Abstract
The protumoral activity of γδT17 cells has recently emerged in a wide variety of solid malignancies, including breast cancer. These cells exert their detrimental functions by promoting tumor growth, angiogenesis, and subsequent metastasis development. However, the intratumoral factors that regulate the biology of γδT17cells within the tumor microenvironment are less well understood. Here, using two experimental models of breast cancer, we reinforced the concept that tumor-infiltrating γδT17 cells are endowed with protumoral functions, which promote tumor progression and metastasis development. More importantly, we demonstrated a critical role for type I IFN signaling in controlling the preferential accumulation in the tumor bed of a peculiar subset of γδT17 cells displaying a CD27− CD3bright phenotype (previously associated with the invariant Vγ6Vδ1+ TCR). Interestingly, this effect was indirect and partially relied on the IFNAR1-dependent control of IL7 secretion, a factor that triggers proliferation and activating functions of deleterious γδT17 cells. Our work therefore identifies a key role of the type I IFN/IL7 axis in the regulation of intratumoral γδT17-cell functions and in the development of primary breast tumor growth and metastasis.
Significance: Tumor-derived IL7 can represent a therapeutic target to prevent accumulation of immune cells endowed with potent protumoral activities. Cancer Res; 78(1); 195–204. ©2017 AACR.
Introduction
Despite significant advances in treatment and earlier detection, cancer remains one of the most devastating diseases in industrialized countries. Over the last decades, immunology has gained considerable attention in the complex discipline of oncology. Indeed, it is now well established that leukocyte composition within tumors is considered significant for patient prognosis in a wide variety of cancers, and therefore some immune populations are highly regarded by clinicians as new biomarkers for prognosis and therapeutic success (1, 2).
Among these cell subsets, tumor-infiltrating IL17A-producing γδT (γδT17) cells represent interesting candidates (3). Mouse γδT17 cells are characterized by the constitutive expression of the orphan nuclear receptor RORγt (4) and the absence of the CD27 marker, a thymic regulator that conditions the cytokine bias (IFNγ vs. IL17A production) of γδT cells in the periphery (5). Moreover, many reports indicate that TCR chain usage is linked with specialized functions (6). In this context, Vγ4+ and Vγ6+ T cells have been described to preferentially produce IL17 (7); a phenomenon that primarily relies on the acquisition of a specific functional programming during their development in fetal thymus (8). Immune system evolves to avoid functional redundancies, and therefore it is likely that γδT17 cells bearing either Vγ4 or Vγ6 TCR have unique properties. We recently reported that bright levels of CD3 signals on γδT cells can be used as a surrogate marker to specifically analyze the prototypic IL17 producer subset Vγ6Vδ1+ T cells (9, 10).
Over the past decade, numerous reports have consensually attributed protumoral functions to γδT17 cells in various experimental transplantable or inducible models of cancer including liver and lung carcinomas (11, 12), melanoma (13), breast cancer (14), fibrosarcoma (15), and peritoneal/ovarian cancer (16). Similar observations have also been reported in tumor-bearing patients such as colorectal cancer (17). Moreover, the frequency of infiltrating γδT cells in breast cancer patients is a relevant predictive marker of clinical outcome; associated with poor prognosis and relapses (18). γδT17 cells exert their detrimental functions through interactions with neighboring cells from the tumor environment, including suppressive immune cells (e.g., myeloid-derived suppressor cells, regulatory T cells, M2 macrophages) and endothelial cells. This ultimately leads (i) to the inhibition of the protective CD8+ T-cell response and (ii) to the initiation of angiogenesis, subsequent increase of tumor growth, and metastasis development.
The factors that regulate γδT17 cell accumulation and activity within the tumor have been mainly attributed to Th17-driving cytokines such as IL1β, IL6, IL23, and TGFβ (11, 15). IL7 has been recently involved in selective expansion and functions of IL17A-producing innate-like T cells, including type I natural killer T (NKT)17 and γδT17 cells (19, 20). Interestingly, the level of IL7 increases with tumor progression in peritoneal exudates from ovarian cancer-bearing mice (16). Moreover, levels of IL7 have been associated with poor prognosis in patients with prostate cancer (21, 22) and positively correlated with tumor aggressiveness in patients with breast cancers (23). In view of this, it is tempting to speculate that the putative regulatory functions of tumor-derived IL7 on γδT17 cell biology can somehow explain the phenotypes stated above.
Type I IFNs are composed of several proteins (IFNα, IFNβ, and atypical IFNs) that exert major antimicrobial functions (24). There is also accumulating evidence, indicating that type I IFNs are pivotal cytokines in cancer immunosurveillance and response to anticancer therapies (25). Type I IFNs have been shown to regulate γδT-cell response. Although type I IFNs increase activity of cytotoxic/IFNγ–producing γδT cells (26, 27), they constrain IL17 production by γδT17 cell subsets in various models of infection (28–30). For instance, this effect is associated with the suppressive activity of type I IFN-induced IL27 (31–36). However, the regulatory role of type I IFNs on γδT17-cell activity in cancer is currently unknown.
