QSAR Pharmacophore-based Virtual Screening, CoMFA and CoMSIA Modeling and Molecular Docking towards Identifying Lead Compounds for Breast Cancer Protease Inhibitors

Aim: This study used QSAR Pharmacophore-based virtual screening and molecular docking to identify lead compounds and determine structural requirements for breast cancer inhibitor development. CoMFA and CoMSIA modeling was employed to design more potential inhibitors. Materials and Methods: 3D-QSAR pharmacophore models were developed using HypoGen Module and validated by Fischer's model and decoy test. The best pharmacophore model was employed to screen ZINC chemical library to obtain reasonable hits. Following ADMET filtering, 18 hits were subjected to further filter through docking. CoMFA and CoMSIA models were built by partial least squares on phenylindole-3-carbaldehydes derivatives. Results: 19 random runs from Fischer's validation and decoy test which led to an enrichment factor of 48.23 and Guner-Henry factor of 0.774 show that the identified pharmacophore model is highly predictive. Top three hits (IC50=0.01 similar to 0.05 mu M, fitness=52 similar to 62) were identified as potential inhibitory candidates from virtual screening and docking, and three new lead compounds were designed with predicted inhibiting potencies by pIC(50) value of 8.55 from CoMFA and CoMSIA modeling and fitness value of similar to 59 from docking. Conclusion: Validation results and decoy test indicate that the developed pharmacophore model is highly predictive. Residue Sep6 and Cys 5 were observed as important active sites for ligand-protein binding. Top three hits were identified as more potential inhibitors, and the designed compounds show more inhibiting potencies. The QSAR and docking results obtained from this work should be useful in determining structural requirements for inhibitor development as well as in designing more potential inhibitors.