MPS World Summit 2026

Maximizing data value in MPS ADME experiments using In Silico modeling

Microphysiological systems (MPS) are increasingly being used to generate human-relevant ADME data, but experimental design is often constrained by limited working volumes, cost, dosing strategies, and in the context of multi-organ systems capturing the kinetics of combined process like absorption and metabolism. To address this and maximize the cost efficiency of MPS-based experiments, we have developed a computational workflow to plan PhysioMimix® ADME studies by combining mechanistic simulations and information-based design.

We designed an application using MATLAB® which simulates compound transport and metabolism across compartments in PhysioMimix® plates using ordinary differential equations (ODEs). The application predicts MPS concentration-time profiles from preclinical parameters, allowing users to make informed decisions about dosing schedules and concentrations.

Optimal sampling times are determined by a machine learning optimizer using an E-optimal objective function [1]. E-optimality ensures that poorly estimated parameter combinations are as well identified as possible. To address uncertainty in input parameters we utilize Monte Carlo simulations, drawing parameters from a uniform distribution around nominal values. Aggregating the information metrics across these simulations yields robust sampling recommendations that reflect the plausible variability in the parameters.

The optimized sampling schedule output by the application was evaluated on 20 theoretical compounds spanning a range of permeabilities, efflux ratios and clearance rates. Compared to the naïve case, where sampling times were evenly spaced, we found an increase in parameter confidence. As expected, when absorption and clearance were fast, earlier timepoints were favored. However, when fast and slow dynamics were combined, the optimal sampling times were distributed to capture both dynamic regimes. This highlights the benefit of using optimized sampling times that adapt to a compound’s kinetic profile, allowing us to capture data for different parameters from a single experiment.

In summary, this workflow uses In Silico modeling to support decisions in MPS ADME planning, reducing trial-and-error and increasing confidence in human parameter estimation from limited, high-value datasets. Ultimately, use of the application lowers the experimental burden and cost for end users by enabling more targeted and efficient MPS study design.

References Manesso, E., Sridharan, S., and Gunawan, R. (2017). Processes 5, 63-78. doi:10.3390/pr5040063

Presenters: Morné Van Wyk

Validation of a high-throughput MPS for 3D culturing of primary hepatocytes

The uptake of, and demands placed upon, microphysiological systems within a broadening range of fields is increasing at an ever-accelerating pace. Increasing experimental throughput, whilst maintaining consistency and reliability, is critical in supporting the expanding role this technology is playing in the space between conventional 2D cell culture and in vivo studies. CN Bio’s PhysioMimix®  System has been used in many laboratories to recreate organ-specific microenvironments in which 3D cell cultures are continually perfused with media by means of pneumatically-driven pump systems, allowing for extended study of viable, functional tissues.In this study we present the development of a higher-through MPS plate, Liver-48, for the provision of a physiologically relevant environment for the 3D culturing of primary liver cells that shows great equivalence to that of the well-established, and widely-adopted, Liver-12 plate, but with an increased experimental capacity within the same footprint. By miniaturisation of Liver-12’s key features, whilst retaining key cell-adjunct architecture, the Liver-48 can support the reliable pre-existing microenvironment with a fourfold increase in experimental replicants, all without necessitating additional laboratory space.By utilising thermal mass flow measurement we show that microfluidic flow conditions are well-controlled and consistent both across and between Liver-48 plates. Performance is validated by extended culturing of primary human hepatocytes from multiple donors within the plate, assessing tissue health and function by soluble biomarker analysis, and gaining further confirmation from endpoint imaging and analysis following the experiment’s conclusion. From these data we show a low level of variance in the fluidic environment provided to  hepatocyte cultures, resulting in consistent culturing of primary cells and a low distribution of analytical data. We are able to show that performance is translatable from the standard Liver-12 plate to the new Liver-48, but with a far richer dataset able to support more complex and reliable experimental matrices.Both the physical and biological validation steps demonstrate that the Liver-48 plate maintains a common environment with the prevalent Liver-12, resulting in well-maintained hepatocyte viability and functionality. The increased number of wells per plate results in a fourfold expansion of data points per experiment, allowing for facile scalability of existing assays.

Recapitulating Immune-Driven Hepatotoxicity Using a Liver Microphysiological Platform

Background

With the April 2025 U.S. Food and Drug Administration (FDA) roadmap promoting New Approach Methodologies (NAMs) for regulatory testing of monoclonal antibodies (mAbs) and other drugs, there is a need for models with greater complexity and human relevance. For biologics such as mAbs, immune-mediated drug-induced liver injury (iDILI) remains a significant concern in drug development. Human-specific immune system interactions with new modalities, such as therapeutic antibodies, can lead to unpredictable hepatotoxicity, which often fail to be captured in traditional animal models and high-throughput cell line models, leading to poor translational outcomes and high attrition rates. The FDA-recognised PhysioMimix® DILI assay was modified to incorporate peripheral blood mononuclear cells (PBMCs) to address this with an immune-competent, human liver relevant MPS model. The model aims to enhance translational relevance and improve safety profiling in drug development pipelines.

Methods

Primary human hepatocytes were cultured under dynamic perfusion in Liver-12 plates by PhysioMimix Core to form 3D liver microtissues. The liver microtissues were analysed for functionality (CYP activity, albumin, urea) and clinically relevant injury markers (ALT, AST). Following four days of hepatocyte culture, HLA-matched PBMCs were added to the circulating media with the liver microtissues for a further four days and functionality assessed (cytokine profiling). Ipilimumab (anti-CTLA-4) or infliximab (anti-TNF-alpha) was dosed into the media of hepatocyte only and co-cultures for 48 hours and samples were taken at 1-, 4-, 8-, 24- and 48-hours post-dosing. The liver microtissues were evaluated for hepatotoxic responses while immune activation and inflammation was assessed by cytokine release.

Results

Immune cell interaction and recovery from the Liver-12 was first completed to determine the ability of the PBMC to transverse the microfluidic channels and scaffolds unhindered (60-70% recovery). Following initiation of co-culture, maintenance of liver microtissue functionality and stable immune phenotypes was demonstrated over four days of co-culture. Upon dosing with the mAbs in the co-culture model, immune-mediated hepatotoxicity was detected with significant elevation in LDH and ALT/AST, and reduction in the functional biomarkers, albumin and urea. Interestingly, significant cytokine

increases were only seen in ipilimumab-treated conditions, aligning with clinical data which demonstrated a pro-inflammatory phenotype via activation of T-cells.

Conclusion

This proof-of-concept study using two clinically relevant monoclonal antibodies demonstrates the ability of the PhysioMimix DILI assay to predict immune-mediated hepatotoxicity. Future studies will look to further profile immune alterations and interactions with the liver microtissues. Together, this assay provides a human-specific solution to the assessment of hepatotoxicity of large molecules. The FDA’s endorsement of NAMs for investigational new drug (IND) applications marks a paradigm shift toward human-relevant, non-animal testing strategies, positioning the PhysioMimix® Core System as a pivotal tool for the future of immunotoxicity assessment and safer drug development.