We have here analyzed the role and regulatory mechanisms of γδT17 cells in cancer immunosurveillance using two syngeneic mouse models of breast cancer. Similar to the human disease, both cell lines enable growth of highly metastatic primary tumors (37). We demonstrate that γδT cells are a prime source of IL17A within the tumor and that T-cell receptor (TCR) δ-deficient mice as well as IL17A depletion (deficient mice or neutralizing mAb treatment) significantly impaired tumor growth. Analysis of intratumoral γδT-cell subsets indicated that IL17 production was almost exclusively due to a subset harboring intense CD3 signals. This subset has recently been associated in the lung tissue with the expression of the clonal Vγ6Vδ1+ TCR (9, 10). Accumulation of CD3bright γδT cells in the tumor bed was largely controlled by type I IFNs. Interestingly, blockade of type I IFN signaling in host led to an increase level of the IL17-promoting factor IL7 within the tumor. Early neutralization of IL7 significantly delayed tumor growth in both experimental syngeneic tumor models. Moreover, IL7 depletion in interferon α/β receptor 1 (IFNAR1)-deficient mice reduced IL17A production by γδT cells as well as their accumulation in the tumor bed. Our study identifies for the first time a cytokine network that strongly regulates the accumulation of protumoral γδT17 cells in breast cancer.
Materials and Methods
Mice
Eight- to 12-week-old female WT C57BL/6J and WT BALB/c mice were purchased from Janvier (Le Genest-St-Isle, France). C57BL/6J IL17A-deficient (Il17a−/−), TCRδ-deficient (Tcrd−/−), and IFNAR1-deficient (Ifnar1−/−) mice were bred in house at the Pasteur institute of Lille. Mice were bred under pathogen-free conditions. Animal studies have been conducted in accordance with an Institutional Animal Care and Use Committee. All animal work conformed to the French “Comité d'Ethique en Experimentation Animale” (C.E.E.A. 75; #00357.01) and committee guidelines.
Reagents
Monoclonal antibodies against mouse CD45 (APCCy7- or AF700-conjugated), CD3 (Pacific Blue-conjugated), TCRδ (PerCpCy5.5-conjugated), TCRβ (FITC-conjugated), Vγ1 (APC-conjugated), Vγ4 (PECy7-conjugated), CD27 (PECy7-conjugated), CD127 (FITC- or PECy7-conjugated), IL17A (PE-conjugated), CD69 (AF700-conjugated), CD44 (PE-conjugated), and appropriated isotype controls were purchased from BioLegend, BD Pharmingen, and eBioscience. PBS-57 glycolipid-loaded and -unloaded control CD1d tetramers (APC-conjugated) were from the National Institute of Allergy and Infectious Diseases Tetramer Facility (Emory University, Atlanta, GA). Propidium iodide was purchased from BD Pharmingen. Mouse monoclonal anti-IL17A (clone 17F3) and anti-IL7 (clone M25) antibodies and their respective isotype controls [MOPC-21 (mouse IgG1) for anti-IL17A and MCP-11 (mouse IgG2b) for anti-IL7] were purchased from Bio X Cell. Recombinant mouse IFNβ and IL7 were purchased from R&D Systems. Recombinant mouse IL1β and IL23 were purchased from Peprotech. Phorbol 12-myristate 13-acetate (PMA) and ionomycin were from Sigma-Aldrich.
Mouse tumor models
The C57BL/6 mouse breast carcinoma cell line AT-3 was obtained from Trina Stewart (Griffith University, Brisbane, Australia; ref. 38), 4T1.2 tumor cells were kindly provided by Robin Anderson (Peter MacCallum Cancer Centre, Melbourne, Australia). Cells were cultured and maintained in DMEM supplemented with 10% FCS, 2mmol/L l-glutamine, penicillin/streptomycin and sodium pyruvate and 55nmol/L 2-mercaptoethanol. For mouse cancer models, mice were inoculated subcutaneously with 5 × 105 AT-3 or 4T1.2 cells on the right flank. For tumor growth studies, tumors were monitored every 2/3 days using a caliper.
Quantification of metastatic burden
Metastasis of AT-3 cells were determined in sections of the lung stained with hematoxylin and eosin. The number of metastases per section was counted.
Cytokine detection by enzyme-linked immunosorbent assay
At time-point as indicated, tumors were weighted and collected in cold PBS until processing. Tumors were then transferred to glass tubes and homogenized in 150 μL PBS. Tubes were centrifuged and supernatants were collected for ELISAs. Cytokine detections were performed using IL17A, IL1β, IL23p19 (eBioscience), and IL7 (R&D Systems) ELISA kits following the manufacturer's instructions.
Flow cytometry
At time as indicated, tumors were harvested from mice and collected in cold PBS until processing. Tumors were then excised and finely minced, followed by enzymatic digestion for 30 to 40 minutes at 37°C in PBS containing 1 mg/mL collagenase type VI (Sigma-Aldrich) 1 mg/mL DNase type I (Roche) and 100 ng/mL Hyaluronidase (Sigma-Aldrich). Tumors were smashed and homogenates were filtered and collected. Cells were centrifuged at 1,200 rpm at 4°C for 10 minutes and pellets washed in PBS 2% FCS for direct surface marker staining or complete IMDM for intracellular staining. For intracellular IL17A detection, cells were incubated in complete IMDM containing Golgi Plug/Golgi Stop (BD Biosciences) in presence of PMA (100 ng/mL) and ionomycin (1 μg/mL) for 4 hours at 37°C before antibody staining. Cells were washed and incubated with the different antibodies for 30 minutes in PBS 2% FCS. For intracellular IL17A detection, cells were then washed with PBS and fixed using IC Fixation Buffer (eBioscience, CliniSciences). Fixed cells were then permeabilized in permeabilization buffer (eBioscience), according to the manufacturer's instructions. Cells were finally stained with anti-17A antibody. Cells were analyzed on a LSR Fortessa (BD Biosciences). FACS analyses were performed using the FlowJo software.
