Usefulness of A Model-Based Approach for Estimating In Vitro P-Glycoprotein Inhibition Potency in a Transcellular Transport Assay

In vitro half-maximal inhibitory concentration (IC50) is a key parameter for accurately predicting the potential risk for P-glycoprotein (P-gp)--mediated drug--drug interactions. We aimed to compare the IC50 values estimated by different approaches and determine the usefulness of model-based approaches. Transcellular transport of digoxin across Caco-2 monolayer was investigated using various concentrations of P-gp inhibitors, quinidine, verapamil, and zosuquidar. To calculate IC50 values, 3 traditional parameters were used: apical-to-basal (AtoB) and basal-to-apical (BtoA) clearance (CL) with inhibitors (CLAtoB,i and CLBtoA,i) and the difference between the efflux ratios (ERs) with P-gp inhibitors (ERi) and those under complete P-gp inhibition [ER(-P-gp)]. Furthermore, a new model-based approach was applied that uses the difference between the reciprocals of CLAtoB with P-gp inhibitors (1/CLAtoB,i) and those under complete P-gp inhibition [1/CLAtoB(-P-gp)] as parameters. IC50 values obtained from 2 model-based approaches [ERi - ER(-P-gp) and 1/CLAtoB,i - 1/CLAtoB(-P-gp)] were comparable, whereas 2.6- to 6.6-fold larger IC50 values were estimated from empirical approaches (CLAtoB,i and CLBtoA,i). The reason for such difference in IC50 values is that indicators for model-based approaches, but not empirical approaches, directly reflect the P-gp function. Our new approach [1/CLAtoB,i - 1/CLAtoB(-P-gp)] based on only AtoB transcellular transport could substitute for current estimation methods using ER.