Trim21 deficiency in mice increases HCC carcinogenesis in a NASH context and is associated with immune checkpoint upregulation

The global pandemic of metabolic diseases has increased the incidence of hepatocellular carcinoma (HCC) in the context of non‐alcoholic steatohepatitis (NASH). The downregulation of the E3 ubiquitin ligase TRIM21 has been linked to poor prognosis in different cancers including HCC. In order to investigate the role of TRIM21 in liver cancer progression on NASH, Trim21+/+ and Trim21−/− male mice were injected with streptozotocin at the neonatal stage. The hypoinsulinemic mice were then fed with a high‐fat high‐cholesterol diet (HFHCD) for 4, 8 or 12 weeks. All mice developed NASH which systematically resulted in HCC progression. Interestingly, compared to the Trim21+/+ control mice, liver damage was worsened in Trim21−/− mice, with more HCC nodules found after 12 weeks on HFHCD. Immune population analysis in the spleen and liver revealed a higher proportion of CD4+PD‐1+ and CD8+PD‐1+ T cells in Trim21−/− mice. The liver and HCC tumors of Trim21−/− mice also exhibited an increase in the number of PD‐L1+ and CD68+ PD‐L1+ cells. Thus, TRIM21 limits the emergence of HCC nodules in mice with NASH by potentially restricting the expression of PD‐1 in lymphocytes and PD‐L1 in tumors.

non-alcoholic steatohepatitis leading to liver cancer in an hypoinsulinemic mouse model with and without TRIM21 expression.TRIM21 deficiency aggravated the development of hepatocellular carcinoma, which correlated with increased expression of the immune checkpoints PD-1 and PD-L1 in the spleen and liver and in hepatocellular carcinoma tumors.The findings suggest that TRIM21 may promote antitumor surveillance via the downregulation of inhibitory immune checkpoints.

