The versatile interplay between steatotic liver disease and liver cancer

Abstract

Steatotic liver diseases (SLDs), including metabolic dysfunction-associated steatotic liver disease (MASLD; formerly known as non-alcoholic fatty liver disease) and alcoholic liver disease, are the leading causes of chronic hepatitis, liver dysfunction, cirrhosis and liver cancer development. Severe manifestations of MASLD and alcoholic liver disease include metabolic dysfunction-associated steatohepatitis (MASH; formerly known as non-alcoholic steatohepatitis) and alcoholic steatohepatitis (ASH) or a combination thereof, termed metabolic dysfunction and alcohol-associated liver disease. While MASH-associated and ASH-associated liver cancers display common histopathological features, the underlying cellular and molecular pathophysiological mechanisms of each are distinct. Recent studies indicate that SLDs encompass previously unrecognized spectra of heterogeneous, metabolic, immunological and genetic diseases that, in combination with lifestyle measures and individual patient comorbidities, affect disease pathogenesis and therapy response. Here, we review the current knowledge of the molecular, genetic and cellular mechanisms underlying the transition of MASH or ASH to liver cancer as well as novel developments in liver cancer risk assessment in SLDs. We further discuss possible obstacles of differential diagnosis and outline current developments in the therapeutic management of both MASH-related and ASH-related liver cancer.

Introduction

Steatotic liver diseases (SLDs) are characterized by pathological lipid accumulation in the liver¹. The underlying causes for this can be manifold, including genetic predisposition, viral infection, lifestyle-related factors, age-induced changes, or a combination thereof². Moreover, SLDs are the leading cause of chronic hepatic inflammation and liver cancer, where liver cancer is the third most common cause of cancer-related deaths worldwide³. In recent decades, the incidence of SLD-associated cancers has steadily increased⁴. SLDs typically lead to the development of hepatocellular carcinoma (HCC), the most common subtype of liver cancer, which represents approximately 80% of all primary liver cancer cases. Less frequently, SLDs can lead to the development of cholangiocarcinoma (CCC) or mixed HCC/CCC, which together account for around 10–20% of cases⁵. Metabolic dysfunction-associated steatotic liver disease (MASLD) and alcoholic liver disease (ALD) are the two most prevalent SLDs. When metabolic dysfunction occurs together with chronic alcohol abuse, these factors act synergistically to induce hepatic steatosis and chronic inflammation, and the result is now classified as metabolic dysfunction and alcohol-associated liver disease (MetALD)^6–8.

MASLD is the hepatic manifestation of metabolic syndrome, which is characterized by obesity, high blood sugar and high cholesterol (Fig. 1a). It has emerged as the most common chronic liver disease worldwide^9,10, likely due to increasing global obesity rates. Increased consumption of highly processed foods and diets rich in lipids and carbohydrates, referred to as a Western diet, combined with a sedentary lifestyle^11,12, contributes to this development¹³. Conversely, ALD is the result of chronic, excessive alcohol intake¹⁴ (Fig. 1b). Both MASLD and ALD are characterized by hepatic steatosis, which is accompanied by lipotoxicity and metabolic reprogramming in hepatocytes, as well as an increase in activated innate and adaptive immune cells within the liver¹⁵. Chronic activation of hepatic immune cells, including lipid-associated macrophages, natural killer cells and CD8⁺ T cells, leads to metabolic dysfunction-associated steatohepatitis (MASH) in MASLD and alcoholic steatohepatitis (ASH) in ALD. Steatohepatitis is characterized by chronic necro-inflammation and develops progressively through a dynamic, cyclic process involving hepatic cell death, inducing an immune response and prompting liver regeneration. The chronic progression of MASH and ASH results in gradual extracellular matrix (ECM) deposition, fibrosis, cirrhosis, DNA damage and, ultimately, liver cancer^10,14,16,17 (Fig. 1a,b). Genetic and epigenetic alterations, such as single nucleotide polymorphisms (SNPs), have been described to predispose to the progression of MASLD and ALD to MASH and ASH, increasing the risk for the development of cirrhosis and liver cancer^18,19.

Fig. 1 |. The pathogenesis of steatotic liver diseases and hepatocellular carcinoma.

a, Metabolic dysfunction-associated steatotic liver disease (MASLD) and its advanced stage, metabolic dysfunction-associated steatohepatitis (MASH), primarily arise in the context of metabolic syndromes and are commonly associated with obesity, arterial hypertension, elevated plasma triglyceride levels, reduced high-density lipoprotein cholesterol and type 2 diabetes mellitus as a comorbidity. These factors lead to hepatic steatosis and metabolically dysfunctional hepatocytes. This process is further altered by genetic predisposition and dysbiosis of the gut microbiome. Crosstalk between hepatocytes and immune cells results in immune cell activation and dysfunction, promoting liver damage, proliferation and hepatocellular carcinoma (HCC) development. b, Alcohol-associated liver disease (ALD) and its advanced-stage disease, alcohol-associated steatohepatitis (ASH), are driven by chronic alcohol consumption. Similar to MASH, its development is also influenced by genetic risk factors and dysbiosis of the gut microbiome. Chronic alcohol intake leads to increased de novo lipogenesis and downstream hepatic steatosis. Acetaldehyde, a byproduct of alcohol metabolism, is elevated in patients with MASH, leading to excessive lipid peroxidation and DNA damage, which promotes disease progression by inducing liver injury, hepatocyte proliferation and, ultimately, HCC. ER, endoplasmic reticulum; HSC, hepatic stellate cell; LPS, lipopolysaccharides; MDSC, myeloid-derived suppressor cell; NK, natural killer; ROS, reactive oxygen species; T_H17 cell, T helper 17 cell; TNF, tumour necrosis factor; T_reg cell, regulatory T cell.

Importantly, many preclinical studies on MASH and ASH focus on earlier disease processes such as hepatic steatosis and inflammation, and thus do not address the mechanisms involved in MASH-to-HCC or ASH-to-HCC transition^20–22. It therefore remains unknown whether the mechanisms promoting MASH or ASH development are also involved in MASH-to-HCC or ASH-to-HCC transition²³. While MASH-associated and ASH-associated HCC are histopathologically similar, the cellular and molecular mechanisms underlying the development and progression of these diseases are partly distinct²⁴; however, the phenotypic and functional factors contributing to these differences are insufficiently understood.

The emergence of novel platforms for single-cell and spatial sequencing has helped to dissect the distinct immune-metabolic microenvironment in these diseases with unprecedented resolution^23,25–27. In both MASH and ASH, immune and metabolic features contribute to the development of primary liver cancers, including HCC and, less frequently, CCC²⁸. Importantly, a growing number of studies revealed that the biology of cancers in the liver is impacted by the immune-metabolic microenvironment in which these malignancies emerge^23,29. Accordingly, disease-specific cellular and molecular patterns have been identified that are linked to survival outcomes and immunotherapy response^17,30.

In this Review, we outline the cellular and molecular mechanisms underpinning SLD-associated HCC pathogenesis. We place particular emphasis on the critical tumour-promoting effects in the context of MASH and ASH such as lipid accumulation, inflammation and metabolic dysregulation. We further discuss how this knowledge translates into possible diagnostic and therapeutic approaches. Finally, we identify and outline how to address the remaining knowledge gaps to mitigate the growing burden of SLDs. Although MASH and ASH are increasingly recognized as systemic conditions with potential oncogenic effects beyond the liver, exploring their roles in the development of extrahepatic cancers is beyond the scope of this Review.

Pathophysiology of SLD-associated cancers

Environmental drivers of MASH-to-HCC and ASH-to-HCC transition

While the relationship between obesogenic diets and developing MASH is well characterized, the correlation of alcohol consumption in promoting the development of MetALD or ASH is less reliable given that patients self-report their alcohol intake³¹. For example, measuring ethyl glucuronide, a metabolite whose presence in the hair correlates positively with alcohol intake, demonstrated that alcohol intake was underreported in 25% of patients with suspected MASLD³². Thus, it can often be difficult to ascertain the direct impact of alcohol versus diet in individuals with an SLD. This challenge emphasizes the importance of using preclinical models where exposure to hepatotoxic compounds, such as ethanol or specific diets, can be controlled to investigate the biology of these diseases (Box 1) — this would allow for the disentanglement of the differences in the pathophysiology of MASH-to-HCC and ASH-to-HCC transitions.

Box 1 | Preclinical models of steatotic liver disease-associated hepatocellular carcinoma.

Preclinical mouse models of metabolic dysfunction-associated steatohepatitis (MASH), MASH-associated hepatocellular carcinoma (HCC), and alcoholic steatohepatitis (ASH) and ASH-associated HCC have been extensively reviewed^284,285. The predominant challenge in adopting these mouse models, like all mouse models, involves the metabolic and genetic differences between mice and humans³⁴. In principle, preclinical mouse models present features of steatotic liver diseases (SLDs) that vary depending on the primary insult and can be categorized based on the underlying cause, for example, dietary, genetic or toxin-induced causes. However, in the context of MASH and MASH-associated HCC, high-caloric diet-induced models do not produce a uniform disease aetiology. The quantity and quality of the dietary components in the high-calorie diet are crucial in dictating the histopathological features of the disease. It is noteworthy that these long-term dietary models display features of metabolic syndrome, such as obesity and liver cancer development, which are characteristic of MASH-associated HCC in humans. Additionally, toxins and carcinogens, such as diethylnitrosamine and carbon tetrachloride, as well as genetic modifications, such as the Foz/Foz, major urinary protein (MUP)–urokinase-type plasminogen activator (uPA) or Pten-null mouse models, are frequently used to induce and accelerate liver damage to more quickly recapitulate aspects of MASH and MASH-associated HCC.

Foz/Foz mice carry a loss-of-function mutation in Alms1 and develop severe obesity, insulin resistance and dyslipidaemia. When challenged with a high-fat or Western diet, they rapidly progress to steatohepatitis with fibrosis and eventually HCC, making this a metabolically driven model of MASH-associated hepatocarcinogenesis²⁸⁶. MUP-uPA mice exhibit transient hepatocyte-specific expression of uPA, which induces early endoplasmic reticulum stress and liver injury. Subsequent high-fat feeding leads to chronic inflammation, fibrosis and a high incidence of HCC, modelling inflammation-driven HCC arising from MASH²⁸⁷. Finally, hepatocyte-specific deletion of Pten results in constitutive PI3K–AKT activation, causing hepatic steatosis, chronic inflammation and spontaneous development of HCC, in the absence of dietary interventions²⁸⁸. Despite their utility, the clinical relevance of these models remains debatable, as disease initiation is driven primarily by strong genetic perturbations rather than by lifestyle-associated factors such as excess calories and physical inactivity that underlie most human MASH and MASH-associated HCC^53,284.

