Although Th17 cells play critical roles in the pathogenesis of many inflammatory and autoimmune diseases, their prevalence among tumor-infiltrating lymphocytes (TILs) and function in human tumor immunity remains largely unknown. We have recently demonstrated high percentages of Th17 cells in TILs from ovarian cancer patients, but the mechanisms of accumulation of these Th17 cells in the tumor microenvironment are still unclear. In this study, we further showed elevated Th17 cell populations in the TILs obtained from melanoma and breast and colon cancers, suggesting that development of tumor-infiltrating CD4+ Th17 cells may be a general feature in cancer patients. We then demonstrated that tumor microenvironmental RANTES and MCP-1 secreted by tumor cells and tumor-derived fibroblasts mediate the recruitment of Th17 cells. In addition to their recruitment, we found that tumor cells and tumor-derived fibroblasts produce a proinflammatory cytokine milieu as well as provide cell–cell contact engagement that facilitates the generation and expansion of Th17 cells. We also showed that inflammatory TLR and nucleotide oligomerization binding domain 2 signaling promote the attraction and generation of Th17 cells induced by tumor cells and tumor-derived fibroblasts. These results identify Th17 cells as an important component of human TILs, demonstrate mechanisms involved in the recruitment and regulation of Th17 cells in tumor microenvironments, and provide new insights relevant for the development of novel cancer immunotherapeutic approaches.
Recent discovery of Th17 cells has resulted in an explosion of immunological research, and it is now widely accepted that the Th17 subset is an independent lineage of Th cells in humans and mice based on their unique cytokine profile, transcriptional regulation, and biological function (1–3). The identification of Th17 cells not only changes the classical Th1/Th2 paradigm of Th cell differentiation but also markedly facilitates our understanding of human immunity under both physiological and pathological conditions (3, 4). Accumulating data suggest that Th17 cells play an important role in host defense against microbial infections, including bacteria, mycobacteria, viruses, and parasites, and appear to be important mediators in the pathogenesis of a broad array of inflammatory and autoimmune diseases such as rheumatoid arthritis, psoriasis, and inflammatory bowel disease (5).
The differentiation and regulation of murine Th17 cells have been extensively studied in the past 3 years. It is now known that transcription factors RORγt and STAT3 are critical and required for the development of Th17 cells (6, 7). Recent studies further demonstrated that IFN regulatory factor-4 is also necessary for Th17 cell differentiation (8). TGF-β and IL-6, or TGF-β and IL-21, are critical cytokines for the initiation of mouse Th17 cell differentiation (9–13). Furthermore, IL-23 is not required in vitro for Th17 cell induction but is critical for in vivo pathogenic effectors of Th17 cells (14, 15). Compared with mouse Th17 cells, requirements for the development of human Th17 cells have not been well studied (16, 17). Recent studies demonstrated that IL-1 is a critical inducer for human Th17 cell differentiation, and the combination of IL-1, IL-6, and IL-23 is an optimal cytokine milieu for human Th17 generation (18). More recently, studies from three groups demonstrated that TGF-β is also required for human Th17 cell differentiation (19–21). In addition to the direct regulation of Th17 cells by cytokines, recent data have also shown that modulation of APCs, such as dendritic cells, through the triggering of TLRs and nucleotide oligomerization binding domain (Nod)2 receptor can promote the differentiation and production of human Th17 cells (22–24). Improved understanding of human Th17 cell regulation is essential for the development of novel therapeutic strategies for treatment of a variety of diseases.
Despite the important role of Th17 cells in host protection against infectious pathogens and in the pathogenesis of various inflammatory and autoimmune diseases, their prevalence and function in human cancer is still under investigation. Although several studies have shown the presence of Th17 cells in some types of human cancers, little is known regarding their generation and regulation within the tumor microenvironment (25–27). In a recent study, we demonstrated high percentages of Th17 cells in TILs from ovarian cancer patients (28). However, whether the prevalence of Th17 cells is a general feature in human cancers other than ovarian cancer is still unclear. Furthermore, the mechanisms of accumulation of Th17 cells and their functional role in the tumor microenvironment remain largely unknown.
In the current study, we extend our observations in ovarian cancer patients to other types of cancers, including melanoma, breast cancer, and colon cancer. We demonstrate elevated CD4+ Th17 cell populations in the TILs from these cancers, suggesting that development of tumor-infiltrating Th17 cells may be a general feature in cancer patients. We further demonstrate that tumor microenvironmental RANTES and MCP-1 secreted by tumor cells and tumor-derived fibroblasts mediate the recruitment of Th17 cells. In addition to their recruitment, we show that tumor cells and tumor-derived fibroblasts produce a proinflammatory cytokine milieu as well as provide cell–cell contact engagement that facilitates the generation and expansion of Th17 cells. Furthermore, we also show that inflammatory TLR and Nod2 signaling promote the attraction and generation of CD4+ Th17 cells induced by tumor cells and tumor-derived fibroblasts. These results identify the potential mechanisms for the accumulation and regulation of Th17 cells in tumor microenvironments and provide the foundation for studies to further investigate the role of Th17 cells in antitumor immunity.
Materials and Methods
Human samples and cell lines
Tumor samples of melanoma, breast cancer, and colon cancer and patient data were obtained from hospitalized patients undergoing surgery in the Department of Surgery at St. Louis University Hospital as approved by the Institutional Review Board and ethics committee of the institution. Tumor-paired normal tissues were also obtained as controls. If we could not obtain paired normal tissue, we used tumor adjacent tissues as tissue controls.
Buffy coats from healthy donors were obtained from the St. Louis Red Cross. Cord blood samples were obtained from the St. Louis Cord Blood Bank. These studies were approved by the Institutional Review Boards. PBMCs or cord blood mononuclear cells were purified from buffy coats or cord blood using Ficoll-Paque. Bulk CD4+ T cells were isolated by either positive or negative selection with microbeads (Miltenyi Biotec, Auburn, CA) according to manufacturer’s instructions. CD4+CD25+ regulatory T (Treg) cells were further purified from CD4+ T cells by FACS sorting after staining with anti-CD25–PE (BD Biosciences, San Jose, CA). Human naive CD4+ T cells (CD45RA+CD45RO−) were purified by EasySep enrichment kits (StemCell Technologies, Vancouver, British Columbia, Canada). The purity of naive CD4+ T cells was >97%, as confirmed by flow cytometry.
Different types of tumor cell lines (breast cancer cell [BC], melanoma cell [MC], and colon cancer cell [CC]), tumor-derived fibroblasts (MF, BF, and CF) and patient-matched normal tissue-derived fibroblasts (MNF, BNF, and CNF), including melanoma, breast cancer, and colon cancer, were established in our laboratory. Fibroblast cell lines for the study were established from primary cells and used within five passages. BC and CC lines were maintained in keratinocyte medium containing 25 mg/ml bovine pituitary extract, 5 ng/ml epidermal growth factor, 2 mM l
Generation of TILs and Th17 cell cloning
Tumor and normal tissue-infiltrating lymphocytes were generated from different tumor and tumor-paired normal tissues, as described previously (29). Briefly, tissues were minced into small pieces followed by digestion with collagenase type IV, hyaronidase, and DNase. After digestion, cells were washed in RPMI 1640 and then cultured in RPMI 1640 containing 10% human serum supplemented with l-glutamine, 2-ME, and 50 U/ml IL-2 for the generation of T cells. The percentages of CD4+ Th17 cells were determined from bulk T cells by FACS analysis after intracellular staining for IL-17. Th17 cell clones were generated from TILs by a limiting dilution cloning method, as described previously (29).
Coculture for generation of Th17 cells
Naive CD4+ T cells were stimulated with plate-bound anti-CD3 Ab (OKT3; 2 μg/ml) in T cell medium containing 50 U/ml IL-2 and 10% human AB serum in 24-well plates for 24 h. OKT3-stimulated naive CD4+ T cells (2 × 105/well) were then cocultured with tumor cell lines, tumor-derived and/or normal tissue-derived fibroblasts at a ratio of 1:1 in T cell culture medium with 20 U/ml IL-2 in 24-well plates in the presence or absence of TLR and Nod ligands, including Pam3CSK4 (200 ng/ml), poly(I:C) (25 μg/ml), LPS (100 ng/ml), flagellin (10 μg/ml), loxoribine (500 μm), R848 (10 μg/ml), CpG-B (3 μg/ml), and muramyldipeptide (MDP) (200 ng/ml) (InvivoGen, San Diego, CA). At days 3 and 6, half of the culture media was collected and replaced with fresh medium. The collected media were stored at −80°C for further cytokine analysis. On day 7, cocultured T cells were harvested and determined the generation of IL-17–producing cells. To elucidate the mechanism mediated by tumor cells and tumor-derived fibroblasts for generation of Th17 cells, we also performed Transwell experiments in 24-well plates (pore size 0.4-μm insert chamber; Corning Costar, Corning, NY), as described previously (29).
Flow cytometry and Abs
28). All stained cells were analyzed on a FACSCalibur flow cytometer (BD Biosciences) and data analyzed with FlowJo software (Tree Star, Ashland, OR).
ELISA and cytokine Ab array
Tumor-infiltrating Th17 cells (1 × 1066
Functional proliferation assay
Proliferation assays were performed as described previously (29, 30). In brief, 1 × 105 naive CD4+ T cells freshly purified from healthy donors were cocultured with Th17 cell clones or control T cells at different ratios (1:0, 1:0.1, 1:0.2, 1:0.5, and 0:1) in anti-CD3–coated (2 μg/ml) 96-well plates in T cell assay medium containing 2% human AB serum (no IL-2). After 56 h of culture, [3H]thymidine was added at a final concentration of 1 μCi/well, followed by an additional 16 h of culture. The incorporation of [3H]thymidine was measured with a liquid scintillation counter.
Chemotaxis assays were performed using 24-well Transwell chemotaxis plates (5-μm pore size; Corning Costar) as described previously (315 cells) were transferred into upper chambers. After 150 min at 37°C, chemotaxis was quantified by detecting the numbers of cells that migrated into the lower chamber. The chemotaxis index was calculated by dividing the numbers of cells migrated in response to test supernatants or recombinant human chemokines by the numbers of cells migrated in response to medium alone.
In some experiments, tumor cell lines and tumor-derived fibroblasts (0.5 × 106
Unless indicated otherwise, data were expressed as means ± SD. Paired or unpaired two-tailed Student t test was used to analyze differences between two groups. Differences were considered significant for values of p < 0.05.
Development of tumor-infiltrating CD4+ Th17 cells is a general feature in cancer patients
The discovery of Th17 cells has advanced our understanding of the immunopathogenesis of inflammatory and autoimmune diseases (5). The existence and role of Th17 cells in tumor immunity remain elusive, although Th17 cells are likely to be important given that tumor development involves chronic inflammatory process (32, 33). We have previously showed the prominence of Th17 cells at ovarian cancer tumor sites (28). This observation prompted us to further determine the generality of development of Th17 cells in other types of cancer. We first generated TIL lines from fresh tumor tissues obtained from breast cancer, melanoma, and colon cancer patients. In parallel, we generated tissue-infiltrating T cell lines from patient-matched normal or tumor-adjacent tissues. We then determined the percentages of IL-17–producing cells in the TILs and tumor-matched normal tissue-infiltrating T lymphocytes. As shown in Fig. 1A, we determined markedly elevated IL-17–producing cell populations in all TILs from breast cancer (1.5–8%, mean 3.8%) and colon cancer (8.6–46.5%, mean 26.3%), and in 7 of the 10 melanoma-derived TILs (0.5–10.4%, mean 2.8%). In contrast, IL-17–producing cell populations in the tumor-matched normal tissue-infiltrating T cells were <1% in breast and melanoid normal tissues and ∼2% in normal colon tissues. Recent studies have suggested that IL-17 is not exclusively produced by CD4+ Th17 cells, and it can also be secreted by other cell types, including CD8+ and γδ T cells (34, 35). Therefore, we further analyzed CD4+IL-17+ T cell populations (Th17) in the TILs from these three types of cancers. Consistent with the results shown in Fig. 1A, we found significantly increased proportions of Th17 cells in CD4+ TILs from all tumor samples of breast cancer (mean 5.8%), melanoma (mean 6.5%), and colon cancer (mean 42%) (Fig. 1B, 1C). In contrast, the CD4+ Th17 cells in tumor-matched normal tissue-infiltrating CD4+ T cells were expressed at very low levels (Fig. 1B, 1C). It is notable that CD4+ Th17 cell proportions varied among tumor types of different origins. Our results revealed that much higher percentages of Th17 cells existed in CD4+ TILs from colon and ovarian cancers compared with breast cancer and melanoma (Fig. 1B, 1C and Ref. 28). Moreover, the percentage of CD4+ Th17 cells in normal colon tissue-infiltrating T cells (mean 4.3%) was also slightly higher than those in normal breast and melanoid tissues (<1%) (Fig. 1C), which suggests that more chronic inflammation may occur in certain organs, such as the gut (36).
Taken together, our studies demonstrated that elevated CD4+ Th17 cell populations were present in the tumor sites of melanoma and ovarian, breast, and colon cancers. This suggests that the development of tumor-infiltrating Th17 cells is a general feature of cancer. In support of this notion, studies from other groups have also revealed increased expression of Th17 cells in patients with gastric, prostate, ovarian, renal cell, and pancreatic cancers (25–27).
Phenotypic and functional features of tumor-infiltrating CD4+ Th17 cells
To date, only limited information exists regarding human Th17 cells that originate from normal or inflamed tissues (18, 36, 37). We next sought to investigate whether tumor-derived Th17 cells have characteristics similar to those found in inflammatory and autoimmune pathologic sites. To better characterize phenotypic and functional features of tumor-infiltrating Th17 cells, we generated Th17 cell clones from bulk TIL lines containing high levels of IL-17–producing CD4+ T cells derived from melanoma and breast and colon cancers using a limiting dilution method, as described previously (29). We obtained ∼50 CD4+ Th17 clones from each bulk TILs; representative Th17 cell clones from melanoma and colon cancer are shown in Fig. 2A. We first evaluated cytokine profiles elaborated by the tumor-infiltrating Th17 cell clones after stimulation with OKT3. Representative data from four tumor-infiltrating Th17 clones are shown in Fig. 2B. We found that these Th17 clones secreted large amounts of IL-8, IL-17, and TNF-α, small amounts of IL-6, but not other cytokines, including IL-2, IL-4, IL-12, or IL-23, which is consistent with previous reports characterizing Th17 cells from other tissue sites (18, 36, 38). Notably, these tumor-infiltrating Th17 cells secreted moderate amounts of IL-10 and TGF-β1 (Fig. 2B), suggesting that Th17 cells may perform regulatory functions in tumor microenvironments (14, 17). In addition, we found only a few Th17/Th1 clones that secreted both IFN-γ and IL-17 among these tumor-derived Th17 cells, based on cytokine profiles determined by ELISA (Fig. 2B). We further confirmed this result by FACS dual staining analyses with the clones (data not shown).
We then examined the phenotypes of tumor-infiltrating Th17 cell clones by flow cytometric analysis, using CD4+CD25+ Treg cells and naive CD4+ T cells as controls. All T cell clones and lines were maintained in the same growth medium containing low levels of IL-2. As shown in Fig. 2C and Supplemental Fig. 1, all Th17 clones uniformly expressed the memory phenotype CCR7+CD62Ldim/− and CD45RA−CD45RO+ (data not shown). Moreover, Th17 clones had minimal or no expression of the cytotoxity-associated markers, CD56, granzyme A, and Fas ligand as well as inhibitory molecule PD-1, similar to those expressed on CD4+CD25+ Treg cells and naive CD4+ T cells. It is surprising that these tumor-derived Th17 clones also expressed CTLA-4, CD25, and Foxp3, which are characteristics of CD4+ Treg cells, suggesting that these Th17 cells may have developmental plasticity and overlap phenotypically with Treg cells. Recent studies have suggested that some chemokine receptors, such as CCR2 and CCR6, are selectively expressed on Th17 cells from certain origins (37–39). To determine whether tumor-infiltrating Th17 cells express unique chemokine receptors, we next analyzed chemokine receptor expression on these Th17 cell clones by FACS analysis. As shown in Fig. 2C, we found that all Th17 clones expressed CCR2, CCR4, CCR5, CCR6, CCR7, and CXCR3, similar to the expression pattern found in CD4+CD25+ Treg cells. These data suggest that tumor-infiltrating Th17 cells express homeostatic chemokine receptors as well as trafficking receptors and share major chemokine receptors with other T cell lineages, including Treg cells (40).
