Historically, bioavailability estimates have relied heavily on animal studies, which are not only expensive but often fail to predict human in vivo outcomes accurately. Additionally, traditional approaches rely on isolated in vitro assays that also often fail to predict human drug absorption and metabolism accurately. This leads to costly late-stage failures with pharmacokinetics and bioavailability being the third most common cause of attrition (16% of failures) in Phase I clinical trials for compounds developed by major pharmaceutical companies (Waring et al 2015). Our blog explores how we are integrating in silico tools with Gut/Liver-on-a-chip to deliver more accurate bioavailability predictions that reduce our reliance on animal use.

The PhysioMimix^® advantage

At CN Bio, we’ve taken a novel approach to address these limitations by combining a primary human Gut/Liver microphysiological system (MPS, also known as Gut/Liver-on-a-chip) with advanced computational modeling. This integrated strategy bridges the gap between in vitro testing and in vivo outcomes, offering a more human-relevant alternative that:

• Improves prediction accuracy
• Reduces reliance on animal studies
• Delivers a more time- and cost-efficient process

Our PhysioMimix multi-organ Gut/Liver model uniquely allows for intestinal absorption and hepatic clearance to be studied in a single, interconnected system, unlike traditional approaches that assess these functions in isolation. This connectivity mimics human physiology and offers a holistic understanding of bioavailability.

A common concern researchers express is whether MPS platforms offer proven advantages over traditional models, and whether transitioning from familiar assays to organ-on-a-chip systems is too complex. This can be a daunting transition, but we are here to help you through it.

Our internal assay validation and adopter experiences have demonstrated that the PhysioMimix Bioavailability assay kit: Human 18 can be implemented without significant hurdles. The all-in-one Bioavailability assay kit contains everything you need to start running MPS experiments as a part of your pipeline in no time. However, we have noticed that certain questions around how to use the assay and its data consistently arise when corresponding with users.

Moreover, regulatory shifts are accelerating the adoption of New Approach Methodologies (NAMs) like the FDA Modernization Act 2.0. The recent announcement by the FDA on its decision to phase out animal testing requirements for monoclonal antibodies, with plans to extend these changes to other drug classes, is cementing MPS as a valuable approach in modern drug discovery pipelines.

Enhancing MPS with computational modeling

Integrating in silico tools with organ-on-a-chip workflows enhances the interpretation of our results and unlocks the ability to use MPS-derived parameters in downstream physiologically based pharmacokinetic (PBPK) modeling. The FDA has emphasized the growing importance of PBPK modeling in their roadmap to reduce animal testing in preclinical safety studies, highlighting the value of integrating experimental data with simulation tools.

Why mechanistic modelling matters

To facilitate the adoption of our advanced in vitro assays, we are developing, with a view to commercializing, computational models. Our computational models enable in silico simulation of experiments before they reach the wet lab. This predictive power allows us to optimize experimental design and refine dosing concentrations, sampling times, and other key variables, enabling our experimentalists to generate meaningful results more efficiently. Eliminating the guesswork in setting drug concentrations and time points for sampling, the PhysioMimix Bioavailability assay kit plus computational tools provides guidance at every step.

Our computational models, which are based on the mechanistic details underlying the Gut/Liver-MPS, allow us to extract meaningful ADME parameters from complex datasets, many of which would be difficult to quantify using traditional approaches. These can be used to better inform PBPK modeling, or in vitro to in vivo extrapolation (IVIVE).

From each replicate of an MPS experiment, we can now identify multiple pharmacokinetic parameters that would typically have required separate assays to determine. Mechanistic modeling also opens the door to deeper biological insights. It allows us to address critical questions like: “When does saturation occur?” or “At what concentration does the system show non-linear behavior?”.

By integrating in silico tools with organ-on-a-chip technology, we can answer these questions with greater precision at a fraction of the cost of the equivalent animal studies.

Integrating in silico tools with organ-on-a-chip: midazolam case study

In a peer-reviewed publication (Abbas et al., 2025), we demonstrate the workflow of using organ-on-a-chip data to determine pharmacokinetic parameters and the bioavailability of midazolam. From a single set of experiments, we were able to quantify the midazolam concentration over 72 hours. The data showed the compounds being absorbed, translocating the gut membrane and entering the liver compartment, where the liver metabolized most of the compound.

Following the acquisition of the experimental data, we developed mathematical models to describe the movement of midazolam throughout the MPS. This process allowed us to generate several feasible models with distinct assumptions, which were all fitted to the experimental dataset. Each model was then ranked according to a performance criterion and the best-performing model was chosen. Using Bayesian methods, we could determine the confidence intervals of key parameters such as intrinsic hepatic and gut clearance (CL_int,liver and CL_int,gut), apparent permeability (P_app) and efflux ratio (Er).

Using the output parameters from the model we were able to generate values for the components of bioavailability, F_a, F_g and F_h (the fraction absorbed, escaping the gut and liver metabolism, respectively) and, by finding the product of these three components, we estimated a value for oral bioavailability of midazolam. Our prediction fell within the observed range of clinical values.

Future outlook

Our method of estimating key ADME parameters allows for organ-on-a-chip experiments to be used in the pipeline of PBPK modeling, as the resulting parameters can be plugged into these more complex models to potentially inform first in human trials.

Conventional approaches to obtaining the necessary parameters for PBPK modeling can be time-consuming, inaccurate and expensive, especially when using animal models. Use of more relevant MPS-based experiments in conjunction with computational modeling offers a cheaper, more translatable method to elucidate important pharmacokinetic parameters of a drug. Further reducing animal studies in the process of lead optimization is another notable benefit of this method.

Our future work integrating in silico tools with organ-on-a-chip is dedicated to further validating the method to evidence the suitability of primary MPS-based assays for lead optimization and as a better alternative to historical methods.

Ready to Get Started?

It is no secret that there is hesitancy around adopting organ-on-a-chip technology into workflows due to experimental costs, relevant internal expertise to perform experiments and proof. However, with regulatory change on the horizon, the time is right to get started with your MPS journey. We have made the onboarding journey easier by developing an all-in-one kit that contains everything required to recreate our dual-organ MPS and bioavailability assay in your laboratory.

Additionally, we have developed and aim to commercialize the in silico tools described above. For now, you can access these tools by outsourcing to our ADME Contract Research Services. Our simulation software helps you to plan your MPS-based experiments, streamline your workflow and reduce experimental costs. The second tool provides an understanding of how to combine data generated using this fully human in vitro Gut/Liver model with in silico computational modeling to improve the quality of input parameters for PBPK modeling and more accurately estimate the in vivo pharmacokinetic parameters of orally administered drugs. A comprehensive overview of this workflow is published in Drug Metabolism and Disposition by Abbas et al., (2025).

How to rapidly integrate MPS into your workflow

There are a variety of options offered by CN Bio. If you are looking to modernize internal workflows, contact us for more information about our industry-proven PhysioMimix technology, plus Bioavailability assay kit: Human18. Alternatively, CN Bio also provides ADME Contract Research Services for experiments to be completed on our premises via our Team. These services can either be used to validate MPS for your specific context of use before investing in PhysioMimix technology in-house. Alternatively, you can outsource your project work to our team of experts to instantly benefit from MPS insights.

Related blogs and resources

Application notes

Connecting the gut and liver

News & Blogs

Evaluating Caco-2 cell’s gold-standard status in absorption, permeability and bioavailability Studies: Are Their Limitations Justified?

News & Blogs

How in vitro human Gut/Liver models increase confidence in ADME estimations before human trials 

In part one of the blog, we unpicked the reasons behind the FDA’s decision to focus on mAbs first, and steps to ensure that a NAM-based approach will offer the same degree of safety as animals before their use is phased out.

If you missed it, read Part one “Why has the FDA loosened its grip on animal testing for mAbs?” here.

In “Microphysiological systems for mAbs”, we continue the commentary by taking a deeper dive into the PhysioMimix^® OOC range of microphysiological systems, also known as organ-on-a-chip (OOC). We explore how microphysiological systems address the shortcomings of animal models for mAb development and the role that CN Bio is taking to support your switch to a NAMs-based approach.

Use of microphysiological systems for mAbs to address animal shortcomings during development

At CN Bio, we have developed, characterized, validated and commercialized many PhysioMimix organ models and assays that have been successfully adopted by 16 of the top 20 Pharma. They can be used to test a range of drug modalities across disease modeling, safety toxicology and ADME applications, and are supported by “all-in-one” reagent kits to support rapid onboarding. Our customers have independently developed many more.

In their roadmap to reducing animal testing in preclinical safety studies, the FDA focuses much of their narrative on immune-related risk and how microphysiological systems can add a “crucial safety net that animal tests struggle to provide”. In principle, use of primary human cell-derived models avoids interspecies differences to reveal immune-related toxicological effects that are more relevant to humans; however, modeling the human immune system is challenging in all experimental setups – not just animals.

There is no doubt that microphysiological systems with closed-loop microfluidic designs, which re-circulate media around cultures are best equipped for mAb immunogenicity testing. Our PhysioMimix Liver-on-a-chip, (also known as a Liver microphysiological system/Liver MPS) is the furthest advanced in terms of immunocompetency. A key customer has already reported the initial characterization of a fully immunocompetent drug-induced liver injury (DILI) model – including circulating peripheral immune cells alongside the liver microtissues – at the Society of Toxicology conference, demonstrating that the basic design of PhysioMimix technology is on point. The development of these fully immunocompetent microphysiological systems for mAbs development is on the horizon. With the ability to report inflammatory responses and resulting organ damage to drugs, they will be game changers!

