Thrombomodulin Mitigates the Progression of Hepatocellular Carcinoma in Steatotic Liver

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a common hepatic disorder in developed countries. MASLD, including the more aggressive metabolic dysfunction-associated steatohepatitis (MASH), is associated with hepatocellular carcinoma (HCC).^1,2 Obesity and liver steatosis promote chemical carcinogen-induced hepatocarcinogenesis and liver metastasis.^3–6 We have previously shown that diet-induced liver steatosis promotes intrahepatic metastases of HCC by activating the hepatic stellate cells in a rodent model.⁷ Moreover, several reports have shown that hepatic steatosis promotes colorectal liver metastasis and intrahepatic metastasis in liver cancer,^8–10 although not uniformly.^11,12 However, the mechanisms underlying hepatic steatosis-induced liver cancer progression remain largely unexplored.

Thrombomodulin (TM) is a transmembrane protein predominantly expressed in the endothelial cells. The extracellular domains of TM comprise a highly charged N-terminal lectin-like domain (D1), a domain with six epidermal growth factor (EGF)-like structures (D2), and serine-and threonine-rich domains (D3).^13,14 The D1 domain traps damage-associated molecular patterns (DAMP), such as the high- mobility group box 1 (HMGB-1) protein, and exerts an anti-inflammatory action.^15,16 Although HMGB-1 promotes cancer cell survival, it has a paradoxical effect on cancer cell death.¹⁷ The D2 domain is a cofactor for thrombin binding, which mediates protein C activation and exhibits anticoagulant activity.^13,14 Recently, several studies have demonstrated that TM expression in tumor tissues is associated with better prognosis in several cancers.^18,19 Other studies have demonstrated that manipulation of TM expression in tumor cells alters biological phenotypes, such as tumor cell proliferation.^20–22 Several studies have demonstrated that recombinant TM (rTM) suppresses the metastasis and progression of pancreatic cancers by blocking neutrophil extracellular traps (NETs)²³ and thrombin-induced protease activated receptor 1²⁴ in animal models. TM has been implicated in tumor suppressive effects in cancer biology.

In steatotic livers, swelling of the hepatocytes caused by lipid accumulation leads to a narrow hepatic sinusoidal space and impairs hepatic microcirculation, thereby inducing hepatic hypoxia.^25,26 Hypoxia-inducible factor 1 (HIF1), a key molecule in hypoxia, is induced in mouse models of diet-induced hepatic steatosis and in patients with MASLD.^27,28 Furthermore, hypoxia promotes HCC progression through the induction of HIF1.²⁹ Hypoxia induces the downregulation of TM expression in endothelial cells.³⁰ Recently, we showed that diet-induced liver steatosis leads to downregulation of TM and reduced production of protein C via TM in mice.³¹

Therefore, liver steatosis-induced hypoxia may enhance the progression of liver cancer via downregulation of TM; however, this possibility has not yet been explored. Thus, we sought to examine whether steatosis-induced hypoxia enhances HCC progression and investigate whether the administration of rTM suppresses HCC progression in a mouse model. We used heterogenous TM-knockout (KO) (TM^+/−) mice in which the functional activity of TM is decreased to approximately 50% compared with that of wild-type mice, to investigate whether downregulation of TM promotes HCC progression. Furthermore, we explored the role of HMGB-1 in the tumor-suppressive effects of TM on HCC. In vitro assays, we investigated the role of HMGB-1, under hypoxia, enhanced the migration of the HCC cells, and rTM and rTM-D1 suppressed the migration of the HCC cells through the absorption of HMGB-1.

Materials and Methods

Animals

Four-week-old male C56BL/6 mice were purchased from the Charles River Breeding Laboratories (Osaka, Japan). For 4 weeks, the mice were fed a choline-deficient diet (CDD, A06071302, New Brunswick, USA), which was purchased from the RESEARCH DIETS (New Brunswick, USA) to encourage the development of a steatotic liver.

