Herpesvirus entry mediator (HVEM) is known to regulate immune response and to be expressed in several human malignancies. However, to the authors's knowledge, the precise role of HVEM in human cancer biology remains unknown. The objective of the current study was to clarify the clinical significance of HVEM in human esophageal squamous cell carcinoma as well as its in vivo functions.
HVEM expression was evaluated in 103 patients with esophageal squamous cell carcinoma to explore its clinical relevance and prognostic value. The functions of HVEM in tumors were analyzed in vitro and in vivo using the small interfering RNA (siRNA) silencing technique.
HVEM expression was found to be significantly correlated with depth of tumor invasion and lymph node metastasis. Furthermore, it was found to be inversely correlated with tumor-infiltrating CD4+, CD8+, and CD45RO+ lymphocytes. It is important to note that HVEM status was identified as an independent prognostic marker. HVEM gene silencing significantly inhibited cancer cell proliferation in vitro and cancer growth in vivo. This antitumor effect was associated with reduced cell proliferation activity. The effect was also correlated with the induction of CD8+ cells and upregulation of local immune response.
Esophageal cancer is one of the most difficult gastrointestinal malignancies to treat and cure.[1, 2] Patients often experience distant metastasis or local disease recurrence even after undergoing curative resection. Although multimodality approaches based on surgery combined with preoperative chemotherapy and/or radiotherapy have been attempted, the efficacy of these treatments is limited, and overall survival remains poor.[3, 4] Therefore, novel strategies against esophageal cancer need to be developed and established to improve the prognosis of patients.
Herpesvirus entry mediator (HVEM; TNFRSF14) is a member of the tumor necrosis factor (TNF) receptor superfamily, which is expressed on several types of cells, including T cells, B cells, natural killer cells, dendritic cells, and myeloid cells, as well as nonlymphoid organs including the lung, liver, and kidney.[5, 6] HVEM is a ligand for the immunoglobulin (Ig) superfamily members B-lymphocyte and T-lymphocyte attenuator (BTLA) and CD160, and is also a receptor for the TNF superfamily members LIGHT and LTα.[7-10] Ligation of HVEM by LIGHT promotes T-cell proliferation and cytokine production by initiating activation of the prosurvival transcription factor nuclear factor-κB (NF-κB).[6, 11] By sharp contrast, HVEM engagement of BTLA and CD160 activates inhibitory signaling in T cells, resulting in decreased T-cell proliferation and cytokine production.[9, 10, 12] Therefore, HVEM is known to display a dual functional activity for T-cell activation depending on the ligands engaged. However, the inhibitory function of HVEM may be dominant over its costimulatory activity, as demonstrated by enhanced activation of T cells in HVEM-deficient mice. Furthermore, HVEM-deficient mice have been shown to be more susceptible to concanavalin A-mediated T cell-dependent autoimmune hepatitis and myelin oligodendrocyte glycoprotein peptide-induced experimental autoimmune encephalomyelitis. To our knowledge to date, many studies have focused on the HVEM pathway in several diseases, such as autoimmune disease, infection, inflammatory bowel disease, and transplantation.[12-15] Conversely, to the best of our knowledge, the role of HVEM in tumors remains largely unknown. HVEM expression has been identified in several human tumor cell lines including colon cancer, breast cancer, and T-cell leukemia, and also in actual human malignancies such as melanoma and hematopoietic malignancies.[16-19] HVEM on tumor cells has been shown to inhibit cytokine production and the proliferation of tumor antigen-specific CD8+ T cells via BTLA in vitro. In addition, several murine studies have demonstrated that blockade of the HVEM/BTLA pathway augments tumor antigen-specific immune responses and inhibits tumor growth.[20-22] Others have reported that LIGHT mediates tumor cell apoptosis via signaling through tumor-expressed HVEM, leading to the suppression of tumor growth.[7, 18] Thus, the functions and roles of HVEM in tumors appear to be complex.
