Discovery of new selective cytotoxic agents against Bcl-2 expressing cancer cells using ligand-based modeling

Abstract

Bcl-2 is an anti-apoptotic protein involved in cancer resistance to cytotoxic therapies making it an interesting target for inhibitors design. Towards this end, we implemented an elaborated ligand-based computational workflow that combines exhaustive pharmacophore modeling and quantitative structure-activity relationship (QSAR) analysis to explore the structural features required for potent Bcl-2 inhibitors employing 98 known Bcl-2 inhibitors. Genetic function algorithm (GFA) coupled with k nearest neighbor (kNN) or multiple linear regression (MLR) analyses were employed to generate predictive QSAR models based on optimal combinations of pharmacophores and physicochemical descriptors. The optimal QSAR-selected pharmacophore models were validated by receiver operating characteristic (ROC) curve analysis and by comparison with crystallographic structures of known inhibitors co-crystallized within Bcl-2 binding pocket. Optimal QSAR models and their associated pharmacophore hypotheses were validated by identification and experimental evaluation of new selective cytotoxic compounds against Bcl-2 expressing cancer cells. The hits were retrieved from the National Cancer Institute (NCI) structural database. Several potent hits were captured. The most potent hits illustrated IC50 values of 4.2 and 2.60 μM against MDA-MB-231 cancer cell-line.

Keywords: Anticancer screening; Bcl-2; MLR; Pharmacophore; QSAR; Virtual screening; kNN.