Homogenous TM-KO (TM^−/−) mice are lethal. Thus, heterogeneous TM-KO (TM^+/−) mice were generated by Setsurotech Inc. (Tokushima, JAPAN) using a previously described genome editing technique with several modifications.^31,32

All animal experiments were performed in accordance with the guidelines for animal experiments of the Kure Medical Center. This animal study (approval number: 2019–02) was reviewed and approved by the Institutional Review Board of the Kure Medical Center.

Cell Lines

The mouse HCC cell line Hepa1-6 and human HCC cell line Huh-7 (RRID: CVCL_0336) were purchased from the Riken BioResource Research Center (Tsukuba, Japan). The human liver sinusoidal endothelial cells (hLSECs) were purchased from ScienCell Research Laboratories (California, USA). The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) in a humidified atmosphere of 5% CO₂ under normoxic state (21% O₂) at 37°C. The cells were transferred to a hypoxic chamber (2% O₂) when needed. All human cell lines have been authenticated using STR (or SNP) profiling within the last three years. All experiments were performed with mycoplasma-free cells.

HCC and Hepatic Ischemia-Reperfusion (IR) Models

The HCC model was developed as follows: the mice were anesthetized with isoflurane and underwent laparotomy with a 1 cm-midline incision. A total volume of 0.2 mL of PBS containing the Hepa1-6 cells (0.5 x 10⁶ cells) was injected into the anterior pole of the spleen using a 27G needle. The tumor cells were allowed to circulate for 10 min. Some mice that received splenic inoculation with tumor cells underwent an additional 15 min of circulation, followed by splenectomy and skin closure. Some mice that received splenic inoculation of tumor cells were subjected to hepatic IR and underwent partial hepatic ischemia (70%); the hepatic partial ischemia was induced by clamping the hepatic artery and portal vein to the left and middle liver lobes with microvascular clips. After 60 min of ischemia, the clamp was removed and reperfusion was performed for 15 min before splenectomy. Mice that did not undergo hepatic IR were blindly divided into two groups: treatment and control groups: treatment mice were administered an intravenous injection of rTM (6 mg/kg) or recombinant N-terminal lectin-like domain1 (r TM-D1) (2.5 mg/kg) dissolved in saline (0.2 mL) 3 times / week for 3 weeks after surgery, and control mice were administered an intravenous injection of saline at the same dose (Figure 1A). Mice that underwent hepatic IR were also blindly divided into two groups: treatment mice were Mice administered an intravenous injection of rTM (6 mg/kg) or recombinant N-terminal lectin-like domain1 (r TM-D1) (2.5 mg/kg) dissolved in saline for 7 days after surgery, and the control group was administered an intravenous injection of saline at the same dose (Figure 1B). Three weeks later, the mice were euthanized under isoflurane anesthesia according to AVMA (American Veterinary Medical Association) Guidelines for the Euthanasia of Animals: 2020 Edition. The liver samples were harvested and preserved in 4% formalin or snap frozen in liquid nitrogen. rTM and r TM-D1 were provided by Asahi Kasei Pharma Corporation (Tokyo, Japan).

Figure 1 HCC and hepatic ischemia-reperfusion (IR) models. (A) Schema of the experimental design of mice with inoculation of Hepa1-6 and treatment of rTM and rTM-D1. (B) Schema of the experimental design of mice with inoculation of Hepa1-6 subjected to IR and treatment of rTM and rTM-D1. Abbreviations: rTM, recombinant thrombomodulin; rTM-D1, recombinant thrombomodulin lectin-like domain; HCC, hepatocellular carcinoma; IR, ischemia/reperfusion; TM-KO, thrombomodulin-knockout.

Functional Assay of TM

Functional activity of TM was assessed based on the activation capacity of protein C. This assay was performed as previously described.³³

Conditioned Medium

Conditioned medium (CM) was harvested from the cultured HCC cells after incubation in serum-free DMEM for 48 h under normoxic (21% O₂) or hypoxic (2% O₂) state. At the end of the incubation period, the medium was stored at −80°C until use.

