Identification of New Inhibitors of the Kinase Activity of CDK2 and CDK9 by Molecular Modeling and High-Efficiency Screening

Kinases are important components of many signaling pathways in a cell, and therefore they are involved in the regulation of such diverse processes as transcription, cell cycle progression, apoptosis, cell differentiation, metabolism, and intercellular communication. Their increased activity often results in the development of various oncological, neurological, and cardiovascular diseases, as well as of immune system disorders. Treatment is most frequently based on an antagonistic approach, when a small molecule compound inhibits excessive kinase activity. However, the ATP-binding site of kinases is highly conserved, which often causes these ATP-competitive inhibitors to cross-react with a variety of other kinases, resulting in low selectivity. Thus, the discovery and development of new safe and effective inhibitors is a challenging task because it requires a comprehensive study of their effects on the human kinome. The experimental validation of possible candidates is an extremely expensive procedure, and narrowing down selection of candidate targets in silico became an attractive approach in early drug discovery. Various approaches to virtual screening were developed, with the goal to increase the cost-effectiveness of drug development manyfold by reducing the need for validation experiments. In this work, we applied a combination of different computational approaches to select 22 candidate inhibitors of the kinase activity of CDK2 and CDK9 and then tested them experimentally. Establishing a pipeline of virtual screening and subsequent validation made it possible to reveal the advantages and limitations of the methods used. About 1/3 of candidates were predicted correctly, including one compound with IC50 < 2 mu M for both kinases.