SOCS3 Transactivation by PPARγ Prevents IL-17–Driven Cancer Growth

Inflammation plays contrasting roles during cancer progression. Acute inflammation can be beneficial for cancer regression as shown by the requirement of HMGB1 and interleukin (IL)-1β for the induction of CD8 T-cell polarization and anticancer immune responses (1, 2). In contrast, proinflammatory mediators can support growth of established cancers by supporting an inflammatory milieu within the tumor microenvironment that favors tumor cell survival and neoangiogenesis (3). IL-17–producing CD4 T cells [T-helper (T_H)17 cells] have recently emerged as a key T-cell subset that produces proinflammatory mediators such as IL-17A, IL-17F, IL-21, and TNF-α and promotes autoimmunity (4). T_H17 cells can be induced from naïve T cells using stimulation with TGF-β and IL-6 and are characterized by the expression of their lineage-specific transcription factor RORγt (5). Although Rorc is indispensable for T_H17 cell induction, STAT3 is also an essential transcription factor for T_H17 cell differentiation as it directly regulates il17a/f and Rorc gene transcription (6, 7). IL-17 secretion from T_H17 cells has been proposed to contribute to tumor progression through the promotion of tumor neoangiogenesis (8).

The beneficial anti-inflammatory properties of dietary intake of (n-3) polyunsaturated fatty acids (PUFA) were initially illustrated by epidemiologic studies reporting a lower incidence of cardiovascular diseases in Eskimos compared with Western populations (9, 10). Administration of the n-3 PUFA docosahexaenoic acid (DHA) was later shown to dampen proinflammatory mediator secretion in patients with cancer (11). Activation of the transcription factor PPARγ has been proposed to account for the ability of DHA to prevent proinflammatory cytokine release (12).

Although DHA anticancer efficacy has been reported in preclinical mouse models of established tumors (13), whether the anticancer activity of DHA relies on its ability to modulate the production of proinflammatory cytokines remains elusive. Here, we report that activation of the PPARγ transcription factor by DHA in developing T_H17 cells induces Socs3 expression and interferes with the STAT3 signaling pathway, thereby inhibiting T_H17 cell differentiation. In addition, DHA prevents in vivo tumor growth of B16F10 melanoma and 4T1 mammary adenocarcinoma tumors in an IL-17–dependent manner. Altogether, not only our study shows that DHA administration prevents tumor angiogenesis and growth in an IL-17–dependent manner, but it also uncovers a previously unrecognized molecular pathway by which PPARγ-induced SOCS3 expression in T_H17 cells prevents IL-17 secretion.

All animals were bred and maintained according to both the Federation of Laboratory Animal Science Associations (FELASA) and the Animal Experimental Ethics Committee Guidelines (University of Burgundy, Dijon, France). Animals were used between 6 and 22 weeks of age. Female C57BL/6, BALB/c and Nude mice were purchased from Centre d'élevage Janvier and from Charles River Laboratories. RORγt-eGFP reporter mice were obtained from Dr. Eberl (Institut Pasteur, Paris, France) and have been previously described (14). IL-17a^−/− mice were obtained from Dr. Y. Iwakura (University of Tokyo, Tokyo, Japan; ref.15). γδTCR-deficient mice were kindly provided by Dr. Bernhard Ryffel (University and CNRS, Orléans, France). Rag2^−/− mice were provided by the CDTA (Cryopréservation, Distribution, Typage et Archivage animal) and distributed by EMMA (European Mouse Mutant Archive, a service funded by the EC FP7 Capacities Specific Programme). PPARβ- and PPARγ-deficient mice were provided by Pr. Béatrice Desvergne (Centre for Integrative Genomics, University of Lausanne, Lausanne, Switzerland) and have been described previously (16, 17).

To investigate the effect of dietary DHA in mice, we constituted 2 diet groups. One was fed a control diet containing sunflower oil, and the other group of mice was fed a DHA-enriched diet containing Omegavie DHA90 TG (Polaris Nutritional Lipids) for at least 3 weeks before starting experiments. Diets were prepared as described by Triboulot and colleagues (Supplementary Fig. S1; ref. 18).

To investigate the effect of dietary DHA in 2 human healthy volunteers (2 of the authors), 860 mg/day of DHA (Nutrixeal) was provided as a diet supplement for 3 weeks. Naïve CD4 T-cell differentiation into T_H17 cells was then assessed.

