Protein kinases as drug targets methods and principles in medicinal chemistry 49

List of Contributors XI Preface XV A Personal Foreword XVII Part One Hit Finding and Profiling for Protein Kinases: Assay Development and Screening, Libraries 1 1 In Vitro Characterizati

and Michael Hamacher

Protein Kinases as Drug Targets

Methods and Principles in Medicinal Chemistry

Edited by R Mannhold, H Kubinyi, G Folkers

Editorial Board

H Buschmann, H Timmerman, H van de Waterbeemd, T Wieland

Previous Volumes of this Series:

Sotriffer, Christopher (Ed.)

Rautio, Jarkko (Ed.)

Prodrugs and Targeted Delivery

Towards Better ADME Properties

Ghosh, Arun K (Ed.)

Aspartic Acid Proteases as

Therapeutic Targets

2010

ISBN: 978-3-527-31811-7

Vol 45

Ecker, Gerhard F / Chiba, Peter (Eds.)

Transporters as Drug Carriers

Structure, Function, Substrates

Sippl, Wolfgang / Jung, Manfred (Eds.)Epigenetic Targets in Drug Discovery

2009 ISBN: 978-3-527-32355-5 Vol 42

Todeschini, Roberto / Consonni, VivianaMolecular Descriptors for Chemoinformatics

Volume I: Alphabetical Listing /Volume II: Appendices, References

2009 ISBN: 978-3-527-31852-0 Vol 41

van de Waterbeemd, Han / Testa,Bernard (Eds.)

Drug BioavailabilityEstimation of Solubility, Permeability,Absorption and Bioavailability

Second, Completely Revised Edition 2008

ISBN: 978-3-527-32051-6 Vol 40

Ottow, Eckhard / Weinmann, Hilmar (Eds.)Nuclear Receptors as Drug Targets

2008 ISBN: 978-3-527-31872-8 Vol 39

Bert Klebl, Gerhard Müller, and Michael Hamacher

Protein Kinases as Drug Targets

Series Editors

Prof Dr Raimund Mannhold

Molecular Drug Research Group

ATP binding site of the Cyclin-dependent protein

kinase 7 (CDK7), a member of the CDK family

involved in the regulation of the cell cycle and

tran-scription The kinase active site is divided in sub-sites

according to its interactions, varying between

indivi-dual enzymes and allowing the indiviual design of

selective inhibitors (Photo courtesy C McInnes)

All books published by Wiley-VCH are carefully produced Nevertheless, authors, editors, and publisher do not warrant the information contained

in these books, including this book, to be free of errors Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Library of Congress Card No.: applied for

British Library Cataloguing-in-Publication Data

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

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliogra fie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de.

# 2011 Wiley-VCH Verlag & Co KGaA, Boschstr 12, 69469 Weinheim, Germany

All rights reserved (including those of translation into other languages) No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers Registered names, trademarks, etc used in this book, even when not speci fically marked as such, are not to be considered unprotected by law.

Typesetting Thomson Digital, Noida, India Printing and Binding betz-druck GmbH, Darmstadt Cover Design Grafik-Design Schulz, Fußgönheim Printed in the Federal Republic of Germany Printed on acid-free paper

ISBN: 978-3-527-31790-5

List of Contributors XI

Preface XV

A Personal Foreword XVII

Part One Hit Finding and Profiling for Protein Kinases: Assay Development

and Screening, Libraries 1

1 In Vitro Characterization of Small-Molecule Kinase Inhibitors 3

Doris Hafenbradl, Matthias Baumann, and Lars Neumann

1.1 Introduction 3

1.2 Optimization of a Biochemical Kinase Assay 4

1.2.1 Step 1: Identification of a Substrate and Controlling of the Linearity

between Signal and Kinase Concentration 4

1.2.2 Step 2: Assay Wall and Optimization of the Reaction Buffer 6

1.2.3 Step 3: The Michaelis–Menten Constant Kmand the ATP

Concentration 10

1.2.4 Step 4: Signal Linearity throughout the Reaction Time

and Dependence on the Kinase Concentration 12

1.2.5 Step 5: Assay Validation by Measurement of the IC50

of Reference Inhibitors 15

1.3 Measuring the Binding Affinity and Residence Time

of Unusual Kinase Inhibitors 15

1.3.1 Washout Experiments 18

1.3.2 Surface Plasmon Resonance 19

1.3.3 Classical Methods with Fluorescent Probes 21

1.3.4 Preincubation of Target and Inhibitor 22

1.3.5 Reporter Displacement Assay 22

1.3.6 Implications for Drug Discovery 25

1.4 Addressing ADME Issues of Protein Kinase Inhibitors in Early

Drug Discovery 26

Protein Kinases as Drug Targets Edited by B Klebl, G Müller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

1.4.2.4 Transporter Assays Addressing P-gp Interaction 33

1.4.3 Experimental Approaches to Drug Metabolism 34

1.4.3.1 Background and Concepts 34

1.4.3.2 Measuring Metabolic Stability 37

1.4.3.3 Measuring CYP450 Inhibition 39

References 39

2 Screening for Kinase Inhibitors: From Biochemical to Cellular Assays 45

Jan Eickhoff and Axel Choidas

2.1 Introduction 45

2.1.1 Kinase Inhibitors for Dissection of Signaling Pathways 46

2.1.2 Cellular Kinase Assays for Drug Discovery Applications 46

2.2 Factors that Influence Cellular Efficacy of Kinase Inhibitors 472.2.1 Competition from ATP 47

2.2.2 Substrate Phosphorylation Levels 51

2.2.3 Ultrasensitivity of Kinase Signaling Cascades 51

2.2.4 Cell Permeability 52

2.2.5 Cellular Kinase Concentrations 53

2.2.6 Effects of Inhibitors Not Related to Substrate

Phosphorylation 54

2.3 Assays for Measurement of Cellular Kinase Activity 55

2.3.1 Antibody-Based Detection 56

2.3.2 High-Content Screening 59

2.3.3 Use of Genetically Engineered Cell Lines 60

2.3.4 Genetically Encoded Biosensors 61

3 Dissecting Phosphorylation Networks: The Use of Analogue-Sensitive

Kinases and More Specific Kinase Inhibitors as Tools 69

Matthias Rabiller, Jeffrey R Simard and Daniel Rauh

3.1 Introduction 69

3.2 Chemical Genetics 71

3.2.1 Engineering ASKA Ligand–Kinase Pairs 71

3.3 The Application of ASKA Technology in Molecular Biology 763.3.1 Identification of Kinase Substrates 76

VI Contents

3.3.3 Alternative Approaches to Specifically Targeting Kinases of Interest 783.4 Conclusions and Outlook 80

References 81

Part Two Medicinal Chemistry 85

4 Rational Drug Design of Kinase Inhibitors for Signal

Transduction Therapy 87

György Kéri, László O´´rfi, and Gábor Németh

4.1 The Concept of Rational Drug Design 88

4.2 3D Structure-Based Drug Design 89

4.3 Ligand-Based Drug Design 92

4.3.1 Active Analogue Approach 92

4.3.2 3D Quantitative Structure–Activity Relationships 92

4.4 Target Selection and Validation 93

4.5 Personalized Therapy with Kinase Inhibitors 96

4.5.1 Target Fishing: Kinase Inhibitor-Based Affinity Chromatography 974.6 The NCLTMTechnology and Extended Pharmacophore Modeling