In the context of ASH, in addition to the metabolic and genetic differences between mice and humans, mice also exhibit psychological behavioural differences²⁸⁹. The classical Lieber–DeCarli mouse model of ethanol feeding, whereby mice obtain a high-caloric liquid composed of ethanol and other dietary components, is frequently used in preclinical studies. However, this model does not fully recapitulate the pathological and clinical features of ASH such as progressive fibrosis. In addition, its use in long-term studies is limited by the reliance on a purely liquid diet, which can alter normal feeding behaviour, affect gut physiology, and lead to reduced compliance and nutritional imbalances over prolonged administration. Similarly to the Lieber–DeCarli diet, chronic administration of alcohol in the drinking water of mice fails to induce typical histopathological features of steatohepatitis if not accompanied by high-caloric feeding²⁹⁰. In the case of the Gao-binge model, where the first phase involves chronic ethanol feeding through a liquid diet followed by a second phase in which an acute binge ethanol gavage is administered, liver injury is initiated in a time-efficient manner. However, it recapitulates only earlier stages of the human disease and does not induce cirrhosis or cancer²⁹¹. Furthermore, as alcohol displays weak carcinogenic activity in mice²⁹², chronic ethanol feeding must be combined with toxin-induced liver damage to induce ASH-associated HCC⁵⁵. A new mouse model has recently been considered where obesity and alcohol administration are combined to induce metabolic dysfunction and alcohol-associated liver disease. The mice are fed a MASH-inducing diet, and ethanol is given either chronically in the drinking water, as binge drinking doses or as a combination of both⁷.

In patients with MASH and mouse models of MASH, the impact of diet on chronic hepatic inflammation and HCC development has been extensively studied over the past few decades³³; however, the mechanisms determining the effects of physical activity or lack thereof on the pathogenesis of MASH-associated HCC remain less well described³⁴. In this context, Piguet et al.³⁵ demonstrated that, in hepatocyte-specific PTEN-deficient mice, a genetically engineered mouse model for MASH and MASH-associated HCC, a regular exercise regime with treadmills leads to a reduction in HCC incidence through phosphorylation and activation of AMP-activated protein kinase (AMPK) and subsequent inhibition of the mTORC1 complex.

In addition to dietary habits and a sedentary lifestyle, environmental risk factors and exposure to toxic compounds (for example, smoking or air pollution) can also influence the initiation of MASH and its progression to HCC^36–40. Recently, Chen et al.⁴¹ demonstrated that nicotine exposure in drinking water aggravates Western diet-induced liver damage by activating intestinal AMP–AMPKα signalling, which leads to the phosphorylation of sphingomyelin phosphodiesterase 3 (SMPD3) and enhances intestinal ceramide formation. Intestinal ceramides contribute to MASH and MASH-associated HCC by acting as gut-derived lipotoxic signals that directly affect the liver through the portal circulation. However, whether inhaling nicotine imparts a different or enhanced effect on HCC development remains unknown.

Moreover, circadian rhythm perturbations have been shown to drive MASH and MASH-associated hepatocarcinogenesis⁴². In a mouse model of chronic jet lag, characterized by circadian disruption, DNA damage and pathological lipid accumulation occurred in hepatocytes. The study further showed that chronic disruption of the circadian clock in mice leads to development of HCC through the dysregulated activity of nuclear receptors involved in bile acid metabolism, especially the constitutive androstane receptor and the farnesoid X receptor (FXR). Circadian rhythm disruption causes bile acid dysregulation and cholestasis in the liver, which leads to indirect activation of the constitutive androstane receptor. Stress signalling pathways triggered by bile acid accumulation, together with increased sympathetic nervous system activity, promote constitutive androstane receptor dephosphorylation and nuclear translocation while suppressing the opposing activity of FXR. Once activated, constitutive androstane receptor induced a transcriptional programme that induced cell-cycle and survival genes. In MASH, this results in forced proliferation and reduced apoptosis of hepatocytes that already harbour metabolic and oxidative DNA damage. By promoting the expansion of damaged cells and worsening bile acid and lipid imbalance, chronic constitutive androstane receptor activation drives MASH-to-HCC transition⁴³.

Chronic alcohol abuse leads to chronic inflammation, fibrosis and development of HCC through enhanced production of acetaldehyde, a product of ethanol metabolism¹⁴. Acetaldehyde, derived mainly by the activity of alcohol dehydrogenase (ADH) and cytochrome p450 2E1 (CYP2E1), can promote oxidative stress, endoplasmic reticulum stress, DNA damage and lipotoxicity^14,44. In both patients and mouse models of ASH-associated HCC, the role of other lifestyle-related aetiologies, such as smoking or high-fat and high-sugar diets, warrants further investigation.

Genetic and metabolic drivers of MASH-associated and ASH-associated HCC

MASH-associated HCC.

There are several shared genetic factors for MASH and ASH progression to HCC⁴⁵. The combination of multiple genetic factors enables the generation of a polygenic risk score (PRS) for HCC development in individuals presenting with risk factors for SLD^46,47. SNPs such as PNPLA3 p.Ile148Met, TM6SF2 p.Glu167Lys, SERPINA1 p.Gly366Lys, HSD17B13 rs6834314 and MTARC1 p.Ala165Thr, and other genetic variants increase the risk of SLD initiation and progression^19,48–50, while the presence of other SNPs, such as HSD17B13 rs6834314, protect against liver fibrosis in individuals presenting with established risk factors for MASH development such as insulin resistance⁵¹. Moreover, a PRS for ASH-associated cirrhosis (PRS_ALC) was generated using a curated set of genetic variants, comprising 20 SNPs located within 12 genes, 3 long non-coding RNAs and 1 intergenic region. SNPs were commonly found in PNPLA3, TM6SF2 and HSD17B13 in both patients with MASH and patients with ASH, and therefore represent major contributors to the PRS_ALC (ref. 52).

While there are overlaps in the genetic alterations seen in MASH-associated and ASH-associated HCC, such as TP53 mutations and TERT promoter mutations, each subtype exhibits unique mutational signatures⁵³. MASH-associated HCC is more strongly associated with CTNNB1 mutations and the activation of the Wnt–β-catenin pathway, reflecting metabolic dysregulation and lipotoxicity intrinsic to the condition⁵³. Similarly, ASH-associated HCC displays high frequencies of mutations in ARID1A and CCND1, which reflect the genotoxic effects of alcohol and the resulting DNA damage and chromatin dysregulation as shown in tissue specimens from patients with ASH-associated HCC^54,55.

Furthermore, MASH-related cirrhosis without HCC often has immunosuppressive molecular features that are likely to predispose the liver to HCC development^53,56. In line, the genetic landscape of MASH-associated HCC is heavily influenced by insulin resistance, lipotoxicity and chronic inflammation, leading to mutations in genes involved in cell proliferation, metabolism and genomic stability^53,57. Interestingly, hepatocarcinogenesis can also occur before the development of cirrhosis in MASH¹⁷. In this context, the mutational signature and number of mutations per tumour are higher in patients with MASH-associated HCC without cirrhosis than in patients with MASH-associated HCC and cirrhotic liver tissue⁵³. The underlying molecular mechanisms that determine the number of mutations in MASH-related hepatocarcinogenesis and the influence of the degree of fibrosis remain elusive.

Mutations in CTNNB1 have been identified as one of the hallmark genetic alterations in MASH-associated HCC, which result in the activation of the Wnt–β-catenin signalling pathway and an immunosuppressive microenvironment⁵³. The accumulation of fat in the liver, coupled with lipotoxicity, induces oxidative stress and inflammatory cytokine release, which further activate these pathways, promoting cellular proliferation and tumorigenesis^58,59. TP53 mutations, prevalent in about 33% of MASH-associated HCCs, are associated with multinodular tumours⁵³. However, how mechanistically TP53 mutations or loss of p53 contribute to hepatocarcinogenesis in MASH remains elusive. Moreover, in MASH-associated HCC, mutations in PIK3CA, which lead to aberrant activation of the PI3K–AKT signalling pathway⁶⁰, are highly prevalent and often associated with insulin resistance and the hyperglycaemic environment characteristic of the metabolic syndrome⁶¹.

Moreover, epigenetic modifications primarily through changes in DNA methylation, histone modifications and non-coding RNA activity, have a pivotal role in liver fibrogenesis and cancer development in MASH⁶². ARID1A, a chromatin remodelling gene, was shown to be mutated in about 6% of MASH-associated HCCs⁵³, contributing to altered chromatin dynamics and disrupting tumour suppressor pathways such as the p53 pathway⁶³. Relatedly, TERT promoter mutations are common in MASH-associated HCC, increasing telomerase activity and enabling continued cellular proliferation and resistance to apoptosis⁵³.

Metabolic risk factors, including arterial hypertension, elevated plasma triglyceride and cholesterol levels, as well as insulin resistance and type 2 diabetes mellitus, are drivers of disease progression in MASH⁶⁴. Chronic hyperglycaemia and insulin resistance promote hepatic de novo lipogenesis, oxidative stress and inflammation, accelerating hepatocellular injury and fibrogenesis⁶⁵. Elevated insulin and insulin growth factor 1 (IGF1) signalling activate oncogenic pathways such as PI3K–AKT and MAPK signalling, enhancing cell proliferation and survival^66,67. Additionally, advanced glycation end products (AGEs), generated by non-enzymatic glycation of proteins and lipids, as well as altered adipokine secretion, exacerbate hepatic inflammation and genomic instability, creating a pro-tumorigenic microenvironment^68,69.

Furthermore, MASH-induced dysregulation of bile acid metabolism⁷⁰ has been shown to promote the accumulation of bile acids in the bloodstream, termed cholaemia, which increases the risk for HCC. While hepatocarcinogenesis in MASH is in most cases linked to obesity, cholaemia can drive HCC development even in the context of patients with a low body mass index⁷¹. Spontaneous cholaemia was recently shown to induce liver tumours in 100% of mice treated with MASH-inducing diets⁷². Furthermore, accumulated bile acids induce hepatotoxicity through oxidative stress and mitochondrial dysfunction through the disruption of mitochondrial membranes, and activate NADPH oxidases as well as inflammatory pathways, including nuclear factor-κB (NF-κB) and the NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome complex. These processes exacerbate liver injury, promoting steatosis, inflammation and fibrosis. Persistent cholestasis further drives progression to MASH-associated HCC by inducing DNA damage⁷³ and epigenetic alterations and activating oncogenic pathways such as YAP–TAZ signalling⁷⁴. Moreover, disruption of bile acid homeostasis⁷² leads to impairment of the function of bile acid transporters, such as the bile salt export pump (BSEP), and activation of bile acid-sensing nuclear receptors such as FXR^73,74, which suppresses hepatic inflammation and fibrosis. Consequently, the loss of FXR function in MASH has been shown to trigger hepatocarcinogenesis in mice⁷⁵. Therefore, targeting bile acid dysregulation by using FXR agonists could mitigate the impact of cholaemia and cholestasis on MASH and its progression to HCC. While the activation of other nuclear receptors, such as peroxisome proliferator-activated receptor-α (PPARα) or PPARγ, to promote fatty acid β-oxidation and therefore reduce hepatic steatosis has been proposed, chronic or excessive activation of these receptors has been associated with hepatomegaly, peroxisome proliferation and increased risk of HCC in preclinical mouse models of MASH⁷⁶.