Total lipids were extracted from the diets according to the method of Bligh and Dyer (19), then transmethylated by BF3/methanol after saponification and fatty acids were analyzed by gas–liquid chromatography using C₁₅:0 as internal standard with a Becker gas chromatograph (Becker Instruments) equipped with a 50 m capillary glass column packed with carbowax 20 m (Spiral-RD) as described previously (18).

Direct quantification of nonesterified fatty acid in plasma was conducted by a one-step method of derivatization and extraction using (trimethylsilyl)diazomethane (Sigma) in the presence of C₁₅:0 as internal standard (20) with subsequent analysis by conventional gas–liquid chromatography as described earlier.

B16F10 melanoma and 4T1 mammary adenocarcinoma cancer cells were cultured at 37°C under 5% CO₂ in RPMI-1640 with GlutaMax-1 (Lonza) supplemented with 10% (v/v) fetal calf serum (Lonza), 1% penicillin, streptomycin, amphotericin B (Gibco), 4 mmol/L HEPES (Gibco), and 1 mmol/L sodium pyruvate (Gibco). 4T1 cells were kindly provided by Dr. Trad (Université de Bourgogne, Dijon, France) and B16F10 cells were obtained from American Type Culture Collection. All cells were routinely tested for Mycoplasma contamination using Mycoalert Mycoplasma Detection Kit (Lonza) and found negative. To induce tumor formation, 3 × 10⁵ B16F10 or 10⁵ 4T1 cancer cells were injected subcutaneously into Nude or immunocompetent mice. In vivo IL-6 and IL-17 neutralization were respectively achieved by injecting 200 μg intraperitoneally (i.p.) of an anti-IL-6 (clone MP5-20F3; BioXCell) or an anti-IL-17 antibody (clone 17F3; BioXCell) on day 0, 1, 2, 3, 4, and 6 following tumor implantation. In vivo IL-1 blockade was carried out by injecting i.p. 30 μg of IL-1 receptor antagonist (IL-1Ra; Kineret from Biovitrum) 3 times a week. All experiments were carried out in accordance with guidelines prescribed by the Ethics Committee at the University of Burgundy.

Tumors were cryosliced from mouse biopsies and permeabilized with 0.1% Triton X-100. Sections were probed with a rat monoclonal against CD31 (BD Biosciences—Pharmingen) followed by a secondary antibody coupled to horseradish peroxidase (HRP) and counterstained with Harris's hematoxylin.

Naïve CD4⁺ T cells (CD4⁺CD62L^hiCD44^lo) were obtained from spleens and lymph nodes of C57BL/6 wild-type (WT) or PPARβ- or PPARγ-deficient mice. CD4⁺ T cells were purified from spleen and lymph nodes with anti-CD4 microbeads (Miltenyi Biotec), then were further sorted as naïve CD4⁺CD62L^hiCD44^lo T cells. Isolated naïve T cells were routinely 98% pure. Isolated naïve CD4⁺ T cells were stimulated with plate-bound antibodies against CD3 (145-2C11, 2 μg/mL; BioXCell) and CD28 (PV-1, 2 μg/mL; BioXCell) and polarized into effector CD4⁺ T lymphocyte subsets with cytokines in presence or not of DHA (Sigma-Aldrich). Mouse IL-4 (20 ng/mL), IL-6 (25 ng/mL), IL-12 (10 ng/mL), and TGF-β (2 ng/mL) were all purchased from MiltenyiBiotec. Anti-IL-4 (clone 11B11) and anti-IFN-γ (clone XMG1.2; 10 μg/mL) antibodies were obtained from Bio-XCell. In some experiments, the PPARγ agonist troglitazone and the antagonist GW9662 have been added at final concentrations of 0.75 and 0.1 μmol/L, respectively (Cayman Chemical). The PPARα and PPARβ antagonists (GW6471 and GSK0660; Sigma-Aldrich) have been supplemented to reach a final concentration of 0.24 and 0.16 μmol/L, respectively. Cells were classically harvested on day 3 (unless otherwise specified) for detection of cytokines by ELISA and quantitative real-time PCR (qRT-PCR) analysis.