From immediate effect, companies that include NAMs-based data into therapeutic submissions to the FDA will be rewarded, for example, toxicology studies using animal and NAMs methods which show no concerning signals in one month can be reduced from six months to three months. Through close collaboration and joint expertise, we are confident in our ability to supply effective microphyiological system-based assays required for these tests, so that you can benefit from the regulatory relief promised by the FDA.

CN Bio’s closely forged links to primary human cell suppliers, including LifeNet Health, position us well to address current market challenges surrounding immunogenicity testing, such as HLA donor matching of all primary cell types required. Plus, our latest collaboration deal with CRO giant Pharmaron provides us with a greater potential to meet market needs for microphysiological systems for mAbs testing more rapidly.

Immune-mediated liver injury associated with mAbs

There are four types of immune-mediated injuries associated with monoclonal antibodies:

Idiosyncratic Reactions

These are unpredictable and often not dose-related, meaning patients may develop DILI even at low drug doses.  

Autoimmune Hepatitis

mAbs can trigger an immune response that damages the liver. 

Other Immune-Mediated Injury

The primary effect of the mAb can trigger an immune or non-immune response that results in liver injury. 

Viral Reactivation

Immunosuppressive mAbs can reactivate latent viral infections, like HBV, which can cause liver damage. 

Use of concomitant medications, like aminosalicylates or methotrexate, can also increase the risk of DILI when used with mAbs, as does the presence of underlying liver disease and genetic predisposition. How can the use of microphysiological systems for mAbs testing reduce the risk?

Liver MPS’ potential to detect immune-mediated liver injury

Human Liver MPS (co-cultures of hepatocytes with non-parenchymal cells under perfusion) have human liver metabolic competency and recapitulate human-specific responses.  Our PhysioMimix Liver MPS has been extensively validated for predicting the DILI effects of small molecules, whilst proof of concept work evaluating their use for therapeutic antibody/small molecule drug-drug interaction responses, and the testing of newer, more human-specific drug modalities (where animal use is less suited) has also been demonstrated. Visit our DILI application page for more info.

A recent CN Bio blog, “Immune-mediated DILI – Predicting the unpredictable!, explores the additional benefits of a fully immunocompetent Liver MPS, where peripheral immune cells are incorporated into the system’s fluidic flow to circulate around liver microtissues. Assays utilizing these models facilitate the detection of cytokine release, T-cell activation, or other immunotoxicity, such as hepatocyte damage via clinical liver function test (LFT) markers. They enable effects that only manifest in human tissue to be detected, overcoming the limitations of animal tests. For more information, visit our immune-mediated liver injury webpage.

Although CN Bio’s fully immunocompetent assay is not yet widely available, the concept is verified, and the approach is built upon proven pedigree. A co-publication with the FDA’s CDER group concluded that data derived using PhysioMimix systems is appropriate for use in drug safety and metabolism applications, evidencing its enhanced performance versus standard techniques. It substantiated CN Bio’s position as a leader in the Organ-on-a-chip (OOC) field with reliable and robust cutting-edge technology, ready for widespread adoption across the pharmaceutical industry (Rubino et al., 2021).

Furthermore, healthy human models can also be induced to common liver diseases such as Metabolic dysfunction-associated steatohepatitis (MASH) to explore increased DILI susceptibility due to the presence of underlying disease. As our PhysioMimix MASH assay also has proven calibre, having supported Inipharm’s INI-822 for metabolic liver disease treatment now in clinical testing, you can rest assured that investing in PhysioMimix OOC microphysiological systems for mAbs development represents a safe future bet.

What role is CN Bio taking in supporting the transition away from animal testing?

CN Bio has been, and remains, actively involved in numerous consortia, groups, and networks that are driving change to facilitate broader adoption of microphysiological systems within our industry (see examples below). This includes direct and indirect work with agencies such as FNIH, ICCVAM, and C-PATH, which are working with the FDA to facilitate interagency coordination and accelerate the process by pooling expertise, data, and resources across the US government and other relevant agencies.

• CEN-CENNELEC – OOC working group  
• The Predictive Safety Testing Consortium (C-PATH) – Qualification workshops 
• FNIH VQN (validation qualification network) – Complement-AIRE programme 
• 3RsC MPS Working Group  & 3RsC-FDA DILI Project to build confidence in Liver MPS for DILI and regulatory applications.
• NICEATM (National Toxicology Program Interagency Centre for Evaluation of Alternative Toxicological Methods) ICCVAM (Interagency Coordinating Committee on the Validation of Alternative Methods) 
• NC3Rs NAMs network 
• CAMs (Cambridge Alliance on Medicines) network 

But ultimately, our success in supporting the transition away from animal testing towards more predictive and cost-effective workflows is entrenched in our customers’ success, and we will never lose sight of that fact! We offer exemplary customer support irrespective of whether customers choose to adopt CN Bio’s validated models and assays or wish to develop their own novel approaches using our platform.

We are proud of our product portfolio’s ability to rapidly and easily onboard customers with a microphysiological system-based approach and are committed to broadening its scope through future model, assay and kit development via our internal R&D efforts and strategic partnerships. Through our inclusion in regulatory and standardization agencies and resultant projects, we will continue to innovate and validate models for applications most urgently needed in the drug discovery/development pipeline.

So, what are you waiting for?

Like the FDA, we believe that integrating more predictive NAMs methods into early decision-making will enable companies like yours to begin the process of reducing non-clinical animal use costs (up to $7.2M in NHP savings alone for mAb development) and be better positioned to make more informed go/no go decisions regarding which therapeutics to advance.

We are here to support your journey so that you can benefit from cost and time savings benefits earlier! So that leaves us with a question. Are you ready to make the change and adopt microphysiologcal systems for mAbs development? Contact us here to get started.

Infographic

Top tips for integrating organ-on-a-chip technology into your workflow

Keep a lookout for part three in this blog series exploring the new era of drug development. Coming soon.

Associated with their announcement was a 5-year roadmap outlining their plan to reduce animal studies from the norm to exceptions in the standard drug development process.

This historic and pivotal moment also included much-anticipated regulatory guidance and incentives for using data from New Approach Methodologies (NAMs), including Microphysiological Systems (MPS) and in silico approaches, for regulatory applications with immediate effect.  

Why has the FDA loosened its grip on animal testing for mAbs?

Several key factors behind the FDA animal testing phase-out for monoclonal antibodies (mAbs) are cited as reasons behind this decision including:

Advancing Public Health

Improve drug safety and accelerate the evaluation process by using more effective, human-relevant methods 

Reducing Animal Experimentation

Due to their limitations in accurately modeling human health and disease and their predictivity for whether a drug may be safe or effective in humans.

Cost Efficiency

Leveraging AI-based computational models and lab-grown human organ-on-a-chip/ organoids, to reduce R&D costs and lower drug prices 

Ethical Considerations

The shift aligns with broader ethical concerns about animal testing and promotes more humane testing methods 

In this first blog in a series of three, we explore highlights from the FDA’s roadmap and why mAbs were chosen as the starting point. We dig into the topic of NAMs and plans to ensure that they offer the same degree of safety as animals before focusing down on the important role of predictive human MPS in the successful outcome of this process.

What does the FDA animal testing phase-out for monoclonal antibodies (and other drugs) roadmap look like?

There’s no doubt that the FDA’s implementation roadmap for reducing toxicity testing in animals in favor of a NAMs-based approach is ambitiously proactive. Their roadmap plans are outlined in a document released alongside the press release, which can be read in full here, however the short- and long- term highlights are summarized below.

Short-term (1-3 year) priorities

Leveraging existing international data, encouraging NAM IND data submissions in parallel with animal data, developing an international open-access data repository of toxicology data from animals and humans, reducing timelines for primate mAb and other drug toxicity testing, and tracking quantifiable changes.

Long-term (3-5 year) efforts

Aim to make animal studies the exception rather than the norm by switching to a “NAM-based default” for pre-clinical safety & toxicity testing.

Despite the initial shock regarding the boldness of the FDA’s decision, when you investigate the detail, their initial focus on monoclonal antibodies represents a cautious first step. mAbs are relatively safe and well-tolerated by humans. But they (and any other new human-specific drug modalities) pose development challenges due to inter-species differences that render the use of animals less suited or even unsuited for their testing. This can result in adverse effects being discovered in the clinic, or during post-approval marketing. But there’s more than just safety concerns behind their decision.

Why did the FDA choose to focus on phasing out animal use for mAbs first?

Monoclonal antibodies specifically target molecules and cells in the body, making them highly effective for treating a plethora of diseases. They are well-characterized and have a long half-life, meaning they can provide sustained therapy over extended periods. And importantly, compared to traditional therapeutics, generally have fewer side effects due to their specificity. For reasons such as these, the mAbs clinical pipeline represents the largest of all new modality classes, with 2,700 mAbs in clinical trials as treatments for a wide variety of diseases according to a 2024 report from BCG.

Although mAbs have a relatively good safety profile, the FDA specifically calls out reasons why mAbs represent a promising area for reducing preclinical animal use in their roadmap. These include the shortcomings of animal models, which render testing workflows vulnerable, plus practical considerations centered around “skyrocketed” costs, as quoted below.

Immunogenicity & interspecies differences

“Animals often mount immune responses to human mAbs, which can alter exposure and confound toxicity interpretation. However, animal immunogenicity is not predictive of human immunogenicity due to interspecies differences in immune systems. In addition to inherent biological differences, stress of laboratory life and use in research can impact immune function, inflammatory responses, metabolism, and disease susceptibility and progression.