Additionally, both metabolic stress and lipotoxicity can induce DNA hypomethylation and the hypermethylation of promoter regions of tumour suppressor genes, such as RARRES3, in hepatocytes^77–79. Similarly, dysregulated expression of non-coding RNAs, such as microRNA-122 (miR-122), can also modulate lipid metabolism, inflammation and cell proliferation through the adaptation of hepatic gene expression by post-transcriptional repression of mRNAs involved in lipid synthesis, inflammatory signalling and cell-cycle control⁸⁰. In this context, blocking miR-122 with an antagomir, a single-stranded antisense oligonucleotide designed to specifically inhibit a mature microRNA, was found to reduce steatosis and weight gain in mouse models of MASLD⁸⁰.

ASH-associated HCC.

In addition to the generation of acetaldehyde and reactive oxygen species (ROS) leading to chronic necro-inflammation⁴⁴, alcohol consumption can promote inflammasome activation, dysregulation of cell death and metabolic reprogramming of bile acids and fatty acids in hepatocytes as well as endoplasmic reticulum stress and the associated activation of the unfolded protein response (UPR)⁸¹. This, together, leads to steatosis, cirrhosis and liver cancer development^82,83. Similarly to MASH-associated HCC, TP53 mutations are the most frequently observed genetic alteration in ASH-associated HCC⁵⁴. Additionally, CCND1 (encoding cyclin D1) amplification and activating mutations in AKT1 are commonly found in ASH-associated HCC⁸⁴, where CCND1 drives uncontrolled cell proliferation, and AKT1 promotes cell survival, migration and resistance to apoptosis⁸⁵.

Additionally, the presence of genetic polymorphisms of enzymes involved in alcohol catabolism is also frequently a determinant of disease progression and hepatic carcinogenesis⁵⁵. For instance, polymorphisms in alcohol-metabolizing enzymes, such as alcohol dehydrogenase 1C (ADH1C)*1 and aldehyde dehydrogenase 2 (ALDH2)*2, can alter acetaldehyde metabolism. The ADH1C*1 variant encodes a higher-activity ADH1C that converts ethanol to acetaldehyde more rapidly, whereas the ALDH2*2 variant produces a catalytically inactive or severely impaired ALDH2, markedly reducing acetaldehyde clearance⁸⁶. The combination of increased acetaldehyde production and impaired detoxification leads to acetaldehyde accumulation, which is associated with an increased risk of developing ASH-associated HCC^87,88.

Moreover, the presence of some epigenetic modifications has been shown to drive ASH progression to HCC⁸⁹. For example, acetaldehyde inhibits methionine adenosyltransferase activity and disrupts folate-dependent methyl group transfer, resulting in reduced hepatic levels of S-adenosylmethionine, the principal methyl donor required for DNA and histone methylation⁹⁰. S-adenosylmethionine depletion promotes global DNA hypomethylation and genomic instability, while paradoxically inducing locus-specific hypermethylation of tumour suppressor genes such as CDKN2A (encoding p16), leading to their transcriptional silencing and loss of cell-cycle control. This selectivity reflects inflammation-driven and stress-driven recruitment of DNA methyltransferases to susceptible promoters and impaired demethylation, resulting in preferential inactivation of key regulatory pathways that facilitates hepatocyte proliferation, malignant transformation and hepatocarcinogenesis^91,92. Beyond DNA methylation, chronic alcohol exposure dysregulates non-coding RNAs, particularly oncogenic microRNAs such as miR-21 and miR-155, which enhance inflammatory signalling and inhibit apoptosis in hepatocytes in ASH⁹³ and ASH-associated HCC^94,95. However, blocking or delivering specific microRNAs as a novel therapeutic approach for both ASH and ASH-associated HCC require further investigation.

Inflammation in MASH and ASH

Pro-inflammatory cytokines, triggering or maintaining chronic inflammation, particularly IL-6 and tumour necrosis factor (TNF), have emerged as critical mediators in the pathogenesis of MASH, ASH and their transition to HCC^96–98. IL-6 was shown to stimulate signal transducer and activator of transcription 3 (STAT3) signalling, and TNF activates canonical and non-canonical NF-κB signalling^99,100 to promote inflammation, cell proliferation and survival, and oncogenesis¹⁰¹. Targeted inhibition of STAT3 in hepatocytes suppressed tumour formation in mice with chemically induced HCC, highlighting the role of STAT3 in liver tumour initiation and growth^102,103. However, chemically induced HCCs in mice do not represent models for patients with MASH-associated or ASH-associated HCC as there is no development of steatohepatitis. It therefore remains to be elucidated whether inhibition of STAT3 also leads to tumour suppression in MASH and ASH. Studies using various mouse models of liver cancer have demonstrated that the pro-tumorigenic or anti-tumorigenic activity of canonical as well as non-canonical NF-κB signalling depends on the cellular context and nature of the initiating insult¹⁰⁴. For instance, transgenic expression of lymphotoxin-α and lymphotoxin-β in hepatocytes — both potent activators of NF-κB — has been shown to be sufficient to trigger HCC development in the context of MASH^105,106. Moreover, NF-κB essential modulator (NEMO) is a subunit of the IκB kinase complex that activates canonical NF-κB signalling in hepatocytes, and its activation can prevent the development of steatohepatitis and HCC in mouse models of MASH, identifying NEMO as a tumour suppressor in the liver¹⁰⁷. However, TAK1 constrains the pro-tumorigenic function of NEMO in parenchymal liver cells independent of canonical NF-κB signalling. Genetic studies in hepatocytes show that loss of TAK1 enables NEMO-dependent hepatocyte death, compensatory proliferation and hepatocarcinogenesis, even when NF-κB transcriptional activity is impaired¹⁰⁸. Thus, these findings support the idea of context-dependent pro-tumorigenic or anti-tumorigenic functions of NEMO signalling in the liver.

MASLD-associated and ALD-associated HCC

Importantly, while the risk for the development of liver cancer rises with the extent of fibrosis and cirrhosis^10,17, it can also evolve in patients with MASLD or ALD, thus bypassing the progression to MASH and ASH and the associated overt chronic inflammation, fibrosis and cirrhosis^109–111. However, the precise mechanisms enabling this remain unknown¹⁷. In tissues from patients with MASLD and in mice fed a high-fat diet without evidence of fibrosis, Fan et al.¹⁰⁹ found that the accumulation of AGEs in ECM enhanced liver viscoelasticity. This ECM stiffening activates mechanotransduction pathways, specifically integrin-β1–tensin 1–YAP signalling, leading to enhanced YAP nuclear localization and transcriptional activity. Through this mechanically driven signalling cascade, AGE-induced ECM remodelling promotes pro-oncogenic responses, thereby increasing the risk for HCC development even in the absence of fibrosis¹⁰⁹. Nonetheless, in vivo models that may closely resemble the biology of cancers in the non-cirrhotic liver in SLD are yet to be established. In the future, identifying patients who are at high risk of developing liver cancer in the absence of advanced fibrosis or cirrhosis will be critical for early detection and prevention strategies. Therefore, it is imperative that mouse models are developed that can enable investigations of MASLD-associated and ALD-associated, non-cirrhotic HCC^65,110. Similarly, research for improving our understanding of the effects of reversing tissue fibrosis and decreasing tissue viscoelasticity is of great importance^112,113.

MASH and MASH-associated HCC

Innate immune cells.

In patients and multiple diet-induced mouse models of MASH, lipid-associated macrophages expressing the lipid receptor triggering receptor expressed on myeloid cells 2 (TREM2) were shown to augment MASH development^114–117 (Fig. 2a). Given the natural function of TREM2, these macrophages were initially described as lipid-scavenging cells. Nevertheless, this myeloid population likely originates from circulating monocytes and is suggested to replace resident lipid-engulfed dysfunctional Kupffer cells in advanced stages of MASH¹¹⁸. Therefore, the replacement of dysfunctional Kupffer cells by TREM2⁺ lipid-associated macrophages represents a critical transition from a compensatory response to a maladaptive macrophage phenotype. Conversely, TREM2⁺ macrophages were shown to mediate resolution of inflammation as well as to promote regression of fibrosis in mouse models of MASH¹¹⁹ and in patients with MASH undergoing bariatric surgery¹²⁰. Moreover, the number of monocyte-derived macrophages expressing pro-tumorigenic molecules (upregulating markers such as CD163, TREM2 and SPP1) and MARCO⁺ Kupffer cells (downregulating genes related to antigen presentation), which are often viewed as ‘restorative’, increases with MASH progression to dampen inflammation and resolve liver damage. While initially these phenotypes limit inflammation, their sustained activation promotes immunosuppression, impaired antigen presentation and secretion of profibrotic mediators (including TGFβ, platelet-derived growth factor (PDGF) and osteopontin), thereby reinforcing hepatic stellate cell (HSC) activation, ECM deposition and fibrosis. Thus, in the context of ongoing metabolic stress, reparative macrophage programmes become maladaptive and actively contribute to disease progression rather than resolution^121–124. However, further studies are needed to characterize the variegated nature of these cells and to validate their modulation as a potential intervention for the treatment of MASH and MASH-to-HCC transition.

Fig. 2 |. The immune landscape of advanced steatotic liver diseases contributes to the development and progression of HCC.

