If you missed part one of this blog series, click here to read A new era of drug development: How regulatory changes to reduce animal testing compare by region.
Introduction
In part 1, we explored how regulatory roadmaps from the US, UK and EU are accelerating the shift away from animal testing in favour of New Approach Methodologies (NAMs). But where do microphysiological systems fit within these evolving frameworks?
In Part 2, we explore how microphysiological systems, also known as Organ-on-a-chip models, generate mechanistic insight into human biology, addressing known limitations of traditional animal and in vitro approaches. As global regulators shift towards defined contexts of use and data quality, MPS will become a critical non-clinical drug development tool.
We will focus on specific microphysiological system contexts of use examples within safety and pharmacokinetics (PK) – as this is where regulators examine data packages most closely before Phase I studies. However, although preclinical packages require very little evidence of efficacy (this is typically evaluated within Phase II & III human trials), it remains the leading reason for drug failures in the clinic. And, since Phases II and III are also the most expensive in drug development, using MPS to de-risk earlier before human testing offers a strategic advantage.
Our examples show how microphysiological systems (MPS) support weight of evidence (WoE) strategies and enable earlier, more confident decision-making ahead of first-in-human studies. However, we will also touch upon the broader use of MPS within other hazard and risk assessment areas, using food safety as an example. It is important to remember – regulators are not looking for 1:1 replacements of animal tests, they are looking for data that answers specific decision-making questions where MPS are fit-for-purpose.
Microphysiological systems for drug safety testing
When safety risks are complex, latent or mechanistically unclear, MPS provide the human-relevant insight to de-risk decisions before human trials.
As the primary organ of drug metabolism, the liver is central to safety assessments; however, many complex and latent toxicities, such as cholestatic or idiosyncratic drug-induced liver injury (DILI), require more insight than traditional in vitro models can provide. In addition, toxicity can present differently in animals, creating gaps in safety assessment due to interspecies differences.
An important consideration when choosing a Liver MPS (or Liver-on-a-chip) to study DILI is how many pathways and endpoints can be explored from each experiment, and how long cultures remain viable. Longer‑lived, mechanistically informative models that allow acute, repeat dosing, and immune‑competent responses add value by filling data gaps that animal models cannot easily address to provide greater assurances of safety.
1. Predicting and understanding the mechanism of human drug-induced cholestasis
Drug‑induced cholestasis involves complex, multi‑mechanistic bile acid dysregulation and is a common cause of DILI. It is clinically important because cholestatic DILI is often prolonged or chronic.
Existing models fail to reliably predict DILI, creating a regulatory and translational blind spot. Animal models are limited by species-specific differences in bile acid differences, and traditional in vitro assays lack core components such as:
- High bile‑acid transport and metabolism fidelity
- Stable long‑term function
- A broad panel of cholestasis-relevant endpoints
- The ability to perform transcriptomics
MPS are a valid primary option for de-risking workflows. However, a recent publication demonstrates that the ability and sensitivity of different MPS technologies to detect cholestasis vary. A comparative study by Nitsche et al., (2025) highlighted that only CN Bio’s PhysioMimix® Liver MPS (primary human hepatocytes or HepaRG cells) detected consistent bile‑acid reductions and early mechanistic changes across cholestatic toxicants, including bosentan and α‑naphthyl isocyanate. This DILI assay used PhysioMimix Core’s Liver-12 plates, to measure bile acids in the media (via kits or LC-MS). The bile acid synthesis, metabolism and transporter expression were explored via ~omics-based analysis of the recovered microtissue.
This challenge highlights why regulators increasingly prioritize human relevance and a clearly defined context of use over default animal testing. The most physiologically relevant MPS are emerging as a critical de‑risking tool within a WoE framework.
Learn more about the PhysioMimix DILI assay here
2. Use of microphysiological systems to predict or investigate conflicting in vivo data
When animal and in vitro data conflict, use MPS to determine which signals are human predictive before a costly decision is made
There are numerous examples in the literature where decisions based on animal data have resulted in post-approval withdrawal due to severe liver injury in humans. To de-risk effectively, it is important to remember that NAMs don’t have to be human, nor do they need to replace an animal model to be useful. Comparative cross-species MPS studies inform in vivo study design or clarify which animal species better predicts the human response where animal data conflict. Both the FDA in their roadmap and the IQ MPS Affiliate have highlighted that animal MPS are an important tool for building confidence in the in vitro to in vivo extrapolation (IVIVE) of MPS before refining, reducing, or replacing in vivo studies.
