TNF plays a dual, still enigmatic role in melanoma, either acting as a cytotoxic cytokine or favoring a tumorigenic inflammatory microenvironment. Herein, the tumor growth of melanoma cell lines expressing major histocompatibility complex class I molecules at high levels (MHC-Ihigh) was dramatically impaired in TNF-deficient mice, and this was associated with enhanced tumor-infiltrating CD8+ T lymphocytes. Immunodepletion of CD8 T cells fully restored melanoma growth in TNF−/− mice. Systemic administration of Etanercept inhibited MHC-Ihigh melanoma growth in immunocompetent but not in immunodeficient (IFNγ−/−, nude, or CD8−/−) mice. MHC-Ihigh melanoma growth was also reduced in mice lacking TNF-R1, but not TNF-R2. TNF−/− and TNF-R1−/− mice as well as Etanercept-treated WT mice displayed enhanced intratumor content of high endothelial venules surrounded by high CD8+ T-cell density. Adoptive transfer of activated TNF-R1–deficient or –proficient CD8+ T cells in CD8-deficient mice bearing B16K1 tumors demonstrated that TNF-R1 deficiency facilitates the accumulation of live CD8+ T cells into the tumors. Moreover, in vitro experiments indicated that TNF triggered activated CD8+ T cell death in a TNF-R1–dependent manner, likely limiting the accumulation of tumor-infiltrating CD8+ T cells in TNF/TNF-R1–proficient animals. Collectively, our observations indicate that TNF-R1–dependent TNF signaling impairs tumor-infiltrating CD8+ T-cell accumulation and may serve as a putative target to favor CD8+ T-cell–dependent immune response in melanoma. Cancer Res; 75(13); 2619–28. ©2015 AACR.

TNFα is a major inflammatory cytokine that has been initially identified in 1975 for its ability to trigger necrosis of mouse cutaneous fibrosarcoma in vivo (1). At variance with prior knowledge, a growing body of evidence indicates that TNF produced by cancer and/or stromal cells may favor the establishment of a proinflammatory microenvironment, modulating immune response, and enhancing cancer cell proliferation, tumor angiogenesis, tumor progression, and metastasis (2, 3). Different studies have illustrated the role of TNF in skin tumor formation. As compared with their wild-type (WT) counterparts, TNF-deficient mice, TNF-R1–deficient mice, and, albeit to a lesser extent, TNF-R2–deficient mice are reluctant to carcinogen-induced benign skin tumor (i.e., papillomas) formation (4–6).

The role of TNF in melanoma remains controversial. Injection of high levels of recombinant TNF triggers necrosis of melanoma, not only in mice, but also in humans and is currently used in isolated limb perfusion in the clinic (2). In sharp contrast, it has been recently shown that TNF, which is produced in patients treated with BRAF V600E inhibitors (7, 8), may confer treatment resistance of human melanoma by increasing Twist1 levels (9). The role of TNF in melanoma has been further investigated in mice using B16 melanoma cells, which do not express TNF endogenously (10). Whereas ectopic membrane TNF on B16 cells triggers TNF-R2–dependent myeloid cell death (11), and subsequent impaired in vivo melanoma growth, ectopic expression of soluble TNF at low levels by B16 has opposite effect on melanoma growth in mice, most likely through its ability to enhance tumor angiogenesis (10). Both B16 cell growth and tumor angiogenesis are reduced in TNF-R2–deficient mice (12). Moreover, lung invasion of intravenously injected B16 melanoma cells is decreased in TNF−/− and TNF-R2−/− mice, indicating that TNF likely enhances melanoma dissemination in a TNF-R2–dependent manner (13). Vaccination toward TNF triggers self anti-TNF antibodies and inhibits lung invasion of intravenously injected B16 melanoma cells (14). Similar data have been obtained by injecting anti–TNF-neutralizing antibodies (14) or soluble TNF-R1 (15), indicating that TNF blockade may represent a useful strategy to prevent melanoma metastasis. A recent study showed that TNF deficiency can delay tumor growth in a spontaneous mouse model of BRAF V600E melanoma (16). In humans, whereas different case reports have documented the occurrence of melanoma in patients with autoimmune disorders treated with anti-TNF, recent meta-analyses did not confirm the association of anti-TNF treatments and an increased melanoma incidence (17–20).

TNF is involved in the modulation of both innate and adaptive immune responses. Conflicting results have been published on the role of TNF in the T-cell immune response toward cancer cells. On the one hand, TNF acts as an effector molecule in CD8+ T-cell–triggered cell death of cancer cells (21) and a costimulatory cytokine able to enhance naive CD8+ T-cell proliferation and cytokine secretion (22, 23). In association with IFNγ, TNF induces senescence in cancer cell lines (24). In addition, TNF is required for the establishment of antitumor immune response by facilitating dendritic cell maturation as well as CD8+ T-cell activation and tumor infiltration (25, 26). On the other hand, TNF triggers activation-induced cell death in CD8+ T cells (27), thus likely limiting immune response duration. Moreover, TNF may facilitate the increased number of regulatory T (Treg; refs. 13, 28) and B cells (29) as well as myeloid-derived suppressor cells (MDSC; refs. 30, 31). In an adoptive transfer therapy protocol of CD8+ T cells in mice, TNF could induce dedifferentiation of melanoma cells associated with a decrease of melanocytic antigen expression, likely contributing to tumor relapse (32).

The present study aimed at re-evaluating the role of TNF in CD8+ T-cell–dependent immunity in experimental melanoma. Our data indicate that host TNF-R1–dependent TNF signaling impairs the accumulation of tumor-infiltrating CD8+ T cells in mouse melanoma, facilitating melanoma immune escape. TNF neutralization enhances CD8+ T-cell–dependent immune response, limiting melanoma progression.

