3D QSAR Pharmacophore, CoMFA and CoMSIA Based Design and Docking Studies on Phenyl Alkyl Ketones as Inhibitors of Phosphodiesterase 4

Abstract

In the present study, three dimensional quantitative structure activity relationship (3D QSAR) models using 3D Pharmacophore, CoMFA and CoMSIA approaches were developed for a series of phenyl alkyl ketone derivatives as PDE4 receptor antagonists. An ideal 3D QSAR pharmacophore model was developed and validated using external test set, Fischer's randomization method and decoy set screening. The top scoring four feature pharmacophore model, Hypo1, includes two hydrogen bond acceptors, two hydrophobic features. Amongst the developed models, Hypo1 has the maximum correlation coefficient (0.9658), cost difference (349.593), low RMS (1.41), and high goodness of fit. CoMFA and CoMSIA models were developed based on the alignment obtained using the pharmacophore (Hypo1), substructure alignment and by application of region focusing. The robustness of CoMFA and CoMSIA model was confirmed with the help of leave one out cross-validation method, while the predictive ability of models was tested using a test set. 3D-QSAR models with high squared correlation coefficient of up to 0.9720 for CoMFA and 0.9610 for CoMSIA were established. Robustness of the models is demonstrated by R(2)(cv) values of up to 0.7582 and 0.8539 for CoMFA and CoMSIA, respectively. Predictive ability of the models is reflected by R(2)(pred) values of 0.9630 and 0.9470 for CoMFA and CoMSIA respectively. Novel molecules were designed on the basis of results of 3D QSAR studies. Designed molecules were evaluated by Docking and Lipinski filters. Predicted activity of the designed molecules correlated well with the docking scores and the molecules also passed the Lipinski filters.