Cell sorting and in vitro assay
Lungs were harvested from naive mice and cell suspensions were prepared as previously described (10). To purify γδT cells, lung mononuclear cells were labeled with anti-CD45, anti-CD3, and anti-TCRδ antibodies (conjugated as previously indicated). Cells were sorted using an ARIA cell sorter (BD Biosciences). A total of 5 × 103 γδT cells were then stimulated for 20 hours with a combination of mouse recombinant IL1β/IL23 (1 ng/mL for each cytokine) with or without mouse recombinant IFNβ (20 ng/mL) in IMDM 10 % FCS media. Supernatants were collected for IL17A detection by ELISA.
RNA, cDNA, and real-time quantitative PCR
RNA was extracted and purified using the NucleoSpin RNA Extraction Kit (Macherey-Nagel) following the manufacturer's instructions. The cDNA was synthesized from total RNA using the High Capacity RNA-to-cDNA Kit (Thermofisher Scientific). Gene expression was determined on the QuantStudio 12K Flex Real-Time PCR System (Life Technologies) using specific primers for Ifnb (5′-CAGGTGGATCCTCCACGCT-3′ and 5′-CATTCAGCTGCTCCAGGAGC-3′), Ifna4 (5′-TGATGAGCTACTACTGGTCAGC-3′ and 5′-GATCTCTTAGCACAAGGATGGC-3′), Il7 (5′-GCGGACGATCACTCCTTCTG-3′ and 5′-AGCCCCACATATTTGAAATTCCA-3′), and Gapdh (5′-GCAAAGTGGAGATTGTTGCCA-3′ and 5′-GCCTTGACTGTGCCGTTGA-3′) and QuantiTect SYBR Green PCR Master Mix (Qiagen) according to the manufacturer's instructions. Gene expression was normalized to Gapdh.
Statistical analysis
Data are presented as means ± SEM and are cumulative of two to four independent experiments. Unpaired T test or one-way ANOVA followed by Bonferroni's post T test were used for statistical analysis when two groups or multiple groups were analyzed, respectively. When data did not follow a Gaussian distribution, Mann–Whitney U test or Kruskal–Wallis followed by Dunn's post T test were used for statistical analysis when two groups or multiple groups were analyzed, respectively. We used two-way ANOVA when multiple parameters were studied. Statistical significance was set at *, P < 0.05; **, P < 0.005; and ***, P < 0.001.
Results
γδT cells are a prime source of protumoral IL17A in breast cancer models
To investigate the role of the pro-inflammatory cytokine IL17A in tumor progression, we used two transplantable syngeneic models of breast cancer namely AT-3 (C57BL/6) and 4T1.2 (BALB/c). To this end, these tumor cell lines were implanted into the right flank of either Il17a−/− (AT-3) or IL17A–depleted (4T1.2) mice. Compared with controls, a significant delay in tumor growth was observed in absence or upon neutralization of IL17A (Fig. 1A; Supplementary Fig. S1). In addition, a decrease number of spontaneous lung metastasis was numbered in AT-3 tumor-bearing Il17a−/− mice (Fig. 1B). Of interest, IL17A protein was mainly found at early-stage (day 14) of tumor development in the primary tumor that may indicate a contribution for innate components in IL17A production (Fig. 1C). To identify the cellular sources of IL17A within the tumor, infiltrating leukocytes from AT3 tumor-bearing mice were analyzed 14 days after inoculation. IL17A production by lymphocyte subsets (Supplementary Fig. S2A) was analyzed by intracellular cytokine staining after ex vivo stimulation with PMA/ionomycin. We observed that intratumoral innate-like γδT cells (43 %) rather than conventional CD4+ T cells (10 %) or NKT cells (9 %) represented the major contributors in IL17A production (Fig. 1D; Supplementary Fig. S2B). Remaining IL17A producers comprised other T-cell subsets (15 %) and non–T-cell populations (14 %; Supplementary Fig. S2B). Moreover, intrinsic capacity of IL17A production as well as intensity of fluorescence in IL17A+ cells suggested that γδT cells had a higher intrinsic capacity to produce this cytokine compared with the other intratumoral IL17A–producing cells, including NKT and CD4+ T cells (Supplementary Fig. S2C and S2D). Interestingly, a similar tendency was observed in the 4T1.2 model 14 days post-inoculation (Supplementary Fig. S3A), which further confirmed γδT cells as a major source of IL17A in the breast tumor microenvironment (Supplementary Fig. S3B and S3C). Consistent with this, we detected a marked decrease in IL17A proteins within the tumors of γδT-cell–deficient (Tcrd−/−) mice (Fig. 1E). To assess the contribution of γδT17 cells in tumor outgrowth, we monitored AT-3 tumor progression in controls versus Tcrd−/− mice. Interestingly, γδT-cell deficiency was accompanied with a reduced tumor progression (Fig. 1F). However, the number of lung metastases in Tcrd−/− mice was similar to control mice (Fig. 1G). Altogether, these results suggest that IL17A mainly produced by γδT cells has protumoral functions in breast cancer models.
