Jialin Mao (1), Utkarsh Doshi (2), Albert P. Li (2), Matthew Wright (1) and Yuan Chen (1)
(1). Department of Drug Metabolism and Pharmacokinetics, Genentech, A Member of the Roche Group.1 DNA Way, South San Francisco, California, CA 94080, USA (2). In Vitro ADMET Laboratories Inc (IVAL), 9221 Rumsey Road, Columbia, MD 21045, USA
Objectives: In the field of in vitro –to-in vivo (IVIVE) translation of CYP3A induction, there is increasing understanding of how a PBPK approach helps to predict the magnitude of drug-drug interactions (DDI) caused by strong CYP3A inducers such as rifampicin [1]. In drug development, many drug candidates are identified as weak-moderate inducers on the basis of in vitro data. Although there are reports using a static model to predict the CYP3A driven DDI magnitude[2], the translation of in vitro data to in vivo using a PBPK approach is not well established. Therefore, there is a need for the systematic investigation beyond strong inducers or inhibitors. This work focuses on how the magnitude of DDI caused by the moderate CYP3A inducers can be predicted using a PBPK approach.
Methods: Three CYP3A moderate inducers armodafinil, pleconaril and modafinil were selected for this investigation. In Vitro: A range of incubation concentrations was tested in plated human hepatoctyes for 3 and 5 days, including the positive control rifampicin. The concentration of each inducer was measured at the end of the in vitro study, and incorporated in generating the EC50 and Emax based on both mRNA and activity measurement. PBPK: The PBPK model was constructed in Simcyp® version 15 using a mixed “bottom-up” and “top-down” approach for each inducer. The predicted PK profile was verified using clinical PK data. The in vitro Emax and EC50 were then incorporated into the models to predict the reported clinical DDI related to armodafinil, pleconaril and modafinil, specifically between CYP3A substrate midazolam (IV and PO) and armodafinil, midazolam (IV and PO) and pleconaril, and CYP3A substrate triazolam (PO) and modafinil.
Results: The constructed PBPK models were able to predict the PK profile of armodafinil, pleconaril and modafinil at the dose where the clinical DDI studies were conducted. For the induction-based DDI prediction, it is observed that the predicted magnitudes of DDI were closer to the clinical observations for all three drugs and related studies when 1) the measured inducer concentrations were used to generate the in vitro Emax and EC50 data (considering the loss of the inducer during the incubation time) and 2) the activity/mRNA data were expressed relative to the rifampicin data (calibration with the positive control). The Emax/EC50 generated with 3-day treatment provides better prediction compared those generated with 5-day treatment. In addition, the activity data demonstrated a better IVIVE than the mRNA data.
Conclusion: The current work demonstrated an approach that successfully predicted the magnitude of DDI caused by moderate CYP3A inducers. It also demonstrates the utility of PBPK modelling to provide recommendation on possible in vitro study designs. It also provided some insights to in vitro study design, including duration of hepatocyte treatment, monitoring of concentration loss and use of positive control, which could increase confidence in the DDI prediction using PBPK approach.
References:
[1]. Almond, L.M., et al., Prediction of Drug-Drug Interactions Arising from CYP3A induction Using a Physiologically Based Dynamic Model. Drug Metab Dispos, 2016. 44(6): p. 821-32.
[2]. Fahmi, O.A. and S.L. Ripp, Evaluation of models for predicting drug-drug interactions due to induction. Expert Opin Drug Metab Toxicol, 2010. 6(11): p. 1399-416.
Reference: PAGE 27 (2018) Abstr 8438 [www.page-meeting.org/?abstract=8438]
Poster: Drug/Disease Modelling - Absorption & PBPK