Abstract
Structural and electronic factors influencing the inhibition of cyclooxygenase-1 and -2 (COX-1/COX-2) were studied by means of Electronic-Topological Method combined with Neural Networks (ETM-NN), molecular docking and Density Functional Theory (DFT). A series of structurally diverse compounds containing 209 molecules were classified in accordance with their inhibiting properties, as selectively inhibiting and non-selectively inhibiting COX-2 receptor agents (110 and 99 molecules, correspondingly). The results obtained from the ETM-NN calculations gave us possibility of selecting those pharmacophoric molecular fragments, which allow for the search of new selective inhibitors of COX-2 with high probability of realization. The final selection of pharmacophores and anti-pharmacophores found was taken as a basis for a system designed for the COX-2 inhibitory activity prediction. Analysis of the electron density distribution showed that more effective binding with COX-2 receptor was observed for selective inhibitors. To make an assessment of these interactions, calculations of stabilization energies were carried out for the ligand-receptor complexes. From the results of the docking and from the analysis of electronic structures of active sites of enzymes, some peculiarities of ligand-receptor binding and its influence on the selectivity of the COX-2 relative to COX-1 inhibition were elucidated. 95% of compounds were recognized correctly, as the most active ones, by the system of prediction designed. Thus, the system being the result of the study is capable of predicting the selective inhibitory activity of COX-2 successfully. As a consequence, it can be used both for computer screening and synthesis of potent inhibitors of COX-2 with molecular skeletons that may vary considerably.
Keywords: COX-1/COX-2, DFT, electronic-topological approach, molecular docking, NSAID.