Fragment-based diffusion modeling and molecular dynamics simulation validation for the discovery of PD-L1 small-molecule inhibitors.

The programmed cell death-1/programmed cell death-ligand 1 (PD-1/PD-L1) pathway is a key target in cancer immunotherapy. Although monoclonal antibodies (mAbs) have demonstrated remarkable clinical efficacy, their application is limited by poor tissue penetration, high production costs, and the need for intravenous administration. Small-molecule inhibitors provide a promising complementary strategy, but designing them remains challenging due to the large, relatively flat PD-1/PD-L1 interface. In this study, we integrated fragment-based drug design (FBDD) with conditional diffusion modeling to overcome these obstacles. Core scaffolds consisting of key fragments identified through protein-ligand interaction analysis were used as conditional inputs. Considering the relatively conserved binding mode and limited pocket flexibility of reported PD-L1/small-molecule inhibitor complexes, seven representative co-crystal structures were selected to capture the major binding features and guide molecular generation. Structurally plausible candidate inhibitors were generated using diffusion modeling and screened by molecular docking. After 500 ns molecular dynamics (MD) simulations, we identified four candidates (bo1-bo4), which were selected for MD-based evaluation. The predicted binding free energy (BFE) values of three compounds (bo1, bo2, and bo3) were lower than - 40 kcal/mol, as calculated by the molecular mechanics-Poisson Boltzmann surface area (MM-PBSA) method with interaction entropy (IE) correction, suggesting their potential to stabilize the PD-L1 dimer interface in silico and serve as computationally prioritized candidates for further experimental evaluation of PD-1/PD-L1 blockade. Overall, this work suggests that fragment-based diffusion modeling is an efficient and interpretable strategy for the discovery of computationally prioritized PD-L1 small-molecule candidate inhibitors and offers a promising framework for tackling challenging targets in cancer immunotherapy.