Migration Assay

For studies of the HCC cell migration, 8-μm-pore size Transwell chambers (Corning, NY, US) were used. In total, 5×10⁵ Huh-7 cells were seeded into the upper chamber with 750 µL of RPMI supplemented with 10% FBS, and cultured for 24 h under normoxic (21% O₂) or hypoxic (2% O₂) state. Under hypoxic conditions, rTM or rTM-D1 was added to the medium, as indicated. After incubation for 24 h, the non-migrating cells on the upper surface of the membrane were removed using a cotton swab. Cells were fixed in 4% paraformaldehyde and stained with propidium iodide solution (Dojindo, Kumamoto, Japan). The migrating cells were counted at 200× magnification in nine adjacent microscopic fields on each membrane.

Adhesion Assay

To evaluate the Huh-7 cell adhesion to hLSEC, 5×10⁵ hLSEC were plated in a 24-well tissue and allowed for 1 h at 37°C in a 5% CO2-humidified incubator under normoxic (21% oxygen) or hypoxic (2% oxygen) condition. The Huh-7 cells were stained with CFSE for 10 min, and 1×10⁵ Huh-7 cells were then added to the wells containing rTM or rTM-D1 at the indicated doses. Following incubation for 24 h at 37°C in 5% CO₂ or 1% O₂ condition, wells were washed with PBS. Adhesive cells were quantified as the number of CFSE-labeled Huh-7 cells at five random HPF at ×20.

Biochemical Assessment

Blood samples were collected from the inferior vena cava. Serum HMGB-1 concentrations were measured using an HMGB-1 EIA kit (Shino-Test, Tokyo, Japan), according to the manufacturer’s instructions.

Histological and Immunohistological Study

For histological analysis, formalin-fixed liver tissue sections were stained with hematoxylin and eosin (H&E) and examined microscopically. Zudan III staining was performed to assess the lipid droplet area. Immunohistology was performed on paraffin-embedded sections using primary antibody; hypoxia-inducing factor (HIF)-alfa (Cell Signaling Technology) was used to detect hypoxia. To detect tissue hypoxia, pimonidazole (Hypoxiprobe-1 60 mg/kg; Natural Pharmacia International, Burlington, MA, USA) was injected intraperitoneally into the mice 1 h before death. Pimonidazole was stained with an anti-pimonidazole mouse IgG1 monoclonal antibody (1:50 dilution).

Statistical Analysis

Data were expressed as average values (standard error [SE]). Differences between the two experimental groups were evaluated using t-tests. Multiple comparison of between the groups were made with a one-way analysis of variance with Turkey’s multiple comparisons. Statistical significance was set at P < 0.05. Statistical analyses were performed using the JMP software version 14 (SAS Institute Japan Inc., Tokyo, Japan).

Results

CDD Induces Liver Steatosis, Hypoxic Microenvironment, and Decreased Function of TM

Feeding CDD for 4 weeks resulted in severe steatosis (60% macrosteatosis) (Figure 2A). Increased lipid accumulation in the hepatocytes was confirmed by Zudan III staining (Supplementary Figure 1). In CDD-fed mice, the accumulation of pimonidazole was enhanced in the liver compared with that in normal diet-fed mice (Figure 2B). Immunohistochemical analyses revealed increased staining for HIF1A in the CDD-fed mice compared to the normal diet-fed mice (Figure 2C). The proteolytic activity of activated protein C, a critical biological function mediated by TM, was significantly decreased to less than 25% in the CDD-fed mice, as compared with the mice with normal livers (Figure 2D).

Figure 2 Choline-deficient diet induces steatosis, a hypoxic state, and dysfunction of TM in the liver. The mice were fed a choline-deficient or normal diet for 4 weeks. (A) H&E, (B) pimonidazole, and (C) HIF1A staining of the sections from the normal and steatotic livers are shown. Original magnification, 100×, scale bar: 200 µm. (D) Bar diagrams showing the functional capacity of activated protein C (% p-nitroaniline production) from the normal livers of the wild-type and TM-KO mice and the steatotic livers of the wild-type and TM-KO mice. Liver weight: 100 mg (n = 5). *P < 0.05, **P < 0.01. Abbreviations: H&E, hematoxylin and eosin; HIF1A, Hypoxia-inducible factor1A; TM-KO, thrombomodulin-knockout.