| INTRODUCTION
Hepatocellular carcinoma (HCC) is the 6th most common cancer and the 3rd cause of cancer-related deaths worldwide. 1While cancer incidence and death-related cancer are decreasing in the world, the incidence of HCC has not diminished. 2The majority of HCC occurs in an inflammatory context induced by recurrent liver insults such as alcohol overconsumption, viral hepatitis infections or metabolic syndrome. 1obally, liver lesions induced by hepatitis B (HBV) or C (HCV) viruses account for the majority of HCC cases in some countries.However, with treatments available, the number of HCC cases with viral etiology has steadily decreased. 1Today, diabetes and obesity pandemics are on the rise, and with it, nonalcoholic fatty liver diseases (NAFLD).Ranging from simple steatosis to nonalcoholic steatohepatitis (NASH), NAFLD is becoming the main contributor to liver diseases, and by 2030 will be the major determinant of HCC. 3 Accumulation of lipids in hepatocytes has been shown to lead to lipotoxicity, metabolic reprogramming, and over time, to NASH.Moreover, chronic inflammation induced by lipid overload favors tumorigenesis in the liver. 4e E3 ubiquitin ligase tripartite motif-containing protein 21 (TRIM21) is a member of the structurally diverse tripartite motif protein family implicated in a broad range of biological processes. 5so known as Ro-52, TRIM21 was first identified as an important auto-antigen in auto-immune diseases such as the Sjögren syndrome and systemic lupus erythematosus. 6Previous studies reported that TRIM21 is induced by type I and II interferons (IFNs).Besides, TRIM21 has been shown to negatively regulate NF-κB activation in murine embryonic fibroblasts, resulting in a reduction of proinflammatory cytokine induction. 7TRIM21 also down-regulates IL-23-dependent Th17 differentiation and suppresses IFN-mediated immune activation by a negative-feedback loop. 8These different data suggest that TRIM21 is a negative regulator of the inflammatory response.Aside from its role in inflammation and immunity, studies have shown that TRIM21 mediates the degradation of two metabolic enzymes, the glyceronephosphate O-acyltransferase and the fatty acid synthetase, inhibiting lipid metabolism, lipid toxicity and thus limiting HCC progression. 9Besides, previous studies also found that patients expressing low Trim21 mRNA in their HCC tumors had the worst survival prognosis, suggesting that TRIM21 is a tumor suppressor. 10The authors hinted at the probable implication of this protein in cell death resistance and in the inflammatory response.However, another group found conflicting results.Thus, Trim21 À/À mice were protected from hepatocarcinogenesis in a chemically induced murine model. 11It has been hypothesized that TRIM21 disrupts the p62-Keap1-Nrf2 pathway by blocking p62 oligomerization, leading to the downregulation of the antioxidant response, a pathway implicated in carcinogenesis. 11merous mice models have been developed to study HCC.None can exactly replicate human disease.However, mice models developing spontaneous tumors after chronic lesions tend to best mimic the human pathophysiology. 12It is noteworthy that the model used in this study based on type 1 diabetic mice subjected to a high-fat diet (HFD) 13,14 led to the rapid development of HCC without inducing obesity.These tumors, which develop in a context of liver inflammation and fibrosis, have been shown to best recapitulate molecular characteristics of the human HCC with similar oncogenic mutations altering the Wnt, cell cycle and chromatin-modification pathways. 15Male mouse pups are injected with streptozotocin (STZ) to induce hypoinsulinemic diabetes.After weaning, diabetic mice are fed HFD.After 8 weeks, mice developed liver adenomas, and after 12 weeks, HCC. 13 Our group had previously shown, using this model, how HCC emergence varies with diet composition, using a high-fat high-cholesterol diet (HFHCD) or a high-fat high-sugar diet (HFHSD).Thus, HFHCD triggered higher immune cell recruitment in the liver, promoting inflammatory response and anti-tumor immune surveillance. 14ysiologically, the liver plays an important role in immunological surveillance, as both portal vein and arterial blood pass through.In this context, the liver has to harbor a tolerogenic microenvironment to avoid triggering a response against exogenous non-pathogenic compounds such as gut-derived products.Thus, the immune system is tightly regulated by immune-activating and suppressing cells. 16In parallel, the immune system has been identified over the last decade as a major player in HCC progression. 16However, the intrahepatic immune environment is affected differentially according to the HCC etiology, highlighting the importance of better deciphering the specific-etiological underlying mechanisms. 16Nonetheless, the intrahepatic immunosuppressive cells are often implicated in HCC progression, such as the resident macrophages (Kupffer cells), or the regulatory T cells. 16 the last decade, immune checkpoint inhibitors (ICIs) are becoming a promising therapeutic solution in immunosuppressive tumors, and understanding the underlying mechanisms to modulate immune checkpoints are a major challenge.Due to its involvement in many cancer-specific pathways, 5 TRIM21 is attracting interest.As described above, its role in the immune system has been studied, however, no evidence has been found yet for its implication in the anti-tumoral response in HCC progression.In the present work, we compared the pathophysiological evolution of NASH leading to liver cancer in a hypoinsulinemic context in mice expressing or not TRIM21.Challenged mice were monitored for up to 12 weeks to evaluate inflammation, NASH severity and HCC progression.

| Biochemical parameters
After weaning and slaughtering, blood sugar was systematically measured using a glucometer (Freestyle Optium Nep, Abbott) to confirm the diabetic status of the mice included in the experimental protocol.
Only mice with glucose levels higher than 300 mg/dl were considered diabetic.Levels of plasma alanine aminotransferase (ALT) were determined according to the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) primary reference procedures using the Olympus AU2700 Chemistry Analyser (Olympus Optical, Tokyo, Japan).

| Quantitative real-time PCR
Total RNA was extracted from frozen liver tissues using the NucleoSpin RNA kit (Macherey-Nagel, #740955).First-strand cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, #4368813, USA).Real-time quantitative PCR was performed using the double-strand specific SYBR Green system (Applied Biosystems, #4367659) on CFX384 Touch Real-Time PCR Detection System (Biorad).Each measurement was performed in triplicate.The relative gene expression was normalized against the 18S gene expression.Samples from diabetic SD-fed mice were used as a reference for mRNA expression (the control mRNA level was arbitrarily set at 1).Primer sequences can be requested from the authors.

| Histology and immunochemistry
Liver and spleen fragments were fixed in 4% paraformaldehyde for 24 h at room temperature before being embedded in paraffin.For the liver, 4 μm sections were used for hematoxylin-and-eosin (H&E), Sirius Red (SR), Reticulin (MICROM, F/010211) or immunochemistry staining.For fresh frozen sections, samples were embedded in OCT embedding Medium (M1-Embedding matrix # Epredia 1310, USA).
Seven μm frozen sections were used for Oil Red O (ORO) staining with a hematoxylin counterstain.High-resolution images (Â20) were used to study H&E and glutamine synthetase (GS) stained sections.
The same pathologist blindly scored all livers for NAFLD activity score (NAS) and quantified the number of nodular lesions.Tissue sections from paraffin blocks were dried for 1 h at 58 C, followed by antigen retrieval and incubation with chosen primary antibody in an automated staining platform (Discovery XT, Roche).
Histology analysis was performed using the HALO software (Indica Labs, NM) (CD45; CD68; SR and ORO).Multiplex IHF image analysis was performed using the open-source image analysis Qupath software (Qupath v0.3.2) 17 and ImageJ/Fiji. 18r image analysis, we followed a specific workflow: Positive cells quantifications on Qupath: 1. Automatic cell segmentation using a StarDist script: 20,21 whole liver image nuclei were recognized by an algorithm trained to recognize different types of cell nuclei.From the nuclei segmentation, the cell perimeter was defined for every liver image.

Training classifier individually for each marker:
Using the "Train a classifier" function from Qupath, single-channel liver images were used to train a classifier to recognize positive marked cells and negative marked cells.This method was used for every single channel (marker): CD4 + , CD8 + , PD-1 + , PD-L1 + , FOXP3 + , CD3 + and CD68 + .

Apply classifier on all images:
The trained classifier was then used in all multichannel liver images to count the number of positive cells for the different markers.

| Spleen immune cell analysis by flow cytometry
Mouse spleens were crushed on a 40 μm cell strainer (BD Falcon).

| Statistics
Results are expressed as the means ± SEM for each group of mice.

| Trim21 helps to protect the liver from damage during NASH
To assess the role of TRIM21 during HCC development in a NASH background, diabetic Trim21 À/À mice and their Trim21 +/+ littermate controls were subjected to HFHCD for 4, 8 or 12 weeks, allowing a follow-up of the disease over time.This protocol has already been described to induce steatosis and chronic hepatitis leading to the subsequent development of HCC. 13,14The different clinical manifestations were investigated.Although the mice did not lose or gain weight during the course of the liver pathology (Figure 1A), hepatomegaly, a symptom expected in fatty liver disease progression, 22 was found from week 4 with progressive worsening over time in all mice regardless of genotype (Figure 1B).To monitor liver damage, plasma alanine aminotransferase (ALT) levels, which correlate with hepatic cell mortality, 23 were measured (Figure 1C).While diabetic control mice under standard diet (STZ + SD) showed normal transaminase levels, high ALT levels were detected from 4 weeks in the plasma of both Trim21 +/+ and Trim21 À/À mice under HFHCD.Only a significant difference was distinguished at the late time of 12 weeks, which reflected greater liver damage in Trim21 À/À mice.
Steatosis creates a lipotoxic environment and participates in hepatocyte death.Moreover, the chronicity of liver damage during NASH promotes the establishment of fibrosis. 23These clinical parameters were therefore compared between the two murine genotypes throughout the pathology.While no steatosis was revealed in diabetic control mice, hematoxylin and eosin staining (HES) revealed the appearance of macrovesicular steatosis as early as 4 weeks under HFHCD (data not shown).To complete this analysis, the NAFLD activity score (NAS) was determined on liver tissue samples.This histological scoring system integrates the status of macrovesicular steatosis, hepatocellular ballooning and lobular inflammation.While NAS scores were typical of NASH-affected livers in HFHCD mice, no differences were detectable between Trim21 À/À mice and their Trim21 +/+ littermate controls (Figure 1D).Similarly, analysis of the oil red O staining of liver sections did not reveal differences in their steatosis status (Figure 1E).Both approaches, Sirius Red staining (Figure 1F) and mRNA quantification of Col1a1 and Tgfbi (Figure 1G) showed the establishment of severe hepatic fibrosis within both genotypes without detecting any significant difference between them.These data indicated that TRIM21 did not influence the severity of either steatosis or fibrosis in the liver of diabetic mice on HFHCD but did alter the late-stage hepatocyte death.
lesions had histological characteristics of tumors such as the absence of portal tracts, broad trabecular growth, as well as atypical nuclei (Figure 2D).Their enumeration at the microscopic scale on these HES-stained sections showed that more tumors dotted the livers of Trim21 À/À mice at 12 weeks (Figure 2E) while no tumors were found in control diabetic mice under SD (Figure 2C).The analysis of different mRNA markers of HCC (Ccnd1, Igf2, Gpc3), oncofetal (Afp) and cancer stem cells (Klf4, Aldh1, Cd133) allowed to distinguish the tumor tissue (T) from the peri-tumoral tissue (NT) with a significantly greater overexpression of Igf2, Gpc3, Afp, Klf4 and Aldh1 in the tumor tissue exclusively in the livers from Trim21 À/À mice (Figure 3A).Furthermore, tumors in Trim21 À/À mice expressed higher levels of these markers compared to tumors found in Trim21 +/+ mice.
At the protein level, the labeling of the glutamine synthetase (GS) and the reticulin helped to further characterize tumors (Figure 2C).Thus, tumors were either positive or negative for GS (Figure 3B).In the literature, it has been described that the tumoral upregulation of the GS marker correlates with the mutation and the constitutive activation of the β-catenin gene, 25 a hallmark of tumorigenesis in some HCCs. 3 However, in humans, NASH-HCC tumors tend to be GS-negative. 26,27Consistently, more GS-negative tumors were systematically found in the livers of mice engaged in the used NASH-HCC model (Figure 3C).Interestingly, at 12 weeks under HFHCD, Trim21 À/À diabetic mice developed more GS-negative tumors than their Trim21 +/+ littermates (Figure 3C).
Altogether, these findings suggest that the deficiency of TRIM21 promotes the emergence of more tumors in the liver of diabetic mice under HFHCD.