The complex interplay between immune and parenchymal cells in metabolic dysfunction-associated steatohepatitis (MASH) and alcohol-associated steatohepatitis (ASH) is depicted, highlighting both shared and distinct immunological features of each aetiology. a, In MASH, steatosis leads to excessive lipid accumulation in hepatocytes, driving metabolic dysfunction and triggering innate and adaptive immune cell activation. This results in a hyper-aggressive immune response and a pro-tumorigenic inflammatory state. The innate immune compartment is characterized by Kupffer cells and infiltrating monocyte-derived macrophages adopting pro-inflammatory and profibrogenic phenotypes that drive cytokine and chemokine production. In parallel, conventional dendritic cells (cDCs) act as key immune sentinels that sense metabolic stress and cellular damage. Upon activation, cDCs capture lipid-modified and oxidation-modified antigens from injured hepatocytes and present them to T lymphocytes in the liver and draining lymph nodes. This antigen-presenting function links metabolic injury to adaptive immune activation. Adaptive immune cells, like CD4⁺ and CD8⁺ T cells, are chronically activated and demonstrate features of auto-aggression and exhaustion, contributing to hepatocyte damage, inflammation and preneoplasia development. B cells and plasma cells expand in response to lipid-derived and oxidation-derived antigens, producing antibodies that can sustain inflammation. b, In ASH, chronic alcohol consumption similarly disrupts the intestinal barrier, leading to microbial translocation, liver injury and fibrosis. The innate immune system is activated, with macrophages having central roles by sensing danger signals and producing pro-inflammatory cytokines such as IL-6, while neutrophils promote oxidative stress, protease release and the formation of neutrophil extracellular traps. Ongoing oxidative stress leads to the accumulation of oxidation-specific epitopes (OSEs), which act as danger signals and activate B cells and their differentiation into plasma cells, resulting in sustained antibody responses against oxidized self-antigens that can amplify hepatic inflammation. The adaptive immune compartment is also dysregulated in response to chronic alcohol exposure, with CD8⁺ and CD4⁺ T cells showing features of chronic activation and functional impairment, leading to an expansion of regulatory T (T_reg) cells and myeloid-derived suppressor cells (MDSCs). Additionally, alcohol exposure promotes altered cytotoxicity and cytokine production in natural killer (NK) and NKT cells, activating hepatic stellate cells (HSCs) and ultimately promoting fibrosis. Together, these immune alterations sustain chronic inflammation, promote hepatocellular damage and drive progression towards fibrosis, cirrhosis and, ultimately, hepatocellular carcinoma (HCC). c, In established HCC, the tumour tissue contains activated macrophages and dendritic cells (DCs) that produce inflammatory cytokines such as C-X-C motif chemokine 10 (CXCL10). In addition, CXCL13⁺CD4⁺ T cells are present and promote local immune cell recruitment and tertiary lymphoid structure-like niches. Together, these populations sustain intratumoural inflammation, support angiogenesis, and enhance tumour cell proliferation and tissue remodelling. At the same time, the HCC microenvironment is infiltrated by immunosuppressive cell populations. Tumour-associated macrophages (TAMs) often exhibit regulatory phenotypes and secrete IL-10 and TGFβ. T_reg cells are enriched and suppress effector T cell responses. CD8⁺ T cells frequently display an exhausted phenotype with impaired cytotoxicity and high expression of inhibitory receptors. Collectively, these mechanisms weaken immune surveillance and enable tumour persistence and progression. CAF, cancer-associated fibroblast; CXCR3, C-X-C chemokine receptor 3; GZMK, granzyme K; ICOS, inducible T cell co-stimulator; MHC, major histocompatibility complex; ROS, reactive oxygen species; SPP1, secreted phosphoprotein 1; T_H17 cell, T helper 17 cell; TNF, tumour necrosis factor; TREM2, triggering receptor expressed on myeloid cells 2; XCR1, XC receptor 1.

In established HCC tumours, TREM2⁺ macrophages, alone or in clusters with PLVAP⁺ endothelial cells and POSTN⁺ cancer-associated fibroblasts (conforming an onco-fetal niche that is associated with relapse after resection^125,126), hinder patient response to immunotherapy and are associated with worse prognosis^127,128 (Fig. 2c). Consistently, in samples from patients with MASH-associated HCC, immunosuppressive PDL1⁺ inducible T cell co-stimulator (ICOS)⁺ myeloid-derived suppressor cells and tumour-associated macrophages are linked to tumour progression, highlighting their role in orchestrating immune evasion²⁵. In the progression to MASH-associated HCC, this immunosuppressive character is likely to dominate, contributing to the recruitment of regulatory T (T_reg) cells and hindering the effector activity of CD8⁺ T cells^128,129. Furthermore, pro-inflammatory C-X-C motif chemokine 10 (CXCL10)⁺ macrophages promote C-X-C chemokine receptor type 3 (CXCR3)⁺ effector memory CD8⁺ T cell infiltration and are enriched in tumours from patients with advanced HCC that are responding to immunotherapy¹²⁷.

Dendritic cells are also important drivers of liver disease, and their numbers increase among infiltrating immune cells in MASH-associated livers¹³⁰. However, the role of these cells in the progression of MASH and MASH-associated HCC remains poorly understood. In mice fed a methionine-choline-deficient diet, selective depletion of conventional dendritic cells (cDCs) halted MASH progression, indicating that cDCs promote rather than protect against disease development. Heier et al.¹³¹ showed that the depletion of CD103⁺ cDCs (cDC1s) similarly resulted in a pro-inflammatory response in the same mouse model of MASH. By contrast, Deczkowska et al.¹³⁰ found that chemokine XC receptor 1 (XCR1)⁺ cDC1s drive liver pathology in mouse models of MASH. Genetic targeting of XCR1⁺ cells and pharmacological inhibition of its ligand, XCL1, improved liver damage in various mouse models of MASH. Further analysis of the relationship between cDC1s and CD8⁺ T cells revealed that, together, these cells promoted progression of MASLD to MASH in both mouse models and patients¹³⁰. Importantly, cDC1–CD8⁺ T cell interactions are necessary for antigen presentation, T cell priming, and natural killer cell activation and therefore to mounting an antitumour response in MASH-associated HCC¹³². Indeed, lipid accumulation in dendritic cells in the context of MASH was shown to impair stimulation of allogeneic T cells and antigen-presenting cell functions¹³³.

Similarly, neutrophils can exert either protumoural or antitumoural effects, depending on their maturation state¹³⁴. Mature neutrophils are found in HCC, especially in the immunosuppressive state¹³⁵, and are associated with worse prognosis, potentially due to their high expression of PDL1, inhibiting antitumour T cell effector function¹³⁵. Neutrophils can also promote MASH development and MASH-associated HCC through the formation of neutrophil extracellular traps (NETs)¹³⁶. In MASH livers, NETosis, a specialized form of cell death of neutrophils characterized by the release of NETs, promotes the establishment of a pro-tumorigenic environment¹³⁷. Xia et al.¹³⁸ demonstrated that NETs can activate quiescent HSCs via a Toll-like receptor 3 (TLR3)-dependent pathway, promoting fibrogenesis and increasing the risk for HCC development. Moreover, Wang et al.¹³⁶ reported that NETs induce a TLR4-dependent differentiation of naive CD4⁺ T cells into T_reg cells, potentially contributing to impaired immune surveillance and tumour progression. Conversely, in mouse models of MASH-associated HCC, the infiltration of immature and peripheral blood-like neutrophils enhances response rates to immune-checkpoint inhibitors (ICIs)¹³⁹. Interestingly, infiltration of immature neutrophils can be promoted therapeutically by targeting CXCR2 in mouse models of MASH-associated HCC^134,139.

Platelets have also been identified as an important initiator and maintaining factor of MASH and MASH-associated HCC¹⁴⁰. In the MASH liver, platelets are activated by metabolism-associated molecular patterns (for example, hyaluronic acid) and release α-granules filled with pro-inflammatory cytokines, such as TNF, IL-6 and TGFβ, exacerbating hepatic inflammation and promoting activation of HSCs^141,142, furthering liver damage, fibrosis and HCC development^140,143. Additionally, activated platelets can secrete PDGFs, which promote the proliferation of fibroblasts and myofibroblasts, further enhancing the fibrotic response¹⁴⁴. Because platelet aggregation in the injured liver leads to platelet activation and degranulation, increased platelet aggregation is associated with higher levels of fibrosis and inflammation — key drivers in SLD progression¹⁴⁰. Anti-platelet therapy has therefore emerged as an approach to lower the risk for liver cancer development in patients with MASLD and MASH¹⁴⁵; however, future studies are needed to ensure it can reduce platelet-derived inflammation without affecting platelet aggregation.

Adaptive immune cells.

Wolf et al.¹⁰⁶ demonstrated that, in mouse models of MASH, activated CD8⁺ T cells and natural killer T cells secrete cytokines, such as LIGHT, that induce chronic activation of the NF-κB pathway within hepatocytes, subsequently leading to steatosis and MASH-to-HCC transition. Antibody-mediated depletion of CD8⁺ T cells can therefore halt progression to HCC by decreasing the non-alcoholic fatty liver disease (NAFLD) activity score (NAS)^23,146. Disruption of mitochondrial function in intrahepatic CD4⁺ T lymphocytes with linoleic acid, a fatty acid that accumulates in these cells in MASH, causes excessive oxidative damage, triggering selective loss of this cell population¹⁴⁷. Blocking ROS generation has been shown to reverse the MASH-induced decrease of hepatic CD4⁺ T lymphocytes and delay the onset of MASH-associated HCC in mouse models of MASH¹⁴⁷. Additionally, T helper 17 (T_H17) lymphocytes, a subset of IL-17A-secreting, CD4⁺ T cells, drove infiltration of neutrophils into white adipose tissue, triggering insulin resistance and MASH in mice⁹⁶. MASH development and subsequent transition to HCC were diminished by pharmacological suppression of T_H17 cell differentiation, IL-17A-blocking antibodies and genetic ablation of the IL-17A receptor in myeloid cells⁹⁶.

In mouse models of MASH and MASH-associated HCC, CXCR6⁺ PD1^hi CD8⁺ T cells accumulate in the liver^23,139. These cells exhibit uncontrolled immune activation¹⁴⁶ through excessive release of granzyme B and TNF. They also induce ATP-dependent FasL-mediated hepatocyte death, thereby exacerbating liver injury^23,148. Moreover, IL-15 released by the altered hepatic metabolic microenvironment induces CD8⁺ T cell reprogramming by downregulating FOXO1, upregulating CXCR6, and increasing major histocompatibility complex (MHC) class I-independent cytotoxic activity of T cells through increased release of granzmye B and perforin. In mouse models of MASH-associated HCC, ICIs increase intrahepatic maintenance of CXCR6⁺ PD1^hi CD8⁺ T cells and thus promote inflammation and hepatic damage^23,148,149. Pfister et al.²³ proposed that this subpopulation is also found in patients with MASH. In concordance with this, steatohepatitis impairs T cell-directed immunotherapies against secondary liver tumours in mouse models of colorectal liver metastasis and MASH¹⁵⁰. Moreover, the presentation of MASH may limit a patient’s response to immunotherapy by increasing the number of exhausted and auto-aggressive CXCR6⁺ PD1^hi CD8⁺ T cells, which preferentially target and damage hepatocytes and inflamed liver tissue instead of mounting effective antitumour immune responses^149,151,152. However, it remains unknown whether the number of tumour-infiltrating CXCR6⁺ PD1^hi CD8⁺ T cells can serve as a predictive biomarker for immunotherapy response and patient stratification in MASH-associated HCC.

