Cyclin dependent kinase 4 inhibitory activity of Thieno[2,3-d] pyrimidin-4-ylhydrazones – Multiple QSAR and docking studies

Abstract

Overexpression of CyclinD1-pRb pathway is observed in many types of cancers. Discovery of novel lead molecules as CDK4 (a key regulator of the cell cycle) specific inhibitors is, therefore, a leading research area in cancer therapeutic field. The present study is focused on evolving multiple 2D and 3D QSAR models with good statistical fit based on the structures of 59, Thieno[2,3-d]pyrimidin-4-yl hydrazones with diverse CDK4 inhibitory activity profile. Statistical validation (both internal and external) based on Organization of Economic Co-operation and Development principles resulted in five comparable models that can furnish a valid correlation between their structure and activity quantitatively. The resultant models have R2 > 0.70, very low standard deviation (SD < 0.4) and external validation scores with Q2 > 0.65. The consensus prediction using the three top-ranked kernel based partial least square models revealed to be more useful for discriminating CDK4 inhibitory molecules from others. Docking analysis was beneficial in establishing the crucial hydrogen bonding interactions with hinge region residue Val96, Asp99 and Glu144. Salt bridge formation also found to be enhancing the CDK4 inhibition. High negative binding energy difference during solvation establishes the stability of the docked poses. The studies reveal that consensus prediction using the KPLS models along with docking observations can be used for screening databases of molecules with various structural features for finding out potential lead molecules with CDK4 inhibition.