CHI's Kinase

Session Chairs:
Dr. Paul Lyne, AstraZeneca R&D
Dr. Mark Namchuk, Vertex Pharmaceuticals, Inc.
Dr. Chris Vlahos, Lilly Research Laboratories

Additional Speakers:
Dr. Ansu Bagchi, Merck & Co., Inc.
Dr. Barry A. Bunin, Libraria, Inc.
Dr . Sean Buchanan, Str uctural GenomiX, Inc. (SGX)
Dr. Laura DeForge, Genentech, Inc.
Ms. Debra Gallant, Millennium Pharmaceuticals, Inc.
Dr. Arthur Kluge, GPC Biotech, Inc.
Dr. Campbell McInnes, Cyclacel, Ltd.
Dr. K. Parang, University of Rhode Island
Dr. Christopher Pargellis, Boehringer Ingelheim Pharmaceuticals
Dr. David J. Parry-Smith, Purely Proteins Ltd.
Dr. Sabita Sankar, Celgene Corporation
Dr. Eric Springman, Locus Pharmaceuticals, inc.
Dr. Andrew Whitney, Cellular Genomics, Inc.
Dr. Patrick P. Zarrinkar, Ambit Biosciences Corporation

Functional Understanding of Kinases
Structural Classification of Kinases
Functional Classification and Comparisons
Functional Validation of Kinases

Structure-Based Design Strategies
A Dual Substrate Inhibitor Design
Co-crystallization based Design
Virtual Screening and Design
Design based on Ligand-Induced Disordering

Screening Strategies
Kinase Families and Screening Strategies
Targeted Libraries and SAR
Kinase Assay Development
Selectivity Panels for Hit to Lead
Kinase Inhibitor Cross-Reactivity

Indication-Specific Kinase Applications
Inflammation and Cell Death
Cardiovascular
Oncology

Novel Binding Sites for Kinases
Targeting p38 MAP Kinase
Alternative Design Sites

1:00 Chair's Opening Remarks
Dr. Paul Lyne, Principal Scientist, Chemistry Department, Cancer Discovery, Astrazeneca R&D

1:10 Classification of Non-alpha Protein Kinases
Dr. Ansu Bagchi, Sr. Research Associate, Applied Computer Science and Mathematics, Merck & Co., Inc.
Classification of a set of 296 non-alpha protein kinases from the Steven Hanks' protein kinase data set based on sequence features will be presented. The classifications are based on similarity matrices, which were generated by: (a) the conventional method employing BLAST2 to generate alignments and then BLOSUM62 and GONNET matrices to "score" associations; and (b) a novel method using sequence-based descriptors that do not require alignment and involves LaSSI (Latent Semantic Structure Indexing) to calculate similarities. The clustering analysis shows general agreement with functional classes as defined by Steven Hanks. In addition, our classification provides putative assignments of other protein kinase (OPK) groups to these functional classes.

1:40 Kinase Proteomics: Knowledge led Informatics Solutions
Dr. David J. Parry-Smith, Chief Information Officer, Purely Proteins Ltd.
The classification of kinases has long been an area of uncertainty, largely because of the different methods used for analyzing the family. A recent analysis of the Kinome (Manning et al, 2002) includes biochemically confirmed functions for certain kinases that are difficult to classify by sequence alone. We have integrated all the available information in the kinase area within TargetBASE®Plus, an extensible, relational database schema that underpins the company's protein purification R&D, as well as providing infrastructure support for our layered proteomics and chemoinformatics databases. A side-by-side comparison of the Manning analysis with the TargetBASE®Plus classification will be presented.

2:10 Chemical Genetic Approaches to Functional Validation of Kinases
Dr. Andrew Whitney, Director, Bioinformatics, Cellular Genomics, Inc.
Analog Sensitive Kinase Alleles (ASKAs) are fully functional genetically engineered kinases that can be potently and selectively inhibited by proprietary small molecule compounds. The ASKA technology has comprehensive applications to the drug discovery process, including: ASKA-based kinase substrate identification for pathway mapping and discovery of novel drug targets; ASKA mice for pharmacologically relevant kinase target validation and therapeutic index determination; and ASKA-based cellular assays for lead identification. Examples of how CGI uses the ASKA technology for target identification, validation and kinase drug discovery will be presented.

3:20 Designing Bi-substrate Analog Inhibitors for Protein Kinases
Dr. K. Parang, Department of Biomedical Sciences, University of Rhode Island
Bi-substrate analog inhibitors form a special group of protein kinase inhibitors that mimic two natural substrates/ligands and that simultaneously associate with two regions of given kinases. This presentation will discuss the design of bi-substrate analog inhibitors that could provide for the enhancement of specificity and potency in protein kinase inhibition.