Given the prominence of Th17 cells in the different types of tumor microenvironments and their expression of features shared by Treg cells, we questioned whether tumor-derived Th17 cells might have negative regulatory effects on the immune system. To address this question, we performed proliferation assays to determine whether tumor-infiltrating Th17 clones could suppress proliferation of naive CD4+ T cells (29, 30). As shown in Fig. 2D, we found that in fact these Th17 clones proliferate very well and can also increase the proliferation of naive T cells stimulated with OKT3, suggesting that these cells perform helper rather than suppressive effects on immune cells. Furthermore, CD4-C1, a CD4+ Th cell line derived from a melanoma TIL serving as an effector T cell control, can also increase the proliferation of naive T cells. In contrast, the CD4+CD25+ Treg cell lines served as suppressive controls that strongly inhibit the proliferation of naive CD4+ T cells. Thus, it will be important for future studies to investigate the functional roles of Th17 cells in antitumor immunity using different tumor models in vivo.
Tumor microenvironmental RANTES and MCP-1 mediate the recruitment of CD4+ Th17 cells
On the basis of our observations that CD4+ Th17 cells accumulate in tumor microenvironments and that tumor-infiltrating Th17 cells express homeostatic chemokine receptors and trafficking receptors, we reasoned that tumor microenvironmental factors can recruit CD4+ Th17 cells from the peripheral blood into tumor sites. To test this possibility, we first investigated whether culture supernatants from tumor cell lines and tumor-derived fibroblasts established from tumor tissues of breast cancer, melanoma, and colon cancer can attract Th17 cells. As shown in Fig. 3A, we found that supernatants from both tumor cells and tumor-derived fibroblasts derived from the three types of tumors uniformly induced significant migration of two different tumor-infiltrating Th17 clones (CTh17-C1 and MTh17-C1) compared with culture medium in the functional Transwell chemotaxis assays and that this chemotactic capacity was dose dependent (data not shown). In contrast, supernatants from these tumor cell lines and fibroblasts had no or little chemotactic activity for control CD4+ Th cell line (Th1-C1) (Fig. 3A), suggesting that tumor cells and tumor-derived fibroblasts secrete molecules that selectively recruit Th17 cells to tumor microenvironments.
We next sought to identify the chemokines secreted by tumor cell lines and tumor-derived fibroblasts that can specifically recruit CD4+ Th17 cells. We collected cell culture supernatants from breast cancer, melanoma, and colon cancer cell lines and determined the chemokines released into culture supernatants using a human chemokine Ab array. As shown in Fig. 3B, we found that all three types of tumor cells secreted large amounts of IL-8 and midlevel amounts of RANTES but not eotaxin, MIP-α, or MIP-β. In addition, melanoma tumor cells secreted large amounts of MCP-1 as well as some IP-10, breast cancer cells secreted large amounts of IP-10 and moderate amounts of MCP-1, and colon cancer cells secreted large amounts of MCP-1 and moderate amounts of IP-10. We obtained similar chemokine profiles from the three types of tumor-derived fibroblasts (data not shown). Recent studies have shown that inflammatory microenvironmental production of CCL20 was involved in the preferential recruitment of Th17 cells in autoimmune arthritis and other chronic inflammatory diseases (40, 41). However, we found that only melanoma and colon cancer cells secreted some CCL20, but breast cancer cells as well as the three types of tumor-derived fibroblasts did not secreted detectable CCL20 (Fig. 3B and data not shown). To further determine the chemokines secreted by tumor cells and tumor-derived fibroblasts that were responsible for the recruitment of Th17 cells, we used specific neutralizing Abs against the identified chemokines in functional Transwell chemotaxis assays. We found that saturating concentrations of neutralizing Abs against MCP-1 and RANTES partially abolished the chemotactic activities of the supernatants from a representative MC (MC1) and colon cancer-derived fibroblast cell (CF1) lines for the tumor-infiltrating Th17 clones (Fig. 3C and Supplemental Fig. 2A). However, none of the neutralizing Abs against IL-8, IP-10, eotaxin, MIP-1α, MIP-1β, and CCL20 affected the chemotactic activities of the supernatants (Fig. 3C and data not shown). In addition, combinations of neutralizing Abs against MCP-1 and RANTES completely abolished the capacity of supernatants from tumor cells and tumor-derived fibroblasts to attract Th17 clones (Fig. 3C). In support of these results, as shown in Supplemental Fig. 2B, we found that recombinant MCP-1 and RANTES exhibited a similar chemotactic activity for tumor-infiltrating Th17 cell clones in a dose-dependent manner. However, recombinant MCP-1 and RANTES had little or no chemotactic activity for a control Th cell line (Supplemental Fig. 2B).
