Morning Session: Success Stories

Session Chairs: Martin Drysdale, Ph.D., Director of Chemistry and Structural Science, Vernalis and Dan Erlanson, Associate Director Medicinal Chemistry, Sunesis

9:00 Session Intro

Dan Erlanson, Associate Director Medicinal Chemistry, Sunesis

9:10 Integrating Fragments into Structure Based Medicinal Chemistry -Discovery and Development of Hsp90 Inhibitors for the Treatment of Cancer

Martin Drysdale, Ph.D., Director of Chemistry and Structural Science, Vernalis

The use of weak binding "fragments" of molecules is now recognised as an efficient and robust method of hit identification in the drug discovery process. The use and integration of fragment hits into successful lead optimization is the critical determinant of whether this technology will become accepted as a significant tool in drug discovery. We have developed a collection of methods for fragment based drug discovery which we call SeeDs (Structural exploitation of experimental Drug startpoints) which have become integral in our structure based drug discovery efforts. Heat shock protein (Hsp) 90 is a molecular chaperone that is responsible for the correct folding of a large number of proteins allowing them to achieve their functional conformation. Client proteins of Hsp90 include many key overexpressed or mutated oncogenes which are known to be critical for the transformed phenotype observed in tumors. 17-AAG and 17-DMAG are Hsp90 inhibitors derived from the prototypical ansamycin natural product inhibitor geldanamycin, which have shown pre-clinical efficacy in mouse xenograft models, and are now in phase I and II clinical trials. I will discuss our experience in developing and applying the SeeDs technology to target Hsp90, where optimization has been fundamentally informed by the fragments uncovered in our hit identification phase and has led to the identification of clinical and pre-clinical development candidates.

9:50 Fragment-to-Clinic: Astex Experience

David Rees, Ph.D., Vice President of Medicinal Chemistry, Astex Therapeutics

This presentation will outline examples of fragment-based projects that have led to candidates being selected for preclinical development within oncology. For example AT7519 and AT9283 have been progressed into clinical trials in patients.

10:30 Break, Exhibits

11:00 Successes and Surprises with Fragments

Wolfgang Jahnke, Ph.D., Senior Research Investigator, Novartis Institutes for BioMedical Research

Fragment-based screening (FBS) is a lead finding technology that is complementary to high-throughput screening (HTS), but leads from FBS often comprise more information than leads from HTS. By probing the protein surface with different fragments, hot spots on the protein surface are identified, and preferential types of interactions are determined separately for each sub-site. In addition, allosteric pockets can be identified which allow the design of novel types of inhibitors. This talk will summarize some of our experiences in fragment-based screening using NMR spectroscopy and X-ray crystallography.

11:40 From a Scaffold to the Clinic

Rick Artis, Ph.D., Vice President, Lead Generation, Plexxikon Inc.

Plexxikon's Scaffold-based Drug Discovery approach has yielded two clinical programs to date, in diabetes (Phase 2) and oncology (Phase 1). In each case, the time from initiation to first-in-human study was less than two years. In part, success of these efforts are derived from the ability to pick starting points with tractable enabling chemistry in the structural context of a given binding site. With this structural analysis in place, the subsequent design strategy focused on generating compact molecules with high atomic economy. The quality of compounds generated from this process has also allowed for early pharmacokinetic screening and resulted in generally favorable properties when compounds are introduced to in vivo profiling. Examples from clinical programs and some early discovery efforts will be discussed.

12:30-1:30 Lunch

1:30-3:00 Posters, Exhibits, FBLD Scavenger Hunt

3:00-4:00 Free Time

Afternoon/Evening Session: Chemical Biology and Drug Discovery

Session Chairs: Duncan McRee, Ph.D., President, ActiveSight and Maurizio Pellechia, Ph.D., Professor, Burnham Institute for Medical Research

4:00 Session Intro

Duncan McRee, Ph.D., President, ActiveSight

4:10 Fragment-based Approaches to Targeting Enzymes from Parasitic Organisms

Gabriele Varani, Ph.D., Professor, University of Washington

Parasitic diseases cause large number of deaths and morbidity in developing countries. Current treatment options are missing altogether for many parasitic diseases, while long standing chemotherapeutic agents are either too expensive or the parasite population has become resistant to treatment. Fragment-based ligand screening and development provides an ideal strategy for working within these therapeutic conditions. In cases where no inhibitor is known, fragment-based screening provides an entry point for new chemical entities to be discovered. When homologous enzymes have been well-characterized and previously targeted by traditional medicinal chemistry, a structure-based approach can help to isolate chemical moieties that are critical for inhbition, while exploring rational alternatives. Even for enzymes that have previously been targeted, such an approach can yield novel scaffolds with more favorable characteristics, in regards to cost and resistance mechanisms, that can generate new classes of inhibitors. We are applying these methods to critical enzymes from Vibrio cholerae, Trypanosoma brucei, and Plasmodium falciparum, as well as a series of homologous human targets, and will report on the results of the application of this approach to this area of pharmacology.

