GENE FAMILY TARGETED MOLECULAR DESIGNdownload.e-bookshelf.de/download/0000/5724/70/L-G... · You should consult with a professional where appropriate. Neither the publisher nor author

GENE FAMILYTARGETEDMOLECULAR DESIGN

Edited by

KAREN E. LACKEY

InnodataFile Attachment9780470423929.jpg


Edited by

KAREN E. LACKEY

Copyright # 2009 by John Wiley & Sons, Inc. All rights reserved

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

Published simultaneously in Canada

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

form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise,

except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without

either the prior written permission of the Publisher, or authorization through payment of the

appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers,

MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to

the Publisher for permission should be addressed to the Permissions Department, John Wiley &

Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at

http://www.wiley.com/go/permission.

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

efforts in preparing this book, they make no representations or warranties with respect to the

accuracy or completeness of the contents of this book and specifically disclaim any implied

warranties of merchantability or fitness for a particular purpose. No warranty may be created or

extended by sales representatives or written sales materials. The advice and strategies contained

herein may not be suitable for your situation. You should consult with a professional where

appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other

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

damages.

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

our Customer CareDepartment within theUnited States at (800) 762-2974, outside theUnited States

at (317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print

may not be available in electronic formats. For more information about Wiley products, visit our

web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

ISBN: 9780470412893

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

http://www.copyright.comhttp://www.wiley.com/go/permissionhttp://www.wiley.com

CONTENTS

Preface xi

Contributors xv

1 Drug Discoveries by Gene Family 1

Karen E. Lackey

1.1 General Drug Discovery Components, 1

1.2 Further Reading for Expert Knowledge, 10

References, 13

2 G-Protein-Coupled Receptors 15

Stephen L. Garland and Tom D. Heightman

2.1 Introduction, 15

2.2 GPCR Structure and Function, 18

2.2.1 Subfamilies, 18

2.2.2 Structural Information and Homology Models, 22

2.2.3 Mechanisms of Receptor Modulation, 27

2.2.4 G-Protein Coupling and Assay Formats, 33

2.3 Challenges Facing the Area of GPCR Drug Design, 35

2.3.1 Hit Generation Strategies: Chemogenomics

and Privileged Structures, 36

2.3.2 Case History: Calcitonin Gene-Related Peptide

Receptor Antagonist (MK-0974, telcagepant), 40

2.3.3 Case History: Mixed Dopamine/Serotonin

Receptor Antagonist As An Atypical Anti-Psychotic, 43

v

2.3.4 Case History: Chemokine Receptor CCR5

Antagonist Maraviroc (CelsentriTM), 45

2.3.5 Case History: The Discovery of Cinacalcet (Sensipar1/

Mimpara1), a CaSR-Positive Allosteric Modulator, 47

2.4 Conclusions and Outlook, 49

References, 50

3 Ion Channels Gene Family: Strategies for DiscoveringIon Channel Drugs 53

Maria L Garcia and Gregory J. Kaczorowski


3.2 Ion Channel Subfamily Descriptions, 54

3.2.1 Voltage-Gated Ion Channels, 54

3.2.2 Inward Rectifier Potassium Channels, 55

3.2.3 Voltage-Gated Potassium Channels, 58

3.2.4 Calcium-Activated Potassium Channels, 61

3.3 Structure of Potassium Channels, 64

3.4 Criteria for Selection of Targets and Establishing Screens, 68

3.5 A Case Study in Ion Channel Drug Discovery, 70

3.6 Perspective on Ion Channels as Drug Targets, 77

References, 78

4 Integrins 85

David D. Miller


4.2 Integrin Inhibitor Discovery, 89

4.2.1 Cyclic Peptides, 91

4.2.2 Peptidomimetic Chemistry, 92

4.2.3 Preferred Peptidomimetic Fragments, 102

4.2.4 I-Domain Integrin Inhibitors, 103

4.2.5 Protein Structure-Based Design, 110

4.3 Challenges—Past and Future, 112

References, 113

5 Strategies for Discovering Kinase Drugs 119

Jerry L. Adams, Paul Bamborough, David H. Drewry,

and Lisa Shewchuk


5.2 Protein Kinase Structural Features, 120

5.3 Generating and Optimizing Kinase Inhibitors, 124

5.3.1 ATP Binding Pocket, 124

5.3.2 Non-ATP Binding Pockets, 133

vi CONTENTS

5.4 Establishing Screens for Understanding Kinase Activity

and Selectivity, 135

5.5 Case Studies of Successful Kinase Drug Discovery, 140

References, 150

6 Protease-Directed Drug Discovery 159

Richard Sedrani, Ulrich Hommel, and Jörg Eder


6.2 Aspartic Proteases, 162

6.2.1 HIV Protease Inhibitors, 164

6.2.2 Renin, 166

6.3 Metalloproteases, 168

6.3.1 Angiotensin-Converting Enzyme, 168

6.3.2 Matrix Metalloproteases, 170

6.4 Serine Proteases, 175

6.4.1 Dipeptidyl Peptidase 4 (DPP4), 176

6.4.2 Trypsin-Like S1 Serine Proteases of the Coagulation

Cascade, 179

6.5 Cysteine Proteases, 183

6.5.1 Cathepsin K, 184

6.5.2 Caspases, 188

6.6 Perspective on Proteases as Drug Targets, 190

References, 191

7 Small-Molecule Inhibitors of Protein–Protein Interactions:Challenges and Prospects 199

