Inflammation and Immunity in Liver Neoplasms: Implications for Future Therapeutic Strategies

Abstract

Over the past two decades, the “hallmarks of cancer” have revolutionized cancer research and highlighted the crucial roles of inflammation and immunity. Pro-tumorigenic inflammation promotes cancer development along with inhibition of anti-tumor immunity, shaping the tumor microenvironment (TME) towards a tumor-permissive state and further enhancing the malignant potential of cancer cells. This immunosuppressive TME allows tumors to evade immunosurveillance. Thus, understanding the complex interplay between tumors and the immune system within the TME has become pivotal, especially with the advent of immunotherapy. Although immunotherapy has achieved notable success in many malignancies, primary liver cancer, particularly hepatocellular carcinoma (HCC), presents unique challenges. The hepatic immunosuppressive environment poses obstacles to the effectiveness of immunotherapy, along with high mortality rates and limited treatment options for patients with liver cancer. In this review, we discuss current understanding of the complex immune-mediated mechanisms underlying liver neoplasms, focusing on HCC and liver metastases. We describe the molecular and cellular heterogeneity within the TME, highlighting how this presents unique challenges and opportunities for immunotherapy in liver cancers. By unraveling the immune landscape of liver neoplasms, this review aims to contribute to the development of more effective therapeutic interventions, ultimately improving clinical outcomes for patients with liver cancer.

Introduction

Since Hanahan and Weinberg proposed the “hallmarks of cancer” two decades ago, the landscape of cancer research has evolved through increased understanding of the complex biology of cancer(1). Among the 14 cancer hallmarks, two are related to inflammation and immunity: tumor-promoting inflammation and avoidance of immune surveillance. Robust innate and acquired immunity eliminate cancer cells or suppress their malignant capacity, whereas inflammation promotes the onset and progression of cancers. The tumor microenvironment (TME) plays a critical role in tumor progression, metastasis, and response to treatment(2). Moreover, the tumor immune microenvironment (TIME) is an important determinant of the effectiveness of anticancer therapy(3,4). Understanding the intricacies of cancer-associated inflammation is essential for advancing the field of cancer biology and optimizing therapeutic strategies.

Immunotherapy, originally developed to enhance the host’s natural defenses against cancer cells, has revolutionized the treatment of many cancers. While surgery, radiotherapy, and chemotherapy are the traditional anticancer strategies, immune checkpoint inhibitors (ICIs) are an effective option for many malignancies. Unlike conventional therapies, ICIs counter immunosuppression induced by tumors and the TME, enabling the immune system to combat cancer cells. The era of modern immunotherapy began in 2011 with approval of the first ICI, ipilimumab (an anti-cytotoxic T-lymphocyte antigen 4 [CTLA-4] antibody), for the treatment of advanced melanoma(5,6). Subsequently, many additional ICIs have been developed, including programmed death-1/programmed death-ligand 1 (PD-1/PD-L1) inhibitors. While ICIs are paradigm-shifting cancer therapy, certain tumors remain unresponsive to these therapies, and mechanism(s) underlying ICI resistance remain unclear. The immune context, encompassing types, density, and spatial distribution of cells in the TIME, correlates well with cancer progression and prognosis. Thus, improving our understanding of the TIME is essential for developing strategies to enhance therapeutic responsiveness and for innovating new cancer therapy targets.

Globally, liver cancer is ranked sixth in incidence and third in mortality among cancers(7). Contrary to the improved survival rates of most cancers in recent decades, survival has not substantially increased for liver cancer(8). Hepatocellular carcinoma (HCC) accounts for approximately 90% of primary liver cancers. It is usually diagnosed at an advanced stage, when treatment options are limited. Over 90% of HCCs develop in patients with advanced chronic liver disease or cirrhosis, making HCC a leading cause of death in these patients. HCC typically arises in the setting of chronic liver conditions with persistent inflammation, such as chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, alcohol-associated liver disease (ALD), or metabolic dysfunction–associated steatotic liver disease (MASLD). Additionally, the liver is a frequent site of metastasis from other malignancies, including gastric, colorectal, pancreatic, breast, and prostate cancers, with hepatic metastasis portending a poor prognosis(9). Surgical resection and chemotherapy are standard treatment for liver metastases, but the presence of unresectable tumors, drug resistance, and post-treatment recurrence pose limit the effectiveness of current therapeutic strategies(10). ICIs may offer superior treatment efficacy for liver cancers, including HCC and metastatic tumors. However, mortality rates for liver cancer remain high because of its etiologic heterogeneity, intrinsic liver immune tolerance, low responsiveness to therapy, and acquired drug resistance(11,12). In-depth understanding of how the immune system regulates liver tumor growth is thereby crucial to develop more effective treatments and improve outcomes.

In this review, we summarize current knowledge of immune-mediated mechanisms in liver neoplasms, including HCC and metastases, by highlighting the TME and TIME. We explore TME heterogeneity and describe key challenges regarding immunotherapy resistance in liver cancer. Finally, we discuss future directions and potential strategies to advance anticancer immunotherapy for HCC and liver metastases.

Inflammation and immunity driving liver cancers

Inflammation and the immune system play pivotal roles in the onset and progression of primary and secondary liver cancers. Under normal conditions, the liver is continuously exposed to a myriad of microbial and food-derived antigens from the gut. Therefore, the hepatic immune system must be tolerogenic to prevent spontaneous inflammation and tissue damage by gut-derived factors(13). However, when this immune system is dysregulated, inflammation occurs and becomes a key driver for neoplasm growth(11,14). Furthermore, hepatic intrinsic tolerogenic immunity impairs cancer surveillance, allowing cancer cells to evade immune responses, thereby promoting oncogenesis and metastasis(11). Thus, the interplay between inflammation and hepatic immunity contributes to cancer progression and metastasis.

