The Future of Liver Health: Can Human Models and Machine Learning Reduce Disease Rates?

October represents a month of liver awareness, marked by many countries as a time to reflect on an incredibly resilient, yet vulnerable, organ. Liver disease is a growing problem worldwide, but particularly in Western countries, with the UK seeing a more than fourfold increase in deaths over the past 50 years¹. At a time when deaths from other diseases are seeing a steady fall, this makes for a particularly concerning statistic.

Sadly, the vast majority of liver disease deaths could have been prevented by lifestyle changes alone, had they been diagnosed early enough. By the time symptoms appear, many patients are at an advanced stage, earning these diseases the label “silent killers”. With more people than ever at risk of liver-related illness, the time is right to ensure we are spreading awareness around the rising prevalence of conditions such as metabolic dysfunction-associated steatotic liver disease (MASLD), also known as metabolic dysfunction-associated fatty liver disease (MAFLD), type 2 diabetes and alcoholic liver disease, as well as the complications arising from obesity.

Insulin resistance in the liver is a common factor in many of these conditions. Over time, liver cells can stop responding properly to the hormone insulin, which supports blood sugar regulation, contributing to type 2 diabetes and fat buildup in the liver. Left undiagnosed, fatty livers can become inflamed, fibrotic and lead to severe conditions including liver failure and cancer.

Human metabolic disorders are complex, multifactorial and challenging to study. In the past, researchers had to choose between animal models, which do not adequately reflect human liver metabolism (rodents, for example, have a catabolic metabolism, whereas humans have an anabolic one), and simple cell cultures that lack the complexity of a real organ or tissue. Thankfully, new innovations are changing the game, bringing fresh hope for the future of liver health in the form of predictive, human-relevant research tools that can be used to understand disease and accelerate the development of novel therapies.

Growing human livers in the lab to model disease

A recent collaborative pre-print publication, between Novo Nordisk and organ-on-a-chip (OOC) pioneers at Massachusetts Institute of Technology (MIT), utilized an advanced microphysiological system (MPS) to culture three-dimensional liver microtissues from primary human hepatocytes². These miniaturized, lab-grown organs were perfused by cell culture media to mimic the bloodstream, maintaining their physiology and function for weeks at a time.

In the study, the authors exposed liver microtissues to certain metabolic stressors in order to simulate poor diet and metabolic overload. In humans, the drivers of progressive hepatic insulin resistance interact in complex ways and are traditionally difficult to study, however, the liver MPS microtissues provide an ideal test system to isolate and understand their effects under controlled conditions:

• Normal vs high insulin (to mimic hyperinsulinemia, an early diabetes hallmark)
• Normal vs high glucose (to simulate hyperglycemia)
• Normal vs elevated fatty acids (to reflect excess dietary fat)
  
  

Using this model, high insulin levels in the liver microtissues were found to trigger insulin resistance in just one week, and the insulin-resistant phenotype worsened when high insulin levels were combined with high glucose and fatty acids. Molecular level investigations indicated that these conditions effectively “rewired” gene expression patterns, disrupted insulin signaling and caused fat and bile acid to accumulate inside hepatocytes. Notably, once the metabolic stressors were removed, many of these adverse effects were partially reversed, as is seen with patients who improve their diet and levels of exercise.

The findings showcased that:

• Liver MPS can model human hepatic insulin resistance.
• It is possible to dissect the individual and synergistic contributions of hyperinsulinemia, hyperglycemia and elevated lipids to the disease pathophysiology in vitro.
• Liver MPS can facilitate testing novel therapeutics aimed at reversing or preventing insulin resistance.

Additionally, the liver MPS reversibility clearly aligns with advice from medical practitioners that lifestyle changes to remove poor diet and metabolic stress factors may reverse early-stage changes.

Bridging the translation gap from lab to clinic with machine learning

Recreating human diseases in the laboratory isn’t the end of this transformative journey. Since last year’s liver awareness month, researchers at MIT have gone a step further towards bridging the translation gap between laboratory data and clinical outcomes. A recent pre-print publication demonstrates the use of a systems biology and machine learning framework to map how MPS model experimental conditions relate to human liver disease biology. ³

The authors describe the use of a MAFLD model, studied using liver MPS, as a case study to introduce the LIV2TRANS (latent in vitro to in vivo translation) machine learning framework. They demonstrate the framework’s potential for determining optimal experimental conditions, such as growth factors, cytokines and matrix proteins to make MAFLD laboratory models more predictive.

The study uncovered key pathways, TGF-β, JAK-STAT, androgen and EGFR signaling that, when manipulated in the model, made it more representative of the in vivo disease process. The authors anticipate that this approach will help to:

• Address the question of how complex an MPS model needs to be to recreate a disease of interest accurately.
• Accelerate the process of in vitro model design.
• Enhance the translatability of MPS-generated data to in vivo human contexts.

Why is this new research important?

Liver awareness month is about recognizing and raising awareness for the serious healthcare challenges associated with diseases of this vital organ, but it should also celebrate innovations from our research community. Ultimately, only through collaboration will we be able to bring about a future where we can effectively reduce the global impact of liver disease.

AI-Enabled Imaging Solutions for Complex 3D Organoid Screening

This application note demonstrates how next-generation confocal imaging technology combined with AI-powered analysis tools enables researchers to capture high-quality images from 3D organoids.
View App Note / Case Study

Advertisement

The studies cited above demonstrate that human liver disease research is entering an exciting new era. In vitro models are becoming ever more human-relevant, enabling researchers to more accurately study the onset and mechanism of disease, identify opportunities for intervention and predict the outcomes of novel therapeutics.

As the disparity between laboratory-based data and clinical outcomes narrows through the application of AI-optimized MPS models, the theoretical risk of unsuccessful clinical trials is likewise reduced. In conjunction with enhanced public education efforts, there is renewed optimism for achieving a pivotal moment at which liver disease can be anticipated, prevented and reversed, to decrease its prevalence.

References

1.  Williams R, Aspinall R, Bellis M, et al. Addressing liver disease in the UK: a blueprint for attaining excellence in health care and reducing premature mortality from lifestyle issues of excess consumption of alcohol, obesity, and viral hepatitis. The Lancet. 2014;384(9958):1953-1997. doi: 10.1016/S0140-6736(14)61838-9

2.  Hellen DJ, Ungerleider J, Tevonian E, et al. A microphysiological model of progressive human hepatic insulin resistance. bioRxiv. Preprint posted online January 8, 2025:2025.01.08.631261. doi: 10.1101/2025.01.08.631261

3. Cadavid JL, Meimetis N, Griffith LG, Lauffenburger DA. Systems biology framework for rational design of operational conditions for in vitro / in vivo translation of microphysiological systems. bioRxiv. Preprint posted online January 22, 2025:2025.01.17.633624. doi: 10.1101/2025.01.17.633624

This article is based on research findings that are yet to be peer-reviewed. Results are therefore regarded as preliminary and should be interpreted as such. Find out about the role of the peer review process in research here. For further information, please contact the cited source.