Adrian Whitty


7.2 Structure and Properties of PPI, 201

7.2.1 Constitutive versus Transient PPI, 201

7.2.2 Physicochemical Properties and Residue Propensities

of PPI, 202

7.2.3 Binding Energetics and ‘‘Hotspots’’, 204

7.3 Structural and Physicochemical Challenges to Inhibiting

PPI with Small Molecules, 207

7.3.1 Key Role of Adaptivity at the Interface, 210

7.3.2 Constraints of ‘‘Drug-Like’’ Chemical Space, 212

7.4 Identifying Hits and Leads Against PPI Targets, 213

7.4.1 Fragment-Based Screening, 214

7.4.2 Validating and Optimizing Hits and Leads, 218

7.5 Assessing the Druggability of New PPI Targets, 225

References, 227

CONTENTS vii

8 Transporters 235

Anne Hersey, Frank E. Blaney, and Sandeep Modi


8.2 Methodologies in Transporter Drug Design, 239

8.2.1 Structure-Based Methods, 239

8.2.2 Ligand-Based Methods, 245

8.3 Therapeutic Transporter Targets in Drug Discovery, 248

8.3.1 Vacuolar ATPases, 248

8.3.2 Gastric (P-) ATPases, 251

8.3.3 Neurotransmitter Transporters as Drug Targets, 252

8.4 Transporters as Liability Targets, 256

8.4.1 P-glycoprotein, 259

8.4.2 OATP1B1, 261

8.5 Application of Methods for Designing Interactions

with Liability Targets, 262

8.6 Perspective, 267

References, 267

9 Nuclear Receptor Drug Discovery 275

Hiroyuki Kagechika and Aya Tanatani


9.2 Nuclear Receptor Superfamily and Their Functions, 276

9.3 Agonism and Antagonism in Nuclear Receptor Functions, 279

9.3.1 AR Antagonists Effective Toward Mutated Receptors, 283

9.3.2 VDR Agonists and Antagonists, 286

9.3.3 Carboranes as Novel Hydrophobic Pharmacophores, 290

9.4 Medicinal Chemistry of Retinoid Nuclear Receptors, 293

9.4.1 Retinoid and Their Nuclear Receptors, 293

9.4.2 Retinobenzoic Acids, 300

9.4.3 RXR-Selective Ligands, 306

9.5 Clinical Application of Retinoids, 309


References, 312

10 Summary and Comparison of Molecules Designed to ModulateDruggable Targets in the Major Gene Families 317

Karen E. Lackey

10.1 Target Class Concept, 317

10.2 Summary of the Unique Features of Each Target Class, 318

10.2.1 GPCR/7TM, 318

10.2.2 Ion Channels, 320

10.2.3 Integrins, 322

viii CONTENTS

10.2.4 Kinases, 323

10.2.5 Proteases, 324

10.2.6 Protein–Protein Interactions, 325

10.2.7 Transporters, 327

10.2.8 Nuclear Receptors, 328


References, 331

Appendix 333

Index 341

CONTENTS ix

PREFACE

Approximately half of the anticipated small-molecule drug targets fall into just six

gene families: G-protein-coupled receptors (GPCRs), protein kinases, zinc metallo-

peptidases, serine proteases, nuclear hormone receptors, and phosphodiesterases. A

system-based research approach groups proteins into classes based on sequence and

common motifs that form 3D space receptive for small-molecule interactions. The

goal of small-molecule drug discovery is to modulate the activity of a biological

target via interactions with an externally administered molecule at optimal drug

intervention points in disease pathology to afford the maximum therapeutic

index. Key to achieving this mechanistic approach to drug discovery is the design,

synthesis, and evaluation of biologically effective compounds. Working within a

system of related targets allows scientists to apply learning from one member to

accelerate identification of ligands for other disease-associated targets.

The purpose of this book is to provide a description of compound design meth-

ods for generating small molecules that interact with important biological targets in

the following major gene families: G-protein-coupled receptors/7-transmembrane

receptors, ion channels, integrins, kinases, proteases, protein–protein interactions,

transporters, and nuclear receptors. Each chapter will cover affinity for the intended

target, the mechanism of the interaction, small molecule examples, and ways to

change the molecule to attenuate the activity. We hope to provide a solid foundation

of information that allows readers to then approach more expert technical literature

with a greater understanding. At the end of the introductory chapter, I have sum-

marized some books that provide an in-depth coverage of the functional areas of

drug discovery that contribute to this stage of research.

xi

The reader will readily come up to speed in a new area of research or be able to

compare their work across other gene families. Nature has developed classes of

related proteins and has borrowed similar motifs to keep all of the different func-

tions of a cell and complex system working. The selectivity achieved by nature is

impressive, and these are the lessons we can exploit to repair, enhance, diminish, or

eliminate an activity using a small molecule. By understanding the major gene

families, the scientist can design molecules that target the intended protein and

minimize interactions with the other protein classes. Beyond the scope of this

book, but nonetheless important in selection of targets and molecule design, are

the subsequent steps of drug development. The stages of drug generation beyond

discovery include extensive toxicity measurements, development, clinical evalua-

tion, registration, and commercialization and marketing of effective medicines

for specific disease treatments or prevention.

