Research and development of new drugs is a protracted and expensive endeavor. The typical drug discovery process evolves from lead identification for a disease target to lead optimization, in vitro/in vivo evaluation, preclinical and clinical testing, and FDA approval. Several studies estimate the average cost of development of a single, approved new drug in excess of $800 million. A large portion of this expense is attributed to abandoned lead candidates that are obtained initially from screening vast, unfocussed libraries of compounds for activity against the target but which fail for various reasons along the pipeline. Thus, having a well stocked pipeline of drug candidates is integral to guaranteeing success in any drug discovery endeavor. Recent vast strides in unraveling the human proteome and interactome have allowed mapping of the complex network of protein-protein interactions (PPIs). PPIs are involved in all cellular processes, including growth, maintenance, and death, and it is documented that the dysregulation of certain PPIs underlies the pathology of various diseases. The identification of modulators of these PPIs, and consequently protein function, and the process of transforming these into high-content lead series are key activities in modern drug discovery. We have proposed a rapid, knowledge-based methodology to develop several high-affinity peptide drug candidates able to modulate key protein interactions, and having high potential for success as drug leads. Peptides are high value targets in drug discovery and peptide-based leads derived from PPI sites currently comprise >50% of pharmaceutical pipelines. In Phase I studies, Lynntech and the Garner group at VBI have demonstrated unequivocally that it is feasible to obtain biologically active peptide ligands to target proteins, from their primary sequence alone, using our unique approach that combines systems biology and bioinformatics tools with an advanced, high-density peptide microarray for high-throughput screening of candidate ligands. Several high- affinity peptide ligands (of nM range affinity) were obtained from array based affinity maturation, and a subset of these ligands displayed a clear proclivity to modulate ESRRG interactions. Our Phase I efforts also have resulted in the successful development of a web-accessible discovery engine and database which enables user-assisted pseudo-automation of the various steps involved in the approach, thereby vastly expediting the process. Further enhancements are required to make this a potent drug discovery engine for lead generation, lead optimization, and lead explosion. The Phase II proposal will not only develop graphic-user interface tools for data analysis and informed down-selection of ligands in a pipeline but also elucidate selection rules that will inform the quickest way to obtain a high value lead drug candidate from protein sequence.
The high cost of modern drugs in the US is related directly to the various wasted efforts spent chasing compounds selected from random, unfocussed libraries with only a promise of 'drug-like'properties. Our methodology starts from an 'informed'and defined starting point and takes rapid, meaningful strides across the mountainous 'fitness'landscape of peptide ligand space to efficiently reach the summit of ligand fitness. The surfeit of high-value hit-to-lead candidates from our approach will doubtlessly enhance the probability of success for finding new drugs, and will drastically change the approach currently taken for drug discovery.