In the current study, we hypothesized that HVEM on tumor cells might play a role in an intractable human malignancy, esophageal squamous cell carcinoma (ESCC). Therefore, we attempted to clarify its clinical significance in human ESCC. Furthermore, we investigated the biological roles of HVEM, using an RNA interference method both in vitro and in vivo.
MATERIALS AND METHODS
We examined 103 patients with esophageal cancer who underwent curative esophagectomy in the Department of Surgery at Nara Medical University between 1995 and 2008. Patients had not received chemotherapy or radiotherapy before surgery. All esophageal cancers evaluated in the current study were pathologically diagnosed as squamous cell carcinoma. Tumors were classified according to the TNM staging system. The median follow-up for all patients was 25.7 months. Written informed consent was obtained from all patients according to our institutional guidelines.
Sections were stained using a Dako EnVision system (Dako Cytomation, Kyoto, Japan) as previously described. As primary antibodies, anti-HVEM antibody (94804; R&D Systems, Minneapolis, Minn), anti-CD4 antibody (4B12; Dako Cytomation), anti-CD8 antibody (C8/144B; Dako Cytomation), anti-CD45RO antibody (UCHL1; Dako Cytomation), and anti-FoxP3 (forkhead box P3) antibody (ab22510; Abcam, Tokyo, Japan) were used. Sections were incubated with primary antibodies overnight at 4°C. For the staining of mouse tissue, anti–Ki-67 antibody (SP6; Spring Bioscience, Fremont, Calif), anti-CD4 antibody (Novus Biologicals, Littleton, Colo), and anti-CD8 antibody (Novus Biologicals) were used and incubated overnight at 4°C. Staining was performed using the VectaStain ABC kit (Vector Laboratories, Burlingame, Calif). Immunohistochemistry for HVEM was evaluated by authorized pathologists who had no knowledge of the patients' clinical status. At least 1000 tumor cells were scored, and the percentage of tumor cells demonstrating positive staining was calculated. To count each T-cell subset, 3 randomly selected areas were counted and an average number was scored.
Animals and Cell Lines
Female BALB/c mice aged 5 weeks were obtained from CLEA Japan Inc (Tokyo, Japan). All experiments were conducted under a protocol approved by our institutional review board. The human ESCC cell lines TE-1 and TE-6 and a murine colon adenocarcinoma cell line, Colon 26, were obtained from the RIKEN BioResource Center (Wako, Saitama, Japan).
Extraction of Total RNA and Real-Time Reverse-Transcriptase Polymerase Chain Reaction Analysis
Total RNA was isolated using RNAspin Mini (GE Healthcare Ltd, Little Chalfont, Buckinghamshire, UK), and cDNA was synthesized using a ReverTra Ace qPCR RT Kit (Toyobo Co, Ltd, Osaka, Japan). For real-time reverse-transcriptase polymerase chain reaction (PCR) analysis, cDNA was amplified in TaqMan Fast Universal PCR Master Mix (Applied Biosystems, Foster City, Calif) with gene-specific primers and probes on the StepOnePlus Real-Time PCR System (Applied Biosystems). The expression level of the housekeeping gene, β-2-microglobulin, was measured as an internal reference.
Preparation of Cell Lysates and Western Blot Analysis
We resolved the cell lysates in sodium dodecyl sulfate -polyacrylamide gels and transferred them onto polyvinylidene difluoride membranes (Merck Millipore, Billerica, Mass). Antihuman HVEM antibody (94804; R&D Systems) and antimouse HVEM antibody (R&D Systems) were used. The membranes were incubated with the indicated primary antibody overnight at 4°C and then incubated with horseradish peroxidase-conjugated IgG (Santa Cruz Biotechnology [Shanghai] Co., Ltd, Shanghai). We detected peroxidase activity on x-ray films using an enhanced chemiluminescence detection system.