4:40 NMR Spectroscopy in Hit Identification and Optimization Process and Reverse Chemical Genetics

Maurizio Pellecchia, Ph.D., Professor, Burnham Institute for Medical Research

Recently we reported on an NMR-based approach, named SAR by ILOEs (structure activity relationships by interligand nuclear Overhauser effect), that makes use of protein mediated ligand-ligand NOEs (ILOEs) in complex mixtures to identify initial weak hits that are converted by synthetic chemistry approaches into bi-dentate compounds with higher affinity. In addition, we also reported on combining this approach with pharmacophore based searches of possible linked molecules from large data bases of commercially available compounds (Pharmacophore by ILOEs). Combined with functional studies using the resulting ligands, these methods represent ideal approaches to hit identification and to reverse chemical-genetics studies. Reverse chemical-genetics entails selecting a protein of interest, screening for a ligand for the protein, and finally determine the eventual phenotypic alterations that the ligand induces in a cellular context. Likewise, these methods enable the identification of protein’s hot spots by using small molecules, regardless of the knowledge of the function of the protein, and the development of a specific assay. Subsequently, such small organic molecules can be used in cellular assays to investigate the possible role of the target. In particular, the approaches were applied to the identification of the first inhibitor of the pro-apoptotic protein Bid and to find highly selective protein kinase inhibitors . We also will report on the use of paramagnetic probes for the design of potent and selective bi-dentate compounds against kinases and phosphatases.

5:10 Break

5:20 BMSC Fragment Cocktails: Development and Experience in MSGPP

Christophe Verlinde, Ph.D., Associate Professor, University of Washington

From a collection of nearly 700 small compounds, called fragments, carefully selected from the MDL ACD database, we created 68 cocktails of 10 compounds that are shape-wise diverse. We have explored the utility of these cocktails for initiating lead discovery in structure-based drug design by soaking numerous protein crystals obtained by the MSGPP (Medical Structural Genomics of Pathogenic Protozoa) consortium. We will report on the fragment selection and cocktail design procedures, and give examples of the successes obtained. The BMSC Fragment Cocktail recipes are available free of charge upon request under the provisions of a university agreement, visit faculty.washington.edu/verlinde for more information.

5:50 Dynamics in Drug Design: Application of Freeze-Frame Click Chemistry in the Design of Acetylcholinesterase and Nicotinic Receptor Selective Ligands

Palmer Taylor, Ph.D., Professor, University of California, San Diego

The recent elucidation of the crystal structures of the acetylcholine binding protein (AChBP), a surrogate for the extracellular domain of the nicotinic acetylcholine receptor, have provided critical templates for analyzing interaction of ligands with the extracellular domain of the receptor molecule. Since multiple binding proteins with distinct, but homologous, sequences exist and templates resembling the receptor can be constructed through mutagenesis, considerable flexibility is accorded to developing templates. Moreover, the sites of potential interaction extend beyond the classical agonist-antagonist site to non-competitive sites at the non-α-subunit interfaces and sites within the extracellular channel vestibule. AChBP, being a soluble protein, also provides a structure for examining solution dynamics and possible changes in conformation. We have employed H/D exchange and various steady-state and relaxation spectroscopic methods to examine conformation and segmental motion of domains of the molecule. Solution dynamics become important, since ligand binding is associated with conformational changes in the C loop, and some of these changes correlate with agonist-antagonist behavior of the ligands. Using AChBP as a template, we have employed freeze-frame, click chemistry for the in situ synthesis of novel ligands that we hope will confer selectivity with respect to receptor subtype and unique agonist-antagonist behaviors. A large series of lead compounds generated by conventional synthesis and through the click-chemistry approach have now been generated by the Sharpless group in our collaborative studies. The compounds have been screened for affinity at the agonist-antagonist site, and a few have been taken as leads for crystallographic studies. Several of the dissociation constants approach sub-nanomolar values, and the lead complexes reveal the orientation of the acetylenic and azide precursors that form the triazole in situ.