The hepatic TME is heterogeneous, composed of cellular and noncellular components (Figure 1). The main cellular components are cancer cells, immune cells (e.g., macrophages, natural killer [NK] cells, dendritic cells, T and B lymphocytes, neutrophils), and stromal cells (e.g., endothelial cells, hepatic stellate cells [HSCs], cancer-associated fibroblasts [CAFs]). Noncellular components include the extracellular matrix (ECM), cytokines, chemokines, and metabolites. The diverse cell populations interact with each other and noncellular components to regulate cancer growth. While TME composition varies depending on HCC etiology, tumor-associated macrophages (TAMs), CAFs, myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Tregs) contribute to the generation of an immunosuppressive TME. Rapid technological development of single-cell and spatial transcriptomic, proteomic, and metabolomic platforms has provided comprehensive high-resolution datasets on tumor-context changes, heterogeneity of the immune landscape, and multifaceted intercellular interactions within the TME. In this section, we review current understanding and discuss unanswered questions regarding the roles of inflammation and immunity in liver cancer development (Figure 2).

Figure 1. Components of tumor microenvironment in the liver.

The hepatic tumor microenvironment (TME) is heterogeneous, composed of both cellular and noncellular components. Major cellular components include cancer cells, immune cells (such as macrophages, natural killer [NK] cells, dendritic cells, T and B lymphocytes, myeloid-derived suppressor cells (MDSCs), or neutrophils), as well as stromal cells (such as endothelial cells [ECs], hepatic stellate cells [HSCs], or fibroblasts). Noncellular components consist of the extracellular matrix (ECM), chemokines, cytokines, and metabolites. These diverse cell populations interact directly with each other or with noncellular components to regulate cancer growth.

Figure 2. The dynamic interactions of inflammation and cancers in the tumor microenvironment.

The tumor microenvironment (TME) is regulated by inflammation and immunity during liver cancer development. (1) Inflammatory pathways connect inflammation with hepatocarcinogenesis. Pro-inflammatory cytokines, like IL-6, TNFɑ, and IL-1β, or pathogen-derived molecules such as LPS, drive inflammation-associated hepatocarcinogenesis. The IL6/JAK/STAT3 pathway is oncogenic and promote hepatocellular carcinoma (HCC) development. Additionally, the NF-κB pathway regulates hepatocarcinogenesis and contributes to drug resistance. (2) Tumor intrinsic pathway plays a critical role in reshaping the tumor immune microenvironment (TIME) during liver cancer progression. Genetic mutations, including loss of TP53 or PTEN, and activation of oncogenic pathways such as Wnt/β-Catenin, Ras-MAPK-ERK, or YAP, contribute to inflammation, immunosuppression, or suppression of anti-tumor immunity. These reprogramming events collectively create a microenvironment that supports tumor growth and progression. (3) Premetastatic niche (PMN) in the liver is established by factors released from primary cancers. These factors include TIMP-1, VEGFA, IL-6 and extracellular vesicles (EVs). Underlying liver disease is an important factor that modifies PMN. Hepatic steatosis resulting from metabolic dysfunction-associated liver disease or alcohol-associated liver disease predisposes the liver to develop a PMN by creating immunosuppressive TIME. (4) Gut dysbiosis is linked to HCC development. The gut microbiome contributes to hepatic stellate cell (HSC) senescence, immunosuppressive TIME formation, tumor-favorable PMN formation, and pro-tumorigenic inflammation (e.g. LPS-induced), enhancing tumor growth.

Key inflammatory pathways during hepatocarcinogenesis

Multiple signaling pathways, including nuclear factor kappa B (NF-κB), Janus kinase/signal transducers and activators of transcription (JAK-STAT), toll-like receptor (TLR), and mitogen-activated protein kinase (MAPK) pathways, have been identified as pivotal regulators of the initiation and resolution of inflammation during hepatocarcinogenesis. The NF-κB pathway is among the most studied pathways linking inflammation and hepatocarcinogenesis, but its actions are context-dependent. This pathway is activated by pathogen-derived molecules, such as lipopolysaccharide (LPS), or inflammatory cytokines, such as tumor necrosis factor (TNF)-ɑ and interleukin (IL)-1β. At basal levels, NF-κB signaling exhibits anti-apoptotic properties in hepatocytes, whereas dysregulated, constitutively active NF-κB leads to spontaneous tumorigenesis. Opposing roles of the NF-κB pathway in hepatocarcinogenesis have been described(11,15). In a mouse model of metabolic dysfunction–associated steatohepatitis (MASH)–induced HCC, TNFɑ from infiltrated macrophages activated NF-κB in hepatocytes, resulting in HCC development(16). However, hepatocyte NF-κB can also prevent HCC development. Inhibiting NF-κB activation by NF-κB essential modulator (NEMO) or MAP3-kinase TGF-β-activated kinase 1 (TAK1) deficiency in hepatocytes leads to spontaneous HCC formation(17,18). This discrepancy may arise from context-dependent differences in timing of NF-κB activation and liver damage. NF-κB activation also participates in sorafenib resistance in patients with HCC. NF-κB and its downstream cytokine lipocalin-2 regulate hepatocarcinogenesis and sorafenib resistance by regulating ferroptosis, an iron-dependent form of cell death(19).

IL-6, a proinflammatory cytokine produced by TME cells, including immune, stromal, and cancer cells, promotes the induction of HCC progenitor cells and the proliferation and survival of HCC cells. IL-6 is associated with sex disparity-associated and obesity-promoted hepatocarcinogenesis(20–22). In HCC, the IL-6–mediated signal transducer and activator of transcription (STAT) 3 pathway is oncogenic, promoting tumor proliferation, survival, anti-apoptosis, invasion, metastasis, angiogenesis, and immune escape(11,23). STAT3 also protects against hepatic inflammation(24,25). In the HCC-TME, STAT3 exerts tumor-promoting effects through tumor-stromal crosstalk(26–28). Single-cell RNA sequencing (scRNA-seq) revealed that HSCs and macrophages produce IL-6 and activate STAT3 in HCC to promote tumor growth(26). Hepatocyte IL-6/STAT3 activation and its downstream serum amyloid A (SAA) suppress T cell infiltration and immune surveillance to promote HCC growth(29), suggesting that IL-6/STAT3 signaling regulates remodeling of the TIME. STAT3 is also activated by IL-22 (produced mainly by CD4⁺ T cells) to promote HCC development(29). Hepatocyte-specific IL-22 overexpression increases hepatocarcinogenesis by promoting cell survival and proliferation without affecting inflammation(30). Also, IL-22–mediated STAT3 enhances HCC resistance to sorafenib(31). IL-22 exerts minimal effects on immune cells because IL-22 receptors are expressed mainly in epithelial cells and HSCs.

Tumor intrinsic pathway links to inflammation

Cancer cell–derived intrinsic factors shape the tumor immune landscape(32). Many genetic events that drive liver tumorigenesis by activating inflammatory and immune signaling contribute to the tumor-promoting TIME. Loss of TP53 or phosphatase and tensin homolog (PTEN) or activation of Wnt/β-catenin, Ras-MAPK-ERK pathway, or Myc is frequently observed in HCC. These changes reprogram the liver microenvironment to promote inflammation, enhance immunosuppression, or dampen anti-tumor immunity, fostering a tumor-favorable microenvironment(33–36). An scRNA-seq study used genetically engineered Myc^OE/Trp53^KO, Myc^OE/Pten^KO, Nras^G12D/Pten^KO, and Nras^G12V/Pten^KO HCC mouse models to demonstrate that cancer cell-intrinsic factors shape the TIME in HCC(36). Myc^OE and Nras^G12D mice had “excluded/immunological cold” tumors with limited T cell infiltration, whereas Nras^G12V/Pten^KO mice had “inflamed/immunological hot” tumors with abundant effector T cells, indicating that distinct Nras mutations are associated with different pro-tumorigenic, immunosuppressive environments with different T cell phenotypes. Cancer-intrinsic yes-associated protein 1 (YAP) signaling is another mechanism for modulating the TIME. In MASLD with liver metastasis, YAP activation was augmented in cancer cells that induced TAM polarization to M2 macrophages and CD8⁺ T cell exhaustion through cancer-derived cysteine-rich protein 61 (CYR61), creating a tumor-favorable TIME(37). Higher-throughput transcriptomic studies may help develop tailored immunomodulating treatment strategies based on cancer molecular profiles.

Premetastatic niche formation in liver metastasis

Premetastatic niche (PMN), a microenvironment permissive for metastatic cancer growth, can be formed by primary cancer-derived factors, including soluble factors and extracellular vesicles (EVs) that reach the liver before metastasis. Tissue inhibitor matrix metalloproteinase 1 (TIMP-1) produced by primary cancers, such as colorectal cancer (CRC) and pancreatic ductal adenocarcinoma (PDAC), contributes to hepatic PMN establishment through HSC activation, chemokine (C-X-C motif) ligand 12 (CXCL12) production, and C-X-C chemokine receptor type 4 (CXCR4)⁺ neutrophil infiltration(38,39). Vascular endothelial growth factor (VEGF) A from primary cancers stimulates TAMs to produce CXCL1, thereby attracting CXCR2⁺ MDSCs to create a hepatic PMN(40). Also, IL-6 from PDAC induces hepatocyte STAT3 activation and SAA production to promote myeloid cell recruitment and hepatic fibrosis, leading to PMN formation(41). EVs from primary cancers play a major role in forming a hepatic PMN. PDAC-derived small EVs containing macrophage migration inhibitory factor (MIF) are internalized by Kupffer cells that produce TGF-β and upregulate fibronectin in HSCs, which together with bone marrow-derived macrophages, create a hepatic PMN(42). PDAC-derived EVs also carry palmitic acid, which stimulates Kupffer cells to induce TNFα-mediated inflammation(43). Furthermore, CRC-derived EVs contain microRNA (miR)-203, miR-21, and integrin subunit β-like 1 (ITGBL1), which induce M2 polarization, activate Kupffer cells to produce IL-6 via TLR7, and activate fibroblasts via NF-κB to produce IL-6, IL-8, and IL-1β, thereby promoting PMN formation and metastasis growth(37,44,45). The hepatic PMN is also induced by primary CRC-derived bacteria. Tumor-resident Escherichia coli disrupt the gut vascular barrier (associated with increased plasmalemma vesicle-associated protein-1 expression), resulting in bacterial translocation to the liver and subsequent PMN development(46).

