Purinergic Signaling as a Regulator of Th17 Cell Plasticity

T helper type 17 (Th17) lymphocytes, characterized by the production of interleukin-17 and other pro-inflammatory cytokines, are present in intestinal lamina propria and have been described as important players driving intestinal inflammation. Recent evidence, supporting the notion of a functional and phenotypic instability of Th17 cells, has shown that Th17 differentiate into type 1 regulatory (Tr1) T cells during the resolution of intestinal inflammation. Moreover, it has been suggested that the expression of CD39 ectonucleotidase endows Th17 cells with immunosuppressive properties. However, the exact role of CD39 ectonucleotidase in Th17 cells has not been studied in the context of intestinal inflammation. Here we show that Th17 cells expressing CD39 ectonucleotidase can hydrolyze ATP and survive to ATP-induced cell death. Moreover, in vitro-generated Th17 cells expressing the CD39 ectonucleotidase produce IL-10 and are less pathogenic than CD39 negative Th17 cells in a model of experimental colitis in Rag-/- mice. Remarkably, we show that CD39 activity regulates the conversion of Th17 cells to IL-10-producing cells in vitro, which is abrogated in the presence of ATP and the CD39-specific inhibitor ARL67156. All these data suggest that CD39 expression by Th17 cells allows the depletion of ATP and is crucial for IL-10 production and survival during the resolution of intestinal inflammation.

Th17 cells are endowed with remarkable functional plasticity, being able to differentiate into Th1 cells in lymphopenic hosts, during EAE and in antitumor immune responses [11][12][13][14][15]. Some groups have also reported the presence of IFN-γ -producing Th17 cells in humans and patients with Crohn's disease [16,17]. Th17 cells can also differentiate into IL-10-producing cells during the resolution of inflammation [18][19][20]. Interestingly, using fate mapping mouse models, Flavell and colleagues have recently demonstrated that in the course of an inflammatory immune response, Th17 cells can transdifferentiate into type 1 regulatory T (Tr1) cells [21].
Two distinct types of Th17 cells have been defined: pathogenic or encephalitogenic Th17 cells and the so-called non-pathogenic Th17 cells. Some reports have demonstrated that Th17 cells generated with TGF-β1 and IL-6 are not pathogenic in the setting of EAE [18], whereas Th17 cells produced with IL-23 and without TGF-β1 are highly pathogenic in the same model [22][23][24]. Depending on the experimental setting, different cytokines produced by Th17 cells seem to drive their effector function. For instance, in tumor immunity, the production of IFNγ by Th17 cells has proved to be determining the potential of Th17 cells to eradicate an established tumor [14]. Other studies revealed that the encephalitogenic properties of Th17 cells depend on GM-CSF production [23,25], whereas IL-10 production has been strongly related to non-pathogenic Th17 cells in EAE [18]. Moreover, IL-10 production by Th17 cells has been strongly related to the acquisition of regulatory properties by Th17 cells and the resolution of intestinal inflammation [20,21].
Extracellular ATP (eATP) is a danger signal released by dying and damaged cells, and it functions as an immunostimulatory signal that promotes inflammation [26,27]. eATP can be sensed by purinergic P2 receptors including the cation-selective receptor channels (P2X) and metabotropic G protein-coupled receptors (P2Y) [28]. eATP has been shown to engage P2 receptors on T cells to induce Th17 differentiation [29]. On the other hand, high doses of eATP induces necrotic lysis through P2X7 receptor signaling in T cells [30,31].
CD39 (NTPDase-1) and CD73 (ecto-5'-nucleotidase) ectonucleotidases are two cell-surface ectoenzymes that dephosphorylate ATP into its metabolites ADP, AMP and adenosine thus shifting the balance from inflammatory to suppressive microenvironments [32]. Recently, it has been reported that ATP signaling through P2X7 receptor inhibits the conversion of naive T cells to Tr1 cells and that CD39 can promote Tr1 differentiation through eATP hydrolysis [33].
It is have been described that Th17 cells generated with TGF-β1 and IL-6 express CD39 and CD73 ectonucleotidases endowing these cells with the capacity to produce adenosine and promote immunosuppressive microenvironments [34]. However, the role of CD39 expression in other aspects such as in depleting eATP to reduce P2X7 receptor-mediated cytotoxicity or promoting the plasticity of this subset of T cells has not been explored. In the current study, we present a comprehensive comparison of in vitro-generated Th17 cells expressing high and low levels of CD39 and CD73 ectonucleotidases. We found that Th17 cells generated with TGF-β1 (Th17 TGF-β1 ) which express the CD39 and CD73 ectonucleotidases and Th17 cells generated with IL-23 (Th17 IL-23 ) which do not express these ectonucleotidases constitute two fundamentally different Th17 subsets, evidenced at the transcriptional level and by the types of cytokines they secrete. In agreement with their transcriptional profile, Th17 TGF-β1 cells were significantly less pathogenic than Th17 IL-23 cells in a model of experimental colitis, as they induce a transient reduction in weight loss in Rag -/mice and the production of IL-10 in the intestine. Interestingly, Th17 TGF-β1 cells can hydrolyze ATP in a CD39-dependent manner, providing Th17 TGF-β1 cells a superior survival capacity when exposed to high levels of ATP. Furthermore, here we show that Th17 TGF-β1 cells produce higher levels of IL-10 than Th17 IL-23 cells when reactivated in vitro in the presence of Tr1-polarizing cytokines. Finally, we report that CD39 activity is important for IL-10 production by Th17 TGF-β1 cells since CD39 inhibition using the specific inhibitor ARL67156 reduced IL-10 production by in vitro re-activated Th17 cells.