5 Kinase Inhibitors in Signal Transduction Therapy 115

György Kéri, László O´´rfi, and Gábor Németh

5.1 VEGFR (Vascular Endothelial Growth Factor Receptor) 115

5.2 Flt3 (FMS-Like Tyrosine Kinase 3) 116

5.3 Bcr-Abl (Breakpoint Cluster Region–Abelson Murine Leukemia

Viral Oncogene Homologue) 118

5.4 EGFR (Epidermal Growth Factor Receptor) 118

5.5 IGFR (Insulin-Like Growth Factor Receptor) 120

5.6 FGFR (Fibroblast Growth Factor Receptor) 120

5.7 PDGFR (Platelet-Derived Growth Factor Receptor) 121

5.16 Auroras 127

5.17 Akt/PKB (Protein Kinase B) 129

5.18 Phosphoinositide 3-Kinases 129

5.19 Syk (Spleen Tyrosine Kinase) 130

5.20 JAK (Janus Kinase) 130

5.21 Kinase Inhibitors in Inflammation and Infectious Diseases 1315.21.1 Inflammation 131

6.4 Common Features of Type II Inhibitors 154

6.5 Design Strategies for Type II Inhibitors 155

7 From Discovery to Clinic: Aurora Kinase Inhibitors as Novel

Treatments for Cancer 195

Nicola Heron

7.1 Introduction 195

7.2 Biological Roles of the Aurora Kinases 195

7.3 Aurora Kinases and Cancer 196

7.4 In Vitro Phenotype of Aurora Kinase Inhibitors 197

7.5 Aurora Kinase Inhibitors 203

7.5.1 The Discovery of AZD1152 203

7.5.1.1 Anilinoquinazolines: ZM447439 203

7.5.1.2 Next-Generation Quinazolines: Heterocyclic Analogues 2047.5.1.3 Amino-Thiazolo and Pyrazolo Acetanilide Quinazolines 2087.5.2 MK-0457 (VX-680) 214

Indication Areas 229

8 Discovery and Design of Protein Kinase Inhibitors:

Targeting the Cell cycle in Oncology 231

Mokdad Mezna, George Kontopidis, and Campbell McInnes

8.1 Protein Kinase Inhibitors in Anticancer Drug

Development 231

8.2 Structure-Guided Design of Small-Molecule Inhibitors

of the Cyclin-Dependent Kinases 233

8.3 Catalytic Site Inhibitors 234

8.4 ATP Site Specificity 236

8.5 Alternate Strategies for Inhibiting CDKs 239

8.6 Cyclin Groove Inhibitors (CGI) 240

8.7 Inhibition of CDK–Cyclin Association 242

8.8 Recent Developments in the Discovery and the Development

of Aurora Kinase Inhibitors 242

8.9 Development of Aurora Kinase Inhibitors through Screening

and Structure-Guided Design 244

8.10 Aurora Kinase Inhibitors in Clinical Trials 248

8.11 Progress in the Identification of Potent and Selective Polo-Like

Kinase Inhibitors 250

8.12 Development of Small-Molecule Inhibitors of PLK1 Kinase

Activity 252

8.13 Discovery of Benzthiazole PLK1 Inhibitors 254

8.14 Recent Structural Studies of the Plk1 Kinase Domain 255

8.15 Additional Small-Molecule PLK1 Inhibitors Reported 256

8.16 The Polo-Box Domain 257

8.17 Future Developments 259

References 259

9 Medicinal Chemistry Approaches for the Inhibition

of the p38 MAPK Pathway 271

Stefan Laufer L, Simona Margutti, Dowinik Hauser

9.1 Introduction 271

9.2 p38 MAP Kinase Basics 271

9.3 p38 Activity and Inhibition 275

9.4 First-Generation Inhibitors 278

9.5 Pyridinyl-Imidazole Inhibitor: SB203580 278

9.6 N-Substituted Imidazole Inhibitors 282

9.7 N,N0-Diarylurea-Based Inhibitors: BIRB796 286

9.8 Structurally Diverse Clinical Candidates 288

9.9 Medicinal Chemistry Approach on VX-745-Like Compounds 297

9.10 Conclusion and Perspective for the Future 301

References 302

10 Cellular Protein Kinases as Antiviral Targets 305

Luis M Schang

10.1 Introduction 305

10.2 Antiviral Activities of the Pharmacological Cyclin-Dependent

Kinase Inhibitors 310

10.2.1 Relevant Properties of CDKs and PCIs 310

10.2.2 Antiviral Activities of PCIs 327

10.2.2.1 Antiviral Activities of PCIs against Herpesviruses 327

10.2.2.2 Antiviral Activities of PCIs against HIV 332

10.2.2.3 Antiviral Activities of PCIs against Other Viruses 335

10.2.3 PCIs Can be Used in Combination Therapies 336

10.2.4 PCIs Inhibit Viral Pathogenesis 337

10.3 Antiviral Activities of Inhibitors of Other Cellular Protein Kinases 33810.4 Conclusion 339

11.3 Drug Target Validation by Genetic Inactivation 351

11.4 STPK Mechanisms, Substrates, and Functions 352

List of Contributors

Protein Kinases as Drug Targets Edited by B Klebl, G Müller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

Division of Infectious Diseases

Vancouver, British Columbia

44227 DortmundGermanyJan EickhoffLead-Discovery Center GmbHEmil-Figge-Straße 76a

44227 DortmundGermanyDoris HafenbradlBioFocus AGGewerbestrasse 16

4123 AllschwilSwitzerlandNicola HeronDevices for DignitySheffield Teaching Hospitals NMSFoundation Trust

Royal Hallamshire HospitalGlossop Road Sheffield, S10 2YFUK

György KériVichem Chemie Research Ltd

Herman Ottó u 15

1022 BudapestHungary

Semmelweis University

Hungarian Academy of Sciences

Pathobiochemical Research Group

1022 BudapestHungaryLars NeumannProteros Biostructures

Am Klopferspitz 19

82152 MartinsriedGermany

László OrfiVichem Chemie Research Ltd.Herman Ottó u 15

1022 BudapestHungaryandSemmelweis UniversityDepartment of PharmaceuticalChemistry

Ho''gyes Endre u 9

1092 BudapestHungaryMatthias RabillerChemical Genomics Centre of theMax Planck Society

Otto-Hahn-Str 15

44227 DortmundGermanyDaniel RauhChemical Genomics Centre of theMax Planck Society

Otto-Hahn-Str 15

44227 DortmundGermany

XII List of Contributors

wwwwwww

Protein kinases are a huge group of evolutionary and structurally related enzymes,which by phosphorylation of certain amino acids, infirst-line serine/threonine andtyrosine, activate a multitude of proteins In this manner, they mediate signaltransduction in cell growth and differentiation The therapeutic potential of kinaseinhibitors results from the crucial role kinases (as well as some kinase mutants andhybrids resulting from chromosomal translocation) play in tumor progression and

in several other diseases With a group size of more than 500 individual members,the‘‘kinome,’’ that is, the sum of all kinase genes, constitutes about 2% of the humangenome Since the isolation of the first Ser/Thr-specific kinase in the muscle in

1959, it took another 20 years until tyrosine protein kinases were discovered andanother 20 years before thefirst 3D structure of a kinase was determined Startingwith the 3D structure of protein kinase A in 1991, many more structures wereelucidated in the meantime, in their active and inactive forms, without and withligands other than ATP These structures show not only the close structural relation-ship between all kinases but also the high complexity of their allosteric regulation.Today, the term‘‘protein kinase’’ retrieves almost 2000 entries from the Protein DataBank of 3D structures; most of these structures are protein–ligand complexes withabout 1000 different ligands All kinases show a highly conserved binding site forATP, and for this reason they were for long time considered nondruggable targets.This view was supported by the fact that the natural product staurosporine inhibits ahuge number of kinases in a nonspecific manner Still today, staurosporine is themost promiscuous kinase inhibitor, despite its large size However, with increase instructural knowledge, additional pockets were discovered in direct vicinity of thebinding motif of the adenine part of ATP (the‘‘hinge region’’) Step by step, thesepockets were explored and kinase inhibitors of higher specificity emerged Finally,the optimization of a PKC inhibitor to the bcr/abl tyrosine kinase inhibitor imatinib(Gleevec1, Novartis) marked a breakthrough in specific tumor therapy Althoughinitially designed for the treatment of chronic myelogenous leukemia, the drugturned out to be beneficial also for the treatment of gastrointestinal stromal tumors(GISTs) Several other kinase inhibitors followed, with significantly different speci-ficity profiles Even nonspecific inhibitors, such as sunitinib (Sutent1, Pfizer), are

