Galectin-9 expression predicts poor prognosis in hepatitis B virus-associated hepatocellular carcinoma

Introduction

With more than 900 thousand new cases and 830 thousand deaths in 2020, liver cancer ranked as the 6^th most commonly diagnosed malignancies and third cancer-related cause of death worldwide [1]. As the main pathological type of liver cancer, hepatocellular carcinoma (HCC) accounts for 75-85% of all cases and chronic hepatitis B virus (HBV) infection is one major risk for HCC [2]. Across the world, more than 250 million people are HBV carriers, and new-born children from HBV-infected mothers also become chronic carriers, they are destined to develop HCC [3]. Thus, it is essential to identify prognostic biomarkers in HBV-associated HCC.

Galectins belong to the lectin family that bind β-galactoside through highly conserved carbohydrate recognition domains (CRDs). Galectin-9 (Gal-9) is the subgroup of tandem repeat-type galectins, since it contains 2 conserved CRDs [4, 5]. So far, Gal-9 has been identified as the ligand of several receptors, such as T cell immunoglobulin- and mucin-domain-containing molecule (Tim-3), CD44 and Dectin 1. Gal-9 could inhibit immune response through different mechanisms [6]. For example, Gal-9 could induce Th1 cell death through Tim-3, thus inhibiting Th1 immunity [7]. Gal-9 could increase iTreg cell stability and function through the interaction with CD44 [8]. And Gal-9 could induce macrophage mediated adaptive immune suppression through its binding to Dectin 1 [9].

Recent years, Gal-9 has been found to have prognostic value in certain cancer types. For example, in pancreatic cancer, Sun et al. reported that higher Gal-9 expression was associated with better disease-free survival (DFS) and overall survival (OS) [10]. For breast cancer, Irie et al. found that patients with Gal-9⁺ tumors had a more favorable DFS than patients with Gal-9^- tumors, and lower Gal-9 expression could predict a relatively higher risk of metastasis [11]. In ovarian cancer, Schulz. et al. found that moderate expression of Gal-9 could predict poorer outcome than Gal-9 negative cases, however, it seemed that patients with strong Gal-9 expression predicted best outcome [12]. In lung adenocarcinoma, patients with high Gal-9 expression had shorter survival time than patients with low or negative Gal-9 expression [13]. In sum, the prognostic value of Gal-9 is inconsistent in different cancer types.

In HCC, several studies have shown that Gal-9 expression also have prognostic value. For example, in one study, Li et al. reported that Gal-9 was highly expressed in Kupffer cells (KCs), which were colocalized with Tim-3 positive T cells in HBV-associated HCC patients. They also found that positive Tim-3 expression predicted poor survival in HBV-associated HCC patients, and blocking the interaction between Tim-3 and Gal-9 could increase the function of Tim-3 positive T cells [14]. However, they didn’t evaluate the prognostic value of Gal-9 in their study. Several other studies showed different results. Sideras et al. found that 79% HCC patients had Gal-9 expression, while Gal-9 mainly expressed in tumor cells, they also found that low Gal-9 and PD-L1 expression level and low number of CD8⁺ tumor-infiltrating lymphocyte (TILs) were associated with reduced outcome in HCC [15]. Besides, they also found that high level of circulating Gal-9 could also predict better survival in HCC. Furthermore, combined analysis of intra-tumoral Gal-9 expression and its circulating level could more confidentially predict survival [16]. Zhang et al. found that Gal-9 expression could be detected in 56.5% of HCC patients, Gal-9 was mainly expressed in tumor cells and its expression was closely correlated with histopathological grade, vascular invasion, lymph node metastasis and intrahepatic metastasis. They also found that patients with Gal-9⁺ tumors had longer survival time than those with Gal-9^- tumors [17]. According to these studies, the cell type that express Gal-9 was not consistent, and the prognostic value of Gal-9 seemed also controversial. Thus it is necessary to further evaluate the prognostic value of Gal-9 in HCC.

