T. Nguyen, K. S. Frederick, S. D. Campbell,
H. L. Hatch, Y. A. Bi, D. C. Kazolias, R. E. Davidson, R. J. Mireles,
D. B. Duignan, E. F. Choo and S. X. Zhao, Drug Metab. Dispos., 2007,
35, 2111.
221. A. S. Kalgutkar, E. Choo, T. J. Taylor and A. Marfat, Xenobiotica, 2004,
34, 755.
222. B. Hsiang, Y. Zhu, Z. Wang, Y. Wu, V. Sasseville, W. P. Yang and T. G.
Kirchgessner, J. Biol. Chem., 1999, 274, 37161.
223. P. J. Neuvonen, J. T. Backman and M. Niemi, Clin. Pharmacokinet.,
2008, 47, 463.
224. Y. Sai and A. Tsuji, Drug Discov. Today, 2004, 9, 712.
225. C. W. Holtzman, B. S. Wiggins and S. A. Spinier, Pharmacotherapy, 2006,
26, 1601.
226. O. Ozdemic, M. Boran, V. Gokce, Y. Uzun, B. Kocak and S. Korkmaz,
Angiology, 2000, 51, 695.
227. T. K. Lau, D. R. Leachman and R. Lufschanowski, Tex. Heart Insti. J.,
2001, 28, 142.
228. W. Muck, I. Mai, L. Fritsche, K. Ochmann, G. Rohde, S. Unger,
A. Johne, S. Bauer, K. Budde and I. Roots, et al., Clin. Pharmacol. Ther.,
1999, 65, 251.
229. Y. Shitara, M. Hirano, H. Sato and Y. Sugiyama, J. Pharmacol. Exp.
Ther., 2004, 311, 228.
230. J. S. Wang, M. Neuvonen, X. Wen, J. T. Backman and P. J. Neuvonen,
Drug Metab. Dispos., 2002, 30, 1352.
231. B. W. Ogilvie, D. Zhang, W. Li, A. D. Rodrigues, A. E. Gipson,
J. Holasapple, P. Toren and A. Parkinson, Drug Metab. Dispos., 2006, 34,
191.
232. J. T. Backman, C. Kyrklund and K. T. Kivistö, Clin. Pharmacol. Ther.,
2000, 68, 122.
233. C. Kyrklund, J. T. Backman, K. T. Kivistö, M. Neuvonen, J. Laitila and
P. J. Neuvonen, Clin. Pharmacol. Ther., 2000, 68, 592.
234. C. Kyrkland, J. T. Backman, K. T. Kivistö, M. Neuvonen, J. Laitila and
P. J. Neuvonen, Clin. Pharmacol. Ther., 2001, 69, 340.
235. A. J. Bergman, G. Murphy, J. Burke, J. J. Zhao, R. Valesky, L. Liu, K. C.
Lasseter, W. He, T. Prueksaritanont, Y. Qiu, A. Hartford, J. M. Vega and
J. F. Paolini, J. Clin. Pharmacol., 2004, 44, 1054.
236. A. Asberg, Drugs, 2003, 63, 367.
237. H. C. Maltz, D. L. Balog and J. S. Cheigh, Ann. Pharmacother., 1999,
33, 1176.
164 Chapter 3
238. L. Gullestad, K. P. Nordal, K. J. Berg, H. Cheng, M. S. Schwartz and
S. Simonsen, Transplant Proc., 1999, 31, 2163.
239. M. F. Segaert, C. De Soete, I. Vandewiele and J. Verbanck, Nephrol. Dial.
Transplant., 1996, 11, 1846.
240. A. Lasocki, B. Vote, R. Fassett and E. Zamir, Ocul. Immunol. Inflamm.,
2007, 15, 345.
241. D. Williams and J. Feely, Clin. Pharmacokinet., 2002, 41, 343.
242. J. W. Deng, I. S. Song, H. J. Shin, C. W. Yeo, D. Y. Cho, J. H. Shon and
J. G. Shin, Pharmacogenet. Genomics, 2008, 18, 424.
243. M. K. Pasanen, H. Fredrikson, P. J. Neuvonen and M. Niemi, Clin.
Pharmacol. Ther., 2007, 82, 726.
244. I. Leiri, S. Suwannakul, K. Maeda, H. Uchimaru, K. Hashimoto,
M. Kimura, H. Fujino, M. Hirano, H. Kusuhara, S. Irie, S. Higuchi and
Y. Sugiyama, Clin. Pharmacol. Ther., 2007, 82, 541.
245. M. Niemi, K. A. Arnold, J. T. Backman, M. K. Pasanen, U. Gödtel-
Armburst, L. Wojnowski, U. M. Zanger, P. J. Neuvonen, M. Eichelbaum,
K. T. Kivistö and T. Lang, Pharmcogenet. Genomics, 2006, 16, 801.
246. G. D. Vladutiu and P. J. Isackson, N. Engl. J. Med., 2009, 360, 304.
247. SEARCH Collaborative Group, E. Link, S. Parish, J. Armitage,
L. Bowman, S. Heath, F. Matsuda, I. Gut, M. Lathrop and R. Collins,
N. Engl. J. Med., 2008, 359, 789.
248. E. Kamiyama, Y. Yoshigae, A. Kasuya, M. Takei, A. Kurihara and
T. Ikeda, Drug Metab. Pharmacokinet., 2007, 22, 267.
249. Z. H. Israili, J. Hum. Hypertension, 2000, 14(Suppl 1), S73.
250. D. A. Sica, T. W. B. Gehr and S. Ghosh, Clin. Pharmacokinet., 2005, 44,
797.
251. C. H. Gleiter and K. E. Mörike, Clin. Pharmacokinet., 2002, 41, 7.
252. L. J. Scott and P. L. McCormack, Drugs, 2008, 68, 1239.
253. M. Sharpe, B. Jarvis and K. L. Goa, Drugs, 2001, 61, 1501.
254. C. I. Furtek and M. W. Lo, J. Chromatogr., 1992, 573, 295.
255. F. Waldmeier, G. Flesch, P. Müller, T. Winkler, H. P. Kriemier,
P. Buhlmayer and M. De Gasparo, Xenobiotica, 1997, 27, 59.
256. J. Stangier, J. Schmid, D. Türck, H. Switek, A. Verhagen, P. A. Peeters,
S. P. Can Marle, W. J. Tamminga, F. A. Sollie and J. H. Jonkman, J.
Clin. Pharmacol., 2000, 40, 1312.
257. H. Stenhoff, P. O. Lagerström and C. Andersen, J. Chromatogr. B
Biomed. Sci. Appl., 1999, 731, 411.
258. D. Farthing, D. Sica, I. Fakhry, A. Pedro and T. W. Gehr, J. Chromatogr.
B Biomed. Sci. Appl., 1997, 704, 374.
259. A. Nohara, H. Kuriki, T. Ishiguro, T. Saijo, K. Ukawa, Y. Maki and
Y. Sanno, J. Med. Chem., 1979, 22, 290.
260. S. W. Huskey, G. A. Doss, R. R. Miller, W. R. Schoen and S. H. Chiu,
Drug Metab. Dispos., 1994, 22, 651.
261. R. A. Stearns, G. A. Doss, R. R. Miller and S. H. Chiu, Drug Metab.
Dispos., 1991, 19, 1160.
Carboxylic Acids and their Bioisosteres 165
262. R. A. Stearns, R. R. Miller, G. A. Doss, P. K. Chakravarty, A. Rosegay,
G. J. Gatto and S. H. Chiu, Drug Metab. Dispos., 1992, 20, 281.
263. H. Davi, C. Tronquet, G. Miscoria, L. Perrier, P. Dupont, J. Caix,
J. Simiand and Y. Berger, Drug Metab. Dispos., 2000, 28, 79.
264. T. J. Chando, D. W. Everett, A. D. Kahle, A. M. Stewart, N. Vachhar-
ajani, W. C. Shyu, K. J. Kripalani and R. H. Barbhaiya, Drug Metab.
Dispos., 1998, 26, 408.
265. T. Kondo, K. Yoshida, Y. Yoshimura, M. Motohashi and S. Tanayama,
J. Mass Spectrom., 1996, 31, 873.
266. G. D. Bowers, P. J. Eddershaw, S. Y. Hughes, G. R. Manchee and
J. Oxford, Rapid Commun. Mass Spectrom., 1994, 8, 217.
267. S. E. Huskey, R. R. Miller and S. H. L. Chiu, Drug Metab. Dispos., 1993,
21, 792.
268. A. Alonen, M. Finel and R. Kostiainen, Biochem. Pharmacol., 2008, 76, 763.
269. T. Unger and E. Kaschina, Drug Safety, 2003, 26, 707.
270. C. H. Yun, H. S. Lee, J. K. Rho, H. G. Jeong and F. P. Guengerich, Drug
Metab. Dispos., 1995, 23, 285.
271. U. Yasar, G. Tybring, M. Hidestrand, M. Oscarson, M. Ingelman-Sundberg,
M. L. Dahl and E. Eliasson, Drug Metab. Dispos., 2001, 29, 1051.
272. C. R. Lee, J. A. Pieper, A. L. Hinderliter, J. A. Blaisdell and J. A.
Goldstein, Pharmacotherapy, 2003, 23, 720.
273. U. Yasar, C. Forslund-Bergengren, G. Tybring, P. Dorado, A. Llerena,
F. Sjöqvist, E. Eliasson and M. L. Dahl, Clin. Pharmacol. Ther., 2002, 71,
89.
274. M. Bourrié, V. Meunier, Y. Berger and G. Fabre, Drug Metab. Dispos.,
1999, 27, 288.
275. S. Uchida, H. Watanabe, S. Nishio, H. Hashimoto, K. Yamazaki,
H. Hayashi and K. Ohashi, Clin. Pharmacol. Ther., 2003, 74, 505.
276. A. Nakashima, H. Kawashita, N. Masuda, C. Saxer, M. Niina, Y. Nagae
and K. Iwasaki, Xenobiotica, 2005, 35, 589.
277. A. Nishino, Y. Kato, T. Igarashi and Y. Sugiyama, Drug Metab. Dispos.,
2000, 28, 1146.
278. P. Laeis, K. Püchler and W. Kirch, J. Hypertens. Suppl., 2001, 19, S21.
279. E. Kamiyama, Y. Yoshigae, A. Kasuya, M. Takei, A. Kurihara and
T. Ikeda, Drug Metab. Pharmacokinet., 2007, 22, 267.
280. M. Takayanagi, N. Sano and H. Takikawa, J. Gastroenterol. Hepatol.,
2005, 20, 784.
281. R. Nakagomi-Hagihara, D. Nakai, K. Kawai, Y. Yoshigae, T. Tokui,
T. Abe and T. Ikeda, Drug Metab. Dispos., 2006, 34, 862.
282. A. Yamada, K. Maeda, E. Kamiyama, D. Sugiyama, T. Kondo,
Y. Shiroyanagi, H. Nakazawa, T. Okano, M. Adachi, J. D. Schuetz,
Y. Adachi, Z. Hu, H. Kusuhara and Y. Sugiyama, Drug Metab. Dispos.,
2007, 35, 2166.
283. W. Yamashiro, K. Maeda, M. Hirouchi, Y. Adachi, Z. Hu and
Y. Sugiyama, Drug Metab. Dispos., 2006, 34, 1247.
166 Chapter 3
284. K. Maeda, I. Ieiri, K. Yasuda, A. Fujino, H. Fujiwara, K. Otsubo,
M. Hirano, T. Watanabe, Y. Kitamura, H. Kusuhara and Y. Sugiyama,
Clin. Pharmacol. Ther., 2006, 79, 427.
285. N. Ishiguro, K. Maeda, W. Kishimoto, A. Saito, A. Harada, T. Ebner,
W. Roth, T. Igarashi and Y. Sugiyama, Drug Metab. Dispos., 2006, 34,
1109.
286. E. J. Murphy, T. J. Davern, A. O. Shakil, L. Shick, U. Masharani,
H. Chow, C. Freise, W. M. Lee and N. M. Bass, Dig. Dis. Sci., 2000, 45,
549.
287. J. Kohlroser, J. Mathai, J. Reichheld, B. F. Banner and H. L. Bonkovsky,
Am. J. Gastroenterol., 2000, 95, 272.
288. S. K. Herrine and C. Choudhary, Ann. Intern. Med., 1999, 130, 163.
289. P. J. Cox, D. A. Ryan, F. J. Hollis, A. M. Harris, A. K. Miller,
M. Vousden and H. Cowley, Drug Metab. Dispos., 2000, 28, 772.
290. M. Hanefeld, Int. J. Clin. Pract. Suppl., 2001, 121, 19.
291. K. Kassahun, P. G. Pearson, W. Tang, I. McIntosh, K. Leung, C. Elmore,
D. Dean, R. Wang, G. Doss and T. A. Baillie, Chem. Res. Toxicol., 2001,
14, 62.
292. K. He, R. E. Talaat, W. F. Pool, M. D. Reily, J. E. Reed, A. J. Bridges
and T. F. Woolf, Drug Metab. Dispos., 2004, 32, 639.
293. R. Alvarez-Sanchez, F. Montavon, T. Hartung and A. Pahler, Chem. Res.
Toxicol., 2006, 19, 1106.
294. L. Guo, L. Zhang, Y. Sun, L. Muskhelishvili, E. Blann, S. Dial, L. Shi,
G. Schroth and Y. P. Dragan, Mol. Divers., 2006, 10, 349.
295. Y. Masubuchi, S. Kano and T. Horie, Toxicology, 2006, 222, 233.
296. T. Jaakkola, J. Laitila, P. J. Neuvonen and J. T. Backman, Basic Clin.
Pharmacol. Toxicol., 2006, 99, 44.
297. S. J. Baldwin, S. E. Clarke and R. J. Chenery, Br. J. Clin. Pharmacol.,
1999, 48, 424.
298. C. M. Loi, M. Young, E. Randinitis, A. Vassos and J. R. Koup, Clin.
Pharmacokinet., 1999, 37, 91.
299. A. J. Scheen, Clin. Pharmacokinet., 2007, 46, 1.
300. M. Niemi, J. T. Backman, M. Granfors, J. Laitila, M. Neuvonen and
P. J. Neuvonen, Diabetologia, 2003, 46, 1319.
301. T. Jaakkola, J. T. Backman, M. Neuvonen and P. J. Neuvonen, Clin.
Pharmacol. Ther., 2005, 77, 404.
302. J. Y. Park, K. A. Kim, M. H. Kang, S. L. Kim and J. G. Shin, Clin.
Pharmacol. Ther., 2004, 75, 157.
303. B. W. Ogilvie, D. Zhang, W. Li, A. D. Rodrigues, A. E. Gipson, J.
Holsapple, P. Toren and A. Parkinson, Drug Metab. Dispos., 2006, 34, 191.
304. L. M. Moriarty, M. N. Lally, C. G. Carolan, M. Jones, J. M. Clancy and
J. F. Gilmer, J. Med. Chem., 2008, 51, 7991.
305. D. Boyer, J. N. Bauman, D. P. Walker, B. Kapinos, K. Karki and A. S.
Kalgutkar, Drug Metab. Dispos., 2009, 37, 999.
306. R. P. Remmell, B. C. Crews, K. R. Kozak, A. S. Kalgutkar and L. J.
Marnett, Drug Metab. Dispos., 2004, 32, 113.
Carboxylic Acids and their Bioisosteres 167
307. N. N. Kim, J. D. Cox, R. F. Baggio, F. A. Emig, S. K. Mistry, S. L.
Harper, D. W. Speicher, S. M. Morris Jr., D. E. Ash and A. Traish,
Biochemistry, 2001, 40, 2678.
308. J. M. Fevig, M. M. Abelman, D. R. Brittelli, C. A. Kettner, R. M. Knabb
and P. C. Weber, Bioorg. Med. Chem. Lett., 1996, 6, 295.
309. T. A. Kelly, J. Adams, W. W. Bachovchin, R. W. Barton, S. J. Campbell,
S. J. Coutts, C. A. Kennedy and R. J. Snow, J. Am. Chem. Soc., 1993, 115,
12637.
310. T. Pekol, J. S. Daniels, J. Labutti, I. Parsons, D. Nix, E. Baronas, F.
Hsieh, L. S. Gans and G. Miwa, Drug Metab. Dispos., 2005, 33, 771.
311. J. Labutti, I. Parsons, R. Huang, G. Miwa, L. S. Gans and J. S. Daniels,
Chem. Res. Toxicol., 2006, 19, 539.
312. O. M. Bakke, M. Manocchia, F. de Abajo, K. I. Kaitin and L. Lasagna,
Clin. Pharmacol. Ther., 1995, 58, 108.
313. M. Fung, A. Thornton and K. Mybeck, Drug Info J., 2001, 35, 293.
314. W. Tang, R. A. Stearns, R. W. Wang, S. H. Chiu and T. A. Baillie, Chem.
Res. Toxicol., 1999, 12, 192.
315. J. P. O’Donnell, D. K. Dalvie, A. S. Kalgutkar and R. S. Obach, Drug
Metab. Dispos., 2003, 31, 1369.
316. Q. Chen, G. A. Doss, E. C. Tung, W. Liu, Y. S. Tang, M. P. Braun, V.
Didolkar, J. R. Strauss, R. W. Wang, R. A. Stearns, D. C. Evans, T. A.
Baillie and W. Tang, Drug Metab. Dispos., 2006, 34, 145.
317. A. S. Kalgutkar, I. Gardner, R. S. Obach, C. L. Shaffer, E. Callegari, K.
R. Henne, A. E. Mutlib, D. K. Dalvie, J. S. Lee, Y. Nakai, J. P.
O’Donnell, J. Boer and S. P. Harriman, Curr. Drug Metab., 2005, 6, 161.
318. A. S. Kalgutkar and J. R. Soglia, Expert Opin. Drug Metab. Toxicol.,
2005, 1, 91.
CHAPTER 4

Primary, Secondary and

Tertiary Amines and their
Isosteres
D. K. WALKER*, R. M. JONES, A. N. R. NEDDERMAN
AND P. A. WRIGHT

Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and

Development, Ramsgate Road, Sandwich, Kent, CT13 9NJ, UK

4.1 Introduction
Amines are one of the most well-represented functional groups amongst
small drug molecules. These compounds cover a wide range of therapeutic
applications and possess a broad spectrum of physicochemical properties.
In the majority of cases, the amine provides a positively charged function
that is important for interaction with the target receptor and thus provides
potency and selectivity. Whilst this ionic interaction may not be essential for
binding to the target, it clearly represents an opportunity to provide higher
affinity binding than may be achieved through weaker intermolecular interac-
tions. There are also a number of examples where the incorporation of an
amine confers pharmacokinetic advantages and some of these are discussed in
this chapter.

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

168
Primary, Secondary and Tertiary Amines and their Isosteres 169
4.1.1 Amines that Interact with Aminergic Receptors
Foremost amongst compounds where the target receptor favours binding to an
amine are the drugs targeting the aminergic receptors that play such a central
role in physiologic function. The natural ligands of these receptors (see
Figure 4.1) all contain amines and hence, for drugs to interact with these
receptors or their re-uptake mechanisms, an amine function is clearly indicated.
Noradrenaline, serotonin (5-hydroxytryptamine), dopamine and histamine
are primary amines, with adrenaline being the only secondary amine amongst
the group. Drugs that interact with these targets comprise a mixture of primary,
secondary and tertiary amines and are employed for a variety of pharmaco-
logical treatments. The discovery of early available antidepressants arose from
chance observations1 including the tricyclic antidepressants (e.g. imipramine,
chlorpromazine and amitryptiline) which are, in the majority, tertiary amines
and exert their pharmacological effect through interaction with multiple ami-
nergic receptor targets.2 Monoamine oxidase inhibitors (e.g. moclobemide,
tranylcypromine, selegiline and indeloxazine) are also used as antidepressants
and were also initially discovered by chance. These agents are inhibitors of
monoamine oxidases (MAO-A and MAO-B), the enzymes responsible for the
breakdown of serotonin, dopamine, adrenaline and noradrenaline. From the
1980s onward, there was a more deliberate design of antidepressants to target
specific single or combined pharmacological action. Hence there are a vast
number of amine containing drugs including:

 selective serotonin re-uptake inhibitors (SSRIs), e.g. fluoxetine, sertraline

and paroxetine;
 noradrenaline re-uptake inhibitors (NSRIs), e.g. desipramine and
reboxetine;
 5HT antagonists, e.g. trazadone and nefazodone;

OH OH HN
H
HO NH2 HO N
CH3
NH2
HO HO
OH
Noradrenaline Adrenaline Serotonin (5HT)

HO NH2
HN NH2
N
HO

Dopamine Histamine

Figure 4.1 Structure of aminergic transmitter substances that are drug targets.
170 Chapter 4
 serotonin and noradrenaline re-uptake inhibitors (SNRIs), e.g.
venlafaxine;
 dopamine and noradrenaline re-uptake inhibitors, e.g. bupropion;
 serotonin and noradrenaline antagonists, e.g. mirtazapine.

Cardiac drugs that operate via action on aminergic receptors include

b-adrenergic antagonists (e.g. propranolol and atenolol) and a-adrenergic
antagonists (e.g. doxazosin and tamsulosin). b-Adrenergic agonists containing
an amine function are utilised as bronchodilators in the treatment of pul-
monary disease (e.g. salbutamol and salmeterol). Treatment of migraine is
achieved by amine drugs targeting blockade of a-adrenoreceptors (e.g. ergo-
tamine) and serotonin (e.g. sumatriptan and eletriptan). H1-antihistamines are
used for the treatment of allergies and include the amine containing drugs,
loratidine and terfenadine, whilst H2 antagonists are used to inhibit gastric acid
production and include the amine containing drugs, ranitidine and famotidine.

4.1.2 Amines that Interact with Acetylcholine

In addition to the aminergic drugs, there are amine containing drugs that are
antagonists of acetylcholine (ACh). Acetylcholine (Figure 4.2) is a quaternary
ammonium neurotransmitter, which dictates the requirement for a positively
charged function within a molecule to interact with the acetylcholine receptor.
There are a set of secondary (e.g. terodiline), tertiary (e.g. atropine, oxybu-
tynin and pirenzipine) and quaternary (e.g. ipratropium and oxitropium) amine
containing drugs that represent the class of antimuscarinic agents as antago-
nists of acetylcholine. There are also a number of amines that inhibit the
breakdown of acetylcholine by acetylcholinesterase—and thus potentiate its
effect—including the tertiary amine, physostigmine and the quaternary amines,
neostigmine and pyridostigmine.

4.1.3 Amines that Interact with Opioid Receptors

Several opioid agonists and antagonists also contain amine functions to bind to
the various opioid receptors. Whilst the endogenous ligands of these receptors
are peptidic in nature, the presence of a primary amine in endomorphins (e.g.
endomorphin-1: Tyr-Pro-Trp-Phe-NH2) demonstrates the role of the amine in
the ligand–receptor interaction for this pharmacological target as exemplified in

O
CH3
CH3
H3C N+ O
H3C
Acetylcholine

Figure 4.2 Structure of acetylcholine.

Primary, Secondary and Tertiary Amines and their Isosteres 171
prototypical opioid drugs such as morphine. Many of these agents (e.g. bupre-
norphine and pethidine) show mixed activity as agonists and antagonists at m, k
and d-opioid receptors. Opioid antagonists are used for central pharmacological
effects for example in the treatment of opioid overdose (e.g. naloxone) or for
peripheral action in the treatment of diarrhoea (e.g. loperamide).

4.1.4 Amines that Interact with Ion Channels

Another group of amine containing drugs are those intended to block the
sodium and potassium channels, the positive charge of the amine function
mimicking the positive charge of the sodium or potassium ion to elicit this
effect. Amine containing sodium channel blockers include encainide and fle-
cainide, whilst examples of potassium channel blockers include dofetilide and
amiodarone.
It is the non-specific interaction of amines with a specific potassium channel
that gives rise to the most common clinical safety concern amongst amine
containing drugs. Cardiac arrhythmia due to blockade of the hERG channel—
which conducts the repolarisation of cardiac muscle via the rapid delayed
rectifier K1 current (IKr)—may lead to Torsades de Pointes and sudden cardiac
death, and is a serious side effect for a number of amine containing drugs (e.g.
cisapride, terfenadine and grepafloxacin). Whilst incidence of this adverse event
is low (about 1 in 120 000 for cisapride), the potential severity of the outcome
has meant that several amine containing drugs have been withdrawn from the
market or relegated to restricted use. This toxicity is widespread across different
drug classes3 and reflects unusual susceptibility of the hERG channel to
blockade by drug molecules.4
A number of calcium channel blockers also contain amine functions (e.g.
verapamil and amlodipine), although this is far from a requirement for this
drug class which contains many neutral molecules (e.g. felodipine, nifedipine).
Amines are also represented amongst local anaesthetics, some of which func-
tion through sodium channel blockade (bupivacaine, lignocaine, prilocaine) or
via unknown mechanisms (ketamine). Lignocaine (also known as xylocaine and
lidocaine) may be used either locally as an anaesthetic or intravenously as an
anti-arrhythmic agent.

4.1.5 Amine Antimalarial Drugs

Alkaloid antimalarial drugs (e.g. amodiaquine, chloroquine, hydroxy-
chloroquine, halofantrine, mefloquine, primaquine and quinine) have a
mechanism of action that depends on the amine function of these molecules.
The drug molecule diffuses down a pH gradient to accumulate in the acidic
food vacuole (pH 4.7) of the parasite. The high concentration of drug inhibits
the polymerisation of haem, hence the haem that is released from the break-
down of haemoglobin builds up to toxic levels killing the parasite with its own
waste product.5
172 Chapter 4
4.1.6 Miscellaneous Amine Drugs
In addition to this substantial number of amine containing drugs where the
amine can be seen to provide interaction, there are a number of miscellaneous
amines including several antibiotics (enoxacin, erythromycin, azithromycin,
clarithromycin, aclarubicin, rifampicin, tetracycline, arbekacin and tobramy-
cin), an antifungal (terbinafine) and an antiparasitic (diethylcarbamazine).
Amines are also represented amongst oestrogen antagonists (clomiphene and
tamoxifen) and antitussive agents (dextromethorphan and pholcodine). Amines
in other drug classes include almitrine (respiratory stimulant), ticlopidine
(antithrombotic agent), sildenafil and vardenafil (PDE5 inhibitors), maraviroc,
aplaviroc and vicriviroc (CCR5 antagonists), pinacidil (potassium channel
opener) and vinblastine and vincristine (tubulin inhibitors).

4.1.7 Amine Isosteres

Amidine and guanidine functions (see Figure 4.3) are present in a number of
drug molecules as alternatives to simple primary, secondary or tertiary amines
and provide an equivalent basic function. Drugs that contain a guanidine
function include the adrenergic blocker, debrisoquine, the antiseptic, chlor-
hexidine and the H2-antagonist, cimetidine. Amidine containing drugs include
the antiparasitic agent diminazene. Related to amidine is the imidazoline
function, a heterocycle derived from imidazole and containing an imine bond.
The a-adrenergic agonist, naphazoline, is an example of an imidazoline con-
taining compound.
Hydrazine provides another alternative chemical function in place of simple
primary secondary or tertiary amines. There are a number of drugs containing
hydrazine functions including the antituberculosis compound, isoniazid, the
antihypertensive, hydralazine, and the MAO inhibitor, phenelzine.

R2 R3
N N
R3 R1 R5
R1 N N N
R4 R2 R4

Amidine Guanidine

R
H
N
N NH H2N R

Imidazoline Hydrazine

Figure 4.3 General structure of amine isosteres.

Primary, Secondary and Tertiary Amines and their Isosteres 173

H H H3C H H3C CH3 H3C CH3

N N N N

H H H CH3

Ammonia Primary Amine Secondary Amine Tertiary Amine

Figure 4.4 Structures of ammonia and primary, secondary and tertiary amines.

..
N

Figure 4.5 Chirality of trimethylamine.

4.2 Physicochemical Properties of Amines

Amines are hydrocarbon derivatives of ammonia consisting, in the neutral
state, of a nitrogen atom that forms three bonds and a lone pair of electrons.
Amines can be classified as primary, secondary and tertiary depending on the
degree of hydrocarbon substitution (Figure 4.4).
Amines are sp3 hybridized and thus are chiral (Figure 4.5), although rapid
pyramidal inversion means that the individual enantiomers are not usually
separable (resolvable).
The primary physicochemical properties of importance of the amino group
are polarity (see Section 4.2.1) and basicity (see Section 4.2.2). As a result of
these properties, amino groups are incorporated into drug molecules to facil-
itate solubility (in water of other vehicles for drug administration) and to
enhance binding to many drug targets including receptors and enzymes (see
Section 4.1).

4.2.1 Polarity of Amines

Due to the electronegativity of the nitrogen atom, C–N and N–H bonds possess
polarity with the partial negative charge situated on the nitrogen. Thus, most
amine containing compounds will have a dipole which will facilitate aqueous
solubility via dipole–dipole interactions with water molecules (Figure 4.6). The
impact of amino groups on solubility and hence on oral absorption is discussed
further in Section 4.3.1.
The holistic lipophilicity of a compound, defined by the partition coefficient
(P, normally presented as log P) between organic and aqueous solvents is
clearly dependent on the relative polarity of the functional groups present in the
molecule. The introduction of a primary amine group into a molecule lowers its
log P by 1.23, reflecting an approximately 17-fold change in the partitioning
coefficient itself.6 It is of course important to remember that whilst amines
174 Chapter 4

H
O H
H
O
H3C H H
N

O
H H

Figure 4.6 Hydrogen bonding interactions between methylamine and water.

Figure 4.7 Structure of amitriptyline.

confer polarity to a molecule, their incorporation will not inevitably result in a

molecule of high aqueous solubility. As a simple example, the tricyclic anti-
depressant drug amitriptyline. (Figure 4.7), which contains a tertiary amine
group, is extremely lipophilic with a log P of 4.9 due to the large number of
non-polar functional groups.

4.2.2 Basicity of Amines

The second key physicochemical property of amines is their basicity, defined as
the ability to donate a pair of electrons and hence accept a proton. The relative
basicity of amine functional groups is defined by the dissociation constant (Ka),
typically described using a log scale (pKa) such that, the higher the pKa value,
the greater the basicity. As compounds are essentially fully ionised when the pH
is 2 or more units below the pKa, then the basicity of a molecule determines its
ionisation at physiological pH which in turn affects many aspects of molecular
behaviour, including target binding, distribution, absorption and solubility.

4.2.2.1 Factors Affecting Basicity

The basicity of amine groups depends on the ability of the amine functionality
to donate its lone pair which in turn is dependent on the chemical environment
of the nitrogen atom. One factor that affects basicity is induction, whereby
Primary, Secondary and Tertiary Amines and their Isosteres 175
Table 4.1 pKa values for some simple amines.
Amine pKa

Ammonia 9.21
Primary amines Methylamine 10.62
Ethylamine 10.63
Benzylamine 9.34
Ethanolamine 9.50
Hydrazine 8.10
Hydroxylamine 5.97
Acetamidine 12.52
Guanidine 13.71
Secondary amines Dimethylamine 10.64
Diethylamine 10.98
Benzylethylamine 9.68
Morpholine 8.36
Pyrolidine 11.27
Piperidine 11.22
Tertiary amines Trimethylamine 9.76
Triethylamine 10.65
Dimethylethylamine 9.99
Anilines Aniline 4.62
o-Hydroxyaniline 4.72
m-Hydroxyaniline 4.17
p-Hydroxyaniline 5.50
o-Chloroaniline 2.62
m-Chloroaniline 3.32
p-Chloroaniline 3.81
o-Methylaniline 4.38
m-Methylaniline 4.67
p-Methylaniline 5.07

proximal electron-donating groups (such as alkyl groups) increase electron

density on the nitrogen and hence increase basicity whilst electron-withdrawing
groups tend to reduce basicity. For this reason, secondary amine groups tend to
be more basic (higher pKa) than primary amines (Table 4.1).
In theory, tertiary amines should be more basic still. However, tertiary
amines introduce steric factors, which may hinder the ability of the amine
group to donate its lone pair. Thus, trimethylamine (pKa 9.76) is in fact
less basic than dimethylamine (pKa 10.64) or methylamine (pKa 10.62)
(Table 4.1).
The third factor that affects basicity is resonance, whereby nitrogen lone
pairs are delocalised across more than one functional group, thereby reducing
the ability of the amine to donate its lone pair and hence reducing basicity. This
explains why aromatic amines (anilines) are significantly less basic than ali-
phatic amines (Figure 4.8) and also why amide functional groups have low
basicity.
The basicity (pKa values) of a selection of simple amines is shown in
Table 4.1. Additional reference data on amine basicity are available from a
number of sources.7–11
176 Chapter 4

Figure 4.8 Electron delocalisation of aromatic amines.

4.3 Absorption Properties of Amine Containing Drugs

The majority of drugs are administered by the oral route and hence absorption
from the gastrointestinal tract is needed for these agents to reach the systemic
circulation and exert their pharmacological effects. The physicochemical
properties of a drug that influence absorption are solubility, particle size,
crystal form, lipophilicity, dissociation constant and molecular weight. The
presence of an amine function within a drug molecule has direct influence on
solubility, lipophilicity and dissociation constant and thus impacts on the
process of gastrointestinal absorption.
For absorption to occur it is first necessary for the drug molecule to be in
solution. There are then essentially two pathways by which drug can cross the
gastrointestinal membrane and enter the circulation. These are the transcellular
pathway where drug passes through the membranes and across the cells of the
gut wall and the paracellular pathway where drug passes through the tight
junctions between the cells. Transcellular transport includes active carrier
mediated processes; however, the majority of drug absorption occurs by the
passive transcellular pathway.
The weakly basic nature of primary, secondary and tertiary amines has two
fundamentally opposing effects on oral absorption via the transcellular
route. On the one hand, the presence of a polar and ionisable function
enhances aqueous solubility in the gastrointestinal fluid whilst, on the other
hand, the same function can hinder diffusion across the membrane which is
largely restricted to the non-ionised form of the drug and is enhanced by
lipophilicity.

4.3.1 Solubility and Absorption

The benefit that an amine can provide in terms of facilitating aqueous solu-
bility is illustrated in the case of the HIV-1 protease inhibitor, indinavir. Early
chemical leads in this series such as L-685434 (Figure 4.9) were shown to be
potent inhibitors of HIV-1 protease but possessed poor oral bioavailability.
Primary, Secondary and Tertiary Amines and their Isosteres 177

OH OH
H
Me N
Me L-685434
Me O O

N OH OH
H
N N
N Indinavir
O
O NH

Me Me
Me

Figure 4.9 Structure of the protease inhibitors L-685434 and the amine containing
indinavir.

Only through incorporation of the basic amine (piperazine) was adequate

solubility achieved and oral bioavailability realised.12 It should be noted that
whilst the amine constituent of indinavir provided adequate solubility, the
intrinsic solubility of the molecule remains relatively low and exhibits marked
pH dependency with solubility of only 0.03 mg ml1 at pH 6 rising to
4100 mg ml1 at pH below 3.5. This results in pH dependent oral absorption
of the compound13 and is a source of pharmacokinetic variability and sensi-
tivity to food intake.14

4.3.2 Membrane Permeability and Absorption

The pH of the stomach is typically acidic with a pH range of 1.4 to 2.1 in the
fasted state and pH 3.0 to 7.0 in the fed state.15 Amine containing drugs, or
salts of these compounds, will be expected to exhibit their maximum solubility
at these pH values16 and thus will most easily enter solution in the upper
regions of the gastrointestinal tract. However, the ionised form of a drug is less
lipophilic than the unionised form and is thus less able to diffuse across the
lipoidal membrane. Hence quaternary ammonium compounds generally have
extremely poor oral absorption due to their permanently charged nature
compared with partially ionised tertiary amine analogues as illustrated by
atropine and ipratropium bromide (Figure 4.10). The tertiary amine, atropine
exhibits 90% absorption17 whilst the quaternary ipratropium shows low
absorption and only 2% oral bioavailability.18
178 Chapter 4

OH OH

O O O O

Br −
N N
+

Me Me Me

Me

Atropine Ipratropium

Figure 4.10 Structure of the tertiary and quaternary amine containing antimuscarinic
agents, atropine and ipratropium bromide.

The rate of absorption of an amine containing drug is thus dependent upon

the concentration of the unionised species at the site of absorption. The fraction
in unionised form (f ) is predicted by the Henderson–Hasselbalch equation and
is dependent upon the pKa of the compound and the pH of the medium.

1
f ¼ ð4:1Þ
1 þ 10ðpKapHÞ

Thus for an amine with a pKa of 8.0, the fraction in the unionised form in the
stomach (pH B2.0) will be only 0.0001% compared to about 25% in the ileum
(pH B7.5). Based on the theory that only the unionised form is able to pene-
trate biological membranes (pH partition theory), it can be expected that amine
drugs are better absorbed in the lower regions of the gastrointestinal tract with
higher pH values when a significant proportion will always be present in the
unionised form. This theory is nicely illustrated by the example above featuring
atropine; no compound is absorbed from the stomach whereas absorption
reaches 90% during passage to the distal part of the jejenum.17
The limitation to this theory derives from an assumption that only unionised
drug is able to be absorbed by passive diffusion when in fact the ionised form of
drug is able to be absorbed by the paracellular route, albeit at a relatively slow
rate. Paracellular absorption is particularly effective for small drug molecules
(MWo250) and can be the major route for some low molecular weight
hydrophilic compounds. Hence for a small, quaternary ammonium compound
such as pyridostigmine, which is permanently charged but is a very small
molecule (MW ¼ 181), oral bioavailability is as high as 15%;19 this is presumed
to be due to paracellular absorption. The clear difference in terms of physi-
cochemical properties between the two quaternary ammonium compounds,
pyridostigmine and ipratropium, is molecular size (MW values of 181 versus
332) and is reflected in the markedly higher oral bioavailability of the former
Primary, Secondary and Tertiary Amines and their Isosteres 179
due to its ability to pass through the restricted access afforded by the aqueous
pores (diameter typically 3–6 Å in humans).
Effective transcellular absorption requires molecules to possess sufficient
lipophilicity to partition into and through the lipophilic environment of the cell
membranes that line the gastrointestinal tract. The distribution coefficient (D)
is a convenient parameter for describing the lipophilicity of a molecule and
reflects the contribution of the degree of ionisation at the pH at which the
determination is carried out to the overall expressed lipid affinity of a mole-
cule.20 Generally speaking, a distribution coefficient greater than 1 (log D40)
will indicate adequate lipid affinity for a compound to diffuse into and through
cell membranes.21
Examination of the absorption properties of a series of b-adrenoceptor
antagonists demonstrates a number of factors that influence absorption.
Figure 4.11 shows the relationship between lipophilicity and the estimated oral
absorption for 16 b-adrenoceptor antagonists ranging from log D7.4 values of -
1.9 (atenolol) to 2.0 (carvedilol). The majority of these compounds show high
and essentially complete (480%) absorption with the exception of atenolol,
nadolol, xamoterol and talinolol. Absorption of the hydrophilic agents, sotalol
and practolol, is complete despite their log D7.4 values of 1.4 and 1.3,
respectively, and indicates that these molecules are sufficiently small (MW 272
and 266, respectively) to pass through the aqueous pores. Three other hydro-
philic members of this series show incomplete absorption in man with values of
9% for xamoterol, 30% for nadolol and 44% for atenolol. For these three
compounds, the increase in molecular size from atenolol (MW ¼ 266) to

practolol bisoprolol timolol propranolol

100
sotalol pindolol oxprenolol penbutolol
metoprolol
80
betaxolol
acebutolol carvedilol
% absorbed

60
talinolol
atenolol
40
nadolol
20

xamoterol
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Log D7.4

Figure 4.11 Relationship between lipophilicity (Log D7.4) and extent of absorption in
man for a series of 16 b-adrenoceptor antagonists.
180 Chapter 4
nadolol (MW ¼ 309) to xamoterol (MW ¼ 339) would appear to correlate with
their ability to enter the systemic circulation via the paracellular route. It should
be recognised that in vitro systems for permeability studies, such as Caco-2 cell
monolayers, may not provide permeability assessments that are representative
of in vivo permeability for compounds which are absorbed via the paracellular
pathway. For example, sotalol displays low permeability in the Caco-2 cell line
despite its greater than 90% absorption in vivo.22 This in vitro/in vivo difference
is considered to be due to differences between the tightness of the intercellular
junctions.

4.3.3 Impact of P-glycoprotein on Absorption

The other b-antagonist to show incomplete absorption is talinolol, which
despite its high lipid affinity (log D7.4 ¼ 1.1), does not appear to show good
membrane permeability. The lipophilicity of talinolol is similar to other
b-antagonists that have complete absorption (e.g. propranolol, log D7.4 ¼ 1.1)
and, whilst its molecular weight (339) is toward the high end for this series, it is
certainly not of a size that would be expected to restrict transcellular permea-
tion. Several studies have shown that the incomplete absorption of talinolol
is a result of the compound being a substrate for the efflux transporter
P-glycoprotein.
Preclinical and clinical studies have demonstrated that P-glycoprotein acts to
prevent the compound from traversing the intestinal membrane by active efflux
back into the gut lumen. The P-glycoprotein inhibitor, verapamil, has been
shown to increase oral bioavailability of talinolol in rats23 and also to decrease
the intestinal secretion of intravenously administered talinolol in humans.24
The affinity of talinolol for this efflux transporter would appear to be due to the
additional hydrogen bonding functionality within this molecule which is
imparted by the urea function, with hydrogen bond acceptor functions having
been identified as important for P-glycoprotein mediated efflux.25 However,
several other b-adrenoceptor antagonists, including propranolol, have similar
affinity for P-glycoprotein26 without suffering from incomplete absorption.
Thus it is likely the combination of P-glycoprotein affinity and low intrinsic
membrane permeability that results in the relatively poor absorption of
talinolol.27
Whilst affinity for P-glycoprotein and a resultant effect on absorption is not
directly attributed to an amine function, there are several other amine con-
taining compounds that have been shown to exhibit dose dependent absorption
potentially due to P-glycoprotein affinity including maraviroc,28 the NK2
antagonist, UK-224 67129 and the PDE5 inhibitor, UK-343 664.30 For each of
these three examples, physicochemical properties are less than ideal for good
membrane permeability31 with molecular weight values above 500 and rela-
tively high numbers of hydrogen bond acceptors. These features result in
borderline membrane permeability and the molecules are thus susceptible to
P-glycoprotein mediated efflux.
Primary, Secondary and Tertiary Amines and their Isosteres 181
4.4 Systemic Behaviour of Amine Containing Drugs
4.4.1 Tissue Affinity and its Impact on Distribution
It is well recognised that the physicochemical properties of drug molecules have
a fundamental impact on their distribution throughout the body and thus are
essential considerations in drug design. Such distribution properties impact the
pharmacological activity of drugs as a result of their affect on pharmacokinetics
and also the ability of compounds to reach intended cellular targets by crossing
cell membranes. The extent to which a drug is distributed outside of the cir-
culation is dependent on the relative affinity of the compound for tissue pro-
teins relative to plasma proteins and this determines the observed volume of
distribution according to eqn (4.2):

Vd ¼ Vp þ Vt :f u =f ut ð4:2Þ

where Vd is the distribution volume , Vp is the volume of plasma, Vt is the

volume of tissues, fu is the unbound fraction in plasma and fut is the unbound
fraction in tissues.
Hence the higher the tissue binding (and smaller the free fraction in tissues)
relative to plasma binding, the larger will be the apparent volume of tissues in
which the molecule resides.

4.4.2 Distribution and Duration

A key component of the systemic behaviour of amine containing drugs is the
high affinity with all types of tissue resulting from the electrostatic interaction
between the basic amine group of the drug and the anionic functionalities of the
phospholipid membranes. As a result, whilst compounds of all physicochemical
classes demonstrate an increase in volume of distribution (Vd) with increasing
lipophilicity (log D), basic compounds show higher Vd values than acids or
neutral molecules of comparable lipophilicity—depending on the basicity of the
amine (as defined by its pKa; see Section 4.2.2).
Furthermore, it is important to recognise that free (unbound) volumes of
distribution are also higher for basic compounds, demonstrating that holistic
affinity for tissue membranes is a more important factor than increased inter-
actions with plasma proteins.21 The general principle of high volumes of dis-
tribution for basic drugs is exemplified by the SSRIs paroxetine, sertraline and
fluvoxamine, all of which have large volumes in humans and animals.32,33 As a
component of drug design, therefore, the addition of a basic centre will often
result in an increase in Vd and concomitant increase in elimination half-life
(T1/2 ¼ 0.693Vd/Cl) as highlighted by the pharmacokinetics properties of
antibacterial agents based on the rifamycin structure (Figure 4.12).
The original agents (e.g. rifamycin SV; Figure 4.12) had short elimination
half-lives and low oral bioavailability.34 Incorporation of a basic centre in
rifampicin (Figure 4.12) increased volume of distribution35 and further
182 Chapter 4

CH3 CH3 CH3 CH3

HO HO
H3C H3C
OH O OH O
CH3COO OH OH CH3 CH3COO OH OH CH3
H3C NH H3C NH
H3C H3C
CH3O CH3O N
O O N
OR OH N
O O CH3
CH3 O CH3 O

Rifamycin Rifampicin
(R=H = Rifamycin SV)
(R=CH2COOH = Rifamycin B)

Figure 4.12 Structure of rifamycin and rifampicin.

O Me
N
Me Me Me Me
Me Me
HO OH HO OH
Me OH Me O Me OH Me O
Me Me Me Me
O O N O O N
Me OH Me Me OH Me
O O O O
Me Me
O O
O O
Me Me Me Me
Me Me
OH OH

Erythromycin Azithromycin

Figure 4.13 Structures of monobasic erythromycin and dibasic azithromycin.

modifications to the basic centre in rifabutin and rifapentine have contributed

to increased distribution volumes, longer elimination half-lives and extended
duration of action.36
Further impact of a basic centre on volume of distribution is shown by the
dramatic increase in Vd observed for the macrolide antibiotic, azithromycin,
when compared to the first generation drug, erythromycin (Figure 4.13). Ery-
thromycin contains one basic centre and has a Vd of 0.5 L kg1 in humans and
a half-life of about three hours. Introduction of a second basic centre in azi-
thromycin increases the volume to 23 L kg1 and half-life to 48 h, leading to a
significant prolongation of pharmacological action.37,38
As well as adding a basic centre to a molecule to affect distribution volume,
there is also evidence that manipulation of the distribution volume can be
achieved by altering the precise nature of the basic centre, as shown by a series of
Primary, Secondary and Tertiary Amines and their Isosteres 183
NHSO2Me NHSO2Me

O
N N
Me MeSO2NH
MeSO2NH
Dofetilide Compound I
pKa 8.0; Vd 4.0 L/kg pKa 8.2; Vd 3.9 L/kg

NHSO2Me
NHSO2Me
O O
N N
Me
MeSO2NH MeSO2NH

Compound II Compound III

pKa 7.8; Vd 3.9 L/kg pKa 7.3; Vd 1.2 L/kg

Figure 4.14 Structures of four potassium channel blockers with varying pKa values
and volumes of distribution.

four potassium channel blockers, including the anti-arrhythmic drug dofetilide

(Figure 4.14). Whilst three amines with pKa values in the range 7.8–8.2 show very
similar volumes in dog (3.9–4.0 L kg1), the fourth compound, with a lower pKa
of 7.3 (due to delocalisation of the tertiary amine) has a Vd of only 1.2 L kg1.39
This reduced volume reflects the increased plasma protein binding of the less polar
analogue restricting compound distribution out of the circulation.

4.4.3 Additional Specific Interactions Enhancing Tissue Affinity

Whilst generically high tissue affinity explains to a large extent the high volumes
of distribution for amine containing compounds, specific interactions between
the basic centre and tissue membranes can result in further increases in the
volume of distribution. An extreme example of this principle is the calcium
channel blocker, amlodipine, where the long plasma half-life (35 h) in humans is
largely driven by a Vd of 21 L kg1. That the basicity of the drug is the key
driver for the large Vd is demonstrated by the fact that the neutral analogue,
felodipine, which has a higher intrinsic lipophilicity than amlodipine (log D7.4
of 4.8 compared to 1.8), has a markedly lower Vd and thus a shorter elim-
ination half-life of 10 h. Further studies have shown that amlodipine binds in
the phospholipid bilayer in a manner that enables additional electrostatic
interactions which contribute to the large Vd.40,41
Another example is provided in the long duration of b-agonist activity for
salmeterol42,43 compared to many other b-agonists such as fomoterol. This is
thought to result from the interaction between the drug and a specific ‘exosite’
on or in the vicinity of the membrane-bound receptor in the lung.44,45

4.4.4 Distribution Dependent on pH

As well as tissue affinity resulting from electrostatic interactions between
amines and membranes, the basicity of amines can be responsible for other
184 Chapter 4
observed distribution phenomena. Most notably, basic compounds are known
to migrate to acidic environments in vivo, driven by pH gradients, rationalising
such effects as passive gut secretion. A recent example is provided by the CCR5
inhibitor maraviroc, which possesses a tertiary amine basic centre and was
shown to distribute 15% of the total dose into the gastrointestinal tract fol-
lowing intravenous administration to bile duct cannulated rats.28 Although
postulated as direct secretion into the gut, this phenomenon could also be
facilitated by active transport as has been observed for the beta-blocker tali-
nolol, where P-glycoprotein is known to be responsible for excretion of the
compound into the gut.46 The potential exploitation of the pH gradient phe-
nomenon has also been demonstrated for amine containing chemotherapeutic
agents such as mitoxantrone, where systemic alkalinisation of mice with sodium
bicarbonate increases ion trapping of the weak base within tumour tissue due to
the increased pH gradient.47

4.4.5 Plasma Protein Binding

The observation that many amine containing compounds exhibit large volumes
of distribution reflects the fact that the general affinity of amines for tissues
outweighs specific interactions with plasma proteins. However, it is important
to consider such interactions to fully understand the impact of the basic centre
on drug disposition. Whilst lipophilic acids show high affinity for human serum
albumin (HSA), amine containing drugs have a tendency to bind strongly with
the other most abundant plasma protein, a1-acid glycoprotein (AAG).
Although lipophilic amines will also bind to albumin via non-specific binding
effects, this specific interaction with AAG explains to a large extent the high
protein binding of many amine containing compounds such as the antiar-
rhythmic drug amiodarone, which is 96.3% bound in human plasma,48 and
most notably the antimuscarinic agent zamifenacin, which demonstrates
binding in excess of 99% in all species.49
An important consideration regarding the high affinity of amine containing
compounds for AAG is the lower concentration of this protein (10–30 mM)
compared to HSA (640 mM).50 In theory, this reduced concentration increases
the potential to saturate protein binding of amine containing compounds.
However, given the significant tissue distribution of these compounds and the
fact that the AAG concentration provides a molar equivalent concentration of
4 mg ml1 for a compound of molecular weight 250, protein binding saturation
is rarely observed in practice. Saturation of binding to AGP has been shown for
the If channel inhibitor YK754 in vitro with an estimated Bmax value of
7.8 mg ml1 (MW ¼ 469).51 This concentration is vastly in excess of plasma
concentrations observed in animal studies.
An important consideration for basic drugs is the variations in AAG con-
centration within the human population. AAG is an acute phase protein such
that several inflammatory states (e.g. infections, rheumatic disorders and sur-
gical injury) and pathological conditions (e.g. myocardial infarctions,
Primary, Secondary and Tertiary Amines and their Isosteres 185
malignancies and nephritis) can raise its serum concentration by up to three- or
four-fold. Furthermore, pregnancy and various disease states can affect the
concentration of AAG.
Thus, it can be important to understand the affinity of potential drugs for
AAG and HSA, and the impact of disease and/or special population on AAG
concentration such that the effect on the degree of protein binding can be taken
into account. To illustrate this phenomenon, the free concentration of pro-
pranolol and chlorpromazine were shown to be inversely correlated to AAG
concentration in studies with plasma from patients with Crohn’s disease,
inflammatory arthritis and renal failure as well as healthy control samples.52
Furthermore, the higher incidence of toxic side effects of bupivicaine in preg-
nant women may be explained by the increased free exposure resulting from
saturation of AAG.53

4.4.6 Brain Distribution

Although the same principles of interactions between basic centres and cell
membranes apply equally well to all tissue types, it is important to acknowledge
that differences exist between the blood–brain barrier and the gastrointestinal
tract. Thus, whilst many CNS drugs contain basic centres (e.g. tricyclic anti-
depressants and SSRIs), good oral absorption does not necessarily correlate
with good brain penetration. Whilst lipophilicity impacts tissue affinity, there is
evidence to suggest that the blood–brain barrier is more sensitive to log D
changes than the gastrointestinal tract. Indeed, it has been proposed that polar
surface area (PSA) should be considered in early discovery as a potential means
of separately impacting oral absorption and brain penetration.54
An example of using lipophilicity to reduce brain penetration for amine
containing drugs is atenolol, which is a less penetrant b-antagonist than pro-
pranolol due to its lower log D, and thus has a reduced side effect profile but
maintains high absorption following oral administration.55,56 Increased size
may also reduce brain penetration, as evidenced by maraviroc57 where low
permeability of the blood–brain barrier in rats was observed. The low pene-
tration of this compound is likely a result of P-glycoprotein affinity which is
often encountered with large amine molecules containing polar functionalities
(see Section 4.3.3).

4.5 Clearance of Amine Containing Drugs

4.5.1 Metabolic Clearance
Amine containing compounds may undergo a number of biotransformations,
of which the most common phase I pathways are N-oxidation and N-deal-
kylation, with deamination and methylation occurring less frequently. Amines
may also undergo the phase II metabolic clearance routes of glucuronidation,
sulfation and acetylation. The ability for an amine to be cleared by a particular
186 Chapter 4
metabolic route depends on the enzymes present in the tissues to which it is
exposed, the propensity of the compound to be metabolised at sites other than
the amine, and variation in the expression of the metabolic enzymes due to
gender, species, age or polymorphism.

4.5.2 Phase 1 Metabolism

4.5.2.1 N-Oxidation and N-Dealkylation
Two groups of enzymes, with overlapping substrate specificities, catalyse the
formation of N-oxide metabolites. The cytochrome P450 (CYP 450) enzymes
have a broad range of substrates with diverse metabolic products, while flavin
monooxygenases (FMOs) are more specific in their substrates, being involved
in oxidation of soft nucleophiles such as nitrogen and sulfur (and also selenium
and phosphorus). Both enzymes are NADPH-dependent and membrane-
bound, being located within the endoplasmic reticulum.
The mechanism by which the P450s and FMOs form N-oxides differ in that
CYP450 employs two sequential stage one electron oxidations whilst FMOs
utilise a single two electron oxidation directly on the nitrogen. It is difficult to
predict if an amine will be the substrate for P450s or for FMOs, but it does
appear that primary amines and charged species seem to be poor substrates for
FMOs.58,59
Isoforms of FMO differ in their distribution between species. FMO1 is the
major hepatic isozyme in most mammalian species (including rodents), whereas
FMO3 is the major form in adult human liver and hence is responsible for most
of the FMO-mediated drug metabolism. Therefore caution should be applied
when using animal metabolism data to extrapolate to man.
FMOs are not induced or inhibited (with the exception of FMO2),60 reducing
the propensity for drug–drug interactions in cases where the drug is mainly an
FMO substrate. For example, itopride and cisapride are both used in the
treatment of gastro-oesophageal reflux disease. Cisapride is metabolised by
CYP3A4. Co-administered drugs that inhibit CYP3A4 lead to an elevated
plasma level of cisapride and hence to adverse events such as arrhythmia.
Itopride, however, is metabolised by FMO3 and so is not subject to such drug-
drug interactions.61
FMO3 exhibits genetic polymorphism, which can effect metabolism and
hence the efficacy of drugs.62,63 This is particularly important as a significant
proportion of the population are at least mildly FMO deficient.64 Benzyda-
mine, a non-steroidal anti-inflammatory (NSAID), is almost fully N-oxidised
by FMO3 such that greater than 94% is excreted as the N-oxide in human urine
in the majority of the population. In subjects with severe FMO3 deficiency,
however, less than 36% N-oxide is found in urine.65 This translates to a 10%
increase in the Cmax of circulating benzydamine in deficient subjects relative to
the control population.
As FMOs and CYP450 have overlapping substrates, many substrates are
metabolised by both enzyme systems. Moclobemide (Figure 4.15) is a
Primary, Secondary and Tertiary Amines and their Isosteres 187

O O
FMO / CYP2C19 / CYP2E1
Cl Cl
N N
+
N O N O
Moclobemide
O

Figure 4.15 Oxidation of moclobemide by cytochrome P450 and FMO.

monoamine oxidase-A inhibitor prescribed for the treatment of depression.

Although predominately metabolised by FMO, it is also a substrate for
CYP2C19 and CYP2E1.66 Similarly, the tricyclic antidepressant clozapine is
N-oxidised by both FMO and CYP3A4.66,67
P450s are responsible for both N-oxidation and N-dealkylation of amines.
Usually, N-dealkylation takes precedent over N-oxidation in molecules where
an abstractable (highly acidic) hydrogen on the a-carbon is available.
Abstraction of this hydrogen results in formation of an intermediate carbon
radical which rearranges to give the aminium radical, which is stabilised by the
lone pair of electrons on the adjacent nitrogen. There are exceptions to this
‘rule’ with examples of amines that are N-oxidised via P450s even though an
abstractable hydrogen on the a-carbon is available. These exceptions include
methamphetamine and the alkaloid, senecionine, both of which are N-oxidised
in preference to being N-dealkylated.68 In general, N-oxidation prevails over
N-dealkylation where the aminium ion can be stabilised within the active site of
the P450.58,59
Physicochemical properties of the molecule dictate its potential to interact
with and hence to be metabolised via P450. Basicity does have an effect, but
pKa is not the predominate factor in active site binding and the relationship
between binding to P450 and basicity is not straightforward. A series of
benzphetamines analogues show increase affinity for CYP2B4 with decreasing
basicity.69 A similar observation was made for propranolol and its fluorinated
analogues, which showed increased interaction with CYP2D6 with decreasing
pKa.70 However, benzodiazepines (diazepam, medazepam and flurazepam,
Figure 4.16) exhibit an increased rate of N-dealkylation with increasing basicity
of the N-alkyl nitrogen.71
Lipophilicity has a major effect in the ability of a substrate to be metabolised
by P450s, with increasing lipophilicity often leading to increased metabolic
clearance. This has been clearly shown for the substrates of CYP2B6.72
Additional examples include sodium channel blockers such as lidocaine,
tocainide and mexilitene, which exhibit decreasing N-dealkylation by CYP3A4
with decreasing lipophilicity.71 An analogous observation was made for the
N-dealkylation of propranolol analogues by CYP1A2,70 where the enzyme
affinity increased with increasing lipophilicity. In contrast, however, these same
propranolol analogues exhibited reduced affinity for CY2D6 with increasing
lipophilicity. This results from the differences in binding mechanism between
CYP2D6 and other P450s such as 3A4 which rely on hydrophobic interactions
188 Chapter 4

N
O
N
N N
Cl
N
N
Cl
F
N
Cl
O

Diazepam Medazepam Flurazepam

pKa 3.3 pKa 6.2 pKa 8.1

N-dealkylation

Figure 4.16 Structures of benzodiazepines, diazepam, medazepam and flurazepam

showing increased rate of metabolism with increasing basicity.

for enzyme affinity; CYP2D6 binding is via ionic interactions such as ion
pairing or hydrogen bonding, and is not predominately driven by
lipophilicity.21
It has been reported that CYP2D6 is involved in metabolism of 20% of all
drugs and its substrate binding specificity, with the requirement of a positively
charged centre, means that amine containing drugs are often substrates.73
CYP2D6 is polymorphic with approximately 10% of caucasians and less than
1% of asians being poor metabolisers, while 5% of caucasians are ultra-rapid
metabolisers. Therefore, the genotype of subjects dosed with CYP2D6 sub-
strates has a major impact on how they respond to therapy and if they
experience adverse events. For example, the tricyclic antidepressant ami-
triptyline is hydroxylated by CYP2D6. Poor metabolisers exhibit elevated
plasma levels of amitriptyline and may be at risk of adverse effects such as
nausea or hypotension.74 In general, the development as medicines of com-
pounds that are highly dependent on clearance by CYP2D6 is avoided in order
to reduce the risks associated with significant inter-subject variability. For
example, development of the calcium channel antagonist UK-84 149, which is a
CYP2D6 substrate, was abandoned because the variability in drug exposure
between poor and extensive metabolisers was considered too great.75

4.5.2.2 Deamination
Amines may be oxidatively deaminated to aldehydes or ketones by P450s or
amine oxidases. The aldehyde or ketone formed is often not the final metabolic
product as further oxidation to the corresponding alcohol or acid may follow.
The best characterised of the amine oxidases are the monoamine oxidases
Primary, Secondary and Tertiary Amines and their Isosteres 189

OH
N
HO OH
Sertraline
HO
CYP2B6 O O
CYP2C19 O
CYPC9 UGT2B7 N O
NH2 UGT1A3
CYP3A4
CYP2D6 UGT1A6
UGT2B4
Cl
Cl
CYP2C19
CYP3A4
MAO-A
Cl
Cl MAO-B
Cl
Cl O
CYP2C19
CYP2E1
CYP3A4
MAO-A
MAO-B

Cl
Cl

Figure 4.17 Metabolism of sertraline by MAO and cytochrome P450.

(MAOs). MAOs are located in the outer surface of the mitochondrial mem-
brane and are found in most mammalian tissues. In humans, the highest MAO
levels are found in liver and lowest levels in the spleen. The two major forms of
MAO are MAO-A and MAO-B.
MAOs have affinity for polar substrates such as the endogenous neuro-
transmitters, dopamine and serotonin, which are outside the lipophilicity range
of many P450 enzymes. To be a substrate for MAO, there must be an available
hydrogen on the a-alpha to the amine and thus aniline is not a substrate. There
is some overlap in specificity between MAOs and P450s, however, as exem-
plified by sertraline (Figure 4.17), which is deaminated to sertraline ketone by
CYP3A4, 2C19 and MAO-A and MAO-B.76 Other compounds such as
sumatriptan are deaminated by MAO-A only with no P450 involvement.77
P450-mediated oxidative dealkylation can activate secondary and tertiary
amines towards deamination by MAO. For example, zolmitriptan (Figure 4.18)
is converted to its indole acetic acid metabolite via a P450-mediated secondary
amine intermediate.78
The role of other mammalian amine oxidases in xenobiotic clearance is less
well characterised. For example, SSAO (semicarbazide-sensitive amine oxidase)
is a copper containing enzyme found predominantly in the plasma, aorta, lung
and duodenum, but exhibiting low hepatic activity.79 SSAO displays different
190 Chapter 4

O O
H H
N CYP1A2 N
NH NH
O O

Zolmitriptan
N NH
MAO-A

O O
H H
N N
NH NH
O O
OH O

O
H
N
NH
O
OH

Figure 4.18 Sequential metabolism of zolmitriptan by P450 followed by MAO.

substrate specificities from MAO; for example, methylamine is a substrate for

SSAO but not for MAO.80 There have been few studies of the involvement of
SSAO in metabolism of pharmaceuticals, the exception being amlodipine81
(Figure 4.19), which is extensively deaminated in the dog but not the rat. This
reflects the high SSAO activity levels in dog plasma relative to that in rat.

4.5.2.3 N-Methylation
Amine N-methyltransferases are responsible for the biotransformations of
endogenous substrates; for example, the methylation of norepinephrine to
epinephrine and histamine and glycine to their methylated forms. They are also
involved in the metabolism of xenobiotics. This initially seems an illogical
metabolic route in that methylation often results in a less polar product
reducing the metabolites propensity to be excreted. When methylation intro-
duces a quaternary ammonium ion, however, the metabolite is more polar than
the parent. The methylation of nicotine (Figure 4.20) is an example of this.
The tricyclic antidepressants, imipramine and amitriptyline, are metabolised
by demethylation to give the active metabolise desipramine and nortriptyline
(Figure 4.21). However, a recent study has shown that 9–15% of subjects
showed methylation of desipramine and nortriptyline back to their parent
Primary, Secondary and Tertiary Amines and their Isosteres 191

N NH2
O N OH
O
O O O
O O
O O O O
Cl SSAO
Cl

Amolodipine

Figure 4.19 Metabolism of amlodipine by SSAO in dog plasma.

N N
+
N N

Nicotine

Figure 4.20 Methylation of nicotine by amine N-methyltransferase to yield a more

polar quaternary amine.

N
N
N
HN

Imipramine Desipramine

N N
H
Nortriptyline
Amitripyline

Figure 4.21 Reversible metabolism of imipramine and amitrryptyline to their

N-desmethyl metabolites involving P450 and N-methyl transferase.

compounds, imipramine and amitriptyline.82 The N-methylation observed

in a minority of subjects may possibly result from polymorphism of the amine
N-methyltransferases. However, polymorphism is known for other methyl-
transferases but has not been reported for amine N-methyltransferase.
192 Chapter 4
An alternative hypothesis is that these tricyclic antidepressants are also sub-
strates for CYP2D6 and CYP2C19, which are known to exhibit polymorphism.
In poor metabolisers, it may be that the low rate of turnover by CYP2D6 and
CYP2C19 gives the N-methyltransferase an opportunity to re-methylate the
secondary amine. This cyclic metabolism has implication in adjusting doses and
avoiding toxicity.

4.5.3 Phase 2 Metabolism

4.5.3.1 N-Glucuronidation
N-glucuronidation is catalysed by the enzyme UDP-glucuronyltransferase.
This enzyme is located in the endoplasmic reticulum but its cofactor, UDP
glucuronic acid, is found in the cytosol; thus, glucuronidation is not observed in
microsomal incubations unless the cofactor is added. Two groups of com-
pounds are N-glucuronidated:

 sulfonamides, aryl amines, amides, and acyclic and cyclic amines which
give secondary or tertiary amine glucuronides;
 tertiary amines which give rise to quaternary N-glucuronides.

Man and the common preclinical species (rat, dog rabbit, guinea pig, rabbit
and non-human primates) exhibit the ability to form secondary and tertiary
N-glucuronides, the species differences tending to be quantitative rather than
qualitative.83 Sulfadimethoxine (Figure 4.22) is N-glucuronidated in primates
to the extent of 4–27% of the administered dose and 1–6% in non-primates.84
Glucuronidation of tertiary amines is seen in non-human primates and man,
but occurs to a lesser extent in other preclinical species. For example, chlor-
promazine is N-glucuronidated extensively in vivo in man, but not in rat or
dog.85
Where there are two possible sites of N-glucuronidation, both of similar
basicity, glucuronidation often occurs on the nitrogen with the least bulky

O
O O O
N
Primates and preclinical S
O O
N species N
S N
O
N OH
H N H2N
O O
H2N HO
OH
Sulfadimethoxine
O OH

Sulfadimethoxine N-glucuronide

Figure 4.22 N-Glucuronidation of the sulphonamide of sulphadimethoxine.

Primary, Secondary and Tertiary Amines and their Isosteres 193

N
N N

+ +
N N
OH OH
O O
HO HO
OH OH
O OH O OH

Mitazapine N-glucuronide Mianserin N-glucuronide

Figure 4.23 N-Glucuronides of mitazapine and mianserin.

substituents.86 Otherwise the propensity for one amine to be glucuronidated

over another can be difficult to rationalise.
N-Glucuronides can undergo enterohepatic recirculation which can have
significant clinical implications. Both the antihistamine, cyclizine,86 and the
anticonvulsant, N-(2-amino-4-(4-fluorobenzylamino)-phenyl) carbamic acid
ethyl ester,87 undergo enterohepatic recirculation resulting in a prolonged half-
life. The tricyclic antidepressant, mirtazapine is converted to its quaternary
N-glucuronide in the gastrointestinal tract. This acts as a prodrug as the N1-
glucuronide is better absorbed than the parent drug, resulting in a doubling of
bioavailability (after deconjugation) of mirtazapine relative to the related
compound mianserin (Figure 4.23) which is N-glucuronidated to a lesser
extent.88

4.5.3.2 N-Sulfation
N-sulfates are formed by the cytosolic enzyme N-sulfotransferase (NST), which
transfers the sulfate group from its cofactor, 3 0 -phosphoadenosine 5 0 -phos-
phosulfate, to the substrate. A wide range of amines such as aniline and
octylamine89 have been shown be sulfated by NST in vitro. Reports of in vivo
xenobiotic substrates are limited, but desipramine (Figure 4.24) and 4-phe-
nylpiperazine,4-phenyl-1,2,3,6-tetrahydropyridine are both N-sulfated in vivo,
rat and man.90,91

4.5.3.3 N-Acetylation
The enzyme arylamine N-acetyltransferases (NAT) are cytosolic enzymes that
catalyse the transfer an acetyl group from acetyl coenzyme A to the nitrogen or
oxygen of primary aromatic amines, aryl hydroxyl amines or hydrazine.
Examples of N-acetylation of drugs in each these categories include hydrala-
zine, N-biphenyl-4-yl-hydroxylamine and procainamide.
194 Chapter 4

N N

O O

HN S
N

Desipramine Desipramine N-sulfate

Figure 4.24 N-Sulphation of desipramine occurs in rat and man.

N N N N
H H
N N

N NAT2 N N
N N N

N N

N N
H
O
UK-469413
Major metabolite in rat (in vivo)
and human (cytosol).

Figure 4.25 N-Acetylation of UK-469413 by NAT2 as a major route of clearance.

Recently unsubstituted piperazines have also shown to be substrates, e.g.

UK-469 413,92 where the N-acetylation was the major route of clearance
(Figure 4.25). This was unexpected as acetylation of aliphatic amines has
previously been reported to be a minor route of metabolism. The piperazine
nitrogen must be unhindered for N-acetylation to occur as the introduction of a
methyl group adjacent to the piperazine nitrogen blocks N-acetylation.
Partial charge on the nitrogen influences whether or not a substituted aniline
will be N-acetylated,93 with the introduction of electron-withdrawing groups
para to the amino group resulting in decreased N-acetylation.
NAT has been known to be polymorphic since 1954 when a proportion of the
population (50% of caucasians and 10% asians) were observed to be slow
acetylators of the anti-tubercular drug isoniazid, leading to adverse events.94
There are two major isozymes in humans, NAT-1 and NAT-2. Slow acet-
ylation, at least partially, results from reduced expression of NAT-2. NAT-1 is
also polymorphic but this is less well characterised.95 NAT1 and NAT2 have
overlapping specificities; NAT-1 preferentially acetylates p-aminobenzoic acid,
sulfamethoxazole and sulfanilamide, whilst isoniazid, hydralazine, procaina-
mide, dapsone and sulfamethazine are better substrates for NAT-2.92
Primary, Secondary and Tertiary Amines and their Isosteres 195
4.5.4 Non-metabolic Clearance
Amines may be cleared unchanged in urine and faeces (bile), and less com-
monly in saliva, sweat, expired air and breast milk.

4.5.5 Renal Clearance

Renal clearance is regulated by the kidney by passive and active processes.
Passive clearance involves filtration and reabsorption. Polar compounds are
readily cleared in the kidney because they can diffuse freely across the mem-
brane in the glomerulus and cannot diffuse back despite the concentration
being in favour of reabsorption. Conversely, lipophilic compounds are exten-
sively reabsorbed.
This relationship between lipophilicity and renal clearance of amines has
been exemplified for the b-adrenergic antagonists with polar compounds such
as atenolol being mainly cleared by renal clearance and lipophilic compounds
such as propranol showing negligible renal clearance.21 Urine pH can vary
widely between pH 4.5 to 8.0 and this can have a marked affect on the renal
excretion of amines. For example, when flecainide is co-administered with
sodium bicarbonate to make the urine alkaline, only 7% of the dose is renally
cleared compared with 45% in acidic urine over 32 h, increasing the exposure
(AUC) by 70% at the higher urinary pH. Similarly, only low levels of
amphetamine are excreted unchanged in alkaline urine but more than 50% is
excreted in acidic urine.
Active secretion by the kidney involves transport proteins in the proximal
tubule that can actively secrete cations such as pramipexole, dofetilide and
cimetidine. These transporters include the organic cation transporter, OCT2,
which is involved in the renal secretion of metformin used in the treatment of
type 2 diabetes. Orally administered metformin is renally eliminated (80% of
dose) without any significant metabolism. Expression of the transport protein is
polymorphic and, in individuals with reduced OCT2 activity, there is reduced
renal clearance and increased circulating plasma levels of metformin, impacting
on the control of plasma glucose levels.96

4.5.6 Biliary Clearance

Amine containing drugs often undergo a degree of biliary excretion, although it
is generally not a major clearance pathway, as may be the case for some acidic
compounds. Excretion of xenobiotics into the bile is via active transport.
Hepatocytes display a variety of active transport proteins that may be
responsible for the transport of amines. For example, in humans OCT1 is a
cation uptake transporter which is involved in the extraction of small hydro-
philic molecules such as cimetidine and verapamil97,98 from the blood into the
liver. MDR1 and MDR3 (also known as P-glycoproteins or P-gp) and MRP2
are ATP-dependent efflux pumps that transport amphiphilic cations and neu-
tral compounds such as tamoxifen, vincristine and ceftriaxone99 from the liver
196 Chapter 4
to the bile. Transport proteins may be inhibited and so may be subject to drug–
drug interactions. Co-administration of digoxin, a P-gp substrate with other P-
gp substrates or inhibitors, including amines such as quinidine or verapamil,
can reduce biliary excretion.100,101

4.6 Amines as Toxicophores and Toxicity of Amine

Containing Drugs
Structural alerts are a means to identify substructures that are considered to
pose a toxicity risk and are thus seen as undesirable groups for drug discovery
programmes. A number of amine containing structural alerts have been iden-
tified and are listed in Table 4.2.
Toxicity of small pharmaceutical molecules may arise from accumulation of
the unchanged compound in tissue. This is particularly true for very lipophilic
compounds including the amine, amidarone, which accumulates in the lung109
leading to toxicity.
The ability of a compound to form a reactive metabolite does not in itself
mean that it will give rise to adverse reactions in clinical use. The amount and
activity of toxic species formed relative to its extent of removal (detoxification)
dictates whether or not toxicity will be observed. In particular, consideration
should be given to the size of the administered dosed. Low dose drugs are
significantly less likely to give rise to adverse reactions. For example, procai-
namide gives raise a series of idiosyncratic toxicities as a result of formation of a
reactive nitroso metabolite formed from an N-hydroxy intermediate. Meto-
clopramine forms a similar nitroso metabolite, but does not display the same
range of adverse reactions as procainamide because it is dosed at 10 mg day1
rather than the gram quantities at which procainamide is dosed.110 The nitroso
metabolites (Figure 4.26) of both molecules go on to react with the tripeptide
glutathione, which acts to detoxify the molecule. At doses in excess of 1 g day1
of procainamide, there is insufficient glutathione to mop up the reactive nitroso
species.
Rigid planar primary amines (e.g. 2-aminofluorene) may be metabolised to
carcinogenic species by N-oxidation catalysed by CYP1A. This appears to be
specific to these flat structures, which can interchelate with DNA, and does not
occur with flexible molecules which may be similarly metabolised.
In other situations, N-oxidation may represent a detoxification pathway. For
example, tamoxifen (used in the treatment of breast cancer) is a-hydroxylated
by CYP3A4 which is then sulfated to form a species that binds to DNA. The N-
oxide of tamoxifen, formed by FMO1 or FMO3, is not reactive. This has been
linked to the observation that tamoxifen toxicity is lower in kidney where
FMO1 expression is high relative to CYP3A4 than in the liver where CY3A4
activity is higher.111
In addition to dose considerations, the observation of adverse drug reactions
also depends on the balance between generation of the toxic (reactive) species
and its removal by various detoxification mechanisms. One route of
Primary, Secondary and Tertiary Amines and their Isosteres 197
Table 4.2 Structural alerts featuring amines and their associated toxicities.
Substructure Toxic effects
Aniline ortho- or R1 R2 R1 R2 Methaemoglo
N N
para-hydro- binemia
xyanilides/anilines, HO R3 Agranulocytosis
e.g. procainamide, Aplastoc anaemia
propanil Hepatotoxicty
Skin hypersensi-
OH R1 R2 tivity
N Carcinogenic102

Dibenzazepines, R1
e.g. clozapine, N R2
N
mianserin N
Agranulocytosis103
X
N
H R2
Aminothiazoles or R R
N
thiazoles, N
e.g. cefepime, Hepatotocicity
pramipexole
R S Neurotoxicity104
R

4-Substituted-N-
alkyl-tetra- A N R Neurotoxicity105
hydropyridine,
e.g. MPTP

Hydrazines, hydra- H
zides, e.g. phe- N Hepatotoxicity
R N R
nelzine, H Carcinogenic106
procarbazine

1,3-Disubstituted
R N
piperazine, e.g.
MB243 HN N Microsomal
binding107
O

Diazine-piperazine/ Y
X z
morpholine, e.g.
BAY-41-8543 Mutagenic108
N N
A

detoxification is via the endogenous tripeptide glutathione, which can react

with and clear electrophiles, as exemplified previously with metoclopramine.
Another mechanism of detoxification is further metabolism of reactive species
to remove the toxicophore. For example, N-hydroxy amphetamine is toxic
198 Chapter 4

O O
O
O
N N HO N N
H2N H N N N H
N H
H
Procainamide

Cl O O O
Cl Cl

N N N N
H 2N H N N
HN H N H
O O O
HO O
Metoclopramine

Figure 4.26 Formation of nitroso metabolites of procainamide and metoclopramine.

FMO3
H
NH2 N
OH
Amphetamine N-hydroxy intermediate (toxic)

OH
+
N N OH
OH
Oxime
(not toxic)

Figure 4.27 Formation of the toxic N-hydroxy metabolite of amphetamine.

(Figure 4.27), capable of modifying biomolecules, but can be further metabo-

lised to form the non-toxic oxime.112
Conjugates are often considered innocuous in that they rarely give rise to
active metabolites and the increased solubility resulting from conjugation often
increases renal clearance. This is not always the case however. For example,
sulfation of 2-aminofluorene N-oxide, leads to an unstable N-O-sulfates, which
may decompose to a reactive species that can adduct to DNA and proteins.113
Another example is amitriptyline, which forms a quaternary N-glucuronide
that has been linked to flushing and tachycardia in man.86
N-Acetylation reactions are involved in the detoxification/activation of sev-
eral amine containing xenobiotics. In the case of hydralazine, N-acetylation
represents a detoxifying route as it reduces the amount of parent available to
undergo an alternative metabolic route that results in the formation of a
reactive species involved in the development of the autoimmune condition
lupus. Thus, the incidence of lupus is higher in slow acetylators on chronic
treatment than in fast acetylators.114
Less commonly, acetylation may be involved with the creation of a tox-
icophore. An example of this is that of 2-aminofluorene (Figure 4.28), where
Primary, Secondary and Tertiary Amines and their Isosteres 199

CYP450
H
NH2 N
OH
2-Aminofluorene
NAT1

H
NH+ N O
O
Adducts with DNA

Figure 4.28 Formation of a mutagenic cation from 2-aminofluorane.

the parent molecule is first N-hydroxylated and then O-acetylated as an

intermediate in the formation of a mutagenic cation.
Avoiding cardiotoxicity is a major concern when developing compounds as
potential medicines. This concern arises from the need to withdraw a number of
drugs from market on the basis of cardiotoxicity, due to the prolongation of
ventricular repolarisation, or as commonly referred to QT prolongation.115 QT
prolongation may lead to arrhythmia and possibly to Torsade de Pointes, a
form of arrhythmia that can cause sudden death. Not all drugs that cause QT
prolongation result in either of these adverse effects, but development of
medicines that show QT effects is preferably avoided and causes complexity
and delay.
Repolarisation of cardiac tissue is controlled by potassium channels, known
as rectifier K1 channels or IKr channels. Drug-induced QT prolongation results
from blockage of the IKr channels, changing the distribution of potassium ions
across the membrane. Because of the potential toxicity of these potassium
channel blockers, considerable effort has gone into modelling what structural
aspects of compounds give rise this form of cardiotoxicity, highlighting the
particular risk of amine containing compounds. One model suggests that a
pyramidal structure with four hydrophobic groups and a basic centre at the
apex is required for channel blocking.116 An example of this is astemizole which
consists of three aromatic rings and a basic nitrogen which make up the four
point of the pyramid.117 Other drugs that cause QT prolongation include
dofetilide, terfenadine, erythromycin and ketoconazole.

4.7 Zwitterions
Combining basic amine and carboxylic acid functions in the same molecule
provides zwitterions which have very different physicochemical properties to
molecules containing either functional group alone. These properties of
200 Chapter 4

O
Cl O Cl O
N OH N OH
N N

Hydroxyzine Cetirizine

Figure 4.29 Structures of the H1 antihistamines, the basic molecule hydroxyzine and
its zwitterionic metabolite cetirizine.

zwitterions may be worthy of medicinal chemistry design considerations when

working with acid, basic or neutral leads.
The change from a basic molecule to a zwitterionic molecule has significant
impact on pharmacokinetic characteristics. This is exemplified by the H1
antagonist class of compounds. First generation H1 antihistamines were lipo-
philic basic molecules such as hydroxyzine (Figure 4.29), which penetrated the
brain and caused CNS side effects such as drowsiness. Many of the second
generation H1 antihistamines were zwitterions, which maintained the phar-
macologically important amine, but were more polar due to the presence of a
carboxylic acid function and showed low CNS penetration and hence improved
side effect profiles. These agents include cetirizine (Figure 4.29), the carboxylic
acid metabolite of hydroxyzine. Cetirizine has a log D7.4 value of around 1.5
compared with 3.1 for hydroxyzine;118 this reduction in lipophilicity is less than
for other zwitterionic compounds, such as acrivastine, and reflects the ability of
hydroxyzine to form an internal hydrogen bond between the acid and amine
functions. In addition to the improved CNS side effect profile of cetirizine, the
physicochemical profile confers additional advantages such as slow receptor
dissociation and negligible cytochrome P450 interaction.119
Another example which contrasts the pharmacokinetic properties of a basic
compound and zwitterion is the fibrinogen receptor antagonist L-767 679 and
its carboxyl ester prodrug.120 Whilst the zwitterion, L-769 679 is highly polar
(log Po-3), it is absorbed via the paracellular route. Attempts to improve
absorption by the carboxyl ester prodrug found that this basic compound was a
substrate for P-glycoprotein—a property not shared with the zwitterion. Most
P-glycoprotein substrates are identified as moderately lipophilic bases and
hence raising lipophilicity of zwitterions by a prodrug approach may not
always result in improved absorption.

4.8 Prodrugs of Amines to Change Physicochemical

Properties
Prodrugs are employed to enable more effective delivery of a pharmacological
active molecule by overcoming one or more barriers to drug delivery through
alteration of physicochemical properties. Ideally, prodrugs are pharmacologically
Primary, Secondary and Tertiary Amines and their Isosteres 201
inactive, but once they have overcome the delivery barrier for which they are
intended, they should be rapidly transformed into the drug molecule. On this
basis there are a number of potential applications of prodrugs to facilitate
improved drug therapy:

(1) Improved solubility to enhance absorption through incorporation of

polar functionality
(2) Improved permeability to enhance absorption through increased
lipophilicity
(3) Targeted delivery to tissues or organs through tissue specific delivery or
cleavage of pro-moiety.

Given that inclusion of an amine function in itself benefits the aqueous

solubility properties of a molecule, there is not surprisingly no clear benefit
from masking this function with the intent of enhancing solubility. On the other
hand, inclusion of an amine function may compromise membrane permeability
due to the overall effect of reducing lipophilicity and the potentially large
fraction that will exist in ionised form for amines with high pKa values.
Consequently amine derivatives that prevent ionisation and add lipophilicity
have the potential to increase membrane permeability and hence absorption.

4.8.1 Prodrugs to Enhance Absorption

The challenge for potential prodrugs of amines is the release of the active
moiety once the prodrug has fulfilled its role of aiding absorption. Unlike ester
prodrugs of carboxylic acid compounds, simple amides do not generally
undergo rapid hydrolysis to yield the free amine. The non-enzymatic hydrolysis
of amides is generally so slow under physiological conditions that such com-
pounds may only be contemplated if enzymatic hydrolysis reactions are
available to them. Species differences in hydrolytic enzyme capabilities need to
be considered for such prodrugs, with generally much greater hydrolytic cap-
ability in rodents compared to larger mammalian species. Indeed a simple
diacetyl prodrug of an antimalarial agent (Figure 4.30) was found to be active
against malarial infection in mice but not rats and this was concluded to be a
result of the inability of rats to activate the prodrug.121
A number of studies utilising a double prodrug approach to provide
advantageous physicochemical characteristics for amines have been explored,

O O
CH3CONH S NHCOCH3 H2N S NH2
O O

+ 2 CH3COOH

Figure 4.30 An antimalarial prodrug 4,4 0 -diacetylamidodiphenylsulphone that was

active in mice but not rats.
202 Chapter 4

OH OH
O O
O
OH N
Me O N H
H Ester
hydrolysis Me
Me Me
Me
Me Me Me Me

OH OH

Lactonisation

O
NH2
O
+
Me
Me
Me Me
OH
OH

Figure 4.31 Schematic for proposed conversion mechanism for esterase and redox-
sensitive double prodrugs of amines (ref. 120).

although these have not to date yielded viable agents. The strategy with such an
approach is to utilise the established esterase activation as a first pro-moiety
forming a chemically reactive intermediate that undergoes spontaneous con-
version to the active compound. This is illustrated by the example of esters
of chemically reactive hydroxy amides122 to liberate 4-methoxyaniline
(Figure 4.31). Formation of the amine was demonstrated to be mediated by
enzymatic catalysis by a serine esterase.
Alternative proposals for double prodrug approaches for amines have
involved N-acyloxyalkoxycarbonyl derivatives of primary and secondary
amines which undergo enzymatic hydrolysis of the ester moiety leading to a
(hydroxyalkoxy) carbonyl derivative that undergoes spontaneous decomposi-
tion to yield the parent amine via an unstable carbamic acid.123

4.8.2 Prodrugs to Achieve Tissue Specificity

As mentioned previously, another potential use of prodrugs is to achieve tissue
specific delivery of the active species. A prodrug that is subject to activation by
an enzyme which is restricted to the target tissue is one method by which this
may be achieved. This approach has been explored for application of renal
specific vasodilation using a prodrug of dopamine. The enzyme, g-glutamyl-
transpeptidase, has highest concentrations in the kidney. The hypothesis
therefore was that L-g-glutamyl dopamine would be preferentially hydrolysed
in the kidney (Figure 4.32) and hence exert a local vasodilator effect; the
Primary, Secondary and Tertiary Amines and their Isosteres 203

H
N NH2
NH2
O
HO HO
OH OH

Figure 4.32 L-g-Glutamyl dopamine, an experimental kidney specific prodrug of

dopamine.

CH3 O O
N NH2
N H3C N N
H
O NH2
Cl O

Cl

Avizafone Diazepam

Figure 4.33 Avizafone, a water soluble prodrug of diazepam.

dopamine would be rapidly metabolised and excreted without producing gen-

eral systemic adrenergic stimulation.124 This approach has shown some pro-
mise in achieving a level of kidney specificity though significant adrenergic
stimulation has been observed in other tissues, suggesting that hydrolysis is not
restricted entirely to the kidney.

4.8.3 Prodrugs Utilising Amine Functionality

As discussed previously, the presence of an amine function can be advanta-
geous in enhancing the solubility of a molecule due to its hydrophilic nature.
Therefore adding a pro-moiety that contains an amine can facilitate the
absorption of a poorly soluble agent—provided of course that it is then readily
able to release the active compound once in the systemic circulation.
An example where this has been applied is in the use of the anticonvulsant
diazepam in the treatment of nerve agent poisoning.125 Such treatment requires
the rapid attainment of pharmacologically effective plasma levels in order to
prevent the extreme consequences of poisoning. Drugs are therefore adminis-
tered by intravenous or intramuscular injection which, in the case of diazepam,
is not facilitated by its poor aqueous solubility and requirement for an organic
solvent. Avizafone is a water soluble prodrug of diazepam that is hydrolysed by
an aminopeptidase to liberate lysine and diazepam (Figure 4.33). The aqueous
solubility is advantageous from a formulation and delivery aspect. Bioavail-
ability of diazepam has been shown to be 62–66% from avizafone in primates,
204 Chapter 4

O O CH3
N
Cl O O CH3 Cl OH
O O

esterase
Cl Cl Cl Cl

Triclosan

Figure 4.34 An amine containing prodrug of the antimalarial agent triclosan with
enhanced cellular uptake.

indicating high but incomplete conversion of the prodrug. In addition, the

prodrug provided an earlier Tmax (35 vs. 55 minutes) after intramuscular
administration and superior clinical efficacy.126
An amine prodrug of the antibacterial agent triclosan has been shown to
enhance hydrophilicity of the molecule and the weakly basic nature has
facilitated accumulation inside bacterial cells providing a four-fold increase in
in vitro potency.127 This prodrug utilised an ester linkage between triclosan and
dimethylaminoethylglutaric acid (Figure 4.34). Such a prodrug approach has
clear potential for a topical agent such as triclosan, but would appear more
challenging for systemic application where targeted delivery will require pro-
drug stability in the gastrointestinal tract and circulation.