Synthetic, structural, computational, and medicinal chemists in academia, bio-

medical companies, and the pharmaceutical industry will benefit from a gene

family-focused description of molecular design. There is a tremendous amount of

literature on the topic and yet very little work has been done to condense the infor-

mation into a manageable format as practical guidance for a chemist to get started

in the area of designing compounds that intervene in important points of disease

pathology. Also, overwhelming amounts of information are available in each

research area making it a daunting task to do cross-comparisons of the different

gene families. The pros and cons of different discovery methods (i.e., use of

high-throughput screening, protein–ligand crystal structures, transient transfection

assays, etc.) are included to help the reader understand the value and context of the

biological evaluation of compounds currently available within each gene family.

Another benefit added to this book is the biographies of the contributing authors

compiled in Appendix. By reviewing the educational backgrounds and careers of

each of the experts in the field, the reader can peruse the many paths of study

one can take in a scientific journey for drug discovery. The reader can also learn

how the scientific fields are integrated to design molecules for drug discovery. I sin-

cerely hope you enjoy reading the fascinating way by which the scientists have

managed to create small molecules that effectively modulate their intended biolo-

gical targets in the major target classes covered in the chapters of this book.

April 2008 KAREN E. LACKEY

xii PREFACE

ACKNOWLEDGMENT

Carol A. Petrick assisted in essential administrative aspects in the preparation of

this work.

xiii

CONTRIBUTORS

Jerry L. Adams, GlaxoSmithKline, 1250 South Collegeville Road, Collegeville,PA 19426, USA.

Paul Bamborough, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood

Road, Stevenage, Hertfordshire SG1 2NY, UK.

Frank E. Blaney, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood


David H. Drewry, GlaxoSmithKline, PO Box 13398, Five Moore Drive, Research

Triangle Park, NC 27709, USA.

Jörg Eder, Novartis Institute for BioMedical Research, Novartis Pharma AG,

WKL-136.6.93, CH-4002 Basel, Switzerland.

Maria L. Garcia, Merck Research Laboratories, R80N-C31, PO Box 2000, Rah-way, NJ 07065, USA.

Stephen L. Garland, GlaxoSmithKline, Medicines Research Centre, Gunnels

Wood Road, Stevenage, Hertfordshire SG1 2NY, UK.

Tom D. Heightman, Structural Genomics Consortium, University of Oxford, Old

Road Campus Research Building, Old Road Campus, Roosevelt Drive, Heading-

ton, Oxford OX3 7DQ, UK.

Anne Hersey, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood


xv

Ulrich Hommel, Novartis Institute for BioMedical Research, Novartis Pharma

AG, WKL-136.6.93, CH-4002 Basel, Switzerland.

Gregory J. Kaczorowski, Merck Research Laboratories, R80N-C31, PO Box

2000, Rahway, NJ 07065, USA.

Hiroyuki Kagechika, School of Biomedical Sciences, Tokyo Medical and Dental

University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8549, Japan.

Karen E. Lackey, GlaxoSmithKline, PO Box 13398, Five Moore Drive, Research

Triangle Park, NC 27709, USA.

David D. Miller, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood


Sandeep Modi, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood


Richard Sedrani, Novartis Institute for BioMedical Research, Novartis Pharma

AG, WKL-136.6.93, CH-4002 Basel, Switzerland.

Lisa Shewchuk, GlaxoSmithKline, PO Box 13398, Five Moore Drive, ResearchTriangle Park, NC 27709, USA.

Aya Tanatani, Department of Chemistry, Ochanomizu University, 2-1-1 Ohtsuka,

Bunkyo-ku, Tokyo 112-8610, Japan.

Adrian Whitty, Department of Chemistry, Boston University, Metcalf Center for

Science and Engineering, 590 Commonwealth Avenue, Room 299, Boston MA

02215, USA.

xvi CONTRIBUTORS

1DRUG DISCOVERIES BY GENEFAMILY

KAREN E. LACKEY

I am fascinated by the notion that a small molecule with a fraction of the molecular

weight and size of a comparatively enormous biological protein can change the

activity of that protein in directed ways. For years, the scientific community dedi-

cated to drug discovery has developed a knowledge base of overwhelming propor-

tions to hone these interactions of small molecules to seek out the intended target

and seemingly avoid all other proteins in its path. We expect these molecules to

enter the body in a convenient manner, interrupt the disease pathology, and quietly

exit the body upon job done. While there are certainly successes out there, some of

which are described in this book, there are so many more unmet medical needs that

require our sense of urgency.

When the human genome was solved, the expectations were that the targets to

affect, avoid, or modulate human diseases would be tackled with drugs emerging at

an unprecedented rate. Sometimes, I believe, the human genome project merely

uncovered just how much we do not know and how far we need to go to understand

the role of genes and protein products in normal and diseased tissues. Approxi-

mately 22,000 genes can be organized into the major protein classes they code

and that are related by common structural and protein sequence features, called

gene families, some of which are depicted in Fig. 1.1 (Hopkins and Groom, 2002).

1.1 GENERAL DRUG DISCOVERY COMPONENTS

Some areas of drug discovery methods transcend all gene families and will be

briefly described here for general background to the subsequent chapters that are

Gene Family Targeted Molecular Design, Edited by Karen E. LackeyCopyright # 2009 John Wiley & Sons, Inc.

