Rethinking how we fight a silent epidemic
Liver disease is a rapidly growing global challenge that often advances without obvious symptoms until it becomes life-threatening. According to the British Liver Trust, mortality rates from liver disease have quadrupled since 1970, a stark contrast to declining death rates seen in other major non-communicable illnesses such as heart and lung diseases. Even more concerning, premature deaths from liver disease and liver cancer in England have risen by 64% over the past two decades. When the COVID-19 pandemic hit, liver-related deaths increased by an additional 21% between 2019 and 2021 – evidence that external stressors can amplify an already dire situation.[1]
Yet the reality is that most chronic liver conditions remain undetected until they are in advanced stages, leaving patients with few treatment options and the looming possibility of needing an organ transplant.[2] Unfortunately, a limited donor pool and lengthy waiting times mean that, for too many, the window for intervention closes before suitable solutions are found. While traditional animal models have played a helpful role in the early exploration of liver disease, they often fail to capture subtle shifts in human cellular behaviour – the very changes that might unlock novel preventive or therapeutic strategies.[3]
Why human-first models are essential
Many promising therapies stumble during clinical trials because preclinical testing rarely reflects the intricacies of human liver disease.[4] Subtle fibrotic processes and small yet pivotal shifts in hepatic cell populations often go unnoticed in these models – early cues that could pave the way for effective early intervention. By the time these types of changes are detected in patients, the disease may already be too advanced for treatments to make a meaningful impact.
Shifting to human-centric platforms such as organ-on-a-chip systems and donor organ perfusion devices – combined with computational modelling of gene behaviour in human models, powered by large-scale causal datasets – offers much-needed insight into the real-time dynamics of liver function. In these near-physiological platforms, researchers can track early markers of liver injury – signals that can be overlooked in animal studies. This perspective is especially key for advanced conditions like cirrhosis, where a human-focused approach might allow us to intervene long before irreversible scarring takes hold.
Turning data into early interventions
One of the most promising human-focused tools is the ex vivo perfused liver system. In this setup, a donated liver is connected to a machine that circulates an oxygen-rich blood substitute, preserving the organ’s architecture and metabolic functions for several days. Researchers can monitor a wide range of health indicators, including pH, nutrient levels, and markers of liver function such as bilirubin or lactate. If any signs of inflammation or reduced liver function are measured, they know immediately that the liver is under stress.
This environment also provides a platform to test RNA-based therapies in conditions that closely resemble real human biology. Fibrogenic markers such as collagen I and alpha-smooth muscle actin can be measured through histology and gene expression assays, giving a window into how well new treatments suppress the scarring process. Meanwhile, tools such as RNA sequencing can map out shifts in thousands of genes at once, revealing broader mechanisms of action and highlight any off-target effects. By collecting and analysing these data, researchers can observe precisely how a therapy performs in a human liver before entering clinical trials.
Marrying machine learning with lab insights
Then comes translating these organ-level insights into effective therapies through precision therapeutics that target gene modifiers of disease. Precision therapeutics demand a holistic approach that unites patient-derived tissues, computational modelling, and advanced analytics. Machine learning (ML) can help sift through enormous volumes of transcriptomic and imaging data, pinpointing which genes and pathways might drive disease progression.
The real power, however, emerges when ML insights are tested iteratively in physiologically relevant models. Instead of a single pass – where a computational pipeline yields a list of possible targets – investigators can take those leads, evaluate them using ex vivo liver perfusions, organ-on-a-chip systems or multicellular co-cultures, then feed the results back into the algorithm, creating a ‘lab-in-a-loop’ approach. Over time, this cycle refines the most promising targets, discarding less viable options early and focusing resources on therapies most likely to succeed in clinical settings.
Targeting the hardest, underserved stages of liver disease
Cirrhosis, one of the final stages of liver disease, emphasises why earlier intervention is critical. At this point, the organ is heavily scarred, making it extremely difficult to reverse damage. Research efforts that centre on perfused donor livers with varying degrees of fibrosis, aiming to pinpoint what drives scarring at the molecular level, are not common. But, by examining real-time indicators – tying histopathological scores to advanced transcriptomics – scientists can identify precisely where and when RNA-based therapies might disrupt the fibrotic cascade.
Additionally, the scarring process makes delivery more challenging due to the reduction in liver function, a consequence of which is to impact the accessibility of therapies to their target cells. Studying failing livers and testing potential therapies on machine perfusion devices enables the distribution and efficacy of the therapy to be quantified within the disease setting, improving the chances of success in subsequent clinical trials.
A dynamic, human-focused system
However, moving from traditional drug development to precision therapeutics requires an ecosystem that is both technologically advanced and continuously adaptable. Clinical data can be fed back into computational models, which inform further ex vivo experiments, sharpening each subsequent round of candidate screening. This ongoing, iterative process stands in contrast to older pipelines that risk committing to the wrong targets for too long.
This approach can bring a new standard of care to liver disease – one where advanced platforms and data-driven human models collaborate to intervene earlier, specialise treatment, and minimise transplant dependency. Rather than hoping for incremental gains, the field can drive towards genuinely transformative outcomes.
Changing the odds for patients
Ultimately, the aim is to re-assess how we address this global crisis in liver disease treatments. By fusing human-focused science, AI-driven insights, and rigorous ex vivo validation, we can detect and treat liver disease sooner, tailoring interventions to the genes involved. This comprehensive approach offers the best chance of improving survival rates and easing the burden on transplant systems.
The question now is how quickly and successfully pharmaceutical companies will seize this momentum and open up new possibilities for precision liver care. Done correctly, these breakthroughs promise to change the odds for millions of people worldwide facing one of healthcare’s most pressing challenges.
By Kenny Moore, Head of R&D at Ochre Bio
References
[1] https://britishlivertrust.org.uk/information-and-support/statistics/
[2] https://pmc.ncbi.nlm.nih.gov/articles/PMC7829073/
[3] https://www.frontiersin.org/journals/drug-discovery/articles/10.3389/fddsv.2024.1355044/full
