Investigation of Diazepam Sequential Metabolism with Hepatocytes via a Microfluidic Device

Abstract

Objectives and hypothesisThe purpose of this study is to investigate the sequential metabolism of diazepam (DZP) in a dynamic system. In rat liver, DZP can be metabolized into primary metabolite temazepam (TZP), nordiazepam (NDP), p‐hydroxydiazepam (PHD), and secondary metabolite oxazepam (OXP) and temazepam glucuronide (T‐G). It is reported that a lag time of DZP metabolite formation can be observed in rat liver single‐pass perfusion and PK study. We are using a microfluidic device to represent liver sinusoid and to investigate the contribution of metabolism and drug transport in the disposition of diazepam and its metabolite along the sinusoid. Ordinary differential equation (well‐stirred model) and partial differential equation (PDE) models are being developed to characterize the disposition of DZP and its metabolites. We hypothesize that characterization of sequential metabolism in a dynamic system compared to a static system (microsomal incubation or static hepatocyte incubation) can provide a more accurate prediction of in vivo sequential metabolism.MethodsThe perfusion study was conducted with rat hepatocytes. Rat hepatocytes were seeded in collagen I coated microscopy slides to form a monolayer, followed by three‐day enzyme induction. An ibidi® μ‐sticky slide was mounted on the microscopy slide to create a microfluidic device. Hepatocyte culture medium containing 5 μM DZP was perfused through the device and subsequently perfused by DZP‐free culture medium. The concentration of diazepam and its metabolites in the perfusate was measured by LC/MS/MS. Mathematical modeling was applied to predict the concentration‐time (C‐t) profile of the perfusion study via Mathematica version 12.3 (software). The model parameters include metabolic intrinsic clearance (CLint,met) and passive diffusion clearance (CLdiff), partitioning into membranes, hepatocyte volume, and device volume. Other relevant model parameters were collected by literature search.ResultsAt 10 μL/min flow rate, three primary metabolites all show similar lag time (pre device dead volume plus device volume) to that of DZP. OXP shows a longer lag time compared to T‐G. The lag time of T‐G is similar to that of primary metabolites. However, a significant decline of TZP and NDP formation at a steady‐state is observed, but not PHD. One possible rationale is constant enzyme loss for CYP3A (responsible for TZP, NDP formation), but not CYP2D (responsible for PHD formation). For the DZP‐free perfusion phase, a similar lag time was observed for DZP and its metabolites. A well‐stirred model and a PDE model to characterize the dimension and spatial information are being optimized.ConclusionsGenerally, this microfluidic device can characterize the sequential metabolism of diazepam. The C‐t profiles of DZP and its metabolites are similar to that of the liver single‐pass perfusion study. Due to the existence of transporters, the lag time of T‐G is similar to that of primary metabolites. As a future study, the distribution of DZP and its metabolites along the device will be measured by MALDI imaging. The information of spatial distribution is expected to further help optimize and validate the mathematical models.