3:50 Structure-Based Drug Discovery for Protein Kinase Targets in Human Cancers
Dr . Sean Buchanan, Senior Dir ector , Bioinfor matics , Structural GenomiX, Inc.
(SGX)A high-throughput gene-to-structure platform that integrates mechanistic enzymology with medicinal and computational chemistry to accelerate drug discovery will be presented. Focused combinatorial libraries, iterative co-crystal structures and enzyme assays of protein kinases implicated in human diseases are yielding quantitative insights into enzyme inhibition. By tackling the entire human kinome at the level of molecular biology, we have rapidly identified tractable representatives from target kinase subfamilies that are amenable to purification for enzyme activity assays and co-crystallization trials. This Target Pipeline feeds a Lead Discovery operation currently examining 12 well-validated targets in detail and two targets are the focus of our in-house Lead Optimization efforts. Examples will be presented from each of these three programs, illustrating how the process contributes to improved lead/drug candidate quality. Particular emphasis will be given to the interplay between co-crystallography, kinase inhibition studies, and medicinal and computational chemistry.

4:20 Structure-based Approaches for the Identification and Optimization of Leads for Protein Kinases
Dr. Paul Lyne
There is a wealth of structural information available for protein kinase-inhibitor complexes, and this information can be capitalized on to use structure-based computational approaches in virtual screening and in design approaches to identify lead series and to optimize the potency of initial hits in a rational approach. Several examples of these approaches will be presented.

4:50 Discovery and Structure Determination of a Novel Family of Cyclin Dependent Kinase Inhibitors: Structural Basis for Ligand-Induced Disordering of the Activation Loop
Dr. Campbell McInnes, Head, Structure Based Design Group, Cyclacel, Ltd.
A novel family of CDK2 inhibitors has been discovered using LIDAEUS, a novel high throughput data-mining program. This series has been optimised for activity in vitro and cellular assays using structure based design methods. X-ray structures of this series indicate that two hydrogen bonding modes are observed for the ligands depending on the substitution and demonstrate a relationship between potency and disordering of the kinase activation loop (T-loop). Subsequent analysis of the CDK2 inhibitor complex structures from the protein data bank reveals that the majority of ligands adopt one of these two binding modes and that this is also correlated with disorder of the T-loop and level of inhibition. This novel observation has significant implications in the design of CDK2 inhibitors and we will discuss how this can be exploited for the further optimisation of ATP antagonists to be used in proliferative disorders.

8:30 Chair's Remarks
Dr. Mark Namchuk, Head, High Throughput Screening, Vertex Pharmaceuticals, Inc.

8:35 Parallel Drug Discovery within the Kinase Gene Family
Dr. Mark Namchuk
There is a growing trend to direct the early portion of the drug discovery process toward target gene families, rather than therapeutic areas. The benefits of this approach are numerous, but most dramatically exemplified in the increased ability to transfer information between targets of the same class and reuse chemistry effort from one program to facilitate execution on subsequent targets. The Vertex kinase effort has been structured to allow easy flow of information and compounds between different kinase targets. The impact on the drug discovery process for the kinase gene family will be discussed.

9:05 Using Kinase SAR Knowledgebases and eScreens™ to Help Scientists Rapidly Scaffold-Jump to Novel Chemotypes
Dr. Barry A. Bunin, President & CSO, Libraria, Inc.
Libraria has captured the majority of the kinase SAR from the scientific and patent literature and built kinase-specific eScreens™ to help scientists identify novel kinase inhibitors. Large "Actual-Virtual" libraries of kinase inhibitors were enumerated, archived, and prioritized based on over 14,000 validated Markush literature reaction transforms in Lucia™ (Libraria's Unique Chemically Intelligent Archive). Case studies will be presented where novel chemotypes were identified against Abl Kinase which is a commercially important target at this time due to the clinical resistance to Gleevec™ observed in patients with chronic myeloid leukemia (CML).

9:35 Development of Kinase Inhibitor Assays
Dr. Laura DeForge, Scientist, Assay & Automation Technology, Genentech, Inc.
Many different approaches can be taken to developing assays for screening of kinase inhibitors, and issues such as the size of the compound library to be screened, the required throughput, and cost must be taken into consideration. We have developed several assays for tyrosine and serine/threonine kinases using both ELISA and homogeneous solution phase technologies. In addition, we have developed a high throughput cell based assay for evaluation of inhibitors of a receptor tyrosine kinase. The advantages, disadvantages, and comparative performance of various approaches will be discussed.