References
1. S. H. Preskorn and R. Ross, in Handbook of Experimental Pharmacology,
ed. S. H. Preskorn, J. P. Feighner, C. Y. Stanga and R. Ross, Springer-
Verlag, Berlin, 2001, pp. 171–183.
2. J. Vetulani and I. Nalepa, Eur. J. Pharmacol., 2000, 405, 351.
3. V. Calderone, L. Testai, E. Martinotti, M. Del Tacca and M. C. Breschi,
J. Pharm. Pharmacol., 2005, 57, 151.
4. M. C. Sanguinetti and M. Tristani-Firouzi, Nature, 2006, 440, 463.
5. M. Foley and L. Tilley, Int. J. Parasitol., 1997, 27, 231.
6. S. Soloway and H. Lipschitz, J. Org. Chem., 1958, 23, 613.
7. A. Bryson, J. Am. Chem. Soc., 1960, 82, 4862.
8. H. K. Hall, J. Am. Chem. Soc., 1957, 79, 5441.
9. T. C. Bissot, R. W. Parry and D. H. Campbell, J. Am. Chem. Soc., 1957,
79, 796.
10. G. W. Stevenson and D. Williamson, J. Am. Chem. Soc., 1958, 80,
5943.
11. M. M. Tuckerman, J. R. Mayer and F. C. Nachod, J. Am. Chem. Soc.,
1959, 81, 92.
12. J. P. Vacca, B. D. Dorsey, W. A. Schleif, R. B. Levin, S. L. McDaniel,
P. L. Darke, J. Zugay, J. C. Quintero, O. M. Blahy, E. Roth, V. V.
Sardana, A. J. Schlabach, P. I. Graham, J. H. Condra, L. Gotlib, M. K.
Primary, Secondary and Tertiary Amines and their Isosteres 205
Holloway, J. Lin, I. W. Chen, K. Vastag, D. Ostovic, P. S. Anderson,
E. A. Emini and J. R. Huff, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 4096.
13. J. H. Lin, I. W. Chen, K. J. Vastag and D. Ostovic, Drug Metab. Dispos.,
1995, 23, 730.
14. C. Csajka, C. Marzolini, K. Fattinger, L. A. Decosterd, A. Telenti,
J. Biollaz and T. Buclin, Antimicrob. Agents Chemother., 2004, 48, 3226.
15. J. B. Dressman, G. L. Amidon, C. Reppas and V. P. Shah, Pharm. Res.,
1998, 15, 11.
16. A. T. M. Serajuddin, Adv. Drug Deliv. Rev., 2007, 59, 603.
17. B. Beerman, K. Hellstrom and A. Rosen, Clin. Sci., 1971, 40, 95.
18. K. Ensing, R. A. de Zeeuw, G. D. Nossent, G. H. Koeter and
C. Cornelissen, Eur. J. Clin. Pharmacol., 1989, 36, 189.
19. U. Breyer-Pfaff, U. Maier, A. M. Brinkmann and F. Schuman, Clin.
Pharmacol. Ther., 1985, 37, 495.
20. C. N. Manners, D. W. Payling and D. A. Smith, Xenobiotica, 1988, 18,
331.
21. D. A. Smith, B. C. Jones and D. K. Walker, Med. Res. Rev., 1996, 16,
243.
22. Y. Yang, P. J. Faustino, D. A. Volpe, C. D. Ellison, R. C. Lyon and L. X.
Yu, Mol. Pharm., 2007, 4, 608.
23. H. Spahn-Langguth, G. Baktir, A. Radschuweit, A. Okyar, B. Terhaag,
P. Ader, A. Hanafy and P. Langguth, Int. J. Clin. Pharmacol., 1998, 36,
16.
24. T. Gramatté and R. Oertel, Clin. Pharmacol. Ther., 1999, 66, 239.
25. A. Seelig, Int. J. Clin. Pharmacol. Ther., 1998, 36, 50.
26. S. Neuhoff, P. Langguth, C. Dressler, T. B. Andersson, C. G. Regardh
and H. Spahn-Langguth, Int. J. Clin. Pharmacol. Ther., 2000, 38, 168.
27. S. Doppenschmitt, H. Spahn-Langguth, C. G. Regardh and P. Langguth,
J. Pharm. Sci., 1999, 88, 1067.
28. D. K. Walker, S. Abel, P. Comby, G. J. Muirhead, A. N. R. Nedderman
and D. A. Smith, Drug Metab. Dispos., 2005, 33, 587.
29. K. Beaumont, A. Harper, D. A. Smith and J. Bennett, Eur. J. Pharm. Sci.,
2000, 12, 41.
30. S. Abel, K. C. Beaumont, C. L. Crespi, M. D. Eve, L. Fox, R. Hyland,
B. C. Jones, G. J. Muirhead, D. A. Smith, R. F. Venn and D. K. Walker,
Xenobiotica, 2001, 31, 665.
31. C. A. Lipinski, F. Lombardi, B. W. Dominy and P. J. Feeney, Adv. Drug
Deliv. Rev., 1997, 23, 3.
32. S. R. Grimsley and M. W. Jann, Clin. Pharm., 1992, 11, 930.
33. L. M. Tremaine, W. M. Welch and R. A. Ronfeld, Drug Metab. Dispos.,
1989, 17, 542.
34. N. Bergamini and G. Fowst, Arzneimittelforschung, 1965, 15, 951.
35. D. A. Smith, H. van de Waterbeemd and D. K. Walker, in Pharmaco-
kinetics and Metabolism in Drug Design, ed. D. A. Smith, H. van de
Waterbeemd and D. K. Walker, Wiley-VCH, Weinheim, 2nd edn., 2006,
pp. 55–65.
206 Chapter 4
36. W. J. Burman, K. Gallicano and C. Peloquin, Clin. Pharmacokinet., 2001,
40, 327.
37. G. Foulds, R. M. Shepard and R. B. Johnson, J. Antimicrob. Chemother.,
1990, 25, 73.
38. N. J. Lalak and D. L. Morris, Clin. Pharmacokinet., 1993, 25, 370.
39. D. K. Walker, K. C. Beaumont, D. A. Stopher and D. A. Smith, Xeno-
biotica, 1996, 26, 1101.
40. R. P. Mason, S. F. Campbell, S. D. Wang and L. G. Herbette, Mol.
Pharmacol., 1989, 36, 634.
41. D. Alker, R. A. Burges, S. F. Campbell, A. J. Carter, P. E. Cross, D. G.
Gardiner, M. J. Humphrey and D. A. Stopher, J. Chem. Soc. Perkin
Trans., 1992, 2, 1137.
42. G. R. Manchee, A. Barrow, S. Kulkarni, E. Palmer, J. Oxford, P. V.
Colthup, J. G. Maconopchie and M. H. Tarbit, Drug Metab. Dispos.,
1993, 21, 1022.
43. M. Cazzola, R. Testi and M. G. Matera, Clin. Pharmacokinet., 2002, 41,
19.
44. R. A. Coleman, M. Johnson, A. T. Nials and C. J. Varday, Trends Pharm.
Sci., 1996, 17, 324.
45. D. Jack, Br. J. Clin. Pharmacol., 1991, 31, 501.
46. K. Westphal, A. Weinbrenner, M. Zschiesche, G. Franke, M. Knoke,
R. Oertel, P. Fritz, O. von Richter, R. Warzok, T. Hachenberg, H. M.
Kauffmann, D. Schrenk, B. Terhaag, H. K. Kroemer and W. Siegmund,
Clin. Pharmacol. Ther., 2000, 68, 345.
47. N. Raghunand, B. P. Mahoney and R. J. Gillies, Biochem. Pharmacol.,
2003, 66, 1219.
48. P. Neyroz and M. Bonati, Experimentia, 1985, 41, 361.
49. K. C. Beaumont, A. G. Causey, P. E. Coates and D. A. Smith, Xeno-
biotica, 1996, 26, 459.
50. F. Zsila and Y. Iwao, Biochim. Biophys. Acta, 2007, 1770, 797.
51. K. I. Umehara, T. Nakamata, K. Suzuki, K. Noguchi, T. Usui and
H. Kamimura, Eur. J. Drug Metab. Pharmacokinet., 2008, 33, 117.
52. K. M. Piafsky, O. Borga, I. Odar-Cederlof, C. Johansson and F. Sjoqvist,
New Eng. J. Med., 1978, 299, 1435.
53. H. Wulf, P. Munstedt and C. Maier, Acta Anaesthesiol. Scand, 1991, 35,
129.
54. J. Kelder, P. D. J. Grootenhuis, D. M. Bayada, L. P. C. Delbressine and
J. P. Ploemen, Pharm. Res., 1999, 16, 1514.
55. P. R. Reeves, D. J. Barnfield, S. Longshaw, D. A. D. McIntosh and M. J.
Winrow, Xenobiotica, 1978, 8, 305.
56. A. Hayes and R. G. Cooper, J. Pharmacol. Exp. Ther., 1971, 176, 302.
57. D. K. Walker, S. J. Bowers, R. J. Mitchell, M. J. Potchoiba, C. M.
Schroeder and H. F. Small, Xenobiotica, 2008, 38, 1330.
58. P. Hlavica, Drug Metab. Rev., 2002, 34, 451.
59. D. M. Ziegler, Drug Metab. Rev., 2002, 34, 503.
60. B. Furnes and D. Schlenk, Toxicol. Sci., 2004, 78, 196.
Primary, Secondary and Tertiary Amines and their Isosteres 207
61. T. A. Mushiroda, R. Douya, E. Takahara and O. Nagata, Drug Met.
Dispos., 2000, 28, 1231.
62. M. S. Motika, J. Zhang and J. R. Cashman, M, Exp. Opin. Drug Metab.
Toxicol., 2007, 3, 831.
63. I. M. Hisamuddin and V. W. Yang, Pharmacogen., 2007, 8, 635.
64. J. R. Cashman and J. Zhang, Drug Metab. Dispos., 2002, 30, 1043.
65. E. Mayatepek, B. Flock and J. Zschocke Pharmacogen., 2004, 14, 775.
66. J. Hoskins, G. Shenfield, M. Murray and A. Gross, Xenobiotica, 2001, 31,
387.
67. B. Eiermann, G. Engel, I. Johansson, U. M. Zanger and L. Bertilsson, Br.
J. Clin. Pharmacol., 2003, 44, 439.
68. Y. Seto and P. J. Guengerich, Biol. Chem., 1993, 268, 9986.
69. P. Hlavica, Biochim. Biophys. Acta, 2006, 1764, 645.
70. A. L. Upthagrove and W. L. Nelson, Drug Metab. Dispos., 2001, 29, 1389.
71. D. A. Smith, in Computer-Assisted Lead Finding and Optimization,
ed. H. van de Waterbeemd, B. Testa and G. Folkers, Verlag Helvetica
Chimica Acta, Basel, 1997, pp. 267–276.
72. D. F. V. Lewis and M. Dickins, Drug Metab. Rev., 2003, 35, 1.
73. S. A. Islam, C. R. Wolf, M. S. Lennard and M. J. E. Sternberg, Carci-
nogen., 1991, 12, 2211.
74. J. Halling, P. Wehe and K. Brosen, Br. J. Clin. Pharmacol., 2007, 65, 134.
75. D. K. Walker, Br. J. Clin. Phamacol., 2004, 58, 601.
76. R. S. Obach, L. M. Cox and L. M. Tremaine, Drug Metab. Dispos., 2005,
33, 262.
77. C. M. Dixon, G. R. Park and M. H. Tarbit, Biochem. Pharmacol., 1994,
47, 1253.
78. M. J. Wild, Xenobiotica, 1999, 29, 847.
79. M. S. Benedetti, K. F. Tipton, R. Whomsley and E. Baltes, J. Neural
Transm., 2007, 114, 787.
80. B. Gong and P. J. Boor, Exp. Opin. Drug Metab. Toxicol., 2006, 2,
559.
81. A. P. Beresford, P. V. Macrae and D. A. Stopher, Xenobiotica, 1988, 18,
169.
82. M. P. Kurpius and B. Alexander, Pharmacotheraphy, 2006, 26, 505.
83. S. H. L. Chiu and S. E. W. Huskey, Drug Metab. Dispos., 1998, 26, 838.
84. R. H. Adamsom, J. W. Bridges, M. R. Kibby, S. R. Walker and R. T.
Williams, Biochem. J., 1970, 118, 41.
85. H. B. Hucker, S. C. Stauffer, A. J. Balletto, S. D. White, A. G. Zacchei
and B. H. Arison, Drug Metab. Dispos., 1978, 6, 659.
86. E. M. Hawes, Drug Metab. Dispos., 1998, 26, 830.
87. H. Schupke, C. Bauer, T. Kronbach, P. J. McNeilly, R. Hempel, J. M.
Strong, H. F. Kupferberg and J.. Engel, Pharmazie, 1997, 52, S22.
88. J. Kelder, C. Funke, T. de-Boer, L. Delbressine, D. Leysen and
V. Nickolson, J. Pharm. Pharmacol., 1997, 49, 403.
89. S. G. Ramaswamy and W. B. Jakoby, J. Biol. Chem., 1987, 262,
10039.
208 Chapter 4
90. K. Iwasaki, T. Shiraga, K. Noda, K. Tada and H. Noguchi, Xenobiotica,
1986, 16, 651.
91. K. O. Wong and K. P. Wong, Xenobiotica, 1996, 26, 17.
92. J. Rawal, R. Jones, A. Payne and I. Gardner, Xenobiotica, 2008, 38,
1219.
93. G. B. Scarfe, I. D. Wilson, M. A. Warne, E. Holmes, J. K. Nicholson and
J. C. Lindon, Xenobiotica, 2002, 32, 267.
94. H. B. Hughes, J. P. Biehl, A. P. Jones and L. H. Schmidt, Am. Rev.
Tuberc., 1954, 70, 266.
95. M. A. Payton and E. Sim, Biochem. Pharmacol., 1998, 55, 361.
96. M. K. Cho and I. S. Song, Drug Metab. Pharmacokinet., 2008, 23, 243.
97. H. Keosell, B. M. Schmitt and V. Goboulev, Rev. Physiol. Biochem.
Pharmacol., 2003, 150, 36.
98. K. I. Umehara, T. Watsububo, K. Noguchi and H. Kamimura, Xeno-
biotica, 2007, 37, 818.
99. M. Li, H. Yuan, N. Li, G. Song, Y. Zheng, M. Baratta, F. Hua,
A. Thurston, J. Wang and Y. Lai, Eur. J. Pharm. Sci., 2008, 35, 114.
100. B. Angelin, A. Arvidsson, R. Dahlqvist, A. Hedman and K. Schenck-
Gusatafsson, Eur. J. Clin. Invest., 1987, 17, 262.
101. A. Hedman, B. Angelin, A. Arvidsson, O. Beck, R. Dahlqvist, B. Nilsson,
M. Olsson and K. Schenck-Gustafsson, Clin. Pharmacol. Ther., 1991, 49,
256.
102. G. Sabbioni and O. Sepai, Chimica, 1995, 49, 374.
103. M. Hummer, M. Kurz, I. Kurtzhaler, H. Oberbauer, C. Miller and W. W.
Fleischhacker, J. Clin. Pharmacol., 1997, 17, 314.
104. D. K. Dalvie, A. S. Kalgutkar, S. C. Khojasteh-Bakht, R. S. Obach and
J. P. O’Donnell, Chem. Res. Toxicol., 2002, 15, 269.
105. T. Obata, Toxicol Lett, 2002, 132, 83.
106. P. M Gannet and J. H. Powell, J. Environ. Pathol. Toxicol. Oncol., 2002,
21, 1.
107. G. A. Doss, Chem. Res. Toxicol., 2005, 18, 271.
108. A. Straub, Bioorg. Med. Chem., 2002, 10, 1711.
109. D. A. Smith, H. Van de Waterbeemd and D. K. Walker, in Pharmaco-
kinetics and Metabolism in Drug Design, ed. R. Mannhold, H. Kubinyi
and H. Timmerman, Wiley-VCH, Weinheim, 2001, pp. 121–152.
110. J. Uetrecht, Drug Metab. Rev., 2002, 34, 651.
111. S. Shibutani, N. Suzuki, Y. R. S. Laxmi, L. J. Schild, R. L. Divi, A. P.
Grollman and M. C. Poirier, Cancer Res., 2003, 63, 4402.
112. J. R. Cashman, Y. N. Xiong, L. Xu and A. Janowsky, J. Pharmacol. Exp.
Therap., 1999, 288, 1251.
113. B. Burchell and M. H. Coughtrie, Environ. Health Persp., 1997, 105, 739.
114. P. Meisel, Pharmocogen., 2002, 3, 349.
115. J. Morganroth, J. Electrocardiol., 2004, 37, 25.
116. S. Ekins, J. Pharmacol. Exp. Therap., 2002, 301, 427.
117. A. Cavalli, E. Poluzzi, F. De Ponti and M. Recanatini, J. Med. Chem.,
2002, 45, 3844.
Primary, Secondary and Tertiary Amines and their Isosteres 209
118. G. Plemper van Balen, G. Caron, G. Ermondi, A. Pagliara, T. Grandi,
G. Bouchard, R. Fruttero, P. A. Carrupt and B. Testa, Pharm. Res., 2001,
18, 694.
119. C. Chen, Curr. Med. Chem., 2008, 15, 2173.
120. T. Prueksaritanont, P. Deluna, L. M. Gorham, B. Ma, D. Cohn, J. Pang,
X. Xu, K. Leung and J. H. Lin, Drug Metab. Dispos., 1998, 26, 520.
121. P. E. Thompson, Int. J. Leprosy, 1967, 35, 605.
122. K. L. Amsberry, A. E. Gerstenberger and R. T. Borchardt, Pharm. Res.,
1991, 8, 455.
123. J. Alexander, R. Cargill, S. R. Michelso and H. Schwamm, J. Med.
Chem., 1988, 31, 318.
124. J. J. Kyncl, F. N. Minard and P. H. Jones, in Peripheral Dopaminergic
Receptors, ed. J.-L. Imbs and J. Schwartz, Pergamon Press, Oxford,
pp. 369–380.
125. T. C. Marrs, Toxicol. Rev., 2004, 23, 145.
126. G. Lallement, F. Renault, D. Baubichon, M. Peoc’h, M. F. Burckhart,
M. Gallonier, D. Clarencon and N. Jourdil, Arch. Toxicol., 2000, 74, 480.
127. S. Mishra, K. Karmodiya, P. Parasuraman, A. Surolia and N. Surolia,
Bioorg. Med. Chem., 2008, 16, 5536.
CHAPTER 5

Sulfonamide as an Essential
Functional Group in Drug
Design
AMIT S. KALGUTKAR,a RHYS JONESb AND AARTI
SAWANTa
a
Pharmacokinetics, Dynamics and Metabolism Department, Pfizer Global
Research and Development, Eastern Point Road, Groton, Connecticut,
06340, USA; b Pharmacokinetics, Dynamics and Metabolism Department,
Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent,
UK, CT139NJ

5.1 Introduction
The sulfonamide, acylsulfonamide/sulfonimide and sulfonylurea functionalities
are found within numerous marketed agents for a wide range of therapies. In
2008 there were 112 marketed drugs in the United States that contained a
sulfonamide group (Table 5.1). The drugs differ in chemical structure, mole-
cular weight (MW) and lipophilicity, and act at different receptors/enzymes via
distinct biochemical mechanisms of action. In some instances, the presence of
the sulfonamide fragment is merely a circumstantial occurrence but in the vast
majority of cases, these drugs can be categorised into distinct groups based on
the role of the sulfonamide motif in the primary pharmacology. By definition a
sulfonamide is a molecule containing a sulfonyl group attached to an amine.
This yields the possibility of a number of sulfonamide expressions by virtue of

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

210
Table 5.1 Drugs containing a sulfonamide group marketed in the USA, 2008.
Acetazolamide Clorexolone Glyburide Penflutizide Sulprostone
Acetohexamide Co-trimoxazolea Glymidine Piretanide Sultiame
Almotriptan Cyclopenthiazide Hydrochlorothiazide Piroxicam Sumatriptan
Althiazide Cyclothiazide Hydroflumethiazide Polythiazide Tamsulosin
Amprenavir Darunavir Ibutilide Probenecid Teclothiazide
Amsacrine Delavirdine mesylate Indapamide Quinethazone Tenoxicam
Argatroban Diazoxide Isoxicam Rosuvastatin Tezosentan
Avitriptan Dichlorphenamide Levosulpiride Sildenafil citrate Thiothixene
Bemetizide Dofetilide Lornoxicam Sotalol Tipranavir
Bendroflumethiazide Dorzolamide Mafenide Sulfacetamide Tirofiban
Benzthiazide Droxicam Mebutizide Sulfacytine Tolazamide
Benzylhydrochlorothiazide Epithiazide Mefruside Sulfadiazine Tolbutamide
Bosentan Ethoxzolamide Meloxicam Sulfadimethoxine Torsemide
Brinzolamide Fenquizone Methazolamide Sulfafurazole Trichlormethiazide
Bumetanide Flosulide Methyclothiazide Sulfamazone Tripamide
Butizide Fosamprenavir Meticrane Sulfameter Valdecoxiby
Carbutamide Furosemide Metolazone Sulfamethizole Vardenafil
Celecoxib Glibornuride Mezlocillin Sulfamethoxazole Xipamide
Sulfonamide as an Essential Functional Group in Drug Design

Chlorothiazide Gliclazide Napsagatran Sulfanilamide Zafirlukast

Chlorpropamide Glimepiride Naratriptan Sulfaphenazole Zonisamide
Chlorthalidone Glipizide Nimesulide Sulfapyridine
Cinnoxicam Gliquidone Paraflutizide Sulfasalazine
Clopamide Glisoxepide Parecoxib sodiumb Sulfisoxazole
a
Co-trimoxazole is combination of the tetrahydropteroic acid synthetase and dihydrofolate reductase inhibitors sulfamethoxazole and trimethoprim, respectively.
b
COX-2 inhibitor withdrawn in 2005 due to concerns about possible increased risk of heart attack and stroke. Parecoxib, which is a prodrug of valdecoxib, is
marketed as Dynastat in the European Union; it has not been approved in the USA.
211
212 Chapter 5
substituents attached to the sulfur and nitrogen atoms. The pivotal role of
derivatives of sulfonamide containing drugs in opening up new areas of
pharmacology was discussed earlier in Chapter 1. These developments are
reviewed in detail in 5.1.1, 5.1.2, 5.1.3, and 5.14.

5.1.1 Sulfanilamide Antibacterial Agents

The term ‘sulfanilamide’ is used to describe a family of molecules containing
the sulfonamide functional group attached to an aniline ring in the para-
position. The first sulfanilamide derivative was a prodrug called prontosil red
(5.1) and was made by Bayer Laboratories in 1932. The finding that the sul-
fanilamide (5.2) metabolite, as opposed to the parent compound 5.1 from
which it was derived (Figure 5.1), was responsible for reversing streptococcal
infections in mice paved the way for the antibiotic revolution in medicine.1,2
The biochemical basis for the antimicrobial/antibacterial activity of sulfanila-
mides is due to competitive inhibition of bacterial tetrahydropteroic acid
synthetase.3 The enzyme catalyses the incorporation of para-aminobenzoic
acid (5.3) into dihydropteroate diphosphate resulting in the formation of
dihydopteroic acid, which is ultimately converted into dihydrofolic acid
(Figure 5.1). Dihydrofolic acid and its two-electron reduction product tetra-
hydrofolic acid are essential for cell division in bacteria; inhibition of the
formation of these pterin derivatives leads to antibacterial activity. Gerhard
Domagk, Jacques and Therese Trefouel are generally credited with the
discovery of the parent sulfanilamide as an antibacterial agent and Domagk
was awarded the Nobel Prize for his work in this field.
The serendipitous finding that replacement of the carboxylic acid moiety in
5.3 with a sulfonamide group affords a compound with a significantly higher
affinity for the enzyme than the natural substrate4,5 constitutes one of the first
examples of a non-classical bioisosteric replacement for the carboxylic acid
moiety. In the case of 5.2 and 5.3, the sulfonamide anion (RSO2N) closely
resembles the carboxylate anion (RCOO); the distance between the two oxygen
atoms of the carboxylic acid and sulfonamide anions are virtually identical (2.1–
2.3 Å) and, like the carboxylic acid group, the sulfonamide group has the
potential for multiple hydrogen bonding interactions.6,7 In contrast with the
carboxylic acid moiety, however, unsubstituted sulfonamides are neutral at
physiological pH (pKa B10.5–11.0); inclusion of appropriate substituents can
ensure that the sulfonamide moiety is anionic at physiologic pH. Structure–
activity relationship (SAR) studies indicate that N- and S-substituents, which
influence the ionisation of the sulfonamide group, are known to increase
bacteriostatic activity; the para-amino group, however, remains optimal for
pharmacological activity. Figure 5.1 also illustrates representative sulfanila-
mide-based anti-bacterial agents in clinical use. The structure of dapsone, which
is used in the treatment of leprosy, is distinct from other sulfonamide-based
antibacterials in that a phenyl ring separates the sulfonyl and amine fragment.
 NH2 O O
N O O N N
N O O O O S
O O
S S N N
S N N N H
N H H
H HOOC N H2N
NH2 N
H2N
H2N H Sulfaphenazole
N Sulfadiazine Sulfasalazine
N NH2 NH2 Sulfamethoxazole
CH3
O H3C
O N O O O
N N O O
O O N S
S S
S N N N O
O S O S N S H H
NH2 NH2 O OH H
O O H2 N H2N H2N NH2
5.1 5.2 5.3 H2 N
Sulfametoxydiazine Sulfafurazole Dapsone
Sulfamethizole

O COOH
COOH
OH OH OH OH OH N COOH
O O H 2N COOH H
N P P N N
N O O OH N N N N
H H
Sulfonamide as an Essential Functional Group in Drug Design

H2N N N H2N N N H2N N N

H
Sulfonamides H H
Dihydrofolic acid

Figure 5.1 Biochemical basis for antibacterial action of sulfanilamides.

213
214 Chapter 5
5.1.2 Sulfonamide-based Anti-inflammatory Agents
The recent discovery of the coxibs (celecoxib, valdecoxib and rofecoxib) as the
next generation of non-ulcerogenic non-steroidal anti-inflammatory drugs
(NSAIDs) provides a fascinating example of the value of the sulfonamide
group in medicinal chemistry.

5.1.2.1 Prostaglandin Biosynthesis and Inflammation

The committed step in prostaglandin and thromboxane biosynthesis involves
the conversion of fatty acid arachidonic acid to prostaglandin H2 (PGH2), a
reaction catalysed by the sequential action of the cyclooxygenase (COX) and
peroxidase activities of prostaglandin endoperoxide synthase or cycloox-
ygenase (PGHS or COX, EC 1.14.99.1) (Scheme 5.1).8 COX activity originates
from two distinct and independently regulated enzymes, termed COX-1 and
COX-2.9,10 COX-1 is the constitutive isoform and is mainly responsible for the
synthesis of cytoprotective prostaglandins in the gastrointestinal tract. COX-2
is inducible and short-lived; its expression is stimulated in response to pro-
inflammatory mediators.11,12 COX-2 plays a major role in prostaglandin bio-
synthesis in inflammatory cells (monocytes/macrophages).13,14
These observations suggest that COX-1 and COX-2 serve different physio-
logical and pathophysiological functions. Classical NSAIDs inhibit both COX-
1 and COX-2 to varying extents; inhibition of the former isozyme is thought to
be responsible for the gastrointestinal liabilities.15 The differential tissue dis-
tribution of COX-1 and COX-2 provided a rationale for the development of
selective COX-2 inhibitors as anti-inflammatory and analgesic agents that lack
the side effects exhibited by currently marketed NSAIDs. This hypothesis was
validated in animal models and has led to the introduction of celecoxib, val-
decoxib (sulfonamides) and rofecoxib (methylsulfone) into the marketplace.16

5.1.2.2 Mode of Inhibition of COX Isozymes by NSAIDs and

Coxibs
Traditional NSAIDs and the selective COX-2 inhibitors bind in the COX active
site but not the peroxidase active site of the isozymes. Kinetic analysis indicate

R1 R1 R1
2 O2 O AH2 O
R2 R2
Peroxidase
R2 COX O O
OOH OH
PGG2 PGH2
Arachidonic acid

R1 = CH2CH=CH(CH2)3COOH
R2 = C5H11

AH2 = Reducing substrate / electron donor for peroxidase catalysis

Scheme 5.1 Oxygenation of arachidonic acid by COX enzymes.

Sulfonamide as an Essential Functional Group in Drug Design 215
that most COX inhibitors are slow, tight-binding inhibitors that conform to the
minimal two-step mechanism depicted in eqn (5.1).17,18 The first step involves
the formation of a rapidly reversible (E  I) complex leading to competitive
inhibition. The second step is the time-dependent conversion of the initial (E  I)
complex to one, [E  I]*, in which the inhibitor is bound more tightly. Forma-
tion of the [E  I]* complex occurs in seconds to minutes and is thought to reflect
the induction of a subtle protein conformational change. It is of great impor-
tance to note that all of the selective COX-2 inhibitors are actually competitive
inhibitors of both COX-1 and COX-2, but exhibit selectivity for COX-2 in the
time-dependent step by binding tightly at the active site and causing a con-
formational change in the isozyme structure.18

k1 k2
E + I (E.I) (E.I)* ð5:1Þ
k-1

5.1.2.3 Origins of the Coxibs–the Diarylheterocycle Class of

Selective COX-2 Inhibitors
The chemical structures of diarylheterocycles can be traced back to clinical
candidates discovered over 30–40 years ago.19 Examination of the structures of
diarylheterocycle-based selective COX-2 inhibitors (e.g. celecoxib, valdecoxib
and rofecoxib) reveals a striking resemblance to the older COX inhibitor
phenylbutazone (Figure 5.2).20 The exceptional anti-inflammatory properties
of phenylbutazone undoubtedly provided much of the impetus that led che-
mists to exploit this chemotype.21
The 4-methylsulfonyl (CH3SO2)-diarylheterocycle combination, which is a
critical component for potent and selective COX-2 inhibition, was first discerned
in the structure of the 2,3-diarylthiophene, DuP 697 (Figure 5.2), disclosed in the
late 1980s as a non-ulcerogenic anti-inflammatory agent.22 The COX-2-selective
inhibitory properties of DuP697 were established by Copeland et al. in the early
1990s18 and reaffirmed the COX-2 selective inhibition hypothesis for designing
safer NSAIDs. Studies on methylsulfone replacements in DuP 697 indicated
that only the 4-sulfonamido (-SO2NH2) group maintained COX-2 inhibitory
potency and selectivity observed with Dup 697.23,24 Incorporating the 4-sulfo-
namido group also results in increased COX-2 inhibition, albeit with a loss of
COX-2 selectivity. This SAR trend appears to be a common theme in all of the
diarylheterocycles studied to date and is illustrated in Figure 5.2 with the
1,5-diarylpyrazole-based methylsulfone SC 58125 (COX-1/COX-2 selectivity
41000) and the sulfonamide celecoxib (COX-1/COX-2 selectivity B325).25
The oxidation state of the sulfur in methylsulfones is crucial for selective
COX-2 inhibition; its reduction to sulfoxide or sulfide reverses isozyme selec-
tivity. For instance, methyl-sulfone-containing 4,5-diarylthiazole (SC-8092) is a
selective COX-2 inhibitor, whereas the corresponding methylthioether deriva-
tive (SC-8076) exhibits COX-1-selective inhibition (Figure 5.2).26 Furthermore,
N-methylation or N,N-dimethylation of the sulfonamide group or reversal of
 216

H3C
O O
S
H3C O N CF3
O O
S N N
H
N Br O
CH3
N
H3C H2 N S
O S
H2 N S
O O
O O O O
Phenylbutazone F Valdecoxib Celecoxib
DuP 697 Rofecoxib

H3 C
SC-8092
Celecoxib R = NH2 H3CO R = SO2
CF3 IC (COX-2) = 0.04 µM Cl IC50 (COX-2) = 0.06 µM
50 R = SO2NH2 COX-2 selective
N N IC50 (COX-1) = 13 µM S IC50 (COX-1) > 100 µM
R = SO2NHCH3 Inactive
SC 58125 R = CH3 SC-8076 R = SO2N(CH3)2 Inactive
N R=S R = NO2, CF3, COOH Inactive
R IC50 (COX-2) = 0.1 µM R = OCH3
S IC50 (COX-1) > 100 µM H3C IC50 (COX-2) > 100 µM R COX-1 selective
R IC50 (COX-1) = 0.1 µM
O O

Figure 5.2 Pivotal role of the methylsulfone/sulfonamide group in selective COX-2 inhibition by diarylheterocycles.
Chapter 5
Sulfonamide as an Essential Functional Group in Drug Design 217
the sulfonamide group to a methanesulfonamido moiety in celecoxib results in
inactive compounds (Figure 5.2).25 Likewise, bioisosteric replacement of the
methylsulfone or sulfonamide groups with nitro, trifluoromethyl, methoxy or
carboxylic acid substituents either reverses isozyme specificity or results in
inactive compounds (see Figure 5.2).25,27
Crystal structures of complexes of sheep COX-1, mouse COX-2, and human
COX-2 with non-selective and selective inhibitors have been solved at 3–3.5 Å
resolution.28,29 Most traditional NSAIDs (e.g. diclofenac, ibuprofen, aspirin
and indomethacin) contain a free carboxylic acid group, which ion pairs to an
active site Arg120 residue. The Arg120 residue also ion pairs with the carboxylate
of the fatty acid substrate arachidonic acid. Site-directed mutagenesis of the
arginine residue in COX-1 to glutamine or glutamate renders the protein
resistant to inhibition by carboxylic acid containing NSAIDs.30,31 Arg120 is part
of a hydrogen bonding network with Glu524 and Tyr355 which stabilises sub-
strate–inhibitor interactions and closes off the upper part of the COX active site
from the spacious opening at the base of the channel referred to as the lobby.
Disruption of this hydrogen bonding network opens the constriction and
enables substrate–inhibitor binding and release to occur.
Co-crystal structures of COX-2 with methylsulfonyl- and/or sulfonamide-
diarylheterocycles have shown that the selective inhibitors bind to regions
accessible in COX-2 but not COX-1. For instance, solution of the COX-2
crystal structure co-complexed with the celecoxib derivative SC-558 (the 4-
methylphenyl group in celecoxib is replaced with the 4-bromophenyl sub-
stituent in SC-558) demonstrates that the sulfonamide moiety wedges into a
hydrophobic ‘side pocket’ of COX-2 bordered by a valine residue (Val523) and
hydrogen bonds with a neighbouring arginine residue (Arg513) and the peptide
bond of Phe518.29 A similar hydrophobic side pocket off the main channel in
COX-1 is not accessible because of the presence of an isoleucine instead of
valine at position 523, which sterically hinders inhibitor approach. The COX-2
mutant V523I is resistant to time-dependent inhibition by diarylheterocycles
but not carboxylic acid type NSAIDs.32,33 Conversely, the COX-1 mutant
I523V is sensitive to time-dependent inhibition by diarylheterocycles.33
Movement of Val523 and insertion of the sulfonamide or methylsulfone moiety
into the side pocket is thought to contribute to the time dependence of inhi-
bition by diarylheterocycles.

5.1.2.4 N-Alkylsulfonanilide Class of Selective COX-2 Inhibitors

As was the case with diarylheterocycles, investigations on the anti-inflamma-
tory properties of N-alkylsulfonanilides (Figure 5.3) such as the marketed
NSAID nimesulide also began in the 1960s.34–36 In the mid-1990s, the obser-
vations on the potent and selective COX-2 inhibition by the structurally related
sulfonanilide analog NS-39811,37 provided a biochemical rationale for the anti-
inflammatory effects of alkylsulfonanilides like nimesulide38 and flosulide,39
and led to a general resurgence in this class of compounds. Upon re-evaluation
218 Chapter 5

O O O O
H S H S
N CH3 N CH3

O O
R

R1 = H (crucial for potency) F F

R2 = CH3, CF3 NO2

O
O O
R1 S NS-398: R = Flosulide
N R2
Nimesulide: R =
O
R3
O O O O
H S H S
X N CH3 N CH3
R3 = cycloalkyl, aromatic O
or heterocyclic S

X = electron withdrawing group O

(can be part of a ring) F F
CONH2 O
N
FR115068 H
O
T-614

Figure 5.3 SAR on N-alkylsulfonanilides as selective COX-2 inhibitors.

for COX inhibition, both nimesulide and flosulide revealed their selective
COX-2 inhibitory properties.40
The common structural features of the alkylsulfonanilides are depicted in
Figure 5.3. The alkyl substituent is typically methyl, but halogenated methyl
substituents such as trifluoromethyl have been reported.41 The ortho-anilide
substituent typically includes aryl, heterocyclic or cycloalkyl ethers and thioe-
thers. The para-anilide substituent invariably bears an electron-withdrawing
group that may be incorporated as part of a ring; absence of the electron-
withdrawing group results in inactive compounds. A variety of methane-
sulfoanilides with different para-electron-withdrawing groups have been
evaluated as selective COX-2 inhibitors and as orally active anti-inflammatory
agents. Substituents include para-acetyl, para-cyano, para-carboxamido
(FR115068), nitro (NS-398 and nimesulide) and para-sulfonamido.42 Another
structural variant is the incorporation of the para-electron-withdrawing
group as part of a ring system as seen in flosulide and the sulfonanilide analog
T-614.43 The phenyl ring in the alkylsulfonanilide scaffold can be replaced with
the electron-withdrawing pyridine ring.42,43
The crucial requirement of the electron-withdrawing group in COX-2
inhibition by alkylsulfonanilides is most likely related to a lowering of the pKa
of the sulfonamide moiety in the range of the carboxylic acid group (the pKa of
nimesulide is B6.4–6.8, whereas that of the corresponding pyridinyl derivative
is B2.9844,45), which then allows for an efficient ion pairing interaction with a
complementary active site amino acid residue(s). This is demonstrated in the
Sulfonamide as an Essential Functional Group in Drug Design 219
NS-398-COX-2 crystal structure complex wherein the sulfonamide group ion
pairs to Arg120 in a manner similar to carboxylic acid containing NSAIDs
rather than inserting into the Val523 side pocket like the diarylheterocycles.46
Consistent with the crystal structure findings are the SAR findings that N-
methylation of the sulfonamide nitrogen in alkylsulfonanilides generates
inactive compounds.41 Unlike the methylsulfone- or sulfonamide-based dia-
rylheterocycles, the structural basis for COX-2 selectivity by NS-398 is not
evident from the COX-2 co-crystal structure.

5.1.3 Sulfonamide-based Carbonic Anhydrase Inhibitors

Carbonic anhydrases (CAs, EC 4.2.1.1) are ubiquitous zinc enzymes which are
encoded by three distinct, evolutionarily unrelated gene families; in humans, 16
different CAs isozymes (a-CAs) or CA-related proteins have been described
with very different sub cellular localisation and tissue distribution.47 The iso-
zymes catalyse a very simple physiological reaction, i.e. the introversion
between carbon dioxide and the bicarbonate ion—see Eqn (5.2). Because CO2
is generated in high amounts in all living organisms, CAs are involved in crucial
physiological processes connected with respiration and transport of CO2/
bicarbonate between metabolising tissues and lung, pH and CO2 homeostasis,
electrolyte secretion in a variety of tissues/organs, biosynthetic reactions (e.g.
gluconeogenesis, lipogenesis and ureagenesis), bone resorption, calcification,
tumorigenicity, and many other physiologic and pathologic processes.

O
+ H2O HCO3 + H ð5:2Þ
C
O

5.1.3.1 Mode of Inhibition of CA by Sulfonamides

X-ray crystal structure has been determined for several a-CA isozymes.48–55
The metal ion (which is Zn21 in all a-CAs) is essential for catalysis.47 Crys-
tallographic analysis indicates that the active site metal ion is situated at the
bottom of a 15Å long active site cleft and is coordinated to three histidine
residues (His94, His96 and His119) and a water molecule/hydroxide ion. The
zinc-bound water is also hydrogen bonded with the hydroxyl group of Thr199,
which in turn is bridged to the carboxylic acid moiety of Glu106; these inter-
actions enhance the nucleophilicity of the zinc-bound water molecule, and
orient the substrate (CO2) in a favourable position for nucleophilic attack by
the hydroxide ion bound to Zn21 (Figure 5.4).
Primary (and in some cases secondary) sulfonamides attached to an aromatic
or heterocyclic scaffold are potent inhibitors of CA enzymes. X-ray crystal-
lographic structures are available for many adducts of sulfonamide inhibitors
with CA isoforms.48–55 In all cases, sulfonamides bind in a tetrahedral geometry
of the Zn21 ion (see Figure 5.4) in a deprotonated state, with the sulfonamide
nitrogen coordinated to Zn21 and an extended network of hydrogen bonds,
220 Chapter 5

Hydrophilic part Hydrophobic part

of active site of active site

O H O R
Thr199 S
N O
N HN
H
O H Zn2+
H Thr199 O His119
H
O O H His94 His96
O
Glu106 Zn2+
O His119
His94 His96 O

Glu106

Catalytic conformation Inhibitory mode

Figure 5.4 a-CA inhibition by primary aromatic and heterocyclic sulfonamides.

involving residues Thr199 and Glu106. In such a conformation, the ionised

sulfonamide NH group displaces the zinc-bound hydroxide ion. The aromatic/
heterocyclic portion of the sulfonamide molecule interacts with hydrophilic and
hydrophobic amino acid residues within the active site cavity. This exquisite
network of interactions explains why sulfonamides selectively interact with
CAs, in contrast to other functional groups (e.g. carboxylic acid, hydroxamic
acid, phosphates and thiols) that are also capable of coordinating with zinc
metal in other enzymes. Quantitative SAR (QSAR) for CA inhibition by sul-
fonamides has been studied in detail by Clare & Suparan.56 They concluded
that effectiveness of ionisation of the SO2NH2 group itself is a key contributor
to inhibitory potency. The more easily the functional group can ionise the more
potent the inhibitor. Otherwise lipophilicity is a modifying influence, as is the
nature of the aromatic ring system if present in the molecule.

5.1.3.2 Clinical Applications of CA Inhibitors

There are over 20 clinically used primary sulfonamide drugs that possess sig-
nificant CA inhibition. The structures of representative compounds are shown
in Figure 5.5. In addition, several novel chemotypes that contain sulfonamide,
sulfamate or sulfamide groups also have been reported to inhibit CA iso-
zymes.57–62
Many CA inhibitors such as acetazolamide, methazolamide, ethoxzolamide,
sulthiame and dichlorophenamide have been used in the clinic for decades as
antiglaucoma agents. They were initially developed in the search for novel
diuretic, antihypertensives or antiepileptic agents in the 1950s and the 1960s.
Their discovery stemmed from the observation of metabolic acidosis (lowered
blood pH due to excess production of H1 or inability of the body to form
HCO3) as a side effect of sulfanilamide therapy, which then led to the
Sulfonamide as an Essential Functional Group in Drug Design 221

N H3 C N
O N O N N N SO2NH2
SO2NH2 SO2NH2
SO2NH2 S O
H3C N S H3 C N S
H O S O
Methazolamide Ethoxzolamide Sulthiame
Acetazolamide

SO2NH2 NH
NH

SO2NH2
SO2NH2
Cl SO2NH2 H3C S S H3CO N
S S
Cl O O
O O
Dichlorophenamide Dorzolamide Brinzolamide

SO2NH2 O
H
N Cl
O
N NH
HOOC SO2NH2 O S
O
O
Furosemide Zonisamide Saccharin

Figure 5.5 Sulfonamide-based inhibitors of CA enzymes.

synthesis of additional sulfonamide derivatives to exploit and maximise this

effect as a potential therapy. By producing bicarbonate-rich aqueous humor
secretion (mediated by ciliary CA isozymes) within the eye, CAs are involved in
vision, and their malfunctioning leads to high intraocular pressure (IOP) and
glaucoma.63 It is well-established that effective reduction of IOP can be
achieved by the systemic administration of CA inhibitors, which act by redu-
cing the rate of aqueous humor secretion. However, systemic therapy, which
requires large dosages of CA inhibitors to obtain reductions in IOP, con-
comitantly evokes a wide array of undesirable side effects resulting in poor
patient compliance.64,65 The hypothesis that the undesirable side effects occur
via CA inhibition in extraocular tissues led to the search for topical CA inhi-
bitors with direct ocular administration. In the 1990s two novel compounds,
dorzolamide and brinzolamide (see Figure 5.5), were developed for topical
application for glaucoma.66 Additional examples of drugs that inhibit CA
enzymes include the diuretic furosemide, the anticonvulsant zonisamide, and
even the artificial sweetener and cyclic acylsulfonamide, saccharin.67–69 All
these drugs contain the sulfonamide motif required for binding to the metal ion
in the CA isozymes.
Of late, selective COX-2 inhibitors celecoxib and valdecoxib (but not rofe-
coxib) have been shown to function as potent CA inhibitors.70 The solution of
the crystal structures of these coxibs with CA indicates that the mode of
inhibition is identical to that discerned with other sulfonamides.71,72 As such,
the inhibition of CA isozymes by sulfonamides are diverse and non-specific,
which may provide a rationale for the different clinical applications of the CA
inhibitors ranging from diuretics and antiglaucoma agents, and more recently
as anticancer and anti-obesity drugs.73
222 Chapter 5
5.1.4 Sulfonylurea-based Hypoglycemic Agents
Sulfonylurea derivatives are a class of orally active hypoglycemic drugs that have
been used in the management of diabetes mellitus type 2 (‘adult onset’) for several
decades. They act by increasing insulin release from the b cells in the pancreas.
The discovery of sulfonylureas as hypoglycemic agents stemmed from a
serendipitous clinical finding in the 1940s by Janbon and Loubatières who
reported the blood sugar-decreasing effect of a sulfanilamide derivate, sulfoi-
sopropyl thiadiazole, while treating bacterial infections.74,75 Although this
sulfonamide derivative was not proven to be useful in the treatment of diabetes,
the information was exploited further in the discovery/development of novel
sulfonamide derivatives, which eventually led to the discovery of carbutamide
(Figure 5.6) as the first orally active sulfonylurea derivative for the treatment of
hypoglycemia.76 The structural similarity of carbutamide with prototypic sul-
fanilamide derivatives is obvious.

5.1.4.1 Pharmacological Mechanism of Action

Sulfonylureas bind to the complex of the ATP-dependent K1 (KATP) ion and
sulfonylurea receptor (SUR1) on the cell membrane of pancreatic b-cells. This
results in the inhibition of a tonic, hyperpolarising efflux of potassium, thus
causing the electric potential over the membrane to become more positive. This
depolarisation opens voltage-gated Ca21 channels.77,78 The rise in intracellular
calcium leads to increased fusion of insulin granulae with the cell membrane
and therefore increased secretion of (pro)insulin. There is some evidence that
sulfonylureas also sensitise b-cells to glucose, limit glucose production in the
liver, decrease lipolysis (breakdown and release of fatty acids by adipose tissue)
and decrease clearance of insulin by the liver. Recently, evidence has been
presented which demonstrates a weak peroxisome proliferator-activated
receptor g agonism by some sulfonylurea-based drugs.79,80

5.1.4.2 SAR Relationships for the Hypoglycemic Activity

All sulfonylureas contain a central S-phenylsulfonylurea scaffold with para-
substitution on the phenyl ring and various groups terminating the urea
nitrogen end group. Sulfonylureas are classified into first- and second-genera-
tion compounds (Figure 5.6). The drugs of these two generations have in
common the sulfonylurea group; they differ mainly by the potency of their
insulin-releasing and anti-diabetic actions, the second generation compounds
being the more potent.
The observation that the sulfonylurea group is not essential for drug activity
was first demonstrated with the sulfonylurea glibenclamide; replacement of the
sulfonylurea motif with a carboxylic acid bioisostere led to the discovery of
meglitinide, a second generation hypoglycemic agent without the sulfonylurea
group.81 Non-sulfonylurea derivatives also display insulin-secreting and anti-
diabetic actions.82 These drugs (glinides) act on the same cellular targets as do
 First Generation Sulfonylureas

O O O O O O O
O O
S (CH2)3CH3 S (CH2)3CH3 S (CH2)3CH3
N N N N N N
H H H H H H
H2N H3C Cl
Carbutamide Tolbutamide Chlorpropamide

Second Generation Sulfonylureas

CH3
O O O O
O O
S H3 C S
O O N N H3C COOH
H H O O N N O O
H H
N N N
H N
Sulfonamide as an Essential Functional Group in Drug Design

H H
H3C
Glimepiride Glibenclamide meglitinide
H3 C
Cl Cl

Figure 5.6 Representative examples of first and second generation sulfonylurea-based hypoglycemic agents.
223
224 Chapter 5
the sulfonylureas, namely they close the KATP channels, although they are
presumed to act in different regions of the SUR1 receptor.

5.1.5 Miscellaneous Applications of the Sulfonamide Group in

Medicinal Chemistry
5.1.5.1 Sulfonamides as Phenol Bioisosteres
The alkysulfonamide group is commonly used as a non-classical bioisostere for
the phenol substituent as it has similar pKa values (B8.0) to that of the phe-
nolic hydroxyl group. For instance, during the lead optimisation of a gona-
dotropin-releasing hormone antagonist, methanesulfonamide replacement of
the phenol in compound 5.4 resulted in bioisostere 5.5, with a four-fold increase
in binding affinity (Figure 5.7).83 However, the lack of generality when trans-
ferring a certain type of bioisosteric transformation between lead compounds
for different therapeutic targets becomes apparent in the next example. When
the methanesulfonamide replacement was applied to the opioid antagonist
naltrexone (5.6) and its receptor agonist 5.7, the corresponding sulfonamide
derivatives 5.8 and 5.9, respectively, were totally inactive in vitro, despite having
similar pKa values (see Figure 5.7).84 The lack of binding affinity has been
attributed to the steric bulk of the sulfonamide group, which makes this motif,
an unsuitable bioisostere in this particular example.

5.1.5.2 Sulfonamides as a-Ketoamide Bioisosteres

Sulfonamides also appear to function as a-ketoamide isosteres based on a
recent study dealing with the optimisation of azaindole derivatives as human

R
R R
N N
OH OH

(H2C)4 O O
NH
O O
CH3 OH HN CH3
S
O O
N 5.6: R = CH2CH(CH3)2 5.8: R = CH2CH(CH3)2
H 5.7: R = CH3 5.9: R = CH3
CH3
5.4: R = OH
5.5: R = NHSO2CH3

Figure 5.7 Successes and failures in the utility of the methanesulfonamide group as a
phenol bioisostere.
Sulfonamide as an Essential Functional Group in Drug Design 225
O
O
O CH3 O N
O N
N S

H3 C
N
N N
H N N
H
O
5.10 5.11
IC50 (antiviral) = 30 nM IC50 (antiviral) = 860 nM

O O
O O
H3C O O N N
O H3C O F O N
O N S
N S

N
N N
N N N
H N H
H
O N
5.12 5.13 N 5.14
IC50 (antiviral) = 8 nM IC50 (antiviral) = 9200 nM N IC50 (antiviral) = 7 nM

Figure 5.8 The sulfonamide group as an a-ketoamide bioisostere.

immunodeficiency virus entry inhibitors.85 Flexible overlay calculations on

a-ketoamide and sulfonamide derivatives, 5.10 and 5.11 (Figure 5.8) revealed
that the sulfonamide group was in close proximity to the ketoamide group, with
one sulfonamide oxygen tightly overlaid on the amide nitrogen. The other
sulfonamide oxygen lies B1.1Å from the ketone oxygen. These results also
suggest that energetically accessible conformations of a-ketoamide and sulfo-
namide have similar dispositions of potentially pharmacophoric aromatic,
lipophilic and H-bond acceptor groups. Although direct replacement of the
a-ketoamide group in the azaindole derivative (e.g. compound 5.12) with a
sulfonamide (e.g. compound 5.13) led to a loss in potency, subsequent SAR
studies led to sulfonamide derivative 5.14 that possessed pharmacology
comparable to 5.12 (see Figure 5.8).

5.1.5.3 Acylsulfonamides (Sulfonimides)

Replacement of the hydroxyl moiety of a carboxylic acid with a phenylsulfo-
namide results in the formation of a sulfonimide derivative. The estimated pKa
values for sulfonimides are similar to that of an aryl carboxylic acid (pKa
B3.5–4.0). An example of the utility of this bioisostere becomes evident in the
case of the indole-based leukotriene antagonists where there was little difference
between sulfonamide 5.16 and its parent compound 5.15 (Figure 5.9) with
respect to their ability to antagonise the cysteinyl leukotriene (cysLT) recep-
tor.86 Likewise, sulfonamide replacement of the hydroxyl group in carboxylic
acid derivatives and serine protease inhibitors 5.17 and 5.18 led to 5.19 and
5.20, respectively, with significantly improved inhibitory potency (Ki values of
0.2–0.6 mM for the sulfonimides versus 1.8–1.9 mM for the parent acids) against
the hepatitis C virus (HCV) NS3 protease.87 A similar observation was noted in
a study on CXCR2 receptor antagonists where acylsulfonamide replacement
226 Chapter 5

O
O O
N H
C5H11 N N
H O N R
(CH2)5 H
O NH O
O HO (CH2)2
X
5.15: R = OH N COOH
5.16: R = NHSO2C6H5 O
H
H
N COOH
O 5.17: R =
R
H
(CH2)3 N COOH
NC 5.18: R =
F
N
H
5.21: R = OH O O
5.22: R = NHSO2CH3 H
N
O N R
H
O CH3 O NH O
HO (CH2)2
N

R O N COOH
H
Cl O O O
H
Zomepirac: R = COOH N S
5.19: R = N
H
IC50 (COX-2) = 2.0 µM
IC50 (COX-1) = 0.3 µM O O O
I
H
H N S
N 5.20: R = N
S
5.23: R = O
H
O O
IC50 (COX-2) = 0.9 µM
IC50 (COX-1) = 242 µM

Figure 5.9 Acylsulfonamides (sulfonimides) as carboxylic acid bioisosteres.

(5.22) for the carboxylic acid derivative (5.21) and CXCR2 receptor antagonist
improved antagonist potency.88
A noteworthy example of the impact on primary pharmacology upon bio-
isosteric replacement of the carboxylic acid OH group with a sulfonamide
moiety was evident during studies on the COX-1-selective NSAID zomepirac.
Conversion of COX-1-selective inhibitor zomepirac into a COX-2-selective
inhibitor was achieved by simply replacing its carboxylic acid group with an
sulfonimide bioisostere (5.23) (Figure 5.9).89 Although 5.23 has been co-
crystallised with COX-2, the crystal structure does not provide a rationale for
COX-2-selective inhibition by the bioisostere.89 The acylsulfonamide motif in
5.23 breeches the constriction at the mouth of the COX active site and projects
Sulfonamide as an Essential Functional Group in Drug Design 227
into the sterically unconjugated lobby region. The sulfonamide portion of the
inhibitor hydrogen bonds to the Arg120, Glu524 and Tyr355 in COX-2 in a
manner similar to the carboxylic acid-based NSAIDs. Because the three amino
acid residues are conserved in COX-1 and COX-2, their importance in deter-
mining the selectivity of the 5.23 for COX-2 inhibition remains uncertain. As
such, replacement of the carboxylic acid moiety in several NSAIDs (e.g.
indomethacin, sulindac, meclofenamic acid, etc.) with bioisosteres (e.g. esters
and amides) has provided a facile strategy for converting non-selective or COX-
1-selective inhibitors into potent and selective COX-2 inhibitors.90–92

5.1.5.4 Discovery of Zafirlukast–a Cysteinyl Leukotriene

Receptor Antagonist and Sulfonimide Derivative
Designing antagonists of the cysLT receptors as orally active anti-asthma drugs
has paid off significantly as evident from the commercial success of mon-
telukast, zafirlukast and pranlukast (Figure 5.10). Peptidoleukotrienes LTC4,
LTD4 and LTE4 are potent constrictors of smooth muscle and, as such, have
been implicated since the 1970s in the development of asthma.93,94 FPL 55712,
a chromane carboxylic acid, which was discovered as a prototype cysLT
receptor antagonist, provided medicinal chemists with a starting point in
structure-based drug design.95 A visual examination of the structures of LTD4
and FPL 55712 suggests that the hydroxyacetophenone region in FPL 55712
mimics the olefinic region of the leukotriene, while the chromane carboxylic
acid segment mimics either the backbone C1–C5 carboxylic acid region or the
peptidic component of LTD4 (Figure 5.10). Exhaustive SAR studies to opti-
mise antagonist potency of FPL 55712 led to several FPL 55712 derived
compounds that were then clinically evaluated.96 Of these, tomelukast
(Figure 5.10) was arguably the most important for many reasons, foremost
amongst which was the demonstration of anti-asthmatic activity in the clinic.
Furthermore, from an SAR perspective, tomelukast demonstrated that sub-
stitution of the carboxylic acid motif in FPL 55712 with the tetrazole bioi-
sostere could potentially result in compounds that exhibit increases in in vitro
and in vivo potency.97 The 600 mg B.I.D. dosing regimen required for clinical
efficacy, however, suggested that even further increases in potency would be
required for a low daily dose anti-asthmatic agent.98 The search for more
potent cysLT receptor antagonists led to pranlukast, the structure of which, can
be considered to be a hybrid between FPL 55712 and tomelukast due to the
presence of the chromane tetrazole motif (Figure 5.10).99 Pranlukast was the
first cysLT receptor antagonist approved for marketing and is currently sold in
Japan.100
The evolution of sulfonimide derivative zafirlukast also proceeded through
initial SAR analyses on FPL 55712-like compounds and led to the fabrication of
an indole-containing lead compound (5.24) (Figure 5.11).101 Modifications were
simultaneously made on three regions of this molecule: the lipid-like tail region,
the acidic head region, and the indole backbone. With regards to exploration of
 228

lipid OH acid
OH
COOH
HO O
S H
O O HO O N
O N
C5H11 H2N O N N
peptide O
NH
O COOH
COOH
FPL 55712 Tomelukast
LTD4

O
H CH3
COOH
N N
N O O
H N CH3
N OH O H3C
N S
O O N
H H
(CH2)4 Cl N N S

O O
O
Pranlukast Montelukast
Zafirlukast

Figure 5.10 Evolution of cysteinyl leukotriene receptor antagonists for the treatment of asthma.
Chapter 5
 CH3
N
O CH3
O H3 C
O N
H H
N S
O CH3
O O
C4H9 O Zafirlukast O
N N
H H
Lipid-like tail Indole backbone N S
O
O CH3 O
O
C 4 H9 O
N N CH3
H
O CH3 N
O CH3
COOH O O
O N N
Acid head group O N
H H H H
N S
5.24 N S
O O
Sulfonamide as an Essential Functional Group in Drug Design

O O
O O

Figure 5.11 SAR studies that led to the discovery of zafirlukast.

229
230 Chapter 5
the acidic head region, SAR studies demonstrated that a preferred linkage
occurred with a 3-methoxy aryl group; however, replacement of the carboxylic
acid head group with phenylsulfonimide produced an approximately 100-fold
increase in potency. More importantly, this change was additive to other
alterations made in the amide region of the tail with in vitro potency B41000-
fold relative to FPL 55712. Although a number of variations of this theme
demonstrated potent in vitro antagonist activity, as such, these substituents had
a negative impact on oral activity.102,103 Further SAR optimisation indicated
that the inverted indole backbone also maintained potent activity against the
receptor, albeit with very poor rat oral bioavailability (o1%); incorporation of
ortho-tolyl sulfonamide substituent led to zafirlukast with oral bioavailability of
B68% and 67% in rats and dogs, respectively.102–105

5.2 Absorption, Distribution, Metabolism and

Excretion of Sulfonamides
5.2.1 Oral Absorption
With the exception of the topical antiglaucoma agents, most sulfonamide-based
drugs are administered by the oral route and hence absorption into the systemic
circulation is essential for their pharmacological action in target tissue (e.g.
kidney for the thiazides, and loop diuretics and pancreas for the sulfonylureas).
A common approach towards gaining an understanding of the balance of
physiochemical properties that leads to a suitable pharmacokinetic profile after
oral administration is to examine whether or not the compounds ‘fit’ with
Lipinski’s ‘rule of five’.106 Using a dataset of 2245 compounds from the World
Drug Index, Lipinski found that approximately 90% of the compounds
(excluding phosphates, polymers and quaternary ammonium ions) had a
molecular weight o500, cLogP o5, sum of hydrogen bond donors (as a sum of
NH and OH) o5, and sum of hydrogen bond acceptors (as a sum of N and O)
o10.106 Lipinski proposed that poor absorption and permeation are more
likely when two or more of these limits are exceeded.106 In addition to the
molecular properties discussed by Lipinski, descriptors such as polar surface
area (PSA) have been shown to reliably predict oral bioavailability, particularly
for organic anions (e.g. carboxylic acid).107 For instance, the negative impact of
a high PSA 4150Å2 and excessive (410) hydrogen bond donor or acceptor
properties on intestinal absorption/permeability has been recognised.107–110
An assessment of the physicochemical properties of 111 marketed sulfona-
mide-based drugs is depicted in Table 5.2. Approximately 90% of the
drugs comply with Lipinski’s rule of five for successful orally active agents.
Examination of the structures of the individual compounds reveals that
the sulfonamide motif is present within a vast chemical space with
neutral molecules being most common followed by an equal number of
acids and bases. The mean range of PSA, hydrogen bond donors and hydrogen
bond acceptors for the marketed sulfonamides lies within the confines of the
Sulfonamide as an Essential Functional Group in Drug Design 231
Table 5.2 Physicochemical properties of the 111 marketed sulfonamide con-
taining drugs.
Property Mean Range

Molecular weight 368 172–606

CLogP 1.8 2.0–7.8
cLogD @ pH 7.4 0.5 4.9–4.2
Polar surface area (PSA) 106Å2 44–200
Hydrogen bond donors (HBD) 3 0–7
Hydrogen bond acceptors (HBA) 4 1–12
Rule of five compliant 90% –
Acids 20% (22 out of 111)
Bases 15% (17 out of 111)
Neutrals 63% (70 out of 111)
Zwitterions 2% (2 out of 111)

physiochemical space required for good oral absorption of acidic, basic or

neutral molecules.

5.2.1.1 Sulfonamide-based Prodrugs–Discovery of Parecoxib

In terms of aqueous solubility, neutral sulfonamides celecoxib and valdecoxib
are slightly more soluble than the methylsulfone rofecoxib (Table 5.3).111,112
However, upon comparison of the solubility parameters with carboxylic acid
containing NSAIDs derivatives, neutral sulfonamides possess considerably
lower solubility (see Table 5.3). The absorptive apical to basolateral perme-
ability (AB) as measured in the Caco-2 cell line is comparable for the coxibs and
the various sub-classes of NSAIDs—a feature consistent with good oral
absorption.111 Lack of asymmetry in the basolateral to apical (BA) direction
suggests that efflux transport proteins such as P-glycoprotein or breast cancer
resistant protein will not negatively impact oral absorption. The PSA estimates
for celecoxib (86.4), valdecoxib (94.6) and rofecoxib (68.8) lie within the range
that predicts moderate to good oral absorption.
Consistent with the in vitro observations, the coxibs are rapidly absorbed
(Tmax B2.0–3.0 h) in humans following oral administration at their effica-
cious doses. Differences in physiochemical attributes (e.g. solubility and
lipophilicity) do appear to play a role in the oral absorption as illustrated
with celecoxib and valdecoxib. Systemic exposure as judged from Cmax
and AUC0N for valdecoxib (at its efficacious dose of 10–20 mg) is
294  118 ng mL1 and 3306  1300 ng hr1 mL1, respectively.113 The
findings from the human mass balance studies on valdecoxib that less than
1% of the administered radioactive dose is recovered in unchanged form in
faeces suggests near complete oral absorption of valdecoxib.114 In the case of
celecoxib, Cmax and AUC0N at the efficacious dose of 200 mg are 797 -
498.8 ng mL1 and 7600  5500 ng hr1 mL1, respectively.115 The lower oral
systemic exposure of celecoxib relative to valdecoxib relates to a greater
 232

Table 5.3 Structure, physiochemical properties, dose and Caco-2 cell permeability of carboxylic acid based NSAIDs and coxibs.
PCaco-2106 cm/sec Equilibrium solubility
Compound MW pKa Dose (mg) A to B B to A (pH 7.4) (mg/mL)

Celecoxib 381 11.1 100–200 17.6 15.1 0.005

H2NO2S

N
N CF3

H3 C

Valdecoxib 314 11.0 10 20.0 21.1 0.010

H2NO2S
CH3

O
N
Chapter 5
Rofecoxib 314 – 12.5–25 24.1 19.7 0.0009

H3CO2S

Diclofenac 295 4.2 50 20.2 21.3 15.9

COOH
Cl
H
N

Cl
Sulfonamide as an Essential Functional Group in Drug Design

Ibuprofen 206 4.4 200–800 10.1 19.8 2.30

H3C CH3

COOH

CH3
233
 234

Table 5.3 (continued )

PCaco-2106 cm/sec Equilibrium solubility
Compound MW pKa Dose (mg) A to B B to A (pH 7.4) (mg/mL)

Indomethacin 357 4.5 25–50 23.8 33.0 1.30

COOH
H3CO
CH3
N

O Cl

Mefenamic acid 241 4.2 250 17.9 22.2 0.10

CH3 COOH
H
N

CH3
Chapter 5
Piroxicam 331 1.8, 5.1 10–20 24.1 19.7 0.26

OH O

N N
H
N
S CH3
O O

Meloxicam 351 1.1, 4.2 7.5–15 17.6 15.1 0.46

OH O N
CH3
N S
H
N
S CH3
O O
Sulfonamide as an Essential Functional Group in Drug Design
235
236 Chapter 5
propensity to undergo first pass metabolism, which is apparent upon com-
parison of the oral clearance values (valdecoxib: CL/F ¼ 1.7 mL min1 kg1;
celecoxib: CL/F ¼ 9.5 mL min1 kg1).113–116
The relatively poor aqueous solubility of COX-2 inhibitors like valdecoxib
can be considerably improved by conversion of the primary sulfonamide to the
corresponding sulfonimide derivative 5.25 (Figure 5.12).117 Conversion to the
sulfonimide derivative imparts acidic character to the sulfonamide motif and
consequently provides the options of preparing salts to improve upon the
aqueous solubility. Compared with valdecoxib, whose aqueous solubility is
0.010 mg mL1, the solubility of 5.25 (as the corresponding sodium salt) is
44 mg mL1. Although 5.25 is inactive as a selective COX-2 inhibitor in vitro,
the compounds displays potent anti-inflammatory activity following oral or
intravenous administration in rats due to rapid amidase-mediated hydrolysis of
the sulfonimide moiety (T1/2 ¼ 15 min) to the active ingredient valdecoxib.
However, resistance of 5.25 towards hydrolytic cleavage in dogs, monkeys and
human liver preparations precluded further studies on this compound. SAR
studies revealed that extension of the alkyl group attached to the sulfonimide
nitrogen led to compounds (e.g. 5.26 and 5.27) (Figure 5.12) that retained the
aqueous solubility characteristics of 5.25 while undergoing facile hydrolytic
cleavage in animals and human liver preparations. Of these compounds, 5.26
(parecoxib sodium) was chosen for clinical development; a randomised, double-
blind, placebo-controlled, multiple-dose study in healthy human volunteers
demonstrated that parecoxib was rapidly absorbed (TmaxB30 min) and con-
verted to valdecoxib (elimination T1/2 ¼ 0.69 h).118 Steady state drug levels were
achieved within seven days and linear pharmacokinetics discerned for both the
prodrug and the active compound. Parecoxib sodium has been shown to be well
tolerated with no clinically significant adverse events observed in this study.
Parecoxib sodium is marketed in the European Union as Dynastat for treat-
ment of post-operative pain.119

N N
O O

CH3 Na CH3

H2N N
S R S
O O
O O
O
Valdecoxib 5.25: R = CH3
5.26: R = CH2CH3
5.27: R = (CH2)2CH3

Figure 5.12 Sulfonimide prodrugs of valdecoxib: discovery of parecoxib sodium for

parental administration for post-surgical pain management.
Sulfonamide as an Essential Functional Group in Drug Design 237
In the case of sulfonylurea hypoglycemic drugs, it is interesting to note that
the SUR1 receptor is a member of the ABC (adenosine-5 0 -triphosphate (ATP)-
binding cassette). Consequently, some sulfonylureas have been shown to inhibit
ABC efflux transporters such P-glycoprotein or multidrug resistance protein
(MRP1);120,121 furthermore, there is evidence that glyburide is a breast cancer
resistant protein and MRP3 substrate in transfected cell lines overexpressing
these transporters.122 The impact of these findings on sulfonylurea oral phar-
macokinetics and drug–drug interaction potential is unknown.

5.2.2 Distribution
Because sulfonamide groups are often used as non-classical carboxylic acid
bioisosteres, the affinity of the functional group towards distribution in tissue
proteins relative to plasma proteins is often similar to that discerned with
carboxylic acid containing drugs. In other words, the distribution volumes (Vd)
of drugs containing the two functional groups in a bioisosteric relationship are
often comparable. Modulation of Vd for sulfonamide-containing drugs can be
influenced by lipophilicity (clogP and/or clogD) and acidity (pKa) in a manner
similar for compounds of all physiochemical classes (acids, bases and neu-
trals).123 Thus, increases in lipophilicity generally coincide with increases in Vd,
and a more acidic character of the sulfonamide group will generally lead to a
lower Vd owing to extensive plasma protein (e.g. serum albumin) as discerned
with carboxylic acid analogs. These principles were outlined in Chapter 2. The
pharmacokinetic attributes of structurally diverse sulfonamide derivatives124 in
humans are shown in Table 5.4 and provide a glimpse of the relationship
between physiochemical properties and pharmacokinetic parameters such as
plasma free fraction (fu), Vd and clearance.

5.2.2.1 Preferential Red Blood Cell Partitioning of Sulfonamide-

based CA Inhibitors
An interesting pharmacokinetic attribute of some sulfonamide derivatives is
their extensive binding to red blood cells, which typically results in blood to
plasma partitioning ratios 41. Preferential distribution into red blood cells is
more commonly discerned with primary sulfonamides due to their affinity for
CA isozymes, which are present within the erythrocyte and may act as a site for
sequestration of drug. In contrast with primary sulfonamides, hypoglycemic
sulfonylureas such as glibenclamide have been shown to preferentially partition
to plasma compartment (blood to plasma ratio B0.5).125 Because of this
property, it is possible that sulfonamides could also exhibit non-linear phar-
macokinetics as illustrated with the CA inhibitor L-693 612 (Figure 5.13).126 At
blood concentrations achieved in the linear dose range (0.05–0.25 mg kg1),
binding to CA results in extensive red blood cell partitioning of L-693 612, with
a constant low free fraction in plasma for elimination (measured in vitro blood
to plasma ratio of B400). In contrast, higher doses (5–25 mg kg1) saturate CA
 238
Table 5.4 Physiochemical and pharmacokinetic parameters of representative sulfonamide-containing drugs
Drug MW cLogP Log D7.4 fu CL(mL/min/kg) Vdss(L/kg)

Loop diuretics
Acetazolamide 222 1.13 0.48 0.04 0.65 0.37

N
N
SO2NH2
S
H3COCHN

Furosemide 330 1.90 0.13 0.01 1.60 0.12

COOH

O
H2NO2S N
H
Cl

Chlorthalidone 338 0.45 0.74 0.24 3.90 1.50

OH SO2NH2

NH Cl

O
Chapter 5
Thiazides
Hydroflumethiazide 330 0.21 0.33 N/A 9.70 2.20

O2
H2NO2S S
NH

F 3C N
H

Sulfonylureas
Chlorpropamide 276 2.35 0.32 0.03 0.05 0.19

O O
O
S
N N
H H
Cl

Glizipizide 445 2.57 0.14 0.02 0.56 0.16

O O
Sulfonamide as an Essential Functional Group in Drug Design

O
S
N N
H H
(H2C)2
O NH

N
N
239
 240

Table 5.4 (continued )

Drug MW cLogP Log D7.4 fu CL(mL/min/kg) Vdss(L/kg)
O O O
S
N N
H H
(H2C)2
Glyburide O NH
494 4.24 1.90 0.02 0.82 0.08
H3CO

Cl
Tolbutamide 270 2.50 0.39 0.05 1.90 0.12

O O O
S
N N
H H
H3C

Sulfanilamides
Dapsone 248 0.89 0.94 0.25 0.48 0.83
H 2N NH2

S
O O
Chapter 5
Sulfadiazine 250 0.10 0.66 0.44 0.55 0.29

H2 N
H
N N
S
O O N

Sulfamethoxazole 253 0.56 0.20 0.02 0.36 0.30

H 2N
H
N N
S
O
O O

CH3

Sulfisoxazole 267 0.22 0.22 0.08 0.30 0.17

H 2N
CH 3
Sulfonamide as an Essential Functional Group in Drug Design

H
N
S CH 3
O O O N
241
242 Chapter 5

CH3
Cl N
HN O
HO O HN
O NH2 H O
S O N S S
S O NH O O NH2 O
H3CO
S S NH2 CH3 N N S O
O O O NH2
Cl
L-693,612 Chlorthalidone Acetazolamide Indapamide

Figure 5.13 Examples of sulfonamides with preferential red blood cell partitioning
due to carbonic anhydrase binding.

binding sites in red blood cells; consequently, the free fraction available in
plasma for elimination is much greater (blood to plasma ratio reduces to B6),
resulting in significant increases in blood clearance. Likewise, the observed
dose-dependent increases in Vd of L-693 612 are consistent with the hypothesis
that high-affinity binding to CA confines the compound largely to the blood
compartment at low doses, but saturation of CA binding sites at high doses
increases availability to peripheral tissues.
Preferential partitioning of sulfonamides into red blood cells renders them
susceptible to significant pharmacokinetic interactions with agents that may
compete for binding to CA. For instance, a significant drug–drug interaction
has been noted in the clinic with the CA inhibitors and anti-hypertensive agents
chlorthalidone and acetazolamide (Figure 5.13), wherein acetazolamide was
able to displace chlorthalidone from blood cells following administration of the
two medications in humans.127 Thus, intravenous administration of acet-
azolamide to humans who had previously received [14C]-chlorthalidone resul-
ted in a marked drop in red blood cell radioactivity, whereas that in plasma
increased. A likely explanation for the pharmacokinetic interaction is the
relatively higher affinity of acetazolamide (versus chlorthalidone) towards
binding to CA as reflected from the greater than six-fold higher blood to
plasma ratio of acetazolamide; the blood-to-plasma ratio of acetazolamide and
chlorthalidone has been estimated to be 467 and 70, respectively.128 Drug–drug
interactions due to drug displacement from red blood cells have also been noted
with the anti-hypertensive agents and CA inhibitors indapamide (Figure 5.13),
chlorthalidone and acetazolamide in the rat. Both chlorthalidone and acet-
azolamide were able to displace indapamide from rat erythrocytes due to
competition for CA binding sites in vivo.129

5.2.3 Metabolism
Presence of the sulfonamide moiety in a molecule in a secondary or tertiary
amide expression usually provides an inert, metabolically robust group that is
not typically vulnerable to phase I or phase II metabolising enzymes. Contrary
to facile enzyme-catalysed hydrolysis of esters and amides, sulfonamides are
generally resistant to cleavage by amidases. An exception to the rule is sulfo-
nimide (acylsulfonamide) derivatives such as valdecoxib, which can undergo
Sulfonamide as an Essential Functional Group in Drug Design 243

N
O O N
N
S
N N
H
NH
H3C O

N O
O O N
N N
S O O N
N N N
S
N N N
H3C O
HN
Sildenafil O

Scheme 5.2 Biotransformation of sildenafil in preclinical species and human.

hydrolysis to the parent sulfonamide. A characteristic biotransformation

reaction associated with secondary or tertiary sulfonamides include N-deal-
kylation of the substituent attached to the sulfonamide nitrogen as highlighted
in the metabolism of the PDE5 inhibitor sildenafil in preclinical species and
human (Scheme 5.2).130

5.2.3.1 Biotranformation Pathways of Sulfonamides

In contrast with carboxylic acids, which are metabolised by phase II processes
such glucuronidation and amino acid conjugation, primary sulfonamide bio-
transformation can involve N-hydroxylation and/or N-glucuronidation via cyto-
chrome P450 (CYP) and uridine glucuronosyl transferease (UGT) enzymes,
respectively.
An example of this phenomenon becomes evident with valdecoxib, wherein a
significant component of its metabolic elimination in mice and humans involves
hydroxylation and glucuronidation of the sulfonamide moiety to yield meta-
bolites 5.28 and 5.29, respectively (Scheme 5.3).114,131 The N-hydroxylated
metabolite 5.28 of valdecoxib is further metabolised to the corresponding O-
glucuronide 5.30 (Scheme 5.3). Mass balance studies in humans reveal that
B25% of the administered dose of valdecoxib is eliminated via the combined
oxidative and glucuronidation pathway depicted in Scheme 5.3.114 The sulfo-
namide group in valdecoxib also appears to be susceptible to hydrolytic clea-
vage, considering that trace amounts (B0.5% of administered dose) of the
corresponding sulfinic and sulfonic metabolites 5.31 and 5.32, respectively,
have been detected in human urine.114
The mechanism of sulfonamide degradation to sulfinic and sulfonic acid
derivatives has been examined in some detail and does not appear to be a
straightforward enzymatic hydrolysis of the motif in a manner similar to that
244 Chapter 5

? HO S HO S

O O O
5.31 5.32

HOOC O O
HO O
HO O S
CYP UGT HO N
HN S OH H
N
O O
O 5.30
5.28

CH3
HOOC
H2 N S O H
UGT HO N
O O HO S
OH
Valdecoxib O O
5.29

Scheme 5.3 Metabolism of the sulfonamide group as illustrated with the selective
COX-2 inhibitor and anti-inflammatory agent valdecoxib.

discerned with esters and amides. Instead, the studies implicate N-hydroxyla-
tion of the sulfonamide group as the rate-limiting step in the formation of the
sulfinic and sulfonic acid metabolites.132–135 For example, the decomposition of
benzenesulfohydroxamic acid (Piloty’s acid) (Scheme 5.4) and some of its
derivatives under anaerobic and strong alkaline conditions has been reported to
yield benzenesulfinic acid and nitrous oxide (N2O).132,133 Under physiological
conditions (aerobic environment, phosphate buffer, pH 7.4), Piloty’s acid
decomposes to form benzenesulfinic acid and nitric oxide (NO), which can be
intercepted with a nitrone free radical trap.134 A mechanism, which depicts the
formation of NO in the oxidative decomposition of Piloty’s acid is shown
in Scheme 5.4.135 Consistent with the general instability of N-hydroxy-
sulfonamides under neutral aerobic conditions, attempts to isolate the N-
hydroxy metabolite of valdecoxib, i.e. 5.28, in pure form by chemical synthesis
were met with failure due to facile decomposition at room temperature to
afford the corresponding sulfinic acid derivative 5.31.136 However, work-up and
crystallisation of crude 5.28 using EDTA and ascorbic acid to remove sources
of chelating metal ion traces and/or oxidants has afforded a stable mono-
hydrate form of 5.28 whose water content remained constant at room tem-
peratures (under standard humidity conditions) over a period of two years.136
It is interesting to note that 5.28 is an active metabolite of valdecoxib; however,
the COX-2 inhibitory potency and selectivity of 5.28 is significantly lower than
that of the parent compound.
Finally, it is noteworthy to point out that the sulfonamide group in celecoxib
is resistant to N-hydroxylation or N-glucuronidation in animals and
humans.116,137 In valdecoxib, CYP mediated hydroxylation occurs on the sul-
fonamide nitrogen, the 5-isooxazylmethyl group as well on the unsubstituted
phenyl ring (Scheme 5.5); in contrast, celecoxib is exclusively metabolised via
hydroxylation of the para-tolyl methyl group (Scheme 5.5). The difference in
 O
O OH O OH
O O O OH
O O S O S
S O S O N OH
S O N N
S OH N
N NO
H H
H

Scheme 5.4 Decomposition of N-hydroxysulfonamides to sulfinic acid and nitric oxide.

Sulfonamide as an Essential Functional Group in Drug Design
245
 246

HOOC
O H
HO N
HO S
OH
O O
O
H3 C
N HO
UGT HO
O
CF3 CF3
CYP
OH N N CF3
N N
HO N N
N
O CYP
H2 N S H2 N S H2 N
CH3 S
O O Celecoxib O O
O O
H2 N S HO
O O HN
Valdecoxib S
O O

Scheme 5.5 Differences in metabolic profile of the selective COX-2 inhibitors and diarylheterocycle derivatives valdecoxib and celecoxib.
Chapter 5
Sulfonamide as an Essential Functional Group in Drug Design 247

CH3 CH3

O O N O O N
S S
N N CH3 N N O
H H
CH3
H2 N H2N
Sulfasomidine Sulfamethomidine

Figure 5.14 An example of structure–activity relationship for sulfonamide N-glu-

curonidation in sulfanilamides. The arrow indicates the site of
glucuronidation in sulfamethomidine, which is not discerned with
sulfasomidine.

metabolic profile between the two compounds is fairly intriguing and suggests
difference(s) in CYP active site binding modes.
Besides valdecoxib, sulfanilamide anti-bacterial drugs are also prone to N-
glucuronidation on the sulfonamide nitrogen. The degree to which N-glucur-
onidation occurs is highly dependent on the functional group attached to the
sulfonamide nitrogen atom as well as the species under consideration. For
example, the process of N-glucuronidation on the sulfonamide nitrogen in
sulfamethomidine is not observed with sulfasomidine (Figure 5.14).138 The
major structural difference is that the pendant 2-methoxy-4-methylpyrimidine
substituent in sulfamethomidine is replaced by the 2,4-dimethylpyrimidine
motif in sulfasomidine. Furthermore, while sulfonamide glucuronidation con-
stituted a major route of metabolism of sulfamethomidine in humans and
primates, the N-glucuronide metabolite was not detected in rats and rabbits.138

5.2.3.2 Role of CYP2C9 in the Oxidative Metabolism of

Sulfonamides
Considering the bioisosteric relationship between sulfonamides and carboxylic
acids, it is not surprising that the polymorphic CYP2C9 is involved in the
oxidative metabolism of numerous sulfonamide-based compounds in a manner
similar to that discerned with carboxylic acid drugs. Structural characteristics
of prototypic CYP2C9 substrates include the presence of an anionic group and
a hydrophobic zone between the substrate hydroxylation site and the anionic
site. Several noteworthy examples of drugs from the antibacterial (e.g. sulfa-
methoxazole and dapsone), antidiabetic (e.g. sulfonylureas) and anti-inflam-
matory (e.g. celecoxib and valdecoxib) class of compounds have been reported
to undergo CYP2C9-catalysed oxidation.139–144 Consequently, genetic poly-
morphisms of CYP2C9 markedly affect the pharmacokinetic (and pharmaco-
dynamic) attributes of the drugs145–147 Like carboxylate-containing drugs,
sulfonamide-based CYP2C9 substrates can be prone to pharmacokinetic
interactions with inhibitors or inducers of this isozyme.148–150
248 Chapter 5
The structural basis for CYP2C9-mediated oxidation of acidic substrates has
been deduced via the solution of the crystal structure of this isozyme complexed
with the anti-coagulant drug warfarin.151 For the most part, acidic function-
alities (e.g. carboxylic acid, sulfonamides, etc.) in drugs bind to a active site
Arg108 residue, which then positions the molecule in proximity of the heme
prosthetic group for subsequent oxidation. Using this basic binding mode,
molecular docking and/or homology modelling studies have been used to
elucidate the structural basis for CYP2C9 mediated metabolism of the sulfo-
nylurea gliclazide.152 However, oxidative metabolism of all sulfonamide (or for
that matter carboxylic acid) derivatives cannot be simply rationalised by the
hydrogen bond interaction between the anionic group and the arginine residue
as illustrated with celecoxib. While molecular docking studies on celecoxib
using the sulfonamide–arginine ion pair as a starting point is consistent with
orientation of the para-tolyl group towards the heme iron for hydroxylation
to occur, experimental studies on the metabolism of neutral derivatives of
celecoxib suggest otherwise.153
As shown in Table 5.5, in an attempt to disrupt the sulfonamide-arginine
interaction, the sulfonamide group of celecoxib was replaced with the corre-
sponding methyl sulfide (5.33), methyl sulfoxide (5.34) and methyl sulfone
(5.35), which represent different oxidation states on the sulfur atom. In contrast
with the expectation that replacement of the sulfonamide with methylsulfone or
methylsulfide would result in a lower propensity towards metabolism by
recombinant CYP2C9, in vitro studies showed no significant decrease in the
extent of metabolism for the sulfide or the sulfone derivative and only a modest
decrease (B48% consumption of sulfoxide). Thus, removing apparently
important interactions with the enzyme had only little to no influence on the

Table 5.5 SAR studies on sulfonamide replacements in celecoxib as CYP2C9

substrates.

R3

N N
R2

R1
% Metabolised by
Compound R1 R2 R3 CYP2C9 (1 mM)

Celecoxib -CH3 -CF3 -SO2NH2 94

5.33 -CH3 -CF3 -SCH3 87
5.34 -CH3 -CF3 -SOCH3 48
5.35 -CH3 -CF3 -SO2CH3 81
5.36 -CH3 -CF3 -COOH 29
Sulfonamide as an Essential Functional Group in Drug Design 249
metabolism of celecoxib. These results indicated that the hydrophobic core
structure of celecoxib has a more significant impact on the total interaction
with CYP2C9. What was even more surprising in this SAR exercise is the fact
that replacement of the sulfonamide group in celecoxib with a carboxylic acid
functionality (e.g. compound 5.36) resulted in a decrease in CYP2C9 catalysed
metabolism.

5.2.4 Renal Elimination

The process of secreting organic anions through the proximal tubule cells is
achieved by unidirectional transcellular uptake of organic anions into the cells
from the blood across the basolateral membrane, followed by extrusion across
the brush–border membrane into the proximal tubule fluid. The process is
catalysed by the human organic anion transporters (hOATs) 1, 2, 3 and 4.154
OAT1, OAT2 and OAT3 are localised on the basolateral side of the proximal
tubule, while OAT4 is localised on the apical domain of the proximal tubule.155
Thiazides and loop diuretics, which are widely used for the clinical man-
agement of hypertension and oedema, exhibit their diuretic effect by inhibiting
Na1–Cl and Na1–K1–2Cl co-transporters at the distal tubule and loop of
Henle, respectively.156 Likewise, the CA inhibitor acetazolamide has a strong
diuretic effect, although it is principally given for the treatment of glaucoma. It
has been hypothesised that all of these compounds are subject to active
transport by hOATs to their sites of action primarily based upon evidence from
drug-drug interactions studies. Thus, in the case of diuretics, concomitant
administration with probenecid, a potent inhibitor of OATs, significantly
diminishes their renal clearance in animals and human.157–159 Likewise, com-
bined use of acetazolamide and the NSAID salicylic acid in humans has been
shown to cause severe metabolic acidosis, implying that the agents compete for
protein binding and that a common transport pathway mediates their tubular
secretion.160 In vitro studies using cell lines overexpressing hOAT isoforms have
demonstrated the interaction with representative thiazides, loop diuretics and
the CA inhibitor acetazolamide (Figure 5.15); hOAT1 exhibits the highest
affinity for thiazides, whereas the affinity for loop diuretics was the greatest
with hOAT3.161 Detailed uptake and efflux studies with bumetanide suggested
that the compound is taken up by hOAT1 and hOAT3, and is excreted in the
urine by hOAT4.162 Overall, the findings on the active OAT mediated transport
of thiazides, loop diuretics and acetazolamide is consistent with their weakly
acidic nature.
Most first and second generation sulfonylurea hypoglycemic agents are
eliminated via metabolism.163,164 Chlorpropamide and glisoxepide are the
exceptions; while metabolism represents a significant component of elimina-
tion, B22–40 % of the administered doses of the compounds is renally excreted
in the unchanged form.163 However, the values for percentage of dose excreted
in the unchanged form need to be interpreted with caution. For example, while
the mean amount of chlorpropamide excreted unchanged in the urine is B22%,
250 Chapter 5

Thiazides
Cl
H H H
Cl N Cl N Cl N Cl N
Cl
H2N NH H2N NH H2N NH H2N NH
S S S S S S S S
O O O O O O O O O O O O O O O O
Chlorothiazide Hydrochlorothiazide Cyclothiazide Trichloromethiazide

Loop diuretics Carbonic Anhydrase Inhibitors

HN O
O Cl NH N NH2 H3C
O N N NH2
S O N
O S O
H2N H2N S O
S COOH S COOH H3 C N S
H H3C N O
O O O O

Bumetanide Furosemide Acetazolamide Methazolamide

Figure 5.15 Sulfonamide-based substrates for the human renal OAT transporters.

a variation from 10–60% has been discerned in human volunteers.162 From an

SAR perspective, it is interesting that chlorpropamide but not tolbutamide is
subject to renal excretion as parent. Noteworthy structural differences are the
replacement of the chlorine atom in chlorpropamide with a methyl group in
tolbutamide and extension of the terminal alkyl chain in chlorpropamide by
one methylene in tolbutamide (see Table 5.4).
It is tempting to speculate that the difference in renal excretion pattern
occurs because of the difference in plasma free fraction (range of chlorpropa-
mide and tolbutamide bound to albumin is 57–87% and 86–99%, respectively)
of the two drugs which, in turn, governs the process of passive filtration and
tubular secretion.165,166 The involvement of active renal transport of chlor-
propamide by human OAT isoforms remains uncertain; although hOAT1 can
be potentially ruled out on the basis of the findings that most sulfonylureas
(including chlorpropamide and tolbutamide) are not substrates for the rat
OAT1 transporter.167
Sulfonamide antibacterials are known to be eliminated by hepatic metabolism
(N-acetylation and CYP oxidation of the N4 amine) as well as renal processes.
Renal excretion (mostly likely transport mediated) appears to be dependent on
the acidity of the sulfonamide group, which in turn is influenced by the hetero-
cyclic substituent attached to the sulfonamide nitrogen.168–171 This is illustrated
with three sulfanilamide derivatives, sulfisomidine, sulfametomidine and sulfadi-
methoxine, respectively (Figure 5.16).172 The rank order for renal clearance
of the three compounds in humans is: sulfisomidine 444 sulfametomidine
4 sulfadimethoxine. Thus, replacement of the 2,6-dimethylpyrimidine group in
sulfisomidine with 2-methyl-6-methoxy- or 2,6-dimethoxypyrimidine groups
dramatically lowers renal clearance. Of interest is the observation that these
functional group changes also significantly increase the half-life of sulfisomidine
from 8 h to 28–35 h. It has been speculated that the difference in half-life may be
Sulfonamide as an Essential Functional Group in Drug Design 251
CH3 CH3
CH3 O O

O O N N O O N N O O N N
S S S CH3
N CH3 N CH3 N O
H H H
H2N H2N H2N
Sulfisomidine Sulfametomidine Sulfadimethoxine
CLR = 231 mL/min CLR = 21 mL/min CLR = 10.8 mL/min

Figure 5.16 SAR for renal excretion of sulfanilamides.

due to a reduction in phase I/II metabolism, active renal secretion or by increased

tubular reabsorption. The individual renal transporters that mediate renal
transport of sulfanilamide remain largely unexplored.
Finally, human mass balance studies on sulfonamide-based coxibs (e.g.
celecoxib and valdecoxib) indicate little to no renal excretion of the parent
compounds.114,116

5.3 Adverse Drug Reactions (ADRs) with Sulfonamide

Drugs
Sulfanilamide antibacterials have been implicated in causing hypersensitivity
reactions. The most common manifestation of a hypersensitivity reaction is
rash and hives. However, there are several life-threatening manifestations of
hypersensitivity to sulfa drugs, including Stevens-Johnson syndrome (also
referred to as toxic epidermal necrolysis), agranulocytosis, haemolytic anaemia,
thrombocytopenia and fulminant hepatic necrosis, among others. Approxi-
mately 3% of the general population has ADRs when treated with sulfonamide
antibacterials. Of note is the observation that patients with HIV have a much
higher prevalence, at about 60%.173

5.3.1 Types of Hypersensitivity Reactions with Sulfanilamide

Antibacterials
At least two types of ADRs have been described that are related to the sulfa-
nilamide-type structure.174,175
The first is a type 1 hypersensitivity reaction (immediate allergy) which is
immunoglobulin (Ig) E mediated and presents most commonly as either a
maculopapular eruption or an urticarial rash that develops within 1–3 days of
initial medication and resolves spontaneously upon discontinuation of the
drug. Anaphylaxis may develop upon repeated exposure. The heterocyclic ring
system that is attached to the sulfonamide nitrogen is believed to be the
immunological determinant of type 1 hypersensitivity response.174,175
The second is a type 2 hypersensitivity reaction that is cytotoxic and
immune-mediated and initially presents as a fever and a non-urticarial rash that
252 Chapter 5
may progress to erythema multiforme and multi-organ toxicity. The onset is
delayed (usually 7–14 days) and resolves upon withdrawal of the medication.
Type 2 hypersensitivity reactions are typically mediated by IgM and IgG
antibodies.

5.3.2 Mechanism of Type 2 Hypersensitivity by Sulfonamide

Antibacterials
Although the biochemical mechanisms involved in type 2 hypersensitivity
associated sulfonamide antibacterials have not been fully elucidated, reactive
metabolites, as opposed to the parent compounds, appear to play a pivotal role.
The obligatory step in the bioactivation of aniline derivatives involves N-
hydroxylation on the primary amine nitrogen leading to the formation of the
hydroxylamine metabolite (Scheme 5.6).176 Sulfamethoxazole hydroxylamine
(5.37), a metabolite generated from the CYP2C9 and myeloperoxidase cata-
lysed oxidation of sulfamethoxazole in humans177 (and putative analogs
thereof) are toxic to cells of the immune system in vitro.178,179 Sulfamethoxazole
metabolites are thought to play an additional role in the pathogenesis of
hypersensitivity reactions because they have been shown to activate dendritic
cells and to be immunogenic.180,181 Once formed, hydroxylamine derivatives

O
HO SG S
HO O N HN G
NH2 NH N
CYP2C9
or
myeloperoxidase [O] GSH

cytochrome b5/ thiols O S

O S O S
O S b5 reductase O S NH NH
NH NH O O
NH O
O O
N N
N N N
O O
O O O
CH3 CH3
CH3 CH3 CH3
5.40 5.41
Sulfamethoxazole 5.37 5.38

NAT 2 [O]

O
NO2
H3C NH

O S
NH
O S O
NH
O N
N O
O CH3
CH3 5.39

Scheme 5.6 Bioactivation of the aniline motif in sulfonamide antibacterials—sulfa-

methoxazole as an example.
Sulfonamide as an Essential Functional Group in Drug Design 253
such as 5.37 can autoxidise to the corresponding nitroso intermediate (5.38).
The nitroso metabolite of sulfamethoxazole and related anilines are unstable in
solution and undergoes further oxidation to the stable nitro metabolite 5.39.182
The nitroso metabolite of sulfamethoxazole has been showed to react with
glutathione in vitro to produce a semimercaptal intermediate 5.40, which can
rearrange to a stable sulfinamide 5.41 (Scheme 5.6).183,184 Covalent adduction
of the sulfamethoxazole nitroso metabolite 5.38 with cysteinyl residues in
peptides and proteins has recently been demonstrated, which then implicates
the potential for intracellular protein haptenisation by 5.38.185 Consistent with
these overall findings, glutathione and related thiols have been shown to reduce
toxicity and protein binding associated with sulfamethoxazole in vitro.186,187
Furthermore, the finding that sulfamethoxazole and related compounds such as
dapsone can also be oxidised in keratinocytes suggests that bioactivation and
haptenisation may occur directly in the skin.188 Support for this hypothesis is
provided by the findings that covalent adduct formation by sulfamethoxazole
hydroxylamine has been detected in human epidermal keratinocytes with
protein complexes observed in the region of 160, 125, 95 and 57 kDa.189

5.3.2.1 Risk Factors Associated with Type 2 Hypersensitivity of

Sulfonamides
Slow acetylation by the polymorphic N-acetyltransferase (NAT) 2 has been
considered as a risk factor for hypersensitivity to sulfonamide antibacterials
because the principal route of elimination of these drugs in humans involves N-
acetylation of the aniline moiety by NAT2 resulting in the neutral amide
metabolites (see Scheme 5.6).190 In a NAT2-deficient population the aniline
motif has the potential to undergo bioactivation via the N-hydroxylation route.
However, given the low incidence of type 2 hypersensitivity reactions to
sulfonamide antibacterials (1 in 1000), slow acetylation cannot be the sole
mechanism of predisposition in the population. Interindividual variability in
the formation and detoxication of reactive metabolites of the sulfonamide
antibacterials could be another possibility with regards to individual risks to
type 2 hypersensitivity reactions, especially since the formation of the reactive
intermediate precursor is mediated by the polymorphic CYP2C9 isozyme.
However, the reason why certain individuals are susceptible to these adverse
effects is at present still largely unknown, as the nature of the process from
formation of the reactive metabolite and downstream consequences leading to
toxicity remains largely undetermined.191,192

5.3.2.2 Are All Sulfonamides Created Alike?

It is commonly believed that people who have a hypersensitivity reaction to one
member of the sulfonamide class of antibacterials are likely to have a similar
reaction to other non-antibacterial sulfonamide-containing drugs. The clinical
significance of cross-reactivity of medications in a patient with a ‘sulfa allergy’
254 Chapter 5
continues to perplex clinicians and complicates decisions regarding patient
safety. Historically, the term ‘sulfa’ refers to a derivative of the antibacterial
agent sulfanilamide. More recently, the term has been broadly applied to
include numerous and chemically diverse set of drugs that contain the sulfo-
namide motif including NSAIDs such as celecoxib, valdecoxib and meloxicam.
For instance, the package insert for the COX-2 inhibitor celecoxib states that
‘celebrex should not be given to patients who have demonstrated allergic-type
reactions to sulfonamides’.
The potential of sulfonamide (e.g. celecoxib, valdecoxib and meloxicam) and
methylsulfone (e.g. rofecoxib) derived NSAIDs to cause Steven-Johnson syn-
drome has been examined using the US Food and Drug Administration (FDA)
adverse event reporting system (AERS). No instance of severe epidermal
necrolysis has been reported for meloxicam, and incidences for the other three
agents were depicted to be in the order valdecoxib 4 celecoxib 4 rofecoxib.193
There is no doubt that Steven-Johnson syndrome is a low-incidence ADR of
most, if not all, NSAIDs, but the studies conducted to date have suggested that
there is no association with the sulfonamide moiety. For instance, a meta-
analysis has explored the incidence of allergic reactions associated with
celecoxib in: 194

 patients in North America;

 patients taking part in international arthritis trials;
 patients with a history of hypersensitivity reactions to sulfonamide
antibacterials;
 patients receiving medications containing sulfonamide antibacterials.

Data from the 11 008 patients studied demonstrated that the incidence of
hypersensitivity with celecoxib was statistically indifferent from that observed
with placebo or active comparators (NSAIDs) when the entire cohort was
examined. Although skin reactions occurred with greater frequency in patients
with a history of sulfonamide functionality, the trend was consistent across all
three treatment groups (celecoxib, NSAIDs and placebo), indicating a general
patient susceptibility rather than a functional group-specific effect.
Furthermore, a prospective study assessed the incidence of cross-reactivity in
28 patients with a self-reported history of sulfonamide antibacterial allergy who
then received celecoxib.195 Sulfonamide allergy was assessed by either skin
prick, intradermal testing, and/or lymphocyte toxicity assay (LTA) using sul-
famethoxazole. Positive reactions occurred in four out of 28 patients admi-
nistered a skin test and two out of 10 patients evaluated by LTA. Patients with
a negative skin test received an oral challenge with sulfamethoxazole. All
patients underwent a low- and high-dose oral challenge test with celecoxib
without any adverse events. Follow-up in 25 patients showed that 15 (60%)
continued to tolerate therapy with celecoxib, five did not take celecoxib after
the oral challenge, and five discontinued celecoxib therapy secondary to
adverse effects. In the six patients with a positive skin test, four continued to
take celecoxib, one discontinued therapy secondary to gastrointestinal adverse
Sulfonamide as an Essential Functional Group in Drug Design 255
effects, and one did not take the drug after the oral challenge upon the advice of
the primary care physician. The authors concluded that the risk for cross-
reactivity of celecoxib in patients with a sulfonamide antibacterial allergy is
low.

5.4 Bioactivation Pathways Involving the Sulfonamide

Motif
There are several examples in bioorganic chemistry where the sulfonamide
group has been shown to participate in reactive metabolite formation. In
addition, there are instances of 2-sulfonamidoheterocycles, which owing to
their electrophilic nature, undergo nucleophilic displacement reactions with
endogenous nucleophiles such as glutathione either under chemical or enzy-
matic conditions.

5.4.1 Bioactivation of Sulfonanilides

Reactive metabolite formation can occur via a two-electron oxidation process
on aryl sulfonamides, which contain electron rich functionalities in an ortho
and/or para framework relative to the sulfonamido group. Reactive inter-
mediates derived from such bioactivation reactions can be generally categorised
as quinone (quinone-imine, quinone-methide and diiminoquinones) derivatives
that react with glutathione and/or protein nucleophiles in a 1,4-Michael fash-
ion. The overall two-electron oxidation process with aryl sulfonamides
resembles the one observed with the acetanilide derivative acetaminophen,
which yields the electrophilic quinone-imine NAPQI.196 The aryl sulfonamide-
based selective COX-2 inhibitor nimesulide illustrates the concept.
After widespread clinical use of nimesulide, hepatotoxicity—including both
acute and hepatitis and the more severe fulminant hepatic failure—were
reported.197,198 The relatively high occurrence of these adverse events in Fin-
land and Spain caused the drug to be withdrawn from the market in these
countries. Circumstantial evidence linking reactive metabolite formation with
idiosyncratic hepatotoxicity has been presented by Li et al.199
As shown in Scheme 5.7, the rate-limiting step in nimesulide bioactivation
involves reduction of the nitrobenzene group to the corresponding electron-rich
aniline metabolite 5.42, which has been detected in human urine. Two-electron
oxidation of 5.42 by CYP and/or peroxidase enzymes generates the reactive
diiminoquinone species 5.43, which adducts with sulfydryl nucleophiles or to
human serum albumin.
Recent studies from our laboratory have shown a similar bioactivation
pathway on 5-aminooxindole-and 5-aminobenzosultam-based inhibitors of
proline rich tyrosine kinase (PYK2) exemplified by analogs 5.44 and 5.45,
respectively, which are metabolised by CYP3A4 to form electrophilic diimi-
noquinone species 5.46 and 5.47, respectively, amenable to trapping with glu-
tathione (Scheme 5.8).200 Interestingly, 5-aminobenzosultam 5.45 was less
 256

O
O O
H O O HS N COOH
O N O H H
N O O H
S S CYP N N O NH
O S GSH S
O
CH3 CH3 O O GSH =
NO2 NH2 CH3 CH3
NH NH2
Nimesulide
5.42 5.43 SG
H2N COOH

Scheme 5.7 Bioactivation of the aryl sulfonamide derivative nimesulide to reactive metabolites.
Chapter 5
Sulfonamide as an Essential Functional Group in Drug Design 257
H CF3
N N H
CYP N N
X X GSH
R R X R=
N N R N NH
N CH3
H
GS H
N
5.44: X = CO 5.46: X = CO CH3
5.45: X = SO2 5.47: X = SO2

Scheme 5.8 Bioactivation of 5-aminooxindole and 5-aminobenzosultam derivatives

to reactive diiminoquinones

susceptible towards bioactivation than the 5-aminooxindole derivative 5.44 as

judged from the dramatically lower peak area ratio of the glutathione conjugate
of 5.45 (relative to the glutathione conjugate of 5.44). Estimates of the relative
rates of oxidation of the 5-aminooxindole and 5-aminobenzosultam fragments
using theoretical quantum chemical calculations suggest that 5-aminobenzo-
sultam fragment is more difficult to oxidise, with approximately +4.5 kcal
mol1 more energy required for oxidation of this component.201

5.4.2 Intrinsic Electrophilicity of ‘Activated’ Sulfonamides

The presence of the sulfonamido group in the a position of certain heterocyclic
rings generates electrophiles that can react with glutathione. The phenomenon
was first demonstrated with the carbonic anhydrase inhibitor benzothiazole-2-
sulfonamide 5.48 (Scheme 5.9). Although 5.48 inhibits carbonic anhydrase
enzyme(s) in vitro with potency greater than the traditional inhibitors of this
enzymes (e.g. acetazolamide and methazolamide, see Figure 5.5), 5.48 is devoid
of in vivo pharmacological activity in preclinical species even after adminis-
tration of very high doses.
The complete absence of 5.48 in circulation and/or urine after administration
of 35S-radiolabeled 5.48 to rats provides a rational explanation for the in vitro–
in vivo pharmacological disconnect. While 480% of the total radioactivity was
eliminated in urine, none of the characterised metabolites (benzothiazole-2-
mercapturic acid 5.50, benzothiazole-2-mercaptan 5.51 and benzothiazole-2-
mercaptoglucuronide 5.52) were radioactive implicating the loss of the 35S label
(see Scheme 5.9).202,203 The proposed mechanism, which collectively accounted
for these observations, involves a quantitative nucleophilic displacement of
the 2-sulfonamide group in 5.48 by the endogenous pool of glutathione in the
mammal to first afford the sulfydryl conjugate 5.49, 35SO2 and ammonia. The
demonstration that all the urinary radioactivity was due to inorganic sulfate
confirms the nucleophilic displacement of the sulfonamide group by glu-
tathione. Subsequent breakdown of 5.49 affords the mercapturic acid conjugate
5.50. A carbon–sulfur bond cleavage in 5.49 (or 5.50) can generate the mer-
captan metabolite 5.51, which can undergo S-glucuronidation to afford 5.52.
A side-by-side comparison on the in vitro reactivity of acetazolamide,
methazolamide and some 6-hydroxybenzothiazole-2-sulfonamide derivatives
with glutathione at 37 1C (pH 7.4 buffer) revealed that acetazolamide and
258 Chapter 5

N O N O COOH N
GSH
S O S NH S COOH
S * NH2 H S O S O
35
SO2 HN HN
GSH NH3 5.49 5.50
5.48 CH3
35
*= S label COOH
H2N
HOOC
O
HO S S N
HO
OH SH
N
S
5.52 5.51

Scheme 5.9 In vivo nucleophilic displacement of the sulfonamide group in ben-

zothiazole-2-sulfonamide by glutathione.

O HO O O
N O
O S
S
R NH2
O S NH2
R
5.53: R = H O O
O 5.58: R = CH3
S S 5.60: R = H
NH2
5.54: R = R
HO N O
HO O
O S
O
O NH2
S 5.57
R
O S NH2 5.59: R = CH3
5.61: R = H
5.55: R = H
O

5.56: R =

Figure 5.17 SAR studies on the in vitro reactivity of glutathione with 2-sulfonamido-
heterocycles.

methazolamide possessed significantly diminished capability to undergo

nucleophilic displacement with glutathione.204 Likewise, the parent 6-hydro-
xybenzothiazole-2-sulfonamide 5.53 was the slowest to react with glutathione;
conversion of the hydroxy group to lipophilic ester derivatives increased the
rate of nucleophilic displacement by glutathione (Figure 5.17).204 Interestingly,
the lipophilic pivalic acid ester derivative of 6-hydroxybenzothiazole-2-sulfo-
namide 5.54 was shown to be a potent allergen in guinea pig model of dermal-
sensitisation potential.204 While the molecular basis for this adverse effect
remains unclear, it is likely a nucleophilic displacement of the sulfonamide by
protein nucleophiles in the skin can trigger the process of haptenisation leading
to an immune response.
The finding that de-aza versions of 5.53 and 5.54, i.e. 2-benzo[b]thiophene-
sulfonamide analogs (e.g. compounds 5.55 and 5.56, Figure 5.17), are
Sulfonamide as an Essential Functional Group in Drug Design 259
practically inert towards reaction with nucleophiles such as glutathione sug-
gests that the electrophilic nature of benzothiazole-2-sulfonamide is derived
from the presence of the imino group in the thiazole ring system.205 Consistent
with this general argument, thiophene-2-sulfonamide analogs (e.g. compound
5.57, Figure 5.17) have also been shown to be latent towards reaction with
glutathione.206 Although benzofuran- and indole-2-sulfonamide derivatives
(e.g. 5.58 and 5.59), which do not contain the heterocyclic imine group, showed
much lower levels of reactivity toward glutathione than benzothiazole-2-sul-
fonamides, they were measurably more electrophilic than the corresponding
benzothiophene-2-sulfonamides.207 Concurrent with these findings, several
benzofuran- and indole-2-sulfonamide were noted to cause skin sensitisa-
tion.207 In addition, the 3-methyl- and N-methyl-substituted derivatives of the
benzofuran- and indolesulfonamides (e.g. compounds 5.58 and 5.59) were
appreciably more reactive towards glutathione than the desmethyl compounds
5.60 and 5.61, respectively. Likewise, electron-donating substituents in benzo-
furan- and indolesulfonamides did not affect or accelerate the rate of dis-
placement reaction. This is in contrast to the benzothiazole series, where a
6-hydroxy substituent markedly reduced the electrophilicity of the system.204
A final example of the in vitro and in vivo reactivity of glutathione
with activated sulfonamides is evident with the HIV-1 protease inhibitor
PNU-10912.208 From an SAR perspective, PNU-140690 that is obtained by
substituting the cyanopyridinyl group in PNU-10912 with a trifluoro-
methylpyridinyl moiety (Scheme 5.10), was shown to be stable towards reaction
with glutathione in vitro and in vivo studies in animals.208 Furthermore, these
studies were the first to demonstrate the catalytic role of glutathione-S-trans-
ferase (GST) enzymes in the sulfonamide nucleophilic displacement reaction.
Additional SAR studies on pyridine-2-sulfonamides further confirmed the
requirement of an electrophilic centre (pyridinyl nitrogen) a to the sulfonyl
group for reaction with glutathione.209 Sulfonamide reaction with glutathione
was not dependent on the substituents attached to the nitrogen, however,
pyridine ring substituents markedly influenced the process. Thus, ortho- and
para-pyridine ring substituents capable of withdrawing sufficient electron
density from the carbon atom a to the sulfonyl group were an absolute
requirement for the GST-mediated sulfonamide displacement. A proposed
mechanism in which the role of electron-withdrawing groups in facilitating
sulfonamide displacement by glutathione is shown in Scheme 5.10 for PNU-
10912.
It is of great importance to note that the nucleophilic displacement reaction
is not limited to sulfonamides; 2-alkylsulfonylheterocyclic compounds also
succumb to this pathway as exemplified with the 2-methylsulfonylthiadiazole
derivative 5.62. The compound is a leptin receptor agonist in vitro but does
demonstrate any pharmacology (spontaneous reduction of food intake) in vivo
in rats or mice, a phenomenon which appears to be related to the lack of
systemic drug levels following intravenous and oral administration of 5.62 to
rodents. In phosphate buffer in the presence of glutathione, 5.62 was shown
to be unstable and underwent quantitative conversion to the glutathione
 260

OH OH

CN CF3
O O O O
HN HN
S N S N
O O O O

PNU-109112 PNU-140690

N OH
N C N
N
GS
GST/GS GS
GS +
H H N
HN N N S GS N O O
S N N S
O O O SO2
O O O NH2
PNU-109112

Scheme 5.10 SAR and mechanistic studies on the glutathione mediated nucleophilic displacement of pyridinyl-2-sulfonamides.
Chapter 5
Sulfonamide as an Essential Functional Group in Drug Design 261

N COOH
Cl N N H
Cl N N CH3 S NH
S GSH S
S O O
O pH 7.4
phosphate Cl
Cl buffer
5.62 5.63 COOH
H2 N

Scheme 5.11 Nucleophilic displacement of a 2-methylsulfonylthiadiazole derivative

by glutathione.

conjugate 5.63, which is obtained via nucleophilic displacement of the methyl-

sulfone group in 5.62 (A. S. Kalgutkar, unpublished findings) (Scheme 5.11).
Consistent with the in vitro observations, rat plasma sample analysis after
intravenous or oral administration of 5.62 indicated the presence of 5.63 as the
exclusive component in the biological matrix.