In current study, we examined the expression of Gal-9 in 140 HBV-associated HCC tumor tissues, evaluated the association between Gal-9 expression and clinicopathological features, and also evaluated its prognostic value in HBV-associated HCC. Besides, we also examined the cell types that have positive Gal-9 expression.

Materials and Methods

Results

Discussion

Gal-9, a S-type lectin that belongs to Galectin family, plays an essential role in the regulation of immune response. Gal-9 was identified as a ligand of Tim-3, and its binding with Tim-3 could induce Th1 cell death, thus inhibiting Th1 immunity [7, 20]. Besides direct interaction with Tim3 on T cells, Gal-9 could promote the expansion of CD11b⁺Ly-6G⁺ cells, which represent the granulocytic myeloid-derived suppressor cells (MDSCs), thus indirectly regulate Th1 immune responses [21].

Besides Tim3, Gal-9 has also been identified to interact with other receptors, such as CD44 and Dectin 1. Wu et al. found that Gal-9 was highly expressed in induced regulatory T cells (iTreg), and Gal-9 could bind with CD44 to activate Smad3, thus increase the stability and function of iTreg cells [8]. Gal-9 could also bind with Dectin 1, resulting in macrophage mediated suppression on adaptive immune response in pancreatic ductal adenocarcinoma (PDA) [9]. Recently, Gal-9 was identified as a program death-1 (PD-1)-binding molecule, and the interaction between Gal-9 and PD-1 could attenuate TIM-3/Gal-9-induced cell death, thus regulate the exhaustion of T cells and the efficacy of immunotherapy [22].

Recent years, Gal-9 were found to be specifically expressed in various tumor cells and play a critical role in antitumor immunity. For instance, in gastrointestinal stromal tumors (GISTs), Komita et al. found that the direct interaction between Gal-9 positive tumor cells and Tim-3 positive NK cells may be involved in the suppression of antitumor immunity [23]. In colon cancer, Wang et al. found that Gal-9 could stimulate NK cell migration, thus promote immune surveillance [24]. In glioma, Yuan et al. found that Gal-9 was highly correlated with immune checkpoint molecules and M2 macrophages, indicating a suppressive immune response [25]. In nasopharyngeal carcinoma (NPC), Klibi et al. found that blockade of Galectin-9/Tim-3 interaction could alleviate the Th1-suppressive effect, thus improve antitumoral T-cell responses [26].

Besides tumor cells, Gal-9 has also been found to be specifically expressed in other cell types in different tumors. For example, in early-stage small cell lung cancer (SCLC), Chen et al. demonstrated that Gal-9 was mainly expressed in TILs, and Gal-9 expression was markedly correlated with immune infiltration in the tumor microenvironment. They also found that high expression level of Gal-9 on TILs predicted better recurrence-free survival (RFS) in stage I–III SCLC [27]. In muscle-invasive bladder cancer, Qi et al. found that Gal-9 mainly expressed on tumor-associated macrophages (TAMs), which predicted poor OS and RFS [28]. In osteosarcoma, Gal-9 was found to be expressed in regulatory T cells (Tregs), and its interaction with Tim3⁺ T cells resulted in the suppression in Th1 responses, thus inhibit immune response [29]. In metastatic differentiated thyroid cancer (mDTC), Severson et al. found that Gal-9 was expressed in tumor-involved lymph nodes (TILNs), and Gal-9 may contribute to the dysfunction of PD-1⁺Tim3⁺CD8⁺ T cells, thus promote disease progression [30]. Recently, Isobel et al. showed that expression of Gal-9 was upregulated in both CD8⁺ and CD4⁺T cells in the TILs and peripheral blood in patients with virus-associated solid tumors (VASTs), and Gal-9 expression was also related to the impaired function of T cells [31]. Since Gal-9 expression was indicated to be associated with the formation of suppressive immune microenvironment, Gal-9 was considered as a potential ideal target for cancer therapy [5, 32, 33].