10:40 The Utilization of a Kinase Selectivity Panel in the Steering of Hit to Lead Programs
Ms. Debra Gallant , Research Investigator, High Throughput Biochemistry, Millennium Pharmaceuticals, Inc.
The use of a kinase selectivity panel to direct kinase inhibitor series selection for medicinal chemistry will be presented. The panel is invoked immediately after confirming HTS hits and used throughout lead optimization. The early availability of potency and selectivity data alongside purity and structure information has enabled us to intelligently prioritize chemistry exploration of the large numbers of hits from primary screens. The general implementation and use of the kinase selectivity panel, as well as specific examples of successes, will be discussed.

11:10 A New Approach To Specificity Profiling of Kinase Inhibitors
Dr. Patrick P. Zarrinkar, Head, Molecular Biology and Technology Development, Ambit Biosciences Corp.
Ambit has developed a new, quantitative method to experimentally determine the specificity of kinase inhibitors. Because most kinase inhibitors target the ATP site, specificity is a central issue. Binding specificity and affinity are not readily predicted based on available sequence and structural information, and current profiling methods are limited by the difficulty of developing large numbers of kinase activity assays. Ambit's new approach circumvents this limitation and allows testing of compounds against many kinases from different families to identify off-target interactions. Primary and secondary targets are identified and binding affinities quantitatively determined without the need to purify individual proteins or develop activity assays. A significant fraction of the human kinome is accessed, and selective compounds are readily distinguished from promiscuous compounds. The technology has been used successfully to reveal a significant number of unexpected targets for multiple classes of kinase inhibitors, including several currently in Phase I, II and Phase III clinical trials. In a number of cases, these newly identified interactions suggest novel uses for the compounds.

12:00 Luncheon Technology Workshop	Sponsored By
Cell-free expression of the MAP Kinase family: Implications for Drug Discovery Presented by Dr. George Martin, Principal Scientist, The Roche Protein Expression Group

1:10 Chair's Remarks
Dr. Chris Vlahos, Head, Atherosclerosis Vessel Wall Inflammation Drug Discovery Team, Cardiovascular Research, Lilly Research Laboratories

1:15 Validation of a Novel IKK-related Kinase which acts via a non-NFKappa-Beta Pathway
Dr. Sabita Sankar, Senior Scientist, Cell Signaling and Target Discovery Department, Signal Research Division, Celgene Corporation
We have identified a TNF-alpha inducible kinase IKKi that plays an important role in modulating a subset of NF-kB regulated genes in response to mediators of inflammation and cell death. These effects appear to be independent of IKK1/2 activation, IkB-alpha degradation and translocation of NFkB to the nucleus. Hence the data suggests that IKKi signals via a parallel but NF-kB independent pathway.

1:45 PI 3-Kinase Isoforms and Their Role in Cardiac Function and Disease
Dr. Chris Vlahos
Phosphoinositide 3-kinases (PI3K) are a conserved family of lipid kinases that phosphorylate the 3'-OH group of the inositol ring of membrane-bound phosphoinositides. Two isoforms appear to play distinct roles in the regulation of cardiac viability and function. Growth and survival of cardiac cells are regulated by one isoform, while the other plays a role in calcium handling and cardiac contraction. PI3K-mediated pathways are suitable targets for a PI3K isoform-targeted strategy, in which activators and inhibitors of specific components are required to selectively manipulate different cellular responses.

2:15 GPC286199: A Novel Cyclin-Dependent Kinase Inhibitor with Therapeutic Potential in Oncology
Dr. Arthur Kluge, Vice President, Drug Discovery, GPC Biotech, Inc.
GPC286199 was identified as an advanced lead in our target-based drug discovery program for identifying inhibitors of the cell cycle as potential oncology drugs. Optimization of SAR within a family of lead inhibitors gave a series of very potent inhibitors of cyclin dependent kinases with good growth inhibitory effects on cancer cells consistent with inhibition of the proposed cell cycle targets. Further optimization based in vivo studies led to the identification of GPC286199, a compound with outstanding efficacy in xenograft models.

3:00 The Design and Development of the Partially Allosteric p38 MAP Kinase Inhibitor, BIRB 796, for the Treatment of Inflammatory Diseases
Dr. Christopher Pargellis, Senior Principle Scientist, Department of Biology, Boehringer Ingelheim Pharmaceuticals

3:30 De Novo Design of Non-ATP Competitive P38 Modulators Using a Computational Fragment Based Approach
Dr. Eric Springman, Director, Department of Biochemistry, Locus Pharmaceuticals, Inc.
Two novel classes of non-ATP competitive P38 modulators have been designed de novo. One class targets a novel computationally predicted small molecule-binding site and preferentially inhibits phosphorylation of P38. This binding site is different from both the ATP binding site and the recently reported allosteric site, which is the target of the second class of de novo designed small molecules.

Sponsoring Publications:

Web Partner: