Background: Liver cancer primarily develops in patients with chronic liver disease (CLD), yet most cases are diagnosed at an advanced stage with poor prognosis. While clinical surveillance of patients with CLD generates extensive longitudinal data, its unstructured free-text nature hinders large-scale research. To unlock this real-world evidence, we developed a scalable framework using open-source Large Language Models (LLMs) to transform unstructured clinical text into structured data. Methods: We conducted a multi-stage evaluation of LLM-based extraction from multi-source clinical documentation of liver transplant recipients. A calibration set comprising 507 reports (414 radiology, 65 pathology, and 28 liver transplant assessment reports) from 30 patients was manually annotated to benchmark four open-source LLMs (Llama-3.1-8B, Llama-3.3-70B, OpenBioLLM-70B, DeepSeek-R1-8B) against a regular expression baseline across 73 tasks. To ensure structured outputs, we compared constrained decoding (Guidance and Ollama packages) against unconstrained prompting across 5,590 prompt-output pairs. The finalised pipeline was then applied to the full cohort of 835 patients transplanted in our centre over the past decade. Results: Among the models tested, Llama-3.3-70B performed best, exceeding 90% accuracy on 59/73 tasks, outperforming both a medically fine-tuned model (OpenBioLLM-70B) and a smaller variant (Llama-3.1-8B). Constrained decoding achieved 99.9% format adherence, far surpassing unconstrained prompting (87.4%). Applied to the full cohort, the pipeline successfully analysed 22,493 reports to generate 37,125 datapoints (45 variables, 835 patients) without manual annotation. Further analysis confirmed known liver cancer risk factors (male sex, viral hepatitis, smoking, diabetes), and allowed for reconstruction of longitudinal disease timelines. Conclusions: This work provides a scalable blueprint for transforming real-world clinical free-text into structured formats, paving the way for accelerated, data-driven research into complex pre-cancerous diseases like CLD. ### Competing Interest Statement H.P. has received research funding from AstraZeneca. Z.G. has received research funding from GE healthcare. M.A.F. is a current employee and stockholder of AstraZeneca. M.H. has received speakers fees from Sirtex medical, consultancy fees from Quotient Therapeutics, Ensocell, Boston Scientific and Spliceor, in addition to unrestricted grant support from AstraZeneca and Pfizer. M.C.O. is a co-founder and employee of 52 North Health Ltd, has received research funding from GE HealthCare, and speaking fees from GSK. The other authors declare no competing interests. ### Funding Statement This study was funded by AstraZeneca UK Limited and the NIHR BioResource (G127831). ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The Health and Social Care Research Ethics Committee A of the Health and Social Care Business Services Organisation gave ethical approval for this work (REC reference 20/NI/0109, IRAS 285521) I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes The data underlying this article cannot be published due to confidentiality considerations.

Senescent cells accumulate in chronically diseased liver tissues and are known to actively contribute to disease pathology. To date, these studies have predominantly focussed on senescence in epithelial cells, such as hepatocytes and biliary epithelial cells, and senescence in liver endothelial cells remains largely understudied. Here, we utilise publicly available single-cell RNA-sequencing data, immunohistochemical and immunofluorescent staining to detect senescent endothelial cells within chronically diseased human liver tissues. Next, we develop a novel protocol for the induction of paracrine senescence in primary human liver endothelial cells and explore their functionality. We demonstrate that senescent liver endothelial cells exhibit a reduced scavenging capacity but are still able to support lymphocyte recruitment under physiological flow conditions in vitro. Mechanistically, we determine that inducible T cell costimulator ligand (ICOSL) is an important factor in the specific recruitment of CD4+ T cells, but antibody blockade, genetic knockdown and genetic overexpression of ICOSL in endothelial cells has no effect on CD8+ T cell recruitment. Finally, we show that ICOSL gene expression is upregulated in chronically diseased tissues, present in scar-associated endothelial cells and correlates to CD4+ T cell infiltration. This is the first study to demonstrate that senescent human liver endothelial cells can potentially shape the liver immune microenvironment in chronic liver disease. Targeting senescent endothelial cells could present new therapeutic opportunities to treat chronic liver diseases.