6:20 Targeting drug-resistant variants of HIV reverse transcriptase: high resolution crystal engineering and fragment screening

Eddy Arnold, Ph.D., Professor of Chemistry and Chemical Biology, CABM & Rutgers University

HIV-1 reverse transcriptase (RT) is a key target for anti-AIDS drug treatment.. Emergence of drug-resistant mutations complicates therapy and development of inhibitors effective against a wide range of HIV variants is highly desirable. We participated in the design and discovery of TMC125 and TMC278, non-nucleoside RT inhibitors which have shown promise in clinical trials for treating HIV-1 infections that are resistant to existing treatments. We have proposed that inhibitor flexibility can be useful in evading drug-resistant mutations and recent structural studies of HIV RT complexed with TMC278 have confirmed these concepts (Das, Bauman, et al., PNAS, in press). Engineered versions of HIV-1 RT have yielded crystals diffracting to 1.8 Å resolution and we are pursuing fragment cocktail screening by crystallography. Progress in identifying fragments bound at multiple sites on HIV-1 RT will be described.

7:00-9:00 Buffet Dinner, Posters, Exhibits

End of Day One

Tuesday February 19^th

Morning Session: Methods & Emerging Technologies, Part I

Session Chairs: Wolfgang Jahnke, Ph.D., Senior Research Investigator, Novartis Institutes for BioMedical Research, and Vicki Nienaber, Ph.D., Chief Scientific Officer, ActiveSight

8:00 Session Introduction

Vicki Nienaber, Ph.D., Chief Scientific Officer, ActiveSight

8:10 Fragment-based Screening in Real-time

David Myszka, Director, Center for Biomolecular Interaction Analysis, University of Utah

Hear ye! Hear ye! Today's biosensor technology (like Biacore) is capable of detecting the binding of the smallest of molecules (less than 150 Da) interacting with the largest of proteins (>300 kDa). This opens up the possibility of applying biosensors as a primary tool for fragment screening. Believe it or not, biosensors can be used to tell not only which compounds interact with a target, but they root out the badly behaved compounds as well. What's that? You want more? Well, biosensors can also provide detailed information about the affinity of the interaction. Let's see structural methods do that! But unfortunately, biosensor assays can't show you where a compound binds. So, we need to keep structural methods around for a while. However, since biosensors can screen a library of about ~1000 compounds in 3 days, it makes sense to put one upstream of structural analysis, thereby saving you time and money. And it whitens teeth. Step right up and join the biosensor-based fragment-screening drug-discovery bandwagon. No shoving . . . there's plenty of room for everyone.

8:50 Fragment screening of drug targets by SPR and subsequent X-ray structural analysis

Michael Hennig, Ph.D., Vice Director, Molecular Structure Research, F. Hoffmann - La Roche

The ability of rapid gain in potency of compounds by structure based drug design together with the high sensitivity of biophysical assays like surface plasmon resonance spectroscopy (SPR) enable the use of fragment molecules as starting points for drug discovery efforts. Examples of the application of fragment screening will be presented in order to illustrate the Roche process for this route for hit and lead generation. There will be emphasis on the interplay of biophysical methods for the screening of the fragment library, the subsequent X-ray structural analysis of the hits after affinity ranking and analysis by computational chemistry methods to facilitate the discovery of novel chemical entities. For example, binding of a tyramine molecule to the Alzheimer target β-secretase was detected and the binding mode analyzed by X-ray complex structure analysis. It binds to the S1 pocket and a follow up chemistry program was initiated to explore this finding. The benefits and challenges of the fragment screening approach will be discussed.

9:30 Application of Enthalpy Arrays to Fragment-based Screening

Michael Recht, Ph.D., Member of Research Staff II, Palo Alto Research Center

Enthalpy arrays are arrays of nanocalorimeters that enable measurements of the thermodynamics of molecular interactions using small sample volumes and short measurement times. We are investigating the use of enthalpy arrays for screening FBS hits and ranking their ligand efficiencies. The measurements do not require immobilization of the reactants, an attractive feature compared with surface-based methods such as SPR. In principle, isothermal titration calorimetry (ITC) can be used to characterize the thermodynamics of fragment binding to targets, but its use in FBS is severely hampered by the need for large samples (≈1.5 mL), long measurement times, and high fragment solubility in the injectant. Our enthalpy array technology addresses this problem by enabling measurements with 250 nL drops that only take a few minutes.

Fragments identified in FBS typically exhibit low binding affinity (0.1 to 1 mM), and it is beneficial to identify the fragments with a high ligand efficiency to take through lead optimization. We have used enthalpy arrays to determine Kd > 0.1 mM, as shown by titration of BaCl₂ with 18-crown-6, demonstrating that the approach has adequate sensitivity for characterizing FBS interactions. In this talk we will discuss the application of enthalpy arrays to confirm and rank hits identified in an X-ray crystallographic FBS.