Materials and Methods Mice
C57BL/6, B6SJL-PTPRC (CD45.1), OT-II, IL-17-GFP, Rag1 -/-, P2X7R -/mice were purchased from The Jackson Laboratory. All mice were kept in an animal facility under standard housing guidelines. Animal work was carried out under institutional regulations of Fundación Ciencia & Vida and was approved locally by the ethical review committee of the Facultad de Ciencias, Universidad de Chile.

Induction of colitis in Rag -/mice
For experimental colitis experiments, 1.3x10 6 Th17 TGF-β1 or Th17 IL-23 cells were sorted based on IL-17 production (GFP+) and then transferred into Rag -/mice. The body weight was measured every 2 days. Six weeks after adoptive transfer, the mice were sacrificed, and the entire colon was removed from cecum to anus. The colon length was measured as an indicator of inflammation. Clinical score was calculated based on weight loss and colon length. Weight-loss scores were determined as 0 = 0-2.5% weight loss; 1 = 2.5-5% weight loss; 2 = 5-7.5% weight loss; 3 = 7.5-10% weight loss; and 4 = >10% weight loss. This score was calculated using the weight of each mouse at the end point. Each weight data was compared to the average weight of control group. Colon length scores were determined as 0 = no colon size reduction; 1 = 0-5% colon size reduction; 2 = 5-10% colon size reduction; 3 = 10-15% colon size reduction; and 4 = >15% colon size reduction. This score was calculated using colon length normalized by the weight of each mouse. For each mouse, these scores were combined and divided by two to give an overall clinical score ranging from 0 (healthy) to 4 (maximal colitis).

Analysis of transferred cells in Rag -/mice
Six to eight weeks after adoptive transfer of Th17 TGF-β1 or Th17 IL-23 cells into Rag-/-mice, the mice were sacrificed and lymphoid organs and lamina propria were dissected. The cells were analyzed by flow cytometry to assess the percentage of the transferred cells (CD3+ CD4+) within a lymphoid gate and the production of cytokines by intracellular cytokine staining.

Intracellular staining and flow cytometry
Cells obtained from lamina propria, lymph nodes and in vitro-generated Th17 cells were restimulated with 0.25 μM PMA (Sigma-Aldrich) and 1 μg/ml ionomycin (Sigma-Aldrich) in the presence of GolgiPlug (BD Biosciences) for 4 h. Cells were stained with antibodies against the cell surface markers CD4, CD39, CD73, and then resuspended in a fixation/permeabilization solution (Cytofix/Cytoperm; BD Pharmingen). Following fixation and permeabilization, the cells were incubated with antibodies against IFN-γ, IL-17 and IL-10 for 30 min at 4°. The cells were then washed with permeabilization buffer and resuspended in PBS + 2% FCS for FACS analysis (FACSCanto II; BD Bioscience). In some cases, Fixable Viability Dye (eBioscience) was used to discard dead cells from the analysis. Analysis of FACS data was performed using the FLOWJO software (Tree Star Inc., Ashland, OR).
To analyze the cytokines produced in the intestine of Rag -/mice transferred with Th17 cells, the intestine was cut into fragments of 1 cm of length and incubated in 1 ml IMDM + 10% FCS for 24 h at 37°C and 5% CO 2 . The medium was collected and centrifuged at 600 x g for 7 min. The supernatant was analyzed using the CBA Kit Mouse Th1/Th2/Th17 (BD Biosciences).
qPCR Th17 cells were isolated by cell sorting, and total RNA was obtained using EZNA Total RNA Kit I (O Bio-Tek). 1 μg of RNA was reverse transcribed using M-MLV reverse transcriptase (Invitrogen). The PCR reaction was performed using Brilliant II SYBR Green QPCR Master Mix (Agilent Technologies) in a Stratagene Mx3000P real-time PCR machine. For relative quantitation, the amplified fragments were normalized according to constitutive transcription of the housekeeping gene GAPDH. The sequences of the primers used for quantification of each measured transcript were the following:

Determination of ATP/AMP hydrolysis by HPLC
The enzymatic activity of CD39 and CD73 was evaluated based on the percentage of hydrolysis of ATP or AMP respectively by HPLC. Briefly, Th17 cells generated in vitro were diluted in Hanks' balanced salt solution (HBSS) and incubated in a 96-well flat-bottom plates at 0.5x10 5 cells/well with 10 μM ATP (Sigma-Aldrich) in the presence or absence of the CD39 inhibitor ARL67156 (Sigma-Aldrich) at a concentration of 50 μM or with 10 μM AMP (Sigma-Aldrich), with or without the CD73 inhibitor APCP (Adenosine 5'-(a,b-methylene) diphosphate) (50 μM) (Sigma-Aldrich). After 1 h, the cells were harvested, transferred to ice for 15 min, and then centrifuged at 1000 x g for 10 min. Supernatants were collected and stored at -20°C until further analysis. HPLC analysis was carried out in a Water Breeze system using an anion exchanger column (Mono Q; GE Healthcare, Chalfont St Giles, UK). The mobile phase used consisted of a linear gradient from buffer A (Tris-HCl 100 mM, pH 7.8) to buffer B (Tris-HCl 100 mM, NaCl 1 M, pH 7.8). The effluent was monitored at 257 nm using an online UV detector. The column was calibrated using ATP, AMP and adenosine as standards.

Determination of Th17 cell survival in high concentrations of ATP
Th17 cells isolated by cell sorting were resuspended at 1x10 5 cells/well in HBSS medium. Cells were cultured in the presence or absence of ATP (100, 500 or 1000 μM) (Sigma-Aldrich) for 30 min at 37°C and 5% CO 2 . Cells were then harvested and centrifuged at 600 x g for 7 min and resuspended in 100 μL of Binding Buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2, pH 7.4) containing 0.5 μL of Annexin V APC (Biolegend) and 2 μL of propidium iodide (50 μg/ mL) (Sigma). Cells were incubated for 20 min at room temperature and 300 μL of Binding Buffer was added. Live and dead cells were analyzed by flow cytometry.

Statistical analysis
Data are presented as mean ± SEM. Differences between groups were determined using Mann-Whitney test or two-tailed t-test. Where indicated, differences were analysed using Kruskal-Wallis or two-way analysis of variance paired with Bonferroni post-tests. Statistical analysis and graphs were obtained with GRAPHPAD PRISM (GraphPad Software Inc., La Jolla, CA).

