Quantitative prediction of P-glycoprotein-mediated drug–drug interactions and intestinal absorption using humanized mice

Abstract

Background and Purpose: P-glycoprotein (P-gp) exhibits a broad substrate specificity and affects pharmacokinetics, especially intestinal absorption. However, prediction, in vivo, of P-gp-mediated drug–drug interaction (DDI) and non-linear absorption at the preclinical stage, is challenging. Here we evaluate the use of human MDR1 mouse artificial chromosome (hMDR1-MAC) mice carrying human P-gp and lacking their own murine P-gp to quantitatively predict human P-gp-mediated DDI and non-linear absorption. Experimental Approach: The P-gp substrates (aliskiren, betrixaban, celiprolol, digoxin, fexofenadine and talinolol) were administered orally to wild-type, Mdr1a/b-knockout (KO) and hMDR1-MAC mice, and their plasma concentrations were measured. We calculated the ratio of area under the curve (AUCR) in mice (AUCMdr1a/b-KO/AUCwild-type or AUCMdr1a/b-KO/AUChMDR1-MAC) estimated as attributable to complete P-gp inhibition and the human AUCR with and without P-gp inhibitor administration. The correlations of AUCRhuman with AUCRwild-type and AUCRhMDR1-MAC were investigated. For aliskiren, betrixaban and celiprolol, the Km and Vmax values for P-gp in hMDR1-MAC mice and humans were optimized from different dosing studies using GastroPlus. The correlations of Km and Vmax for P-gp between human and hMDR1-MAC mice were investigated. Key Results: A better correlation between AUCRhuman and AUCRhMDR1-MAC (R2 = 0.88) was observed. Moreover, good relationships of Km (R2 = 1.00) and Vmax (R2 = 0.98) for P-gp between humans and hMDR1-MAC mice were observed. Conclusions and Implications: These results suggest that P-gp-mediated DDI and non-linear absorption can be predicted using hMDR1-MAC mice. These mice are a useful in vivo tool for quantitatively predicting P-gp-mediated disposition in drug discovery and development.