Experiments on human CD4⁺ T cells were carried out using peripheral blood from healthy volunteers (provided by the “Etablissement Français du Sang” Besançon, France). Written informed consent was obtained from all healthy blood donors. Naïve CD4⁺ T cells were isolated with the human CD4⁺ T Cell Isolation Kit II (MiltenyiBiotec), followed by stimulation with the Expansion/Activation Kit (MiltenyiBiotec), and differentiated for 3 days into T_H0 cells without cytokines or T_H17 cells in presence of TGF-β (2 ng/mL), IL-6 (10 ng/mL), and IL-23 (10 ng/mL).

For transfection experiments, naïve CD4⁺ T cells were transfected in vitro with Silencer Select Predesigned siRNA specific for murine SOCS3 (ID: s72013; Ambion, Life Technologies) or murine PPARγ (ID: 160219; Ambion) or Silencer Negative Control N°1 (Ambion) with transfection reagent TransIT-TKO (Mirus Bio LLC) according to the manufacturer's instructions. Twenty-four hours after transfection, CD4⁺ T cells were restimulated with anti-CD3 and anti-CD28, differentiated in T_H0 or T_H17 conditions as described earlier and cultured for an additional 24 or 48 hours before analysis.

After 72-hour polarization, cell culture supernatants were assayed by ELISA for mouse IL-4 (BD Biosciences), IL-17a (Biolegend), IFN-γ (BD Biosciences), or human IL-17 (Biolegend) according to the manufacturer's protocol.

For intracellular cytokine staining, cells were cultured for 5 days and then stimulated for 4 hours at 37°C in culture medium containing phorbol 12-myristate 13-acetate (PMA; 50 ng/mL; Sigma-Aldrich), ionomycin (1 μg/mL; Sigma-Aldrich), and monensin (GolgiStop; 1 μL/mL; BD Biosciences). After staining for surface markers and 7-amino-actinomycin D (7-AAD) to exclude dead cells, cells were fixed and permeabilized according to the manufacturer's instructions (Cytofix/Cytoperm Kit; BD Biosciences), then stained for intracellular products. Monoclonal antibodies (mAb) used for flow cytometry analyses were as follows: fluorescein isothiocyanate (FITC)–conjugated anti-CD4 (GK1.5; BD Biosciences), Alexa Fluor 647–conjugated anti-IL-17 (eBio 17B7; eBiosciences), and phycoerythrin (PE)-conjugated anti-IL-4 (554435; BD Biosciences) or PE-conjugated anti-IFN-γ (554412; BD Biosciences).

All events were acquired by a BD LSR-II cytometer equipped with BD FACSDiva software (BD Biosciences) and data were analyzed using FlowJo software (Tree Star).

Purified naïve T cells were differentiated for 24 to 48 hours into T_H0 or T_H17 cells with or without DHA (20 μmol/L), then collected and pelleted by centrifugation (5 minutes, 1,500 × g). Cells were lysed in boiling buffer [1% SDS, 1 mmol/L sodium orthovanadate, and 10 mmol/L Tris (pH 7.4)] containing protease inhibitor cocktail for 20 minutes at 4°C. Cell lysates were subjected to sonication (10 seconds at 10%) and protein concentration was assessed using the Bio-Rad DC Protein Assay Kit. Proteins were then denaturated, loaded, and separated on SDS-PAGE and transferred on nitrocellulose membranes (Schleicher & Schuell). After blocking with 5% bovine serum albumin (BSA) in TBS–0.1% Tween 20 (TBST), membranes were incubated overnight with primary antibody diluted in TBST containing 1% BSA, washed and incubated for 1 hour with secondary antibody diluted in TBST–1% BSA. After additional washes, membranes were incubated with luminol reagent (Santa Cruz Biotechnology) and exposed to X-ray films. The following mouse mAbs were used: anti-STAT3 (Cell Signaling Technology), anti-SOCS3 (Novus Biologicals), and anti-β-actin (Sigma-Aldrich). The following rabbit polyclonal antibody was used: anti-phospho-STAT3-Tyr705 (Cell Signaling Technology). Secondary antibodies HRP-conjugated polyclonal goat anti-mouse (Jackson ImmunoResearch) and polyclonal rabbit anti-goat immunoglobulins (Dako) were also used.