Results
In vitrogenerated Th17 cells expressing high levels of CD39 ectonucleotidase present a regulatory phenotype To evaluate a putative role of CD39 and CD73 in Th17 cells we generated Th17 cells expressing high or low levels of these ectonucleotidases. It has been reported that TGF-β1 induces CD39 and CD73 expression in Th17 cells; however after a second round of restimulation in the absence of TGF-β1, Th17 cells lose ectonucleotidase expression [34]. For this reason, we generated Th17 cells using two rounds of activation. Th17 cells expressing high levels of CD39 and CD73 were generated with TGF-β1, IL-6, and IL-1β, and reactivated in the presence of TGF-β1 and IL-6 (Th17 TGF-β1 ). Th17 cells expressing low levels of ectonucleotidases were generated with TGF-β3, IL-6, and IL-1β, and reactivated in the presence of IL-6, IL-1β, and IL-23 (Th17 IL-23 ). As shown in Fig 1A and 1B, a similar percentage of IL-17 producing cells were obtained in Th17 TGF-β1 and Th17 IL-23 culturing conditions. Accordingly, Th17 TGF-β1 and Th17 IL-23 cells expressed similar levels of RORγt, the master transcription factor of Th17 cells and neither expressed GATA-3 nor Foxp3 transcription factors. Interestingly, although it has been reported that Th17 cells generated with IL-23 express higher levels of T-bet than cells generated with TGF-β1 [24,35], in our setting Th17 TGF-β1 cells expressed higher levels of this master transcription factor compared to Th17 IL-23 cells (Fig 1C and 1D). Next, we analyzed CD39 and CD73 expression by these in vitro-generated Th17 cells. As expected, Th17 TGF-β1 cells acquired high levels of CD39 and CD73 ectonucleotidases expression whereas Th17 IL-23 cells presented low expression of both ectonucleotidases (Fig 1E and  1F). A fraction of Th17 TGF-β1 cells (approx. 25%) also expressed significant levels of CD49b and Lag-3, markers of Tr1 cells [36], whereas a fraction of Th17 IL-23 cells only expressed Lag-3. Although a fraction of Th17 TGF-β1 cells showed Tr1 markers, these do not constitute bona fide Tr1 cells, since they express RORγt transcription factor and produce IL-17. Although both Th17 cell subsets expressed CCR6, Th17 TGF-β1 cells showed lower levels of this chemokine receptor compared to Th17 IL-23 cells (Fig 1G and 1J).
We further investigated the expression of Th17-associated genes in Th17 TGF-β1 and Th17 IL-23 cells. The expression of mRNAs encoding T-bet, IFN-γ, IL-9, AHR, c-Maf and IL-10 was higher in Th17 TGF-β1 cells whereas expression of mRNAs encoding GM-CSF and IL-22 was higher in Th17 IL-23 cells (Fig 2A). Interestingly, following in vitro reactivation, Th17 TGF-β1 cells secreted higher levels of IL-10, and lower levels of IL-17, GM-CSF, TNF, IL-2 compared to Th17 IL-23 cells (Fig 2B-2D). Taken together, these results strongly suggest that Th17 TGF-β1 cells rapidly lose their inflammatory potential and present a regulatory phenotype upon reactivation.

In vitrogenerated Th17 cells expressing high levels of CD39 and CD73 can hydrolyze ATP and survive to ATP-induced death
We next tested the enzymatic activity of CD39 and CD73 ectonucleotidases in Th17 cells. As shown in Fig 3A-3D, only Th17 TGF-β1 cells, which express the CD39 and CD73 ectonucleotidases, can hydrolyze ATP and AMP. Moreover, ATP hydrolysis was partially blocked in Th17 TGF-β1 cells by inhibiting CD39 enzymatic activity using the ecto-ATPase inhibitor ARL67156 while adenosine production was blocked using the CD73 inhibitor APCP in these cells.
As Th17 TGF-β1 cells express the CD39 and CD73 ectonucleotidases and produce adenosine, we tested whether these cells present suppressive capacity in vitro. For this, Th17 TGF-β1 or Th17 IL-23 cells were sorted based on IL-17-GFP expression and co-cultured with Violet-labeled CD4+ effector T cells from OT-II mice activated with OVA 323-339 and antigen presenting cells for 3 days. As shown in S1 Fig, both Th17 TGF-β1 and Th17 IL-23 cells displayed a low suppressive capacity only when co-cultured in direct contact with effector cells at high Th17/Teff ratios (1:1). No suppression was observed when the cells were cultured using transwell chambers (S1 Fig), suggesting that soluble factors such as adenosine are not involved in this process.
Since Th17 TGF-β1 cells can hydrolyze ATP, we sought to determine whether these cells were more resistant to ATP-induced cell death. As shown in Fig 3E and 3F, Th17 TGF-β1 cells were more resistant than Th17 IL-23 cells to rapid cell death induced by high doses of ATP  >500μM). Since both types of Th17 cells express similar levels of the mRNA encoding the P2X7 receptor (Fig 3G), this suggests that ATP signaling through the P2X7 receptor is reduced   in Th17 TGF-β1 cells due to CD39 activity. These results also raise the possibility that CD39 expression confers Th17 TGF-β1 cells with a superior survival capacity when faced with toxic doses of ATP.

