Integration of ligand- and structure-based cheminformatics tools with protein dynamic modeling for drug design

Abstract

Cheminformatics and computer-aided drug design (CADD) have matured substantially in the last decade. CADD is no longer the niche approach of a subset of specialists, but rather an integral component of the drug discovery process in both academia and industry. Today, we strive for precision in drug design, such that our molecules bind to specific conformations of flexible proteins and are selective against homologous receptors and/or mutants. The need for such precision in drug design is apparent when we consider epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC). Proteins can be conformationally dynamic biomolecules. These dynamics give rise to function, and aberrant dynamics can lead to disease. Changes in protein dynamics of EGFR caused by amino acid mutations drive oncogenic behavior in NSCLC. Herein, we characterized the mechanisms of oncogenesis and tyrosine kinase inhibitor (TKI) resistance in several new EGFR, HER2, and HER3 variants, as well as identified structure-function relations that may be responsible for variable outcomes in NSCLC patients with different EGFR exon 19 deletion variants. Subsequently, we describe new cheminformatics technologies developed to address unmet challenges in drug design.