Cells were differentiated for 24 hours into T_H17 cells with or without DHA. Chromatin immunoprecipitation (ChIP) was conducted according to the manufacturer's instructions (ChiP-IT Express Enzymatic; Active Motif). Briefly, cells were fixed in a solution containing 37% formaldehyde for 10 minutes and quenched with 0.125 mol/L glycine. Chromatin was isolated and sheared to an average length of 300 to 500 bp by Enzymatic Cocktail. Twenty-five micrograms of DNA were immunoprecipitated with a PPARγ-specific antibody (H-1000; Santa Cruz Biotechnology). After chromatin elution, cross-links were reversed by a Reverse Cross-linking Buffer and qRT-PCR was conducted. Data were normalized to Actin-b C_t values and expressed in fold enrichment according to immunoglobulin (Ig) values. Primers designed to assess ChIP assay are as follows: Actb 5′-actctttgcagccacattcc-3′ and 3′-agcgtctggttcccaatact-5′; Socs3 1679 5′-gtcgacattccttctcaggttt-3′ and 5′-gcaccccttcccttttctttt-3′; and Socs3 576 5′-cccaggtcctttgcctgatt-3′ and 5′-tgagagaggggaccaggagaaa-3′.

Total RNA from T cells was extracted with TriReagent (Ambion), reverse transcribed using M-MLV Reverse Transcriptase (Invitrogen), and was analyzed by qRT-PCR with the SYBR Green method according to the manufacturer's instructions using the 7500 Fast Real-Time PCR System (Applied Biosystems). Expression was normalized to the expression of mouse Actb. Primers designed to assess gene expression are as follows: Actb 5′-atggaggggaatacagccc-3′ and 5′-ttctttgcagctccttcgtt-3′; Tbx21 5′-atcctgtaatggcttgtggg-3′ and 5′-tcaaccagcaccagacagag-3′; Gata-3 5′-aggatgtccctgctctcctt-3′ and 5′-gcctgcggactctaccataa-3′; Foxp3 5′-ctcgtctgaaggcagagtca-3′ and 5′-tggcagagaggtattgaggg-3′; Rorc mus musculus 5′-ggtgataaccccgtagtgga-3′ and 5′-ctgcaaagaagacccacacc-3′; Rora 5′-cccctactgttccttcacca-3′ and 5′-tgccacatcacctctctctg-3′; Ahr 5′-ctccttcttgcaaatcctgc-3′ and 5′-ggccaagagcttctttgatg-3′; il17f 5′-ttgatgcagcctgagtgtct-3′ and aattccagaaccgctccagt-3′; il21 5′-aaaacaggcaaaagctgcat-3′ and 5′-tgacattgttgaacagctgaaa-3′; il22 5′-tcgccttgatctctccactc-3′ and gctcagctcctgtcacatca-3′; cd4 5′-cctgtgcaagaagcagagtg-3′ and 3′-gttctgctgattccccttcc-5′; Tgfb1 5′-caacccaggtccttcctaaa-3′ and 3′-ggagagccctggataccaac-5′; Rorc homo sapiens 5′-aagcaggagcaatggaagtg-3′ and 3′-gcaatctcatcctcggaaaa-5′; Socs3 mus musculus 5′-aacttgctgtgggtgaccat-3′ and 5′-aaggccggagatttcgct-3′; Socs3 homo sapiens 5′-tggatggagcgggaggct-3′ and 5′-acgggacatctttcacctcaggctcct-3′; gpr120 5′-gatttctcctatgcggttgg-3′ and 5′-cccctctgcatcttgttcc-3′; mmp9 ctgtcggctgtggttcagt-3′ and agacgacatagacggcatcc-3′; cyp27 5′-gggcactagccagattcaca-3′ and 5′-ctatgtgctgcacttgccc-3′, clu 5′-ccattgtcccagatcagca-3′ and 5′-aggaggagcgcactggag-3′; vegfr 5′-aagagagtctggcctgcttg-3′ and 5′-ctgctcgggtgtctgctt-3′; and vegf 5′-aatgctttctccgctctgaa-3′ and 5′-gcttcctacagcacagcaga-3′.

Statistical analysis was conducted using Prism software (Graph Pad software). For the analysis of experimental data, comparison of continuous data was achieved by the Mann–Whitney U test and comparison of categorical data by Fisher exact test, as appropriate. All P values are two-tailed. P values less than 0.05 were considered significant. Data are represented as mean ± SD.

