Qsar Studies Of The Anti-cancer Drugs Based On Heat Shock Protein90and Cyclin-dependent Kinase2Receptors

Nowadays, with the development of science and technology, and the changes of environment, the occurrence of cancer is still in an upward trend, which has become a kind of commonly and frequently encountered disease. Recently, anticancer drugs have been developed quickly; however, experimental methods for developing anti-cancer drugs are considerably time-consuming and labor-intensive.Computer-aided drug design (CADD) can greatly increase the efficiency and correction of drug analysis and evaluation, shorten the period of drug design by providing more information, for example, the structure of drug candidate. CoMFA、CoMSIA、molecular docking and molecular dynamics have been commonly used in CADD.In this paper, different series of HSP90and CDK2inhibitors were regarded as research objects. Some important factors improving the activity of the ligands can be analyzed and discussed by the steric, electrostatic, hydrophobic, hydrogen bond donor and hydrogen bond acceptor contour maps which can be obtained by CoMFA and CoMSIA. The inhibitory mechanism and binding modes were further studied by molecular docking and molecular dynamics. At last, some significant results can be gained as follows:(1)3D-QSAR models are performed on three different series of Hsp90inhibitors, which are ligand-based or receptor-based methods. The optimum3D-QSAR models exhibite reasonable statistical characteristics with averaging internal q2>0.60and external r2pred>0.66for Benzamide tetrahydro-4H-carbazol-4-one analogs (BT), AT13387derivatives (AT) and Dihydroxylphenyl amides (DA). The results indicate that steric effects contribute the most to the BT model, whereas H-bonding effects are more important to AT model, and electrostatic, hydrophobic, H-bond donor almost contribute equally to the DA model. The molecular docking and molecular dynamics analysis show that Asp93, Tyr139and Thr184in Hsp90are important for the three different series of inhibitors.(2) A molecular modeling study is performed to understand the binding mode of these modulators at the allosteric site of the CDK2enzyme. The resultant optimum QSAR models exhibite q2=0.73, r2pred=0.94for6,6-dimethyl pyrrolo [3,4-c] pyrazoles analogs (DPPs), q2=0.62, r2pied=0.63for Imidazole pyrimidine amides analogs (IPAs) and q2=0.56, r2Pred=0.58for4-(Pyrazol-4-yl)-pyrimidines analogs (PYPs), indicating that the obtained models possess a good predictive ability in both internal and external validation. Furthermore, a comparison between3D-contour map, molecular docking and molecular dynamics simulation have validated the detailed binding modes, i.e., for DPPs and PYPs, the overall preferred binding modes are consistent with the initial optimal docked structures, demonstrating the rationality of the docking models. For IPA analogs, the molecular dynamics simulation optimizes the docked structure model, and the reasonable binding model shows that the key amino acid residues impacting the interactions are Asp86, Asp145and Lys33.The obtained results should be useful in the structure optimization, and binding mode analysis for these two kinds of anti-tumor compounds.