Moreover, some safety risks may go undetected in animals – a notable example is the mAb TGN1412, which caused a life-threatening cytokine release syndrome in human volunteers despite appearing safe in preclinical monkey studies. That tragedy highlighted the limitations of animal models for certain immune-activating mAbs and spurred efforts to develop in vitro assays to better predict human-specific responses”.

Practical challenges

“The cost of drug development can vary by therapeutic class, with a market report noting the cost to develop a mAb at $650-$750 million and taking up to 9 years. Typical mAb development programs typically use 144 non-human primates (NHPs). In recent years, costs of NHPs have skyrocketed, up to $50,000 per NHP. The time and cost of long-term animal studies slow down delivery of new therapies to patients.”

This is where it gets interesting, by forcing the hand to “plug” this weakness – alongside animals initially – testing workflows can be made even safer than today!

Before animals are fully replaced, the FDA plan also allows time for lessons to be learnt, gaps to be identified and NAMs to become further developed, validated and entrenched into workflows. Plus, the FDA acknowledges the importance of updating international guidelines so that companies do not face different rules in different regions to alleviate adoption barriers.

This approach also begins the important process of reducing both the cost, and the time required to bring novel mAb therapeutics to patients to lower the price of drugs, whist also preparing workflows for the inclusion of additional drug types going forward.

What are NAMs and how do we ensure that they deliver the same degree of drug safety as animals?

In their roadmap, the FDA provides an overview of key NAM categories and their applicability to drug development. There are many definitions for NAMs but the FDA refers to them as:

“tools to assess safety, efficacy, and pharmacology of drugs and therapeutics without traditional animal models. NAMs include in vitro human-based systems such as organs-on-chips, “in silico”, or computer-based modeling, as well as other innovative platforms that can collectively evaluate immunogenicity, toxicity, and pharmacodynamics with high relevance to human biology. The FDA and the broader scientific community recognize NAMs as a means to obtain faster and more accurate human risk assessments while reducing animal use”.

The FDA acknowledges that the transition to a NAMS-based approach will require careful planning, robust science, and collaboration. It lists specific actions that the FDA is considering for validating and integrating NAMS safely into their regulatory process.

They also recognize that, while computational models can be used to predict human-relevant outcomes, in vitro test systems will be needed to confirm and improve predictions. Amongst the in vitro-derived systems cited by the FDA are organoids and microphysiological systems (MPS), often called organ-on-a-chip (OOC).

MPS can be thought of as next-generation organoids because they more accurately recapitulate human organs. The differences between organoids and MPS are explored in more detail in our blog “Taking organoids to the next level”, however, the most important differentiators are the incorporation of microfluidic flow, mechanical forces, and multi-cell primary human co-cultures on a bioengineered chip to recapitulate the in vivo environment.

Compared to self-assembling organoids cultured in static conditions, MPS deliver higher functioning microtissues that deliver greater assay sensitivity, the reporting of clinically relevant biomarkers for enhanced in vitro to in vivo extrapolation (IVIVE), plus prolonged culture longevity (for up to a month using PhysioMimix^® OOC solutions) to facilitate repeat-dosing studies in vitro. Furthermore, they can be interconnected to form multi-organ networks that simulate pharmacodynamic effects systemically, human processes such as drug absorption and metabolism, or used to understand interactions between organs, such as inflammation, which drive disease and cause unexpected toxicities.

Many of the organ systems affected by mAbs are available, including liver, heart, lung and kidney. However, despite these technological advances, certain scientific challenges are yet to be overcome before animal tests can be fully replaced. For example, in vitro or in silico models are not able to represent a complete biological system, and further advances are required in the field to recreate the full complement of the human immune system.

Despite these challenges, it is important to acknowledge that, although animal use is a current mainstay of drug development, when first utilized they were never fully validated before being adopted. By comparison, the ongoing validation of non-animal methods, including MPS, for predicting human drug responses is highly rigorous. It is also a well-understood fact that, for certain read-outs, animal responses are no better than the toss of a coin on whether a drug may be safe or effective in humans. Frequently, different answers will be reported by different species, which makes the final assessment of drug safety difficult. In addition to safety concerns, this lack of clarity also causes safe human therapeutics to be unnecessarily dropped from the pipeline out of caution.

In our extensive experience, MPS-based PhysioMimix assays are more predictive of human responses than animal tests (see example publications and application notes for more info).  Our findings are echoed by the FDA’s CDER group who concluded that data derived using PhysioMimix is appropriate for use in drug safety and metabolism applications, evidencing its enhanced performance versus standard techniques. Their publication (Rubino et al., 2021) substantiates CN Bio’s position as a leader in the field with reliable and robust cutting-edge technology, ready for widespread adoption across the pharmaceutical industry.

To provide further confidence in the human predictivity of MPS models, we are also developing animal equivalent MPS models/assays. They enable developers to recreate in vivo outcomes in vitro alongside human assays to prevent good assets from being abandoned. However, rather than just using these to troubleshoot, our future vision for MPS-based cross-species assays is for upstream lead optimization to reduce unnecessary in vivo animal use.

Despite the positive outcomes of our MPS validation tests using “fallen angels” it is also imperative to remember that, like animals, no one NAMs method will be 100% accurate or applicable for every question, especially given the fact that no human is identical. Therefore, we believe it is important to adopt a broad range of NAMs to predict human behavior. The skill to acquire here is determining which combination of alternative approaches should be used in combination to determine a drug’s safety. At CN Bio we will continue to explore and demonstrate the advantages of combining MPS plus in silico tools ourselves, with further announcements planned later this year.

Concluding thoughts

So, as we wrap up this first (of three) blog post FDA announcement, three things are certain. Firstly, it is well known fact that drug development failures are due to lack of efficacy or unexpected safety issues that were not evident in standard in vitro or animal tests. Secondly, the largest global regulator has sent a clear message regarding their plans to phase out animal testing to save significant costs (up to $7.2M for NHP use alone/mAb) and reduce development cycle times. Thirdly, the FDA’s timeline to modernize drug development workflows is short.

The key way for pharmaceutical companies to enhance the human predictivity of workflows and keep pace with these changes is to invest in several NAM approaches. By becoming familiar with the types of data they generate and understanding where their limitations lie, you can decide which combinations of NAMs-based tests are required to fully investigate the safety and efficacy of a new drug entity.

So, having explored the FDA animal testing phase-out for monoclonal antibodies (and other drugs) roadmap, and what is being done to ensure that the collective use of NAMs offer the same degree of safety as animals before their use is reduced and replaced – are you ready to get started now and incorporate OOC/MPS into your workflows to avoid being left behind?

Contact us here, to find out how we can support your journey with PhysioMimix so that you can reap the cost and time-saving benefits earlier.

Infographic

Top tips for integrating organ-on-a-chip technology into your workflow

In part two of this blog series, we take a deeper dive into how MPS address the shortcomings of animal models for mAbs development and the role that CN Bio is taking to support the switch to a NAMs-based approach. Continue reading here.

New modality drugs, or advanced oligonucleotide therapeutics, refer to innovative approaches that go beyond traditional small molecule approaches to target diseases more precisely and effectively.

They are designed to address complex diseases and conditions that are difficult to treat with conventional therapies and may have been classed as “undruggable”.

According to The 2024 New Drug Modalities Report | BCG, new modalities represent $168 billion in Pharma/Biotech projected pipeline value, up 14% from 2023. By 2029, analysts project that 9 of the top 10 drugs will be new modalities. Although antibodies make up most of the later-stage preclinical pipeline, a plethora of other modalities can be found in the earlier phases including oligonucleotides which are more amenable to intracellular targets.

However, oligonucleotides have unique development challenges that require adaptations to preclinical testing workflows. A major challenge relates to their in vivo testing. Due to the high degree of conservation of nucleic acid base sequences between humans and non-human primates (NHPs), their use is common but, expensive. However, even NHPs can incorrectly predict human responses – as exemplified in the phase 1 clinical trial of TGN1412, a monoclonal antibody, in 2006. Six healthy volunteers were left with life-threatening morbidities due to the lack of CD28 expression on CD4+ effector memory T cells in macaques (Eastwood et al., 2010). Consequently, ensuring you have as much confidence in oligonucleotide delivery, safety and efficacy ahead of in vivo studies is advantageous to minimize risk. To achieve this, the pre-clinical toolbox requires modernization to become more human-centric. New Approach Methodologies (NAMs), including Organ-on-a-Chip (OOC), provide a viable path forward to address many development challenges. And, following the passing of the FDA’s Modernization Act 2.0 in 2022, their use is supported by the largest global regulator for evaluating drug safety and effectiveness in place of animal testing – where appropriate.

In this blog, we explore why oligonucleotide-based therapeutics are increasing in popularity, their ADMET discovery and development challenges, and how in vitro human OOC models help to circumvent these by providing data-rich insights that better inform/ justify the expense of in vivo NHP or human studies.

Why are Oligonucleotide-based Therapies Rising in Popularity?

Oligonucleotide-based therapeutics, or RNA-based therapeutics, are short oligonucleotide sequences that interfere with a specific RNA. They include antisense oligonucleotides (ASOs), RNA interference (RNAi), small interfering RNA (siRNA), microRNA (miRNA), and aptamers. Interest in oligonucleotide therapeutics arose from our understanding of the human genome and the genetic link to many diseases, combined with their relatively fast drug discovery process compared to traditional small molecules and accessibility to organizations of any size.

Around 18 oligonucleotide-based therapeutics have received market approval from 1998-2024 (Table 2). Significant advances over more recent years are due to the use of delivery systems, such as lipid nanoparticles (LNPs) and the N-acetylgalactosamine (GalNAc)-conjugations, that enable drug targeting to the liver for the treatment of hepatic disorders. For example, in 2016 Patisiran (marketed as Onpattro) for transthyretin amyloidosis (ATTR) uses LNPs to deliver siRNA to the liver. In 2019, Givosiran (marketed as Givlaari), which uses GalNAc-conjugation to target the liver was approved for acute hepatic porphyria.

Liver targeting using the sugar molecule GalNAc is made possible because of the asialoglycoprotein receptor (ASGPR), which is found primarily on the cell surface of hepatocytes. GalNAc-conjugated oligonucleotides bind with high affinity to ASGPRs and are rapidly removed from the systemic circulation via endocytosis. Once inside the cell, the therapeutic oligonucleotide is released from the receptor to exert its therapeutic effect by targeting specific messenger RNA (mRNA) sequences. This approach improves potency and effectiveness at lower doses, reduces off-target effects, and potentially enables longer dosing intervals with lower immunogenicity. With continued research and development, GalNAc-based therapeutics hold immense future potential for treating a wide range of diseases, including genetic disorders, metabolic diseases, and infectious diseases.

In 2021 Novo Nordisk acquired Dicerna to capitalize on this potential. A subsequent report in Bioprocess International, in 2023, citing Marcus Schindler, executive vice president of research and early development at Novo Nordisk, stated “a significant portion of my pipeline now is oligonucleotide”. The reason, he explained, “it’s really to do with building a toolbox to accelerate everything that we do in drug discovery”. Regarding how Novo Nordisk is supporting oligo-development, he said, “so first and foremost, we are taking what we call a human-centric approach to understanding complex biology. We start with datasets and experimental models that mimic the one species that we’re interested in. And that is humans.”

So, why is it that this type of drug requires a more human-centric approach, and what are the challenges in its discovery and development that require a rethinking of the preclinical toolbox?

The ADMET Challenges of Oligonucleotide Therapeutic Discovery & Development for Liver Diseases

The Absorption, Distribution, Metabolism, Excretion and Toxicology (ADMET) challenges relate to the point above, oligonucleotides aren’t traditional small molecule compounds and animals aren’t humans. Oligonucleotides have multiple modes of activity by which they can achieve their intended pharmacological effect, however, most commonly, they contain complementary nucleotide base sequences to a targeted sequence of an intracellular nucleic acid. Some may be highly specific to humans and consequently, this either forces the use of NHPs, or in some instances negates the use of non-human species. Test species-active surrogate oligonucleotides are an option but this doesn’t get around the fact that animals are not humans. In silico and in vitro methodologies help to identify potential off-target hybridization effects, however, in vitro models are limited by their physiological relevance (explored in more detail below) and in silico models are only as good as the input data.

Many challenges remain in the development of oligonucleotide-based therapeutics for liver disease and their preclinical ADMET profiling, the most pertinent of which are summarized below (Table 1).

Table 1

The key challenges in ADMET profiling of oligonucleotide therapeutics.

How OOC Models Support Advancements in the ADMET Profiling of Oligonucleotide-therapeutics for Liver Disease

These challenges underscore the need for innovative new approaches, such as OOC, to improve the predictive power of preclinical oligonucleotide testing. OOC platforms, also known as microphysiological systems (MPS), condense all the technologies needed to build functional tissues that model specific organ-level responses into a miniaturized format. To replicate the liver, primary human hepatocytes (PHHs) and the liver’s resident immune cell population (Kupffer cells, as necessary) are cultured to form 3D microtissues in bespoke engineered scaffolds. Microtissues are continuously perfused by media to mimic blood flow, providing gas exchange and nutrients. This dynamic 3D microenvironment promotes the viability and functionality of liver cells (quantified by albumin and lactate dehydrogenase) over multiple weeks, compared to just a few days for 2D in vitro hepatocyte cultures. They aim to help circumvent traditional preclinical model limitations by improving human and physiological relevance. Their insights support the responsible use of animal models, by justifying the progression of only the most promising candidates, and better-informed decision-making ahead of clinical trials where no pharmacologically relevant model exists.

So, how do these in vitro models address ADMET challenges? Firstly, the longevity of OOC cultures versus static PHH cultures enables the concentration effects of oligonucleotides to be measured. This is achieved by dosing liver microtissues and subsequently measuring oligonucleotide uptake and gene knockdown – the latter of which takes time to manifest. Additionally, their longevity permits repeat oligonucleotide dosing to inform the most effective strategy before in vivo animal, or first-in-human (FIH) studies. Furthermore, the health of Liver-on-a-chip cultures can be monitored throughout the experiment to flag any potential adverse effects, including activation of innate immune cells. Finally, their human-relevant data can be used to “feed” in silico models to improve the performance of predictions.

Liver-on-a-chip use for small molecule drug metabolism, accumulation and toxicity testing has already been recognized by the US FDA as superior to traditional approaches. Now, their ability to improve the ADMET profiling of oligonucleotide therapies is being explored by Pharma and Biotech via PhysioMimix^® OOC platform adoption, or our contract research services. The following case studies highlight recent findings.

Case studies

Case Study 1:  Human-Relevant ASO Delivery and Uptake

In a recent publication by GSK, Majer et al., (2024), explored OOC’s potential for addressing one of the major hurdles in therapeutic ASOs, delivering oligonucleotides in sufficient concentration to the target tissue and the cells of interest. Their targeted delivery is important for reducing off-target effects and improving the specificity of gene silencing, both of which are essential for safe and effective therapeutics.

Their study evaluated the cellular uptake and localization of Alexa488 (A488)-labelled non-conjugated ASO and GalNAc-conjugated ASO by a PhysioMimix Liver-on-a-chip model using label/label-free imaging techniques.

The authors cite “by replicating key aspects of the liver function, such as drug metabolism, bile production and immune responses in a controlled in vitro environment, these models provide a cost-efficient, robust, and human biology-relevant ethical alternative to traditional in vivo animal testing that can accelerate drug discovery and development”.

An interesting observation made by the publication is that Liver-on-a-chip narrows the gap between in vitro assay and in vivo human liver environment because most PHHs in the model were cuboidal. Cuboidal PHHs were found to exhibit different chemical composition, morphology, and ASO distribution compared to circular hepatocytes, which are more common/less physiologically relevant in 2D culture. For this reason, the authors discounted data from circular hepatocytes These results call into question the usefulness of 2D, or 3D spheroid models for this purpose.

The study observed that uptake of non-conjugated ASOs by Liver-on-a-chip PHHs was slower compared to GalNAc-conjugated ASOs at early time points, although a greater amount of non-conjugated ASOs was taken up over time. This is probably caused by the high binding affinity of GalNAc and insufficient receptor recycling.

Case Study 2: ASO Uptake vs. Knockdown

In an independent CN Bio study, using a PHH-only model, similar results to the Majer et al., study were observed i.e., greater hepatocyte uptake (signal intensity) over time by the non-GalNAc conjugated ASOs versus GalNAc-conjugated – likely due to different processing of the ASOs by PHHs. However, the gene knockdown data tells a different story. This hint into the efficacy of ASOs highlights why it is important to measure knockdown and uptake. Register for our on-demand webinar to find out more, or read the Application Note.

Case Study 3:  Resolving ASO Toxicity Data Discrepancies

A further study in collaboration with AstraZeneca utilized our PhysioMimix Liver-on-a-chip and DILI assay to resolve conflicting ASO toxicology data between animal models and 2D hepatocytes. In their publication, Sewing et al., (2016), cite “A series of reports describe changes observed in rodent and non-rodent studies related to SSO (single-stranded oligo)-induced toxicity, but as the safest preclinical candidates progress into the clinic, the human relevance of the observed effects remained unclear so far.” CN Bio was asked to test two ASOs to ascertain if a human system can provide more insight. Human-relevant markers including alanine aminotransferase (ALT) and CYP3A4 activity were able to clearly show and predict human outcomes between ASO-1 (toxic) and ASO-2 (safe). This gave the authors further clarity regarding which would be safest for humans.

Going forward, recent advancements in OOC technology enable the linking of individual organs together into systems. As more models become available to fluidically interconnect, these systems can be utilized to explore off-target effects and potentially explore renal excretion.

Summary

The rise of oligonucleotide therapeutics represents a significant advancement for the treatment of complex diseases, offering targeted and effective solutions where traditional therapies fall short. Despite their potential, oligonucleotides face unique challenges that necessitate a shift towards more human-centric preclinical testing approaches.

Use of human OOC models offers a solution to many of these challenges. These models replicate human organ functions in a controlled in vitro environment, providing more relevant and predictive data compared to animal models. By improving the physiological relevance of preclinical testing, OOC platforms help to better predict human responses, justify the use of expensive NHP studies or provide a path forward where in vivo models aren’t suited, and ultimately accelerate the development of safe and effective oligonucleotide therapies.

Published case studies from AstraZeneca and GSK, highlight the effectiveness of OOC models in addressing key hurdles in oligonucleotide development. These studies demonstrate how OOCs provide clearer insights into human-specific responses.  

Furthermore, as OOC technology continues to advance, it holds the potential to further revolutionize the field by enabling the integration of multiple organ systems, thereby offering a comprehensive understanding of drug behavior and safety.

To conclude, the adoption of human-centric approaches, such as OOC, is crucial for overcoming the development challenges of oligonucleotide therapeutics and unlocking their full potential.

Find out more about the possibilities offered by PhysioMimix^® Liver-on-a-chip models in our next webinar: Register here or contact us for more information.

