Rapid Parallel Virtual Screening Aids the Discovery of Novel P-Glycoprotein Inhibitors

P-glycoprotein (P-gp), a membrane-bound ATP-binding cassette (ABC) transporter, plays a crucial role in drug absorption, distribution, and therapeutic effectiveness. Despite extensive efforts, many P-gp inhibitors are large, hydrophobic molecules that bind the substrate-binding domain (SBD) in the transmembrane region and have failed in clinical trials. This motivates alternative strategies to discover smaller, more drug-like inhibitors. Here, we used virtual screening to identify novel P-gp inhibitors with improved physicochemical properties by targeting multiple potential allosteric sites in parallel. To accelerate candidate selection, we developed Ramp Up Active Learning (RUAL), a virtual screening protocol that prioritizes molecules with favorable docking scores from large, enumerated libraries while preserving sampling diversity. Based on the virtual-screening output, 33 molecules were synthesized and tested for their ability to inhibit P-gp-mediated Calcein-AM efflux in intact MDCKII MDR1 cells. Of these, six diverse molecules (18%) exhibited single-digit micromolar IC₅₀ values while maintaining low molecular weight and hydrophobicity, representing promising starting points for optimization. This study demonstrates how rapid, parallel virtual screening can enhance exploration across multiple docking templates and mitigate the risk of relying on a single docking model.