Targeting cryptic-orthosteric site of PD-L1 for inhibitor identification using structure-guided approach

Approved mAbs that block the protein-protein interaction (PPI) interface of the PD-1/PD-L1 immune checkpoint axis have led to significant improvements in cancer treatment. Despite having drawbacks of mAbs only few a compounds are reported till date against this axis. Inhibiting PPIs using small molecules has emerged as a significant therapeutic opportunity, demanding for the identification of drug-like molecules at an accelerated pace under the hit-to-lead campaigns. Due to the PD-L1's cross-talk with PD-1/CD80 and its overexpression on cancer cells, as well as the availability of its crystal structures with small molecules, it is an enticing therapeutic target for structure-assisted small molecule design. Furthermore, the selection of chemical databases enriched with focused designing for PPI interfaces is crucial. Therefore, in this study we have utilized the Asinex signature library for structure-assisted virtual screening to find the potential PD-L1 inhibitors by targeting the cryptic PD-L1 interface, followed by induced fit docking for pose refinements in the pocket. The obtained hits were then subjected to interaction fingerprinting and ligand-based drug-likeness investigations in order to evaluate and analyze their drug-like qualities (ADME). Twelve compounds qualified for molecular dynamics simulations, followed by thermodynamic calculations for evaluation of their stability and energetics inside the pocket. Two novel compounds with different chemical moieties have been identified that are consistent throughout the simulation, mimicking the interactions and binding energies with BMS-1166. These compounds appear as potential therapeutic candidates to be explored experimentally, thereby paving the way for the development of novel leads as immunomodulators.