Table 2

FDA-approved oligonucleotide therapeutics from 1998 to 2024

These human epithelial colorectal adenocarcinoma cells are prized for their ability to differentiate into enterocyte-like cells, forming tight junctions that mimic the intestinal barrier.

However, as with any model system, Caco-2 cells come with their own set of limitations. This raises important questions: Do their limitations undermine their gold-standard status? Is it time to improve estimations with more physiologically relevant approaches? What new approaches are available and most importantly, do their benefits sufficiently outweigh the effort required to adopt a new approach?

The strengths of Caco-2 cells

The undeniable strengths and simplicity of Caco-2 cell models are testament to why they have remained a valuable tool in drug development for so long. So often we hear the phrase “If it isn’t broken, why fix it?” Therefore, innovations must be much more beneficial to drive change.  

So, what’s the driver for change when Caco-2 cells are well known for their:

1. Predictive Power: Their ability to form a monolayer with tight junctions on a Transwell® insert enables reliable prediction of drug permeability of passively diffused compounds.
2. Reproducibility: The use of Caco-2 cells allows for consistent and reproducible results, which is crucial for comparative studies and regulatory submissions.
3. Ease of Use: These cells are relatively easy to culture because they are derived from a human colon cancer, making them accessible to laboratories.

We believe the driver for change relates to drug metabolism. When a drug enters the body via the oral delivery route, its absorption and permeability through the gut is merely the first step in its journey. The gut is a metabolically active organ plus, before an orally administered drug has a chance to elicit its actual target effect, there’s the impact of a second organ to consider too – the liver.

Currently, we derive in vitro absorption & permeability data from Caco-2 models, then utilize in vitro liver clearance assays. The disconnect here is that this approach fails to account for the pharmacokinetics of a drug after it is absorbed and metabolized by the gut to estimate what proportion reaches the liver. Therefore, the use of isolated assays leaves a worrying gap in our knowledge when estimating a drug’s bioavailability.

Why is estimating human oral bioavailability a preclinical challenge?

Key to safe and efficacious dose setting ahead of clinical studies, and a regulatory requirement, is estimating a drug’s bioavailability. Data derived from isolated in vitro tests can be used to infer a drug’s bioavailability but only in extreme cases does the approach deliver actionable insights. Whilst these limitations can be countered to some extent by in vivo animal bioavailability studies and in silico models they too suffer limitations. In silico models are only as good as the data input, whilst the translatability of animal predictions is hindered by interspecies differences due to differences in physiology and metabolic capacity.  A seminal study by Musther et al in 2014 investigated 184 compounds and found that bioavailability data from commonly used animal models weakly correlates with human data (mouse R² = 0.25, rat =0.28, and dog = 0.37). Non-human primate (NHP) models provide an improved bioavailability correlation with humans (R² =0.69). However, these models are not commonly used due to significant ethical considerations, stringent regulatory requirements, and high cost.

As a result, the preclinical estimation of human oral bioavailability is a complex process involving the “piecing together” of data derived from many sources. Improvements to the process via more humanized approaches represents an area of “quick win” potential.  

Use of Caco-2 cells to mimic the process of first-pass metabolism in vitro

Fluidically interconnecting a gold-standard approach to measuring drug absorption/permeability with a gold-standard approach to measuring liver clearance would represent the obvious step towards improving the value of data derived from Caco-2 cell assays. At CN Bio, we explored this approach via the development of a gut-on-a-chip model by culturing a mixture of epithelial (Caco-2) and goblet (HT29-MTX) intestinal cells on the basolateral side of a Transwell® insert exposed to circulating media.

We fluidically connected the Caco-2 gut -on-a-chip with our highly characterized and validated primary human liver-on-a-chip model. The latter of which has been recognized by the US FDA (in the first publication between an Organ-on-a-chip provider and regulator) as delivering improved performance versus standard static liver culture approaches (Rubiano et al., 2021).

By fluidically linking these two models together to form a system, we combined previously isolated assays into one (using the PhysioMimix Multi-organ System & Multi-chip Dual-organ plates) to mimic the process of first pass metabolism in vitro. The advantages of this approach are that you can simulate IV and oral dosing to evaluate the individual and combined contributions of each model to derive the area under the curve, i.e., the pharmacokinetics of a drug after it is absorbed and metabolized by the gut, and estimate what proportion reaches the liver to provide the missing piece in the bioavailability estimation puzzle.  

The performance of this novel Gut/Liver-on-a-chip model, also known as a Gut/Liver microphysiological system (MPS), was evaluated by Roche using the prodrug mycophenolate mofetil (Milani et al., 2022). However, in their publication Roche noted the limitations of the Caco-2 cell model when working with a drug of this type due to carboxylesterase 1 & 2 (CES1 and CES2) expression. In Caco-2 cells, expression of these enzymes is neither physiological nor human relevant. These enzymes are crucial for the hydrolysis of ester prodrugs and can affect the accurate prediction of the metabolism and absorption of ester-containing drugs.

But this is just one of many Caco-2 cell limitations that restrict the accuracy of estimations derived using the dual-organ model in our PhysioMimix Bioavailability assay.

The innovation leap: when two MPS providers are better than one

In January 2024 we announced a close strategic partnership with Altis Biosystems. This partnership enabled us to replace our perfused Caco-2 gut-on-a-chip with their primary human RepliGut^®-Planar Jejunum model in our dual-organ system. The RepliGut model utilizes human intestinal stem cells to create an in vitro model that more closely mimics the human to overcome the additional limitations of Caco-2 cells including:

1. Lack of Complexity: While Caco-2 cells mimic the intestinal barrier, they do not fully replicate the complexity of the human intestine. They lack the diversity of cell types found in the gut, which play significant roles in absorption and permeability.
2. Mucus Secretion: Caco-2 cells do not secrete mucus. RepliGut® models include mucus-secreting goblet cells, better mimicking the in vivo environment.
3. Barrier Integrity: The RepliGut model has a distinct cell expansion phase, with mostly stem and progenitor cells and a differentiation phase, mostly differentiated cells. As the cells mature, the barrier integrity increases monitored using TEER (Transepithelial Electrical Resistance). This more closely mimics the human intestine, which is highly regenerating.   
4. Overexpression of Efflux Transporters: Caco-2 lack key efflux transporters like P-glycoprotein, which can lead to an underestimation of drug absorption. Gene expression data from RepliGut models indicates transporters undergo differential regulation as cells mature with the Caco-2 cells looking more like immature cells. 
5. Metabolic enzyme expression: Caco-2 cells have limited expression of Phase 1 & 2 metabolic enzymes. This can lead to an incomplete understanding of a drug’s metabolic profile and potential bioavailability. RepliGut models express more physiologically relevant levels of metabolic enzymes, including cytochrome P450 (CYP), UGT, and CES enzymes, improving the accuracy of drug metabolism studies.

Improved ability to profile human bioavailability in vitro with a primary human RepliGut/Liver-on-a-chip versus Caco-2 cells in isolation

The performance of the fluidically linked RepliGut^®-Planar Jejunum/Liver-on-a-chip (RepliGut^®-Planar Jejunum/Liver MPS) was investigated using two compounds (temocapril and midazolam) who’s human ADME properties were not predicted by existing models. Midazolam is a drug that intestinal and liver clearance by the CYP3A4 enzyme which has negligible expression in Caco-2 cells. Temocapril is a pro-drug designed to be resistant to intestinal hydrolysis that is metabolized to temocaprilat by CES1. CES1 is more predominantly expressed in Caco-2 cells versus humans leading to an over estimation of drug clearance.

The results of this study demonstrated that the primary dual-organ model offers a viable alternative to circumvent the human-relevance limitations of the Caco-2 cell line for uniquely profiling human bioavailability in vitro.

More information can be found in our application note.

Conclusion

Despite their limitations, the ease of use of Caco-2 cells make them indispensable for initial screening and comparative studies. However, it is crucial to complement Caco-2 studies with other more physiologically relevant models. By doing so, you obtain a holistic understanding of the combined effects of the human gut and liver to more accurately profile oral drug bioavailability in vitro.

Generating more human-relevant data earlier in drug discovery means that observed issues can be flagged and addressed before costly preclinical in vivo studies. As the primary RepliGut/Liver-on-a-chip is entirely made up of primary human cells, there are no interspecies differences to account for, therefore, the model can be utilized to help overcome the poor correlation between animal model ADME predictions and human outcomes. Bridging the gap between in vitro assays and in vivo studies, the model enables researchers to confidently progress only the most promising drug candidates to support reductions in cost and the number of animals required. For this reason, we believe that the benefit of change significantly outweighs the benefit of maintaining the status quo and foresee that the gold standard will become redefined by the integration of dynamic models into the toolkit of ADME/DMPK scientists.

So, don’t delay tackling challenges in ADME drug development! The longer you wait the greater the risk of cascading problems down the line from miscalculated pre-clinical estimations.

Our primary human in vitro Gut/Liver model and Bioavailability assay can be accessed in two ways:

1. Via our ADME Contract Research Services
2. In your own laboratory using our recently launched PhysioMimix Bioavailability assay kit: Human 18 and PhysioMimix^® Multi-organ System.

However, the ADME data derived from these models is only as good as the test’s ability to mimic the human response – all have limitations leading to challenges in ADME drug development. Animals suffer from interspecies differences, whilst in vitro assays are simplistic and operate in isolation. In silico models are only as good as the input data sets, frequently requiring first-in-human study data to be “taught”. The problem is further compounded by advanced drug modalities that require a rethink regarding the best tool for the job from the ADME experimental toolbox. Plus, we humans are all different! Our genetic polymorphisms, disease states, age, gender and the drugs we already take also play a part in the way that our bodies interact with drugs.