Protein Kinases as Drug Targets Edited by B Klebl, G Müller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

valuable anticancer drugs, in this case for the therapy of advanced kidney cancer and

as the second-line treatment of GIST, in cases where Gleevec1 fails Due to themultitude of tumor forms, resulting from various mechanisms, research on kinaseinhibitors is now one of the hottest topics in pharmaceutical industry Resistance tosome kinase inhibitors forces the industry to also search for analogues with abroader spectrum of inhibitory activity As of today, nine small-molecule kinaseinhibitors for the treatment of oncological diseases have reached the market andmany more are in different phases of clinical development Even thefirst kinaseinhibitors targeted toward nononcological applications, such as inflammatory dis-ease states, have reached late-stage clinical development

We are very grateful to Bert Klebl, Gerhard Müller, and Michael Hamacher whoassembled a team of leading scientists for discussion of various topics of proteinkinase inhibitors, including assay development, hitfinding and profiling, medicinalchemistry, and application of kinase inhibitors to various therapeutic areas We arealso very grateful to all chapter authors who contributed their manuscripts on time

Of course, we appreciate the ongoing support of Frank Weinreich and NicolaOberbeckmann-Winter, Wiley-VCH, for our book series‘‘Methods and Principles

in Medicinal Chemistry’’ and their valuable collaboration in this project

Hugo Kubinyi, Weisenheim am SandGerd Folkers, Zürich

XVI Preface

A Personal Foreword

Kinase inhibitors are one of the fastest emerging fields in pharmaceuticalresearch, reigning at “No 2” in terms of overall spending for discovery anddevelopment of pharmaceuticals, when split according to target family classes Inour own professional histories, we still witnessed the dogma in pharmaceuticalindustry claiming that protein kinases are considered to be nondruggable targets.This dogma was all around during the 1990s of the last millennium Some braveindividuals nevertheless pursued the idea of identifying and developing kinaseinhibitors for biologically highly interesting targets, such as p38 kinases [3] andprotein kinase C (PKC) isoforms [4] Although these were groundbreaking efforts indrug discovery in those early days, p38 and PKC inhibitors have never really made itbeyond the status of tool compounds for biological research and chemical biology sofar At the end, a rather serendipitousfinding started the race toward the competitivegeneration of kinase inhibitors in oncology The introduction of a simple methylgroup into a diaminopyrimidine scaffold of a known protein kinase C inhibitor led tothe generation of a relatively specific Bcr-Abl inhibitor, called imatinib or GleevecÔ.The fusion protein Bcr-Abl has been known as the driving oncogene in chronicmyeloid leukemias (CML) with a mutation on the Philadelphia chromosome [5],which is mediated by the elevated Abl activity of the mutant Subsequently, imatinibhas shown convincing efficacy in treating CML patients [6] A new era started whenimatinib was launched in 2001 as the first specifically designed small-moleculekinase inhibitor The second beneficial serendipity during the generation anddevelopment of imatinib was understood only slowly Imatinib is not just a plainand simple ATP competitor as most kinase inhibitors were designed to be It binds tothe inactive form of Bcr-Abl and keeps the kinase in its inactive conformation [7].Today, this phenomenon is not only much better understood but also considered to be

an important design element when synthesizing novel kinase inhibitors Bothserendipitous features of imatinib, inhibition of Bcr-Abl and binding to the inactivekinase, paved the way for the establishment of its clinical efficacy However, thissuccess gave birth to another dogma that kinase inhibitors will be useful only fordeveloping anticancer therapies This second dogma was based on two assumptions:(1) since 2001, imatinib has been considered to be among the most selective kinaseinhibitors although it potently inhibits at least a dozen other protein kinases [8]; (2)

“ATP-competitive inhibitors are never going to be highly selective, because they bind

Protein Kinases as Drug Targets Edited by B Klebl, G Müller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

to the highly conserved active domain of kinases” Especially, the second point on thelack of selectivity was and still is highly speculative and led to the conclusion thatnonselective kinase inhibitors cannot be used as treatment options in indicationareas outside cancer because of their naturally invoked off-target mediated adverseeffects This assumption vice versa also led to the conclusion that nonselective butpotent kinase inhibitors will be effective cancer killing agents We would like tochallenge these hypotheses for a number of reasons:

. Kinase inhibitor technologies quickly advanced, especially compound designtechnologies, facilitated by the development of molecular modeling and X-rayresolutions of a large number of kinase inhibitor cocrystals (www.pdb.org/pdb/home/home.do)

. Exploitation of inhibitor binding to the inactive form of a kinase (type II inhibitors)has become an accepted design strategy and leads to a number of advantages inthe pharmacological development of kinase inhibitors

. Monoselective ATP competitors (type I inhibitors) have been generated, despitethe fact that they bind only to the active site of a kinase [9]

. A fair number of scaffolds are known to compete with ATP for binding to thekinase active site, allowing a quick screening effort to identify potential startingpoints for a subsequent optimization program on practically any kinase

. Allosteric kinase inhibitors have been reported to be an option for furtherdevelopment [10]

. The correlation between kinase homology and parallel structure–activity tionship tends to be understood much better [11]

rela-. Nowadays, kinase inhibitor design can be envisioned as the molecular game withLego bricks– and it really works

Over these past years, we have been able to generate highly specific kinaseinhibitors [12] Since kinases play a role not only in carcinogenesis but also in allsorts of physiologically relevant signaling pathways [13], we are convinced that bothoncology and any other medical indication might represent an important playgroundfor the application of selective and safe kinase inhibitors Future will demonstratethat kinase inhibitors are going to be applied to treat chronic conditions and not only

in life-threatening settings Therefore, we have chosen contributions to this book thatdescribe the generation and application of kinase inhibitors also outside theimportantfield of anticancer drug discovery Broadly specific kinase inhibitors, such

as sunitinib, will not have a chance for development for indications other than cancer.Instead, monoselective kinase inhibitors or multikinase inhibitors with a narrowprofile will turn out to be efficacious if the chosen target is critical enough in aparticular pathophysiological process It is more about the validation of the target(s)and the underlying target(s) rationale In that respect, it remains to be seen if p38aturns out to be a valid target for rheumatic arthritis or to be valid only for some distinctinflammatory diseases The odds are that p38a inhibitors will not reach the status of ageneral anti-inflammatory agent due to target-mediated toxicities [14] Although allp38a inhibitor research might then be considered a lost investment, it has none-theless contributed enormously to the general strategies in developing kinase

XVIII A Personal Foreword

highly selective kinase inhibitors, as well as their translation into pharmacologicallyactive substances (e.g., [15]) These efforts significantly helped to pave the way for thedevelopment of highly selective future kinase inhibitors for different kinase targetswithout target-mediated toxicities The world of protein kinases consists of morethan, 500 individual members, the human kinome [16], therapeutically relevantparasitic kinase targets even not considered Therefore, our prediction is that we willsee many more novel drug candidates and pharmaceutical products arising from thislarge and important family of enzymes.

This gives hope to millions of patients suffering not only from various cancers butalso from inflammatory, metabolic, and neurological disorders and infectiousdiseases, where a distinct kinase is out of control and must be tamed by a highlyspecific and potent kinase inhibitor But what makes a good inhibitor? Which stepshave to be taken for identifying a target and successfully making a drug with, ifpossible, no side effects? Which kinase inhibitors have been developed so far by usingwhich design strategy? Can we already define lessons learned?

Small molecules and their apparently endless modularity andflexibility to produceall necessary structures are the perfect source for developing kinase inhibitors.Libraries of thousands to millions of compounds can be screened easily in high-throughput screens (HTS) or even in silico Detected hits can be optimized step-by-step in iterative cycles toward highly potent and specific preclinical candidates andwell-tolerated drugs on the market (or toward specific probes and tools in basicresearch) Thus, this book is dedicated to small-molecules kinase inhibitors and theirvarious contributions to medical application

Literature is exploding in the kinase inhibitorfield, particularly when dealing withappropriate tools and design In order to give a comprehensive overview about thisspecial but diverse inhibitor species, this book covers the most important criteriafrom assay development to profiling and from medicinal chemistry-based optimi-zation to a potential application This book has been arranged in a logical order invarious parts to highlight

. hitfinding and profiling for protein kinases, describing the Dos and Don’ts whileidentifying and (cellular) profiling of active small-molecule kinase inhibitors

. chemical kinomics to detect phosphorylation networks

. medicinal chemistry, offering a detailed summary of existing kinase inhibitors,available technologies, and design principles that might be considered

. application to therapeutic indication areas, discussing in detail success stories andunmet needs in medical application including cancer, inflammatory diseases, andinfections

Thanks to the enthusiasm and the perseverance of the authors and the publisher ofthis book, wefinally made it Somehow, the genesis of this small compendium onkinase inhibitor research resembles the field of small-molecule-based kinase in-hibitors itself Some brave individuals quickly wrote and delivered their contributionswithin a short period of time, some others took more time to develop their chapters,andfinally, some opted out of the project and were replaced by others who maybe

considered newcomers to thefield This process seemed to reflect the development ofthe field of kinase inhibitor research over the past 15 years in nice analogy Onpurpose, we have selected contributions on kinase inhibitor drug discovery fromearly-stage discoveries since there have been a lot of writing and comprehensivereviews on successfully launched kinase inhibitors, such as Gleevec, Iressa, Tarceva,Sorafenib, Sutent, Dasatinib, Lapatinib, and others ([1, 2] and references therein).There is also a good body of literature available on kinase inhibitors in cancer drugdiscovery So, we rather focused both on the technologies for the discovery of kinaseinhibitors and on the optimization of these inhibitors, and we included novelpotential therapeutic applications of kinase inhibitors, especiallyfields outside thecancer research Therefore, this collection of articles is quite unique, albeit highlyrepresentative when it comes to the identification and generation of novel kinaseinhibitors with biological and pharmacological activity.

In the different chapters, experts in theirfield summarize the historical evolution,the trends, and a good part of their own experience gained while working in theirrespectivefields After reading the book, it will become clear how much promisesmall-molecule kinase inhibitors really hold, not only for the described therapeuticindications but also beyond, when obeying basic, intrinsic rules

We are convinced that small-molecule kinase inhibitors will become ever moreimportant in the years to come and are going to celebrate new success stories forresearch and patients– despite or even because of the current dramatic changes inpharmaceutical industry Enjoy reading!

References

1 Pytel, D., Sliwinski, T., Poplawski, T.,

Ferriola, D., and Majsterek, I (2009)

Tyrosine kinase blockers: new hope

for successful cancer therapy

Anti-Cancer Agents in Medicinal Chemistry,

9, 66–76.

2 Natoli, C., Perrucci, B., Perrotti, F., Falchi,

L., Iacobelli, S., and Consorzio

Interuniversitario Nazionale per

Bio-Oncologia (CINBO) (2010) Tyrosine

kinase inhibitors Current Cancer Drug

Targets, 10, 462–483.

3 Lee, J.C., Laydon, J.T., McDonnell, P.C.,

Gallagher, T.F., Kumar, S., Green, D.,

McNulty, D., Blumenthal, M.J., Heys, J.R.,

Landvatter, S.W., Strickler, J.E.,

McLaughlin, M.M., Siemens, I.R., Fisher,

S.M., Livi, G.P., White, J.R., Adams, J.L.,

and Young, P.R (1994) A protein kinase

involved in the regulation of in flammatory

cytokine biosynthesis Nature, 372,

739 –746.

4 Kawamoto, S and Hidaka, H., (1984) Isoquinolinesulfonyl)-2-

1-(5-methylpiperazine (H-7) is a selective inhibitor of protein kinase C in rabbit platelets Biochemical and Biophysical Research Communications, 125, 258–264.

5 Heisterkamp, N., Stam, K., Groffen, J., de Klein, A., and Grosveld, G (1985) Structural organization of the bcr gene and its role in the Ph’ translocation Nature, 315, 758–761.

6 Druker, B.J., Talpaz, M., Resta, D.J., Peng, B., Buchdunger, E., Ford, J.M., Lydon, N.B., Kantarjian, H., Capdeville, R., Ohno- Jones, S., and Sawyers, C.L (2001) Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia The New England Journal of Medicine, 344, 1031–1037.

7 Dietrich, J., Hulme, C., and Hurley, L.H (2010) The design, synthesis, and evaluation of 8 hybrid DFG-out allosteric

XX A Personal Foreword

the binding interactions of Gleevec,

Nexavar, and BIRB-796 Bioorganic and

Medicinal Chemistry, 18, 5738–5748.

8 Fabian, M.A., Biggs, W.H., 3rd, Treiber,

D.K., Atteridge, C.E., Azimioara, M.D.,

Benedetti, M.G., Carter, T.A., Ciceri, P.,

Edeen, P.T., Floyd, M., Ford, J.M., Galvin,

M., Gerlach, J.L., Grotzfeld, R.M.,

Herrgard, S., Insko, D.E., Insko, M.A.,

Lai, A.G., Lelias, J.M., Mehta, S.A.,

Milanov, Z.V., Velasco, A.M., Wodicka,

L.M., Patel, H.K., Zarrinkar, P.P., and

Lockhart, D.J (2005) A small

molecule-kinase interaction map for clinical molecule-kinase

inhibitors Nature Biotechnology, 23,

329–336.

9 Walburger, A., Koul, A., Ferrari, G.,

Nguyen, L., Prescianotto-Baschong, C.,

Huygen, K., Klebl, B., Thompson, C.,

Bacher, G., and Pieters, J (2004) Protein

kinase G from pathogenic mycobacteria

promotes survival within macrophages.

Science, 304, 1800–1804.

10 Lindsley, C.W., Zhao, Z., Leister, W.H.,

Robinson, R.G., Barnett, S.F.,

Defeo-Jones, D., Defeo-Jones, R.E., Hartman, G.D.,

Huff, J.R., Huber, H.E., and Duggan, M.E.

(2005) Allosteric Akt (PKB) inhibitors:

discovery and SAR of isozyme selective

1697, 243–257.

12 Wabnitz, P et al (2008) 4,6-Disubstituted aminopyrimidine derivatives as inhibitors

of protein kinases WO/2008/129080.

13 Cohen, P (2002) Protein kinases: the major drug targets of the twenty- first century? Nature Reviews Drug Discovery,

1, 309–315.

14 Hammaker, D and Firestein, G.S (2010)

“Go upstream, young man”: lessons learned from the p38 saga Annals of the Rheumatic Diseases, 69 (Suppl I), i77–i82.

15 Pargellis, C., Tong, L., Churchill, L., Cirillo, P.F., Gilmore, T., Graham, A.G., Grob, P.M., Hickey, E.R., Moss, N., Pav, S., and Regan, J (2002) Inhibition of p38 MAP kinase by utilizing a novel allosteric binding site Nature Structural Biology, 9,

268 –272.

16 Manning, G., Whyte, D.B., Martinez, R., Hunter, T., and Sudarsanam, S (2002) The protein kinase complement of the human genome Science, 298, 1912–1934.