CDD-Induced Steatotic Liver Has a Permissive Microenvironment for HCC Progression, Which Was Attenuated by rTM Administration

First, we investigated whether steatotic livers created a prometastatic environment in mice. When the mice were sacrificed, the normal diet-fed mice showed sparse liver tumor nodules, whereas the CDD-fed mice showed numerous confluent liver tumor nodules (Figure 3). The liver weight and liver tumor replacement area (LTRA) were significantly higher in steatotic livers than in normal livers (Figure 3). The serum concentration of HMGB-1 3 weeks after inoculation with the Hepa1-6 cells was significantly higher in the CDD-fed mice than in the normal diet-fed mice (Figure 4). Next, the administration of rTM showed a significantly reduced liver tumor burden in mice with steatotic livers, compared to mice without rTM (Figure 3). At the time of sacrifice, the serum concentration of HMGB-1 was significantly higher in mice without rTM administration than in the mice with rTM administration (Figure 4). We assessed the anti-tumor effect of the D1 domain (rTM-D1). Mice treated with r TM-D1 showed sparse liver tumor nodules. rTM-D1 significantly suppressed liver weight and LTRA (Figure 3).

Figure 3 Choline deficient diet-induced fatty liver has a permissive microenvironment for HCC metastasis, and rTM and lectin-like domain (D1) ameliorate HCC progression. The graph shows the liver weights (A) and liver tumor replacement area (tumor volume/liver volume) (B). Results are presented as means (± SE). *P < 0.05, **P < 0.01 (n=15 per group). (C) Representative macrography (upper) and micrography (bottom) of liver tumors in the normal liver, steatotic liver, and steatotic liver treated by rTM, and steatotic livers treated by rTM-D1. Abbreviations: HCC, hepatocellular carcinoma; rTM, recombinant thrombomodulin; SE, standard error; LTRA, liver tumor replacement area.

Figure 4 The serum concentration of HMGB-1 3 weeks after inoculation with the Hepa1-6 cells in the normal diet-fed mice, the CDD-fed mice without rTM, and the CDD-fed mice with rTM. Measurements of the serum HMGB-1 levels 3 weeks after surgery. Results are presented as means (± SE). (n=5 per group) *P < 0.05, **P < 0.01. Abbreviations: rTM, recombinant thrombomodulin; SE, standard error; HMGB-1, high motility group box 1.

IR Promotes HCC Progression in Heterogeneous TM-KO Mice with Steatotic Livers

To assess the direct effect of TM downregulation in the liver on HCC progression, heterogeneous TM-KO mice were generated. The functional productive activity of activated protein C was significantly decreased by approximately 50% in the heterogeneous TM-KO mice compared with the wild-type mice (Figure 2D).

First, we investigated whether the heterogeneous TM-KO mice show accelerated HCC progression. TM-KO mice fed with a normal diet showed sparse liver tumor nodules (data not shown). IR injury promotes HCC progression and liver metastasis after liver surgery, including liver transplantation.^34,35 Thus, we investigated whether IR injury accelerates HCC progression in the heterogeneous TM-KO mice with normal livers. However, the TM-KO mice with normal livers subjected to IR developed liver tumor nodules 3 weeks after inoculation with tumor cells; however, there was no significant difference in tumor burden between the wild-type and TM-KO mice (Supplementary Figure 2).

Next, we investigated whether IR injury accelerated HCC progression in the heterogeneous TM-KO mice with steatotic livers. The TM-KO mice with steatotic livers were partially subjected to partial IR. The heterogeneous TM-KO mice with steatotic livers developed numerous confluent liver tumor nodules (Figure 5). Liver weight and LTRA levels were significantly higher in the TM-KO mice with steatotic livers than in the wild-type mice (Figure 5). Furthermore, in TM-KO mice with steatotic livers, rTM significantly reduced the liver tumor burden compared to the mice without rTM (Figure 5) rTM or r-TM-D1 also significantly reduced liver tumor burden in wild-type mice with steatotic livers subjected to partial IR compared to the mice receiving neither rTM nor rTM-D1 (Figure 5). These results indicate that downregulation of endogenous TM contributes to the progression of HCC in steatotic livers.