In this article, we explore six current challenges in profiling preclinical human ADME cited by publications highlighting where advances in the human relevance of preclinical models can improve estimations for safer and more efficient drug development.

#1 How to gain greater insights into human ADME profiles through a broader approach.

Addressing Today’s Absorption, Distribution, Metabolism, and Excretion (ADME) Challenges in the Translation of In Vitro ADME Characteristics to Humans: A Case Study of the SMN2 mRNA Splicing Modifier Risdiplam. Fowler et al (2022).

Risdiplam is a small molecule approved to treat spinal muscular dystrophy with complex ADME properties including high metabolic stability. Conventional in vitro drug metabolism assays failed to predict the human response to Risdiplam, underestimating in vivo turnover. The authors used a modelling and simulation-based approach (in silico physiological-based pharmacokinetic (PBPK)) in combination with in vitro data and early clinical studies to elucidate a comprehensive picture of the drug’s ADME profile. They argue this “combination approach” provides greater insights when investigating small molecule drugs with complex ADME properties as conventional preclinical approaches alone aren’t good enough.

Is it time to supplement your preclinical toolbox further with more advanced in vitro multi-organ gut/liver models to iteratively improve ADME estimations before first-in-human studies?

#2 How to account for human and population differences in intestinal metabolism

Contribution of Intestinal Cytochrome P450-Mediated Metabolism to Drug-Drug Inhibition and Induction Interactions. Galetin et al (2010).

This review focuses on the role of intestinal cytochrome P450 (CYP) enzymes in drug-drug interactions (DDIs). Intestinal CYP enzymes are present in the gut where they play a crucial role in metabolizing drugs before they reach systemic circulation – which can significantly affect a drug’s bioavailability. The review discusses how current DDI prediction models often don’t fully account for intestinal CYP activity, potentially leading to inaccurate predictions. The authors recommend incorporating data on intestinal CYP metabolism into DDI prediction models to improve their accuracy and highlight the need for better methods to assess intestinal CYP activity and variability among individuals.

Whilst traditional Caco-2 cell assays offer a quick and easy method of estimating drug absorption, some CYPs are missing and the levels of expressed CYPs are generally lower and varied compared to the human intestine. This can contribute to estimation discrepancies. Additionally, this cell line approach does not offer the potential to study donor-to-donor population differences. Advanced in vitro models utilizing primary human intestinal cells that are fluidically linked to human liver models provide a unique path to overcome these limitations for a more accurate estimation of a drug’s first-pass metabolism and bioavailability in humans.

#3 Overcoming the weak correlation between animal and human bioavailability data

Animal versus human oral drug bioavailability: do they correlate? Musther et al (2014).

Estimating the bioavailability of drugs in humans during preclinical development is necessary to set safe and effective doses in the clinic. This seminal study investigating 184 compounds found that bioavailability data from commonly used animal models weakly correlates with human data (mouse R² = 0.25, rat =0.28, and dog = 0.37). Non-human primate (NHP) models provide an improved bioavailability correlation with humans (R² =0.69), although these models are not commonly used due to significant ethical considerations, stringent regulatory requirements, and high cost. The overall correlation between animal and human bioavailability is poor mainly due to differences in physiology and metabolic capacity. Whilst these data do not eliminate the usefulness of animal models for pharmacokinetic analysis, it highlights that these studies are best used as qualitative indicators, rather than quantitative data.

A new human-relevant approach, which combines oral absorption and hepatic metabolism, is required to more accurately estimate human drug bioavailability in vitro to help circumvent this challenge in ADME drug development.

#4 How to rationalize drug design with earlier use of in silico modeling

In silico ADME/T modelling for rational drug design. Wang et al (2015).

One of the most important innovations in modern drug development has been the adoption of computational or in silico tools to screen for new compounds and predict ADME behavior from a compound’s chemical structure. Clinical behavior can be forecasted using PBPK modelling, which combines a compound’s chemical properties, the physiology-based properties of the population being modelled, and data from ADME studies.

In this article, the authors argue that in-depth ADME testing is carried out too late, only when a limited number of candidate compounds have been identified. To improve R&D efficiency, in silico methods should be used in early drug discovery to enable parallel optimisation of compound efficacy and “druggability properties”. The future implementation and success of in silico models will in part, however, depend on the quality of models and the quality of in vitro data fed into the models. Organ-on-a-chip (OOC) technology provides a tool to generate more physiologically relevant ADME data compared to traditional in vitro approaches. This presents an opportunity for OOC and PBPK to be integrated into a workflow, where higher-quality input data leads to improved population predictions.

#5 Does the increasing complexity of advanced drug modalities require a rethink of the ADME toolbox?

Recent advances in the translation of drug metabolism and pharmacokinetics science for drug discovery and development. Yurong et al (2022).

The authors highlight the tremendous progress made in the last decade to understand ADME properties of candidate drugs with progress, in part, driven by the adoption of high-throughput screening in the ADME workflow. As the landscape of compounds changes beyond traditional small molecules to large molecules and biologics, they argue the need for further innovation in experimental tools to improve the characterization of their ADME properties.

This is particularly important for newer modalities designed for oral dosing, as they tend to suffer from poor bioavailability. A good example is PROteolysis TArgeting Chimera  (PROTACs), which have emerged as a promising new chapter in drug discovery but are challenging to develop because they are molecules with high molecular weight, poor solubility and low permeability. Here emerges a need for drug developers to improve PROTAC bioavailability to unlock their full therapeutic potential. Tools such as OOC that uniquely offer the ability to profile oral bioavailability in vitro enable drug developers to test strategies to improve oral bioavailability. Example strategies may include, developing a prodrug form of the PROTAC and improvements to cellular permeability through changes to the PROTAC chemistry.

#6 Navigating the preferred (oral) administration route for new modality drugs

Oral absorption of middle-large molecules and its improvement, with a focus on new modality drugs. Asano et al (2023).

This article explores challenges in orally delivering middle-to-large molecules, a desirable route for drug administration. These molecules are typically poorly absorbed due to their size. The authors discuss the limitations of traditional methods and the need for improved absorption enhancers or modifications to the drugs themselves. They highlight recent successes like Rybelsus and Mycapssa, tablets containing absorption enhancers that allow oral delivery of peptide drugs for diabetes and acromegaly. However, even with these advancements, oral bioavailability remains low. The article emphasizes the ongoing efforts to develop new strategies for effective oral delivery of these promising therapeutic molecules and discrepancies in the amount of absorption enhancers required to improve oral administration between non-clinical in vivo animal studies and clinical human studies.

The use of advanced in vitro human models to further their development could offer a path forward to ensure that dosages are feasible in humans before clinical studies commence.

In the last decade, microphysiological systems (MPS), also known as organ-on-a-chip (OOC), have shown their potential to address ADME challenges in drug development and improve the human translatability of ADME studies. MPSs are designed to recapitulate the structural and functional biomarkers of cells and tissues in a physiologically relevant manner through the culture of primary human cells on perfused scaffolds. Efforts to improve the in vitro to in vivo translation of drug efficacy and safety data have led to more complex MPS from CN Bio where multiple organs, such as the gut and liver, are fluidically linked to simulate processes such as drug absorption and first-pass metabolism. Through this unique approach, it is now possible to profile human drug bioavailability in vitro. Broadening the ADME toolbox with advanced human in vitro assays enables better-informed decisions to reduce the risk of identifying poor oral bioavailability during first-in-human studies.

So, don’t delay tackling challenges in ADME drug development! The longer you wait, the greater the risk of cascading problems from miscalculated pre-clinical estimations. The PhysioMimix^® primary human in vitro Gut/Liver model and Bioavailability Assay can be recreated in your laboratory using the Bioavailability assay kit: Human 18, or accessed via our ADME Contract Research Services.

By unlocking a deeper understanding of how our bodies Absorb, Distribute and Metabolize drugs (three of the four ADME pillars) and their potential bioavailability, these advanced models provide a unique path forward to mitigate risk before clinical trials. Pre-clinical miscalculations of these critical pharmacokinetic factors can cause a cascade of problems that impact trial outcomes, drug development timelines, and costs.  

The most cited reasons for drugs failing in the clinic are lack of efficacy and safety concerns, both of which are influenced by dose. Could success rates improve if preclinical estimations were more accurate? 

New approach methodologies (NAMs), such as Organ-on-a-chip (OOC), are being readily adopted in drug discovery and development pipelines to address inherent ADME workflow challenges and to improve predictivity. In this blog, we will first explore why traditional approaches constrain our ability to deliver accurate human ADME estimations, followed by the benefits of human in vitro Gut/Liver models, and their potential to move the needle towards safer and more effective clinical trials. 

Why is estimating human drug ADME a challenge? 

Currently, a combination of simple in vitro assays that either model the gut (Caco-2 cell line) or the liver (liver microsomes and suspension hepatocytes), and in vivo animal models are used. However, significant limitations exist with both approaches.  

Caco-2 cells have been the workhorse for assessing in vitro intestinal permeability for a long time, but they cannot account for liver metabolism. The cell line also has absent or low levels of enzyme and transporter expression. Conversely, liver microsomes and suspension hepatocytes are used for in vitro drug metabolism screening studies, but they do not consider intestinal absorption. Collectively these limit the accuracy of in vitro-derived estimations.  