5.4.3 Bioactivation of Sulfonylureas

Although diarylsulfonylureas such as sulofenur have demonstrated utility as
antitumor agents against a wide range of cancers, dose-limiting toxicities in
humans (including methemoglobinemia and haemolytic anaemia), have ham-
pered wide usage.210 In paediatric patients with refractory malignant solid
tumors, methemoglobinemia was observed at all doses during a phase I clinical
trial.211 Likewise, the antidiabetes attributes of tolbutamide has been offset by
its teratogenic properties in several animal species and humans, and its use is
contraindicated in pregnant diabetic patients.212 In both cases, circumstantial
evidence has been presented which links toxicity with bioactivation of the
sulfonylurea motif to protein-reactive metabolites. Following in vivo adminis-
tration of sulofenur to rats, Jochheim et al.213 characterised a novel glutathione
conjugate 5.65 derived from conjugation of the thiol nucleophile with p-
chlorophenylisocyanate (5.64) (Scheme 5.12). Similarly, analysis of rat bile
following administration of tolbutamide led to the identification of S-(n-
butylcarbamoyl)glutathione (5.67), a glutathione conjugate derived from the
reaction with n-butylisocyanate (5.66) (see Scheme 5.12).214
The biochemical mechanism(s) for sulfonylurea decomposition to the cor-
responding isocyanates remains unclear since sulfonylureas have been found to
be stable towards decomposition at physiological pH and increasing tem-
perature. Likewise, no glutathione–isocyanate conjugate formation is dis-
cernible in incubations of the parent sulfonylureas and glutathione under
chemical or biochemical (GST catalysis) conditions.
Finally, it is interesting to note that the isocyanate and corresponding glu-
tathione conjugate, derived from tolbutamide metabolism, are time-dependent
 262

O H
H H GS N
R1 N N ? C GSH
N R2
S R2
Sulofenur : R1 = R2 = R2
O O O
Cl O R1 NH2
S
H3C O O 5.64: R2 = 5.65: R2 =
Cl Cl
Tolbutamide : R1 = R2 =
5.66: R2 = 5.67: R2 =

Scheme 5.12 Bioactivation of sulfonylureas to reactive isocyanate intermediates.

Chapter 5
Sulfonamide as an Essential Functional Group in Drug Design 263
inactivators of glutathione reductase (GR) enzyme that presumably cause
inactivation via protein carbamoylation.213,214 It has been speculated that the
teratogenic effects of tolbutamide in rats could be linked to tolbutamide-
mediated depletion of glutathione through inhibition of GR in rat embryos.214
GR catalyses the reduction of the oxidised disulfide form of glutathione to
glutathione and thus maintains intracellular levels of the thiol antioxidant in
the embryo, which is critical for cell viability and normal growth. Glutathione
depletion by exogenous compounds has been reported not only to enhance the
embryo-lethal, teratogenic and growth-retarding properties of these chemicals,
but also to cause an increase in the number of dead and malformed rat embryos
in vivo.215

5.5 Conclusions
The juxtaposition of terminologies around the sulfonamide-containing drugs
and sulfonamide drugs can cause confusion around the issue of hypersensitivity
reactions (sulfa allergy). Historically, the term ‘sulfa’ refers to a derivative of
the antibacterial agent sulfanilamide. More recently, the term has been applied
generously to a diverse group of drugs, all of which contain the sulfonamide
motif. This is largely for historic reasons, as many of the drugs were approved
prior to any toxicity mechanism investigations. In addition, some non-anti-
bacterial sulfonamides (e.g. sulfonylureas and loop diuretics) are classified as
derivatives of the antibacterials because their discovery has been influenced in
large part by the polypharmacology of the original sulfa drugs.
The sulfa allergy association raises concerns for medicinal chemistry and the
use of the sulfonamide group in drug discovery. However, to date, a thorough
examination of the literature does not support the dogma of sulfonamide
antibacterial cross reactivity with non-antibacterial sulfonamides. In fact, rig-
orous studies examining cross reactivity between sulfonamide antibiotics and
sulfonamide non-antibiotics suggest that predisposition to allergic reactions is
the primary risk factor for the association, not a true cross reactivity.95,194
In retrospect, the sulfonamide group should remain an essential part of the
medicinal chemist’s arsenal in the search of new chemical entities with phar-
macological action. There are numerous examples of blockbuster drugs which
contain this motif and yet are not associated with any significant incidence of
sulfa allergy type adverse drug reactions. In the current climate, which is
focused on a heightened awareness of toxicophores (structural alerts) and
toxicity, it is recommended that the sulfonamide motif should not be treated as
a toxicophore, although appropriate caution needs to be exercised in terms of
novel bioactivation pathways involving this functional group.
Finally, the finding that metabolic elimination of many acidic sulfonamide
derivatives is principally mediated by CYP2C9 needs to be taken into con-
sideration from a drug design perspective, given the polymorphic expression of
this isozyme in humans and its contribution towards variability in pharmaco-
kinetic and pharmacodynamic response.
264 Chapter 5

References
1. G. Domagk, Deut. Med. Wochschr., 1935, 61, 250.
2. J. Tréfouel, J. Mme. Tréfouel, F. Nitti and D. Bovet, C. R. Soc. Bio.
Paris., 1935, 120, 756.
3. N. Anand, in Burger’s Medicinal Chemistry, 4th edn, ed. M. E. Wolff,
Wiley-Interscience, New York, 1979, Part II, Chapter 13, p. 13.
4. D. Woods, Brit. J. Exptl. Pathol., 1940, 21, 74.
5. D. Woods and P. Fildes, Chem. Ind., 1940, 59, 133.
6. P. H. Bell and R. O. Roblin Jr., J. Am. Chem. Soc., 1942, 64, 2903.
7. A. Burger, in Progress in Drug Research, ed. H. von E. Jucker,
Birkhäuser-Verlag, Basel, 1991, vol. 37, pp. 287–371.
8. C. J. Hawkey, Lancet, 1999, 353, 307.
9. D. L. DeWitt and W. L. Smith, Proc. Natl. Acad. Sci. U. S. A., 1988, 85,
1412.
10. T. Hla and K. Neilson, Proc. Natl. Acad. Sci. U. S. A., 1992, 89,
7384.
11. S. H. Lee, E. Soyoola, P. Chanmugam, S. Hart, W. Sun, H. Zhong,
S. Liou, D. Simmons and D. Hwang, J. Biol. Chem., 1992, 267,
25934.
12. D. A. Kujubu, B. S. Fletcher, B. C. Varnum, R. W. Lim and H. R.
Herschman, J. Biol. Chem., 1991, 266, 12866.
13. J. R. Vane, J. A. Mitchell, I. Appleton, A. Tomlinson, D. Bishop-Bailey,
J. Croxtall and D. A. Willoughby, Proc. Natl. Acad. Sci. U. S. A., 1994,
91, 2046.
14. C. J. Smith, Y. Zhang, C. M. Kobolt, J. Muhammad, B. S. Zweifel,
A. Shaffer, J. J. Talley, J. L. Masferrer, K. Seibert and P. C. Isakson,
Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 13313.
15. T. D. Warner, F. Giuliano, I. Vojnovic, A. Bukasa, J. A. Mitchell and
J. R. Vane, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 7563.
16. L. J. Marnett, Annu. Rev. Pharmacol. Toxicol., 2009, 49, 265.
17. L. H. Rome and W. E. M. Lands, Proc. Natl. Acad. Sci. U. S. A., 1975, 72,
4863.
18. R. A. Copeland, J. M. Williams, J. Giannaras, S. Nurnberg, M.
Covington, D. Pinto, S. Pick and J. M. Trzaskos, Proc. Natl. Acad. Sci.
U. S. A., 1994, 91, 11202.
19. J. J. Talley, Prog. Med. Chem., 1999, 36, 201.
20. R. Flower, R. Gryglewski, K. Herbaczynska-Cedro and J. R. Vane,
Nature New Biol., 1972, 238, 104.
21. L. J. Marnett and A. S. Kalgutkar, in Milestones in Drug Therapy–COX-2
Inhibitors, ed. M. Pairet and J. van Ryn, Birkhäuser, Verlag, Basel, 2004,
pp. 15–40.
22. K. R. Gans, W. Galbraith, R. J. Roman, S. B. Haber, J. S. Kerr, W. K.
Schmidt, C. Smith, W. E. Hewes and N. R. Ackerman, J. Pharmacol. Exp.
Ther., 1990, 254, 180.
Sulfonamide as an Essential Functional Group in Drug Design 265
23. Y. Leblanc, J. Y. Gauthier, D. Ethier, J. Guay, J. Mancini, D. Riendeau,
P. Tagari, P. Vickers, E. Wong and P. Prasit, Bioorg. Med. Chem. Lett.,
1995, 5, 2123.
24. S. R. Bertenshaw, J. J. Talley, D. J. Rogier, M. J. Graneto, R. S. Rogers,
S. W. Kramer, T. D. Penning, C. M. Koboldt, A. W. Veenhuizen and
Y. Zhang, et al., Bioorg. Med. Chem. Lett., 1995, 5, 2919.
25. T. D. Penning, J. J. Talley, S. R. Bertenshaw, J. S. Carter, P. W. Collins,
S. Doctor, M. J. Graneto, L. F. Lee, J. W. Malecha and J. M. Miyashiro,
et al., J. Med. Chem., 1997, 40, 1347.
26. A. S. Kalgutkar, B. C. Crews and L. J. Marnett, Biochemistry, 1996, 35,
9076.
27. J. S. Carter, Expert Opin. Ther. Pat., 2000, 10, 1011.
28. D. Picot, P. J. Loll and R. M. Garavito, Nature, 1994, 367, 243.
29. R. G. Kurumbail, A. M. Stevens, J. K. Gierse, J. J. McDonald,
R. A. Stegeman, J. Y. Pak, D. Gildehaus, J. M. Miyashiro, T. D.
Penning, K. Seibert, P. C. Isakson and W. C. Stallings, Nature, 1996, 384,
644.
30. D. K. Bhattacharyya, M. Lecomte, C. J. Rieke, R. M. Garavito and W. L.
Smith, J. Biol. Chem., 1996, 271, 2179.
31. J. A. Mancini, D. Riendeau, J.-P. Falgueyret, P. J. Vickers and G. P.
O’Neill, J. Biol. Chem., 1995, 270, 29372.
32. J. K. Gierse, J. J. McDonald, S. D. Hauser, S. H. Rangwala, C. M.
Koboldt and K. Seibert, J. Biol. Chem., 1996, 271, 15810.
33. E. Wong, C. Bayly, H. L. Waterman, D. Riendeau and J. A. Mancini,
J. Biol. Chem., 1997, 272, 9280.
34. J. Harrington, J. E. Robertson, D. C. Kvam, R. R. Hamilton, K. T.
McGurran, R. J. Trancik, K. F. Swingle, G. G. Moore and J. F. Gerster,
J. Med. Chem., 1970, 13, 137.
35. K. F. Swingle, G. G. Moore and T. J. Grant, Arch. Int. Pharmacodyn.,
1976, 221, 132.
36. R. Davis and R. N. Brogden, Drugs, 1994, 48, 431.
37. J. L. Masferrer, B. S. Zweifel, P. T. Manning, S. D. Hauser, K. M. Leahy,
W. G. Smith, P. C. Isakson and K. Seibert, Proc. Natl. Acad. Sci. U. S. A.,
1994, 91, 3228.
38. N. Futaki, S. Takahashi, T. Kitagawa, Y. Yamakawa, M. Tanaka and
S. Higuchi, Inflamm. Res., 1997, 46, 496.
39. I. Wiesenberg-Bottcher, A. Schweizer, J. R. Green, Y. Seltenmeyer and
K. Muller, Agents Actions, 1989, 26, 240.
40. T. Vago, M. Bevilacqua and G. Norbiato, Arzneim.-Forsch., 1995, 45,
1096.
41. C. S. Li, W. C. Black, C. C. Chan, A. W. Ford-Hutchinson, J. Y. Gauthier,
R. Gordon, D. Guay, S. Kargman, C. K. Lau and J. Mancini, et al.,
J. Med. Chem., 1995, 38, 4897.
42. K. Tsuji, K. Nakamura, N. Konishi, H. Okumura and M. Matsuo, Chem.
Pharm. Bull., 1992, 40, 2399.
266 Chapter 5
43. K. Tanaka, T. Shimotori, S. Makino, Y. Aikawa, T. Inaba, C. Yoshida
and S. Takano, Arzneimittelforschung, 1992, 42, 935.
44. A. K. Singh, M. Chawla and A. Singh, J. Pharm. Pharmacol., 2000, 52,
467.
45. F. Julémont, X. de Leval, C. Michaux, J.-F. Renard, J.-Y. Winum, J.-L.
Montero, J. Damas, J.-M. Dogne and B. Pirotte, J. Med. Chem., 2004, 47,
6749.
46. L. J. Marnett, S. W. Rowlinson, D. C. Goodwin, A. S. Kalgutkar and
C. A. Lanzo, J. Biol. Chem., 1999, 274, 22903.
47. C. T. Supuran, Curr. Pharm. Design, 2008, 14, 603.
48. T. Stams and D. W. Christianson, in The Carbonic Anhydrases–New
Horizons, ed. W. R. Chegwidden, Y. Edwards and N. Carter, Birkhäuser
Verlag, Basel, 2000, pp. 159–174.
49. A. Di Fiore, S. M. Monti, M. Hilvo, S. Parkkila, V. Romano, A. Scaloni,
C. Pedone, A. Scozzafava, C. T. Supuran and G. De Simone, Proteins,
2009, 74, 164.
50. C. Temperini, A. Cecchi, A. Scozzafava and C. T. Supuran, Bioorg. Med.
Chem. Lett., 2008, 18, 2567.
51. C. Y. Kim, D. A. Whittington, J. S. Chang, J. Liao, J. A. May and D. W.
Christianson, J. Med. Chem., 2002, 45, 888.
52. D. A. Whittington, A. Waheed, B. Ulmasov, G. N. Shah, J. H. Grubb,
W. S. Sly and D. W. Christianson, Proc. Natl. Acad. Sci. U. S. A., 2001,
98, 9545.
53. S. K. Nair, J. F. Krebs, D. W. Christianson and C. A. Fierke, Biochem-
istry, 1995, 34, 3981.
54. K. Fridborg, K. K. Kannan, A. Liljas, J. Lundin, B. Strandberg,
R. Strandberg, B. Tilander and G. Wirén, J. Mol. Biol., 1967, 25, 505.
55. A. Liljas, K. K. Kannan, P. C. Bergstén, I. Waara, K. Fridborg,
B. Strandberg, U. Carlbom, L. Järup, S. Lövgren and M. Petef, Nature
New Biol., 1972, 235, 131.
56. B. W. Clare and C. T. Supuran, Expert Opin. Drug Metab. Toxicol., 2006,
2, 113.
57. F. Z. Smaine, F. Pacchiano, M. Rami, V. Barragan-Montero, D. Vullo,
A. Scozzafava, J. Y. Winum and C. T. Supuran, Bioorg. Med. Chem.
Lett., 2008, 18, 6332.
58. O. Guzel, A. Innocenti, A. Scozzafava, A. Salman, S. Parkkila, M. Hilvo
and C. T. Supuran, Bioorg. Med. Chem. Lett., 2008, 16, 9113.
59. K. D’Ambrosio, R. M. Vitale, J. M. Dogné, B. Masereel, A. Innocenti,
A. Scozzafava, G. De Simone and C. T. Supuran, J. Med. Chem., 2008,
51, 3230.
60. A. Thiry, S. Rolin, D. Vullo, A. Frankart, A. Scozzafava, J. M. Dogné,
J. Wouters, C. T. Supuran and B. Masereel, Eur. J. Med. Chem., 2008, 43,
2853.
61. C. T. Supuran and A. Scozzafava, Bioorg. Med. Chem., 2007, 15, 4336.
62. B. L. Wilkinson, L. F. Bornaghi, T. A. Houston, A. Innocenti, C. T.
Supuran and S. A. Poulsen, J. Med. Chem., 2006, 49, 6539.
Sulfonamide as an Essential Functional Group in Drug Design 267
63. J. Caprioli, in Aldler’s Physiology of the Eye–Clinical Application, ed.
R. A. Moses and W. M. Hart, The C. V. Mosby Company, St Louis,
1987, pp. 204–222.
64. P. R. Lichter, L. P. Newman, N. C. Wheeler and O. V. Beal, Am. J.
Ophthalmol., 1978, 85, 495.
65. D. L. Epstein and W. M. Grant, Arch. Ophthalmol., 1977, 95, 1378.
66. F. Mincione, A. Scozzafava and C. T. Supuran, Curr. Pharm. Des., 2008,
14, 649.
67. C. Temperini, A. Cecchi, A. Scozzafava and C. T. Supuran, Org. Biomol.
Chem., 2008, 6, 2499.
68. Y. Masuda and T. Karasawa, Arzneimittelforschung, 1993, 43, 416.
69. K. Köhler, A. Hillebrecht, J. Schulze-Wischeler, A. Innocente, A. Heine
and C. T. Supuran, Angew. Chem., Int. Ed. Engl., 2007, 46, 7697.
70. J.-M. Dogné, A. Thiry, D. Pratico, B. Masereel and C. T. Supuran, Curr.
Topics Med. Chem., 2007, 7, 885.
71. A. Di Fiore, C. Pedone, K. D’Ambrosio, A. Scozzafava, G. De Simone
and C. T. Supuran, Bioorg. Med. Chem. Lett., 2006, 16, 437.
72. A. Weber, A. Casini, A. Heine, D. Kuhn, C. T. Supuran, A. Scozzafava
and G. Klebe, J. Med. Chem., 2004, 47, 550.
73. C. T. Supuran, Nat. Rev. Drug Discov., 2008, 7, 168.
74. J. C. Henquin, Diabetologia, 1992, 35, 907.
75. M. M. Loubatières-Mariani, J. Soc. Biol., 2007, 201, 121–125.
76. H. Kleinsorge, Exp. Clin. Endocrinol. Diabetes, 1998, 106, 149.
77. A. Farret, L. Lugo-Garcia, F. Galtier, R. Gross and P. Petit, Fundam.
Clin. Pharmacol., 2005, 19, 647.
78. F. M. Gribble and F. M. Ashcroft, Metabolism, 2000, 49, 3.
79. S. Fukuen, M. Iwaki, A. Yasui, M. Makishima, M. Matsuda and
I. Shimomura, J. Biol. Chem., 2005, 280, 23653.
80. M. Scarsi, M. Podvinec, A. Roth, H. Hug, S. Kersten, H. Albrecht,
T. Schwede, U. A. Meyer and C. Rucker, Mol. Pharmacol., 2007, 71, 398.
81. G. R. Brown and A. J. Foubister, J. Med. Chem., 1984, 27, 79.
82. A. Dornhorst, Lancet, 2001, 358, 1709.
83. P. Lin, D. Marino, J. L. Lo, Y. T. Yang, K. Cheng, R. G. Smith, M. H.
Fisher, M. J. Wyratt and M. T. Goulet, Bioorg. Med. Chem. Lett., 2001,
11, 1073.
84. C. R. McCurdy, R. M. Jones and P. S. Portoghese, Org. Lett., 2000, 2,
819.
85. R.-J. Liu, J. A. Tucker, T. Zinevitch, O. Kirichenko, V. Konoplev,
S. Kuznetsova, S. Sviridov, J. Pickens, S. Tandel, E. Brahmachary and
Y. Yang, et al., J. Med. Chem., 2007, 50, 6535.
86. Y. K. Yee, P. R. Bernstein, E. J. Adams, F. J. Brown, L. A. Cronk, K. C.
Hebbel, E. P. Vacek, R. D. Krell and D. W. A. Snyder, J. Med. Chem.,
1990, 33, 2437.
87. M. Bäck, P. O. Johansson, F. Wangsell, F. Thorstensson, I. Kvarnström,
S. Ayesa, H. Wähling, M. Pelcman, K. Jansson, S. Lindström and
H. Wallberg, et al., Bioorg. Med. Chem. lett., 2003, 11, 2551.
268 Chapter 5
88. M. P. Winters, C. Crysler, N. Subasinghe, D. Ryan, L. Leong, S. Zhao,
R. Donatelli, E. Yurkow, M. Mazzulla, L. Boczon, C. L. Manthey,
C. Molloy, H. Raymond, L. Murray, L. McAlonan and B. Tomczuk,
Bioorg. Med. Chem. Lett., 2008, 18, 1926.
89. C. Luong, A. Miller, J. Barnett, J. Chow, C. Ramesha and M. F.
Browner, Nat. Struct. Biol., 1996, 3, 927.
90. A. S. Kalgutkar, B. C. Crews, S. W. Rowlinson, A. B. Marnett, K. R.
Kozak, R. P. Remmel and L. J. Marnett, Proc. Natl. Acad. Sci. U. S. A.,
2000, 97, 925.
91. A. S. Kalgutkar, A. B. Marnett, B. C. Crews, R. P. Remmel and L. J.
Marnett, J. Med. Chem., 2000, 43, 2860.
92. A. S. Kalgutkar, S. W. Rowlinson, B. C. Crews and L. J. Marnett, Bioorg.
Med. Chem. Lett., 2002, 12, 521.
93. R. C. Murphy, S. Hammarström and B. Samuelsson, Proc. Natl. Acad.
Sci. U.S.A., 1979, 76, 4275.
94. J. M. Drazen, K. F. Austen, R. A. Lewis, D. A. Clark, G. Goto,
A. Marfat and E. J. Corey, Proc. Natl. Acad. Sci. U.S.A., 1980, 77,
4345.
95. J. Augstein, J. B. Farmer, T. B. Lee, P. Sheard and M. L. Tattersall,
Nature (New Biol.), 1973, 245, 215.
96. F. J. Brown, P. R. Bernstein, L. A. Cronk, L. L. Dosset, K. C. Hebbel,
T. P. Maduskuie, H. S. Shapiro, E. P. Vacek, Y. K. Lee, A. K. Willard,
R. D. Krell and D. W. Snyder, J. Med. Chem., 1989, 32, 807.
97. R. W. Fuller, P. N. Black and C. T. Dollery, J. Allergy Clin. Immunol.,
1989, 83, 939.
98. M. L. Cloud, G. C. Enas, J. Kemp, T. Platts-Mills, L. C. Altman,
R. Townley, D. Tinkelman, T. King Jr, E. Middleton and A. L. Sheffer,
et al., Am. Rev. Respir. Dis., 1989, 140, 1336.
99. H. Nakai, M. Konno, S. Kosuge, S. Sakuyama, M. Toda, Y. Arai,
T. Obata, N. Katsube, T. Miyamoto, T. Okegawa and A. Kawasaki,
J. Med. Chem., 1988, 31, 84.
100. Y. Taniguchi, G. Tamura, M. Honma, T. Aizawa, N. Maruyama,
K. Shirato and T. Takishima, J. Allergy Clin. Immunol., 1993, 92, 507.
101. F. J. Brown, Y. K. Yee, L. A. Cronk, K. C. Hebbel, R. D. Krell and
D. W. Synder, J. Med. Chem., 1990, 33, 1771.
102. Y. K. Yee, P. R. Bernstein, E. J. Adams, F. J. Brown, L. A. Cronk, K. C.
Hebbel, E. P. Vacek, R. D. Krell and D. W. Synder, J. Med. Chem., 1990,
33, 2437.
103. V. G. Matassa, F. J. Brown, P. R. Bernstein, H. S. Shapiro, T. P.
Maduskuie Jr., L. A. Cronk, E. P. Vacek, Y. K. Yee, D. W. Synder, R. D.
Krell, C. L. Lerman and J. J. Maloney, J. Med. Chem., 1990, 33, 2621.
104. V. G. Matassa, T. P. Maduskuie Jr, H. S. Shapiro, B. Hesp, D. W.
Snyder, D. Aharony, R. D. Krell and R. A. Keith, J. Med. Chem., 1990,
33, 1781.
105. P. R. Bernstein, Am. J. Respir. Crit. Care Med., 1998, 157, S220.
Sulfonamide as an Essential Functional Group in Drug Design 269
106. C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeny, Adv. Drug
Deliv. Rev., 1997, 23, 3.
107. Y. C. Martin, J. Med. Chem., 2005, 48, 3164.
108. K. Palm, P. Stenberg, K. Luthman and P. Artursson, Pharm. Res., 1997,
14, 568.
109. D. F. Veber, S. R. Johnson, H.-Y. Cheng, B. R. Smith, K. W. Ward and
K. D. Kopple, J. Med. Chem., 2002, 45, 2615.
110. P. D. Leeson and A. M. Davis, J. Med. Chem., 2004, 47, 6338.
111. M. Yazdanian, K. Briggs, C. Jankovsky and A. Hawi, Pharm. Res., 2004,
21, 293.
112. S. G. Vijaya Kumar and D. N. Mishra, Chem. Pharm. Bull. (Tokyo),
2006, 54, 1102.
113. N. Sarapa, M. R. Britto, M. B. Mainka and K. Parivar, Eur. J. Clin.
Pharmacol., 2005, 61, 247.
114. J. J. Yuan, D.-C. Yang, J. Y. Zhang, R. Bible Jr., A. Karim and J. W. A.
Findlay, Drug Metab. Dispos., 2002, 30, 1013.
115. N. M. Davies, A. J. McLachlan, R. O. Day and K. M. Williams, Clin.
Pharmacokinet., 2000, 38, 225.
116. S. K. Paulson, J. D. Hribar, N. W. K. Liu, E. Hajdu, R. H. Bible Jr.,
A. Piergies and A. Karim, Drug Metab. Dispos., 2000, 28, 308.
117. J. J. Talley, S. R. Bertenshaw, D. L. Brown, J. S. Carter, M. J. Graneto,
M. S. Kellog, C. M. Koboldt, J. Yuan, Y. Y. Zhang and K. Seibert,
J. Med. Chem., 2000, 43, 1661.
118. V. A. Mehta, R. Johnston, R. Cheung, A. Bello and R. M. Langford,
Clin. Pharmacol. Ther., 2008, 83, 430.
119. S. F. Barton, F. F. Langeland, M. C. Snabes, D. LeComte, M. E. Kuss,
S. S. Dhadda and R. C. Hubbard, Anesthesiology, 2002, 97, 306.
120. P. E. Golstein, A. Boom, J. van Geffel, P. Jacobs, B. Masereel and
R. Beauwens, Pflügers Arch–Eur. J. Physiol., 1999, 437, 652.
121. L. Payen, L. Delugin, A. Courtois, Y. Trinquart, A. Guillouzo and
O. Fardel, Br. J. Pharmacol., 2001, 132, 778.
122. C. Gedeon, J. Behravan, G. Koren and M. Piquette-Miller, Placenta,
2006, 27, 1096.
123. D. A. Smith, B. C. Jones and D. K. Walker, Med. Res. Rev., 1996, 16,
243.
124. R. S. Obach, F. Lombardo and N. J. Waters, Drug Metab. Dispos., 2008,
36, 1385.
125. W. Rupp, O. Christ and W. Heptner, Arzneimittelforschung, 1969, 19,
1428.
126. B. K. Wong, P. J. Bruhin and J. H. Lin, Pharm. Res., 1994, 11, 438.
127. B. Beerman, K. Hellström, B. Lindström and A. Rosén, Clin. Pharmacol.
Ther., 1975, 17, 424.
128. S. Yu, S. Li, H. Yang, F. Lee, J.-T. Wu and M. G. Qian, Rapid Commun.
Mass Spectrom., 2005, 19, 250.
129. J. T. Lettieri and S. T. Portelli, J. Pharmacol. Exp. Ther., 1983, 224, 269.
270 Chapter 5
130. D. K. Walker, M. J. Ackland, G. C. James, G. J. Muirhead, D. J. Rance,
P. Wastall and P. A. Wright, Xenobiotica, 1999, 29, 297.
131. J. Y. Zhang, J. J. Yuan, Y. F. Wang, R. H. Bible Jr. and A. P. Breau,
Drug Metab. Dispos., 2003, 31, 491.
132. P. C. Wilkins, H. K. Jacobs, M. D. Johnson and A. S. Gopalan, Inorg.
Chem., 2004, 43, 7877.
133. F. T. Bonner and Y. Ko, Inorg. Chem., 1992, 31, 2514.
134. R. Zamora, A. Grzesiok, H. Weber and M. Feelisch, Biochem. J., 1995,
312, 333.
135. A. Grzeslok, H. Weber, R. P. Zamora and M. Feelisch, in Biology of
Nitric Oxide, ed. M. Moncada, M. Feelisch, R. Busse and E. A. Higgs,
Portland Press, London, 1994, pp. 238–241.
136. P. Erdélyi, T. Fodor, A. K. Varga, M. Czugler, A. Gere and J. Fischer,
Bioorg. Med. Chem., 2008, 16, 5322.
137. S. K. Paulson, J. Y. Zhang, A. P. Breau, J. D. Hribar, N. W. K. Liu, S. M.
Jessen, Y. M. Lawal, J. N. Cogburn, C. J. Gresk, C. S. Markos, T. J.
Maziasz, G. L. Schoenhard and E. G. Burton, Drug Metab. Dispos., 2000,
28, 514.
138. J. W. Bridges, S. R. Walker and R. T. Williams, Biochem. J., 1969, 111,
173.
139. P. M. Vyas, S. Roychowdhury, F. D. Khan, T. E. Prisinzano, J. Lamba,
E. G. Schuetz, J. Blaisdell, J. A. Goldstein, K. L. Munson, R. N. Hines
and C. K. Svensson, J. Pharmacol. Exp. Ther., 2006, 319, 488.
140. J. M. Hutzler, D. Kolwankar, M. A. Hummel and T. S. Tracy, Drug
Metab. Dispos., 2002, 30, 1194.
141. J. H. Shon, Y. R. Yoon, M. J. Kim, K. A. Kim, Y. C. Lim, K. H. Liu,
D. H. Shin, C. H. Lee, I. J. Cha and J. G. Shin, Br. J. Clin. Pharmacol.,
2005, 59, 552.
142. D. J. Elliott Suharjono, B. C. Lewis, E. M. Gillam, D. J. Birkett, A. S.
Gross and J. O. Miners, Br. J. Clin. Pharmacol., 2007, 64, 450.
143. C. Tang, M. Shou, Q. Mei, T. H. Rushmore and A. D. Rodrigues,
J. Pharmacol. Exp. Ther., 2000, 293, 453.
144. K. D. Rainsford, in Milestones in Drug Therapy–COX-2 Inhibitors, ed.
M. Pairet and J. van Ryn, Birkhäuser Verlag, Basel, 2004, pp. 67–131.
145. M. L. Ecker, L. E. Visser, P. H. Trienekens, A. Hofman, R. H. van Schaik
and B. H. Stricker, Clin. Pharmacol. Ther., 2008, 83, 288.
146. J. Kirchheiner and J. Brockmoller, Clin. Pharmacol. Ther., 2005, 77,
1–16.
147. J. Kirchheiner, E. Störmer, C. Meisel, N. Steinbach, I. Roots and J.
Bröckmoller, Pharmacogenetics, 2003, 13, 473.
148. V. Kumar, J. L. Wahlstrom, D. A. Rock, C. J. Warren, L. A. Gorman
and T. S. Tracy, Drug Metab. Dispos., 2006, 34, 1966.
149. K. Komatsu, K. Ito, Y. Nakajima, S. Kanamitsu, S. Imaoka, Y. Funae,
C. E. Green, C. A. Tyson, N. Shimada and Y. Sugiyama, Drug Metab.
Dispos., 2000, 28, 475.
Sulfonamide as an Essential Functional Group in Drug Design 271
150. H. Malhi, B. Atac, A. K. Daly and S. Gupta, Postgrad. Med. J., 2004, 80,
107.
151. P. A. Williams, J. Cosme, A. Ward, H. C. Angove, D. Matak Vinkovic
and H. Jhoti, Nature, 2003, 424, 464.
152. Y. Yao, W. W. Han, Y. H. Zhou, Z. S. Li, Q. Li, X. Y. Chen and D. F.
Zhong, Eur. J. Med. Chem., 2009, 44, 854.
153. M. M. Ahlström, M. Ridderström, I. Zamora and K. Luthman, J. Med.
Chem., 2007, 50, 4444.
154. S. Y. Ahn and V. Bhatnagar, Curr. Opin. Nephrol. Hypertens., 2008, 17,
499.
155. E. babu, M. Takeda, S. Narikawa, Y. Kobayashi, A. Enomoto, A. Tojo,
S. H. Cha, T. Sekine, D. Sakthisekaran and H. Endou, Biochem. Biophys.
Acta, 2002, 1590.
156. H. E. Ives, in Basic and Clinical Pharmacology, ed. B. G. Katzung,
McGraw-Hill, New York, 2001, pp. 245–264.
157. J. B. Hook and H. E. Williamson, J. Pharmacol. Exp. Ther., 1965, 149,
404.
158. J. H. Gustafson and L. Z. Benet, J. Pharmacokinet. Biopharm., 1981, 9,
461.
159. P. Chennavasin, R. Seiwell, D. C. Brater and W. M. Liang, Kidney Int.,
1979, 16, 187.
160. K. R. Sweeney, D. J. Chapron, J. L. Brandt, I. H. Gomolin, P. U. Feig
and P. A. Kramer, Clin. Pharmacol. Ther., 1986, 40, 518.
161. H. Hasannejad, M. Takeda, K. Taki, H. J. Shin, E. Babu, P. Jutabha, S.
Khamdang, M. Aleboyeh, M. L. Onozato, A. Tojo, A. Enomoto, N.
Anzai, S. Narikawa, X.-L. Huang, T. Niwa and H. Endou, J. Pharmacol.
Exp. Ther., 2004, 308, 1021.
162. J. A. Taylor, Clin. Pharmacokinet., 1972, 13, 710.
163. P. Marchetti, R. Giannarelli, A. D. Carlo and R. Navalesi, Clin. Phar-
macokinet., 1991, 21, 308.
164. L. Balant, Clin. Pharmacokinet., 1981, 6, 215.
165. J. Judis, J. Pharm. Sci., 1973, 62, 1906.
166. M. J. Crooks and K. F. Brown, J. Pharm. Pharmacol., 1974, 26, 304.
167. Y. Uwai, H. Saito, Y. Hashimoto and K.-I. Inui, Eur. J. Pharmacol.,
2000, 398, 193.
168. T. B. Vree, Y. A. Hekster, J. E. Damsma, M. Tijhuis and W. T. Friesen,
Eur. J. Clin. Pharmacol., 1981, 20, 283.
169. T. B. Vree, Y. A. Hekster, J. E. Damsma, R. van Dalen, J. C. Haf-
kenscheid and W. T. Friesen, Ther. Drug Monit., 1981, 3, 129.
170. T. B. Vree, Y. A. Hekster, M. W. Tijhuis, M. Baakman, M. J. Oosterbaan
and E. F. Termond, Pharm. Weekbl. Sci., 1984, 6, 150.
171. A. Despoppulos and P. X. Callahan, Am. J. Physiol., 1962, 203, 19.
172. L. Z. Cooper, M. A. Madoff and L. Weinstein, Antibiot. Chemother.,
1962, 12, 618.
173. S. A. Tilles, South. Med. J., 2001, 94, 817.
272 Chapter 5
174. A. E. Cribb, B. L. Lee, L. A. Trepanier and S. P. Spielberg, Adverse Drug
React. Toxicol. Rev., 1996, 15, 9.
175. G. Choquet-Kastylevsky, T. Vial and J. Descotes, Curr. Allergy Asthma
Rep., 2002, 2, 16.
176. A. S. Kalgutkar, I. Gardner, R. S. Obach, C. L. Shaffer, E. Callegari,
K. R. Henne, A. E. Mutlib, D. K. Dalvie, J. S. Lee, Y. Nakai,
J. P. O’Donnell, J. Boer and S. P. Harriman, Curr. Drug Metab., 2005, 6,
161.
177. A. E. Cribb, S. P. Spielberg and G. P. Griffin, Drug Metab. Dispos., 1995,
23, 406.
178. A. Carr, B. Tindall, R. Penny and D. A. Cooper, Clin. Exp. Immunol.,
1993, 94, 21.
179. S. N. Lavergne, J. R. Kurian, S. U. Bajad, J. E. Maki, A. R. Yoder, M. V.
Guzinski, F. M. Graziano and L. A. Trepanier, Toxicology, 2006, 222, 25.
180. J. P. Sanderson, D. J. Naisbitt, J. Farrell, C. A. Ashby, M. J. Tucker, M.
J. Rieder, M. Pirmohamed, S. E. Clarke and B. K. Park, J. Immunol.,
2007, 178, 5533.
181. D. J. Naisbitt, S. F. Gordon, M. Pirmohamed, C. Burkhart, A. E. Cribb,
W. J. Pichler and B. K. Park, Br. J. Pharmacol., 2001, 133, 295.
182. D. J. Naisbitt, J. Farrell, S. F. Gordon, J. L. Maggs, C. Burkhart,
W. J. Pichler, M. Pirmohamed and B. K. Park, Mol. Pharmacol., 2002, 62,
628.
183. A. E. Cribb, M. Miller, J. S. Leeder, J. Hill and S. P. Spielberg, Drug
Metab. Dispos., 1991, 19, 900.
184. D. J. Naisbitt, P. M. Neill, M. Pirmohamed and B. K. Park, Bioorg. Med.
Chem. Lett., 1996, 6, 1511.
185. H. E. Callan, R. E. Jenkins, J. L. Maggs, S. N. Lavergne, S. E. Clarke,
D. J. Naisbitt and B. K. Park, Chem. Res. Toxicol., 2009, 22, 937.
186. M. J. Rieder, J. Uetrecht, N. H. Shear and S. P. Spielberg, J. Pharmacol.
Exp. Ther., 1988, 244, 724.
187. A. E. Cribb, C. E. Nuss, D. W. Alberts, D. B. Lamphere, D. M.
Grant, S. J. Grossman and S. P. Spielberg, Chem. Res. Toxicol., 1996, 9,
500.
188. P. M. Vyas, S. Roychowdhury, F. D. Khan, T. E. Prisinzano, J. Lamba,
E. G. Schuetz, J. Blaisdell, J. A. Goldstein, K. L. Munson, R. N. Hines
and C. K. Svensson, J. Pharmacol. Exp. Ther., 2006, 319, 488.
189. S. Roychowdhury, P. M. Vyas, T. P. Reilly, A. A. Gaspari and C. K.
Svensson, J. Pharmacol. Exp. Ther., 2005, 314, 43.
190. P. Wolkenstein, V. Carriere, D. Charue, S. Bastuji-Garin, J. Revuz, J. C.
Roujeau, P. Beaune and M. Bagot, Pharmacogenetics, 1995, 5, 255.
191. T. A. Baillie, Chem. Res. Toxicol., 2008, 21, 129.
192. C. Ju and J. P. Uetrecht, Curr. Drug Metab., 2002, 3, 367.
193. D. A. Smith and R. M. Jones, Curr. Opin. Drug Discov. Develop., 2008,
11, 72.
194. R. Patterson, A. E. Bello and J. Lefkowith, Clin. Ther., 1999, 21, 2065.
Sulfonamide as an Essential Functional Group in Drug Design 273
195. L. E. Shapiro, S. R. Knowles, E. Weber, M. G. Neuman and N. H. Shear,
Drug Saf., 2003, 26, 187.
196. D. C. Dahlin, G. T. Miwa, A. Y. Lu and S. D. Nelson, Proc. Natl. Acad.
Sci. U.S.A., 1984, 81, 1327.
197. H. Tan, W. M. C. Ong, S. H. Lai and W. C. Chow, Singapore Med. J.,
2007, 48, 582.
198. M. A. Macia, A. Carvajal, J. G. del Pozo, E. Vera and A. del Pino, Clin.
Pharmacol. Ther., 2002, 72, 596.
199. F. Li, M. D. Chordia, T. Huang and T. L. Macdonald, Chem. Res.
Toxicol., 2009, 22, 72.
200. D. P. Walker, F. C. Bi, A. S. Kalgutkar, J. N. Bauman, S. X. Zhao, J. R.
Soglia, G. E. Aspnes, D. W. Kung, J. Klug-McLeod, M. P. Zawistoski,
M. A. McGlynn, R. Oliver, M. Dunn, J. C. Li, D. T. Richter, B. A.
Cooper, J. C. Kath, C. A. Hulford, C. L. Autry, M. J. Luzzio, E. J. Ung,
W. G. Roberts, P. C. Bonnette, L. Buckbinder, A. Mistry, M. C. Griffor,
S. Han and A. Guzman-Perez, Bioorg. Med. Chem. Lett., 2008, 18,
6071.
201. H. Sun, R. Sharma, J. Bauman, D. P. Walker, M. P. Zawistoski, G. E.
Aspnes and A. S. Kalgutkar, Bioorg. Med. Chem. Lett., 2009, 19, 3177.
202. J. W. Clapp, J. Biol. Chem., 1956, 223, 207.
203. D. F. Colucci and D. A. Buyske, Biochem. Pharmacol., 1965, 14, 457.
204. O. W. Woltersdorf Jr., H. Schwam, J. B. Bicking, S. L. Brown, S. J.
deSolms, D. R. Fishman, S. L. Graham, P. D. Gautheron, J. M. Hoff-
man, R. D. Larson, W. S. Lee, S. R. Michelson, C. M. Robb, N. N. Share,
K. L. Shepard, A. M. Smith, R. L. Smith, J. M. Sondey, K. M. Stroh-
maier, M. F. Sugrue and M. P. Viader, J. Med. Chem., 1989, 32, 2486.
205. S. L. Graham, K. L. Shepard, P. S. Anderson, J. J. Baldwin, D. B. Best,
M. E. Christy, M. B. Freedman, P. Gautheron, C. N. Habecker, J. M.
Hoffman, P. A. Lyle, S. R. Michelson, G. S. Ponticello, C. M. Robb,
H. Schwam, A. M. Smith, R. L. Smith, J. M. Sondey, K. M. Strohmaier,
M. F. Sugrue and S. L. Varga, J. Med. Chem., 1989, 32, 2548.
206. G. D. Hartman, W. Halczenko, R. L. Smith, M. F. Sugrue, P. J. Mal-
lorga, S. R. Michelson, W. C. Randall, H. Schwam and J. M. Sondey,
J. Med. Chem., 1992, 35, 3822.
207. S. L. Graham, J. M. Hoffman, P. Gautheron, S. R. Michelson, T. H.
Scholz, H. Schwam, K. L. Shepard, A. M. Smith, R. L. Smith, J. M.
Sondey and M. F. Sugrue, J. Med. Chem., 1990, 33, 649.
208. K. A. Koeplinger, Z. Zhao, T. Peterson, J. W. Leone, F. S. Schwende, R. L.
Heinrikson and A. G. Tomasselli, Drug Metab. Dispos., 1999, 27, 986.
209. Z. Zhao, K. A. Koeplinger, T. Peterson, R. A. Conradi, P. S. Burton,
A. Suarato, R. L. Heinrikson and A. G. Tomasselli, Drug Metab. Dispos.,
1999, 27, 992.
210. B. Forouzesh, C. H. Takimoto, A. Goetz, S. Diab, L. A. Hammond,
L. Smetzer, G. Schwartz, R. Gazak, J. T. Callaghan, D. D. Von Hoff and
E. K. Rowinsky, Clin. Cancer Res., 2003, 9, 5540.
274 Chapter 5
211. C. B. Pratt, L. C. Bowman, N. Marina, A. Pappo, L. Avery, X. Luo and
W. H. Meyer, Invest. New Drugs, 1995, 13, 63.
212. A. Saili and M. S. Sarna, Indian Pediatr., 1991, 28, 936.
213. C. M. Jochheim, M. R. Davis, K. M. Baillie, W. J. Ehlhardt and T. A.
Baillie, Chem. Res. Toxicol., 2002, 15, 240.
214. X. Guan, M. R. Davis, C. Tang, C. M. Jochheim, L. Jin and T. A. Baillie,
Chem. Res. Toxicol., 1999, 12, 1138.
215. B. F. Hales and H. Brown, Teratology, 1991, 44, 251.
CHAPTER 6

Influence of Aromatic Rings on

ADME Properties of Drugs
DEEPAK DALVIE, SAJIV NAIR, PING KANG AND
CHO-MING LOI

Pharmacokinetics, Dynamics and Metabolism and Department of Medicinal

Chemistry, Pfizer Global Research and Development, San Diego, CA 92121,
USA

6.1 Introduction
Aromatic rings are commonly found in most drugs. Historically, these rings have
been a classic structural feature of drugs ever since the discovery of aspirin in the
late 1800s. Since then several drug molecules that span all therapeutic areas have
been designed and contain a combination of these rings along with other side
chains and functional groups. The best selling drugs shown in Figure 6.1, as well
as other marketed drugs, contain at least one aromatic ring system.
Carbocyclic six-membered aromatic rings are widely present in both natural
products and endogenous ligands, and are therefore commonly incorporated
by medicinal chemists into new chemical entities. Aromatic rings are also
preferred replacements of linear and branched alkyl/cycloalkyl groups. Such a
replacement not only imparts greater rigidity to the molecule but can poten-
tially increase the non-covalent interactions between the macromolecule and
the small molecules. These interactions generally involve aromatic amino acid
side chains of the receptor and/or aromatic and heteroaromatic rings of the
ligand. Analysis of the ligand–protein complexes reveals that aromatic rings are

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

275
276 Chapter 6

O O
O O HN
O NH
N O
N
S O
Cl O
HO HO ClO NH2
O HO
Clopidrogel
OH Atorvastatin Amlodipine
F
O N

O N F
F
N
Aspirin HN
Cl NH F
N S O
H Cl
Olazapine Sertraline Fluoxetine

Figure 6.1 Some drugs containing aromatic rings.

primary involved in p–p stacking, O–H/p, and cation–p interactions.1 From a

drug design perspective, aromatic rings and their functionalised counterparts
can be considered as versatile pharmacophores that increase these individual
non-bonding and electrostatic interactions between the ligand and macro-
molecules. Hence, incorporation of these rings into a molecule can improve the
binding affinity and therefore the potency of newly synthesised analogs relative
to their aliphatic or alicyclic ring congeners.
The presence of an aromatic ring in a molecule also provides a platform for
functionalisation of a compound. It serves as a core that helps to orient other
groups in a structure in the right direction and therefore enhance the inter-
actions with functionalities in the receptor. This chapter reviews the influence of
phenyl and substituted phenyl rings on the adsorption, distribution, meta-
bolism and excretion (ADME) properties of the molecules. Wherever possible,
calculated physicochemical properties (clogP or clogD) are used to illustrate the
effect of the aromatic substituent on ADME properties.

6.2 Physicochemical Properties of Aromatic and

Substituted Aromatic Rings
The most basic structure in aromatic rings is the benzene ring (or a phenyl
group). It is essentially a fully unsaturated analog of a cyclohexyl ring with a
hydrogen to carbon ratio equal to one (lower than most other rings and
functionalities). In spite of its low H : C ratio, these rings confer exceptional
chemical stability compared with simple unsaturated congeners (e.g. the
cycloalkene or the open chain hexatriene). The high thermodynamic and che-
mical stability of this system is attributed to the delocalisation of p-orbitals
containing 6p-electrons (resonance stabilisation of a conjugated cyclic triene).
Influence of Aromatic Rings on ADME Properties of Drugs 277

O O O O
O O S
H3C S N N N N
H H H H
H3C

Compound 6.1 Tolbutamide

clogP = 3.09 clogP = 2.36

Figure 6.2 Comparison of cLogP of alicylic and aromatic analogs.

The phenyl ring is a neutral moiety and therefore its presence in a molecule
does not affect its pKa. Further, in the absence of any hydrogen bonding
capabilities of the phenyl ring, the inclusion of this group in a molecule does not
have an effect on its polar surface area (PSA). However, addition of a phenyl
ring into a molecule addresses another important physicochemical property
such as lipophilicity and therefore influences the absorption, distribution and
clearance (excretion and metabolism) properties of a compound. Although less
lipophilic relative to cyclohexane ring (clogP of cyclohexane is 3.16 and clogP
of benzene is 2.18), the clogP of a phenyl ring is still sufficiently high to change
the lipophilicity of the whole molecule. For example, a hypothetical methyl-
cyclohexyl analog of tolbutamide (compound 6.1) has a clogP of 3.09 while
tolbutamide has a clogP of 2.36 (Figure 6.2).
Addition of a substituent or a functional group in an aromatic ring in place
of hydrogen atom can significantly perturb the electronic, steric, hydrophobic,
hydrophilic and hydrogen-bonding parameters of the ring depending on the
characteristics of that group. These changes can affect the interactions of the
compound with the receptors and therefore alter its biological activity. For
example, the replacement of an aniline group in carbutamide with a tolyl group
in tolbutamide results in separation of the antibacterial effect from the anti-
diabetic effect (Figure 6.3).2 Similarly chlorothiazide is an antihypertensive
agent with a strong diuretic effect. Replacement of the sulphonamide moiety in
chlorothiazide with hydrogen atom results in an antihypertensive without the
diuretic activity (Figure 6.3).2
An isosteric change of a functional group or a substituent on a phenyl ring
has a major effect on the lipophilicity of the molecule and can therefore
influence the disposition (absorption, distribution or excretion) of compounds.
In addition to lipophilicity, the change in polarity brought about by pertur-
bation of electron distribution of the aromatic ring due to inclusion of func-
tional groups can also influence the metabolism of the compound in the body.
Hammett was the first to explore the effect of substituents on reactivity of the
phenyl ring.3,4 He demonstrated that both the inductive and resonance effect of
a substituent could affect the electron density of the ring. Based on these studies,
an electronic parameter, (s), also called as the Hammett constant, was assigned
to each substituent. The s value for each substituent is dependent on the elec-
tronic property of that substituent (electron withdrawing or electron donating)
and the position of the substituent on the ring (para or meta relative to other
group). The more electron withdrawing a substituent, the more positive is its
278 Chapter 6

R
H H
N N
S
OO O
R = NH2; Carbutamide
R = CH3; Tolbutamide

Cl N N

NH NH
H2NO2S S Cl S
O O O O
Chlorothiazide Diazoxide

Figure 6.3 Effect of functional group modification on the biological activity.

s value (relative to H, which is set at 0); conversely, the more electron donating,
the more negative is its s value. Table 6.1 shows the Hammett constants (smeta
and spara) for some commonly used groups. A list of Hammett constants for
several groups is given in the review by Hansch and co-workers.5
The Hammett relationship does hold true for the substituents at the ortho
position where it is complicated due to steric interactions and polar effects. As
the substituent becomes more electron withdrawing, it potentially reduces the
electron density of the phenyl ring, making the ring less reactive. In contrast,
the electron donating substituents enhance the reactivity of the ring. Thus,
substitution of the phenyl ring with a cyano or the nitro group deactivates the
phenyl ring and makes it less susceptible to oxidative metabolism, while
replacement of hydrogen atom with substituents such as the methyl group can
make the phenyl ring more prone to P450 catalysed oxidation.
Like the substituent constants derived by Hammett, Hansch and co-workers
derived constants for the contribution of individual functional groups to the
partition coefficient of the phenyl ring (and eventually to the logP of the overall
molecule).6,7 The lipophilicity substitution constant (p), dictates the influence of
a particular substituent on the partition coefficient (clogP) of the compound. The
positive value indicates that the substituent makes the phenyl ring more lipo-
philic and therefore increases the logP of the molecule. Likewise, the negative
value indicates that the substituent makes the molecule more polar by decreasing
its lipophilicity. Table 6.1 also shows the p values for various substituents that
are commonly incorporated in molecules. A comprehensive listing of all p values
for various substituents is depicted in the review by Hansch and co-workers.5
The third important parameter which addresses the interaction of a drug
with the receptor or an enzyme is the steric effect. The most widely used
parameter for steric substituent effects is the steric parameter, Es. This was
defined by Taft in much the same way as the Hammett’s electronic effects, but
was standardised to the methyl group (Es methyl ¼ 0.0) as opposed to the
hydrogen by Hammett.8,9 While the Hammett equation accounts for how field,
Table 6.1 Aromatic substituent constants of some commonly used functional groups

R
Molar refrac- No. of hydrogen No. of hydrogen
R Formula p sp sm tion (MR) bonddonors bond acceptors

H Hydrogen 0.00 0.00 0.00 1.03 0 0

F Fluoro 0.14 0.06 0.34 0.92 0 0
Cl Chloro 0.71 0.23 0.37 6.03 0 0
OH Hydroxy 0.67 0.37 0.12 2.85 1 1
OCH3 Methoxy 0.02 0.27 0.12 7.87 0 1
SH Mercapto 0.39 0.15 0.25 9.22 1 0
SCH3 Methylmercapto 0.61 0.00 0.15 13.82 0 0
SOCH3 Sulfoxide 1.58 0.49 0.52 13.70 0 1
SO2CH3 Methylsulfone 1.63 0.72 0.60 13.49 0 1
NO2 Nitro 0.28 0.78 0.71 7.36 0 1
NH2 Amino 1.23 0.66 0.16 5.42 1 1
NHCH3 Methylamino 0.47 0.84 0.30 10.33 1 1
N(CH3)2 Dimethylamino 0.18 0.83 0.15 15.55 0 1
NHSO2CH3 Methylsulfonamido 1.18 0.03 0.20 18.18 1 1
Influence of Aromatic Rings on ADME Properties of Drugs

COOH Carboxy 0.32 0.45 0.37 6.93 0 1

COOCH3 Methylcarboxylate 0.01 0.45 0.37 12.87 0 1
OCOCH3 Acetate 0.64 0.31 0.39 12.47 0 1
CONH2 Carboxamido 1.49 0.36 0.28 9.81 1 1
CN Cyano 0.57 0.66 0.56 6.33 0 1
COCH3 Methylketone 0.55 0.50 0.38 11.18 0 1
279
 280

Table 6.1 (continued )

R
Molar refrac- No. of hydrogen No. of hydrogen
R Formula p sp sm tion (MR) bonddonors bond acceptors

Phenyl 1.96 0.01 0.06 25.36 0 0

CRCH Ethynyl 0.40 0.23 0.21 9.55 0 0

HC¼CH2 Ethenyl 0.82 0.02 0.05 10.99 0 0
CH2CH3 Ethyl 1.02 0.15 0.07 10.30 0 0
CH3 Methyl 0.56 0.17 0.07 5.65 0 0
CH2CH2CH3 n-Propyl 1.55 0.13 0.07 14.96 0
0
CH(CH3)2 Isopropyl 1.53 0.15 0.07 14.96 0 0
Chapter 6
(CH2)3CH3 n-Butyl 0.02 0.27 0.12 7.87 0 0
CH2CH(CH3)2 iso-Butyl 2.05 0.12 19.59 0
0
CH(CH3)C2H5 sec-Butyl 2.04 0.12 19.59 0
0
C(CH3)3 tert-Butyl 1.98 0.20 0.10 19.62 0 0
SO2NH2 Sulfonamido 1.82 0.57 0.46 12.28 1 1
CF3 Trifluoromethyl 0.88 0.54 0.43 5.02 0 0
OCF3 Trifluoromethoxy 1.04 0.35 0.38 7.86 0 1
SO2CF3 Trifluoromethyl- 0.55 0.93 0.79 12.86 0 1
sulfone
CF2CF3 Pentafluoroethyl 1.68 0.52 0.47 9.23 0 0
CH2CN Cyanomethyl 0.57 0.01 0.16 10.11 0 1
CH2OH Hydroxymethyl 1.03 0.00 0.00 7.19 1 1
NHCONH2 Urea 1.30 0.24 0.03 13.72 1 1
NHCSNH2 Thiourea 1.40 0.22 0.16 22.19 1 1
NHCOCH3 Acetamido 0.97 0.0 0.21 14.93 1 1
NHCO2CH3 Methylcarbamate 0.37 0.15 0.07 16.53 1 1
CH2OCH3 Methoxymethyl 0.78 0.03 0.02 12.07 0 1
Cyclopropyl 1.14 0.21 0.07 13.53 0 0
Influence of Aromatic Rings on ADME Properties of Drugs
281
282 Chapter 6
inductive and resonance effects influence reaction rates, the Taft equation also
describes the steric effects of a substituent. This parameter is useful for studying
intramolecular steric effects.
However, since Es constants have not been determined for the majority of
substituents and since biochemical–biomedical ‘steric’ requirements are often
of the ‘bulk’ type, two parameters readily available for each substituent; molar
refraction (MR) and molecular weight (MW) are commonly employed to
understand the biological quantitative structure–activity relationship (QSAR).
The molar refraction values not only address the steric bulk of substituent but
also measure the electronic effects and therefore the dipole–dipole interactions
at the active site.10 The greater the MR value of a substituent, the larger is its
steric or bulk effect. Table 6.1 also gives MR values for some commonly used
substituents; a comprehensive listing of values for all substituents is presented
by Hansch and co-workers.10
Several other parameters have been described to assess the QSAR of com-
pounds. However, s, p and molecular weight are the three parameters com-
monly used in practice by most medicinal chemists to modulate the overall
biological and ADME properties of a molecule.
The following example illustrates the concept of the influence of substitution
of a phenyl ring or interchange of a functional group on lipophilicity and
solubility. Replacement of a hydrogen atom in synthesis of NPY5 receptor
antagonists11 6.2 with an ethyl in compound 6.3 or a CF3 group in compound
6.4 results in the change of clogP and calculated solubility (Table 6.2). This in

Table 6.2 Effect of substitution on the lipophilicity and solubility.

No. Structure cLogP cSolubility (mg/mL)

6.2 NC H 2.43 0.01

N S O
N

6.3 NC H 3.51 0.002

N S O
N

6.4 NC H 4.90 0.0001

N S O
CF3
N
Influence of Aromatic Rings on ADME Properties of Drugs 283
turn can affect the absorption and permeability of the molecule, and also affect
the distribution characteristics of the compound. The introduction of a CF3
group on the phenyl ring can possibly make the phenyl ring less reactive and
therefore less susceptible to P450 mediated oxidative metabolism.

6.2.1 Importance of Fluorine Substitution on Phenyl Rings

Fluorine group is the most commonly used functional group in the recent
years.12–15 Approximately 5–15% of the launched drugs contain a fluoro group
as a part of their structure. Some of the top-selling fluorinated pharmaceuticals
include the antidepressant fluoxetine (which contains fluorines in the form of a
CF3 group), the cholesterol-lowering drug atorvastatin (Figure 6.1).
Inclusion of a fluorine atom in a drug molecule can, in theory, influence both
the disposition of the drug and the interaction of the drug with its pharma-
cological target; hence it deserves a special mention with respect to impact and
influence on the physicochemical properties of a molecule.
The electronic properties of fluorine make it a unique atom. It is both
comparable to hydrogen and the most highly electronegative element in the
periodic table(3.98 Pauling scale). Even though it is considered as an isostere of
hydrogen, the size and stereoelectronic influences of the two atoms are quite
different. The van der Waal radius of fluorine is larger than hydrogen and
(1.47 versus 1.20Å) and much closer to that of oxygen (1.52Å), as is its elec-
tronegativity (3.98 for F versus 3.44 for O).16 The C–F bond length (1.39Å) is
also in the same range as the C–O bond length (1.40Å), but much bigger than
C–H (1.09Å).17,18
When fluorine is substituted for hydrogen in aryl ring, the resulting change in
the electron distribution affects the pKa, dipole moments, and overall reactivity
and stability of neighbouring functional groups within the molecule.14,15 For
instance, replacement of a hydrogen atom with one fluorine group can affect
the basicity of amines including aromatic amines (aniline) and the acidity of the
phenol.18 The magnitude of the change in the pKa depends on the distance
between the fluorine atom and the functional group (Table 6.3). Thus, the
presence of a fluorine atom ortho to a phenolic group or the amine is associated
with a reduced pKa, whereas meta fluoro substitution has much less effect in
both rings.
In terms of lipophilicity, H/F exchange leads to a more lipophilic molecule.
Whereas fluorination of alkanes will actually decrease lipophilicity, the addi-
tion of fluorine group on an aryl ring increases lipophilicity.14 Although the
value of clogP may differ depending upon the substitution of fluorine at the
ortho, para or meta position, there is an overall increase in the clogD as depicted
in Table 6.3.
Due to effect of fluorine on the lipophilicity and pKa, this group can influence
a number of different parameters in lead optimisation including solubility and
binding affinities (potency, selectivity), and therefore the absorption and dis-
tribution properties of a drug. Enhanced lipophilicity due to the replacement of
284 Chapter 6
Table 6.3 Effect of fluorine on aniline and phenol.
Substrate Structure pKa cLogD7.4

Aniline NH2 4.61 1.14

p-Fluoroaniline NH2 4.66 1.2

o-Fluoroaniline NH2 3.2 1.47

F

m-fluoroaniline NH2 3.6 1.47

Phenol OH 9.86 1.54

p-Fluorophenol OH 9.92 1.84

o-Fluorophenol OH 8.71 1.82

F

m-Fluorophenol OH 8.97 2.06

F
Influence of Aromatic Rings on ADME Properties of Drugs 285
hydrogen with a fluoro group can potentially favour the membrane penetration
and therefore an increased absorption and improved blood–brain barrier
penetration of a compound.
Fluorine also forms a strong bond with carbon. The C–F bond energy is
higher than the C–H or C–O bond (116, 99 and 85 kcal1, respectively).18 This
in turn increases the thermal and oxidative stability of the ring and makes the
C–F bond less sensitive to metabolic degradation.18 The electronic change also
makes the overall ring less reactive to P450 mediated oxidation.

6.3 Influence of Aromatic and Substituted Aromatic

Rings on ADME Properties of Compounds
As discussed earlier, incorporation of phenyl and substituted phenyl rings into
drugs can affect their physical properties and hence influence their pattern on
the disposition and toxicity. The following sections illustrate the influence of
these aromatic rings on the absorption, distribution and clearance (metabolism
and excretion) of drugs.

6.3.1 Absorption
A majority of drugs are administered orally. Two main factors influencing
intestinal absorption are the solubility of a compound in the gastrointestinal
fluid and its permeability through the gastrointestinal wall.19,20 Both these
parameters are dependent on several physicochemical properties of a molecule
such the molecular size, lipophilicity, polar surface area (PSA) and hydrogen
bond donating and accepting capacity of the polar atoms.21 Although mole-
cular size impacts the transport of molecules through biological membranes,
lipophilicity has been proven to be the most important parameter affecting both
the permeability and aqueous solubility of compounds.
The logP value of a molecule primarily governs its ability to partition into
biological membranes. Since addition of a phenyl ring to a molecule increases
its lipophilicity, it can have a positive influence on the membrane permeability
of the compound. Modulation of the log P/log D value of a compound by
adding a substituent on the ring will further enhance or reduce its absorption.
Substituents such as alkyl and trifluoromethyl groups or the halogen atoms (Cl
or F), which commonly increase the logP value of a molecule, can increase its
membrane permeability. Alternatively, inclusion of the polar groups such as
carboxy, carboxamido, cyano groups (which have a negative p value) can
negatively influence the permeation of the compound through the gastro-
intestinal membrane.
Fichert and co-workers have explored structure–permeability relationships,
particularly the effect on cell permeability of some commonly used functional
groups attached to the drug-like core structures in lead optimisation efforts
286 Chapter 6
Table 6.4 Effect of commonly used functional groups on cell permeability.22
R

N N
PCaco2106 cm/sec Log D

H 71.1 2.91
mCN 65.8 2.47
PCO2H 3.19 0.61
PSO2NH2 23.4 1.04
pSO3H 0.08 1.78
mNH2 57.4 1.92
mNHSO2CH3 23.4 2.08
mCONH2 25.1 1.70
mCO2H 1.56 0.72
mCH2NH2 27.6 0.21

(Table 6.4).22 For this set of compounds, those with log D values within the
range 40 to o3 showed high permeability, while compounds with log D values
o0 display variable permeability.
For simple drugs like b-adrenoceptor antagonists, the log D7.4 values are
remarkably predictive of the absorption potential. All these compounds have a
molecular weight that ranges from 250 to 310; however, their log D values differ
depending on the substituents on the phenyl ring. As seen in Table 6.5,
incorporation of the cyclohexane diol or the methyl carboxamido group in the
molecule results in a negative clogD7.4 value, which in turn affects the
absorption of these two drugs through the gastrointestinal tract. On the other
hand, introduction of the alkyl or phenyl group in betaxolol and propranolol
increases the lipophilicity of these compounds, which results in their 100%
absorption following oral administration23(Table 6.5)
Aqueous solubility dictates the amount of drug in solution and hence the
amount of drug available for absorption from the gastrointestinal tract.
Compounds with low solubility may suffer from dissolution-limited absorption.
Although an increase in lipophilicity enhances the permeability, a higher logP
value has a deleterious effect on the aqueous solubility of compounds. In
general, aqueous solubility is inversely proportional to lipophilicity, and
compounds with a high logD7.4 value (43) may show poor dissolution and
resultant poor bioavailability. For instance, promazine (Figure 6.4) has been
classified as a high permeability high solubility drug in the biopharmaceutics
classifications system while chlorpromazine belongs to high permeability-low
solubility class of compounds.24 The only difference in the two compounds is
the replacement of one of the hydrogen atoms in promazine with a chloro
group in chlorpromazine (Figure 6.4). This subtle change in chlorpromazine
results high permeability but low aqueous solubility (Figure 6.4). The effect on
solubility is probably due to the increase in the clogP of chlorpromazine relative
to promazine.
Influence of Aromatic Rings on ADME Properties of Drugs 287

S S

N Cl N

Promazine
Chlorpromazine
clogP = 4.69 N N
clogD7.4 = 2.73 clogP = 5.18
cSol = 0.4 mg/ml clogD7.4 = 3.24
cSol = 0.09 mg/ml

Figure 6.4 Structures and physicochemical properties of promazine and chlor-

promazine.

Even though incorporation of polar groups increases the aqueous solubility

of compounds, the increased hydrogen bonding capabilities and therefore the
increased polar surface area of these groups can affect the transfer of com-
pounds across the gastrointestinal membrane.21 Desolvation and the breaking
of hydrogen bonds becomes the rate-limiting step to transfer across the
membrane. Thus, in order optimise the absorption of a compound, it is
important to strike a balance between the hydrophilic and lipophilic parameters
of a compound.
The impact of hydrophilic substituents on absorption can be observed by
comparing the pharmacokinetics of two D2-type receptor antagonists, sulpiride
and remoxipride (Table 6.6). Sulpiride is a sulfonamide derivative which is
incompletely absorbed through the gastrointestinal tract and has a bioavail-
ability of only 27%.25 Remoxipride on the other hand is well absorbed (bioa-
vailability of 90%) and is comparable to other standard neuroleptics.26
Comparison of the physicochemical properties of these two drugs indicates that
sulpiride has a negative value (logD7.4 -1.07) and high PSA (110.1 Å2) which is
due to its increased number of hydrogen bond donors and acceptors. On the
other hand, remoxipride has a logD7.4 of 0.38 and a PSA of 50.8 Å2 (Table 6.6).
Hence the low absorption of sulpiride is attributed to its high polarity in
addition to low lipophilicity.
The drug efflux transporter P-glycoprotein (P-gp), which is known to confer
multidrug resistance in cancer chemotherapy, is also expressed in the apical
surface of intestinal epithelial cells.27 Hence, it can play an important role in
oral absorption of a drug. Although numerous drugs have been identified as P-
gp transporter substrates, no clear structure–activity relationships with respect
to the influence of phenyl and substituted phenyl rings can be drawn. Based on
clinical and animal studies with P-gp modulators, it has become apparent that
the effect of P-gp on drug absorption is not as important as generally believed.
It is important to note that, regardless of whether a compound is a substrate for
P-gp or not, a drug molecule has to passively gain intracellular access prior to
being acted upon by this transporter. Overall, this makes permeability and
solubility the most important factors for intestinal absorption, as described
above.
 288

Table 6.5 Physicochemical properties of representative beta-blockers and its influence on absorption.
Name Structure MW cLogP clogD7.4 PSA % Absorbed23

Nadolol OH 309 0.56 2.1 81.95 30

H
HO N

O
OH

Atenolol O 266 0.33 1.76 84.5 50

H 2N
O NH

OH

Acebutolol O 336 1.77 0.3 87.6 90

H
O N

N
H
OH
Chapter 6
Betaxolol 307 2.53 0.43 50.72 100
O
O NH

OH

Propranolol 259 2.9 0.79 41.49 100

HO
O

N
H
Influence of Aromatic Rings on ADME Properties of Drugs
289
290 Chapter 6

6.3.2 Distribution
Distribution of drugs entails passage of a compound from circulation into the
tissues. Like absorption, there is a trend for increasing tissue permeability with
increasing lipophilicity.21 Hence the incorporation of the phenyl rings in a
molecule is likely to increase its affinity for the tissues. As in the case of
absorption, the substituents that increase the logP value of the compounds will
enhance distribution into the tissues more than the polar groups. For example,
propranolol and alprenolol with a clogP value of 1.2 and 1.0, respectively, show
good penetration into the cerebrospinal fluid (CSF) from the plasma as shown
by the CSF–plasma ratio (Table 6.7), while nadolol and atenolol are not
permeable through the blood–brain barrier.28 Clearly, incorporation of
cyclohexanediol in the ring system makes the compound quite polar for it to
cross the blood–brain barrier, whereas very hydrophilic nature of atenolol
hinders its permeability into the brain as is the case in the absorption of these
two drugs.
A similar relationship between lipophilicity and transfer across the blood–
brain barrier has been observed among the benzodiazepines midazolam,
alprazolam and triazolam.29 As seen in Table 6.8, the increase in lipophilicity
increases the brain to serum ratio.
Plasma protein binding is also an important parameter that influences tissue
distribution. Only free drug in plasma can cross the membrane and equilibrate
with free drug in the tissues. Incorporation of neutral lipophilic substituents on
the phenyl ring tends to increase the binding of the drug to plasma proteins and
therefore reduce the free fraction that is available to transfer to the tissues.
Comparison of the brain to unbound serum concentration ratio of lorazepam,
in which one of the hydrogen atom on the phenyl ring is replaced by a chloro
group (Table 6.9), versus oxazepam suggested that extraction of the latter into
the brain was slightly enhanced.29 Since the change in lipophilicity is minimal
between the two compounds, the increase in brain exposure is attributed to
more unbound fraction of oxazepam.
It has been demonstrated that other factors such as the hydrogen bonding
capabilities of a compound can also affect its penetration across the blood–
brain barrier.30,31 Hence, the established relationship between permeability
across the blood–brain barrier, protein binding and lipophilicity may not
necessarily hold in all cases of lipophilic drugs. For example, in spite of
similar clogD and protein binding values of flunitrazepam compared to
diazepam (Table 6.10), its brain to serum ratio is much lower than diaze-
pam. This disconnect is probably attributed to the increase in PSA of
flunitrazepam (78.94 Å2) due to presence of a nitro group in one of its aromatic
rings.
The effect of hydrogen bond donor and polar functionality in the phenyl
ring on brain exposures was also demonstrated in the following example
(Table 6.11). Compound 6.5 was synthesised as one of the analogs of mGluR2
potentiators but exhibited poor brain exposure following administration to
rats.32 Replacement of the phenolic group with an ethoxy group (compound 6.6)
Table 6.6 Comparison of physicochemical properties of sulpiride and remoxipride.25,26
Name Structure MW clogD7.4 PSA Å2 CSolubility mg/mL %F

Sulpiride O 341 1.07 110.1 126 27

O

NH
O
N S
O
NH2

Remoxipride 371 0.38 50.8 84 90

O
Influence of Aromatic Rings on ADME Properties of Drugs

H
N
Br N
O O
291
292 Chapter 6
Table 6.7 Brain penetration of b-blockers and its relationship to lipophilicity.28
Compound Structure cLogP CSF/Plasma

Propranolol 1.2 1
HO
O

N
H

Alprenolol 1.0 1

O N
H
OH

Nadolol OH 1.5 ND
H
HO N

O
OH

Atenolol O 1.1 ND
H2N
O NH

OH

gave better brain levels and higher brain-to-plasma ratio (1.2:1). Detectable
levels for 6.6 were attributed to the drop in the polar surface area by 16 Å2
compared with compound 6.5, in addition to the increase in the log D of the
molecule.
Another factor that affects brain penetration is the efflux of the compound
from blood–brain barrier by P-gp transporter.27 In contrast to absorption, the
P-gp transporter plays an important role in brain uptake as it limits cellular
uptake of drugs from the blood circulation into the brain. As mentioned above,
the structure–activity relationships in terms of the role of the substituted phenyl
ring systems are not fully understood. However, an increase in brain exposure
via modulation of PSA, which in turn affects the P-gp efflux properties, has
been demonstrated by Roberts and co-workers.33 This group focused on the
reduction of polar surface to influence the efflux properties of compounds.
2-Imidazole a1A partial agonist (6.7) was synthesised for its potential utility in
treating stress urinary incontinence (Table 6.12). However, 6.7 showed poor
brain penetration (rat free plasma/CSF 1:0.08), which was attributed to high
P-gp-mediated efflux when assessed in MDCK MDR-1 cells (AB/BA ratio
of 7:29). The medicinal chemistry strategy focused on reducing the PSA of new
Influence of Aromatic Rings on ADME Properties of Drugs 293
Table 6.8 Comparison of brain to serum (B/S) ratio of midazolam, alprazo-
lam and triazolam and its physicochemical properties.29
Name Structure cLogD7.4 B/S ratioa

Midazolam 3.78 33.91

F
Cl N

N
N

Alprazolam 1.92 2.62

CI N

N
N
N

Triazolam 2.08 19.52

Cl
Cl N

N
N
N
a
B/S ratio ¼ brain to unbound serum ratio.

analogs in order to reduce the potential for P-gp recognition and therefore
improve blood–brain barrier penetration.
As shown in Table 6.12, the replacement of the sulfonamide with dialky-
lethers such as the methoxymethyl ether 6.8 resulted in retention of the a1A
partial agonist potency with no evidence of P-gp efflux. Since the log D values
of all the analogs made were similar, the modulation of the P-gp mediated efflux
of 6.7 by addition of the ether functionalities was attributed to the decrease in
PSA of the molecule. The brain penetration was further enhanced by increasing
the lipophilicity of the compounds via incorporation of the chloro and fluoro
groups (Table 6.12).
Although incorporation of polar groups in phenyl rings has a negative
influence on the permeability across the blood–brain barrier, the strategy of
substituting the aromatic hydrogen with polar moieties can be used in pre-
venting serious central nervous system (CNS) side effects that may be caused by
some lipophilic compounds. This has been demonstrated in the following two
294 Chapter 6
Table 6.9 Comparison of the brain to unbound serum concentration and
plasma protein binding of lorazepam and oxazepam.29
Compound Structure clogD7.4 Unbound fraction B/S ratioa

Lorazepam HO O 2.38 0.157 16.01

N NH
Cl

Cl

Oxezapam HO O 2.22 0.171 20.30

N NH

Cl
a
B/S ratio ¼ brain to unbound serum ratio.

Table 6.10 Relationship between unbound fraction, physicochemical prop-

erties and brain to serum ratios of diazepam and flunitrazepam.29
Name Structure Total B/Sa B/unbound Sb clogD7.4 PSA (Å2)

Diazepam O 3.34 26.05 2.8 32.67

N N

Cl

Flunitrazepam O 1.15 5.97 2.2 78.94

N N
F

NO2
a
Total B/S ¼ total brain to serum ratio,
b
B/unbound S ¼ brain to unbound serum concentration.

examples. Compound 6.11 (Figure 6.5), which was designed as a cardiotonic

agent, resulted in ‘bright visions’ upon administration to patients. The side
effect was probably due to CNS activity and was a result of permeability of the
drug across the blood–brain barrier. Kutter and Austel optimised the structure
Influence of Aromatic Rings on ADME Properties of Drugs 295
Table 6.11 Effect of replacing a phenolic group with an ethoxy group on brain
to plasma ratio.32
OH O
H3C

O
R O

No. R MW PSA (Å2) clogD HBDa B/Pb

6.5 HO 473 102 3.8 2 BQLc

HO2C

6.6 C2H5O 424 86 5.6 1 1.2

HO2C
a
HBD ¼ hydrogen bond donor,
b
B/P ¼ brain to plasma ratio,
c
BQL ¼ below quantitation limits.

by substituting the methoxy group in 6.11 with more hydrophilic substituents

that could potentially prevent blood–brain penetration.34 Replacement of a
methoxy group with a bioisosteric methylsulfoxide moiety in sulmazole (6.12)
sufficiently modified the lipophilicity of 6.11 from a clogP of 2.59 to 1.17
(Figure 6.5). Furthermore, the increase in the PSA of 6.12 from 60.03 to 87.08
prevented it from crossing the blood–brain barrier, thus eliminating the
unwanted CNS related side effects but retaining the desired activity.
Rubenstein and co-workers have also used a similar approach for limiting the
brain exposure of a class of antimitotic agents.35 Compounds 6.13 and 6.14
were synthesised as irreversible inhibitors of tubulin polymerisation and were
effective against a variety of tumors, including those that express the multidrug
resistant (MDR) phenotype. However, CNS toxicity was observed for both
compounds possibly due to the high brain levels of 6.13 and 6.14 (Table 6.13).
Synthesis of less lipophilic analogues of 6.13 was therefore undertaken in order
reduce brain exposure of these agents and therefore increase the therapeutic
window. Acyl derivatives of aniline 6.14 were designed that could potentially
increase the polar surface area and decrease log D of both 6.13 and 6.14 (Table
6.13). Introduction of the polar acyl groups (compounds 6.15–6.17) successfully
prevented the compounds from crossing the blood–brain barrier. Unfortu-
nately, analogs 6.15–6.17 were less potent and displayed weak concentration-
dependent inhibition of tubulin polymerisation in vitro than 6.13 and 6.14.

6.3.3 Clearance
Metabolism and excretion are the two primary routes of elimination of a drug
from the body. Like absorption and distribution processes, the physicochemical
296 Chapter 6
Table 6.12 Modulation of P-gp efflux by reducing total polar surface area.33
No. Structure MDCK-MDR1 ratio PSA (Å2) logD7.4

6.7 N NH
7/29 83 1.5

Cl
NH
H3CO2S

6.8 N NH
40/36 38 1.8

6.9 N NH
39/37 38 2.0

6.10 N NH
40/36 38 2.6

CI

properties of a drug and its structure can primarily dictate its route of elim-
ination.21 Among all the physicochemical properties, lipophilicity is the key
factor determining the ultimate route by which the drug will be eliminated.21 It
governs the accessibility of the drug to many of the drug metabolising enzymes
as well as the ability of a molecule to cross the renal tubular membrane. The
latter governs the urinary excretion of a compound. Since substituted aromatic
rings play a major role in modulating a drug’s lipophilicity, incorporation of a
Influence of Aromatic Rings on ADME Properties of Drugs 297

H H
N N O
O S
N N N
N
O O

Compound 6.11
Sulmazole 6.12
cLogP = 2.59
logP = 1.17
PSA = 60.03
PSA = 87.08

Figure 6.5 Structures and physicochemical properties of cardiotonic agents 1, and

sulmazole 2.

Table 6.13 Prevention of brain permeability by addition of polar acyl

groups.35
O O
O O O
S S O
NH NH S
F3C F3C NH
F3C
O H
N
F NH 2 N
OCH3 H R O
OCH3 OCH3
1 2
3 R = CH3
4R=H
5 R = CHCH3OH
No. MW PSA (Å2) clogD Brain levels

6.13 371 55 2.0 860

6.14 368 81 1.2 1200
6.15 481 114 0.36 BQLa
6.16 467 134 0.020 BQL
6.17 497 134 0.040 BQL
a
BQL ¼ not detected

substituted aromatic ring in a molecule influences the route of elimination of a

drug.

6.3.3.1 Excretion
Molecules that have a relatively small size (approximate molecular weight of
200–300) and are polar are generally excreted unchanged in the urine. Since
incorporation of substituted phenyl rings adds a certain degree of lipophilicity
to the molecule, most molecules with one or more phenyl rings generally
undergo metabolism. However, replacement of a hydrogen atom in an aro-
matic ring with polar functionalities can decrease the logD7.4 and can result in
excretion of the drug renally. For example, comparison of structures of various
beta blockers and their physicochemical properties (Table 6.14) indicates that
analogs with polar functionalities which increase the PSA and decrease the
 298

Table 6.14 Influence of physicochemical properties on renal clearance of b-blockers.36

Name Structure MW clogD7.4 PSA (Å2) % Excreted unchanged in urine

Sotalol 272 1.7 86.81 80

HN
NH
O
HO S
O

Practolol O 266 1.5 70.59 95

HN O NH

OH

Atenolol O 266 1.7 84.5 85

H2N
O NH

OH
Chapter 6
Metoprolol O 267 0.47 50.72 10
O NH

OH

Betaxolol 307 0.43 50.72 15

O
O NH

OH
Influence of Aromatic Rings on ADME Properties of Drugs
299
300 Chapter 6

O
S O
O HN S
O O
NH
N

Dofetilide
clogD7.4 = 0.52 (clogP = 1.38)
2
PSA = 121.57 Å

Figure 6.6 Structure and physicochemical properties of dofetilide.

clogD value (e.g. atenolol, sotalol and practolol) are more prone to undergo
renal excretion than those with lipophilic groups such as betaxolol and
metoprolol.36
The effect of polar functionality has also been demonstrated in the phar-
macokinetics of the neuroleptic sulpiride. Assessment of unchanged sulpiride in
the urine indicates that metabolism is a minor route of clearance for this
compound and about 70% of the dose is excreted unchanged in the urine.25 In
contrast, the analog of sulpiride, remoxipride is extensively metabolised and
only 25% of the dose is excreted in the urine.26 This can be correlated to the
physicochemical properties of these two drugs (Table 6.6). The highly polar
nature of sulpiride (clogD7.4 ¼ 1.07) is linked to the high renal clearance of the
drug relative to remoxipride.
Even though it is commonly believed that hydrophilic compounds are
excreted unchanged renally and that the log D value of renally excreted drugs is
generally o0, there are some discrepancies. For example, the class III antiar-
rhythmic agent, dofetilide, is an amine with a log D7.4 value of 0.52 (Figure 6.6).
The moderate lipophilicity of this compound suggests that it has the potential
to undergo oxidative metabolism by P450. However, assessment of the renal
clearance of dofetilide in humans indicates that 80% of the compound is
excreted unchanged in the urine.37 This is probably due to the presence of the
two phenyl sulfonamide groups in the molecule, which increase the hydrogen
bonding potential and therefore the PSA of the molecule (121.57 Å2). Taken
together, dofetilide more closely resembles atenolol or sotalol as reflected by the
renal clearance of the compound.

6.3.3.2 Biliary Excretion

Several uptake and efflux transporters are present in the hepatobiliary system.38
These are quite broad in their specificity and have a high capacity for elim-
inating metabolites of drugs via the bile. In some instances, these transporters
are responsible for the excretion of unchanged drug in the bile. Although the
structure–activity relationship of these transporters is unknown, a rough cor-
relation between the physicochemical parameters of a drug and its substrate
property for the biliary transporters is exemplified in the following example.
Compound 6.18 is a prostaglandin D2 receptor antagonist which is primarily
Influence of Aromatic Rings on ADME Properties of Drugs 301
excreted unchanged in the bile after administration to bile-duct cannulated rats
(Table 6.15).39 This result suggested a role for active transport of parent from
plasma circulation into the bile. Modification of 6.18, with substituents that
decreased its PSA (Table 6.15) reduced the biliary clearance significantly, as
observed for 6.19 and 6.20. Similarly, there was a trend towards increasing log
D7.4 value and decreasing biliary concentration of the parent.

6.3.3.3 Metabolism
As mentioned earlier, metabolism is a primary route of clearance for most
compounds containing one or more aromatic rings. Lipophilic compounds are
susceptible to oxidation by liver enzymes (namely cytochrome P450) as well as
by other conjugative enzymes such as uridinylglucuronyltransferase (UGT).21
Incorporation of phenyl rings into a molecule affects its metabolism in two
ways. Firstly, the presence of unsubstituted or substituted aromatic rings in the
molecule increases its lipophilicity; this facilitates an easy transfer of the
molecule across the membrane and brings it into close proximity to the xeno-
biotic metabolising enzymes. Secondly, the aromatic rings themselves become a
site of metabolism. The latter, however, is dependent upon the reactivity of the
ring. As discussed in Section 6.2, an aromatic ring is a delocalised system with
6p electrons. Although aromatic molecules display enhanced stability due to
the delocalisation of these electrons, these rings can undergo electrophilic
aromatic substitution reactions that involve replacement of hydrogen atom in
the ring by an electrophilic substituent. In the case of CYP450 catalysed oxi-
dations, the hypervalent iron–oxene complex is widely accepted as ‘the oxidant
species’ that is ultimately responsible for the oxidation of compounds.40 This
high-energy oxo-species can add into the aromatic ring to produce a tetrahedral
intermediate that subsequently rearranges via an epoxide (an arene oxide) or a
ketone intermediate and ultimately results in an arenol (Figure 6.7).
Like electrophilic aromatic substitution reactions, the substituents on the
ring influence the oxidation of the phenyl ring.41 Hence, aromatic rings that are
activated (contain electron rich substituents) are more susceptible to aromatic
hydroxylation.42 For example, aniline containing analogs are hydroxylated
extensively. The final location of the oxygen is further explained by simple
organic principles. Ring opening of the arene oxide of compounds containing
electron rich rings is expected to occur in the direction that results in a more
stable carbocation. Thus, aniline derivatives are hydroxylated in the ortho and
para positions of the amino group as observed in atorvastatin (Figure 6.8).43,44
On the other hand, deactivated rings (rings with electron withdrawing
groups) are hydroxylated at a slower rate or not at all. For example, the uri-
cosuric agent, probenecid, undergoes no detectable aromatic hydroxylation
(Figure 6.9).45,46 The presence of two deactivating groups on the ring—the
carboxy group and the sulfonamide moiety—results in a significant decrease in
the electron density of the phenyl ring. The same principle applies to drugs that
contain two phenyl rings. In case of chlorpromazine or fluphenazine, the
 302

Table 6.15 Correlation of biliary clearance of D2 prostaglandin receptor antagonists with their physicochemical properties.39
No. Concentration in rat bile (mM) clogD7.4 cLogP PSA MW pKa

6.18 H3CO2S 1100 0.16 2.96 101.82 459.9 4.56

CO2H
N
COCH3 Cl

6.19 F 81 1.19 3.98 59.3 399.8 4.56

CO2H
N
COCH3 Cl

6.20 F 103 0.85 3.66 84.7 435 4.54

CO2H
N
SO2 CH3 Cl
Chapter 6
Influence of Aromatic Rings on ADME Properties of Drugs 303

H
O
H
a
R R Epoxide
Intermediate R
b H
O b
a Fe 3+ OH
O
b Arenol
Fe 4+
R

O
HH
Ketone
Intermediate

Figure 6.7 Oxidation of aromatic rings by iron-oxo species.

OH
OH
N O
H
F N OH
O

para-Hydroxyatorvastatin
HO
OH
OH
N O F
H
N OH
O
Atorvastatin OH
OH
N O
H
N OH
O
OH
ortho-Hydroxyatorvastatin

Figure 6.8 Structure of atorvastatin and its CYP3A4 mediated metabolites.

unsubstituted phenyl ring is hydroxylated while no oxidation is observed on the

less reactive substituted ring (Figure 6.9).47,48
In light of these observations, several strategies to reduce the metabolic
clearance of compounds have been employed. One strategy involves
304 Chapter 6

S S
HO2C

N N CI
S N Cl
O O
Probenecid
N
N
Chlorpromazine N Fluphenazine
HO

Figure 6.9 Structures of probenecid, chlorpromazine and fluphenazine. The arrow

indicates the site of metabolism

CF3 CF3 CF3

O
H H
N N O O O
CF3 CF3 CF3

N N N N
H H
N N
O O F
CP-99994 L-733060 HN NH L-741671 HN NH Aprepitant

Figure 6.10 Structures of NK1 receptor antagonists.

inactivation of aromatic rings towards oxidation by substituting them with

strongly electron withdrawing groups (e.g. CN, SO2NH2, SO3H). Since such
groups reduce the lipophilicity of the molecule, it becomes necessary to com-
pensate by adding a lipophilic group at another site in the molecule to retain
activity. Alternatively, a lipophilic trifluoromethyl group (CF3) is frequently
incorporated into the aromatic ring in the compound. Not only does the CF3
group add lipophilicity to the molecule but it also leads to the deactivation of
the ring.
The following example illustrates the use of trifluoromethyl groups to
improve metabolism. CP-99994 (Figure 6.10) was one of the lead NK1 receptor
antagonists but was discontinued from phase II clinical trials due to poor
bioavailability,49 which was attributed to the compound’s high metabolic lia-
bility. To improve bioavailability, the molecule was modified by replacing the
methoxyphenyl moiety with a 3,5-bistrifluoromethylphenyl moiety. This
modification resulted in enhanced activity and improved metabolism. Further
improvement in metabolic liability and basicity finally resulted in aprepitant
(Figure 6.10).49
A more frequently employed strategy to circumvent the problem of meta-
bolic instability involves blocking the reactive site by the introduction of a
halogen atom (Cl or F). The design of fenclofenac (an analog of diclofenac)
illustrates the role of a halogen atom in potentially blocking the site of meta-
bolism. Diclofenac is metabolised relatively rapidly via hydroxylation of the 4-
position of the dichlorophenyl ring (Figure 6.11), yielding 4-hydroxydiclofenac
Influence of Aromatic Rings on ADME Properties of Drugs 305

CO2H CO2H CO2H

HN HN HN
Cl CI Cl Cl Cl

OH Cl
Diclofenac Hydroxydiclofenac Fenclofenac

Figure 6.11 Metabolism of diclofenac and structure of fenclofenac

and resulting in a short half-life in humans. A change in the position of the

chloro group from the 6-position to the 4-position in fenclofenac not only
improves its metabolic stability but also increases its half-life to 20 h.50
Given the unique properties of fluorine (see Section 6.2) and the convenient
methods of introducing it, this group is more popular and is commonly used to
alter the rate and route of metabolism, and therefore improve the bioavail-
ability and half-lives of compounds of interest.14,17,18 As described earlier, the
carbon–fluorine bond is much more resistant to direct chemical attack by
cytochrome P450 in comparison to the carbon–hydrogen bond. The intro-
duction of a fluorine on a phenyl ring is also said to decrease the rate of
interaction between the p system and the iron-oxo intermediate.17 Thus even
though the aromatic rings with a fluorine group are metabolised, the rate of
metabolism is slower than the unsubstituted aromatic ring.
Some examples of where introduction of the fluoro group in the molecule
alters the metabolic rate are demonstrated in the discovery of ezetimibe, a
cholesterol absorption inhibitor, and a calcium entry blocker, flunarizine, used
in the treatment of cerebral and vascular insufficiency.51,52 Both cases involve
the blocking of unwanted metabolic oxidation of the aromatic ring by fluorine
(Figure 6.12). This altered the metabolic route of the compounds and increased
their stability and half-lives.
In another example 3-piperidinylindole, which was synthesised as an anti-
psychotic agent, had poor bioavailability due to high metabolic instability and
poor permeability.53 Although g-fluorination of the piperidine group decreased
the pKa of 3-piperidinylindole and increased its permeability, it suffered from
high metabolic instability due to hydroxylation of the 6-position on the indole
ring. Structural modification via incorporation of fluorine at this position led to
a greater improvement in bioavailability and increased binding affinity by an
order of magnitude (Table 6.16).
Discovery of celecoxib presents an interesting case where substituents were
introduced on the aromatic ring of the molecule to incorporate metabolic soft
spots and decrease the half-life of the compound.54 The fluoro- and chlor-
ophenyl- analogs (6.21 and 6.22) were discovered as an early lead in the
synthesis of cyclooxygenase-2 (COX-2) inhibitors. Although both compounds
306 Chapter 6

OCH3 OH
OH

F
N N
O O
SCH 48461 Ezetimibe
OCH3 F

F
N N
N N

Cinnarizine Flunarizine
F

Figure 6.12 Examples of fluorine substitution in rational drug design results in

alteration of metabolic routes.,

Table 6.16 Modification of 3-piperidinylindole, an antipsychotic, to improve

bioavailability.53

NH
Y

X N
H
X Y F (%) cLogP pKa

H H Poor 10.4
H F 18 8.5
F F 80

showed good efficacy in models of inflammation, they had an unacceptably

long plasma half-life in rats (221 h and 117 h, respectively). Replacement of
chloro group in 6.22 with a methyl group (celecoxib) resulted in a minor change
in its physicochemical properties but decreased its half-life to 3.5 h in rats. The
rapid elimination of celecoxib was attributed to oxidation of the tolyl group to
the corresponding carboxylic acid (Figure 6.13).
Appropriate incorporation of substituents on aromatic rings of a molecule
can also alter the rate of the primary metabolic routes and therefore influence
their duration of action. For example, the electron donating capability of the
two methyl groups at the 2- and 6-positions in lidocaine is speculated to reduce
Influence of Aromatic Rings on ADME Properties of Drugs 307

F
Cl

CF3
N N CF3
N N

H3CO2S
H2NO2S
6.21 6.22
t1/2 = 221 hr t1/2 = 117 hr

H3C H2 O
C C
HO HO
CF3
N N CF3 CF3
N N N N

H2NO2S
Celecoxib H2NO2S H2NO2S
t1/2 = 3.5 hr

Figure 6.13 Introduction of metabolically labile substituents to decrease half-life of

cylcooxygenase-2 (COX-2) inhibitors and discovery of celecoxib.

O
N
O

Lidocaine

O O CI O
N N
O O
H2N H2N
Propoxycaine Chloroprocaine

Figure 6.14 Structures of lidocaine, propoxycaine and chloroprocaine.

its rate of hydrolysis by esterases by chemically reducing the electrophilicity of

the carbonyl carbon to hydrolytic attack relative to its desmethyl analog, and
therefore the enhance its anaesthetic potency (Figure 6.14).55 A similar ratio-
nalisation has been put forth to explain the different duration of actions of
procaine analogs. Prolonged action of propoxycaine (Figure 6.14) is due to
reduced rate of hydrolysis of the ester linkage as a result of inductive effect of
the o-propoxy group as well as the steric hindrance due to its location on the
ring. In contrast, the short of duration of action of chloroprocaine is attributed
308 Chapter 6
to the electron withdrawing effect of the o-chloro group, which increases the
electrophilicity of the carbonyl group and makes the ester group more sus-
ceptible to hydrolysis by plasma and tissue esterases.55

6.4 Toxicity
Aromatic rings can play a major role in influencing the toxicity of drugs. As
discussed above, hydrogen atoms in the aromatic rings of compounds are
generally replaced with lipophilic functional groups to improve their ADME
properties and potency. These changes in the physicochemical properties of
compounds also enhance their interactions with targets and proteins, and result
in side effects. For example, amiodarone (Figure 6.15) is a lipophilic compound
belonging to this category. This antiarrhythmic agent is quite efficacious but
causes a number of side effects including pulmonary toxicity and phospholi-
pidosis.56,57 The toxicity is attributed to the presence of a diiodophenyl moiety
in the molecule which increases the lipophilicity of the compound (clogP
¼ 7.81).58 This enhances its permeation and subsequent accumulation into
several tissues and results in a very long half-life in humans (B9 to 77 days).59
In another example, the alkyl and cyclohexyl groups present in proxicromil, an
anti-allergy agent, allow this compound absorb readily through the gastro-
intestinal tract (Figure 6.15).60 However, this highly lipophilic, detergent-like
drug tends to accumulate in the biliary canaliculus, which may be a reason for
its hepatotoxicity.61

O OH O
I
O
N OH
O O
I O

Amiodarone Proxicromil
clogP = 7.81 clogP = 4.89
F

HO HO
N N O

Cl Cl
CF3
F F
Penfluridol Haloperidol
clogP = 6.95 clogP = 3.76

Figure 6.15 Structures of amiodarone, proxicromil, penfluridol and haloperidol.

Influence of Aromatic Rings on ADME Properties of Drugs 309
Other examples exist where the introduction of lipophilic groups has resulted
in increased volume of distribution or accumulation into tissues, which ulti-
mately result in long half-lives. A classic example is haloperidol versus pen-
fluridol (Figure 6.15). Introduction of various blocking groups to improve the
metabolism of haloperidol lead to the discovery of penfluridol. The compound
is more potent than haloperidol and penetrates the blood–brain barrier very
readily, probably due to its high lipophilicity (clogP ¼ 6.95). Although toxicities
have not been observed with penfluridol, its potential for accumulating in the
tissues and the resulting long half-life (199 h) raises concern, especially in case
of an overdose of the drug.62,63
Metabolic activation or bioactivation is another way in which the aromatic
rings can influence toxicity of a compound. Most xenobiotic metabolising
enzymes are not only responsible for detoxication of drugs but can induce the
formation of reactive intermediates. The electrophilic intermediates thus
formed are capable of covalently reacting with nucleophiles found in proteins
and nucleic acids, which in turn may be responsible for the drug’s toxic man-
ifestations including mutagenesis, carcinogenesis and hepatic necrosis.64,65
As described above, most lipophilic aromatic compounds undergo P450
mediated metabolic oxidation via epoxide formation. The reactivity and the
importance of epoxides was first realised with polyaromatic hydrocarbons.
These compounds underwent oxidative metabolism to reactive epoxide that
covalently bound to DNA.66,67 Furthermore, there was a good correlation
between the DNA adduct formation and carcinogenesis.68 Drugs possessing
structural features prone to metabolic epoxidations are abundant.69 Aro-
matic anticonvulsants such as phenytoin, phensuximide and phenobarbital
are known to exert side effects such as hepatic necrosis and aplastic anaemia
via the chemically reactive epoxide metabolites formed by P450 oxidation
(Figure 6.16).
It is important to realise, however, that not all epoxides exert toxic effects.
The reactivity of the epoxide intermediate varies greatly with its molecular
geometry, stability, electrophilic reactivity and relative activity as substrates of
epoxide-transforming enzymes (epoxide hydrolase, glutathione transferase and
others). In most cases, epoxides are quite unstable and rearrange rapidly to the
arenols (see above). Only the ones that are stable enough and capable of tra-
versing membranes are exposed to macromolecules and form adducts.
Aromatic compounds containing phenolic moieties and aromatic amines are
susceptible to bioactivation by several oxidative enzymes such as the perox-
idases and cytochrome (CYP) enzymes. Phenolic compounds generally
undergo a two electron oxidation to quinoid intermediates namely, o and p
quinone imines or o and p quinone methides (Figure 6.17).70 Quinone imine
and quinone methides are reactive electrophiles and may adduct to cysteine and
lysine residues of proteins, or they may act as haptens and initiate immuno-
logical responses.71 These intermediates are congeners of a,b-unsaturated
carbonyl compounds and are capable of undergoing 1,4-addition of a nucleo-
phile (Michael addition) to form conjugates as shown in Figure 6.17. In vivo
these intermediates react with macromolecules and form covalent adducts,
310 Chapter 6

O O O
HN HN HN
NH NH NH
O O O

OH
Phenytoin O Hydroxyphenytoin
O
O O

H3C N
H3C N H3C N
O
O O
O OH
Phensuximide Hydroxyphensuximide
H H
H O N O O N O
O N O
HN HN
HN
O O
O

O OH
Phenobarbital Hydroxyphenobarbital

Covalent Binding

Hepatic Necrosis

Figure 6.16 Bioactivation of compounds containing aromatic rings to epoxides.

which may trigger mechanisms that finally lead to cell necrosis, immune
responses or blood toxicities.
Metabolic activation of acetaminophen represents a classic example in this
category.72–74 Although safe at therapeutic doses, this widely used analgesic
causes fulminant hepatic necrosis when taken in overdose. The cause of
hepatotoxicity is ascribed to P450 catalysed 2-electron oxidation of acet-
aminophen to a quinone imine intermediate (NAPQI) and subsequent covalent
binding of this electrophilic intermediate with various proteins in the liver
(Figure 6.18). Some additional examples include bioactivation of amodiaquine
(Figure 6.18) and troglitazone (Figure 6.19).75–78 The hepatotoxicity and
agranulocytosis observed following administration of antimalarial amodia-
quine is triggered by formation of reactive quinone imine intermediate (similar
to that observed for acetaminophen) which is produced in the liver as well as
other sites in the body (Figure 6.18).75
Troglitazone was developed for the treatment of type 2 diabetes but was
withdrawn from the clinic due to rare cases of liver failure and deaths. The
Influence of Aromatic Rings on ADME Properties of Drugs 311

OH O OH
-
2 e Oxidation
R1 R1 Nu R1
Nu
R R R
Quinone Methide

OH O OH
-
R1 2 e Oxidation Nu
R1 R1
Nu
NH N NH
R R R
Quinone Imine

Nu = Nucleophile (proteins, Nucleic acids)

Figure 6.17 Two electron oxidation of phenolic compounds to quinone methides and
quinone imines and their reaction with nucleophiles.

