Artificial intelligence-driven approach to understand and predict drug-drug interactions related to ABC transporters and metabolizing enzymes

Drug efficacy/toxicity prediction is a major concern in modern research and development. To prevent adverse drug reactions, in silico approaches assessing ADMET (absorption, distribution, metabolism, excretion, toxicity) have become an integral part of early drug discovery. We will develop an original Artificial Intelligence (AI) approach for prediction of ADMET related to chemicals/drugs interactions with major ABC transporters and drug metabolizing enzymes (DME). We will focus on the two ABC transporters, P-glycoprotein (P-gp, ABCB1) and BCRP (ABCG2), and two DME, cytochrome P450 and UGT, key for drug-drug interactions. Conformational space of these proteins will be explored using molecular modeling simulations. Our approach will combine structure-based and machine learning approaches as well as pharmacogenetics considerations. We will employ the developed AI approach to identify new drugs substrates and/or inhibitors of P-gp, BCRP and UGT, that will be experimentally validated using in vitro and PBPK methods. Pharmacogenetics studies will be undertaken in order to understand and predict structural bases involved in drug inefficacy/toxicity due to the studied proteins’ polymorphisms. To the best of our knowledge, this will be the first AI approach for prediction of drug-drug interactions specifically related to ABC transporters and DMEs. Finally, we will implement the developed AI methodology in a new software for the scientific community.