We also examined whether culture supernatants from tumor cells and tumor-derived fibroblasts could attract Th17 cells in bulk TILs and peripheral blood. As shown in Fig. 3D, supernatants from MC1 and CF1 cell lines significantly attracted Th17 cells from colon cancer TILs in a Transwell chemotaxis assay. In addition, we found that the culture supernatants from tumor cells and tumor-derived fibroblasts also induced marked migration of Th17 cells in bulk naive CD4+ T cells freshly purified from healthy donors (data not shown). Collectively, these results indicate that tumor cells, as well as tumor microenvironmental fibroblasts, secrete MCP-1 and RANTES that can recruit Th17 cells into tumor sites.
Tumor cells and tumor-derived fibroblasts facilitate the generation of CD4+ Th17 cells
Recent studies have suggested that IL-1, IL-6, IL-23, and TGF-β are key cytokines for human Th17 generation and differentiation (18–20). Given the accumulation of Th17 cells in tumor sites, it is possible that tumor cells and tumor microenvironmental stromal cells create an optimal cytokine milieu to facilitate the generation of Th17 cells. To address this issue, we first determined whether tumor cells and tumor-derived fibroblasts, derived from melanomas, colon cancer, and breast cancer, secrete cytokines required for human Th17 cell development. As shown in Fig. 4A, the three types of tumor cells secreted IL-23, which is consistent with previous findings of significantly upregulated IL-23 in human tumor tissues (42). Furthermore, all three types of tumor-derived fibroblasts secreted large amounts of IL-6. In addition, tumor cells secreted varied amounts of TGF-β and little IL-1β, whereas tumor-derived fibroblasts secreted some IL-1β, IL-23, and TGF-β. These results suggest that tumor microenvironments contain an optimal cytokine milieu that promotes the differentiation and expansion of human Th17 cells.
We next determined whether these tumor cells and tumor-derived fibroblasts could generate Th17 cells from naive CD4+ T cells. Freshly purified naive CD4+ T cells from PBMCs of healthy donors were cultured in OKT-3–coated plates for 24 h, followed by further coculture with medium alone, tumor cells, or tumor-derived fibroblasts for 6 d in the presence of low concentrations of IL-2 (28). Representative data are shown in Fig. 4B. We found that significantly increased numbers of Th17 cells were generated from naive CD4+ cells following their coculture with MC1 melanoma tumor cells (3.9–10.9%, mean 6.9%) and colon cancer-derived CF1 fibroblasts cells (7.6–20%, mean 13.2%). However, the percentage of Th17 cells was relatively low when stimulated naive CD4+ T cells were cultured in medium alone (2.5–5.5%, mean 3.8%). We extended these experiments to other tumor cell lines and tumor-derived fibroblasts and obtained similar results (data not shown). It is of note that the capacity of tumor-derived fibroblasts to induce Th17 cell generation was more potent than that of tumor cells (p = 0.02) (Fig. 4B). In addition, as shown in Fig. 4C, we demonstrated that breast tumor-derived fibroblasts, but not patient-paired normal breast tissue-derived fibroblasts, could significantly induce the generation of Th17 cells from naive CD4+ T cells. Cytokine profiles further showed that breast tumor-derived fibroblasts secreted much higher amounts of IL-1β and IL-6 than those from patient-paired normal tissue-derived fibroblasts (Supplemental Fig. 3). To further confirm these FACS analyses of Th17 cell generation in Fig. 4B, we harvested coculture supernatants at various time points and determined that IL-17 was released by CD4+ T cells. As shown in Fig. 4D, we measured little or no IL-17 in the supernatants at day 0, whereas marked increases of IL-17 were detected in supernatants from MC1 and CF1 cell cocultures at days 3 and 6, compared with cocultures in medium alone. Furthermore, IL-17 levels in supernatants from fibroblast cocultures were higher than those of tumor cell cocultures (Fig. 4D). However, we did not observe any IL-17 production by tumor cells or tumor-derived fibroblasts cultured in medium alone using FACS and ELISA analysis (data not shown).
Given donor variability in the existence of Ag-experienced cells and pre-existing memory Th17 cells in naive CD4+ T cells of adult PBMCs, it is difficult to distinguish the role of tumor cells and tumor-derived fibroblasts in the differentiation of Th17 cells from naive CD4+ T cells or in the expansion of Th17 cells from pre-existing Th17 populations. To obtain relatively more naive CD4+ T cells for study, we purified naive CD4+ T cells from human cord blood by FACS sorting (Supplemental Fig. 4) and then investigated the generation of Th17 cells by tumor cells and tumor-derived fibroblasts. As shown in Fig. 4E, we found very few pre-existing Th17 cells in naive CD4+ T cells purified from cord blood. In addition, minimal Th17 cell induction was observed following coculture with medium alone, MC1 tumor cells, or tumor-derived CF1 fibroblasts in the presence of OKT3 and low concentrations of IL-2 (<1%). Furthermore, we did not find enhanced IL-17 expression in cord blood-derived naive CD4+ T cells induced either by MC1 cells or CF1 cells compared with culture in medium alone, suggesting that tumor cells and tumor-derived fibroblasts play a critical role in the expansion rather than differentiation of human Th17 cells.
