Machine learning-driven identification of drugs inhibiting cytochrome P450 2C9

Author summaryCytochrome P450 (CYP) is a superfamily of heme-containing oxidizing enzymes responsible for the metabolism of a wide variety of drugs, xenobiotics and endogenous molecules. Five of the human CYPs (1A2, 2C9, 2C19, 2D6, and 3A4) are involved in similar to 95% of the CYP-mediated metabolism of drugs representing similar to 75% of drug metabolism. CYP inhibition leads to decreased drugs/chemicals elimination, which is a major cause of drug-drug interactions provoking adverse drug reactions. We developed an original integrated structure-based and machine learning approach for the prediction of CYP2C9 inhibitors. It exhibited excellent performance and its application allowed to demonstrate for the first time that the drugs vatalanib, piriqualone, ticagrelor and cloperidone are strong inhibitors of CYP2C9. Cytochrome P450 2C9 (CYP2C9) is a major drug-metabolizing enzyme that represents 20% of the hepatic CYPs and is responsible for the metabolism of 15% of drugs. A general concern in drug discovery is to avoid the inhibition of CYP leading to toxic drug accumulation and adverse drug-drug interactions. However, the prediction of CYP inhibition remains challenging due to its complexity. We developed an original machine learning approach for the prediction of drug-like molecules inhibiting CYP2C9. We created new predictive models by integrating CYP2C9 protein structure and dynamics knowledge, an original selection of physicochemical properties of CYP2C9 inhibitors, and machine learning modeling. We tested the machine learning models on publicly available data and demonstrated that our models successfully predicted CYP2C9 inhibitors with an accuracy, sensitivity and specificity of approximately 80%. We experimentally validated the developed approach and provided the first identification of the drugs vatalanib, piriqualone, ticagrelor and cloperidone as strong inhibitors of CYP2C9 with IC values <18 mu M and sertindole, asapiprant, duvelisib and dasatinib as moderate inhibitors with IC50 values between 40 and 85 mu M. Vatalanib was identified as the strongest inhibitor with an IC50 value of 0.067 mu M. Metabolism assays allowed the characterization of specific metabolites of abemaciclib, cloperidone, vatalanib and tarafenacin produced by CYP2C9. The obtained results demonstrate that such a strategy could improve the prediction of drug-drug interactions in clinical practice and could be utilized to prioritize drug candidates in drug discovery pipelines.

Automated summary

This study developed a novel approach combining structure-based and machine learning methods to predict CYP2C9 inhibitors, and successfully predicted several drugs as inhibitors of CYP2C9. This method is of great importance for improving the prediction of drug-drug interactions in clinical practice and prioritizing drug candidates.