Furthermore, human bioavailability estimations from in vivo animal models weakly correlate with human-derived data, as demonstrated in a seminal study investigating 184 compounds, (R²=0.34)¹. Human-relevant approaches that combine oral absorption and hepatic metabolism are required to estimate drug bioavailability more accurately before first-in-human (FIH) trials. 

Why does the problem of estimating drug ADME and bioavailability in humans remain? 

One of the most important innovations in the field has been the adoption of computational or in silico tools to screen for new compounds and predict their ADME behavior using physiological-based pharmacokinetic (PBPK) modeling. These models are becoming increasingly important prior to trials in humans. However, their ability to accurately predict outcomes in the clinic is often constrained by the quality of the input data from prior ADME studies. 

OOC technologies, also known as microphysiological systems (MPS), are designed to recapitulate the structural and functional biomarkers of cells and tissues in a more physiologically relevant manner through the culture of primary human cells on perfused scaffolds. But their potential in the field of ADME has been held back because of several challenges. 

Firstly, maintaining differentiated intestinal enterocytes from human donors is challenging. By their very nature, intestinal epithelial cells have a short life span with a renewal cycle of 4-5 days. The intestinal lining is highly dynamic; enterocytes move towards the intestinal villus tip, become fully differentiated, and only then gain absorptive function. Because of the high turnover rate, isolated enterocytes de-differentiate quickly, making it challenging to conduct absorption studies in vitro. Additional challenges also come from the inherent microbial population in the intestinal lining, which can lead to contamination by bacteria isolated alongside human enterocyte cells².  

Secondly, generating enough metabolic capacity or “horsepower” to reproduce phase 1 and 2 drug metabolism by in vitro human liver models requires continual culture perfusion (to mimic the bloodstream) and scale. Half a million primary cells perfused by large volumes of media (ranging between 1-2 mL) are required to achieve this goal which we have made possible via our FDA-recognized and industry-proven PhysioMimix^® Liver-on-a-chip model^3,4. 

Finally, the kinetics of drug absorption and metabolism by the gut can impact the fraction of a drug that is absorbed and available to the liver. Therefore, efforts to improve the in vitro to in vivo translation (IVIVE) of drug efficacy, safety and pharmacokinetic data have led to the emergence of more complex models where multiple organs, such as gut and liver, are fluidically linked to simulate drug absorption and first-pass metabolism.  

One of the major challenges when co-culturing two or more tissues together is establishing the conditions that maintain the functionality of both tissue types. We initially linked the established Caco-2 gut model to our industry-proven Liver-on-a-chip model⁵  using a chemically defined media that maintains hepatic metabolic functionality and intestinal barrier integrity.

Enabling human oral bioavailability to be profiled in vitro, rather than inferred from isolated gut and liver assays, the model provides a step forward. However, it remains limited by absent or low levels of enzyme and transporter expression. For some drug types, this limitation of the Caco-2 cell line results in failure to accurately profile a drug’s ADME behavior and bioavailability in humans. 

What’s the solution? An in vitro human Gut/Liver model 

To address the limitations of the Caco-2 cells, we have successfully developed a PhysioMimix primary in vitro human Gut/Liver-on-a-chip model and Bioavailability assay kit: Human 18 in collaboration with Altis Biosystems. Altis Biosystems’ RepliGut^®-Planar Jejunum model recreates the intestinal barrier using primary cells isolated from the human jejunum and expanded on a biomimetic scaffold. Primary gut tissue is then interconnected with our Liver-on-a-chip model using dual-organ supporting media, which is added to the basolateral side of the RepliGut’s Transwell^® and to the liver compartment of our PhysioMimix Multi-chip Dual-organ plate. Multi-chip Dual-organ plates enable six RepliGut/Liver-on-a-chip models to be cultured per plate. Up to three plates can be run simultaneously by one PhysioMimix Multi-organ System (18 replicates/run). 

Using well-studied drug compounds, we have demonstrated the improved predictive capacity of this primary in vitro human Gut/Liver model for profiling the ADME behavior of oral drugs compared to an equivalent Caco-2 Gut/Liver model. For the full study results, please read our Application Note.  

This in vitro human Gut/Liver model now enables intestinal absorption and hepatic clearance to be studied within a system that more precisely recapitulates the human process – versus traditional approaches operating in isolation. The added benefit is that the approach enables bioavailability estimations to be made too.  

By combining data generated using this fully human in vitro Gut/Liver model with in silico computational modelling tools, the in vivo pharmacokinetic parameters of orally administered drugs can now be more accurately estimated⁵.  

Here’s our recommendation for using the human in vitro Gut/Liver model 

We recommend utilizing this advanced human in vitro Gut/Liver model to change the risk calculus and gain confidence in your lead candidates. Using more human and physiologically relevant in vitro test models in lead optimization, or before clinical trials, ensures that only the candidate compounds with optimal drug absorption and bioavailability profiles are progressed into in vivo animal or FIH studies. Additionally, the approach can be used to refine the preclinical design, save costs, and align with the 3Rs (to support reductions in the number of animals required). 

Acknowledgement of where OOC is today

OOC/MPS models do have their limitations as they operate in small systems rather than a whole functioning organism, however, they significantly move the needle towards recapitulating the way that a human interacts with drugs before in vivo studies.  

OOC/MPS models should be used in a complementary manner to overcome the physiological-relevance limitations of traditional in vitro approaches and the human-relevance limitations of animal studies. By doing so, they enable a “bigger picture” view for more informed decision-making about the optimal therapeutic window (maximal efficacy/minimized toxicity), and to reduce the risk of identifying poor oral bioavailability during first-in-human studies for increased confidence in clinical success. 

So, don’t delay! The longer you wait, the greater the risk of cascading problems from miscalculated pre-clinical estimations. The PhysioMimix primary human in vitro Gut/Liver model and Bioavailability assay can be recreated in your own laboratory using the Bioavailability assay kit: Human 18 or accessed via our ADME Contract Research Services.

Useful links

Application notes

Connecting the gut and liver

ADME Contract Research Services

Applications: Bioavailability

Gut/Liver-on-a-chip models

References: 

1. Musther et al. (2014). Animal versus human oral drug bioavailability: Do they correlate?  

2. Dutton et al. (2019). Primary Cell-Derived Intestinal Models: Recapitulating Physiology 

3. Rubiano et al. (2021). Characterizing the reproducibility in using a liver microphysiological system for assaying drug toxicity, metabolism, and accumulation. 

4. Docci et al (2022). Exploration and application of a liver-on-a-chip device in combination with modelling and simulation for quantitative drug metabolism studies. 
5. Milani et al. (2022). Application of a gut–liver-on-a-chip device and mechanistic modeling to the quantitative in vitro pharmacokinetic study of mycophenolate mofetil.  

Drug discovery is becoming more complex as we move from a small molecule world to large molecules and new modality drugs. This shift brings new challenges, particularly in pre-clinical toxicology testing, where traditional animal models often fall short because of differences in genetics, metabolism or immunological response that make animal models unsuitable.

Pharmaceutical companies face increasing pressure to improve the efficiency of their drug discovery pipeline by reducing attrition rates. A significant portion of drug failures, up to 30% (Giri et al., 2015), are due to toxicity issues that were not predicted during pre-clinical testing. This high attrition rate highlights the need for more reliable and predictive toxicology testing methods that can narrow the current gap between pre-clinical and clinical outcomes.

The FDA Modernization Act 2.0 supports the increased use of alternative approaches to in vivo testing, by expanding the range of tools available to drug developers. Ultimately the FDA aims to ensure that companies are using the best tools available as well as the right combination of tools to avoid any adverse effects during clinical trials. The key to successfully including Organ-on-a-chip (also known as Microphysiological Systems) data in regulatory submissions lies in establishing a clear context of use. By positioning these tools often before in vivo studies, it is possible to be more confident that the right dose of a molecule will hit its intended target, delivering maximal efficacy without inducing toxicity. A more human-centric approach to toxicology testing may prevent potentially successful molecules from being lost in the clinic or before.

PhysioMimix Organ-on-a-chip adoption into toxicology workflows

Technologies, such as CN Bio’s PhysioMimix®  OOC, offer a reliable, reproducible and higher throughput (see our new Multi-Chip Liver-48 plate) option to accurately assess the same biomarkers in preclinical testing as the clinic. These more sophisticated in vitro models, which incorporate immune components, can detect highly specific and human-translatable endpoints compared to in vivo animal models.

One significant application of OOC technology is in pre-clinical toxicology testing. A recent survey by CN Bio and CiteLine showed that more than half of drug developers are using or planning to use OOCs for safety toxicology testing. Since in vivo safety tox studies are often the most complex and costly part of preclinical development, it makes sense to start here.

What are the benefits of Organ-on-a-chip adoption within toxicology workflows?

These technologies have a crucial role to play in reducing the riskiness of developing new drugs by providing:

• detailed assessments of acute and chronic exposure to gain deep mechanistic insights into a compound’s toxicological profile to guide the selection of the most appropriate animal species. More detailed and relevant data upfront will help improve in vivo study design, reducing the number of animals needed and lowering associated costs.
• improved translatability of pre-clinical data to clinical outcomes will ensure that drug developers can confidently take fewer, safer drugs forward.
• data for better patient study design; based on more clinically-relevant data, and data from different populations (e.g. the highly prevalent liver disorder MASLD and its more advanced form MASH) before drugs enter first-in-human (phase I) clinical trials reducing clinical risk.

Organ-on-a-chip versus animal models: reduce, refine or replace?

