Mechanistic modeling of hepatic transport from cells to whole body: application to napsagatran and fexofenadine.

A mechanistic model was applied to quantitatively derive the kinetic parameters from in vitro hepatic uptake transport data. These parameters were used as input to simulate in vivo elimination using a fully mechanistic physiologically based pharmacokinetic (PBPK) model. Fexofenadine and napsagatran, both BDDCS class 3 drugs, were chosen as model compounds. In rat, both compounds are hardly metabolized and are eliminated unchanged mostly through biliary excretion. Uptake was estimated in this study based on plated rat hepatocytes, and a mechanistic model was used to derive the active and passive transport parameters, namely Michaelis-Menten uptake parameters (V(maxI) and K(mI,u)) together with passive diffusion (P(dif)) and nonspecific binding. Maximum transport velocity and passive diffusion were scaled to in vivo parameters (J(maxI) and PS(TC)) using hepatocellularity. Biliary excretion, through passive and active transport, was assessed from in vivo studies. These transport parameters were then used as input in a whole body physiologically based model in which the liver compartment was parametrized for the different passive and active transport processes. Each of the processes was linked to the free concentration in the relevant compartment. For napsagatran hepatic uptake, no passive diffusion and no binding were detected in vitro besides the active transport (K(mI,u) = 88.4 +/- 8.1 microM, V(maxI) = 384 +/- 19 pmol/mg/min). Fexofenadine was rapidly taken up into rat hepatocytes (K(mI,u) = 271 +/- 35 microM, V(maxI) = 3162 +/- 274 pmol/mg/min), and some contribution of passive diffusion to the uptake (P(dif) = 2.08 +/- 0.67 microL/mg/min) was observed. For fexofenadine, the biliary export rate was found to be slower than the uptake, leading to drug accumulation in liver. No accumulation was observed for napsagatran where excretion was faster than hepatic uptake. Observed plasma, liver and bile concentration time profiles were compared to PBPK simulations based on scaled in vitro transport kinetic parameters. An uncertainty analysis indicated that for both compounds the scaled in vitro uptake clearance had to be adjusted with an additional empirical scaling factor of 10 to match the plasma and liver concentrations and biliary excretion profiles. Applying this model, plasma clearance (CL(P)) and half-life (t(1/2)), maximum liver concentration (C(maxL)) and fraction excreted in bile (f(bile)) were predicted within 2-fold. In vitro uptake data had most impact on the simulated plasma and biliary excretion profiles, while accurate simulations of liver concentrations required also quantitative estimates of biliary excretion transport. This study indicated that the mechanistic model allowed for accurate evaluation of in vitro experiments; and the scaled kinetic parameters of hepatic uptake transport enabled the prediction of in vivo PK profiles and plasma clearances, using PBPK modeling.