Despite reports suggesting that DHA could dampen the production of proinflammatory cytokines from innate immune cells (21, 22), whether DHA affects naïve CD4 T-cell differentiation in a cell-intrinsic manner has not been explored. To test this, we first examined the ability of naïve CD4⁺ CD62L⁺ T cells isolated from mice fed a control or DHA-enriched diet for 3 weeks to differentiate into effector T-cell subsets following activation in absence of antigen-presenting cells. DHA-enriched diet led to accumulation of DHA within splenocytes as previously described (Supplementary Fig. S2; ref. 18). Upon examining T-cell polarization, DHA failed to significantly alter the expression of the transcription factors Tbx21, Gata3, and Foxp3 that respectively specify the T_H1, T_H2, and regulatory T cell (Treg) lineages (Fig. 1A). However, naïve CD4⁺ T cells isolated from mice under a DHA-enriched diet featured a poor capacity to differentiate into T_H17 cells under T_H17-polarizing conditions, as illustrated by their weak expression of the T_H17-cell transcription factor Rorc (Fig. 1B). By assessing IL-17a secretion from T_H17 cells generated from mice under a DHA-enriched diet by ELISA, we found reduced IL-17a secretion in comparison with controls, whereas IFN-γ or IL-4 production from T_H1 and T_H2 cells were not affected (Fig. 1C). Intracellular staining data confirmed dampened IL-17a secretion from T_H17 cells from mice receiving DHA in vivo, whereas the respective production of IFN-γ or IL-4 from T_H1 and T_H2 cells remained unaffected (Fig. 1D). Overall, these results suggest that DHA selectively affects T_H17 cell differentiation in vivo.

To test whether DHA directly affects T_H17 cell development, we differentiated naïve CD4⁺ T cells without antigen-presenting cells under T_H17 cell-polarizing conditions in the presence of DHA. We observed a marked reduction of Rorc expression in T_H17 cells upon DHA treatment, confirming that DHA impairs T_H17 cell differentiation (Fig. 2A). This inhibition is specific for DHA as other PUFAs such as eicosapentaenoic or arachidonic acids did not impair T_H17 cell differentiation (Fig. 2A). To study further the impact of DHA on T_H17 differentiation, we generated T_H17 cells from Rorγt-eGFP reporter mice with increasing doses of DHA. DHA reduced the proportion of Rorγt-expressing cells in a dose-dependent manner, corroborating the DHA capacity to inhibit T_H17 cell differentiation (Fig. 2B). In addition, we also explored the ability of DHA to affect other T_H17-related genes and found that DHA impaired the expression of Rora, Ahr, i17f, il21, and il22, suggesting that DHA globally affects the T_H17 cell differentiation program (Fig. 2C). IL-23 is essential to maintain the proinflammatory T_H17 cell program both in vitro and in vivo (23). Because DHA blunted the primary differentiation of naïve CD4 T cells into effector T_H17 cells, we explored the ability of DHA to influence IL-23 capacity to restimulate engaged T_H17 cells. We found that DHA was also able to reduce IL-17a secretion from engaged T_H17 cells restimulated with IL-23 (Fig. 2D). To test whether DHA skewed T_H17 cell differentiation toward another CD4 effector T-cell subset, we monitored the expression of Tbx21, Gata3, and Foxp3 genes in T_H17 cells treated with DHA. We found that DHA failed to induce any of the transcription factors defining the T_H1, T_H2, or Treg lineages, indicating that DHA restricts but does not skew T_H17 cell differentiation in vitro (Fig. 2E). We tested whether DHA also affects human T_H17 cell differentiation. For this, we first differentiated naïve human CD4⁺ T cells isolated from the blood of healthy volunteers into T_H17 cells in the presence of DHA. In line with our findings with mouse T cells, we found that DHA also prevented human T_H17 cell differentiation (Fig. 2F). Similarly, upon assessing T_H17 cell induction from naïve human CD4 T cells obtained from healthy individuals before or after 3 weeks of DHA intake as a diet supplement, we found that oral DHA intake decreased T_H17 cell differentiation (Fig. 2G). Altogether, these findings indicate that DHA restrains mouse and human T_H17 cell differentiation.