Abstract Metabolic dysfunction-associated steatotic liver disease (MASLD), previously known as NAFLD (non-alcoholic fatty liver disease), is a growing global concern, affecting nearly a third of the world’s population. This umbrella term covers a range of liver pathologies, from reversible disease stages like simple steatosis, to irreversible conditions such as cirrhosis and hepatocellular carcinoma (HCC). MASLD, which may result from metabolic risk factors like obesity and type 2 diabetes, is projected to increase due to a rise in sedentary lifestyles. This review addresses genetic influences that predispose to disease development, including the role of risk-conferring and protective/preventive alleles. The epistatic relationships between genetic variants can significantly influence the development and progression of MASLD. Key genetic variants, such as those located in the PNPLA3, TM6SF2, and MBOAT7 genes, often interact to exacerbate MASLD severity and play key roles in lipid metabolism and liver inflammation. For example, the co-expression of certain PNPLA3 and TM6SF2 variants increases the risk of advanced fibrosis and HCC. Some variants located in HSD17B13 and MTARC1 offer protective effects, reducing the risk of severe liver disease despite comorbidities such as obesity, and can mitigate the harmful effects of these risk alleles. Additionally, the potential of polygenic risk scores (PRS) to predict MASLD development and its complications is also discussed, although challenges remain, particularly in underrepresented populations due to the lack of comprehensive catalogues of genetic variation. Understanding these complex gene-gene interactions and the role of the environment underscores the importance of considering epistatic relationships when assessing MASLD risk and developing personalized therapeutic strategies, which could ease the future burden on healthcare systems.

Somatic variants accumulate in non-malignant tissues with age. Functional variants, leading to clonal advantage of hepatocytes, accumulate in the liver of patients with acquired chronic liver disease (CLD). Whether somatic variants are common to CLD from differing etiologies is unknown. We analyzed liver somatic variants in patients with genetic CLD from alpha-1 antitrypsin (A1AT) deficiency or hemochromatosis. We show that somatic variants in SERPINA1, the gene encoding A1AT, are strongly selected for in A1AT deficiency, with evidence of convergent evolution. Acquired SERPINA1 variants are clustered at the carboxyl terminus of A1AT, leading to truncation. In vitro and in vivo, C-terminal truncation variants reduce disease-associated Z-A1AT polymer accumulation and disruption of the endoplasmic reticulum, supporting the C-terminal domain swap mechanism. Therefore, somatic escape variants from a deleterious germline variant are selected for in A1AT deficiency, suggesting that functional somatic variants are disease-specific in CLD and point to disease-associated mechanisms.

In mice, glucagon regulates lipid metabolism by activating receptors in the liver; however, its role in human lipid metabolism is incompletely understood. Here we describe three normal-weight individuals from a consanguineous family with early-onset hepatic steatosis and/or cirrhosis. Using exome sequencing, we found they were homozygous for two missense variants in the glucagon receptor gene (GCGR). In cells, the double GCGR mutation reduced cell membrane expression and signaling, resulting in an almost complete loss of function. Carriers of pathogenic GCGR mutations had substantially elevated circulating glucagon and amino acid levels and increased adiposity. Introducing the double GCGR mutation into human induced pluripotent stem cell-derived hepatocytes using CRISPR/Cas9 caused increased lipid accumulation. Our results provide an explanation for increased liver fat seen in clinical trials of GCGR antagonists and reduced liver fat in people with obesity and steatotic liver disease treated with GCGR agonists. ARTICLE HIGHLIGHTS:In this study, we investigated a consanguineous family in whom normal-weight individuals had hepatic steatosis and cirrhosis. Using whole-exome sequencing we found two rare homozygous variants in the glucagon receptor (GCGR) gene that cosegregated with the phenotype. In cells, the GCGR mutations result in a loss of function and increased lipid accumulation. These results highlight the potential risks associated with GCGR antagonists and the benefits of GCGR agonists, currently in clinical trials.