Figure 5 Dysfunction of endogenous TM promotes HCC progression in the mice with steatotic livers after IR, and exogenous TM mitigates HCC progression. The graph shows the liver weights (A) and liver tumor replacement area (tumor volume/liver volume) (B). Results are presented as means (± SE). *P < 0.05 (n=10 per group). Representative macrography and micrography of liver tumors in the wild-type and TM-KO mice (C). Abbreviations: TM, thrombomodulin; HCC, hepatocellular carcinoma; IR, ischemia/reperfusion; TM-KO, thrombomoduline-knockout; SE, standard error; LTRA, liver tumor replacement area.

Hypoxia Induces HMGB-1 Release from the HCC Cells and Promotes Migration of the HCC Cells, Which Is Attenuated by rTM

The relation between hypoxia and HCC progression has been identified via HMGB-1.³⁶ We have previously shown that fatty liver induces a hypoxic microenvironment in the liver,³¹ and in this study, we observed an increased concentration of serum HMGB-1 in mice with HCC progression (Figure 4). Therefore, we examined whether hypoxia promotes the migration of the HCC cells and whether HMGB-1 is involved in HCC progression in vitro. Migration assay showed that exposure to 2.0% oxygen for 24 h significantly increased the migrated Hepa1-6 cells, and the addition of rTM to the hypoxic medium decreased the migrated cells (Figure 6A). The concentration of HMGB-1 in the CM was higher after exposure to 2.0% oxygen for 24 h than that after normoxic conditions (Figure 6B). The CM from the cultured HCC cells under hypoxic conditions increased the number of migrated Hepa1-6 cells compared to that under normoxic conditions (Figure 6C), and the addition of rTM to the CM significantly decreased the concentration of HMGB-1 and the number of migrated Hepa1-6 cells (Figure 6B and C). Furthermore, addition of the D1 domain protein to the CM decreased the concentration of HMGB-1 and migrated the Hepa1-6 cells (Figure 6D). The addition of anti-HMGB-1 antibodies to the CM, but not to control antibodies, significantly decreased the migrated Hepa1-6 cells (Figure 6E). Finally, the addition of HMGB-1 to the medium significantly increased the migrated Hepa1-6 cells, and pretreatment with rTM reversed the migratory effect of HMGB-1 (Figure 6F). These results provide evidence that hypoxia induces the release of HMGB-1 from the HCC cells and the increased release of HMGB-1 promotes the migration of the HCC cells. Recombinant TM and D1 domain proteins suppress migration of the HCC cells via absorption of released HMGB-1.

Figure 6 Hypoxia induces HMGB-1 release from the HCC cells and promotes migration of the HCC cells, which is attenuated by rTM. (A) Migration in the Hepa1-6 cells. Hypoxia (2% oxygen, 24 h) induces tumor migration in the Hepa1-6 cells and rTM attenuates tumor migration. Images are taken using a microscope. Results are presented as means (± SE). *P < 0.05 (n=6 per group). (B) The concentration of HMGB-1 in the conditioned media harvested from the Hepa1-6 cells. Hypoxia (2% oxygen, 24 h) induces the secretion of HMGB-1 from the Hepa1-6 cells and rTM attenuates the secretion of HMGB-1. Results are presented as means (± SE). *P < 0.05 (n=5 per group), **P < 0.01 (n=5 per group). (C) Migration of the Hepa1-6 cells in addition of the conditioned media harvested from the Hepa1-6 cells under hypoxia (2% O2, 24 h). Images are taken using a microscope. Results are presented as means (± SE). *P < 0.05 (n=6 per group). (D) Migration in the Hepa1-6 cells. Hypoxia (2% oxygen, 24 h) induces tumor migration in the Hepa1-6 cells and rTM-D1 attenuates tumor migration. Images are taken using a microscope. Results are presented as means (± SE). *P < 0.05 (n=6 per group). (E) Migration in the Hepa1-6 cells. Hypoxia (2% oxygen, 24 h) induces tumor migration in the Hepa1-6 cells and the addition of anti-HMGB-1 antibody (upper) attenuates tumor migration, but control antibody (lower) did not. Images are taken using a microscope. Results are presented as means (± SE). *P < 0.05 (n=6 per group). (F) Migration in the Hepa1-6 cells. The addition of HMGB-1 induced tumor migration in the Hepa1-6 cells and rTM attenuates tumor migration. Images are taken using a microscope. Results are presented as means (± SE). *P < 0.05 (n=6 per group). Abbreviations: rTM, recombinant thrombomodulin; SE, standard error; HMGB-1, high-motility group box 1; HCC, hepatocellular carcinoma.