Gerhard Müller, (Planegg)Michael Hamacher, (Dortmund)

Part One

Hit Finding and Profiling for Protein Kinases: Assay Development and Screening, Libraries

Protein Kinases as Drug Targets Edited by B Klebl, G M€uller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

identifi-of their inhibitory potential against the target kinase, the intellectual propertysituation around the small-molecule inhibitor class, the potential for further chem-ical optimization, and other criteria Once the optimization process has started,several parameters have to be considered and continuously monitored In most drugdiscovery programs, the optimization of the inhibitory activity of the small moleculesagainst the target kinase represents the center of activities While this parameterseems to be a straightforward and measurable parameter, there are a variety ofpossibilities of how an inhibitor might be binding to a protein kinase Potentially,these different binding modes can cause modifications of the kinetic bindingbehavior of the compound For the full assessment of an inhibitor, a detailed analysis

of binding modes and kinetic consequences is required

The optimization of a specific protein kinase inhibitor requires the constantassessment of a wide range of kinases to reduce the risk of possible side effects

It is therefore important to use comparable conditions in each protein kinase assay

A successful drug candidate also requires a balanced physicochemical profile thatdetermines the pharmacokinetic (PK) behavior of a small-molecule inhibitor inanimals In the past 10 years, a variety of in vitro assays have been developed andproven to be useful for the prediction of the PK parameters of an inhibitor

Here, we describe in detail a selection of in vitro assays that are critical for theoptimization process of small-molecule kinase inhibitors For an appropriate start, athorough optimization of the biochemical kinase assay is needed In addition, oneneeds to consider the mode of inhibition and should be prepared for unexpectedexceptions from the general rules Besides the rationalization of the measurement ofthe biochemical activity and the selectivity that influence the pharmacodynamicbehavior of a small-molecule inhibitor, we will give an in-depth overview of options

Protein Kinases as Drug Targets Edited by B Klebl, G M€uller, and M Hamacher

Copyright Ó 2011 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

for in vitro measurement of parameters that determine the pharmacokinetic behavior

of small-molecule inhibitors

1.2

Optimization of a Biochemical Kinase Assay

At thefirst glance, a biochemical kinase assay seems to be a very straightforwardenterprise with only very few parameters that can be modified: the concentration ofATP, substrate, and protein kinase, the composition of the reaction buffer, and thereaction time Nevertheless, a detailed optimization process is needed and severalconsiderations have to be taken into account In the following, we will give guidancefor the evaluation of each step of the assay optimization and how this information isused to achieve the goal: a biochemical screening assay that yields reliable andreproducible information about the inhibitory activity of a small molecule as one ofthe most critical parameters throughout an entire drug discovery project Theoptimization process for the AGC kinase Rock II is used as an example for describing

in detail the considerations and evaluation of the results

1.2.1

Step 1: Identification of a Substrate and Controlling of the Linearity between

Signal and Kinase Concentration

Finding a substrate that is recognized and efficiently phosphorylated by the kinase ofinterest is thefirst essential step in developing a biochemical kinase assay Equallyimportant is the identification of the kinase concentration to start the assay optimi-zation that guarantees a sufficiently high signal and at the same time good linearitybetween signal and kinase activity

When the concentration of kinase is low, the concentration changes in ATP, ADP,and phosphorylated substrate are very small after a given reaction time As aconsequence, the associated assay signal is low and inaccurate (Figure 1.1, region

of low assay signal) At moderate kinase concentrations, a sufficiently high assaysignal can be detected and at the same time linearity between signal and kinaseactivity is observed Thus, for example, a doubling of kinase activity is directlytranslated into the doubling of the assay signal (Figure 1.1, linear region) At highkinase concentrations, the bulk of ATP or substrate transforms into phosphorylatedsubstrate after the given reaction time and the linearity between kinase activity andassay signal is lost (Figure 1.1, nonlinear region) At very high kinase concentration,all ATP or substrate is converted into ADP and phosphorylated substrate and an evenhigher kinase concentration cannot increase the signal further (Figure 1.1, insen-sitive region) Thus, as soon as the assay is depleted of either ATP or substrate, theassay is blind to changes in kinase activity This situation is detrimental for tworeasons First, if the goal is to improve the assay conditions in order to increase thekinase activity, the assay cannot deliver an answer since changes in the kinase activity

do not translate into a change in signal A further increase in kinase activity cannot be

detected because even less kinase activity is sufficient to consume all ATP orsubstrate Second, in the opposite scenario the question is whether or not acompound reduces the activity of a kinase If the compound blocks 50% of thekinase activity, no change of the assay signal can be detected, as even 50% kinaseactivity is sufficient to consume all ATP or substrate within the given reaction time.Therefore, the activity of an inhibitor would be underestimated or the inhibitionwould not be detected at all.

In thefirst assay optimization step, both the substrate that yields in the highestkinase activity and the kinase concentration that combines sufficient assay signal andsignal linearity are identified Therefore, a series of potential substrates are tested in

Figure 1.1 Assay signal is plotted against

kinase activity The assay signal can be derived

from the concentration change of ATP, ADP, or

phosphorylated substrate after the given

reaction time Kinase activity is adjustable, for

example, by varying the kinase concentration or

reaction time The plot is separated in four

distinct regions (1) Region of low assay signal at

very little kinase activity In this region, the

kinase activity is so low that only very little

concentration changes in either educts or

product have occurred Usually, this region

yields signals that are too weak to generate

reliable data (2) Linear region At higher kinase

activities, the concentration changes are larger

and therefore the assay signals are generally

strong enough to give robust data quality This

region is the optimal to perform kinase assays

since a change in kinase activity is translated

linearly in a signal change (3) Nonlinear region.

At even higher kinase activities, most of the ATP and/or substrate is transformed into ADP and phosphorylated substrate, respectively In this region, high assay signals can be achieved, but kinase activity and signal do not depend linearly

on each other anymore (4) Insensitive region If all ATP or substrate is consumed after the investigated reaction time, the maximal possible change of signal has been reached Increased kinase activity cannot modulate the signal anymore because ATP and/or substrate has been completely consumed Thus, neither

an increase nor a decrease in kinase activity can

be detected In this region, the assay is insensitive to both an improvement of kinase activity (e.g., by optimizing the buffer components) and the inhibition of kinase activity (e.g., by the presence of a kinase inhibitor) and should therefore be avoided implicitly.

1.2 Optimization of a Biochemical Kinase Assayj5

the presence of increasing kinase concentrations In Figure 1.2, step 1 of an assayoptimization for the kinase Rock II is exemplified Seven potential Rock II substrateswere incubated with increasing concentrations of Rock II In addition, Rock II wasincubated in the absence of substrate After 1 h, the reaction was terminated and theamount of phosphorylated substrate quantified As shown in Figure 1.2, S6-derivedpeptide is phosphorylated most efficiently yielding the highest assay signal at lowkinase concentrations The generic peptide 3 was recognized with lowest efficiency.

In the absence of substrate, consistently no assay signal was detected at all In thepresence of the S6-derived peptide, the linear assay region is found between 0.5 and 7

nM Rock II Below 0.5 nM Rock II, only very small amounts of S6-derived peptide arephosphorylated and the assay signal is too small to be reliable At Rock II concen-tration above 7 nM, the majority of the S6-derived peptide is phosphorylated and theassay reaches its nonlinear region Thus, from step 1 the following information can

be taken into account for the next optimization step: (1) S6-derived peptide is selected

to be the substrate that is recognized most efficiently and (2) for the next optimizationstep, a Rock II concentration of 0.5 nM should be used to guarantee strict linearitybetween assay signal and kinase activity

1.2.2

Step 2: Assay Wall and Optimization of the Reaction Buffer

In the second assay optimization step, a reaction buffer is identified that enables thekinase to work at its maximal capacity In other words, the reaction buffer is

Figure 1.2 10 mM ATP and 12.5 mCi/ml

33 P-y-ATP are incubated with increasing

concentrations of Rock II and 10 mM of various

potential Rock II substrates in 40ml 20 mM Tris

pH 7.5, 10 mM MgCl 2 , 1 mM DTT for 1 h at

room temperature After 1 h, the reaction was

terminated by adding 10ml 0.5 M EDTA The

reaction mixtures are transferred to phosphor

cellulose filters and incubated with 60 ml 0.75%

H 3 PO 4 for 15 min Remaining 33 P-y-ATP was removed from the filters by three washes with

200 ml 0.75% H 3 PO 4 each The filter-associated substrate-incorporated 33 P was quantified by scintillation counting and plotted against Rock II concentration The error bars are given in standard deviations of duplicates.

optimized to obtain a sufficiently high assay signal at the lowest possible kinaseconcentration Beside the cost considerations, a low kinase concentration is essentialsince the kinase concentration limits the lowest IC50values that can be determined.The lowest IC50value that can be measured equals half the kinase concentration inthe assay (see Figure 1.3) [1, 2] For example, in an assay that uses 10 nM kinase, thelowest IC50value that can be measured is 5 nM Even if the real IC50value would be0.5 nM, the observed IC50revealed by the assay would be 5 nM This phenomenon iscalled “assay wall.” No IC50value can be measured below this wall defined by thekinase concentration This behavior is self-evident if one considers that half thekinase molecules have to be bound by an inhibitor to reduce the kinase activity by50%.

