A Study on Pharmacokinetics of Bosentan with Systems Modeling, Part 2: Prospectively Predicting Systemic and Liver Exposure in Healthy Subjects

Abstract

Predicting human pharmacokinetics of novel compounds is a critical step in drug discovery and clinical study design but continues to be a challenging task for hepatic transporter substrates, particularly in predicting their liver exposures. In this study, using bosentan as an example, we prospectively predicted systemic exposure and the (pseudo) steady-state unbound liver-to-unbound plasma ratio (p) in healthy subjects using 1) a mechanistic approach solely based on in vitro hepatocyte assays and 2) an approach based on hepatic process rates from monkey in vivo data but Michaelis-Menten constants from in vitro data. Both methods reasonably match the observed human systemic time course data, but the second method leads to better prediction accuracy. In addition, the second method can predict a human p value that is close to the value deduced using clinical data. We also generated rat and monkey liver p values in terminal studies. However, these directly measured animal values are different from the deduced human value.

A Study on Pharmacokinetics of Bosentan with Systems Modeling, Part 1: Translating Systemic Plasma Concentration to Liver Exposure in Healthy Subjects

“…We also tried to estimate a Kp u scaler by including it as another fitted parameter; however, this parameter cannot be precisely estimated. In a Translating Bosentan Systemic Exposure to Liver Exposure monkey study presented in part 2 of this work (Li et al, 2018), in which both systemic and liver exposure are determined experimentally, we can confidently estimate the nonliver Kp u scaler as 1.47, which justifies a value of 1 in the present exercise.…”

“…There are hepatic active uptake, active basal efflux, and passive diffusion between plasma and tissue, and metabolism within the tissue. The biliary excretion was assumed to be minimal for bosentan based on the fact that 1) in vitro sandwich cultured human hepatocytes showed no biliary excretion (data provided in part 2 of this study; Li et al, 2018), and 2) minimal compound is excreted into feces after intravenous dosing in humans (Weber et al, 1999b). With the exception of passive diffusion clearance (CL liver,pass ), hepatic processes were assumed to follow Michaelis-Menten kinetics.…”

“…Blood binding parameters share the same values as those in circulating blood. Intracellular binding parameters are fixed at values estimated from in vitro hepatocyte assays in part 2 of this work (Li et al, 2018).…”

“…An in vitro hepatocyte induction study was performed to understand whether the P450 induction could be accurately predicted using primary hepatocytes. The data were analyzed using a mechanistic model that combines the sandwich cultured human hepatocyte model (Li et al, 2018) and the P450 turnover model mentioned above. Details are provided in the Supplemental Material.…”

In vitro and in vivo methods to assess pharmacokinetic drug– drug interactions in drug discovery and development

“…Both static or dynamic models are available for human DDI prediction. A static model assumes that the concentration of a perpetrator is at the maximum concentration (C max ) (Brown, Galetin, Hallifax, & Houston, ), whereas a dynamic model assumes the perpetrator concentration for DDI is the dynamic concentration reached within a given tissue (Li et al, ; Li et al, ; Pang & Durk, ). The accurate prediction of a drug's human PK early in drug discovery is important.…”