10:10 Break, Exhibits

10:40 HT Fragment Screening using Fluorescence Correlation Spectroscopy

Thomas Hesterkamp, Ph.D., Vice President, Fragment Based Drug Discovery, Evotec AG

The most widely described fragment screening methods are NMR and X-ray crystallography but the use of high concentration biochemical assays has emerged as a viable alternative. In particular the use of high concentration screening of fragments has been successfully demonstrated using single-molecule Fluorescence Correlation Spectroscopy detection techniques to ensure both high data precision and reproducibility and pharmacological sensitivity. In Evotec's fragment screening process bioassay hits are confirmed by NMR (or vice versa), followed by X-ray and structure driven optimization of the weak binders.

11:20 In vivo Incorporation of NMR-active Unnatural Amino Acids

Bernhard Geierstanger, Ph.D., Group Leader, Protein Sciences, Genomics Institute of the Novartis Research Foundation

Recently developed methods for genetically encoding the incorporation of unnatural amino acid in E. coli, yeast, and mammalian cells can be used to introduce NMR-active labels site-specifically into proteins (Deiters, A., Geierstanger, B. H., Schultz, P. G. (2005) ChemBioChem 6, 55-8). The process requires encoding the incorporation site by an amber non-sense codon, TAG, in the gene of interest, and an orthogonal tRNA/tRNA synthetase pair evolved specifically for each unnatural amino acid. We will report on our progress to incorporate fluorinated, ¹⁵N and ¹³C labeled unnatural amino acids into proteins for the characterization of small molecule-protein interactions.

12:00-1:00 Lunch

1:00-3:00 Posters, Exhibits, FBLD Scavenger Hunt

Afternoon Session: Methods & Emerging Technologies, Part II

3:00 From Fragments to Leads: Seed, Assemble and Grow

Gerhard Klebe, Prof. Dr., Philipps-Universität

Virtual screening has been developed as an alternative for lead discovery. It departs from the 3D structure of a target protein and tries to predict putative ligands by docking and molecular similarity analyses. Large collections of candidate molecules are screened which exhibit the size of usual pharma molecules. Successive hierarchical filtering strategies are applied to reduce the initial sample of several million entries to some hundred prospective hits. The properties of the target protein binding pocket are considered in terms of hot spots appropriate to accommodate functional groups of putative ligands. Once such interaction sites are detected, a protein-based pharmacophore is defined to anchor a ligand in the binding pocket. This concept has been transferred to a screening sample of fragments with MW >250 Da. Special care is needed to define a relevant pharmacophore model and to tailor the applied docking and scoring tools to fragment-based screening. Once discovered as a seed, a fragment is further assembled and grown to molecules of usual drug size. To obtain a more detailed insight into the different contributions gained by adding molecular portions to a given lead fragment, systematic studies by stepwise optimizing ligands in the thrombin binding pocket were performed. This process of ligand growing has been studied by recording isothermal titration calorimetry, crystal structure analysis and computational simulations.

3:40 Integrating Biophysical Methods into Fragment-Based Screening

Glyn Williams, Ph.D., Director of Biophysics, Astex Therapeutics

There are a number of established and emerging biophysical methods which are capable of detecting fragment which bind to protein targets with low affinities. The relative strengths and weaknesses of these methods are now better understood and there are a number of ways in which they can be integrated into an effective and efficient screening process. This talk will focus on factors which influence the selection of techniques and the ways in which they can be integrated, with particular reference to X-ray crystallography, NMR and thermal methods.

4:20 Pulling it all together: Faster Better Lead Discovery through Integration of Technologies

Vicki Nienaber, Ph.D., Chief Scientific Officer, ActiveSight

"We have the capabilities to make higher quality clinical candidates . . . Better than they were before.   Better.   Stronger.   Faster." Since fragment-based lead discovery (FBLD) was first introduced in the mid-1990s, significant progress has been made in the field. Technologies have progressed both in the detection of fragment binding and in their optimization to lead compounds. Biophysical techniques such as SPR and calorimetry are becoming central assets to an FBLD program. Furthermore, the line between a traditional FBLD process and optimization of these early leads into clinical candidates is becoming blurred. Technologies such as parallel synthesis, SPR and calorimetry may be used hand-in-hand with high-resolution crystal structures to rapidly advance leads through optimization into clinical candidates. Fragment co-crystal structures may be used to scaffold hop towards the goal of expanding or generating IP or improving the drug-like properties of a late-stage lead series. ActiveSight's LENS™ technology pulls together multiple biophysical techniques with proprietary software tools to focus synthetic efforts in the transformation of fragment hits into drugs

