Integrative Computational Approaches for TRPV1 Ion Channel Inhibitor Discovery: An Integrated Machine Learning, Drug...

The transient receptor potential vanilloid 1 (TRPV1) ion channel is a key mediator of pain and inflammation, making it a crucial target for developing new analgesics. Despite progress in understanding TRPV1's role, novel modulators that effectively inhibit nociceptive transduction are needed. Additionally, robust drug design protocols capable of exploring vast chemical space remain imperative. In this study, we present a computational framework combining machine learning (ML), virtual screening, ensemble molecular docking, molecular dynamics (MD) simulations, and MM-GBSA calculations to identify potential TRPV1 modulators among FDA-approved drugs. ML models trained on bioactivity data from the ChEMBL database classified 670 candidates from a library of FDA-approved drugs. Ensemble docking simulations, conducted on four TRPV1 cryo-EM structures representing different functional states, assessed binding interactions at the vanilloid site, a critical modulation domain. The top 20 candidates were further analyzed using MD simulations and MM-GBSA calculations to evaluate their stability and binding energetics. Among these, CYM-5442 (CA6), Rociletinib (CA7), and SC-51089 (CA9) demonstrated strong binding affinities and thermodynamic stability, outperforming known modulators such as Capsazepine and Capsaicin. These findings highlight the effectiveness of combining ML and molecular simulations in drug discovery, offering valuable insights into the identification of novel TRPV1 modulators as a starting point for further experimental validation and optimization in the development of next-generation analgesics targeting the TRPV1 channel.