| Trim21 regulates the splenic and hepatic immune environment
The immune response, which could be influenced by the diet, plays a pivotal role during the disease course. 14As TRIM21 plays an important role in the immune system, it prompted us to analyze the effect of TRIM21 knockout on the liver environment during the pathology evolution.First, liver inflammation was evaluated by measuring mRNA levels of inflammatory cytokines.An important increase in mRNA levels was detected for TNFα and IL-6 from weeks 4 to 12 in both genotypes (Figure 3D).It should be noted that at week 4, the mRNA levels of IFNγ, TNFα and IL-6 were significantly higher in the livers of Trim21 À/À mice compared to those of Trim21 +/+ under HFHCD.Systemic inflammation was confirmed with high levels of TNFα and IL-6 at week 12 in the plasma of both genotypes (data not shown).At week 4, the level of TNFα was significantly higher in the plasma of Trim21 À/À mice than in those of Trim21 +/+ under HFHCD (data not shown).This difference in pro-inflammatory cytokine expression at week 4 may indicate a more favorable microenvironment to initiate tumor development in the liver of Trim21 À/À mice under HFHCD.Furthermore, immunostaining of CD45, a marker of leucocytes, revealed a progressive infiltration in the livers for both genotypes (Figure 3E).At 12 weeks of HFHCD, a drop of hepatic CD45 + cells occurred only in the livers of Trim21 À/À mice.Similar profiles were found for CD68, a marker of macrophages (Figure 3F).
The monitoring of the spleen index (SBW, spleen to body weight ratio) revealed the gradual development of splenomegaly, which may suggest activation of the systemic immune response in both Trim21 +/+ and Trim21 À/À mice.Nonetheless, this spleen index appeared significantly lower in Trim21 À/À mice even before their inclusion in the HFHCD.This difference persisted throughout the protocol (Figure 4A).An immunohistochemistry approach was used to study the different populations of immune cells inside spleens (Figure 4B,C).The analysis of the number of B (B220 + ) and T (CD3 + ) cells showed similar evolutions within both Trim21 +/+ and Trim21 À/À mice during the treatment, with nevertheless a tendency to a greater increase in the number of T cells after 12 weeks of HFHCD in the spleen of Trim21 À/À mice (Figure 4B, right graph).More in-depth investigations on the subpopulations of T lymphocytes only revealed a significant upregulation of the number of T helper cells (CD4 + ) in Trim21 À/À mice at 12 weeks of HFHCD (Figure 4D,E).Indeed, no significant difference between the tested genotypes was detected in the number of cytotoxic T cells (CD8 + ), regulatory T cells (CD4 + FOXP3 + cells) or Ki67 + cells (Figure 4D; data not shown).
Interestingly, TRIM21 has been recently described to interact with PD-L1 in lung adenocarcinomas. 28,29Its high expression level induces the downregulation of PD-L1 after its ubiquitinylation. 29This immune checkpoint plays an important role in anti-tumor immunity, and its expression highly correlates with its PD-1 receptor in different cancers. 30,31Hence, the proportion of T cells expressing PD-1 in the spleen was investigated.Actually, at 12 weeks of HFHCD, the proportion of PD-1 + cells in the T CD4 + and T CD8 + lymphocyte populations were higher in the spleen of Trim21 À/À mice (Figure 4F).These significant differences within the populations of helper and cytotoxic F I G U R E 2 More hepatic tumors emerged in Trim21 À/À diabetic mice under a high-fat high-cholesterol diet.Diabetic (streptozotocin, STZ) male mice were fed either a standard diet (SD) or a high-fat high-cholesterol diet (HFHCD) for 4, 8 or 12 weeks.T lymphocytes were also found in the liver (Figure 5A,B).No differences were found between genotypes in regulatory T cells (CD4 + FOXP3 + cells) numbers (data not shown).In parallel, the study of cells expressing PD-L1 led to the discovery that more PD-L1 + cells were present in the liver of Trim21 À/À mice at 12 weeks of HFHCD (Figure 5C,D).This was confirmed in particular for intrahepatic macrophages (CD68 + PD-L1 + cells) (Figure 5D).By adopting the same type of analysis but this time inside hepatic tumors, more PD-L1 + and CD68 + PD-L1 + macrophages were also interestingly observed in Trim21 À/À mice (Figure 6A,B).Furthermore, it appeared that more infiltrated-CD3 + cells were present in tumors found in the liver of Trim21 À/À mice, with higher proportions of CD4 + and CD4 + PD-1 + cells (Figure 6C).Conversely, the proportion of intra-tumoral CD8 + cells tended to be decreased in Trim21 À/À mice (Figure 6C).Similar analysis only in GS-negative tumors revealed higher significant infiltration of CD4 + and CD4 + PD-1 + cells in Trim21 À/À tumors (Figure 6D).
Our findings show that diabetic mice on HFHCD, the lack of TRIM21 would directly avoid the downregulation of PD-L1 and indirectly of PD-1, explaining the higher observed levels of CD4 + PD-1 + cells and PD-L1 + cells in the liver of Trim21 À/À mice, and particularly more CD68 + /PD-L1 + macrophages described as tumor promoters.Since cell surface expression of PD-L1 has been shown to inactivate PD1 + T cells, the anti-tumor immune response is most likely impaired in Trim21 À/À mice, thus promoting the development of more HCC tumors.