The transcription factor STAT3 was shown to transactivate the Il17a/f and the Rorc promoters (24, 25). Because DHA dampened IL-17a/f secretion and Rorc expression in developing T_H17 cells, we hypothesized that DHA negatively regulates STAT3 activation during T_H17 cell differentiation. To test this, we induced T_H17 cells from naïve CD4⁺ T cells in vitro in presence or absence of DHA. As expected, we noted a robust STAT3 (p-STAT3) phosphorylation in T_H17 cells generated without DHA (Fig. 3A). In contrast, CD4 T cells incubated under T_H17-polarizing conditions with DHA featured a weak level of p-STAT3, confirming that DHA reduces STAT3 signaling in T_H17 cells (Fig. 3A). Because SOCS3 is a major regulator of STAT3 phosphorylation in T_H17 cells (24), we hypothesized that the DHA-mediated downregulation of p-STAT3 may be due to increased SOCS3 expression. We found that SOCS3 mRNA and protein expression was induced by DHA treatment in mouse and human T_H17 cells, suggesting that the DHA-mediated downregulation of T_H17 cell differentiation could be related to SOCS3 induction (Fig. 3B). To investigate the involvement of SOCS3 in the capacity of DHA to control T_H17 cell differentiation, we downregulated SOCS3 expression in T cells differentiated under T_H17-cell skewing conditions with DHA using siRNA. We found that SOCS3 silencing blunted the capacity of DHA to restrain IL-17 and RORγt expression in T_H17 cells (Fig. 3C and D). Collectively, our findings suggest that DHA restrains T_H17 cell differentiation by interfering with STAT3 signaling through SOCS3 induction.

To unravel the mechanism accounting for SOCS3 induction in developing T_H17 cells upon DHA treatment, we investigated the intracellular signaling pathways engaged by DHA on T_H17 cells. GPR120 was recently identified as an omega-3 fatty acid receptor (26). To test the involvement of GPR120 in the DHA-driven impairment of T_H17 cell differentiation, we downregulated GPR120 expression with siRNA in naïve T cells and analyzed the effect of DHA during T_H17 differentiation. Our results showed no significant differences in T_H17 polarization between cells transfected with control or GPR120 siRNA, indicating that the decreased T_H17 cell induction upon DHA treatment does not depend on GPR120 (Fig. 4A). PPARγ is another proposed receptor for omega-3 fatty acid. Synthetic PPARγ ligands have been reported to induce Socs3 expression and inhibit STAT3 activation (27). Given that DHA interfered with T_H17 cell differentiation through STAT3 signaling pathway blockade, we hypothesized that the effects of DHA on T_H17 cells relied on the transcription factor PPARγ. Gene expression analysis revealed that the known PPARγ target genes MMP9, Cyp27, and Clu were induced in DHA-treated T_H17 cells, suggesting that PPARγ signaling is activated by DHA treatment on T_H17 cells (Fig. 4B). In addition, the DHA effect on T_H17 cell differentiation was markedly reduced in the presence of a PPARγ inhibitor GW9662, but not with a PPARβ inhibitor GSK0660 or PPARα inhibitor GW6471, as measured by IL-17 secretion and Rorc expression (Fig. 4C). Moreover, PPARγ agonist troglitazone neutralized IL-17 secretion from T_H17 cells to a comparable extent to DHA, suggesting that PPARγ activation is involved in DHA effects on T_H17 cells (Fig. 4C). To validate the elective involvement of PPARγ in the intracellular effects of DHA on T_H17 cells, we downregulated PPARγ expression with siRNA and tested the ability of DHA to inhibit T_H17 cell differentiation in absence of PPARγ. We found that silencing PPARγ completely abolished the ability of DHA to restrain T_H17 cell induction (Fig. 4D). Although PPARγ agonists induce Socs3 expression (27, 28), the direct link between PPARγ and SOCS3 remains unclear. To test whether the DHA-driven Socs3 induction is directly due to PPARγ activation, we conducted a bioinformatic analysis of the Socs3 promoter and found 2 putative binding sites for PPARγ at positions −1679 and −576 from the transcription start site. By conducting ChIP, we found that PPARγ binds to the Socs3 promoter only at position −576 in DHA-treated T_H17 cells as compared with control T_H17 cells (Fig. 4E). We further checked the specificity of this interaction by downregulating PPARγ expression by siRNA and confirmed that the amount of PPARγ-binding Socs3 promoter was accordingly reduced (Fig. 4E). Finally, we tested the effects of DHA on naïve T cells from WT-, PPARγ-, or PPARβ-deficient mice. In line with our results obtained with the PPAR inhibitors, we found that while DHA could suppress T_H17 cell differentiation from WT- and PPARβ-deficient T cells, PPARγ-deficient T cells were totally unaffected by DHA treatment (Fig. 4F). Altogether, these findings show that the ability of DHA to hamper T_H17 cell differentiation is strictly dependent on PPARγ-dependent Socs3 expression.