1

dedicated to the larger gene families. Broadly speaking, genomics is the study of all

genes in an organism including the sequence, structure, function, and regulation.

When referring to genetics, it is typically the inherited variation of genes that is

being considered and is one component of genomics. Figure 1.2 summarizes the

gene expression process with the general concepts of molecular biology where

the rapid technology advances have made a major impact over the last few decades

providing significant value to drug screening (Watson et al., 2003). In the nucleus,

the DNA is transcribed into messenger RNA, which is transported to the cytoplasm.

Translation occurs when transfer RNA (tRNA) reads the genetic code where three

base pairs code for each amino acid of a polypeptide chain that forms the protein

FIGURE 1.1 The gene families of proteins are classified by function and common structure

motifs as can be seen by the representative structures for kinases, integrins, GPCRs, ion

channels, proteases, and nuclear receptors. (See the color version of this figure in the Color

Plates section.)

N C

P

PCHO

Product

DNA

Primary RNA

Messenger RNA

Protein

Functional protein

Transcription

RNA splicing

RNA transporttranslation

N C

Processing

FIGURE 1.2 The normal process and products of gene expression.

2 DRUG DISCOVERIES BY GENE FAMILY

product. Protein products are the biological machinery where the normal functions

are carried out, such as activation, inactivation, scaffolding, and signaling in a

healthy cell. Drug intervention can occur at any part of this process, as well as

in the modulation of the activity of the proteins produced.

The discovery of drugs for diseases is typically selected based on medical need

and research feasibility. A biological target is a component within a normal func-

tioning cell or tissue where aberrant activity has given rise to a disease. For exam-

ple, a genetic instability leads to mutations in cell signaling that can cause cancer.

The target would become the biological component of the cascade to stop the

unwanted or unchecked signal. A thorough understanding of disease pathology

and/or the underlying genetics underpins the successful selection of targets. In

some cases, genetic information can be used in patient selection in a field called

pharmacogenomics. Whether the target is selected via genetic factors or by deter-

mining an ideal intervention point in the cellular protein activities, the ability to

modulate the intended target using a small molecule is called tractability or

druggability. The assessment of the tractability is shown in an oversimplified

scheme in Fig. 1.3. Only a portion of the protein products of the human genome

and infectious agents can be modulated by a small molecule or biological product

(vaccine, antibody, etc.), and thus named the druggable genome (Imming et al.,

2006). One way to increase the capability and understanding of biological target

modulation involves organizing the protein products of the gene into families

based on similar structure and function. This strategy allows teams of scientists

to characterize the biology of the drug targets and build a knowledge base rele-

vant to diseases that can readily be transferred from one member of the gene

family to another.

Gene

mRNA

Protein

Disease link(target)

Test small moleculeson target or relatedprotein

Lead or tool created

Drug

Confirmormodify approach

I

FIGURE 1.3 The process of target selection and validation is linked with assessing

tractability or druggability.

GENERAL DRUG DISCOVERY COMPONENTS 3

It seems that target validation (defined as the ability to effectively treat the dis-

ease via modulation of the intended biological target) is an ongoing process from

the time of target selection all the way until the desired activity is observed in a

patient. The drug discovery industry continuously challenges the data through a ser-

ies of increasingly more complex biological evaluations. A compound evaluation

pathway is typically used to assess the ability of the small molecules and to assist

with decision making, as summarized in Fig. 1.4. The labels in the diagram also

help with nomenclature used in the field to describe the stages of identifying small

molecules that interact with the biological target of interest.

Treatment strategies from the molecular knowledge of a disease are used to

prioritize targets and drug intervention points. Sometimes, it is necessary to create

test compounds for multiple points in a particular disease pathology to determine

the one that is the most tenable for a treatment of a patient. It could be that targeting

one component of a pathway provides a more tractable target (i.e., a knowledge

base exists for the gene family) or that using a test molecule in the assays that

mimic the disease shows that it is insufficient to be an effective medicine. It still

remains elusive to always predict which parts of a disease pathology are the

most likely to be the best for producing the desired effects. Sometimes it is the

wrong target, and sometimes the problem lies with compound’s activity profile.

In early stages of research when little is known about the disease at the molecular

level, it is very difficult to discern between the two types of failures.

To begin the drug discovery process, the protein target is isolated and a screen is

devised to test specific compounds, focused sets, and/or enormous sets of com-

pounds of over 500,000 compounds often referred to as high-throughput screening

or diversity screening. A compound must be potent toward the intended biological

target, and the affinity is determined by an in vitro assay. In vitro assays are

designed to measure antagonism (inhibition) of activity to stop an unwanted activ-

ity or agonism to provide more of an activity that has been lost or subdued by a

In vitro selectivityTarget assay

Cellular and functional assays

In vivo studies: PK/PD

Toxicity assessment

Compounds

FIGURE 1.4 A generic compound evaluation pathway for a drug discovery project.


disease. There are many mechanisms that will be considered in the following chap-

ters to achieve complete or partial antagonism or agonism that are specific to a gene

family. Fundamental to all of them is a need to be able to assess the activity at the

intended target in a way that allows a chemist to compare the inherent activity to

either increase or decrease it depending on the drug intervention strategy. The

screens often reflect the biological activity of the target, for example, by measuring

the inhibition of an enzymatic reaction. Screens are also built to measure binding

affinity—how tight the compound sticks to the target, with the assumption that

by binding to certain parts of a protein, the target can be modulated. Some targets

cannot be screened as an isolated assay and must be screened in a cellular setting.

