Immune-competent human liver organoids model adaptive immunity-mediated idiosyncratic DILI

26 September 2025

Hepatotoxicity, including drug-induced liver injury (DILI), is a leading cause of patient attrition, drug development discontinuation and post-marketing withdrawal, resulting in significant morbidity and mortality. Conversely to direct DILI, idiosyncratic DILI (iDILI) is caused by drugs with little or no intrinsic hepatotoxicity, is not dose-dependent, and occurs with a longer latency period. The majority of DILI cases encountered in the clinical setting are idiosyncratic by nature.

Crucially, iDILI is not reproducible in animal models. Owing to numerous species-specific differences in metabolic rate, hepatic metabolizing enzymes and transporters, stress response capacity and immune system, animal testing is poorly predictive of hepatoxicity in humans. Modelling of hepatoxicity in vivo causes severe suffering in animals, and yet, the precise mechanisms of liver injury remain poorly understood.

Vulnerability to DILI is believed to arise from an interplay of drug and host factors. Host risk factors for DILI risk include age, gender, comorbidities, microbiome dysbiosis, immune status, and genetic polymorphism. In iDILI, both innate and adaptive immunity were found to participate to liver damage. Importantly, individual and combined effects of risk factors cannot be recapitulated in animals, resulting in underestimation or overestimation of risk of DILI in certain patient populations.

The genetic susceptibility to iDILI primarily originates from single nucleotide polymorphisms (SNPs) in immune-related genes, such as human leukocyte antigens (HLA) and cytokine-related genes. Certain HLA variants may increase the iDILI risk for one drug and be protective for another. 

For example, SNP in HLA can alter the binding affinity of HLA to peptides, as well as the binding affinity of peptide-HLA class I antigen complexes to T cell receptor (TCR), triggering cytotoxic CD8 T cell activation against hepatocytes. SNPs in cytokine genes can induce dysregulation of pro-inflammatory/anti-inflammatory cytokines, potentially amplifying immune-mediated liver damage. Epigenetic, transcriptional, and post-translational DNA transformations are also believed to play a role in genetic susceptibility to liver injury. However, due to significant inter-species differences in genetics and gene expression regulation, animals are poorly suited for assessing its effects on iDILI in humans.

For all these reasons, establishing advanced human-based in vitro platforms is vital for reliably predicting immune-mediated iDILI, obtaining human-relevant insights into the molecular and cellular pathways that lead to liver injury, and marketing compounds that are safer by design.

To faithfully recapitulate key features of patient-specific immune-mediated iDILI, Soussi et al. from the Cincinnati Children’s Hospital Medical Center have engineered an immuno-competent human liver organoid (HLO) microarray platform in which HLA-genotyped, donor iPSC-derived liver organoids are co-cultured with autologous CD8 T cells.

Matrigel being an ill-defined, mouse tumor–derived extracellular matrix with important batch variability, the team has seeded cells onto a fully human, matrix-free micropatterned hydrogel as to improve physiological relevance, uniformity and experimental reproducibility of formed organoids, all while maintaining their structural and functional characteristics. Given that primary human hepatocytes are limited in supply, the team has employed HLA genotyping in donor PBMC, PBMC reprogramming into iPSC and iPSC differentiation into donor-derived liver organoids to obtain a larger, more renewable, genetically diverse and HLA-defined panel of human liver models. By isolating CD8 T cells and antigen-presenting dendritic cells from the same donor’s PBMC, the researchers have ensured these T cells were already HLA‑restricted to donor’s HLA and that antigen‑specific CD8 T cells were expanded. In this manner, Soussi and colleagues have succeeded in engineering a scalable, immuno-competent platform for a controlled, reproducible testing of adaptive immunity-mediated iDILI.

In a proof-of-concept demonstration with the reference compound flucloxacillin, this immune-competent in vitro platform successfully supported antigen-specific CD8 T cell activation, cytokine secretion, and hepatocyte apoptosis in a genetically defined context, enabling cell donor-specific prediction of iDILI. Indeed, only flucloxacillin-primed CD8 T cells from genetically susceptible HLA-B*57:01 donors triggered hepatocyte death in HLOs pretreated with flucloxacillin, as evidenced by clinically relevant biomarkers of DILI such as cytokeratin-18 release, TNF-α and Granzyme B secretion, and DRAQ7-positive cells.

The autologous configuration of this immune-competent platform makes it possible to further investigate how patient-specific risk factors affect CD8 T cell priming and cytotoxic effector responses, leading to inter-individual variation in vulnerability to idiosyncratic drug toxicity.

Future research directions will include scaling the platform through automation, integration of iPSC-derived T cells, expansion of donor diversity, and single-cell transcriptional profiling to reveal signatures associated with heterogeneity in immune responses.

What Puts It on the Frontier

• Engineering of a physiologically-relevant human immune-competent liver organoid platform that models adaptive immunity-mediated idiosyncratic DILI (iDILI)

• The platform, composed of HLA-genotyped liver organoids co-cultured with autologous CD8 T cells, supports antigen-specific CD8 T cell activation and effector function

• Use of Matrigel-free culture for better physiological relevance, reproducibility and immune specificity

• Large, renewable, genetically diverse and HLA-defined panel of human liver models that reflects real‑world HLA variation and connects genetic vulnerability to adaptive immune responses

Impact Snapshot

• Improved prediction of immune-mediated iDILI

• Elucidation of the mechanisms responsible for idiosyncratic adverse drug effects. Mechanism-based evaluation of immunogenicity risk of compounds

• Patient-specific prediction of drug-induced immune-mediated adverse effects

Reference

Soussi FEA, Brusilovsky M, Buck E, Bacon WC, Dadgar S, Fullerton A, Durban VM, Barrile R, Helmrath MA, Takebe T, et al. Autologous Organoid‐T Cell Co‐Culture platform for modeling of Immune‐Mediated Drug‐Induced liver injury. Advanced Science. Sep 2025;12(43):e08584. https://doi.org/10.1002/advs.202508584