These scenarios require cross-species MPS with:
- Consistent culture conditions and stable function over more than 10 days
- The mechanistic ability to detect inter‑species divergence
- Clinically relevant biomarker reporting
- The ability to perform transcriptomics
Our PhysioMimix Application Note explores the use of human, rat and dog Liver MPS-based DILI assays to unlock interspecies differences in DILI response using known examples. These include tolcapone, buspirone and troglitazone, and their safe enantiomers to demonstrate the assay’s sensitivity, specificity and utility.
Download our Application Note: Enhanced IVIVE with cross-species Liver MPS DILI assays
To show the benefits of MPS over traditional approaches, Negi et al., (2025) compared primary hepatocyte function and drug effects in human, monkey, dog, and rat using a PhysioMimix Liver MPS versus 2D assays. The authors showed that the MPS maintained longer‑term liver microtissue function and captured species‑specific toxicity responses that simple 2D systems could not. While MPS is more expensive and has lower throughput, the number of drugs requiring this level of scrutiny is nominal compared to the value of the deeper insights generated.
The translational clarity and WoE offered by MPS arguably justify the cost of supporting more confident clinical development, preventing the loss of safe assets from the pipeline and reducing unnecessary animal use.
3. Use of microphysiological systems to predict immune-mediated toxicities
Building on this foundation, we are now tackling a key unmet need highlighted in the regulatory roadmaps: assessing immune‑mediated events associated with new modalities such as mAbs.
Traditional animal models are often poor at assessing human‑specific immune‑mediated risks, especially for complex therapeutics like siRNA or cell and gene-based therapies.
Our first step was to integrate human peripheral blood mononuclear cells into our FDA‑recognized liver MPS DILI assay, to create an immune‑competent platform that captures human‑specific inflammatory and immune‑driven liver injury mechanisms. In our poster at Society of Toxicology, we showed proof‑of‑concept studies using clinically relevant antibodies to demonstrate the detection of immune‑mediated hepatotoxicity alongside cytokine‑driven responses. The results are aligned with clinical observations. We have also demonstrated similar principles for lung models.
This work aligns directly with the latest FDA and MHRA guidance. Both agencies recognize the limitations of animal models for human‑specific immune responses and encourage the use of fit‑for‑purpose NAMs to support a more confident and targeted reduction in animal use for human‑specific therapeutics.
ISTAND program: A clear insight into the regulatory use of micrphysiological systems for DILI assessments
In collaboration with the FDA‑CDER, NIH‑NIEHS, and the 3Rs Collaborative, CN Bio is actively contributing to a cross‑platform regulatory evaluation of commercial Liver MPS, which has been accepted into the FDA’s iSTAND program. This blinded, multi‑site study of eight liver MPS uses matched hepatotoxic and non‑hepatotoxic compounds dosed across clinically relevant exposure ranges. Using primary human liver microtissue under continuous perfusion, liver function, injury and metabolic competence have been assessed over extended dosing. All raw data have been independently analyzed for transparent and standardized comparisons with publications to follow. This initiative provides a critical framework for building regulatory confidence and advancing MPS qualification for DILI risk assessment. Initial data can be viewed in this poster presented at the Society of Toxicology 2026
4. Use of liver microphysiological systems for evaluating genotoxicity and mutagenicity in vitro
The area of genotoxicity represents another area specifically called out for change in the European Commission’s accompanying Staff Working document.
Current pre-clinical drug safety assessments lack a single, comprehensive test system for genotoxicity hazards identification. In 2024, our collaborators at Charles River Laboratories (Kopp et al) published “Liver-on-a-chip model and application development in predictive genotoxicity and mutagenicity of drugs”, published their findings from a study aimed to develop an in vitro model to address this predictive genotoxicity gap using PhysioMimix technology. Their findings demonstrated robust human metabolic activity and the potential to assess all required endpoints for genotoxicity hazards identification:
- DNA strand breaks (comet assay) in hepatocytes
- chromosome loss/damage (micronucleus assay) in TK6 cells,
- mutation (Duplex Sequencing) in TK6 cells
Their proof-of-concept study demonstrates the potential of the PhysioMimix Liver-on-a-chip model (Liver MPS) for in vitro genotoxicity hazard identification, paving the way for a more streamlined and animal-free pre-clinical drug safety assessment process.
Microphysiological systems for pharmacokinetics and ADME profiling
MPS enable more reliable predictions that directly inform dose selection and progression decisions.
