Accelerating Drug Development with Innovative Discovery Platforms

In recent years, the drug discovery and development process has been revolutionized, with the pharmaceutical industry shifting the focus of drug discovery to targeted therapeutics designed to attack genetic alterations that drive the onset and progression of disease. Whereas the industry had been largely chemically driven, with massive capacity to create new chemical entities, this revolution has led to a change that integrates biology discovery and validation efforts in the drug development process.Specifically, evaluation of genes involved in driving disease progression and biomarkers to indentify target patient populations, as well as response to new therapies, is becoming routine.

The pharma industry's focus is now to design and synthesize new chemicals to take advantage of this new biological understanding. Today, drug discovery begins with the identification and validation of molecular targets that serve as the basis of new chemical synthesis and screening. However, the genomic revolution has generated new potential targets at a pace that exceeds the industry's ability to validate and exploit them such that the industry is now rich with potential targets but challenged by the ability to identify drugs of sufficient potency and specificity to fully exploit the biology.

New approaches to drug discovery have been developed and implemented to face these challenges and to best exploit the tremendous amount of information being generated by these genomic approaches. The application and utility of several innovative drug discovery platforms designed to shorten the time from target validation to identification of drug leads will be described in this symposium. The evidence suggests that these new approaches promise to transform the time and cost of identifying and developing new drugs.

The BPDG at the New York Academy of Sciences represents a diverse group of scientists and others with an interest in biochemistry, molecular biology, biomedical research, and related areas. Members are from pharmaceutical and biotechnology companies, and university and medical center research facilities across the Eastern United States. The group also serves as the Biochemical Topical Group for the American Chemical Society's New York Section. The purpose of the BPDG is to bring together diverse institutions and communities, industrial and academic, to share new and relevant information at the frontiers of research and development.

Organizers: David Bearss, SuperGen, Inc. and George Zavoico, Westport Capital Markets, LLC

Speakers: David Bearss, SuperGen, Inc.; Randall Peterson, Harvard Medical School & Massachusetts General Hospital; Thomas Chan, ArQule, Inc.; Richard Friesner, Columbia University

Agenda

1:00 - 1:05  PM
Introduction
George Zavoica, Westport Capital Markets, LLC

1:10 - 1:50 PM
Rapid Identification and Optimization of Novel Drug Candidates using CLIMB^TM
David Bearss, SuperGen, Inc.

1:55 - 2:35 PM
Applying High-throughput Chemical Screens in Intact Zebrafish for Small Molecule Drug Discovery Randall Peterson, Harvard Medical School & Massachusetts General Hospital

2:40 - 3:10 PM
Coffee Break

3:15 - 3:55 PM
Accelerating the Drug Discovery and Development Process using Computer-aided Structure-based Drug Design
Thomas Chan, ArQule, Inc.

4:00 - 4:40 PM
Small Molecule Docking Approaches to Drug Discovery
Richard Friesner, Columbia University

4:45 - 5:15 PM
Panel Discussion:
Expectations and Reality: How Much More Can Drug Development Be Accelerated?

Abstracts

Rapid Identification and Optimization of Novel Drug Candidates using CLIMB^TM
David Bearss, SuperGen, Inc.

Over the past two decades there has been an exponential increase in the knowledge of the basis for many human diseases which has created unprecedented opportunities for the discovery of novel therapeutics for the treatment of unmet medical needs. Yet, companies involved in drug discovery and development face the most challenging time in history, with new drug development costs increasing dramatically and new drug approvals at or near record low levels. There is an urgent need to improve the success rate of drug development by incorporation of new and innovative methods to discover and develop new medicines. Although computational tools have become increasingly important in drug discovery, they still remain largely in a supporting role to high-throughput screening approaches in most companies' drug discovery processes. High-throughput small molecule screening has allowed for as many as several million compounds to be tested, identifying the few that interact potently with a disease-related protein target.

We have developed a computationally-driven drug discovery process called CLIMBTM which we use to identify new drug leads by screening as few as several hundred computationally selected compounds. Using CLIMBTM we identify leads more efficiently than high-throughput screening and we also find leads with better potential drug-like properties. CLIMBTM screening is based the use of structural biological information of a potential target protein and the prediction of small molecule ligand conformations and orientations within a targeted binding site. Using models for guidance in drug design instead of brute force or chemical intuition we have been able to develop a process which we are utilizing to discover and develop new potential treatments for life-threatening diseases. Examples of how CLIMBTM was utilized to discover inhibitors of Pim-1 and Jak2 protein targets will be presented.

Applying High-throughput Chemical Screens in Intact Zebrafish for Small Molecule Drug Discovery
Randall Peterson, Harvard Medical School & Massachusetts General Hospital

The divide between in vitro and in vivo research remains a major impediment to drug discovery. The in vitro assays that dominate modern discovery efforts are often poor predictors of a compound's ultimate efficacy and safety. As a result, compounds with promising in vitro activity often fail upon transition to the in vivo setting.

The unique attributes of the zebrafish enable in vivo experimentation to be conducted inexpensively and in high throughput. As a result, zebrafish allow in vivo studies to be incorporated into most steps in the drug discovery process, including high-throughput screening. We have used screening in zebrafish to discover compounds with in vivo activity against pathways of diseases ranging from congenital birth defects to anemia and psychiatric diseases. Because the discovery process occurs in vivo, effective compounds can be discovered even in the absence of a validated target. And, because the in vivo screens eliminate compounds with significant in vivo toxicity, the compounds discovered may be less likely to produce unexpected toxicities at later stages of development. Thus, the zebrafish and other emerging model organisms may accelerate drug discovery by reducing dependence on target validation and by providing in vivo data on compound efficacy and safety at the earliest stages of discovery.

Accelerating the Drug Discovery and Development Process using Computer-aided Structure-based Drug Design
Thomas Chan, ArQule, Inc.

Protein kinases are important signal transducers in normal and malignant cells. Collectively they represent almost a quarter of the "druggable" human genome. These enzymes catalyze reversible phosphorylation of protein substrates in cells of different tissues to regulate many functions in the human body. The typical Type 1 kinase inhibitors that bind to the highly conserved ATP recognition region of these enzymes tend to have poor selectivity, which can translate to substantial toxic side effects. Off target toxic effects of multi-kinase inhibitors tend to limit their utility in oncology patients. New approaches in structure-based design may yield more selective and less toxic kinase inhibitors that have the potential to treat diseases outside the field of oncology.

Small Molecule Docking Approaches to Drug Discovery
Richard Friesner, Columbia University

Docking algorithms have reached the point where the docking of a flexible ligand into a rigid receptor with the correct structure is successful for a very high percentage of protein-ligand complexes. However, two fundamental problems remain in robustly predicting protein-ligand interactions: (1) accounting for induced fit effects of the protein (2) reliably predicting binding affinity from the structure of the complex. These two problems are in fact related, because modeling protein reorganization energy (which arises from induced fit effects) is a major unsolved problem in the development of scoring function for binding affinity prediction.

We will discuss our latest developments of the Glide XP (extra precision) scoring function, in which we have now incorporated significant rank ordering capabilities and greatly enhanced abilities to identify known active compounds from a random database. Several qualitatively new terms have been added to account for strain energy of the complex and to improve modeling of water displacement by ligands, the main source of binding affinity in most complexes. Improvements in sampling algorithms, and refinement of parameters, combined with these new terms, leads to major advances in performance as calibrated using ~20 sets of PDB derived data with 10 or more ligands associated with a given receptor. The great majority of the model is global in nature, but some target specific parameter optimization is required to properly incorporate reorganization and strain effects.