Cell lines

B16F10 cell line was purchased from the ATCC. B16K1 is a genetically modified cell line obtained from B16F10 cells, which stably express MHC-I molecule H-2Kb (33, 34). Cells were cultured in DMEM medium containing 10% heat-inactivated FCS. B16K1 cells were authenticated in February 2012 by Leibniz-Institut DSMZ GmbH (34). To guarantee cell line authenticity, B16F10 and B16K1 cell lines were used for a limited number of passages and routinely tested for the expression of melanocyte-lineage proteins such as tyrosinase-related protein 2 (TRP2).

Mice

TNF-deficient, TNF-R1-deficient, TNF-R2–deficient, and IFNγ-deficient C57BL/6 mice were purchased from The Jackson Laboratories. WT C57BL/6 and Nude (RjHan:NMRI strain) mice were from Janvier laboratories. CD8-deficient C57BL/6 mice were a gift from Prof. J. van Meerwijk (INSERM U1043, Toulouse, France). Mice were housed in temperature-controlled rooms in the specific pathogen-free animal facility (Anexplo platform, Toulouse, France), kept on a 12-hour light/dark cycle, and had unrestricted access to food and water. All animal studies were conducted according to national and international policies and were approved by the local committee for animal experimentation.

In vivo tumorigenesis

B16F10 or B16K1 cells (3 × 105) were intradermally and bilaterally injected in WT, TNF−/−, and CD8−/− C57BL/6 mice. In some experiments, Etanercept (3 mg/kg), anti–CD8-depleting antibody (200 μg), or vehicle (NaCl 0.9%) was intraperitoneally injected every third day, the first injection being done 3 days before (preventive treatment) or 10 days after (curative treatment) B16K1 graft. Tumor volume was calculated using a caliper at the indicated days.

Analysis of lymphocyte and high endothelial venule content in tumors

One million B16K1 cells were intradermally and bilaterally injected in WT, TNF−/−, or TNF-R1−/− C57BL/6 mice. At days 10 to 12, mice were sacrificed and tumors were collected and digested with collagenase D (Roche Diagnostics) for 30 minutes at 37°C. Cells were stained with the indicated antibodies or MHC-I dextramers (APC-conjugated H-2Kb/SVYDFFVWL) and live-dead reagents before flow cytometry analysis. Alternatively, tumors were embedded in OCT (Tissue-Tek OCT compound, Sakura), frozen under nitrogen, and cut using a cryostat. Slides were stained with DAPI and the indicated antibodies for immunohistochemistry and analyzed by confocal microscopy (Zeiss; LSM510). At least two slides per tumor were analyzed to determine High endothelial venule (HEV) frequency. CD8+ cells exhibiting lymphoid morphology were quantified in a 100-μm diameter circle around HEV (CD31+ MECA79+) and non-HEV (CD31+ MECA79) vessels by using the Zeiss software.

Analysis of purified CD8 T cells

Spleen CD8+ T cells were purified from naive WT, TNF-R1−/−, and TNF-R2−/− mice using a mouse CD8 T cell purification kit (Miltenyi Biotec). CD8+ T cells were activated, or not, with anti-CD3 and anti–CD28-coated beads (Life Technologies) in the presence of IL2 (Invitrogen; 200 U/mL). To assess CD8+ T-cell proliferation, cells were stained with CFSE (5 μmol/L), and, 3 days later, CFSE dilution was analyzed by flow cytometry. Proliferative index was calculated by using the Modfit software. Expression of TNF-R1 and 2 was determined by flow cytometry. Fresh and activated CD8+ T cells were incubated for 72 hours in the presence of recombinant mouse TNF (1–50 ng/mL; Peprotech). Cell death was evaluated by flow cytometry by monitoring plasma membrane permeability increase toward 7-AAD (eBioscience; 1:30) or caspase-3 activation with an antiactive mouse caspase-3 antibody (BD Pharmingen; 10 μg/mL).

Adoptive transfer in CD8-deficient mice

CD8+ T cells were isolated from WT and TNFR1−/− mice and activated as above described. After 7 days, cell purity and viability averaged 90% for both WT and TNFR1-deficient CD8+ T cells. Three millions of these WT or TNFR1−/− CD8+ T cells were injected into the B16K1 tumors from CD8-deficient mice, which were injected 10 days before with 1 million B16K1 cells. Three days later, animals were sacrificed and the tumor CD8+ T-cell content was analyzed by flow cytometry.

Statistical analyses

Results are expressed as mean of at least three independent determinations per experiment. Mean values were compared using the Student t test with Prism software (Graph-Pad). Differences were considered to be statistically significant when P < 0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant).

Reagents and antibodies are indicated in the Supplementary Information.

