Organ-Chips: Enhancing our ability to translate science into medicines

Written by:

Dominic Williams

Associate Director, Clinical Pharmacology and Safety Sciences, R&D

Rhiannon David

Microphysiological Systems (MPS) Scientific Lead, Clinical Pharmacology and Safety Sciences, R&D

Predicting how patients will respond to candidate drug molecules is one of the greatest challenges facing pharmaceutical researchers. No new molecule can proceed, no matter how effective, without an acceptable safety profile. This is the principle behind ‘right safety’, one of the critical ‘5Rs’ that underpin the success of AstraZeneca’s R&D productivity. ‘Right Safety’ also spurs us to seek new technology and ways of working which can improve how we model likely toxicities of novel compounds. Traditional pre-clinical animal models provide a good indication of the safety and effectiveness of a compound but cannot be fully translatable to humans. By using humanised models that predict patient response more accurately than traditional animal models, we are aiming to further enhance our success rates for taking molecules from candidate nomination to completion of Phase III trials – already up from 4% in 2010 to more than 19% in 2018.1

Three recent publications in high-impact journals, to which we were authors, outline the critical importance of more predictive and human-relevant models. The first, in collaboration with Emulate, Inc., is published in Science Translational Medicine. This study applied cutting-edge Organ-on-Chips (Organ-Chips) technology to create multi-species Liver-Chip models. This has enormous potential value for helping us decide whether a toxicity signal in pre-clinical research should affect whether we pursue development of a compound in humans.

The second, as part of a wider EU consortium that was published in Nature Reviews Drug Discovery, reviewed the emerging body of evidence that drug induced-liver injury (DILI), as it occurs in humans, can be a multi-step and complex multicellular disease process. DILI can be caused by various chemical insults and present as an array of different pathologies dependent on the specific function of the liver that is impaired. This means that prediction of all forms of DILI may be inherently too complex for simple models, such as single cell culture screening strategies.3 More complex models such as Organ-Chips have the potential to help overcome these challenges.

We are both part of Clinical Pharmacology and Safety Sciences (CPSS), R&D at AstraZeneca. The Global Head of CPSS is Senior Vice President Stefan Platz. Speaking to media around these publications he made clear the importance of Organ-Chips: 

Organ-Chips technology has the potential to enhance and accelerate our ability to translate science into innovative medicines for patients. This collaborative Liver-Chip work points the way to broader applications of this technology to improve our predictions of adverse drug reactions before drug candidates enter clinical trials.

Organ-Chips are one element of our predictive science approach, which aims to secure quicker and more accurate predictions of drug toxicity in humans, while allowing for more targeted use of animal experiments. This allows us to make better decisions earlier in the drug discovery process.

These advances are transforming how we do science and create potential new medicines. We know we cannot advance science alone and collaborate with other pioneers, including Emulate, Inc., the Wyss Institute at Harvard University and TissUse, to combine our expertise and drive forward innovation.

Our third recent publication in a high impact journal, this time in Nature Biomedical Engineering, was a collaboration with the Wyss Institute and focused on reproducing Human Bone Marrow on a Chip.4

This immunofluorescent image shows the multiple cell types that develop within the human bone marrow chip (magenta: erythroid cells, yellow: megakaryocytes, blue: other hematopoietic cells). Credit: Wyss Institute at Harvard University

In traditional bone marrow tissue assays, cells are extracted from healthy volunteers or patients, placed in culture medium and often are bathed in the drug candidate for extended periods. These static models cannot reflect the dynamic interaction of the different cells in the bone marrow, and do not replicate the natural environment, leading to cell loss over time. As a result, these studies can only offer limited predictions on the impact of drug candidates, radiation or genetic mutation on bone marrow.  

Using refined Organ-Chips models developed in collaboration with the Wyss Institute, we produced a ‘mini human bone marrow’ in the lab, which has the potential to be used to enable drug discovery, basic research and translational studies for a wide range of haematopoietic disorders and toxicities, and potentially offer a human-specific alternative to animal testing for regulatory assessment.

Taking just one finding, the vascular channel of our bone marrow chip allows for more natural perfusion of culture medium, sustaining cells for an extended period compared to traditional cell culture. This would permit experiments with potential new drugs over a sustained period of time and in a context that is more similar to how a human patient would receive the potential drug.

So what does the future hold? The concept of a ‘body on a chip’ is being investigated, where multiple Organ-Chips are connected to ensure each is taken out of isolation and interacts with other organs. We are already starting to investigate multi-organ systems, such as Liver-Chip and Pancreas-Chip combinations to model Type 2 diabetes. Preliminary results suggest that interactions between an insulin resistant Liver-Chip and a Pancreas-Chip affect insulin release in the same way as in Type 2 diabetes.5 Once the system is fully tested, our aim is to use this as a diabetes model for target validation.

We believe that in maybe 5-10 years, we will see progress made on ‘disease on a chip’ models, where instead of healthy tissue, we will be able to test novel compounds directly against the diseases they are designed to treat.

The field of predictive science is advancing rapidly. Humanised models bridge the gap between animals and humans and are a big step forward compared to conventional human cell cultures which have been used for many years. They provide an environment in which human cells behave more like they would in the body, generating data about toxicity, efficacy and other key effects that are more relevant to patients than previous methods.

Organ-Chips are just one of a host of exciting innovations at the fingertips of AstraZeneca scientists. If you share our belief that the best science doesn’t happen in isolation, then AstraZeneca is keen to hear from you. Search for open positions in Clinical Pharmacology and Safety Sciences here, and for more information on our partnering opportunities visit:


  1. Morgan P et al. Impact of a five-dimensional framework on R&D productivity at AstraZeneca. Nature Reviews Drug Discovery, 2018; 17,167-181
  2. KJ Jang et al. Reproducing human and cross-species drug toxicities using a Liver-Chip. Science Translational Medicine, 2019; 11, 517
  3. Weaver RJ et al. Managing the challenge of drug-induced liver injury: a roadmap for the development and deployment of preclinical predictive models. Nature Reviews Drug Discovery, 2019;
  4. Chou DB et al. On-chip recapitulation of clinical bone marrow toxicities and patient-specific pathophysiology. Nat Biomed Eng, 2020;
  5. Bauer S et al. Functional coupling of human pancreatic islets and liver spheroids on-a-chip: Towards a novel human ex vivo type 2 diabetes model. Nature Scientific Reports 7: 14620(2017)