O O
HN HN

UGT OR
O ST CO2H
O
O OH OSO3H
HN OH
HO

O O
OH N HN
Acetaminophen
CYP450 Protein-SH
[O] Protein
S
O OH
Quinone imine

OH
O
N
HN N
N

Cl N
Cl N
Amodiaquine Quinone imine Formation

Figure 6.18 Metabolic activation of acetaminophen and amodiaquine to a quinone

imine.
312 Chapter 6

O O

O HO
SG =
GS glutathione
Quinone Methide

S OH
O O O O
O NH
O O
O O O
HO OH

Troglitazone
O OH O
S S S
O HO O
NH N C O
O O O N SG
H

SG
S

HN2
O

Figure 6.19 Metabolic activation of troglitazone.

observed hepatotoxicity has been associated with the formation of reactive

intermediates. Three different mechanisms for troglitazone bioactivation that
results in covalent binding to proteins have been proposed. While two
mechanisms involve metabolic activation of the thiazolidinedione moiety, one
of the proposed mechanisms involves the oxidation of the chromane moiety of
troglitazone to a reactive o-quinone methide (Figure 6.19).79 The formation of
the quinone methide has been confirmed by trapping the intermediate with
glutathione or N-acetylcysteine, and the structure of the adduct has been
determined unequivocally by NMR.
Phenolic compounds can also form reactive intermediates via catechols.
Most often phenolic compounds are cleared via glucuronidation or sulfation of
the hydroxyl group (as shown for acetaminophen, Figure 6.18). However, given
the electronic distribution of the hydroxyphenyl ring, these compounds can be
oxidised to catechol or hydroquinone derivatives. These hydroxylated deriva-
tives of phenols are readily oxidised to chemically reactive quinones and
semiquinones, which form covalent bonds with proteins and DNA.80,81
Alternatively, these can generate toxic reactive oxygen species that could
damage the cellular mechanisms. The natural estrogens such as estrone and
estradiol and drugs such as 17a-ethynylestradiol undergo extensive CYP-
mediated oxidation in the 2- or 4-positions.82 One pathway implicated in the
tumorgenic process involves the metabolism of estrogens to catechols mediated
by P450s and further oxidation of these catechols to estrogen o-quinones
(Figure 6.20). The estrogen quinones react with DNA and form N3-adenine or
N7-guanine DNA adducts, resulting in mutagenic apurinic sites.83 A similar
bioactivation pathway is also demonstrated for equilin and equilenin, which are
used for estrogen replacement therapy.84
Influence of Aromatic Rings on ADME Properties of Drugs 313
O O

HO O
OH O
O

DNA Covalent
Catechols o-Quinones Binding
HO
O
Estrone O

HO
O

HO
O

OH OH O O

HO HO
HO HO
Estradiol Equilin Equilenin
17α-Ethynylestradiol

Figure 6.20 Metabolic activation of estrogens via the catechols.

Hydroxylated metabolites of substituted or unsubstituted aromatic com-

pounds are also susceptible to formation of quinone methide or quinone imine
intermediates that are capable of binding to proteins in vivo and induce hepta-
totoxicity. For instance, diclofenac is a non-steroidal anti-inflammatory agent
(NSAID) that undergoes metabolism via glucuronidation of the carboxylic
acid.85 However, a minor pathway of diclofenac metabolism involves CYP2C9
or 3A4 mediated hydroxylation of the two aromatic rings (Figure 6.21). The
hydroxy group is perfectly positioned to undergo a two electron oxidation to the
corresponding quinone imine metabolite (Figure 6.21). The reactivity of these
metabolites is implicated in the hepatotoxicity of this compound.86,87 Similarly,
the toxicity of high incidence of tacrine hepatotoxicity has been correlated with
the formation of 7-hydroxytacrine.88 The position of the hydroxy group on the
aromatic ring in the hydroxytacrine metabolite is suggestive of the formation of
a quinone imine moiety as depicted in Figure 6.21.89
Likewise, a causative role of the quinone imine intermediate in nefazo-
done hepatotoxicity has also been speculated.90 Following incubation of
nefazodone with microsomes or recombinant P4503A4 in the presence of sulfy-
dryl nucleophiles, conjugates derived from the addition of thiol to a mono-
hydroxylated nefazodone metabolite have been identified (Figure 6.22).
Tamoxifen is an effective drug in the prevention and treatment of breast cancer,
but long-term usage of this anti-estrogenic drug has been linked to an increased
risk of uterine cancer. A potential pathway leading to toxicity of tamoxifen
314 Chapter 6

CO2H CO2H
Cl Cl
H
Acyl Glucuronide N N
Covalent
HO Binding
CO2H Cl O Cl
Cl
H
N 2C9
Hydroxydiclofenac
Cl
CO2H
Diclofenac 3A4 Cl CO2H
H Cl
N
N Covalent
Binding
Cl OH
Cl O

NH2 NH2 NH2

HO O
Covalent
N N Binding
N
Tacrine hydroxytacrine

Figure 6.21 P450 catalyzed bioactivation of diclofenac and tacrine to quinone imines
via their hydroxylated metabolites.

involves bioactivation of 4-hydroxytamoxifen via several pathways as shown in

Figure 6.22.91 These involve formation of the o-quinone via the catechol or
extended quinone methides (Figure 6.22). All these electrophilic species react
with glutathione (GSH) and DNA, and may contribute to the genotoxicity of
tamoxifen. However, the one that is actually responsible for genotoxicity has
not been explicitly identified.
Like epoxides and even though phenolic derivatives are prone to oxidation
and generation of reactive metabolites, not all compounds that contain a
phenolic moiety are toxic. Hence factors that affect the amount of reactive
metabolite formed in vivo can influence the toxicity elicited by these compounds
via metabolic activation. Some factors that can play a major role in controlling
the amount of reactive metabolite formed include the dose of the compound
and the efficiency of the pathways that detoxify the reactive metabolites or the
parent compound.
The quinoid intermediates formed by metabolic activation of phenolic
compounds or the phenolic metabolites can also lead to inactivation of CYP
enzymes that produce them. For example, nefazodone is an inactivator of
CYP3A4.90 The mechanism of inactivation possibly involves alkylation of the
nucleophiles in the active site of the molecule. Raloxifene, a very popular
selective estrogen receptor modulator (SERM) used in treatment of osteo-
porosis and breast cancer, also inactivates CYP3A4 via bioactivation
(Ki ¼ 9.9 mM; kinact ¼ 0.19 min1).92 The inactivation of the enzyme is ascribed
to the formation of extended quinoid species that is formed by CYP3A4
(Figure 6.23).92,93 The reactive species has been shown to alkylate the thiol
moiety in the active site of 3A4.94
Influence of Aromatic Rings on ADME Properties of Drugs 315
O
N
N N N N
N N O

Nefazodone HO
Cl
Cl

N N
GS O GS
N N
OH

HO O O
HO
Cl Cl Cl
Cl

Possibly leads
to inactivation of
CYP3A4

Ar Ar Ar Ar

N
O
Ar Ar P450
HO OH O O
P450

Ar Ar Ar Ar
OH

HO
Tamoxifen P450

OH O
Direct 2e- oxidation

Figure 6.22 Metabolic activation of nefazodone and tamoxifen.

N O N
O O O

O O Inactivation of
CYP3A4

OH O
S O S
HO

Raloxifene Extended quinoid

Figure 6.23 Metabolic activation of raloxifene to an extended quinoid intermediate.

Fluorine substitution has been commonly used as a tool to block metabolic

activation of compounds and in the development of safer drugs (Figure 6.24).
Liehr found that 2-fluoro-estradiol was as estrogenic as estradiol but non-
carcinogenic in the Syrian hamster.95 In vivo studies indicate that the presence
316 Chapter 6

O
OH
HN
F F
F

HO
OH
2-Fluoroestradiol 2, 6- Difluoroacetaminophen

F
SG
OH
OH
N
HN N
HN

SG = glutathione
Cl N
Cl N
6-Fluoroamodiaquine

N
HN

Cl N
Fluoroamodiaquine

Figure 6.24 Introduction of fluorine into the molecule prevents bioactivation.

of a 2-fluoro substituent diverts the metabolism of the steroid from 2-hydro-

xylation to glucuronidation and therefore prevents its bioactivation to the
quinone intermediates. Introduction of fluorine into acetaminophen also affects
its NAPQI formation by altering the oxidation potential.18 This influence is
dependent on the number of hydrogen atoms that are replaced in the mole-
cule and their positions. The presence of fluorine at the 2- and 6-positions
(Figure 6.24) increased the oxidation potential of acetaminophen from 1.14 to
1.52, and reduced the propensity of the molecule to undergo oxidative bioac-
tivation thereby reducing hepatotoxicity.18 However, the glucuronidation
and sulfation pathways are not affected by this change and therefore do not
affect the elimination of the compound from the body. As is the case for
acetaminophen, 6-fluorination of the aminophenol ring of amodiaquine
(Figure 6.24) produced an analogue with significantly raised oxidation poten-
tial and reduced bioactivation in vivo.18 However, some bioactivation was
observed due to defluorination followed by formation of the quinone imine, as
observed in amodiaquine. Similarly, isosteric replacement of the hydroxyl
function with fluorine also provided an analogue, fluoroamodiaquine
Influence of Aromatic Rings on ADME Properties of Drugs 317

O O O O
N N
N N
H H
OH
6.23

O O
N
3A4 Inactivation N
H
O
Quinone Methide

O O
N
N
H
6.24 F

Figure 6.25 Inactivation of CYP3A4 by KCNQ2 potassium channel opener 6.23.

(Figure 6.24), which had similar antimalarial potency to the parent drug but
was not bioactivated to toxic metabolites in vivo.18
The following example demonstrates modification of the lead compound to
prevent mechanism based inhibition by introducing fluorine into the molecule.
The lead compound (6.23) was designed to be a KCNQ2 potassium channel
opener and had excellent oral bioavailability and pharmacological activity
(Figure 6.25).96 However, 6.23 was an inactivator of CYP3A4. P450 inactiva-
tion was consistent with a bioactivation pathway involving the initial aromatic
hydroxylation ortho to the morpholine ring (or para to the benzylamine
methine). Further two electron oxidation of this initial metabolite by P450 can
result in the formation of the reactive quinone methide intermediates. The
potential formation of either reactive intermediate was avoided by the intro-
duction of the fluorine atom on the position ortho to the morpholine ring.
Compound 6.24 is not only devoid of P450 inactivation liability, but also
retains the pharmacological and pharmacokinetic properties of the prototype
compound 6.23.
Although fluorine atoms have been commonly used in blocking metabolism
and bioactivation (as described earlier), it is important to note that fluorination
of the aromatic ring does not necessarily prevent the compound from under-
going bioactivation or from inactivating the enzyme.18 More recently, dasati-
nib, a tyrosine kinase inhibitor, has been shown to inactivate CYP3A4 and this
inactivation proceeds through a reactive intermediate.97 The major mechanism
of inactivation proceeds through hydroxylation at the 4-position of the 2-
chloro-6-methylphenyl ring in the molecule followed by further oxidation
forming a reactive quinone-imine (Figure 6.26). Interestingly, blocking the para
318 Chapter 6

N N
N N
Cl N N [O] Cl N N
H N H N
N N N N
S H OH S H OH
O HO O
Dasatinib
[O]
[O]
Cl N
Cl N
N N
N N S H
S H
O O
O
Quinoneimine
Iminomethide
Inactivation of CYP3A4

Figure 6.26 Bioactivation of Dasatinib.

Table 6.17 Effect of fluorine on inactivation of CYP3A4 by dasatinib and its

analogs.97

N
N N N
N OH
H NH
N S
Ar
O

Compound Ar KI (lM) Kinact (min1)

Dasatinib Cl 6.3 0.034

6.25 Cl 5.4 0.039

6.26 F No inactivation

Cl
Cl

position with fluorine did not prevent inactivation of CYP3A4 (Table 6.17).
The rates of inactivation were virtually identical and the KI for the fluorinated
analog was lower, implying that it is more efficiently bioactivated possibly by
virtue of increase in the clogP value due to the incorporation of fluorine.
Influence of Aromatic Rings on ADME Properties of Drugs 319

O O O O O O
N N
N N HO N
H N
H H
O OH OH
Br Br Br NCQ344
Remoxipride

O O
O N
N
Covalent Adducts H
with proteins Peroxidase O
Br

Figure 6.27 Metabolic activation of remoxipride via O-dealkylation.

However, introduction of three halogens on the molecule prevented the

bioactivation and therefore the inactivation of the compound.97
Other functional groups present on the aromatic rings that generate phenolic
metabolites also play a role in mediating toxicity via metabolism. For instance,
an O-dealkylated metabolite resulting in a phenolic species has been implicated
in aplastic anaemia observed for the neuroleptic remoxipride.98,99 Elegant
mechanistic studies conducted by Erve and co-workers have demonstrated that
NCQ344, a hydroquinone metabolite formed via O-dealkylation and sub-
sequent oxidation, is capable forming a reactive para-quinone (Figure 6.27) by
stimulated human neutrophils or myeloperoxidase.100
A similar metabolic activation pathway has been observed in metabolism
studies of an investigational maxi-K channel opener, MaxiPost, being devel-
oped to treat ischemic stroke.101 The metabolite M1 of MaxiPost is bioacti-
vated via M1 to form a quinone methide, which has been shown to covalently
bind to proteins (Figure 6.28). Although major toxic events have not been
associated with this compound, covalent binding of the compound in vivo in
humans poses a major risk of possible idiosyncratic reactions.
Compounds containing aromatic amines (anilines) induce a variety of tox-
icological responses including carcinogenicity and hepatotoxicity. Several drugs
containing an aniline moiety, which have been withdrawn from the market,
have a black box warning on their labels. Therefore, anilines have been put on
the black list of functional groups that the medicinal chemists generally avoid.
Anilines can also undergo a two electron oxidation to dimines or quinone
imines depending on the functionality at the 4-position of the amino group
(some of the drugs discussed above are derivatives of aniline).
However, the primary pathway of bioactivation of an aromatic amine is
hydroxylation of the amino group.102,103 N-oxidation leads to generation
of the reactive species via conjugation (sulfation or acetylation) of the corre-
sponding hydroxylamine. The electrophilic intermediate thus formed by elim-
ination of the sulfate moiety or the acetoxy moiety can covalently bind to
cellular macromolecules (Figure 6.29). Alternatively, the hydroxylamine is
320 Chapter 6

H H
F3C N F3C N
O O
F F
O CH3 OH

M1
Cl Cl
MaxiPost
H
H F3C N
F3C N O
O
NHLys-Proteins O
OH

Cl
Cl
Covalent binding ortho-Quinone Methide
to Lysine residue

Figure 6.28 P450-catalyzed metabolic activation of MaxiPost.

OH OX
NH 2 NH NH NH 2
R R Nu
R R Nu
Hydroxylamine
X = Acetyl
X = SO3H

O
N Reaction with
Nucleophiles on
R the Macromolecule

Nitroso
Intermediate

Figure 6.29 Activation of aromatic amines.

oxidised to a nitroso derivative, which reacts with the nucleophile on the

macromolecule.104
Some drugs that contain an aniline group and have been withdrawn from the
clinic include nomifensine (Figure 6.30) and carbutamide (Figure 6.3).
Nomefensine was withdrawn from the market due to acute immune haemolytic
anaemia. The aniline moiety in this drug has not been implicated in its toxicity,
but its role in inducing toxic events via reactive metabolites cannot be ruled
out.105 Carbutamide, an oral antidiabetic, was withdrawn from the market due
to life-threatening bone marrow toxicity in man.106 The toxicity was attributed
to the activation of the aniline moiety since tolbutamide, in which the amino
group was replaced with a methyl group (Figure 6.3), was devoid of toxicity.
Influence of Aromatic Rings on ADME Properties of Drugs 321

NH2
O O CF3
N
N H
N N
H F3C O N
H2N H
O
Procainamide Flecainide

Nomifensine

X
X = SO2, Dapsone
H2N NH2 X = S, O, CH2, CO

Figure 6.30 Structures of aniline containing compounds and its analogs.

Procainamide (Figure 6.30) is another drug where removal of the aniline moiety
translates into a markedly improved safety profile. Bone marrow aplasis and
lupus syndrome observed upon administration of procainamide is not observed
upon administration of flecainide, which is devoid of aniline moiety.107
Dapsone (Figure 6.30), an antibacterial, is also shown to induce agranulo-
cytosis, aplastic anaemia and cutaneous adverse drug reactions; all these
toxicities are associated with N-oxidation of aniline nitrogen.108–110 The
influence of electronics on bioactivation potential is evident in structure–toxi-
city assessment on dapsone analogues. Replacement of the sulfone group in
dapsone with sulfur, oxygen, methylene or carbonyl substituents significantly
reduce methemoglobinaemia in human erythrocytes.111 Although the reason
for the influence of these substituents is unclear, a good correlation between
haemotoxicity and the Hammett constant (sp) has been demonstrated in this
analysis indicating that the substituent in the 4-position can potentially affect
the rate of oxidation of the aniline nitrogen.
Compounds containing nitroaromatic rings, which can be metabolically
reduced to anilines, can also elicit toxicity. Tolcapone, a catechol-O-methyl-
transferase inhibitor (Figure 6.31), has been associated with a number of
problems including abnormalities in liver function tests and three cases of fatal
hepatotoxicity.112 These problems have led to withdrawal of the drug from the
market in some countries and the introduction of a black box warning and
intensive monitoring requirements in the United States. A significant portion of
tolcapone biotransformation proceeds via reduction of the nitrobenzene group
to the aniline derivative, which is transformed to the corresponding anilide by
N-acetyltransferase. Both the aniline and the anilide metabolites of tolcapone
undergo facile two electron oxidation to the corresponding quinone-imine
metabolites that are trapped with GSH (Figure 6.31).113 Interestingly, the
structural analog of tolcapone, entacapone, does not undergo similar bioacti-
vation in spite of the presence of the same structural motif. This is attributed to
minimal (or lack of) reduction of the nitro group to the aromatic amine in
322 Chapter 6

O SG O
HO HO

HO CH 3 HO CH 3
NHCOCH3 NHCOCH 3
NAT
O O O
HO HO HO
Reduction

HO CH 3 HO CH3 O CH3
NO 2 NH2 NH Quinone imine
Tolcapone
O O
SG O
HO O
N HO
CN O CH3
HO HO CH3
NO2 NH 2
NH2
Entacapone ortho-Quinone
SG = Glutathione

Figure 6.31 Metabolic activation of tolcapone via reduction of the nitro group.

humans. Instead, its principal clearance pathway involves N-de-ethylation of

the tertiary amide substituent and isomerisation of the active E-isomer to the
inactive Z-isomer, followed by glucuronidation of the catechol moiety.114

References
1. R. K. Castellano, F. Diederich and E. A. Meyer, Angew. Chem., Int. Ed.,
2003, 42, 1210.
2. R. B. Silverman, in The Organic Chemistry of Drug Design and Drug
Action, Academic Press, San Diego, 2004, pp. 7–99.
3. L. P. Hammett, J. Am. Chem. Soc., 1937, 59, 96.
4. L. P. Hammett, Physical Organic Chemistry, McGraw-Hill, New York,
1940.
5. C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 165.
6. T. Fujita, J. Iwasa and C. Hansch, J. Am. Chem. Soc., 1964, 86, 5175.
7. C. Hansch and T. Fujita, J. Am. Chem. Soc., 1964, 86, 1616.
8. R. W. Taft, in Steric Effects in Organic Chemistry, ed. M. S. Neuman,
Wiley, New York, 1956, pp. 556–675.
9. S. H. Unger and C. Hansch, Prog. Phys. Org. Chem., 1976, 12, 91.
10. C. Hansch, A. Leo, S. H. Unger, K.-H. Kim, D. Xikaitani and E. J. Lien,
J. Med. Chem., 1973, 16, 1207.
11. W. Guba, W. Neidhart and M. Nettekoven, Bioorg. Med. Chem. Lett.,
2005, 15, 1599.
Influence of Aromatic Rings on ADME Properties of Drugs 323
12. S. Purser, P. R. Moore, S. Swallow and V. Gouverneur, Chem. Soc. Rev.,
2008, 37, 320.
13. H.-J. Böhm, D. Banner, S. Bendels, M. Kansy, B. Kuhn, K. Müller,
U. Obst-Sander and M. Stahl, ChemBioChem, 2004, 5, 637.
14. W. K. Hagmann, J. Med. Chem., 2008, 51, 4359.
15. K. Müller, C. Faeh and F. Diederich, Science, 2007, 317, 1881.
16. A. Bondi, J. Phys. Chem., 1964, 68, 441.
17. B. K. Park, N. R Kitteringham and P. M. O’Neill, Annu. Rev. Pharmacol.
Toxicol., 2001, 41, 443.
18. B. K. Park and N. R. Kitteringham, Drug Metab. Rev., 1994, 26, 605.
19. G. L. Amidon, H. Lennernas, V. P. Shah and J. R. Crison, Pharm. Res.,
1995, 12, 413.
20. O. H. Chan and B. H. Stewart, Drug Discov. Today, 1996, 1, 461.
21. D. A. Smith, B. C. Jones and D. K. Walker, Med. Res. Rev., 1996, 16, 243.
22. T. Fichert, M. Yazdanianb and J. R. Proudfoota, Bioorg. Med. Chem.
Lett., 2003, 13, 719.
23. R. Griffith, in Foye’s Principles of Medicinal Chemistry, ed. D.A. Williams
and T. L. Lemke, Lippincott Williams and Wilkins, Baltimore, 2002,
pp. 292–312.
24. J. M. Custodio, C.-Y. Wu and L. Z. Benet, Adv. Drug Deliv. Rev., 2008,
60, 717.
25. F.-A. Wiesel, G. Alfredsson, M. Ehrnebo and G. Sedvall, Eur. J. Clin.
Pharmacol., 1980, 17, 385.
26. G. Movin-Osswald and M. Hammarlund-Udenaes, Br. J. Clin. Pharma-
col., 1991, 32, 355.
27. J. H. Lin and M. Yamazaki, Drug Metab. Rev., 2003, 35, 417.
28. F. Atkinson, S. Cole, C. Green and H. van de Waterbeemd, Curr. Med.
Chem. Cent. Nerv. Syst. Agents, 2002, 2, 229.
29. R. M. Arendt, D. J. Greenblatt, D. C. Liebisch, M. D. Luu and S. M.
Paul, Psychopharmacology, 1987, 93, 72.
30. J. Kelder, P. D. J. Grootenhuis, D. M. Bayada, L. P. C. Delbressine and
J.-P. Ploemen, Pharm. Res., 1999, 16, 1514.
31. S. A. Hitchcock and L. D. Pennington, J. Med. Chem., 2006, 49, 7559.
32. R. V. Cube, J. -M. Vernier, J. H. Hutchinson, M. F. Gardner, J. K. James,
B. A. Rowe, H. Schaffhauser, L. Daggett and A. B. Pinkerton, Bioorg.
Med. Chem. Lett., 2005, 15, 2389.
33. L. R. Roberts, J. Bryans, K. Conlon, G. McMurray, A. Stobie and G. A.
Whitlock, Bioorg. Med. Chem. Lett., 2008, 18, 6437.
34. E. Kutter and V. Austel, Arzneim.-Forsch., 1981, 31, 135.
35. S. M. Rubenstein, V. Baichwal, H. Beckmann, D. L. Clark, W. Frank-
moelle, D. Roche, E. Santha, S. Schwender, M. Thoolen, Q. Ye and J. C.
Jaen, J. Med. Chem., 2001, 44, 3599.
36. D. A. Smith, H. van de Waterbeemd and D. K. Walker, Pharmacokinetics
and Metabolism in Drug Design, Wiley-VCH, Weinheim, 2006, p. 83.
37. D. A. Smith, H.S. Rasmussent, D. A. Stopher and D. K. Walker,
Xenobiotica, 1992, 22, 709.
324 Chapter 6
38. H. Kusuhara, H. Suzuki and Y. Sugiyama, J. Pharm. Sci., 1998, 87, 1025.
39. C. F. Sturino, G. O’Neill, N. Lachance, M. Boyd, C. Berthelette,
M. Labelle, L. Li, B. Roy, J. Scheigetz, N. Tsou, Y. Aubin, K. P. Bate-
man, N. Chauret, S. H. Day, J. L. Le 0 vesque, C. Seto, J. H. Silva, L. A.
Trimble, M. C. Carriere, D. Denis, G. Greig, S. Kargman, S. Lamon-
tagne, M. C. Mathieu, N. Sawyer, D. Slipetz, W. M. Abraham, T. Jones,
M. McAuliffe, H. Piechuta, D. A. Nicoll-Griffith, Z. Wang, R. Zamboni,
R. N. Young and K. M. Metters, J. Med. Chem., 2007, 50, 794.
40. B. Meunier, S. P. de Visser and S. Shaik, Chem. Rev., 2004, 104, 3947.
41. C. M. Bathelt, L. Ridder, A. J. Mulholland and J. N. Harvey, Org.
Biomol. Chem., 2004, 2, 2998.
42. L. K. Low and N. Castagnoli Jr., in Burgers Medicinal Chemistry: Basis
of Medicinal Chemistry, ed. M. E. Wolff, Part 1, J. Wiley and Sons,
Hoboken, NJ, USA, 1979, pp. 107–226.
43. D. V. Parke, Biochem. J., 1960, 77, 493.
44. W. Jacobsen, B. Kuhn, A. Soldner, G. Kirchner, K.-F. Sewing, P. A.
Kollman, L. Z. Benet and U. Christians, Drug Metab. Dispos., 2000, 28,
1369.
45. P. G. Dayton, J. M. Perel, R. F. Cummingham, Z. H. Israeli and I. M.
Weiner, Drug Metab. Dispos., 1973, 1, 742.
46. P. G. Dayton and J. M. Perel, Ann. N. Y. Acad. Sci., 1971, 179, 399.
47. M. W. Dysken, J. l. Javai, S. S. Chang, C. Schaffer, A. Shahid and J. M.
Davis, Psychopharmacology, 1981, 17, 205.
48. T. L. Perry, C. F. A. Culling, K. Berry and S. Hansen, Science, 1964, 146,
81.
49. L. Quartara and M. Altamura, Curr. Drug Targets, 2006, 7, 975.
50. R. K. Verbeck, J. L. Blackburn and G. R. Loewen, Clin. Pharmacokin.,
1983, 8, 297.
51. S. B. Rosenblum, T. Huynh, A. Afonso, H. R. Davis Jr, N. Yumibe, J. W.
Clader and D. A. Burnett, J. Med. Chem., 1998, 41, 973.
52. S. Kariya, S. Isozaki, K. Uchino, T. Suzuki and S. Narimatsu, Biol.
Pharm. Bull., 1996, 19, 1511.
53. M. Rowley, D. J. Hallett, S. Goodacre, C. Moyes, J. Crawforth, T. J.
Sparey, S. Patel, R. Marwood, S. Patel, S. Thomas, L. Hitzel,
D. O’Connor, N. Szeto, J. L. Castro, P. H. Hutson and A. M. Macleod,
J. Med. Chem., 2001, 44, 1603.
54. T. D. Penning, J. J. Talley, S. R. Bertenshaw, J. S. Carter, P. W. Collins, S.
Docter, M. J. Graneto, L. F. Lee, J. W. Malecha, J. M. Miyashiro,
R. S. Rogers, D. J. Rogier, S. S. Yu, G. D. Anderson, E. G. Burton, J. N.
Cogburn, S. A. Gregory, K. M. Koboldt, W. E. Perkins, K. Seibert, A. W.
Veenhuizen, Y. Y. Zhang and P. C. Isakson, J. Med. Chem., 1997, 40, 1347.
55. M. C. Lu, in Foye’s Principles of Medicinal Chemistry, ed. D.A. Williams
and T. L. Lemke, Lippincott Williams and Wilkins, Baltimore, 2002,
pp. 338–353.
56. D. K. Ernawati, L. Stafford and J. D. Hughes, Br. J. Clin. Pharmacol.,
2008, 66, 82.
Influence of Aromatic Rings on ADME Properties of Drugs 325
57. K. N. S. Sirajudeen, P. Gurumoorthy, H. Devaraj and S. N. Devaraj,
Drug Chem. Toxicol., 2002, 25, 247.
58. C. Lafuente-Lafuente, J. C. Alvarez, A. Leenhardt, S. Mouly, F. Extra-
miana, C. Caulin, C. Funck-Brentano and J. F. Bergmann, Br. J. Clin.
Pharmacol., 2009, 67, 511.
59. M. D. Freedman and J. C. Somberg, J. Clin. Pharmacol., 1991, 31, 1061.
60. B. Clark, D. A. Smith, C. T. Eason and D. V. Parke, Xenobiotica, 1982,
12, 147.
61. D. A. Smith, K. Brown and M. G. Neale, Drug Metab. Rev., 1985, 16,
365.
62. J. M. Grindel, B. H. Migdalof and W. A. Cressman, Drug Metab. Dispos.,
1979, 7, 325.
63. W. A. Cressman, J. R. Bianchine, V. B. Slotnick, P. C. Johnson and
J. Plostnieks, Euro. J. Clin. Pharmacol., 1974, 7, 99.
64. A. S. Kalgutkar and J. R. Soglia, Expert Opin. Drug Metab. Toxicol.,
2005, 1, 91.
65. J. C. Erve, Expert Opin. Drug Metab. Toxicol., 2006, 2, 923.
66. M. Cosman, C. D. L. Santos, R. Fiala, B. E. Hingertyt, S. B. Singht,
V. Ibanezt, L. A. Margulist, D. Live, N. E. Geacintov, S. Broyde and D. J.
Patel, Proc. Natl. Acad. Sci. U. S. A., 1992, 89, 1914.
67. A. Dipple, Carcinogenesis, 1995, 16, 437.
68. C. Heidelberger, Ann. Rev. Biochem., 1975, 44, 79.
69. F. Oesch, Biochem. Pharmacol., 1976, 25, 1935.
70. J. A. Hinson, N. R. Pumford and S. D. Nelson, Drug Metab. Rev., 1994,
26, 395.
71. J. A. Hinson and D. W. Roberts, Annu. Rev. Pharmacol. Toxicol., 1992,
32, 471.
72. W. Chen, L. L. Koenigs, S. J. Thompson, R. M. Peter, A. E. Rettie, W. F.
Trager and S. D. Nelson, Chem. Res. Toxicol., 1998, 11, 295.
73. J. A. Hinson, A. B. Reid, S. S. Mccullough and L. P. James, Drug Metab.
Rev., 2004, 36, 805.
74. B. K. Park, N. R. Kitteringham, J. L. Maggs, M. Pirmohamed and D. P.
Williams, Annu. Rev. Pharmacol. Toxicol., 2005, 45, 177.
75. H. Jewell, J. L. Maggs, A. C. Harrison, P. M. O’Neill, J. E. Ruscoe and B.
K. Park, Xenobiotica, 1995, 25, 199.
76. J. N. Tettey, J. L. Maggs, W. G. Rapeport, M. Pirmohamed and B. K.
Park, Chem. Res. Toxicol., 2001, 14, 965.
77. M. T. Smith, Chem. Res. Toxicol., 2003, 16, 679.
78. S. Prabhu, A. Fackett, S. Lloyd, H. A. McClellan, C. M. Terrell, P. M.
Silber and A. P. Li, Chem-Bio. Interact., 2002, 142, 83.
79. K. Kassahun, P. G. Pearson, W. Tang, I. McIntosh, K. Leung, C. Elmore,
D. Dean, R. Wang, G. Doss and T. A. Baillie, Chem. Res. Toxicol., 2001,
14, 62.
80. E. L. Cavalieri, K.-M. Li, N. Balu, M. Saeed, P. Devanesan, S. Higgin-
botham, J. Zhao, M. L. Gross and E. G. Rogan, Carcinogenesis, 2002, 23,
1071.
326 Chapter 6
81. E. Cavalieri, K. Frenkel, J. G. Liehr, E. Rogan and D. Roy, J Natl.
Cancer Inst. Monogr., 2000, 27, 75.
82. C. P. Martucci and J. Fishman, Pharmacol. Ther., 1993, 57, 237.
83. J. L. Bolton and G. R. J. Thatcher, Chem. Res. Toxicol., 2008, 21, 93.
84. J. L. Bolton, Adv. Mol. Toxicol., 2006, 1, 1.
85. W. Tang, Curr. Drug Metab., 2003, 4, 319.
86. W. Tang, R. A. Stearns, S. M. Bandiera, Y. Zhang, C. Raab, M. P.
Braun, D. C. Dean, J. Pang, K. H. Leung, G. A. Doss, J. R. Strauss, G. Y.
Kwei, T. H. Rushmore, S.-H. L. Chiu and T. A. Baillie, Drug Metab.
Dispos., 1999, 27, 365.
87. G. K. Poon, Q. Chen, Y. Teffera, J. S. Ngui, P. R. Griffin, M. P. Braun,
G. A. Doss, C. Freeden, R. A. Stearns, D. C. Evans, T. A. Baillie and W.
Tang, Drug Metab. Dispos., 2001, 29, 1608.
88. S. Madden, T. F. Woolf, W. F. Pool and B. K. Park, Biochem. Phar-
macol., 1993, 46, 13.
89. T. F. Woolf, W. F. Pool, S. M. Bjorge, T. Chang, O. P. Goel, C. F.
Purchase, M. C. Schroeder, K. L. Kunze and W. F. Trager, Drug Metab.
Dispos., 1993, 21, 874.
90. A. S. Kalgutkar, A. D. N. Vaz, M. E. Lame, K. R. Henne, J. Soglia, S. X.
Zhao, Y. A. Abramov, F. Lombardo, C. Collin, Z. S. Hendsch and C. E.
C. A. Hop, Drug Metab. Dispos., 2005, 33, 243.
91. P. W. Fan and J. L. Bolton, Drug Metab. Dispos., 2001, 29, 891.
92. Q. Chen, J. S. Ngui, G. A. Doss, R. W. Wang, X. Cai, F. P. DiNinno, T.
A. Blizzard, M. L. Hammond, R. A. Stearns, D. C. Evans, T. A. Baillie
and W. Tang, Chem. Res. Toxicol., 2002, 15, 907.
93. B. R. Baer, L. C. Wienkers and D. A. Rock, Chem. Res. Toxicol., 2007,
20, 954.
94. J. T. Pearson, J. L. Wahlstrom, L. J. Dickmann, S. Kumar, J. R. Halpert,
L. C. Wienkers, R. S. Foti and D. A. Rock, Chem. Res. Toxicol., 2007, 20,
1778.
95. J. G. Liehr, Mol. Pharmacol., 1983, 23, 278.
96. Y.-J. Wu, C. D. Davis, W. C. Dworetzky, W. C. Fitzpatrick, D. Harden,
H. He, R. J. Knox, A. E. Newton, T. Philip, C. Polson, D. V. Sivarao,
L.-Q. Sun, S. Tertyshnikova, D. Weaver, S. Yeola, M. Zoeckler and
M. W. Sinz, J. Med. Chem., 2003, 46, 3778.
97. X. Li, Y. He, C. H. Ruiz, M. Koenig and M. D. Cameron, Drug Metab.
Dispos., 2009, 37, 1242.
98. S. Ahlenius, E. Ericson, V. Hillegaart, L. B. Nilsson, P. Salmi and
A. Wijkström, J. Pharmacol. Exp. Ther., 1997, 283, 1356.
99. S. T. Laidlaw, J. A. Snowden and M. J. Brown, Lancet, 1993, 342,
1245.
100. J. C. Erve, M. A. Swensson, H. Von Euler-Chelpin and E. Klasson-
Wehler, Chem. Res. Toxicol., 2004, 17, 564.
101. D. Zhang, R. Krishna, L. Wang, J. Zeng, J. Mitroka, R. Dai, N.
Narasimhan, R. A. Reeves, N. R. Srinivas and L. J. Klunk, Drug Metab.
Dispos., 2005, 33, 83.
Influence of Aromatic Rings on ADME Properties of Drugs 327
102. P. Hlavica, I. Golly, M. Lehnerer and J. Schulze, Hum. Exp. Toxicol.,
1997, 16, 441.
103. R. B. Silverman, in The Organic Chemistry of Drug Design and Drug
Action, Academic Press, San Diego, 2004, pp. 405–495.
104. J. P. Sanderson, D. J. Naisbitt, J. Farrell, C. A. Ashby, M. J. Tucker,
M. J. Rieder, M. Pirmohamed, S. E. Clarke and B. K. Park, J. Immunol.,
2007, 178, 5533.
105. P. D. Stonier, Pharmacoepidemiol. Drug Saf., 1992, 1, 177.
106. R. L. Hare, B. Holcomb, O. C. Page and J. W. Stephens, N. Engl. J. Med.,
1957, 256, 74.
107. J. P. Uetrecht, J. Pharmacol. Exp. Ther., 1985, 232, 420.
108. Z. H. Israili, S. A. Cucinell, J. Vaught, E. Davis, J. M. Lesser and P. G.
Dayton, J. Pharmacol. Exp. Ther., 1973, 187, 138.
109. S. J. Grossman and D. J. Jollow, J. Pharmacol. Exp. Ther., 1988, 244, 118.
110. J. Uetrecht, N. Zahid, N. H. Shear and W. D. Biggar, J. Pharmacol. Exp.
Ther., 1988, 245, 274.
111. R. Mahmud, M. D. Tingle, J. L. Maggs, M. T. Cronin, J. C. Dearden and
B. K. Park, Toxicology, 1997, 117, 1.
112. N. Borges, Expert Opin. Drug Saf., 2005, 4, 69.
113. K. S. Smith, P. L. Smith, T. N. Heady, J. M. Trugman, W. D. Harman
and T. L. Macdonald, Chem. Res. Toxicol., 2003, 16, 123.
114. T. Wikberg, A. Vuorela, P. Ottoila and J. Taskinen, Drug Metab. Dispos.,
1993, 21, 81.
CHAPTER 7

Influence of Heteroaromatic
Rings on ADME Properties of
Drugs
DEEPAK DALVIE, PING KANG, CHO-MING LOI,
LANCE GOULET AND SAJIV NAIR

Departments of Pharmacokinetics, Dynamics and Metabolism and Medicinal

Chemistry, Pfizer Global Research and Development, San Diego, CA 92121,
USA

7.1 Introduction
Aromatic heterocycles play a critical role in medicinal chemistry and drug
design.1–3 More than half of known drugs contain at least one heterocyclic
component in their structure. New chemical entities are generally designed as
analogues of endogenous ligands that are vital to biochemical processes or, in
some cases, as simplified congeners of complex natural products. Since most of
these substrates contain heterocyclic rings, these rings by default become core
structures of the newly designed compounds. These drugs compete with
endogenous ligands and mimic (agonist) the ligands in their action. Alter-
natively, they block (antagonist) the active site in the receptor and therefore its
normal function. Some common endogenous ligands that have heteroaromatic
rings as their core structure include histamine, 5-hydroxytryptamine (sero-
tonin), adenosine 5 0 -triphosphate (ATP) and nucleic acid bases such as the

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

328
Influence of Heteroaromatic Rings on ADME Properties of Drugs 329
uracil, thymine, cytosine, guanine and adenine (Figure 7.1). Figure 7.2 repre-
sents some drugs that are structurally similar to these ligands.
Cimetidine, an antiulcer drug (Figure 7.2), was one of the first compounds
developed through rational drug design.4 This is a classic example of where the
endogenous ligand, histamine, was used as a lead to design a H2 receptor

NH2 NH2
N N
HO O O O
N NH2
O P O P O P O N N
N O
HN O O O
H
Histamine 5-Hydroxytryptamine (5HT)
HO OH
Adenosine5′-Triphosphate (ATP)

NH2 NH2 O NH2 O

N N N N NH NH NH
N N N N O N O
H H N O N O
H H H H
Adenine Guanine Thymine Cytosine Uracil

Figure 7.1 Structures of endogenous ligands that contain heterocyclic rings.

NH2
N
H H
H3C N NHCH3
N N
H3C S
S N
NCN O2
HN N N
N H N
H
Cimetidine Sumatriptan Rizatriptan
H2 Receptor Antagonist 5-HT1B and 5-HT1D Agonist 5-HT Agonist

Cl
O N
N N N HN
N
N O
N N S O
O HN Cl
N H
Gefitinib HO Dasatinib
F Tyrosine kinase Inhibitor
Tyrosine kinase Inhibitor

O
F
NH

N O
H
5-Fluorouracil
Anticancer Agent

Figure 7.2 Drugs that structurally mimic endogenous ligands.

330 Chapter 7
antagonist that could block the undesired effects of this endogenous ligand,
which is a mediator of gastric hypersecretory disorders and allergic responses.
Like histamine, cimetidine also has an imidazole ring and can bind to the active
site of the H2 receptor. It competitively inhibits histamine at the H2 receptors of
parietal cells and therefore reduces acid secretion.
Since then, several endogenous transmitters have been used to design drugs
that can either mimic or antagonise the effects of the natural ligand. For
instance, triptans mimicked serotonin and were designed as specific and
selective agonists for the 5HT receptors for the treatment of migraine and
cluster headaches.5 Since the discovery of sumatriptan, several analogues with
improved potency/selectivity and/or oral bioavailability have been made. All
triptans contain an indole nucleus and basic nitrogen at a similar distance from
the nucleus like serotonin but differ in their side chain (Figure 7.2).
Another example involves designing small molecules that inhibit various
protein kinases.6,7 These compounds compete with ATP and prevent the
phosphorylation of important physiological substrates. Although not structu-
rally similar to ATP, these compounds possess heterocyclic cores that interact
with various residues in the ATP binding site of the enzymes.8 Figure 7.2
represents two such receptor tyrosine kinase inhibitors that have been recently
approved in treatment of cancer.
Several natural products containing heteroaromatic rings have been used as
leads to design new molecules that can mimic their pharmacological activity.9
One example includes quinine, an antimalarial, which was used as a lead for
two drugs—primaquine and chloroquine. The latter two drugs remain ther-
apeutic standards for the treatment of malaria (Figure 7.3). Other historical
examples of analog design from natural products include the design of local
anaesthetic procaine from the alkaloid cocaine, or the synthesis of potent 3-
hydroxyl-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase inhibitor
(‘statins’) such as rosuvastatin, inspired from lovastatin.
Even where the natural substrate or ligand for a biological target does not
contain a heterocycle, compounds designed to act on that target frequently
contain heterocyclic groups. These compounds can function as agonists or
antagonists of the receptor depending upon the presence of the heterocyclic
motifs incorporated in them. Figure 7.4 shows some examples of this class.
In some cases, compounds that resemble essential metabolites can provide false
synthons in biosynthetic processes. Some examples include anti-metabolites such as
methotrexate (an analog of folic acid) or gemcitabine, cladribine or 5-fluorouracil,
which resemble deoxycytidine, deoxyadenosine or uracil, respectively (Figure 7.5).
Heteroaromatic rings have been used as bioisosteres by medicinal chemists to
design new candidates. Several functional groups including phenyl rings or
carboxylic acid and its ester analogues have been replaced by heterocyclic rings
and have resulted in therapeutically attractive compounds.10,11 These replace-
ments often confer structural integrity to the molecules and hold functionalities
in a particular geometry. Alternatively, the presence of a heteroatom in the ring
has been shown to increase interactions with the receptor or the enzyme, and
thus render greater pharmacological activity or specificity to resulting
Influence of Heteroaromatic Rings on ADME Properties of Drugs 331

HO O
N N
HN
O N
HN
N NH2
Cl N
Quinine Primaquine Chloroquine

O
N
O O
N
O
O
NH2
O
Cocaine Procaine

HO O F

O O
N N
S
O N
O O
H
OH

OH OH O

Lovastatin Rosuvastatin

Figure 7.3 Structures of some naturally occurring ligands and their synthetic analogs.

NH2
CH3
H3C O
N
CH3
O H3C N
Acetylcholine
Endogenous Ligand Tacrine
OH H
H O
HO N HO N
R
N O
HO
R = CH3, Adrenaline N N
R = H, Noradrenaline S Timolol

Figure 7.4 Examples endogenous ligands that lack heteroaromatic rings and their
structural mimics.

compounds. For example, the increased potency of pioglitazone (Figure 7.6)

compared with its predecessor ciglitazone is primarily attributed to fortuitous
incorporation of an additional receptor binding group (pyridine moiety).12
Introduction of heteroaromatic rings into drug molecules also affects their
physicochemical properties, which in turn can alter their absorption, distribution,
332 Chapter 7

O CO2H O CO2H

OH N CO2H NH2 N CO2H

H N H
N N N
N N
H
H2N N N H2N N N
Methotrexate
Folic Acid Dihydrofolate reductase Inhibitor

NH2
NH2
N
N N
O N
HO N N
HO
O O

OH H OH H
Deoxycytidine Deoxyadenosine

NH2 NH2
N N N O
O N N N Cl F
HO HO NH
O O
N O
F H
OH H
OH F

Gemcitabine Cladribine
5-Fluorouracil
A nucleoside analog A nucleoside analog
Anticancer Agent
used in chemotherapy used in chemotherapy

Figure 7.5 Structures of antimetabolites.

O O
N O
S
NH HN
S
O
O O
Pioglitazone
Ciglitazone

Figure 7.6 Structure of pioglitazone and ciglitazone.

metabolism and excretion (ADME) profiles. In the following sections, the

influence of the heteroaromatic rings on the physicochemical and ADME
properties of a molecule are reviewed. Attempts are made to correlate the effect
of the change in the physicochemical properties such as log P/log D, acidity/
basicity (pKa), hydrogen bonding capability of molecules, and polar surface area
(PSA) on their ADME properties. Calculated physicochemical parameter values
using ACD software are used to describe the effect of a structural change
(incorporation, removal or replacement of an aromatic heterocycle) on ADME
characteristics.
Influence of Heteroaromatic Rings on ADME Properties of Drugs 333
7.2 Types of Heteroaromatic Rings and their
Physicochemical Properties
Two types of heteroaromatic rings that are commonly found in drugs are the
six-membered heteroaromatic ring and five-membered heteroaromatic ring
systems.13–15
All six-membered heterocycles and their benz-fused derivatives are analogs
of carbocyclic rings (phenyl or the napthyl ring), where one or more methine
(CH) groups in the rings are replaced by a sp2-hybridised nitrogen atom. These
rings maintain the same electronic structure as the carbocyclic rings (they have
a complete cycle of p-orbitals containing 6p-electrons). The most commonly
used six-membered rings and their benz-fused derivatives that constitute sub-
structures of drugs are shown in Table 7.1.

Table 7.1 Physicochemical properties of commonly used six-membered rings.a

Heterocycle Name MW LogP pKa PSA

Benzene 78.1 2.18

Pyridine 79.1 0.84 5.23 12.89

N

N
Pyrimidine 80.1 0.26 1.78 25.78
N

N Pyridazine 80.1 –0.51 2.82 25.78

N
N
Pyrazine 80.1 0.002 0.92 25.78
N
N
N 1,2,3-Triazine 81.1 0.73 1.33 38.67
N
N

N
1,2,4-Triazine 81.1 0.92 1.32 38.67
N
N
1,3,5-Triazine 81.1 0.73 1.59 38.67
N N
N
N
N
Tetrazine 82.1 1.47 1.6 51.56
N

Quinoline 129.2 2.13 4.97 12.89

N

N
Quinazoline 130.2 1 3.44 25.78
N

N Cinnoline 130.2 1.14 2.79 25.78

N
N
Pthalazine 130.2 1.42 0.29 25.78
N

a
All values were calculated using ACD software (version 11.0).
334 Chapter 7
Although the presence of an electronegative nitrogen atom in the ring does not
alter the character of the p-orbitals of these rings, it distorts the distribution of
the p electrons and therefore lowers the energies of p-orbitals relative to the
phenyl ring. This coupled with the inductive effect of the electronegative nitrogen
atom makes it less reactive and renders properties that corroborate with con-
jugated imines or conjugated carbonyls.15 As the number of sp2 nitrogen atoms
in the ring increase, there is a further decrease in the reactivity of the rings. This
makes it less susceptible to P450 mediated oxidation. However, the lone pair of
electrons on the nitrogen atom that are planar to the ring provide a site for
protonation and other reactions such as N-oxidation (equivalent to a tertiary
amine), which is not analogous to the carbocyclic ring system.
The presence of sp2 nitrogen atoms in the ring also affects the physicochemical
properties of these rings. These rings are weak bases in comparison with their
alicyclic counterparts, piperidine or piperazines. As shown in Table 7.1, pyridine
and quinoline are the most basic of them all, with a pKa of B5. As the number of
nitrogen atoms in the ring increase, the pKa decreases. The pKa is also affected
by the position of the second nitrogen atom in the ring (Table 7.1). The repla-
cement of a methine group of the phenyl ring with nitrogen atoms also affects the
lipophilicity (log P value) and the polarity of heteroaromatic rings compared
with the benzenenoid rings. As the number of ring nitrogens increase, the log P of
the fragment decreases and the polarity increases (Table 7.1). Thus, introduction
of one nitrogen atom lowers the log P of a phenyl ring from 2.13 to 0.65, while
the log P value of pyrimidine and triazine ring is 0.3 and 0.7, respectively.
Five-membered heterocyclic rings form the second popular class of hetero-
aromatic ring systems that are commonly used as substructures in compound
synthesis. The more commonly used five-membered heterocyclic rings and
their benz-fused analogs are shown in Table 7.2 and Table 7.3, respectively.
Like six-membered rings, all five-membered rings and their benz-fused deri-
vatives share the same electronic structure of the phenyl ring, though the pre-
sence of only five atoms in the ring instead of six makes these rings smaller in
size relative to the carbocyclic and six-membered heteroaromatic rings.
Five-membered rings with a single heteroatom are the simplest ring systems
that belong to this class. The most common rings in this class are the pyrrole,
furan and the thiophene and their respective benz-fused analogs (indole, ben-
zofuran and benzothiophene); they contain a nitrogen, an oxygen and a sulfur
atom, respectively, in the ring system. Given the electron rich nature of these
rings, the five-membered rings are more reactive than the six-membered rings
and can undergo oxidative metabolism like the benzenoid rings (see Section
7.3.3). However, the degree of reactivity and the physicochemical properties of
these rings are dependent on the nature and the electronegativity of the het-
eroatom present in the ring (Table 7.2). Thus, thiophene and furan is the most
lipophilic of all with a log P value that is quite similar to benzene (1.74 and 1.95
versus 2.13) while the log P value of pyrrole is 0.85. The pyrrole moiety pos-
sesses an NH group that is typical of a secondary amine; however, the basicity
of pyrrole, pKa ¼ -3.8 for a conjugated acid, is much less than that of a sec-
ondary amine (pKa ¼ 10.87). This large difference is primarily due to the
Influence of Heteroaromatic Rings on ADME Properties of Drugs 335
Table 7.2 Physicochemical properties of commonly used five-membered rings.
Heterocycle Name MW LogP pKaa PSA pKab

Benzene 78.1 2.18

N Pyrrole 67.1 0.85 0.27 15.79 17

H

Furan 68.1 1.74 13.14

O

Thiophene 84.1 1.95 28.24

S
N
Imidazole 68.1 0.1 7.1 28.68 14
N
H
N
Oxazole 69.1 0.12 0.98 26.03
O
N
Thiazole 85.1 0.44 2.72 41.13
S
N Pyrazole 68.1 0.36 2.83 28.68 14
N
H

N Isoxazole 69.1 0.51 2.03 26.03

O

N Isothiazole 85.1 1.14 4 41.13

S
N
N 1,2,3-Triazole 69.1 0.63 1.2 41.57 8.7
N
H
N NH
1,2,4-Triazole 69.1 0.58 2.7 41.57 10.5
N
N N

N
4H-1,2,4-Triazole 69.1 0.89 2.7 41.57 10.2
H
N
NH 2H-1,2,4-Triazole 69.1 0.009 1.2 41.57 8.7
N

N N 1,2,5-Oxadiazole 70.1 0.16 1.47 38.92

O
N
N 1,2,4-Oxadiazole 70.1 0.69 0.28 38.92
O
N
N 1,2,3-Oxadiazole 70.1 0.68 4.12 38.92
O
N N
1,3,4-Oxadiazole 70.1 0.69 4.12 38.92
O

N N 1,2,5-Thiadiazole 86.1 0.34 0.28 54.02

S
N
N 1,2,4-Thiadiazole 86.1 0.2 0.46 54.02
S
336 Chapter 7
Table 7.2 (Continued )
Heterocycle Name MW LogP pKaa PSA pKab
N
N 1,2,3-Thiadiazole 86.1 0.96 3.39 54.02
S
N N
1,3,4-Thiadiazole 86.1 0.2 0.63 54.02
S
N N
N Tetrazole 70.1 0.6 0.8 54.46 4.73
N
H
a
pKa for conjugate acid,
b
pKa for conjugate base.

Table 7.3 Physicochemical properties of commonly used benz-fused 5-mem-

bered rings.
Heterocycle Name MW LogP pKaa PSA pKab

N
Indole 117 2.59 2.4 15.79
H

Benzofuran 118 2.68 13.14

O

Benzothiophene 134 3.2 28.68

S
N
Benzimidazole 118 1.32 5.26 28.68 13.2
N
H
N
Benzoxazole 119 1.59 1.17 26.03
O
N
Benzothiazole 135 1.9 0.85 41.13
S

N
N
Indazole 118 1.77 1.26 28.68
H
N Benzisoxazole 119 2.02 2.03 26.03
O

N Benzoisothiazole 135 1.85 4.37 41.13

S
N
N
1,2,3-Benzotriazole 119 1.44 1.17 41.57
N
H
a
pKa for conjugate acid,
b
pKa for conjugate base.

incorporation of the non-bonding electron pair of the N-atom into the cyclic
conjugated system of the pyrrole molecule. The pyrrole ring is also a weak acid
(about the same strength as the simple alcohol) and the pKa of its conjugate
base is 17 in aqueous solution.
Influence of Heteroaromatic Rings on ADME Properties of Drugs 337
These three rings form a platform for rest of the heteroaromatic rings that
contain an additional sp2 nitrogen atom (Table 7.2). The variation in the
position of the extra nitrogen atom further adds to the structural diversity of
the groups especially for heterocycles containing the two or three heteroatoms.
Extra nitrogen atoms and their placement in the ring (relative to the first
heteroatom) also have an important effect on the properties of the ring system.
The additional nitrogen atom has lone pair of electrons that is in plane with the
ring and therefore has attributes that are quite similar to six-membered het-
eroaromatic rings. Like pyridine and its six-membered analogs, the increase in
number of nitrogen atoms causes a decrease in the energy levels of the p orbitals
so that the heterocycles are less electron rich compared with their counterparts
containing only one heteroatom. This makes them less reactive relative to the
pyrrole, thiophene and furan.
The effect of additional sp2 nitrogen atoms can also be seen on the acidity
and basicity of these heterocycles (Table 7.2). Even though all ‘azoles’ are weak
bases like the pyrrole, the lone pair of electrons adds sites of protonation and
therefore most ‘azoles’ are stronger bases than pyrrole (Table 7.2).
The position of the sp2 nitrogen atom relative to the first heteroatom also
influences the basicity of the rings. For example, the imidazole ring has a pKa
of around 7.1 while pyrazole (nitrogen atom at the two position relative to the
first nitrogen) has a pKa of 2.5. The azoles containing additional NH groups
(imidazole, pyrazole, triazoles and tetrazoles) are stronger acids than the pyr-
role. Thus, imidazoles and pyrazoles are stronger acids (pKa of 14) than pyr-
role (pKa B17), while the triazoles and the tetrazole have pKa values of 9.5
and 4.92 (which is equivalent to a carboxy group), respectively.
The addition of nitrogen atoms also affects the log P value of these rings. The
logP values of all rings are listed in Table 7.2. The additional nitrogen atom
adds polarity to the ring and hence lowers the lipophilicity of all these rings
relative to their one heteroatom analogs. However, the trend is similar to that
observed in the rings with single heteroatom, i.e. logPthiazole or isothiazole 4
logPoxazole or isoxazole 4 logPimidazole or pyrazole.
An illustration of the effect of incorporation of different heterocyclic rings on
physicochemical properties of compounds is depicted in Table 7.4. Replace-
ment of a pyridine ring in sulfapyridine with pyrazine or pyrimidine results
changes the clogP and PSA values significantly. Similarly, replacement of six-
membered rings with thiazole, isoxazole or thiadiazole in sulfathiazole, sulfa-
methoxazole and sulfamethizole affects the lipophilicity and the PSA of the
molecule depending upon the ring incorporated in the compound. Also, the
addition of the type of substituents and the position of attachment can some-
times affect the physicochemical properties of a compound.
In addition to monocyclic heterocycles and their benz-fused derivatives, the
‘hybrid’ bicyclic rings resulting from annulation of two aromatic heterocycles
have also become very popular among medicinal chemists.2 These rings are
comprised of either five-membered rings fused with five-membered rings or five-
membered rings fused with six-membered rings. Given the number of hetero-
cycles that belong in the five-membered rings and five-membered rings classes,
338 Chapter 7
Table 7.4 Effect of heteraromatic rings on the physicochemical properties of
sulfa drugs.
Compound Structure MW clogP clogD7.4 PSA
H2N N
O
Sulfapyridine S
249 0.47 0.41 93.46
N
O H
H2N N
O
Sulfadiazine N 250 0.07 0.99 106
S N
O H N
H2N O
Sulfapyrazine 250 0.05 1.68 106
S N N
H2N O H
O S
Sulfathiazole 255 0.05 0.33 121.7
S
N N
O H
O H
N
Sulfamethoxazole S 253 0.66 0.54 106
O N
H2N O
O H
N N
S
Sulfamethizole N 270 0.52 1.16 134
O S
H2N O H
N
S
Sulfisomidine O N 278 0.3 0.004 106
N
H2N
O H N
N
S
Sulfadimidine O N
278 0.58 1.0 106
H2N
H2N
O O
Sulfamoxzole S
267 0.68 0.32 106
N N
O H

several bicyclic rings can be envisioned in this hybrid class. Some members of
this group are commonly used in drug design and are considered bioisosteres of
the benz-fused six-membered and five-membered rings. Others are commonly
observed in the natural and endogenous ligands, such as the purine ring (imi-
dazopyrimidine) observed in DNA bases or the pteridine (pyrazinopyrimidine)
observed in folic acid. Drugs containing these rings are shown in Figure 7.7.

7.3 Influence of Heteroaromatic Rings on ADME

Properties of Compounds
One goal of introducing a heterocycle as a bioisostere in a compound is to
improve its absorption, distribution (e.g. brain penetration) and clearance
Influence of Heteroaromatic Rings on ADME Properties of Drugs 339

O
N
N N N
N N O N
N
O N N
N N O
H2N
N
Zaleplon Zolpidem Temozolomide

Figure 7.7 Examples of drugs with ‘hybrid’ bicyclic ring systems.

characteristics while maintaining the activity of that compound. As discussed

earlier, incorporation of heterocyclic rings into drugs can affect their physical
properties and hence modify their pattern of absorption, metabolism or toxi-
city. The following sections present some case studies that illustrate the influ-
ence of the heteroaromatic rings on the absorption, distribution and clearance
(metabolism and excretion) of drugs.

7.3.1 Absorption
For most molecules, the rate and extent of absorption is largely dependent on
the dissolution of the compound in the aqueous contents of the gastrointestinal
tract and the permeability through the membrane. These in turn depend upon
the molecular characteristics of a drug such as molecular weight/size, aqueous
solubility, lipophilicity and the hydrogen bonding capacity of the molecule.16
Since aqueous solubility dictates the amount of drug that will be available for
absorption, most efforts are often directed towards the optimisation of solu-
bility of molecules.17 One approach involves introduction of polar solubilising
groups in a molecule to modulate the lipophilicity of the compound and hence
increase its absorption. Replacement of phenyl rings in the molecule with
aromatic heterocycles does not change the molecular weight (MW) of the
compound to a significant extent. However, it can modify the log D and pKa of
compounds to improve the absorption properties of a molecule.
Pyridine ring is one of the heterocycles that is commonly used to modify the
absorption properties of molecules. Replacement of a phenyl ring with this ring
not only imparts unique ionising properties at the acidic pH (therefore
increasing solubility), but also provides a handle for the introduction of various
groups that help in modulating the solubility and lipophilicity characteristics of
new molecules, which can assist in absorption of the compound. Khanna and
coworkers exemplify this in their continued efforts to seek new diaryimidazoles
as COX-2 inhibitors with better oral activity.18 Replacement of a tolyl group in
compound 7.1 (Table 7.5) with a pyridine ring (compound 7.2) yielded het-
eroaromatic analogues with improved water solubility which resulted in
excellent bioavailability (490%) and high plasma levels in the rat following a
dose of 20 mg kg1 relative to 7.1. The key to enhanced absorption of 7.2 was
probably ascribed to improved physicochemical properties of this compound
340 Chapter 7
18
Table 7.5 Physicochemical properties of COX-2 inhibitors.
CF3 CF3
N N

N N
N

SO2CH3 SO2CH3
Property Compound 7.1 Compound 7.2

MW 380 367
ClogP 2.68 1.13
PSA 60.34 73.23
clogD6.5 2.68 1.13
Sol6.5 (mg/ml) 0.005 0.05

(Table 7.5). Replacement of tolyl ring in 7.1 with a polar bioisosteric pyridyl
ring in 7.2 lowered the lipophilicity of 7.2 (clogP of 7.1 ¼ 2.68 and clogP of
7.2 ¼ 1.13) sufficiently to increase its aqueous solubility and therefore it
bioavailability.
Incorporation of a pyridine ring has also been shown to alter the absorption
characteristics of the glitazones. Assessment of the pharmacokinetic properties
of rosiglitazone indicated complete absorption following oral administration of
[14C]rosiglitazone solution (8 mg) and rosiglitazone tablets (4 mg, B.I.D.),
resulting in an absolute bioavailability of 95% and 99%, respectively.19 In
contrast, the oral bioavailability of troglitazone was estimated to be 40–50%
following oral administration of a 400 mg dose to humans.20 In addition,
administration of troglitazone (400 mg) with high fat meal increased the plasma
AUC by 30% to 80%, suggesting that oral absorption was incomplete in the
fasted state. Comparison of the physicochemical properties of these two gli-
tazones is shown in Table 7.6. Enhanced absorption of rosiglitazone is surmised
in part to the improved solubility at the clinically relevant doses. The key
modification is the replacement of the chroman moiety in troglitazone with an
aminopyridine ring in rosiglitazone. This change resulted in lower clogP of
rosiglitazone relative to troglitazone (3.02 versus 4.69) and increase in solubility
(0.05 versus 0.003 mg/ml) (Table 7.6). It is important to note that the dose,
which to a degree is also dependent upon its potency, also affects the solubility
of the compound in the gastrointestinal tract. Thus, in addition to lipophilicity,
increased solubility of rosiglitazone in contrast to troglitazone can be ascribed
to a lower dose (4 mg B.I.D., or 8 mg QD) of rosiglitazone in contrast to
troglitazone (200–400 mg, QD). The primary difference between the two
structurally related glitazones is its affinity for the PPAR-g receptor (332 nM
for troglitazone and 36 nM for rosiglitazone).
Optimum pKa and logP value of imidazole rings also makes this ring very
attractive in imparting solubilising properties in a molecule and modulating the
lipophilicity. These rings with appropriate substituents are also commonly used
Influence of Heteroaromatic Rings on ADME Properties of Drugs 341
19,20
Table 7.6 Comparison of troglitazone and rosiglitazone.
O
S O
HN
S NH
O HO O
N O
O
N
O

Property Rosiglitazone Troglitazone

MW 357.4 441.5
clogP 3.02 4.69
PSA 96.83 110.16
clogD6.5 2.35 4.31
Sol6.5 (mg/ml) 0.05 0.003
% Oral 95–99% 40–50%
bioavailability

in the optimisation of the pharmacokinetic properties of candidates in early

discovery. Quan and co-workers have demonstrated the use of imidazole ring
to enhance permeability of the lead compound, which is a potent Xa inhi-
bitor.21 These efforts have resulted in the discovery of a clinical candidate,
razaxaban (Table 7.7), which is an orally active factor Xa inhibitor. In this
programme, although the lead compound (compound 7.3, Table 7.7) was
highly selective as a Xa inhibitor, it had poor permeability (Caco-2 Papp is
o0.1106 cm sec1) and solubility (o0.1 mg mL1) resulting in only 2%
bioavailability in dogs. Since the clogP of the compound was moderate (3.12),
low permeability characteristics were probably attributed to the high PSA of
the molecule (142 Å2). In an effort to address the permeability issue, the phe-
nylsulfonamide moiety was replaced with several heterocycles as shown in
Table 7.7. Although the pyridylsulfonamide group (7.4) showed slight
improvement in the permeability in Caco-2 studies (0.54106 cm sec1),
replacement of the phenylsulfonamide group in the molecule with solubilising
heterocycles such as 1-methylimidazol-2-yl (7.5) and 2-methylimidazol-1-yl
(7.6) groups improved the permeability in Caco-2 cells significantly (4.7–
7.4106 cm sec1). The replacement of phenylsulfonamide with imidazoyl
groups also resulted in the lowering of PSA. The enhanced permeability of the
compound was potentially due to the reduced PSA following incorporation of
the imidazoyl moieties as shown in Table 7.7. The introduction of the dime-
thylaminomethyl imidazole group in razaxaban further aided in optimising
potency and permeability, and resulted in 84% bioavailability in dogs.
Sometimes, incorporation of a heteroaromatic moiety in a molecule can
affect the pharmacokinetic properties of drugs. This is illustrated by the dif-
ferences observed in the oral bioavailability of nilotinib (Tasigna) relative to
imatinib (Gleevec) (Table 7.8). Both nilotinib and imatinib are tyrosine kinase
inhibitors used in the treatment of chronic myelogenous leukemia (CML).
Nilotinib appears to offer benefits in patients who are resistant to imatinib,
342 Chapter 7
21
Table 7.7 Properties of Razaxaban and its analogs.
F3C F
H
N N
N R
O

H2N
N O
Caco-2 Papp
R MW clogP PSA cm sec1106
SO2CH3
7.3 559 3.1 142 o0.1

SO2CH3
7.4 560 3.0 154 0.54
N

N
7.5 485 1.9 117 4.6
N

7.6 485 4 117 7.4

N N

N
Razaxaban 528 3.24 120 5.6
N N

which is used as a first line treatment for CML. In humans, the absolute oral
bioavailability of imatinib is 98%, indicating that absorption is almost com-
plete.22 In contrast, the systemic exposure of nilotinib is increased by 82% when
administered 30 minutes after a high fat meal, suggesting that oral absorption
of this drug in the fasted state is incomplete.23 The variation in their absorption
properties can be attributed to the differences in the lipophilicity and solubility
(at pH 6.5) between the two compounds (Table 7.8). Incorporation of a 4-
methylimidazole group in nilotinib increases the clogD6.5 to 4.36 and reduces
the solubility to 0.0004 mg ml1 relative to imatinib (clogD6.5 ¼ 1.77 and
solubility 0.11 mg ml1). Thus, replacement of the substituted imidazole with
the piperazine side chain appears to be associated with decreased lipophilicity,
increased solubility and an increase in the extent of oral absorption.
The tetrazole moiety is comparable in size and in acidity to the -COOH
group and is therefore commonly used as a bioisosteric replacement of car-
boxylic acids.24 Tetrazoles as isosteres of carboxylic acids are also discussed in
detail in Section 3.7.2. The tetrazole moiety is 10-fold more lipophilic and
Influence of Heteroaromatic Rings on ADME Properties of Drugs 343
Table 7.8 Comparison of nilotinib (Tasigna) and imatinib (Gleevec).
O O
HN NH
N
N NH NH
N N
F
N N N N
F N
F N

Property Nilotinib Imatinib

clogP 4.42 2.89

PSA 97.62 86.28
clogD6.5 4.36 1.77
Sol6.5 (mg ml1) 0.0004 0.11

Table 7.9 Comparison of physicochemical properties of losartan and

eposartan
Cl
OH
N
N
N
N
OH

O
O N
NH
OH
S N N
Property Eprosartan Losartan

ClogP 2.96 4.36

PSA 120 92.5
clogD6.5 0.53 2.6
Sol6.5 (mg ml1) 4.6 0.04
% Oral bioavailability 1325 3726,27

masks the polarity of the carboxylate group, and hence tends to increase the
bioavailability of compounds. The effect of replacing a carboxylic acid func-
tionality with a tetrazole ring on absorption has been demonstrated in angio-
tensin II converting receptor antagonists, eprosartan and losartan (Table 7.9).
Assessment of the pharmacokinetic properties of losartan and eprosartan
indicates that, even though the bioavailability of losartan in humans was only
37%, the drug is completely absorbed through the gut following oral admin-
istration.25 In contrast, only 15% of eprosartan was absorbed through the gut,
resulting in bioavailability of 13%.26,27 Low absorption is inherent in the
physicochemical properties of eprosartan rather than incomplete dissolution.
As shown in Table 7.9, the tetrazole ring in losartan and the carboxy group in
344 Chapter 7

HN N
N N
CO2H
O CO2H O CO2H

O H O H
H H
O N N O N N
N N
H O H O
O O
7.8 7.9
clogD7.4 = 0.76 clogD7.4 = 0.2
Papp = 5.1 x 10-7 cm/sec Papp = <1 x 10-7 cm/sec

Figure 7.8 Structures of protein tyrosine phosphatases 1B (PTP1B) inhibitors and

their clogD.

eprosartan have pKa values of 6 and B4, respectively; hence both compounds
are almost completely ionised at the pH in the gastrointestinal tract (B6.5).
Although losartan is less soluble than eprosartan (0.04 mg/ml versus 4.6 mg/
ml), the negative clogD6.5 of eprosartan (log D6.5 of 0.53) hinders the per-
meability of the compound through the membrane. On the other hand,
replacement of dicarboxylic acid with a tetrazole moiety and a hydroxymethyl
substituent in losartan results in a clogD6.5 of 2.6, thereby enhancing the per-
meability through the gastrointestinal tract in spite of poorer solubility. In
addition to lipophilicity, the difference in the polar surface area of the two
compounds (92.5 Å2 for losartan versus 120 Å2 for eprosartan) may favour
permeability of losartan through the membranes compared to eprosartan.
In another example, Liljebris and co workers have reported the influence of
tetrazole ring on the bioavailability of small molecular weight peptidomimetics
that were designed and synthesised as competitive inhibitors of protein tyrosine
phosphatases 1B (PTP1B).28 Because of the low cell permeability of this com-
pound class, the possibility of replacing one or both of the remaining carboxyl
groups while maintaining PTP1B inhibitory activity was explored. An impor-
tant discovery was the ortho tetrazole analogue 7.8, which was equipotent to the
lead dicarboxylic acid analogue 7.9 (Figure 7.8). This novel monocarboxylic
acid analogue revealed higher Caco-2 cell permeability and this was attributed
to the increase in lipophilicity compared to the lead (Figure 7.8).

7.3.2 Distribution
Central nervous system (CNS) penetration is an important consideration in
drug discovery and development for many therapeutic areas. Thus, examples
that demonstrate the influence of aromatic heterocycles on the permeability of
compounds to cross the blood–brain barrier (BBB) are discussed in this section.
Like absorption, permeability of a drug across the blood–brain barrier is
affected by its PSA and logP, i.e. the polarity of a compound.16 Since the
Influence of Heteroaromatic Rings on ADME Properties of Drugs 345
Table 7.10 Comparison of physicochemical properties and ratio of the brain
to blood concentrations for mianserin and mirtazapine.29

N N

N N N

Property Mianserin Mirtazapine

clogP 1.14 0.28

PSA 6.48 19.47
Brain–blood 9.7 3.4

presence of one or more heteroatoms in the aromatic heterocyclic ring results in

variation of both parameters, this can potentially influence the permeability
across the blood–brain barrier. Most often, inclusion of heteroatoms in a ring
system results in an increase in the number of hydrogen bond donors (HBDs)
or hydrogen bond acceptors (HBAs) (and therefore an increase in PSA) in a
molecule. Therefore incorporation of aromatic heterocyclic rings into a mole-
cule or replacement of carbocyclic rings with aromatic heterocycles can
somewhat hinder the permeation of a drug into the CNS.
For example, a comparison of the ratio of concentrations in the brain to
blood of the two antidepressants, mianserin (Tolvon) and mitrazapine
(Remeron), indicates that mianserin has Bthree-fold higher concentrations in
the brain than mirtazapine (Table 7.10).29 The only structural difference
between these two compounds is the replacement of one of the phenyl rings
with a pyridine ring (Table 7.10). Comparison of log P and PSA values of
mianserin (clogP ¼ 1.14; PSA ¼ 6.48 Å2) and mirtazapine (clogP ¼ 0.28; PSA
¼ 19.47 Å2) indicates that isosteric replacement of a methine group in mianserin
with a sp2 nitrogen atom lowers the lipophilicity and increases the PSA of
mirtazapine. Even though the molecules with small PSA (o20 Å2) are com-
pletely absorbed via passive diffusion,30 the relatively low permeability of
mirtazapine compared to mianserin is possibly associated with the lowering of
clogP value and a slight increase in PSA by isosteric replacement of a phenyl
with a pyridyl group in mirtazapine.
Although lipophilicity plays an important role in BBB permeability, some
reports have indicated that changes in PSA of a molecule can affect BBB
permeability more than the changes in lipophilicity.29 The presence of addi-
tional heteroatoms in heteroaromatic rings generally increases its potential for
hydrogen bonding (and therefore the PSA) of the molecules. Hence, this can
negatively impact the permeability of the molecule across the blood–brain
barrier.
The following example illustrates this point. As shown in Table 7.11,
replacement of a pyridine ring in compound 7.10 with a thiazole ring in com-
pound 7.11 introduces another heteroatom into the molecule. Even though the
346 Chapter 7
Table 7.11 Effect of PSA on the blood brain permeability of compounds.29

S
N N
O N N O N N
H H
Compound 7.10 Compound 7.11

clogP 3.75 3.93

PSA 37.39 65.63
Brain–blood 4.9 2.8

Table 7.12 Comparison physicochemical properties of a7 neuronal receptor

agonists 1 and 2.33
O O
O O S
NH N Cl

N N

Property Compound 7.12 Compound 7.13

clogP 1.5 0.57

clogD7.4 0.5 2.1
PSA 42 33
HBD 1 0
Brain to plasma ratio 0.76 59

sulfur atom is a weak hydrogen bond acceptor, its presence in the ring can
affect the PSA of the molecule as observed in 7.11 (PSA for 7.10 is 37.39 versus
65.63 for 7.11). However, no significant change in the lipophicity is observed
with such a replacement (clogP for 7.10 is 3.75 and clogP for 7.11 is 3.93).
Determination of brain to blood ratio of compounds 7.10 and 7.11 shows a 1.8-
fold decrease when the aminopyridine group in 7.10 is replaced with ami-
nothiazole moiety in compound 7.11.29 Since the clogP for both compounds are
similar, the decrease in the brain to blood ratio in 7.11 is ascribed to the change
in PSA of the two compounds.
Recent studies have shown that the permeability across the blood–brain
barrier is primarily affected by the number of hydrogen bond donors than
hydrogen bond acceptors present in the molecule.31,32 Reduction in the number
of hydrogen bond donors in a molecule has been attempted by addition of
chlorothienyl moiety to the cyclic carbamate 7.12 (Table 7.12) by Tatsumi and
co-workers.33 Compound 7.12 was synthesised as a part of developing a7
neuronal nicotininc acetylcholine receptor agonists for the treatment of schi-
zophrenia. Addition of the chlorothienyl group led to an increase in brain to
Influence of Heteroaromatic Rings on ADME Properties of Drugs 347

O
N
OCH3
HN N
HN
HN O
HN O
N

N N

Compound 7.14 Compound 7.15

clogP = 4.4 clogP = 3.4
clogD = 2.6 clogD = 3.4
PSA = 59 PSA = 102
HBD = 2 HBD = 2

HBD = no of Hydrogen Bond Donors

Figure 7.9 Intramolecular hydrogen bonding in the most polar compound increases
the brain penetration.

plasma ratio from 0.76 for 7.12 to 59 for 7.13. The change was rationalised in
part to the reduction in hydrogen bond donating capability of 7.13, resulting in
a slightly lower PSA relative to 7.12. This approach also led to an overall
improvement of clogD7.4 of the compound (clogD7.4 ¼ 0.57 for 7.13 versus -1.5
for 7.12). The calculated physicochemical parameters are shown in Table 7.12.
A group of oxerin-1 and 2 (OX1 and OX2) receptor antagonists, designed
and synthesised by Porter and his co-workers, represents an interesting example
where introduction of a polar napthyridine group in a molecule increased its
brain to plasma ratio in spite of lower log P and higher PSA relative to the
lead.34 The lead compound 7.14 (Figure 7.9) showed good affinity for the
receptors but exhibited poor blood–brain barrier permeability following an
intravenous infusion to rats. The compound was essentially undetectable in the
CNS even though it exhibited good lipophilicity (clogP, 4.4; clogD, 2.6) and a
moderate polar surface area of 59 Å2. Structure–activity relationship (SAR)
studies to optimise potency and selectivity resulted in compound 7.15 that
showed considerably higher brain to plasma ratio (0.4) relative to 7.14. Higher
ratios were observed despite the lower clogP than 7.14 (clogP 3.4) and sig-
nificantly higher PSA (102 Å2). So in theory, even though 7.15 has two
hydrogen bond donors like the lead and unfavourable PSA, potential intra-
molecular hydrogen bonding between the nitrogen atom at the 5-position of
napthyridine and the proton on the proximal urea nitrogen can mask its proton
donor ability and therefore reduce the PSA. This in turn can help in enhancing
BBB penetration.
The previous section describes the advantages of introducing tetrazolyl
moiety as a bioisostere of carboxylic acid in an effort to increase in the bioa-
vailability of losartan. However, introduction of this ring has been shown to
negatively affect the brain exposure of compounds. This has been observed with
some tetrazolyl analogs that were synthesised as CCK-B antagonists.35 In this
report, replacement of the methyl group on the phenyl urea (7.16) with a
348 Chapter 7
Table 7.13 Comparison of the physicochemical properties of CCK-B
antagonists (1 and 2) and mGlu2 modulators (3 and 4).35,36

H
N N
O O
N NH O O N N
NH N NH
N NH
Ph 7.16 N
7.17
Ph
OH O
OH O

O
O S
O
N
N N
7.18 7.19
N NH
Compound MW clogP clogD PSA HBD HBA

7.16 398 3.71 3.71 73.8 2 6

7.17 452 2.63 0.65 128 3 10
7.18 410 5.4 3.4 110 2 8
7.19 374 5.4 5.3 84.7 1 4

tetrazole moiety (7.17) (Table 7.13) resulted in poor brain exposure of 7.17.
Similarly, enhanced brain to plasma (B/P) ratios in rat were observed following
replacement of tetrazole terminus in 7.18 (B/P ¼ 0.01) with a 4-thiopyridyl
group (47.19) (B/P ¼ 1.1) for the modulators of mGlu2 receptor (Table 7.13).36
In both examples, replacement of the substituent with the tetrazole moiety
increased the number of hydrogen bond donors and hydrogen bond acceptors,
and therefore increased PSA. Although there was a decrease in the clogD in
both cases, the authors attributed the poor permeability to cross the blood–
brain barrier to a high PSA and an increase in hydrogen bond donors.

7.3.3 Metabolism
In addition to enhancing absorption properties, heterocyclic rings are often
incorporated into a molecule to improve the metabolic stability of newly syn-
thesised compounds. Replacement of a carbocyclic ring with an aromatic
heterocycle is a common strategy to reduce the metabolic liability and some-
times block the sites of metabolism.
Structure–metabolism relationship studies have shown that incorporation of
one or more heteroatoms in an aromatic ring of a compound modulates its
chemical and biochemical reactivity (in addition to changing lipophilicity), and
therefore alters its metabolism. The degree of influence of heteroaromatic rings
on the metabolism of a compound depends on the type of ring system that is
Influence of Heteroaromatic Rings on ADME Properties of Drugs 349
incorporated in a molecule and the number of nitrogen atoms present in the
ring. As described in Section 7.2, compounds containing six-membered ring
systems are less reactive than their corresponding carbocyclic analogs. Incor-
poration of one or more nitrogen atoms in the rings decreases the electron
density in the aromatic ring carbons, thereby decreasing P450-mediated aro-
matic oxidation. Thus, the lower reactivity of six-membered rings with multiple
nitrogen atoms makes these prime fragments to improve metabolic stability
such as higher resistance to chemical or enzymatic degradation.
Despite low reactivity, compounds containing these rings are not devoid of
P450 catalysed metabolic transformations. For example, voriconazole, an
antifungal agent containing a pyrimidine ring, primarily undergoes CYP2C19
or 3A4-mediated N-oxidation of the pyrimidine ring (Figure 7.10).37,38 Simi-
larly, compounds containing a pyridine ring such as indinavir or rosiglitazone
are subject to oxidative metabolism either on the pyridine ring or on the rest of
the molecule by P450 (Figure 7.11).19,39 Binding of these compounds to the
P450 active site is attributed to their overall lipophilicity (clogD7.4 for indina-
vir ¼ 3.4; clogD7.4 for rosiglitazone ¼ 2.5 and clogD7.4 for voriconazole ¼ 1.9).
Alternatively, the presence of the lone pair of electrons on the nitrogen atom
of the aromatic heterocycles provides a site for glucuronidation by UDP-glu-
curonyl transferases (UGT) for most six-membered rings. Pyridine containing
compounds are readily prone to glucuronidation. Indinavir, which undergoes
extensive oxidative metabolism, is also subject to N-glucuronidation (Figure
7.11).39,40 Other examples include compounds such as nicotine that form the N-
glucuronide in humans (Figure 7.12).41 Lamotrigine, an anti-epileptic agent
containing a triazene ring, is also metabolised via N-glucuronidation (Figure
7.12).42 This suggests the importance of nucleophilicity of the nitrogen atom
rather than the reactivity of the heteroaromatic ring in glucuronidation of the
molecule. Not much is known about the interactions of the substrates with
UGTs.
A unique metabolic pathway for compounds containing electron deficient
azaheterocycles is oxidation by molybdenum hydroxylases such as aldehyde
oxidase (AO) and xanthine oxidase (XO).43 Unlike P450s that promote the

N N
N N
N N
OH F OH F
CYP3A4

N F N F
F F
N N N-Oxide
Voriconazole O Metabolite

Figure 7.10 Metabolism of the pyrimidine moiety in voriconazole by CYP2C19 and

3A4 to its N-oxide metabolite.
350 Chapter 7

N
O N N-Oxide metabolite
HN
O
OH O
N
N N-Glucuronide
N OH N conjugate
H N CO2H
O
Indinavir N HO OH
OH

Other Oxidative
Metabolites

O
O
NH OH Glucuronides
HN
O S
N O S
N N
O N
O Sulfates
Rosiglitazone

Figure 7.11 Metabolism of the pyridine ring in indinavir and rosiglitazone.

CH3
N
CO2H
O OH
N N
N H3C HO
OH
Nicotine
N-Glucuronide of
Nicotine

HO2C
NH2 NH2 O OH
OH
N N N N HO
N N
NH2 H2N
Cl Cl N-Gluruconide
of Lamotrigine
Cl Cl
Lamotrigine

Figure 7.12 N-Glucuronidation of nicotine and lamotrigine.

electrophilic attack via the Fe¼O complex on aromatic substrates, these

enzymes oxygenate heteroaromatic rings via nucleophilic attack of Mo–OH on
the electron-deficient carbon atom adjacent to nitrogen atoms in the hetero-
cycles. Both AO and XO can oxidise several single or polyaromatic hetero-
cycles, but AO shows broader substrate specificity than XO. Some of the
Influence of Heteroaromatic Rings on ADME Properties of Drugs 351

HN NH
N Br
O N N
N O N
N N
Metyrapone Quinine
Brominidine
HN
O
H O
N O O
N

N N
N
N N O
DACA N
N N
N-(2-(dimethylamino)ethyl)
Zaleplon CN O
acridine-4-carboxamide
Carbazeran

F HN NH2 NH2
N
N O
N
O N H2N N OH
H N N HN
5-fluoro-2-pyrimidone N N
Hydralazine O O
HO
Methotrexate
HN
O NH2 O
NH
HN N
N
N H2N N
N

Acyclovir
N
OH
Zoniporide

Figure 7.13 Aldehyde oxidase mediated oxidation of various drugs. Arrows indicate
the site of oxidation.

representative compounds that undergo AO and XO mediated metabolism are

shown in Figure 7.13. An example is the bioactivation of famciclovir to the
active antiviral agent penciclovir which requires AO mediated metabolism.44
Famciclovir is converted to deoxypenciclovir, which is oxidised by AO to
penciclovir (Figure 7.14).
Hydrophobicity and electron deficiency appears to be two important factors
in predicting metabolism by AO. As the number of nitrogen atoms in the ring
increases, the ring becomes more electron deficient and more hydrophobic. As a
result, the potential for the compound to undergo AO-mediated oxidation
increases. This has been demonstrated by Hall and Kretinistsky who studied
the capability of the rabbit AO to oxidise purines and their analogs.45 In this
study, the chemical nature of the 6-substituent of purine markedly influenced
substrate efficiency. Substituents that were hydrophobic and electron
352 Chapter 7

O NH2
NH2 NH2

O N N
N N OH N NH
OH
N Deacetylation AO
N N O
N
N N
O Famciclovir OH Deoxypenciclovir OH Penciclovir
O

Figure 7.14 Metabolic activation of famciclovir to penciclovir.

NH2 NH2
O N O N
N N N N
O N O O N O

Prazosin O O
Terazosin

Figure 7.15 Structures of Prazosin and Terazosin.

withdrawing enhanced AO oxidation, whereas 6-hydroxy and 6-amino sub-

stituent virtually abolished substrate activity.45
Compounds containing five-membered heterocyclic rings behave quite
differently than those containing six-membered rings. Since all five-membered
heterocycles are electron rich in nature, compounds with these rings can
undergo similar P450 mediated oxidative metabolism like their carbocyclic
congeners. In some instances, the five-membered rings serve as alternative sites
for metabolic attack and, at times, have the potential to undergo unusual
metabolic transformations that can result in toxic events. The metabolic
transformations for some well-known drugs containing five-membered rings
result in electrophilic intermediates (see Section 7.3.4). The metabolism of five-
membered heterocycles is also extensively reviewed elsewhere.46
There is a good correlation between lipophilicity and the metabolism of five-
membered rings. Thiophenes, pyrroles and furan containing compounds, which
are quite lipophilic, undergo extensive metabolism to ring opened products.
Lowering lipophilicity by subtle changes in these heterocyclic rings can
significantly improve the pharmacokinetic properties of the compound. For
example, prazosin (Figure 7.15), a furan containing compound, is a selective
a1 blocker that is extensively metabolised by CYP450 resulting in a half-life of
2–3 h and a bioavailability of 45–65%.47,48 However, terazosin, in which the
furan ring is replaced by a tetrahydrofuran moiety, increases the bioavailability
to 90% with a half-life of 12 h.49 The increase in exposure of terazosin is
possibly attributed to the lowering in the clogP value (0.8) compared to pra-
zosin (clogP 2.14) by replacement of a furan with a hydrophilic tetrahydrofuran
moiety (Figure 7.15).
Influence of Heteroaromatic Rings on ADME Properties of Drugs 353

O OH OH
N NH O
O AO O O
S NH2 S NH2 S NH2
O O O
Zonisamide Ketimine Hydroxyphenyl Metabolite

F 3C F F3C F
H N H N
N N
N N
N N N N
O AO N N
O

H2N H2N
Razaxaban
N O NH OH

O O

HN HN
N AO N
N N
N S S-Methyl transferase NH S
Cl Cl CH3
Ziprasidone

Figure 7.16 Reductive metabolism of zonisamide, razaxaban and ziprasidone by

aldehyde oxidase.

Compounds containing five membered heterocycles, especially those contain-

ing two heteroatoms, undergo an AO-mediated reduction and subsequent ring
opening of the heterocyclic ring. The 1,2-benzisoxazole moiety in zonisamide
(Figure 7.16) is reductively cleaved to a hydroxyphenyl derivative via the ketimine
intermediate.50 A similar reductive cleavage has been observed in the metabolism
of the anticoagulant razaxaban as well as the benzothiazole derivative ziprasi-
done (Figure 7.16).51–53 Since the mechanism of AO-mediated reduction path-
ways is unknown, a proper correlation of the physicochemical properties and the
AO reducing activity has not been reported. Comparison of the clogP values
(0.72 for zonisamide and B3.0 for ziprasidone and razaxaban) and molecular
weights (ranging from 212 to 528) of the three drugs suggests that lipophilicity
could be a determinant in the reductive metabolism of these compounds by AO.
Like six-membered rings, the compounds are less susceptible to metabolic
attack by P450 or other enzymes as the number of heteroatoms in the ring
increase. For example, bioisosteric replacement of a carboxylic ester group in a
molecule with an oxadiazole or a thiadiazole moiety increases the stability to
hydrolytic degradation. This approach has been employed in the design of
several arecoline analogs (Figure 7.17).54,55 The ester functionality is prone to
hydrolysis by a number of esterases that are present throughout the body and
also subject to chemical degradation at acidic pH in the gastrointestinal tract.
This liability prevents the ester containing drugs from being used orally.
354 Chapter 7

O O N S N
O N N
N N N
CH3 CH3 CH3
Arecoline

Figure 7.17 Structure of arecoline and its analogs.

H
N N N N N
N
N Cl N
N NH O
N N OH
Losartan Irbesartan

Figure 7.18 Structures of losartan and irbesartan.

Compounds containing rings with two or more nitrogen atoms increase

compound polarity and lower the log P of the molecule. Most often, these
compounds are excreted unchanged in the urine as observed with diflucan, a
triazole derivative (see Section 7.3.4).56 Alternatively, more lipophilic com-
pounds are eliminated via the glucuronidation pathway. Formation of a N2-
glucuronide has been identified as a major route of metabolism for the tetra-
zole-containing losartan (Figure 7.18).57,58 Within the same therapeutic drug
class, irbesartan (Figure 7.18) has also been shown to undergo tetrazole N2-
glucuronidation in humans.59

7.3.4 Excretion
As mentioned earlier, introduction of polar heterocyclic rings systems in
molecules can influence the routes of clearance. Substitution with appropriate
polar heterocyclic rings during drug design that allows the drug to get absorbed
and confers metabolic stability to the molecule can sometimes favour the drug
to be cleared renally.
Fluconazole (Diflucan) (Table 7.14) is a classic example in which introduc-
tion of heterocyclic rings has influenced the route of elimination of this drug.
Fluconazole belongs to a triazole class of antifungal agents used in the treat-
ment and prevention of superficial and systemic fungal infections. As described
before, triazoles are ring systems that impart significant polarity to a molecule
and make them resistant to metabolism. However, these rings are unionised at
physiological pH and therefore allow complete absorption of the drug in the
Influence of Heteroaromatic Rings on ADME Properties of Drugs 355
Table 7.14 Physicochemical properties of fluconazole, voriconazole, itraco-
nazole and posaconazole
N
N N
N N N
F
O O
N
N N Cl N
O O N N N
F HO Cl N
Fluconazole
Itraconazole
N
N N
N
OH F N N
O O
OH
F F N
N O O N N N
F F N
N Posaconazole
Voriconazole
Compound MW clogD7.4 % Excreted in urine

Fluconazole 306.2 0.45 80

Voriconazole 349.3 1.21 2.1
Itraconazole 705.6 4.93 0.01
Posaconazole 702.7 3.68 Not detected

gastrointestinal tract (oral bioavailability is 490%).60 The presence of two

triazole rings in fluconazole imparts polarity to the molecule (clogD7.4 0.45)
(Table 7.14). Poor lipophilicity of the molecule renders the molecule stable to
P450 and UGT mediated metabolism and results in about 80%of the drug
being primarily excreted in the urine unchanged.56 In contrast, other triazole
containing compounds such as itraconazole and pasoconazole are quite lipo-
philic and have a clogD7.4 value of 4.93 and 3.68, respectively (Table 7.14).
Both compounds are primarily cleared via metabolism (P450 mediated for
itraconazole and glucuronidation for pasoconazole) and no unchanged drug is
excreted in the urine.61,62
Sometimes, subtle changes in the molecular structure can result in significant
changes in the routes of clearance despite the low molecular weight and low
log D. For example, replacement of the triazolomethyl moiety in fluconazole
with a fluoropyrimidinyl benzyl group in voriconazole (Table 7.14) results
in a clogD7.4 of 1.2. Regardless of the low molecular weight like flucona-
zole (Table 7.14) and only a slight change lipophilicity (DlogD7.4 ¼ 0.76 rel-
ative to fluconazole), the compound is extensively metabolised (2% of the
unchanged drug being excreted in the urine) to the pyrimidine N-oxide by CYP
3A4 (Table 7.14).63,64
Even though a correlation between physicochemical properties (log D and
MW) and the renal excretion of the drugs has been well established,16 good
356 Chapter 7

N F
N
F N N NH2 F
F O
F
Sitagliptin
MW 407.3
clogD7.4 1.93
clogP 2.06

Figure 7.19 Structure of sitagliptin (Januvia), a DPP-4 inhibitor.

substrate properties of compounds for some efflux transporters can also lead to
renal excretion of lipophilic drugs. Sitagliptin (Januvia) (Figure 7.19), an orally
active triazolopyrazine derivative and an inhibitor of the dipeptidyl peptidase-4
(DPP-4) enzyme, presents an interesting case.65
Despite modest lipophilicity (clogD7.4 ¼ 1.93) and molecular weight
(MW ¼ 407), approximately 79% of sitagliptin is excreted unchanged renally
and metabolism is a minor pathway of elimination.65 Following a [[14]C]sita-
gliptin oral dose, only 16% of the radioactivity was excreted as metabolites of
sitagliptin. Metabolic stability is achieved by the presence of triazolopyrazine
and trifluorophenyl group in the molecule. Reports indicate that sitagliptin is a
low affinity substrate of human organic anion transporter-3 (hOAT-3).66
Organic anion transport systems such as human organic anion transporters
(hOAT1 and 3) are predominantly expressed in the kidney and play an
important role in the transport of organic anions across the basolateral
membrane of human proximal tubules.67 Hence, clearance of sitagliptin pri-
marily involves its active tubular secretion from the body into the urine.
Like the kidney, many transporters are expressed on the canalicular mem-
brane of the liver. These transporters can mediate the excretion of selected
compounds into the bile. Lipophilic compounds and high molecular weight
compounds with a considerable hydrogen bonding functionality and which
show poor permeability can be potentially cleared via this active efflux process.
Incorporation of heterocyclic rings in molecules can alter the route of clearance
for high molecular weight lipophilic compounds from metabolism to active
efflux (and therefore biliary excretion of unchanged drug). The change is
dependent upon the hydrogen bonding capability that is incorporated into the
molecule by the newly introduced heterocyclic ring system. Gardner and co-
workers have demonstrated this for inhibitors of thromboxane A2 synthase
(7.20 and 7.21, Table 7.15).68 Both the compounds showed high hepatic
extraction ratio (E ¼ 0.9) in the isolated perfused rat liver. Comparison of their
physicochemical properties showed that both compounds had molecular
weight close to 500 (Table 7.15). Compound 7.20 was mainly eliminated from
the body by metabolism whereas 7.21 was excreted into the bile unchanged.
This suggested that 7.21 was a substrate for one of the hepatic efflux
Influence of Heteroaromatic Rings on ADME Properties of Drugs 357
Table 7.15 Effect of replacement of heterocyclic ring systems in a compound
on the route of elimination.68
Compound Structure PSA clogD7.4 clogP HBD HBA
N

HO2C
CH3
N

7.20 109 0.053 2.78 2 7

O NH
S
O

F
N
N
HO2C
CH3
N

7.21 114 0.73 1.93 2 8

O NH
O S

transporters. The change in the route of clearance for 7.21 was attributed to its
high polarity (negative clogD7.4) relative to 7.20 and increase in the number of
hydrogen bond donors and acceptors in the molecule. It should be noted that
the above inference is merely speculative since the influence of the physico-
chemical properties on the SAR of efflux transporters is not fully understood.

7.4 Influence of Heteroaromatic rings on Toxicity of

Compounds
Heteroaromatic rings can influence the toxicity of compounds in a desirable or
an undesirable manner. Lipophilic compounds containing phenyl rings (espe-
cially phenolic or aromatic amines) are bioactivated to electrophilic quinoid
intermediates that covalently bind to macromolecules and possibly cause
untoward side effects. A strategy to mitigate the potential risk of bioactivation
involves replacement of the labile motifs with polar and less reactive bioisos-
teres. Since the aromatic heterocycles can influence the reactivity and lipophi-
licity of a molecule, these rings are attractive fragments that can be
incorporated into new compounds to prevent metabolite activation.
358 Chapter 7
Table 7.16 Incorporation of substituted aromatic heterocycles reduces the
level of covalent binding of a molecule to microsomal proteins.69
Covalent binding
Compound Analog (pmol-equiv/mg protein)
O
7.22 3870
O F

7.23 1690
F
7.24 911
O N
Cl
7.25 303
O N
CF3
7.26 88
O N

An example to illustrate this approach was reported by Samuel and co-

workers.69 Studies with radiolabelled phenoxy analog (7.22) exhibited extensive
covalent binding (3870 pmol-equiv mg1 protein) to microsomal proteins
following incubation with NADPH supplemented human liver microsomes
(Table 7.16). Replacement of this group with a pyridinyl group resulted in a
significant decrease in covalent binding to the microsomal protein when the
analog was incubated under similar conditions as the lead. Additional decrease
in covalent binding was observed when the pyridine ring was substituted with
electron withdrawing group (CF3 group). Although not described by the
authors, the decrease in covalent binding of the analogs is probably related to
the decrease in electron density (therefore the reactivity) of the moiety in the
molecule that is susceptible to metabolic activation.
Other possible applications of heterocyclic rings have been proposed to avoid
bioactivation of compounds. For instance, metabolic activation of carboxylic
acids via formation of acyl glucuronide and subsequent acyl migration is a
matter of concern due to the potential for these metabolites to react covalently
with other biological molecules including proteins. These reactions have been
implicated in severe toxicity, resulting in the withdrawal of multiple drugs from
the marketplace in the USA and other countries. To reduce the risk of meta-
bolism related liabilities associated with acyl glucuronides, tetrazoles can be
used to replace carboxylic acid moieties.70 Though the tetrazolyl derivatives are
metabolised via glucuronidation like the carboxylic acids, the corresponding
glucuronide conjugates do not undergo acyl rearrangement to form a reactive
metabolite.
Although heterocyclic rings have been widely used to reduce undesirable
effects of a compound, these rings can also result in unwanted adverse effects
such as heptotoxicity or inactivation of P450. Compounds containing thiophenes
Influence of Heteroaromatic Rings on ADME Properties of Drugs 359

Cl Cl Cl
S O
S OH N
O O
O
O S
OH
Tienilic Acid (-)-Suprofen Ticlopidine

O
O
N N
HO NH H
O S
N
N O N
S H

Ritonavir

Figure 7.20 Structures of drugs containing thiophene and thiazole rings.