BACKGROUND & AIMS:Somatic and germline CIDEB mutations are associated with protection from chronic liver diseases. The mechanistic basis and whether CIDEB suppression would be an effective therapy against fatty liver disease remain unclear. METHODS:Twenty-one CIDEB somatic mutations were introduced into cells to assess functionality. In vivo screening was used to trace Cideb mutant clones in mice fed normal chow, western diet (WD), and choline-deficient, L-amino acid-defined, high-fat diet (CDA-HFD). Constitutive and conditional Cideb knockout mice were generated to study Cideb in liver disease. Isotope tracing was used to evaluate fatty acid oxidation and de novo lipogenesis. Transcriptomics, lipidomics, and metabolic analyses were utilized to explore molecular mechanisms. Double knockout models (Cideb/Atgl and Cideb/Pparα) tested mechanisms underlying Cideb loss. RESULTS:Most CIDEB mutations impaired function, and loss-of-function clones were positively selected under CDA-HFD but not all steatogenic diets. Cideb knockout mice were protected from WD-, CDA-HFD-, and alcohol-induced liver disease, with the strongest effect in CDA-HFD models. Hepatocyte-specific Cideb deletion ameliorated disease after MASLD (metabolic dysfunction-associated steatotic liver disease) establishment, modeling the impact of therapeutic small-interfering RNAs. Cideb loss protected livers via increased β-oxidation, specifically through ATGL and PPARα activation. CONCLUSIONS:Cideb deletion is more protective in some types of fatty liver disease. β-oxidation is an important component of the Cideb protective mechanism. CIDEB inhibition represents a promising approach, and somatic mutations in CIDEB might predict the patient populations who will benefit the most. IMPACT AND IMPLICATIONS:It is not clear why somatic and germline CIDEB mutations are protective in metabolic dysfunction-associated steatotic liver disease (MASLD). Cideb mutations are predominantly loss of function, and Cideb-deficient clones selectively expand in specific dietary contexts such as choline-deficient, L-amino acid-defined, high-fat diet-induced MASLD. Consistently, liver-wide deletion of Cideb ameliorates MASLD most profoundly after choline-deficient, L-amino acid-defined, high-fat diet feeding. Mechanistically, Cideb deficiency enhances hepatic fatty acid β-oxidation via ATGL and PPARα activation. These findings suggest that CIDEB inhibition might be most effective in patients with the subtypes of MASLD that promote the expansion of CIDEB mutant clones.

Senescence is a nonproliferative survival state that cancer cells can enter to escape therapy. In addition to soluble factors, senescence cells secrete extracellular vesicles (EV), which are important mediators of intercellular communication. To explore the role of senescent cell (SC)-derived EVs (senEV) in inflammatory responses to senescence, we developed an engraftment-based senescence model in wild-type mice and genetically blocked senEV release in vivo, without significantly affecting soluble mediators. SenEVs were both necessary and sufficient to trigger immune-mediated clearance of SCs, thereby suppressing tumor growth. In the absence of senEVs, the recruitment of MHC-II+ antigen-presenting cells (APC) to the senescence microenvironment was markedly impaired. Blocking senEV release redirected the primary target of SC signaling from APCs to neutrophils. Comprehensive transcriptional and proteomic analyses identified six ligands specific to senEVs, highlighting their role in promoting APC-T cell adhesion and synapse formation. APCs activated CCR2+CD4+ TH17 cells, which seemed to inhibit B-cell activation, and CD4+ T cells were essential for preventing tumor recurrence. These findings suggest that senEVs complement the activity of secreted inflammatory mediators by recruiting and activating distinct immune cell subsets, thereby enhancing the efficient clearance of SCs. These conclusions may have implications not only for tumor recurrence but also for understanding senescence during de novo carcinogenesis. Consequently, this work could inform the development of early detection strategies for cancer based on the biology of cellular senescence.Significance: Chemotherapy-treated senescent tumor cells release extracellular vesicles that trigger an immune response and suppress tumor recurrence.See related commentary by Almeida and Melo, p. 833

Background and aims Steatotic liver disease (SLD), which encompasses various causes of fat accumulation in the liver, is a major cause of liver fibrosis. Understanding the specific mechanisms of lipotoxicity, dysregulated lipid metabolism, and the role of different hepatic cell types involved in fibrogenesis is crucial for therapy development. Methods We analysed liver tissue from SLD patients and 3 mouse models. We combined bulk/spatial lipidomics, transcriptomics, imaging mass cytometry (IMC) and analysis of published spatial and single-cell RNA sequencing (scRNA-seq) data to explore the metabolic microenvironment in fibrosis. Pharmacological inhibition of sphingolipid metabolism with myriocin, fumonisin B1, miglustat and D-PDMP was carried out in hepatic stellate cells (HSCs) and human precision cut liver slices (hPCLSs). Results Bulk lipidomics revealed increased glycosphingolipids, ether lipids and saturated phosphatidylcholines in fibrotic samples. Spatial lipidomics detected >40 lipid species enriched within fibrotic regions, notably sphingomyelin (SM) 34:1. Using bulk transcriptomics (mouse) and analysis of published spatial transcriptomics data (human) we found that sphingolipid metabolism was also dysregulated in fibrosis at transcriptome level, with increased gene expression for ceramide and glycosphingolipid synthesis. Analysis of human scRNA-seq data showed that sphingolipid-related genes were widely expressed in non-parenchymal cells. By integrating spatial lipidomics with IMC of hepatic cell markers, we found excellent spatial correlation between sphingolipids, such as SM(34:1), and myofibroblasts. Inhibiting sphingolipid metabolism resulted in anti-fibrotic effects in HSCs and hPCLSs. Conclusions Our spatial multi-omics approach suggests cell type-specific mechanisms of fibrogenesis involving sphingolipid metabolism. Importantly, sphingolipid metabolic pathways are modifiable targets, which may have potential as an anti-fibrotic therapeutic strategy.

