IL-17 can promote tumor growth through an IL-6–Stat3 signaling pathway

We assessed the role of IL-17 in modulating tumor growth in mice with genetically ablated IL-17 alleles (IL-17 KO). Because IFN-γ–producing Th1 T cells generally provide antitumor immunity (Kaplan et al., 1998; Dunn et al., 2006) and are known to restrain Th17 cell development (Harrington et al., 2005), we also evaluated tumor growth in IFN-γ KO mice and in IL-17IFN-γ double KO mice. Sex- and age-matched WT, IL-17^−/−, IFN-γ^−/−, and IFN-γ^−/−IL-17^−/− B6 mice were challenged with B16 melanoma. Results from these experiments showed that tumor growth rate was reduced in the B6 IL-17^−/− mice when compared with WT B6 controls (Fig. 1 A). As expected from the antitumor role of IFN-γ (Kaplan et al., 1998; Dunn et al., 2006), in B6 mice lacking IFN-γ alleles, B16 tumors appeared earlier and the tumor growth rate was drastically increased relative to the WT (Fig. 1 A). Importantly, in IFN-γ^−/−IL-17^−/− mice, tumor growth was significantly reduced compared with IFN-γ^−/− mice (Fig. 1 A), supporting a tumor-promoting role of Th17 cells in the B16 subcutaneous tumor model.

We also evaluated growth of the MB49 bladder carcinoma in WT, IL-17^−/−, IFN-γ^−/−, and IFN-γ^−/−IL-17^−/− mice. Similar to B16, we observed growth inhibition of MB49 tumors in IL-17^−/− mice relative to WT controls (Fig. 1 B). Furthermore, MB49 tumor growth was significantly accelerated in IFN-γ^−/− relative to WT mice. Ablating IL-17 diminished accelerated tumor growth caused by a lack of IFN-γ (Fig. 1 B). However, the tumor growth rate in the IFN-γ^−/−IL-17^−/− B6 mice was higher than in IL-17^−/− mice. Thus, the general pattern of tumor growth in IL-17^−/−, IFN-γ^−/−, and IFN-γ^−/−IL-17^−/− mice was similar for B16 and MB49 tumors despite their different tissue origins. Subcutaneous MB49 tumors also naturally metastasize to lungs; we therefore assessed whether levels of lung metastasis correlate with tumor size in the four groups of mice. Ablating IL-17 in both WT and IFN-γ^−/− mice was associated with a decreased number of lung metastases (Fig. 1, C and D).

To assess whether T cells were a primary source of IL-17 for promoting tumor growth, we reconstituted Rag2^−/− mice by adoptive T cell transfer with purified T cells from IFN-γ^−/− mice that display high levels of IL-17 production (Fig. 2), or from IFN-γ^−/−IL-17^−/− mice. Tumor growth in mice receiving IFN-γ^−/−IL-17^−/− T cells was significantly reduced compared with mice given IFN-γ^−/− T cells (Fig. 1 E), suggesting that the increased tumor growth, caused by elevated IL-17 production, was mainly contributed by T cells.

We performed intracellular cytokine staining and in vitro cytokine production assays to compare the cytokine profile of CD4⁺ and CD8⁺ T cells in tumors harvested from WT, IL-17^−/−, IFN-γ^−/−, and IFN-γ^−/−IL-17^−/− mice. IL-17 production in B16 tumors was markedly augmented in the absence of IFN-γ, with an increase in the number of tumor-infiltrating Th17 cells (Fig. 2 A, top). IL-17 expression was also elevated in MB49 tumors grown in IFN-γ^−/− mice (Fig. 2 A, bottom). Further, ELISA analyses indicated a >10-fold increase in secreted IL-17 by B16 tumor–infiltrating CD4⁺ T cells in IFN-γ^−/− mice, which was again abrogated in IFN-γ^−/−IL-17^−/− double KO mice (Fig. 2 B). Notably, we observed only minimal levels of IL-17 production by CD8⁺ T cells in both B16 and MB49 tumors (Fig. S1).

We observed an increased infiltration of CD4⁺ and CD8⁺ T cells in tumors grown in IL-17^−/− mice (Fig. 2 C, top two rows). Furthermore, tumor-infiltrating CD8⁺ and CD4⁺ T cells from IL-17^−/− mice produced more IFN-γ relative to WT control mice, as assessed by intracellular cytokine staining (Fig. 2 C). The increase in IFN-γ production by both tumor-infiltrating CD4⁺ and CD8⁺ T cells from IL-17^−/− mice was confirmed by ELISA (Fig. 2 D). IL-17 ablation only modestly increased the percentage of tumor-infiltrating regulatory T cells (Fig. S2 A). It has been reported that IL-17 ablation does not affect the function of regulatory T cells (Yi et al., 2008). No correlation was observed with Th2-type immune responses as measured by CD4⁺ T cell–mediated IL-4 production caused by IL-17 and IFN-γ ablation (Fig. S2 B).