In vitrogenerated Th17 cells expressing high levels of CD39 ectonucleotidase convert to IL-10 producing cells during intestinal inflammation
It has been reported that Th17 cells generated with TGF-β1 and IL-6 are less pathogenic than Th17 cells differentiated with IL-23 [18]. To evaluate the pathogenicity of our Th17 cells in a model of experimental colitis, we transferred Th17 TGF-β1 or Th17 IL-23 cells into Rag -/mice and measured body weight loss and colon length in these mice. As shown in Fig  4A, mice treated with Th17 TGF-β1 cells started losing body weight 3 weeks after adoptive transfer whereas mice transferred with Th17 IL-23 cells presented a delay in the onset of weight loss and started losing weight 5 weeks following adoptive transfer. Moreover, Rag -/mice treated with Th17 TGF-β1 cells presented a transient weight loss and began to recover by week 5, whereas mice treated with Th17 IL-23 cells presented severe weight loss and had to be sacrificed by 6 weeks after the adoptive transfer (Fig 4A). In agreement with the severe and persistent weight loss, mice treated with Th17 IL-23 cells presented a significant reduction in the colon length compared to mice receiving PBS (Fig 4B). Although not statistically significant, mice treated with Th17 TGF-β1 cells presented a lower clinical score compared to mice treated with Th17 IL-23 cells (Fig 4C). H&E and alcian blue staining of distal colonic sections revealed severe wall thickening, extensive leukocyte infiltration, and disruption of intestinal crypts and goblet cells in mice treated with Th17 TGF-β1 or Th17 IL-23 cells compared to controls (Fig 4D). This suggests that although mice treated with Th17 TGF-β1 cells recover and gain weight, the damage to the colon is not reversed in these mice at this time point.
In agreement with the less severe body weight loss, mice treated with Th17 TGF-β1 cells secreted higher levels of IL-10 as determined directly in their small intestine compared to mice treated with Th17 IL-23 cells (Fig 4E). We next analyzed the stability of Th17 TGF-β1 and Th17 IL-23 cells 6 weeks after adoptive transfer and found that both Th17 cell subsets lost IL-17 production and a significant fraction of transferred cells produced IFN-γ (Fig 4G). Notably, a small fraction of transferred Th17 TGF-β1 cells, but not of Th17 IL-23 cells, were able to produce IL-10 ( Fig 4F). The differences observed in body weight loss were not due to a differential capacity of of Th17 TGF-β1 and Th17 IL-23 cell subsets to survive or persist in the host, since both populations were found at similar percentages within spleen, mesenteric lymph nodes and small intestine lamina propria up to 8 weeks following adoptive transfer into Rag -/mice (S2 Fig). Taken together, these data suggest that Th17 TGF-β1 cells are able to induce the production of IL-10 in the intestine of Rag -/mice and can convert into IL-10-producing cells during the recovery from intestinal inflammation, which may explain the transient weight loss observed in mice treated with Th17 TGF-β1 cells.

CD39-mediated ATP hydrolysis is crucial for in vitro Th17 conversion to IL-10 producing cells
To test the possibility that Th17 TGF-β1 cells differentiate into IL-10-producing cells more efficiently than Th17 IL-23 cells, we reactivated IL-17 (GFP+) Th17 cells with anti-CD3 and anti-CD28 antibodies for 3 days and evaluated IL-10 production in vitro. As shown in Fig 5A, Th17 TGF-β1 cells produced higher levels of IL-10 than Th17 IL-23 cells following activation. Moreover, the addition of TGF-β1, IL-21, and IL-27 (an improved cytokine cocktail to stimulate IL-10 production and Tr1 cell differentiation) during the reactivation increased the levels of IL-10 production by Th17 TGF-β1 cells.
Since it has been described that ATP inhibits naive T cell conversion to Tr1 cells via P2X7 receptor signaling [33], we tested whether CD39 and ATP hydrolysis is necessary for the conversion of Th17 cells to Tr1 cells. For this, both subsets of Th17 cells were tested in their ability to convert to IL-10-producing cells in the presence of Tr1 polarizing cytokines (TGF-β1/IL-21/ IL-27), ATP and the CD39 specific inhibitor, ARL67156. As shown in Fig 5A, the addition of ATP reduced the ability of Th17 TGF-β1 cells to produce IL-10 in the presence of ARL67156, demonstrating that CD39 enzymatic activity is important for IL-10 production. Interestingly, Th17 IL-23 cells generated from P2X7 receptor knockout mice, produced higher levels of IL-10 when compared to Th17 IL-23 cells generated from wild-type mice (Fig 5B). These results suggest that CD39 enzymatic activity is crucial for limiting P2X7 receptor signaling in Th17 cells, promoting the production of IL-10 and the conversion of Th17 cells into Tr1-like cells.

