Establishment and evaluation of prediction models for the discovery of CYP1A2 inhibitors

CYP1A2 enzyme plays a crucial role in drug metabolism, and its inhibition may cause low metabolic rates and increased plasma concentrations of the drugs metabolized by CYP1A2, thus leading to drug toxicity. Therefore, distinction between CYP1A2 inhibitors and non-inhibitors becomes important topic on the early selection of new drug candidates and drug safety assessment. In this study, a CYP1A2 inhibitor-ligand library was built with 674 compounds known to possess CYP1A2 inhibitory activity. From the point of receptor-based and ligand-based view, we have built an automatic screening protocol with Pipeline Pilot, using molecular docking and pharmacophore modeling methods, so as to predict inhibitors from CYP1A2 inhibitor library quickly and accurately. The final model predicted 16 target compounds from the library, and 14 of which were CYP1A2 inhibitors. At last, we used the final model to screen the American Medicine Database, and four drugs were found to possess CYP1A2 inhibitory activity. The combination of prediction models can improve the efficiency of CYP1A2 inhibitor discovery significantly.