A computer aided drug discovery based discovery of lead-like compounds against KDM5A for cancers using pharmacophore modeling and high-throughput virtual screening.

Abstract

KDM5A over-expression mediates cancer cell proliferation and promotes resistance toward chemotherapy through epigenetic modifications. As its complete mechanism of action is still unknown, there is no KDM5A specific drug available at clinical level. In the current study, lead compounds for KDM5A were determined through pharmacophore modeling and high-throughput virtual screening from Asinex libraries containing 0.5 million compounds. These virtual hits were further evaluated and filtered for ADMET properties. Finally, 726 compounds were used for docking analysis against KDM5A. On the basis of docking score, 10 top-ranked compounds were selected and further evaluated for non-central nervous system (CNS) and CNS drug-like properties. Among these compounds, N-{[(7-Methyl-4-oxo-1,2,3,4-tetrahydrocyclopenta [c] chromen-9-yl) oxy]acetyl}-l-phenylalanine (G-score: -11.363 kcal/mol) was estimated to exhibit non-CNS properties while 2-(3,4-Dimethoxy-phenyl)-7-methoxy-chromen-4-one (G-score: -7.977 kcal/mol) was evaluated as CNS compound. Docked complexes of both compounds were finally selected for molecular dynamic simulation to examine the stability. This study concluded that both these compounds can serve as lead compounds in the quest of finding therapeutic agents against KDM5A associated cancers.