Throughout each of the following chapters, details regarding the type of assays

that are used in each gene family of targets and reasons for the choices will be

discussed.

The compound evaluation pathway is designed to enable decisions on what com-

pounds to synthesize. Designing compounds based on the data that are generated in

each of the assays used in a screening scheme involves computational chemistry,

structural biology, in silico predictions, and mathematical analyses. Different

gene families rely on specialized design tools in combination with physicochemical

assays. The overall process is an iterative learning cycle in which medicinal che-

mists generate ideas for structural changes to simultaneously optimize activities

across all of the assays employed. Chemists working on a gene family system

use their knowledge of the target class assisted by computational techniques to

design compounds to find biological activity. Sets of compounds, often called

arrays, are created to test hypotheses regarding structure–activity relationships

(SAR). Where there is more protein structure information or a long history of

designing active compounds, the sets of compounds are generally small as the che-

mist hones in on the best molecule. When less is known or available to guide the

design, much larger sets of compounds are generally synthesized (>500). Takingadvantage of special pockets or distinct binding interactions can help to narrow

the scope of what the drug will modulate. By focusing on gene families, methods

to design selectivity for one family member compared with another are discussed. It

is noteworthy that building in this selectivity still requires extensive toxicity testing

to verify minimal off-target effects.

After a screening event, the data are analyzed to determine if any compounds

provide the readout intended. The sets may be tested in a general format to cast

as wide a net to find as many active starting points as possible. Testing many com-

pounds at a single concentration is often used. Anything that provides a hint of

activity is typically called an assay ‘‘hit.’’ The hits are analyzed to check that

they are truly active and interesting to work on. They are often narrowed to only

include chemical series that will be further worked on via synthetic modifications.

Depending on the gene family, the hit validation can be as simple as generating a

value that is called the effective dose to give a 50% response. The reason this value

is used is because it makes comparing compounds much more reliable based on

using a reproducible point on the curve. It is expressed as a concentration or a

pXC50 (the negative log of the IC50 for better numeric comparisons of activity).


It is also important to understand the error range of the data. Often replicates are

averaged to provide values for comparison among series of chemically related com-

pounds and that gives a fairly good representation of what activity can be used for

analyses. However, to assess how significant an improvement or decrease in activity

is, it is important to get an understanding of the variability of the data. Often, pED50values are useful for preventing overinterpretation of the data. For example, if you

compare compound Awith an IC50 value of 5 nM with compound B that is 20 nM,

you might be tempted to say the compound A is fourfold more active than B. How-

ever, if the average range of the assay data on repeat tests is �0.2 log units, then thecompounds would be considered equipotent. The data interpretation aspects are

different among the various gene families, given the state of the art of the assays

systems, and some discussion of the reliability of the data is covered in each chap-

ter. Also, the term high-content lead series is often used to describe a chemical

series that meets many of the criteria for modulating the target of interest and

often requires the parallel use of several hit generation and follow-up techniques

(Bleicher et al., 2003).

The activity against an isolated target in an in vitro assay is often verified or

further evaluated in a cellular environment. In cases where it is not possible to cre-

ate an assay for a desired target, whole cells or more complex systems are used to

find active small molecules that incorporate the desired cellular outcome, or pheno-

type, as the assay system. These assays are typically lower throughput (less com-

pounds are screened because the assay is either more difficult to run or too

expensive) but more information rich. Also, it might be necessary to run control

assays to ensure that the activity that is observed is related to the desired biological

target modulation. For many drug discovery projects, it is also important to test for

selectivity against other similar targets or important conflicting biological path-

ways. Each of the gene families has specific selectivity challenges that will be dis-

cussed for greater understanding of how to design a compound that is effective for

the disease area without unwanted or intolerable side effects. For example, often

screening is done on the human protein target, but the activity must also be checked

on the animal orthologue (the protein found in the animal that is analogous to the

human version) to support the in vivo animal studies of efficacy and toxicity.

Data generated from evaluating compounds in cell-based assays are important to

integrate with any isolated enzymatic or binding assay. A compound’s activity is

affected by cellular context that includes multiprotein complexes, cell-type specific

pathways, branching pathways, intracellular localization, and multiple target iso-

forms or states (phosphorylated, open, closed, etc.). Of course, compensatory

bypass or feedback loop mechanisms cannot be detected in an isolated target assay

and thus allow the cell assay to be used as a way to judge the tractability of the drug

intervention point. In this way, cellular assays can better measure efficacy, not just

potency. As a generalized screening tool, cell-based assays can also be set up to

select for only permeable, nonmetabolized, and nontoxic compounds. This method

of identifying hits and optimizing compounds can be useful in limiting the number

of chemical series. On the contrary, it automatically eliminates potentially potent

compound interactions that may need properties such as solubility or permeability


built in. Each target class strategy of molecular design balances these aspects of

cellular screens based on the typical properties or needs of the drug discovery

research efforts.

