A MULTI-ORGAN MPS-PBPK WORKFLOW FOR PROFILING CYP2D6 POPULATION ACTIVITY THROUGH DEXTROMETHORPHAN METABOLISM

New approach methodologies (NAMs) are emerging as powerful alternatives to accelerate and improve preclinical drug discovery 1,2. Conventional in vitro and in vivo models used to estimate absorption, clearance, and oral bioavailability often lack translational reliability, resulting in poor prediction of clinical dosing for Investigational New Drug applications 3,4. While physiologically based pharmacokinetic (PBPK) modelling can help bridge this gap by integrating preclinical data into a systems-level framework, its predictive performance is fundamentally limited by the quality, mechanistic scaling, and human relevance of the input parameters 5. This work presents a microphysiological system (MPS) to PBPK workflow, where ADME parameters can be successfully extracted from MPS time course data and used to predict clinical dosages through PBPK modelling. Cytochrome P450 2D6 (CYP2D6) is a highly polymorphic enzyme exhibiting substantial interindividual variability in metabolic activity 6. Consequently, drugs metabolised by CYP2D6 frequently require dose adjustment based on patient-specific activity scores. Accurately capturing this variability in preclinical systems remains a key challenge for predicting human pharmacokinetics and informing dosing strategies. In this study, a primary human Gut/Liver MPS was utilised to generate human-relevant parameters. This system utilises primary cells using RepliGut® Planar jejunal epithelial monolayers and liver microtissue derived from human hepatocyte donors.  Experimental concentration-time profiles from the MPS were used to construct an in silico compartmental model, enabling estimation of key parameters including intestinal absorption, hepatic intrinsic clearance, first-pass extraction and oral bioavailability. These experimentally derived parameters were subsequently scaled and incorporated into a PBPK model for prediction of clinical pharmacokinetics and dose requirements. Primary human hepatocyte donors representing a range of CYP2D6 activity scores were incorporated to enable direct assessment of genotype-phenotype relationships. Dextromethorphan was selected as a model CYP2D6 substrate, and both parent compound and metabolite formation were quantified to characterise metabolic activity across hepatocyte donor groups. Distinct activity-dependent metabolic profiles were observed, with clear differences in intrinsic clearance and metabolite formation rates corresponding to CYP2D6 activity score. Integration of gut and liver compartments enabled simultaneous assessment of first-pass metabolism and systemic exposure, providing a more physiologically relevant representation of oral drug disposition compared to conventional static systems. The experimentally derived parameters were subsequently scaled and incorporated into a PBPK model to predict human pharmacokinetics. Model outputs demonstrated good agreement with clinical datasets across CYP2D6 phenotypes, supporting the translational relevance of the generated data. This combined MPS-PBPK approach improves the prediction of human pharmacokinetics and captures CYP2D6-driven interindividual variability. This work highlights the potential of NAMs and the Gut/Liver MPS to enhance dose prediction, support precision-medicine strategies, and reduce reliance on animal models in drug development.

