A 3-step static mechanistic modeling approach to predicting transporter- and enzyme-mediated interaction signals of direct oral anticoagulants.

Abstract

Direct oral anticoagulants (DOACs) are susceptible to drug-drug interactions involving both transport and metabolism. The relative contribution of intestinal transporters in DOAC pharmacokinetics, however, is uncharacterized, hindering the prediction of their drug interactions. In this study, we examined potential interactions of select precipitants with apixaban, dabigatran, edoxaban, and rivaroxaban using a 3-step static mechanistic combinatorial modeling approach. In this approach, we (1) calculated apparent estimates of the fractions of efflux transport (f_e) of the 4 DOACs by the intestinal breast cancer resistance protein and intestinal permeability glycoprotein (P-gp), (2) validated our estimates, and (3) used them to simulate unstudied drug interactions with the anticoagulants. We estimated P-gp and breast cancer resistance protein transport to play a similar role in the interactions of apixaban and rivaroxaban (f_{e_P-gp} = 0.21, f_{e_breast cancer resistance protein} = 0.13). Additionally, the role of P-gp was prominent in the interactions affecting dabigatran etexilate (f_{e_P-gp} = 0.61) and less so in those affecting edoxaban (f_{e_P-gp} = 0.38). Using our apparent estimates, we predicted 144 interaction signals of 47 drugs with the 4 anticoagulants. All signals showed potential change >25% in the area under the curve of DOACs, warranting further investigation. Using our combined model, we were able to uncover new drug interaction signals, including previously unrecognized signals by kinase inhibitors. SIGNIFICANCE STATEMENT: Multiple enzymes and transporters affect direct oral anticoagulant pharmacokinetics. Drugs interacting with the same proteins may increase the levels of direct oral anticoagulants, leading to bleeding complications. Current methods predicting such interactions focus only on individual enzymes or transporters, limiting the prediction of cases where both may be involved. In this study, we combined multiple models to capture the complexity of interactions affecting direct oral anticoagulants. Using our validated model, we predicted 144 drug-drug interaction signals with potential clinical significance.