Immunogenic melanoma cell growth is impaired in TNF-deficient mice

To evaluate the role of TNF in antimelanoma CD8+ T-cell–dependent immune response, we challenged WT and TNF-deficient mice with orthotopic (i.e., intradermal) injection of B16 mouse melanoma cell lines. We initially selected B16F10 and B16K1, which constitutively express MHC-I at low and high levels, respectively (Fig. 1A). Both B16 melanoma cell lines exhibited similar in vitro cell proliferation rates (Fig. 1B) and expressed TNF-R1, but not TNF-R2 (Supplementary Fig. S1A). Exogenous TNF did not alter in vitro B16 cell proliferation (Supplementary Fig. S1B) and growth of spheroids (Supplementary Fig. S1C). Thus, in agreement with a previous report (10), it is unlikely that TNF directly exerts a potent pro- or antiproliferative effect toward B16 melanoma cells. B16F10 cells grew much faster in WT mice than B16K1 cells, whereas both cell lines displayed quite similar tumor growth in CD8-deficient mice (Fig. 1C), indicating that B16K1 cells are indeed more immunogenic. Whereas the in vivo tumor growth of B16F10 was minimally reduced in TNF−/− mice (Fig. 1D), that of B16K1 was dramatically impaired at all times (Fig. 1E). The contribution of host TNF to tumor growth was next evaluated using two additional melanoma cell lines. B16-BL6 (MHC-Ilow) tumor growth was not significantly impaired in TNF−/− mice (Supplementary Fig. S2A). In sharp contrast, BrafV600E/+, Pten−/−, Cdkn2a−/− C57BL/6 primary mouse melanoma cells (Yumm cells; ref. 35), which constitutively expressed MHC-I at high levels, displayed significant tumor growth reduction in TNF−/− mice (Supplementary Fig. S2B). Collectively, our data indicate that MHC-I expression level at the cell surface of melanoma cells conditioned the extent of tumor growth inhibition induced by TNF deficiency.

CD8+ T-cell–dependent immune response and tumor infiltration are enhanced in TNF-deficient mice

Because the MHC-I restricts antigen recognition by CD8+ T cells, the consequences of host TNF deficiency on the adaptive immune response were next investigated. To this end, WT and TNF-deficient mice were orthotopically injected with 1 million of B16K1 cells and sacrificed 12 days later. Under these conditions, B16K1 tumor weights were significantly reduced in TNF-deficient mice (Fig. 2A), further confirming the tumor-promoting role of host TNF toward immunogenic mouse melanoma. Analysis of the immune response in spleens showed no significant changes in absolute numbers of CD45+ leukocytes, CD4+, and CD8+ T cells between the two groups (Supplementary Fig. S3A). In addition, the spleen cell–mediated cytotoxicity toward B16K1 cells was not significantly affected by TNF deficiency, further arguing that TNF is unlikely a potent cytotoxic cytokine toward B16K1 cells (Supplementary Fig. S3B). In sharp contrast, the absolute numbers of CD45+ leukocytes as well as CD4+ and CD8+ T cells were significantly increased in draining lymph nodes of TNF-deficient mice (Supplementary Fig. S4A). Analysis of CD69 expression, a T-cell activation marker, indicated that the number of activated CD4 and CD8 T cell was greater in TNF-deficient mice (Supplementary Fig. S4B). Upon restimulation of lymph node cells with irradiated B16K1 cells, cytokine concentration was determined by cytometric bead array, enabling the measurement of Th1, Th2, and Th17 cytokines. The production of IFNγ, IL6, and IL10 was enhanced by TNF deficiency (Supplementary Fig. S4C); IL17 and IL4 were not detected under our experimental conditions.

Analysis of leukocyte tumor content by flow cytometry indicated a major increase of CD45+ and CD3+ cell tumor infiltration in TNF-deficient mice as compared with WT mice (Fig. 2B). In addition, the proportion of different T-cell subpopulation content was altered in tumors derived from TNF−/− mice, with a significant increase of CD4+, including Treg (CD4+ Foxp3+), and CD8+ tumor-infiltrating lymphocytes (TIL; Fig. 2C and Supplementary Fig. S5A). Of note, the ratio of CD8+ T cells to CD4+ T cells was significantly higher in tumors from TNF-deficient mice (Fig. 2C). Moreover, the proportion of intratumor MDSC (Gr1+ CD11b+) was not affected by TNF deficiency (Supplementary Fig. S5B). We next evaluated the tumor content of CD8+ T cells, which are specific for TRP2, a differentiation antigen of melanocytic cells. By using MHC-I dextramer technology, we showed that TRP2-specific CD8+ T-cell content was higher in tumors derived from TNF-deficient animals (Fig. 2D).

To evaluate whether host TNF deficiency impairs B16 melanoma cell growth by enhancing CD8+ T-cell tumor infiltration, CD8+ T cells were depleted both in WT and TNF-deficient mice. As a matter of fact, anti-CD8 antibody injection depleted CD8+ T cells efficiently (Fig. 3A) and fully restored B16K1 tumor growth in TNF-deficient mice (Fig. 3B). Importantly, B16K1 tumor growth was identical in TNF-deficient and -proficient mice depleted for CD8+ T cells, indicating that melanoma-promoting effects of host TNF likely rely on the inhibition of CD8+ T-cell–dependent immune response.

HEVs are more abundant in melanoma from TNF-deficient mice

Recent studies indicated that HEV density is directly correlated to CD8+ T-cell infiltration in human and mouse tumors, including melanomas (36–39). We sought to evaluate by immunohistochemistry whether CD8+ T-cell infiltration was associated with the presence of HEVs in B16 mouse tumors developed in WT and TNF-deficient mice (Fig. 4A). Twelve days after B16K1 cell injection, whereas only 20% of tumors from WT mice displayed HEVs (denoted as MECA79+ vessels), 65% of tumors from TNF-deficient mice were positive for HEV (Fig. 4B). We next quantified the number of CD8+ T cells that surrounded HEV (CD31+ MECA79+) and non-HEV (CD31+ MECA79) vessels (Fig. 4C). In both WT and TNF-deficient mice, the CD8+ T-cell number was higher next to HEV vessels than around MECA79 CD31+ vessels. This finding indicates that HEVs represent a major gateway for CD8+ T-cell tumor infiltration in melanoma, regardless TNF expression in stromal cells. Moreover, the density of CD8+ T cells was higher next to non-HEV vessels from TNF−/− mice than in their WT counterparts, further arguing that host TNF reduces CD8+ T-cell melanoma infiltration.