The underlying mechanisms by which Th17 promoted tumor growth was still unknown. Because Stat3 activation in tumor cells and tumor-associated inflammatory cells plays a critical role in tumor progression by augmenting tumor survival and tumor angiogenesis, and suppressing antitumor immunity (Yu et al., 2007), we explored the possibility that IL-17 mediates tumor growth via activation of Stat3. Stat3 activity in B16 tumors from WT, IL-17^−/−, IFN-γ^−/−, and IFN-γ^−/−IL-17^−/− mice was examined by immunofluorescence staining of phosphorylated Stat3 (phospho-Stat3) in situ. We observed reduced phospho-Stat3 levels in B16 tumors grown in IL-17^−/− mice compared with WT controls. In addition, Stat3 activity was drastically increased in tumors grown in IFN-γ^−/− mice, suggesting an inhibitory role of IFN-γ on tumor Stat3 activation. Stat3 phosphorylation was diminished in tumors harvested from IFN-γ^−/−IL-17^−/− compared with IFN-γ^−/− mice (Fig. 3 A). Immunofluorescence staining results were confirmed by Western blotting using an antibody for phospho-Stat3 (Fig. 3 B). Collectively, these results indicated that the levels of IL-17 positively correlated with Stat3 activity in growing tumors.

We next tested whether increased IL-17 expression affected downstream Stat3 target genes. We found that Bcl family members, including Bcl-2 and Bcl-x_L, were up-regulated in tumors from IFN-γ^−/− mice (Fig. 3 C). In addition, vascular endothelial growth factor and MMP9 were up-regulated in tumors from IFN-γ^−/− mice, and their expression was reduced in tumors from IFN-γ^−/−IL-17^−/− mice (Fig. 3 D). These results suggested that IL-17 promoted Stat3 activity in tumors, leading to up-regulation of antiapoptotic and angiogenic genes.

To investigate whether IL-17 signals directly through Stat3, we cultured B16 tumor cells, normal fibroblasts, endothelial cells, splenic DCs, and non-DC myeloid cells in the presence of recombinant mouse IL-17 or control. Results from these experiments showed that IL-17 indeed induced Stat3 activity (Fig. 4 A). Although IL-17 induced Stat3 activity in all tested target cells, it did so with significantly different kinetics (Fig. 4 A). Normal fibroblasts and endothelial cells were included to test IL-17 response, as they are major components of the tumor microenvironment. Similarly, we found that human recombinant IL-17 significantly stimulated Stat3 phosphorylation in several human cancer cell lines (Fig. S3 A). We further confirmed the expression of IL-17 receptor in tumor cells and fibroblasts (Fig. S4). These results suggested that induction of Stat3 activation in tumor cells and elements of the tumor microenvironment by IL-17 occurred through induction of an intermediary factor rather than directly.

IL-6 is a Stat3 activator and is elevated in diverse cancers (Hirano et al., 2000). IL-17 has recently been reported to stimulate production of IL-6 and Stat3 activation through a positive feedback loop, in inflammatory cells as well as fibroblasts, in an autoimmune disease setting (Ogura et al., 2008). We therefore determined whether IL-6 mediated IL-17–driven Stat3 activation in a tumor setting. We found that IL-17 stimulated IL-6 production by B16 and MB49 tumor cells in vitro, in addition to other cell types that constitute the tumor stroma such as fibroblasts, endothelial cells, and DCs (Fig. 4, B and C). The levels of IL-6 per cell produced by the tested tumor cells were lower than those produced by fibroblasts and endothelial cells. Although treating the tumor cells with recombinant IL-17 only modestly increased their proliferation (Fig. S5 A), which is consistent with another study (Numasaki et al., 2005), similar treatment of endothelial cells led to increased expression of several Stat3 downstream genes that are involved in cell migration (Fig. S5 B). In addition, IL-17–mediated increased secretion of angiogenic factors by endothelial cells induced endothelial cell migration in a Stat3-dependent manner (Fig. S5 C). Importantly, neutralizing IL-6 but not IL-10 (another Stat3 activator) with a blocking mAb reduced Stat3 activation in (mouse embryonic fibroblasts) MEFs, DCs, and MB49 tumor cells after IL-17 stimulation (Fig. 4 D). Similarly, anti–IL-6 antibodies abrogated IL-17–induced Stat3 activation in human cancer cells (Fig. S3 B). These results indicated that IL-17 activated Stat3 through IL-6 in diverse types of cancer and nontransformed cells constituting the tumor stroma.