Nu O
S R S R H O
Nu S R
O O
H+ OH

Nu S R Nu S R
Cl Cl
O O
R= O O

OH

Figure 7.21 Mechanism of metabolic activation of thiophene ring containing tienilic

acid.

and furans commonly fall into this category.46,71 Several therapeutic agents with
these rings have been withdrawn from the market or have warnings on their
labels due to specific organ toxicities and related idiosyncratic reactions, and are
inactivators of drug metabolising enzymes. Tienilic acid, suprofen and ticlopi-
dine (Figure 7.20) are examples of thiophene containing compounds that have
been associated with adverse drug reactions as a consequence of metabolic
activation.
Tienilic acid was withdrawn from the market because of hepatotoxicity. It
has been proposed that tienilic acid undergoes sulfoxidation to a reactive
electrophilic, tienilic acid-S-oxide by CYP2C9, which inactivates the enzyme
(Figure 7.21).72
Tienilic acid specific autoantibodies (called anti-LKM2) directed against
CYP2C9 have been detected in patients treated with tienilic acid and suffering
360 Chapter 7
with tienilic acid-induced hepatitis. More recently, the structural analog ()-
suprofen was shown to also be a mechanism-based inactivator of P450 2C9.73
Tienilic acid and ()-suprofen are reported to cause mechanism-based inacti-
vation by a similar mechanism, leading to covalent modification of the
CYP2C9 apoprotein within the active site.74 Although ticlopidine is still utilised
clinically as an inhibitor of ADP-induced platelet aggregation, its use is asso-
ciated with a relatively high incidence of agranulocytosis, aplastic anaemia and
thrombocytopenia. The thiophene ring in ticlopidine is known to undergo
CYP2C19 and 2B6 mediated catalysis to reactive intermediate(s) that causes
enzyme inactivation.72,75 Other examples of sulfur containing heterocycles
include the potent mechanism-based inactivator of CYP3A4 by the protease
inhibitor, ritonavir (Figure 7.20).72 Ritonavir has been speculated to undergo
bioactivation of one or both of its thiazole rings and therefore result in inac-
tivation of the enzyme.
Furan-containing compounds such as furosemide, ipomeanol and L-739010
(Figure 7.22) also cause hepatic and renal necrosis in mouse and humans or
develop potentially lethal pulmonary lesions in rat.46 Oxidative ring opening of
furan and irreversible protein binding of the corresponding substituted ketoe-
nals has been reported during metabolism studies on many furan-containing
biologically active compounds (Figure 7.23).76–79
Several compounds containing furan rings are inactivators of P450
enzymes.76,77 Evidence for P450 inactivation by reactive intermediate(s) derived
from furan ring scission has also been presented for the experimental HIV

Cl O O
S NH2 O CN
OH N
O O
N
H O O
O CO2H 4-Ipomeanol
L-739010
Furosemide O

Figure 7.22 Structures of compounds containing a furan ring system.

R R R
O
O O O OH
H
Reactive
Intermediates

Covalent Binding R
To Macromolecules
O
O

Figure 7.23 Metabolic activation of ring opening of furan rings.

Influence of Heteroaromatic Rings on ADME Properties of Drugs 361

N OH O
H OH N
O N N N P450
N O N
O Inactivation
O NH O NH

L-745,394

N OH OH
H
O N N
N
O
O NH

Compound 7.27

Figure 7.24 Proposed mechanism for inactivation of CYP450 by L-745,394.

protease inhibitor L-745 394 which inactivates rat CYP2C11 and human
CYP3A4 enzymes (Figure 7.24).80,81 Removal of the furan ring or reduction of
2,3-double bond in L-754,394 led to compound 7.27 (Figure 7.24) which did
not inactivate CYP3A4, implicating that the furan ring was involved in the
bioactivation sequence leading to enzyme inactivation.82 For this reason furan,
thiophenes and thiazoles (especially the amino thiazole ring) are considered as
structural alerts and are generally excluded by the medicinal chemists when
considering the design of new drugs candidates.
Some six-membered heterocycles with substituted alkyl, amino or hydroxyl
substituents can also undergo metabolic activation. For instance, 2,3-diami-
nopyridine containing compounds (7.28) (Figure 7.25) that were designed as
bradykinin B1 receptor antagonists undergo CYP3A4-catalysed bioactivation
and covalent binding to liver microsomal proteins and glutathione.83 The
presence of two amino groups on the pyridine ring increases its susceptibility to
P450-mediated two-electron oxidation and consequent formation of the reac-
tive pyridine-2,3-diimine 7.29 (Figure 7.25). Reaction of this intermediate with
glutathione afforded the glutathione adduct 7.30. The bioactivation liability of
the 2,3-diaminopyridine moiety was addressed by replacement of the 2-amino
group on the pyridine nucleus with an oxygen atom (7.31) or addition of a
methyl group on the 2-aminopyridine (7.32) (Figure 7.25).
Compounds possessing a sterically unhindered nitrogen heterocycles are
known to act as reversible inhibitors of P450 enzymes. These compounds have
been shown to coordinate with the heme iron inside the CYP catalytic pocket.
Known as type-II ligands, these compounds can inhibit CYP by displacing a
sixth (weak) ligand, water, and stabilising the iron in its low spin state. This
spin state change is accompanied by an increase in the redox potential of P450,
which makes the P450 reduction (by NADPH P450 reductase) more difficult.
362 Chapter 7

Cl H
N
CN Cl Cl H
O N GS N
N NH CYP450 CN CN
GSH
O O
CO2CH3 N N N NH
F
Ar Ar
7.28 7.29 7.30

Cl CN
H
N

Cl O
H N NCH3
N
CN
O Ar
N O
7.32
Ar
7.31

Figure 7.25 Metabolic activation of bradykinin B1 receptor antagonist (1) containing

2,3-diaminopyridine ring.

N
N N
N Cl N N
N
O
Cl O
O O N N
Cl NC CN

Ketoconazole Anastozole
Clotrimazole

Figure 7.26 Structures of additional azole containing compounds that can inhibit
CYP450.

The formation of type-II complexes may inhibit the metabolism of co-admi-

nistered drugs and result in drug–drug interactions.
A few classic inhibitors that are known to inhibit P450 by coordinating with
the heme include azole antifungals such as clotrimazole (Figure 7.26), ketoco-
nazole (Figure 7.26), itraconazole (Table 7.14) and fluconazole (Table 7.14).84
The primary mode of action of all antifungal agents is to inhibit the fungal P450
(CYP 51s). Since the selectivity of the compounds to inhibit CYP51s over
human drug metabolising P450s is low, these compounds are associated with
significant drug interactions. Aromatase inhibitors are another class of com-
pounds that inhibit CYP450. Triazole derivatives in this class such as anastro-
zole (arimidex) (Figure 7.26) can inhibit CYP1A2, 2C9, and 3A-mediated
Influence of Heteroaromatic Rings on ADME Properties of Drugs 363
Table 7.17 Influence of structural modification of indinavir on P450 binding
spectra and CYP3A4 inhibitory potency.86

X
N OH OH
H
N N

O
O NH

Compound X P450 binding spectra IC50 (mM)

Indinavir II 0.45
N

7.33 II 0.87
N

7.34 I 15.1
N

7.35 I 8.7
N

catalytic activities.85 The strength of the bond between their heteroatomic lone
pair electrons and the prosthetic heme iron governs the inhibitory character of
these compounds in addition to hydrophobicity.
Minor structural modifications dramatically change the CYP3A4 potency of
these compounds as illustrated by some analogs of indinavir (Table 7.17).86
Although addition of a gem-dimethyl group (7.33) did not change the
inhibitory characteristics of indinavir, incorporation of a methyl group on the
carbon atom adjacent to nitrogen atom of the pyridine ring (7.33 to compounds
7.34 or 7.35) dramatically decreased CYP3A4 inhibitory potency by 410-fold
relative to indinavir (Table 7.17).86 The decrease in enzyme inhibition
was presumably due to the change in the interaction of substrate and P450
active site.
In another example, Smith and co-workers have conducted SAR relationship
studies focused on bioisosteric replacements of 2-pyridyl group (7.36) mGlu5
receptor antagonists to reduce inhibition of CYP1A2 (Table 7.18).87 Compared
to 7.36, which inhibited CYP1A2 with an IC50 of 3.8 mM, the thiazolyl deri-
vative 7.37 showed an undesired increase in CYP1A2 inhibition. On the other
hand and more importantly, CYP1A2 inhibition was greatly reduced (IC50
414 mM) with 2-imidazolyl derivative 7.38 compared with 7.36. The change in
the potency correlated with the log P value of the three compounds. Thus, the
364 Chapter 7
Table 7.18 Influence of clogP on the inhibition of CYP1A2 by mGlu5
receptor antagonists.87
Compound Structure CYP1A2 IC50 (mM) clogP

CN
7.36 N N 3.8 2.16
N
N N
S
CN
7.37 N N 1 2.31
N
N N
NH
CN
7.38 N N 414 1.63
N
N N

least lipophilic compound with an imidazole ring was the least potent while the
incorporation of a thiazole ring increased the lipophilicity of the compound and
hence its affinity for P450.
Many N-substituted azoles also have the ability to induce hepatic micro-
somal mixed-function oxidases. Some N-substituted imidazoles such as clo-
trimazole have been characterised as high magnitude inducers of rat hepatic
CYP. Although the number of clinically used drugs which induce P450 enzymes
in humans is limited, studies with nilotinib (Table 7.8) have shown that this
drug may induce CYP2B6, 2C8 and 2C9 in humans and thereby decrease the
concentrations of drugs eliminated by these enzymes.
It is now well recognised that that adverse events (either idiosyncratic toxicity
or induction) caused by drugs correlates well with the dose administered.
Reducing the dose size either by increasing the potency of the compound or by
improving its pharmacokinetic properties can potentially reduce the toxic
events of a compound. Thus the impact of the administered dose should not be
underestimated.
A classic example of the impact of dose on the toxicity of compounds can be
observed with glitazones. As described earlier, troglitazone, which is adminis-
tered at a relatively high dose was withdrawn from the market due to hepa-
totoxicity. The toxicity was attributed to the ring opening of thiazolidinone ring
in the molecule. Furthermore, this drug was associated with induction of
CYP3A4. In contrast rosiglitazone, which has the same thiazolidinone func-
tionality and is administered at a much lower dose, is devoid of hepatotoxicity;
neither does it show evidence of enzyme induction.

7.5 Summary
This chapter illustrates the influence of heteroaromatic rings on the ADME
properties of a compound. The examples reviewed here demonstrate that
Influence of Heteroaromatic Rings on ADME Properties of Drugs 365
incorporation of heteroaromatic rings in a compound can have major effect on
its disposition in the body. However, it is important to note that such a cor-
relation is not straightforward. The disposition of a compound can be quite
complex and is affected not only by addition of heterocyclic rings but other
groups and substituents on the molecule. Further, other physiological factors
(e.g. protein binding and transporters) can also affect the ADME properties of
compounds. Nevertheless, it is hoped that this overview provides a flavour
of the impact that various aromatic heterocycles have on the ADME properties
of the compound.

References
1. H. B. Broughton and I. A. Watson, J. Mol. Graphics Modelling, 2005, 23,
51.
2. S.-W. Zhao, L. Liu, Y. Fu and Q.-X. Guo, J. Phys. Org. Chem., 2005, 18,
353.
3. J. B. Sperry and D. L. Wright, Curr. Opin. Drug Discov. Dev., 2005, 8,
723.
4. R. B. Silverman, in The Organic Chemistry of Drug Design and Drug
Action, Academic Press, San Diego CA, 1992, pp. 88–95.
5. P. P. A. Humphrey, Headache, 2008, 48, 685.
6. P. Cohen, Curr. Opin. Chem. Biol., 1999, 3, 459.
7. M. E. M. Noble, J. A. Endicott and L. N. Johnson, Science, 2004, 303,
1800.
8. S. P. Davies, H. Reddy, M. Caivano and P. Cohen, Biochem. J., 2000, 351,
95.
9. Y.-Z. Shu, J. Nat. Prod., 1998, 61, 1053.
10. L. M. Lima and E. J. Barreiro, Curr. Med. Chem., 2005, 12, 23.
11. P. H. Olesen, Curr. Opin. Drug Discov. Dev., 2001, 4, 471.
12. B. C. C. Cantello, M. A. Cawthome, D. Haigh, R. M. Hindley, S. A. Smith
and P. L. Thurlby, Bio. Med. Chem. Lett., 1994, 4, 1181.
13. J. A. Joule and K. Mills, in Heterocylic Chemistry, Blackwell Publishing,
Oxford, 2000, pp. 1–15.
14. T. Eicher, S. Hauptmann and A. Speicher, in The Chemistry of Hetero-
cycles, Wiley-VCH, Weinheim, 2003, pp. 5–16.
15. T. L. Gilchrist, Heterocyclic Chemistry, Longman Scientific & Technical,
Harlow, 1985, pp. 5–30.
16. D. A. Smith, B. C. Jones and D. K. Walker, Med. Res. Rev., 1996, 16,
243.
17. O. H. Chan and B. H. Stewart, Drug Discov. Today, 1996, 1, 461.
18. I. K. Khanna, Y. Yu, R. M. Huff, R. M. Weier, X. Xu, F. J. Koszyk, P. W.
Collins, J. N. Cogburn, P. C. Isakson, C. M. Koboldt, J. L. Masferrer,
W. E. Perkins, K. Seibert, A. W. Veenhuizen, J. Yuan, D.-C. Yang and
Y. Y. Zhang, J. Med. Chem., 2000, 43, 3168.
366 Chapter 7
19. P. J. Cox, D. A. Ryan, F. J. Hollis, A.-M. Harris, A. K. Miller,
M. Vousden and H. Cowley, Drug Metab. and Dispos., 2000, 28, 772.
20. C.-M. Loi, M. Young, E. Randinitis, A. Vassos and J. R. Koup, Clin.
Pharmacokinet., 1999, 37, 91.
21. M. L. Quan, P. Y. S. Lam, Q. Han, D. J. P. Pinto, M. Y. He, R. Li, C. D.
Ellis, C. G. Clark, C. A. Teleha, J.-H. Sun, R. S. Alexander, S. Bai, J. M.
Luettgen, R. M. Knabb, P. C. Wong and R. R. Wexler, J. Med. Chem.,
2005, 48, 1729.
22. B. Peng, C. Dutreix, G. Mehring, M. J. Hayes, M. Ben-Am, M. Seiberling,
R. Pokomy, R. Capdeville and P. Lloyd, J. Clin. Pharmacol., 2004, 44, 158.
23. Tasigna (Nilotinib) product label.
24. R. J. Herr, Bioorg. Med. Chem., 2000, 10, 3379.
25. M.-W. Lo, M. R. Goldberg, J. B. McCrea, H. Lu, C. I. Furtek and T. D.
Bjornsson, Clin. Pharmacol. Ther., 1995, 58, 641.
26. P. J. Cox, B. D. Bush, P. D. Gorycki, G. Y. Kuo, D. Kenworthy, D. W.
Law, D. Murphy, P. C. Shardlow, A. Taylor, J. W. Upward, C. H.
Compton and R. D. Murdoch, Exp. Tox. Pathol., 1996, 48, 75.
27. D. Teneroa, D. Martinb, B. Ilsonb, J. Jushchyshync, S. Boikeb, D.
Lundberga, N. Zariffab, D. Boylea and D. Jorkaskyb, Biopharm.
Drug Dispos., 1998, 19, 351.
28. C. Liljebris, S. D. Larsen, D. Ogg, B. J. Palazuk and J. E. Bleasdale,
J. Med. Chem., 2002, 45, 1785.
29. J. Kelder, P. D. J. Grootenhuis, D. M. Bayada, L. P. C. Delbressine and
J.-P. Ploemen, Pharm. Res., 1999, 16, 1514.
30. K. Palm, P. Sternberg, K. Luthman and P. Artursson, Pharm. Res., 1997,
14, 568.
31. K. M. Mahar Doan, J. E. Humphreys, L. O. Webster, S. A. Wring, L. J.
Shampine, C. J. Serabjit-Singh, K. K. Adkison and J. W. Polli, J. Phar-
macol. Exp. Ther., 2002, 303, 1029.
32. S. A. Hitchcock and L. D. Pennington, J. Med. Chem., 2006, 49, 7559.
33. R. Tatsumi, M. Fujio, H. Satoh, J. Katayama, S.-I. Takanashi,
K. Hashimoto and H. Tanaka, J. Med. Chem., 2005, 48, 2678.
34. R. A. Porter, W. N. Chan, S. Coulton, A. Johns, M. S. Hadley, K. Widdowson,
J. J. C. Jerman, S. J. Brough, M. Coldwell, D. Smart, F. Jewitt, P. Jeffrey and
N. Austin, Bioorg. Med. Chem. Lett., 2001, 11, 1907.
35. J. L. Castro, R. G. Ball, H. B. Broughton, M. G. N. Russell, D. Rathbone,
A. P. Watt, R. Baker, K. L. Chapman, A. E. Fletcher, S. Patel, A. J.
Smith, G. R. Marshall, W. Ryecroft and V. G. Matassa, J. Med. Chem.,
1996, 39, 842.
36. A. B. Pinkerton, R. V. Cube, J. H. Hutchinson, J. K. James, M. F.
Gardner, H. Schaffhauser, B. A. Rowe, L. P. Daggett and J.-M. Vernier,
Bioorg. Med. Chem. Lett., 2004, 14, 5867.
37. N. Murayama, N. Imai, T. Nakane, M. Shimizu and H. Yamazaki, Bio-
chem. Pharmacol., 2007, 73, 2020.
38. R. Hyland, B. C. Jones and D. A. Smith, Drug Metab. Dispos., 2003, 31,
540.
Influence of Heteroaromatic Rings on ADME Properties of Drugs 367
39. S. K. Balani, B. H. Arison, L. Mathai, L. Kauffman, R. R. Miller,
R. A. Steams, I.-W. Chen and J. H. Lin, Drug Metab. Dispos., 1995, 23,
266.
40. M. Chiba, M. Hensleigh and J. H. Lin, Biochem. Pharmacol., 1997, 53,
1187.
41. O. Ghosheh and E. M. Hawes, Drug Metab. Dispos., 2002, 30, 1478.
42. J. Magdalou, R. Herber, R. Bidault and G. Siest, J. Pharmacol. Exp. Ther.,
1992, 260, 1166.
43. S. Kitamura, K. Sugihara and S. Ohta, Drug Metab. Pharmacokinet., 2006,
21, 83.
44. M. R. Rashidi, J. A. Smith, S. E. Clarke and C. Beedham, Drug Metab.
Dispos., 1997, 25, 805.
45. W. W. Hall and T. A. Krenitsky, Arch. Biochem. Biophys., 1986, 251,
36.
46. D. K. Dalvie, A. S. Kalgutkar, S. C. Khojasteh-Bakht, R. S. Obach and J.
P. O’Donnell, Chem. Res. Toxicol., 2002, 15, 269.
47. D. C. Hobbs, T. M. Twomey and R. F. Palmer, J. Clin. Pharmacol., 1978,
18, 402.
48. R. Griffith, in Foye’s Principles of Medicinal Chemistry, ed. D. A. Williams
and T. L. Lemke, Lippincott Williams and Wilkins, Philadelphia, 2002,
pp. 292–312.
49. J. J. Kyncl, R. C. Sonders, W. D. Sperzel, M. Winn and J. H. Seely,
Cardiovasc. Drug Rev., 1986, 4, 1.
50. K. Sugihara, S. Kitamura and K. Tatsumi, Comp. Biochem. Physiol., 1996,
24, 1996.
51. C. Beedham, J. J. Miceli and R. S. Obach, J. Clin. Psychopharmacol., 2003,
23, 229.
52. Z. Miao, A. Kamel and C. Prakash, Drug Metab. Dispos., 2005, 33,
879.
53. D. Zhang, N. Raghavan, S.-Y. Chen, H. Zhang, M. Quan, L. Lecureux,
L. M. Patrone, P. Y. S. Lam, S. J. Bonacorsi, R. M. Knabb, G. L. Skiles
and K. He, Drug Metab. Dispos., 2008, 36, 303.
54. P. Sauerberg, P. H. Olesen, S. Nielsen, S. Treppendahl, M. J. Sheardown,
T. Honor, C. H. Mitch, J. S. Ward, A. J. Pike, F. P. Bymaster, B. D.
Sawyer and H. E. Shannon, J. Med. Chem., 1992, 35, 2274.
55. P. Sauerberg, J. W. Kindtler, L. Nielsen, M. J. Sheardown and T. Honor,
J. Med.Chem., 1991, 34, 687.
56. K. W. Brammer, A. J. Coakley, S. G. Jezequel and M. H. Tarbit, Drug
Metab. Dispos., 1991, 19, 764.
57. R. A. Stearns, R. R. Miller, G. A. Doss, P. K. Chakravarty, A. Rosegay,
G. J. Gatto and G.-H. Chiu, Drug Metab. Dispos., 1992, 20, 281.
58. S. W. Huskey, R. R. Miller and S.-H. Chiu, Drug Metab. Dispos., 1993, 21,
792.
59. T. J. Chando, D. W. Everett, A. D. Kahle, A. M. Starrett, N. Vachhar-
ajani, W. C. Shyu, K. J. Kripalani and R. H. Barbhaiya, Drug Metab.
Dispos., 1998, 26, 408.
368 Chapter 7
60. K. W. Brammer, P. R. Farrow and J. K. Faulkner, Rev. Infect. Dis., 1990,
12(Suppl 3), S318.
61. T. C. Hardin, J. R. Graybill, R. Fetchick, R. Woestenborghs, M. G.
Rinaldi and J. G. Kuhn, Antimicrob. Agents Chemother., 1988, 32, 1310.
62. P. Krieter, B. Flannery, T. Musick, M. Gohdes, M. Martinho and
R. Courtney, Antimicrob. Agents Chemother., 2004, 48, 3543.
63. D. Levêque, Y. Nivoix, F. Jehl and Raoul Herbrecht, Int. J. Antimicrob.
Agents, 2006, 27, 274.
64. S. J. Roffey, S. Cole, P. Comby, D. Gibson, S. G. Jezequel, A. N. R.
Nedderman, D. A. Smith, D. K. Walker and N. Wood, Drug Metab.
Dispos., 2003, 31, 731.
65. N. A. Thornberry and A. E. Weber, Curr. Topics Med. Chem., 2007, 7, 557.
66. X.-Y. Chu, K. Bleasby, J. Yabut, X. Cai, G. H. Chan, M. J. Hafey, S. Xu,
A. J. Bergman, M. P. Braun, D. C. Dean and R. Evers, J. Pharmacol. Exp.
Ther., 2007, 321, 673.
67. W. Lee and R. B. Kim, Annu. Rev. Pharmacol. Toxicol., 2004, 44, 137.
68. I. B. Gardner, D. K. Walker, M. S. Lennard, D. A. Smith and G. T.
Tucker, Xenobiotica, 1995, 25, 185.
69. K. Samuel, W. Yin, R. A. Stearns, Y. S. Tang, A. G. Chaudhary, J. P.
Jewell, T. Lanza, L. S. Lin, W. K. Hagman, D. C. Evans and S. Kumar,
J. Mass Spectrom., 2003, 38, 211.
70. Y.-Z. Shu, B. M. Johnson and T. J. Yang, AAPS J., 2008, 10, 178.
71. A. S. Kalgutkar, I. Gardner, R. S. Obach, C. L. Shaffer, E. Callegari, K. R.
Henne, A. E. Mutlib, D. K. Dalvie, J. S. Lee Y. Nakai, J. P. O’Donnell,
J. Boer and S. P. Harriman, Curr. Drug Metab., 2005, 6, 161.
72. A. S. Kalgutkar, R. S. Obach and T. S. Maurer, Curr. Drug Metab., 2007,
8, 407.
73. J. P. O’Donnell, D. K. Dalvie, A. S. Kalgutkar and R. S. Obach, Drug
Metab. Dispos., 2003, 31, 1369.
74. J. M. Hutzler, L. M. Balogh, M. Zientek, V. Kumar and T. S. Tracy, Drug
Metab. Dispos., 2009, 37, 59.
75. E. Fontana, P. M. Dansette and S. M. Poli, Curr. Drug Metab., 2005, 6, 413.
76. T. M. Alvarez-Diez and J. Zheng, Drug Metab. Dispos., 2004, 32, 1345.
77. T. M. Alvarez-Diez and J. Zheng, Chem. Res. Toxicol., 2004, 17, 150.
78. D. P Williams, D. J Antoine, P. J. Butler, R. Jones, L. Randle, A. Payne,
M. Howard, I. Gardner, J. Blagg and B. K. Park, J. Pharmacol. Exp. Ther.,
2007, 322, 1208.
79. Y. Sahali-Sahly, S. K. Balani, J. H. Lin and T. A. Baillie, Chem. Res.
Toxicol., 1996, 9, 1007.
80. J. H. Lin, I. W. Chen, M. Chiba, J. A. Nishime and F. A. Deluna, Drug
Metab. Dispos., 2000, 28, 460.
81. J. H. Lin, M. Chiba and I. W. Chen, J. Pharmacol. Exp. Ther., 1995, 274,
264.
82. M. Chiba, J. A. Nishime and J. H. Lin, J. Pharmacol. Exp. Ther., 1995,
275, 1527.
83. J. J. Chen and K. Biswas, Prog. Med. Chem., 2008, 46, 173.
Influence of Heteroaromatic Rings on ADME Properties of Drugs 369
84. W. Zhang, Y. Ramamoorthy, T. Kilicarslan, H. Nolte, R. F. Tyndale and
E. M. Sellers, Drug Metab. Dispos., 2002, 30, 314.
85. S. W. Grimm and M. C. Dyroff, Drug Metab. Dispos., 1997, 25, 598.
86. M. Chiba, L. Jin, W. Neway, J. P. Vacca, J. R. Tata, K. Chapman and
J. H. Lin, Drug Metab. Dispos., 2001, 29, 1.
87. N. D. Smith, S. F. Poon, D. Huang, M. Green, C. King, L. Tehrani, J. R.
Roppe, J. Chung, D. P. Chapman, M. Cramera and N. D. P. Cosforda,
Bio. Med. Chem. Lett., 2004, 14, 5481.
CHAPTER 8

Peptidomimetics and Peptides as

Drugs: Motifs Incorporated to
Enhance Drug Characteristics
TRACEY BOYDEN, MARK NIOSI AND ALFIN VAZ*

Department of Pharmacokinetics Dynamics and Metabolism, Pfizer Global

Research and Development, Eastern Point Road, Groton, CT 06340, USA

8.1 Introduction
The human genome is projected to have 5000–10 000 potential drug targets, of
which it is speculated that 3000–6000 are amenable to small molecule drugs and
another 1500–3000 to biopharmaceuticals.1–3 Currently marketed drugs are
directed towards less than 500 targets encoded by the genome.3,4 Thus, it would
appear that a large pool of potential therapeutic targets remain untapped.
Much of this has to do with a significant lack in detailed understanding of the
biological roles played by most of these genome products. For some, where
roles have been established, their mechanisms of action can involve protein–
protein or peptide–protein interactions for which a detailed understanding of
the epitopes and specifics of their molecular interactions remain unknown. For
a few peptide–protein and protein–protein interactions, such molecular spe-
cifics have been resolved from crystal structures of complexes or approaches
using targeted mutagenesis to identify critical amino acids or peptide sequences
that are involved in an interaction. The interactions are generally based on
complementary structural features at the interaction site and can include

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

370
Peptidomimetics and Peptides as Drugs 371
hydrogen bonding, polar and hydrophobic effects between proximal amino
acids, and restricted conformations in the bound state. For systems where a
detailed understanding of the interactions is known, peptides or proteins as oral
pharmaceuticals have not been successfully compared with small molecule
pharmaceuticals due to their intrinsic properties which include molecular size,
absorption across biological membranes, and rapid systemic digestion by
proteolytic enzymes.
For over three decades, medicinal chemists have taken a peptidomimetic
approach to overcoming these limitations. Several definitions have been
ascribed to peptidomimetics. Giannis and Kolter5 gave a functional definition
of a peptidomimetic as: ‘a compound that, as the ligand of a receptor, can
imitate or block the biological effect of a peptide at the receptor level’. Wiley
and Rich6 defined peptidomimetics as ‘chemical structures designed to convert
the information contained in peptides into small non-peptide structures’ and
Gante7 defined a peptidomimetic as ‘a substance having a secondary structure
as well as other structural features analogous to that of the original peptide
from receptors or enzymes’. These definitions confine peptidomimetics to
design features of molecules related to optimal binding to a receptor or enzyme.
The progress of research in the area of peptidomimetics as pharmaceutical
molecules encompasses a broader definition to include physicochemical prop-
erties compatible with current understanding of the parameters that constitute
a drug molecule.
Here we define a ‘peptidomimetic’ as any small molecule whose structural
base is derived from the peptide and defined by the minimum number of
interactions that provide binding at a receptor or enzyme active site with equal
or better affinity than the original peptide sequence, effects the same physio-
logical response as the original peptide, and possesses drug-like properties for
absorption and pharmacokinetics consistent with an appropriate therapeutic
index.
In this chapter we have selected a few examples from the literature where drug
design originated from a peptidomimetic approach to yield therapeutic agents,
and examples where peptidomimetics show promise with in vitro systems but
need further testing for drug properties consistent with therapeutic agents. We
examine a few functional groups used as isosteres for the peptide bonds, amino
acid side chains and spatial interaction within binding sites, and finally the drug
properties of the resulting molecules. This chapter includes some topics, tran-
sition state analogues, which are also covered in Chapter 9, although here we
study the absorption, distribution, metabolism and excretion (ADME) prop-
erties from the viewpoint of the peptidomimetic properties of the drugs.

8.2 Peptidomimetics for Aspartic Acid Proteases

Aspartic acid proteases are particularly interesting because of the development
of marketed antiviral drugs from peptidomimetic approaches in the past two
decades and the potential for drugs to address mammalian aspartic acid
372 Chapter 8
proteases important in physiological processing of the cell. The human
immunodeficiency virus-1 (HIV-1) and hepatitis C retroviruses use virally
encoded aspartic acid proteases which digest the pol precursor proteins from
the immature virions to produce the mature viral particles which are then
released from the cell.8 The HIV-1 protease (HIV-PR) is a homo-dimer with 99
amino acids per monomer. NMR and X-ray crystal structures of HIV-PR for
the unbound and substrate-based inhibitor-bound complexes have provided
insights into the dynamics associated with catalysis and inhibition.9–11
Figure 8.1(a) shows the sequence of an HIV-PR substrate peptide with a
scissile amide bond between the two methionines. Replacing the methionine
residues with norleucine resulted in a peptidomimetic with an affinity to the
HIV-PR comparable to the substrate peptide. To produce a non-hydrolysable
peptide, methylene-amine was chosen as the isostere to replace the scissile
amide bond. The transition state for aspartic acid proteases involves a tetra-
hedral intermediate from nucleophilic addition of water to the scissile amide
bond (Figure 8.1b), catalysed by the carboxylate residues of Asp25 in the
protease sequence. Further modification of the substrate peptide by replacing
the scissile amide bond with the methylene-amine isostere (Figure 1c, boxed)
resulted in MVT-101, the first non-hydrolysable peptidomimetic inhibitor of
the HIV-PR with a Ki of about 0.8 mM which is substantially higher than the
Km for the hydrolysable substrates.12 The methylene-amine isostere provides a
tetrahedral carbon adjacent to the amide nitrogen and a molecular volume that
is smaller than either the amide bond or its hydrated transition state. However,
the methylene hydrogens lack the hydrogen bonding capability required for
stabilization of a tetrahedral geminal diol transition state. Furthermore, this
isostere has free rotation about the C–N bond which is lacking in the amide
bond, thus allowing MVT-101 greater conformational flexibility than the ori-
ginal substrate peptide. These factors may account for the lower affinity of
MVT101 for the HIV-PR than the substrate peptides.
In the ligand-free structure, two flexible b-hairpin loops between amino acids
45 and 55 of each monomer were identified in the vicinity of the active site.
These loops showed considerable motion that allowed the substrate or sub-
strate-based inhibitor access to the active site aspartates (residues 25 and 25 0 ).
Molecular dynamic simulations suggested a flexible active site with an ‘induced
fit’ that drives the catalytic power of this protease.9,11 This interpretation was
consistent with the crystal structure, which showed that the b-hairpin loops
become ordered and closed; placing residues Ile 50/50 0 over the cleavage site.12
Peptidomimetics of the pol protein sequence between Leu165 and Ile169, where
the scissile amide bond between Phe167 and Pro168 of the penta-peptide
sequence was replaced by either the methylene-amine or hydroxyethylamine
amide bond isosteres, showed good HIV-PR inhibition but failed to advance
due to poor bioavailability.13
Figure 8.2 shows the hydroxyethylamine transition state isostere for the
amide bond hydrolysis. While this isostere is larger in volume than the tran-
sition state for hydrolysis of the amide due to the methylene group interspaced
between the geminal diol and the amine in the transition state, it has a
Peptidomimetics and Peptides as Drugs 373

(a) NH2

S HN

O O O
H H H
N N N NH2
N N N
H H H
O O O O
OH

S
H2N O
P3 P2 P1 P1′ P2′ P3′

O − H
O OH H
(b)

N N
N H H
H non-hydrolyzable
hydrolysis
transition state amide bond isostere

(c) HN NH 2
Gly27′ Asp29′
Gly27 NH
Asp29 NH
Asp25 Asp25′ O N HN
O H
CO2- N
H
O
COOH −
OOC
O O
H H H
N N N NH2
N N N
H H H
O O O O
OH O NH
H
N O H2N O Gly48′
Gly48 H H
O
H H
O N N O
Ile50 Ile50′

Figure 8.1 Schematic representations of: (a) the substrate hexapeptide with the arrow
showing the scissile amide bond; (b) the transition state for hydrolysis of
the amide bond by aspartic acid proteases and the non-hydrolysable
methylene-amine amide bond isostere; and (c) the peptidomimetic
MVT101 that has the methylene-amine amide bond isostere between
residues P1 and P1 0 (boxed) bound to the enzyme, showing the flap bound
water and hydrogen bonds from protein residues to the peptide.

tetrahedral carbon with a hydroxyl group capable of hydrogen bonding

equivalent to the tetrahedral geminal diol intermediate of the transition state,
but incapable of hydrolysis. This isostere has been successfully incorporated
into the nine peptidomimetic anti HIV-PR drugs shown in Figure 8.2 and
374 Chapter 8

-
O HO O HO H
H
N
N
N H
H hydrolysis non-hydrolysable
transition state hydroxyethyl isostere

H H
O N S O N
O O
H
N N HO
N N N N
H H H H
O CONH2 OH OH
Saquanavir H Nelf inavir H

NH2 NH 2
O O O
O O O
S
S O N N O
O N N O H
H O
OH O P
O- Ca2+
O-
Amprenavir Fosamprenavir calcium salt

H NH2
O N O
O O O
O O
O S
O N NH O
O N N H
H H OH
OH
Darunavir
compound x H

N OH OH
H O
N N H
N N O
N N
O H
O NH O OH
N O
H
Indinavir Lopinavir

O O O OH O
H H H
N N N N N O
N N N O S
H H O N N
O OH H H
S N O O
Ritonavir Atazanavit

Figure 8.2 Hydroxyethyl transition state isostere for hydrolysis of an amide bond: ten
approved peptidomimetic HIV-1 protease inhibitors incorporating the
hydroxyethyl isostere.

which were approved by the US Food and Drug Administration (FDA)

between 1995 and 2006.
The common motifs among the nine peptidomimetic drugs shown in
Figure 8.2 are the hydroxyethylamine transition state isostere (dashed box)
Peptidomimetics and Peptides as Drugs 375
and a phenethyl equivalent group as an isostere of the phenylalanyl residue in
the pentapeptide (dashed circle). These two groups constitute the core phar-
macophore of this class of HIV-PR inhibitors (solid box). Except for darunavir,
the affinities of these drugs for the HIV-PR are comparable, yet their overall
structures have distinct differences at the flanks of the core pharmacophore.
The nine peptidomimetic structures show little resemblance to the original
peptide from which they were derived and contain variable numbers of amide
bonds, yet they possess affinity for the protease that is comparable or higher
than the original peptide. In this series, the hydroxyethylamine transition state
isostere appears crucial for the tight binding with the protease. Its hydroxyl
group is suggested to be hydrogen bonded to the catalytic aspartic acid residue
in the active site, mimicking the bound water molecule in the native enzyme
which serves as the nucleophile for hydrolysis of the amide bond. In saquinavir,
a bicyclic amine serves as an isostere for the prolyl residue in the peptide
sequence. In addition to providing the structural constraint that proline pro-
vides in peptides, the bicyclic amine isostere is believed to provide additional
hydrophobic interactions. A quinaldic acid serves as a hydrophobic replace-
ment for the carboxy terminal leucine.
Nelfinavir was structurally derived from saquinavir by truncating the peptide
at the N-terminus by replacing the leucine–asparagine sequence with the less
peptidic hydroxymethyl benzoyl moiety, resulting also in a decrease in mole-
cular weight and an increase in solubility. Nelfinavir showed comparable affi-
nity for HIV-PR to saquinavir. In compound-X, the leucine–asparagine
sequence was replaced by a bicyclic bis-tetrahydrofuranyl carbamate. The
compound was similar in affinity to saquinavir with the protease but with
greater solubility, suggesting that the bis-tetrahydrofuranyl isostere provided
comparable binding properties as the quinaldic acid moiety in saquinavir.
Amprenavir is structurally distinct from saquinavir and nelfinavir. It con-
tains a p-aminosulfonamide in place of the prolyl isostere of saquinavir and
nelfinavir. Additionally, the asparaginyl end was replaced by a fur-
anocarbamate. Amprenavir is significantly more soluble than nelfinavir, with
comparable affinity for the protease.
Indinavir is structurally distinct from saquinavir, nelfinavir and amprenavir
at both flanks of the core pharmacophore. An amidohydroxy-indan group
replaces the prolyl-isoleucine motif of the pentapeptide and a pyridyl-methyl-
piperazine motif replaces the asparaginyl end. These substitutions provide
indinavir with even greater solubility.
Ritonavir contains a methyl thiazole carbamate at the prolyl end and, at the
asparaginyl end, a substituted urea containing an isopropylthiazole group.
Atazanavir, approved by the FDA in 2003, contains this core pharmacophore
except that the hydroxyethyl amine is replaced with the hydroxyethyl hydrazine
isostere, which results in a decreased number of stereochemical centres in the
drug structure and increased ease of manufacture. The core pharmacophore in
this molecule is symmetrically flanked by carbamates.
Modification of amprenavir with the bicyclic bis-tetrahydrofuranyl isostere
of the asparaginyl moiety in compound X yielded darunavir which was
376 Chapter 8
approved by the FDA for therapy in 2006. Despite the very high structural
similarity of darunavir to amprenavir, it shows almost two orders of magnitude
tighter binding to the wild-type protease and a 1.5 order of magnitude tighter
binding to the multi-drug resistant mutants of the HIV-PR.14–17
Since all the peptidomimetic structures approved for drug use (except dar-
unavir) demonstrate comparable affinity for the HIV-PR, the large structural
diversity flanking the core pharmacophore in these molecules suggests a high
degree of tolerance in the active site cavity for the prolyl–isoleucyl carboxyl end
and the leucyl–asparaginyl amino end of the peptide sequence from which they
were derived. These structures vary in amide bond content from one to four,
which appears to be the only common feature to the peptide from which they
were designed. Thus, the differences in therapeutic potential observed for these
compounds must derive from differences in their physicochemical properties
which impact absorption, distribution and metabolism, and suggests room for
further enhancements based on physicochemical and drug metabolism prop-
erties of groups attached to either flank of the core pharmacophore.
From a drug design strategy point of view, it is of interest to mention
tipranavir (Figure 8.3), the only non-peptidomimetic HIV-PR inhibitor
approved by the FDA for therapy. This compound was designed by a tradi-
tional structure activity approach from 4-hydroxycoumarin and 4-hydro-
xypyrone leads identified from traditional library screening approaches.18 This
structure lacks amide bonds and has no apparent similarity to any of the
peptidommetically derived structures shown in Figure 8.2.
From a drug metabolism perspective, saquinavir shows poor bioavailability;
this may possibly be due to poor absorption resulting in low solubility and
extensive metabolism by CYP3A4. Nelfinavir, in contrast, shows greater
bioavailability than saquinavir which may be a consequence of its decreased
molecular weight and increased solubility. Nelfinavir and saquinavir have short
in-vivo half lives due to extensive metabolism by CYP3A4.
The primary metabolic pathway for these drugs involves hydroxylation at the
bicyclic prolyl isostere and at the N-t-butyl methyl carbon. It is unclear if the
metabolites are pharmacologically active and if they contribute to the overall
potency of these drugs. Amprenavir has higher solubility than either saquinavir
or nelfinavir and has higher plasma exposure with a longer half-life (10 h),
permitting twice daily dosing regimens. Amprenavir is also metabolised

OH

CF3
O O
HN
S
Tipranavir O O

Figure 8.3 Structure of the approved anti HIV-PR drug, tipranavir, designed by a
non-peptidomimetic strategy.
Peptidomimetics and Peptides as Drugs 377
primarily by CYP 3A4. Fosamprenavir is a prodrug for amprenavir with higher
solubility and is rapidly hydrolysed at the intestinal brush border to release
amprenavir, thus allowing formulation of higher strength tablets. Indinavir
shows high bioavailability and plasma exposure, probably due to the hydro-
xyindan and pyridyl-methylpiperazine motifs which offer greater solubility.
Indinavir is a substrate and inhibitor of CYP 3A4. Ritonavir is a potent CYP
3A4 inhibitor; consequently, its primary use in antiretroviral therapy has
evolved as a low dose co-drug to increase the plasma exposure of co-admi-
nistered antiretrovirals.17,19 Atazanavir has high bioavailability providing high
oral exposure and a longer in-vivo half-life that can support once daily dosing
for treatment. Darunavir is metabolised by CYP 3A4 but is a weak inhibitor of
3A4. It is co-administered with low dose ritonavir to achieve plasma exposure
necessary for therapy. The high affinity of darunavir for the HIV-PR and its
higher solubility than the other anti HIV-PR drugs allows for doses in the range
of 300–400 mg twice daily as opposed to doses of 800–1000 mg twice or thrice
daily for the other protease inhibitors.
As these anti HIV-PR drugs are generally metabolised by CYP3A4, a clear
understanding of drug interactions between co-administered antiretrovirals in
clinical treatment of HIV is critical where multi-drug therapy is considered the
standard of care. For example, darunavir/ritonavir co-administered with
saquinavir resulted in a significant decrease in steady state exposure of dar-
unavir, suggesting that these drugs should not be co-administered.20 The
development of anti HIV-PR antivirals is an excellent example of successful
application of the peptidomimetic approach to the discovery of new therapeutic
agents.
Renin is another example of an aspartic acid protease where peptidomimetic
approaches have been tried for the development of antihypertensives. However,
with renin, the peptidomimetic approach has not met with success comparable
to that with the HIV-PR. Renin cleaves angiotensinogen to the decapeptide
angiotensin I, which is further digested by a non-specific dipeptidyl carbox-
ypeptidase, agiotensin-converting enzyme (ACE), to the octapeptide angio-
tensin II which is a potent vasopressor. Renin and ACE are popular targets for
the development of antihypertensive drugs.
Figure 8.4 shows a commonly used retro inversion of an amide bond as a
strategy in peptidomimetic design to prevent protease digestion. This strategy
was used to design the three peptidomimetic inhibitors of renin shown in
Figure 8.4. Histidine is the only residue that was not changed in the inhibitors,
as this residue provides specific binding and hydrogen bonding interaction with
Ser233. The P4 prolyl residue was replaced with piperidinyl or morpholino
rings to better fit the S4 renin binding site.21 The P3 phenylalanine was replaced
by 2-methyl-naphthyl-succinic acid, with a retro inverso amide to the amino
group of piperidine or morpholine at P4 and the a-amino group of histidine.
The naphthyl aromatic ring was selected to prevent recognition of a Phe–His-
like amide bond by chymotrypsin. The P1 isoleucine was replaced with nor-
statine (3-amino-2-hydroxy propionic acid) having either an isobutyl or
methylcyclohexyl group at P3. Compound 8.3 (Figure 8.4) was found to be
378 Chapter 8

O H
N R2
R2 R1
R1 N
H O
retro inversion

O O O
H H
N N O O O
N N N H
H H N
O O N N O
N H
O OH
N
N
P4 P3 H P2 P1 P1′ 8.1 IC50 = 41 nM
Pro Phe His Ile Val N
H

O O O O O O
H H
N N
N N O N N O
H H
O O OH O O OH
N N

N 8.2 IC50 = 6.7 nM N 8.3 IC50 = 2.4 nM

H H

Figure 8.4 Retro inversion of the amide bond used in peptidomimetics to decrease
protease susceptibility of peptidomimetic inhibitors of renin derived from
the angiotensinogen pentapeptide sequence (arrow shows the scissile
amide bond).

stable in monkey liver homogenates, human plasma and chymotrypsin, and

was pharmacologically active in monkey at an oral dose of 10 mg kg1 causing
a 10–20 mm Hg fall of blood pressure over a five-hour period.22 Clearly, the
pharmacological effect for 5 h at a 10 mg kg1 dose suggests that the pharma-
cokinetic properties of the peptidomimetic are not optimal for potential ther-
apeutic use.
Much effort has been spent by several pharmaceutical companies in devel-
oping peptidomimetics for renin.23 The frustrations described by Ciba-Geigy
scientists with the peptidomimetic approach for the search of renin inhibitors
starting in the early 1980s with CGP29287, a nona-peptide peptidomimetic
(FW 1495) that had high affinity (7 nM) for renin but poor absorption and
pharmacokinetics. Its next generation peptidomimetic derivative CGP38560 in
the late 1980s, with a lower molecular weight (FW 729) and an order of
magnitude higher affinity (0.7 nM) but equally poor bioavailability and phar-
macokinetics, led the Ciba-Geigy/Novartis group to abandon the traditional
peptidomimetic approach in favour of a structure-based approach. But, as
noted by Cohen,23 the lowest energy conformational model of the peptidomi-
metic, CPG38560, docked to the homology model of renin served as the
starting point for the de novo structure-based design approach used in the
Peptidomimetics and Peptides as Drugs 379
O

O NH
H
HN NH2 N

NH N O
O OH O H
H H O
N N N
H N N N
N H H
HN O O O O
O
N
O
O O
N
CGP29287 H

H
N

N O OH
O OH H
H H H 2N N NH2
N N
S N
O H O O O
O O O
CGP60536 (Aliskiren)
O

CGP38560

Figure 8.5 Potent in vitro peptidomimetics for human renin.

development of non-peptide based structures that ultimately yielded aliskerin,

which was approved in 2007 as the first renin targeting antihypertensive drug.
Figure 8.5 shows the structures of aliskerin and the peptidomimetics that failed
as pharmaceuticals because of their ADME properties, not because of their
ability to inhibit renin.
The efforts undertaken by many pharmaceutical companies to develop renin
inhibitors since the mid-1970s reinforces the critical need to incorporate phy-
sicochemical properties, components of drug metabolism such as metabolic
stability, permeability and pharmacokinetics early in any of several strategies in
used for the development of pharmaceuticals. While aliskerin is the first com-
mercialised renin inhibitor, its human bioavailability is only 3%. This suggests
room for significant enhancements in next generation renin inhibitors based on
knowledge of aliskerin’s ADME properties in humans.

8.3 Anticancer Peptidomimetics

Mammalian farnesyl transferase plays a critical role in translocation of ras gene
products from the cytosol to the plasma membrane where they play an
important role in mitogen-activated cell proliferation. Inhibition of the path-
way is considered a target for anticancer therapy.24,25 CAAX is the signal
recognition sequence for farnesylation of these gene products where cysteine is
the fourth residue in from the carboxy terminal, the second and third residues
(AA) are typically valine or isoleucine, and X can be any amino acid but is
usually serine or methionine. The tetrapeptide CVIM is a potent in vitro
380 Chapter 8
inhibitor of farnesyltransferase, but has poor activity in cell culture due to its
lack of intracellular access.26 Crystal structure analysis of CVIM bound to the
transferase shows that the tetrapeptide is bound in an extended conformation
through hydrophobic interactions with the Ile–Met residues and a hydrogen
bond to the isoleucine carbonyl group.27 A peptidomimetic approach has been
used to develop potent inhibitors of farnesyl transferase.
As shown in Figure 8.6, when the Val-Ile residues were replaced by p-ami-
nobenzoyl spacer the peptidomimetic 8.4 showed a higher affinity than the
inhibitory peptide CVIM (IC50 340 nM). Given the extended conformation in
which the peptide is bound within the active site and the enhanced affinity of
peptidomimetic 8.4, analog 8.5 was examined and found to bind with com-
parable potency as 8.4, eliminating the chiral components of the P1 methionine.
Crystal structure analysis showed a hydrophobic space adjacent to the biphenyl
spacer. Attachment of a phenyl substituent to the amino phenyl ring of
structure 8.5 resulted in a more potent inhibitor (8.7, IC50 15 nM). Placement of
the phenyl ring on the 4-amino-benzoic acid spacer at the 3-position resulted in
the most potent peptidomimetic in this series.28,29 While these compounds show
excellent in vitro potency against farnesyl transferase and would be expected to
enter cells because of their lipophilic characteristics, lack of drug metabolism

HS
O O
H H
N N
H 2N N OH
H
O O

S
P4 P3 P2 P1
Cys Val Ile Met
CVIM IC50 340 nM

O S
HS HN HS HN
HN
H 2N O OH H2 N OH
O O
8.4 IC50 150 nM 8.5

O S
HS HN HS HN
HN
H 2N OH H 2N OH
8.6 IC50 0.6 nM O 8.7 IC50 15 nM O

Figure 8.6 Peptidomimetics of farnesyltransferase.

Peptidomimetics and Peptides as Drugs 381
information makes judgment of these structures as potential therapeutic agents
impossible.
Cancer cells are compromised in their ability to undergo apoptosis;30 many
cancer cell lines and tumors isolated from patients have high levels of expres-
sion levels of the X-linked inhibition of apoptosis protein (XIAP).31 XIAP
functions by binding and inhibiting three members of the caspase family of
enzymes.32 Mitochondrial proteins, Smac/DIABLO, released into the cytosol
in response to an apoptotic stimulus such as TNFa, bind to XIAP and caspase-
9 reversing the inhibition by XIAP of caspases.33–36 As XIAP blocks apoptosis
at the convergence of multiple signalling pathways, it is particularly attractive
as a target for the design of drugs aimed at reversing the apoptotic resistance of
tumor cells. Crystal and NMR solution structures reveal that the N-terminal
tetrapeptide (AlaValProIle) of Smac binds to a surface groove on XIAP.
Figure 8.7 shows a schematic representation of the interaction between the
tetrapeptide with residues on the BIR3 domain of the XIAP protein.
As with peptides in general, Smac based peptides show good in vitro binding
but have poor cell permeation. However, when tethered to a carrier peptide,
they were shown to sensitise tumor cells establishing the validity of the target
for cancer therapy.37–39 The peptidomimetics shown in Figure 8.6 show strong
but comparable binding to the XIAP protein and represent the early stage in
the discovery of novel drugs for targeted chemotherapy similar to the early
stages of anti HIV-protease anti retroviral peptidomimetics. Lessons from the
failure to develop peptidomimetic inhibitor drugs for renin should help guide
approaches to the development of a new generation of anticancer drugs
directed at these novel targets.

8.3.1 Summary
As shown by the examples above, peptidomimetics hold promise as another
means of developing novel therapeutics based on a clear understanding of the
biochemical pathway and the mechanisms underlying peptide–protein or pro-
tein–protein interactions. As was shown for the development of HIV-PR
inhibitors, the structures of the final drug products had little in common
structurally with the peptide from which they were derived, and with exception
of the core pharmacophore, have little in common with each other. Further-
more, a potent antiretroviral, tirpanavir, was developed by a traditional
iterative structure–activity–drug properties screening approach from lead
matter identified in high throughput screening, without structural knowledge of
the binding to the active site but a robust assay for activity. At the core of
discovering new chemical matter for therapeutic use is a clear understanding of
the biochemical basis for target selection, clear design strategies for new che-
mical matter with high potency and early evaluation of drug properties for
absorption, pharmacokinetics, and metabolism, which lead to the selection of
optimal structures to progress through the drug discovery pipeline.
382 Chapter 8

2 4 Trp
Trp3
23 Tyr3 3 23

small
hydrophobic O N
H
pocket N
N
H O
O
Gl NH 2 CONH 2
n3
19
08 06
Thr3 y3
Glu314 Gl
P1 P2 P3 P4
Ala Val Pro Leu

O
O O
HN N HN
N N
N
H H
O N O NH
O H
O
8.8 Kd = 12 nM 8.9 Kd = 5 nM

O O
O O
HN HN N
N
N N
H H
O N O N
O H O H
9.0 Kd = 12 nM 9.1 Kd = 12 nM

Figure 8.7 Schematic representation of the N-terminal tetrapeptide (AlaValProLeu)

of Smac bound in the BIR3 binding groove of XIAP showing the inter-
actions of peptide backbone and side chains with amino acids of the XIAP
protein and four potent Smac peptidomimetics with dissociation
constants.

8.4 Peptide Drugs

8.4.1 Insulin and Insulin Analogs
The discovery of insulin in the first quarter of the 20th century40 has led to the
treatment of diabetes mellitus for over 70 years, with early treatments using
partially purified extracts from porcine or bovine spleen that had highly vari-
able efficacy. With progress in purification of animal spleen-derived insulin,
efforts focused on enhancing the action duration of insulin preparations.
Treatment initially involved subcutaneous injections of crystalline insulin just
prior to meals and typically showed an action onset between 0.5 and 1.0 hour,
with peak activity between 2 and 3 h and between 6 and 8 h of duration.
Peptidomimetics and Peptides as Drugs 383
Early efforts to increase the action duration were achieved by a complex
between insulin and protamine, with further enhancements achieved by adding
zinc ions to stabilize the protamine–insulin complex (PZ insulin) and further
modifications to produce a neutral protamine insulin complex termed NPH
insulin.41 Proteolysis of protamine in the complex at the deposition site resulted
in the slower release of insulin and a doubling of the duration of insulin’s
action. Another form of extending the action duration of insulin was by the
addition of excess variable amounts of zinc ions, which produced amorphous
crystalline suspensions called lente insulins.42
In the 1980s, through recombinant technology human insulin replaced
bovine and porcine pancreatic insulin. Crystalline insulin, PZ-insulin, NPH-
insulin and lente insulins were the main therapeutic forms of insulin up to the
1990s. In the 1990s, two major clinical studies, the UK Prospective Diabetes
Study Group (UKPDS33) and the Diabetes Control and Complications Trial
Research Group, established a clear link between glycemic control and
microvascular complications in diabetics.43,44 These findings, along with the
observation of the basal and postprandial secretion pattern of insulin in normal
subjects,45 suggested that the goal of insulin therapy should be to mimic the
rhythmic pattern observed for insulin secretion.46 The finding that pancreatic
insulin secretion was directly into the hepatic portal circulation and knowledge
of the structure and physicochemical properties of native insulin made it evi-
dent that subcutaneous deposition of insulin was not ideal to mimic the
observed physiological secretion pattern of insulin. Efforts were directed
towards developing short acting analogs to mimic postprandial secretion and
long-acting analogs to mimic basal secretion.
Figure 8.8 shows the sequence of human insulin. Soluble human insulin is a
hexameric aggregate which dissociates slowly to monomers within the

S S

G IV E QC CT SIC SL YQ LE NY CN
S S

S S

FV NQ HL C GSH LV E AL YL V CG E RG FFYT PKT

Short acting insulin analogs

Insulin lispro: B chain P28 K29 inversion to KP
Insulin aspart: B chain P28 replaced by D
Insulin glulisine: B chain N3 replaced with K;K29 replaced with E

Long acting insulin analogs

Insulin glargine: A chain N21 replaced with G, RR added to position 30 in the B chain
Insulin determir: B chain T30 removed; K29 acylated with myristic acid

Figure 8.8 Structure of human insulin and the structure modifications to create short-
and long-acting analogs.
384 Chapter 8
subcutaneous tissue with the resultant slow appearance in blood. Structural
characterization of human insulin showed that residues 26–30 (YTPKT) were
not critical for binding of insulin to the insulin receptor, but were critical for
aggregation.47 Inversion of proline–lysine at position 28–29 in the B chain of
human insulin results in an analog (insulin lispro) with comparable insulin
receptor affinity but decreased self association. This in turn results in faster
dissociation of aggregates into monomers and consequently faster absorption
from the subcutaneous deposition site, leading to an earlier and greater peak
serum level for a shorter duration of time compared with unmodified human
insulin. Pharmacokinetics of insulin lispro shows an onset of activity between
0.2 and 0.5 h, with maximal activity between 0.5 and 2 h and a duration of 3–
4 h.48,49 Insulin lispro was the first insulin analog approved for human use in
1996.
Insulin aspart was derived from human insulin by substituting proline 28 in
the B chain with aspartic acid. Like insulin lispro, this analog results in weak
dimeric and hexameric aggregation leading to comparable rapid absorption
into serum after subcutaneous injection. The pharmacokinetic behaviour of
insulin aspart is similar to insulin lispro50 and it was approved for human use
in 2000.
Insulin glulisine is the third approved rapid onset insulin analog approved for
human use. Insulin glulisine was obtained by substituting B chain lysine 29 and
aspargine 3 with glutamic acid and lysine, respectively. The pharmacokinetic
properties of insulin glulisine are comparable with those of the other two fast
acting insulin analogs, with rapid onset of activity within 20 minutes and
maximal activity in 1.5 h.51,52
The pharmacological properties of the insulins lispro and aspart are com-
parable to human insulin in terms of insulin receptor affinity, receptor off rate,
metabolic potency, insulin like growth factor-1 receptor affinity, and mitogenic
potency. Insulin glulisine has somewhat lower affinity for the human insulin
receptor and is significantly weaker in affinity to the insulin-like growth factor-1
receptor and mitogenic potency.53,54 In general, the three fast-acting insulin
analogs are very similar in pharmacokinetic and physiological function and
more closely mimic the normal physiological insulin response following pre-
prandial subcutaneous administration when compared to subcutaneously
administered human insulin.
Long-acting insulin analogs have been developed by shifting the isoelectric
point towards neutral pH and by increasing the hydrophobicity by covalent
modification with a fatty acid. Insulin glargine was developed by replacing
asparagine 21 in the A chain with glycine and adding two arginine residues at
position 30 in the B chain. These amino acid substitutions result in an iso-
electric point shift from pH 5.4 to 6.7.55 This analog is injected subcutaneously
as an acidic solution at pH 4.0 and forms a precipitate. The precipitated insulin
glargine dissolves slowly and is absorbed into the serum without any significant
peak. It has a duration of action of about 20 h at physiological doses and is
significantly longer acting than NPH insulin.56 Insulin glargine (lantus) was the
first long-acting insulin analog to be approved for human use.
Peptidomimetics and Peptides as Drugs 385
Another long-acting insulin analog approved for human use is insulin
determir. This analog is formed by removing threonine 30 in the B chain and
acylating the lysine at position 29 in the B chain with myristic acid. Insulin
determir forms soluble hexameric and di-hexameric complexes at the site of
injection via aggregation of the fatty acid chains. These self-associated com-
plexes are thought to equilibrate with serum albumin. The equilibrium between
self-association and albumin binding are thought to be the reason for the long
depot residence time. The binding to albumin is also thought to slow its dis-
tribution to peripheral tissues resulting in an increased duration of activity.57
Binding of insulin determir to serum albumin does not appear to compete with
other albumin bound compounds or to show kinetic differences in low albumin
states, suggesting that albumin may serve as a reservoir for insulin determir
forming a buffer for any rapid changes in insulin absorption.58,59
Long-acting insulin analogs show comparable reductions in HbA1c to NPH
insulin.60,61 However, these analogs show a significantly lower risks of noc-
turnal hypoglycemia.62–64 Combination therapies with long- and short-acting
insulin analogs, as well as long-acting insulin analogs and oral antidiabetic
drugs are being considered as therapeutic options to achieve the rhythm of
insulin observed in normal humans.

8.4.2 Incretin Hormones

Plasma insulin levels are substantially increased upon oral glucose adminis-
tration compared with intravenously administered glucose. This effect is termed
the incretin effect65 and is caused by the release of peptides from endocrine cells
in response to ingestion of food. Several such incretin peptides have been
identified and their plasma levels are known to be differentially affected by
nutrients in the digestive track. These include:

 glucagon-like peptide-1 (GLP-1)—this is released by intestinal L-cells,

plasma levels of which remain elevated for a while after cessation of
feeding;
 amylin—a 37 amino acid polypeptide co-secreted with insulin by pan-
creatic b-cells;
 cholecystokinin (CCK)—a peptide released by I-cells in the upper intest-
inal track in response to the presence of digestive products of fats and
proteins;
 peptide YY—like GLP-1 this is released by intestinal L cells slowly during
meals and remains elevated in plasma for several hours;
 ghrelin—produced by gastric endocrine cells is a 28 amino acid peptide
that serves as a signal of energy depletion and acts to initiate food intake.

GLP-1 and amylin analogs have become therapeutic agents whereas CCK,
PYY and ghrelin receptors are still investigational targets for therapy.
386 Chapter 8
Most of the incretin effect on pancreatic function is derived from
GLP-1.66 Glucose homeostasis in the postprandial period is regulated by
GLP-1 through several mechanisms that include inhibition of glucagon secre-
tion, increase of insulin synthesis, delay in gastric emptying and promotion of
satiety.67 Type 2 diabetic patients have a decreased incretin effect due to a
decrease in secretion of GLP-1, which results in lowered insulin secretion
and disruption of glucose homeostasis.68 Administration of GLP-1 to type 2
diabetic patients results in an increase in insulin secretion and lowering of
postprandial blood glucose.69 GLP-1 is, however, rapidly degraded by dipep-
tidyl peptidase 4 (DPP-4), a protease that is widely distributed in multiple cell
types including the capillary bed of the gut mucosa and so rapidly inactivates
GLP-1.
Exendin-4 (Bayetta, Exenatide) a 39 amino acid peptide extracted from the
venom of the Gila monster (Heloderma suspectum) with a high affinity for
GLP-1 receptors and has 53% homology with GLP-1. Exenatide is the
first GLP-1 receptor agonist approved for adjunctive therapy for type 2
diabetic patients receiving sulfonyl ureas or methformin treatment but who
have not achieved adequate control. Clinical trials have demonstrated that
exenatide improves glycemic control when combined with methformin and
sulfonylureas. It may be an alternative to insulin glargine for patients who
require non-traditional therapy.
Amylin’s site of action is the area posterma—a hind brain circumventricular
organ with a porous blood–brain barrier. Peripheral administration of amylin
analogs have been shown to reduce food intake in non human primates.70
Pramlintide is a human amylin analog that has been approved for human
therapy co-administered with mealtime insulin for patients who have not
achieved glucose control. Clinical studies have demonstrated reductions in
HbA1c and body weight.71
Most currently used peptide-based drugs are currently administered par-
enterally by subcutaneous or intramuscular injection. This is a major limitation
to the administration of such medications. Efforts to develop inhaled forms of
such drugs have met with limited success. Peptides and peptide-based drugs
show no oral bioavailability due to gastric digestion as well as poor absorption
across the intestinal wall.
Some efforts to develop methods for oral delivery of such drugs including
protein drugs have focused on the dietary uptake mechanism of vitamin B12.72
In this approach, vitamin B12 is either conjugated directly with the protein–
peptide of interest for oral administration or is encapsulated in a vitamin B12
coated dextran nanocapsule. Success has been achieved using this technology
for some proteins and peptides including insulin and erythropoietin.73,74 The
development of such approaches including exploring other carrier-mediated
transport processes across the intestinal wall needs further research and tech-
nological innovations to circumvent the oral gastric barrier when delivering
peptide and protein drugs. Such developments will dramatically enhance
the use and development of peptide-based drug molecules based on protein–
protein or protein–peptide interactions.
Peptidomimetics and Peptides as Drugs 387
References
1. J. Drews, Science, 2000, 287, 1960.
2. A. L. Hopkins and C. R. Groom, Nat. Rev. Drug Discov., 2002, 1, 727.
3. P. Imming, Nat. Rev. Drug Discov., 2007, 5, 821.
4. J. Drews and S. Ryser, Nat. Biotechnol., 1997, 15, 1318.
5. A. Giannis and T. Kolter, Angew. Chem., Int. Ed., 1993, 32, 1244.
6. R. A. Wiley and D. H. Rich, Med. Res. Rev., 1993, 13, 327.
7. J. Gante, Angew. Chem., Int. Ed., 1994, 33, 1699.
8. A. Frankel and J. A. T. Young, Annu. Rev. Biochem., 1998, 67, 1.
9. R. Ishima, D. Freedberg, Y.-X. Wang, J. M. Louis and D. A. Torchia,
Structure, 1997, 7, 1047.
10. G. Lange-Savage, H. Berchtold, A. Liesum, K.-H. Budt, A. Peyman, J.
Knolle, J. Sedlacelk, M. Fabry and R. Hilgenfeld, Eur. J. Biochem., 1997,
249, 912.
11. S. Piana, P. Carloni and M. Paminello, J. Mol. Biol., 2002, 319, 567.
12. M. Miller, J. Schneider, B. K. Sathyanarayana, M. V. Toth, G. R. Mar-
shall, L. Clawson, L. Selk, S. B. H. Kent and A. Slodawer, Science, 1989,
246, 1149.
13. N. A. Roberts, J. A. Martin, D. Kinchington, A. V. Broadhurst, J. C.
Craig, I. B. Duncan, S. A. Galpin, B. K. Handa, J. Kay, A. Krohn, R. W.
Lambert, J. H. Merrett, J. S. Mills, K. E. B. Parkes, S. Redshaw, A. J.
Ritchie, D. L. Taylor, G. J. Thomas and P. J. Machin, Science, 1990, 248,
358.
14. N. M. King, M. Prabu-Jeyabalan, E. A. Nalivaika, P. Wigerinck, M.-P. de
Bethune and C. A. Schiffer, J. Virol., 2004, 78, 12012.
15. D. L. Surleraux, A. Tahri, W. G. Verschueren, G. M. E. Pille, H. A. de
Kock, T. H. M. Jonckers, A. Peeters, S. De Meyer, H. Azijn, R. Pauwels,
M.-P. de Bethune, N. M. King, M. Prabu-Jeyabalan, C. A. Schiffer and P.
Wigerinck, J. Med. Chem, 2005, 48, 1813.
16. S. De Meyer, H. Azijn, D. Surleraux, D. Jochmans, A. Tahri, R. Pauwels,
P. Wigerinck and M.-P. de Bethune, Antimicrob. Agents Chemother., 2005,
49, 2314.
17. M. Rittweger and K. Arastéh, Clin. Pharmacokinet., 2007, 46, 739.
18. S. M. Poppe, D. E. Slade, K.-T. Chong, R. R. Hinshaw, P. J. Pagano, M.
Markowitz, D. D. Ho, H. Mo, R. R. Gorman III, T. J. Dueweke, S.
Thaisrivongs and W. G. Tarpley, Antimicrob. Agents Chemother., 1997, 41,
1058.
19. F. M. Uckun and O. J. D’Cruz, Expert Opin. Ther. Pat., 2006, 16, 1354.
20. V. Sekar, E. Lefebvre, K. Marien, M. De Pauw, T. Vangeneugden and R.
M. W. Hoetelmans, Ther. Drug Monit., 2007, 29, 795.
21. R. Kato, O. Takahashi, Y. Kiso, I. Moriguchi and S. Hirono, Chem.
Pharm. Bull., 1994, 42, 176.
22. K. Iizuka, T. Kamijo, H. Harada, K. Akahane, T. Kubota, H. Umeyama,
T. Ishida and Y. Kiso, J. Med. Chem., 1990, 33, 2707.
23. N. C. Cohen, Chem. Biol. Drug Des., 2007, 70, 557.
388 Chapter 8
24. M. Barbacid, Ann. Rev. Biochem., 1987, 56, 779.
25. R. J. A. Grand and D. Owen, Biochem. J., 1991, 279, 609.
26. Y. Reiss, J. L. Goldstein, M. C. Seabra, P. J. Casey and M. S. Brown, Cell,
1990, 62, 81.
27. C. L. Strickland, W. T. Windsor, R. Syto, L. Wang, R. Bond, Z. Wu, J.
Schwartz, H. V. Le, L. S. Beese and P. C. Weber, Biochemistry, 1998, 37,
16601.
28. Y. Qian, M. A. Blaskovich, M. Saleem, C. M. Seong, S. P. Wathen, A. D.
Hamilton and S. M. Sebti, J. Biol. Chem., 1994, 269, 12410.
29. M. A. Kothare, J. Ohkanda, J. W. Lockman, Y. Qian, M. A. Blaskovich,
S. M. Sebtib and A. D. Hamilton, Tetrahedron, 2000, 56, 9833.
30. P. T. Lansbury, Nat. Rev. Neurosci., 2004, 5, S51.
31. I. Tamm, S. M. Kornblau, H. Segall, S. Krajewski, K. Welsh, S. Kitada, D.
A. Scudiero, G. Tudor, Y. H. Qui, A. Monks, M. Andreeff and J. C. Reed,
Clin. Cancer Res., 2000, 6, 1796.
32. E. N. Shiozaki and Y. Shi, Trends Biochem. Sci., 2004, 29, 486.
33. C. Du, M. Fang, Y. Li and X. Wang, Cell, 2000, 102, 33.
34. A. M. Verhagen, P. G. Ekert, M. Pakusch, J. Silke, L. M. Connolly, G. E.
Reid, R. L. Moritz, R. J. Simpson and D. L. Vauz, Cell, 2000, 102, 43.
35. G. Wu, J. Chai, T. L. Suber, J. Wu, C. Du, X. Wang and Y. Shi, Nature,
2000, 408, 1008.
36. Z. Liu, C. Sun, E. T. Olejniczak, R. P. Meadows, S. F. Betz, T. Oost, J.
Herrmann, J. C. Wu and S. W. Fesik, Nature, 2000, 408, 1004.
37. C. R. Arnt, M. V. Chiorean, M. P. Helderbrant, G. J. Gores and S. H.
Kaufmann, J. Biol. Chem., 2002, 277, 44236.
38. L. Yang, T. Mashima, S. Sato, M. Mochizuki, H. Sakamoto, T. Yamon, T.
Oh-Hara and T. Tsuruo, Cancer Res, 2003, 63, 831.
39. S. Fulda, W. Wick, M. Weller and K. M. Debalin, Nat. Med. (N. Y.),
2002, 8, 808.
40. F. G. Banting, C. H. Best, J. B. Collip, W. R. Campbell and A. A. Fletcher,
Can. Med. Assoc. J., 1922, 12, 141.
41. H. C. Hagedorn, B. N. Jensen, N. B. Krarup and I. Wodstrug, J. Am. Med.
Assoc., 1936, 106.
42. K. Hallas-Moller, Diabetes, 1956, 5, 7.
43. Diabetes Control and Complications Trial Research Group, N. Eng. J.
Med., 1993, 329, 977.
44. UK Prospective Diabetes Study Group, Lancet, 1998; 352, 837.
45. Y. T. Kruszynska, P. D. Home, I. Hanning and K. G. M. M. Alberti,
Diabetologia, 1987, 30, 16.
46. R. H. Roscamp and G. Park, Diabetes Care, 1999, 22, S2B109.
47. Z. Vajo, J. Fawcett and W. Duckworth, Endocr. Rev., 1994, 22, 706.
48. D. C. Howey, R. R. Bowsher, R. L. Brunelle and J. R. Woodworth,
Diabetes, 1994, 43, 396.
49. E. Torlone, C. Fanelli, A. M. Rambotti, G. Kassi, F. Modarelli, A. Di
Vincenzo, L. Epifano, M. Ciofetta, S. Pampanelli and P. Brunetti, Dia-
betologia, 1994, 37, 713.
Peptidomimetics and Peptides as Drugs 389
50. S. R. Mudaliar, F. A. Lindberg, M. Joyce, P. Beerdsen, P. Strange, A. Lin
and R. R. Henry, Diabetes Care, 1999, 22, 1501.
51. S. K. Garg, S. L. Ellis and H. Ulrich, Expert Opin. Pharmacother., 2005, 6,
643.
52. R. H. Becker, A. D. Frick, F. Burger, H. Scholtz and J. H. Potgieter, Exp.
Clin. Endocrinol. Diabetes, 2005, 113, 292.
53. P. Kurtzhals, L. Schaffer, A. Sorensen, C. Kristensen, I. Jonassen, C.
Schmid and T. Trub, Diabetes, 2000, 49, 999.
54. G. Seipke and I. Stammberger, Diabetes, 2005, 54(Suppl. 1), A339.
55. G. B. Bolli and D. R. Owens, Lancet, 2000, 356, 443.
56. T. Heise, S. Bott, K. Rave, A. Dressler, R. Rosskamp and L. Heinemann,
Diabetes Med., 2002, 19, 490.
57. S. Havelund, A. Plum, U. Ribel, I. Jonassen, A. Volund, J. Markussen and
P. Kurtzhals, Pharm. Res., 2004, 21, 1498.
58. P. Kurtzhals, S. Havelund, I. Jonassen and J. Markussen, J. Pharm. Sci.,
1997, 86, 1365.
59. P. Kurtzhals, Int. J. Obes., 2004, 28(Suppl. 2), S23.
60. K. Hermansen, M. Davies, T. Derezinski, R. G. Martinez, P. Clauson and
P. Home, Diabetes Care, 2006, 29, 1269.
61. T. Haak, A. Tiengo, E. Draeger, M. Suntum and W. Waldhausl, Diabetes
Obes. Metab., 2005, 7, 56.
62. S. R. Heller, S. A. Amiel and P. Mansell, Diabetes Care, 1999, 22, 1607.
63. S. R. Heller, S. Colagiuri, S. Vaaler, B. H. R. Wolffenbuttel,
K. Koelendorf, H. H. Friberg, K. Windfeld and A. Lindholm, Diabet.
Med., 2004, 21, 769.
64. D. Russell-Jones, H. Kim, S. Heller and P. Clauson, Diabetologia, 2005, 48,
A92.
65. M. A. Nauck, E. Homberger, E. G. Siegel, R. C. Allen, R. P. Eaton, R.
Ebert and W. Creutzfeldt, J. Clin. Endocrinol. Metab., 1986, 63, 492.
66. A. Barnett, Int. J. Clin. Pract., 2006, 60, 1454.
67. D. J. Drucker and M. A. Nauck, Lancet, 2006, 368, 1696.
68. I. Idris and R. Donnelly, Diabet. Obes. Metab., 2007, 9, 153.
69. C. F. Deacon, Diabetes, 2004, 53, 2181.
70. N. T. Bello, M. H. Kemm and T. H. Moran, Am. J. Physiol. Regul. Integr.
Comp. Physiol., 2008, 295, R76.
71. D. Singh-Franco, G. Robles and D. Gazze, Clin. Ther., 2007, 29, 535.
72. A. K. Petrus, T. J. Fairchild and R. P. Doyle, Angew. Chem., Int. Ed., 2009,
48, 1022.
73. G. J. Russell-Jones, S. W. Westwood and A. D. Habberfield, Bioconjug.
Chem., 1995, 6, 459.
74. K. B. Chalasani, G. J. Russell-Jones, A. K. Jain, P. V. Diwan and S. K.
Jain, J. Control. Release, 2007, 122, 141.
CHAPTER 9

Pharmacokinetics and
Metabolism of Compounds that
Mimic Enzyme Transition States
IAIN GARDNER,a CHRIS BARBER,b MARTIN
HOWARD,d AARTI SAWANTc AND KENNY WATSONa
a
Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and
Development, Sandwich, Kent, CT13 9NJ, UK; b Worldwide Medicinal
Chemistry, Pfizer Global Research and Development, Sandwich, Kent, CT13
9NJ, UK; c Pharmacokinetics, Dynamics and Metabolism, Pfizer Global
Research and Development, Eastern Point Road, Groton, CT 06340, USA;
d
Drug Metabolism and Pharmacokinetics Discovery Research,
Pharmaceuticals Division F. Hoffmann-La Roche LTD, CH-4070 Basel,
Switzerland

9.1 Enzyme Transition States

Enzymes are catalysts that function by enhancing the rate of covalent bond
formation and breaking. Inhibition of enzymes has been a fruitful source of
drugs effective in treating disease, with 300 drugs targeting more than 70 dif-
ferent enzymes approved for therapeutic use in humans.1
A central tenant of the catalysis of biological reactions by enzymes is that the
enzyme–substrate complex exists in equilibrium with a higher energy transition
state which sits at the top of an energy maximum where the chances of the
enzyme–substrate complex reverting to substrates or proceeding to products

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

390
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 391

ES*

Free
Energy
(ΔG°)

E+S

ES

E+P
EP

Figure 9.1 Graphical representation of energy changes along the reaction coordinate
as enzyme (E) binds to substrate (S) forming an enzyme-substrate complex
(ES), followed by formation of a high energy transition state (ES*) with
eventual formation of an enzyme–product complex (EP) and release of
product from the enzyme (E+P).

are equal (Figure 9.1). In reality, enzymatic transition states are dynamic with
lifetimes (a fraction of a picosecond) defying direct physical observation.2
The catalytic power of enzymes can be viewed to lie in the very high affinity
that enzymes have for the transition state relative to the affinity with which they
bind substrate in the ground state.3 As an example, calf intestinal adenosine
deaminase has a dissociation constant for the transition state approximately
1012 times smaller than it has for the substrate adenosine.
Transition state analogue inhibitors (TSAIs) are a class of competitive enzyme
inhibitors that are designed to take advantage of this very high affinity interaction
between an enzyme and the transition state of the reaction it catalyses. By
mimicking the chemically unstable transition state, whilst being chemically stable
themselves, TSAIs can bind to the enzyme with much higher affinity than sub-
strate.4–8 Although in most cases substrates and transition state analogues bind in
a similar fashion, the transition state analogue can not be converted to a product
and as a result is a potent enzyme inhibitor. The theoretical potency of a perfect
transition state analogue inhibitor can be calculated as shown in eqn (9.1):6

Theoretical maximum Ki ¼ substrate Km=catalytic rate enhancement ð9:1Þ

As catalytic rate enhancements by enzymes are typically of the order of 1010–15

and can be as high as 1019, then a perfect transition state analogue inhibitor can
bind to an enzyme with a dissociation constant of approximately 1022 M.8
As it is not possible to exactly replicate the bond angles, bond distances and
electronic distribution of a transition state in a stable molecule, it is unusual to
see TSAIs with potency increases of 1015 or more compared to substrate.
However, even picking up some of the structural features of the transition state
392 Chapter 9
can yield potent inhibitor compounds that combine sufficient chemical stability
to be therapeutically useful as enzyme inhibitors.
To prove that an enzyme inhibitor is acting as a transition state analogue
inhibitor and not simply as a potent competitive or covalent inhibitor is not
trivial because binding kinetics alone (including slow inhibition and/or offset)
are not diagnostic for a compound being a transition state analogue inhibitor.2
Experimental techniques to confirm that a compound is a transition state
analogue inhibitor include:9–14

(1) Deriving the free energy relationship for the interaction between inhi-
bitor and enzyme
(2) Kinetic isotope experiments where individual atoms close to the bonds
being broken are replaced with isotopes (e.g. H is replaced with D or T,
and 13C replaces 12C and 15N replaces 14N) and the rate of catalysis
determined
(3) X-ray structural analysis
(4) NMR.

The potential to inhibit enzymes with high affinity and specificity using
transition state analogue inhibitor approaches has attracted much attention
and TSAIs have been described for all major classes of enzymes and for
more than 130 individual enzymes.7,15 However, the number of TSAIs that
have been successfully tested in humans as potential therapeutic agents is
relatively small.
In this chapter we discuss the impact of the structural features needed to
mimic the transition state of an enzymatic reaction on the disposition of drugs
in the body and some of the strategies used to successfully develop TSAIs as
therapeutic agents. In considering the effects of chemical structure and physi-
cochemical properties on oral bioavailability (F), we have used eqns (9.2) and
(9.3) to separate the effects on absorption and first pass clearance/metabolism:

F ¼ Fa  Fg  Fh ð9:2Þ

where:

F ¼ bioavailability
Fa ¼ fraction absorbed
Fg ¼ fraction escaping gastrointestinal tract first-pass metabolism
Fh ¼ fraction escaping first-pass clearance in the liver.

Fh ¼ ½1  ðCLh =QÞ ð9:3Þ

where:

CLh ¼ hepatic blood clearance

Q ¼ hepatic blood flow.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 393
1 1
Q is assumed to be 90, 70, 40, 40 and 20 ml min kg in mouse, rat, dog,
monkey and human, respectively. Unfortunately, Fg is unknown for many
compounds and in these cases we have assumed Fg to be 1.
As the majority of pharmacokinetic studies in the literature report plasma
(rather than blood) clearance data, we have converted the plasma clearance to
blood clearance by dividing it by the blood : plasma ratio (BP). Wherever
possible, we have used measured values of BP; if an assumed value is used this is
noted.

9.2 Physicochemical Properties of Transition State

Analogues
The atoms that mimic the bond(s) breaking/forming in the transition state of
the enzymatic reaction actually represent only a small fraction of those present
in TSAIs. The rest of the transition state analogue inhibitor molecule functions
as a scaffold to position the transition state mimicking group correctly within
the active site of the target enzyme and makes interactions with the amino acid
residues present in the active site of the enzyme to increase the potency and
specificity with which the TSAI binds the target enzyme. In this way, selective
inhibition of the target enzyme can be achieved without affecting the function
of closely related enzymes. In addition, the scaffold of the TSAI can be used to
modify the pharmacokinetics and metabolism of transition state analogue
inhibitor compounds.
The general physicochemical properties of TSAIs such as molecular weight
(MW) and lipophilicity are largely reflective of the scaffold of the transition
state analogue inhibitor and vary considerably depending on the lipophilicity
and size of the active site of the enzyme target. The TSAIs discussed in this
chapter have molecular weights ranging from 228 (zebularine) to 1500 (CGP-
29827), with clogP values spanning a 16 log unit range (from -5.6 for zanamivir
to +11 for BAY-793) and encompass acidic, basic and neutral compounds.
The one physicochemical property shared by TSAIs is a relatively high polar
surface area (PSA). Figures 9.2–9.4 show the calculated PSA vs. molecular
weight, the hydrogen bonding count normalised by molecular weight vs. PSA
and the PSA vs. clogP, respectively, for the TSAIs discussed in this chapter
together with the same data for a number of approved small molecule human
drugs.
Only 14% of the approved drug dataset have a PSA 4140 Å2 (a value often
used as a cut-off for predicting good oral absorption); this value increases to
52% when TSAIs are considered. That TSAIs have a high PSA is not unex-
pected since the majority of transition states involve the intermediacy of high
energy, highly polar, charged species such as a tetrahedral intermediate fol-
lowing nucleophilic attack on an amide bond. By mimicking these polar, short-
lived transition states, TSAIs can achieve great potency, but the cost is that
TSAIs have high PSAs that can limit the passive membrane permeability and
hence the oral absorption of the compounds.
394 Chapter 9

Figure 9.2 Molecular weight vs. polar surface area (PSA) for human drugs (triangles)
and the transition state analogues discussed in this chapter (squares).

This apparent conundrum has been tackled by a number of strategies:

 the use of prodrugs;

 the intermediacy of hemi-acetals or internal hydrogen bonding networks
that mask polarity and allow better lipid solubility;
 the use of functional groups that only form the polar transition state
analogue following attack by the enzyme (e.g. some boronic acids).

Another strategy utilised to increase absorption has been to increase clogP

(and, in turn, the non-polar surface area of the molecules) to try and overcome
the high polar surface area of TSAIs and achieve sufficient oral exposure for the
compounds to have therapeutic efficacy. While this strategy has improved the
absorption of some transition state analogue inhibitor compounds, it also tends
to result in rapid metabolism and high clearance of the compound; so although
the compound’s absorption is improved, the bioavailability is often still low due
to high first-pass clearance. Achieving a balance of permeability and absorp-
tion, whilst maintaining low enough clearance to attain oral exposure of the
drug, has been a major challenge for a number of classes of TSAI.
The volume of distribution at steady state (Vss) of the TSAIs discussed in this
chapter is shown in Figure 9.5. As with other drugs, lipophilicity (which largely
reflects the property of the scaffold of the molecule) and charge (which in some
cases is provided by the transition state mimicking group) are the two main
factors influencing Vss. Acid compounds generally have low Vss even at high
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 395

0.06

0.05
(HBA+HBD)/mw

0.04

0.03

0.02

0.01

0 100 200 300 400 500 600

PSA

Figure 9.3 Sum of hydrogen bond acceptors (HBA) and donors (HBD) divided by
molecular weight vs. PSA for human drugs (white triangles) and TSAI
discussed in this chapter (black squares).

600

500

400
PSA

300

200

100

0
-8 -6 -4 -2 0 2 4 6 8
CLogP

Figure 9.4 Polar surface area (PSA) vs. clogP for human drugs (white triangles) and
TSAI discussed in this chapter (black squares).
396 Chapter 9

20

10
8
6
4
Vss (I/kg)

1
0.8
0.6
0.4

0.2

-6 -4 -2 0 2 4 6 8
clogP

Figure 9.5 Vss (L kg1) of transition state analogue inhibitors vs. clogP. Triangles
represent human data and squares rat data. Basic compounds are light
grey, neutral compounds (basic pKa o5; acid pKa 48) are dark grey and
acidic compounds are coloured black.

clogP values, whilst the basic and neutral TSAIs tend to have larger Vss than
acidic compounds. The basic compounds show a trend for increased Vss with
an increase in clogP whilst the trend is not as obvious for neutral compounds.
The other thing to note from Figure 9.5 is that there are relatively few acidic
TSAI compounds.
The widely differing physicochemical properties of transition state analogues
present vastly different challenges to the drug discovery scientist depending on
the physicochemical properties of the TSAI needed to potently inhibit the
enzyme target. These are illustrated in the next section where specific enzyme
targets that have been successfully targeted by TSAI therapeutics are
considered.

9.3 ADME Properties of Transition State Analogue

Inhibitors against Different Enzyme Targets
9.3.1 Proteases
Human proteases cleave peptide (amide) bonds of polypeptides with specificity
that is controlled by the accommodation of specific peptide side chain residues
into the pockets neighbouring the active site.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 397
Asp
O His Asp O
His
O H NH2
O H N
N +H2O 1′
O N R
N Ser
N Ser
H H
+
O1 1 1
O R O O O R O R O
N N HO
N N N N N
1′
R
1′
O R O O

Figure 9.6 The first step in the catalytic mechanism is attack of the active site serine
onto the substrate peptide’s scissile bond, forming a tetrahedral inter-
mediate which is stabilised by further interactions by the protease prior to
hydrolysis to release amine and acid products.

The mechanism of proteolysis by serine proteases relies upon a catalytic triad of

amino acids comprising aspartic acid, histidine and serine. Figure 9.6 shows the
initial steps of this process—the deprotonation of the hydroxyl of serine by his-
tidine’s imidazole assisted by a neighbouring aspartic acid. The resultant
nucleophile attacks the carbonyl of the scissile bond to form a tetrahedral tran-
sition state which subsequently collapses following hydrolysis by water. An
equivalent mechanism using a nucleophilic thiol from cysteine is exploited by
cysteine proteases, whilst aspartyl proteases and metalloproteases can use a
nucleophilic water molecule to generate the tetrahedral transition state.16 The
protease enzymes lower the energy required for amide hydrolysis by stabilising the
(high energy) tetrahedral transition state. Figure 9.7 shows a number of chemo-
types that have been used to mimic the tetrahedral transition state of proteases in
TSAI. Each of these chemical classes is discussed in the following sections.