Altogether, our data indicate that TNF reduces the content of HEVs and CD8+ T cells within melanoma under our experimental conditions.

CD8+ TIL content is increased in TNF-R1–deficient mice

The role of host TNF receptors was next evaluated by comparing B16K1 tumor growth in WT and TNF-R1−/− or TNF-R2−/− mice. B16K1 tumor growth was significantly reduced in mice lacking TNF-R1 but not TNF-R2 (Supplementary Fig. S6A), indicating that the host TNF/TNF-R1 axis promotes the development of immunogenic mouse melanoma. We hypothesized that host TNF-R1–dependent TNF signaling impairs CD8+ T-cell–dependent immune response. In accordance with this tenet, CD8+ TIL content was increased in TNF-R1–deficient mice similarly to TNF-deficient mice (Supplementary Fig. S6B), including TRP2-specific CD8+ T cells (Supplementary Fig. S6C), as evaluated by using MHC-I dextramer technology. Moreover, abundant HEV vessels surrounded by high CD8 T-cell density were frequently found in tumors derived from TNF-R1−/− mice (5 of 6 tumors exhibited HEVs; Supplementary Fig. S6D). Thus, host TNF-R1 reduces the content of both HEVs and CD8+ T cells within melanoma.

TNF induces activated CD8+ T-cell death in a TNF-R1–dependent manner

Analysis of TNF receptor expression on purified CD8+ T cells showed that freshly isolated splenic CD8+ T cells expressed TNF-R2 but not TNF-R1 (Fig. 5A). Upon polyclonal stimulation, TNF-R2 expression was strongly upregulated. Moreover, TNF-R1 became slightly, yet significantly, expressed at the surface of activated CD8+ T cells (Fig. 5A). Neither cell proliferation nor cell death was affected by TNF-R1 deficiency upon polyclonal activation (Fig. 5B). Seven days after polyclonal activation, TNF-R1–deficient or –proficient CD8+ T cells were injected into the B16K1 tumors developed in CD8-deficient mice. Three days later, mice were sacrificed, and CD8+ T-cell content was analyzed in the tumors. As a matter of fact, TNF-R1 deficiency significantly enhanced the accumulation of live CD8+ T cells into the tumors (Fig. 5C). The latter data suggest that TNF impairs CD8+ TIL content through a direct effect on activated CD8+ T cells in a TNF-R1–dependent manner. One putative underlying mechanism could be the induction of cell death of activated CD8+ T cells. Because this hypothesis is technically difficult to evaluate in vivo, we next examined the in vitro sensitivity of purified CD8+ T cells to exogenous TNF by monitoring 7AAD uptake to analyze plasma membrane permeability (Fig. 5D). Freshly isolated WT CD8+ T cells were fully resistant to TNF at all doses. Ten days after polyclonal stimulation, TNF triggered activated WT CD8+ T-cell death (Fig. 5D). Of note, the cytotoxic effect of TNF on activated WT CD8+ T cells was already detected at concentration as low as 5 ng/mL (Fig. 5D). Moreover, TNF-induced CD8+ T-cell death was associated with caspase-3 activation, as evaluated by flow cytometry (Fig. 5E). TNF-R1 deficiency impaired exogenous TNF-induced cell death (Fig. 5D) and caspase-3 activation (Fig. 5E) in activated CD8+ T cells. In sharp contrast, TNF-R2 deficiency minimally affected TNF-induced cell death in activated CD8+ T cells. These findings suggest that TNF-triggered activated CD8+ T-cell death, mainly in a TNF-R1–dependent manner, likely contributes to the resolution of CD8+ T-cell immune response in TNF-proficient hosts.

Etanercept impairs B16K1 melanoma growth in immunocompetent but not immunodeficient mice

To further investigate the role of TNF in melanoma growth and immune response, WT mice were intraperitoneally injected with Etanercept, a soluble human TNF-R2 that efficiently neutralizes mouse TNF (40). Etanercept reduced TNF serum concentration by 50% to 70% upon LPS administration (Supplementary Fig. S7A), further indicating its efficiency to block mouse TNF. Etanercept significantly impaired B16K1 growth in WT mice (Fig. 6A). Whereas Etanercept augmented the number of CD8+ T cells both in lymph nodes (Supplementary Fig. S7B) and tumors (Fig. 6B), CD4+ T-cell content was not significantly affected by TNF neutralization (Supplementary Fig. S7B and Fig. 6B). Interestingly, Etanercept had no effect on B16K1 growth in IFNγ-deficient and nude mice (Supplementary Fig. S7C) as well as in CD8-deficient mice (Fig. 6C), indicating that melanoma growth inhibition mediated by TNF blockade likely requires CD8+ T-cell–dependent immune response. We next evaluated whether Etanercept may represent a putative therapeutic agent toward immunogenic melanoma. To this end, we developed a curative protocol by treating WT mice with Etanercept 10 days after implanting B16K1 cells. Interestingly, Etanercept significantly reduced tumor growth in WT mice having established tumors (Fig. 6D), indicating that blocking TNF may be a useful therapeutic strategy to fight against immunogenic melanoma. Of note, a single Etanercept injection triggered the increased content of HEV and CD8+ T cells in melanoma (Supplementary Fig. S7D).

Collectively, our data demonstrate that the tumor-promoting action of host TNF toward melanoma cells is entirely dependent on its ability to impair CD8+ T-cell homeostasis and tumor infiltration. Blocking TNF enhances CD8+ TIL content and reduces immunogenic melanoma growth in mice.