Hypoxia Promotes the Adhesion of the HCC Cells to hLSEC, Which Is Attenuated by rTM

Hypoxia induces adhesion of the HCC cells to endothelial cells.²⁹ Adhesion assay showed that exposure to 2.0% oxygen for 24 h significantly increased the adhesion of Hepa1-6 cells to hLSECs, and the addition of rTM to the hypoxic medium, dose dependently, decreased the adhered cells (Supplementary Figure 3A). Moreover, addition of rTM-D1 decreased the number of adhered cells (Supplementary Figure 3B).

Discussion

A rapidly growing body of literature indicates that MASLD, including MASH, is associated with accelerated liver metastasis, especially intrahepatic liver metastasis, in liver cancer.^4–10 However, the mechanisms underlying the hepatic steatosis-induced promotion of liver cancer remain unclear. In the present study, we observed endogenous TM dysfunction in livers with CDD-induced steatosis. We demonstrated that CDD-induced hepatic steatosis accelerates the progression of liver cancer, and rTM and r TM-D1 suppressed the progression of liver cancer. We also revealed that the serum concentration of HMGB-1 was increased in mice with massive liver cancer and that the concentration of HMGB-1 was decreased by treatment with rTM. Furthermore, we showed that IR exacerbates HCC progression in the TM-KO mice with steatotic livers. In vitro assays showed that HMGB-1, which is released from the cells under hypoxia, enhanced the migration of the HCC cells, and rTM and r TM-D1 suppressed the migration of the HCC cells through the absorption of HMGB-1. These results indicate that the steatotic liver accelerates the progression of the HCC cells and that both exogenous and endogenous TM regulate HCC progression in the steatotic liver through the binding of HMGB-1.

Our previous study showed that a steatotic liver microenvironment favors HCC progression and metastasis through the activation of the hepatic stellate cells. However, inhibiting the activation of the hepatic stellate using a Rho inhibitor did not suppress the progression of HCC in the steatotic livers.⁷ These results suggest that hepatic stellate cell activation is probably one of several mechanisms associated with the tumorigenic environment in the steatotic livers, and that the key mechanisms underlying steatosis-promoting HCC progression remain unclear. Endogenous TM expression in the cancer cells is associated with cancer prognosis. In HCC, decreased TM expression increases postoperative recurrence.¹⁸ In pancreatic cancer, increased TM expression is a favorable factor for survival after surgical resection.¹⁹ However, the function of TM in noncancerous liver tissues remains unknown. The TM-KO mice with steatotic livers exhibited HCC progression after IR in our study. This is the first study to show that the expression of TM in non-cancerous liver tissues is involved in the cancer microenvironment and that the decreased functional activities of TM contribute to HCC progression.