This assay wall can cause a severe impact on drug discovery projects In thebeginning of the project, usually the IC50values are high and far above the kinaseconcentration During the course of the project, the IC50values typically decreasewith every cycle of compound optimization At a certain level, the IC50values cannot

be decreased anymore A project course such as this is indicative of having reachedthe assay wall and it should be constantly monitored if the IC50values have reachedthe kinase concentration used in the assay

In the second optimization step, the composition of the reaction buffer isevaluated The potential addition of detergents, the optimal pH, and ion compositionare evaluated to ensure maximal kinase activity Since kinases are most dependent on

Mg2þand Mn2þ, these ions should be investigated in great detail In addition, the

Figure 1.3 The observed IC 50 (IC 50obs ) is given

by the sum of the real IC 50 and 0.5-fold the

kinase concentration Consequently, the

minimal IC 50 that can be measured equals 0.5

times the kinase concentration even if the real

IC 50 value is lower Assays requiring low kinase

concentrations – lower than the IC 50 values of

the examined inhibitors – yield IC 50 values that

are very close to the real IC 50 The ratio between the observed IC 50 and the real IC 50 (IC 50obs /IC 50 )

is close to 1 In contrast, assays that need high kinase concentrations – as high as or even higher than the IC 50 values of the inhibitors – will measure IC 50obs larger than the real IC 50 values The ratio IC 50obs /IC 50 is above 1 and increases with rising kinase concentrations.

1.2 Optimization of a Biochemical Kinase Assayj7

influence of NaCl and CaCl2is examined Figure 1.4a shows how different tions of MgCl2and MnCl2influence the Rock II activity A clear maximum in Rock IIactivity is detected at 10 mM MgCl2in the absence of MnCl2 Increasing or decreasingthe MgCl2reduces the Rock II activity Also, addition of MnCl2results in the loss ofRock II activity Similarly, the presence of NaCl (Figure 1.4b), CaCl2, and thephosphate inhibitors sodium-o-vanadate andb-glycerol phosphate reduces the assaysignal (Figure 1.4c) In contrast, the presence of 0.01% detergent such as Brij35,Tween 20, Triton X-100, or NP40 has no significant influence on Rock II performance(Figure 1.4c) Figure 1.4d shows the pH dependence of Rock II Various buffersystems were used to cover the pH range from 5.5 to 8.5 While Rock II is nearly

combina-Figure 1.4 1 mM ATP, 12.5 mCi/ml 33 P-Y-ATP,

10 mM S6-derived substrate peptide are

incubated with 0.5 nM Rock II for 1 h in 40 ml

(a) 20 mM Tris pH 7.5, 1 mM DTT, and the

indicated amounts of MgCl 2 and MnCl 2 ;

(b) 20 mM Tris pH 7.5, 10 mM MgCl 2 , 1 mM

DTT, and the indicated concentration of NaCl;

(c) 20 mM Tris pH 7.5, 10 mM MgCl 2 , 1 mM

DTT, and the indicated concentrations of

detergent, CaCl 2 , or phosphate inhibitors; or

(d) 10 mM MgCl 2 , 1 mM DTT, and 20 mM of the

indicated buffer at the given pH values After 1 h, the reaction was terminated by adding 10 ml 0.5 M EDTA The reaction mixtures were transferred to phosphor cellulose filters and incubated with 60 ml 0.75% H 3 PO 4 for 15 min Remaining33P-Y-ATP, was removed from the filters by three washes with 200 ml 0.75% H 3 PO 4

each The filter-associated incorporated 33 P was quantified by scintillation counting Error bars are given in standard deviations of duplicates.

substrate-inactive at acidic pH, maximal activity is reached at around pH 7.5 At pH valuesabove 7.5, Rock II loses activity In summary, on the basis of these results (Figure 1.4),

20 mM Mops pH 7.5, 10 mM MgCl2, 0.01% Triton X-100, 1 mM DTT were chosen asoptimal Rock II reaction buffer and were used in the following optimization steps.The MgCl2/MnCl2preferences of kinases can widely vary (Figure 1.5) While Rock

II prefers 10 mM MgCl2and the absence of MnCl2, the kinase PknG, for example, isalmost inactive under these conditions PknG shows maximal activity at 50 mMMnCl2in the absence of MgCl2 On the other hand, PDGFRb shows highest activity at

a combination of 10 mM MgCl2and 0.4 mM MnCl2 Thus, the optimization of theMgCl2and MnCl2concentration for each kinase usually allows to dramatically reducekinase concentrations in the assay

In addition to the high diversity in MgCl2/MnCl2preference, the tolerance forvarious detergents, CaCl2, and phosphatase inhibitors widely differs between kinases(Figure 1.6), so do pH optima (Figure 1.7) Thus, using a generic kinase reactionbuffer for all kinases would result in significantly higher kinase assay concentrationsand therefore unnecessarily high assay wall and high assay costs

Figure 1.5 Activity of six kinases in the presence of various combinations of MgCl 2 and MnCl 2

Step 3: The Michaelis–Menten Constant Kmand the ATP Concentration

After the identification of a good substrate and the optimal reaction buffer, the nextstep is the determination of the ATP concentration that should be used Since themajority of all kinase inhibitors are ATP competitive, the ATP concentrationdetermines the ability of an assay to identify the potential of a given small-moleculekinase inhibitor Generally, there are three options of choosing the ATPconcentration

Thefirst is to use a standard ATP concentration that is identical in all differentprotein kinase assays The main advantage of a standard ATP concentration is theease of the experimental procedure, especially if a large number of differentkinases are regularly screened The main disadvantage of a kinase assay with astandard ATP concentration, for example, 30mM or 100 mM, is that the IC50valuescannot be used to rank the potency of a given inhibitor between different kinases ForATP-competitive inhibitors, the dependencies between IC50and ATP concentrationsare described by the Cheng–Prusoff equation (Figure 1.8) [3] The IC50and the ATPconcentration are linearly connected The slope is given by the ratio betweenthe inhibitor constant Ki and the Michaelis–Menten constant for ATP Km They-intercept is defined by the Kivalue The Kivalue describes the affinity betweeninhibitor and kinase, while the Kmvalue approximates the affinity between ATP andkinase Since a given inhibitor has different Kivalues for every kinase and sinceevery kinase has a different Kmfor ATP, slope and y-intercept are different for eachkinase As a consequence, the lines of the Cheng–Prusoff plot intersect each other.Thus, for example, at an arbitrary assay ATP standard concentration of 10mM, asmaller IC50will be measured for a given inhibitor against a theoretical kinase 1(KmATP¼ 1.5 mM, Ki¼ 0.002 mM) than against kinase 2 (KmATP¼ 10 mM, Ki¼ 0.01mM) (Figure 1.8) At an arbitrary ATP standard concentration of 30 mM ATP, theopposite ranking would be observed At 30mM, the IC50of the given inhibitor would

be measured to be smaller for kinase 2 than for kinase 1 (Figure 1.8) While at oneATP concentration the given inhibitor seems to be more specific for kinase 1, itappears to be more specific for kinase 2 at another ATP concentration Thus,selectivity ranking based on assays using standard ATP concentrations is arbitraryand therefore should be avoided

Figure 1.7 Activity of six kinases at various pH values In order to cover a pH range from 5.5 to 8.5, different buffer systems were used.

The second option is to choose an ATP concentration at the cellular ATP level that isseen by the kinase of interest in the pathologic situation This approach requires exactknowledge of the ATP concentration in the relevant cellular location within thepatient Unfortunately, little is known about the exact cellular ATP concentrations.Even less is known about fluctuations of ATP concentrations between differentlocations within a cell, between cells in different tissues, between cancer andnoncancer cells, between cells in different stages of their development, and so on.