The SAR is the study of the correlation between the biological activity and the

chemical structure. There can be SAR at every level of the screening cascade, and it

guides the entire process of designing an effective medicine. The knowledge

derived from SAR forms an understanding of interaction features such that predic-

tions can be made on what kinds of molecules will retain activity and which ones

will not within a chemical series. There are nearly always compromises that can be

made in potency to add solubility and to create an improved therapeutic index (ratio

of the dose required for biological efficacy to the dose at which unacceptable toxi-

city occurs). All along the project path, it is important to be aware of the SAR to

know where on the molecule changes can be made to adjust for newly uncovered

risks or to be aware of what cannot be changed without losing all target affinity.

This book should help the reader know what types of compounds to synthesize

in each gene family, and examples are provided to understand the scope of the gen-

eric properties of the small molecules. Each family of targets appears to depend on

a few major design techniques based on the properties of the biological targets.

However, paramount to the success of creating a drug candidate with optimized

gene family protein interactions is the ability to incorporate drug developability

characteristics. Requirements of developability properties are different for the vari-

ety of intended routes of administration (e.g., oral, inhaled, and intravenous) and

typically cover absorption, distribution, metabolism, and excretion, which are

abbreviated as ADME properties. There is a whole branch of science supporting

the necessary ADME features and components that affect them related to disease

pathology and normal body functions. Although the details are beyond the scope of

the target class design focus, these developability properties must be optimized

while generating potent interactions with the intended biological targets.

Developability, metabolism, and pharmacokinetics (DMPK) encompass many

types of assays that are used to determine the in vivo properties of potential drug

molecule. While the emphasis in this book focuses on designing molecules for spe-

cific biological targets within gene families, the ultimate design of the compounds

must take into consideration the properties that allow the compound to get to the

tissue or site of action. In vitro tests to measure solubility at various pH levels can

often be done to predict absorption levels or the handling properties of the drug

product. The activity of compounds is measured in a variety of in vitro assays

with CYP isoforms (e.g., CYP3A4 or CYP2D6) to predict which compounds might

have better or worse in vivo properties based on predictions on metabolism of the

parent molecule.

The principal evaluation of in vivo properties comes by testing representative

compounds in rodent models looking for data to determine half-life, bioavailability,

volume of distribution, and clearance of the compound. A thorough discussion of

this topic is beyond the scope of the book, but it is important to note that each target

class discussed may require certain types of compound classes for proper interac-

tions to modulate the target. These features may offer particular challenges in


designing an effective medicine and will be pointed out in the chapters that follow.

A high-level example, though, is to look at the properties of many drugs that mod-

ulate nuclear receptors. They tend to be lipophilic in nature and, in contrast, the

kinase compounds that interact at the ATP binding domain are frequently ‘‘flat’’

heterocycles with inherently poor solubility. These features are addressed in drug

discovery research projects and it helps to have a knowledge base in the area to

facilitate the appropriate design focus.

More recently, in silico techniques have made huge advances in the area of pre-

dictive properties. Many software packages are available on the computer desktop

for either straightforward analyses or more complex analyses for a trained compu-

tational chemist. They are used to identify potential issues with active series iden-

tified from a screening event to help prioritize the hits with the highest likelihood of

generating high-quality compounds. The in silico methods are also used to priori-

tize a synthetic plan by taking chemistry that is designed to functionalize one or

more sites on a central scaffold and create the structures by building the final pro-

ducts as a virtual set (existing only as an electronic collection of structures). The

virtual compounds are then put through predictive models in order to select the sub-

set of compounds that would be predicted to retain the desired activity. For exam-

ple, if the intended target is in the central nervous system (CNS), then an in silico

method can be used to virtually screen a large set of possible compounds for CNS

penetration property predictions. The purpose of such an exercise is to reduce the

number of compounds required for synthesis, thus reducing time, chemical reagent

costs and waste, and biological screening costs. To predict oral absorption using

in silico predictions, the focus is on oral bioavailability (how much drug reaches

the blood compared to the original dose) and predicting interactions with metabo-

lizing enzymes (e.g., enzymes in the liver designed to rid the body of toxins might

degrade a compound rapidly or too slowly).

Virtual screening techniques are often used for the gene family target interac-

tions. Chemical descriptors form the basis of the structure input and often account

for the nature of the compounds, such as molecular size, polarity, hydrogen bonding

capacity, lipophilicity, and acidity/basicity. As a medicinal chemist builds knowl-

edge in each of the target classes, the data from virtual and measured screening

are used together to form the project strategy in creating a drug.

Lipophilicity is measured by log P, which is a partition coefficient that describes

the ratio of the concentrations of a compound between two phases: octanol and

water. Why this matters in drug design is due to the fact that the cell membrane

permeability (ability of compounds to pass through the cell membrane to reach

intracellular targets) requires more lipophilic nature, which has a tendency to be

less soluble in blood/aqueous media. Conversely, hydrophilic compounds tend to

be water soluble and amenable to blood transport, but are poorly cell permeable.

This measure of lipophilicity requires a balance of design between the various

effects to create a drug that gets to the site of action and retains the key functional

groups necessary for target interaction for the desired biological effect. A calculated

log P value, designated clog P, has become fairly easy to calculate and is often used

to prioritize which of the compounds should be synthesized given the ideal range


of 2–4 (for oral drugs), where the larger the number the greater the lipophilicity

(Wenlock et al., 2003).