MDSC, TAM, and Treg infiltration is important for hepatic PMN formation. Single-cell and spatial transcriptomics analyses showed that metastatic liver tumors are enriched with immunosuppressive, metabolically activated MRC1⁺CCL18⁺ M2 macrophages(47). In liver metastatic TME, infiltrated immunosuppressive CD11b⁺F4/80⁺ macrophages induce antigen-specific T cell apoptosis(48). Accumulated M2-TAMs enhance stabilization of immunosuppressive fibrinogen-like protein 1 to facilitate CRC liver metastasis(49). Immunosuppressive Foxp3⁺ Tregs produce IL-10, which upregulates PD-L1 in monocytes to dampen CD8⁺ T cell infiltration and antitumor immunity in liver metastases(50).

Efferocytosis (engulfment of apoptotic cells in macrophages) has been linked to PMN formation(51). In early-stage PDAC liver metastasis, cancer cells induce low-grade hepatocyte injury, and dead hepatocytes are cleared by efferocytosis. During this process, efferocytosis induces macrophages to an immunosuppressive tumor-promoting phenotype. Interfering with efferocytosis inhibits macrophage immunosuppressive properties and restores antitumor immunity, reducing liver metastasis.

Underlying liver disease is an important modulator of PMN. Hepatic steatosis resulting from MASLD or ALD predisposes to a PMN. In MASLD, hepatic steatosis creates an immunosuppressive TIME by infiltrating M2-TAMs(37,52). These TAMs activate the NLR family CARD domain containing 4 (NLRC4) inflammasome and upregulate IL-1β, VEGF, TGF-β, and PD-L1. The metastatic TIME is enriched with CD8⁺ T cells with high PD-1 expression, which suppresses anticancer immunity. Neutrophils, MDSCs, and scar/lipid-associated macrophages contribute to an immunosuppressive TIME in MASLD-associated liver metastasis(53). In ALD-associated hepatic steatosis, IL-6/STAT3 in hepatocytes leads to secretion of lipocalin-2, which promotes neutrophil infiltration and T cell exhaustion to create a PMN(54).

Intestinal microbiome: an emerging cancer hallmark

Gut dysbiosis is frequently observed in patients with HCC and is associated with HCC development(55–57). Gut microbiome promotes anti-cancer NKT cell activity through CXCL16–CXCR6 interaction, which is mediated by secondary bile acid conversion by gut microbiota(58). Conversely, gut microbiome also contributes to immunosuppression. In primary sclerosing cholangitis and colitis-associated cholangiocarcinoma (CCA) mouse models, translocated LPS induces CXCL1 expression in hepatocytes, resulting in CXCR2⁺ polymorphonuclear-MDSC infiltration to form an immunosuppressive TIME(59). In obesity-associated HCC, gut microbiota–derived lipoteichoic acid induces IL-1β and IL-33 production in a TLR2- and gasdermine D-dependent manner, resulting in HSC senescence. Senescent HSC–derived IL-33 infiltrates Treg cells in the liver TIME to promote HCC progression(60). Gut dysbiosis induces secretion of IL-25 from intestinal epithelial tuft cells, leading to hepatic M2-TAM polarization to create an immunosuppressive TIME(61). Intriguingly, intestinal dysbiosis in NLRP6-null mice induced TLR4-dependent expansion of monocytic MDSCs and suppression of CD8⁺ T cells in the liver, enhancing genetically induced HCC development. This enhancement correlated with loss of Akkermansia muciniphila, and Akkermansia supplementation restored intestinal barrier function and suppressed hepatic inflammation and fibrosis(62).

Similarly, gut microbiome facilitates the development of a hepatic PMN. In mice, antibiotic-induced microbiome depletion reduced tumor burden in CRC and PDAC liver metastasis(63). Gut microbiota–derived LPS triggers the formation of neutrophil extracellular traps via TLR9-dependent MAPK activation, promoting CRC liver metastasis(64). LPS also promotes TLR4-dependent M2-TAM polarization through cancer cell–derived cathepsin K in liver metastases(65), whereas blocking LPS prevents liver metastasis by reducing immunosuppressing cells(66). Pathogenic gut microbiota, including Fusobacterium nucleatum, Bacteroides fragilis, and E. coli, promote liver metastasis(46,67–69). Formate, a Fusobacterium-derived metabolite, induces T helper 17 (Th17) cell infiltration to enhance CRC metastasis(70).

Fungi are another human commensal microbe. Aflatoxin, an Aspergillus-derived toxin that can contaminate food, is among the most significant causes of HCC. Candida albicans and Malassezia are abundant in advanced HCC, whereas Saccharomyces is found in early-stage HCC(71,72). Current data on fungi in HCC and liver metastases are limited, but fungal effects should not be overlooked(73). Relationships between gut microbiota and cancer-permissive liver TME and PMN deserve further confirmation and exploration.

Immunometabolic reprogramming: another emerging cancer hallmark

Metabolic changes in tumors affect the composition and function of immune cells, potentially supporting tumor growth and impairing anticancer immunity(74,75). Reprogramming methionine metabolism in tumors induces dysregulation of methionine recycling machinery and leads to increases in 5-methylthioadenosine and s-adenosylmethionine (SAMe) (methionine metabolites). This process is mediated by MAT2A (a SAMe-synthesizing enzyme) in HCC cells and results in T cell exhaustion(76). Expression of MAT1A, another SAMe-producing enzyme, in hepatocytes has opposite effects: it inhibits CD8⁺ T cell infiltration and suppresses HCC growth and liver metastasis(77). These findings highlight the opposing functions of MAT2A expression in tumors and MAT1A expression in hepatocytes. Succinate, a tricarboxylic acid cycle intermediate, regulates the TAM phenotype. Lung cancer-derived succinate facilitates M2-TAM polarization, migration, and invasion through the succinate receptor SUCNR1, contributing to liver metastasis(78). Metabolic reprogramming in immune cells also impacts cancer development. Lipid metabolism has divergent effects on NK cell function. Obesity-induced NK cell uptake of free fatty acids results in metabolic and functional defects, impairing their antitumor responses(79). In contrast, cholesterol accumulation in NK cells enhances their antitumor activity in HCC by promoting lipid raft formation(80). Exhaustion of CD8⁺ T cells is also associated with lipid metabolic reprogramming in T cells, characterized by enhanced lipolysis and fatty acid oxidation(81). Thus, metabolic reprogramming is crucial for immune cell functions in the TME.