9.3.1.1 Alcohol Containing TSAI of Proteases

A number of aspartyl protease enzymes (including HIV protease and renin) are
important therapeutic targets and have been targeted with alcohol (and enol)
containing TSAIs.16,17 The first step in the catalytic mechanism and the pro-
posed tetrahedral transition state of aspartyl proteases,18 along with the dif-
ferent alcohol (and enol) groups that have been used to mimic the transition
state, are shown in Figure 9.8. An alternative catalytic mechanism has also been
proposed for HIV protease whereby cleavage occurs in a one-step process in
which nucleophile (water) and electrophile (acidic proton) attack the scissile
bond in a concerted manner.19
The substrates of HIV protease (GP-160, gag, gag-pol) and renin (angio-
tensinogen) are fairly large polypeptides; consequently TSAI of these enzymes
also tend to be large and have high clogP (average 5.1) in order to make enough
interactions with the enzyme to effectively compete with substrate (see Table 9.1
for structures, physicochemical properties and pharmacokinetic data). This
398 Chapter 9

Transition state
Classification Drug precursor
intermediate

R R
alcohol R′ R′
OH OH
Enz

O
aldehyde R′ R′

OH O
Enz

O
R R
activated R′ R′ R = CF3, het, carbonyl,
ketone O phosphonate
OH
Enz

O OH
boronate R′ B OH R′ B
OH OH

OH OH
phosphonic P R
acid R′ P R
O R′
O

Figure 9.7 Tetrahedral transition state mimetics used in protease inhibitors

(het ¼ heterocycle).

high lipophilicity together with the high PSA (average 174 Å2) and high
hydrogen bond acceptor (HBA) and donor (HBD) count has a large impact on
the pharmacokinetics of the compounds. So whilst these TSAIs are highly
potent inhibitors of their enzyme target (e.g. the renin inhibitors discussed have
Ki in the 0.7–14 nM range and the HIV protease inhibitors have Ki’s of
o10 nM),20 achieving potency and oral bioavailability in the same molecule is
often challenging.
The pharmacokinetic behaviour of the peptidic renin inhibitors such as
remikiren is fairly typical for compounds with these physicochemical proper-
ties. Although oral administration of the peptidic renin inhibitors to pre-clinical
species often resulted in pharmacological efficacy, oral F% was usually low
requiring high doses to be used, plasma levels were often highly variable (for
instance the plasma levels of remikiren varied up to 100-fold in rats given the
same dose) and the concentration time profiles often showed secondary max-
ima peaks or were an irregular shape. In addition, there is evidence that the
renin inhibitor compounds could inhibit their own metabolism after oral
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 399

R N NH2
R N
R1 R1 1′
O -O OH R

H O H
O H H O Asp O HO Asp +
Asp Asp
HO R
O O O O
O

Initial catalytic step transition state intermediate products

OH
R R1 O
R R
R1 R R1
OH O OH R1
OH O O
R2

Figure 9.8 Proposed catalytic mechanism and transition state for aspartyl proteases
where the active site aspartic acids and substrate (RCONR1) are shown,
together with alcohol and enol groups that have been used to inhibit
aspartyl proteases.

dosing.21–23 The human pharmacokinetics of remikiren24 were characterised by

high blood CL (1500 ml min1; 21 ml min1 kg1) that approximates to liver
blood flow. Consequently when remikiren was dosed orally in humans, plasma
concentrations were highly variable and bioavailability was very low (o1%)
and variable.25 Ultimately, none of the peptidic renin inhibitor compounds had
a sufficient balance of potency and absorption, distribution, metabolism and
excretion (ADME) properties to be developed as a drug. As the chemical
synthesis of many of the peptidic renin transition state analogue inhibitor
compounds was complicated, the low bioavailability and subsequent high doses
needed for efficacy meant that the cost of these compounds also became too
high to be commercially attractive.
The three properties that need to be balanced in order to achieve high oral
bioavailability for alcohol containing TSAIs are solubility (as only drug in
solution can be absorbed), permeability and first-pass clearance. The interplay
of these three factors can be seen with the HIV protease inhibitors. The solu-
bility of HIV protease inhibitors is low, leading to erratic absorption (often
with multiple absorption peaks being observed) and high variability after oral
dosing. Other consequences of the low solubility are that the absorption of the
compounds is heavily influenced by whether subjects are fed or fasted and on
the particular drug formulation administered. The effects of food on the HIV
protease inhibitors vary from compound to compound with atazanavir, dar-
unavir, saquinavir (F% increases 18-fold), lopinavir, nelfinavir and tipranivir
showing an increased AUC in the fed state (presumably due to increased
solubility in the presence of the gastric secretions and increased gastric motility
of the fed state),248 while the AUC of indinavir and amprenavir are decreased
 400
Table 9.1 Chemical structure, physicochemical properties and clearance, bioavailability and absorption estimates of alcohol (and
enol) containing TSAI.
a
TPSA
HBA
Compound Structure Target MW clogP HBD pKa Vss (L/kg) PPB (%) CLp (ml/min/kg) F% Fa (%)

Amprenavir HIV 506 3.3 140 Neutral 6–10 h 90–96 h B11–20 h (CL/F) 490 h
Refs. 20, 33, 270 R protease 3 (V/F)
N
O 4
O N N
O S
O O
O

OH
R = H amprenavir R= P O fosamprenavir
OH

Atazanavir O HIV 705 5.2 171 Neutral 86h 6–8 h (CL/F) 480 h
H
Refs 26, 271 O N protease 13 (4.7 B)
N
5
O N N

O HO
NH
O N
O

Darunavir HIV 548 2.9 149 Neutral 1.3 h 95 h 8h 37 h 490 h

27, 264, 272 O O protease 4 1.4 with 82 with
O O O 4 ritonavir ritonavir
S
O N N
H
OH
NH2

Indinavir Chiral HIV 614 3.7 118 Weak 2.2 r 70 r* 107 r 16–24 r 81 r
Refs. 31, 52, protease 7 base 0.7 d 70 d* 16 d 16 d (pH 6.5) 100 d
246, 263, N N O 4 (5.2) 1.5 mk 70 m* 36 mk 72 d (pH 2.5) 100 mk
273–275 N O 0.8 h 64h 10–18 h 19 mk 480 h
N

O
60–65 h
N O
Chapter 9
KNI-272 Chiral HIV 668 4.7 201 Weak base 5–21 r 98–99 h 18–48 r 12 hp 32–100 h
Refs. 276–278 protease 9 (4.9) 0.9 mk 42 mk 16–55 h
N
4 0.11 hp 11 hp
10–11 h
O

O N
O

S N

O
O O
N
N
S
Lopinavir HIV 629 6.1 120 Neutral 98–99 h 1.4–1.7 h 25 r 470 h
Ref 30 protease 5 (CL/F in
O O 4 presence of
N N N ritonavir)
O N
O O

Nelfinavir ABS HIV 568 5.8 127 Basic(7.5) 3–7 r 98–99 h 27–60 r 30–43 r 100 r
O
Refs 82, 279, H protease 6 1.7 – 4 d 22–60 d 29–43 r 100 d
280 N 4 2.7 mk 20–28 mk 30–59 d 63 mk
H 10 h (V/F) 10–12 h (CL/F) 9–42 mk 480 h
N
H H
N
HO OH
O
S

Ritonavir Chiral HIV 720 5.2 202 Neutral 1r 97–99 r 11–17 r 71–78 r 470 h
Refs 34, protease 11 0.3 d 99 d 4–5 d 21–100 d
281–284 S O O N 4 2.2 mk 96–99 mk 11–12 m 30–70 mk
N N N O 0.4–0.5 h 99–99.5 h 1–2.1 h (CL/F)
N N S
O O (V/F)

Saquinavir O HIV 670 4.7 167 Basic (7.6) 4–10 h 98–99 h 13–17 h 4 h hgc B30 hge
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme

H2N
Refs 34, 51, 246 protease 7 12 h sgc B100 sgc
O OH
H 6
N N N
N
H N
O
O H

SDZ-PRI-053 HIV 723 5.6 158 Weak base 2.7 mo 19 m 47–61 m 60–77 m
Ref 285 O O O protease 7 (6) 1.6 r 34 r 22–54 r 43–100 r
N 6 9.5 d (CL/F)
O N N
O N O
401

O
 402
Table 9.1 (continued )
a
TPSA
HBA
Compound Structure Target MW clogP HBD pKa Vss (L/kg) PPB (%) CLp (ml/min/kg) F% Fa (%)

Tipranavir F HIV 603 7.8 114 Acid (4.5) 0.5 r 4 99.9 20 m 11 m 53–60 r
F
Refs 29, O F protease 5 0.13 d in all 2.8 r 28–30 r 418 h
286–288 N 2 0.11–0.14 h species 1.3 d 0.2–0.3 h 10 rab
S N
O O (V/F with (CL/F with 7–8d
O O ritonavir) ritonavir)

A-74273 Renin 787 5.5 139 Basic(7.5) 12–20 d (CL/F) 26 r

Ref. 289 O O O 7 31 f
O
N N N N 3 54 d
O O O O 1.9 mk

Aliskiren Chiral Renin 552 3.5 146 Basic (9.8) 2 h 50 h B2 h B2.6 h B5 h

Refs. 36, O 7
53, 290 6

O
O

N
O

N
O N
O

BW-175 Renin 688 6.6 154 Basic (6.8) 1.5 r 22 r 3–10 r o15 c r
Ref. 23 O O
N
S N N
O O
O O O

CGP-29827 O Renin 1495 0.3 528 Neutral

ON 20
N
N N N 20
N O O O
O
N N N N N N O
N
N N O O N O
OO
O OO N
N
Chapter 9

N N
CGP-38560 N Renin 730 5.0 179 Base (6.7) o2 h
Ref. 291 N 7
O O 5
N N
S N
O O O O

Enalkiren O Renin 657 3.7 192 Basic (8.9) 0.2 h 94 h Not Only phar-
Ref. 291 9 measurable macologi-
8 cally active
O O O by IV route
N
N N N
O O
N

N
Remikiren Renin 631 3.6 170 Basic(6.7) 1.8 m 86–96 h 119 r 3r B30 b r
Ref. 22, 24, 7 4.4 r 48 ma 6 ma B56 b d
291,292 O 5 1.5 d 25 d 1.1 d o2 d mk
O O H OH
S N
2.7 ma 20 mk o0.3 mk
N 0.9 h 13 h o2 h
H
O OH
N

N
H
YM-21095 Renin 732 3.5 206 Basic (6.7) 3.9 d 22 r 3–10 r o15 c r
Ref. 293 9 37 d 0.2 d
4
O O N N
H
N N
N N S N
H
O O OH

N NH
Zankiren S Renin 706 5.6 189 Basic (6.2) 148 r 24 r 100b r
N
28 f 32 f
N O OH 10 d 53 d 72 c d
N N 27 mk 8 mk 25 c mk
S N
O O O OH
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme

d ¼ dog, f ¼ ferret, h ¼ human, hp ¼ human paediatric population, ma ¼ marmoset, mk ¼ monkey, mo ¼ mouse, r ¼ rat, rab ¼ rabbit, hgc ¼ hard gelatin capsule,
sgc ¼ soft gelatin capsule
a
Topological polar surface area
b
Based on urinary and biliary recovery of radioactivity
c
Assuming blood : plasma ratio ¼ 1
d
Assuming blood : plasma ratio ¼ 0.6 (as in human)
403

* ¼ blood binding.
404 Chapter 9
26–30
by food. The requirement for some compounds to be given with food and
others to fasted subjects complicates the usage of these compounds in the
multiple drug regimes needed to treat HIV infection. In addition to food, the
absorption of some compounds is very dependent on intestinal pH. For
instance the low solubility of atazanavir at high pH means that this compound
can not be co-administered with antacids or proton pump inhibitors that raise
the pH of the gastrointestinal tract26 and similar pH dependent effects were
seen when indinavir was dosed orally in the dog.31
The commercial formulations of the protease inhibitors often require organic
excipients to maintain the compounds in solution and to achieve sufficient
exposure after oral dosing; the oral solution formulation of lopinavir contains
40% v/v ethanol, tipranavir is formulated as a fatty emulsion29 and the com-
mercial dosage form of amprenavir requires the use of organic excipients
(including PEG 400 and propylene glycol) and a low drug loading (150 mg),
meaning that pill burden for the patient was high (eight capsules have to be
taken twice a day).32
To try and overcome the problems caused by in solubility, the use of more
soluble prodrugs has been investigated.32 The alcohol group that mimics the
tetrahedral transition state in amprenavir has been conjugated with phosphate
to give a prodrug (fosamprenavir) that is 10-fold more soluble (0.4 mg ml1)
than parent amprenavir (Table 9.1). The advantage of this increased solubility
was that fosamprenavir could be formulated at a higher drug loading (700 mg
tablet), reducing the pill burden for the patient (two capsules dosed once or
twice a day). Once dosed orally, fosamprenavir is converted almost entirely to
amprenavir at or near the intestinal epithelium by alkaline phosphatase in both
animals and humans; the amprenavir is then absorbed and equivalent systemic
exposures are achievable using fosamprenavir and the standard amprenavir
capsules.33 In contrast to amprenavir, fosamprenavir pharmacokinetics are
unaffected by food and this together with the lower pill burden mean that
fosamprenavir has been commercially successful.32 Another advantage for
fosamprenavir is that it should be possible to combine fosamprenavir in a single
tablet with other classes of HIV drugs; the organic excipients and low drug
loading meant it was not easy to do this with amprenavir.
Saquinavir also has low solubility and formulation changes have been used
to increase its bioavailability. Originally marketed as a hard gelatine capsule
(HGC) formulation, saquinavir had low bioavailability (F% 4) partly as a
result of incomplete absorption. When saquinavir was reformulated in a soft
gelatine capsule as a lipid formulation containing mono- and di-glycerol
medium chain fatty acids (SGC) the F% increased 3–4 fold (12–16%) and was
further increased in the presence of food.34 The SGC formulation presumably
increases the solubility and dispersion of saquinavir and in some ways mimics
the effect of food on the compound. However, it is also possible that the
excipients have other effects (e.g. changing intestinal leakiness, inhibiting P-gp)
that contribute to the higher F%. Some evidence to support this theory is that,
when boosted with ritonavir, the plasma levels from the HGC formulation are
higher than those with the SGC formulation;35 this is hard to explain if the
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 405
difference between formulations is only due to solubility effects. Adding a basic
centre to molecules is another strategy that has been employed to try and
increase the solubility and hence absorption of HIV protease inhibitors (the
discovery of indinavir is discussed in Chapter 4).
Once in solution the absorption of compounds depends on the passive per-
meability across the intestine and the action of drug transporters that can act to
increase (influx) or decrease (efflux) the rate and extent of drug passage across
the intestine. Ideally absorption of compounds is estimated from a comparison
of intravenous and oral pharmacokinetics of the compound. Many of the
compounds in Table 9.1 are too lipophilic to formulate in a vehicle that can be
administered intravenously. For these compounds, an estimate of human
absorption has been made by assuming that the compounds are not metabo-
lised in the luminal contents of the gastrointestinal tract and calculating the Fa
from the recovery of unchanged drug in faeces after oral dosing. Although this
may underestimate absorption (as material absorbed and excreted unchanged
in bile or by exsorption of unchanged drug back into the intestine is considered
not to have been absorbed), it allows the absorption of more of the compounds
in Table 9.1 to be compared.
PSA is one factor known to limit passive membrane permeability and hence
oral absorption of compounds. A plot of Fa (%) against PSA for the com-
pounds in Table 9.1 is shown in Figure 9.9. Clearly there are compounds with
high PSA and low absorption in this dataset. For example aliskiren (the only
marketed renin inhibitor) is highly soluble (4350 mg ml1) at acidic pH but has
high molecular weight (552) and PSA (146 Å2). Despite low clearance in
humans (9 L hr1; B2 ml min1 kg1), the compound has low oral

100

80

60
FA

40

20

120 140 160 180 200

PSA

Figure 9.9 Absorption (FA as percentage) of alcohol TSAI compounds in human

(black) and rat (white) vs. PSA (larger symbols indicate larger clogP).
406 Chapter 9
36,37
bioavailability (B2.6%) and absorption (B5%). Following a 300 mg oral
dose of 14C-labelled aliskiren, the majority (91%) of the dose was eliminated in
the faeces (77.5% as unchanged drug reflecting elimination of unabsorbed
drug).37
But there are also compounds in this dataset with high PSA and high
absorption in both rats and humans. This is true even for compounds (ataza-
navir, KNI-272 and ritonavir) with very high PSA (4170 Å2). Where these
compounds do have low bioavailability, it is due more to high metabolic
clearance than to poor absorption per se. The number of HIV protease inhi-
bitors that can be administered effectively in combination with ritonavir (see
later) also suggests that they have sufficient intrinsic membrane permeability to
be orally absorbed when the contribution of P-gp and CYP 3A to first-pass
clearance of the compounds is inhibited. One possible explanation for the high
absorption of compounds despite high PSA is that the molecules are large
enough to adopt conformations where intramolecular hydrogen bonds form,
effectively hiding some of the polarity of the molecule whilst exposing the non-
polar groups of the molecule to the lipophilic membrane. This would act to
minimise the free energy of solvation and facilitate the passive membrane
permeability of the compounds.38
Small changes in structure and physicochemistry can also tip the balance in
favour of oral exposure. Enalkiren and zankiren are structurally related renin
inhibitors. The compounds have similar molecular weight (706 vs. 657) and
PSA (192 vs. 189 Å2) but zankiren is more lipophilic (clogP 5.6 vs. 3.7; log D7.4
4.6 vs. 2.6) and has lower numbers of HBA (six vs. nine) and HBD (four vs.
eight) groups. Whilst enalkiren has no oral bioavailability and only showed
pharmacological efficacy after intravenous dosing, zankiren has high oral
bioavailability and absorption in pre-clinical species.39
P-glycoprotein (P-gp) is expressed in the intestine and acts to efflux com-
pounds from the intestinal cells back into the gastrointestinal tract contents and
to limit absorption/bioavailability of compounds (particularly at low doses).
Structural features that typically result in high affinity for P-gp are clogP 42.9,
an 18 atom or longer molecular axis, a basic charge (although many neutral
compounds are also P-gp substrates) and a molecular weight o800. In addi-
tion, most P-gp substrates possess two or three electron donor groups with a
fixed spatial separation of 2.5–4.6 Å. Compounds with more of these elements
tend to have higher affinity for the transporter.40–41,260
The alcohol transition state analogue inhibitor compounds contain many of
these structural and physicochemical features (Table 9.1) and many are sub-
strates for P-gp. Some of the compounds also interact with other intestinal
efflux transporters that may limit the absorption/bioavailability of the com-
pounds. For example, lopinavir is a substrate for MRP-2 and saquinavir is
known to inhibit BCRP (IC50 75 mM).
The effect of P-gp on disposition of HIV protease inhibitors has been studied
in vitro and in vivo. Efflux of these compounds (basolateral - apical transport
4 than apical - basolateral transport) is seen in Caco-2 cell monolayers, in
cell lines over-expressing human P-gp, and in human and rat intestine tissue
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 407
42
mounted in Ussing chambers. The efflux ratios in these systems can be
reduced with P-gp inhibitors (e.g. GF-120918, cyclosporin) and often saturate
as substrate concentration is increased.31,43 In addition the plasma levels of
saquinavir (1–1.5 fold increase), indinavir and nelfinavir are all higher fol-
lowing oral dosing in P-gp knockout mice than they are in wild-type mice.42,43
P-gp also plays a major role in limiting saquinavir absorption in the rat in
vivo.261
Even though many compounds are shown to be substrates in in vitro P-gp
expressing cell line experiments, the degree of efflux and the rate of apical to
basolateral flux vary considerably showing that this parameter is also influ-
enced by the passive membrane permeability of the compound.44 This becomes
important in vivo because many compounds identified as P-gp substrates
in vitro are still well absorbed in humans.262
Aliskiren is an example of a compound where human absorption in vivo
appears to be influenced by efflux transporters. The AUC of aliskiren increases
in a supra-proportional manner with increasing oral doses suggesting satura-
tion of a gut efflux mechanism, and compounds known to inhibit P-gp increase
the oral exposure of aliskiren in humans in vivo (AUC is increased 50, 80 and
500% by atorvastatin, ketoconazole and cyclosporine, respectively),36,45
Although these two findings are suggestive of a role for P-gp in the disposition
of aliskiren, the inhibitors that increase aliskiren AUC are known to inhibit
other ADME enzymes/transporters (e.g. CYP 3A4, BCRP OATPs) and these
other enzymes/transporters might also play a role in the non-linear disposition
of aliskiren.
Clearly other compounds that are P-gp substrates in vitro are still well
absorbed in vivo. This probably reflects a high enough intrinsic membrane
permeability of some compounds, allowing them to overcome the effects of
P-gp, and also the fact that many of the compounds are given at high doses and
the concentrations achieved in vivo saturate P-gp and limit its impact.
Many of the compounds that are P-gp substrates are also substrates of CYP
3A4 (and 3A5); P-gp and CYP 3A in the intestine act in consort to limit the oral
bioavailability of compounds that are substrates for both entities. P-gp slows
down the absorption of compounds and the repeated absorption and efflux
cycles increase the likelihood of metabolism by the intestinal CYP 3A isozymes
resulting in high first-pass intestinal extraction.46 Because of their affinity for P-
gp and CYP 3A, many of the compounds shown in Table 9.1 have the potential
to undergo significant intestinal first-pass metabolism (Fg {1). Although it is
not easy to accurately estimate Fg for many of the compounds in Table 9.1,
there are enough data to do this for some of them.
Pre-clinical studies have been conducted with indinavir dosed intravenously,
orally and via infusion into the hepatic portal vein.31 Based on data in the rat,
hepatic extraction was 70% (Fh ¼ 0.3) and intestinal first-pass extraction was
o10% (Fg 4 0.9). Studies in the dog showed that the intestine had minimal
contribution to saquinavir metabolism after intravenous or oral dosing and
that F% was low (8–20%), mainly due to high hepatic extraction (Fh ¼ 0.14–
0.25).249
408 Chapter 9
Obtaining estimates of human intestinal first-pass extraction is more diffi-
cult,47,48 but methods to estimate this parameter from in vitro data have been
used with some success.49 Another method that can be used, although it is not
without caveats, is to compare the effects of grapefruit juice (a relatively
selective inhibitor of intestinal CYP 3A) on the pharmacokinetics of the
compound in question after intravenous and oral dosing.48 Using these
methods led to estimates of human intestinal first-pass extraction of 0.3–0.5 for
saquinavir (Fg ¼ 0.5–0.7) and 0.05 for indinavir (Fg40.9).31,43,48–51
The high clogP of the alcohol containing TSAI of proteases often results in
rapid hepatic clearance (low values of Fh) and low bioavailability (Table 9.1).
As would be expected from their physicochemical properties, the compounds
tend to be cleared by metabolism or biliary excretion in the liver, with only
indinavir having appreciable renal clearance (12% of total clearance) which
slightly exceeds GFR, suggesting an active tubular secretion component.52 The
actual enzymes and transporters responsible for the clearance of alcohol pro-
tease TSAI vary from compound to compound and in many cases their identity
is not known. Early renin inhibitors (e.g. CPG-29827) were peptidic in nature
and rapidly hydrolysed by proteases, while later renin inhibitors were meta-
bolised by CYP isozymes and underwent biliary excretion. When alcohol TSAI
do undergo metabolism, it generally does not involve the atoms that mimic the
transition state of the enzyme but more usually reflects metabolism of the
lipophilic scaffold that holds the transition state mimic in place.
Compounds that undergo significant biliary excretion include remikiren
(B30% biliary excretion of unchanged drug in rats and dogs) and CGP-38560,
a compound that had poor oral absorption and rapid biliary clearance (half-life
B7 min) in humans,53 although no information on the transporters involved is
published. As well as parent drug, an additional 60% of a remikiren dose was
excreted in to bile of rats (18 radioactive peaks identified) and dogs as meta-
bolites.54 The AUC of radiolabelled material was lower in the bile duct can-
nulated dog (about 30%) than in intact animals, suggesting that the
radiolabelled material also undergoes significant enterohepatic circulation.
Aliskiren is less lipophilic and more basic than many of the other renin
inhibitors and is known to be a substrate for at least one member of the organic
anion polypeptide (OATP) family in vitro; it is likely that aliskiren undergoes
active uptake in to the liver in vivo. Following uptake into the liver, aliskiren is
slowly metabolised by CYP 3A4.45
The HIV protease inhibitors are generally metabolised by CYP 3A4 in
humans though, for some compounds, other isozymes are involved (Table 9.2).
The HIV protease inhibitors also interact with a number of uptake transporters
(OATP family) in both rats and humans. In vitro data suggest that transporters
control the clearance (and subsequent metabolism) of saquinavir, nelfinavir
and ritonavir in rat hepatocytes and by extension are likely to play a similar
role in vivo.55
The influence of these transporters on human pharmacokinetics is less clear
but indinavir, nelfinavir, ritonavir and saquinavir all inhibit uptake of estra-
diol-17b-glucuronide by human OATP 1B156–58and there is some evidence that
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 409
Table 9.2 Effect of ritonavir co-administration on plasma AUC of HIV
protease inhibitors.
Enzymes involved
in human Ritonavir dose Fold-increase in
Compound metabolism (mg, frequency) AUC Reference

Amprenavir CYP 3A4 100, BD 2–4 63–65

200, QD 4
300, BD 3.3
Atazanavir CYP 3A4 100, QD 3-4 26,66–68
100, BD 5.9 (Cmin)
2.7 (Cmin)
Darunavir CYP 3A4 100, BD 6.6 69
Indinavir CYP 3A4 100, BD 1.6–2.4 70–74
200, BD 2.8
300, BD 3.3–5
400, BD 5.5
Lopinavir CYP 3A4 50 mg single 77 30
dose
Nelfinavir CYP 3A4 (52%) 100, BD 1.3 75
also 2C19, 2C9, 200, BD 1.5
2D6
Saquinavir CYP 3A4 100, single dose 30 76–79
200, BD 18
300, single dose 58
300, BD 22
400, BD 22
600, single dose 112
Tipranavir 100, BD 4.6–5.5 80,81
200, BD 12.4–15.7

the HIV protease inhibitors can also be transported by human OATPs. To date
no information has been published showing transport by the major human liver
OATPs (1B1, 1B3 and 2B1),59 but saquinavir is transported by OATP 1A2
which is present in the central nervous system (CNS) and cholangiocytes in the
liver and HIV protease inhibitors also appear to be substrates of OATP 3A1
and 4A1.60
A number of HIV protease inhibitors including saquinavir (F% 4–12),
indinavir (F% 60–65) and nelfinavir (F% 45 fasted, 30% fed) have higher oral
F% than would be expected if all the systemic clearance of the compounds
occurred only in the liver, so it is possible that other organs are also involved in
the systemic clearance of these compounds. The low and variable F% and short
half-lives (mainly due to high clearance rather than low volume) of many of the
HIV protease inhibitors means that they need to be dosed two or three times a
day at high doses to have therapeutic efficacy. As well as being CYP 3A4
substrates, most of the HIV proteases are also potent inhibitors and often
inducers of the enzyme and, at the high doses used, these compounds inhibit
and/or induce their own metabolism (resulting in non-linear pharmacokinetics
with increasing dose and time) and alterations in the pharmacokinetics of many
410 Chapter 9
61,263
other compounds co-administered with them. Because of its potent CYP
3A inhibition, ritonavir has been used at low doses to enhance (boost) the
pharmacokinetics of other protease inhibitors allowing a lower pill burden and
less frequent administration of the inhibited compound in HIV patients.
This is currently the main clinical use of ritonavir as, at the high doses needed
for the compound to exert its antiviral effect, ritonavir has an unacceptable
safety and toleration profile. The changes in AUC when protease inhibitors are
co-dosed with ritonavir are shown in Table 9.2 and in some instances can be
quite dramatic; for example, the AUC of saquinavir can increase up to 100-fold
in the presence of ritonavir. Although plasma levels are higher when co-dosed
with ritonavir, the pharmacokinetics of saquinavir are still highly variable.
The pharmacokinetic boosting strategy for HIV proteases is so well accepted
that a number of protease inhibitors (e.g. lopinavir, darunavir, tipranavir) are
only used in combination with ritonavir as they do not achieve sufficiently high
plasma exposure alone. As well as increasing F%, ritonavir also increases the
half-life of a number of compounds including saquinavir and darunavir.27,28
Interestingly ritonavir has been shown to change the disposition of darunavir in
humans. When radiolabelled darunavir was dosed alone, it was rapidly meta-
bolised and recovery of unchanged drug was low (7% faeces, 1% urine), in the
presence of ritonavir excretion of unchanged darunavir was extensive (41% in
faeces and 8% in urine), suggesting that in the presence of ritonavir, trans-
porters play a role in darunavir clearance.27,264
Although ritonavir boosts the levels of all the HIV protease inhibitors, its
interactions with other drugs are more complex. In addition to inhibiting CYP
3A4, ritonavir also inhibits 2D6 and 2C9 and 2C19 and P-gp, and induces levels
of a number of CYP isozymes (including 2B6, 2C9, 2C19) and some UGT iso-
zymes with full induction requiring two weeks of ritonavir dosing.62 Depending
on the length and dose of ritonavir administration and the co-administered
substrate, administration with ritonavir can result in inhibition or induction.
The distribution of the alcohol TSAI of proteases within the body is largely
as would be expected on the basis of the physicochemical properties of the
compounds (Figure 9.5, Table 9.1). The high lipophilicity of these compounds
means that they tend to show extensive binding to plasma proteins. All the HIV
proteases apart from the least lipophilic compound, indinavir, show extensive
protein binding in humans (498%). Amprenavir, darunavir and nelfinavir
bind mainly to a1-acid glycoprotein (AAG).20,27,28,82 Amprenavir also changes
the levels of AAG on chronic dosing and so changes its own protein binding
with time. Atazanavir and lopinavir bind to both albumin and AAG.26,30
The distribution of alcohol transition state analogue inhibitor compounds
has been investigated using whole body autoradiography techniques. Imme-
diately following intravenous dosing of 14C-remikiren in the rat,83 high levels of
radioactivity were seen in the liver, kidney, stomach, small intestine, choroid
plexus, thyroid, pancreas, pituitary and myocardium of all three species studied
(rat, guinea pig and monkey). After one hour, tissue radioactivity levels were
still high but radioactivity was not detectable in the blood. At 24 h post-dose,
radioactivity was still seen in the gastrointestinal tract, kidney and some brain
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 411
areas. The kidney appears to act as a reservoir for remikiren because further
analysis of radioactive material in the kidney after 24 h showed that the
majority of radioactivity was unchanged remikiren. A possible explanation for
this retention is the material enters the kidney (possibly via an uptake trans-
porter), but unlike in the liver, there is no comparable excretory transporter to
remove the compound from the kidney. As metabolic capacity for this com-
pound is low in the kidney, the compound remains in the kidney predominantly
as unchanged drug.
HIV protease inhibitors in general have limited and variable ability to enter
the cerebrospinal fluid (CSF) of humans treated with the drugs (Table 9.3). One
reason for this is that P-gp (and possibly other transporters such as MRP 1 and
2) limits the permeability of the compounds across the blood–brain bar-
rier.267,268 Whilst in the intestine compounds can reach high concentrations and
inhibit (and overcome) P-gp, the free systemic concentrations of these com-
pounds are much lower than those seen in the intestine and thus P-gp at the
blood–brain barrier is not inhibited.
The effect of P-gp on limiting CNS penetration of HIV protease inhibitors
has been extensively studied in mice. Increases of between 5–36 fold were seen
in the brains of P-gp knockout mice compared to wild-type mice dosed with
amprenavir, saquinavir, indinavir, ritonavir and nelfinavir .31,44,84,85 Increases
of brain concentrations of a similar magnitude are seen for these compounds
when wild-type mice are dosed with P-gp inhibitors such as GF-120918 or
cyclosporine.31,44 As well as influencing the CNS penetration of protease
inhibitors, transporters (including P-gp and MRP proteins) also play a role in
modulating the levels of HIV protease inhibitors in lymphocytes, a pre-
dominant site of HIV replication.89–92
Inhibition of transporters may play a role in some of the side effects seen with
HIV proteases inhibitors. Saquinavir (1.2 mM) and indinavir (6.8 mM) both
inhibit human OATP 1B1, which has a physiological role in bilirubin transport.
When the Fu in plasma and therapeutic plasma concentrations were considered
it was concluded that, in vivo, the inhibition of OATP-1B1 would be 40% for
indinavir and 7% for saquinavir, and that this may explain the unconjugated
hyperbilirubinemia seen in some patients dosed with indinavir.58 Ritonavir and

Table 9.3 CSF : free plasma concentrations of HIV protease inhibitors in

human.
Compound Average CSF : free plasma ratio Reference

Atazanavir B0.05 26,86

Amprenavir B0.02 87
Darunavir 0.17 88
Indinavir B0.2a 87
Saquinavir 0.07–0.2 87
Ritonavir 0.05–0.25 295
Lopinavir Undetectable, 0.8 30, 294
a
Occasional instances of CSF : free plasma 41 reported for indinavir.
 412

Table 9.4 Chemical structure, physicochemical properties and clearance, bioavailability and absorption estimates of aldehyde
containing TSAI.
Compound Structure Target MW clogP TPSAa pKa HBA HBD F%

Pralnacasan Caspase-1 524 2.2 147 Neutral 7 2 40–60 r

O
Ref. 100 N
N
O N N
O
O
O N O

RU-36384 O caspase-1 496 1.2 166 Acid 8 3 4r

Ref. 100 N N O
(4.2)
N
O
O N
O
O N
O

Efegatran N N Thrombin 417 -0.5 140 Neutral 7 6 3r

Ref. 269 N (phase 2) 30 d
O O
N
N N
O
Chapter 9
CVS-1123 N N Thrombin 511 0.8 184 Neutral 8 6 79 d
Ref. 269 N (phase 2) 36 mk
O O
N
N N
O O
O

AL 1 Thrombin 509 0.1 168 Weak acid 7 6 15 d

H
Ref. 98, 99 N N (9.0)
O
S N O N
O N N
O
O
AL 2 N N Thrombin 481 0 183 Weak acid 8 6 66 d
Ref. 98, 99 N
(9.0)
O
S N O
O N N
O
O
AL 3 N N Thrombin 467 -0.6 183 Weak acid 8 6 35 d
Ref. 98, 99 N
(9.0)
O
S N O
O N N
O
O

d ¼ dog, mk ¼ monkey, r ¼ rat.

a
TPSA ¼ topological polar surface area
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme
413
414 Chapter 9
saquinavir both inhibit the uptake (via NTCP) and excretion (via BSEP) of bile
acids in human liver.93 If the balance of these competing actions in vivo results
in bile acid accumulation in hepatocytes, it may lead to toxicity.
A side effect common to many HIV protease inhibitors is a partial lipody-
strophy characterised by a redistribution of adipose, with a loss of fat in the
face, arm and legs and the gain of fat around the back of the neck leading to a
characteristic buffalo hump. This has been linked to a lack of specificity of some
of the HIV protease inhibitors which also inhibit ZMPSTE24, a human pro-
tease involved in the conversion of farnesyl prelamin A to lamin A, a key
nuclear structural protein.94–96

9.3.1.2 Aldehyde Containing TSAI of Proteases

Aldehydes are not frequently incorporated into drugs as a result of their poor
chemical stability and because they carry the risk of toxicity; for example, from
the in vivo generation of covalent protein adducts through Schiff base forma-
tion. However, there are some examples of medicinal chemistry programmes
that have focused on aldehyde containing compounds as potential protease
inhibitors since reaction with the catalytic triad of serine proteases leads to the
formation of tetrahedral (transition state analogue) intermediates (Figure 9.7).
The enzymes for which aldehyde TSAI have been tested in vivo include
thrombin and caspase-1. Caspases hydrolyse peptide bonds on the carboxyl
side of an aspartate residue using a thiolate anion formed from Cys285 by
His237 through a mechanism equivalent to the serine protease shown in Figure
9.6.97 Caspase-1 cleaves pro-IL-1b and IL-18 to release the active cytokine.
Two aldehyde containing thrombin inhibitors (efegatran and CVS-1123) have
reached the clinic (Table 9.4)269 and pre-clinical data on a related set of aldehyde
thrombin inhibitors (AL 1, 2 and 3) has been published.98,99 It is likely that these
compounds exist as hemi-aminals rather than true aldehydes, through reversible
cyclisation between aldehyde and guanidine. This is illustrated for efegatran in
Figure 9.10. The cyclisation stabilises the aldehyde functionality and could
enhance oral bioavailability by reducing both the basicity of the guanidine and
the overall PSA and may explain the reasonably high bioavailability of these
compounds in the dog (15–79%). It is interesting that changing the ring size in
AL 2 and 3 resulted in a doubling of bioavailability, although no further phar-
macokinetic data were presented to explain this finding.
The free aspartate hemi-acetal caspase inhibitors have poor bioavailability
due to poor absorption, but this can be improved through the use of prodrugs.
Pralnacasan is an ethyl acetal prodrug of RU-36384 and increased oral bioa-
vailability in rodents from 4% to 40–60%.100 Pralnacasan also has good
absorption in humans and is in phase II trials for osteoarthritis.101

9.3.1.3 Activated Ketone Containing TSAI of Proteases

The dipole moment of the carbonyl group makes a ketone susceptible to
nucleophilic attack. If correctly presented within the active site of a protease,
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 415

Open chain efegatran Cyclic hemi-acetal efegatran

TPSA 140 92
cpKa 14 12
HBD* 6 3

* excluding internal H-bonding network from cyclic acetal

Figure 9.10 Open chain and cyclic hemi-acetal forms of efegatran.

it can be converted to a tetrahedral intermediate. The addition of an a-electron

withdrawing group (heterocycles, carbonyl and fluorinated alkyls) will both
enhance nucleofugicity and stabilise the tetrahedral transition state thereby
increasing binding affinity of the activated ketone compound. Provided that
these activating groups are poor leaving groups, this can result in potent
inhibitors that are not substrates of the protease.
The enzymes targeted with activated ketones include cathepsin-K, elastase,
HCV NS3-protease and tryptase. The chemical structure, physicochemical
properties and pharmacokinetic data for the activated ketone compounds
discussed in this chapter are shown in Table 9.5.
ZD-8321 and ZD-0892 are lipophilic (clogP 3-4) trifluoromethyl activated
ketone TSAIs with a PSA in the region associated with good oral absorption.
In pre-clinical species these compounds, as would be expected, had high oral
absorption and bioavailability was also high. ONO-6818 is a less lipophilic
elastase inhibitor (clogP) with a higher PSA (144 Å2), but the compound also
shows reasonable bioavailability in pre-clinical species.102
Relacatib, a potent cathepsin inhibitor (Ki 0.04 nM), is an example of a
heterocycle activated ketone and boceprevir is a carbonyl activated ketone
inhibitor of the NS3-4A protease enzyme of hepatitis C virus (HCV). Both
compounds are large, lipophilic, neutral compounds with a PSA on the edge of
where good oral absorption is expected (relacatib PSA 147 Å2, boceprevir PSA
151 Å2). Despite this both compounds show reasonable exposure following oral
dosing (Table 9.5). Relacatib is well absorbed in both rat and monkey, and has
high bioavailability in the rat and moderate bioavailability in the monkey.
Removal of the methyl substituent on the azepane ring of relacatib led to a
significant poorer pharmacokinetic profile in the rat, with clearance doubling
and bioavailability halving to 42% whilst protein binding was unchanged at
B97%. Addition of the methyl group may change the conformation of the
 416
Table 9.5 Chemical structure, physicochemical properties and clearance, bioavailability and absorption estimates of activated
ketone containing TSAI.
TPSA
HBA Vss CL (ml/min/
Compound Structure Target MW clogP HBD pKa (L/kg) PPB (%) kg) F% % absorbed

Relacatib O O Cathepsin- 541 3.6 147 Neutral 1.8 r 97 r 20 r 89 r 100 r

O S N
Ref. 103 O N k 4 1 mk 12 mk 28 mk B40
N 2 (BP ¼ 1)mk
N
O O

ICI-200,880 O Elastase 687 5.6 167 Acid (3.7)

Ref. 105, 106 Cl O 7
N O F
3
N N F
S N
F
O O O O
K-1 O Elastase 583 2.8 125 Neutral
O N O
5
N 2
N N F
O
O F
O F
K-2 O Elastase 633 3.4 125 Neutral
O N O 5
N F F
N N F 2
O
O F
O F
ZD-0892 O Elastase 500 3.9 105 Neutral 122 ham 82 r
Ref. 109 5 98 r 39 d
N O 2 52 d 26 ham
N
N F
O
O O F
F
ZD-8321 O Elastase 423 2.7 105 Neutral 58 ham 84 r
Ref. 109 4 25 r 75 ham
N O 2 18 d 70 d
O N
O N F
O F
Chapter 9

F
FK-706 Chiral Elastase 593 2.7 162 Acid (3.5) 86 ha
Ref. 104 (phase 2) 7 73 hsa
O 4
F
N F
N F
O
O
O N

N O
O

O
ONO-6818 Elastase 453 51 r
Ref. 102 31 d
N 18 mk
N O
H2N
O N N
H N

O O

Boceprevir NS-3 pro- 520 3.3 151 Neutral 0.5 r B33 r 34 mo 36 r

Ref. 10, 111 tease 5 26 r
O O (phase 3) 5 30 d
N N N 4–11 mk
N N
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme

O O O

Telaprevir H NS-3 pro- 680 5.4 180 Neutral 0.5–5.8 r 18–53 rb 25–35 r 445 rc
N
(VX-950) O O tease 8 1–1.8 d 42 d 41 d Complete d
Ref. 111, N N HN N (phase 3) 4
N N
265, 296, O O
O O
297
417
 418

Table 9.5 (continued )

TPSA
HBA Vss CL (ml/min/
Compound Structure Target MW clogP HBD pKa (L/kg) PPB (%) kg) F% % absorbed

K3 NS-3 740 AUC

Ref. 113 O protease (1 mM h1) at 740 ng h1 ml1
H H
O N N 10 mg kg1
N N N O O
O
O O

K4 NS-3 726 43 r 22 r B80

Ref. 114 O protease 21 d 35 d
H H
O N N 32 mk 4 mk
N N N O O
O
O O

d ¼ dog, h ¼ human, ham ¼ hamster, hs ¼ human smokers, mo ¼ mouse, mk ¼ monkey, r ¼ rat.

a
After inhalation administration 100 mg.
b
non-linear with increasing dose (0.95–5 mg kg1). c Telaprevir BP ratio is 0.56 (unpublished observations).
Chapter 9
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 419
103
azepane ring and increase its metabolic stability. In the monkey the com-
pounds had similar Cl, but again the desmethyl analogue had poorer F% (7%).
Telaprevir (VX-950), like boceprevir, is a carbonyl activated ketone inhibitor
of HCV NS3-4a protease but it has physicochemical properties considerably
outside those typically seen for orally bioavailable compounds (MW 680, clogP
5.4, measured log D7.4 4, PSA 180 Å2). Yet telaprevir has reasonable bioa-
vailability and absorption (445–100%) in pre-clinical species (Table 9.5) and
flux through Caco-2 cells is observed in vitro.265,266 The poor solubility of tel-
aprevir does lead to variable plasma levels following high oral doses in the rat
(30 mg kg1), and in both rats and dogs the oral half-life (B3 h) was longer than
after intravenous dosing (r1 h) suggesting that the rate of absorption tela-
previr is slower than the rate of elimination. One possible explanation for the
good absorption of telaprevir despite its physicochemical properties is that the
highly peptidic structure of the compound results in a well-defined secondary
structure held together with intramolecular hydrogen bonding. In turn this will
reduce the hydrogen bonding with solvent and lowers the desolvation penalty
required for the compound to penetrate membranes.
FK-706 (clogP 2.7, MW 593, PSA 162 Å2) and ICI-200 880 (clogP 5.6, MW
687, PSA 167 Å2) are high molecular weight elastase TSAI with similar (par-
ticularly ICI-200 880) physicochemical properties to telaprevir. However, both
FK-506 and ICI-200 880 have little oral bioavailability and need to be admi-
nistered by the intravenous or inhaled route to show biological effects in
humans. In the case of ICI-200 880, the low F% is due in part to high clearance
while absorption of FK-706 in pre-clinical species is low.104 Following inha-
lation of nebulised FK-706 in otherwise healthy humans, the AUC was similar
in smokers and non-smokers; however, the Cmax was higher, the Tmax earlier,
the half-life shorter and the rate of absorption about 10-fold higher in smokers
reflecting the higher pulmonary permeability in people who smoke.104 ICI-200
880 shows flip-flop kinetics following inhalation administration in the hamster
(intravenous half-life 1 h, inhaled half-life 10 h)105,106 possibly reflecting low
solubility and tissue permeability of the compound.
Incorporation of electron withdrawing groups a- to a carbonyl in order to
gain potency also has consequences for the behaviour of these compounds in
biological systems. As well as reducing the pKa of the carbonyl group and
increasing its reactivity towards nucleophiles, the electron withdrawing group
increases the propensity for racemisation leading to the formation of a mixture
of diasteroisomeric compounds.107 For example, the half-life for epimerisation
of K-2 in human blood serum at 37 1C is 30 min,108 ZD-0892 and ZD-8321 are
also susceptible to rapid epimerisation in plasma, in this case to give a 6 : 4
mixture in favour of the SSS isomer,109 FK-706 exists as two epimers (the
pharmacologically active SSS epimer and an SSR epimer104) and ICI 200,880
exists as two epimers that undergo facile interconversion in aqueous
solution.106
It has also been proposed that differences in oral efficacy of K-1 and K-2
could be due to the measurably different levels of hydration of the two com-
pounds.108 Despite equal in vitro potencies, oral efficacy was only observed with
420 Chapter 9
19
pentafluoroethyl derivative K-2. F NMR studies in buffered saline showed
that, while K-1 was fully hydrated, a significant proportion (20%) of K-2
existed in the ketone form. It was not ascertained whether these differing levels
of hydration affected clearance rates or oral absorption, but since further
examples highlighted that merely incorporating a pentafluoroethyl group is not
sufficient to guarantee in vivo efficacy, it is more likely that oral bioavailability is
driven by the overall physiochemical properties of the molecule (most notably
the hydrogen bond burden and the lipophilicity) than by the degree of
hydration of the ketone group.110
In addition to undergoing epimerisation, the lipophilic activated ketone
transition state analogue inhibitor compounds are also susceptible to meta-
bolism, often involving the activated ketone group. For instance, the major
metabolite of ICI-200 880, ZD-0892 and ZD-8321 was the trifluoromethyl
alcohol.105,109 Other metabolites of ICI-200 880 were also identified in rat bile
including a metabolite where hydrolysis of the right hand side amide group
had occurred to liberate the whole of the trifluoromethylketone group.105
Both Boceprevir and Telaprevir are metabolised by CYP 3A isozymes in vitro
and the metabolism of these compounds in rat and human liver microsomes
can be inhibited with the HIV protease inhibitor compound ritonavir.111 Cmax
(43-fold) and AUC (48-fold) of both compounds were increased when the
compounds were dosed orally (5 mg kg1) to rats with ritonavir (5 mg kg1).111
As well as primary structure, the conformation of peptides is also
important for recognition by protease enzymes. The finding that only the
extended ‘saw-tooth’ conformation of peptides appears to bind to the active
sites of proteases has led to the proposal that ‘macrocyclisation’ can be used as
a general strategy to design protease inhibitors. Macrocyclic inhibitors
mimicking the transition state of many enzymes (including HIV protease,
renin) have been described.112
In terms of ADME properties, macrocyclic compounds also have some
advantages. Unlike linear peptides, the macrocyclic peptides have high resis-
tance to proteolytic degradation and, if sufficiently lipophilic, the compounds
can show membrane permeability and oral bioavailability. K3 (Table 9.5) is a
potent (Ki 2 nM, HCV replicon EC90 20 nM) macrocyclic ketoamide inhibitor
of HCV NS3 protease and shows oral exposure in rat following oral dosing.113
Although there is no intravenous pharmacokinetic data for K3, the exposure
after oral dosing in the rat (1 mM h1) is very similar to that achieved with the
structurally related but non-macrocyclic compound K4 (Ki 4, replicon EC90
30 nM) (Table 9.5), which is a back-up compound to boceprevir and has high
absorption and good bioavailability in the rat.114

9.3.1.4 Boronic Acid Containing TSAI of Proteases

Boronic acids are Lewis acids with the ability to accept electrons thereby
completing the open shell of boron resulting in a shift from a neutral trigonal
planar sp2 boron to an anionic tetrahedral sp3 boron (Figure 9.11). Phenyl
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 421

OH
OH H 2O
R B R B OH
OH OH

Figure 9.11 Conversion from neutral sp2 boron to tetrahedral anionic sp3 boron.

boronic acids typically have a pKa in the 4.5–8 range and butylboronic acid has
a pKa of 10.6.250
As amide (peptide) cleavage requires conversion of a neutral sp2 carbonyl
carbon to a tetrahedral sp3 carbon, boronic acid compounds can mirror this
transition and make the same interactions with the protease enzyme as the
tetrahedral transition state thereby acting as potent TSAI of a number of
protease enzymes.115 Boronic acid inhibitors of serine proteases were first
synthesised in the early 1970s.116,117 Since that time boronic acid compounds
have been shown to be effective inhibitors of a number of enzymes including
thrombin,118,119 Factor Xa,120 trypsin, chymotrypsin, pancreatic elastase, leu-
kocyte elastase, thrombin, b-lactamases,121–125 DPP IV and HCV NS3 pro-
tease,126,127 and have been explored for use as therapeutic agents in various
disease states.115,128
There are two different mechanisms by which the tetrahedral boronate anion
is formed. In the case of arginase inhibitors, the tetrahedral boronate anion is
formed by reaction with water.129 In contrast for other enzymes such as serine
proteases, the hydroxyl anion of the catalytic amino acid typically attacks the
boronic acid to form the boronate intermediate, resulting in the formation of a
covalent bond to the enzyme which, while reversible, typically gives slow off-
rates of hours or days thereby giving a significant pharmacodynamic advantage
to this class of inhibitor.130 Because of this requirement for boronic acids to
form boronates to mimic the transition state of protease enzymes, it has been
argued that boronic acids are not formal transition state analogues but should
be designated as either ‘reaction coordinate analogues’ or covalent inhibitors
depending on the mechanism of boronate formation (water vs. protein).
The boronic acid TSAI of proteases discussed further in this chapter are
shown in Table 9.6. The enzyme targets of these compounds include thrombin,
dipeptidylpeptidase IV (DPP-IV) and the 20S proteasome. DPP-IV cleaves
dipeptides from the N-terminus of numerous polypeptides including incretins
such as GLP-1 that have an important role in blood glucose homeostasis.
Small, polar (clogP 2.2 to 5.3) boronic acid containing compounds are
effective TSAIs of DPP-IV. Proteasomes cleave peptides through threonine-
dependent nucleophilic attack and transition state inhibitors have been
designed to inhibit this process. The boron containing TSAI of proteasomes
need higher lipophilicity (clogP 0.81.3) to effectively bind in the active site of
the proteasome. Bortezomib is a potent boronic acid inhibitor of the chymo-
tryptic-like site of the 20S proteasome (Ki 0.62 nM) and establishes a slowly
reversible (half-life B20 min) tetrahedral intermediate with the active site
N-terminal threonine residue.131
 422

Table 9.6 Chemical structure, physicochemical properties and clearance, bioavailability and absorption estimates of boronic acid
containing TSAI.
TPSA
HBA Vss PPB CL
Compound Structure Target MW clogP HBD pKa (L/kg) (%) (ml/min/kg) F% % absorbed

Dutogliptin DPP-IV 241 -2.2 85 Zwitterion

3 (9.7 A, 10.5
N N O B)
N B
4
O O
Flovagatran Thrombin 524 2.2 140 Weak acid 0.6 h Dosed IV
Ref 298 (9.7)
4
O O 3
N N B
O N O
O O

O
Tri-50b Thrombin 608 5.0 115 Neutral 24 r
Ref 134
4
O O 2
N N B
O N O
O O

O
DuP-714 N N Thrombin 460 2.4 183 Weak acid
5 (9.6)
N 8

N
N
O
O N B
O O
Chapter 9

O
S-18326 Thrombin 488 1.3 181 Weak acid 27 dfa
Ref 299 N N (9.6)
6 6 dfe
N 8
O O
N O
N N B
O O
Bortezomib Proteasome 384 0.8 124 Weak acid B3–13 h 93 r 14 mo 11 mo
(approved) N (9.7) Vz B
Ref. 136–138, O 24–7 h
247 4 94 d 13–27 hsd
N O
N N B 4 83 h B3.5–8.5 hmd

O O

CEP-18770 Proteasome 413 1.6 132 Weak acid 10 mo 99.9 r 23 mo 39 mo B80a

(9.7)
O 4 18 r 99.9 d 3.2 r 54 r
N O 5 99.8 h
N N B
H
O O
O

dfa ¼ dog fasted, dfe ¼ dog fed, hmd ¼ human multiple dose, hsd human single dose, r ¼ rat.
a
Assuming blood : plasma ¼ 1.
mo ¼ mouse.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme
423
424 Chapter 9
In general the boronic acid compounds discussed have reasonable PSA
values for oral absorption (six out of seven compounds have PSA in the range
85–140 Å2). One reason for this is that the boronic acid compounds can be
considered as prodrugs as they are converted to the final, more polar tetra-
hedral transition state analogue inhibitor in the body.
Polar boronic acids are poorly absorbed transcellularly as would be expected
from their physicochemistry; for instance DUP-714, a thrombin inhibitor
containing both a guanidine and boronic acid group (clogP 2.4, PSA 183 Å2),
has low permeability through rat jejunum (Papp 1.5106 cm sec1, pH 7.4).
Some analogues of DUP-714 had significantly higher permeability (Papp
18106 cm sec1), which was reduced in the presence of known inhibitors of
the oligopeptide transporter,132 suggesting that while passive permeability was
not high this could be overcome if the compounds cross the membrane by
active influx mechanisms. Dutogliptin is a small (MW 241), basic, polar (clogP
2.2) DPP IV inhibitor. It is orally bioavailable (CL/F 10–30 ml min1 kg1) in
humans and Tmax was 3–5 h. Given the physicochemical properties of duto-
gliptin, it is likely that this compound is absorbed via the paracellular route.
The more lipophilic protease inhibitor CEP-18770 (clogP 1.6) has reached
phase I trials. In pre-clinical species it has high absorption and oral bioavail-
ability, low clearance (although the protein binding and hence CLu are high)
and a long half-life.133 The data with CEP-18770 demonstrate that the presence
of the boronic acid moiety per se does not preclude good transcellular
absorption.
In an effort to increase the absorption of poorly absorbed boronic acid
TSAIs, prodrug strategies using boronic esters (which decrease the number
of HBD by two) have been employed. TRI-50b is a pinanediol ester of
flovagatran which showed modest bioavailability in the rat and is cur-
rently in clinical trials—phase I (oral as prodrug) and phase II (iv as
flovagatran).12,134,135
The distribution of boronic acid compounds in vivo can be influenced by their
ability to form reversible covalent bonds. For example, bortezomib in addition
to having moderate binding to human plasma proteins, binds slowly to the
proteasome in red blood cells. This binding has been shown to compli-
cate the bioanalysis of bortezomib in blood with differential recoveries
between incurred and spiked standard samples being noted.251 Bortezomib has
marked multiphasic pharmacokinetics after intravenous dosing with a long
terminal elimination phase (half-life B9–15 h) and a large terminal
volume of distribution (410 L kg1).136–137,247 Some studies have reported
a change in volume on multiple dosing,136 whilst others have shown it to be
unchanged.138 Extensive distribution of bortezomib related radioactivity
into tissues (apart from the brain) was also seen in monkeys dosed with
radiolabelled bortezomib.
The elimination and clearance pathways of boronic acid compounds vary
from compound to compound. After intravenous dosing of flovagatran, 14%
of the dose was excreted unchanged in the urine; in contrast after intravenous
administration bortezomib is extensively metabolised with little excretion of
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 425
unchanged bortezomib in the urine and none in the faeces of humans dosed
with the compound.139,252 The initial route of metabolism of bortezomib is
oxidative deboronation (none of the thirteen metabolites in human plasma
contain the boronic acid group), which removes the functional group
mimicking the transition state producing metabolites inactive against the
proteosome.139 Extensive deboronation is likely to be a major metabolic
pathway for any boronic acid compound lipophilic enough to undergo CYP
metabolism and is consistent with the finding that boronate compounds in
general are susceptible to undergoing Baeyer–Villiger-type oxidative debor-
onation reactions.139–142
In human liver, CYP 3A4 is the predominant isozyme metabolising borte-
zomib with quantitatively important contributions also from CYP 2C19 and
CYP 1A2.143 In vivo drug–drug interaction (DDI) studies showed that clear-
ance of bortezomib was decreased; AUC increased by a mean value of 35%
(CI90% 1.03 to 1.8) in the presence of ketoconazole (a CYP 3A4 inhibitor) but
not in the presence of omeprazole (a potent CYP 2C19 inhibitor).247 As well as
being metabolised by CYP 2C19, bortezomib is also a weak inhibitor of the
isozyme (IC50 ¼ 18 mM),144 although the inhibition is related to the scaffold of
the compound rather than to the presence of the boronic acid group per se.
Mechanistic studies145 suggest that human P450 can deboronate bortezomib via
multiple mechanisms including CYP mediated formation of reactive oxygen
species (e.g. superoxide) in addition to direct oxidation by the hypervalent oxo-
iron species [Fe(V) ¼ O].
Bortezomib and CEP-18770 can also theoretically cause CYP-mediated DDI
by inhibiting the degradation of CYP isozymes by the 20S proteasome (their
pharmacological target). In rats, bortezomib inhibits the degradation of CYP
2E1 following induction of the isozyme by ethanol but paradoxically decreased
the levels of CYP 3A1/2 and total CYP content.144 In humans, bortezomib
administration caused no changes in results from the erythromycin breath test
(an in vivo measure of CYP 3A4 activity).
The physical form and stability of boronic acid TSAIs can also pose a pro-
blem in their development as therapeutics. For example in solid form borte-
zomib exists in its cyclic anhydride form as a trimeric boroxine. To overcome
this, the bortezomib drug product is a mannitol ester which, when recon-
stituted, exists in equilibrium with the monomeric boronic acid form (solubility
B3 mg ml1).146 Bortezomib is also chemically unstable in a number of
matrices and acidification was needed to stabilise bortezomib during protein
precipitation extractions.139

9.3.1.5 Phosphonic Acid Containing TSAI of Proteases

Angiotensin converting enzyme (ACE) is a Zn21 metalloproteinase that con-
verts angiotensin I (ANG I; 10AA) to angiotensin II (ANG II; 8AA) a major
regulator of fluid and sodium balance and haemodynamics, but also of cellular
growth and cardiovascular remodelling.253 The tetrahedral transition state in
the conversion of ANGI to II by ACE is shown in Figure 9.12.254
426 Chapter 9
H
H
N H R2
E B
NH OH
O N N
H O N
R2 H O R O H O
+H2O H H O R
N O
N N H O N O +
OH N N
H O R1 H O H2N K O R
O R1 H
R2 H2N O
HO Y N
Zn HO Y OH
R1 H

Transition state

Figure 9.12 Proposed tetrahedral intermediate in the conversion of Angiotensin I to

Angiotensin II by Angiotensin converting enzyme.

Three different structural classes of ACE inhibitors have been approved as

drugs:

 sulfhydryl (e.g. captopril);

 dicarboxylate (e.g. enalaprilat and ramiprilat);
 phosphonic (fosinoprilat).

Although potent inhibitors of ACE, the sulfhydryl (Zn21 chelation) and

carboxylate compounds are not TSAI (Figure 9.13). The phosphonic acid
compounds such as fosinoprilat, however, mimic the tetrahedral nature of the
transition state.
Due to the highly polar nature of fosinoprilat it has limited
membrane permeability and oral bioavailability is low. To improve bioavail-
ability, fosinoprilat is administered as a phosphinic ester prodrug, fosinopril.
Although fosinopril has physicochemical properties more typical of an oral
drug (log D 2.63 and PSA 70 Å2), it is predominantly in its ionised state in the
gastrointestinal tract and this contributes to its incomplete absorption
(B35%). As well as passive transcellular permeability, the influx transporter
PEPT-1 may play a role in absorption of fosinopril.255,256 The polar, acidic
nature of fosinoprilat limits its distribution outside the extravascular space in
humans (Vss 0.14 L kg1). Clearance of fosinoprilat is low (0.56 ml min1 kg1)
with a high renal component (0.24 ml min1 kg1). After intravenous dosing of
fosinoprilat, excretion is only as unchanged drug. After oral administration of
the more lipophilic fosinopril, additional metabolites of fosinoprilat (including
glucuronide and oxidised metabolites) are observed. This may arise from
metabolism in the intestine or because the more lipophilic prodrug is able to
undergo metabolism in the liver prior to cleavage and liberation of
fosinoprilat metabolites.147,148 Studies in the dog showed that there was sig-
nificant hydrolysis of fosinopril in both the intestine and liver (extraction
ratios of 0.6–0.9), but because the intestine sees drug first, it is responsible
for metabolising 475% of an oral dose of fosinopril. In addition it was
shown that following systemic administration of fosinopril, extrahepatic sites
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 427

Chemical group Example ACE inhibitors Key interactions with

the active site Zn2+
Sulfhydryl O OH H H
O
-S
HS N CH3
CH3 Zn Y
Captopril HO

Carboxylic acid O OH
H B
OH O E
O
N H
O N O O
CH3 H O H
N R
O
..
Enalaprilat R′
Zn
H
O Y

Phosphonic acid OR O O OH E
O
P
N O H
O
H
O
+
R= H Fosinoprilat P
R= Fosinopril
O O

O Zn HO Y

Figure 9.13 Example of sulfhydryl, carboxylic and phosphonic acid ACE inhibitors.

(including the kidney) were responsible for about 50% of the systemic
clearance.147

9.3.2 Neuraminidase TSAI

The neuraminidase enzyme of the influenza virus plays a key role in viral
infectivity and has emerged as a useful therapeutic target.149 The substrate of
viral neuraminidase is the extremely polar terminal sialic acid residue of gly-
coproteins.150 The neuraminidase catalytic mechanism is thought to involve
two chemical steps—cleavage of the C2 glycosidic bond and subsequent
hydroxylation of a planar oxocarbonium ion intermediate.150 To bind effi-
ciently in the polar active site of viral neuraminidase and to mimic the oxo-
carbonium intermediate, the neuraminidase TSAI also need to be very polar151
428 Chapter 9

OH OH
O OH O OR

O δ+ O

HO HO OH
OH
HN HN
HO OH HO OH
O O

Sialic Acid Proposed transition state

Figure 9.14 Structure of sialic acid and the proposed transition state of sialic acid
within viral neuraminidase.

(Figure 9.14, Table 9.7). Although, the interactions of sialic acid, zanamivir and
oseltamivir acid with viral neuraminidase are similar, the TSAI bind to the
enzyme 0.5–1.5106 times more tightly than sialic acid and show time-depen-
dent and slowly reversible inhibition kinetics.150,152
Zanamivir is a highly polar zwitterionic compound and as a result has poor
membrane permeability, limited absorption and low bioavailability in rat and
human (Table 9.7). Zanamivir has higher F% when administered via the
intranasal (B10%) or inhaled routes (B15–20%) in humans.153 Inhaled
administration of zanamivir Tc99m (10 mg) in healthy volunteers showed that
4–17% of the dose was deposited in the lungs. After dry powder inhalation
administration, the Tmax of zanamivir in the systemic circulation was 1–2 h;
the systemic half-life of zanamivir was longer than after intravenous adminis-
tration (3.6–5 vs. 1.7 h) suggesting that absorption through the lung is slow
enough to become the rate limiting step in the elimination of zanamivir (flip-
flop kinetics).
Oseltamivir acid is 4 log units more lipophilic than zanamivir and has a PSA
more typical of an oral drug (o140 A2). However, oseltamivir acid is still quite
hydrophilic clogP 1.2 and in vitro has limited ability to cross cell monolayers
(Papp o2106 cm s1). When administered orally to rats or humans, osel-
tamivir acid has low bioavailability, primarily due to low absorption.154 To
overcome this limitation, the ethyl ester prodrug (oseltamivir) of oseltamivir
acid was prepared. Oseltamivir is about 3 log units more lipophilic than osel-
tamivir acid and has high aqueous solubility (4500 mg ml1) and a moderate
ability to cross cell monolayers in vitro ( Papp 6–12106 cm s1 in Caco-2 and
LLC-PK1 cell lines).155,156 Despite its relatively small size, oseltamivir is a
substrate for human P-gp.
In pre-clinical pharmacokinetic studies, the bioavailability of oseltamivir
acid following oral dosing of oseltamivir was 11–73% (Table 9.7). The lowest
F% was seen in the ferret which had a relative inability to cleave oseltamivir to
the acid metabolite suggesting that, even in this species, absorption of oselta-
mivir was good.157 The improvement in F% (o5 to B75–80%)154 of
Table 9.7 Chemical structure, physicochemical properties and clearance, bioavailability and absorption estimates of neur-
aminidase TSAI.
Compound Structure MW clogP TPSA pKa HBA HBD CL (ml/min/kg) Oral F%
O O
Zanamivir 332 5.6 198 3.8 A 10 9 1.4–32 r 3–4 r
Ref 153 N O O
O 11.3 B 1.6 h 2h
N N
O N O

Oseltamivir O O 312 2.1 91 8.8 B 4 3

Ref 154

O N
N

O
O O
Oseltamivir acid 284 1.2 102 2.2 A 5 4 25 r 4.3 r
Ref 154
9.7 B 5–6 h o5 h
O NH2
11 fa
N
30–73* m, r,
O
d, ma
75–80 ha
O O
GS–4116 326 1.7 138 4.3 A 7 6 22 4r
Ref 157
N N
12.8 B 2 ra
O
N
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme

N
O
Dimeric zanamivir 943 5.5 412 19 17 0.7 o1
O O
Ref 172
O O N N O O
O O O
O O
O
N O N
N N N N
429

O O
N N

d ¼ dog, f ¼ ferret, h ¼ human, m ¼ mouse, ma ¼ marmoset, r ¼ rat. aFollowing dosing of ethyl ester prodrug.
430 Chapter 9
oseltamivir acid seen in humans when dosed as oseltamivir is one of the highest
seen with an ester prodrug 158 and it is unusual to get such a big increase in F%
with such a simple modification. In a study to look at regional absorption of
oseltamivir in humans, the prodrug was well absorbed from the stomach,
jejunum and ileum.156 Oseltamivir is absorbed at least in part transcellularly, as
slow absorption was also seen from the colon (a region where tight junctions
have low porosity and paracellular absorption is minimal).
Although oseltamivir was a successful prodrug, attempts to ‘prodrug’ a
more polar analogue (GS-4116: clogP 1.7, PSA 138 Å2) were unsuccessful
(Table 9.7). The ethyl ester prodrug of GS-4116 showed no improvement
in oral bioavailability (F% B2) in the rat compared with oral administration of
the acid compound itself.157 As illustrated by this example, prodrugging of
compounds is unlikely to be successful if the PSA of the parent molecule is too
far from ideal oral drug space.159
The distribution of zanamivir and oseltamivir acid within the body is heavily
influenced by their polar physicochemical characteristics. Plasma protein
binding of zanamivir is low and that of oseltamivir negligible, whilst the more
lipophilic oseltamivir has moderate plasma protein binding in humans
(B42%). Distribution of zanamivir into red blood cells is also negligible
(BP ¼ 0.4–0.6).160,161 The steady-state volume of distribution (Vss) in humans
of both zanamivir and oseltamivir acid is low (0.2–0.4 L kg1) and similar to the
volume of extracellular fluid. The volume of distribution in dog was similar to
that in human (0.3 L kg1) but higher volumes of distribution (1, 3 and
9 L kg1) were seen in rat, mouse and ferret, respectively.157
Some concerns about neuropsychiatric side effects of oseltamivir arising in
Japan162 have led to a number of studies looking at concentrations of oselta-
mivir and oseltamivir acid in the CNS. In rats following oral dosing of oselta-
mivir, the brain: plasma AUC ratio was 0.15 for oseltamivir and 0.01 for
oseltamivir acid. In vitro studies showed that rat and human S9 brain homo-
genates had little ability to convert oseltamivir to oseltamivir acid.162 The lim-
ited brain penetration of oseltamivir is in part due to transporter-mediated
efflux. Oseltamivir is a substrate for both human and murine P-gp and the brain
to plasma ratio of oseltamivir was increased (B5-fold) in P-gp knockout mice or
in wild-type mice dosed with the P-gp inhibitor GF-120918. The brain pene-
tration of oseltamivir is also lower in older rats that are known to have increased
P-gp expression in the blood–brain barrier than younger animals. The limited
brain penetration of oseltamivir acid is largely due to the poor membrane
permeability of the compound as the brain to plasma ratio of oseltamivir acid
was unaffected by the P-gp inhibitor and was similar in P-gp knockout and wild-
type mice,155,163 although other transporters may have a role in limiting osel-
tamivir acid brain penetration. The penetration of oseltamivir and oseltamivir
acid into the CNS of humans was also low (CSF: free plasma concentration
ratio was o0.05) and CSF Cmax was delayed relative to that in the plasma
suggesting a barrier to free permeability of the compounds in the CNS.164
Zanamivir and oseltamivir acid are highly polar and are almost exclusively
(87–90%) excreted into urine as unchanged drug following intravenous
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 431
160
dosing. The renal clearance of zanamivir was similar to the glomerular fil-
tration rate (GFR) in humans suggesting minimal reabsorption of zanamivir
(a highly polar; water soluble compound) occurs in the kidney presumably due
to its poor membrane permeability. Renal clearance of oseltamivir acid
(B4.5 ml min1 kg1) in humans is greater than GFR suggesting active secre-
tion into the urine occurs. In vitro studies have shown that oseltamivir acid is a
substrate and weak inhibitor (Ki 45 mM) of human renal transporter
hOAT1165 and renal Cl in humans in vivo is inhibited by probenecid (a known
hOAT inhibitor).166 In the presence of probenecid, the renal Cl of oseltamivir
acid is roughly equal to GFR suggesting that the active component of oselta-
mivir acid is completely inhibited by probenecid. Oseltamivir itself is also
actively secreted in to human urine (mean renal Cl 360–480 ml min1, B6 ml
min1 kg1), although the responsible transporter has not been identified.167
Oseltamivir acid is the only metabolite of oseltamivir identified in humans. In
rats, however, some additional minor metabolites have been identified of which
the R-omega (o) carboxylic acid metabolite was the most abundant (metabo-
lism occurs on the terminal carbon of the pentoxy group). Rat liver microsomes
in the presence of NADPH metabolise oseltamivir to the o-hydroxy metabo-
lite—a first step on the pathway to the acid.168
As with other ester prodrugs,158 the tissues hydrolysing oseltamivir to osel-
tamivir acid differ across species. The hydrolysis of oseltamivir is rapid in rat
plasma but relatively slow in ferret, dog and human plasma.157,169 Oseltamivir
is also hydrolysed to oseltamivir acid in human liver. There are two major
carboxylesterase isozymes with differing substrate specificity in human liver,
HCE-1 and 2.170 Hydrolysis of oseltamivir in human liver correlates with HCE-
1 activity and the Km in human liver microsomes (187 mM) and recombinant
HCE-1(177 mM) were similar. The rate of hydrolysis of oseltamivir was shown
to vary when some naturally occurring variants of HCE1 were assessed for
catalytic activity. HCE1S58N slightly increased the hydrolysis (B25%)
whereas variants HCE1C70F and HCE1R128H markedly decreased the
hydrolysis.171 To date the impact of these mutations on oseltamivir human
pharmacokinetics has not been reported.
Once formed from orally administered oseltamivir, the elimination half-life
of oseltamivir acid in humans (6–10 h) is considerably longer than that seen
following intravenous administration of oseltamivir acid itself (2 h);154 a similar
phenomenon is seen in pre-clinical species.154 As oseltamivir itself is rapidly
absorbed, the longer plasma half-life of oseltamivir acid seen after oral
administration of oseltamivir probably reflects the rate at which the acid
metabolite is formed by hydrolysis and crosses the sinusoidal membrane of the
liver to re-enter the blood. Given the polarity of oseltamivir acid and its poor
membrane permeability, it is likely that transporters play a role in the re-entry
of the compound in to the circulation from the liver.
Attempts have been made to increase the half-life and retention of zanamivir
in the lung using polymers of di-, tri- and tetrameric zanamivir molecules linked
by lipophilic spacer chains, and by making a dextran linked polymer of zana-
mivir. These modifications have resulted in molecules with greater potency and
432 Chapter 9
172
a longer residence time in in vivo models of influenza infection. The zana-
mivir dimer shown in Table 9.7 is 100 times more potent than zanamivir in a
mouse influenza model, and levels in rat lung one week after dosing are 100-
fold higher than monomeric zanamivir.172 In rat studies where the pharma-
cokinetics of zanamivir and the dimer were directly compared, the dimer had
a two-fold lower Cl than zanamivir and a larger although still small volume
(0.2 vs. 0.08 L kg1) and has high F% (67) after intratracheal dosing.

9.3.3 N-Ribosyltransferase TSAI

N-Ribosyltransferase enzymes catalyse the transfer of ribose or deoxyribose
moieties from nucleosides and nucleotides.173 Purine nucleoside phosphorylase
(PNP) (E.C. 2.4.2.1) is a key enzyme in the purine salvage pathway; it catalyses
the reversible cleavage of purine nucleosides to the corresponding purine free
base and ribo- or deoxyribose-1-phosphate.174–177 Inherited PNP deficiency in
children leads to selective depletion of their T-lymphocytes making PNP
inhibition an attractive target for T-cell proliferative disorders (including organ
transplant rejection and T-cell lymphoma and leukaemia).178–185
The catalytic mechanism of PNP is summarised in Figure 9.15. The doubly
ionised phosphate, which is stabilised by His86, attacks the anomeric centre of
the ribose. The ribose–purine bond is weakened by charge separation, resulting
in oxonium cation formation enabling facile attack by the phosphate dia-
nion,174,186 The transition states of human, bovine and Plasmodium falciparum
PNP have been studied in great detail with active site mutagenesis and kinetic
isotope experiments used to design a series of more potent TSAI of PNP with
drug-like attributes.179 The key features incorporated into the TSAI were the
oxocarbenium ion character of the ribosyl group and a C1 substitution without
the potential to become a leaving group.186 Some of the compounds arising
from this initiative are shown in Figure 9.16.187,188
Forodesine is a potent inhibitor of bovine (23 pM) and human (66 pM)
PNP175,188 due to favourable hydrogen bonding interactions in the active site
and effective mimicry of the developing positive charge of the early transition
state.186,188 The human transition state of PNP was found to have a larger
C1 0 –N distance and increased cationic character compared to the bovine
enzyme; this explained the potency drop for forodesine in the human enzyme.
To effectively act as inhibitors of human PNP, a series of 4-deaza-1-aza-2-
deoxy-1,9- methylene immucillin compounds were designed where the cationic
charge was moved to the 1 0 - position of the iminoribitol ring and a methylene
was added between the ribosyl ring and the nucleoside base.
Of these analogues, BCX-4208 was a potent (Kd 16 pM) inhibitor of human
PNP. BCX-4208 is orally available in mice and a single oral dose causes inhi-
bition of murine red blood cell PNP. Once inhibited by BCX-4208, the return of
PNP activity to pre-dose levels principally reflects red blood cell replacement
(half-life for recovery of PNP activity 11.5 days),191,192
 Asn243 Asn243
Asn243
H H
H O N O N
O N H H
N H N N
NH2 NH2 NH2
H
N N N N
N N N N
- N
H N H N H
His257 N His257 N His257 N N
O O + O
O O O
O O O O
O O P O O O OH OH P O
P
HO O HO O HO O
H H
N N H
N N N N
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme

His86 His86 His86

Figure 9.15 Proposed mechanism for PNP-catalysed phosphorolysis.

433
434 Chapter 9

O O
N N N N
OH N OH
N
H
N
N
OH OH OH OH

Forodesine BCX-4208

Figure 9.16 TSAIs of PNP.174,188–190

Table 9.8 Pharmacokinetics of forodesine.183,193,195,196

Species CL (ml/min/kg) Vss (L/kg) F% Fa

Mouse a 63
Rat B7 B0.6 (Vz) 7–22
Dog 100
Monkey 5–10 3–13
Human B1.5 B1.3 30–50 32–53
a ¼ terminal volume.

Forodesine (BCX-1777, immucillin-H) is the most advanced of the PNP

TSAI and has been studied in phase I/II clinical trials for T- and B-cell
mediated disorders. Forodesine is polar (clogP -2.4, PSA 134 Å2) and inhibits
the proliferation of activated human T-cells (IC50 ¼ 0.05 mM).181–183 The
pharmacokinetics of forodesine are summarised in Table 9.8.
Absorption of forodesine was high in mice and dog after oral administration,
although flip-flop kinetics were observed with the oral half-life (2.2 h) being
longer than the intravenous half-life (0.8 h). Given the physicochemistry of
forodesine, it is likely that the compound is absorbed either paracellularly or via
an active transporter-mediated transcellular pathway.183 Oral F% was lower in
the rat and monkey193 and absorption was slow (Tmax 4–11 h) and incomplete
in both species and saturated as the dose of compound was increased. Changing
the formulation [5 mg kg1 in citrate buffered (pH 6.7) sodium lauryl sulfate] of
forodesine failed to improve the oral absorption and bioavailability in the
monkey.193
Despite the high polarity of forodesine, the compound was also well absor-
bed in humans. In clinical trials in cancer patients, the bioavailability of for-
odesine is 30–50% with some evidence of saturation of bioavailability at higher
doses of the drug. In addition to oral exposure, pharmacodynamic effects of
forodesine were elicited after oral dosing with all patients demonstrating ele-
vated levels of the dGuo biomarker.194
The clearance of forodesine was low in rats and humans, moderate in mouse
and high in monkey. In keeping with the physicochemical properties of for-
odesine, in vitro studies using liver microsomes from mice, rats, primates and
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 435
humans did not produce measurable metabolites suggesting that CYP meta-
bolism will have a minor role in clearance of the compound. In humans, 40–
70% of an intravenous forodesine dose was recovered in the urine unchanged in
the first 24 h after dosing suggesting that this is the major route of clearance of
the compound.83,193,195,196
The volume of distribution of forodesine was Z body water in rat, humans
and mouse suggesting that, despite its small polar nature, forodesine can dis-
tribute into the tissues of the body. The half-life of forodesine was short in all
pre-clinical species (0.8–3 h) but reasonably long in humans (median half-life
of 11 h) (Table 9.8 and references cited therein).

9.3.4 Nucleoside Deaminase TSAI

9.3.4.1 Purine Deaminase Inhibitors
Pentostatin (2 0 -deoxycoformycin) (Figure 9.17) is a very potent (2.5 pM against
human erythrocyte enzyme) TSAI of adenosine deaminase (ADA) and is
approved for the treatment of hairy cell leukaemia. In addition pentostatin has
been used to boost the pharmacokinetics of other cytotoxic drugs that are
metabolised by ADA (e.g. vidaribine).197
Pentostatin undergoes rapid acid catalysed glycosidic cleavage (T1/2o6 min
at pH 1), which makes oral administration challenging due to the low pH in the
stomach.197 As would be expected for a small, polar compound, the Vss of
pentostatin following intravenous dosing in humans is low (B0.6 L kg1) and
close to the volume of total body water. Pentostatin is believed to enter cells via
nucleoside transporters and shows extensive distribution into red blood cells
(blood to plasma ratio is 2–4).259 Pentostatin is essentially free in human plasma
(PPB 3.8%) (PPB ¼ plasma protein binding) 198 and, not surprisingly for a polar
compound, CSF penetration of pentostatin is low (10–25% of serum levels). 197
However, studies in rats show that 90–95% inhibition of ADA in the brain can
be achieved following intraperitoneal injection of pentostatin.199 Clearance of
pentostatin is also low (B1.4–1.9 ml min1 kg1) and the human half-life is
about 6 h in patients with normal renal function. Clearance (total and renal) is

OH
NH2 HO NH2
N
N N
N N
N
N N N O
O N O N OH
H N OH
OH
HO OH
HO OH
OH

pentostatin adenosine ADA transition state intermediate

Figure 9.17 Chemical structure of pentostatin, the substrate of ADA (adenosine) and
the proposed transition state intermediate of ADA.
436 Chapter 9
198
lower and half-life longer in patients with impaired renal function and a large
proportion of pentostatin is excreted unchanged in the urine in humans (esti-
mates vary from 20–66, 32–48, 50–73 up to 90% in different studies) and ani-
mals (mice 91–100%, rats 81–86%, dogs 54–83%).198,199,197,259

9.3.4.2 Pyrimidine Deaminase Inhibitors

Important anticancer agents which are cytidine analogues (including cytosine
arabinoside (Ara-C), gemcitabine, capecitabine, decitabine and 5-azacyti-
dine)200 are converted to inactive or toxic metabolites by cytidine deaminase
(CDA) resulting in short in vivo half-lives and poor bioavailabilities of these
drugs.201,202 As a result, inhibition of CDA has emerged as a target to enhance
the efficacy and reduce the toxicity of chemotherapeutic cytidine
analogues.203,204
Crystallographic studies of CDA–ligand complexes have shown that there
are four key features of the enzymatic mechanism of CDA:205

(i) The zinc atom in the active site activates water to OH, which subse-
quently attacks the C4 position of substrate.
(ii) The resulting tetrahedral transition state is stabilised by a strong
interaction with the carboxylate group of the enzyme present in the
vicinity.
(iii) Intramolecular proton transfer is mediated entirely by the carboxylate
group.
(iv) Stabilisation of the leaving NH3 by hydrogen bonds to the protein
backbone carbonyl oxygen of T127.206

A schematic of the transition state intermediate of cytidine deaminase is

shown in Figure 9.18.
Zebularine and 3,4,5,6-terahydrouridine (THU) are rationally designed
inhibitors of cytidine deaminase that share structural features with the pro-
posed transition state. Replacing the NH3 group of cytidine with a proton
(zebularine) produces a compound that reacts with water to produce a tran-
sition state analog inhibitor of CDA in which the conversion of the tetrahedral
transition state into product is prevented. The 4-OH compound binds to the
enzyme l07-fold more potently than zebularine itself.207 It is also likely that
zebularine can form a stable alkoxide-like transition state analogue inhibitor in
the active site of CDA following nucleophilic attack by the Zn21-hydroxide
group of the enzyme with two concerted proton transfers.208 As well as acting
as an inhibitor of CDA, it has also been shown that zebularine can also form a
covalent complex with DNA methyltransferases and has gene silencing
activity.209,210
The amine group of cytidine is replaced by a hydroxyl group in THU which
mimics the transition state of CDA yielding a potent, competitive inhibitor [Ki
28–240 nM (7–60 ng ml1)].211,212 Complete inhibition of CDA in vitro is
achieved with THU concentrations of 1 mg ml1.213
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 437

NH2 O
H2N OH
N H2O
NH NH

N O N O N O
NH3
R R R
cytidine uridine

H OH
H OH
N + H2O NH
NH

N O N O N O
R R R
zebularine tetrahydrouridine

Figure 9.18 Proposed scheme of hydrolytic deamination of cytidine deaminase and

structures of zebularine, 4-hydroxyzebularine and tetrahydrouridine
(R ¼ ribose).206

Like the transition state of CDA that they mimic, zebularine (MW 228, clogP
2.1) and THU (MW 248, 1.90) are small polar compounds. Both com-
pounds have low oral F% due to a combination of extensive first-pass meta-
bolism and limited passive permeability and poor absorption.201,214–216 In
contrast to the low oral bioavailability, subcutaneous bioavailability of THU
was high (F% 84) in humans.201,216 The pharmacokinetics of zebularine and
THU are summarised in Table 9.9.
Despite the compounds having similar physicochemistry, their clearance
mechanisms are quite different. THU is excreted in the urine predominantly as
unchanged drug (80–90%) after intravenous dosing in humans, rats, dogs and
monkeys.216,218 In the mouse renal excretion of unchanged drug is lower (55%)
and renal clearance (5.0–6.2 ml min1 kg1) is less than GFR in this species
(B10 ml min1 kg1)219 suggesting that there is some renal reabsorption of this
polar compound in the mouse following glomerular filtration.217 The clearance
of THU in mice is low compared to other pyrimidine analogues because the
compound is not a substrate for the enzymes (thymidine phosphorylase or
uridine phosphorylase) that typically cleave the ribose moiety of pyrimidine
analogues.217 In humans, no metabolites of THU were seen after intravenous
or subcutaneous dosing, but after oral dosing of 14C THU, the radiochemical
and biochemical assays did not give the same results suggesting that a sig-
nificant part of the dose was metabolised to inactive species. This may be
due to metabolism by intestinal microflora as well as to first-pass intestinal
metabolism.216
In contrast to THU, zebularine is extensively metabolised. Enzymatic
degradation of zebularine occurs in both rat and human plasma, although the
 438

Table 9.9 Pharmacokinetics of zebularine and THU.214–218

Compound CL (ml/min/kg) PPB (%) Volume (L/kg) Oral F% Fa (%) Half-life (hours)
b
Zebularine 63 m 43 m 1.1 m 7m 0.7 m
14 m (100 mpk) 73 rb 0.2–0.3 r B3–6 r 6r
4r 34 d 0.4-0.7 mk o1 mk 1.2–2.5 mk
4–11 mk 42 mk
57 h
THU 9m B0 m, h 0.9 m 15–23 m 18–29 ma 1.2 m
B1.6 h B0.18 h 10-14 h B14 ha 1–1.5 r
1.1 d
1.2 mk
7.2 h
h ¼ human, m ¼ mouse, mk ¼ monkey, r ¼ rat.
a
based on urinary recovery of radioactivity
b
@ 1 mg ml1
mpk ¼ mg per kg
Chapter 9
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 439
Table 9.10 Metabolism of zebularine by aldehyde oxidase in hepatic cytosol
from different species.214,257
Vmax Clint
(nmol/min/mg (ml/min/mg In vivo CL
Species Sex Km (mM) protein) protein) (ml/min/kg) F%

Mouse M 102 11 108 14 7

F 27 0.2 7
Rat M 11 0.9 81 4 3–6
F 11 0.9 81
Dog M No metabolic activity
Monkey M 15 4.6 307 6–10 o1
Human M 7.3 2.1 288
F 8.4 2.7 321

responsible enzymes were not identified.214 The 2-hydroxypyrimidine base of

zebularine is also a substrate for aldehyde oxidase (AO).220,221
The ability of liver cytosol from different species to metabolise zebularine has
been studied in vitro (Table 9.10).220 The intrinsic clearance of zebularine in
liver cytosol was highest in monkey and human. Reasonably high levels of
metabolism was also seen in the rat and male mouse liver cytosol with low levels
of metabolism seen in the female mouse liver cytosol and no observable
metabolism seen in the dog liver cytosol. AO is also known to be expressed in
the intestine222 and as such may play a role in intestinal first-pass metabolism
(Fg) of zebularine limiting it’s oral bioavailability.214
Quantitatively, AO was shown to convert 38% of an intravenous dose of
[14C]-zebularine in the mouse to uracil with another 30% of the dose excreted as
unchanged drug in the urine via an active process (CLr4GFR),215,220 although
the transporters involved have not been identified. Once zebularine is oxidised
by AO, the radioactive material is further metabolised by endogenous meta-
bolic pathways and a significant proportion of the dose is ultimately lost as
expired 14CO2.
The metabolism of zebularine in vivo proceeds as shown in Figure 9.19 with
uridine, uracil and dihydrouracil identified as the primary metabolites of
zebularine in mice.215 There was no evidence of 2-pyrimidinone as a metabolite
in either plasma or urine suggesting that, like cytidine, zebularine is not a
substrate for uridine phosphorylase and must first be converted to uridine
before the N-glycosidic bond can be cleaved.
Zebularine and THU have low to moderate volumes of distribution (Table
9.9)214,216 suggesting the compounds do not distribute widely into tissues.
Studies in mice with radiolabelled zebularine also showed a modest distribution
of zebularine to tissues.215 In keeping with the polar nature of zebularine, the
compound showed minimal distribution into red blood cells223 and fat cells,
and the compound had low penetration of the blood–brain and the blood–
testes barrier.215
One use of zebularine and THU is to prolong the activity of anticancer drugs
that are substrates for CDA (e.g. Ara-C) and this aspect has been studied in
440 Chapter 9
O
O
N NH
HO NH
N O HO Uridine
O N O
AO O Phosphorylase N O
H

HO OH OH
HO
Uracil
Zebularine Uridine
Dihydropyrimidine DH

NH
CO2 + NH3 + β-alanine
N O
H

dihydrouracil

Figure 9.19 Metabolic conversion of zebularine to uridine by aldehyde oxidases.220

animals and humans. Zebularine increases (two-fold) the steady state plasma
levels of 5-aza-2 0 -deoxycytidine when the compounds are co-infused in mice,
whilst concomitant administration of THU markedly increases the myelosup-
pressive potency of Ara-C in monkeys.224 In humans, simultaneous oral
administration of THU (10–50 mg kg1) and Ara-C resulted in a two-fold
increase of blood levels of Ara-C at all time points over a four hour period and
a slight increase in the half-life of Ara-C. The inhibitory effect of THU on CDA
metabolism of Ara-C was also reflected in a considerably increased ratio of
THU/uracil arabinoside excretion in the urine.225 Although there is a theore-
tical risk of toxicity from zebularine administration due to the large amount of
uridine formed, in reality this does not seem to be an issue.226–230

9.3.5 Inosine 5-monophosphate Dehydrogenase TSAI

Inosine 5-monophosphate dehydrogenase (IMPDH) catalyses the conversion
of inosine monophosphate to xanthine monophosphate with reduction of
NAD; this is the rate limiting step in de novo guanine nucleotide biosynthesis.
As cell proliferation is strongly linked to the size of the guanine nucleotide pool,
IMPDH inhibition has emerged as an attractive therapeutic target.231
Mizoribine (formerly known as bredinin) is an imidazole nucleoside anti-
biotic (Figure 9.20) that is phosphorylated to produce mizoribine 5 0 -mono-
phosphate, a transition state analogue inhibitor of IMPDH.232 Mizoribine has
been increasingly used in Japan since 1978 as a substitute to azathioprine for
clinical renal transplantations.231,233–236
Mizoribine is a polar nucleoside analogue (clogP 1.4, TPSA 151 Å2) and as
such is expected to have limited passive membrane permeability. However,
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 441

N NH2

O N
O
OH
HO OH
HO

Figure 9.20 Chemical structure of mizoribine.

when used as a therapeutic in humans, mizoribine is administered orally and is

almost completely absorbed.237 In the presence of inosine, the absorption rate
of mizoribine is significantly decreased238 suggesting a role for the nucleoside
transporter CNT2/N1 (model substrates are inosine, guanosine and formycin-
B) in the active absorption of mizoribine.239
As expected for a polar compound, renal clearance (21 mL min1) accounts
for 56% of total clearance of mizoribine in the dog. Infusion of mizoribine into
the renal artery showed that renal clearance in the dog decreases with
increasing dose, indicating that the renal elimination pathway was a saturable
transporter-mediated pathway. The half-life was longer in renal-transplanted
animals than in healthy dogs, reflecting the better renal function in healthy
animals.240
The pharmacokinetics of mizoribine has also been determined in healthy
humans and in renal transplant patients.237 Following oral doses (3–12 mg kg1)
to healthy volunteers, the half-life was B3 h and CL/F was B3.2 ml min1 kg1.
Renal excretion of unchanged drug in to the urine accounted for 65–100% of
the dose in humans. In renal transplant patients, there was significant variation
in the terminal half-life and trough concentrations of mizoribine that were
dependent on the degree of renal function of the individual.241,242
Mizoribine has a moderate volume of distribution in dogs (0.5 l kg1) and
humans (0.8 l kg1) in keeping with its polar nature.240,243 An oral study with
14
C-mizoribine (radioactivity Tmax ¼ 1.5 h) in rats also showed moderate tissue
distribution of radioactivity with high levels of radioactivity observed in the
gastrointestinal tract, kidneys, liver, spleen and thymus.258

9.3.6 Aspartate Carbamyl Transferase TSAI

Phosphonoacetyl-L-acetate (PALA) inhibits aspartate carbamyl transferase
(ACT) and blocks de novo synthesis of pyrimidines in vitro and in vivo.244
Although cytotoxic to tumour cells in vitro, it was shown to be devoid of
antitumour effect in vivo and more recently has been assessed for its ability to
enhance the antitumour activity of 5-fluorouracil. PALA binds about 1000
times more potently to the enzyme (Ki 27 nM) than carbamyl phosphate for
which it is a competitive inhibitor. However, in PALA the phosphate group is
replaced by a C-linked phosphonate which cannot act as a leaving group
(Figure 9.21) and PALA may act as a multi-substrate analogue rather than as a
true TSAI.245
442 Chapter 9

O HO2C O
HO2C H2
N O
NH2 O H C
+ HN
2 P
P CO2H O
CO2H O O O
O

PALA
HO2C O
N O
H O
H2N P
CO2H O
O

HO2C O
N O O
H + P
NH2 O
CO2H O

Figure 9.21 Reaction catalysed by aspartate carbamyl transferase, the proposed

transition state and the structure of PALA.

As would be expected for such a polar compound (clogP 2.2, clogD 8.9),
the pharmacokinetics in animals and humans are characterised by poor oral
bioavailability, elimination unchanged by the kidney, little metabolism and
distribution into extracellular water with little tissue affinity.244 In humans,
clearance was B1.6 ml min1 kg1 and renal excretion of unchanged drug was
80–85%. The renal CL of PALA approximates to the glomerular filtration rate,
suggesting that CL may be by passive filtration with no reabsorption (though
equivalent rates of active renal secretion and reabsorption can not be ruled
out). The volume of distribution was B0.3 L kg1 reflecting the limited dis-
tribution of PALA into tissues and the half-life in man was 5–12 h.

9.3.7 Glycosidase Inhibitor TSAI

Acarbose is a polar (clog D 4.2), high molecular weight (646) potent inhibitor
of a-glycosidase inhibitor. Structurally, acarbose has an amine group in the
position of the protonated exocyclic oxygen in the transition state (Figure 9.22).
As expected from its complex polysaccharide structure, acarbose is a very polar
compound (PSA 329 Å2) with high solubility. Acarbose has 19 hydrogen bond
acceptors and 14 hydrogen bond donor groups.
The main site of action of acarbose is in the luminal side of the intestine and
therefore absorption and systemic exposure of acarbose is not needed for its
efficacy, which is just as well as the reported bioavailability in humans for
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 443

OH H
+
HO O
R
O
HO

HO

Putative transition state with

protonated exocyclic oxygen

OH
OH
O
HO
OH OH OH

HO N O O
O

HO HO O OH
OH OH
HO
Acarbose

Figure 9.22 Chemical structure of acarbose and the putative transition state of the
glycosidase enzyme.

acarbose is 0.5–1.7% and across a number of pre-clinical species absorption

ranged from 1–4%.
Following intravenous administration to humans, Vss is 0.32 L kg1 and a
similar Vss is seen in other species. This suggests that acarbose has limited
distribution beyond the extracellular fluid volume. As would be expected for
such a polar compound, protein binding of acarbose is negligible.
Acarbose undergoes significant degradation in the gut due to hydrolysis of the
glucosidic linkages. It is likely that the gut microflora plays a role in this meta-
bolism. When dosed intravenously, acarbose is predominantly excreted unchan-
ged in the urine (79% rat, 95% dog and 94% human). The clearance of acarbose
is about 2 mL min1 kg1 in humans, which is similar to GFR.21–23,246,4

9.4 Conclusions
In general, the group of transition state analogues discussed in this chapter
have a relatively high PSA, particularly when this value is compared with the
PSA value of the top 200 selling drugs. This high polar surface area imparts a
number of pharmacokinetic properties to this group of compounds.
Compounds with low clog P and high TPSA typically had poor membrane
permeability and low transcellular oral absorption. This has led to a number of
444 Chapter 9
compounds either being dosed as prodrugs or being given by other routes such
as intravenous, subcutaneous or inhaled administration. Even when this type of
transition state analogue compound shows good absorption and high bioa-
vailability, this can often be attributed to paracellular or transport-mediated
absorption rather than to good transcellular permeability.
One strategy to improve oral absorption of polar transition state analogues is
to increase clogP significantly and there are examples such as the HIV protease
inhibitors where the compounds have both high clogP and high TPSA yet still
show good absorption. The downside is that the extra lipophilicity added to
increase absorption also makes this class of compounds susceptible to rapid
metabolic clearance and frequently F% is low because the higher absorption is
counteracted by higher Cl and first-pass extraction.
Other properties of this group of compounds such as clearance and volume
of distribution are as would be expected from their physicochemical properties.
Polar compounds tend to have a low volume of distribution and are often
excreted unchanged into the urine. More lipophilic compounds have relatively
higher volumes of distribution and tend to be rapidly cleared by metabolism. In
some cases (e.g. boronic acid and activated ketones), the metabolism involves
the functional groups mimicking the transition states whilst in many other
compounds the metabolism occurs on the molecular scaffold.
The high PSA of these compounds also suggests that attempting to inhibit
central (behind blood–brain barrier) targets with TSAI will be challenging. The
affinity of many of the compounds to P-glycoprotein will exacerbate the
challenge of inhibiting central targets.
Striking a balance between absorption and clearance is the main challenge
for scientists looking to develop transition state analogues as therapeutic
agents. However, the high potency (and often slow offset from the target) of
these compounds means that, if the balance between absorption and clearance
can be achieved, useful therapeutic agents can be discovered.

