GENE FAMILYTARGETEDMOLECULAR DESIGN
Edited by
KAREN E. LACKEY
InnodataFile Attachment9780470423929.jpg
GENE FAMILYTARGETEDMOLECULAR DESIGN
GENE FAMILYTARGETEDMOLECULAR DESIGN
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
3.1 Introduction, 53
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
4.1 Introduction, 85
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
5.1 Introduction, 119
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
6.1 Introduction, 159
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
7.1 Introduction, 199
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
8.1 Introduction, 235
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
9.1 Introduction, 275
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
9.6 Perspective, 311
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
10.3 Perspective, 330
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
Road, Stevenage, Hertfordshire SG1 2NY, UK.
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
Road, Stevenage, Hertfordshire SG1 2NY, UK.
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
Road, Stevenage, Hertfordshire SG1 2NY, UK.
Sandeep Modi, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood
Road, Stevenage, Hertfordshire SG1 2NY, UK.
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.
4 DRUG DISCOVERIES BY GENE FAMILY
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).
GENERAL DRUG DISCOVERY COMPONENTS 5
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
6 DRUG DISCOVERIES BY GENE FAMILY
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
GENERAL DRUG DISCOVERY COMPONENTS 7
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
8 DRUG DISCOVERIES BY GENE FAMILY
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
GENERAL DRUG DISCOVERY COMPONENTS 9
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
10 DRUG DISCOVERIES BY GENE FAMILY
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.
12 DRUG DISCOVERIES BY GENE FAMILY