Small Interfering RNA Transfection of HVEM
For transfection analyses, TE-1 and TE-6 cells were transfected either with control RNA or with 80 nmol/L of small interfering RNA (siRNA) of HVEM. Colon 26 cells were also transfected either with control RNA or with 20 nmol/L of siRNA of HVEM. Transfections were performed using the Lipofectamine system (Invitrogen, Carlsbad, Calif). The human HVEM siRNA duplexes, generated with 3′-dTdT overhangs and prepared by Qiagen (Venlo, the Netherlands), were chosen against the DNA target sequence as follows: 5′-CACCTACATTGCCCACCTCAA-3′. For the mouse HVEM, the DNA target sequence was as follows: 5′-CTGTATGTGCTGACTGCCTAA-3′.
Cell Viability Assay and Cell Cycle Analysis
Cell viability was determined using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega Corporation, Madison, Wis). The absorbance at 490 nanometers (nm) was recorded. Analysis of the cell cycle was performed using the Cycletest Plus DNA Reagent Kit (BD Biosciences, San Jose, Calif). The cellular DNA content of at least 20,000 cells was analyzed using a FACSCalibur instrument (BD, Franklin Lakes, NJ) and the percentage of cells in the different phase of the cell cycle was determined using CellQuest software (BD).
Animal Experimental Protocol
In an in vivo model, Colon 26 cells were subcutaneously injected into the lower flank region of mice. Treatment was initiated 3 days after tumor implantation when a small palpable lump was evident. We locally injected either control RNA or HVEM siRNA with AteloGene Local Use (Koken Company Ltd, Tokyo, Japan) twice a week for 2 weeks, as previously described. The tumor volume was calculated according to the following formula: V = A × B2/2 (mm3), in which A is the largest diameter (in mm) and B is the smallest diameter (in mm).
Categorical variables were presented as numbers and percentages, and groups were compared using the chi-square test or Fisher exact test. Continuous variables were expressed as means and standard errors, and were compared using the Student t test. The survival curves were calculated by the Kaplan-Meier method and were analyzed using the log-rank test. A multivariate survival analysis was performed using the Cox proportional hazards model. A P value < .05 was considered to be statistically significant.
HVEM Expression in Human ESCC
We first compared the relative expression of HVEM between ESCC tissues and noncancer tissues using available frozen tissues. Real-time PCR analysis demonstrated that ESCC tissues expressed much higher levels of HVEM mRNA than noncancer tissues (P < .001) (Fig. 1A). Furthermore, the HVEM expression of cancer tissues was consistently found to be higher than that of noncancer tissues in each individual patient with esophageal cancer (Fig. 1A). We next examined HVEM expression in actual ESCC tissues by immunohistochemistry. Positive staining for HVEM was noted on both the cell membrane and in the cytoplasm of cancer cells in 91 of 103 patients (88.3%) (Fig. 1B). The mean percentage of HVEM-positive tumor cells was 42.8%. In noncancer tissues, some mononuclear cells were also found to be positive for HVEM.
Clinicopathological Significance of HVEM Expression in Human ESCC
To further investigate the clinical relevance of HVEM expression, all specimens were classified into 2 groups according to the percentage of HVEM-positive tumor cells as follows: 42 tumors (40.8%) with high expression (≥ 50% of HVEM-positive tumor cells) and 61 tumors (59.2%) with low expression (< 50% of HVEM-positive tumor cells) (Fig. 1B). We then evaluated the correlation between HVEM status and various clinicopathological findings (Table 1). The tumors with high HVEM expression were found to be significantly larger in size compared with the tumors with low HVEM expression (P < .001). Furthermore, the tumors with high HVEM expression had a significantly deeper invasion of the wall and more commonly had lymph node metastasis (P < .001). Thus, the data from the current study suggested that expression of HVEM might be involved in disease progression in patients with ESCC.
Table 1. Clinicopathological Characteristics According to Tumor HVEM Expression
Abbreviation: HVEM, herpesvirus entry mediator.