We first assessed whether IL-17 affected IL-6 production by tumors in vivo. Freshly harvested B16 tumor cells from WT mice produced relatively high levels of IL-6, which were reduced after IL-17 ablation (Fig. 5 A). In addition, tumors grown in IFN-γ^−/− mice produced a markedly increased level of IL-6 when compared with the IFN-γ^−/−IL-17^−/− mice (Fig. 5 A). More detailed analysis of growing tumors indicated that tumor-associated immune cells (CD45⁺) were the main source of IL-6 when compared with the tumor cells (Fig. 5 B). To directly test whether IL-17 could promote tumor growth through IL-6–mediated Stat3 activation, IFN-γ^−/− mice were challenged with B16 melanoma, followed by administration of IL-6–neutralizing mAbs or control rat IgG. IFN-γ^−/−IL-17^−/− mice challenged with B16 melanoma but without anti–IL-6 antibody treatment were included as an additional control. Because IL-6 was shown to promote Th17 cell differentiation, we did not start anti–IL-6 treatment until 6 d after tumor inoculation, a time at which tumor-infiltrating Th17 cells were detectable (unpublished data). We found that administration of monoclonal anti–IL-6 but not rat IgG antibodies significantly inhibited tumor growth in IFN-γ^−/− mice (Fig. 5 C), suggesting that IL-17–induced tumor progression is at least in part mediated by IL-6. Consistent with a role for IL-6 in mediating tumor Stat3 activation and tumor growth promotion, neutralizing IL-6 in vivo reduced Stat3 activity and the expression of its downstream antiapoptotic gene, Bcl-x_L, in tumors (Fig. 5 D).

In summary, we have shown that IL-17 can promote tumor growth. Our results suggested that the IL-17–mediated tumor-promoting role involves a direct effect on tumor through IL-6 induction, which in turn activates Stat3 in both tumor as well as nontransformed cells constituting the tumor microenvironment. IL-6 activation of Stat3 in tumor cells results in increases in antiapoptotic, proproliferation, and proangiogenic genes, as well as suppression of certain proinflammatory genes (Yu et al., 2007). IL-6–induced Stat3 signaling in T cells is also critical in promoting Th17 cell differentiation and expression (Chen et al., 2006; Harris et al., 2007). Given that TGF-β, the other cytokine involved in Th17 cell differentiation, is also expressed at high levels in tumors, IL-17–induced IL-6 production can generate an autoamplification loop for Th17 cells in the tumor microenvironment. Consistent with our findings, several publications suggested a role of IL-17 in promoting tumor growth (Numasaki et al., 2003; Numasaki et al., 2005; Nam et al., 2008; Xiao et al., 2009), although the mechanism was not well addressed in these studies. However, a recent paper showed that, in a different tumor system (MC38 sarcoma), tumors growth was increased in IL-17^−/− mice (Kryczek et al., 2009). These findings suggest that the role of IL-17 in cancer is context and system dependent, like many other cytokines such as TNF-α. Additional studies are required to reveal why IL-17 can have opposite roles in modulating growth in different tumor systems.

The opposing functions of IFN-γ and IL-17 as anticarcinogenic and protumor growth effector cytokines extend an emerging concept that the qualitative nature of the immune response determines whether it will predominantly inhibit tumor growth or promote cancer development. Recently, analogous opposing effects of the two IL-12 family members, IL-12 and IL-23, in tumor inhibition versus promotion have been ascertained based on enhanced tumor formation in IL-12p35 KO mice versus diminished tumor formation in IL-23p19 KO mice (Langowski et al., 2006). We subsequently reported that Stat3 signaling in tumors and their microenvironment determined the balance between these cytokines by transcriptionally activating IL-23p19 and inhibiting IL-12p35 (Kortylewski et al., 2009). The current work is among the first examples by which a qualitatively distinct (Th17) endogenous T cell immune response can be cancer promoting. Further, it adds a new element to the link between immune responses that either depend on or induce Stat3 signaling and Stat3-mediated protumor growth.

WT and IFN-γ^−/− C57BL/6 mice were purchased from the National Cancer Institute and the Jackson Laboratory, respectively. The generation of C57BL/6 IL-17^−/− mice has been previously reported (Nakae et al., 2002), and the mice were provided by Y. Iwakura (University of Tokyo, Tokyo, Japan). Mouse care and experimental procedures were performed under pathogen-free conditions in accordance with established institutional guidance and approved protocols from the Research Animal Care Committee of the City of Hope Medical Center. For tumor challenge, 10⁵ B16 tumor cells were injected subcutaneously into 8–12-wk-old WT or transgenic mice, and tumor growth was monitored every other day. For IL-6 neutralization experiments in vivo, mice were treated with 125 µg anti–IL-6 antibodies (eBioscience) or control rat IgG (Jackson ImmunoResearch Laboratories) injected i.v. on days 6, 8, 10, 12, and 14 after tumor inoculation. Mice were sacrificed after 2–3 wk from tumor inoculation or when tumor volume exceeded 1.5 cm in diameter.