We next performed Transwell experiments to determine whether the expansion of Th17 cells by tumor cells and tumor-derived fibroblasts are due to the secreted cytokines. As shown in Fig. 5, we found that cocultures with MC1 tumor cells and tumor-derived CF1 fibroblasts significantly enhanced the generation of Th17 cells compared with medium alone, which was consistent with the results shown in Fig. 4B. However, the numbers of Th17 cells generated from naive CD4+ T cells were dramatically decreased when naive CD4+ T cells were separated from tumor cells or tumor-derived fibroblasts in a Transwell system, although the percentages of Th17 cells were still much higher than those in the culture medium only group (p < 0.05). Taken together, our present data combined with those of our previous report of tumor-infiltrating Th17 cells in ovarian cancer indicate that tumor cells and tumor environmental stromal cells such as fibroblasts produce a proinflammatory cytokine milieu and provide cell–cell contact mechanism(s) as well that facilitate the generation of Th17 cells.
TLR and Nod2 signaling promote the attraction and generation of CD4+ Th17 cells induced by tumor cells and tumor-derived fibroblasts
Mounting evidence suggests that chronic infection and inflammation are significant environmental factors for tumorigenesis; ∼15% of human tumors are initiated by infection-induced inflammation (32, 33). Given the existence of a proinflammatory cytokine milieu suitable for the generation of Th17 cells in tumor microenvironments, we reasoned that chronic infection also contributes to the accumulation of Th17 cells in tumor sites. Because infection-induced inflammation is triggered by interactions between pathogens and TLRs expressed on immune cells and other types of tissue cells (43), we first examined the expression of TLRs (TLR1-9) in tumor cells and tumor-derived fibroblasts by performing RT-PCR screening. We found that these different types of tumor cells as well as tumor-derived fibroblasts expressed nearly all TLRs (Supplemental Fig. 5). We next sought to determine whether TLR and Nod2 signaling on tumor cells and tumor-derived fibroblasts influence their chemotactic activity for Th17 cells. The three types of tumor cell lines and tumor-derived fibroblast cell lines were pretreated with different TLR ligands and MDP (a ligand for Nod2) for 48 h, and then the culture supernatants were harvested and assessed for chemotactic activity for Th17 cells using the functional Transwell chemotaxis assays. As shown in Fig. 6A and 6B, supernatants from both tumor cell lines and tumor-derived fibroblasts induced significant migration of the CTh17-C1 cells, which is consistent with the results shown in Fig. 3A. It is notable that supernatants from the three types of tumor cells pretreated with multiple TLR ligands and MDP had markedly enhanced capacity to attract CTh17-C1 cells (Fig. 6A); these included supernatants from MC1 cells treated with Pam3CSK4 (TLR2), poly(I:C) (TLR3), loxoribine (TLR7), and MDP, supernatants from CC1 cells treated with ligands of Pam3CSK4, poly(I:C), LPS (TLR4), R848 (TLR7/8), CPG (TLR9), and MDP, and supernatants from BC1 cells treated with ligands of Pam3CSK4, poly(I:C), loxoribine, R848, and MDP. However, supernatants from the three tumor-derived fibroblast cell lines treated only with poly(I:C) but not other ligands significantly enhanced chemotactic activity for Th17 cells (Fig. 6B). We next determined whether the increased chemotactic capacity of the supernatants from TLR ligand and MDP-treated tumor cells and tumor-derived fibroblasts was due to the enhanced expression of MCP-1 and RANTES, which we had found to be responsible for the migration for CD4+ Th17 tumor-infiltrating cells. As expected, we found that stimulation of these tumor cell lines with TLR ligands and MDP led to enhanced MCP-1 and RANTES expression (Supplemental Fig. 6A). In addition, treatment of fibroblast cell lines with poly(I:C) also induced higher expression of MCP-1 and RANTES (Supplemental Fig. 6B). To exclude the possibility that these TLR ligands could directly influence Th17 cell function resulting in their enhanced migration, we pretreated Th17 cells with TLR ligands and MDP for 48 h and then determined their chemotactic activity. As shown in Supplemental Fig. 7A, we observed that pretreated Th17 cells with TLR ligands and MDP did not enhance chemotactic activity of Th17 cells. In addition, TLR ligands and MDP cannot directly induce migration for Th17 cells (Supplemental Fig. 7B). These results suggest that inflammation triggered by TLR and Nod signaling can increase the expression of chemokines by tumor cells and tumor-derived fibroblasts that could lead to the enhanced migration of Th17 cells to tumor sites.
We then asked whether the signaling triggered by different TLR and Nod2 on tumor cells and tumor-derived fibroblasts influence the generation of Th17 cells from naive CD4+ T cells as described in Fig. 4. We cultured OKT3-activated naive CD4+ T cells purified from PBMCs of healthy donors with MC1 tumor cells and CF1 fibroblasts in the presence or absence of MDP and different TLR ligands for 6 d, and then determined IL-17 levels in culture supernatants by ELISA. As shown in Fig. 7A, only small amounts of IL-17 were detected in the supernatants of naive CD4+ T cells cocultured with MC1 and CF1 cells, whereas the levels of IL-17 were significantly increased in supernatants from naive CD4+ T cells cocultured with MC1 cells following treatment with Pam3CSK4, poly(I:C), LPS, and MDP. In addition, IL-17 was also increased in the supernatants of naive CD4+ T cells cocultured with CF1 cells treated with R848 and MDP. However, these TLR ligands and MDP did not directly promote the production of IL-17 from naive CD4+ T cells alone, suggesting that the increased production of IL-17 from the cocultured naive CD4+ T cells was due to the effects of TLR ligands and MDP on tumor cells and tumor-derived fibroblasts. These results were also confirmed by intracellular staining for IL-17–producing CD4+ T cells in the cocultured naive CD4+ T cells (Fig. 7B). We further studied whether TLR ligands and MDP could enhance the production of proinflammatory cytokines IL-1β, IL-6, IL-23, and TGF-β1 by tumor cells and tumor-derived fibroblasts, which then promoted the generation of Th17 cells from naive CD4+ T cell. We observed enhanced production of these four cytokines by tumor cells and tumor-derived fibroblasts treated with multiple TLR ligands and MDP (Supplemental Fig. 8). However, the results were variable and depended on tumor type, which did not directly correlate with the IL-17 production induced by ligand treatment. These results further support the notion that tumor cells and tumor-derived fibroblasts provide not only soluble cytokines but also undetermined cell-contact signaling for Th17 cell generation.