Values are expressed as the mean ± standard deviation.
All 9 patients had a supraclavicular lymph node metastasis.
We then investigated the prognostic value of HVEM expression. The 5-year survival rate was significantly lower in patients with tumors with high HVEM expression compared with patients with low HVEM expression (18.3% vs 49.6%; P < .001) (Fig. 1C). Furthermore, the multivariate analysis demonstrated that HVEM status as well as tumor status, lymph node metastasis, and distant metastasis were independent prognostic factors (P = .041) (Table 2). Taken together, HVEM expression in tumor cells might play a critical role and also be a promising potential therapeutic target in patients with ESCC.
Data were obtained from the Cox proportional hazards model.
Tumor size, mm
T3, T4a/T1, T2
Lymph node status
Inverse Correlation Between HVEM Expression and Tumor-Infiltrating Lymphocytes
Because the HVEM-BTLA-CD160 pathway is known to inhibit T-cell function, we evaluated the tumor-infiltrating lymphocytes (TILs) by immunohistochemistry to clarify the correlation between HVEM status and TILs. HVEM expression levels were found to be inversely correlated with the number of CD4+ lymphocytes (P = .02) and CD8+ lymphocytes (P = .004) (Fig. 1D). There was also a significant inverse correlation noted between HVEM expression and CD45RO+ lymphocytes (P = .049). By contrast, no significant correlation with FoxP3+ lymphocytes was observed (data not shown). Data indicated that HVEM expression in tumor cells might inhibit T-cell infiltrations into ESCC tissues.
HVEM Silencing Inhibits the Proliferation of Esophageal Cancer Cells in Vitro
To further investigate the precise function of HVEM in the tumor environment, we next investigated the role of HVEM expression in ESCC in vitro. We used the human ESCC cell lines TE-1 and TE-6 and examined the effects of HVEM downregulation using the siRNA knockdown approach. At 72 hours after transfection, siRNA knockdown significantly reduced HVEM expression compared with control specimens (Figs. 2A and 2B). We then examined its role in the regulation of cancer cell proliferation by MTS assay. Cell proliferation was significantly suppressed by HVEM gene silencing in both cells (Fig. 2C). Thus, tumor-expressing HVEM might play a direct role in ESCC proliferation.
HVEM Silencing Induces Cell Cycle Arrest
To determine the underlying mechanisms in the inhibition of cell proliferation observed with HVEM knockdown, we analyzed cell cycle profiles. The cell cycle analysis demonstrated a significant increase in the S-phase and the G2/M-phase cell populations in both TE-1 and TE-6 cells treated with HVEM siRNA compared with control specimens (TE-1, control RNA vs HVEM siRNA, S-phase: 13% ± 0.3% vs 15.4% ± 0.1% [P = .001] and G2/M-phase: 22.1% ± 0.4% vs 28.9% ± 0.2% [P < .001]; TE-6, control RNA vs HVEM siRNA, S-phase: 12.9% ± 0.4% vs 15.2% ± 0.3% [P = .013] and G2/M-phase: 22.1% ± 0.9% vs 29.7% ± 0.8% [P = .003]) (Fig. 2D). Furthermore, the percentage of apoptotic cells was measured by Annexin V/propidium iodide double staining. There was no significant difference noted between the cells treated with HVEM siRNA or control RNA (data not shown). These data suggest that HVEM silencing might induce cell cycle arrest but not apoptosis, leading to the inhibition of cancer cell proliferation in vitro.
