Exploring QSAR and pharmacophore mapping of structurally diverse selective matrix metalloproteinase-2 inhibitors

Objectives and Methods Matrix metalloproteinase-2 (MMP-2) is a potential target in metastases. Regression (conventional 2D QSAR) and classification (recursive partitioning (RP), Bayesian modelling) QSAR, pharmacophore mapping and 3D QSAR (comparative molecular field analysis and comparative molecular similarity analysis) were performed on 202 MMP-2 inhibitors. Key findings Quality of the regression models was justified by internal (Q(2)) and external (R-Pred(2)) cross-validation parameters. Stepwise regression was used to develop linear model (Q(2) = 0.822, R-Pred(2) = 0.667). Genetic algorithm developed linear (Q(2) = 0.845, R-Pred(2) = 0.638) and spline model (Q(2) = 0.882, R-Pred(2) = 0.644). The RP and Bayesian models showed cross-validated area under receiver operating characteristic curve (AUC(ROC_CV)) of 0.805 and 0.979 respectively. QSAR models depicted importance of descriptors like five-membered rings, fractional positively charged surface area, lipophilocity and so on. Higher molecular volume was found to be detrimental. Pharmacophore mapping was performed with two tools - Hypogen and PHASE. Both models indicated that one hydrophobic and three hydrogen bond acceptor features are essential. The Pharmacophore-aligned structures were used for CoMFA (Q(2) of 0.586 and R-Pred(2) of 0.689) and CoMSIA (Q(2) of 0.673 and R-Pred(2) of 0.758), results of which complied with the other analyses. Conclusions All modelling techniques were compared to each other. The current study may help in designing novel MMP-2 inhibitors.