While less is known about B cells in MASH and MASH-associated HCC, in the jejunal mucosa, B cells were found to activate CD8⁺ T cells, which subsequently translocate to the liver, supporting liver inflammation in mouse models of MASH²⁹. B cells have also been shown to secrete antibodies, primarily IgA, that form immune complexes and engage Fc receptor-expressing (FcR⁺) myeloid cells, including macrophages and neutrophils. This interaction promotes sustained activation of these cells, leading to the release of pro-inflammatory cytokines, chemokines, ROS and profibrotic mediators such as TGFβ and PDGFR. These factors amplify hepatocellular injury, activate HSCs and drive ECM deposition, thereby exacerbating MASH-associated fibrosis and creating a tumour-promoting inflammatory microenvironment²⁹. However, further work is needed to evaluate the role of B cells in established MASH-associated HCC as well as whether and how they might be targeted in a therapeutic context.

Non-parenchymal HSCs.

Beyond immune cells, non-parenchymal liver cells like HSCs also have crucial roles in the progression of inflammation, fibrosis and HCC development in MASH. Under conditions of chronic liver inflammation, HSCs become activated, driving fibrogenesis^153,154. As reviewed recently, HSCs contribute to disease progression by promoting angiogenesis and modulating hepatic inflammation¹⁵⁵. Activated HSCs secrete pro-inflammatory cytokines and chemokines, such as CCL2 and IL-6, which recruit and sustain inflammatory immune cells, particularly monocytes and macrophages¹¹³. Through reciprocal crosstalk with immune cells and persistent activation by a stiffening ECM, HSCs function as amplifiers that couple chronic inflammation to fibrogenesis and disease progression.

ASH and ASH-associated HCC

The immune microenvironment in the context of the ASH-to-HCC transition is less well understood (Fig. 2b). Chronic alcohol abuse can impair the antitumour immune response in the liver and thereby promote HCC development and progression⁸⁹. The population of hepatic immune cells is qualitatively and quantitatively altered depending on the different stages of ASH⁸². It is likely that the amount of alcohol consumption and the chronicity of liver damage proportionally imprint a systemic immunosuppressive character to promote cancer⁹³.

The ASH liver is enriched in immunosuppressive cells, such as myeloid-derived suppressor cells and T_reg cells^51,62, and antigen presentation is downregulated in macrophages⁵³, reducing antigen-specific CD8⁺ T cell proliferation and activation⁶². Single-cell analysis comparing healthy livers to livers with viral-related and ASH-related cirrhosis revealed an increase of monocytes and macrophages, most likely resembling Kupffer cells, and a decrease in infiltrating CD3⁺ T cells in the liver of patients with ASH-related cirrhosis¹⁵⁶. A more recent analysis comparing livers from healthy individuals, patients with ASH and patients with MASH similarly identified an enrichment of a metabolically active APOE⁺ macrophage population with Kupffer cell-like characteristics in patients with ASH¹⁵⁷. This observation suggests that, despite differing upstream insults, steatotic liver diseases converge on shared macrophage adaptation programmes driven by metabolic stress. The presence of APOE⁺ Kupffer cell-like macrophages in ASH indicates a compensatory response aimed at lipid handling and tissue homeostasis. However, sustained activation of this metabolically specialized macrophage state may become maladaptive, promoting chronic inflammation and disease progression. Moreover, neutrophils are also likely to contribute to ASH pathogenesis; a recent retrospective analysis demonstrated that binge drinking induces NETosis in the liver, promoting tissue damage, inflammation and, thus, cancer¹⁵⁸.

The initiation of alcohol-driven steatosis is characterized by an overall increase in neutrophil and macrophage infiltration and activation in the liver⁹⁰. Concomitantly, the progression of MASH is accompanied by a selective enrichment of CD4⁺ T cells, CD8⁺ T cells and innate immune cells, including natural killer cells and dendritic cells^91,92. Alongside enhanced infiltration, chronic alcohol consumption also leads to a disruption of natural killer cell function, reducing natural killer cell-induced clearance of activated HSCs, particularly in the setting of cirrhosis⁹⁴.

Intriguingly, in the context of ASH, a population of cytotoxic granzyme K (GZMK)⁺ CD4⁺ T cells was identified as a possible disease-specific driver subset associated with hepatic fibrosis⁹⁵ and was related to impaired immunosurveillance and HCC progression^96,97. Further, the involvement of effector CD4⁺ T cells (T_H1, T_H2 and T_H17) in the context of ASH has been shown, where they orchestrate hepatic inflammation and tissue remodelling⁸². T_H1 cells promote macrophage activation and cytotoxic immune responses through the production of IFNγ and TNF, while T_H2 cells contribute to fibrogenesis by inducing alternative macrophage polarization and stimulating HSCs via IL-4 and IL-13. Together, these subsets sustain immune-mediated hepatocellular injury, promote fibrotic progression and shape a microenvironment conducive to disease advancement. Nevertheless, it remains to be elucidated whether these cell populations can be therapeutically targeted to halt disease progression and prevent HCC development.

Chronic alcohol use may further increase liver damage and tumorigenesis by activating IL-17 secretion in the liver and inducing its cognate receptor in hepatocytes. IL-17 activates HSCs, promoting fibrosis and HCC development¹⁰⁰, and regulates key inflammatory responses in macrophages and myeloid cells as well as cholesterol biosynthesis in hepatocytes through activation of the caspase 2–SP1–SREBP1–SREBP2–DHCR7 pathway⁹⁹. Blocking IL-17 signalling is being explored for the treatment of ASH and may have potential for reducing future risk of HCC.

Inter-organ crosstalk in MASH and ASH

The role of the adipose tissue–liver axis

Like the liver, adipose tissue is an organ that has a key role in regulating metabolic and energetic homeostasis. Therefore, communication between these compartments is bidirectional and occurs under physiological and pathological conditions¹⁵⁹. Conditions such as obesity, insulin resistance and systemic inflammation negatively affect the capacity of adipocytes to store free fatty acids (FFAs) through impairment of their endocrine functions and repression of PPARγ, preventing transcription of genes involved in triglyceride synthesis^160,161. Correspondingly, adipose tissue lipolysis leads to hepatic lipid accumulation and progression of inflammation in MASH and ASH. FFAs, like non-esterified fatty acids, released from adipose tissue trigger several biological responses within the liver, including lipotoxicity, oxidative stress, recruitment of Kupffer cells and HSC activation, ultimately creating an environment permissive to oncogenesis¹⁶². Suppression of lipolysis in adipose tissue through inhibition of vascular endothelial growth factor B (VEGFB) signalling was reported to halt MASLD progression in mice and correlates with reduced hepatic steatosis in patients with MASLD¹⁶³. In this context, VEGFB signalling may alter lipid availability, insulin sensitivity or local metabolic cues that secondarily suppress HSC activation. Furthermore, in mouse models of MASH, IL-17A secretion was shown to accelerate MASH-to-HCC transition by inducing lipolysis in white adipose tissue, consequently increasing systemic FFAs⁹⁶. Additionally, Boesch et al.¹⁶⁴ identified transcriptionally dysfunctional macrophages within white adipose tissue in patients with MASLD, which further promote adipose tissue inflammation and downstream liver inflammation.

Clinical observational studies indicate that the visceral adipose tissue index, defined as a cross-sectional area of adipose tissue by computed tomography normalized to height, is an independent risk factor for developing HCC in patients with liver cirrhosis and patients with MASH as well as for recurrence of HCC after liver transplantation^165,166. Increased adiposity and visceral fat are also high-risk factors for developing HCC in patients with MASH or ASH¹⁶⁷. Moreover, adipose tissue depots secrete and release a wide range of adipokines that are not only necessary for the regulation of systemic energetic homeostasis but also tightly orchestrate the hepatic immune microenvironment¹⁵⁹. Among these, SPARCL1 was identified as an adipose tissue-derived secreted glycoprotein in a mouse model of MASH, promoting hepatic inflammation by inducing expression of CCL2 in hepatocytes¹⁶⁸. Furthermore, circulating adiponectin correlated positively with HCC incidence in patients with MASLD¹⁶⁹. It was recently shown that adipocyte-specific loss of SWELL1 results in excessive lipolysis, exacerbation of MASLD and the development of HCC in mouse models of MASH¹⁷⁰. Conversely, the adipocyte-derived endocrine factor neuregulin 4 (NRG4) ameliorated metabolic reprogramming and restrained tumour burden and progression in MASH¹¹⁵. However, the molecular mechanisms underlying the protective effects of NRG4 remain to be elucidated. These data suggest that the accumulation of visceral adipose tissue and the metabolic alteration of its functionality contribute to the MASH-to-HCC progression.

It is not yet entirely understood how ethanol affects adipose tissue depots in ASH and how this may contribute to liver cancer development. In a mouse model of binge drinking that was fed a high-fat diet, consumption of alcohol in high concentrations directly induced adipocyte death and lipolysis, resulting in aggravation of liver injury and hepatic endoplasmic reticulum stress¹⁷¹. However, whether alcohol-induced adipocyte death and lipolysis also affect ASH-to-HCC transition remains unknown.

The gut–liver axis

Obesity and alcohol consumption have both been shown to change gut microbial composition and function^172,173. Gut microbial dysbiosis and the associated metabolic output, including short-chain fatty acids, such as acetate, butyrate, formate and propionate, secondary bile acids, and other bacteria-specific metabolites, are increasingly recognized as determinants of the immune microenvironment of MASH and MASH-associated HCC¹⁷⁴. Interestingly, transplantation of faeces from high-fat and high-cholesterol diet-fed mice into germ-free mice was sufficient alone to induce MASLD and eventually promote tumorigenesis¹⁷⁵. Concurrently, MASH progression is associated with the leakage of bacterial products (pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns) and metabolites from the gut into the portal circulation and, thus, the liver^176–179. Translocation of PAMPs, such as serum lipopolysaccharides, was shown to contribute to tumour progression by activating liver-intrinsic macrophages in a TLR-dependent manner¹⁸⁰.