Considering the major, yet not completely solved, functions of TNF in immune response, our study goes beyond a central dogma of immunology on the role of TNF in anticancer immune response. We provide evidence that host TNF contributes to melanoma immune escape, limiting CD8+ T-cell–dependent antimelanoma immune response. We investigated the molecular and cellular mechanisms by which host TNF inhibits CD8+ T-cell immune response.

The B16K1 tumor growth was significantly reduced in TNF-R1–deficient mice, and this was associated with the accumulation of CD8+ T cells into the tumors. Herein, TNF triggers activated CD8+ T-cell death in a TNF-R1–dependent manner. Whereas such a mechanism remains to be demonstrated in vivo, it likely contributes, at least in part, to limit the establishment of CD8+ T-cell–dependent immune responses in host TNF-proficient animals. Accordingly, TNF-R1 deficiency led to an accumulation of live CD8+ TIL after adoptive transfer of activated CD8+ T cells in CD8-deficient recipients bearing B16K1 tumors.

The tumorigenic TNF-R1–dependent TNF signaling may not be restricted to a direct effect on CD8+ T cells; it could also involve additional cell types of the tumor microenvironment. TNF has been recently shown to facilitate the accumulation of Treg (41) and MDSC (30, 31), the latter exhibiting granulocytic phenotype in cancer (42). Upon B16K1 cell injection, the tumor content of Treg and MDSC did not decrease in TNF-deficient mice, indicating that under our experimental conditions, TNF is unlikely a critical cytokine for accumulating those immunosuppressor cells into the tumors. Whereas we cannot rule out that the activity of both Treg and MDSC is reduced in TNF-deficient mice, TNF-R2 is required for this process (30, 41, 43). Because B16K1 tumor growth was not affected in TNF-R2–deficient mice, this pathway appears to have a minimal function for limiting the CD8+ T-cell immune response. Our observation seems at odds with studies pointing a major effect of TNF on tumor microenvironment, facilitating the accumulation and biologic activity of MDSC (30, 31). One should note, however, that our experimental setting (i) is based on transplanted melanoma cells, (ii) is very short since TILs were analyzed 12 days after cell injection, and (iii) may not totally reflect the chronic inflammation associated with melanoma progression in spontaneous mouse melanoma models (44) and patients.

Another major finding of our study that likely accounts, at least in part, for CD8+ T-cell accumulation in tumors is the increased HEV frequency in B16 melanoma tumors developed in TNF and TNF-R1–deficient mice, indicating that the TNF signaling pathway negatively modulates HEV content in melanoma. Moreover, CD8+ T cells are more abundant around HEV vessels than non-HEV vessels in both TNF−/− and TNF-R1−/− mice. Lymphotoxin (LT) α and β heterotrimers have been shown to trigger LTβR-dependent HEV differentiation in lymph nodes (45) and possibly in melanoma (36). Because LTα is also a ligand for TNF-R1 (46), the possibility that disruption of TNF-R1 signaling may somehow enhance the ability of LTα to interact with LTβ and trigger LTβR-dependent HEV differentiation in melanoma cannot be ruled out. Moreover, because skin inflammation is exacerbated in TNF−/− mice (29), proinflammatory cytokines may favor HEV differentiation in melanoma developed in mice with impaired TNF signaling.

Our study thus indicates that TNF may serve as a putative target for melanoma treatment. Accordingly, systemic administration of Etanercept significantly enhanced CD8+ T-cell–dependent immune response and reduced B16K1 tumor growth in mice with established tumors. It is tempting to speculate that anti–TNF-neutralizing molecules could be used in humans for treating immunogenic melanoma. Anti-TNF strategies may exert, at least, three beneficial effects in human melanoma by (i) preventing melanoma dedifferentiation (32), (ii) limiting the accumulation and biologic activity of immunosuppressor cells such as MDSC (30, 31), which produce TNF (47) and play a key role in melanoma (44), and (iii) facilitating CD8+ T-cell survival and melanoma infiltration, as documented here.

To our knowledge, only one clinical trial has shown stabilization of disease progression in 1 patient with metastatic melanoma treated by anti-TNF monoclonal antibodies (48). A critical finding of our study is that (i) TNF deficiency had no, or minimal, effect toward poorly immunogenic B16 melanoma cells and (ii) Etanercept failed to alter B16K1 melanoma growth in immunodeficient mice. Therefore, TNF-neutralizing molecules might be useful in nonimmunocompromized patients affected with immunogenic melanoma. Interestingly, HLA-I expression was detected on more than 50% of melanoma cells in 69% of melanoma metastases (49). Evaluating HLA-I expression in melanoma biopsies would help define eligibility of patients for TNF blockade treatment. In addition, emerging therapies for the treatment of melanomas, such as anti-CTLA4 (50), and BRAF V600E inhibitors (7, 8), are associated with CD8+ T-cell immune response and TNF production, even in patients with disease progression. Macrophage-derived TNF is important for BRAF V600E melanoma growth in vivo and confers resistance to MAPK pathway inhibitors (16). Whether a combination of anti-TNF and emerging therapies may exert a clinical benefit by enhancing CD8+ T-cell–dependent immune response toward melanoma remains to be investigated.

No potential conflicts of interest were disclosed.

