Identification of new potential cyclooxygenase-2 inhibitors: insight from high throughput virtual screening of 18 million compounds combined with molecular dynamic simulation and quantum mechanics

Abstract

The cyclooxygenase isoenzymes (COX-1 and COX-2) have a critical role in inflammation, fever, and pain. In contrary to COX-1, COX-2 is specifically expressed in inflamed tissues. Because of the subtle difference between both enzyme active sites, targeting COX-2 represents an efficient strategy for the development of novel inhibitors against inflammation with fewer side effects. In order to identify potential inhibitors of COX-2, more than 18,000,000 small molecules were retrieved from the ZINC database and virtually screened against it with a gradual increase in the precision through combined multistep docking. The results were sorted according to the rank-by-rank, induced-fit docking, and MM-GBSA evaluation. Subsequently from the final hit list, two top hits along with an approved selective inhibitor (celecoxib) were further investigated by the molecular dynamics (MD) simulations. The results were indicated that ZINC16934653 and ZINC40484701 demonstrate the highest affinity for the COX-2 binding pocket. Both ligands were bound to the important active-site residues, which are necessary for the correct orientation of inhibitors inside the binding cavity. Their binding free energies were comparable to celecoxib. 100 ns MD simulation is revealed that ZINC40484701 is more preferred in comparison with ZINC16934653 and celecoxib. In addition, non-covalent interactions between the compounds and key residues located in 6 Å distance from the COX-2 binding site show similar patterns of bonding by the reduced density gradient and the independent gradient model. Therefore, ZINC40484701 can be a potential candidate for further in vitro and in vivo analysis after lead-optimization efforts. Communicated by Ramaswamy H. Sarma