Discovery and biological evaluation of novel TLR2 antagonists through structure-based virtual screening and molecular dynamics simulations

Toll-like Receptor 2 (TLR2) is a key pattern recognition receptor in the innate immune system, and its aberrant activation is implicated in numerous inflammatory and immune-related disorders, making it an attractive therapeutic target. However, the development of potent and drug-like TLR2 antagonists remains challenging. In this study, we report the discovery and biological evaluation of novel small-molecule TLR2 antagonists targeting its extracellular domain through a structure-based virtual screening campaign. A hierarchical docking protocol (HTVS, SP and XP) was performed against the TLR2 crystal structure (PDB ID: 2Z7X) using the TopScience compound database containing approximately 1.2 million compounds. After Lipinski filtering (∼680,000 compounds) and stepwise docking, 18 hit compounds were identified. Top hits were prioritized by MM/GBSA binding free energy calculations and comprehensive ADMET predictions. Experimental validation using a HEK-Blue™ hTLR2 reporter cell assay confirmed that five hit compounds (T2, T5, T9, T12, T15) exhibit concentration-dependent antagonistic activity, with T9 demonstrating the most potent IC₅₀ of 13.32 μM, comparable to the reference MMG-11 (IC₅₀ = 5.36 μM). Notably, all novel hits displayed superior predicted metabolic stability compared to MMG-11, overcoming its major developmental bottleneck. Extensive 200 ns molecular dynamics simulations, complemented by PCA and free energy landscape analyses, revealed that T9 and T2 induce a more rigid and thermodynamically stable binding mode than MMG-11. MM/PBSA calculations identified van der Waals interactions as the dominant driving force, with per-residue decomposition highlighting PHE-322, PHE-349, and ILE-319 as critical binding hotspots. Density functional theory and interaction region indicator analyses further confirmed that the novel scaffolds possess extensive neutral, lipophilic surfaces and are intrinsically pre-organized for receptor binding. This integrated computational experimental approach has successfully identified promising new chemotypes with validated biological activity and improved drug-like properties, providing a robust foundation for developing next-generation TLR2 antagonists.