Conception and design: F. Bertrand, C. Colacios, T. Levade, H. Benoist, B. Ségui

Development of methodology: F. Bertrand, J. Rochotte, C. Colacios, A.F. Tilkin-Mariamé, C. Touriol, I. Lajoie-Mazenc

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Bertrand, J. Rochotte, C. Colacios, A. Montfort, A.F. Tilkin-Mariamé, P. Rochaix

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Bertrand, J. Rochotte, C. Colacios, A.F. Tilkin-Mariamé, C. Touriol, P. Rochaix, N. Andrieu-Abadie, H. Benoist, B. Ségui

Writing, review, and/or revision of the manuscript: F. Bertrand, C. Colacios, N. Andrieu-Abadie, T. Levade, H. Benoist, B. Ségui

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): I. Lajoie-Mazenc

Study supervision: B. Ségui

The authors thank the flow cytometry, microscopy, and animal facilities of the I2MC (INSERM U1048, Toulouse, France) and the CRCT (INSERM U1037, Toulouse, France) for their technical assistance, Prof. J. van Meerwijk (INSERM U1043, Toulouse, France) for the kind gift of CD8-deficient mice, Drs. J.P Girard (IPBS, Toulouse, France), O. Micheau (INSERM U866, Dijon, France), P. Legembre (INSERM U1085, Rennes, France), V. Douin-Echinard (INSERM U1048, Toulouse, France), J.C. Guéry (INSERM U1043, Toulouse, France), and Prof. H. Prats (INSERM U1037, Toulouse, France) for fruitful discussions. They also acknowledge the assistance from R. Gence (INSERM U1037, Toulouse, France) and thank Drs. I. Escargueil-Blanc (Sanofi) and S. Tartare Deckert (INSERM U1065, Nice, France) for the kind gift of B16-BL6 and Yumm melanoma cell lines, respectively.