As a consequence, the ATP concentration that would be assumed to mimic thecellular ATP concentration in vivo does most likely not reflect the reality The chosenATP concentration is more likely to represent another form of an arbitrary ATPstandard concentration with the associated problem discussed in the beginning ofthis section

The third option for choosing the ATP concentration to measure IC50values forATP-competitive inhibitors is to use ATP at a concentration that equals its Kmvaluefor the individual kinase The Kmvalue is defined by the ATP concentration thatallows half maximal reaction velocity Thus, the ATP concentration would be differentfor every kinase assay In addition, the determination of the K value for every kinase

Figure 1.8 The Cheng –Prusoff equation

describes the dependencies between IC 50

value and ATP concentration for

ATP-competitive inhibitors The IC 50 values for

one inhibitor against three kinases are

calculated for an ATP concentration range

from 0 to 40 mM All three kinases have different

K m values and the K i values describing the

interaction between the theoretical inhibitor

and the three kinases vary from 0.002 to

0.02 mM At ATP concentrations below 12 mM,

kinase 1 has the lowest IC 50 and kinase 3 has the highest IC 50 At ATP concentrations between 12 and 17 mM, kinase 2 has the highest and kinase 1 the lowest IC 50 Between 17 and 24 mM, kinase 3 has the lowest

IC 50 Above 24mM, the IC 50 ranking is completely the opposite compared to the IC 50

ranking at ATP concentrations below 12 mM Thus, the selectivity ranking of the theoretical inhibitor depends on the selected ATP concentration.

1.2 Optimization of a Biochemical Kinase Assayj11

is required during the assay development, thereby complicating the assay ment and screening workflow On the other hand, IC50values determined at an ATPconcentration that represents its Kmvalue reflect 2
Kivalue (Figure 1.9) Thus, the

develop-IC50value is a direct measure of affinity between the inhibitor and the investigatedkinase As a consequence, the selectivity of an inhibitor against various kinases can beranked on the basis of its binding affinity for different kinases

Comparing the three options of choosing an ATP concentration, the Kmvalue forATP represents the most advantageous choice when more than one kinase are tested.Since this situation will be found in the majority of all discovery projects, the Km

determination has to be included as the essential step in the assay development Sincethe Kmvalue is defined by the ATP concentration that allows half maximal reactionvelocity, the assay signal in the presence of increasing ATP concentrations ismeasured andfitted to the Michaelis–Menten equation (Figure 1.10) [4] The ATP

Kmfor Rock II was determined to be 25mM In further assay optimization, 25 mMATP will be used

Determination of the ATP Kmof kinases is complicated by the fact that kinaseshave two substrates, ATP (the phosphate donor) and what we have called the substrate(the phosphate acceptor) so far Therefore, the ATP Kmdepends on the phosphateacceptor concentration Only if the concentration of the phosphate acceptor is at leastfive times above its own Kmvalue, the ATP Kmvalue is independent of the phosphateacceptor concentration and can be determined precisely At lower phosphate acceptorconcentrations instead of the real ATP Km, an apparent ATP Kmresults from an ATP

Kmdetermination experiment Under these circumstances, the measured ATP Kmisvalid only for the given phosphate acceptor (substrate) concentration

1.2.4

Step 4: Signal Linearity throughout the Reaction Time and Dependence

on the Kinase Concentration

After identifying an appropriate substrate, an optimal reaction buffer, and a ingful ATP concentration, the selection of assay components is now complete Step 4

mean-of the assay optimization controls signal linearity if the optimized reaction buffer andadjusted ATP concentration have shifted the range of kinase concentration thatguarantees signal linearity compared to assay optimization step 1 (Section 1.1.1) Thesignificance of signal linearity was discussed in Section 1.1.1 (Figure 1.1) In addition,

Figure 1.9 The Cheng–Prusoff equation describes the relation between IC 50 value and ATP concentration for ATP-competitive inhibitors If the ATP concentration equals the K m value for ATP, the IC 50 represents twice the K i value.

step 4 examines the dependency between signal linearity and reaction time Signallinearity has to be maintained regarding both kinase concentration and reaction time

to ensure that the measured IC50(IC50obs) reflects the real IC50(Figure 1.11) [5] Thehigher the kinase activity is, regardless whether due to a high kinase concentration ordue to a long reaction time, the more the substrate is converted (Figure 1.11a) At highsubstrate conversion, the measured IC50(IC50obs) is significantly larger than the real

IC50(Figure 1.11b) Thus, in order to measure meaningful IC50values, it is essential

to identify a combination of kinase concentration and reaction time that has asufficiently high assay signal and minimal substrate conversion

In order to identify the Rock II concentration and the Rock II reaction time thatguarantees signal linearity,five different Rock II concentrations were incubated at sixdifferent reaction times (Figure 1.12) From this experiment, the scientist can pick theoptimal combination between Rock II concentration and reaction time If short

Figure 1.10 The Michaelis –Menten equation

describes the dependencies between reaction

velocity and ATP concentration The K m value is

defined by the ATP concentration that results in

half maximal reaction velocity Increasing

concentrations of ATP were incubated for 1 h

with 0.5 nM Rock II, 2.5 mCi/ml 33

P-Y-ATP, and

10 mM S6-derived substrate peptide in 40 ml

20 mM Mops pH 7.5, 10 mM MgCl 2 , 0.01%

Triton X-100, 1 mM DTT The reaction was

terminated by adding 10 ml 0.5 M EDTA and

transferred to phosphor cellulose filters

followed by an incubation with 60 ml 0.75%

H 3 PO 4 for 15 min Remaining 33 P-Y-ATP was removed from the filters by three washes with

200 ml 0.75% H 3 PO 4 each The filter-associated substrate-incorporated 33 P was quantified by scintillation counting The assay signal was corrected for the dilution of33P-Y-ATP in nonradioactive ATP and plotted against the ATP concentration The data were fitted

to the given Michaelis–Menten equation, thereby determining the Rock II ATP K m

to be 25 mM.

1.2 Optimization of a Biochemical Kinase Assayj13

reaction times are needed, for example, 10 nM Rock II and a reaction time of 60 mincan be selected If low Rock II concentrations are required, for example, to shift theassay wall to lower IC50values (see Section 1.1.2), a Rock II concentration of 2.5 nMand a reaction time of 240 min could be chosen without changing the intensity of the

Figure 1.11 The given equation describes the

dependency between measured IC 50 (IC 50obs )

value and substrate conversion By increasing

kinase concentration at a constant reaction

time, or by increasing the reaction time at a

constant kinase concentration, more and more

substrate will be converted At very high kinase

concentrations or at very long reaction times, 100% of the substrate is converted (a) Using the given equation, the ratio between observed

substrate conversion (b) At substrate conversions above 70%, the observed IC 50obs

becomes significantly higher than the real IC 50

Figure 1.12 Different concentrations of Rock II

were incubated for the indicated reaction time

with 25 mM ATP, 2.5 mCi/ml 33 P-Y-ATP, and

10 mM S6-derived substrate peptide in 40 ml

20 mM Mops pH 7.5, 10 mM MgCl 2 , 0.01%

Triton X-100, 1 mM DTT Reactions were

terminated by adding 10ml 0.5 M EDTA The

reaction mixtures were transferred to phosphor

cellulose filters and incubated with 60 ml 0.75%

H 3 PO 4 for 15 min Remaining 33 P-Y-ATP was removed from the filters by three washes with

200 ml 0.75% H 3 PO 4 each The filter-associated substrate-incorporated 33 P was quantified by scintillation counting Raw data were plotted either against Rock II concentration (a) or reaction time (b).

assay signal (Figure 1.12) Here, we have chosen 2.5 nM Rock II and a reaction time of