In HCC, the cell types that express Gal-9 is still controversial, and the prognostic value of Gal-9 has not been well studied. In Li’s study, Gal-9 was found to be highly expressed in KCs, while in two other studies, Gal-9 was found to be mainly expressed in tumor cells. In addition, they also got different conclusions in the prognostic value of Gal-9 in HCC. In Li’s study, although they didn’t directly evaluate the prognostic role of Gal-9, they did find that positive Tim-3 expression could predict reduced survival in patients with HBV-associated HCC. Since they found that when the interaction between Tim-3 and Gal-9 was blocked, the function of Tim-3⁺ T cells could increase, indicating that positive expression of Gal-9 in KCs may predict poor survival. While in other three studies, Sideras et al. and Zhang et al. found that patients with Gal-9⁺ tumors seemed to have longer survival time than patients Gal-9^- tumors [14–17]. These discrepancies might come from the different patients they studied, Li et al. specifically studied HBV-associated HCC patients, while the other two groups didn’t specify a subclass of the HCC patients. Because of these discrepancies, more studies are needed to confirm the cell types that express Gal-9 and the prognostic value of Gal-9 in HCC patients.

In our study, we examined Gal-9 expression in 140 HBV-associated HCC patients and found that Gal-9 was specifically expressed in certain patients, and Gal-9 expression was closely correlated with some clinicopathological features, that is, lymph node metastasis (P =0.029) and Ki-67 proliferation index (P =0.009). We also found that patients with Gal-9⁺ tumors tended to have shorter survival time than patients with Gal-9^- tumors. Our Cox regression analysis results showed that Gal-9 could be an independent prognostic marker for HBV-associated HCC. In addition, by using dual IF staining, we confirmed that Gal-9 was mainly expressed in CD68⁺CD163⁺ KCs rather than tumor cells or M2 macrophages, indicating that Gal-9 may be involved in the formation of immune suppressive microenvironment in HBV-associated HCC.

Our data are more consistent with Li’s results, we think it’s because that we all used tissue samples from HBV-associated HCC patients. In Li’s study, they also showed CD68+ KCs had highest Gal-9 expression in HCC by flow cytometry. Yet, in their study, they only evaluated the prognostic value of Tim-3, but not Gal-9 [14]. Here, in this study, we confirmed that positive expression of Gal-9 was significantly associated with poor prognosis, that is, Gal-9 had prognostic value in HBV-associated HCC. Since Li’s study showed that the Tim-3⁺ T cell function can be increased by blocking the interaction between Tim-3 and Gal-9, we proposed that expression of Gal-9 in KCs may mediate a suppressive function on anti-tumor T cell response.

Our data is not consistent with the other two studies, which might be caused by the different group of samples we used. We specifically studied Gal-9 expression in HBV-associated HCC, but the other two studies didn’t specify the subclass of HCC. More studies are still needed to examine the prognostic value of Gal-9 in HCC patients that are not HBV-associated. Besides, although our results demonstrated that KCs specifically expressed Gal-9 in HBV-associated HCC, however, how Gal-9 expression was regulated in KCs is not known. Recently, Selno et al. reported that transforming growth factor beta type 1 (TGF-β)/Smad3 pathway could regulate the expression of Gal-9 in colorectal cancer [34]. And Chen et al. reported that EZH2 could promote Gal-9 expression through miR-22 in HCC cells [35]. Further study was also needed to verify whether Gal-9 expression is also regulated through similar mechanisms in KCs in HBV-associated HCC.

In brief, in this study we confirmed that Gal-9 is specifically expressed in certain HBV-associated HCC patients, and positive Gal-9 expression is closely correlated with some clinicopathological features, that is, lymph node metastasis and proliferation index, and also poor prognosis. Furthermore, Gal-9 is mainly expressed in CD68⁺CD163⁺ KCs, indicating it may promote tumor progression through the induction of immunosuppressive microenvironment. Our results suggested that Gal-9 could be a new marker for prognosis and a potential target for the therapy of HBV-associated HCC.

Supplementary Materials

Supplementary Methods