Discussion
Extracellular ATP (eATP) is a danger signal released by dying and damaged cells, and it functions as an immunostimulatory signal that promotes inflammation [26,27]. CD39 and CD73 are two cell-surface ectoenzymes that dephosphorylate eATP into its metabolites, ADP, AMP, and adenosine, in a tightly regulated process. CD39 catalyzes the conversion of eATP into AMP, whereas CD73, catalyzes the dephosphorylation of AMP into adenosine [37][38][39]. The coordinated action of these ectonucleotidases results in the generation of extracellular adenosine, a molecule known for its immunosuppressive properties [40].
It has been described that the addition of TGF-β and IL-6 during the in vitro differentiation of Th17 cells induces the expression of CD39 and CD73 ectonucleotidases by Th17 cells [34] and that tumor-infiltrating Th17 cells may express these ectonucleotidases. Although the expression of CD39 and CD73 ectonucleotidases has been related to their potential to generate adenosine and create an immunosuppressive microenvironment [34], they may also serve to deplete eATP. In this line of evidence, Falk and colleagues have reported that CD39 expression by Foxp3+ regulatory T cells (Tregs) is involved in eATP depletion and reduces the cytotoxic effects of this molecule [41]. Thus, CD39 expression may endow Tregs with the capacity to enter into inflamed sites and mediate immunosuppression by preventing P2X7 receptor-mediated cell death. Our results demonstrate that this mechanism may also be active on Th17 cells, where ATP-induced cell death is reduced in the population of Th17 cells (Th17 TGF-β1) expressing the CD39 ectonucleotidase. This result puts forward the idea that Th17 cells expressing CD39 may survive in ATP-rich sites such as an inflamed tissue.
Th17 cells have been defined as a plastic subset of T cells, being able to differentiate into several other T cell population during inflammation [42]. We and others have reported that Th17 cells can differentiate into IFN-γ-producing cells when transferred into lymphopenic hosts, during EAE or in a murine melanoma model [11,[13][14][15]. Moreover, Flavell and colleagues have shown that Th17 cells are also able to differentiate into IL-10-producing Tr1 cells during the resolution of inflammation and that this population of Tr1 cells present regulatory properties as they abolished Th17 cell-mediated colitis [20,21]. In this study, we confirmed that Th17 cells generated with TGF-β1 are a highly plastic subset of T helper cells, as they can differentiate into IFN-γ-and IL-10-producing cells following transfer into Rag -/mice.
Tr1 cells constitute an important subset of CD4+ T cells that help to control excessive inflammatory responses mainly through the production of IL-10 [43]. It has been reported that AHR and c-Maf transcription factors physically interact enabling the transactivation of the IL-10 promoter and thus are involved the differentiation of Tr1 cells [44,45]. Importantly, eATP and hypoxia have been shown to suppress the generation of Tr1 cells by triggering AHR inactivation through HIF1-α [46]. It has been demonstrated that eATP increases the interaction of HIF1-α to ARNT decreasing AHR binding to ARNT [33], resulting in the reduction of the transcription of AHR controlled genes. Interestingly, the group of Quintana reported that CD39 expression is important for the production of IL-10 by Tr1 cells as it allowed the depletion of eATP favoring AHR/ARNT interaction [33]. Based on this evidence, we tested whether CD39 may also play a role in promoting Th17 cell differentiation into IL-10-producing cells. In agreement with this hypothesis, Th17 cells expressing the CD39 ectonucleotidase (Th17 TGF-β1 ) induce an IL-10-rich microenvironment when transferred in a setting of intestinal inflammation. Furthermore, Th17 TGF-β1 cells produce higher levels of IL-10 compared to Th17 IL-23 cells when re-activated in vitro. Importantly, in the presence of eATP and the CD39 inhibitor ARL67156 Th17 TGF-β1 cells reduced their ability to produce IL-10.
Although we demonstrated adenosine production by Th17 TGF-β1 cells, we could not detect a strong suppressive activity of Th17 cells over effector T cells, which could be due to the strong activation stimulus we used to induce effector T cell activation. These results are different from those reported by the group of Ghriringelli who suggest that adenosine produced by Th17 cells have a suppressor activity [34]. Our data strongly argue in favor of the idea that CD39 may not only be involved in generating a suppressive microenvironment, but also may deplete eATP allowing these cells to survive into inflamed tissues. On the other hand, CD39 expression may be a determining factor in the differentiation of Th17 cells to Tr1-like cells.