Secondary bile acids, gut-derived metabolites whose composition is influenced by the microbiome, are known to have a causal role in promoting carcinogenesis in both MASH-associated and ASH-associated HCC^70,181,182. In mouse models of HCC without MASH, Ma et al.¹⁸³ found that the ratio of primary-to-secondary bile acids is highly dependent on the gut microbiome composition. Further, they could show that microbiome-mediated metabolism of bile acids can induce CXCL16 secretion by liver sinusoidal endothelial cells. Subsequently, CXCL16 dampens natural killer T cell-mediated antitumour response¹⁸³. Moreover, in a high-fat diet-fed mouse model of MASLD, deoxycholic acid, a secondary bile acid, was shown to drive liver cancer formation through induction of the senescence-associated secretory phenotype in HSCs and subsequent fibrosis¹⁸².

Yang et al.¹⁸⁴ reported that damage-associated molecular patterns or PAMPs, such as 2-oleoglycerol, a bacterial metabolite, act as a key mediator of Western diet-induced MASH through macrophage-dependent activation of HSCs. Mechanistically, 2-oxoglutarate activates hepatic macrophages through GPR119-dependent signalling, leading to TAK1 phosphorylation and activation of NF-κB. This cascade induces the expression of pro-inflammatory mediators and reinforces an activated macrophage phenotype. However, bacterial metabolites can also exert protective effects against MASH-associated HCC pathogenesis. For example, acetate produced by Bifidobacterium pseudolongum was shown to reach the liver via the portal circulation and suppress tumour formation by inhibiting the IL-6–JAK1–STAT3 signalling axis in two mouse models of MASH-associated HCC¹⁸⁵. Moreover, gut-to-liver bacterial translocation and intratumoural microbiota are increasingly recognized as having a role in disease progression¹⁸⁶. Recently, Wang et al.¹⁸⁷ demonstrated that translocation of Klebsiella pneumoniae from the gut to the liver promotes HCC proliferation through TLR4 activation by the bacterial surface protein PBP1B. However, the full spectrum of mechanisms linking microbial gut-to-liver translocation to liver tumorigenesis has yet to be defined.

Alcohol intake can also directly affect the viability of cells lining the intestine, namely enterocytes, the function of the intestinal barrier, and the biodiversity of the microbiome¹⁸⁸. The disruption of the intestinal barrier through alcohol-induced enterocyte damage promotes translocation of bacteria and bacterial components and establishes an inflammatory liver microenvironment^6,82,149,189. This results in the recruitment and activation of neutrophils, pro-inflammatory macrophages and Kupffer cells, which produce ROS, leading to an exacerbation of DNA damage and promoting lipogenesis and steatosis in hepatocytes through acetaldehyde in ASH^82,190–192. Moreover, gut-derived lipopolysaccharides were increased in patients with ASH-associated HCC as compared to patients who do not drink alcohol¹⁹³. Notably, compositional changes in the microbiome in the context of patients with obesity and MASH have also been shown to lead to the production and systemic release of endogenous ethanol, adding another layer of complexity within the gut–liver communication^194,195. Interestingly, ethanol can be endogenously produced as a normal, low-level byproduct of microbial fermentation of dietary carbohydrates by intestinal bacteria and yeasts, and under healthy conditions, it is efficiently metabolized by intestinal and hepatic alcohol dehydrogenases, resulting in minimal systemic exposure¹⁹⁶. However, in obesity and MASLD, overrepresentation of ethanol-producing microorganisms (including certain Enterobacteriaceae and yeasts), increased substrate availability, and impaired intestinal barrier function can markedly elevate luminal ethanol concentrations and promote its translocation into the portal circulation^197,198. This phenomenon has been demonstrated primarily in animal models and supported by human observational studies, in which obesity-associated or MASLD-associated microbiomes produced higher ethanol levels ex vivo and were associated with detectable circulating ethanol in the absence of alcohol consumption^194,195. These findings emphasize the role of bacterial metabolites in shaping hepatic immunity, with implications also extending to a role in cancer^199,200.

The brain–liver and heart–liver axes

In addition to well-established inter-organ communication between the liver and the gut or adipose tissue, recent studies have suggested the possibility of other crosstalk mechanisms that contribute to MASH progression. Interestingly, a brain–adipose–liver axis in the context of ALD may exist. Shen et al.²⁰¹ illustrated that alcohol could activate the hypothalamus and sympathetic nerves innervating brown adipose tissue and increase the expression and activity of UCP1 in mice; this induced thermogenesis and thus inhibited fatty acid accumulation in the liver. Moreover, alcohol increased the activity of sympathetic nerves through insulin resistance in the central nervous system, resulting in unrestrained adipose tissue lipolysis responsible for hepatic steatosis in rodent models of ethanol consumption²⁰². However, further studies are needed to determine the role of the brain–adipose–liver axis in ALD-to-HCC transition.

Xie et al.²⁰³ demonstrated that cardiovascular disease, particularly myocardial infarction, can promote the recruitment of Ly6C^hi monocytes into the liver and steatosis, fostering a pro-inflammatory microenvironment and promoting progression of MASLD to MASH. Simultaneously, elevated levels of cardiac periostin, a secreted ECM protein, following myocardial infarction drives hepatic fibrosis by activating HSCs through integrin β3, exacerbating steatosis by suppressing PPAR signalling²⁰³. These findings underscore the multifaceted systemic mechanisms that contribute to and underlie the development of SLDs as well as the necessity to study SLDs as systemic conditions.

Clinical management of HCC in patients with MASH and ASH

The clinical management of MASH-associated and ASH-associated HCC remains a major challenge as they are frequently diagnosed in advanced stages and are refractory to many standard anticancer treatments^17,55. Moreover, liver cell damage and liver dysfunction narrow the therapeutic window for treatment interventions and enhance the risk for treatment-associated complications as the regenerative capacity of the liver is limited and cannot recompensate for the loss of liver tissue (for example, through surgical resection)^204,205. Therefore, expanding our understanding of the mechanisms involved in MASH-associated and ASH-associated development and progression represent important unmet needs for the development of novel prophylactic and therapeutic interventions.

Preventing MASH-to-HCC and ASH-to-HCC transition

In addition to reducing or halting the progression of steatosis, inflammation and fibrosis, the prevention of hepatocarcinogenesis represents an important challenge in MASH and ASH. The complex interplay of metabolic dysregulation, inflammation and fibrosis in MASH progression has driven efforts to develop therapies aimed at regulating these pathways. These include PPAR agonists, FXR agonists, glucagon-like peptide 1 receptor agonists (GLP1RAs), sodium–glucose cotransporter 2 (SGLT2) inhibitors and thyroid hormone receptor-β (THRβ) agonists, which have shown promising results in reducing liver fat and markers of fibrosis in clinical trials of patients with MASH²⁰⁶. However, the mechanism of action and influence on disease pathophysiology are distinct. PPAR agonists, such as pioglitazone and lanifibranor, improve insulin resistance, lipid metabolism and hepatic inflammation while reducing fibrosis by limiting the activation of quiescent HSCs²⁰⁷. FXR agonists, including obeticholic acid, regulate bile acid metabolism, decrease lipotoxicity primarily by limiting hepatic lipid accumulation rather than by directly promoting lipid breakdown, and suppress fibrosis by inhibiting HSC activation⁷⁴. GLP1RAs, such as semaglutide and liraglutide, lead to weight loss and thus improve systemic metabolism, and reduce hepatic steatosis by enhancing insulin sensitivity and lowering systemic inflammation²⁰⁸. SGLT2 inhibitors, such as empagliflozin and dapagliflozin, are primarily used for type 2 diabetes but have also demonstrated benefits in reducing hepatic fat content and improving systemic metabolism in mouse models of and in patients with MASH²⁰⁹. Resmetirom selectively targets THRβ in the liver, therefore avoiding systemic hormone effects, to reduce steatosis and lipid accumulation²¹⁰. It is the first FDA-approved drug for the treatment of MASH.

Furthermore, for ASH-associated HCC, therapies aimed at reducing inflammation, oxidative stress and fibrosis may slow disease progression. PPAR agonists, FXR agonists and THRβ agonists may mitigate liver damage in ASH by addressing pathways that are also involved in MASH pathophysiology. However, the direct hepatotoxic effects of ethanol and its metabolites, including acetaldehyde, may limit the efficacy of treatments such as PPAR agonists, FXR agonists, THRβ agonists and GLP1RAs, as these drugs do not target ethanol-induced liver damage. Importantly, it remains to be elucidated if the improvement in steatosis, inflammation and fibrosis outcomes induced by the abovementioned therapeutic agents in patients with SLD will also translate into a lower risk for the development of liver cancer.

Lifestyle interventions are another approach for treating MASH-associated and ASH-associated HCC (Box 2). In both mouse models of MASH and clinical trials, caloric restriction could lower inflammation and fibrosis and prevent transition to HCC²¹¹. Identifying the molecular mechanisms of how lifestyle interventions, such as dietary restriction, intermittent fasting or physical activity, prevent progression of SLDs is of increasing importance. This may also allow the development of therapeutic mimetics that induce the beneficial effects of these interventions.

Box 2 | Lifestyle interventions to prevent and treat SLD-associated HCC.

Lifestyle interventions, such as dietary changes, increasing physical activity and attenuating alcohol intake, have pivotal roles in modifying metabolic reprogramming and improving clinical outcomes in steatotic liver diseases (SLDs) and SLD-associated hepatocellular carcinoma (HCC)²⁹³. Lifestyle interventions that induce weight loss, such as intermittent fasting, caloric restriction and increased physical activity, help correct metabolic imbalances that drive chronic liver inflammation and fibrosis²⁹⁴ by improving insulin sensitivity, reducing hepatic fat content, and attenuating lipotoxicity and oxidative stress²⁹⁵. Weight loss exceeding 7–10% of an individual’s body weight can reduce liver inflammation and fibrosis in earlier stages of the disease²⁹⁵. Caloric restriction and physical activity can induce the expression and release of adipokines such as adiponectin and leptin²⁹⁶. Adiponectin levels, which are typically reduced in patients with metabolic dysfunction-associated steatohepatitis (MASH), increase with lifestyle changes, exerting anti-inflammatory and anti-fibrotic effects via the AMP-activated protein kinase (AMPK) pathway²⁹⁷. Emerging data suggest that intermittent fasting enhances autophagy, which prevents tumour initiation by mitigating oxidative stress and inflammation in mouse models of and patients with MASH²⁹⁸. In this regard, intermittent fasting, for example, the 5:2 regimen, where patients fast on two non-consecutive days per week, can prevent MASH progression and MASH-to-HCC transition through the activation of peroxisome proliferator-activated receptors (PPARs) and PCK1, as shown in diet-based mouse models of MASH and MASH-associated HCC²¹¹. First clinical trials have demonstrated efficacy of intermittent fasting for the treatment of metabolic dysfunction-associated SLD (MASLD), halting worsening of steatosis, inflammation and fibrosis²⁹⁹.