| DISCUSSION
Through its involvement in immunity and inflammation 7,8 the E3-ubiquitin ligase TRIM21 has been suggested to be a negative regulator of inflammation.Studies have also shown that the loss of TRIM21 induced resistance to different programmed cell death such as apoptosis, 32 necroptosis, 33 pyroptosis, 34 and ferroptosis, 35 a change that could for example confer a selective advantage to cancer cells.Furthermore, its implication in metabolic pathways 9 or cell proliferation 5 prompted investigations on the role of TRIM21 in cancer.Thus, the downregulation of this E3-ubiquitin ligase correlates with a worse survival prognostic in breast cancer, 36 colitisassociated cancer, 37 diffuse large B-cell lymphoma, 38 or HCC. 9,10However, in other studies, conflicting results have been found.For instance, TRIM21 overexpression has been described as pro-oncogenic in HCC. 39,40These contradictions could be nonetheless explained by differences according to HCC etiology.
To understand the role of TRIM21 in HCC with NASH etiology, we subjected diabetic Trim21 À/À mice and their Trim21 +/+ littermate controls to an HFHCD for up to 12 weeks.This protocol, described previously, 13,14 follows the pathological pattern observed in patients during NASH progression, namely from steatosis to steatohepatitis, with the onset of fibrosis and finally HCC, without reproducing the obesity disorder.Nonetheless, the inter-and intra-tumoral heterogeneity is recapitulated, 15 thus making this model the one that best reproduces human pathology regarding the molecular characteristics of human HCC. 15Besides, liver tumors found in NASH patients have been classified as poorly proliferating tumors. 41Accordingly, Ki67 immunolabeling led to very weak or no staining inside or outside liver tumors of mice included in our experimental protocol (data not shown).
Therefore, this mouse model of HCC induced by DEN and PB shows more similarities with human HCC of alcoholic etiology. 46These data again clearly illustrate the divergence of mechanisms involved in the onset of HCC depending on the etiology.
In patients with NASH, a higher proportion of tumors are negative for the glutamine synthetase marker (GS). 26,27This feature was found in diabetic mice fed the HFHCD, both in Trim21 À/À and Trim21 +/+ .However, while the number of GS-positive tumors remained similar between the two genotypes over time, a greater number of GS-negative tumors emerged in the liver of Trim21 À/À mice after 12 weeks of HFHCD.Besides the fact that GS-positive tumors are associated with mutations in the gene encoding β-catenin, 25 GS-negative tumors usually exhibit more tumorinfiltrating lymphocytes (TILs). 47Nonetheless, tumors rich in TILs have been associated, with some cancers, such as those affecting the breast, with high expression levels of PD-1 and PD-L1 and a poor prognosis, but could potentially respond better to immune checkpoint inhibitors (ICIs). 31PD-1 is mainly expressed by exhausted lymphocytes and is representative of inactive anti-tumor immunity.Cancer cells acquire the ability to express PD-L1 on their surface, thus inhibiting the activity of T lymphocytes expressing PD-1.Interestingly, TRIM21 has been described to interact with PD-L1 and its high expression level induces the downregulation of PD-L1 after its ubiquitinylation. 29After 12 weeks of HFHCD, Trim21 À/À mice displayed higher proportions of PD-1 + T cells in the spleen and the liver than Trim21 +/+ mice.This was associated with a higher number of PD-L1 + cells and a higher proportion of PD-L1 + macrophages (CD68 + PD-L1 + , tumor-promoting macrophages) in the liver.These same features were also found within the tumors themselves.Further characterizations of the tumoral immune infiltrate also revealed differences.Thus, in TRIM21 À/À mice, a greater proportion of CD4 + T cells expressed PD-1, especially in GS-negative tumors.Conversely, fewer CD8 + T lymphocytes were found, whereas the expression of PD-1 by these cells did not appear to be modified.Hence, the high expression of PD-1 + and PD-L1 + and the decrease of effector CD8 + T cell infiltration in Trim21 À/À might enhance tumor emergence.It's worth noting that other immune cells such as dendritic cells, macrophages and others could play a role in this tumorigenesis process and will be the subject of further studies.