These findings suggest reprogramming tumor metabolism as a potential strategy for cancer treatment. However, the origin of cells producing metabolites remains unclear, and targeting key metabolites may also inhibit immune cells, as they share similar metabolic requirements and compete for common nutrient resources within the TME(82). Additional research is required to understand the metabolic processes of tumor and TIME cells.

Etiologic heterogeneity in liver cancer-associated inflammation

HCC usually develops in patients with underlying liver disease, including chronic HBV or HCV infection or nonviral etiologies (e.g., ALD, MASLD). While viral-associated HCC is decreasing because of vaccination and antiviral therapy, the incidence of MASLD-related HCC is increasing(83,84). Approximately 80% of patients with MASLD-HCC have hepatic cirrhosis, whereas 20% do not(85). As HCC is associated with diverse etiologies, its etiology could influence the immune landscape and efficacy of immunotherapy.

Emerging evidence indicates that the effectiveness of HCC immunotherapy varies according to the underlying liver disease. A meta-analysis of three clinical trials (IMbrave150, KEYNOTE-240, CheckMate459) revealed that the survival advantage of immunotherapy is lower in nonviral-HCC than in viral-HCC(86). In the IMbrave150 study, survival was improved with atezolizumab (anti–PD-L1) plus bevacizumab (anti–VEGF-A), compared with sorafenib (a multikinase inhibitor), in patients with viral-HCC but not nonviral-HCC(86). The KEYNOTE-240 study demonstrated greater survival benefit from pembrolizumab (anti–PD-1) in patients with HBV-HCC than in those with HCV-HCC or nonviral-HCC(87). Differential sensitivity to immunotherapy between nonviral- and viral-HCC may be attributed to different immune responses or immune surveillance mechanisms. HBV-HCC and HCV-HCC are associated with T cell-mediated immune tolerance, characterized by increased Tregs, exhausted T cells, MDSCs, and IL-10–producing immature B cells and reduced effector T cells(88–91). scRNA-seq analyses identified enrichment of CD8⁺ T cell clusters with CD8⁺ T cell exhaustion in viral-HCC, compared with nonviral-HCC(92). These T cell findings were associated with high ICI efficacy in viral-HCC. Conversely, exhausted PD-1⁺CXCR6⁺CD8⁺ T cells were increased in MASLD-HCC and were associated with ineffective anti-PD-1 immunotherapy(86). In a preclinical MASLD-HCC model, CXCR2 inhibition improved ICI effectiveness, suggesting that neutrophils contribute to reduced ICI efficacy in MASLD-HCC(93). In MASLD-HCC, selective depletion of CD4⁺ T cells and aberrant activation and exhaustion of CD8⁺ T cells promote impaired anti-tumor surveillance and enhanced tumor growth, respectively(86,94). Furthermore, liver-resident IgA-producing cells in MASLD-HCC express PD-L1 and IL-10, which suppresses liver cytotoxic CD8⁺ T cells and interferon-γ production(95). MASLD-HCC immune evasion is associated with β-catenin mutation, TNF receptor superfamily, member 19 (TFNRSF19) upregulation, and a senescence-associated secretory phenotype(96). Moreover, Th17 cells induce adipose tissue inflammation, thereby promoting MASLD-HCC development(97). Spatial proteomics analyses revealed that close interactions between CD8⁺ T cell and immunosuppressive cell populations contribute to impaired antitumor immunity in MASLD-HCC(98). Exhausted PD-1⁺CD8⁺ T cells interact with PD-L1⁺ MDSCs and TAMs in MASLD-HCC (but not viral-HCC), potentially impeding antitumor immunity.

Nevertheless, the HIMALAYA trial demonstrated improved survival benefits with tremelimumab (anti–CTLA-4) plus durvalumab (anti–PD-L1) versus sorafenib in patients across all viral and nonviral etiologies(99). Hence, more data are required to fully understand the impact of different HCC etiologies on immunotherapy.

Evidence suggests that intestinal microbiome alterations may be associated with etiologic differences between viral- and nonviral-HCC(100). Patients with HBV-HCC have higher bacterial diversity than those with nonviral-HCC. Furthermore, individuals with nonviral-HCC have more pathogenic bacteria (e.g., Escherichia, Shigella, Enterococcus) and fewer bacteria associated with anti-inflammatory properties, whereas the opposite bacterial composition has been observed in patients with HBV-HCC.

The liver cancer TME is influenced by the underlying liver disease. As HCC typically arises within the context of chronic liver disease, the role of nonresolving inflammation and fibrosis/cirrhosis in HCC tumorigenesis has been well studied(101,102). HSCs in tumor-adjacent liver could create an immunosuppressive TIME by promoting the expansion of MDSC and the differentiation of Tregs(103,104). In tumor-adjacent liver, CX₃CL1⁺HSCs interact with CX₃CR1⁺macrophages to suppress CD8⁺T cell activity within the TIME(105). Considering that the adjacent liver exhibits a unique immune landscape distinct from the HCC-TME(105,106), further research is needed to understand the complex interplay between the HCC-TIME and the adjacent liver.