Partaking in physical activity was shown to enhance mitochondrial function and reduce oxidative stress by upregulating genes involved in mitochondrial biogenesis, such as peroxisome proliferator-activated receptor-γ coactivator 1α (PGC1α) in mouse models of MASLD³⁰⁰. Exercise also improves systemic inflammation by decreasing pro-inflammatory cytokines like tumour necrosis factor (TNF) and IL-6 in mouse models of MASLD³⁰¹. However, the type and regimen of physical activity most suitable to block MASH-to-HCC transition and to treat MASH-associated HCC remains unknown. Further research is needed to uncover how exercise influences metabolic reprogramming, antitumour immune surveillance and immunotherapy efficacy in the context of MASH and MASH-associated HCC.

In the case of alcohol-associated steatohepatitis (ASH), cessation of alcohol intake is a cornerstone of therapy, directly preventing alcohol-induced hepatoxicity³⁰². Alcohol abstinence reduces cytochrome p450 2E1 (CYP2E1) activity, decreasing reactive oxygen species production and oxidative damage³⁰³. Over time, abstinence also dampens TGFβ signalling in the liver, slowing fibrosis progression and allowing the liver to regenerate. Further lifestyle interventions in patients with ASH-associated HCC focus on mitigating the metabolic dysfunction that often accompanies ASH. Incorporating a balanced diet and exercise can improve systemic metabolic health, further reducing HCC risk. In patients with ASH, it has been shown that, even in advanced disease stages, alcohol abstinence and lifestyle improvements can extend survival by limiting ongoing disease progression³⁰⁴.

Early diagnosis and biomarkers for HCC detection in MASH and ASH

Several cohort studies support the value of HCC surveillance as it is associated with better access to curative treatments, in turn improving overall survival^212–214. Better screening tools and surveillance methods are needed to identify patients with MASH-related or ASH-related cirrhosis as well as those with earlier-stage disease at high risk of developing HCC.

Currently, HCC surveillance is based on ultrasound; however, the accuracy is hampered by excessive body weight, moderate-to-severe fatty liver and fibrosis²¹⁴, limiting its effective use in patients with SLDs²¹⁵. While MRI exhibits high accuracy for detecting HCC, it is not feasible as a diagnostic tool due to its high cost and poor patient compliance. To solve this, abbreviated MRI, either contrast-enhanced or non-contrast, has shown high sensitivity for detecting HCC lesions below 2 cm in diameter in a meta-analysis of 917 patients with HCC^216,217. Its performance was comparable across aetiologies and fibrotic scores of chronic liver disease. However, concerns regarding the potential nephrotoxicity associated with repeated administration of contrast agents may limit its feasibility as a surveillance strategy.

Additionally, biomarker analysis and alternative screening strategies, such as liquid biopsies, are under consideration for HCC surveillance. Concomitant evaluation of α-fetoproetin (AFP) is recommended for surveillance purposes (not diagnosis) over ultrasonography alone by the American Association for the Study of Liver Disease and the Asian Pacific Association for the Study of the Liver but not by the European Association for the Study of the Liver^218–220. Additionally, the GALAD score, which incorporates gender, age, AFP-L3 (a glycoform of AFP with a high affinity for Lens culinaris agglutinin), AFP and Des-γ-carboxy prothrombin (an isoform of AFP with defective post-translational carboxylation produced in HCC), has shown excellent performance in predicting HCC prevalence regardless of stage^221–223. It is also effective in patients with non-cirrhotic MASH²²⁴, and exhibits high sensitivity and specificity as well as better area under the receiver operating characteristic curve (AUROC) than ultrasound^225,226. The GALAD score improves the sensitivity for HCC detection but with modest specificity in a phase III clinical trial²²⁵. Analysis of cell-free DNA also showed greater sensitivity and specificity than ultrasound in detecting early and very-early HCC in patients with SLDs²²⁷. Similarly, methylome analyses of cell-free DNA may also be used in the future as non-invasive tools for HCC surveillance, displaying increased AUROC compared to AFP evaluation or the GALAD score^228–231. Moreover, transcriptomic analysis of the RNA of circulating peripheral blood-derived extracellular vesicles represents an additional non-invasive approach for HCC surveillance^232,233. However, none of these methods has yet reached phase III trials. Finally, the emergence of artificial intelligence-based surveillance tools for liver cancer has shown promising results in the context of MASH and ASH²³⁴. Artificial intelligence-based approaches have been shown to outperform traditional regression models associated with HCC surveillance²³⁵. Imaging-based artificial intelligence enhances ultrasound or MRI performance by improving lesion detection and reducing operator dependence, which is particularly relevant in patients with obesity and MASH^234,235. In parallel, artificial intelligence-driven liquid biopsy models combine cell-free DNA, epigenetic markers and serum proteins to support early tumour detection when imaging is suboptimal.

The treatment of MASH-related and ASH-related HCC

Treatment decisions for HCC are guided by the Barcelona Clinic Liver Cancer staging system, which is the framework recommended by most international guidelines^218,219,236. Curative approaches (resection, liver transplantation and local ablation) are available for patients with early-stage HCC²³⁷, although patients with MASH-associated HCC are at risk for treatment-related complications due to the high likelihood of having comorbidities. However, among patients who are eligible for and successfully undergo curative treatment, meta-analyses have shown that those with MASH-associated HCC have equal or superior overall and recurrence-free survival compared with patients with HCC arising from viral hepatitis or ASH¹⁷. In this context, adjuvant and neoadjuvant immunotherapies are being explored to reduce recurrence rates as 70% of patients with HCC develop disease recurrence at 5 years²³⁸.

Patients with intermediate-stage HCC are candidates for locoregional therapies such as transarterial chemoembolization (TACE) and transarterial radioembolization (TARE)²³⁹. Clinical trials are now assessing the efficacy of systemic therapies, such as ICIs, in combination with locoregional therapies compared with locoregional therapies alone²⁴⁰. The results of a recent phase III trial demonstrated a better progression-free survival for patients treated with TACE in combination with ICIs²⁴¹. In patients with intermediate-stage HCC, TACE and TARE can allow downstaging of the tumours and subsequent curative strategies²¹⁸. Moreover, in patients with MASH-associated HCC, outcomes after TACE and TARE, including time-to-progression, overall survival and treatment-related adverse events, have been reported to be comparable to those observed in patients with HCC arising from chronic viral hepatitis^17,242,243.

Patients with advanced-stage HCC and with contraindications for locoregional therapies are eligible for systemic therapy. The standard-of-care first-line systemic therapy for patients with HCC is administered irrespective of the underlying aetiology and consists of the combination of an ICI, such as the anti-PDL1 antibody atezolizumab, with tyrosine kinase inhibitors such as the anti-VEGFA antibody bevacizumab²⁴⁴. Other approved ICIs target the immune-checkpoint molecules CTLA4 (tremelimumab) or PD1 (nivolumab, ipilimumab)^244–246 (Fig. 3a). In patients with contraindications for immunotherapies, either multi-kinase inhibitors such as sorafenib²⁴⁷, or VEGFR and fibroblast growth factor receptor (FGFR) inhibitors, such as lenvatinib, are approved for use²⁴⁸ (Fig. 4a,b). Additional combinations of other ICIs are under evaluation²³⁷, as is the use of bispecific antibodies that target multiple immune-checkpoint molecules simultaneously^249–253 (Fig. 3a,b). Preclinical studies and retrospective patient analyses suggest that immunotherapies may be less effective in non-viral HCC, particularly in ASH-associated or MASH-associated tumours, compared with viral hepatitis-related HCC^17,254,255. Although some clinical datasets show a trend towards higher immunotherapy response rates in viral aetiologies, these differences have not reached statistical significance and therefore remain inconclusive²⁵⁶. Notably, the impact of underlying MASH on immunotherapy efficacy has not been prospectively evaluated as no clinical trial to date has stratified patients by aetiology^30,149. Importantly, it remains unclear whether anti-metabolic drugs that block MASH pathogenesis and progression could be useful in treating established MASH-associated HCC (Fig. 4b). It is likely that these therapies could alter tumour cell phenotypes and the antitumour immune response through the activation of glycolysis, autophagy and fatty acid oxidation. For example, the antidiabetic medication metformin has been shown to rescue the efficacy of anti-PD1 therapy in mouse models of liver cancer by restoring CD8⁺ T cell motility and immunometabolism, improving antitumour responses^{23,146,257,258}. Furthermore, establishing the use of predictive biomarkers may improve response rates. While tumour mutational burden has been proposed as a biomarker for ICI efficacy in other tumour types^259–264, the low median tumour mutational burden in MASH-associated or ASH-associated HCC limits its use as a predictive biomarker²⁶⁵.

Fig. 3 |. Mechanisms of action of current and emerging systemic treatment interventions for hepatocellular carcinoma.