The immune system plays an important role in promoting or suppressing cancer progression.In recent years, various studies have shown that antitumor surveillance is impaired in NASH etiology via dysfunctions affecting CD8 + T cells. 16,48,49While some works have reported the efficacy of anti-PD-L1 treatment in slowing the progression of HCC in genetically modified MUP-uPA mice fed with HFD, 50 others have claimed the ineffectiveness of immune checkpoint inhibitor (ICI)-based treatments (anti-PD-1) both in NASH-HCC HFD-fed mouse model and in humans. 51Besides, a recent report confirmed the involvement of defective CD8 + T cells in the onset of HCC and the ineffectiveness of ICI in different mice models.The authors used mice fed with different HFD types and then injected tumor cells into the mice's livers to study the effect of ICI treatment on NASH-HCC.They found that the defective motility of CD8 + T lymphocytes would render ICI ineffective in NASH-HCC due to the inability of active CD8+ T cells to infiltrate tumors. 52In the diabetic mice model of HCC progression on NASH, a correlation was evidenced between a 23-gene signature of fibrosis and CD8 + localization in the tumor.An anti-TGFβ treatment (1D11) could facilitate the redistribution of CD8 + lymphocytes into tumors. 53However, in our study, we did not observe differences in fibrosis between both genotypes up to 12 weeks.Longer-term studies, up to 16 weeks, would be needed to address this point.
In sum, our data showed that the percentage of CD8+ T cells was lowered in Trim21 À/À tumors while the proportion of CD68 + /PD-L1 + tumor-promoting macrophages was higher suggesting that TRIM21 contributes to limit the occurrence of HCC during the NASH disease in this particular mouse model, probably by promoting antitumor surveillance via the downregulation of inhibitory immune checkpoints.
for 30 min with the LIVE/DEAD fixable yellow stain (#L34959, Life Technologies) to later exclude dead cells during analysis.Cells were also incubated with an anti-CD16/CD32 antibody (#2.4G2, BD Pharmingen) to block non-specific binding.Cells were then labeled with the appropriate fluorochrome-conjugated antibodies/reagents according to the manufacturer's instructions.Stained cells were analyzed on an LSR X-20 Fortessa flow cytometer (BD Biosciences) and data were analyzed using Flowlogic (MACS Miltenyi Biotec).Doublets and dead F I G U R E 1 Legend on next page.cells were excluded based on forward/side scatter and LIVE/DEAD labeling, respectively.The immuno-phenotyping used was as follows: T lymphocytes: CD3 + (BD Pharmingen #553066)/NK1.1 (BioLegend #562921)-on which CD4 + (BD Pharmingen #565974)/PD-1 + and Trim21 +/+ diabetic mice under a high-fat high-cholesterol diet developed the same levels of steatosis or fibrosis.Diabetic (streptozotocin, STZ) male mice were fed either a standard diet (SD) or a high-fat high-cholesterol diet (HFHCD) for 4, 8 or 12 weeks.(A) Body weight.(B) Liver to body weight (LBW) ratio.(C) Plasma alanine transaminase (ALT) concentrations (IU/L).(D) NAS score (evaluation of hepatocyte ballooning, inflammation and steatosis), and Hematoxylin and eosin (H&E) staining of liver sections.Scale bars: 50 μm, original magnification Â40.(E) Signal quantification of Oil-Red O-stained liver tissues and Oil Red O (ORO) staining of liver sections.Scale bars: 50 μm, original magnification Â40.

F
I G U R E 3 Legend on next page.