5:00 Break, exhibits

5:30 Scientific Advisory Board Round Table

7:00-9:00 Dinner Reception, Poster & FBLD Scavenger Hunt Winners Announced

End of Day Two

Wednesday February 20^th

Morning Session: Lessons Learned

Session Chairs: Roderick Hubbard, Professor, University of York and Senior Fellow, Vernalis and Harren Jhoti, CEO & Founder, Astex Therapeutics

8:00 A Decade of Thinking Small: Highlights and Lessons Learned from Building a Fragment Discovery Company

Harren Jhoti, CEO & Founder, Astex Therapeutics

8:20 Fragmentology

Roderick Hubbard, Professor, University of York and Senior Fellow, Vernalis

Over the past six years, we have developed, refined and applied fragment-based methods in a number of structure-based discovery projects. In this presentation, I will review some of the phenomena observed. These include:

• hit rates for different target classes and relationship to library design methods
• conformational change induced by fragments
• virtual screening and docking of fragments
• similarities and differences in fragment binding to members of the same protein family
• comparison of NMR and SPR for identifying fragments

From these experiences, I will review what we believe are the next steps for developing and applying the methods. These include improved characterisation of library content, design of novel fragments, decision support for fragment evolution and targeting protein-protein interactions.

9:05 How then Shall we Screen?

Phil Hajduk, Ph.D., Project Leader, Abbott Laboratories

Fragment-based screening is only one of many tools that can be brought to bear on the discovery of new drug leads. Especially at large pharmaceutical companies, the proper balance of resources between conventional HTS, fragment-screening, and fast-follower strategies is not always clear. This presentation will discuss the integration of fragment-based screening into the array of screening technologies available at Abbott, the optimal application of FBDD, and opportunities for continued contributions to development candidates.

9:50 Break, Exhibits

10:05 Integrating Fragment-Based Methods into Drug Discovery: Some Examples

Siegfried Reich, Ph.D., Vice President of Drug Discovery, SGX Pharmaceuticals

Fragment-Based Methods continue to evolve as a powerful tool in the identification and optimization of useful leads for Drug Discovery. Integrated into an overall drug discovery process which includes iterative structure-based design and a focus on retaining intrinsic favorable drug-like properties of the original hit, viable development candidates can be discovered. We have optimized our FAST approach to allow rapid crystallographic screening of our fragment library against oncology targets of high therapeutic value. Several examples of the application of FAST toward the identification of advanced leads targeting BCR-Abl, HCV Pol, and MET will be described.

10:50 Fragment-based discovery for challenging targets

James Wells, Ph.D., Professor in Pharmaceutical Chemistry and Molecular & Cellular Pharmacology, UCSF

Drug discovery has been challenged when targets do not present small molecule precedents for binding such as protein-protein interfaces and novel allosteric regulatory sites in enzymes. Fragment discovery approaches have been showing promise in addressing of these targets. I’ll review some of this work and the general outlook for attacking what has been considered “undruggable” territory.

11:35 Closing Plenary Presentation

One Discovery Shoe Size Paradigm Does Not Fit All and Could Well Leave Us Barefoot

Christopher Lipinski, Ph.D., Scientific Advisor, Melior Discovery

For orally active drugs we live in a target poor environment. There is nothing wrong with using high-throughput screening (HTS) to find a superbly selective drug for a new target. It is just that we do not know which the really good targets are. So we search for complements to the one shoe size fits all mentality of the single ligand for a novel target discovered by HTS paradigm. All the alternatives have a common feature; they expand opportunity space on the biology side, the chemistry side or the computational side. On the biology side we have phenotypic screening; letting the experimental observation rather than biology/genomics opinion point us to the correct target. We have non reductionist approaches such as systems biology. We have the most relevant experiment; keen clinical observation and we even have the revolutionary concept that maybe it is not the target per se that is critical but the physiological set point/state of a cell. On the chemistry side we have fragment screening with advantages of chemistry space coverage and better interrogation of dubious targets. We have the search for multiple activities in a single compound or a mixture of compounds and we have the rediscovery that medicinal chemistry pattern recognition and expertise often works better than automated technology. On the computational side we have new software to assist in fragment generation and reassembly and we have software and databases in the commercial and increasingly in the public domain to predict biological activity from chemistry structure alone. Finally, we are trying approaches to fix a major disconnect. The innovative biology resides in academia but the drug development capability resides in industry.

End of Conference