There are many in vitro and in silico tools for small molecule PK profiling that are fit-for-purpose; however, they often fall short in scenarios that require longer experimental windows. For example, for studies on clearance beyond a week, or enzyme induction (which typically requires multiple dosing cycles), we need novel ways to reduce reliance on animal models where interspecies differences in metabolism and transport limit their ability to predict human PK reliably. There is also a need for more sophisticated systems to assess the individual and combined effects of first-pass metabolism by the gut and liver to improve oral bioavailability predictions, where animals are known to have poor alignment (Musther et al., 2014)
These unresolved questions require a different approach. Complementary perfused Liver MPS offers the highest possible metabolic relevance and dynamic flow conditions that extend performance over weeks. They also enable human gut and liver models to be interconnected to simulate the process of first-pass metabolism (Abbas et al., 2025).
Newer drug modalities also bring a unique set of exposure-relevant PK challenges. For those that have highly human-specific targets, like Antisense oligonucleotides (ASOs), a more human-relevant approach is required. Here, MPS models can be used to decipher the mechanisms of uptake and delivery (Majer et al., 2024) as well as provide an insight into pharmacodynamics/efficacy via gene knockdown.
By using MPS alongside traditional approaches, the translatability gap between preclinical PK data and clinical outcomes can be narrowed. This improves human predictivity and more confident first‑in‑human dose setting whilst reducing reliance on animal studies, where they offer limited translational value.
Read our blogs to learn more about the ADME/PK challenges that can be addressed through MPS use.
Learn more about our MPS-based ADME assay solutions
Disease modeling and efficacy: where microphysiological systems add strategic value
While not always required by regulators, MPS provides early evidence of human efficacy for asset prioratization and to help avoid costly late-stage failures.
Although very little evidence of efficacy is required in preclinical packages, if you dig a little deeper, a large proportion of UK and NIH funding supporting the reduction in animal use in science is to accelerate the development and validation of critical disease models. In the UK, metabolic dysfunction-associated steatohepatitis (MASH) is one of five diseases specifically called out.
In a recent blog, we discussed the challenges in modelling complex metabolic diseases such as MASH and insulin resistance. Interspecies differences in animals and oversimplified in vitro systems fail to adequately replicate disease pathophysiology which has hindered progress. It includes case studies showing how human‑relevant MPS disease models using the PhysioMimix platform provide critical workflow advancements, including how MPS data supported a successful regulatory filing for Inipharm’s INI-822.
Examples of CN Bio’s fully validated disease models can be found here. Given the sheer number and diversity of human diseases, no single platform is pre‑validated for every indication. That’s why the PhysioMimix Core is built for adaptability, allowing you to configure disease models for specific human‑relevant mechanisms and regulatory questions, rather than predefined disease labels. This approach aligns with regulatory moves toward context‑of‑use–driven, human‑relevant NAMs that reduce risk. However, this scenario reduces the risk of drug failure due to poor efficacy in humans versus their safety.
Use of microphysiological systems to assess environmental health risks
Whilst the FDA and MHRA are focused on NAMs use to reduce and eventually replace animal testing within drug development, the European Commission is taking a broader systems-level approach—reshaping how all chemical risks, including pharmaceuticals, are assessed without animals.
Read more about how multi-organ PhysioMimix models are being utilized in a European Food Safety Assessment (EFSA) project to translate complex biological data into clear, usable outputs for hazard and risk assessment at Luxembourg Institute of Science and Technology (LIST).
Summary
Recent regulatory announcements from the USA, UK and Europe to reduce, refine, and replace animal testing make it clear that the goal is to generate more predictive data that supports more confident decision making, not to replace animal tests in a like-for-like manner.
Here, we have demonstrated that MPS delivers the greatest value when applied to decision-critical gaps where traditional approaches fail to provide clear human-relevant insight within drug development. Whilst most models will tell you that something is wrong, MPS goes further, helping you to understand why early enough to change decisions. This includes complex safety liabilities, reducing translational uncertainty and answering more complex ADME/DMPK questions to de-risk.
MPS does not claim to be a standalone replacement; instead, they are most effective when used as part of a weight-of-evidence approach to strengthen data packages and reduce uncertainty before first-in-human studies. Ultimately, the advantage to developers is not only the ethical desire to reduce animal use, but to make earlier, better-informed decisions that improve confidence in which candidates to progress, and which to stop.
Within the EU, the scope for MPS use is broader following the EC announcement in June 2026. The cited EFSA project, which aims to translate complex biological data into something regulators can use for food safety assessment, showcases the growing contexts of use for which MPS technology can be applied as we transition towards more human-centric testing.
Find out more about PhysioMimix:
Organ Models
PhysioMimix® Organ on a chip models
Applications
Organ on a chip applications
Services