Solubility is an important physical property for a drug for both biological effec-

tiveness and drug product formulation and processing. Solubility testing can be

done in a variety of conditions to mimic neutral aqueous, gastric fluid simulation

or intestinal fluid simulation, or any pH necessary for drug delivery. For oral drugs

need to be well absorbed by the body, unless they are designed to specifically com-

bat a GI tract disease for which systemic exposure is unnecessary (e.g., certain bac-

teria infections of the gut). Understanding the solubility characteristics of series can

also assist in interpreting the biological data from screening. Care should be taken

to ensure that the compounds are fully solubilized in the test solution to study the

SAR trends.

In silico methods such as structure-based design and pharmacophore modeling

are best covered under the specific target class chapter. When the three-dimensional

structure of an enzyme or receptor is known, the information can be used to deter-

mine the interactions between it and a bound compound. Computer models can be

generated to mimic other similar molecules bound to the active site so that design-

ing can be based on maximizing interactions (while taking into account all of the

other properties). It is often referred to as finding where the molecule can tolerate a

substitution that would allow the synthetic chemist to attach functional groups to

improve the properties without destroying the positive interactions with the

ligand/protein. Crystallography can be used to design ideal drug molecules and

will be highlighted in the gene families where it currently has the greatest impact.

Often, the three-dimensional structure of a protein target is not known, but using

binding or catalytic activity of compounds from screening events, a pharmacophore

model can be built. A high-level explanation follows: A set of compounds that are

known to have activity for the intended target can be used to gain insight into the

steric and electronic structural features necessary for a compound to bind. Even if

they are poorly active, a model can be built to identify the specific features that

garner any activity. Iterative searching and testing in the in vitro assay allows the

scientists to refine the pharmacophore model and hone in on the key features for

better and better affinity. Also, models can be built based on the similarity of pro-

teins or protein folds from one member of a protein family to another. Often, the

models help scientists understand how changes in the small molecule affect the

activity observed in the assay and designs incorporate better interactions through

appending functional groups.

In typical drug discovery programs, assays and in silico predictions such as solu-

bility, in vitro metabolic stability, and cell permeability are used as filters for deter-

mining ideal pharmacodynamic and pharmacokinetic properties (Leeson et al.,

2004). There seems to be an ongoing discussion in the medicinal chemistry field

over the value of rules in drug design, but most scientists agree that trends can

be derived for general predictions (Leeson and Springthorpe, 2007). However, with-

in a given chemical series, there can be anomalies and those should be considered

when using assays for guidance in progressing compounds through more complex

and resource-intensive biological evaluation systems. Specific examples of screens


for in vitro developability assays include solubility in multiple solvents (neutral

aqueous, simulated gastric fluid, and simulated intestinal fluid), Caco-2 permeabil-

ity, p450 enzyme assays, and microsomal metabolism studies using methods similar

to published reports (Guo and Shen, 2004; Li 2004). It is invaluable to have training

sets of compounds to investigate correlations between in vitro assays designed to

predict drug-like characteristics and actual in vivo data within chemical series.

The entire discussion on protein family affinity and modulation coupled with

incorporating the desired parameters for delivering the small molecule to the site

of action is only a fraction of the drug discovery process. Most researchers will

agree that the more the quality is put into the compounds in the early stages of iden-

tifying a potential drug candidate, the more likely it will be successful in reaching a

patient in need. However, as can be seen by the remarkable financial investments

(Goozner, 2004) made in the field, I still believe that the combination of knowing

which intervention point to modulate (target selection and validation), achieving

selective target interactions with no off-target activities, meeting all of the drug-

like criteria for the intended route of administration, and being able to synthesize

sufficient, pure quantities of material in the correct form is a formidable task to

expect of one small molecule. However, there are success stories out there. We

owe it to ourselves to understand how to modulate every potential biological target

to enable successful drug discovery for the patients still in need.

1.2 FURTHER READING FOR EXPERT KNOWLEDGE

1. Textbook of Drug Design and Discovery, 3rd ed., edited by Povl Krogsgaard-

Larsen, Tommy Liljefors, and Ulf Madsen, Taylor & Francis, Inc., New York,

2002.

This book is a thorough guide to drug discovery with a compilation of

chapters going into details on mechanics of small molecule–protein interac-

tions, kinetics, and development of pharmacophore models and descriptors. A

terrific explanation of quantitative SAR with physicochemical parameters and

how to interpret the data is also included. A general explanation of receptor-

based targets in multiple gene families is provided, with subsequent chapters

discussing in more detail on ion channels and therapeutic outcomes from

modulation of targets. Creating and interpreting data from imaging studies

with labeled compounds provides insights into understanding drug distribu-

tion and why so much effort goes into understanding physicochemical proper-

ties of molecules. The chapter on enzymes with mechanism and kinetics of

inhibition offers many drug discovery approaches to these important biologi-

cal reactions. More advanced topics in drug design include prodrugs (a pre-

cursor to the active parent drug typically done to mask an unwanted property

of the compound and dependent upon the human metabolism), peptides and

their mimics, and the use of metals in modulating biological targets that are

dependent on them for their normal processes. The textbook ends with

specific examples of drug discovery in the research area of antiviral and

anticancer agents where many of the concepts discussed throughout the


book are shown in practice of selecting targets in the disease pathology and

determining how to modulate them for a beneficial effect.