HVEM Silencing Inhibits Tumor Growth in Vivo
We next wanted to examine the function of HVEM under physiological conditions. We used a murine colon cancer cell line, Colon 26, for in vivo analysis because no murine esophageal cancer cell line was available. First, we examined the in vitro effect of HVEM silencing in Colon 26. Similar to human ESCC cells, siRNA knockdown significantly reduced HVEM expression (Fig. 3A). Furthermore, HVEM silencing significantly inhibited cell proliferation in vitro (Fig. 3B). In contrast to human ESCC cells, HVEM silencing induced a significant increase in the G1 phase cell population (control RNA vs HVEM siRNA: 39.7% ± 1.3% vs 58.8% ± 0.1% [P < .001]) (Fig. 3C). We then evaluated the in vivo effect of HVEM silencing. Colon 26 cells were subcutaneously injected into syngeneic BALB/c mice and treated with control or HVEM siRNA. HVEM expression was successfully downregulated by in vivo HVEM siRNA transfection (Figs. 3D and 3E). It is interesting to note that HVEM downregulation significantly inhibited tumor growth (Fig. 3F). Furthermore, the percentage of the necrotic area was significantly higher in those tumors treated with HVEM siRNA than in control specimens (35.2% ± 10% vs 10.2% ± 3.1% [P = .038]) (Figs. 3G and 3H). In addition, we examined the proliferation activity of tumor cells by Ki-67 staining. The percentage of Ki-67–positive cells was found to be significantly decreased in tumors treated with HVEM siRNA compared with control specimens (34.5% ± 0.2% vs 64.8% ± 2.7% [P < .001) (Fig. 4A).
HVEM Silencing Enhances CD8+ Lymphocyte Recruitment and Local Immunity
Finally, we evaluated TILs by immunohistochemistry in this in vivo model. As a result, CD8+ but not CD4+ lymphocytes infiltrating into the surrounding area of the tumor were found to be significantly more abundant in tumors treated with HVEM siRNA than in controls (Fig. 4B). We then analyzed local immune status in tumors. The expression levels of interferon-γ (IFN-γ) and interleukin-2 (IL-2) were significantly higher in tumors treated with HVEM siRNA (Fig. 4C). Data indicated that HVEM blockade not only directly reduced cancer cell proliferation but also promoted CD8+ infiltration into tumors and enhanced local immune response, thereby resulting in the inhibition of tumor growth in vivo.
Tumors evade immune surveillance by expressing several ligands that engage inhibitory T-cell receptors and dampen T-cell functions within the tumor microenvironment.[26, 27] The programmed cell death 1 ligand 1 (PD-L1)/programmed cell death protein 1 (PD-1) pathway is known to be one of the major negative regulatory pathways in tumor immunity. We and others have reported that tumor-expressing PD-L1 is correlated with adverse clinicopathological features and has an independent prognostic value in several human cancers, including esophageal cancer.[26, 28-30] Furthermore, targeting this pathway is currently under investigation in clinical trials.[31, 32] However, the clinical efficacy appears to be limited. Therefore, there is still a need to explore other novel therapeutic targets. In the current study, we addressed the clinical significance and functional role of a recently discovered immunoinhibitory ligand, HVEM, in esophageal cancer.
To the best of our knowledge, there are relatively few studies regarding HVEM in cancer biology. Derre et al have shown that HVEM on melanoma cells inhibited IFN-γ production and the proliferation of tumor-specific CD8+ T cells via BTLA in vitro, suggesting that inhibitory interactions of HVEM-BTLA may play a role in the evasion of host antitumor immunity. To our knowledge to date, however, the role of HVEM in actual human cancer remains largely unknown. In the current study, we first confirmed the overexpression of HVEM in human ESCC tissues. We further found that tumors with higher HVEM expression had more advanced features. It is important to note that the multivariate analysis identified tumors expression of HVEM as an independent prognostic factor. We then analyzed the correlation between HVEM status and TILs. It is widely recognized that TILs play some role in the inhibition of tumor progression and disease recurrence, and have prognostic significance in several human malignancies, including esophageal cancer.[33, 34] More recently, we and others have shown that CD45RO+ for memory T cells may be a better prognostic marker in patients with esophageal cancer.[24, 35] As a result, we found that HVEM expression levels were inversely correlated not only with tumor-infiltrating CD4+ and CD8+ T cells but also CD45RO+ memory T cells. It is interesting to note that recent studies have suggested that HVEM regulates the generation and maintenance of memory T cells.[36, 37] Taken together, HVEM expression in tumor cells may play a critical role in the evasion of host antitumor immune responses and contribute to tumor progression. Therefore, these data further emphasized that HVEM could be a promising target for novel cancer therapy against human ESCC.