Collectively, our results indicate that TLR and Nod2 signaling increase MCP-1 and RANTES expression in tumor cells as well as tumor-derived fibroblasts that promotes the migration and trafficking of Th17 cells from extratumoral sites. In addition, these signaling can accelerate the generation and expansion of Th17 cells through the secretion of inflammatory cytokines and providing cell-contact engagement from tumor cells and tumor-derived fibroblasts from intratumoral sites.
Although recent studies suggest that Th17 cells play critical functional roles in the pathogenesis of human chronic inflammatory disorders and autoimmune diseases, their generation and regulation in tumor microenvironments have not been well characterized. In our previous as well current studies, we analyzed T cell populations in TILs from melanoma, ovarian, breast, and colon cancer patients and showed a prominence of CD4+ Th17 cell populations in these TILs. On the basis of these studies, we hypothesized that the development of tumor-infiltrating Th17 cells may be a general feature in cancer patients. In support of our hypothesis, previous studies had shown high IL-17 mRNA expression in tumor samples from prostate and colorectal cancer patients (44, 45). Furthermore, high proportions of Th17 cells have also been observed in the peripheral blood, tumor tissues, and tumor-draining lymph nodes of patients with advanced ovarian and gastric cancers (25, 27). Although these studies have established the generality of development of tumor-infiltrating CD4+ Th17 cells in cancer patients, the role of Th17 cells in antitumor immunity has not been clearly defined. Studies from both experimental tumor models and human cancer patients showed that IL-17, a dominant effector cytokine produced by Th17 cells, favored tumor growth and exhibited a significant angiogenic effect (5, 45, 46). In addition, a recent study in gastric cancer patients showed that increased numbers of Th17 cells in patients were correlated with cancer stages, further suggesting that Th17 cells contribute to cancer pathogenesis (25). However, the role of IL-17 and Th17 cells in antitumor immunity remains controversial. In some tumor models, expression of IL-17 can boost antitumor immunity by promoting the development of Ag-specific cytotoxic T cells (47, 48). Tumor-specific Th17 cells have been shown to eradicate established and advanced B16 melanoma in an adoptive transfer therapeutic tumor model (49). In our present studies, we generated Th17 cells and clones from TILs of different types of tumor samples and determined their functional effects on naive T cell proliferation and tumor growth in vitro. We found that these tumor-derived Th17 cells performed helper rather than suppressive effects on immune cells (Fig. 2). However, our studies showed that these tumor-derived Th17 cells neither killed tumor cells nor inhibited tumor proliferation but promoted tumor growth in an in vitro culture system (data not shown). As a result of the small sample size in the present studies, we did not observed a significant correlation between Th17 cells in cancer patients and disease progression. Further studies will be needed to advance our understanding of the role of Th17 cells in the immunopathogenesis of individual cancers.
Increasing evidence suggests that Th17 cells mediate inflammatory responses through selective migration and accumulative retention at specific sites. Recent studies have demonstrated that inflammatory microenvironmental production of CCL20 preferentially recruits CCR6-expressing Th17 cells in human rheumatoid arthritis, psoriasis, and other chronic inflammatory diseases (40, 41). Our studies showed that Th17 cells accumulate in the tumor microenvironments of melanoma, ovarian, breast, and colon cancer. We further analyzed chemokine receptor expression on tumor-infiltrating Th17 cells and demonstrated that these Th17 cells express homeostatic chemokine receptors as well as trafficking receptors and that they share major chemokine receptors with other T cell lineages, including CD4+CD25+ Treg cells (40). On the basis of these studies, one explanation for the accumulation of Th17 cells in tumor sites may be due to the preferential recruitment of these Th17 cells by the tumor microenvironment. To address this possibility, in our present studies, we showed that tumor cells as well as tumor-derived fibroblasts secrete MCP-1 and RANTES that strongly attract Th17 cell migration, which suggests that tumor microenvironments may use migratory mechanisms to selectively recruit Th17 cells from the periphery into tumor sites. In addition to the recruitment of Th17 cells, recent studies have shown that tumor cells also use this migratory strategy by recruiting Treg cells to block immunosurveillance and immune destruction of cancer cells in the tumor microenvironment. Studies of Hodgkin’s lymphoma and ovarian cancer showed that tumor microenvironmental CCL22 derived from cancer cells as well as microenvironmental macrophages specifically recruits the CCR4-positive CD4+ Treg cells to tumor sites (31, 50). These studies suggest that Th17 cells, various types of Treg cells, as well as other T cell lineages, may share common chemokine receptors, but the entry and retention of these cells in nonlymphoid tissues are controlled by distinct signals from different chemokines/chemokine reports (51). A better understanding of the migratory mechanisms used by the tumor microenvironment to recruit different types of TILs is essential to the development of immunotherapeutic approaches to regulate the trafficking of these cells.
