AI-assisted strategy transforms PD-L1 inhibitory peptides into potent, membrane-permeable small molecule inhibitors

Programmed death-ligand 1 (PD-L1) is a crucial immune checkpoint protein, often overexpressed in cancer, allowing tumors to evade immune surveillance. Inhibiting its interaction with PD-1 is a cornerstone of modern cancer immunotherapy. While macrocyclic peptides can effectively disrupt these protein-protein interactions (PPIs), their poor membrane permeability and metabolic instability limit therapeutic utility. There's a critical need to translate the potency of peptide inhibitors into small molecules with enhanced drug-like properties for broader clinical application.

Researchers developed a novel AI-assisted strategy to transform macrocyclic PD-L1 inhibitory peptides into small molecules. This approach combined amino acid mapping (AAM) descriptors with a generative AI platform. The goal was to identify potent and membrane-permeable protein-protein interaction (PPI) inhibitors. The methodology focused on precision design, moving beyond traditional empirical screening, to optimize activity and stability through structure-based drug design.

The AI-assisted strategy, leveraging amino acid mapping (AAM) descriptors, successfully facilitated the transformation of macrocyclic PD-L1 inhibitory peptides into small molecules. This novel approach enabled the identification of compounds designed to be potent and membrane-permeable protein-protein interaction (PPI) inhibitors. The methodology demonstrated its capability to improve activity through structure-based drug design, moving towards optimized stability and drug-like properties for the derived small molecules. The generative AI platform played a crucial role in this precision design process, shifting from empirical screening to intelligent drug discovery for challenging targets.