In addition to its immunological roles, HVEM functions as either a ligand or receptor in diverse physiological and pathological processes. Recent studies have demonstrated that BTLA, CD160, and glycoprotein D function as activating ligands for HVEM, promoting NF-κB activation and cell survival via HVEM in lymphoid and epithelial cells.[15, 38] However, to our knowledge, no study published to date has addressed the direct effect of HVEM in cancer cell survival. We examined the biological mechanisms of HVEM on tumor cells using the siRNA method, and made several important observations. First, cell proliferation was significantly inhibited by HVEM gene silencing in human ESCC cells and murine colon cancer cells. Second, HVEM silencing induced cell cycle arrest but not apoptosis in vitro. Although HVEM silencing induced the G2/M-phase arrest in human ESCC cells, it mediated the G1 phase arrest in a murine colon cancer cell. The differences may be due in part to the p53 status of each cell line. p53 is known as a key regulator of both the G1/S-phase and G2/M-phase. In fact, the human ESCC cells used in the current study retain mutated p53, whereas Colon 26 cells expressed wild-type p53. Thus, HVEM silencing may induce differential cell cycle alterations depending on p53 status. Taken together, the data from the current study indicate that HVEM might be directly involved in cancer cell proliferation.
Finally, we investigated the immunological and nonimmunological role of HVEM in tumors in vivo under physiological condition. To this end, we used a murine colon cancer cell line, because a murine esophageal cell line was not available in our institution. As a result, HVEM blockade induced by local injection of siRNA was found to significantly inhibit tumor growth of Colon 26 cells in syngeneic immunocompetent mice. We also found that HVEM blockade significantly inhibited the proliferation of tumor cells in vivo. Thus, HVEM blockade had a direct antitumor effect on tumor cells in vivo. We also found that HVEM blockade significantly induced the infiltration of CD8+ TILs. Furthermore, IFN-γ and IL-2 were found to be significantly upregulated in tumors treated with HVEM siRNA. Several recent studies have demonstrated that blockade of HVEM/BTLA pathways using soluble BTLA or the vaccine fused to glycoprotein D enhanced tumor-reactive T-cell activation and led to tumor regression or tumor growth inhibition.[20-22] In addition, the blockade of HVEM/BTLA interactions increased levels of IFN-γ and IL-2 in the tumor microenvironments. The current study data may further corroborate these previous findings. Thus, HVEM blockade might have an indirect antitumor effect induced by the inhibition of a T cell-negative pathway. However, underlying molecular mechanisms present in T cells infiltrating into tumors are still unclear. Therefore, further fundamental studies will be needed. Furthermore, to confirm the in vivo efficacy of the HVEM blockade, different strategies such as monoclonal antibody therapy should be evaluated in the preclinical setting.
In conclusion, we believe the current study is the first to demonstrate that higher HVEM expression is correlated with advanced features of human cancer and fewer TILs, and that HVEM is an independent prognostic marker in human ESCC. Furthermore, HVEM contributes to cancer cell proliferation and impairs antitumor immune responses. It is important to note that HVEM blockade has a significant antitumor effect under physiological conditions. Therefore, the results of the current study may provide the rationale for developing a novel cancer therapy targeting HVEM in human malignancies.
Supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (no. 21591648 and no. 24591887 to Dr. Sho and no. 24791443 to Dr. Migita); a research grant from Takeda Science Foundation; and a research grant from Nakayama Cancer Research Institute (to Dr. Sho).