Increasing evidence suggests that liver steatosis facilitates liver metastasis by reshaping the hepatic immune landscape to form a PMN. Primary CRC accounts for approximately 50% of metastatic liver cancers, and preclinical and clinical studies demonstrated higher rates of CRC liver metastasis in mice and patients with steatotic livers than in those without steatosis(37,52,107–109). Similar findings were observed for metastases from breast, prostate, and lung cancers(110–112). Steatotic liver induces an immunosuppressive TIME consisting of immunosuppressive M2-TAMs and exhausted CD8⁺ T cells, which enhance metastatic tumor growth(37,52). Moreover, liver metastasis induces systemic immunosuppression by depleting CD8⁺ T cells and restoring M2-TAMs(48,113). Given that obesity and MASLD suppress anti-tumor immunity of CD8⁺ T cells(37,114), comprehensive analysis of T cell phenotypes is essential to unravel the unique immunosuppressive mechanisms within the HCC-TME, in both the presence and absence of underlying liver disease.

Fibrotic TME influences the immune milieu of liver cancers

Liver fibrosis/cirrhosis is frequently seen in patients with HCC. Patients with cirrhosis are at high risk of developing HCC and CCA, and cirrhosis limits drug delivery to hepatic tumors, negatively impacting prognosis. A recent cohort study found that the presence of cirrhosis in patients with CCA was associated with increased chemotherapy-induced toxicity and shorter survival(115), suggesting that these patients may require nonstandard therapies.

Hepatic inflammation and fibrosis are closely intertwined. Persistent tissue damage and inflammation activate HSCs, triggering their transition to myofibroblasts. Myofibroblasts produce ECM (including fibrillar collagen), resulting in hepatic remodeling and fibrosis(116). HSCs are the main source of CAFs in primary and metastatic liver cancers(117). Interactions between HSCs, CAFs, and ECM play a significant role in the onset and progression of liver cancers. The heterogenous TME enhances tumor growth, immune evasion, and therapy resistance.

In HCC, quiescent, cytokine-producing HSCs (cyHSCs) with enhanced hepatocyte growth factor (HGF) expression suppress tumor development, whereas activated, myofibroblastic HSCs expressing collagens promote tumor development(118). In CCA and liver metastases, inflammatory CAFs (iCAFs) promote tumor growth via HGF, and myofibroblastic CAFs (myCAFs) promote tumor growth via hyaluronan synthase 2; however, myCAF-derived collagens oppose or do not affect liver tumor growth(119,120). In metastatic tumors, myCAF-derived collagens suppress the growth of liver metastases from desmoplastic CRC and PDAC(120). These findings likely reflect mechanical barriers induced by fibrotic TMEs, which restrict tumor growth and drug delivery. Differential ECM effects may be partially explained by differences in ECM structure, form, stability, and abundance. CAFs and ECM may interact with tumor cells, depending on the degree of fibrosis or desmoplastic stroma. This is supported by observations of opposing roles of collagens according to their size in PDAC, a highly desmoplastic cancer(121). PD-L1 may be a key regulator promoting HSC differentiation into tumor-promoting myCAFs and modulating CCA growth (effects that are independent of PD-L1/PD-1-mediated immunosuppression)(122).

CAFs and ECM remodeling shape the hepatic TIME by attracting immunosuppressive cells. In CCA, FAP⁺ iCAFs recruit MDSCs via STAT3-CCL2 signaling(123,124). In HCC, CD36⁺ CAFs secrete MIF to foster an immunosuppressive TME by recruiting CD33⁺ MDSCs(125). CAFs induce M2 macrophages via the endosialin/ growth arrest-specific protein 6 (GAS6) axis and N2 neutrophils through CAF-derived cardiotrophin-like cytokine factor 1 (CLCF1) and tumor cell-derived CXCL6 and TGF-β in HCC(126,127). In HCC, CAF-derived IL-6 induces PD-L1⁺ neutrophils via the JAK-STAT3 pathway, which impairs T cell function through PD-L1/PD-1 signaling(128). In liver metastasis, CAFs communicate with immune cells to create a PMN in which IL-6 from activated HSCs recruits MDSCs and inhibits CD8⁺ T cell activity, thereby fostering an immunosuppressive milieu(129). HSCs also contribute to activation of dormant metastatic cancer cells. CXCL12 secreted by HSCs interacts with CXCR4 on NK cells to suppress NK cell activity, promoting transition of breast cancer cells from dormancy to active metastatic growth(130). Despite the significant contributions of CAFs in remodeling TIMEs, CAF-derived type 1 collagen does not alter the TIME despite possessing tumor-restraining effects(120). Further studies are required to elucidate the role of collagens and other ECM components (e.g., hyaluronan) in shaping the hepatic TIME.

ECM remodeling in the primary CRC promotes establishment of a perivascular metastatic environment for liver metastasis. scRNA-seq analysis revealed that transcription factor 21 (TCF21)^high “tumor matrix-pericytes” increase perivascular ECM stiffness, collagen remodeling, and basement membrane degradation to establish a perivascular environment in primary CRC promoting liver metastasis(131). Collectively, CAFs foster an immunosuppressive TME by modulating immune systems. ECM remodeling in primary sites also affects the progression of liver metastasis. Modulating the CAF-mediated TIME and the perivascular metastatic environment could be useful approaches for reducing tumor burden and increasing therapeutic efficacy.