9.5 Abbreviations
ACE angiotensin converting enzyme
ACT aspartate carbamyl transferase
ADA adenosine deaminase
ADME absorption, distribution, metabolism and excretion
ANG I angiotensin I
ANG II angiotensin II
AO aldehyde oxidase
Ara-C cytosine arabinoside
BP blood : plasma ratio
Cav0t average concentration from 0 to time t
CDA cytidine deaminase
CLh hepatic blood clearance
CNS central nervous system
CSF cerebrospinal fluid
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 445
CYP cytochrome P450
DDI drug–drug interaction
DPP-IV dipeptidyl peptidase IV
F bioavailability
Fa fraction absorbed
Fg fraction escaping first-pass metabolism in the gastrointestinal
tract
Fh fraction escaping first-pass clearance in the liver
GFR glomerular filtration rate
HBA hydrogen bond acceptor
HBD hydrogen bond donor
HCV Hepatitis C virus
HIV human immunodeficiency virus
IMPDH inosine 5-monophosphate dehydrogenase
MTAP methylthioadenosine phosphorylase
OATP organic anion polypeptide
P-gp P-glycoprotein
PALA phosphonoacetyl-L-acetate
PNP purine nucleoside phosphorylase
Q hepatic blood flow
THU 3,4,5,6-terahydrouridine
TPSA topological polar absorption area
TSAI(s) transition state analogue inhibitors(s)
Vss steady state volume of distribution

References
1. J. G. Robertson, Biochemistry, 2005, 44, 5561.
2. V. L. Schramm, Curr. Opin. Struct. Biol., 2005, 15, 604.
3. L. Pauling, Nature, 1948, 161, 707.
4. V. H. Lillelund, H. H. Jensen, X. Liang and M. Bols, Chem. Rev., 2002,
102, 515.
5. V. L. Schramm, Annu. Rev. Biochem., 1998, 67, 693.
6. R. Wolfenden, Nature, 1969, 223, 704.
7. R. Wolfenden, Annu. Rev. Biophys. Bioeng., 1976, 5, 271.
8. R. Wolfenden and M. J. Snider, Acc. Chem. Res., 2001, 34, 938.
9. B. A. Malcolm, R. Liu, F. Lahser, S. Agrawal, B. Belanger, N. Butkie-
wicz, R. Chase, F. Gheyas, A. Hart, D. Hesk, P. Ingravallo, C. Jiang,
R. Kong, J. Lu, J. Pichardo, A. Prongay, A. Skelton, X. Tong, S. Ven-
katraman, E. Xia, V. Girijavallabhan and F. G. Njoroge, Antimicrob.
Agents Chemother., 2006, 50, 1013.
10. S. Venkatraman, S. L. Bogen, A. Arasappan, F. Bennett, K. Chen, E. Jao,
Y. T. Liu, R. Lovey, S. Hendrata, Y. Huang, W. Pan, T. Parekh, P. Pinto,
V. Popov, R. Pike, S. Ruan, B. Santhanam, B. Vibulbhan, W. Wu,
W. Yang, J. Kong, X. Liang, J. Wong, R. Liu, N. Butkiewicz, R. Chase,
446 Chapter 9
A. Hart, S. Agrawal, P. Ingravallo, J. Pichardo, R. Kong, B. Baroudy,
B. Malcolm, Z. Guo, A. Prongay, V. Madison, L. Broske, X. Cui, K. C.
Cheng, Y. Hsieh, J. M. Brisson, D. Prelusky, W. Korfmacher, R. White,
S. Bogdanowich-Knipp, A. Pavlovsky, P. Bradley, A. K. Saksena, A.
Ganguly, J. Piwinski, V. Girijavallabhan and F. G. Njoroge, J. Med.
Chem., 2006, 49, 6074.
11. J. J. Deadman, S. Elgendy, C. A. Goodwin, D. Green, J. A. Baban,
G. Patel, E. Skordalakes, N. Chino, G. Claeson and V. V. Kakkar, et al.,
J. Med. Chem., 1995, 38, 1511.
12. P. C. Weber, S. L. Lee, F. A. Lewandowski, M. C. Schadt, C. W. Chang
and C. A. Kettner, Biochemistry, 1995, 34, 3750.
13. V. L. Schramm, Nucleic Acids Res. Suppl., 2003, 3, 107.
14. T. L. Amyes and J. P. Richard, ACS Chem. Biol., 2007, 2, 711.
15. A. Radzicka and R. Wolfenden, Methods Enzymo.l, 1995, 249, 284.
16. J. B. Cooper, Curr. Drug Targets, 2002, 3, 155.
17. T. Harrison and D. Beher, Prog. Med. Chem., 2003, 41, 99.
18. A. Brik and C. H. Wong, Org. Biomol. Chem., 2003, 1, 5.
19. M. Jaskolski, A. G. Tomasselli, T. K. Sawyer, D. G. Staples, R. L.
Heinrikson, J. Schneider, S. B. Kent and A. Wlodawer, Biochemistry,
1991, 30, 1600.
20. B. M. Sadler and D. S. Stein, Ann. Pharmacother., 2002, 36, 102.
21. Y. Etoh, M. Miyazaki, H. Saitoh and N. Toda, Jpn. J. Pharmacol., 1993,
63, 109.
22. P. Coassolo, W. Fischli, J. P. Clozel and R. C. Chou, Xenobiotica, 1996,
26, 333.
23. H. Morishima, Y. Koike, M. Nakano, S. Atsuumi, S. Tanaka,
H. Funabashi, J. Hashimoto, Y. Sawasaki, N. Mino and M. Nakano,
et al., Biochem. Biophys. Res. Commun., 1989, 159, 999.
24. C. H. Kleinbloesem, C. Weber, E. Fahrner, M. Dellenbach, H. Welker,
V. Schroter and G. G. Belz, Clin. Pharmacol. Ther., 1993, 53, 585.
25. C. Weber, H. Birnbock, J. Leube, I. Kobrin, C. H. Kleinbloesem and
P. Van Brummelen, Br. J. Clin. Pharmacol., 1993, 36, 547.
26. C. Le Tiec, A. Barrail, C. Goujard and A. M. Taburet, Clin. Pharmaco-
kinet., 2005, 44, 1035.
27. D. Back, V. Sekar and R. M. Hoetelmans, Antivir. Ther., 2008, 13, 1.
28. C. McCoy, Clin. Ther., 2007, 29, 1559.
29. B. J. Dong and J. M. Cocohoba, Ann. Pharmacother., 2006, 40, 1311.
30. R. S. Cvetkovic and K. L. Goa, Drugs, 2003, 63, 769.
31. J. H. Lin, Ernst Schering Res. Found. Workshop, 2002, 37, 33.
32. V. J. Stella and K. W. Nti-Addae, Adv. Drug Deliv. Rev., 2007, 59,
677.
33. M. B. Wire, M. J. Shelton and S. Studenberg, Clin. Pharmacokinet., 2006,
45, 137.
34. J. W. Beach, Clin. Ther., 1998, 20, 2.
35. G. L. Plosker and L. J. Scott, Drugs, 2003, 63, 1299.
36. M. Azizi, J. Mol. Med., 2008, 86, 647.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 447
37. F. Waldmeier, U. Glaenzel, B. Wirz, L. Oberer, D. Schmid, M. Seiberling,
J. Valencia, G. J. Riviere, P. End and S. Vaidyanathan, Drug Metab.
Dispos., 2007, 35, 1418.
38. J. E. John, Chem. Biol. Drug Des., 2009, 73, 367.
39. H. D. Kleinert, S. H. Rosenberg, W. R. Baker, H. H. Stein, V. Klinghofer,
J. Barlow, K. Spina, J. Polakowski, P. Kovar and J. Cohen, et al., Science,
1992, 257, 1940.
40. A. Seelig, Eur. J. Biochem., 1998, 251, 252.
41. S. F. Zhou, Xenobiotica, 2008, 38, 802.
42. G. C. Williams and P. J. Sinko, Adv. Drug Deliv. Rev., 1999, 39, 211.
43. S. J. Mouly, M. F. Paine and P. B. Watkins, J. Pharmacol. Exp. Ther.,
2004, 308, 941.
44. J. W. Polli, J. L. Jarrett, S. D. Studenberg, J. E. Humphreys, S. W.
Dennis, K. R. Brouwer and J. L. Woolley, Pharm. Res., 1999, 16, 1206.
45. S. Vaidyanathan, V. Jarugula, H. A. Dieterich, D. Howard and W. P.
Dole, Clin. Pharmacokinet., 2008, 47, 515.
46. V. J. Wacher, J. A. Silverman, Y. Zhang and L. Z. Benet, J. Pharm. Sci.,
1998, 87, 1322.
47. A. Galetin, L. K. Hinton, H. Burt, R. S. Obach and J. B. Houston, Curr.
Drug Metab., 2007, 8, 685.
48. M. Gertz, J. D. Davis, A. Harrison, J. B. Houston and A. Galetin, Curr.
Drug Metab., 2008, 9, 785.
49. J. Yang, M. Jamei, K. R. Yeo, G. T. Tucker and A. Rostami-Hodjegan,
Curr. Drug Metab., 2007, 8, 676.
50. M. E. Fitzsimmons and J. M. Collins, Drug Metab. Dispos., 1997, 25,
256.
51. H. H. Kupferschmidt, K. E. Fattinger, H. R. Ha, F. Follath and S.
Krahenbuhl, Br. J. Clin. Pharmacol., 1998, 45, 355.
52. K. C. Yeh, P. J. Deutsch, H. Haddix, M. Hesney, V. Hoagland, W. D. Ju,
S. J. Justice, B. Osborne, A. T. Sterrett, J. A. Stone, E. Woolf and S.
Waldman, Antimicrob. Agents Chemother., 1998, 42, 332.
53. N. C. Cohen, Chem. Biol. Drug Des., 2007, 70, 557.
54. R. Jauch, P. Schmid, W. Fischli, W. Meister, R. Maurer and G. Wendt,
Xenobiotica, 1996, 26, 285.
55. A. J. Parker and J. B. Houston, Drug Metab. Dispos., 2008, 36, 1375.
56. R. G. Tirona, B. F. Leake, A. W. Wolkoff and R. B. Kim, J. Pharmacol.
Exp. Ther., 2003, 304, 223.
57. Z. W. Ye, P. Augustijns and P. Annaert, Drug Metab. Dispos., 2008, 36,
1315.
58. S. D. Campbell, S. M. de Morais and J. J. Xu, Chem. Biol. Interact., 2004,
150, 179.
59. B. Hagenbuch and C. Gui, Xenobiotica, 2008, 38, 778.
60. O. Janneh, E. Jones, B. Chandler, A. Owen and S. H. Khoo, J. Anti-
microb. Chemother., 2007, 60, 987.
61. C. L. Cooper, R. P. van Heeswijk, K. Gallicano and D. W. Cameron,
Clin. Infect. Dis., 2003, 36, 1585.
448 Chapter 9
62. M. M. Foisy, E. M. Yakiwchuk and C. A. Hughes, Ann. Pharmacother.,
2008, 42, 1048.
63. B. M. Sadler, P. J. Piliero, S. L. Preston, P. P. Lloyd, Y. Lou and D. S.
Stein, Aids, 2001, 15, 1009.
64. R. Wood, J. Eron, K. Arasteh, E. Teofilo, C. Trepo, J. M. Livrozet, J.
Yeo, J. Millard, M. B. Wire and O. J. Naderer, Clin. Infect. Dis., 2004, 39,
591.
65. M. B. Wire, K. L. Baker, L. S. Jones, M. J. Shelton, Y. Lou, G. J. Thomas
and M. M. Berrey, Antimicrob. Agents Chemother., 2006, 50, 1578.
66. A. Winston, M. Bloch, A. Carr, J. Amin, P. W. Mallon, J. Ray, D.
Marriott, D. A. Cooper and S. Emery, J. Antimicrob. Chemother., 2005,
56, 380.
67. H. Khanlou, L. Bhatti and C. Farthing, J. Acquir. Immune Defic. Syndr.,
2006, 41, 124.
68. J. Molto, J. R. Santos, M. Valle, C. Miranda, J. Miranda, A. Blanco, E.
Negredo and B. Clotet, Ther. Drug. Monit., 2007, 29, 648.
69. V. J. Sekar, E. Lefebvre, M. De Pauw, T. Vangeneugden and R. M.
Hoetelmans, Br. J. Clin. Pharmacol., 2008, 66, 215.
70. D. M. Burger, R. E. Aarnoutse, J. P. Dieleman, I. C. Gyssens, J. Nouwen,
S. de Marie, P. P. Koopmans, M. Stek Jr. and M. E. van der Ende,
Antivir. Ther., 2003, 8, 455.
71. D. W. Haas, B. Johnson, J. Nicotera, V. L. Bailey, V. L. Harris, F. B.
Bowles, S. Raffanti, J. Schranz, T. S. Finn, A. J. Saah and J. Stone,
Antimicrob. Agents Chemother., 2003, 47, 2131.
72. A. Hsu, G. R. Granneman and R. J. Bertz, Clin. Pharmacokinet., 1998,
35, 275.
73. F. S. Rhame, S. L. Rawlins, R. A. Petruschke, T. A. Erb, G. A. Winchell,
H. M. Wilson, J. M. Edelman and M. A. Abramson, Antimicrob. Agents
Chemother., 2004, 48, 4200.
74. R. P. van Heeswijk, A. I. Veldkamp, R. M. Hoetelmans, J. W. Mulder, G.
Schreij, A. Hsu, J. M. Lange, J. H. Beijnen and P. L. Meenhorst, Aids,
1999, 13, F95.
75. M. Kurowski, B. Kaeser, A. Sawyer, M. Popescu and A. Mrozikiewicz,
Clin. Pharmacol. Ther., 2002, 72, 123.
76. A. Hsu, G. R. Granneman, G. Cao, L. Carothers, T. el-Shourbagy, P.
Baroldi, K. Erdman, F. Brown, E. Sun and J. M. Leonard, Clin. Phar-
macol. Ther., 1998, 63, 453.
77. C. Merry, M. G. Barry, F. Mulcahy, M. Ryan, J. Heavey, J. F. Tjia, S. E.
Gibbons, A. M. Breckenridge and D. J. Back, Aids, 1997, 11, F29.
78. N. Buss, P. Snell, J. Bock, A. Hsu and K. Jorga, Br. J. Clin. Pharmacol.,
2001, 52, 255.
79. C. J. la Porte, Y. Li, L. Beique, B. C. Foster, B. Chauhan, G. E. Garber,
D. W. Cameron and R. P. van Heeswijk, Clin. Pharmacol. Ther., 2007, 82,
389.
80. T. R. MacGregor, J. P. Sabo, S. H. Norris, P. Johnson, L. Galitz and S.
McCallister, HIV Clin. Trials, 2004, 5, 371.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 449
81. S. McCallister, H. Valdez, K. Curry, T. MacGregor, M. Borin, W.
Freimuth, Y. Wang and D. L. Mayers, J. Acquir. Immune Defic. Syndr.,
2004, 35, 376.
82. B. Jarvis and D. Faulds, Drugs, 1998, 56, 147.
83. W. F. Richter, B. R. Whitby and R. C. Chou, Xenobiotica, 1996, 26, 243.
84. R. B. Kim, M. F. Fromm, C. Wandel, B. Leake, A. J. Wood, D. M.
Roden and G. R. Wilkinson, J. Clin. Invest., 1998, 101, 289.
85. N. N. Salama, E. J. Kelly, T. Bui and R. J. Ho, J. Pharm. Sci., 2005, 94,
1216.
86. B. M. Best, S. L. Letendre, E. Brigid, D. B. Clifford, A. C. Collier, B. B.
Gelman, J. C. McArthur, J. A. McCutchan, D. M. Simpson, R. Ellis, E.
V. Capparelli and I. Grant, Aids, 2009, 23, 83.
87. H. E. Wynn, R. C. Brundage and C. V. Fletcher, CNS Drugs, 2002, 16,
595.
88. A. Yilmaz, A. Izadkhashti, R. W. Price, P. W. Mallon, M. De Meulder, P.
Timmerman and M. Gisslen, AIDS Res. Hum. Retroviruses, 2009, 25, 457.
89. J. Ford, M. Boffito, A. Wildfire, A. Hill, D. Back, S. Khoo, M. Nelson, G.
Moyle, B. Gazzard and A. Pozniak, Antimicrob. Agents Chemother., 2004,
48, 2388.
90. J. Ford, D. Cornforth, P. G. Hoggard, Z. Cuthbertson, E. R. Meaden, I.
Williams, M. Johnson, E. Daniels, P. Hsyu, D. J. Back and S. H. Khoo,
Antivir. Ther., 2004, 9, 77.
91. J. Ford, E. R. Meaden, P. G. Hoggard, M. Dalton, P. Newton, I. Wil-
liams, S. H. Khoo and D. J. Back, J. Antimicrob. Chemother., 2003, 52,
354.
92. O. Janneh, R. C. Hartkoorn, E. Jones, A. Owen, S. A. Ward, R. Davey,
D. J. Back and S. H. Khoo, Br. J. Pharmacol., 2008, 155, 875.
93. M. P. McRae, C. M. Lowe, X. Tian, D. L. Bourdet, R. H. Ho, B. F.
Leake, R. B. Kim, K. L. Brouwer and A. D. Kashuba, J. Pharmacol. Exp.
Ther., 2006, 318, 1068.
94. M. Caron, M. Auclair, H. Sterlingot, M. Kornprobst and J. Capeau, Aids,
2003, 17, 2437.
95. S. G. Clarke, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 13857.
96. C. Coffinier, S. E. Hudon, E. A. Farber, S. Y. Chang, C. A. Hrycyna, S.
G. Young and L. G. Fong, Proc. Natl. Acad. Sci. U. S. A., 2007, 104,
13432.
97. W. C. Earnshaw, L. M. Martins and S. H. Kaufmann, Annu. Rev. Bio-
chem., 1999, 68, 383.
98. O. E. Levy, J. E. Semple, M. L. Lim, J. Reiner, W. E. Rote, E. Dempsey,
B. M. Richard, E. Zhang, A. Tulinsky, W. C. Ripka and R. F. Nutt, J.
Med. Chem., 1996, 39, 4527.
99. J. E. Semple, D. C. Rowley, T. K. Brunck, T. Ha-Uong, N. K. Minami, T.
D. Owens, S. Y. Tamura, E. A. Goldman, D. V. Siev, R. J. Ardecky, S. H.
Carpenter, Y. Ge, B. M. Richard, T. G. Nolan, K. Hakanson, A.
Tulinsky, R. F. Nutt and W. C. Ripka, J. Med. Chem., 1996, 39, 4531.
100. S. D. Linton, Curr. Top. Med. Chem., 2005, 5, 1697.
450 Chapter 9
101. J. C. Randle, M. W. Harding, G. Ku, M. Schonharting and R. Kurrle,
Expert Opin. Investig. Drugs, 2001, 10, 1207.
102. H. Ohbayashi, Expert Opin. Investig. Drugs, 2002, 11, 965.
103. D. S. Yamashita, R. W. Marquis, R. Xie, S. D. Nidamarthy, H. J. Oh, J.
U. Jeong, K. F. Erhard, K. W. Ward, T. J. Roethke, B. R. Smith, H. Y.
Cheng, X. Geng, F. Lin, P. H. Offen, B. Wang, N. Nevins, M. S. Head, R.
C. Haltiwanger, A. A. Narducci Sarjeant, L. M. Liable-Sands, B. Zhao,
W. W. Smith, C. A. Janson, E. Gao, T. Tomaszek, M. McQueney, I. E.
James, C. J. Gress, D. L. Zembryki, M. W. Lark and D. F. Veber, J. Med.
Chem., 2006, 49, 1597.
104. F. Koizumi, M. Murakami, H. Kageyama, M. Katashima, M. Terakawa
and A. Ohnishi, Clin. Pharmacol. Ther., 1999, 66, 501.
105. J. C. Williams, R. L. Stein, R. E. Giles and R. D. Krell, Ann. N. Y. Acad.
Sci., 1991, 624, 230.
106. P. C. Davis, R. A. Wildonger and H. S. Veale, J. Pharm. Biomed. Anal.,
1993, 11, 549.
107. J. P. Burkhart, J. R. Koehl, S. Mehdi, S. L. Durham, M. J. Janusz, E. W.
Huber, M. R. Angelastro, S. Sunder, W. A. Metz and P. W. Shum, et al.,
J. Med. Chem., 1995, 38, 223.
108. M. R. Angelastro, L. E. Baugh, P. Bey, J. P. Burkhart, T. M. Chen, S. L.
Durham, C. M. Hare, E. W. Huber, M. J. Janusz and J. R. Koehl, et al., J.
Med. Chem., 1994, 37, 4538.
109. C. A. Veale, P. R. Bernstein, C. M. Bohnert, F. J. Brown, C. Bryant, J. R.
Damewood Jr., R. Earley, S. W. Feeney, P. D. Edwards, B. Gomes, J. M.
Hulsizer, B. J. Kosmider, R. D. Krell, G. Moore, T. W. Salcedo, A. Shaw,
D. S. Silberstein, G. B. Steelman, M. Stein, A. Strimpler, R. M. Thomas,
E. P. Vacek, J. C. Williams, D. J. Wolanin and S. Woolson, J. Med.
Chem., 1997, 40, 3173.
110. P. D. Edwards, D. W. Andisik, C. A. Bryant, B. Ewing, B. Gomes, J. J.
Lewis, D. Rakiewicz, G. Steelman, A. Strimpler, D. A. Trainor, P. A.
Tuthill, R. C. Mauger, C. A. Veale, R. A. Wildonger, J. C. Williams, D. J.
Wolanin and M. Zottola, J. Med. Chem., 1997, 40, 1876.
111. D. J. Kempf, C. Klein, H. J. Chen, L. L. Klein, C. Yeung, J. T. Randolph,
Y. Y. Lau, L. E. Chovan, Z. Guan, L. Hernandez, T. M. Turner, P. J.
Dandliker and K. C. Marsh, Antivir. Chem. Chemother., 2007, 18, 163.
112. J. D. Tyndall and D. P. Fairlie, Curr. Med. Chem., 2001, 8, 893.
113. S. Venkatraman, F. Velazquez, W. Wu, M. Blackman, K. X. Chen, S.
Bogen, L. Nair, X. Tong, R. Chase, A. Hart, S. Agrawal, J. Pichardo, A.
Prongay, K. C. Cheng, V. Girijavallabhan, J. Piwinski, N. Y. Shih and F.
G. Njoroge, J. Med. Chem., 2009, 52, 336.
114. K. X. Chen, L. Nair, B. Vibulbhan, W. Yang, A. Arasappan, S. L. Bogen,
S. Venkatraman, F. Bennett, W. Pan, M. L. Blackman, A. I. Padilla,
A. Prongay, K. C. Cheng, X. Tong, N. Y. Shih and F. G. Njoroge, J.
Med. Chem., 2009, 52, 1370.
115. W. Yang, X. Gao and B. Wang, Med. Res. Rev., 2003, 23, 346.
116. K. A. Koehler and G. E. Lienhard, Biochemistry, 1971, 10, 2477.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 451
117. M. Philipp and M. L. Bender, Proc. Natl. Acad. Sci. U.S.A., 1971, 68,
478.
118. J. M. Fevig, J. Buriak Jr., J. Cacciola, R. S. Alexander, C. A. Kettner, R.
M. Knabb, J. R. Pruitt, P. C. Weber and R. R. Wexler, Bioorg. Med.
Chem. Lett., 1998, 8, 301.
119. S. L. Lee, R. S. Alexander, A. Smallwood, R. Trievel, L. Mersinger, P. C.
Weber and C. Kettner, Biochemistry, 1997, 36, 13180.
120. J. Cacciola, J. M. Fevig, P. F. Stouten, R. S. Alexander, R. M. Knabb and
R. R. Wexler, Bioorg. Med. Chem. Lett., 2000, 10, 1253.
121. C. A. Kettner, R. Bone, D. A. Agard and W. W. Bachovchin, Biochem-
istry, 1988, 27, 7682.
122. C. A. Kettner and A. B. Shenvi, J. Biol. Chem., 1984, 259, 15106.
123. I. E. Crompton, B. K. Cuthbert, G. Lowe and S. G. Waley, Biochem. J.,
1988, 251, 453.
124. S. Ness, R. Martin, A. M. Kindler, M. Paetzel, M. Gold, S. E. Jensen, J.
B. Jones and N. C. Strynadka, Biochemistry, 2000, 39, 5312.
125. G. S. Weston, J. Blazquez, F. Baquero and B. K. Shoichet, J. Med. Chem.,
1998, 41, 4577.
126. M. R. Attwood, J. M. Bennett, A. D. Campbell, G. G. Canning, M. G.
Carr, E. Conway, R. M. Dunsdon, J. R. Greening, P. S. Jones, P. B. Kay,
B. K. Handa, D. N. Hurst, N. S. Jennings, S. Jordan, E. Keech, M. A.
O’Brien, H. A. Overton, J. King-Underwood, T. M. Raynham, K. P.
Stenson, C. S. Wilkinson, T. C. Wilkinson and F. X. Wilson, Antivir.
Chem. Chemother., 1999, 10, 259.
127. R. M. Dunsdon, J. R. Greening, P. S. Jones, S. Jordan and F. X. Wilson,
Bioorg. Med. Chem. Lett., 2000, 10, 1577.
128. M. P. Groziak, Am. J. Ther., 2001, 8, 321.
129. N. N. Kim, J. D. Cox, R. F. Baggio, F. A. Emig, S. K. Mistry, S. L.
Harper, D. W. Speicher, S. M. Morris Jr, D. E. Ash, A. Traish and D. W.
Christianson, Biochemistry, 2001, 40, 2678.
130. J. Das and S. D. Kimball, Bioorg. Med. Chem., 1995, 3, 999.
131. T. Mc Cormack, W. Baumeister, L. Grenier, C. Moomaw, L. Plamondon,
B. Pramanik, C. Slaughter, F. Soucy, R. Stein, F. Zuhl and L. Dick, J.
Biol. Chem., 1997, 272, 26103.
132. H. Saitoh and B. J. Aungst, Pharm. Res., 1999, 16, 1786.
133. B. D. Dorsey, M. Iqbal, S. Chatterjee, E. Menta, R. Bernardini, A.
Bernareggi, P. G. Cassara, G. D’Arasmo, E. Ferretti, S. De Munari, A.
Oliva, G. Pezzoni, C. Allievi, I. Strepponi, B. Ruggeri, M. A. Ator, M.
Williams and J. P. Mallamo, J. Med. Chem., 2008, 51, 1068.
134. J. Ruef and H. A. Katus, Expert Opin. Investig. Drugs, 2003, 12, 781.
135. A. Schwienhorst, Cell. Mol. Life Sci., 2006, 63, 2773.
136. E. C. Attar, D. J. De Angelo, J. G. Supko, F. D’Amato, D. Zahrieh, A.
Sirulnik, M. Wadleigh, K. K. Ballen, S. McAfee, K. B. Miller, J. Levine,
I. Galinsky, E. G. Trehu, D. Schenkein, D. Neuberg, R. M. Stone and P.
C. Amrein, Clin. Cancer Res., 2008, 14, 1446.
137. D. Leveque, M. C. Carvalho and F. Maloisel, In vivo, 2007, 21, 273.
452 Chapter 9
138. Y. Ogawa, K. Tobinai, M. Ogura, K. Ando, T. Tsuchiya, Y. Kobayashi,
T. Watanabe, D. Maruyama, Y. Morishima, Y. Kagami, H. Taji, H.
Minami, K. Itoh, M. Nakata and T. Hotta, Cancer Sci., 2008, 99, 140.
139. T. Pekol, J. S. Daniels, J. Labutti, I. Parsons, D. Nix, E. Baronas, F.
Hsieh, L. S. Gan and G. Miwa, Drug Metab. Dispos., 2005, 33, 771.
140. H. C. Brown, M. M. Midland and G. W. Kabalka, J. Am. Chem. Soc.,
1971, 93, 1024.
141. S. B. Mirviss, J. Org. Chem., 1967, 32, 1713.
142. S. Wu, W. Waugh and V. J. Stella, J. Pharm. Sci., 2000, 89, 758.
143. V. Uttamsingh, C. Lu, G. Miwa and L. S. Gan, Drug Metab. Dispos.,
2005, 33, 1723.
144. C. Lu, R. Gallegos, P. Li, C. Q. Xia, S. Pusalkar, V. Uttamsingh, D. Nix,
G. T. Miwa and L. S. Gan, Drug Metab. Dispos., 2006, 34, 702.
145. J. Labutti, I. Parsons, R. Huang, G. Miwa, L. S. Gan and J. S. Daniels,
Chem. Res. Toxicol., 2006, 19, 539.
146. P. F. Bross, R. Kane, A. T. Farrell, S. Abraham, K. Benson, M. E.
Brower, S. Bradley, J. V. Gobburu, A. Goheer, S. L. Lee, J. Leighton, C.
Y. Liang, R. T. Lostritto, W. D. McGuinn, D. E. Morse, A. Rahman, L.
A. Rosario, S. L. Verbois, G. Williams, Y. C. Wang and R. Pazdur, Clin.
Cancer Res., 2004, 10, 3954.
147. R. A. Morrison, S. M. Singhvi, A. E. Peterson, D. A. Pocetti and B. H.
Migdalof, Drug Metab. Dispos., 1990, 18, 253.
148. S. M. Singhvi, K. L. Duchin, R. A. Morrison, D. A. Willard, D. W.
Everett and M. Frantz, Br. J. Clin. Pharmacol., 1988, 25, 9.
149. I. R. McNicholl and J. J. McNicholl, Ann. Pharmacother., 2001, 35, 57.
150. K. Klumpp and B. J. Graves, Curr. Top. Med. Chem., 2006, 6, 423.
151. D. B. Mendel, C. Y. Tai, P. A. Escarpe, W. Li, R. W. Sidwell, J. H.
Huffman, C. Sweet, K. J. Jakeman, J. Merson, S. A. Lacy, W. Lew, M. A.
Williams, L. Zhang, M. S. Chen, N. Bischofberger and C. U. Kim,
Antimicrob. Agents Chemother., 1998, 42, 640.
152. W. M. Kati, A. S. Saldivar, F. Mohamadi, H. L. Sham, W. G. Laver and
W. E. Kohlbrenner, Biochem. Biophys. Res. Commun., 1998, 244, 408.
153. A. W. Peng, E. K. Hussey and K. H. Moore, J. Clin. Pharmacol., 2000, 40,
242.
154. G. He, J. Massarella and P. Ward, Clin. Pharmacokinet., 1999, 37, 471.
155. K. Morimoto, M. Nakakariya, Y. Shirasaka, C. Kakinuma, T. Fujita, I.
Tamai and T. Ogihara, Drug Metab. Dispos., 2008, 36, 6.
156. C. Oo, P. Snell, J. Barrett, A. Dorr, B. Liu and I. Wilding, Int. J. Pharm.,
2003, 257, 297.
157. W. Li, P. A. Escarpe, E. J. Eisenberg, K. C. Cundy, C. Sweet, K. J.
Jakeman, J. Merson, W. Lew, M. Williams, L. Zhang, C. U. Kim,
N. Bischofberger, M. S. Chen and D. B. Mendel, Antimicrob. Agents
Chemother., 1998, 42, 647.
158. K. Beaumont, R. Webster, I. Gardner and K. Dack, Curr. Drug Metab.,
2003, 4, 461.
159. D. A. Smith, Curr. Opin. Drug Discov. Devel., 2007, 10, 550.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 453
160. M. J. Daniel, J. M. Barnett and B. A. Pearson, Clin. Pharmacokinet.,
1999, 36(Suppl 1), 41.
161. T. Yamanaka, M. Yamada, K. Tsujimura, T. Kondo, S. Nagata, S.
Hobo, M. Kurosawa and T. Matsumura, J. Vet. Med. Sci., 2007, 69,
293.
162. S. Toovey, C. Rayner, E. Prinssen, T. Chu, B. Donner, B. Thakrar, R.
Dutkowski, G. Hoffmann, A. Breidenbach, L. Lindemann, E. Carey, L.
Boak, R. Gieschke, S. Sacks, J. Solsky, I. Small and D. Reddy, Drug Saf.,
2008, 31, 1097.
163. A. Ose, H. Kusuhara, K. Yamatsugu, M. Kanai, M. Shibasaki, T. Fujita,
A. Yamamoto and Y. Sugiyama, Drug Metab. Dispos., 2008, 36, 427.
164. S. S. Jhee, M. Yen, L. Ereshefsky, M. Leibowitz, M. Schulte, B. Kaeser,
L. Boak, A. Patel, G. Hoffmann, E. P. Prinssen and C. R. Rayner,
Antimicrob. Agents Chemother., 2008, 52, 3687.
165. G. Hill, T. Cihlar, C. Oo, E. S. Ho, K. Prior, H. Wiltshire, J. Barrett, B.
Liu and P. Ward, Drug Metab. Dispos., 2002, 30, 13.
166. M. Holodniy, S. R. Penzak, T. M. Straight, R. T. Davey, K. K. Lee, M. B.
Goetz, D. W. Raisch, F. Cunningham, E. T. Lin, N. Olivo and L. R.
Deyton, Antimicrob. Agents Chemother., 2008, 52, 3013.
167. K. McClellan and C. M. Perry, Drugs, 2001, 61, 263.
168. D. J. Sweeny, G. Lynch, A. M. Bidgood, W. Lew, K. Y. Wang and K. C.
Cundy, Drug Metab. Dispos., 2000, 28, 737.
169. N. Lindegardh, G. R. Davies, T. H. Tran, J. Farrar, P. Singhasivanon, N.
P. Day and N. J. White, Antimicrob. Agents Chemother., 2006, 50, 3197.
170. D. Yang, R. E. Pearce, X. Wang, R. Gaedigk, Y. J. Wan and B. Yan,
Biochem. Pharmacol., 2009, 77, 238.
171. D. Shi, J. Yang, D. Yang, E. L. LeCluyse, C. Black, L. You, F. Akhlaghi
and B. Yan, J. Pharmacol. Exp. Ther., 2006, 319, 1477.
172. S. J. Macdonald, R. Cameron, D. A. Demaine, R. J. Fenton, G. Foster,
D. Gower, J. N. Hamblin, S. Hamilton, G. J. Hart, A. P. Hill, G. G.
Inglis, B. Jin, H. T. Jones, D. B. McConnell, J. McKimm-Breschkin, G.
Mills, V. Nguyen, I. J. Owens, N. Parry, S. E. Shanahan, D. Smith, K. G.
Watson, W. Y. Wu and S. P. Tucker, J. Med. Chem., 2005, 48, 2964.
173. G. B. Evans, R. H. Furneaux, B. Greatrex, A. S. Murkin, V. L. Schramm
and P. C. Tyler, J. Med. Chem., 2008, 51, 948.
174. M. D. Erion, J. D. Stoeckler, W. C. Guida, R. L. Walter and S. E. Ealick,
Biochemistry, 1997, 36, 11735.
175. M. D. Erion, K. Takabayashi, H. B. Smith, J. Kessi, S. Wagner, S.
Honger, S. L. Shames and S. E. Ealick, Biochemistry, 1997, 36, 11725.
176. R. W. Miles, P. C. Tyler, R. H. Furneaux, C. K. Bagdassarian and V. L.
Schramm, Biochemistry, 1998, 37, 8615.
177. J. D. Stoeckler, A. F. Poirot, R. M. Smith, R. E. Parks Jr., S. E. Ealick, K.
Takabayashi and M. D. Erion, Biochemistry, 1997, 36, 11749.
178. S. Bantia, J. A. Montgomery, H. G. Johnson and G. M. Walsh, Immu-
nopharmacology, 1996, 35, 53.
179. V. L. Schramm, Biochim. Biophys. Acta, 2002, 1587, 107.
454 Chapter 9
180. R. G. Silva, J. E. Nunes, F. Canduri, J. C. Borges, L. M. Gava, F. B.
Moreno, L. A. Basso and D. S. Santos, Curr. Drug Targets, 2007, 8, 413.
181. S. Banti, P. J. Miller, C. D. Parker, S. L. Ananth, L. L. Horn, Y. S. Babu
and J. S. Sandhu, Int. Immunopharmacol., 2002, 2, 913.
182. S. Bantia, S. L. Ananth, C. D. Parker, L. L. Horn and R. Upshaw, Int.
Immunopharmacol., 2003, 3, 879.
183. S. Bantia, P. J. Miller, C. D. Parker, S. L. Ananth, L. L. Horn, J. M.
Kilpatrick, P. E. Morris, T. L. Hutchison, J. A. Montgomery and J. S.
Sandhu, Int. Immunopharmacol., 2001, 1, 1199.
184. J. P. Jenuth, J. E. Dilay, E. Fung, E. R. Mably and F. F. Snyder, Adv.
Exp. Med. Biol., 1991, 309B, 273.
185. M. L. Markert, Immunodefic. Rev., 1991, 3, 45.
186. E. A. Taylor Ringia and V. L. Schramm, Curr. Top. Med. Chem., 2005, 5,
1237.
187. G. B. Evans, R. H. Furneaux, T. L. Hutchison, H. S. Kezar, P. E. Morris
Jr., V. L. Schramm and P. C. Tyler, J. Org. Chem., 2001, 66, 5723.
188. G. B. Evans, R. H. Furneaux, A. Lewandowicz, V. L. Schramm and P. C.
Tyler, J. Med. Chem., 2003, 46, 3412.
189. G. A. Kicska, P. C. Tyler, G. B. Evans, R. H. Furneaux, K. Kim and V.
L. Schramm, J. Biol. Chem., 2002, 277, 3219.
190. A. Lewandowicz, E. A. Ringia, L. M. Ting, K. Kim, P. C. Tyler, G. B.
Evans, O. V. Zubkova, S. Mee, G. F. Painter, D. H. Lenz, R. H. Fur-
neaux and V. L. Schramm, J. Biol. Chem., 2005, 280, 30320.
191. V. L. Schramm, Arch. Biochem. Biophys., 2005, 433, 13.
192. A. Lewandowicz, P. C. Tyler, G. B. Evans, R. H. Furneaux and V. L.
Schramm, J. Biol. Chem., 2003, 278, 31465.
193. J. M. Kilpatrick, P. E. Morris, D. G. Serota Jr, D. Phillips, D. R. Moore,
J. C. Bennett and Y. S. Babu, Int. Immunopharmacol., 2003, 3, 541.
194. C. M. Galmarini, IDrugs, 2006, 9, 712.
195. V. Gandhi, J. M. Kilpatrick, W. Plunkett, M. Ayres, L. Harman, M. Du,
S. Bantia, J. Davisson, W. G. Wierda, S. Faderl, H. Kantarjian and D.
Thomas, Blood, 2005, 106, 4253.
196. H. S. Kezar 3rd, J. M. Kilpatrick, D. Phillips, D. Kellogg, J. Zhang and P.
E. Morris Jr., Nucleosides Nucleotides Nucleic Acids, 2005, 24, 1817.
197. B. J. Kane, J. G. Kuhn and M. K. Roush, Ann. Pharmacother., 1992, 26,
939.
198. C. Lathia, G. F. Fleming, M. Meyer, M. J. Ratain and L. Whitfield,
Cancer Chemother. Pharmacol., 2002, 50, 121.
199. J. D. Geiger, J. L. Lewis, C. J. MacIntyre and J. I. Nagy, Neuro-
pharmacology, 1987, 26, 1383.
200. S. B. Kaye, Br. J. Cancer, 1998, 78(Suppl 3), 1.
201. W. Kreis, T. M. Woodcock, M. B. Meyers, L. A. Carlevarini and I. H.
Krakoff, Cancer Treat. Rep., 1977, 61, 723.
202. S. Urien, K. Rezai and F. Lokiec, J. Pharmacokinet. Pharmacodyn., 2005,
32, 817.
203. S. Grant, K. Bhalla and C. McCrady, Leuk. Res., 1991, 15, 205.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 455
204. G. L. Neil, T. E. Moxley, S. L. Kuentzel, R. C. Manak and L. J. Hanka,
Cancer Chemother. Rep., 1975, 59, 459.
205. S. Xiang, S. A. Short, R. Wolfenden and C. W. Carter Jr, Biochemistry,
1997, 36, 4768.
206. C. W. Carter Jr, Biochimie, 1995, 77, 92.
207. W. M. Kati and R. Wolfenden, Science, 1989, 243, 1591.
208. H. Guo, N. Rao, Q. Xu and H. Guo, J. Am. Chem. Soc., 2005, 127, 3191.
209. C. B. Yoo, J. C. Cheng and P. A. Jones, Biochem. Soc. Trans., 2004, 32,
910.
210. L. Zhou, X. Cheng, B. A. Connolly, M. J. Dickman, P. J. Hurd and D. P.
Hornby, J. Mol. Biol., 2002, 321, 591.
211. R. M. Cohen and R. Wolfenden, J. Biol. Chem., 1971, 246, 7561.
212. J. Laliberte, V. E. Marquez and R. L. Momparler, Cancer Chemother.
Pharmacol., 1992, 30, 7.
213. N. Eliopoulos, D. Cournoyer and R. L. Momparler, Cancer Chemother.
Pharmacol., 1998, 42, 373.
214. J. L. Holleran, R. A. Parise, E. Joseph, J. L. Eiseman, J. M. Covey, E. R.
Glaze, A. V. Lyubimov, Y. F. Chen, D. Z. D’Argenio and M. J. Egorin,
Clin. Cancer Res., 2005, 11, 3862.
215. J. H. Beumer, E. Joseph, M. J. Egorin, R. S. Parker, Z. D’Argenio D, J.
M. Covey and J. L. Eiseman, Clin. Cancer Res., 2006, 12, 5826.
216. D. H. Ho, G. P. Bodey, S. W. Hall, R. S. Benjamin, N. S. Brown, E. J.
Freireich and T. L. Loo, J. Clin. Pharmacol., 1978, 18, 259.
217. J. H. Beumer, J. L. Eiseman, R. A. Parise, J. A. Florian Jr., E. Joseph, D.
Z. D’Argenio, R. S. Parker, B. Kay, J. M. Covey and M. J. Egorin,
Cancer Chemother. Pharmacol., 2008, 62, 457.
218. S. M. el-Dareer, V. White, F. P. Chen, L. B. Mellett and D. L. Hill,
Cancer Treat. Rep., 1976, 60, 1627.
219. Z. Qi, I. Whitt, A. Mehta, J. Jin, M. Zhao, R. C. Harris, A. B. Fogo and
M. D. Breyer, Am. J. Physiol. Renal Physiol., 2004, 286, F590.
220. R. W. Klecker, R. L. Cysyk and J. M. Collins, Bioorg. Med. Chem., 2006,
14, 62.
221. S. Yoshihara and K. Tatsumi, Arch. Biochem. Biophys., 1997, 338, 29.
222. C. Beedham, Prog. Med. Chem., 1987, 24, 85.
223. K. Nagai, K. Nagasawa and S. Fujimoto, Biochem. Biophys. Res. Com-
mun., 2005, 334, 1343.
224. P. P. Wong, V. E. Currie, R. W. Mackey, I. H. Krakoff, C. T. Tan, J. H.
Burchenal and C. W. Young, Cancer Treat. Rep., 1979, 63, 1245.
225. W. Kreis, T. M. Woodcock, C. S. Gordon and I. H. Krakoff, Cancer
Treat. Rep., 1977, 61, 1347.
226. G. J. Peters, C. J. van Groeningen, E. J. Laurensse, J. Lankelma, A. Leyva
and H. M. Pinedo, Cancer Chemother. Pharmacol., 1987, 20, 101.
227. C. J. van Groeningen, A. Leyva, I. Kraal, G. J. Peters and H. M. Pinedo,
Cancer Treat. Rep., 1986, 70, 745.
228. A. Leyva, C. J. van Groeningen, I. Kraal, H. Gall, G. J. Peters, J. Lan-
kelma and H. M. Pinedo, Cancer Res., 1984, 44, 5928.
456 Chapter 9
229. G. P. Connolly and J. A. Duley, Trends Pharmacol. Sci., 1999, 20, 218.
230. D. M. Becroft, L. I. Phillips and A. Simmonds, J. Pediatr., 1969, 75, 885.
231. H. Koyama and M. Tsuji, Biochem. Pharmacol., 1983, 32, 3547.
232. L. Gan, M. R. Seyedsayamdost, S. Shuto, A. Matsuda, G. A. Petsko and
L. Hedstrom, Biochemistry, 2003, 42, 857.
233. T. Inou, R. Kusaba, I. Takahashi, H. Sugimoto, K. Kuzuhara, Y.
Yamada, J. Yamauchi and O. Otsubo, Transplant Proc., 1981, 13, 315.
234. F. Marumo, M. Okubo, K. Yokota, H. Uchida, K. Kumano, T. Endo, K.
Watanabe and N. Kashiwagi, Transplant Proc., 1988, 20, 406.
235. K. Mita, N. Akiyama, T. Nagao, H. Sugimoto, S. Inoue, T. Osakabe, Y.
Nakayama, K. Yokota, K. Sato and H. Uchida, Transplant Proc., 1990,
22, 1679.
236. A. Tajima, M. Hata, N. Ohta, Y. Ohtawara, K. Suzuki and Y. Aso,
Transplantation, 1984, 38, 116.
237. M. Honda, H. Itoh, T. Suzuki and Y. Hashimoto, Biol. Pharm. Bull.,
2006, 29, 2460.
238. M. Okada, K. Suzuki, M. Nakashima, T. Nakanishi and N. Fujioka, Eur.
J. Pharmacol., 2006, 531, 140.
239. S. D. Patil, L. Y. Ngo, P. Glue and J. D. Unadkat, Pharm. Res., 1998, 15,
950.
240. S. A. Gruber, G. R. Erdmann, B. A. Burke, A. Moss, L. Bowers, W. J.
Hrushesky, R. J. Cipolle, D. M. Canafax and A. J. Matas, Transplanta-
tion, 1992, 53, 12.
241. R. Kusaba, O. Otubo, H. Sugimoto, I. Takahashi, Y. Yamada, J.
Yamauchi, N. Akiyama and T. Inou, Proc. Eur. Dia.l Transplant. Assoc.,
1981, 18, 420.
242. Y. Kokado, S. Takahara, M. Ishibashi and A. Okuyama, Transplant.
Proc., 1994, 26, 2111.
243. D. Stypinski, M. Obaidi, M. Combs, M. Weber, A. J. Stewart and H.
Ishikawa, Br. J. Clin. Pharmacol., 2007, 63, 459.
244. P. J. O’Dwyer, Pharmacol. Ther., 1990, 48, 371.
245. K. D. Collins and G. R. Stark, J. Biol. Chem., 1971, 246, 6599.
246. R. S. Obach, F. Lombardo and N. J. Waters, Drug Metab. Dispos., 2008,
36, 1385.
247. www.emea.europa.eu/humandocs/PDFs/EPAR/velcade/velcade.htm.
248. C. H. Gu, H. Li, J. Levons, K. Lentz, R. B. Gandhi, K. Raghavan and R.
L. Smith, Pharm. Res., 2007, 24, 1118.
249. N. Tam-Zaman, Y. K. Tam, S. Tawfik and H. Wiltshire, Pharm. Res.,
2004, 21, 436.
250. J. O. Baker, S. H. Wilkes, M. E. Bayliss and J. M. Prescott, Biochemistry,
1983, 22, 2098.
251. A. H. Brockman, P. Hatsis, M. Paton and J. T. Wu, Anal. Chem., 2007, 4,
1599.
252. Y. Wang, N. Mealy, N. Serradell, J. Bolos and E. Rosa, Drugs Future,
2007, 32, 310.
253. F. Fyhrquist and O. Saijonmaa, J. Intern. Med., 2008, 264, 224.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 457
254. A. G. Tzakos, S. Galanis, G. A. Spyroulias, P. Cordopatis, E. Manessi-
Zoupa and I. P. Gerothanassis, Prot. Engin., 2003, 16, 993.
255. C. Shu, H. Shen, U. Hopfer and D. E. Smith, Drug Metab. Dispos., 2001,
29, 1307.
256. I. Knutter, C. Wollesky, G. Kottra, M. G. Hahn, W. Fischer, K. Zebisch,
R. H. Neubert, H. Daniel and M. Brandsch, J. Pharmacol. Exp. Ther.,
2008, 327, 432.
257. R. W. Klecker, R. L. Cysyk and J. M. Collins, Bioorg. Med. Chem., 2006,
14, 62.
258. J. Murase, K. Mizuno, K. Kawai, S. Nishiumi, Y. Kobayashi, M.
Hayashi, T. Morino, T. Suzuki and S. Baba, Oyo Yakuri, 1978, 15, 829.
259. J. M. Kolesar, A. K. Morris and J. G. Kuhn, J. Oncol. Pharm. Pract., ,
1996, 2, 211.
260. E. M. del Amo, A. T. Heikkinen and J. Monkkonen, Eur. J. Pharm. Sci.,
2009, 36, 200.
261. H. H. Usansky, P. Hu and P. J. Sinko, Drug Metab. Dispos., 2008, 36,
863.
262. W. L. Chiou, S. M. Chung, T. C. Wu and C. Ma, Int. J. Clin. Pharmacol.
Ther., 2001, 39, 93.
263. K. C. Yeh, J. A. Stone, A. D. Carides, P. Rolan, E. Woolf and W. D. Ju,
J. Pharm. Sci., 1999, 88, 568.
264. M. Vermeir, S. Lachau-Durand, G. Mannens, F. Cuyckens, B. van Hoof
and A. Raoof, Drug Metab. Dispos., 2009, 37, 809.
265. C. Li, T. Liu, L. Broske, J. M. Brisson, A. S. Uss, F. G. Njoroge, R. A.
Morrison and K. C. Cheng, Biochem. Pharmacol., 2008, 76, 1757.
266. P. Revill, N. Serradell, J. Bolos and E. Rosa, Drugs Future, 2007, 32, 788–
798.
267. S. Letendre, J. Marquie-Beck, E. Capparelli, B. Best, D. Clifford, A. C.
Collier, B. B. Gelman, J. C. McArthur, J. A. McCutchan, S. Morgello, D.
Simpson, I. Grant and R. J. Ellis, Arch. Neurol., 2008, 65, 65.
268. L. Varatharajan and S. A. Thomas, Antiviral. Res., 2009, 82, A99.
269. W. C. Ripka, Curr. Opin. Chem. Biol., 1997, 1, 242.
270. O. Okusanya, A. Forrest, R. DiFrancesco, S. Bilic, S. Rosenkranz, M. F.
Para, E. Adams, K. E. Yarasheski, R. C. Reichman and G. D. Morse,
Antimicrob. Agents Chemother., 2007, 51, 1822.
271. D. R. Goldsmith and C. M. Perry, Drugs, 2003, 63, 1679.
272. M. Rittweger and K. Arasteh, Clin. Pharmacokinet., 2007, 46, 739.
273. J. H. Lin, M. Chiba, S. K. Balani, I. W. Chen, G. Y. Kwei, K. J. Vastag
and J. A. Nishime, Drug Metab. Dispos., 1996, 24, 1111.
274. J. H. Lin, M. Chiba, I. W. Chen, J. A. Nishime, F. A. deLuna, M.
Yamazaki and Y. J. Lin, Drug Metab. Dispos., 1999, 27, 1187.
275. J. H. Lin, M. Chiba, I. W. Chen, J. A. Nishime and K. J. Vastag, Drug
Metab. Dispos., 1996, 24, 1298.
276. A. Kiriyama, T. Nishiura, H. Yamaji and K. Takada, Biopharm. Drug
Dispos., 1999, 20, 199.
458 Chapter 9
277. R. W. Humphrey, K. M. Wyvill, B. Y. Nguyen, L. E. Shay, D. R. Kohler,
S. M. Steinberg, T. Ueno, T. Fukasawa, M. Shintani, H. Hayashi, H.
Mitsuya and R. Yarchoan, Antiviral Res., 1999, 41, 21.
278. B. U. Mueller, B. D. Anderson, M. Q. Farley, R. Murphy, J. Zuckerman,
P. Jarosinski, K. Godwin, C. L. McCully, H. Mitsuya, P. A. Pizzo and F.
M. Balis, Antimicrob. Agents Chemother., 1998, 42, 1815.
279. B. V. Shetty, M. B. Kosa, D. A. Khalil and S. Webber, Antimicrob. Agents
Chemother., 1996, 40, 110.
280. K. A. Jackson, S. E. Rosenbaum, B. M. Kerr, Y. K. Pithavala, G. Yuen
and M. N. Dudley, Antimicrob. Agents Chemother., 2000, 44, 1832.
281. http://www.accessdata.fda.gov/drugsatfda_docs/nda/99/20-945.pdf_Ritonovir_
Prntlbl.pdf..
282. D. J. Kempf, K. C. Marsh, J. F. Denissen, E. McDonald, S. Vasava-
nonda, C. A. Flentge, B. E. Green, L. Fino, C. H. Park and X. P. Kong
et al., Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 2484.
283. J. F. Denissen, B. A. Grabowski, M. K. Johnson, A. M. Buko, D. J.
Kempf, S. B. Thomas and B. W. Surber, Drug Metab. Dispos., 1997, 25,
489.
284. R. Lledo-Garcia, A. Nacher, L. Prats-Garcia, V. G. Casabo and M.
Merino-Sanjuan, J. Pharm. Sci., 2007, 96, 633.
285. A. Billich, G. Fricker, I. Muller, P. Donatsch, P. Ettmayer, H. Gstach, P.
Lehr, P. Peichl, D. Scholz and B. Rosenwirth, Antimicrob. Agents Che-
mother., 1995, 39, 1406.
286. S. Thaisrivongs and J. W. Strohbach, Biopolymers, 1999, 51, 51.
287. J. D. Courter, J. E. Girotto and J. C. Salazar, Expert Rev. Anti Infect.
Ther., 2008, 6, 797.
288. S. R. Turner, J. W. Strohbach, R. A. Tommasi, P. A. Aristoff, P. D.
Johnson, H. I. Skulnick, L. A. Dolak, E. P. Seest, P. K. Tomich, M. J.
Bohanon, M. M. Horng, J. C. Lynn, K. T. Chong, R. R. Hinshaw, K. D.
Watenpaugh, M. N. Janakiraman and S. Thaisrivongs, J. Med. Chem.,
1998, 41, 3467.
289. H. D. Kleinert, H. H. Stein, S. Boyd, A. K. Fung, W. R. Baker, K. M.
Verburg, J. S. Polakowski, P. Kovar, J. Barlow and J. Cohen et al.,
Hypertension, 1992, 20, 768.
290. M. Azizi, R. Webb, J. Nussberger and N. K. Hollenberg, J. Hypertens.,
2006, 24, 243.
291. G. A. Rongen, J. W. Lenders, P. Smits and T. Thien, Clin. Pharmacoki-
net., 1995, 29, 6.
292. F. H. Derkx, A. H. van den Meiracker, W. Fischli, P. J. Admiraal, A. J.
Man in’t Veld, P. van Brummelen and M. A. Schalekamp, Am. J.
Hypertens., 1991, 4, 602.
293. M. Shibasaki, T. Usui, O. Inagaki, M. Asano and T. Takenaka, J. Pharm.
Pharmacol., 1994, 46, 68.
294. A. Yilmaz, L. Stahle, L. Hagberg, B. Svennerholm, D. Fuchs and M.
Gisslen, Scand. J. Infect. Dis., 2004, 36, 823.
Pharmacokinetics and Metabolism of Compounds that Mimic Enzyme 459
295. Y. Khaliq, K. Gallicano, S. Venance, S. Kravcik and D. W. Cameron,
Clin. Pharmacol. Ther., 2000, 68, 637.
296. R. B. Perni, S. J. Almquist, R. A. Byrn, G. Chandorkar, P. R. Chaturvedi,
L. F. Courtney, C. J. Decker, K. Dinehart, C. A. Gates, S. L. Harbeson,
A. Heiser, G. Kalkeri, E. Kolaczkowski, K. Lin, Y. P. Luong, B. G. Rao,
W. P. Taylor, J. A. Thomson, R. D. Tung, Y. Wei, A. D. Kwong and C.
Lin, Antimicrob. Agents Chemother., 2006, 50, 899.
297. P. Revill, N. Serradell, J. Bolos and E. Rosa, Drugs of the Future, 2007,
32, 310.
298. Y. Wang, N. Mealy, N. Serradell, J. Bolos and E. Rosa, Drugs of the
Future, 2007, 32, 310.
299. M. O. Vallez, A. Rupin, D. Versluys, G. De Nantuil, and T. Verbeuren, J.
Thromb. Haemost., 1999, Suppl, Abst 2284.
300. Y.-J. Chyan and L. Ming, Recent Patents on Endocrine, Metabolic and
Immune Drug Discovery, 2007, 1, 15–24.
CHAPTER 10

Alcohols and Phenols:

Absorption, Distribution,
Metabolism and Excretion
ZHUANG MIAO AND R. SCOTT OBACH

Pfizer Research and Development, Groton Laboratories, CT, USA

10.1 Physicochemical Properties of Alcohols and

Phenols and their Prevalence in Drugs
Many drugs contain a hydroxyl group (aliphatic alcohols and phenols). While
contributing to the polarity and hydrophilicity of a molecule, the addition of a
hydroxyl group (-OH) does not alter (in almost all cases) the charge of a
molecule at physiological pH. The pKa of simple aliphatic alcohols is greater
than 14 and for phenols is typically in the range of 9–12. Thus, in drug design,
an alcohol group can offer a moderate increase in hydrophilicity without
altering the charge. For phenols of rings that possess highly electron deficient
substituents (e.g. nitro), the pKa of the hydroxyl group can be in the range of
physiological pH, thus causing a significant fraction of the drug to be anionic in
the body. And, in some cases, if the deprotonated anionic form can be reso-
nance stabilised (e.g. warfarin, meloxicam, etc.), the hydroxyl group can be
nearly as acidic as a carboxylic acid.
The hydroxyl group possesses simultaneously a hydrogen bond acceptor
(HBA) and a hydrogen bond donor (HBD), which can provide binding energy
to macromolecule targets and increase affinity as well as increase water

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

460
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 461
solubility. Also, by decreasing lipophilicity, the addition of a hydroxyl group
will generally decrease tissue partitioning. The increase in hydrophilicity and
hydrogen-bonding capacity can also decrease membrane penetration. If not
oriented toward direct interaction with substituents on the macromolecule
target, a hydroxyl group can reside in solvent-exposed space. With these diverse
capabilities, the hydroxyl group offers the medicinal chemist a structural
modification that can substantially alter the behaviour of a molecule at the
target and in the body (Table 10.1).
The presence of the hydroxyl group in drugs is high, with 40 of the 150 most
frequently prescribed medications containing an alcohol or phenol function,
most of which are taken orally (Table 10.2). These drugs span a range of
indications including those which require penetration into the central nervous
system (CNS), supporting the notion that a hydroxyl group does not preclude
penetration across membranes or the blood–brain barrier.

Table 10.1 Average physicochemical properties for a set of fifteen drugs and
their hydroxy analogies.
Property Without -OH With -OH

log P 2.7 1.9

log D7.4 1.3 0.6
PSA (Å2) 39 60
Molecular weight 237 253
HBD+HBA 4 6

Table 10.2 Prevalence of drugs containing an OH group among the top 150
most frequently prescribed drugs.
Drug Indication Drug Indication

Acetaminophen Analgesic/Antipyretic Lorazepam Anxiolytic

Albuterol Anti-asthmatic Losartan Antihypertensive
Alendronate Anti-osteoperotic Lovastatin Anti-atherosclerotic
Amoxicillin Antibacterial Meloxicam Anti-inflammatory
Aspirin Analgesic/antipyretic Methylprednisolone Anti-inflammatory
Atenolol Antihypertensive Metoprolol Antihypertensive
Atorvastatin Anti-atherosclerotic Metronidazole Antifungal
Azithromycin Antibacterial Montelukast Anti-asthmatic
Carvedilol Antihypertensive Morphine Narcotic analgesic
Clarithromycin Antibacterial Olmesartan Antihypertensive
Clindamycin Antibacterial Oxycodone Narcotic analgesic
Codeine Narcotic analgesic Prednisone Anti-inflammatory
Digoxin Anti-arrythmic Propranolol Antihypertensive
Doxycycline Antibacterial Quetiapine Antipsychotic
Ethinyl estradiol Oral contraceptive Rosuvastatin Anti-atherosclerotic
Ezetimibe Anti-atherosclerotic Salmeterol Anti-asthmatic
Fexofenadine Antihistamine Simvastatin Anti-atherosclerotic
Fluticasone Anti-inflammatory Tramadol Narcotic analgesic
Hydroxyzine Antihistamine Venlafaxine Antidepressant
Levothyroxine Hypothroidism Warfarin Anticoagulant
462 Chapter 10
Introduction of a hydroxyl group by oxidative metabolism is extremely com-
mon and, in many cases, the hydroxyl metabolite can possess pharmacological
activity similar to the parent drug—albeit that in many cases the intrinsic potency
is altered. In such cases, the absorption, distribution, metabolism and excretion
(ADME) properties of the metabolites have been explored to varying degrees and
compared to that of the parent drug, either to ensure an understanding of the
contribution of the metabolite to drug effect or to explore the possibility that the
metabolite itself could be pharmacologically active. There are, therefore, a
number of reports in the scientific literature in which the ADME properties of a
structure with and without a hydroxyl group have been directly compared. From
these reports, the general impact of hydroxyl on ADME properties can be
derived. These cases are included in this chapter for illustration purposes.

10.2 Comparative Pharmacokinetics of Alcohols,

Phenols and their Counterparts Lacking the
Hydroxy Group
The pharmacokinetic parameters, clearance (CL) and volume of distribution
(VD), are the most important ones that help define the dosing regimen of a
drug. A comparison of these parameters for analogues lacking or possessing a
hydroxyl group is given in Table 10.3.
The addition of a hydroxyl group can either increase or decrease total CL; in
the examples listed, the split between an increase and a decrease is 50/50.
Clearance is a function of intrinsic clearance (i.e. the rate at which drug meta-
bolising enzymes and transporters can act upon the drug) and the free fraction.
Addition of hydroxyl group can cause the compound to be a substrate for
conjugating enzymes [e.g. UDP-glucuronosyltransferase (UGT), sulfotransferase
(SULT) ] or other oxidative enzymes (e.g. alcohol dehydrogenase, peroxidases);
they may also impart greater substrate properties for some drug transport pro-
teins. However, it is important to note that the addition of hydroxyl can cause
the free fraction in plasma to increase—in some cases substantially. Thus, free
clearance for hydroxyl analogues is lower than their corresponding analogues
(Table 10.4). Since it is unbound exposure that is important for drug effect, the
addition of a hydroxyl group is likely to lead to an improvement, even if the total
CL appears to increase (provided that intrinsic target potency is retained as well
as the ability of the hydroxyl containing analogue to penetrate the target tissue).
Addition of a hydroxyl group can yield substantial changes in the volume of
distribution. In many cases, it can decrease markedly (e.g. imipramine vs. 2-
hydroxyimipramine, maprotiline vs. oxaprotiline; Table 10.3). VD is a function
of the relative partitioning of compounds between the central compartment (i.e.
plasma) and the peripheral tissues. This partitioning is a function of the relative
non-specific binding capacity of macromolecular entities in tissues compared
with proteins in plasma that bind drugs (e.g. albumin, alpha-1-acid glycopro-
tein). In general, compounds that are lipophilic tend to partition more exten-
sively into tissues and have relatively higher VD values.1 The addition of a
Table 10.3 Comparison of the parenteral pharmacokinetics of compounds possessing or lacking a hydroxy group.
CL (mL/min/kg VD (L/kg body
Compound Compound body weight) weight) Half-life (h)
lacking -OH possessing -OH Species –OH +OH –OH +OH –OH +OH

Buspirone 6-Hydroxybuspirone Rat 69 47 2.5 2.6 0.9 1.2

Dantrolene 5-Hydroxydantrolene Dog 14 17 1.2 0.9 1.1 0.6
Glimiperide Hydroxglimepiride Human 0.5 1.1 0.19 NF 10 1.2
Maprotiline Oxiprotiline Human 14 13 45 19 51 19
Diazepam Temazepam Rat 82 45 4.5 4.6 1.1 4.0
Ezlopitant Hydroxyezlopitant Rat 65 53 22 8.9 7.7 1.8
Ezlopitant Hydroxyezlopitant Dog 27 24 5.6 3.2 2.5 1.7
Cotinine 3 0 -Hydroxycotinine Human 0.89 1.8 1.1 0.85 17 5.9
Imipramine 2-Hydroximipramine Human 13 9.2 12 6.6 16 10
Nordiazepam Oxazepam Human 0.12 1.1 0.45 0.59 46 6.7
Chloroquine Hydroxchloroquine Human 4.1 11 140 700 570 850
Daunorubicin Adriamycin Dog 210 51 26 7.2 22 26
Tetracycline Oxytetracycline Human 1.5 2.0 1.2 1.7 9.4 10
Amphetamine Phenylpropanolamine Dog 8.0 9.4 2.7 4.3 4.5 6.0
Amphetamine Phenylpropanolamine Rat 58 25 4.1 5.5 1.1 2.2
Amphetamine p-Hydroxyamphetamine Rat 58 88 4.1 3.8 1.1 0.66
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion

Limonene Perillyl alcohol Rat 50 22 12 1.7 4.7 1.4

463
464 Chapter 10
Table 10.4 Comparison of the parenteral pharmacokinetic parameters cor-
rected for plasma protein binding of compounds possessing or
lacking a hydroxy group.
CLu (mL/min/
kg body VDu (L/kg
weight) body weight)
Drug pair Species –OH +OH –OH +OH

Maprotiline/Oxaprotiline Human 140 76 450 110

Ezlopitant/Hydroxyezlopitant Rat 3000 230 1000 39
Ezlopitant/Hydroxyezlopitant Dog 3000 130 620 17
Imipramine/2-Hydroxyimipramine Human 170 26 160 18
Nordiazepam/Oxazepam Human 4.2 28 15 15
Chloroquine/Hydroxychloroquine Human 10 19 330 1200
Tetracycline/Oxytetracycline Human 1.9 2.2 1.5 1.9
Amphetamine/Phenylpropanolamine Rat 97 42 6.8 9.2
CLu ¼ free clearance; VDu ¼ free volume distribution.

hydroxyl group will decrease lipophilicity and increase PSA (Table 10.1), and
thus it is expected that tissue distribution will decrease and plasma protein
binding will also decrease. Hence, the impact on VD will depend on the relative
effect that addition of the hydroxyl will have on these two opposing properties.
After correcting for differences in plasma protein binding for drugs and their
hydroxyl containing analogues, the free VD values for the hydroxyl containing
set are generally lower than their non-hydroxyl counterparts. This reflects the
notion that addition of hydroxyl will decrease tissue partitioning (Table 10.4).
CL and VD determine the half-life (t1/2). Since the addition of a hydroxyl
group can impact both these properties, one may not be able to necessarily
predict the impact on half-life when a hydroxyl group is added during drug
design. The addition of the hydroxyl group could direct a compound to drug
metabolising enzymes that ordinarily would not act upon the parent, while at
the same time decreasing the rate of metabolism by P450 enzymes. For any new
chemical scaffold being used in drug design efforts, this determination needs to
be made experimentally, although there are increasingly improving computa-
tional approaches to predicting VD and plasma protein binding, as well as
interactions with drug-metabolising enzymes. The impact of a hydroxyl group
on t1/2 for a number of drugs is listed in Table 10.3. Comparisons of CL, VD
and t1/2 for hydroxyl containing compounds and their counterparts are shown
in Figure 10.1. It should be noted that, while the impact on CL and VD is
varied, the addition of a hydroxyl more often leads to a decrease in t1/2.

10.3 Biochemical Determinants of ADME

Characteristics of Drugs Possessing Hydroxyl
Groups
The thousands of interactions that occur between a small molecule drug and
macromolecular structures within the body are what determine the
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 465

100

CL for -OH containing analogue

10

(mL/min/kg)

0.1
0.1 1 10 100
CL for non -OH compounds (mL/min/kg)
VD for -OH containing analogues (L/kg)

100

10

0.1
0.1 1 10 100
VD for non -OH compounds (L/kg)

100
t1/2 for -OH containing analogues (hr)

10

0.1
0.1 1 10 100
t1/2 for non-OH containing compounds (hr)

Figure 10.1 Comparison of CL, VD, and t1/.2 for compounds possessing and lacking
a hydroxyl group. The median CL for the OH-containing is approxi-
mately the same as for the non-OH counterparts, albeit there may be a
trend for high CL vs. low CL compound pairs. For VD, the median value
for the OH-containing compounds is 85% of the value for the non-OH
compounds. For half-life, the median value for the OH-containing
compounds is 60% of that of the non-OH counterparts.
466 Chapter 10
pharmacokinetic characteristics of that compound and, together with the
specifics regarding the interaction of that drug and its intended target, will
determine its overall efficacy and usefulness for treating disease. Thus, during
the drug design process, consideration of the potential impact of the addition or
removal of a hydroxyl group on these interactions is important. The questions
that need to be asked include:

 Will the addition/removal of the hydroxyl result in increased or decreased

potency at the target? (This is a case-by-case consideration for each
pharmacological target and is beyond the scope of this chapter.)
 Will addition/removal of the hydroxyl group alter the free fraction?
 Will the addition/removal of the hydroxyl group increase or decrease non-
specific binding in tissues?
 Will the addition/removal of the hydroxyl group increase or decrease the
ability of the molecule to penetrate membranes?
 Will the addition/removal of the hydroxyl group alter interaction(s) with
membrane transporters?
 What impact will the addition/removal of the hydroxyl group have on
interactions with various drug-metabolising enzymes, and will this affect
the rate and routes of metabolic clearance?

Based on our level of understanding, these questions can be addressed to

varying degrees of satisfaction.

10.3.1 Plasma Protein Binding and Tissue Distribution

In general, designing drugs with the intent of altering plasma protein binding
alone is not a fruitful venture, as increases in free fraction will not necessarily
yield higher free concentrations since clearance will also increase. However,
knowledge of the plasma protein binding value is critical to understanding in
vivo disposition and linking that to in vitro data. (For example, relating in vivo
efficacious concentrations to receptor binding affinity requires knowledge of the
free fraction. In addition, understanding whether a drug can freely distribute to
target tissues or whether there is a drug transporter altering this equilibrium
requires knowledge of the free fraction in plasma.) Decreases in plasma protein
binding can also lead to lower dose requirements, provided target affinity is
unchanged.
Comparisons of plasma protein binding of compounds and their analogues
possessing a single additional hydroxyl group are listed in Table 10.5. For this
set of ten drugs, it can be readily seen that free fraction either remains the same
or increases (considerably in some cases) when an hydroxyl is present in the
molecule. Typical increases in free fraction are around two-fold. For the highly
bound compounds, two example sets (both benzodiazepines) did not show a
difference in protein binding between the hydroxyl containing and hydroxyl
lacking analogues; however, for two others (imipramine and ezlopitant) the
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 467
Table 10.5 Comparison of the plasma protein binding of compounds pos-
sessing or lacking a hydroxy group.
Free fraction
Drug pair Species –OH +OH

Tacrine/Velnacrine Human 0.25 0.50

Diazepam/Temazepam Human 0.02 0.02
Maprotiline/Oxaprotiline Human 0.10 0.17
Imipramine/2-Hydroxyimipramine Human 0.075 0.36
Nordiazepam/Oxazepam Human 0.03 0.04
Chloroquine/Hydroxychloroquine Human 0.43 0.57
Tetracycline/Oxytetracycline Human 0.78 0.90
Ezlopitant/Hydroxyezlopitant Rat 0.022 0.23
Dog 0.009 0.19
Risperidone/Paliperidone Human 0.10 0.23
Dog 0.08 0.20
Rat 0.12 0.25
Amphetamine/Phenylpropanolamine Rat 0.60 0.60
Human 0.84 0.45

addition of an hydroxyl yielded marked changes in free fraction. Obviously, for

compounds that are not highly bound, the hydroxyl containing and hydroxyl
lacking analogues will show little difference. The list of examples contains drugs
bound to albumin and alpha-1-acid glycoprotein.
Studies on tissue distribution of hydroxyl and non-hydroxy analogues are rare,
but an excellent demonstration of the role hydroxyl can play in tissue partitioning
is exemplified for a series of progesterone analogues. Using lung tissue, the
partitioning of progesterone, 11b-hydroxy, 11b,21-dihydroxy, and 11b,17a,21-
trihydroxy was measured and shown to be directly related to the number of
hydroxyl groups, with partition coefficients decreasing from 22, 3.7, 1.7, and 0.86,
respectively.2 Since these analogues are unionised at physiological pH, the change
in tissue partitioning can be concluded to be driven solely by the lipophilicity
decreases that each successive hydroxyl group imparts to the structure.

10.3.2 Interactions of Hydroxyl Group Containing Drugs with

Drug Transporters and Impact on Absorption,
Distribution and Excretion
One of the most compelling examples of a comparison of two compounds
that differ by a single hydroxyl group is risperidone vs. its metabolite
9-hydroxyrisperidone (paliperidone) as a result to the desire to better
understand the relative contributions of these two entities to the effects of
risperidone. Tissue distribution of these two compounds was compared in the
rat after administration of risperidone.3,4 The total brain exposures were
similar between the two compounds, but the total plasma exposure to the
hydroxyl metabolite was over four-fold greater (hence the free plasma
exposure is over eight-fold greater), indicating that the brain–plasma ratio was
468 Chapter 10
greater for the non-hydroxyl parent compound. Distribution within specific
brain regions also differed.5 Other tissue–plasma ratios did not differ as
much. It was shown that, while both risperidone and 9-hydroxyrisperidone are
substrates for P-glycoprotein (an important efflux transporter at the blood–
brain barrier), the 9-hydroxy metabolite was a better substrate and had a
greater impact.6–8 Similarly, for ezlopitant metabolites, it was shown that the
cerebrospinal fluid (CSF) to free plasma ratio in subarachnoid catheterised
dogs was much greater for a non-hydroxyl metabolite vs. a benzyl alcohol
metabolite despite the fact that the benzyl alcohol metabolite was present in
much greater concentrations in plasma.9 The 7-hydroxy metabolite of metho-
trexate was considerably less distributed to tissues in the rat relative to the
parent drug.10

10.3.3 Metabolism and Interaction with Drug-metabolising

Enzymes
The metabolism of hydroxy groups in drugs includes both oxidative and con-
jugative reactions. Primary alcohols are oxidised to aldehydes, which are
usually further oxidised to carboxylic acids. Secondary alcohols are oxidised to
ketones. In addition to these oxidative pathways, alcohols also can undergo
conjugative metabolism. The enzymes catalysing the reaction from alcohol to
aldehyde or ketone metabolites include alcohol dehydrogenases, cytochrome
P450s (e.g. CYP2E1), and prostaglandin H synthease. Phenols undergo con-
jugative metabolism including glucuronidation, sulfation and methylation. In
general, glucuronidation and sulfation have greater impacts on the pharma-
cokinetics and bioavailability of hydroxyl containing drugs than oxidative
metabolism or methylation due to the facile elimination and enterohepatic
circulation of conjugates.
Both oxidative and conjugative metabolism of the hydroxy group has the
potential to generate reactive metabolites. The oxidation of alcohols yields
aldehydes which can covalently bond to macromolecules and potentially cause
toxicity. Many phenol drugs and catechols can form quinoid structures that can
react with glutathione (GSH) (a detoxication pathway) or nucleophilic groups
on proteins and nucleic acids. Sulfation of benzylic and allylic alcohols has been
documented as a mechanism of bioactivation of alcohol drugs. These are dis-
cussed in greater detail below.

10.3.3.1 Enzymes of Oxidative Metabolism

10.3.3.1.1 Alcohol Dehydrogenase (ADH). Alcohol dehydrogenases are a
family of zinc-containing isozymes that catalyse the reversible oxidation of
alcohols to their corresponding ketones or aldehydes using NAD1 or NADP1 as
cofactors. (The reverse reaction uses the reduced cofactors.) Mammalian ADHs
are located almost exclusively in the cytoplasm of cells. They have been cate-
gorised into six classes, of which the first five have been identified in humans;11,12
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 469
ADH-1 is the best known among the classes and is the enzyme mainly respon-
sible the metabolism of ethanol. ADH-1 is widely distributed in human tissues,
with the greatest abundance in liver. Class II ADH activity has been found
mainly in liver and shows a high KM value for ethanol. Class III ADH is a poor
ethanol dehydrogenase, but has been proposed to play a significant role in
cytoprotection by metabolism of formaldehyde. Class IV is extra-hepatically
expressed with high ethanol metabolising activity in stomach mucosa.13 Class V
ADH transcripts were found to be at significantly higher levels in fetal liver
compared with adult liver, suggesting that human class V is a predominant fetal
alcohol dehydrogenase.13 Several polymorphisms in genes coding for alcohol
dehydrogenases have been well characterised. Although ADHs are important in
the metabolism of ethanol, susceptibility to alcoholism and alcoholic liver dis-
ease, they are generally not of great importance in the metabolism of commonly
used drugs.14 ADHs have broad substrate specificities; however, the class I iso-
zymes are mostly responsible for the oxidation of ethanol and other small, ali-
phatic alcohols. Class II and Class III enzymes preferentially oxidise aromatic or
medium to long chain alkyl chain aliphatic alcohols.
Endogenous compounds metabolised by ADH include steroids and retinol. A
number of drugs are metabolised by ADH. The primary alcohols are oxidised to
aldehydes which usually are further oxidised to corresponding carboxylic acids.
The subsequent oxidation of aldehyde to carboxylic acid is facile and often
involves other enzymes such as aldehyde dehydrogenase; in some cases, CYPs
are also involved. Therefore, it is difficult to establish the identity of the parti-
cular alcohol-oxidising enzyme in vivo. ADHs appear to be involved in humans
in the metabolism of ethambutol and celecoxib.15,16 In both cases, the end
products are carboxylic acids.17 In the metabolism of the reverse transcriptase
inhibitor abacavir,18 ADH1A is able to carry out not only the step from aba-
cavir (an alcohol) to the corresponding aldehyde, but also the second step from
the aldehyde to the acid (Figure 10.2). Although secondary alcohols are sus-
ceptible to oxidation to their corresponding ketones, this reaction is not often as
important in drug metabolism because the reverse reaction occurs readily and
also because they can form glucuronides that are rapidly eliminated.

10.3.3.1.2 Cytochrome P450. Cytochrome P450s (CYPs) also play a role

in the metabolism of hydroxy groups in drugs as well as ethanol. The micro-
somal oxidation of ethanol has been the subject of extensive research since
its discovery in the 1970s.19 Recently, it has been demonstrated that most
CYP isoforms (except CYP2A6 and CYP2C18) are capable of oxidising
ethanol to acetaldehyde with KM value around 10 mM.20 Cytochrome
P4502E1 (CYP2E1), which is induced by chronic ethanol consumption, was
found to have a prevailing role in this oxidation. P450s are also involved in
the metabolism of many hydroxy containing drugs. At the normal ther-
apeutic doses, acetaminophen is safe; its major metabolic pathways include
glucuronidation and sulfation, and only small amounts of the reactive inter-
mediate, a quinonimine, are formed by the CYP enzymes (see Chapter 6,
470 Chapter 10
OH
OH
H ADH
H3C N H
N CH3 H3C N
H N CH3
H
HO
HO O

Ethambutol

H3C H3C
N N O
N O ADH/ALDH N O
OH OH

Hydroxyzine Cetirizine

OH
O NH2 Hydrolysis

O NH2
O NH2
O
O
ADH
Felbamate
H
H
-CO2, + NH2 O
O

CH 2 O NH 2

O
Atropaldehyde
CF3 CF3
CF3

N N
N HO HO N
N N
CYP2C9 ADH1/ADH2
O

O S O O S O
O S O
NH2 NH2
NH2

Celecoxib
OH H O HO O

N ADH N ADH N
N N N
N N N

H 2N N NH H 2N N H 2N N NH
NH

Abacvir

Figure 10.2 Hydroxy containing drugs metabolised by ADH.

Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 471
section 2). Among several CYP isoforms including CYP3A4, CYP1A2 and
CYP2E1, CYP2E1 was found to play a major role in the bioactivation of
acetaminophen causing hepatotoxicity when large doses are ingested.21,22
P450 enzymes are also frequently involved in the further metabolism of the
intermediary hydroxyl metabolites also generated by these enzymes. For
example, the antihistamine, terfenadine, is sequentially oxidised by CYP3A4 to
an alcohol metabolite that is converted into the circulating active metabolite,
fexofenadine, which is a carboxylic acid. Losartan, an angiotensin II antagonist
possessing a hydroxymethyl group, is metabolised by CYP2C9 to the corre-
sponding aldehyde and carboxylic acid metabolites (the latter is referred to as
E-3174),23,24 with the acid metabolite being largely responsible for efficacy.25
These findings have been verified in vivo, with marked differences in the
urinary ratios of losartan/E-3174 across populations of different CYP2C9
genotypes.26–28 Possession of one *3 allele results in 2–3 times lower plasma
exposure to E-3174 and *3 homozygous individuals have very markedly lower
exposure to E-3174, consistent with the diminished activity of CYP2C9*3 to
convert losartan to E-3174, but not as much of a difference in losartan expo-
sure. It is likely that the carboxylic acid metabolite contributes to the activity
since the blood pressure lowering ability of losartan is diminished in patients
possessing a *3 allele.29,30 From the examples of terfenadine and losartan, it is
clear that alcohol containing drugs could yield pharmacologically active car-
boxylic acid metabolites.

10.3.3.1.3 Prostaglandin H Synthetase. PGHS is a haem-containing pro-

tein with two enzymatic activities: a cyclooxygenase (COX) and a peroxidase.
To date, two structurally related isoforms of PGHS have been identified,
COX-1 and -2. Both isoforms participate in the synthesis of important biolo-
gical mediators called prostanoids (including prostaglandins, prostacyclin
and thromboxane) [PHS-1]. During the synthesis of prostaglandin H2
(PGH2), the precursor of the series-2 prostanoids, COX catalyses bioox-
ygenation of free fatty acid substrates and converts arachidonate to PGG2
(endoperoxide–hydroperoxide); and the peroxidase activity reduces this inter-
mediate to the corresponding alcohol (PGH2) in the presence of reducing co-
substrates. Since phenol and hydroquinone are good reducing co-substrates
as electron donors for the peroxidase of PGHS,31 these phenolic compounds
undergo one electron oxygenation to form reactive species which can bind to
macromolecules (PGHS-2).
The PGHS-catalysed oxidation of paracetamol is an example. The one-
electron oxidation product N-acetyl-p-benzo-semiquinone imine may be fur-
ther oxidised to N-acetyl-p-quinoneimine (NAPQI) or dimerise; it can also
undergo disproportionation, forming paracetamol and NAPQI. PGHS can
catalyse both one- and two-electron oxidations of paracetamol. In contrast,
CYP metabolism occurs by two-electron oxidation and did not produce oli-
gomers, but instead NAPQI (Figure 10.3).
472 Chapter 10

two-electron oxidation
P450/PGHS

NHCOCH3 NHCOCH3 NCOCH3

e- + H+ e- + H+

PGHS PGHS
OH O O
2x semiquinone

GSH GSH
2x

NHCOCH3

NHCOCH3
NCOCH3

OH SG
OH

OH
O

NHCOCH3

Figure 10.3 Bioactivation of acetaminophen by PGHS.

10.3.3.2 Conjugative Metabolism of Hydroxy Containing Drugs

Compared with oxidative reactions, conjugative metabolic reactions such as
glucuronidation and sulfation have a greater prevalence in the metabolism of
alcohol containing drugs. Furthermore, the impact on the pharmacological
activity and ADME properties of the drug are substantially altered following
conjugation. In most cases, the products of phase II metabolism are not
pharmacologically active and are readily eliminated due to their greatly
increased polarity.
Many hydroxy containing drugs have excellent oral absorption, but due to
high ‘first-pass’ effects largely attributable to phase II metabolism, their oral
bioavailability can be poor. For example, the analgesic drug meptazinol (Fig-
ure 10.4) is rapidly and completely absorbed after oral and intramuscular
administration.32 However, the absolute bioavailability of the drug following
oral dosage is low (4.5–8.7%), whereas the drug is totally systemic available
after intramuscular (i.m.) dosage. The elimination of the drug proceeds rapidly
(t1/2 ¼ B2h), largely due to glucuronidation and sulfation of the phenolic
function in the molecule. Following oral administration of 3H-labelled drug,
90% of the radioactivity was excreted through urine where no parent drug was
detected.33 The highly efficient phase II metabolism of meptazinol can be
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 473

OH

Figure 10.4 Structure of meptazinol.

blocked by making benzoyl esters as prodrugs to improve oral bioavailability.34

Many dietary flavonoids and other polyphenols also exhibit poor oral bioa-
vailability, mainly due to highly efficient glucuronic acid and sulfate conjuga-
tion of these mono- or polyhydroxylated agents. Methyl capping of all free
hydroxyl groups of flavones results in dramatically increased metabolic stabi-
lity, as the metabolism is shifted to less efficient CYP-mediated oxidation.35
When hydroxy containing drugs undergo extensive glucuronidation and the
glucuronic acid conjugates are biliarily excreted, enterohepatic circulation can
occur and cause prolonged drug exposure and multiple peaks in plasma con-
centration vs. time profiles.

10.3.3.2.1 Uridine 5 0 -diphospho-glucuronosyltransferase (UGT). UGTs are

one of the most important drug metabolising enzyme families. They catalyse
the addition of glucuronic acid from UDP-glucuronic acid (UDPGA) to a
multitude of endobiotic and xenobiotic compounds.36 The products of glu-
curonidation are generally more polar, less toxic and more easily excreted
from the body through bile or urine. UGTs have a wide tissue distribution
with the liver possessing the highest level of activity. UGT enzymes are sub-
divided into two families designated as UGT1A and UGT2A/UGT2B.37 To
date, more than 60 different UGT enzymes have been isolated in several
mammalian species.38 These transferases exhibit distinct but overlapping sub-
strate specificities and are known to catalyse the glucuronidation of a variety
of alcoholic and phenolic drugs.
Glucuronidation has a profound impact on the ADME properties of hydroxy
containing drugs. This can be illustrated with diazepam and temazepam (Figure
10.5). They both have common metabolic pathways such as N-demethylation,
p-hydroxylation, and glucuronidation of p-hydroxy metabolites. The major
difference is that temazepam has a hydroxy group which undergoes direct glu-
curonidation extensively.
In humans, temazepam is eliminated mainly through urine accounting for
80% of dose, of which 88% was the direct glucuronide.39 While in rats, the
urinary excretion only accounted for 15% of dose; the majority of dose was
eliminated though biliary excretion (485%).39,40 When temazepam glucuronide
is excreted through bile, it undergoes enterohepatic recirculation which gives
rise to the prolonged duration of drug in the body. Therefore in rats, temazepam
has a longer half-life and lower CL than its non-hydroxy counterpart, diazepam
474 Chapter 10

Cl Cl
N N
OH
N N
O O

Diazepam Temazepam

Figure 10.5 Structures of diazepam and temazepam.

(i.e. T1/2 and CL are 4.0 h and 45 mL min1 kg1 for temazepam, and 1.1 h and
80 mL min1 kg1 for diazepam, respectively).41,42 In humans, although the
major metabolite of temazepam is also direct glucuronide, the excretion of the
glucuronide is mainly through urine. The enterohepatic circulation is minimal,
so the half-life is 8–10 h for temazepam and 32–33 h for diazepam.41,43
A single hydroxy group in the molecule makes significant difference in the
pharmacokinetics and pharmacodynamics of temazepam and diazepam. The
former is categorised as a short-acting benzodiazepine used mainly as a hyp-
notic, and the latter is the long-acting one used for the treatment of anxiety.

10.3.3.2.2 Sulfotransferases. Hydroxy containing drugs also undergo sul-

fate conjugation. The reaction is catalysed by sulfotransferase (SULT) using
3 0 -phosphoadenosine-5 0 -phosphosulfate (PAPS) as cofactor. In mammals,
SULTs are classified into five families (SULT 1–5), with SULT1 and SULT2
as the most important for xenobiotic metabolism. Conjugation of xenobiotics
with sulfate mainly occurs with phenols and occasionally with alcohols, aro-
matic amines and hydroxylamines.
Sulfate conjugation of phenolic compounds is less prevalent than glucur-
onidation. When a phenolic compound is the substrate for both glucuronida-
tion and sulfation, these two metabolic pathways compete with each other.
Paracetamol is metabolised by both UGT and SULT, and at normal doses in
adults the O-glucuronide is the major urinary metabolite and the O-sulfate the
minor one. At low doses, the major metabolite is sulfate conjugate. As dose
increases, the proportion of the acetaminophen conjugated with sulfate
decreases, whereas the proportion of glucuronic acid conjugate increases. This
is due to the different affinity and capacity that UGT and SULT possess; UGT
has low affinity but high capacity, with the SULT being the opposite. The low
capacity of sulfation is limited by the co-substrate PAPS availability due to
limited supply of inorganic sulfate.44 Interestingly, in infants and children, the
predominant urinary metabolite of paracetamol is the sulfate conjugate, since
in neonates and young children, glucuronidation capacity is reduced due to the
undeveloped glucuronosyltransferase.45,46 In general, alcohols have less pre-
valence of being conjugated with sulfuric acid than phenols; however, sulfate
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 475
conjugates of benzylic alcohols, allylic alcohols and aromatic hydroxylamines
are known to be toxic and can covalently bind to cellular macromolecules,
DNA and RNA.47,48 These are discussed later.

10.3.3.2.3 Methyltransferases. Catechols encompass a wide range of phy-

siological, biologically and medically importance substances such as catecho-
lamines, catecholestrogens, flavonoids and drugs used for the treatment of
Parkinson’s disease.
Phenolic drugs also undergo oxidative metabolism to form catechols. The
metabolic fate of catechols is frequently via O-methylation catalysed by cate-
chol-O-methyltransferase, which transfers a methyl group from S-adeno-
sylmethionine (SAM) to the m-hydroxyl group of catechol compounds. In
mammals, there is only one gene located on chromosome 22q11 for COMT;
this single gene encodes both the soluble COMT (S-COMT) and the mem-
brane-bound COMT (MB-COMT).49 COMT is widely distributed in mam-
malian tissues with the highest level in liver, followed by kidney and
gastrointestinal tract.50–53 COMT is also presented in spleen, pancreas, lung,
heart, eye and spinal membrane.54–57 The soluble form (S-COMT) are the
dominant form in most rat or human tissues, and only a small fraction is
present in the membrane-bound form (MB-COMT). The exception is human
brain where 70% of the total COMT proteins were found to be MB-COMT
and 30% S-COMT.
COMT plays a major role in the metabolic inactivation of catecholic neu-
rotransmitters such as dopamine, norepinephrine and epinephrine. It also is
responsible for the inactivation of catecholestrogens such as 2-hydroxyestrone
and 4-hydroxyestrone. In addition to inactivating endogenous catecholamines
and catechol estrogens (Figure 10.6), COMT also catalyses the O-methylation
metabolism of a large number of catechol containing xenobiotics. Many dietary
bioflavonoids such as quercetin, fisetin and tea polyphenols are exceptionally
good substrates for O-methylation by cytosolic COMT from porcine liver or
hamster kidney.58 The rates of the COMT-mediated O-methylation metabolism
of these bioflavonoids were up to several orders of magnitude higher than
those for the endogenous catechols. Many phenolic drugs undergo P450-
mediated oxidation to form catechols and are subsequently methylated by
COMT. For example, some neuroactive drugs possessing a catechol structure
[e.g. L-3,4-dihydroxyphenylalanine (L-Dopa)] are also O-methylated by COMT
(Figure 10.7).59 Traxoprodil (TRX), a selective N-methyl-d-aspartate receptor
antagonist, is a phenolic drug and it first undergoes P450-mediated oxidation to
form a catechol, and is then further metabolised by COMT.60

10.3.3.3 Bioactivation of Hydroxy Containing Compounds to

Reactive Metabolites
Primary aliphatic alcohols can be oxidised to aldehydes, which are hard
electrophiles capable of covalently binding to macromolecules; however, the
476 Chapter 10

1) Catecholamin

HO NH2 O NH2
COMT

HO SAM SAH
HO

Dopamine 3-Methoxytyramine

2) Catechoestrogen
O
O
COMT
O
HO
SAM SAH
HO
HO

2-Hydroxyestrone 2-Methoxyestrone

3) Flavonoids
OH O
OH OH
COMT
HO O HO O
SAM SAH
OH OH
OH O OH O

Quercetin Isohamnetin

4) Coumarins

OH O
OH OH
COMT

SAM SAH

O O
Esculetin Scopoletin

Figure 10.6 COMT mediated biotransformation of neurotransmitters, catecholes-

trogens, dietary bioflavonoids and coumarins.
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 477

O O
HO COMT O
OH OH
NH2 SAM SAH NH2
HO HO

Levodopa Methyl Dopa

OH
HO O
HO
HO
N
P450 COMT
N
OH OH N
OH OH
SAM SAH
OH OH

Traxoprodil (TRX) Hydroxy TRX Methoxy TRX

Figure 10.7 COMT-mediated methylation biotransformation of drugs.

capacity of aldehyde dehydrogenase is typically high such that toxic aldehydes

do not accumulate in general. In cases of excess ethanol consumption, acet-
aldehyde can build up and cause toxicity. Allyl alcohol and crotyl alcohol can
be bioactivated by alcohol dehydrogenases to aldehydes that can react with
proteins and cause toxicity.61 Longer chain polyfluorinated alcohols are
environmental contaminants shown to be bioactivated to toxic aldehydes by
P450 enzymes.62 Abacavir is a reverse transcriptase inhibitor marketed for the
treatment of HIV-1 infection.
A small percentage of patients experience a hypersensitivity reaction indi-
cating immune system involvement and bioactivation. Abacavir is metabolised
to an aldehyde by ADH, although the exact mechanism causing hypersensi-
tivity remain unclear. The formation of aldehyde and its covalent binding to
proteins was thoroughly studied by the incubation of 14C labelled abacavir in
human cytosol or human isoforms of alcohol dehydrogenase (ADH) in the
presence of human albumin63 (Figure 10.8).
Phenols can undergo oxidative bioactivation to quinones, quinonemethides
or quinonimines depending on the structure of the drug. Such intermediates are
good Michael acceptors and stable adduct formation is driven by the re-aro-
matisation of the adduct. The bioactivation of paracetamol has been well
characterised and serves as a quintessential example of drug bioactivation
causing hepatic toxicity.64 Overdoses of paracetamol are common and repre-
sent an important healthcare concern. Ordinarily, conjugative pathways handle
the clearance of this drug, but when this capacity is exceeded or if the drug is
taken with ethanol which induces CYP2E1, substantial oxidative metabolism
can occur in which the phenol is oxidised to a chemically reactive quinoneimine
(Figure 10.9).
Remoxipride itself does not contain a phenol, but it is metabolised to two
phenol intermediates, one via O-demethylation and one via hydroxylation,
which are subsequently oxidised to a p-quinone. This pathway has been
478 Chapter 10

OH HO O
H O

N ADH ADH N
N
N N N
N N
N

H2N H2N N NH
N NH H2N N NH

Abacavir Intermediary aldehyde metabolite Acid metabolite

Protein covalent binding

hypersensitivity

Figure 10.8 Bioactivation of abacavir.

O O O

HN OH HN OH HN OH
Sulfotransferase UDP Glucuronide

Transferase

OSO3- OH O-Glu

P450

O
O
N OH
HN OH
Glutathione
Covalent Binding
to SH groups
Transferase
S-glutathione
O
OH
N-acetylbenzoquinone imine
Cell death
NAPQI

Figure 10.9 Metabolism and bioactivation of acetaminophen (paracetamol).

Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 479
O OMe O OMe O OMe
N N OH
N P450 N P450 N OH
1) H H N
H
MeO MeO MeO
Br Br Br
Remoxipride
P450

O OMe O OMe
O OMe
N OH N O [O]
N GSH N OH
N
H H N
H
HO O
HBr HO
SG Br
Br

OH OH
OH
DNA Damage
P450 [O]

2) Protein Damage
HO O
HO
OH O

17-Estradiol

OMe OMe
OMe
OMe OMe
OMe

N N
N

3) P450 GSH
O OH
OH
O O
O OMe
OMe OMe*
MeO MeO
MeO

SG

N N
N

Dauricine

Cl Cl Cl
Cl
P450 [O] N GSH NH
NH
NH
Cl O Cl O Cl O
4) Cl O
OH OH OH
OH GS

OH O OH

Diclofenac

Figure 10.10 Examples of phenol containing drugs that are bioactivated.

implicated as responsible for the idiosyncratic toxicity that led to the with-
drawal of this agent from clinical use65 (Figure 10.10). Other examples of drugs
and endobiotics that undergo bioactivation to quinoid structures include the
estrogens, dauricine and diclofenac.66–68 In these cases, the phenols are oxidised
to highly conjugated quinonemethides. Diclofenac is oxidised to 5-hydro-
xydiclofenac by CYP3A; this p-aminophenol metabolite can then undergo a
second oxidation to a reactive quinoneimine, which may be involved in the
hepatotoxicity of this drug.
480 Chapter 10
Conjugative metabolism can also bioactivate alcohols to electrophiles. Sul-
fation of benzylic and allylic alcohols is a well-established mechanism of
bioactivation. The sulfate ester is reactive by virtue of the excellent leaving
group properties of the sulfate group and reactions occur with tissue nucleo-
philes (e.g. DNA) via SN2 reactions. For example, tamoxifen is well-established
as a hepatic carcinogen and clinical data have shown a link between tamoxifen
therapy and an increased incidence of endometrial cancer.69 The major meta-
bolic pathway is N-demethylation. Tamoxifen also forms two hydroxy meta-
bolites; one metabolite, a-hydroxy-tamoxifen has an allylic hydroxyl group that
undergoes sulfation. The resulting sulfate is hypothesised to be responsible for
the formation of DNA adducts48,70 (Figure 10.11).

10.3.4 Fluorine as an Isostere of Hydroxy Groups

When a lead compound is identified during drug discovery, a common chemical
strategy is to use bioisosteric substitution to synthesise new compounds that
have less undesirable characteristics (e.g. low bioavailability, inadequate half-
life and potential to form reactive metabolites).
One of the most common classical bioisosteric substitution is the incor-
poration of fluorine into a compound in replacement of a hydroxyl group.71
The rationale for such replacement is based on the fact that the size of the
fluorine atom is intermediate between that of hydrogen and oxygen, and that
the substitution of the hydroxy group with fluorine is particularly favoured
when the presence of an electronegative atom is necessary for the interaction of
the ligand with the target protein.72,73 The van der Waals radius of fluorine
(1.35 Å) is intermediate between that of hydrogen (1.2 Å) and oxygen (1.40 Å).
The C–F bond length (1.39 Å) is closest to the C–O bond (1.43 Å) compared
with the C–H bond (1.09 Å) and other halogens.73 Fluorine has the strongest
electronegativity in the periodic table, and the next strongest is oxygen. Both
fluorine and a hydroxy group can form hydrogen bonds, though the difference
between the fluorine atom and hydroxy oxygen is that the former can only act
as the hydrogen bond acceptor and cannot replace the hydrogen bonding donor
role as it is devoid of the acidic hydrogen of the hydroxy group.

N N
N
O O
O

DNA Adducts

O OSO- +

Figure 10.11 Bioactivation of tamoxifen via an allylic sulfate.

Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 481
The substitution of a hydroxy group with fluorine, in theory, will have
profound effect on the rate and extent of metabolism for those compounds
which would undergo predominantly phase II metabolism on the substituted
hydroxy group. Such phase II metabolism includes methylation, acetylation,
sulfation and glucuronidation. Many phenolic hydroxyl compounds have the
potential for bioactivation when they are present either ortho- or para- to a
primary or secondary aryl amine group, or a methylene group. Subsequent
oxidation can generate ortho- or para-quinone imine and quinone methide,
groups which are associated with toxicity in vivo. There are numerous examples
in the literature citing the elimination of such a bioactivation through the
substitution with fluorine.
Amodiaquine (AQ) (Figure 10.12) is a phenolic 4-aminoquinoline anti-
malarial effective against chloroquine-sensitive and resistant strains of Plas-
modium falciparum.74 However, idiosyncratic adverse drug reactions have
severely restricted its use. Both direct cytotoxic and indirect immunogenic
causal mechanisms have been proposed to explain the toxicity observed.75 AQ
is known to undergo extensive bioactivation to a chemically reactive quinone
imine (AQQI) in vivo in the rat.76 Fluorine substitution at the C-4 position of

OH O

N N

[O] Binding to cellular proteins

NH N Hepatotoxicity
P450 Agranulocytosis

Cl N Cl N
Amodiaquine AQQI
GSH

F GSH-conjugate (11.8% of dose)

NH
Blocked Bioactivation (0% of dose)

Cl N

4-deOH-4F-Amodiaquine

Figure 10.12 Fluorine substitution of hydroxy group in amodiaquine blocked the

bioactivation in CD1 mice.
482 Chapter 10
AQ completely blocked bioactivation, since none of the dose resulted in
thioether conjugates. This is in contrast to 10.9% of a dose of AQ excreted as
thioether conjugate in CD1 mice77 (Figure 10.12).

10.4 Conclusions
The hydroxyl group represents an important substituent in drug design. Its
specific hydrogen-bonding capability can enhance or disrupt interactions with
macromolecule targets. The physicochemical properties that it imparts to drug
molecules can have a large impact on dispositional behaviour and result in
alterations in the pharmacokinetics relative to analogues lacking this group.
Furthermore, inclusion of a hydroxyl group in a molecule offers a new handle
which can be acted upon by drug-metabolising enzymes and metabolism can be
shifted from cytochrome P450 enzymes to conjugative enzymes such as UGTs.

References
1. R. S. Obach, F. Lombardo and N. Waters, Drug Metab. Dispos., 2008, 36,
1385.
2. J. Yu and Y. W. Chien, Pharm. Dev. Technol., 2002, 7, 215.
3. M. Aravagiri and S. R. Marder, Psychopharmacology, 2002, 159, 424.
4. T. B. Ejsing, A. D. Pedersen and K. Linnet, Hum. Psychopharmacol., 2005,
20, 493.
5. L. E. C. van Beijsterveldt, R. J. F. Geerts, J. E. Leysen and A. A. H. P.,
Psychopharmacology, 1994, 114, 53.
6. K. M. Kirschbaum, S. Henken, C. Hiemke and U. Schmitt, Behav. Brain
Res., 2008, 188, 298.
7. A. Doran, R. S. Obach, B. J. Smith, N. A. Hosea, S. Becker, E. Callegari, C.
Chen, X. Chen, E. Choo, J. Cianfrogna, L. M. Cox, J. P. Gibbs, M. A.
Gibbs, H. Hatch, C. E. C. A. Hop, I. N. Kasman, J. LaPerle, J.-H. Liu, X.
Liu, M. Logman, D. Maclin, F. M. Nedza, F. Nelson, E. Olson, S. Rahe-
matpura, D. Raunig, S. Rogers, K. Schmidt, D. K. Spracklin, M. Szewc, M.
Troutman, E. Tseng, M. Tu, J. W. van Deusen, K. Venkatakrishnan, G.
Walens, E. Q. Wang, D. Wong, A. S. Yasgar and C. Zhang, Drug Metab.
Dispos., 2005, 33, 165.
8. B. Feng, J. B. Mills, R. E. Davidson, R. J. Mireles, J. S. Janiszewski, M. D.
Troutman and S. M. de Morais, Drug Metab. Dispos., 2008, 36, 268.
9. A. E. Reed-Hagen, M. Tsuchiya, K. Shimada, J. Wentland and R. S.
Obach, Biopharm. Drug Dispos., 1999, 20, 429.
10. L. Slordal, R. Jaeger, J. Kjaeveand and J. Aarbakke, Pharmacol. Toxicol.,
1988, 63, 81.
11. G. Duester, J. Farres, M. R. Felder, R. S. Holmes, J. O. Höög, X. Parés, B.
V. Papp, S. J. Yin and H. Jörnvall, Biochem. Pharmacol., 1999, 58, 3.
12. J. O. Hoog, P. Stromberg, J. J. Hedberg and W. J. Griffiths, Chem-Biol.
Interact., 2003, 143–144, 175.
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 483
13. M. Estonius, S. Svensson and J. O. Hoeoeg, FEBS Letters, 1996, 397, 338.
14. A. Parkinson, Biotransformation of Xenobiotics, in Casarett and Doull’s
Toxicology, ed. Curtis D. Klassen. McGraw-Hill., New York, 2001, Sixth
Edition, pp. 152–154.
15. M. Sandberg, Y. Uemit, P. Stromberg, J. O. Hoeoeg and E. Eliasson, Br. J.
Clin. Pharmacol., 2002, 54, 423–429.
16. M. Breda, M. Strolin-Benedetti, M. Bani, C. Pellizzoni, I. Poggesi, G.
Brianceschi, M. Rocchetti, L. Dolfi, D. Sassella and R. Rimoldi, Phar-
macol. Res., 1999, 40, 351.
17. M. Sandberg, U. Yasur, P. Stromberg, J. O. Hoog and E. Eliasson, Br. J.
Pharmacol., 2002, 54, 423.
18. J. S. Walsh, M. J. Reese and L. M. Thurmond, Chem.-Biol. Interact., 2002,
142, 135.
19. L. K. Low and N. Castangnoli, Metabolic Changes of Drugs and Related
Organic Compounds, Wilson and Gisvold’s Textbook of Organic Medicinal
and Pharmaceutical Chemistry, eds. J. N. Delgado and W. A. Remers, JB
Lippincott Company, PA., USA, 1991, 9th edition, pp. 83–84.
20. S. Hamitouche, J. Poupon, Y. Dreano, Y. Amet and D. Lucas, Toxicol.
Lett., 2006, 167, 221.
21. P. T. Manyike, E. D. Kharasch, T. F. Kalhorn and J. T. Slattery, Clin.
Pharmacol. Ther., 2000, 67, 275.
22. R. L. Rose and P. E. Levi, in A Textbook of Modern Toxicology, ed. E.
Hodgson, 2004, Wiley, Hoboken, NJ, Ch. 8, pp. 149–162.
23. R. A. Stearns, P. K. Chakravarty, R. Chen and S. H. L. Chiu, Drug Metab.
Dispos., 1995, 23, 207.
24. U. Yasar, G. Tybring, M. Hidestrand, M. Oscarson, M. Ingelman-Sund-
berg, M. L. Dahl and E. Eliasson, Drug Metab. Dispos., 2001, 29, 1051.
25. J. Turgeon, Prog. Exper. Cardiol., 1998, 2, 153.
26. U. Yasar, M. L. Dahl, M. Christensen and E. Eliasson, Br. J. Clin.
Pharmacol., 2002, 54, 183.
27. U. Yasar, C. Forslund-Bergengren, G. Tybring, P. Dorado, A. Lerena, F.
Sjoeqvist, E. Eliasson and M. L. Dahl, Clin. Pharmacol. Ther., 2002, 71, 89.
28. M. O. Babaoglu, U. Yasar, M. Sandberg, E. Eliasson, M. L. Dahl, S. O.
Kayaalp and A. Bozkurt, Eur. J. Clin. Pharmacol., 2004, 60, 337.
29. K. Sekino, T. Kubota, Y. Okada, Y. Yamada, K. Yamamoto, R. Horiuchi,
K. Kimura and T. Iga, Eur. J. Clin. Pharmacol., 2003, 59, 589.
30. M. Lajer, L. Tarnow, S. Andersen and H. H. Parving, Diabetic Med., 2007,
24, 323.
31. M. J. Schlosser, R. D. Shurina and G. F. Kalf, Environ. Health Perspect.,
1989, 82, 229.
32. R. A. Franklin, Xenobiotica, 1988, 18, 105.
33. G. R. Murray, G. M. Whiffin, R. A. Franklin and J. A. Henry, Xenobio-
tica, 1989, 19, 669.
34. M. Lu, C. Zhang, J. Hao and Z. Qiu, Bioorg. Med. Chem. Lett., 2005, 15,
2607.
35. T. Walle, Molec. Pharmaceutics, 2007, 4, 826.
484 Chapter 10
36. B. Burchell, Transformation Reactions: Glucuronidation, in Handbook of
Drug Metabolism, ed. T. F. Woolf, Marcel Dekker Inc., New York, 1999,
pp. 153–173.
37. D. W. Nebert, D. R. Nelson, B. Burchell and K. W. Bock, DNA Cell Biol.,
1991, 10, 487.
38. C. Albert, O. Barbier, M. Vallee, G. Beaudry, A. Belanger and D. W. Hum,
Endocrinology, 2000, 141, 2472.
39. H. J. Schwarz, Br. J. Clin. Pharmacol., 1979, 8, 23s.
40. J. Escoriaza, M. C. Dios-Vieitez, I. F. Troconiz, M. J. Renedo and D. Fos,
Chromatographia, 1997, 44, 169.
41. M. Mandelli, G. Tognoni and S. Garattini, Clin. Pharmacokinet., 1978, 3, 72.
42. F. L. S. Tse, F. Balliard and J. M. Jaffe, J. Pharm. Sci., 1983, 72, 31.
43. D. D. Breimer, R. Jochemsen and H. H. von Albert, Arzneim.-Forsch.,
1980, 30, 875.
44. Z. Gregus, H. J. Kim, C. Madhu, V. Liu, P. Rozman and C. D. Klaassen,
Drug Metab. Dispos., 1994, 22, 725.
45. R. P. Miller, R. J. Roberts and L. J. Fischer, Clin. Pharmacol. Ther., 1976,
19, 284.
46. A. Parkinson, in Casaret and Doull’s Toxicology: The Basic Science of
Poisons, ed. C. D. Klaassen, McGraw-Hill, New York, 1996, Ch. 6, pp.
113–186.
47. L. Yi, J. Dratter, C. Wang, J. A. Tunge and H. Desaire, Anal Bioanal
Chem., 2006, 386, 666.
48. E. Banoglu, Curr. Drug Metab, 2000, 1, 1.
49. B. T. Zhu, Curr. Drug Metab., 2002, 3, 321.
50. J. A. Roth, M. H. Grossman and M. Adolf, Biochem. Pharmacol., 1990, 40,
1151.
51. E. Nissinen, R. K. Tuominen, V. Perhoniemi and S. Kaakkola, Life Sci.,
1988, 42, 2609.
52. P. T. Männistö, I. Ulmanen, K. Lundström, J. Taskinen, J. Tenhunen, C.
Tilgmann and S. Kaakkola, Prog. Drug Res., 1992, 39, 291.
53. B. Meister, A. J. Bean and A. Aperia, Kidney Int., 1993, 44, 726.
54. T. Karhunen, C. Tilgmann, I. Ulmanen, I. Julkunen and P. Panula, J.
Histochem. Cytochem., 1994, 42, 1079.
55. C. De Santi, P. C. Giulianotti, A. Pietrabissa, F. Mosca and G. M. Pacifici,
Eur. J. Clin. Pharmacol., 1998, 54, 215.
56. L. J. Bryan-Lluka, Arch.Pharmacol., 1995, 351, 408.
57. Y. S. Ding, S. J. Gatley, J. S. Fowler, R. Chen, N. D. Volkow, J. Logan, C.
E. Shea, Y. Sugano and J. Koomen, Life Sci., 1996, 58, 195.
58. B. T. Zhu, U. K. Patel, M. X. Cai and A. H. Conney, Drug Metab. Dispos.,
2000, 28, 1024.
59. H. C. Guldberg and C. A. Marsden, Pharmacol. Rev., 1975, 27, 135.
60. C. Prakash, D. Cui, M. J. Potchoiba and T. Butler, Drug Metab. Dispos.,
2007, 35, 1350.
61. F. R. Fontaine, R. A. Dunlop, D. R. Petersen and P. C. Burcham, Chem.
Res. Toxicol., 2002, 15, 1051.
Alcohols and Phenols: Absorption, Distribution, Metabolism and Excretion 485
62. J. W. Martin, S. A. Mabury and P. J. O’Brien, Chem.-Biol. Int., 2005
155, 165.
63. J. S. Walsh, M. J. Reese and L. M. Thurmond, Chem.-Biol. Int., 2002, 142,
135.
64. D. C. Dahlin, G. T. Miwa, A. Y. H. Lu and S. D. Nelson, Proc. Natl. Acad.
Sci. U. S. A., 1984, 81, 1327.
65. J. C. L. Erve, M. A. Svensson, H. von Euler-Chelpin and E. Klasson-
Wehler, Chem. Res. Toxicol., 2004, 17, 564.
66. J. L. Bolton, in Advances in Molecular Toxicology, ed. J. C. Fishbein,
Elsevier, Amsterdam, 2006, vol. 1, pp. 1–23.
67. Y. Wang, D. Zhong, X. Chen and J. Zheng, Chem. Res. Toxicol., 2009, 22,
824.
68. U. A. Boelsterli, Toxicol. Appl. Pharmacol., 2003, 192, 307.
69. C. J. Cohen, Semin. Oncol., 1997, 24, S55.
70. L. M. Notley, K. H. Crewe, P. J. Taylor, M. S. Lennard and E. M. J.
Gillam, Chem. Res. Toxicol., 2005, 18, 1611.
71. B. K. Park and N. R. Kitteringham, Drug Metab. Rev., 1994, 26, 605.
72. K. Harada, J. Matulic-Adamic, R. W. Price, R. F. Schinazi, K. Watanabe
and J. J. Fox, J. Med. Chem., 1987, 30, 226.
73. H. G. Howell, P. R. Brodfuehrer, S. P. Brundidge, D. A. Benigni and
C. Sapino, J. Org. Chem., 1988, 53, 85.
74. W. M. Watkins, D. G. Sixsmith, H. C. Spencer, D. A. Boriga, D. M.
Kiriuki, T. Kipingor and D. K. Koech, Lancet, 1984, 1(8373), 357.
75. K. A. Neftel, W. Woodtly and M. Schmid, Br. Med. J., 1986, 292, 721.
76. A. C. Harrison, N. R. Kitteringham, J. B. Clarke and B. K. Park, Biochem.
Pharmacol., 1992, 43, 1421.
77. P. M. O’Neill, A. C. Harrison, R. C. Storr, S. R. Hawley, S. A. Ward and
B. K. Park, J. Med. Chem., 1994, 37, 1362.
CHAPTER 11

Future Targets and

Chemistry and ADME Needs
DENNIS A. SMITH AND DAVID S. MILLAN

Pharmacokinetics, Dynamics and Metabolism and Medicinal Chemistry,

Pfizer Global R & D, Sandwich, Kent, CT13 9NJ, UK

11.1 The Human Genome

Modern molecular biology techniques have enabled cloning and sequencing
which have revolutionised approaches to identify new targets through whole
genome sequencing, with the human genome being the most notable example.
The human genome project1 (HGP) was started in 1990 with a $3 billion funding
from the US National Institutes of Health (NIH) and the US Department of
Energy. It was expected to take 15 years to sequence the entire human genome,
which is 3.2 gigabases in length and 25 times bigger than any previously mapped
genome, and therefore a significant undertaking. The project actually became a
consortium with 20 sequencing centres from several different countries including
the USA, China, France, Japan, Germany and the UK. An initial incomplete
draft2,3 of the genome was announced in 2001, which was generated by a
hierarchical mapping and sequencing strategy. The completed genome1 was
published in April 2004.
Alongside the NIH-backed HGP, another sequencing campaign was run in
parallel by the American researcher J. Craig Venter through the company Cel-
era. This effort started in 1998 and was expected to cost a fraction ($300 million)
of the NIH effort due to the use of a faster more high risk technique known as

RSC Drug Discovery Series No. 1

Metabolism, Pharmacokinetics and Toxicity of Functional Groups: Impact of Chemical Building
Blocks on ADMET
Edited by Dennis A. Smith
r Royal Society of Chemistry 2010
Published by the Royal Society of Chemistry, www.rsc.org

486
Future Targets and Chemistry and ADME Needs 487
the ‘whole-genome random shotgun’ approach. The aim of this effort was to
sequence the human genome in three years and make the information available
as a commercial product. In 2000, Celera scientists publicly announced their first
draft and, in 2001, it was published.4
At this stage, B83% of the human genome had been sequenced, but it was
not until 2003 and 2005 that a fuller picture emerged, with 499% of genome
sequenced. Estimates at the start of both approaches suggested that the genome
may contain up to two million genes, but on completion, the general consensus
appears to be much less at approximately 30000.
This landmark achievement in science created a huge opportunity for the
pharmaceutical industry, academia and governmental agencies to convert this
wealth of information into drugs to treat disease.

