COMPUTATIONAL INVESTIGATION OF NATURAL COMPOUNDS AS POTENTIAL THERAPEUTIC LEADS AGAINST BREAST CANCER: MOLECULAR DOCKING, MOLECULAR DYNAMICS AND ADMET ANALYSIS

Abstract

Breast cancer is a diverse disease with molecular pathways related to DNA repair, tumor suppression, cell proliferation and hormonal signaling. This study performed analysis of the 10 selected natural compounds against the five breast cancer-associated protein targets by using an integrated in silico workflow of molecular docking, protein–ligand interaction analysis, docking validation, 200-ns molecular dynamics (MD) simulation and ADMET/drug-likeness prediction. The targets included: BRCA1, BRCA2, TP53, HER2 and estrogen receptor alpha (ERα). The highest predicted binding among the target proteins was seen with EGCG via docking score with the proteins BRCA1, BRCA2, TP53, HER2 and ERα of −9.6, −9.2, −8.8, −10.1 and −9.4 kcal/mol respectively. HER2-EGCG complex was especially good as it contained six hydrogen bonds in total with the residues ASP863, THR862 and LYS753. The RMSD values obtained in the docking procedure were ranging from 1.23 to 1.67 Å for the docking validation. The average RMSD (1.85 Å), average RMSF (1.12 Å), average hydrogen-bond occupancy (6.4) and radius of gyration (21.45 Å) were the lowest for HER2–EGCG in MD simulations. Review of the predicted profiles by ADMET analysis was generally favorable, with EGCG being one violation of the Lipinski rule and showing low predicted hepatotoxicity. Integrated ranking was used and EGCG came out on top followed by Luteolin and Quercetin. The results show that EGCG (specifically when bound with HER2) represents the strongest computational lead in this data set but that experimental validation is necessary.