Determination of comprehensive in silico determinants as a strategy for identification of novel PI3Kα inhibitors

Abstract

The PI3KCA gene functions by activating cascade signaling pathways leading to cell proliferation, survival, and growth. Being one of the frequently aberrant kinase in various malignancies, isoform selectivity among kinases remains a challenging aspect. In present study, efforts have been made to conceptualize determinants that are responsible for PI3Kα inhibition. Drug design techniques such as 3D-QSAR models, e-pharmacophore models, and shape-based screening utilities were derived from set of molecules and clinical trial candidates. QSAR models were validated using structure-based cross validation technique. Further, ROC analysis and molecular dynamics simulations were performed for the selected crystal structure for its validation. Virtual screening was employed for selection of hits and based on interaction pattern, binding affinity, and energy scores three hits with central scaffold as theino[2,3-d] pyrimidine (SS-RM-03), theino[3,2-d] pyrimidine (SS-RM-04), and oxadiazole (SS-RM-05) have been identified. The screened hits were then subjected to molecular dynamics simulations and quantum mechanical calculations. Further structure-guided methodology was adopted for analyzing prominent features of the hits and was correlated using common site feature analysis. The developed models along with structural features provided by molecular dynamics simulations serve as tools for identification of structural features essential for PI3Kα inhibition. Molecular determinants using diverse in silico tools have been identified which will facilitate drug discovery programs worldwide.