Therapeutic strategies for liver cancer targeting TIME

Management options for HCC vary according to tumor burden, liver function, comorbidity, and patient performance. Surgical resection is indicated for only early-stage HCC(132). Advanced HCC is treated with systemic therapy, including multikinase inhibitors (e.g., sorafenib) and VEGF inhibitors (e.g., bevacizumab), but acquired drug resistance limits long-term benefits in most patients. ICIs were recently approved for advanced HCC as monotherapy or combination therapy. Immune checkpoint molecules in the HCC-TIME include PD-1, CTLA-4, lymphocyte-activation gene 3 (LAG-3), V-domain immunoglobulin suppressor of T cell activation (VISTA), and T cell immunoglobulin mucin domain containing-3 (TIM3). The IMbrave150 trial showed superior overall and progression-free survivals with atezolizumab (anti–PD-L1) plus bevacizumab (anti-VEGF) than with sorafenib in individuals with advanced HCC(133). In the phase 3 HIMALAYA trial, overall survival was better with tremelimumab (anti–CTLA-4) plus durvalumab (anti–PD-L1) than with sorafenib in treatment-naïve patients with advanced HCC(121). The CheckMate459 trial demonstrated that nivolumab (anti-PD-1) is a viable systemic therapy for patients with HCC who are ineligible for anti-angiogenic or tyrosine kinase inhibitor therapy(134). Multiple clinical trials involving ICIs are currently underway(12,135), but treatment efficacy remains limited by HCC etiology and the diverse TME regulating immune responses (Figure 3).

Figure 3. Reshaping the TIME in liver cancer with immune checkpoint inhibitors reduces tumor burden.

The tumor immune microenvironment (TIME) in both primary and secondary liver cancer development is characterized as immunosuppressive or poorly immunogenic, marked by an abundance of immunosuppressive cell types such as myeloid-derived suppressor cells (MDSCs), M2-polarized tumor-associated macrophages, N2-polarized neutrophils, and regulatory T (Treg) cells. These cell types collectively promote tumor growth and progression. Immunotherapeutic strategies, particularly immune checkpoint inhibitors (ICIs), either alone or in combination therapies, have been extensively investigated in preclinical settings. These strategies aim to reshape the liver cancer TIME into a more immunogenic environment characterized by increased presence of CD8 T cells, CD4 T cells, and natural killer (NK) cells, and reduced numbers of immunosuppressive cells, thereby creating a less favorable environment for tumor growth. However, the response and resistance to ICIs are influenced by the heterogeneous nature of the TME. (1) Host-related factors such as cancer etiology, alterations in the gut microbiome, and underlying liver disease contribute to TIME heterogeneity and influence ICI response. The distinct immunosuppressive TIME, shaped by host-related factors, impacts therapy outcomes. (2) Tumor-intrinsic factors including genetic mutations modulate the TIME and affect ICI response. Genetic variations contribute to individual differences in the impact of ICIs on immune responses and resistance. (3) Variability in the composition of TIME among individuals can limit the effectiveness of ICIs, with differences in immune cell numbers and activity, such as the absence of CD8 T cells, impacting therapy outcomes. Understanding these complexities is crucial for optimizing ICI therapies and enhancing treatment outcomes in liver cancer.

HCC-TIME composition can affect the efficacy of ICI therapy. Abundant Tregs and MDSCs contribute to immunotherapy resistance, whereas abundant CD8⁺ T cells are associated with a better response(136). Cytotoxic immunity is frequently suppressed in HCC, hindering the efficacy of immunotherapy. ICIs are often ineffective in “cold tumors” (characterized by minimal T cell infiltration), and absence of CD8⁺ T cells in the HCC-TIME is associated with poor outcomes after ICI therapy(137). Conversely, higher cytolytic activity of CD8⁺ T cells predicts a better response to immunotherapy(138). Combining bevacizumab (anti-VEGF) with ICIs may increase T cell infiltration to enhance treatment efficacy(139). Accumulation of immature CD11b^-CD27^- NK cells (with reduced NK cell function) has also been observed in HCC(140,141).

HCC etiology affects the HCC-TIME and, in turn, ICI efficacy. ICI responsiveness differs between viral- and nonviral-HCC. In MASLD-HCC, anti–PD-1 immunotherapy is ineffective because of impaired tumor surveillance(86). Efficacy of anti–PD-1 therapy can be restored by inhibiting neutrophils with a CXCR2 antagonist. Adding a CXCR2 antagonist reduced tumor burden and prolonged survival in a mouse model of MASLD-HCC(93). Combining an ICI with a CXCR2 antagonist increased intratumoral XCR1⁺ dendritic cells, CD8⁺ T cells, and reprogramming of tumor-associated neutrophils (TANs), which are associated with anti-tumor cytotoxicity. Targeting MDSCs and immunosuppressive TAMs is a potential strategy for overcoming ICI resistance in MASLD-HCC(98). Oncogenic pathways driven by genetic alterations also impact the HCC-TIME and responses to immunotherapy. β-catenin activation by CTNNB1 mutations is associated with enhanced immune escape and resistance to anti–PD-1 via impaired T cell activity and dendritic cell recruitment(35).