Although DHA was proposed to promote cancer cell death (29), we found that DHA did not exert direct cytotoxic effects on mouse melanoma B16F10 and in 4T1 mammary adenocarcinoma tumor cells in vitro at concentrations that could be reached in vivo (60 μmol/L; Fig. 5A). However, we noted that the DHA-enriched diet delays subcutaneous tumor growth in vivo in both cancer models (Fig. 5B). In contrast, DHA did not show any antitumor efficacy in nude mice in both models suggesting that DHA mediates its anticancer function through T cells (Fig. 5C and Supplementary Fig. S3). Given that DHA impedes proinflammatory cytokine release, which has been proposed to contribute to B16F10 and 4T1 tumor growth (8, 30–32), we tested whether DHA affected tumor growth in vivo in these 2 tumor models by affecting IL-1, IL-6, or IL-17 release. DHA still exhibited in vivo anticancer activity against B16F10 tumors in absence of IL-1 and IL-6 (Supplementary Fig. S4). However, DHA failed to exhibit any effect on tumor progression in IL-17a^−/− mice, suggesting that the anticancer effects of DHA are dependent on IL-17 (Fig. 5D). These results were confirmed by neutralizing IL-17 in vivo in the 4T1 tumor model (Fig. 5D). As a control, IL-17 mAb did not exert any direct cytotoxic effect on tumor cells, ruling out any direct effect of IL-17 on tumor cell growth (Supplementary Fig. S5).

Because γδ T cells and T_H17 cells are major IL-17 sources in vivo, we investigated whether DHA anticancer efficacy relied on γδ T cells and IL-17–producing CD4 T cells. We found that the DHA-enriched diet could prevent tumor growth in γδTCR-deficient mice, ruling out any significant contribution of these cells in DHA anticancer activity (Fig. 6A). However, upon testing DHA antitumor effect in T-cell–deficient Rag2^−/− mice reconstituted with either IL-17^+/+ or IL-17^−/− αβ CD4 T cells, we failed to note any therapeutic benefit of the DHA diet in mice that had received IL-17–deficient CD4 T cells (Fig. 6B). We also measured T_H17 cell responses from WT mice under the control or DHA-enriched diet. In line with our previous findings, draining lymph node CD4 T cells from B16F10 tumor-bearing mice under the DHA-enriched diet expressed lower amounts of IL-17 and Rorc than controls (Fig. 6C). Together these data suggest that DHA may affect tumor growth via this capacity to regulate IL-17 secretion from T_H17 cells.

IL-17A is a proangiogenic factor that favors neovascularization and tumor growth in vivo via its capacity to induce VEGF production (8, 33). In line with the documented angiogenic properties of IL-17A, we noted that B16F10 tumors from mice under DHA-enriched diet featured decreased tumor vasculature reflected by reduced vegfr and vegf expression and reduced number of CD31-positive blood vessels (Fig. 6D). Finally, to determine the role of VEGF in the antitumor effect of DHA, we treated tumor-bearing mice fed or not a DHA-enriched diet with the VEGF receptor (VEGFR)-blocking mAb, DC101. We observed that while DHA and anti-VEGF therapy exhibited similar antitumor effects, these treatments failed to synergize (Fig. 6E). Collectively, these data suggest that the activity of DHA relies on the prevention of IL-17–dependent tumor neoangiogenesis.

Although activation of the transcription factor PPARγ has been associated with reduced inflammation, the intracellular series of events consecutive to PPARγ activation remain incompletely understood and the actual involvement of PPARγ in the anti-inflammatory effects of PPARγ ligands remains debated (34, 35). In this study, we propose what we believe to be a novel molecular pathway by which PPARγ contributes to limit the secretion of the proinflammatory cytokine IL-17. DHA activates PPARγ, which binds to the Socs3 promoter and drives Socs3 expression in nascent T_H17 cells, leading to reduced STAT3 phosphorylation, Rorc expression and IL-17 secretion. Finally, DHA prevents tumor growth in vivo through the downregulation of IL-17 production.