11.2 Drug Targets within the Genome

This chapter confines discussion to the discovery and development of small
molecule drugs derived from the human genome and does not cover other
approaches such as biologics, vaccines or non-human targets. This chapter
describes drug-like properties as absorption, distribution, metabolism and
excretion (ADME) space which concentrates on lipophilicity and hydrogen
bonding rather than molecular weight (MW).
With estimates of the human genome comprising 30 000 genes, two important
questions emerge for drug discoverers. Firstly, how many of these are disease-
related genes and secondly how many of the 30 000+ proteins derived from
these genes are druggable with small molecules. Of course, it is the intersection
of these two datasets where the hunt for small molecule drugs should focus.
To address the first question, Drews5,6 estimated the number of disease-
related genes of interest at 5000–10 000. This was based on the assumptions
that:

 there are 100–150 multifactorial diseases in the developed world;

 each disease is caused by ten genes;
 each of these genes is ‘linked’ to 5–10 other genes which offer an oppor-
tunity for therapeutic intervention.

The second question then relates to how many of the 30 000 gene products
may be seen as ‘druggable’. Hopkins and Groom took the gene family approach
and focussed on proteins as the most significant druggable biomolecule com-
pared with the other major classes of biomolecules—DNA, polysaccharides and
lipids.7 They first examined the chemical substrate used to target current gene
families and assumed that the binding site architecture was conserved across a
gene family. Future chemical substrates were then assumed to also be similar.
An assessment of the druggability8 of the small molecule substrate within each
gene family was then defined using the ‘rule of five’ (Ro5). It emerged from this
analysis that, of the 30 000 genes in the human genome, approximately 3000
488 Chapter 11
genes encode proteins with binding sites suitable for binding molecules which
are likely to have acceptable ADME properties.
The estimate of 5000–10 000 disease-related genes and 3000 genes which
encode proteins that are druggable prompts the question ‘how many disease
targets reside at the intersection’. In reality, the answer is that we do not have
the tools to give a precise estimate. An estimate by Hopkins and Groom was
put at 600–1500, and was based on extrapolation from the number of anti-
fungal targets in the yeast genome.7
It is this intersection of druggable proteins with disease-modifying proteins
that the majority of drugs in the current pharmacopeia occupy. An analysis by
Drews in 20006 showed that the modern pharmacopoeia modulates 483 targets.
A different analysis by Hopkins and Groom7 in 2002 put the number of pro-
teins that the pharmacopeia modulates at only 120 drugs with properties
commensurate for acceptable pharmacokinetic properties. These estimates
suggest that as researchers we have only scratched the surface of druggable and
disease-modifying proteins which could lead to important new drugs.

11.3 The Genome Gap

What are universities, government institutes and industry currently doing to
address this gap? In 2006, Zheng and co-workers9 analysed the distribution of
therapeutic targets against disease classes. What was evident from this analysis
is that there is a significant increase in the level of exploration within each
disease class, suggesting that the sequenced human genome is impacting the
number of new targets that are being explored (Figure 11.1.). According to the
therapeutics target database, there were around 500 research targets being
pursued in 1996, with 1535 targets in 2006.
The analysis by Hopkins and Groom7 in 2002 not only gives estimates of the
number of druggable protein targets but also provides a breakdown of the
human genome in terms of gene family (Figure 11.2). This analysis suggests that
G-protein coupled receptors (GPCRs) and protein kinases are two of the largest

600
Successful
Number of targets

500
Research
400
Total
300
200

100
0
Muscloskeletal

Nutritional and
subcutaneous

Infectious and
blood forming

Genitourinary

abnormalities
Inflammation
Respiratory
system and

Neoplasms
Circulatory

Congenital
Blood and

Endocrine

Injury and
poisoning
metabolic
Ill defined
disorders

disorders

disorders
Digestive

Immunity
Skin and

parasitic
systems

Nervous

dieases
system

system

system

system

Mental

Figure 11.1 Therapeutic targets and disease: successful marketed drugs, drugs in
development and total.
Future Targets and Chemistry and ADME Needs 489

GCPRs
26%

Ther 119
genefamilies and
singleton targets Kinases
46% 10%

Zn peptidases
4%
CYPs Nuclear hormone Rs
2% 3%
Gated ion channels Ser proteases
Cys proteases 3%
2%
Cation channels 2%
2%
A
Kinases
22%
Other 114
genefamilies and
singleton targets
40%
GPCRs
15%

Cation channels
Dehydrogenases/ 5%
reductases
Nuclear hormone Rs
2%
2%

Carboxylases Zn peptidases
2% Protein phosphatases 2%
CYPs Ser proteases
4%
2% 4%
B

Figure 11.2 Drug target families shown as A—the molecular targets of experimental
and marketed drugs with physicochemical properties compliant with
druggable and B—the druggable genome.

druggable gene families within the human genome, although smaller than pre-
viously expected. GPCRs have previously been rich hunting grounds for drug
discovery9,10 and protein kinases have begun to yield some success stories.
Alongside these important classes, the total of zinc metallopeptidases, serine
proteases, phosphodiesterases and nuclear hormone receptors make up nearly
half of the gene families within the druggable human genome. In contrast to
what was initially expected, proteases in particular have emerged as a significant
gene family.
490 Chapter 11
A conservative estimate by Russ and Lampel made in 2005 puts the number of
druggable proteases at B230 proteins.11 Together with kinases, GPCRs and
other families, the number of druggable targets reached in this analysis is just
over 3000, which is remarkably similar to that estimated by Hopkins and
Groom.7

11.3.1 New Mechanisms and the Druggable Genome

There are other approaches which have not been extensively investigated such
as epigenetics12 and inhibition of protein–protein interactions,13,14 for which
the small molecule substrate required is only just beginning to emerge.
Protein trafficking15–19 is another potential mechanism for which small
molecule intervention might be appropriate. Evidence is emerging that the
expression of cell surface proteins can be influenced by small molecules, both in
terms of increased expression and decreased expression.
A number of diseases are caused by genetic polymorphisms which code for
proteins that misfold and do not reach the cell surface. An example of this is long
QT syndrome, which is associated with low cell surface expression of the HERG
channel. Antagonists or blockers of the HERG channel can decrease the
expression of the wild-type HERG channel, but also can raise the expression of the
mutated form. One hypothesis suggests that the compounds bind to the same or
similar residues (in the active site) in the nascent folding channel (at the endo-
plasmic reticulum) as they do when the channel is expressed at the cell surface.
These growing insights suggest that compounds closely related to existing
agonists and antagonists may be active in the control of cell surface expression.
Critical here for indications such as long QT is that the structure–activity
relationships (SARs) for antagonism and aiding protein folding are separate.
Importantly, the chemical space of such molecules is unlikely to be sub-
stantially different from existing agents which have good physicochemical and
resultant ADME properties.
One example of such an approach is a treatment for cystic fibrosis. Cystic
fibrosis (CF) is a fatal genetic disease caused by mutations in cftr, a gene
encoding a protein kinase A (PKA) regulated chloride channel. The most
common mutation results in a deletion of phenylalanine at position 508
(DF508-CFTR); this impairs protein folding, trafficking and channel gating in
epithelial cells. In the airway, these defects alter salt and fluid transport, leading
to chronic infection, inflammation and loss of lung function.
Two classes of novel, potent small molecules19 have been identified from
screening compound libraries which restore the function of DF508-CFTR in
both recombinant cells and cultures of human bronchial epithelia isolated from
CF patients. The first class partially corrects the trafficking defect by facilitating
exit from the endoplasmic reticulum and restores DF508-CFTR-mediated
chloride transport to more than 10% of that observed in non-CF human
bronchial epithelial cultures; this level is one expected to result in a clinical
Future Targets and Chemistry and ADME Needs 491
benefit in CF patients. The second class of compounds potentiates cyclic ade-
nosine monophosphate (cAMP) mediated gating of DF508-CFTR and
achieves single-channel activity similar to wild-type CFTR.
The CFTR-activating effects of the two mechanisms are additive and support
the rationale of a drug discovery strategy based on rescue of the basic genetic
defect responsible for CF.

11.4 The Need for New ADME tools

Analysis of the human genome suggests that no large gene families remain
undiscovered and that therefore future drug discovery efforts will reside largely
within gene families with which we are already familiar and with which the
industry as a whole has some experience in discovering and developing new
drugs.
Looking at the current pharmacopoeia and breaking that down into gene
families gives an idea of the properties of the small molecule chemical matter that
is required to modulate these targets. For example, in the area of kinases, several
drugs have made it to market since the turn of the century,20 with imatinib being
the first in 2001. Many of these drugs have similar physicochemical properties,
which although not unique to kinase inhibitors, are indicative of what future
chemical substrates will look like if adenosine triphosphate (ATP) binding site
inhibitors are targeted. Most of these drugs are lipophilic, with many containing
weakly basic centres (Figure 11.3). Therefore future campaigns targeting ATP-
binding site inhibitors will likely be in similar physicochemical space.
Similar trends emerge if we look more widely at other gene families that will
probably be targets of the future. The work of Paolini et al.21 was instrumental
in observing these gene family trends with respect to their chemical matter. The
shift from an average molecular weight of the aminergic GPCR modulators of
376 Daltons (Da) to other gene families such as peptide GPCR modulators
(average MW 477 Da), metalloproteases (average MW 429 Da) and serine
proteases (average MW 463 Da) has meant that, over the last two decades, the
average molecular weight of approved drugs has increased by 420%. This is of
course at odds with the Ro5 and pushes leads within these gene families closer
to the edge of currently perceived successful oral drug space. But is molecular
weight a key determinant of oral absorption? The relevance of molecular
weight with oral absorption is a topic discussed later in this chapter.
Looking at other physicochemical properties of chemical matter related to
future and current gene family targets reveals more worrying trends. Much of
the chemical matter required to modulate gene families of the future also
contain more hydrogen bond donors (HBDs) and hydrogen bond acceptors
(HBAs), including many of the targets within the protease gene family. Simi-
larly, the clogP value of many of these compounds is also moving closer to 5.
These upward trends in physicochemical properties considered important for
oral bioavailability suggest that medicinal chemists will be working closer to the
edge of oral drug space more often. Therefore, there is a growing need to
 N
N O O
492

N
O N O
O N
N O O
O
N N N N
N
N F
Cl N N
Imatinib (Novartis - 2001) Gefitinib (AstraZeneca - 2003) Erlotinib (OSI, Roche - 2004)
MW 494 MW 447 MW 393
clogP 4.5 clogP 5.6 clogP 4.3
TPSA 86Å2 TPSA 69Å2 TPSA 75Å2

Cl
O
O
N Cl S
N N O
N N
N
F N
O O N N N
F N O N
N F N

N O F Sunitinib (Pfizer - 2006)

Sorafenib (Bayer - 2005) MW 494
MW 465 Dasatinib (BMS - 2006) clogP 4.5
clogP 5.5 MW 488 TPSA 86Å2
TPSA 92Å2 clogP 2.5 F
TPSA 135Å2 F F

N
N Cl N N
N

N O O
O O N
S N
O F N N

Lapatinib (GSK - 2007) N

Nilotinib (Novartis - 2007)
MW 581 MW 530
clogP 6 clogP 5.8
TPSA 115Å2 TPSA 98Å2
Chapter 11

Figure 11.3 Structures of marketed kinase inhibitors (TPSA¼topological polar surface area).
Future Targets and Chemistry and ADME Needs 493

O
O O N
O HN
NH
O
O N
O NH
HN N N N N N
N
OH N

Atazanavir (BMS - 2003) Imatinib (Novartis - 2001)

MW 705 MW 494
clogP 5.9 clogP 4.5
TPSA 171Å2 TPSA 86Å2

Figure 11.4 Structures of atazanavir and imatinib—compounds with pharmacoki-

netics not predicted by calculation of their physicochemical properties.

develop new tools to better understand the properties of molecules within this
challenging physicochemical space and ultimately to increase confidence pre-
clinically that the required ADME profiles will be achieved.
Alongside this, a deeper and more rational understanding of the physical
processes that occur during oral absorption is required. These tools should
include improved in silico methods for prospective, property based design of
orally bioavailable compounds, as well as more relevant in vitro methods for
investigating membrane permeability (both gut and blood–brain barrier
penetration).
Analysis of the HIV protease inhibitor, atazanavir, and the first kinase inhi-
bitor, imatinib, illustrates these points well (Figure 11.4). Both are high mole-
cular weight, relatively lipophilic compounds. Atazanavir moves further out of
desirable ADME space due to a high hydrogen bond donor count and high polar
surface area (PSA), yet displays remarkably good human oral pharmacoki-
netics.22,23 Similarly, imatinib24 is highly orally bioavailable despite its high
molecular weight and lipophilic character. If property-based design principles25
were applied to these molecules, both would be at the high end of predicted risk
for poor oral pharmacokinetics. In reality, both are highly successful oral agents.

11.5 The Chemistry Gap

Accessing desirable chemical space for current and future drug discovery efforts
is an ongoing challenge in medicinal chemistry. The best theoretical drug design
in the world is no use if you cannot synthesise the ‘right’ molecule. In this
section we briefly highlight several areas where synthetic chemistry could
potentially impact future drug discovery efforts. We focus on the synthesis and
commercial availability of aryl building blocks, the importance of fluorination
494 Chapter 11
in drug discovery and finally exemplify areas where improved methods in
asymmetric synthesis are required.
A common feature of small molecule drug-like compounds is the presence of
heteroaromatic rings that act as core systems for further elaboration and
diversification. Many of these cores are linked to other heteroaromatics or
functionalised phenyl rings through biaryl like carbon–carbon bond connec-
tions as well as anilinic and aryl ether like connections. This raises an inter-
esting philosophical question with regard to the chemical structure of ‘drug-like
molecules’. Are aromatic ring systems abundant in drug molecules because they
are required to bind to the target protein, or is this merely a consequence of
synthetic ease?
With respect to the synthesis of heteroaromatic rings a recent report26 sug-
gested, through calculation, that B25 000 small heteroaromatic ring systems are
theoretically possible. However, a search of the literature suggested that only
B7% of these heteroaromatic rings have ever been made and published.
Moreover, this report found that only 5–10 novel heteroaromatic ring systems
are appearing in the literature every year. With heteroaromatic chemical space
continuing to be patented year on year, as well as the ability of heteroaromatics
to offer potentially modified ADME properties, the synthesis of novel systems is
becoming increasingly important in drug discovery. This example perhaps
highlights the fact that, although synthetic chemistry may be widely considered
as a ‘mature’ science, there is still a huge challenge open to the organic chemistry
community.
A good example of how synthetic chemistry can impact in the discovery of a
new drug is the case of the phosphodiesterase-5 inhibitor (PDE5), vardenafil.
The first PDE5 inhibitor to reach the market was sildenafil (ViagraTM), which
was later followed by a structural analogue, vardenafil (Figure 11.5). Vardenafil
differs from sildenafil by the position of a nitrogen atom in the heterocyclic ring
system as well as a one carbon homologation of the piperazine substituent,
leading to increased potency for the PDE5 enzyme and a lower dose. Clearly
the ability to synthesise new heterocyclic ring systems, with different properties,
had an impact on the discovery of this drug.
Allied to bespoke syntheses of polyfunctionalised heteroaromatic ring sys-
tems, accessing commercially available functionalised phenyl and heterocyclic

O
O
N N
O HN N
N N O HN
S N N
N S N
O N
O
O
O

Sildenafil Vardenafil

Figure 11.5 Chemical structures of the PDE5 inhibitors, sildenafil and vardenafil.
Future Targets and Chemistry and ADME Needs 495
building blocks for singleton and library syntheses is critical to the rapid
advance of drug discovery programmes. An analysis of commercially available
tri- and tetra-substituted phenyl rings reveals that these molecular building
blocks are under-represented in supplier catalogues relative to the numbers of
combinations and permutations theoretically possible with such highly sub-
stituted ring systems. A similar scenario exists for tri-substituted pyridine ring
systems (Table 11.1). This is presumably due to synthetic complexity and
expense, but again highlights where synthetic chemistry could add value to
advancing medicinal chemistry programmes.
Moving beyond the paradigm of chemical space exploration, that is largely
driven by commercial availability of suitable building blocks, is a must. To
reiterate a point made above, the design and synthesis of the ‘right’ molecule is
key even if this involves significant investment in synthesis.
The synthetic methodology required to covalently join functionalised phenyl
and heteroaromatic ring systems and derivatise them has advanced significantly
in the last 10–15 years. A comprehensive review of this synthetic methodology
is beyond the scope of this chapter, but it has been reviewed extensively else-
where.27 Many of these chemical reactions are now sufficiently robust to be
carried out by parallel synthesis protocols. Moreover, advances in catalysts
have enabled a greater variety in coupling partners which has broadened the
utility of many of the commercially available compounds analysed in Table
11.1. But to reiterate the point made previously, the major limitation in
accessing this chemical space remains with the availability of diverse and highly
substituted compatible building blocks.
With many drug targets favouring lipophilic chemotypes such as the kinase
drugs described in Figure 11.3, P450-mediated oxidative metabolism is often
an inevitable consequence. This can then lead to poor oral bioavailability
and undesirable pharmacokinetic half-lives. Fluorination as a means to block
P450-mediated metabolism, as well as offering the potential to increase
potency, is emerging as a favoured tactic.28,29 Fluorination is also a useful
strategy to modulate the pKa of basic drugs, which is often important for
designing compounds with reduced affinity for the HERG cardiac potassium
channel.
The DPP-4 inhibitor, sitagliptin, is an example of a highly fluorinated drug
that benefits from improved potency and oral bioavailability by judicious use of
its fluorine substituents. Again, fluorination strategies in the NK1 antagonist,
aprepitant led to an improved duration of action, potency and central nervous
system (CNS) penetration (Figure 11.6). The role of fluorine as an isostere is
discussed at length in Chapter 10 section 3.4. With fluorine substitution capable
of imparting such stark pharmacokinetic and pharmacodynamic advantages
for such a small molecular change (often H to F), synthetic methods to
introduce this atom into molecules of interest is therefore high on the wish list
of most medicinal chemists. Advances in synthetic methodology to introduce
fluorine have emerged in recent years, yet fluorine scans still represent an
enormous synthetic challenge and therefore improved methods and/or more
commercially available fluorinated building blocks would have a huge impact.
 496

Table 11.1 Commercially available tri- and tetra-substituted benzenes

Commercially avail- Commercially avail- Commercially avail-
able compounds able compounds able compounds
Tetra-substitutiona (MWr200 Da) Tri-substitutiona (MWr180 Da) Pyridinesa (MWr180 Da)

A 1006 H 4967 A 273

A A A A H H

A H A H A N A
H H
A 1728 A 1574 A 278
A H A H H A

A A A H A N H
H H
H 897 H 595 A 305
A A A A A H

A A H H A N H
H A
a
A ¼ non-hydrogen.
Chapter 11
Future Targets and Chemistry and ADME Needs 497

F F O
F
F O N
F NH
F N N
N F O N
N N H

NH 2 O
F
F F
F Sitagliptin F Aprepitant
F

Figure 11.6 Fluorinated drugs, sitagliptin and aprepitant.

F
OH O
N
N O
O O
N N N
F N
H
NH2 Oseltamivir
Voriconazole
F

Figure 11.7 Voriconazole and oseltamivir.

As the endogenous substrates of many proteins are chiral (e.g. peptides), it is

important for medicinal chemists to access three-dimensional space through
introduction of chirality into molecules. Moreover, introduction of chirality
can often lead to very different pharmacological selectivity and/or pharmaco-
kinetic profiles between stereoisomers. This tactic of course often costs little or
results in no increase in the molecular weight of the molecule. A recent report
has also suggested that compound complexity, as measured by the presence of
chiral centres is directly related to the likelihood of clinical progression of drug
candidates.30 A comprehensive review of the literature concerning asymmetric
synthesis is also beyond the scope of this chapter but the reader is directed
towards several excellent reviews.31,32
Advances in asymmetric synthesis have continued to emerge in recent
years, though often in the context of natural product synthesis. Application to
‘drug-like’ compounds often necessitates different synthetic strategies and
therefore huge challenges still exist for synthetic chemistry in the drug discovery
arena. A noteworthy example of this point is the antifungal drug, voriconazole
(Figure 11.7). This drug contains two chiral centres, with the active enantiomer
being 500 times more potent than the opposite antipode. Current synthetic
methods rely on a chiral resolution towards the end of the synthetic route,
leading to an inefficient and costly process. Clearly an improved method for
installing this chirality is required.
Another illustrative example is the currently important antiviral drug, osel-
tamivir (Figure 11.7). The biological target, the neuraminidase enzyme, is
498 Chapter 11
involved in the hydrolysis of silaic acid (a carbohydrate derived molecule) from
newly formed virus particles. Therefore to inhibit this enzyme, a carbohydrate
like molecule with three contiguous asymmetric centres is required. This
complex molecular structure led to an extremely challenging and costly
synthesis by its makers, Hoffmann La Roche. In fact, the naturally sourced
starting material (shikimic acid) required for the synthesis of oseltamivir con-
tains all three asymmetric centres, with current synthetic methods unable to
produce this molecule in a reliable way. This example and the others high-
lighted above again suggest that the seemingly mature science of synthetic
organic chemistry still does not have all the answers and that significant
challenges exist.

11.6 The Knowledge Gap in Drug Design

In the preceding sections we have referred to druggability, Ro5 and ADME
space as descriptors of a set of physicochemical properties that define good
delivery of a small drug molecule from an oral dose form to its target. We have
also given examples where the compounds behave very differently with what we
would have expected based on those properties. In the following section we
review some of the current concepts and also identify areas where our knowl-
edge and scholarship could be improved.

11.7 Permeability of Membranes: A Pivotal

Role in Drug Disposition
Permeability of membranes is obviously central to the absorption and dis-
tribution of drugs. Drugs can cross the bilayer of the membrane of cells or
traffic through organs or tissues by the tight junctions (aqueous channels) that
are formed at the junctions between cells. The tight junctions allow limited
permeability to small drugs through the gastrointestinal tract and also into
interstitial water through the capillary walls. This aqueous route has been
exploited with aminergic GPCR agonists and antagonists with drugs such as
ranitidine, cimetidine, atenolol and sumatriptan. Passage out of the capillaries
of the brain is much more restricted. The blood–brain barrier is formed by
‘tighter’ tight junctions between the brain microvessel endothelial cells than
comparative junctions elsewhere. In addition, anthracitic end-feet ensheath the
vessel wall and maintain the barrier properties of the brain capillary endothelial
cells. Because of the tight junctions, the aqueous pore pathway is of very limited
availability and only lipophilic compounds can cross by passive diffusion. Most
drugs are absorbed and reach their intended target by crossing the lipid bilayer
of the gastrointestinal tract. When a target is intracellular, the lipid bilayer
generally has to be crossed to access the target.
Transfer through the lipid bilayer necessitates the removal of the molecule
from its solution in water. Drugs are a hybrid of lipophilic and hydrophilic
functionality. Lipophilic functionality has a positive solvation free energy
Future Targets and Chemistry and ADME Needs 499

Dissolution (D) and Absorption (A)

Glomerular filtration

Plasma

systems
Carrier
Carrier
systems

A Lipoidal
D diffusion
Metabolism

Liver Kidney
Gut
Carrier
systems

Bile Faeces

Figure 11.8 Simplified schematic for the processes of drug absorption and clearance.
The figure illustrates the role of passive and active transport, and enzy-
matic processes.

which favours the move from an aqueous environment to the lipophilic

environment of the core of the membrane bilayer. Conversely, the desolvation
of hydrophilic functionality is energetically unfavourable and hinders transfer
of the drug through the lipophilic core of the membrane bilayer.
It is now realised that not only absorption but clearance of drug are governed
by drug permeability. Figure 11.8 provides a simplified schematic for the
processes of drug absorption and clearance. Passive processes include lipoidal
diffusion, aqueous channel diffusion and glomerular filtration. Active transport
processes—P-glycoprotein (P-gp) and multidrug resistance proteins, organic
anion and organic cation transporting polypeptides, etc.—are involved
potentially both ways in the influx and efflux of drugs across the intestine, the
secretion and reabsorption of drugs in the kidney, and the uptake and traf-
ficking of the drug into the bile or back into the circulation. Enzymatic pro-
cesses occur with metabolism principally in the liver and, with some drugs, the
gut. In addition, transport proteins are expressed at brain capillary endothelial
cells and astrocytic end-feet and efflux drugs further creating the ‘blood–brain
barrier’. Not all these processes occur with many drugs. Most transport pro-
cesses and drug-metabolising enzymes have binding sites that depend largely on
the removal of the groups at the binding interface from water. The loss of
solvent is a major contributor to the binding free energy. Lipophilic groups
have a positive solvation free energy; this together with favourable direct van
der Waals interactions and their incorporation makes the drug a more likely
substrate for transporters or drug-metabolising enzymes.
Many existing drug classes have their disposition mainly characterised by
passive diffusion. The small hydrophilic aminergic GPCR agonists (low per-
meability) and antagonists referred to above have low protein binding and are
500 Chapter 11
cleared renally at a rate close to glomerular filtration rate (GFR). More lipo-
philic (high permeability) aminergic GPCR agonists and antagonists may also
undergo some degrees of filtration, but any drug excreted this way is rapidly
reabsorbed as the urine is concentrated on its passage through the kidney
tubules. These drugs are cleared by metabolism. Permeability across mem-
branes is high—the rate equal to or exceeding the rate that drug transporters
can transport drugs across membranes. Such a high rate of diffusion renders the
influence of transporters as negligible, even if the drug is a substrate. These
definitions are not absolute. As drugs increase in lipophilicity, but not sufficient
for substantial membrane bilayer permeability, they can show evidence of
affinity for active transport systems in the liver and the kidney.
Many of the post-genomic drug classes do not fall easily into a hydrophilic or
lipophilic/low permeability or high permeability description. The drugs are
often of higher molecular weight and contain large amounts of both hydro-
philic and lipophilic functionality. Because of the large areas of hydrophilicity
(PSA), these drugs are of low or at best moderate permeability and their pas-
sage across membranes can be attenuated or aided by active transport systems.
Their absorption across the intestine and penetration into the brain is adversely
affected by efflux transporters (P-gp), whilst their passage into the liver may be
facilitated by further active transport. This step may be the actual systemic
clearance mechanism or it could be a combination of this, metabolism and even
efflux into the bile.
The differentiation of drug disposition has been rationalised as ADME space
(which is comparable to target space)—a physicochemical space that encloses
the properties most likely to be associated with permeable drugs. The dimen-
sions of such a space and its boundaries are illustrated as Figure 11.9. The
formula ‘Log P¼MW – PSA’ is conceptual but links the hydrophilic (PSA) and
lipophilic (log P) functionality referred to above with the size of the molecule.

Log P = MW - PSA

500

ADME
space
MW

5
Lipophilicity

140
PSA

Figure 11.9 ADME space bounded by the interconnected physicochemical properties

of molecular weight, polar surface area and lipophilicity. Drugs with
desirable pharmacokinetic properties such as absorption are much more
likely to occupy the space. The cut-off values of PSA¼140 Å2and log P–5
have been determined from the literature.
Future Targets and Chemistry and ADME Needs 501
PSA can be viewed as the energy cost in desolvation for the molecule to enter
the membrane; so, as higher values of PSA are reached, membrane permeability
is energetically unfavourable. Lipophilicity aids drug permeability but
decreases drug solubility to the point where dissolution of the drug is too low to
allow any absorption The interconnection with molecular weight has spawned
a belief in that particular property being important per se. Most likely it is the
fact that drug molecules are made from carbon (lipophilicity), oxygen and
nitrogen (PSA) and that as, molecular weight approaches 500 Da, the chances
of too high a lipophilicity or too great a PSA increase.
An elegant study33 examined P-gp substrates with the MDRI–MDCKII cell
monolayer as a measure of permeability. The ratio of B to A transport over A
to B transport increased initially with intrinsic absorptive permeability, but
later reduced with compounds with high intrinsic absorptive permeability. This
is explained initially by the substrate needing to penetrate the membrane to be a
substrate of P-gp in this study would be penetration and permeability are such
that the rapid bi-directional flux negates any transport. Most of the drugs
shown to be significantly influenced by P-gp were found in the top right-hand
corner of ADME space (Figure 11.9), with molecular weight around or
exceeding 500 Da and PSAs 475 Å2.
To illustrate this region of ADME space, several aminergic GPCR receptor
antagonist and agonist drugs, and their physicochemical properties. are listed
in Table 11.2.
Table 11.2 illustrates how all the compounds sit well within the boundaries
of ADME space and largely divide into high permeability (mainly metabolic
clearance) and low permeability (renal clearance) purely on the measurement
of lipophilicity (log D). Fraction absorbed is a useful index of permeability.
Sumatriptan is unusual in that the drug is metabolised by monoamine
oxidase, an enzyme that has natural hydrophilic substrates. Ranitidine and
cimetidine show evidence for active tubular secretion with renal clearance
values above GFR.
In contrast, a series of drugs against the HIV protease are shown in
Table 11.3. All of the molecules are on the boundary or outside ADME space

Table 11.2 Physicochemical properties of agonists and antagonists of ami-

nergic GCPRs.
Renal Hepatic
clearance clearance
PSA Log Log Fraction (ml/min/ (ml/min/
MW (Å2) P D absorbed kg) kg)

Atenolol 266 85 0.1 2.0 0.56 2.0 0

Nadolol 309 82 1.3 0.83 0.34 2.6 1.0
Betaxolol 307 51 2.7 0.7 1.00 1.5 8.9
Propranolol 259 42 3.1 1.0 1.00 o0.5 122
Sumatriptan 295 74 0.7 1.73 0.3 2.7 16.3 *
Cimetidine 252 114 0.1 0.45 0.8 6.3 3.7
Ranitidine 314 112 1.2 0.1 0.52 8.5 3.7
502 Chapter 11
Table 11.3 Physicochemical properties of HIV protease inhibitors contrasted
with propranolol.
PSA Log Log A-B B-A Fraction
MW (Å2) P D (nm/s) (nm/s) absorbed

Ritonavir 721 202 5.3 5.3 16 852 0.7

Saquinavir 671 167 4.4 3.9 2 395 0.4
Lopinavir 629 120 6.3 6.3
Indinavir 614 118 2.9 2.9 0.6
Nelfinavir 568 127 7.0 6.0 16 852 0.8
Amprenavir 506 140 4.2 4.2 5 30 0.7
Propranolola 259 42 3.1 1.0 450 700 1.0
a
Propranolol is a high permeability drug (see Table 11.2).

O
N N

O O

CH3 HO CH3
HO

O O
N S F N S F
O O

Figure 11.10 Structures of UK-102 333, a combined thromboxane receptor antago-

nist/synthetase inhibitor, and its N-oxide metabolite.

with molecular weight exceeding 500 Da and combining high lipophilicity and
high PSA (118–202 Å2). These drugs have much lower permeabilities with
highly polarised flux due to transporters. Although low flux is evident, rea-
sonable bioavailabilities can be achieved mainly due to the high local con-
centrations of drug present in the gastrointestinal tract after oral
administration and resultant saturation of the transporters.
The properties of the protease inhibitors in Table 11.3 (which are discussed in
detail in Chapter 10) move the drugs outside the boundaries of ADME space
(high PSA and high lipophilicities) and the drugs have all low permeability;
their disposition will be highly influenced by transporters as illustrated by the
flux across Caco-2 cells.
The low permeability also manifests itself in clearance since most transporter
systems work to influx the drug into the hepatocyte or kidney tubule (poor
permeability essentially making it a one-way process) and then efflux it into the
urine or bile. With poor permeability, the process of metabolism may be sec-
ondary to the process of systemic clearance.
Figure 11.10 shows the structure of a combined thromboxane receptor
antagonist/synthetase inhibitor and its major N-oxide metabolite. The physi-
cochemical properties of the molecule are a log P of 4.3 and a PSA of 106 Å2.
Future Targets and Chemistry and ADME Needs 503
The compound is therefore likely to be of lower permeability. As an acid with
lipophilic regions, it is likely also to be a substrate for organic acid transport
protein (OATP) transporters. The biliary excretion and clearance of the com-
pound was monitored in a series of experiments in the isolated perfused rat
liver.34 In the first series of experiments in the untreated rat liver, systemic
plasma clearance was around 15 ml min1 kg1 and the drug was excreted
alongside its N-oxide, which represented the majority portion of the excretion.
In a second series of experiments, clearance remained unchanged following
ketoconazole pretreatment (a CYP450 inhibitor), but the N-oxide metabolite
was absent and only the parent was detected in bile. This compound was
therefore cleared by hepatic uptake (influx) and metabolism was subsequent to
the event (post-clearance metabolism).
Further examples of the interplay between uptake and metabolism with low
permeability compounds are provided by studies with HIV protease inhibitors.
The intrinsic metabolic clearance of saquinavir, nelfinavir and ritonavir35 was
determined in both rat liver microsomes and fresh isolated rat hepatocytes in
suspension. In the absence of metabolism (achieved by pretreating hepatocytes
with a mechanism-based inhibitor of cytochrome P450), the protease inhibitors
were actively and rapidly taken up into hepatocytes; intracellular unbound
drug concentrations were 5- and 12-fold higher than extracellular unbound
concentrations. Comparison of the rate of uptake into hepatocytes with the rate
of metabolism in hepatocytes and microsomes indicates that the former is the
rate-limiting step at low concentrations. These findings explain the rapid
clearance of saquinavir, nelfinavir and ritonavir and also indicate that the rate
of uptake limits the metabolic clearance of these three drugs. Further studies
(using an hepatic cell model, Hep G2 and Xenopus laevis oocytes over-
expressing human OATP-A) strongly indicate that basolaterally located
OATP-A (influx transporter) in human hepatocytes is the influx transporter
and is acting in concert with apically located P-gp and/or MRP2 (efflux
transporters) for the vectorial transport and excretion of saquinavir into bile.36

11.8 Future CNS Targets and ADME Space

Targeting the CNS has been a rewarding area for drug discovery although the
precise nature of the modulations of signalling required, the families of
receptors and transporters direct the medicinal chemist to small molecules of
simple structure—mainly antagonists of aminergic 7TM receptors or inhibitors
of the transporters of the natural agonists of these receptors. In recent years,
there has been a pronounced shift from disease areas such as depression and
schizophrenia to neurodegeneration to try to meet the ‘epidemic’ of such dis-
eases as Alzheimer’s brought on by an ageing population. The targets for these
are ones that, if previously tackled for a particular disease, did not reside in
such protected areas as the CNS.
The presence of transport proteins at the ‘blood–brain barrier’ is of greater
significance than in the gastrointestinal tract. Most drugs when given by the
504 Chapter 11
oral route achieve concentrations in the gastrointestinal tract that saturate the
efflux transporters. The Km value for transporters such as P-gp is usually 10–
100 mM, a concentration that is exceeded by dissolution of the drug in the
nominal 250 ml of gastrointestinal fluid. This saturation leads to a higher flux
across the gastrointestinal tract. In contrast, unbound drug concentrations in
the systemic circulation will rarely achieve concentrations sufficient to saturate
the transporters, with most drugs requiring nanomolar plasma concentrations
to achieve target efficacy. To gain access to the CNS, drugs have to be suffi-
ciently permeable to have a passive diffusion that is not greatly attenuated by
any efflux transport.
This difference largely explains why ADME space for CNS drugs is more
restricted, possibly having an upper limit of PSA around 75 Å2 to ensure suf-
ficient permeability. This has not been a problem for many CNS drugs which
target aminergic GPCRs, monoamine transporters or enzymes such as
acetylcholinesterase.
Table 11.4 lists the physicochemical characteristics and blood–brain barrier
penetration of a number of CNS drugs against these targets. It is becoming a
bigger issue as drug targets for degenerative diseases of the brain are identified,
including proteases, d and b secretases, etc. Table 11.4 includes examples of
some of these drugs and highlights the fact that inhibitors of these targets are
on the edge of ADME space and will have low relative permeability.
One technical problem with CNS penetration is that the existing datasets are
misleading, with many being based on total brain versus total blood con-
centration. This suggests a much larger dynamic scale. Most of the range is
measuring partitioning of the drug into brain lipid rather than the drug that is
available to receptors or enzymes. Ideally, unbound drug in the plasma should
be compared with brain extracellular fluid concentration. Practically, the most
realistic measure is unbound concentration in plasma versus brain cere-
brospinal fluid. Figure 11.11 depicts the data in Table 11.4.

Table 11.4 Physicochemtcical parameters of selected drugs and their perme-

ability as measured by cerebrospinal fluid (CSF) to unbound
plasma concentration ratio.
Concentration
MW PSA (Å2) Log P Log D CSF/Cp free

Erlotinib 393 75 2.4 2.4 0.3

Nelfinavir 568 127 7.0 6.0 0.09
Indinavir 614 118 2.9 2.9 0.015
Nevirapine 266 58 1.8 1.8 1.5
CP-615 003 373 83 1.8 0.44 0.14
CP-141 938 403.5 79 1.11 1.2 0.11
Fluoxetine 309 21 4.1 1.4 1.1
Propranolol 259 42 3.1 1.2 1.3
Caffeine 194 53 0.13 0.13 1.0
Diazepam 284 33 3.0 3.0 1.0
Future Targets and Chemistry and ADME Needs 505

1.6
CSF / Cp (free)

1.2

0.8

0.4

-0.4
-4 -2 0 2 4 6 8
Log D

Figure 11.11 Penetration across the blood–brain barrier for the drugs shown in Table
11.3. The graph plots penetration expressed as CSF to unbound drug in
plasma ratio against lipophilicity (log P) with the size of point indicative
PSA. Note that high molecular weight drugs have penetration
decreased by poor permeability due to polar surface area, despite high
apparent lipophilicities. These trends are illustrative of the concepts
depicted in Figure 11.2.

To access the brain and interact with many of the new drug targets is a
major challenge. Figure 11.12 illustrates the properties of a series of b-secretase
inhibitors. Ab peptide is associated with the damage caused during
Alzheimer’s disease and is formed by processing amyloid precursor protein.
b-secretase (BACE-1) is the first enzyme in the cascade; it is an aspartyl
protease (a target that is already challenging), made even more difficult by
being present inside the CNS. Other aspartyl protease inhibitors such as renin
and HIV protease have traditionally started with transition state isosteres, but
in this approach new templates were screened for with selectivity to the
enzyme.37 Considerable synthetic diversity and large increases in potency were
achieved from this lead. The initial direction was to incorporate a hydro-
xyethylamine transition state isostere in the molecule.38 This dramatically
improves potency but also increases PSA. Despite good cellular activity of
10 nM, brain penetration was negligible. Reduction of PSA39 was attempted by
truncation of the hydroxyethylamine isostere, initially to a primary alcohol and
then to an amine (Figure 11.12). Despite impressive potency, poor pharma-
cokinetics (including limited transfer across the blood–brain barrier) hamper
the series.
The fundamental problem of such targets is the need to maintain consider-
able hydrogen bonding functionality to achieve selectivity and potency. The
resultant high PSA (and consequent permeability issues) attenuates absorption
and access to the target, which may be due to high clearance through efflux.
506 Chapter 11

Figure 11.12 b-Secretase (BACE-1) inhibitors, structure and resultant ADME

properties which exist on the edge or outside ADME space.

11.9 Penetration into the Cell—Intracellular Drug

Targets
In parallel to the above discussion on the blood–brain barrier, any intracellular
target requires passage across a cell membrane and similar constraints apply to
that discussed above for the blood–brain barrier. Depending on the cell system,
efflux and influx transporters may or may not play a significant role.
In attempting to define SAR more directly, many programmes looking at
intracellular targets may use a direct enzyme assay and a cell-based assay. For
instance to target HCV protease, drugs have to cross the liver cell membrane.
Cheng et al.40 used a protease assay to generate intrinsic Ki values and a
replicon assay using a Hu-7 cell line to generate cell-based IC90 values. These
assays often showed a lack of correlation across compounds due to the
variability in permeability encountered. In an examination of permeability
variability, a series of tetrapeptide/pentapeptide HCV NS3 protease inhibitors
were evaluated. The structures included carbamates, esters and acids. The
compounds ranged in molecular weight from 564 to 872 Da. The ratios of Ki :
IC90 ranged from 0.5 to 0.0003. The lowest permeability compounds with the
biggest difference between Ki and IC90 were acids with negative log D7.4 values.
Molecular weight showed no correlation with permeability, while there was a
low correlation with log D7.4 and the Ki : IC90 ratio. However, a parallel
artificial membrane permeability assay (PAMPA) showed a reasonable corre-
lation. PAMPA is a dodecane-impregnated membrane system, dodecane being
a solvent that unlike octanol does not support hydrogen bonding. Although all
Future Targets and Chemistry and ADME Needs 507
2
compounds had a high PSA (Z163 Å ), compounds with the cell-based IC90
closest to the Ki tended to have the lowest PSA and vice versa.

11.10 Permeability and Large Molecules

TC-1 is a BACE inhibitor derivative of the series outlined above that achieves
good membrane permeability (Figure 11.13). Although it has a PSA of 136 Å2,
it achieves good membrane permeability as indicated by its A to B and B to A
CaCo-2 membrane flux and its entry into the CNS. The compound has been
used for proof of concept studies for BACE inhibition and Ab lowering in
rodents and rhesus monkey.41 The permeability may be related to the fact that
its PSA is comprised of substantially fewer hydrogen bond donors than other
protease inhibitors as shown in Table 11.5.
The drug is now cleared almost exclusively by metabolism as described
above, particularly by CYP3A enzymes. The rate of metabolism is very high,
something that plagues many of the larger molecules, probably because of the
flexibility of the active sites of CYP450s and the multiple sites of metabolism
that a large molecule can present. Common features of the drug-metabolising
enzymes include:42

O O
S CH3
H3C N
Cl
N

H3C N N
N
O

O NH2
CH3 CH3

Figure 11.13 TC-1—a BACE inhibitor with good permeability properties. Although
PSA remains high (136 Å2), it is possible that the difference between
hydrogen bond donors and acceptors accounts for its permeability
characteristics.

Table 11.5 Comparison of number of H-bond donors or acceptors for

selected protease inhibitors.
Number of HBAs No. of HBDs PSA MW

Indinavir 9 4 118 614

Saquinavir 11 6 167 671
TC-1 8 2 136 563
508 Chapter 11
 large substrate-binding cavities;
 binding of more than one substrate/effector;
 binding of substrates in alternative orientations and locations;
 rotation of a substrate at the active site;
 substantial substrate-induced conformational changes.

Studies of human CYP3A4 in complex with two well-characterized drugs,

ketoconazole and erythromycin,43 have shown the protein undergoes dramatic
conformational changes upon ligand binding, with an increase in the active site
volume by 480%. Two distinct open conformations of CYP3A4 can be
observed, which ketoconazole and erythromycin induce upon binding. More-
over, two molecules of ketoconazole can be observed in the CYP3A4 active site
and, of particular importance to large molecules, multiple binding modes for
erythromycin.
The example of indinavir41 shows the many orientations a large molecule can
adopt in CYP3A4 and the number of vulnerable sites such a molecule can
contain. Major CYP3A4 metabolites of indinavir include pyridine N-oxidation,
indane hydroxylation, removal of the methylpyridine, and to a lesser extent
para phenyl hydroxylation. These can be carried out as a single oxidation or
multiple consecutive oxidations.
The flexibility of the proteins and their resultant promiscuity has meant that
progress in understanding metabolism from knowledge of the protein structure
has been hampered. The concept that X-ray crystallography of CYPs and their
substrates would lead to better drug design has not been fulfilled. It is likely that,
in future, medicinal chemists will follow the largely empirical process of elim-
inating vulnerable sites of metabolism in compounds wherever possible. We
have described the need for better ways of fluorination to aid in this process. The
use of fluorine (due to its size, electronic characteristics and possible hydrogen
bonding role) is widespread; Figure 11.14 illustrates the use of fluorine to help
stabilise a thrombin inhibitor44 and a bradykinin B antagonist.45

11.11 Conclusion: Beyond PSA and ADME Space

In this chapter we have suggested that most of the future drug targets will come
from known gene families. Some of these have binding sites that dictate the
ligands having properties on the edge of those required for acceptable ADME
properties. Moreover, some of the most attractive targets reside in locations
requiring even greater constraints on physicochemical properties such as the
CNS. One of the problems in the way most physicochemical properties are
calculated is that the fragmental methods used do not allow, in most cases, for
conformational differences and proximity effects. On closer examination of the
drug atazanavir (Figure 11.4), it could be envisaged that this molecule has a
conformation that allows for internalisation of some of its polar groups
through intramolecular hydrogen bonding. The internal hydrogen bonding
network could nullify hydrogen bonding (PSA) particularly when in an aprotic
Future Targets and Chemistry and ADME Needs 509

$
A1 A2 Cl
N O N O

N N N
N N N N
H £ F F H
O O
F
N NH2

O
O
HN CH3
HN CH3

B1
B2
O
S O
O O S
O O
N

Figure 11.14 Stabilisation to oxidation using fluorine (and chlorine, $) and con-
jugation (removal of functionality, d) for A1 and A2, a thrombin
inhibitor, and B1 and B2, a bradykinin receptor antagonist.

lipophilic environment such as a membrane interior. Such behaviour has been

demonstrated for cyclosporin, a drug that significantly changes conformation
moving from solvent into aprotic environments. At present there is a paucity of
methods to predict such behaviour and yet, if available, compounds could be
designed against targets requiring considerable hydrogen bonding interactions
and yet were permeable to membranes.

References
1. International Human Genome Sequencing Consortium, Finishing the
euchromatic sequence of the human genome, Nature, 2004, 431, 931.
2. International Human Genome Sequencing Consortium et al., Initial
sequencing and analysis of the human genome, Nature, 2001, 409, 860.
3. International Human Genome Sequencing Consortium et al., A physical
map of the human genome, Nature, 2001, 409, 934.
4. J. C. Venter, et al., The sequence of the human genome, Science, 2001, 291,
1304.
5. J. Drews, Nat. Biotechnol., 1996, 14, 1516.
6. J. Drews, Science, 2000, 287, 1960.
7. A. L. Hopkins and C. R. Groom, Nat. Rev. Drug. Discov., 2002, 1, 727.
8. C. Lipinski, F. Lombardo, B. Dominy and P. Feeney, P. Adv. Drug Deliv.
Rev., 1997, 23, 3.
9. C. J. Zheng, l. Y. Han, C. W. Yap, Z. L. Ji, Z. W. Cao and Y. Z. Chen,
Pharmacol. Rev., 2006, 58, 259.
510 Chapter 11
10. K. Beaumont, E. Schmid and D. A. Smith, Bioorg. Med. Chem. Lett., 2005,
15, 3658.
11. A. P. Russ and S. Lampel, Drug Discov. Today, 2005, 10, 1607.
12. M. Szyf, Ann. Rev. Pharmacol. Toxicol., 2009, 49, 243.
13. J. A. Gerrardl, C. A. Hutton and M. A. Perugini, Mini Rev. Med. Chem.,
2007, 7, 151.
14. S. Shangary and S. Wang, Annu. Rev. Pharmacol. Toxicol., 2009, 49, 223.
15. G. W. Carlile, R. Renaud, Z. Donglei Zhang, K. A. Teske, Y. Luo, J. W.
Hanrahan and D. Y. Thomas, ChemBioChem., 2007, 8, 1012.
16. H. Takemasa, T. Nagatomo, H. Abe, K. Kawakami, T. Igarashi, T.
Tsurugi, N. Kabashima, M. Tamura, M. Okazaki, B. P. Delisle, C. T.
January and Y. Otsuji, Br. J. Pharmacol., 2008, 153, 439.
17. D. Thomas, J. Kiehn, H. A. Katus and C. A. Karle, Cardiovasc. Res., 2003,
60, 235.
18. J. S. Carew, S. T. Nawrocki, Y. V. Krupnik, K. Dunner, D. J. McConkey,
M. J. Keating and P. Huang, Blood, 2006, 107, 222.
19. F. van Goor, K. S. Straley, D. Cao, J. Gonzalez, S. Hadida, A. Hazlewood,
J. Joubran, T. Knapp, L. R. Makings, M. Miller, T. Neuberger, E. Olson,
V. Panchenko, J. Rader, A. Singh, J. H. Stack, R. Tung, P. D. Grootenhuis
and P. Negulescu, Am. J. Physiol., 2006, 290, L1117.
20. J. Zhang, P. L. Yang and N. S. Gray, Nat. Rev. Cancer, 2009, 9, 28.
21. G. V. Paolini, R. H. B. Shapland, W. P. van Hoorn, J. S Mason and A. L.
Hopkins, Nat. Biotechnol., 2006, 24, 805.
22. P. J. Piliero, Drugs Today, 2004, 40, 901.
23. M. J. Pérez-Elı́as, Expert Opin. Pharmacother., 2007, 8, 689.
24. Drugs@FDA: www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm;
human oral bioavailability value taken from the drug label, available freely
at www.fda.gov.
25. H. van de Waterbeemd, D. A. Smith, K. Beaumont and D. K. Walker, J.
Med. Chem., 2001, 44, 1313.
26. W. R. Pitt, D. M. Parry, B. G. Perry and C. R. Groom, J. Med. Chem.,
2009, 52, 2952.
27. R. Larsen, Curr. Opin. Drug. Discov. Dev., 1999, 2, 651.
28. W. K. Hagmann, J. Med. Chem., 2008, 51, 4359.
29. K. Müller, C. Faeh and F. Diederich, Science, 2007, 317, 1881.
30. F. Lovering, J. Bikker and C. Humblet, J. Med. Chem., 2009, 52, 6752.
31. D. W. C. MacMillan, Nature, 2008, 455, 304.
32. B. M. Trost, PNAS, 2004, 101, 5348–5355.
33. M. V. S. Varma, K. Sateesh and R. Panchagnula, Mol. Pharm., 2005, 2, 12.
34. I. B. Gardner, D. K. Walker, M. S. Lennard, D. A. Smith and G. T.
Tucker, Xenobiotica, 1995, 25, 185.
35. A. J. Parker and J. B Houston, Drug Metab. Dispos., 2008, 36, 1375.
36. Y. Su, X. Zhang and P. J. Sinko, Mol. Pharm., 2004, 1, 49.
37. C. A. Coburn, S. J. Stachel, Y. -M. Li, D. M. Rush, T. G. Steele, E. Chen
Dodson, M. K. Holloway, M. Xu, Q. Huang, M.-T. Lai, J. DiMuzio, M.-
C. Crouthamel, X.-P. Shi, V. Sardana, Z. X. Chen, S. Munshi, L. Kuo, G.
Future Targets and Chemistry and ADME Needs 511
M. Makara, D. A. Annis, P. K. Tadikonda, H. W. Nash, J. P. Vacca and T.
Wang, J. Med. Chem., 2004, 47, 6117.
38. S. J. Stachel, C. A. Coburn, T. G. Steele, K. G. Jones, E. F. Loutzenhiser,
A. R. Grego, H. A. Rajapaske, M. T. Lai, M.-C. Crouthamel, M. Xu, K.
Tugusheva, J. E. Lineberger, B. L. Pietrak, A. S. Espeseth, X.-P. Shi, E.
Chen-Dodson, M. K. Holloway, S. Munshi, A. J. Simon, L. Kuo and J. P.
Vacca, J. Med. Chem., 2004, 47, 6447.
39. M. G. Stanton, S. R. Stauffer, A. R. Grego, M. Steinbeiser, P. Nantermet,
S. Sanmkaranarayanan, E. A. Price, G. Wu, M.-C. Crouthamel, J. Ellis,
M. Y. Lai, A. S. Espeseth, X. P. Shi, L. Jin, D. Colussi, B. Pietrk, Q.
Huang, M. Xu, A. J. Simon, S. L. Gaham, J. P. Vacca and H. Seinick, J.
Med. Chem., 2007, 50, 3431.
40. L. Cheng, L. Nair, T. Liu, F. Li, J. Pichardo, S. Agrawal, R. Chase, X.
Tong, A. S. Uss, S. Bogen, F. G. Njoroge, R. A. Morrison and K. C.
Cheng, Biochem. Pharmacol., 2008, 75, 1186.
41. S. Sankaranarayanan, M. A. Holahan, D. Colussi, M. C. Crouthamel, V.
Devanarayan, J. Ellis, A. Espeseth, A. T. Gates, S. L. Graham, A. R.
Gregro, D. Hazuda, J. Hochman, K. Holloway, L. Jin, J. Kahana, M.-T.
Lai, J. Lineberger, G. McGaughey, K. P. Moore, P. Nantermet, B. Pietrak,
E. A. Price, H. Rajapakse, S. Stauffer, M. A. Steinbeiser, G. Seabrook, H.
G. Selnick, X. P. Shi, M. G. Stanton, J. Swestock, K. Tugusheva, K. X.
Tyler, J. P. Vacca, J. Wong, G. Wu, M. Xu, J. J. Cook and A. J. Simon, J.
Pharm. Exper. Therap., 2009, 328, 131.
42. Q. Ma and A. Y. H. Lu, Curr. Drug Metab., 2008, 9, 374.
43. M. Ekroos and T. Sjoegren, Proc. Natl. Acad. Sci. U.S.A., 2006, 103,
13682.
44. C. S. Burgey, K. A. Robinson, T. A. Lyle, P. E. J. Sanderson, S. D. Lewis,
B. J. Lucas, J. A. Krueger, R. Singh, C. Miller-Stein, R. B. White, B.
Wong, E. A. Lyle, P. D. Williams, C. A. Coburn, B. D. Dorsey, J. C.
Barrow, M. T. Stranieri, M. A. Holahan, G. R. Sitko, J. J. Cook, D. R.
McMasters, C. M. McDonough, W. M. Sanders, A. A. Wallace, F. C.
Clayton, D. Bohn, Y. M. Leonard, T. J. Detwiler, J. J. Lynch, Y. Yan, Z.
Chen, L. Kuo, S. J. Gardell, J. A. Shafer and J. P. Vacca, J. Med. Chem.,
2003, 46, 461.
45. D.-S. Su, J. L. Lim, E. Tinney, B.-L. Wan, K. L. Murphy, D. R. Reiss, C.
M. Harrell, S. S. O’Malley, R. W. Ransom, R. S. L. Chang, D. J. Petti-
bone, J. Yu, C. Tang, T. Prueksaritanont, R. M. Freidinger, M. G. Bock
and N. J. Anthony, J. Med. Chem., 2008, 51, 3946.
Subject Index

A-74273 402 acyl glucuronide metabolites

abacavir 41 inhibition of UGT and transport
bioactivation 478 proteins 129–30
metabolism 470 toxicity 127–9
absorption acylsulfonamides 225–7
aromatic rings 285–9 adenine 329
carboxylic acids 122–5 ADME 61–94
heteroaromatic rings 339–44 alcohols and phenols 464, 466–82
sulfonamides 230–7 fluorine as isostere 480–2
see also individual drugs interactions with drug
absorption, distribution, metabolism transporters 467–8
and excretion see ADME metabolism 468–80
acarbose 442–3 protein binding and tissue
ACE inhibitors 425–7 distribution 466–7
acebutolol aromatic rings 285–308
absorption 179 absorption 285–9
physicochemical properties 288
clearance 296–308
acetaminophen, bioactivation 311,
distribution 290–5
472, 478
carboxylic acids 122–43
acetazolamide 76, 250
distribution and clearance 125–6
blood to plasma ratio 242
hepatobiliary disposition 141–4
physiochemical/pharmacokinetic
characteristics 238 metabolism 126–40
structure 221, 242 oral absorption 122–5
N-acetylation of amines 193–4 drug delivery 65–75
acetylcholine 170, 331 esters and amides 149–50
N-acetyltransferases 90–1 heteroaromatic rings 338–57
aciclovir (Zovirax) 20, 41 absorption 339–44
aclarubicin 172 distribution 344–8
active absorption 69–70 excretion 354–7
acyclovir 351 metabolism 348–54
acyl CoA metabolites, covalent metabolism 84–91
modification of proteins 132–4 new tools 491–3
Subject Index 513

physicochemical principles 62–5 aldehyde oxidase 87–8

tetrazoles 143–6 heteroaromatic rings 351
thiazolidinediones 146–8 aliskiren 379, 407
tissue distribution 75–82 structure, physicochemical properties
toxicology 94 and ADME 402
transition state analogue inhibitors almitrine 172
396–443 alpha-agonists 170
aspartate carbamyl transferase alprazolam 293
analogues 441–2 alprenolol
glycosidase inhibitor analogues brain penetration 292
442–3 lipophilicity 292
inosine 5-monophosphate amides, ADME 149–50
dehydrogenase analogues amidine 172
440–1 amine isosteres 172–3
neuraminidase analogues 427–32 amineptine 153
nucleoside deaminase analogues amines 168–204
435–40 absorption properties 176–80
membrane permeability 177–80
protease analogues 396–427
P-glycoprotein 180
N-ribosyltransferase analogues
solubility 176–7
432–5
antimalarials 171
ADME space, CNS drugs 503–6
biliary clearance 195–6
adrenaline 169, 331
clearance 185–96
adverse drug reactions metabolic 185–94
idiosyncratic 94 non-metabolic 195–6
sulfonamides 251–5 distribution
see also toxicity brain 185
Agenerase see amprenavir and duration of action 181–3
AL1 413 pH dependence 183–4
AL2 413 and tissue affinity 181, 183
AL3 413 interactions
Alberts, Alfred 111 with acetylcholine 170
alcohol dehydrogenase 468–9, 470 with aminergic receptors 169–70
alcohols and phenols 460–82 with ion channels 171
ADME 464, 466–82 with opioid receptors 170–1
clearance 462, 465 metabolism 185–94
fluorine as isostere 480–2 N-acetylation 193–4
interactions with drug deamination 188–90
transporters 467–8 N-glucuronidation 192–3
metabolism 468–80 N-methylation 190–2
protein binding and tissue N-oxidation and N-dealkylation
distribution 466–7 186–8
volume of distribution 462, 464, 465 N-sulfation 193
comparative pharmacokinetics 462–4 physicochemical properties 173–6
half-life 465 basicity 174–6
physicochemical properties 460–2 pKa 175
prevalence as drugs 460–2 polarity 173–4
514 Subject Index

prodrugs 200–4 anti-inflammatory agents

amine functionality 203–4 NSAIDs see NSAIDs
enhanced absorption 201–2 sulfonamide-based 214–19
tissue specificity 202–3 N-alkylsulfonanilide COX-2
protein binding 184–5 inhibitors 217–19
renal clearance 195 diarylheterocycle COX-2 inhibitors
toxicity 196–9 215–17
as toxicophores 196–9 inhibition of COX isozymes 214–15
zwitterions 199–200 prostaglandin biosynthesis 214
amino acid conjugation 130–4 anticancer agents, peptidomimetics
acyl CoA metabolites in covalent 379–81
modification of proteins 132–4 antidiabetics 36–7, 38
mechanism of 132, 133 antimalarials 171
2-aminoflurane 199 prodrugs 201
amiodarone 308 antimetabolites 330, 332
amitriptyline 169, 174 aplaviroc 172
metabolism 191 aprepitant 497
amlodipine structure 304
metabolism 191 Aptivus see tipranavir
structure 276 arbekacin 172
ammonia 173 arecoline, structure 354
amodiaquine 171 argatroban 17
fluorine substitution 481 Aricept (donepezil) 18
metabolic activation 311 aromatase inhibitors 362
amoxicillin 108 aromatic amines, bioactivation 320
amphetamine 198 aromatic rings 275–322
pharmacokinetics 463, 464 ADME 285–308
protein binding 467 absorption 285–9
ampicillin 108, 125 clearance 296–308
amprenavir (Agenerase) 22, 374, 375, distribution 290–5
404 bioactivation to epoxides 310
CSF-plasma ratio 411 drugs containing 276
effect of ritonavir co-administration fluorine substitution 283–5
409 lipophilicity 276, 277, 282
physicochemical properties 400, 502 physicochemical properties 276–85
structure and ADME 400 solubility 282
amylin 385, 386 toxicity 308–22
anastozole 362 arsphenamine (Salvarsan) 6
angiotensin converting enzyme arylesterases 89
(ACE) 425–6 arylsulfonamides 31–7
angiotensin II antagonists 120–2 as antibacterials 31–4
pharmacokinetics 144 as antidiabetics 36–7
physiochemical properties 143 as antileprotics 34–5
transporters in disposition of 146 as diuretics 35–6
aniline 284 aspartate carbamyl transferase
Ansaid (flurbiprofen) 19, 27, 50 analogues 441–2
Subject Index 515
aspartic acid proteases 371–9 benzimidazole 336
asperlicin 32 benzisoxazole 336
aspirin 19, 27, 103–6 benzodiazepines 30–1, 32, 33
metabolism 89 benzofuran 336
physical characteristics 124 benzoisothiazole 336
structure 276 benzothiazole 336
synthesis 105 benzothiophene 336
atazanavir (Reyataz) 23, 374, 375 1,2,3-benzotriazole 336
CSF-plasma ratio 411 benzoxazole 336
effect of ritonavir co-administration beta-agonists 170
409 beta-blockers, brain penetration 292
physicochemical properties and beta-lactam antibiotics
ADME 400 carboxylic acids 107–10
structure 400, 493 penicillins 109
atenolol mechanism of action 109–10
absorption 179 betaxolol
brain penetration 292 absorption 179
lipophilicity 292 physicochemical properties 289, 299, 501
physicochemical properties 288, 298, renal clearance 299
501 bezafibrate 114
renal clearance 298 biliary clearance 91
atorvastatin (Lipitor) 3, 44, 45, 112 amines 195–6
metabolism 303 aromatic rings 300–1, 302
β-oxidation 136 binding sites 25, 26
physiochemical/pharmacokinetic bioactivation
characteristics 126 alcohols and phenols 475–80
structure 276 abacavir 478
ATP 329 acetaminophen 472, 478
atropaldehyde 470 dauricine 479
atropine 170, 178 diclofenac 314, 479
Augmentin 110 17-estradiol 479
avizafone 203 tamoxifen 480
azithromycin 172, 182 aromatic amines 320
aromatic rings 310
bacampicillin 125 dasatinib 318
BAY-41-8543 197 diclofenac 314
Baycol (cerivastatin) 44, 45, 48–9, 112 phenobarbital 310
β-oxidation 136 phensuximide 310
physiochemical/pharmacokinetic phenytoin 310
characteristics 126 tacrine 314
bendrofluazide 35, 36 prevention by fluorine substitution
benoxaprofen 50, 153 316, 318, 481
benzenes sulfonamides 255–63
physicochemical properties 333, 335 intrinsic electrophilicity 257–61
tri- and tetra-substituted 496 nimesulide 256
benzenesulfohydroxamic acid sulfonanilides 255–7
(Piloty’s acid) 244, 245 sulfonylureas 261–3
516 Subject Index

bioavailability 11, 65–75 in drug discovery 115–16

oral 72 drug safety 151–4
bisoprolol 179 drugs containing 100–1
Black Swan events 49 esters and amides 149–50
blood-brain barrier 80, 505 fibrates 113–15
see also CSF-plasma ratio NSAIDs 103–7
boceprevir 415, 417, 420 aspirin 103–5
boronic acid derivatives 150–1, 420–5 pKa values 102
bortezomib 151, 421, 423, 425 prodrugs 124–5
brain penetration statins 110–13
β-blockers 292 tetrazoles 143–6
prevention of 297 thiazolidinediones 146–8
sedatives 293, 294 carbutamide 36, 38, 223, 320
Brazil, pharmaceutical sales 3 structure 278
brinzolamide 221 carvedilol 179
bromfenac 153 carzenide 36
brominidine 351 caspases 414
catechol-O-methyltransferase 475,
bumetanide 35, 36, 250
476, 477
bupivacaine 171
cathepsin-K 415
buprenorphine 171
CCK-B antagonists 31, 32
bupropion 170
physicochemical properties 348
buspirone 463
cefadroxil 69
BW-175 402 cefepime 197
Celebrex see celecoxib
caffeine, CSF-plasma ratio 504 celecoxib (Celebrex) 19, 40, 105, 214, 216
Canada, pharmaceutical sales 3 metabolism 246, 248, 305–6, 307
candesartan celexetil 122, 123 physiochemical properties 232
pharmacokinetics 144 cell permeability
physiochemical properties 143 and functional grouping 286
captopril 426, 427 see also lipophilicity
carbazeran 87–8, 351 central nervous system
carbonic anhydrase inhibitors 35 drug distribution 80–2
sulfonamide-based 219–21 targeting of 503–6
clinical applications 220–1 CEP-18770 423, 424
mode of action 219–20 cerebrospinal fluid see CSF
red cell partitioning 237, 242 cerivastatin (Baycol) 44, 45, 48–9, 112
carboxylesterase 89 β-oxidation 136
carboxylic acids 99–154 physiochemical/pharmacokinetic
ADME 122–43 characteristics 126
distribution and clearance 125–6 cetirizine 200
hepatobiliary disposition 141–4 metabolism 470
metabolism 126–40 CGP-29287 379
oral absorption 122–5 CGP-29827 402
β-lactam antibiotics 107–10 CGP-38560 379
bioisosteres 116–22 structure, physicochemical
boronic acid derivatives 150–1 properties and ADME 403
Subject Index 517
CGP-60536 379 clinical trials 2
CGS-2466 40 clofibrate 114
Chain, Ernst Boris 109 cLogP see lipophilicity
chemical space 9–10 clometacin 153
privileged substructures 44–6 clomiphene 172
chemistry gap 493–8 clopidogrel (Plavix) 3
chemogenomics 7 structure 276
chemokine receptor 5 37, 39 clotrimazole 362
China, pharmaceutical sales 3 cloxacillin 108
chloroprocaine 307 clozapine 197
chloroquine 171, 330, 331 CNS see central nervous system
pharmacokinetics 463, 464 cocaine 331
protein binding 467 compound attrition 47–51
chlorothiazide 35, 36, 250 cotinine 463
structure 278 COX inhibitors 26–8, 39–41
chlorpromazine 169 COX-1 105
physicochemical properties 287 COX-2
structure 287, 304 N-alkylsulfonanilides 217–19
chlorpropamide 38, 223 diarylheterocycles 215–17
physiochemical/pharmacokinetic physicochemical properties 340
characteristics 239 NSAIDs 105–6
chlorthalidone 242 mode of action 106
physiochemical/pharmacokinetic molecular/structural basis 106–7
characteristics 238 COX isozymes 105
cholecystokinin 385 COX-2 105
cholinesterases 88–9 inhibition of 214–15
Cialis (tadalafil) 18, 25–6, 27 coxibs 214
ciglitazone 332 mode of action 214–15
cimetidine 329 origin of 215–17
physicochemical properties 501 CP-99994 304
cinnarizine 306 CP-141938, CSF-plasma ratio 504
cinnoline 333 CP-615003, CSF-plasma ratio 504
ciprofibrate 114 Crestor (rosuvastatin) 44, 48, 112, 330, 331
ciprofloxacin 103 physiochemical/pharmacokinetic
cisalpine 47 characteristics 126
cisapride 47–8, 171 CSF-plasma ratio 411, 504
cladribine 330, 332 CVS-1123 412
clarithromycin 172 cyclooxygenase see COX
clearance 82–4, 85 cyclothiazide 250
alcohols and phenols 462, 465 CYP1A2 471
aromatic rings 296–308 inhibition 364
biliary see biliary clearance CYP2C8, thiazolidinedione
carboxylic acids 125–6 metabolism 148
liver 84, 85 CYP2C9 85
metabolic see metabolism metabolism of alcohols and phenols 471
organ extraction 83–4 metabolism of sulfonamides 247–9
renal see renal clearance metabolism of tetrazoles 145–6
518 Subject Index
CYP2C19 192 dichlorophenamide 221
CYP2D6 85, 188, 192 diclofenac (Voltaren) 19, 27, 104, 153, 313
CYP2E1 471 bioactivation 314, 479
CYP3A4 71–2, 85, 471 metabolism 304–5
inactivation of 317 pharmacokinetic characteristics 126
thiazolidinedione metabolism 148 physical characteristics 124
cystic fibrosis 490–1 physiochemical properties 126, 232
cytochrome P450 84–6, 469, 471 diethylcarbamazine 172
see also CYP diflunisal 124
cytosine 329 dihydrofolic acid 212
dissolution 68
DACA 351 distribution
dantrolene 463 aromatic rings 290–5
dapsone 34, 213 carboxylic acids 125–6
physiochemical/pharmacokinetic heteroaromatic rings 344–8
characteristics 240 sulfonamides 237–42
toxicity 321 diuretics 35–6
darunavir 23, 374, 377 dofetilide 183, 300
CSF-plasma ratio 411 Domagk, Gerhard 33, 212
effect of ritonavir co-administration donepezil (Aricept) 18
409 dopamine 87, 169
structure, physicochemical properties dorzolamide 221
and ADME 400 doxazosin 170
dasatinib 329 drug delivery 65–75
bioactivation 318 buccal 66
structure 492 dermal 66
daunorubicin 463 inhalation 66, 73
dauricine, bioactivation 479 intranasal 66, 73
Dawber, Thomas Royle 111 intravenous 66
N-dealkylation, amines 186–8 oral 66–72
deamination 188–90 rectal 74
debrisoquine 172 subcutaneous/intramuscular 66, 74–5
deoxyadenosine 332 sublingual 73–4
deoxycytidine 332 transdermal 74
desipramine 169 drug design, knowledge gap 498
metabolism 194 drug discovery 2, 115–16
devazepide 32 high throughput screening 29
dextromethorphan 172 parecoxib 231, 236–7
diazepam 32, 188 drug safety 151–4
brain to serum ratio 294 drug transporters see transporters
CSF-plasma ratio 504 drug-drug interaction 129
pharmacokinetics 463 druggability 8–9, 25, 487, 490–1
physicochemical properties 294 drugs 1–53
prodrugs 203 binding to targets 15–28
protein binding 467 launched 6–15
structure 474 chemical space 9–10
diazoxide 278 molecular properties 10–13
Subject Index 519
polypharmacology 13–15 etodolac 104
target space 7–9 excretion see renal clearance
molecular targets 7 exendin-4 386
pharmaceutical sales 3, 4 EXP3174 123
DuP-714 422, 424 ezetimibe 305
dutogliptin 422, 424 structure 306
ezlopitant
efavirenz (Sustiva) 24 pharmacokinetics 463, 464
efegatran 412, 414, 415 protein binding 467
efflux transporters 70–1
Ehrlich, Paul 6 famciclovir, metabolic activation 352
elastase 415 famotidine 170
eletriptan 170 farnesyl transferase 379–81
enalaprilat 426, 427 fatty acids, β-oxidation 134, 135
enalkiren 403, 406 felbamate 470
Enbrel (etanercept) 3 felodipine 171
Endo, Akira 111 fenamic acid 119
endogenous ligands, drug mimicking fenbufen 153
329, 330, 331 fenclofenac 305
enoxacin 172 fenofibrate 114
entacapone 321 fenoprofen 50
metabolic activation 322 fexfofenadine 116
enzyme transition states 390–3 fibrates 113–15
see also transition state analogue first-pass metabolism 71–2
inhibitors gut wall 71–2
eprosartan hepatic 72
physicochemical properties 343 FK-706 417, 419
structure 343 flavin monooxygenase 86, 186
equilenin 313 flecainide 321
equilin 313 Fleming, Alexander 109
ergotamine 170 Florey, Howard 109
erlotinib (Tarceva) 17, 43 flosulide 218
CSF-plasma ratio 504 flovagatran 422
structure 492 fluconazole
erythrocytes, drug distribution into 75–6 physicochemical properties 355
erythromycin 172, 182 renal reabsorption 93
esomeprazole (Nexium) 3 flunarizine 305
esterases 88–9 structure 306
esters, ADME 149–50 flunitrazepam
estradiol 313 brain to serum ratio 294
17-estradiol, bioactivation 479 physicochemical properties 294
estrogens, metabolic activation 313 fluorine substitution
estrone 313 alcohols and phenols 480–2
etanercept (Enbrel) 3 and drug metabolism 306
ethambutol 470 on phenyl rings 283–5
ethoxzolamide 221 prevention of bioactivation 316,
17α-ethynylestradiol 313 318, 481
520 Subject Index

5-fluoro-2-pyrimidone 351 Gleevec see imatinib

fluoroanilines 284 glibenclamide 38, 223
fluorophenols 284 glimiperide 223
5-fluorouracil 329, 332 pharmacokinetics 463
fluoxetine 169 glipizide 38, 239
CSF-plasma ratio 504 glitazones see thiazolidinediones
structure 276 glomerular filtration 92
fluphenazine 304 glucagon-like peptide-1 385, 386
flurazepam 188 glucuronidation
flurbiprofen (Ansaid) 19, 27, 50 amines 192–3
fluticasone/salmeterol (Serentide) 3 carboxylic acids 126–7
fluvastatin 112 heteroaromatic rings 350
β-oxidation 136 tetrazoles 145
physiochemical/pharmacokinetic glyburide 240
characteristics 126 glyceryl trinitrate 73–4
folic acid 332 glycosidase inhibitor analogues 442–3
fomoterol 183 grepafloxacin 171
forodesine 432 GS-4116 429
pharmacokinetics 434 guanidine 172
structure 434 guanine 329
fosamprenavir 374, 377, 404
fosinopril 426, 427 half-life
fosinoprilat 426, 427 alcohols and phenols 465
FPL 55712 228 haloperidol 309
France, pharmaceutical sales 3 halofantrine 171
free drug 78, 79 haloperidol
and receptor occupancy 80 half-life 309
frusemide 36 structure 308
furans Hammett constant 277, 321
physicochemical properties 335 HCV NS3-protease 415
structures 360 Henderson-Hasselbalch equation 178
furosemide 221, 250 HERG channel 490
physiochemical/pharmacokinetic heteroaromatic rings 328–65
characteristics 238 ADME 338–57
structure 360 absorption 339–44
distribution 344–8
G-protein coupled receptors 9, 10–11, excretion 354–7
14, 64, 488–9 metabolism 348–54
ganciclovir 20 physicochemical properties 333–8
gefitinib (Iressa) 17, 25–6, 43, 329 benz-fused five-membered rings 336
structure 492 five-membered rings 335–6
gemcabene 129 six-membered rings 333
gemcitabine 330, 332 sulfa drugs 338
gemfibrozil (Lipur/Lopid) 49, 114 toxicity 357–64
Germany, pharmaceutical sales 3 types of 333–8
ghrelin 385 high throughput screening 29
Subject Index 521

histamine 169, 329 imatinib (Gleevec) 16, 341

HIV protease inhibitors 396–427 physicochemical properties 343
physicochemical properties 502 structure 343, 492, 493
HIV-1 protease 372 imidazole 335
HMG-CoA reductase inhibitors see imidazoline 172
statins imipramine 169
Hodgkin, Dorothy Crowfoot 109 metabolism 191
Hoffman, Felix 105 pharmacokinetics 463, 464
human genome 370, 486–7 protein binding 467
drug targets 487–8 incretin hormones 385–6
genome gap 488–91 indapamide 242
human peptide transporter 1 69 indazole 336
hydralazine 172 indeloxazine 169
metabolism 351 indinavir 21, 176, 177, 374, 375, 407, 507
hydrazine 172 CSF-plasma ratio 411, 504
hydrochlorothiazide 250 effect of ritonavir co-administration 409
metabolism 350
hydroflumethiazide 239
physicochemical properties 400, 502
hydrogen bonds 63
structural modifications 363
carboxylic acids 102
structure and ADME 400
and CNS penetration 81
indole 336
hydrogen bond acceptors 345, 395, 460,
indomethacin 104, 118
507 metabolism 149
hydrogen bond donors 345, 395, 460, 507 phsiochemical and pharmacokinetic
hydrolases 88–9 properties 150
hydroxychloroquine 171 physical characteristics 124
pharmacokinetics 464 physiochemical properties 234
protein binding 467 infliximab (Remicade) 3
hydroxyezlopitant inhaled drug delivery 66, 73
pharmacokinetics 464 inosine 5-monophosphate
protein binding 467 dehydrogenase analogues 440–1
2-hydroxyimipramine insulin 382–5
pharmacokinetics 464 lente 383
protein binding 467 insulin analogs 382–5
3-hydroxyl-3-methyl-glutaryl-coenzyme insulin determir 385
A reductase inhibitors see statins insulin glargine 384
5-hydroxytryptamine 329 insulin glulisine 384
hydroxyzine 200 insulin lispro 384
metabolism 470 intracellular drug targets 506–7
hypoglycemic agents 222–4 intranasal drug delivery 66, 73
ion channels 171
ibufenac 128, 153 4-ipomeanol 360
ibuprofen 19, 27, 49, 104, 128 ipratropium 170, 178
pharmacokinetic characteristics 126 irbesartan 122, 123
physical characteristics 124 pharmacokinetics 144
physiochemical properties 126, 233 physiochemical properties 143
ICI-200,880 416, 420 structure 354
522 Subject Index
Iressa see gefitinib β-blockers 292
isoniazid 172 and drug distribution 75, 76
isothiazole 335 Lipur/Lopid (gemfibrozil) 49, 114
isoxazole 335 liver, drug clearance see metabolism
Italy, pharmaceutical sales 3 log D 12
itraconazole 355 loperamide 171
lopinavir 22, 374
Japan, pharmaceutical sales 3 CSF-plasma ratio 411
effect of ritonavir co-administration
K1 416 409
K2 416 physicochemical properties 401, 502
K3 418 structure and ADME 401
K4 418 loratidine 170
ketamine 171 lorazepam 33
α-ketoamide bioisosteres 224–5 brain to serum ratio 294
ketoconazole 362 protein binding 294
ketoprofen 50, 104
losartan 120–2, 123
physiochemical/pharmacokinetic
pharmacokinetics 144
characteristics 126
physiochemical properties 143, 343
ketorolac
structure 343, 354
physical characteristics 124
lovastatin 45, 112, 331
physiochemical/pharmacokinetic
characteristics 126 β-oxidation 136
kinase inhibitors 41–4
KNI-272 401 maprotiline
Kuroda, Masao 111 pharmacokinetics 463, 464
protein binding 467
L-685434 176, 177 maraviroc 172
L-693612 242 MaxiPost 319, 320
L-733060 304 MB243 197
L-739010 360 meclofenamic acid 118, 119
L-741671 304 medazepam 188
L-745394 361 mefenamic acid 104
lamotrigine 350 physiochemical properties 234
lapatinib 492 mefloquine 171
large molecules 507–8 meglitinide 223
Levitra see vardenafil meloxicam 104
levodopa 477 physiochemical properties 235
lidocaine 187 membrane permeability 498–503
structure 307 amines 177–80
ligand efficiency 9, 65 meptazinol 473
lignocaine 171 metabolism 84–91
limonene 463 N-acetyltransferases 90–1
lipid efficiency 65 alcohols and phenols 468–80
Lipitor see atorvastatin bioactivation 475–80
lipophilicity 62–3, 75 conjugative metabolism 472–5
aromatic ring structures 276, 277, 282 oxidative metabolism 468–72
Subject Index 523

aldehyde oxidase 87–8 midazolam 72

amines 185–94 brain to serum ratio 293
N-acetylation 193–4 mirtazapine (Remeron) 170
deamination 188–90 brain to blood ratio 345
N-glucuronidation 192–3 glucuronidation 193
N-methylation 190–2 physicochemical properties 345
N-oxidation and N-dealkylation mizoribine 440–1
186–8 moclobemide 169
N-sulfation 193 metabolism 187
aromatic rings 301–8 molecular properties of drugs 10–13
carboxylic acids 126–40 molecular weight 10, 11, 63
amino acid conjugation 130–4 monoamine oxidase inhibitors 169
glucuronidation 126–30 monoamine oxidases 87, 188–9
β-oxidation 134, 135 montelukast (Singulair) 3, 227, 228
P450 isoenzymes 140 multidrug resistance-associated
cytochrome P450 84–6 protein 2 141
flavin monooxygenase 86
nadolol
heteroaromatic rings 348–54
absorption 179
hydrolases 88–9
brain penetration 292
monoamine oxidases 87
lipophilicity 292
statins 134, 136–7
physicochemical properties 288, 501
sulfotransferases 90
naloxone 171
tetrazoles 145–6 naphazoline 172
CYP2C9 in 145–6 naproxen 50, 104
N-glucuronidation 145 physical characteristics 124
thiazolidinediones 147–8 nefazodone 169, 313
UDP-glucuronosyltransferases 89–90 metabolic activation 315
valproic acid 137, 139–40 nelfinavir (Viracept) 22, 374, 375, 407
methazolamide 221, 250 CSF-plasma ratio 504
methotrexate 24, 330, 332 effect of ritonavir co-administration
metabolism 351 409
N-methylation of amines 190–1 physicochemical properties 401, 502
methyltransferases 475, 476, 477 structure and ADME 401
metoclopramide 198 neuraminidase analogues 427–32
metoprolol nevirapine (Viramune) 24, 33
absorption 179 CSF-plasma ratio 504
physicochemical properties 299 new chemical entities 1
renal clearance 299 new molecular entities 4
metyrapone 351 Nexium (esomeprazole) 3
mevastatin 44, 45, 112 nicotine 86, 191, 350
mexilitene 187 nifedipine 171
mianserin (Tolvon) nilotinib (Tasigna) 341, 492
brain to blood ratio 345 physicochemical properties 343
glucuronidation 193 structure 343
physicochemical properties 345 nimesulide 218
toxicity 197 bioactivation 256
524 Subject Index

nomifensine 320, 321 1,2,5-oxadiazole 335

non-steroidal anti-inflammatory drugs 1,3,4-oxadiazole 335
see NSAIDs oxaprotiline
noradrenaline 169, 331 pharmacokinetics 464
nordiazepam protein binding 467
pharmacokinetics 463, 464 oxazepam
protein binding 467 pharmacokinetics 464
NSAIDs protein binding 467
Caco-2 permeability 124 oxazole 335
carboxylic acids 103–7, 117–19 oxezapam
aspirin 103–6 brain to serum ratio 294
COX inhibition 105–6 protein binding 294
mode of action 106, 214–15 N-oxidation, amines 186–8
molecular/structural basis 106–7 β-oxidation
dose 124 fatty acids 134, 135
solubility 124 statins 134, 136–7, 138
nucleoside deaminase analogues 435–40 valproic acid 137, 139–40
purine deaminase inhibitors 435–6 oxitropium 170
pyrimidine deaminase inhibitors oxprenolol 179
436–40 oxybutynin 170
oxytetracycline 464
olanzapine (Zyprexa) 3
structure 276 P-glycoprotein 70, 180, 287, 292, 406
olmesartan medoxomil 122, 123 paliperidone, protein binding 467
ONO-6818 417 parecoxib, discovery of 231, 236–7
opioid receptors 170–1 paroxetine 169
oral drug delivery 66–72 penbutolol 179
absorption penciclovir 20, 352
active 69–70 penfluridol 308
gastrointestinal tract 68–9 penicillin G 108
active efflux 70–1 physiochemical/pharmacokinetic
barriers to 67 characteristics 126
bioavailability 72 penicillin V 108
dissolution 68 penicillins 108, 109
first-pass metabolism 71–2 pentostatin 435
organ extraction 83–4 peptide YY 385
organic anion transporting peptides 382–6
polypeptides 141, 408, 503 incretin hormones 385–6
oseltamivir acid 429, 430–1 insulin and analogs 382–5
oseltamivir (Tamiflu) 16, 25, 429, 430, peptidomimetics 370–82
497 anticancer agents 379–81
structure, physicochemical properties aspartic acid proteases 371–9
and ADME 429 definition 371
oxacillin 108 permeability 12
1,2,3-oxadiazole 335 peroxisome proliferator-activated
1,2,4-oxadiazole 335 receptors 114
Subject Index 525

pethidine 171 practolol

pharmaceuticals, sales of 3, 4 absorption 179
pharmacokinetics physicochemical properties 298
alcohols and phenols 462–4 renal clearance 298
angiotensin II antagonists 144 pralnacasan 412
forodesine 434 pramipexole 197
3,4,5,6-tetrahydrouridine 438 pranlukast 227, 228
zebularine 438 pravastatin 44, 45, 112
see also ADME; and individual drugs physiochemical/pharmacokinetic
pharmacological potency 62 characteristics 126
phenelzine 172 prazosin 352
toxicity 197 prilocaine 171
phenobarbital, bioactivation 310 primaquine 171, 330, 331
phenol 284 privileged substructures 28–44
phenol bioisoteres 224 aminoheterocycles 41–4
phensuximide, bioactivation 310 arylsulfonamides and derivatives 31–7
phenyl rings, fluorine substitution 283–5 benzodiazepines 30–1, 32, 33
phenylbutazone 216 and chemical space 44–6
phenylpropanolamine chemokine receptor 5 37–9
pharmacokinetics 464 diaryl heterocycles 39–41
protein binding 467 HMG-CoA reductase inhibitors
phenytoin, bioactivation 310 (statins) 44
pholcodine 172 probenecid 304
phosphodiesterase 5 drugs 25–6 procainamide 198
phosphonoacetyl-L-acetate 441–2 structure 321
physiologically based pharmacokinetic toxicity 197
modelling 78 procaine 331
pinacidil 172 procarbazine 197
pindolol 20 prodrugs 64
absorption 179 amines
pioglitazone 147, 332 amine functionality 203–4
3-piperidinylindole 305, 306 enhanced absorption 201–2
pirenzipine 170 tissue specificity 202–3
piretanide 36 carboxylic acids 124–5
piroxicam 104 sulfonamides 231, 236–7
physiochemical properties 235 profiling 28
pivampicillin 125 promazine
plasma concentration-time curve 77 physicochemical properties 287
Plavix (clopidogrel) 3, 276 structure 287
polar surface area 63, 277, 395 promiscuous targets 14, 15
and blood-brain permeability 346 propanil 197
and P-glycoprotein efflux 296 propoxycaine 307
transition state analogues 393–4, 395 propranolol 170
proteases 405–6 absorption 179
polypharmacology 13–15 brain penetration 292
posaconazole 355 CSF-plasma ratio 504
526 Subject Index

lipophilicity 292 ramiprilat 426

physicochemical properties 289, 501, ranitidine 170
502 physicochemical properties 501
prostaglandin H synthetase 471–2 renal clearance 93
prostaglandins, biosynthesis 214 razaxaban 341
protamine-insulin complex 383 metabolism 353
protease analogues 396–427 physicochemical properties 342
activated ketone containing 414–20 structure 342
alcohol containing 397–414 reboxetine 169
aldehyde containing 412–13, 414 rectal drug delivery 74
boronic acid containing 420–5 relacatib 415, 416
phosphonic acid containing 425–7 Relenza see zanamivir
protein binding 75, 290 Remeron see mirtazapine
alcohols and phenols 466–7 Remicade (infliximab) 3
amines 184–5 remikiren 403, 411
Protein Data Bank 15 remoxipride
protein kinase inhibitors 330, 492 absorption 287
metabolic activation 319
protein kinases 488–90
physicochemical properties 291
protein trafficking 490
renal clearance 92–3
protein tyrosine phosphatase 1B
amines 195
inhibitors 344
aromatic rings 297–300
proteolysis 397
heteroaromatic rings 354–7
proxicromil 308 sulfonamides 249–51
pthalazine 333 renal reabsorption 92–3
purine deaminase inhibitors 435–6 renal secretion 93
pyrazine 333 renin 377, 378
pyrazole 335 research and development, spending
pyridazine 333 on 4–5
pyridine 333 Reyataz see atazanavir
pyrimidine 333 N-ribosyltransferase analogues 432–5
pyrimidine deaminase inhibitors 436–40 rifampicin 172, 182
pyrrole 335 rifamycin 182
rimonabant 40
QT prolongation 47 Risperdal see risperidone
quantitative structure–activity risperidone (Risperdal) 3
relationship 282 protein binding 467
quetiapine (Seroquel) 3 ritonavir 21, 359, 374, 375
quinapril 129 ADME 401
quinaprilat 129 CSF-plasma ratio 411
quinazoline 333 pharmacokinetic boosting of
quinine 171, 330, 331 protease analogues 409–10
metabolism 351 physicochemical properties 401, 502
quinoline 333 structure 359, 401
rizatriptan 329
raloxifene 314 rofecoxib (Vioxx) 40, 49, 105, 214, 216
metabolic activation 315 physiochemical properties 232
Subject Index 527

rosiglitazone 147 Singulair (montelukast) 3, 227, 228

metabolism 350 sitagliptin 356, 495, 497
physicochemical properties 341 Sneader, Walter, Drug Discovery, A
structure 341 History 51
toxicity 364 solubility
rosuvastatin (Crestor) 44, 48, 112, 330, 331 amines 176–7
physiochemical/pharmacokinetic aromatic ring structures 282
characteristics 126 sorafenib 492
routes of administration see drug sotalol
delivery absorption 179
RU-36384 412 physicochemical properties 298
rule of five 9, 10, 63, 230, 487 renal clearance 298
Spain, pharmaceutical sales 3
S-18326 423 statins 44, 330
saccharin 221 carboxylic acids 110–13
salbutamol 170 discovery of 111, 113
salicylic acid 124 hepatobiliary transport 142–3
salmeterol 73, 170, 183 mechanism of action 113
Salvarsan (arsphenamine) 6 β-oxidation 134, 136–7, 138
saquinavir 21, 374, 404, 407, 507 steady state of distribution 78
CSF-plasma ratio 411 structural similarities 28
effect of ritonavir co-administration structure-activity relationships 7, 88, 490
409 subcutaneous/intramuscular drug
physicochemical properties 401, 502 delivery 66, 74–5
structure and ADME 401 sublingual drug delivery 73–4
SB-203580 40 sulbactam 126
SB-271046 81 sulfa drugs 338
SC544 27 sulfacetamide 34
SCH 48461 306 sulfadiazine 34, 213
SDZ-PRI-053 401 pharmacokinetic characteristics 241
β-secretase inhibitors 506, 507 physicochemical properties 241, 338
selective serotonin re-uptake inhibitors sulfadimethoxine 251
(SSRIs) 169 glucuronidation 192
selegiline 169 sulfadimidine 34
Serenading 47 physicochemical properties 338
Serentide (fluticasone/salmeterol) 3 sulfafurazole 213
Seroquel (quetiapine) 3 sulfamethizole 213
serotonin 169 physicochemical properties 338
sertraline 169 sulfamethomidine 247
metabolism 189 sulfamethoxazole 213
structure 276 pharmacokinetic characteristics 241
sialic acid 428 physicochemical properties 241, 338
sildenafil (Viagra) 17, 25–6, 27, 172 sulfametomidine 251
metabolism 243 sulfametoxydiazine 213
structure 494 sulfamidochrosoidine 31, 34
simvastatin 45, 112 sulfamoxzole 338
528 Subject Index
sulfanilamides sulfotransferases 90, 474–5
antibiotics 35, 212–13 sulindac 104
physiochemical/pharmacokinetic physical characteristics 124
characteristics 240–1 sulpiride
sulfaphenazole 213 absorption 287
sulfapyrazine 338 physicochemical properties 291
sulfapyridine 338 sulthiame 221
sulfasalazine 213 sumatriptan 87, 170, 329
sulfasomidine 247 physicochemical properties 501
sulfathiazole 338 sunitinib 492
N-sulfation, amines 193 suprofen 153, 359
sulfisomidine 251 Sustiva (efavirenz) 24
physicochemical properties 338
sulfisoxazole 241 tacrine 331
sulfonamides 210–63 bioactivation 314
adverse drug reactions 251–5 protein binding 467
anti-inflammatory agents 214–19 tadalafil (Cialis) 18, 25–6, 27
N-alkylsulfonanilide COX-2 Taleb, Nassim Nicholas 49
inhibitors 217–19 talinolol 180
diarylheterocycle COX-2 inhibitors absorption 179
215–17 Tamiflu see oseltamivir
inhibition of COX isozymes 214–15 tamoxifen 172, 313–14
prostaglandin biosynthesis 214 bioactivation 480
bioactivation 255–63 metabolic activation 315
intrinsic electrophilicity 257–61 tamsulosin 170
carbonic anhydrase inhibitors 219–21 Tarceva see erlotinib
clinical applications 220–1 target space 7–9
mode of action 219–20 target-hopping 13–15
distribution 237–42 targets
drugs containing 211 binding to 15–28
as α-ketoamide bioisostere 224–5 molecular 7
metabolism 242–9 Tasigna see nilotinib
biotransformation pathways 243–7 telaprevir 417, 419
role of CYP2C9 in 247–9 telmosartan 122, 123
oral absorption 230–7 temazepam
as phenol bioisotere 224 protein binding 467
physicochemical properties 231 structure 474
prodrugs 231, 236–7 temozolomide 339
renal clearance 249–51 terazosin, structure 352
sulfonanilides, bioactivation 255–7 terbinafine 172
sulfonimides 225–7 terfenadine 47–8, 115–16, 170, 171
sulfonylureas 222–4 terodiline 170
bioactivation 261–3 tetracycline 172
mechanism of action 222 pharmacokinetics 463, 464
physiochemical/pharmacokinetic tetrahydrofolic acid 212
characteristics 239–40 3,4,5,6-tetrahydrouridine 436, 437
structure-activity relationships 222–4 pharmacokinetics 438
Subject Index 529
tetrazine 333 tolcapone 321
tetrazoles metabolic activation 322
ADME 143–6 tolmetin 128
angiotensin II antagonists 120–2 Tolvon see mianserin
transporters in disposition of 146 tomelukast 228
metabolism 145–6 topological polar surface area 15
CYP2C9 in 145–6 torsade de pointes 115, 116, 171, 199
N-glucuronidation 145 toxicity
physicochemical properties 336 acyl glucuronide metabolites 127–9
5-substituted-1H 119–20 amines 196–9
1,2,3-thiadiazole 336 aromatic rings 308–22
1,2,4-thiadiazole 335 heteroaromatic rings 357–64
1,2,5-thiadiazole 335 thiazolidinediones 364
1,3,4-thiadiazole 336 see also individual drugs
thiazoles toxicology 94
physicochemical properties 335 transdermal drug delivery 74
structures 359 transition state analogue inhibitors 390–445
thiazolidinediones ADME 396–443
ADME 146–8 aspartate carbamyl transferase
clearance and oral bioavailability 147 analogues 441–2
metabolism 147–8 glycosidase inhibitor analogues 442–3
P450 isozymes 148 inosine 5-monophosphate dehydro-
toxicity 364 genase analogues 440–1
thiophenes neuraminidase analogues 427–32
metabolic activation 359 nucleoside deaminase analogues
physicochemical properties 335 435–40
structures 359 protease analogues 396–427
thymine 329 N-ribosyltransferase analogues 432–5
ticlopidine 172, 359 physicochemical properties 393–6
tienilic acid 153, 359–60 polar surface area 393–4, 395
tifluadom 33 transporters
timolol 331 ADME effects 467–8
absorption 179 alcohols and phenols 467–8
tipranavir (Aptivus) 23, 376 angiotensin II antagonists 146
effect of ritonavir co-administration 409 efflux 70–1
structure, physicochemical human peptide transporter 1 69
properties and ADME 402 inhibition of 129–30
tissue distribution 75–82 organic anion transporting
alcohols and phenols 466–7 polypeptides 141, 408, 503
central nervous system 80–2 tranylcypromine 169
see also volume of distribution traxoprodil 477
tobramycin 172 trazadone 169
tocainide 187 Trefouel, Jacques 212
tolbutamide 36–7, 38, 223 Trefouel, Therese 212
physiochemical/pharmacokinetic Tri-50b 422
characteristics 240 1,2,3-triazine 333
structure 278 1,2,4-triazine 333
530 Subject Index
1,3,5-triazine 333 Viagra see sildenafil
triazolam, brain to serum ratio 293 vicriviroc 172
1,2,3-triazole 335 vinblastine 172
1,2,4-triazole 335 vincristine 172
2H-1,2,4-triazole 335 Vioxx (rofecoxib) 40, 49, 105, 214, 216
4H-1,2,4-triazole 335 physiochemical properties 232
trichloromethiazide 250 Viracept see nelfinavir
triclosan, prodrug 204 Viramune see nevirapine
tricyclic antidepressants 169 Virchow, Rudolph 111
trimethyaminuria 86 Voltaren see diclofenac
troglitazone 147 volume of distribution 76–7, 78
metabolic activation 310, 312 alcohols and phenols 462, 464, 465
physicochemical properties 341 voriconazole 349
structure 341 physicochemical properties 355
toxicity 364 structure 497
trovafloxacin 153
tryptamine 87 X-linked inhibition of apoptosis
tryptase 415 protein 381
xamoterol 179
UDP-glucuronosyltransferases 89–90
UK-102333 502 YM-21095 403
UK-224671 70
UK-343664 71 zafirlukast 227–30
UK-469413 91 zaleplon 339
metabolism 194 metabolism 351
Unasyn 110 zanamivir (Relenza) 16, 429, 430
United Kingdom, pharmaceutical sales 3 structure, physicochemical properties
uracil 329 and ADME 429
uridine 5’-diphospho-glucuronosyl- zankiren 403, 406
transferase 473–4 ZD-0892 415, 416, 419
USA, pharmaceutical sales 3 ZD-8321 415, 416, 419
zebularine 436, 437
valacyclovir 69–70 metabolism 439, 440
valdecoxib 105, 214, 216 pharmacokinetics 438
metabolism 244, 246 ziprasidone 88
physiochemical properties 232 metabolism 353
prodrugs 236 zolmitriptan 190
valproic acid 153 zolpidem 339
β-oxidation 137, 139–40 zomepirac 128, 153, 226
valsartan 122, 123 zoniporide 351
pharmacokinetics 144 zonisamide 221
physiochemical properties 143 metabolism 353
vardenafil (Levitra) 18, 25–6, 27, 172 Zovirax (aciclovir) 20, 41
structure 494 zwitterions
velnacrine, protein binding 467 amines 199–200
venlafaxine 170 carboxylic acids 102, 103
Venter, J. Craig 486 Zyprexa see olanzapine