USING A GUT/LIVER MICROPHYSIOLOGICAL SYSTEM TO PREDICT DRUG-DRUG INTERACTIONS IN HUMANS

Drug-drug interactions (DDIs) are a critical consideration in patients with comorbidities, where the use of multiple medications is common and can significantly affect therapeutic efficacy and safety. Interactions involving metabolising enzymes and drug transporters can alter oral bioavailability, systemic exposure, and ultimately clinical outcomes 1. Despite their importance, predicting DDIs prior to first-in-human studies remains challenging, particularly where both intestinal absorption and hepatic metabolism contribute to compound disposition 2. Microphysiological systems (MPS) that integrate multiple human-relevant tissues offer a promising approach to address this gap. [YA1.1] Such systems are particularly valuable when they enable the effects of DDIs to be delineated both in individual organs separately and within a more complex Multi-Organ system. Here, we present a Gut/Liver MPS comprising Repligut® Planar primary jejunal epithelial cells and liver microtissues derived from human hepatocyte donors. This platform of primary cells enables mechanistic investigation of clinically relevant DDIs by recapitulating key aspects of oral absorption and first-pass metabolism, allowing simultaneous assessment of enzyme and transporter-mediated interactions. DDIs involving CYP3A4 substrates were successfully modelled within the Gut/Liver MPS. [YA2.1]Inhibition of CYP3A4 by ritonavir, which led to a marked increase in the oral bioavailability of darunavir, consistent with the well-characterised in vivo boosting effect used clinically to enhance HIV protease inhibitor exposure 3. [YA3.1]Conversely, induction of CYP3A4 activity using rifampicin resulted in accelerated metabolism of dabrafenib, reducing its bioavailability[YA4.1] and demonstrating the system’s ability to capture induction-driven clearance mechanisms relevant to oncology therapeutics. Development of an in silico compartmental model from the time-course data generated from these MPS experiments can be utilised to predict key ADME parameters including hepatic clearance, intestinal absorption and oral bioavailability with and without DDI. Importantly, the integrated Gut/Liver MPS enabled the contribution of intestinal and hepatic CYP3A4 activity to be evaluated independently while also revealing the overall effect on systemic drug disposition. In addition to metabolic DDIs, transporter-mediated interactions were investigated. P-glycoprotein (P-gp) inhibition kinetics were characterised using the P-gp substrate and prodrug dabigatran etexilate in the presence of the inhibitor zosuquidar. The system enabled quantitative assessment of altered transport and absorption dynamics within Repligut® Planar primary human epithelial intestinal cells. Furthermore, carboxylesterase 1 and 2 (CES1/2) activity was successfully quantified through detection of dabigatran etexilate metabolites and formation of the active drug, dabigatran, demonstrating the model’s capability to resolve prodrug activation pathways alongside transporter effects. Collectively, this work highlights the utility of a physiologically relevant Gut/Liver MPS for integrated assessment of enzyme- and transporter-mediated DDIs. The platform provides mechanistic insight into complex interactions affecting oral drug disposition and supports improved prediction of clinical outcomes. Such approaches have the potential to inform dose selection strategies, mitigate adverse interactions, and reduce reliance on animal studies during preclinical development.[
“

INTER-LABORATORY REPRODUCIBILITY OF MIDAZOLAM ORAL BIOAVAILABILITY ASSESSMENT USING GUT/LIVER MICROPHYSIOLOGICAL SYSTEMS AND MECHANISTIC MODELLING

Evaluating human oral bioavailability (F) remains a critical challenge in drug discovery and development. Microphysiological systems (MPS) offer a promising approach for capturing the intestinal absorption and first-pass metabolism processes that govern oral drug exposure. However, broader adoption of these technologies requires demonstration of reproducible performance across laboratories and compatibility with different biological model systems, and quantitative translation of experimental outputs into pharmacokinetic parameters. Here, we investigated the reproducibility and quantitative translatability of a Gut/Liver MPS workflow using midazolam, a well-characterised CYP3A substrate undergoing both intestinal and hepatic metabolism.
Two independent studies were conducted in separate laboratories using a fluidically coupled Gut/Liver MPS platform. In the first study, the intestinal compartment comprised a primary human intestinal epithelial model from crypt stem/progenitor cells isolated from the jejunum (RepliGut®), alongside a parallel Caco-2 monolayer control. While in the second study, an induced pluripotent stem cell derived small intestinal epithelial model (hiSIECs) was evaluated using the same approach, together with a parallel Caco-2 monolayer. Both studies incorporated a 3D liver microtissue, generated from primary human hepatocytes (PHH) from different donors and employed comparable experimental workflows to simulate oral and intravenous drug administration. Midazolam and its primary metabolite, 1′-hydroxymidazolam, were quantified by LC-MS/MS from system samples collected over 72 hours.
Mechanistic mathematical models integrating intestinal absorption, intestinal metabolism and hepatic disposition were developed to describe parent drug disposition and metabolite formation, enabling estimation of key pharmacokinetic parameters including intestinal permeability, gut and hepatic intrinsic clearance, and the oral bioavailability components Fa, Fg, Fh and F. Model fitting and parameter estimation were applied independently to each dataset.
Despite differences in intestinal cell source, PHH donor and independent execution of the studies, highly comparable concentration-time profiles were observed for both midazolam and 1′-hydroxymidazolam. Mechanistic model analyses generated consistent estimates of oral bioavailability and first-pass extraction parameters, indicating similar contributions of intestinal and hepatic metabolism across the two experimental systems. Predicted oral bioavailability values were comparable between studies and remained within the range reported clinically for midazolam1.
These findings demonstrate that MPS-derived estimates of oral bioavailability can be reproduced across laboratories and intestinal model systems, supporting the broader implementation of Gut/Liver MPS platforms as robust and transferable tools for quantitative ADME assessment and translational pharmacokinetic prediction.
“