2. Biopharmaceutical Drug Design and Development, edited by Susanna Wu-

Pong and Yongyut Rojanasakul, Humana Press, Inc., Totowa, NJ, 1999.

It provides excellent descriptions of using biotechnology in generating very

specific target modulation and the unique properties of these biologicals in

creating useful drugs to combat diseases. It is an easy to read scientific

book that offers a marked contrast to small molecule design in the gene

family focused molecular design. Vaccines and recombinant human products,

for example, have excellent specificity but offer difficult challenges to man-

ufacture, deliver to the site of action in the body, and management of stability

issues throughout the entire process of ‘‘bench to bedside.’’ Many technology

advances have been made and continue to be made to provide effective med-

icines in this class of drugs, and this book provides a foundation of knowledge

to appreciate the more expert literature.

3. Guidebook on Molecular Modeling in Drug Design, edited by N. Claude

Cohen, Academic Press, San Diego, CA, 1996.

There are many computational approaches needed to operate in the poten-

tial chemical space that can be reached with synthetic chemistry to maximize

interactions with the intended biological target(s). This guidebook delves into

the details of the tools used with a variety of molecular capabilities and

applies them to examples in ligand generation. The complexity of drug–bio-

logical target interactions that take into account pharmacophores, desolvation,

and binding properties is discussed from the perspective of computer-assisted

drug discovery and modeling. While the book is over 10 years old, it has a

good foundation for anyone interested in understanding the breadth of contri-

butions from computations chemistry to the design of drugs.

4. Protein–Ligand Interactions. From Molecular Recognition to Drug Design,

edited by H.-J. Böhm and G. Schneider, in: Methods and Principles in

Medicinal Chemistry, Vol. 19, edited by R. Mannhold, H. Kubinyi, and G.

Folkers, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2003, ISBN

3-527-30521-1.

This book is a must-have expert reference for any serious medicinal chemist

who wants to understand the intimate interactions of a small molecule and a

large biological protein. From the list of abbreviations provided in the front

of the book to the well thought-out order of chapters describing details of small

molecule modulation properties, the readers will, at a minimum, understand the

nomenclature of the business partners they work with and, at best, will utilize

the appropriate principles that affect the SAR of the drug they are hoping to

discover. Most of the formulas are broken down into the important features

of the design tools, so even if you were not going to use the calculated

approaches described, the lessons that are included add insights for the medic-

inal chemist. The combination of computational chemistry and compound

screening explanations is useful for understanding all of the factors that go

into compound design at the molecular recognition detailed level.

FURTHER READING FOR EXPERT KNOWLEDGE 11

5. Receptor-Based Drug Design, edited by Paul Leff, Marcel Dekker, New York,

1998, ISBN 0-8247-0162-3.

This book is part of a long series of textbooks dedicated to drugs and phar-

maceutical sciences. It is a useful reference book for a more in-depth look at a

subset of the gene families where the biological target is a receptor. Details

and insights are provided on the functions of a receptor such that thorough

interpretation of assay systems can be achieved. Mechanisms of interactions

range from determining binding affinity to understanding many forms of

modulation. Understanding the outcome of the effect of small molecule on

the receptor in an information-rich assay system should help guide the reader

to more expert advice on judging the tractability of a receptor for drug dis-

covery. While advances in assay systems for receptors have advanced over

recent years along with a large increase in the number of ligand-bound

protein crystal structures, this book will still provide expert details for

more in-depth understanding of the reader’s research project and current

scientific literature.

6. Analogue-Based Drug Design, edited by Janos Fischer and C. Robin

Ganellin, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2006, ISBN

3-527-31257-9.

This book is one in a series of six titles dealing with different aspects of

drug discovery. It covers the general approach that a synthetic chemist can

use to explore chemical space around an established active compound with

the potential for differentiating the final product. Nineteen specific case stu-

dies are provided in detail to demonstrate how the approach works, the para-

meters used to differentiate analogues, and what the properties were of the

final drug candidate.

7. Structure-Based Drug Discovery, edited by Harren Jhoti and Andrew Leach,

Springer, Fordrecht, The Netherlands, 2007, ISBN 1-4020-4406-2.

Understanding the cellular proteins at the molecular level from three-

dimensional protein structures allows researchers to design specific interac-

tions to create potent, small molecule drug candidates. This book details

the process improvements in protein crystallization and structure determina-

tion, and provides an advanced view of scaling the technologies for parallel

production for higher throughput. A general guide to using structure in dis-

covering small molecule fragments and strategies for converting the frag-

ments into lead series is also included. Multiple methods of screening for

fragments of drug-like molecules are covered including NMR and X-ray crys-

tallography. An alternative to linking fragments together to increase affinity is

discussed, whereby the screening event begins with finding the small core

scaffold followed by functionalization to improve affinity. All these methods

of finding ligand-efficient hits depend upon understanding the mechanisms of

the interactions between the small molecules and the protein, so the chapter

on biophysical methods is useful. The final chapters cover computational

ways to extend the value of the structure-based design methods.


Documents

GENE FAMILY TARGETED MOLECULAR DESIGNdownload.e-bookshelf.de/download/0000/5724/70/L-G... · You should consult with a professional where appropriate. Neither the publisher nor author