Biomarkers in human liver cell assays for in vitro preclinical safety testing

22% of clinical trials fail due to hepatotoxicity, and the number rises to 32% when talking about market withdrawals. These figures highlight a persistent gap in our ability to predict liver-related adverse effects before a drug reaches patients. As drug developers place greater emphasis on New Approach Methodologies (NAMs), in vitro hepatic biomarkers have become a key tool for early safety assessment within these human-relevant models.

What are biomarkers in in vitro liver models and why do they matter for preclinical safety?

Biomarkers are measurable characteristics or molecules found in the body that serve as indicators of normal biological processes, disease states, or responses to treatments. In a clinical setting, hepatic biomarkers are typically measured in blood to indicate liver injury or dysfunction. More recently, these biomarkers are being applied in in vitro liver models, where their levels or functional changes can be used to interrogate how liver cells respond to a drug.

Their relevance lies in their ability to indicate stress or injury before overt cytotoxicity becomes apparent, providing a window into the mechanistic toxicity. Classic biochemical markers in the clinical setting such as albumin, alanine aminotransferase (ALT), and aspartate aminotransferase (AST) remain widely used in vitro because they report on synthetic function and hepatocellular injury. The growing adoption of mechanistic indicators, such as α-glutation-S transferase (α-GST) and miR-122, has significantly strengthened the connection between in vitro observations and clinical signals. Both markers rise in a dose- and time-dependent manner during hepatotoxic responses to reference compounds like acetaminophen or fialuridine.

Alongside functional and injury-associated indicators, metabolic biomarkers are becoming increasingly important. Studies examining patient sera during acute and acute-on-chronic liver failure have revealed a pattern of increased acylcarnitines with simultaneous reductions in specific phosphatidylcholines and sphingomyelins. Translating these signatures back into primary human hepatocyte and hepatic cell line cultures shows that even simple models can recapitulate these metabolic shifts when exposed to hepatotoxic compounds. On another level of cell culture complexity, amino acids such as aspartic acid, arginine, glutamine, and phenylalanine have also shown an alteration in cell culture media of organoids exposed to hepatotoxic drugs, compared to nonhepatotoxic ones. Such features expand in vitro readouts beyond single-endpoint assays and bring the metabolic dimension of liver injury into early safety testing.

Biomarkers matter because they offer sensitivity, mechanism, and, increasingly, clinical relevance. When a model captures enzyme activity, transporter competence, and injury markers over extended exposure periods, it mimics the multifactorial nature of human liver injury. 

Integrating advanced liver cell platforms and biomarker endpoints in CRO testing workflows

Contract Research Organizations (CROs) are now incorporating platforms such as spheroids, organoids, and microengineered liver-on-a-chip (LOCs) into their preclinical portfolios to better mimic the hepatic physiology in vitro. When combined with biomarker panels, these systems outperform traditional models in sensitivity and mechanistic insight. 

During repeated-dose testing with acetaminophen or fialuridine, Foster et al. demonstrated that both spheroids and LOCs show progressive declines in albumin secretion, increases in miR-122 release and measurable ATP depletion at concentrations aligned with known human toxicity. On the other hand, Zhang et al. combined patient-derived organoids and liver-on-a-chip systems to enhance sensitivity for modeling hepatic function. They increased sensitivity in detecting drug-induced liver injury (DILI) using biochemical biomarkers compared to traditional two-dimensional (2D) plate cultures. The model displayed significantly increased release of hepatocellular injury biomarkers, like ALT and AST, upon treatment with known hepatotoxic drugs. 

Challenges in validating translational biomarkers for drug-induced liver injury in vitro

Despite the progress, the path from in vitro biomarker discovery to validated translational tools remains complex. Many markers respond differently across species, limiting direct comparison with non-clinical datasets. Even within human systems, primary hepatocytes vary by donor, affecting baseline biomarker levels and response sensitivity. This variability complicates attempts to define universal thresholds that reliably predict clinical outcomes.

Metabolomic biomarkers introduce another layer of difficulty. While acylcarnitines and lipid species differ clearly in patients with acute liver failure, reproducing these patterns in vitro requires cell models capable of sustaining physiologically relevant metabolic flux. Some hepatoma lines lack critical enzyme families, and even primary hepatocytes may lose functionality over time if culture conditions are not carefully maintained. Advanced platforms mitigate these limitations, but do not eliminate them entirely.

Finally, consistency in assay methods, sampling time points, and data interpretation will be essential before these markers can be applied as standard decision-making tools.

In vitro liver biomarkers are reshaping preclinical safety studies by improving sensitivity, extending mechanistic insight, and offering pathways to closer clinical translation. Their integration into advanced liver models brings preclinical predictions closer to real human outcomes, even though challenges in standardization and variability remain. As these tools continue to mature, they provide a more dependable foundation for evaluating risk and supporting decision-making throughout the development process.

Sources

Cuykx M, Rodrigues RM, Laukens K, Vanhaecke T, Covaci A. In vitro assessment of hepatotoxicity by metabolomics: a review. Arch Toxicol. 2018 Oct;92(10):3007-3029. doi: 10.1007/s00204-018-2286-9. 

Foster AJ, Chouhan, B, Regan, SL. et al. Integrated in vitro models for hepatic safety and metabolism: evaluation of a human Liver-Chip and liver spheroid. Arch Toxicol. 2019; 93:1021–1037. https://doi.org/10.1007/s00204-019-02427-4

Kaur G. Chapter 14 – Hepatic toxicity biomarkers. In: Gupta RC, editor. Biomarkers in Toxicology (2nd ed). Academic Press; 2019. p. 251-66. doi:10.1016/B978-0-12-814655-2.00014-1.

Rentschler S, Doss S, Kaiser L, Weinschrott H, Kohl M, Deigner HP, Sauer M. Metabolic Biomarkers of Liver Failure in Cell Models and Patient Sera: Toward Liver Damage Evaluation In Vitro. Int J Mol Sci. 2024 Dec 23;25(24):13739. doi: 10.3390/ijms252413739.

Zhang CJ, Meyer SR, O’Meara MJ, Huang S, Capeling MM, Ferrer-Torres D, Childs CJ, Spence JR, Fontana RJ, Sexton JZ. A human liver organoid screening platform for DILI risk prediction. J Hepatol. 2023 May;78(5):998-1006. doi: 10.1016/j.jhep.2023.01.019.