Daniel Moj(1), Tobias Kanacher(2), Thomas Wendl(2), Stefan Willmann(2), Thorsten Lehr(1)
(1) Clinical Pharmacy, Saarland University, Saarbruecken, Germany, (2) Bayer Technology Services GmbH, Systems Biology & Computational Solutions, Leverkusen, Germany
Objectives: Development of a descriptive and predictive time-dependent PBPK DDI model for clarithromycin (CYP3A4 substrate/inhibitor) and midazolam (CYP3A4 substrate).
Methods: PBPK models were independently developed for clarithromycin and midazolam using physicochemical properties (e.g. pka, logP) and study demographics (age, height, mass and BMI) of various clinical studies in humans. Time dependency was incorporated applying a dynamic CYP3A4 turnover model (intestinal mucosa and liver) using in vitro/vivo data (e.g. kdeg). Parameter identification (e.g. fraction unbound) was performed using plasma concentration-time profiles after intravenous (i.v.) infusion (clarithromycin), i.v. bolus injection (midazolam), single (clarithromycin, midazolam) and multiple oral dosing (clarithromycin). Single PBPK models were linked for DDI simulations using oral clarithromycin and oral or i.v. midazolam. Model evaluation was accomplished predicting external concentration-time profiles of the individual compounds and the DDI. Modeling was performed using PBPK-Software PK-Sim® 5.2.2 and Mobi® 3.2.2.
Results: A time-dependent PBPK DDI model for clarithromycin and midazolam was developed. A CYP3A4 turnover model was applied dynamically accounting for clarithromycin dependent irreversible CYP3A4 (auto-) inhibition. Biotransformation of midazolam was dynamically linked to organ specific CYP3A4 concentrations. An adjustment to the distribution of clarithromycin into the red blood cells compartment was required to account for accumulation in macrophages and leukocytes. Organic anion-transporting polypeptide processes of clarithromycin were incorporated. Renal clearances were estimated using urinary excretion data. Using final model parameter values the DDI model was able to describe and predict the increase in midazolam exposure due to increased degradation of CYP3A4 via clarithromycin.
Conclusion: A whole body PBPK DDI model of clarithromycin and midazolam was successfully developed using clarithromycin-dependent inhibition of CYP3A4 in a time-dependent manner. The model allows predictions of plasma concentration-time profiles of clinically relevant midazolam doses after clarithromycin administration. The presented model may serve as a valuable future tool to describe and predict time-dependent drug-drug-interactions.
Reference: PAGE 23 () Abstr 3222 [www.page-meeting.org/?abstract=3222]
Poster: Drug/Disease modeling - Absorption & PBPK