Prediction of drug-drug interaction risk of P-glycoprotein substrate in drug discovery.

We aimed at predicting the drug-drug interaction (DDI) risk of P-glycoprotein (P-gp) substrates by using P-gp expressing LLC-PK1 cells and its knockout mice (KO). The area under the curve (AUC) of 16 marketed drugs and plasma concentration (C) of 207 screening compounds, with corrected efflux ratio (CER) ≥ 2, were compared between P-gp KO mice and wild type mice (WT). At permeability (Papp) ≥ 10 × 10 cm/s in parent LLC-PK1 cells, AUC ratios (KO/WT) and C ratios (KO/WT) of these compounds were within 3-fold. AUC ratios (KO/WT) of clinical P-gp substrates, with human AUC ratios with and without P-gp inhibitor administration ≥2, were higher than 8.7. These observations led us to establish a work-flow of P-gp substrate assessment with the threshold AUC ratio (KO/WT) ≥ 9 leading to a DDI risk of AUC ratio (human) ≥ 2. A screening compound showing high CER (=57.6) was found, but its AUC ratio (KO/WT) was 3.7, had been presumed to be a weak risk and its AUC ratio (human) was 1.2 in a later clinical DDI study. Our proposed workflow should be useful for predicting the DDI risk of P-gp substrates in drug discovery.