Although our data indicate that DHA inhibits T_H17 cell induction through downregulating the STAT3 signaling pathway, the strong inhibitory effect of DHA on T_H17 cell differentiation suggests that DHA could act at several levels on the T_H17 cell differentiation program. Indeed, activation of PPARγ has been shown to prevent the clearance of corepressors from the Rorγt promoter (36), resulting in repressed T_H17 cell generation. We cannot rule out that this signaling pathway plays an additional role in DHA effect on differentiating T_H17 cells.

The links between inflammation and cancer have been thoroughly documented as illustrated by data showing an increased risk of patients suffering from inflammatory bowel disease to develop cancer compared with healthy individuals (37). Chronic inflammation can indeed contribute to cell transformation and cancer formation. Dietary supplementation of DHA has been shown to alleviate the severity of intestinal inflammation (38, 39) in experimental models of colitis and in inflammatory bowel disease in humans (40, 41). These findings possibly explain why dietary (n-3) PUFAs intake was associated with decreased risk of cancer (42, 43). However, recent data indicate that DHA also improves outcome of chemotherapy in patients with metastatic breast cancer, suggesting that it not only affects cancer incidence, but can be therapeutically used to prevent tumor growth (44). Our study brings up evidence that the anticancer effects of DHA could rely on its ability to dampen IL-17 levels in vivo.

The role of T_H17 cells during cancer growth remains intensely debated. Adoptive transfer experiments have shown the anticancer potential of both mouse and human T_H17 cells (45, 46). It was further shown recently that the remarkable ability of T_H17 cells to eliminate tumors in these settings relied on their ability to self-renew and transdifferentiate into IFN-γ–producing cells in vivo (46). However, in some human cancers IL-17 was shown to be associated with poor prognosis (47–49). Mouse and human studies revealed an association between IL-17 levels in the tumor bed, VEGF, and blood vessels number (33, 50). In addition, IL-17 has the capacity to directly trigger VEGF release from VEGFR-expressing cells such as cancer cells and stromal cells (33). These data support the contention that IL-17 is an important factor, which drives VEGF-dependent neoangiogenesis during cancer growth. Anti-VEGF therapies are being implemented for cancer treatment in humans but are expensive. In this regard, our study supports the antiangiogenic property of a DHA-enriched diet via its capacity to regulate IL-17 production and VEGF-related angiogenesis. If such results could be confirmed in patients with cancer, DHA should be considered as a potential antiangiogenic candidate for future clinical trials.

Altogether, our study unraveled that DHA-mediated PPARγ activation limits T_H17 cell differentiation by inducing SOCS3 expression. Because our results suggest that the activity of DHA is relevant in humans, DHA administration might prove beneficial in diseases where IL-17 cytokine secretion is undesirable.

No potential conflicts of interest were disclosed.

Conception and design: S. Ladoire, A. Hichami, F. Ghiringhelli, L. Apetoh

Development of methodology: F. Végran, B. Ryffel, A. Hichami

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Berger, F. Végran, M. Chikh, F. Gilardi, S. Ladoire, H. Bugaut, G. Mignot, M. Bruchard, B. Ryffel, C. Pot, B. Desvergne, L. Apetoh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Berger, F. Végran, S. Ladoire, L. Apetoh

Writing, review, and/or revision of the manuscript: H. Berger, F. Végran, M. Chikh, F. Gilardi, G. Mignot, C. Rébé, C. Pot, B. Desvergne, F. Ghiringhelli, L. Apetoh

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F. Chalmin, V. Derangère, A. Chevriaux, C. Rébé

Study supervision: L. Apetoh

The authors are supported by grants from the Ligue Nationale contre le Cancer (F. Ghiringhelli and F. Végran), the Fondation de France (F. Ghiringhelli), the Institut National du Cancer (F. Ghiringhelli), the Association pour la recherche sur le cancer (F. Ghiringhelli and G. Mignot), the Conseil Régional Bourgogne/INSERM (H. Berger), the Etat de Vaud and the Swiss National Funds for Research (B. Desvergne), the CNRS, FEDER, Le Studium, Orléans and Fondation pour la Recherche Médicale (B. Ryffel), the Agence Nationale de la Recherche (ANR-10-PDOC-014-01; L. Apetoh), the Ligue Régionale contre le cancer Comité Grand-Est (L. Apetoh) and the European Community (Marie Curie Fellowship PCIG10-GA-2011-303719; L. Apetoh). This work was supported by a French Government grant managed by the French National Research Agency under the program “Investissements d'Avenir” with reference ANR-11-LABX-0021.

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.