B16 and MB49 cell lines were originally obtained from American Type Culture Collection and were maintained in our laboratory. Endothelial cell lines derived from mouse prostate were provided by J. Fidler (M.D. Anderson Cancer Center, Houston, TX).

T cells were enriched from IFN-γ^−/− or IFN-γ^−/−IL-17^−/− mice. Specifically, non–T cells were depleted with biotin-labeled mAbs, including anti-B220, Gr-1, Mac-1, CD11c, DX5, and Ter119, using a magnetic purification system from Miltenyi Biotec. TCRβ⁺ T cell purity was >90%, as determined by flow cytometry.

The procedure for isolating immune cells from tumors has been described previously (Kortylewski et al., 2005). In brief, freshly excised tumor tissues were gently minced and digested for 30 min at 37°C by collagenase D/DNase I (Roche). Cell suspensions were filtered through a 70-µm cell strainer, and dead cells were removed by centrifugation at 600 g over a Histopaque gradient (Sigma-Aldrich).

The FITC-, PE-, allophycocyanin (APC)-, or Cy7-APC–conjugated antibodies specific to CD4, CD8, IL-17, and IFN-γ were purchased from BD; PE-labeled antibody against mouse Foxp3 was purchased from eBioscience, and PE-labeled antibody for IL-17R was purchased from R&D Systems. Intracellular staining for IL-17 and IFN-γ was performed according to the manufacturer’s instructions (BD; Yi et al., 2008). In brief, cells were stimulated for 5 h with PMA and ionomycin (Sigma-Aldrich) in the presence of Golgi-Stop reagent (BD). Cells were harvested, washed, and stained with anti-CD4 or -CD8 in the presence of FcR-Block (BD). After the wash, cells were fixed, permeabilized, and stained with cytokine-specific or control isotype antibodies for 30 min on ice. Multiple-color FACS analysis was performed using a three-laser CyAn Immunocytometry System (Dako). Dead cells were excluded by the Fixable Aqua Dead Cell Stain Kit (Invitrogen).

10⁵ sorted tumor-infiltrating CD4⁺, CD8⁺ T cells were cultured for 24 h in a U-bottom 96-well plate with plate-bound CD3 and 2 µg/ml of soluble CD28. 10⁶ enriched splenic CD11c⁺ or cultured cells/ml were stimulated for 24 h with 10 ng/ml of recombinant IL-17. IFN-γ, IL-17, IL-6, and IL-4 concentrations in culture supernatants were measured using ELISA kits from R&D Systems.

5-µm sections of flash-frozen tumor specimens were fixed in acetone, permeabilized with methanol, stained with antibodies specific to tyrosine–phospho-Stat3 (Santa Cruz Biotechnology, Inc.) and detected with secondary antibodies conjugated Alexa Fluor 488 (Invitrogen), as previously described (Kortylewski et al., 2005). After staining with Hoechst 33342 (Invitrogen) to visualize cell nuclei, slides were mounted and analyzed by fluorescence microscopy.

The unpaired t test was used to calculate the two-tailed p-value. The two-way analysis of variance test was used to calculate the p-value for tumor growth. Data were analyzed using Prism software (GraphPad Software, Inc.).

Fig. S1 shows the production of IL-17 in tumor-infiltrating CD8⁺ T cells, determined by intracellular staining and ELISA. Fig. S2 indicates that ablating IFN-γ or IL-17 did not significantly affect the T reg cell percentage and IL-4 production. Fig. S3 shows that IL-17 activates Stat3 in human tumor cell lines in an IL-6–dependent manner. Fig. S4 confirms IL-17 receptor expression on B16 tumor cells and MEFs. Fig. S5 depicts the effect of IL-17 on tumor cell proliferation and endothelial cell migration.

We thank Dr. Y. Iwakura for generously providing us with IL-17^−/− mice; Dr. Y. Liu for superb assistance; the members of the Flow Cytometry Core, the Pathology Core, and the Animal Facility at the City of Hope Medical Center for their contributions; and Dr. S. da Costa for editing the manuscript.

This study was supported by the National Institutes of Health (grants R01CA122976, R01CA115815, and R01 AI066008).

The authors have no conflicting financial interests.