Identification of novel selective MMP-9 inhibitors as potential anti-metastatic lead using structure-based hierarchical virtual screening and molecular dynamics simulation

Abstract

MMP-9 is an attractive target for the development of new anticancer drugs. In the current study, pharmacophore modeling was employed using two highly active and selective gelatinase inhibitors obtained from two X-ray crystal structures (PDB IDs: 2OVX and 2OW1) to identify novel selective MMP-9 inhibitors. The derived model was refined manually and also validated by the GH scoring method. The refined pharmacophore model, ADRR, was able to retrieve 86% of actives with a GH score of 0.774, indicating that the model was capable of retrieving the active MMP-9 inhibitors. ADRR was used to screen 2 838 166 unique structures. Hit filtration was carried out using a fitness score >1.5 and drug-likeness properties. Hierarchical clustering generates 33 clusters based on diversity. A total of 33 molecules were obtained and these molecules were taken for cross-docking studies with 5 subtype MMPs. Among 33 tested, 2 molecules, P10A-0000088030 (Lig-1) and P10A-0001383812 (Lig-2), were found to have the highest docking scores (−8.59 kcal mol⁻¹ and −8.27 kcal mol⁻¹) towards MMP-9 compared with the other MMPs. Further MM-GBSA analysis was performed for two hits with 5 subtype MMPs to reveal the essential features that contribute to selectivity. The results showed that van der Waals contributions play a central role in determining the selectivity of MMP-9 inhibitors. Molecular dynamics studies were carried out for total time of 330 ns to assess the stability of ligands at the active site. MD analysis showed that binding of Lig-1 with MMP-9 is stable compared to that with Lig-2. Hence, we suggest the Lig-1 compound as a good lead in designing novel potent inhibitors of MMP-9.