Our results showed that TM protects against HCC progression by blocking HMGB-1 in the D1 domain. HMGB-1 promotes both cancer cell survival and death by regulating multiple cancer signaling pathways.¹⁷ Intracellular HMGB-1 has anti-tumor effects, such as acting as a tumor suppressor by binding to Rb, a tumor suppressor protein, and increasing genome stability; HMGB-1-mediated DNA damage repair contributes to genome stability and increased autophagy, and loss of HMGB-1 leading to autophagy deficiency causes genome instability and inflammation.^17,37,38 However, under hypoxic and inflammatory conditions, HMGB-1 is passively and actively secreted by the injured and inflamed cells. Extracellular HMGB-1 promotes tumor invasion and metastasis via NF-kB, ERK, and mitogen-activated protein kinase (MAPK) pathways.^39,40 In the experiment in which the HCC cells were injected into the spleen, we observed that the serum levels of HMGB-1 were significantly increased in the mice with steatotic livers, which showed numerous and confluent liver tumor nodules, compared to the mice with non-steatotic livers. We also observed that the Hepa1-6 cells secreted HMGB-1 under hypoxia, and that HMGB-1 promoted the migration of the HCC cells. These results indicate that extracellular HMGB-1 promotes HCC cell invasion and metastasis, although it remains unclear which intracellular signaling pathways contribute to HCC progression. The D2 domain is a cofactor for thrombin binding, which mediates protein C activation and exhibits anticoagulant activity. However, Ito et al reported that TM not only binds to HMGB-1 through the N-terminal D1 domain but also aids the proteolytic cleavage of HMGB-1 by thrombin through the D2 domain.⁴¹ TM may protect against HCC progression by inhibiting the action of HMGB-1 through the D1 and D2 domains.

We showed that hypoxia leads to increased adhesion of the HCC cells to hLSEC and that TM protects this adhesion. Li et al demonstrated that hypoxia contributes to endothelial migration of the HCC cells via the vascular cell adhesion molecules-1 (VCAM-1)/integrin B1 axis.²⁹ Recently, it was reported that the TM-KO endothelial cells enhance the adhesion of leukocytes through the increased expression of VCAM-1 and endothelial selectin.⁴² Furthermore, the TM-D1 domain can bind extracellular matrix proteins such as fibronectin and collagen and inhibit integrin-dependent binding of human breast cancer cells to fibronectin.⁴³ Thus, TM may protect the HCC cells from adhering to hLSEC via the D1 domain. However, further studies are required to elucidate these mechanisms.

Numerous studies have reported that the IR injury promotes liver cancer progression, although this is not uniform.^44,45 IR injury promotes the recurrence of HCC in fatty livers via the arachidonate-12-lipoxygenase /hydroxyeicosatetraenoic acid pathway, but not in non-steatotic livers.⁴⁶ Aging and steatosis promote IR-induced colorectal liver metastasis.⁴⁴ Ischemia aggravates HCC recurrence via the lipopolysaccharide (LPS)-toll-like receptor 4 (TLR4) pathway.⁴⁷ TLR4 binding of circulating LPS and danger-associated (DAMP) such as HMGB-1 leads to the activation of MAPK and NF-kB signaling and upregulation of various inflammatory cytokines and adhesion molecules.⁴⁸ TM absorbs HMGB-1 and LPS to reduce inflammation. We previously showed that rTM suppresses IR in the non-steatotic and steatotic livers through the TM-D1 domain binding of HMGB-1.³¹ Our current results indicate that IR promotes the progression of HCC in the TM-KO mice, which have a low degradation capacity for HMGB-1 due to decreased expression of TM.

Recently, it has been reported that rTM may be a promising drug for pancreatic cancer.^23,24 In an orthotopic pancreatic cancer implantation model, rTM suppressed tumor growth in pancreatic cancer by blocking thrombin-induced NF-kB activation.²⁴ In a liver metastasis model of pancreatic cancer, rTM suppressed NET-induced liver metastasis by inhibiting the generation of NETs and degrading HMGB-1.²³ We have previously shown that recombinant TM suppresses IR in non-steatotic and steatotic livers through TM-D1 domain binding of HMGB-1,³¹ and that rTM suppresses steatosis-induced HCC progression. TM is used clinically as an anticoagulant in Japan. In the setting of liver resection for liver cancer in steatotic livers, rTM may be a promising candidate for inhibiting not only the IR in steatotic livers but also liver cancer recurrence.

There is a limitation to this study. In this study, we used a CDD-induced steatotic liver model, which developed steatotsis but differs from the complexity of human metabolic dysfunction-associated steatotic liver disease (MASLD). This difference may affect the progression of HCC.

In conclusion, steatotic liver reduces the expression of TM and accelerated the progression of HCC by decreasing the capacity of TM to absorb HMGB-1. rTM may be a useful modality to prevent HCC progression in steatotic livers during hepatic surgery.