In addition to the potential recruitment of Th17 cells into tumor microenvironments, our data further showed that tumor cells and tumor-derived stromal cells, such as fibroblasts and APCs, are responsible for the generation and expansion of Th17 cells (Ref. 28 and current study). First, we demonstrated that tumor cells and tumor environmental stromal cells produce the proinflammatory cytokines IL-1, IL-6, IL-23, and TGF-β, which can form an optimal proinflammatory cytokine milieu that is suitable for human Th17 cell differentiation and expansion (18–20). Second, our data also strongly suggest that tumor cells and tumor environmental stromal cells provide not only soluble cytokines but also cell–cell contact signaling for the generation and expansion of human Th17 cells, with the latter being more critical for the regulation of Th17 cells in the tumor microenvironment. Further identification of this cell-contact mechanism will facilitate our understanding of the regulation and functional role of Th17 cells in antitumor immunity. Finally, our current studies further demonstrated that inflammatory signaling triggered by TLRs and Nod2 promote the attraction and generation of Th17 cells induced by tumor cells and tumor-derived fibroblasts. Increasing evidence suggests that chronic infection and inflammation are important environmental factors for tumorigenesis (32, 33). In addition, recent studies have also shown that activation of dendritic cells, monocytes, and PBMCs by TLR and Nod signaling can potentiate human Th17 cell differentiation and production (22–24). Our results showed that TLR and Nod2 signaling can increase MCP-1 and RANTES expression on tumor cells and tumor-derived fibroblasts that can promote the migration and trafficking of Th17 cells. TLR and Nod2 signaling can also accelerate the generation and expansion of Th17 cells through the secretion of inflammatory cytokines, as well as by providing undetermined cell-contact engagements of tumor cells and tumor-derived fibroblasts. These results indicate that signaling mediated by local chronic infections at tumor sites can directly influence tumor cells and tumor-derived stroma cells, which may also contribute to the accumulation of CD4+ Th17 cells in tumor microenvironments. In addition, our studies demonstrated that tumor-derived fibroblasts have greater potency for the generation of Th17 cells compared with paired tumor cells.
Although different types of T cell lineages, including Th1, Th2, Treg cells, and Th17 cells, have independent and distinct gene-expression and regulation signatures, each subset retains substantial developmental plasticity (2, 52). Recent studies from humans and mice demonstrated a Th1/Th17 subset (IL-17+/IFN-γ+) that exclusively coexpresses IFN-γ and IL-17 and that is often identified in vivo in infectious and autoimmune diseases (17, 36). In addition, human CD4+ Treg cells can also be differentiated into IL-17-producing Th17 cells (IL-17+/FoxP3+), and Th17 cells can express FoxP3 and RoRγt (RoRγt+/FoxP3+) (53, 54). Two recent reports further demonstrated that CD4+ T cell subsets, including Th17 cells, have developmental plasticity and overlapping fates determined by an epigenetic mechanism (55, 56). In our current studies, we found that these tumor-derived Th17 cells do not have suppressive activity. However, these Th17 cells can secrete moderate amounts of IL-10 and TGF-β1 after stimulation with anti-CD3 Ab and express the Treg cell markers CTLA-4, FoxP3, and CD25. These data suggest that tumor-infiltrating Th17 cells may also have developmental plasticity as shown for other T cell lineages and may have dual functions performing regulatory as well effector roles in tumor microenvironments. These studies also suggest that Th17 cells, together with Treg cells, may contribute to the immunopathogenesis of cancer. In support of our concept, recent studies showed that in parallel with Treg cells, IL-17+CD4+ and IL-17+CD8+ T cells are kinetically induced in multiple tumor microenvironments in mice and humans, and the number of these T cells is gradually and synchronically increased during tumor development (27). In addition, a study focused on human gastric cancers showed that an increased proportion of Th17 cells in the peripheral blood of affected patients correlated with advanced clinical stages (25). Further studies are needed to determine whether human Th17 cells generated in vivo have a similar instability and plasticity and to identify the mechanisms underlying regulation and dynamic interaction among Th17 cells and Treg and Th1 cells in human pathological conditions such as autoimmune diseases and cancer.
In summary, our work shows that the development of tumor-infiltrating CD4+ Th17 cells is a general feature in cancer patients. We have further identified potential mechanisms for the generation and regulation of Th17 cells in tumor microenvironments. Our data clearly show that factors derived from tumor microenvironments can recruit, generate, and expand Th17 cells and that local chronic infection and inflammation at tumor sites may facilitate all of these processes. These results provide the basis for further investigations of the functional role of Th17 cells in antitumor immunity.
We thank Dr. Joyce Koenig for providing cord blood samples and helpful discussion.
Disclosures The authors have no financial conflicts of interests.
The online version of this article contains supplemental material.
Abbreviations used in this paper:
- breast cancer cell
- colon cancer cell
- IFN-γ–inducible protein 10
- melanoma cell
- nucleotide oligomerization binding domain
- polyinosinic-polycytidylic acid
- tumor-infiltrating lymphocyte
- regulatory T.
- Received August 26, 2009.
- Accepted November 17, 2009.
- Copyright © 2010 by The American Association of Immunologists, Inc.