This work was supported by RITC (Recherche et Innovation Thérapeutique en Cancérologie), INSERM, and Paul Sabatier University (Toulouse III).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Carswell
EA
,
Old
LJ
,
Kassel
RL
,
Green
S
,
Fiore
N
,
Williamson
B
. 
An endotoxin-induced serum factor that causes necrosis of tumors
.
Proc Natl Acad Sci U S A
1975
;
72
:
3666
70
.
2.
Balkwill
F
. 
Tumour necrosis factor and cancer
.
Nat Rev Cancer
2009
;
9
:
361
71
.
3.
Mantovani
A
,
Allavena
P
,
Sica
A
,
Balkwill
F
. 
Cancer-related inflammation
.
Nature
2008
;
454
:
436
44
.
4.
Moore
RJ
,
Owens
DM
,
Stamp
G
,
Arnott
C
,
Burke
F
,
East
N
, et al
Mice deficient in tumor necrosis factor-alpha are resistant to skin carcinogenesis
.
Nat Med
1999
;
5
:
828
31
.
5.
Suganuma
M
,
Okabe
S
,
Marino
MW
,
Sakai
A
,
Sueoka
E
,
Fujiki
H
. 
Essential role of tumor necrosis factor alpha (TNF-alpha) in tumor promotion as revealed by TNF-alpha-deficient mice
.
Cancer Res
1999
;
59
:
4516
8
.
6.
Arnott
CH
,
Scott
KA
,
Moore
RJ
,
Robinson
SC
,
Thompson
RG
,
Balkwill
FR
. 
Expression of both TNF-alpha receptor subtypes is essential for optimal skin tumour development
.
Oncogene
2004
;
23
:
1902
10
.
7.
Hong
DS
,
Vence
L
,
Falchook
G
,
Radvanyi
LG
,
Liu
C
,
Goodman
V
, et al
BRAF(V600) inhibitor GSK2118436 targeted inhibition of mutant BRAF in cancer patients does not impair overall immune competency
.
Clin Cancer Res
2012
;
18
:
2326
35
.
8.
Wilmott
JS
,
Haydu
LE
,
Menzies
AM
,
Lum
T
,
Hyman
J
,
Thompson
JF
, et al
Dynamics of chemokine, cytokine, and growth factor serum levels in BRAF-mutant melanoma patients during BRAF inhibitor treatment
.
J Immunol
2014
;
192
:
2505
13
.
9.
Menon
DR
,
Wels
C
,
Bonyadi Rad
E
,
Joshi
S
,
Knausz
H
,
Lade-Keller
J
, et al
TGF-beta1 and TNF-alpha differentially regulate Twist1 mediated resistance towards BRAF/MEK inhibition in melanoma
.
Pigment Cell Melanoma Res
2013
;
26
:
912
6
.
10.
Li
B
,
Vincent
A
,
Cates
J
,
Brantley-Sieders
DM
,
Polk
DB
,
Young
PP
. 
Low levels of tumor necrosis factor alpha increase tumor growth by inducing an endothelial phenotype of monocytes recruited to the tumor site
.
Cancer Res
2009
;
69
:
338
48
.
11.
Ardestani
S
,
Li
B
,
Deskins
DL
,
Wu
H
,
Massion
PP
,
Young
PP
. 
Membrane versus soluble isoforms of TNF-alpha exert opposing effects on tumor growth and survival of tumor-associated myeloid cells
.
Cancer Res
2013
;
73
:
3938
50
.
12.
Sasi
SP
,
Yan
X
,
Enderling
H
,
Park
D
,
Gilbert
HY
,
Curry
C
, et al
Breaking the ‘harmony’ of TNF-alpha signaling for cancer treatment
.
Oncogene
2012
;
31
:
4117
27
.
13.
Chopra
M
,
Riedel
SS
,
Biehl
M
,
Krieger
S
,
von Krosigk
V
,
Bauerlein
CA
, et al
Tumor necrosis factor receptor 2-dependent homeostasis of regulatory T cells as a player in TNF-induced experimental metastasis
.
Carcinogenesis
2013
;
34
:
1296
303
.
14.
Waterston
AM
,
Salway
F
,
Andreakos
E
,
Butler
DM
,
Feldmann
M
,
Coombes
RC
. 
TNF autovaccination induces self anti-TNF antibodies and inhibits metastasis in a murine melanoma model
.
Br J Cancer
2004
;
90
:
1279
84
.
15.
Cubillos
S
,
Scallon
B
,
Feldmann
M
,
Taylor
P
. 
Effect of blocking TNF on IL-6 levels and metastasis in a B16-BL6 melanoma/mouse model
.
Anticancer Res
1997
;
17
:
2207
11
.
16.
Smith
MP
,
Sanchez-Laorden
B
,
O'Brien
K
,
Brunton
H
,
Ferguson
J
,
Young
H
, et al
The immune microenvironment confers resistance to MAPK pathway inhibitors through macrophage-derived TNFalpha
.
Cancer Discov
2014
;
4
:
1214
29
.
17.
Moulis
G
,
Sommet
A
,
Bene
J
,
Montastruc
F
,
Sailler
L
,
Montastruc
JL
, et al
Cancer risk of anti-TNF-alpha at recommended doses in adult rheumatoid arthritis: a meta-analysis with intention to treat and per protocol analyses
.
PLoS One
2012
;
7
:
e48991
.
18.
Mercer
LK
,
Lunt
M
,
Low
AL
,
Dixon
WG
,
Watson
KD
,
Symmons
DP
, et al
Risk of solid cancer in patients exposed to anti-tumour necrosis factor therapy: results from the British Society for Rheumatology Biologics Register for Rheumatoid Arthritis
.
Ann Rheum Dis
2015
;
74
:
1087
93
.
19.
Askling
J
,
Fahrbach
K
,
Nordstrom
B
,
Ross
S
,
Schmid
CH
,
Symmons
D
. 
Cancer risk with tumor necrosis factor alpha (TNF) inhibitors: meta-analysis of randomized controlled trials of adalimumab, etanercept, and infliximab using patient level data
.
Pharmacoepidemiol Drug Saf
2011
;
20
:
119
30
.
20.
Leombruno
JP
,
Einarson
TR
,
Keystone
EC
. 
The safety of anti-tumour necrosis factor treatments in rheumatoid arthritis: meta and exposure-adjusted pooled analyses of serious adverse events
.
Ann Rheum Dis
2009
;
68
:
1136
45
.
21.
Ratner
A
,
Clark
WR
. 
Role of TNF-alpha in CD8+ cytotoxic T lymphocyte-mediated lysis
.
J Immunol
1993
;
150
:
4303
14
.
22.
Kim
EY
,
Teh
HS
. 
TNF type 2 receptor (p75) lowers the threshold of T cell activation
.
J Immunol
2001
;
167
:
6812
20
.
23.
Chatzidakis
I
,
Fousteri
G
,
Tsoukatou
D
,
Kollias
G
,
Mamalaki
C
. 
An essential role for TNF in modulating thresholds for survival, activation, and tolerance of CD8+ T cells
.
J Immunol
2007
;
178
:
6735
45
.
24.
Braumuller
H
,
Wieder
T
,
Brenner
E
,
Assmann
S
,
Hahn
M
,
Alkhaled
M
, et al
T-helper-1-cell cytokines drive cancer into senescence
.
Nature
2013
;
494
:
361
5
.
25.
Ding
X
,
Yang
W
,
Shi
X
,
Du
P
,
Su
L
,
Qin
Z
, et al
TNF receptor 1 mediates dendritic cell maturation and CD8 T cell response through two distinct mechanisms
.
J Immunol
2011
;
187
:
1184
91
.
26.
Calzascia
T
,
Pellegrini
M
,
Hall
H
,
Sabbagh
L
,
Ono
N
,
Elford
AR