60 min for further optimization

1.2.5

Step 5: Assay Validation by Measurement of the IC50of Reference Inhibitors

In the last step of the optimization procedure, the assay is validated by themeasurement of the IC50 values of reference inhibitors Besides controllingthe IC50values themselves, it has to be ensured that the Hill coefficients, reflectingthe slope of the IC50curves, are close to a value of 1 (ideally between 0.5 and 1.8) Hillcoefficients deviating significantly from 1 indicate that something unexpected isoccurring in the assay that in most cases will obscure the measured IC50values.Phenomena such as negative or positive cooperativity of kinase, a contamination with

a second kinase that has a different IC50value for the inhibitor from the target kinase,and the presence of different variants of the target kinase (various phosphorylationstates, dimers, splice variants, etc.) would influence the Hill coefficient

In order to validate the optimized Rock II assay, the IC50 values of thereference inhibitors H-89 and Y-27632 were measured The Rock II activity wasquantified in increasing concentrations of the reference inhibitors (Figure 1.13) ForH-89, an IC50value of 0.18mM was determined that is in line with the published value

of 0.27mM [6] The IC50value for Y-27632 was measured to be 0.22mM Literaturereports a Kivalue of 0.14mM for Y-27632 against Rock II [7] Since Y-27632 is an ATP-competitive inhibitor and an ATP concentration was used that equals the ATP Km, themeasured IC50value of 0.22mM translates into a Kivalue of 0.11mM (see Section1.1.3, Figure 1.9) Thus, the value for Y-27632 was also measured correctly by thedeveloped assay In addition, the Hill coefficients were calculated to be 1.0 and 0.8,respectively In conclusion, both the IC50values and the Hill coefficients prove thatthe optimized Rock II assay is able to measure Rock II IC50 values in a reliablemanner Thus, the Rock II assay could be released for a potential Rock II drugdiscovery project

1.3

Measuring the Binding Affinity and Residence Time of Unusual Kinase Inhibitors

Besides the classical binding mode, where small-molecule inhibitors bind into theATP binding cleft forming H-bonds only with the hinge region, there are severalknown exceptions Among these kinase inhibitors are examples such as imatinib (1),sorafenib (2), lapatinib (3), and BIRB 796 (4, see below)

1.3 Measuring the Binding Affinity and Residence Time of Unusual Kinase Inhibitorsj15

These specific inhibitors of protein kinases take advantage of the conformationaldifferences between active and inactive forms of kinases [8] The main determinant ofthese forms is the so-called activation loop that can undergo large conformationalchanges.

Quite often, but not always, these nonclassical inhibitors also show unusualbinding characteristics that require special methods for evaluation The classicalway of IC50determination, which has been described in detail in Section 1.1, does nottake into account the fact that inhibitors might also show nonclassical enzymekinetics Therefore, the activity of these inhibitors might be largely underestimatedduring the course of an optimization program or might be entirely overlooked in ahigh-throughput screening campaign

Several methods have been used in the past to evaluate novel protein kinaseinhibitors To realize the full potential of these nonclassical protein kinase inhibitors,

Figure 1.13 2.5 nM Rock II was incubated for

1 h with 25mM ATP, 2.5 mCi/ml 33 P-Y-ATP, and

10 mM S6-derived substrate peptide in 40 ml

20 mM Mops pH 7.5, 10 mM MgCl 2 , 0.01%

Triton X-100, 1 mM DTT in the presence of

the indicated concentrations of the reference

inhibitor H-89 or Y-27632 Maximal Rock II

activity was measured in the absence of

inhibitor Background signal was determined in

the absence of Rock II Reactions were

terminated by adding 10ml 0.5M EDTA The

reaction mixtures were transferred to phosphor

cellulose filters and incubated with 60 ml 0.75%

H 3 PO 4 for 15 min Remaining 33 P-Y-ATP was

removed from the filters by three washes with

200 ml 0.75% H 3 PO 4 each The filter-associated substrate-incorporated 33 P was quantified by scintillation counting Rock II activity was expressed by calculating the ratio between the background-corrected assay signals

in the absence and presence of the indicated inhibitor concentrations.

The Rock II activity was plotted against the inhibitor concentration and fitted to the given equation For H-89 (a) and Y-27632 (b), IC 50 values of 0.18 and 0.22 mM and Hill coefficients of 1.0 and 0.8 were calculated, respectively.

generic and efficient tools are needed that apply the strengths of diversity-orientedchemical synthesis to the identification and optimization of lead compounds fordisease-associated protein kinase targets.

Inactive conformation wasfirst observed crystallographically for the unliganded

IR kinase [9], but it was not until the structures of Abl in complex with imatiniband analogues were solved that it became clear that this conformation could beexploited by inhibitors [10] (see also Chapter 6) The so-called DFG-out confor-mation creates an additional hydrophobic pocket adjacent to the ATP pocket that

is frequently referred to as the “allosteric site” [11] or the “deep pocket.” Because theamino acids surrounding this pocket are less conserved relative to those in the ATPbinding pocket, it has been proposed that it may be easier to achieve kinaseselectivity with “deep pocket” binding inhibitors [12] compared to the classicalinhibitors

Very often these nonclassical inhibitors show remarkable cellular activity thatcould result from binding to the inactive conformation of kinases that may be moreaccessible in the cellular environment

It has been discussed that the departure from the kinase–ligand equilibriuminteraction comprises an important determinant of the in vivo effectiveness of small-molecule drugs Copeland et al propose that the most crucial factor for sustaineddrug efficacy in vivo is not the apparent affinity of the drug to its target per se, but ratherthe residence time of the drug molecule on its molecular target [13]

The term residence time in thefield of drug target interaction is defined as theperiod for which the receptor is occupied by a ligand A long dissociation half-life of

an intracellular receptor would be expected to translate into sustained efficacy in cellculture after removal of the ligand supply from the extracellular medium Forthe in vivo situation, the duration of efficacy of a ligand is no longer well described

by the in vitro measured dissociation constant, but rather depends on the rate ofreceptor–ligand association (kon) and, most critically, on the dissociation rate con-stant, or off rate (koff), of the receptor–ligand complex The off rate can be simplytranslated into a dissociative half-life for the receptor–ligand complex, and this half-life is a direct measure of the residence time (see Figure 1.14) As demonstrated in asimulation (Figure 1.15), the residence time becomes the driving parameter for theefficacy and the pharmcodynamic behavior of a drug candidate in vivo, especiallywhen the plasma half-life is short Over time, cKIT with low binding affinity but longcompound residence time is more efficiently inhibited than DDR1 with its highbinding affinity but short residence time

Both the improvement of the metabolic stability and the residence time can beused to optimize the efficacy of an inhibitor compound As shown in the simulation,the affinity data alone would be a misleading parameter In addition to the efficacy, the

in vivo selectivity is affected both by the affinity and by the residence time(Figure 1.15)

Various experimental approaches have been considered to analyze new proteinkinase inhibitors In recent years, the pharmaceutical industry has identified theneed both for the discovery of inhibitors with novel modes of inhibition and for thedetailed characterization of their lead compounds in preparation for clinical assess-

1.3 Measuring the Binding Affinity and Residence Time of Unusual Kinase Inhibitorsj17

ment A selection of methods will be discussed and the advantages and disadvantageswill be compared.

1.3.1

Washout Experiments

A recent example of the effects of ligand dissociation half-life comes from the work ofWood et al on inhibitors of epidermal growth factor receptor (EGFR) tyrosine kinaseactivity [14] The Kdvalues, off rates, and the recovery of cellular proliferation afterwashout for three similarly potent inhibitors of EGFR were measured: GW572016(lapatinib (3)), ZD-1839 (Iressa), and OSI-774 (Tarceva) These compounds bind tothe ATP binding pocket of the kinase, but display maximum affinity for differentconformation states of the enzyme For ZD-1839 and OSI-774, a rapid recovery of the

Figure 1.14 Association and dissociation of a receptor –ligand complex and calculations of the parameters residence time, k off , k on , and K d

Figure 1.15 Simulation of in vivo inhibition

of four targets of the bRaf inhibitor sorafenib:

impact of residence time and K d on

pharmacodynamics for three hypothetical

compound plasma half-lives (C ¼ 10 mM).

Especially for drugs with short or medium plasma half-lives, in vivo target inhibition is determined by binding kinetics rather than by binding affinity.