Computational Docking and in Silico Toxicity Prediction of Polydatin as a Potential TNF-α Modulator

Abstract

This study presents a comprehensive computational evaluation of the interactions of Polydatin (piceid) and Cortisone with members of the TNF superfamily, focusing on TNF-α. Blind docking was performed across multiple TNF proteins, while selective docking targeted TNF-α using the crystallographic binding pocket of its co-crystallized ligand (PDB ID: 7JRA). Both compounds displayed similar binding energies (~ −7 kcal/mol) across several TNF family members. Notably, Polydatin exhibited a significantly higher affinity for TNF-α (−9.7 kcal/mol using AutoDock Vina) compared to Cortisone (−6.6 kcal/mol).Redocking of the crystallographic ligand VGY validated the docking protocol, demonstrating accurate pose superimposition and consistent binding energies. Further docking using AutoDock4 (AMDock) confirmed favorable Polydatin binding (−10.87 kcal/mol), with estimated nanomolar inhibitory constants. In silico toxicity predictions via pkCSM suggest that Polydatin possesses a favorable safety profile, including a high predicted maximum tolerated dose and absence of mutagenicity or hepatotoxicity. Overall, these results indicate that Polydatin is a promising natural compound for TNF-α modulation, warranting further in vitro and in vivo studies.