Discovery of novel human COX-2 inhibitors from the lotus natural products database: an integrated in silico approach

Abstract Cyclooxygenase-2 (COX-2) is a key enzyme involved in the inflammatory pathway and represents an important therapeutic target for the treatment of inflammatory diseases. This study aimed to identify potential natural inhibitors of human COX-2 from the LOTUS natural products database using an in silico drug discovery approach. A library of 276,518 natural compounds from the LOTUS database was screened using an e-pharmacophore model derived from the COX-2 crystal structure. Compounds that satisfied the pharmacophoric features were subjected to hierarchical molecular docking and were further evaluated through MMGBSA calculations, ADMET and AutoQSAR predictions. The selected compounds exhibited favorable docking scores ranging from − 10.062 to − 8.637 kcal/mol compared to Diclofenac (− 9.021 kcal/mol) and the reference ligand (− 6.559 kcal/mol). LTS0025171 showed the most favorable MMGBSA binding energy (− 52.12 kcal/mol), suggesting potentially favorable stable interaction within the COX-2 catalytic pocket. Interaction analysis revealed key hydrogen bonding with critical residues such as SER530, TYR385, ARG120, and TYR355. ADMET evaluation indicated favorable ADME and acceptable toxicity profiles for the lead compounds. AutoQSAR prediction identified Diclofenac and LTS0192010 as the compounds with the highest predicted inhibitory activity (5.691 and 5.628 respectively). The integrated computational approach identified candidate compounds from the LOTUS database with predicted COX-2 inhibitory potential characterized by demonstration of strong binding energies, favorable ADMET properties, and promising predicted biological activity. These compounds show potential lead scaffolds for the development of novel anti-inflammatory agents, although further experimental validation through in vitro and in vivo studies is recommended.