Soluble stromal factors also impact immunotherapy. TGF-β signaling promotes an immunosuppressive TIME by inducing tolerogenic dendritic cells and M2 polarization. Accordingly, high plasma TGF-β levels are associated with resistance to anti–PD-L1 therapy(35). Low major histocompatibility class I (MHC-I) expression in cancer cells is associated with reduced CD8⁺ T cell activity and ineffectiveness of ICIs(142,143). Inhibitors of fatty acid synthesis (FASN) or palmitoyl-protein thioesterase-1 can restore MHC-I expression and CD8⁺ T cell cytotoxicity. Combining FASN inhibitors with ICIs dramatically suppresses HCC growth(144,145). YAP upregulates PD-L1 and downregulates MHC-I to promote cancer growth by suppressing CD8⁺ T cell infiltration and activity, whereas YAP inhibition enhances anti–PD-1 therapy by upregulating MHC-I in cancer cells(146).

Gut microbiome composition also impacts immunotherapy effectiveness. Patients responding to anti–PD-1 therapy have higher diversity and abundance of gut microbiome than nonresponders(147). Furthermore, anti–PD-1 therapy effectiveness improved in mice receiving fecal microbiota transplantation (FMT) from patients responding to ICIs but not in mice receiving FMT from nonresponders(148). Responsiveness depended on Akkermansia muciniphila, since oral supplementation with this bacterium restored anti–PD-1 therapy effectiveness in mice receiving FMT from nonresponders. Additionally, gut Bacteroides can influence the effectiveness of anti–CTLA-4 therapy(149). These findings suggest an important role of gut microbiome in the efficacy of ICIs.

Liver metastasis is often associated with immunotherapy resistance. The hepatic immune microenvironment is tolerogenic and immunosuppressive, which could contribute to immunotherapy resistance. Immunosuppressive M2-TAMs and exhausted CD8⁺ T cells are critical for limiting immunotherapy efficacy in liver metastasis(37,52,150). Intriguingly, the presence of liver metastases induces systemic resistance to immunotherapy(48,113). Liver metastasis has been shown to reduce the responsiveness to immunotherapy but not chemotherapy. A recent randomized clinical trial provides additional evidence that liver metastasis may determine immunotherapy resistance in patients with CRC(151–153). Additionally, mice bearing primary tumors without liver metastasis respond to anti–PD-L1 therapy, whereas those with concomitant liver metastases fail to respond. Immunosuppressive CD11b⁺F4/80⁺ macrophages induce Fas ligand–Fas-dependent T cell apoptosis(48). This results in loss of systemic antigen-specific CD8⁺ T cells, contributing to ineffective tumor control and poor immunotherapy responsiveness. Tregs and MDSCs also influence the poor response to anti–PD-1 ICIs, which can be reversed by inactivating Tregs(113).

Distant primary tumors may limit the efficacy of immunotherapy systemically through EVs. Primary tumor-derived EVs may travel to the liver, where they induce inflammation and metabolic remodeling(43). This process promotes liver steatosis and reduces the effectiveness of chemotherapy. Given that MASLD fosters a PMN, hepatic steatosis induced by extrahepatic tumors may promote liver metastasis. Further research is required to understand the impact of tumor-derived EVs on modulating the hepatic TIME and immunotherapy efficacy.

Most immune interventions beyond ICIs, including chimeric antigen receptor (CAR)-T cells, allogenic NK cells, and oncolytic viruses, are still in preclinical or early clinical stages(154–156). Antagonizing IL-10 increases carcinoembryonic antigen–specific CAR-T cell activity and cytotoxicity to inhibit CRC liver metastasis(157). Novel ICIs or ICI combination therapies hold much promise for extending the benefits of immunotherapy to an increasing number of patients(158,159). However, it remains unclear which strategies will supplement or replace current systemic agents.

Conclusions and Future Directions

The complexity of inflammation and immunity and the heterogeneous TME influence the development of primary and secondary liver cancers. With advances in single-cell and spatial transcriptomics analyses, new insights into the cellular heterogeneity and complexities of the TME are abundant. Specific etiologies and underlying liver diseases are not only risk factors for HCC but also important contributors to TME and TIME heterogeneity. Machine-learning technology could help interpret complex datasets derived from single-cell and spatial multi-omics for heterogenous TMEs in different underlying liver disorders. Although immunotherapy is clearly beneficial for many malignancies, resistance to immunotherapy is a major concern in liver cancer. ICI therapy combined with strategies to modulate immune activity, such as enhancing MHC-I expression and cytotoxic CD8⁺ T cell activity or inhibiting pro-tumorigenic myeloid cell activity, may overcome resistance to ICIs(93,144,145,157). Multi-omics analysis, combined with machine-learning platforms, may predict effective ICI combination therapies for HCC and liver metastases(160).

One must also consider the possibility of ICIs inducing immune-related adverse events (irAEs) by triggering autoimmunity. ICI therapy can continue under close monitoring in the presence of mild irAEs. However, moderate to severe irAEs may result in a delay in therapy or lead to life-threatening declines in organ function and quality of life(161). Therefore, early detection and proper management of these toxicities are crucial. The liver is a frequent target of irAEs, with hepatotoxicity occurring in 2–25% of patients(162). Hyperactivation of CD8⁺ T cells and myeloid cells by ICIs results in upregulation of inflammatory cytokines and cytotoxic cell death, creating a pro-inflammatory hepatic environment(163,164). Recent research has highlighted CXCR3⁺CD8⁺ effector memory T cells and CD14⁺ antigen-presenting cells as significant contributors in mediating the response to ICIs and irAEs within the HCC-TME(165). As ICI use continues to expand, understanding the mechanisms underlying irAEs is imperative to differentiate adverse effects from desired anti-tumor responses.