Markus Krauß, Sebastian Wertz, Michaela Bairlein, Michael Gerisch, Frank Hucke, Mark Jean Gnoth
Department of Drug Metabolism and Pharmacokinetics, Bayer AG
Objectives:
Over the last years, a new technology allowed the creation of chimeric mice with a humanized liver consisting of human hepatocytes surrounded by phagocytic, vascular and biliary murine sourced cells. Chimeric mice with a humanized liver can be used to predict human ADME characteristics of new chemical entities such as overall clearance and elimination mechanisms to improve nonclinical to clinical translation. Data from such mice might be helpful especially in cases where in vitro to in vivo extrapolation (IVIVE) is challenging and classic allometric scaling with wildtype nonclinical species results in a wide prediction interval. In a retrospective meta-analysis using in-house data from First-in-Human (FiH) studies, we compared the predicted human clearance and area under curve (AUC) as well as the shape of human pharmacokinetic (PK) profiles based on data from chimeric mice with those predictions from wildtype nonclinical species.
Methods:
Several drugs were incorporated in the analysis where FiH data were available for intravenous administration. Total clearances in humans were predicted by a single-species allometric scaling approach from data obtained in chimeric mice with humanized liver (FRG® knockout mice, Yecuris Corporation). Results were compared to clearance prediction using classical allometric scaling from wildtype nonclinical species and observed clearance in humans. For selected compounds prediction of concentration-time profiles in humans was also performed with single-species scaling using a compartmental modeling approach. Resulting PK parameters such as AUC as well as the shape of predicted concentration-time profiles were compared with human experimental data and predictions from wildtype nonclinical species.
Results:
Overall, superior prediction of human clearance was achieved after single-species allometric scaling using data from chimeric mice. For seven out of ten compounds human clearance prediction was within a twofold range. In comparison, applying classical allometric scaling prediction for six out of ten was within a twofold range. In addition, in three projects human PK was predicted from chimeric mice data using a two compartment model, where a broad AUC interval was predicted from nonclinical species due to interspecies differences. Resulting predictions were at least as good as from the nonclinical species that predicted best the observed human data, especially regarding AUC, but also the PK profile. Notably, the nonclinical species best predicting human PK differed in the three projects.
Conclusions:
The presented results demonstrate that single-species-scaling from chimeric mice with humanized liver predicts human clearance quite well and can also lead to valuable prediction of human pharmacokinetics in general. Especially in projects where a broad range of human CL based on wildtype nonclinical species was predicted and when IVIVE is challenging, chimeric mice clearly strengthen human ADME and PK prediction. Therefore, chimeric mice with a humanized liver can expand the toolbox to improve prediction of human ADME properties, especially by further derivation of mass balance data and information about potential human metabolites. Such data can then be incorporated into mechanistic modeling approaches to further improve predictivity of human PK in future early drug development programs.
Reference: PAGE 28 (2019) Abstr 8988 [www.page-meeting.org/?abstract=8988]
Poster: Methodology - Other topics