a, Blockade of inhibitory checkpoint molecules on T cells can restore their activation and cytotoxic activity. Antibodies against CTLA4 can enhance priming by antigen-presenting cells (APCs). PDL1 interacts with CD80 on APCs, further contributing to immune evasion. Blocking PD1 or PDL1 prevents these inhibitor interactions, restoring T cell effector function and promoting antitumour immunity. Additional inhibitory axes, including TIM3 and LAG3, can be targeted with antibodies such as cobolimab and relatlimab. Inhibition of TIM3 and LAG3 together restores dysfunctional T cell responses by relieving complementary inhibitory pathways, with LAG3 blockade enhancing antigen-driven T cell activation through disruption of the LAG3–major histocompatibility complex (MHC) class II interaction and TIM3 blockade reversing late-stage T cell exhaustion by preventing inhibitory TIM3–Galectin 9 (Gal9) signalling, thereby enabling stronger and more sustained immune effector function. b, Targeting the innate immune cells can overcome tumour-driven immunosuppression and restore effective antitumour activity. Regulatory T (T_reg) cells express glycoprotein-A repetitions predominant (GARP), which anchors latent TGFβ on their surface and is the source of active TGFβ upon activation. Antibodies such as livmoniplimab block the GARP–TGFβ1 complex, preventing activation of latent TGFβ. Without active TGFβ, TGFβ signalling in effector T cells is reduced. Normally, active TGFβ signalling suppresses T cell activation by inducing SMAD-mediated transcription and antagonizing nuclear factor of activated T cells (NFAT)-driven pathways, thereby reinforcing T_reg cell-mediated inhibition of CD8⁺ T cell cytotoxicity. Antibodies targeting C-X-C chemokine receptor type 2 (CXCR2) on neutrophils prevent their recruitment into the tumour microenvironment, reducing secretion of immunosuppressive factors that induce T cell exhaustion. Targeting triggering receptor expressed on myeloid cells 2 (TREM2) and CD163 on tumour-associated macrophages promotes a shift from M2 to M1 polarization and reduces secretion of immunosuppressive IL-10 and TGFβ, restoring CD8⁺ T cell activation and cytotoxic function. Bispecific antibodies targeting CD16 on natural killer (NK) cells and glypican 3 (GPC3) on malignant hepatocytes redirect NK cell cytotoxicity towards tumour cells, leading to enhanced antibody-dependent cellular cytotoxicity (ADCC) and tumour cell apoptosis. c, Engineered cell-based treatments directly mediate tumour cell killing. Chimeric antigen receptor (CAR) T cells can be engineered to target cell-surface antigens on hepatocellular carcinoma such as GPC3 and claudin 6 (CLDN6). T cell receptor (TCR)-transduced T cells recognize intracellular tumour antigens presented on MHC molecules, such as α-fetoprotein (AFP), leading to tumour cell death. This leads to perforin-mediated and granzyme-mediated lysis or enhanced antigen presentation. Expansion of tumour-infiltrating lymphocytes (TILs) can also enrich for tumour-reactive T cells. Bispecific CAR T constructs can also be used to recruit accessory immune cells to further amplify cytotoxic responses. NF-κB, nuclear factor κB; PIP₃, phosphatidylinositol (3,4,5)-trisphosphate; ROS, reactive oxygen species.

Fig. 4 |. Emerging tumour-directed therapeutic approaches for the treatment of hepatocellular carcinoma.

a, Vascular endothelial growth factor (VEGF) signalling is critical for inducing angiogenesis in endothelial cells and promoting cancer cell proliferation and survival through autocrine signalling. Therapeutic strategies target this axis to disrupt tumour vascularization and growth. Antibodies targeting VEGF ligand (VEGFA) and those blocking the VEGF receptor (VEGFR) directly prevent activation of downstream PI3K–AKT and PLCγ–PKC pathways and inhibit angiogenesis. Multi-kinase inhibitors, such as sorafenib, cabozantinib, lenvatinib and regorafenib, target oncogenic receptor tyrosine kinases to block signalling cascades critical for hepatocellular carcinoma progression. Of note, sorafenib and regorafenib additionally inhibit RAF kinase, directly suppressing downstream signalling. b, Microbiome-modulating interventions, such as probiotics and faecal microbial transplantation (FMT), can be used to reshape both systemic and tumour immunity, ultimately influencing treatment efficacy. Oncolytic viruses can induce direct tumour cell lysis while promoting antigen release to stimulate antitumour immunity. Antibody–drug conjugates (ADCs), such as glypican 3 (GPC3)–DM1 or claudin 6 (CLDN6)–DM1, can deliver cytotoxic payloads to tumour cells, triggering cell death. Proteolysis-targeting chimeras (PROTACs) can degrade oncogenic proteins by recruiting E3 ligases for targeted degradation. DC, dendritic cell; FGFR, fibroblast growth factor receptor; PDGFR, platelet-derived growth factor receptor; SCFAs, short-chain fatty acids; TME, tumour microenvironment; Ub, ubiquitination.

Alternative therapies, such as adoptive cell transfer^266,267, chimeric antigen receptor (CAR) T cells or bispecific antibodies²⁶⁸, are also being tested against specific antigens, such as glypican 3 (GPC3)^269–271 or AFP²⁵⁷, in HCC (Fig. 3c). The success of cell-based approaches has been hindered by the microenvironment of the liver as it is immunosuppressive and exhibits a metabolic shift that is detrimental for T cell functionality¹⁴⁹. Consequently, CAR T cells used in mouse models of HCC have been modified to preserve glycolytic metabolism to sustain effector activity within the microenvironment²⁷². However, the clinical efficacy of this approach demands further evaluation. Moreover, vascular normalization, induced by antiangiogenic therapies, such as monoclonal antibodies targeting VEGFA, may enhance the ability of endogenous T cells as well as CAR T cells to infiltrate the HCC microenvironment. Additionally, based on the specific immune infiltrate described above for ASH and MASH livers, therapeutic approaches that block protumoural macrophages (for example, TREM2⁺ or CD163⁺ macrophages), or that reduce mature neutrophil infiltration or skew neutrophils to peripheral blood and/or immature phenotypes¹³⁹ are also promising strategies to propel antitumour immune responses (Fig. 3b). Similarly, vaccine-based therapies in HCC have shown induction of immune responses against neoantigens^273,274. Whether this holds true for all HCC aetiologies, including MASH and ASH, remains to be investigated. In line with this, mRNA-loaded dendritic cell vaccines are providing promising results in phase I clinical trials of patients^{245–247,273}. Peng et al.²⁷⁵ demonstrated, in a phase I trial, that a personalized neoantigen-loaded dendritic cell vaccine in combination with a neoantigen-activated T cell therapy can be safely administered to patients with HCC across all aetiologies and induce neoantigen-specific T cell responses. A DNA-based vaccine encompassing personalized neoantigen sequences in combination with anti-PD1 therapy (pembrolizumab) reached higher objective response rates than pivotal trials with anti-PD1 therapies in advanced HCC²⁷⁶. Hence, vaccine-based approaches remain appealing to ensure immune targeting of malignant cells and remnant micrometastasis after surgical resection. However, it remains elusive whether vaccine-based approaches induce differential response rates in patients with HCC with different underlying aetiologies. Regulating the gut microbiota represents another potential therapeutic avenue for treating MASH-associated and ASH-associated HCCs^277–280 (Fig. 4c). These approaches include the depletion of pathogenic microbiota with antibiotic treatment agents^73,281 and enrichment of bacteria that execute probiotic effects²⁸¹.

Conclusions and future directions

It is expected that the incidence of SLD-associated liver cancers will steadily increase over the next three decades, emphasizing the need to better understand the molecular and cellular mechanisms of pathogenesis and treatment response²⁸². While earlier stages of MASH and ASH have been extensively studied, investigating the mechanisms that drive fibrosis and hepatocarcinogenesis remains essential to reduci the risk of HCC development in these patients.

Dissecting disease-specific molecular patterns with the advent of novel technologies in the field of single-cell and spatial transcriptomics, proteomics, epigenomics and metabolomics, may reveal novel therapeutic targets or prediction markers for HCC development and therapy response tailored by the underlying liver disease. In this context, artificial intelligence-based analyses hold great promise for discovering novel molecular patterns through analysis of large and complex datasets composed of multi-dimensional data from patients with SLD-associated HCC. Prospective clinical trials with patient stratification based on the underlying liver disease will be necessary to better understand how specific aetiologies might impact immunotherapy efficacy. Preclinical studies using aetiology-specific in vivo models for HCC will further help to investigate whether disease-specific mechanisms of antitumour immune response exist and how to address them therapeutically. Moreover, while anti-metabolic drugs and lifestyle interventions, such as caloric restriction or exercise, have been shown to halt disease progression in MASH^35,283, the effect of these therapeutic approaches in the context of established MASH-associated and ASH-associated HCC represents a critical next step in the field.

The next phase of research in SLD-associated HCC must move decisively towards aetiology-aware biology. Integrating mechanistic studies, advanced multi-omics, refined preclinical models and clinically annotated patient cohorts will be critical to disentangling shared versus disease-specific drivers of hepatocarcinogenesis. Such efforts will ultimately enable more precise prevention strategies and improve therapeutic efficacy across the full spectrum of SLDs.

Glossary

Cholestasis

A condition characterized by impaired bile flow, leading to the accumulation of bile acids in the liver, promoting liver injury, inflammation and fibrosis

Chronic necro-inflammation

Persistent inflammation accompanied by cell death (necrosis), often driven by immune responses and tissue damage, contributing to fibrosis and carcinogenesis in liver disease

Cirrhosis

A hallmark of advanced stages of steatotic liver disease characterized by extensive fibrosis and regenerative nodules, leading to impaired liver function and increased cancer risk

Hepatic steatosis

Associated with metabolic dysfunction, this is a hallmark of steatotic liver disease, whereby fat accumulates within liver cells

Hepatomegaly

Abnormal enlargement of the liver that is typically associated with conditions such as liver disease, heart failure, infections or metabolic disorders

Insulin resistance

A pathophysiological condition in which specific cells or tissues (for example, muscle, adipose or liver tissue) demonstrate no response to normal or elevated circulating insulin levels, resulting in impaired glucose uptake and metabolism

Kupffer cells

Specialized resident macrophages located in the liver sinusoids that have a key role in innate immunity by removing pathogens and cellular debris and by regulating inflammatory and immune responses in the liver

Lipolysis

The metabolic process of breaking down stored triglycerides into free fatty acids and glycerol, which are then released into the bloodstream for energy use or further metabolic processing

Lipotoxicity

Cellular damage caused by the accumulation of toxic lipid species, such as free fatty acids, diacylglycerols and ceramides, particularly in non-adipose tissues like the liver, contributing to inflammation and fibrosis

Metabolic reprogramming

The adaptation of cellular metabolic pathways to meet the energy demands of growth or stress, commonly observed in cancer and chronic liver diseases

Neutrophil extracellular traps

(NETs). Extracellular networks of decondensed chromatin–DNA, histones and neutrophil-derived antimicrobial proteins released by activated neutrophils

Onco-fetal niche

A tumour microenvironment that resembles the fetal developmental stage where cancer cells exploit similar signalling pathways and cellular interactions to promote tumour growth, survival and metastasis

Polygenic risk score

(PRS). A numerical estimate of an individual’s genetic predisposition to a particular disease, calculated by summing the effects of multiple genetic variants associated with the condition

Transarterial chemoembolization

(TACE). A minimally invasive procedure used to treat liver tumours in which chemotherapy is delivered directly to the tumour via the hepatic artery, while blocking the blood flow, to induce ischaemia and trap the chemotherapeutic agents within the tumour

Transarterial radioembolization

(TARE). A minimally invasive procedure used to treat liver tumours, where radioactive microspheres are delivered through the hepatic artery to directly irradiate the tumour, while blocking blood flow, causing tumour nutrient starvation and an enrichment of the microspheres in the tumour

Viscoelasticity

A material property characterized by the combined behaviour of viscosity (resistance to flow) and elasticity (ability to return to its original shape), which can be useful to describe liver tissue mechanics, particularly in the context of fibrosis