γδT cells are a major source of protumoral IL17A in AT-3 breast cancer model. A and B, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Il17a−/− mice. A, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of two (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 10 to 12 mice/group of n = 2. B, At end-point, lungs were harvested, sectioned, and metastases determined. One representative section of each group is shown on the left. Individuals and means ± SEM are shown on the right (n = 2). C and D, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT mice. C, At time indicated, tumors were harvested and homogenates were tested for IL17A protein detection by ELISA. Results shown are the means ± SEM of 6 to 8 mice/group of n = 2. D, After 14 days, tumors were collected. Means of cell subset proportions in overall live IL17-producing intratumoral leukocytes (IL17A+ CD45+ cells) cells are represented on a pie chart. Leukocyte populations (CD45+) are gated as follows: conventional CD4+ T cells (CD4+ tetramer CD1d/PBS57− TCRβ+), NKT (tetramer CD1d/PBS57+ TCRβ+), and γδT (CD3+ TCRδ+). Results shown are the means of 15 mice/group of n = 3. E–G, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT or Tcrd−/− mice. E, After 14 days, tumors were collected and homogenates were tested for IL17A protein detection by ELISA. Results shown are the individuals and means ± SEM of 13 mice/group of n = 2. F, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of three (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 17 mice/group of n = 3. G, At end-point, lungs were harvested, sectioned, and metastases determined. Individuals and means ± SEM are shown (n = 2; 13–14 mice/group); ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.
γδT cells are a major source of protumoral IL17A in AT-3 breast cancer model. A and B, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Il17a−/− mice. A, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of two (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 10 to 12 mice/group of n = 2. B, At end-point, lungs were harvested, sectioned, and metastases determined. One representative section of each group is shown on the left. Individuals and means ± SEM are shown on the right (n = 2). C and D, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT mice. C, At time indicated, tumors were harvested and homogenates were tested for IL17A protein detection by ELISA. Results shown are the means ± SEM of 6 to 8 mice/group of n = 2. D, After 14 days, tumors were collected. Means of cell subset proportions in overall live IL17-producing intratumoral leukocytes (IL17A+ CD45+ cells) cells are represented on a pie chart. Leukocyte populations (CD45+) are gated as follows: conventional CD4+ T cells (CD4+ tetramer CD1d/PBS57− TCRβ+), NKT (tetramer CD1d/PBS57+ TCRβ+), and γδT (CD3+ TCRδ+). Results shown are the means of 15 mice/group of n = 3. E–G, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT or Tcrd−/− mice. E, After 14 days, tumors were collected and homogenates were tested for IL17A protein detection by ELISA. Results shown are the individuals and means ± SEM of 13 mice/group of n = 2. F, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of three (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 17 mice/group of n = 3. G, At end-point, lungs were harvested, sectioned, and metastases determined. Individuals and means ± SEM are shown (n = 2; 13–14 mice/group); ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Type I IFN signaling restrains accumulation and activity of tumor-infiltrating γδT17 cells
The factors that positively regulate γδT17 cell activity and proliferation have been extensively studied (3, 39). Conversely, those curbing γδT17-cell functions in cancer are far less understood. Type I IFNs are endowed with potent immostimulatory functions that confer them clinically relevant antitumor effects (25). The capacity of type I IFNs to constrain γδT17-cell activity during viral and bacterial infections has recently been proposed (28–30). Interestingly, we observed high expression of Ifnb and Ifna4 transcripts in tumors at 14 days post-tumor inoculation (Fig. 2A). To evaluate the putative role of type I IFNs on γδT-cell functions, AT-3 tumor cells were inoculated in mice deficient for IFNAR1. We observed a significant increase in the frequency of tumor-infiltrating γδT cells in these animals compared to WT controls (Fig. 2B). Meanwhile, frequencies of NKT cells and conventional CD4+ T cells were unchanged or decreased in this setting (Fig. 2B). Of great interest, accumulating intratumoral γδT cells from IFNAR1-deficient mice was mainly associated with a IL17A producer profile (Fig. 2C). However, these cells had a comparable intrinsic ability to produce IL17A (Supplementary Fig. S4A), which interestingly suggested that type I IFNs might primarly influence the accumulation of these cells. Besides, concomitant analyses on the other IL17A-producing cells showed that type I IFN signalling blockade restrained IL17A+ NKT cells but did not interfere with the ability of CD4+ T cells to express this cytokine (Supplementary Fig. S4B). In line with their TH17-like phenotype, tumor-infiltrating γδT cells from Ifnar1-deficient mice exhibited a pronounced activated/memory phenotype as judged by CD69 and CD44 stainings (Fig. 2D). This increase was parralleled with the detection of higher levels of IL17A protein in the tumors of Ifnar1−/− mice (Fig. 2E). Consistent with the antitumoral effects of type I IFNs, Ifnar1−/− mice displayed a slightly but significant faster progression of AT-3 tumor growth (Fig. 2F). Altogether, these data indicate that type I IFNs control the accumulation and functions of γδT17 cells in breast cancer.
Type I IFN signaling negatively modulates IL17A–producing γδT cells. A, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice. At time indicated, tumors were collected and levels of transcripts for Ifnb and Ifna4 were determined by qRT-PCR. Means ± SEM of two independent experiments are shown (12 mice/group; n = 2). B–F, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. B–E, After 14 days, tumors were collected and TILs prepared. B, Individuals and means ± SEM of the frequency of γδT, NKT, and CD4+ T cells in TILs were compared between groups (15 mice/group; n = 3). C, IL17A-producing intratumoral γδT cells (CD45+ CD3+ TCRδ+) are represented. A representative dot plot for each genotype is shown at the top (concatenate for 6 mice/group). Individuals and means ± SEM of the percentage of IL17A+ γδT within the γδT-cell population (bottom left) and total lymphocyte (bottom right) populations of 11 to 12 mice/group of n = 2 are shown. D, Expression of CD69 and CD44 markers on intratumoral γδT cells was compared between groups. Means ± SEM of 11 to 12 mice/group of n = 2 are shown. E, Tumors were harvested and homogenates were tested for IL17A protein detection by ELISA. Results shown are the individuals and means ± SEM of 12 mice/group of n = 2. F, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of 3 (6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 18 mice/group of n = 3; *, P < 0.05; **, P < 0.01.
Type I IFN signaling negatively modulates IL17A–producing γδT cells. A, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice. At time indicated, tumors were collected and levels of transcripts for Ifnb and Ifna4 were determined by qRT-PCR. Means ± SEM of two independent experiments are shown (12 mice/group; n = 2). B–F, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. B–E, After 14 days, tumors were collected and TILs prepared. B, Individuals and means ± SEM of the frequency of γδT, NKT, and CD4+ T cells in TILs were compared between groups (15 mice/group; n = 3). C, IL17A-producing intratumoral γδT cells (CD45+ CD3+ TCRδ+) are represented. A representative dot plot for each genotype is shown at the top (concatenate for 6 mice/group). Individuals and means ± SEM of the percentage of IL17A+ γδT within the γδT-cell population (bottom left) and total lymphocyte (bottom right) populations of 11 to 12 mice/group of n = 2 are shown. D, Expression of CD69 and CD44 markers on intratumoral γδT cells was compared between groups. Means ± SEM of 11 to 12 mice/group of n = 2 are shown. E, Tumors were harvested and homogenates were tested for IL17A protein detection by ELISA. Results shown are the individuals and means ± SEM of 12 mice/group of n = 2. F, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of 3 (6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 18 mice/group of n = 3; *, P < 0.05; **, P < 0.01.
Type I IFNs preferentially exert its functions on tumor-infiltrating CD27− CD3bright γδT17 cells
Because γδT17 cells are composed of various subsets, we investigated in details whether or not type I IFNs could control specific γδT17 cell subsets in the tumor microenvironment. In line with Fig. 2B, γδT cells in tumors of Ifnar1−/− mice were mainly CD27− (Fig. 3A), a phenotype associated with preferential ability to produce IL17A (5).
Type I IFN signaling controls the accumulation of the IL17A-producing CD27− CD3bright γδT cell subset. A–C, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. A, Individuals and means ± SEM of the frequency of CD27− γδT cells within the γδT-cell population were compared between groups (10–12 mice/group; n = 2). B, Comparison of intratumoral CD3dim versus CD3bright γδT cells. One representative plot of 15 to 16 mice for each genotype is shown (n = 3). C, Proportions of γδT-cell subsets based on Vγ expression within total γδT-cell compartment are shown for each genotype as means (15–16 mice/group; n = 3). D, Capacity of intratumoral γδT-cell subset to produce IL17A based on the CD3 phenotype in each genotype. Individuals and means ± SEM of subsets for each genotype is shown. n.s., not significant; **, P < 0.01; ***, P < 0.001.
Type I IFN signaling controls the accumulation of the IL17A-producing CD27− CD3bright γδT cell subset. A–C, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. A, Individuals and means ± SEM of the frequency of CD27− γδT cells within the γδT-cell population were compared between groups (10–12 mice/group; n = 2). B, Comparison of intratumoral CD3dim versus CD3bright γδT cells. One representative plot of 15 to 16 mice for each genotype is shown (n = 3). C, Proportions of γδT-cell subsets based on Vγ expression within total γδT-cell compartment are shown for each genotype as means (15–16 mice/group; n = 3). D, Capacity of intratumoral γδT-cell subset to produce IL17A based on the CD3 phenotype in each genotype. Individuals and means ± SEM of subsets for each genotype is shown. n.s., not significant; **, P < 0.01; ***, P < 0.001.
Compared with control AT-3 tumor-bearing mice, we observed a significant increase of CD3bright γδT cells within the tumors of Ifnar1−/− mice (Fig. 3B; Supplementary Fig. S5A) to become the main γδT cell subset (13.7 % vs. 58.8 %). This was parralleled with a reduced proportion of the CD3dim γδT-cell population (Fig. 3B; Supplementary Fig. S5A), which include Vγ1+ and Vγ4+ T cells (Fig. 3C). In total TILs, the effect of type I IFN signaling blockade on CD3bright γδT cells was even more striking (Supplementary Fig. S5B). In the meantime, although their proportion within the γδT-cell compartment was decreased (Fig. 3C), the frequency of both Vγ1+ and Vγ4+ T-cell subsets in total TILs was unaffected in Ifnar1-deficient mice (Supplementary Fig. S5B) despite the fact that Vγ4+ T cells have important IL17A–producing capacities (40). CD3bright γδT cells accounted for the majority of IL17A production within the γδT-cell compartment in both control and Ifnar1-deficient mice (Fig. 3D). Collectively, these data support a preferential role for type I IFNs in the negative control of protumoral CD3bright γδT cells in breast cancer.
Type I IFN signaling inhibits accumulation of the IL17-driving cytokine IL7 in the tumor
To investigate the mechanisms involved in the regulatory functions of type I IFNs on γδT17 cells, we first evaluated the possible direct effect of IFNβ on IL17A production by γδT cells. As shown in Fig. 4A, in vitro IFNβ treatment failed to lower IL17A synthesis by purified γδT cells in response to IL1β and IL23, suggesting an indirect mode of action.
Type I IFN signaling inhibits protumoral IL7. A, Purified γδT cells from the lungs of naive WT mice were stimulated in presence or not of recombinant IFNβ (1,000 U/mL) and tested for IL17A production. Mean ±SEM of pooled data from three independent experiments is shown. B and C, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. B and C, After 14 days, tumors were harvested and homogenates were tested for the detection of transcripts or cytokines by qRT-PCR or ELISA, respectively. Results shown for IL1β and IL23 (B) and IL7 (C) are individuals and means ± SEM of 10 to 12 mice/group of n = 2 for ELISA and 16 mice/group of n = 3 for qRT-PCR. D–F, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. D, After 14 days, tumors were collected and homogenates were tested for IL17A protein. Individuals and means ± SEM of 10 to 12 mice/group of n = 2 are shown. E, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of two (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 10 to 12 mice/group of n = 2. F, At end-point, lungs were harvested, sectioned, and metastases were numbered. Individuals and means ± SEM of one experiment out of two are shown in the right (8 mice/group); n.s., not significant; *, P < 0.05; **, P < 0.01.
Type I IFN signaling inhibits protumoral IL7. A, Purified γδT cells from the lungs of naive WT mice were stimulated in presence or not of recombinant IFNβ (1,000 U/mL) and tested for IL17A production. Mean ±SEM of pooled data from three independent experiments is shown. B and C, A total of 5 × 105 AT-3 cells were injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. B and C, After 14 days, tumors were harvested and homogenates were tested for the detection of transcripts or cytokines by qRT-PCR or ELISA, respectively. Results shown for IL1β and IL23 (B) and IL7 (C) are individuals and means ± SEM of 10 to 12 mice/group of n = 2 for ELISA and 16 mice/group of n = 3 for qRT-PCR. D–F, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. D, After 14 days, tumors were collected and homogenates were tested for IL17A protein. Individuals and means ± SEM of 10 to 12 mice/group of n = 2 are shown. E, Primary tumor growth was monitored every 2 to 3 days and areas under the curve were compared. Curves shown are representative of one experiment out of two (5–6 mice/group). Statistics were performed on cumulative data from two independent experiments and results are shown as means ± SEM of 10 to 12 mice/group of n = 2. F, At end-point, lungs were harvested, sectioned, and metastases were numbered. Individuals and means ± SEM of one experiment out of two are shown in the right (8 mice/group); n.s., not significant; *, P < 0.05; **, P < 0.01.
We next turned to analyze the impact of type I IFN signaling blockade on the production of well-known γδT17-driving cytokines in the tumor tissue. While the levels of IL1β and IL23 proteins were not affected by type I IFN signaling disruption (Fig. 4B), we observed a significant increase in IL7 concentration, at both mRNA and protein levels, within the tumors of Ifnar1−/− mice compared with their control counterparts (Fig. 4C). Since certain tumor cells have been previously shown to produce IL7 (41), we tested the ability of AT-3 in culture to secrete IL7. However, in these conditions, we failed to detect a significant increase of IL7 production by the tumor cell line between 4 and 24 hours culture time points (Supplementary Fig. S6A).
Furthermore, in vivo neutralization of IL7 in WT mice significantly reduced the levels of IL17A within the tumors (Fig. 4D). In addition, IL7 neutralization led to a marked tumor growth deceleration in both AT-3– and 4T1.2-bearing mice, thus demonstrating the protumoral activity of this cytokine (Fig. 4E; Supplementary Fig. S6B). In line, the number of spontaneous lung metastases was reduced in anti-IL7 mAb-treated mice (Fig. 4F). Thus, type I IFNs impair the expression of the pro-IL17A cyokine IL7 to favor tumor growth.
IL7 depletion decreases the accumulation of tumor-infiltrating γδT17 cells
We next examined the potential impact of IL7 depletion on γδT17-cell homeostasis in the tumor microenvironment. Indeed, IL7 was recently shown to be a critical cytokine in the regulation and expansion of γδT17 cells (19). In the presence of IL7-neutralizing mAb, we noted a marked decrease in the percentages of IL17A+ tumor-infiltrating γδT cells in both AT-3– and 4T1.2-bearing mice (Fig. 5A; Supplementary Fig. S7A). Interestingly, IL7 depletion concomitantly caused reduction in the proportion of both CD27− and CD3bright γδT cells that clearly demonstrate the importance of IL7 in regulating γδT-cell subsets with great ability to secrete IL17A (Fig. 5A; Supplementary Fig. S7B). As previously shown in peripheral organs (Supplementary Fig. S8A; ref. 9), intratumoral CD3bright γδT cells expressed higher levels of IL7 receptor α (IL7Rα/CD127) at the cell surface compared with their CD3dim counterpart (Fig. 5B). It is also important to mention that type I IFN signalling blockade resulted in a increase expression of γδT-cell subsets to a similar extent (Fig. 5C). Nevertheless, this effect was likely to be indirect because incubation of purified CD3dim or CD3bright γδT cells with recombinant IFNβ did not directly modulate CD127 expression (Supplementary Fig. S8B) whereas the positive control did (Supplementary Fig. S8C).
IL7 induces accumulation of IL17-producing CD27−CD3bright γδT cells. A, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. Individuals and means ± SEM of the percentage of IL17+, CD27−, and CD3bright subset within the γδT-cell compartment of 10 to 12 mice per group of n = 2 are shown. B, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice. At day 14, tumors were harvested and expression of IL7Rα (CD127) on infiltrating CD3dim and CD3bright γδT-cell subsets was analyzed. Means ± SEM of CD127 expression (mean fluorescence intensity, MFI) for each subset from 15 mice of n = 3 are shown. C, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. TILs are prepared from d14 tumors. Modulation of CD127 levels (MFI) on intratumoral CD3dim and CD3bright γδT subsets from WT versus Ifnar1−/− mice was evaluated. Means ± SEM from 6 mice/genotype are shown. D, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. Results show individuals and means ± SEM of the percentage of intratumoral IL17+ γδT cells (left) and CD3bright γδT cells of 6 to 7 mice/group of n = 2; n.s., not significant; *, P < 0.05; **, P < 0.01.
IL7 induces accumulation of IL17-producing CD27−CD3bright γδT cells. A, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. Individuals and means ± SEM of the percentage of IL17+, CD27−, and CD3bright subset within the γδT-cell compartment of 10 to 12 mice per group of n = 2 are shown. B, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT mice. At day 14, tumors were harvested and expression of IL7Rα (CD127) on infiltrating CD3dim and CD3bright γδT-cell subsets was analyzed. Means ± SEM of CD127 expression (mean fluorescence intensity, MFI) for each subset from 15 mice of n = 3 are shown. C, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice. TILs are prepared from d14 tumors. Modulation of CD127 levels (MFI) on intratumoral CD3dim and CD3bright γδT subsets from WT versus Ifnar1−/− mice was evaluated. Means ± SEM from 6 mice/genotype are shown. D, A total of 5 × 105 AT-3 cells was injected subcutaneously into C57BL/6 WT or Ifnar1−/− mice and treated with an anti-IL7 (150 μg/mouse) or isotype control at day 0. Results show individuals and means ± SEM of the percentage of intratumoral IL17+ γδT cells (left) and CD3bright γδT cells of 6 to 7 mice/group of n = 2; n.s., not significant; *, P < 0.05; **, P < 0.01.
Finally, we aimed to investigate the potential role of IL7 in the type I IFN signaling-mediated control of the tumor-infiltrating IL17A-producing CD3bright γδT-cell subset. Strikingly, IL7 neutralization significantly decreased the percentages of IL17A+ as well as CD3bright γδT cells in Ifnar1−/− mice (Fig. 5D). Taken together, the type I IFN/IL7 axis exerts a critical role in the regulation of intratumoral CD3bright γδT17-cell functions, an effect that leads to the restrainement of tumor outgrowth.
Discussion
Understanding the fine-tuning of immune responses in cancer is a useful goal to propose new prognosis biomarkers as well as putative new therapeutic targets. In the present study, we investigated the cytokine network responsible for the regulation of tumor-infiltrating IL17A-producing γδT cells using two syngeneic models of breast cancer. Interestingly, we showed that protumoral γδT17 cells, described herein as a major and early source of IL17A in the tumor microenvironment during cancer progression, were negatively regulated by type I IFN signaling. In addition, type I IFNs specifically control CD3bright γδT cells, a subset that displays remarkable IL17A production capacities. Moreover, type I IFN exerted their regulatory functions via an indirect mechanism involving inhibition of IL7, which was further identified as a critical cytokine for the accumulation of CD3bright γδT17 cells in tumors.
Several studies have started to highlight the protumoral activities of IL17A-producing γδT cells (i.e., tumor burden on primary sites as well as metastasis and/or angiogenesis) in rodents and humans (3). Our observations confirmed previous investigations on both breast and ovarian cancer models showing deceleration of tumor growth and/or decreased lung metastasis burden upon both deficiency and neutralisation of IL17A (14, 16). As previously shown by other groups, we demonstrated the main source of IL17A in early stages of tumor progression to be γδT-cell–dependent in terms of both cell percentage and signal intensity (14, 16). Surprisingly, even if γδT-cell deficiency led to a significant reduction in tumor outgrowth, the number of lung metastasis at end-point was unchanged in comparison to WT AT-3 tumor-bearing mice. Since the γδT cell compartment comprises subsets endowed with either anti- or pro-tumoral functions (3), our observation might be explained by the absence of the protective activity of some tumorocidal γδT cells. In this context, the use of more specific animal models such as the Vγ4/Vγ6−/− mice that are deficient for the two major subsets of γδT17 cells (42) will be worth undertaking. Little is known about the importance of tumor-infiltrating NKT cells in the production of IL17A. Interestingly, we observed substantial IL17A expression in NKT cells although proportions appeared to be lower than γδT cells at the early stages of tumor progression. In any case, the putative role of IL17A–producing NKT cells should be considered in future studies.
We demonstrated in the present study that type I IFN signaling impedes the accumulation of γδT17A cells in the tumor microenvironment. Although our observations are in line with previous investigations indicating the pivotal role of type I IFN in mediating negative regulation of IL17A-producing γδT cells upon infections (28–30), to our knowledge, this is the first time that a similar mechanism is defined in the context of cancer. We herein extended those previous obervations by demonstrating that type I IFNs preferentially exert their negative control on the CD3bright CD27− γδT-cell subset. We recently described this population to express the canonical Vγ6Vδ1 TCR in multiple lymphoid or non-lymphoid tissues (9, 10). However, our investigations on other populations, including both Vγ1+ and Vγ4+ showed that those γδT-cell subsets were unlikely to be controlled by type I IFN. These findings add another substantial level of regulation in type I IFN-mediated control of tumor-infiltrating γδT17 cells in breast cancer.
We next investigated whether type I IFN could directly inhibit IL17A production by γδT cells or whether it requires intermediates. Our data indicate that recombinant IFNβ does not directly modulate IL17A secretion by γδT cells in response to both IL1β and IL23. Interestingly, one report showed that type I IFN could downregulate both IL1β and IL23 production in human DCs in vitro as well as subsequent Th17 response (43). However, we observed that IFNAR1 deficiency in mice did not impair the secretion of both IL1β and/or IL23 in the tumor microenvironment, strongly indicating the involvement of another mechanism. It is well known that secretion of type I IFNs by antigen-presenting cells leads to sequential cellular signaling events leading to the release of immunoregulatory cytokines such as IL27 and IL10 (44). Cao and collegues reported inhibition of IL17A response in γδT cells by type I IFNs through an IL27-dependent mechanism during post-influenza pneumococcal infection (29). Accordingly, we can not exclude in our model the possibility for IL27 to act as an intermediate of type I IFN-mediated impairment of γδT17A enrichment in tumors. The same hypothesis could be formulated in the case of IL10, although, to our knowlegde, the existence of a type I IFN/IL10 regulatory axis on γδT17A cells has never been reported.
Numerous studies have reported the critical role of IL7 for the differentiation and regulation of IL17A-producing γδT cells (19, 45–47). Especially, administration of recombinant IL7 tended to increase the frequency of Vγ6+ cells in a model of ovarian cancer (16). However, we highlighted here, using a neutralizing antibody, the protumoral role of host IL7 on breast cancer tumor growth as well as in the accumulation of CD3bright γδT17-cell population in the tumor bed. It is likely that the preferential effect of IL7 on CD3bright γδT cells can be simply explained by the high level of IL7Rα surface expression on this very subset compared with CD3dim γδT cells. Besides its role in controling IL7 secretion within the tumor microenvironment, type I IFNs also exert regulatory functions on IL7Rα expression on CD3bright γδT cells. This may also play a part in the mechanism defined in the present study.
Importantly, our results contrast with one previous report showing the role of type I IFNs in the induction of IL7 during viral hepatitis (47). Several reasons could explain this apparent discrepancy between the two models such as differences in experimental systems, in particular the nature and location of inflammation. Solid tumors are usually associated with a strong immunosuppressive environment (48) whereas virus infection leads to a generally IFN type I–mediated protective innate response. Then, variation in cellular populations as well as inflammatory status could subsequently lead to distinctive signaling mechanisms involved in the regulation of IL7 production. In addition, the cellular source of IL7 could represent another important factor of modulation.
Keratinocytes, stroma, and epithelium cells have been described as major cellular sources for IL7, although its production was also reported in dendritic cells (49, 50). Interestingly, IL7 expression was observed in tumor cells of epithelial origin (41), which raised the question whether breast cancer cell lines could spontaneously produce IL7. Nevertheless, our in vitro studies suggested no detectable secretion of IL7 from AT-3 cells, although we cannot exclude the ability of those cells to produce this cytokine upon specific conditions in culture or in the appropriate environment in vivo.
To conclude, the development of immunotherapeutic tools to fight solid tumors has been proved to be challenging. Then, a better understanding of the complex regulatory mechanisms involved in the immunosurveillance inside the tumor microenvironment is required. We identified an important type I IFN/IL7 axis involved in the negative control of protumoral γδT17 cells in solid tumors. Those findings could also be potentially relevant in the context of human breast cancer regarding that elevated expression of IL7 is considered usually as a poor prognosis for tumor progression in patients (21–23). Considering the pivotal role of IL17A in regulating tumor growth, angiogenesis, and metastasis during breast cancer in the mouse model, our study suggests that the implementation of novel combined strategies to both specifically activate type I IFN and/or neutralize IL7 activities could benefit the field of immunotherapy research against solid tumors. This could pave the way to the design of new curative approaches that could increase the immunotherapeutic arsenal for oncologists.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: E.C. Patin, F. Trottein, C. Paget
Development of methodology: E.C. Patin, C. Paget
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.C. Patin, D. Soulard, S. Fleury, D. Dombrowicz, C. Paget
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.C. Patin, D. Soulard, F. Trottein, C. Paget
Writing, review, and/or revision of the manuscript: E.C. Patin, M. Hassane, D. Dombrowicz, C. Faveeuw, F. Trottein, C. Paget
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.C. Patin, D. Soulard, M. Hassane, C. Faveeuw
Study supervision: F. Trottein, C. Paget
Acknowledgments
This work was supported by the recurrent annual financial support from INSERM. E.C. Patin was supported by a post-doctoral fellowship from the French Institute of Cancer (INCa). C. Paget, S. Fleury, D. Dombrowicz, and C. Faveeuw were supported by INSERM. F. Trottein was supported by CNRS. M. Hassane was the recipient of a doctoral fellowship from the AZM Foundation. We thank the Pasteur Lille animal facility for excellent mouse husbanding. We also thank the BICeL flow cytometry core facility for technical assistance. This work was supported by the “Institut National du Cancer” (INCa, PLBIO14-155; recipient: C. Paget).
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.