Background & Aims Somatic and germline CIDEB mutations are associated with protection from chronic liver diseases. The mechanistic basis and whether CIDEB suppression would be an effective therapy against fatty liver disease remain unclear. Methods 21 CIDEB somatic mutations were introduced into cells to assess functionality. In vivo screening was used to trace Cideb mutant clones in mice fed normal chow, western (WD), and choline-deficient, L-amino acid-defined, high-fat (CDA-HFD) diets. Constitutive and conditional Cideb knockout mice were generated to study Cideb in liver disease. Isotope tracing was used to evaluate fatty acid oxidation and de novo lipogenesis. Transcriptomics, lipidomics, and metabolic analyses were utilized to explore molecular mechanisms. Double knockout models ( Cideb/Atgl and Cideb/Ppara ) tested mechanisms underlying Cideb loss. Results Most CIDEB mutations showed that they impair function, and lineage-tracing showed that loss-of-function clones were positively selected with some, but not all fatty liver inducing diets. Cideb KO mice were protected from WD, CDA-HFD, and alcohol diets, but had the greatest impact on CDA-HFD induced liver disease. Hepatocyte-specific Cideb deletion could ameliorate disease after MASLD establishment, modeling the impact of therapeutic siRNAs. Cideb loss protected livers via increased β-oxidation, specifically through ATGL and PPARa activation. Conclusions Cideb deletion is more protective in some types of fatty liver disease. β-oxidation is an important component of the Cideb protective mechanism. CIDEB inhibition represents a promising approach, and somatic mutations in CIDEB might predict the patient populations that might benefit the most. ### Competing Interest Statement H.Z. and P.C. are co-founders of Quotient Therapeutics and Jumble Therapeutics, is an advisor for Newlimit, Alnylam Pharmaceuticals, and Chroma Medicines. H.Z. receives research support from Chroma Medicines. H.Z. and L.L. have a patent on CIDEB siRNA for liver disease (patent #63/328,557).

Hepatocellular carcinoma (HCC), the most common form of primary liver cancer, is a leading cause of cancer-related mortality worldwide1,2. HCC occurs typically from a background of chronic liver disease, caused by a spectrum of predisposing conditions. Tumour development is driven by the expansion of clones that accumulate progressive driver mutations3, with hepatocytes the most likely cell of origin2. However, the landscape of driver mutations in HCC is broadly independent of the underlying aetiologies4. Despite an increasing range of systemic treatment options for advanced HCC, outcomes remain heterogeneous and typically poor. Emerging data suggest that drug efficacies depend on disease aetiology and genetic alterations5,6. Exploring subtypes in preclinical models with human relevance will therefore be essential to advance precision medicine in HCC7. Here we generated a suite of genetically driven immunocompetent in vivo and matched in vitro HCC models. Our models represent multiple features of human HCC, including clonal origin, histopathological appearance and metastasis. We integrated transcriptomic data from the mouse models with human HCC data and identified four common human-mouse subtype clusters. The subtype clusters had distinct transcriptomic characteristics that aligned with the human histopathology. In a proof-of-principle analysis, we verified response to standard-of-care treatment and used a linked in vitro-in vivo pipeline to identify a promising therapeutic candidate, cladribine, that has not previously been linked to HCC treatment. Cladribine acts in a highly effective subtype-specific manner in combination with standard-of-care therapy.

Regulatory T cells (T regs ) control adaptive immunity and restrain type 2 inflammation in allergic disease. Interleukin-33 promotes the expansion of tissue-resident T regs and group 2 innate lymphoid cells (ILC2s); however, how T regs locally coordinate their function within the inflammatory niche is not understood. Here, we show that ILC2s are critical orchestrators of T reg function. Using spatial, cellular, and molecular profiling of the type 2 inflamed niche, we found that ILC2s and T regs engage in a direct (OX40L-OX40) and chemotaxis-dependent (CCL1-CCR8) cellular dialogue that enforces the local accumulation of Gata3 high T regs , which are transcriptionally and functionally adapted to the type 2 environment. Genetic interruption of ILC2-T reg communication resulted in uncontrolled type 2 lung inflammation after allergen exposure. Mechanistically, we found that Gata3 high T regs can modulate the local bioavailability of the costimulatory molecule OX40L, which subsequently controlled effector memory T helper 2 cell numbers. Hence, ILC2-T reg interactions represent a critical feedback mechanism to control adaptive type 2 immunity.