OOCs provide cost effective alternatives that complement animal studies to deliver human-translatable results faster. There is a clear advantage to using OOCs before starting in vivo animal testing. However, human OOCs are not attempting to be comparable to animal models especially where animal models are poor predictors of clinical outcomes. Their purpose is to catch what animal models might miss, or falsely flag as toxic due to interspecies differences. Additionally, for human-specific new modality drugs, OOCs offer a viable path forward where animal species are less suited to reduce unnecessary animal use.

Recent experiments by CN Bio using animal-on-a-chip models (or animal MPS) that utilize primary rat, or dog cells show that animal OOC data closely matches published animal testing results. These data demonstrate the system’s ability to recapitulate in vivo physiology, building confidence and trust to facilitate accelerated Organ-on-a-chip adoption. Additional data, guidance on broader context of animal OOC use, and its potential to inform species selection and safeguard research animal will be available in 2025.

By improving the translatability of data between the laboratory and the clinic, OOCs can make more accurate predictions of human responses in the pre-clinical phase, decreasing the risk of unexpected results in clinical trials. This can help to improve the efficiency of the drug development process. More accurate, human-relevant data not only addresses the limitations of traditional animal models and enhances the predictability of clinical outcomes, but also enables more effective drugs to reach the market faster.

To learn more about Organ-on-a-chip adoption and the roadmap to broader use, register here to watch a GEN Live Learning discussion featuring CN Bio’s CSO Dr Tomasz Kostrzewski and industry expert Dr Clive Roper on demand.

Register now

References

1. Giri S., Bader A. (2015). A low-cost, high-quality new drug discovery process using patient-derived induced pluripotent stem cells. Drug Discov. Today 20, 37–49. 10.1016/j.drudis.2014.10.011

Over the last 12 months, much has changed for the better in the field of non-alcoholic steatohepatitis research. New nomenclature has been gaining traction, with MASLD (metabolic dysfunction-associated steatotic liver disease) replacing the previously used term NAFLD (non-alcoholic fatty liver disease) to encompass broader steatosis types. Importantly, the term MASLD also helps to remove the stigma associated with the words “non-alcoholic” and “fatty”, which has been shown to result in reduced consideration in health policy and negative patient behaviors such as decreased likelihood to seek medical attention.

There is also increased hope for patients presenting with MASLD’s more advanced fibrotic form MASH (metabolic dysfunction associated steatohepatitis). This potentially fatal disease arises from fat accumulation, inflammation and fibrosis of the liver. Until recently, there were no approved drug treatments to treat MASH; however, a significant milestone was reached in 2024, with the FDA approval of Resmetirom (Rezdiffra), an oral thyroid hormone receptor beta (THR-β) agonist.

Challenges in MASH drug development

Despite Resmetirom’s approval, many challenges remain. Firstly, the long-term safety profile of the drug is yet to be determined, as adverse effects can develop over time. Secondly, MASLD/MASH is a complex multifactorial disease that is often linked to obesity and metabolic syndromes. Because of this, it is possible that Resmetirom alone may not prove to be an effective treatment for all patients. This is a significant ongoing challenge for the healthcare sector, and one that will likely require combination therapies, targeting different aspects of the disease and extending beyond the liver to other organs, to combat the disease effectively.

Although the pipeline of drugs and drug combinations in clinical trials for MASLD/MASH has been described as “an impending tsunami”, much remains unknown about the onset of this disease, and efficacy testing in humans is both risky and costly. Lack of efficacy remains the leading cause of drug failures in clinical trials so, although the future looks brighter, the need for better drugs remains. In response, a rising tide of research and development has been observed.

Ultimately, late-stage drug failures due to poor efficacy result from the inability of preclinical (in vitro and in vivo) models to accurately predict human outcomes. The reason for this is that their physiological relevance to humans is inherently limited.

Organ-on-a-chip models offer greater relevance

Growing demand for more predictive preclinical tools to plug the “relevance” gap has led to the development of several new approach methodologies (NAMs). Leading the alternative methodology sector are organ-on-a-chip (OOC) models, or microphysiological systems (MPS), that enable a deeper mechanistic understanding of disease and deliver more translatable preclinical research and development insights.

OOC models utilize co-cultures of primary human cells grown on perfused scaffolds to mimic blood flow. These human organ mimics recapitulate the key microarchitecture, phenotypes and functions of healthy or diseased human organs in the laboratory. For MASLD and MASH, the disease phenotype is induced from healthy liver-on-a-chip cultures, comprised of primary human hepatocytes, stellate and Kupffer cells, following treatment with a proprietary HEP-Fat media (Figure 1). The model can be used to study genetic polymorphisms, the mechanisms of disease and drug action.^1,2

Physiologically relevant mechanistic insights are generated from OOC models through the measurement of clinically relevant inflammatory, steatosis and fibrotic biomarkers associated with the MASH phenotype, providing high-content quantitative data that aids data translatability between the laboratory and the clinic. Human organ MASH models have been characterized in various studies and validated using a variety of drugs, including those in recent clinical trials (Figure 2).^1,2,3

Where animals capture the complexity of a whole organism, lab-grown organs-on-chips demonstrate how the disease mechanism, or the effects of drugs, may differ in a human setting. In 2020, Vacca et al. explored the transcriptomic profile of this MASH model, demonstrating a strong profile compared to human data and one that more closely replicates changes found in MASH patients compared to the murine Western diet (WD) model.³ The complementary use of OOC models alongside animals gives a broader insight into a drug’s potential in a human setting.

Murine model testing of drugs targeting MASLD/MASH is commonplace; however, their use is less suited for the development of drugs targeting human metabolism. As metabolism in mice is very different to humans, OOC models offer a viable path forward where translatability to humans is likely to be poor. Likewise, key genetic factors associated with MASH may be expressed differently, or not at all, in animal models. In late 2023, organ-on-a-chip data supported the initiation of Inipharm’s INI-822, targeting human metabolism, into clinical trials.⁴ This was the first example of OOC data supporting the clinical progression of a drug for this complex metabolic disorder, demonstrating the transformative potential of the approach.

Using OOC models to predict bioavailability

In addition to studying this liver disease, or the efficacy of drugs within the liver in isolation, OOC can now link healthy or diseased organs together to mimic human processes such as first-pass metabolism and provide information regarding a drug’s bioavailability – a key parameter required for dose setting in the clinic.

Animals are notoriously poor predictors of human bioavailability. In a seminal study investigating 184 compounds, animal models were found to have a weak correlation with bioavailability in humans (R²=0.34).⁵ A new approach that is physiologically relevant to humans, which combines oral absorption and hepatic metabolism, is therefore required to provide more accurate estimations and dose settings ahead of the clinic. This approach would give drug developers greater confidence that an optimal concentration of a drug will reach its intended target, to elicit maximal efficacy whilst minimizing toxicity and decrease the risk of late-stage failures.

Furthermore, interlinked multi-organ systems facilitate inter-organ cross-talk, opening a world of future possibilities to study shared risk factors involved in MASLD/MASH, such as diabetes, obesity and metabolic syndrome. They offer the potential to explore key bidirectional axes, such as the gut–liver axis, to identify how poor diet causes dysbiosis of the gut microbiome and how this affects liver health and causes inflammation.

A deeper understanding of interactions such as these, and other influencing axes, will help to gradually unlock the unknowns of this multifaceted disease, holding promise for future treatments in years to come. But let’s not forget that drug treatments are a last resort, as most forms of MASLD and its progression into MASH are entirely preventable.

Alongside the development of new treatments, it is key that we support education efforts to improve disease awareness and promote healthy lifestyle modifications such as exercise and dietary changes. Changes such as these can prevent, halt or even reverse disease onset and remain an essential factor to curbing the disease’s growing prevalence.

About the author:

Dr. Oliver Culley is a Senior Scientist at CN Bio where his work focuses on organ-on-a-chip modelling of NASH disease using the PhysioMimix OOC System. Oliver completed a BSc (Hons) Biomedical Sciences at University of Manchester and a PhD in Cell Biology at King’s College London, before undertaking postdoctoral work in Bioengineering at Queen Mary University of London. Oliver has over 10 years of experience in academia and industry with extensive experience in high-content imaging and in vitro assay design within the fields of cell and molecular biology.

References:

1. Kostrzewski T, Maraver P, Ouro-Gnao L, et al. A Microphysiological System for Studying Nonalcoholic Steatohepatitis. Hepatol Commun. 2020;4(1):77–91. doi: 10.1002/hep4.1450
2. Kostrzewski T, Snow S, Battle AL, et al. Modelling human liver fibrosis in the context of non-alcoholic steatohepatitis using a microphysiological system. Commun Biol. 2021;4:1080. doi: 10.1038/s42003-021-02616-x
3. Vacca M, Leslie J, Virtue S, et al. Bone morphogenetic protein 8B promotes the progression of non-alcoholic steatohepatitis. Nat Metab. 2020;2:514–531. doi: 10.1038/s42255-020-0214-9
4. CN Bio. CN Bio PhysioMimix® Organ-on-a-Chip data supports Inipharm’s INI-822 for metabolic liver disease treatment now in clinical testing. https://cn-bio.com/physiomimix-data-supports-inipharms-ini-822-for-metabolic-liver-disease-treatment/. Published December 4, 2023. Accessed May 30, 2024.
5. Musther H, Olivares-Morales A, Hatley OJD, Liu B, Rostami Hodjegan A.  Animal versus human oral drug bioavailability: Do they correlate? Eur J Pharm Sci. 2014;57(100):280–291. doi: 10.1016/j.ejps.2013.08.018

For more information about our products and services visit:

Our PhysioMimix OOC range of microphysiological systems

Our NASH Application page

Our NASH-in-a-box kit

Our NASH Contract Research Service