, et al
TNF-alpha is critical for antitumor but not antiviral T cell immunity in mice
.
J Clin Invest
2007
;
117
:
3833
45
.
27.
Zheng
L
,
Fisher
G
,
Miller
RE
,
Peschon
J
,
Lynch
DH
,
Lenardo
MJ
. 
Induction of apoptosis in mature T cells by tumour necrosis factor
.
Nature
1995
;
377
:
348
51
.
28.
Okubo
Y
,
Mera
T
,
Wang
L
,
Faustman
DL
. 
Homogeneous expansion of human T-regulatory cells via tumor necrosis factor receptor 2
.
Sci Rep
2013
;
3
:
3153
.
29.
Schioppa
T
,
Moore
R
,
Thompson
RG
,
Rosser
EC
,
Kulbe
H
,
Nedospasov
S
, et al
B regulatory cells and the tumor-promoting actions of TNF-alpha during squamous carcinogenesis
.
Proc Natl Acad Sci U S A
2011
;
108
:
10662
7
.
30.
Zhao
X
,
Rong
L
,
Li
X
,
Liu
X
,
Deng
J
,
Wu
H
, et al
TNF signaling drives myeloid-derived suppressor cell accumulation
.
J Clin Invest
2012
;
122
:
4094
104
.
31.
Sade-Feldman
M
,
Kanterman
J
,
Ish-Shalom
E
,
Elnekave
M
,
Horwitz
E
,
Baniyash
M
. 
Tumor necrosis factor-alpha blocks differentiation and enhances suppressive activity of immature myeloid cells during chronic inflammation
.
Immunity
2013
;
38
:
541
54
.
32.
Landsberg
J
,
Kohlmeyer
J
,
Renn
M
,
Bald
T
,
Rogava
M
,
Cron
M
, et al
Melanomas resist T-cell therapy through inflammation-induced reversible dedifferentiation
.
Nature
2012
;
490
:
412
6
.
33.
Porgador
A
,
Feldman
M
,
Eisenbach
L
. 
H-2Kb transfection of B16 melanoma cells results in reduced tumourigenicity and metastatic competence
.
J Immunogenet
1989
;
16
:
291
303
.
34.
Pequeux
C
,
Raymond-Letron
I
,
Blacher
S
,
Boudou
F
,
Adlanmerini
M
,
Fouque
MJ
, et al
Stromal estrogen receptor-alpha promotes tumor growth by normalizing an increased angiogenesis
.
Cancer Res
2012
;
72
:
3010
9
.
35.
Pencheva
N
,
Buss
CG
,
Posada
J
,
Merghoub
T
,
Tavazoie
SF
. 
Broad-spectrum therapeutic suppression of metastatic melanoma through nuclear hormone receptor activation
.
Cell
2014
;
156
:
986
1001
.
36.
Martinet
L
,
Le Guellec
S
,
Filleron
T
,
Lamant
L
,
Meyer
N
,
Rochaix
P
, et al
High endothelial venules (HEVs) in human melanoma lesions: major gateways for tumor-infiltrating lymphocytes
.
Oncoimmunology
2012
;
1
:
829
39
.
37.
Avram
G
,
Sanchez-Sendra
B
,
Martin
JM
,
Terradez
L
,
Ramos
D
,
Monteagudo
C
. 
The density and type of MECA-79-positive high endothelial venules correlate with lymphocytic infiltration and tumour regression in primary cutaneous melanoma
.
Histopathology
2013
;
63
:
852
61
.
38.
Cipponi
A
,
Mercier
M
,
Seremet
T
,
Baurain
JF
,
Theate
I
,
van den Oord
J
, et al
Neogenesis of lymphoid structures and antibody responses occur in human melanoma metastases
.
Cancer Res
2012
;
72
:
3997
4007
.
39.
Schrama
D
,
Voigt
H
,
Eggert
AO
,
Xiang
R
,
Zhou
H
,
Schumacher
TN
, et al
Immunological tumor destruction in a murine melanoma model by targeted LTalpha independent of secondary lymphoid tissue
.
Cancer Immunol Immunother
2008
;
57
:
85
95
.
40.
Popivanova
BK
,
Kitamura
K
,
Wu
Y
,
Kondo
T
,
Kagaya
T
,
Kaneko
S
, et al
Blocking TNF-alpha in mice reduces colorectal carcinogenesis associated with chronic colitis
.
J Clin Invest
2008
;
118
:
560
70
.
41.
Chen
X
,
Baumel
M
,
Mannel
DN
,
Howard
OM
,
Oppenheim
JJ
. 
Interaction of TNF with TNF receptor type 2 promotes expansion and function of mouse CD4+CD25+ T regulatory cells
.
J Immunol
2007
;
179
:
154
61
.
42.
Youn
JI
,
Collazo
M
,
Shalova
IN
,
Biswas
SK
,
Gabrilovich
DI
. 
Characterization of the nature of granulocytic myeloid-derived suppressor cells in tumor-bearing mice
.
J Leukoc Biol
2012
;
91
:
167
81
.
43.
Hu
X
,
Li
B
,
Li
X
,
Zhao
X
,
Wan
L
,
Lin
G
, et al
Transmembrane TNF-alpha promotes suppressive activities of myeloid-derived suppressor cells via TNFR2
.
J Immunol
2014
;
192
:
1320
31
.
44.
Meyer
C
,
Sevko
A
,
Ramacher
M
,
Bazhin
AV
,
Falk
CS
,
Osen
W
, et al
Chronic inflammation promotes myeloid-derived suppressor cell activation blocking antitumor immunity in transgenic mouse melanoma model
.
Proc Natl Acad Sci U S A
2011
;
108
:
17111
6
.
45.
Moussion
C
,
Girard
JP
. 
Dendritic cells control lymphocyte entry to lymph nodes through high endothelial venules
.
Nature
2011
;
479
:
542
6
.
46.
Hohmann
HP
,
Remy
R
,
Poschl
B
,
van Loon
AP
. 
Tumor necrosis factors-alpha and -beta bind to the same two types of tumor necrosis factor receptors and maximally activate the transcription factor NF-kappa B at low receptor occupancy and within minutes after receptor binding
.
J Biol Chem
1990
;
265
:
15183
8
.
47.
Sevko
A
,
Michels
T
,
Vrohlings
M
,
Umansky
L
,
Beckhove
P
,
Kato
M
, et al
Antitumor effect of paclitaxel is mediated by inhibition of myeloid-derived suppressor cells and chronic inflammation in the spontaneous melanoma model
.
J Immunol
2013
;
190
:
2464
71
.
48.
Brown
ER
,
Charles
KA
,
Hoare
SA
,
Rye
RL
,
Jodrell
DI
,
Aird
RE
, et al
A clinical study assessing the tolerability and biological effects of infliximab, a TNF-alpha inhibitor, in patients with advanced cancer
.
Ann Oncol
2008
;
19
:
1340
6
.
49.
Turcotte
S
,
Gros
A
,
Hogan
K
,
Tran
E
,
Hinrichs
CS
,
Wunderlich
JR
, et al
Phenotype and function of T cells infiltrating visceral metastases from gastrointestinal cancers and melanoma: implications for adoptive cell transfer therapy
.
J Immunol
2013
;
191
:
2217
25
.
50.
Yuan
J
,
Gnjatic
S
,
Li
H
,
Powel
S
,
Gallardo
HF
,
Ritter
E
, et al
CTLA-4 blockade enhances polyfunctional NY-ESO-1 specific T cell responses in metastatic melanoma patients with clinical benefit
.
Proc Natl Acad Sci U S A
2008
;
105
:
20410
5
.