Research ArticleImmunology

Glutamine-dependent α-ketoglutarate production regulates the balance between T helper 1 cell and regulatory T cell generation

See allHide authors and affiliations

Sci. Signal.  29 Sep 2015:
Vol. 8, Issue 396, pp. ra97
DOI: 10.1126/scisignal.aab2610

Feeding immune responses

When a naïve CD4+ T cell is activated by antigen, cytokines determine whether it differentiates into an effector T cell, such as a T helper 1 (TH1) cell, to mount an immune response, or a regulatory T (Treg) cell, which suppresses immune responses. Klysz et al. found that depletion of glutamine from culture medium enhanced the differentiation of naïve CD4+ T cells into Treg cells even in the presence of cytokines that promote TH1 cell generation. Adding a cell-permeable analog of the glutamine-derived metabolite α-ketoglutarate to the glutamine-deprived cells inhibited the generation of Treg cells. Together, these results suggest that glutamine deprivation, which occurs in tumor microenvironments, shifts the balance of the immune response to become more suppressive.

Abstract

T cell activation requires that the cell meet increased energetic and biosynthetic demands. We showed that exogenous nutrient availability regulated the differentiation of naïve CD4+ T cells into distinct subsets. Activation of naïve CD4+ T cells under conditions of glutamine deprivation resulted in their differentiation into Foxp3+ (forkhead box P3–positive) regulatory T (Treg) cells, which had suppressor function in vivo. Moreover, glutamine-deprived CD4+ T cells that were activated in the presence of cytokines that normally induce the generation of T helper 1 (TH1) cells instead differentiated into Foxp3+ Treg cells. We found that α-ketoglutarate (αKG), the glutamine-derived metabolite that enters into the mitochondrial citric acid cycle, acted as a metabolic regulator of CD4+ T cell differentiation. Activation of glutamine-deprived naïve CD4+ T cells in the presence of a cell-permeable αKG analog increased the expression of the gene encoding the TH1 cell–associated transcription factor Tbet and resulted in their differentiation into TH1 cells, concomitant with stimulation of mammalian target of rapamycin complex 1 (mTORC1) signaling. Together, these data suggest that a decrease in the intracellular amount of αKG, caused by the limited availability of extracellular glutamine, shifts the balance between the generation of TH1 and Treg cells toward that of a Treg phenotype.

INTRODUCTION

When T cells interact with their cognate foreign antigens, they undergo rapid activation, which requires considerable energy and cellular resources. These metabolic needs are secured by the augmented uptake and use of nutrients. Indeed, a sine qua non for optimal T cell proliferation and effector function is the T cell receptor (TCR)–stimulated increase in the cell surface amounts of glucose and glutamine transporters. The resulting increase in glucose and glutamine metabolism (19) results in a metabolic shift from the fatty acid oxidation that characterizes quiescent T cells (1012).

Studies have shown that distinct T lymphocyte subsets exhibit disparate metabolic profiles; effector T (Teff) cells are highly glycolytic and even lipogenic, whereas suppressive regulatory T (Treg) cells display a mixed metabolism with increased lipid oxidation (1316). Cellular metabolism is regulated, at least in part, by the mammalian target of rapamycin (mTOR) pathway. The mTOR complex 1 (mTORC1) is critical for the differentiation of naïve T cells into T helper 1 (TH1) and TH17 cells, as well as for the cytolytic activity of CD8+ memory T cells, whereas mTORC2 signaling promotes the differentiation of naïve CD4+ T cells into TH2 cells (1721). Conversely, inhibition of mTOR activity blocks the generation of Teff cells, instead promoting the generation and function of Foxp3+ (forkhead box P3–positive) Treg cells (2224). Consistent with these observations, the activation of mTOR blocks the differentiation of naïve CD4+ T cells into Treg cells and blocks Treg function (2527). Furthermore, decreasing mTOR activity by genetic deletion of nutrient transporters that are responsible for the uptake of glucose, leucine, or glutamine (8, 9, 28) inhibits the generation of Teff cells without affecting Treg cell generation.

Because pathological microenvironments can alter the nutrients available to a T cell, it is important to determine whether the external nutrient environment regulates the intrinsic differential potential of that cell. Indeed, nutrient concentrations within tumor microenvironments are generally reduced compared to those in normal tissues. Specifically, quantitative metabolomics profiling has revealed that the intratumoral concentrations of glucose and glutamine are reduced in patients with hepatocellular carcinomas and stomach and colon tumors (29, 30). Furthermore, alkylating chemotherapies decrease the intracellular generation of the antioxidant glutathione because of limited glutamine availability (3134).

Here, we demonstrated that nutrient availability plays a major role in regulating the differentiation of naïve CD4+ T cells into different subsets. Stimulation of the TCR under conditions in which the amount of glutamine was limited resulted in the conversion of naïve CD4+ T cells into Foxp3+ T cells. This phenomenon was specifically a result of glutamine catabolism because it was recapitulated by a glutaminase inhibitor. The converted Foxp3+ T cells exhibited increased proliferation in vivo as compared to conventional Foxp3 T cells, and they protected recombinase activating gene (Rag)–deficient mice from the development of Teff cell–mediated autoimmune colitis. This de novo generation of Treg cells occurred even under TH1-polarizing conditions, abrogating the differentiation of naïve cells into TH1 cells. This block in differentiation was associated with attenuated mTOR signaling and decreased expression of the gene encoding glutaminase 2 (GLS2), the enzyme that catalyzes the first step in the generation of α-ketoglutarate (αKG) from glutamine. Because αKG is incorporated into the tricarboxylic acid (TCA) cycle, the major anaplerotic step in proliferating cells (35, 36), we assessed whether the glutamine-derived production of αKG was required for the generation of TH1 cells. We found that supplementing glutamine-deprived CD4+ T cells with a cell-permeable αKG ester under TH1-polarizing conditions resulted in a marked increase in the abundance of Tbet, a transcription factor required for TH1 cell differentiation, as well as interferon-γ (IFN-γ), a signature cytokine of TH1 cells. This was associated with the activation of mTORC1, as demonstrated by increased phosphorylation of the ribosomal protein S6. Together, these data suggest that extracellular glutamine availability governs the concentration of intracellular αKG, which in turn functions as a metabolic regulator that determines whether naïve T cells differentiate into TH1-type Teff or Treg cells.

RESULTS

Teff cell function is differentially modulated by the deprivation of glucose or glutamine

To assess whether the relative availability of nutrients in the external environment regulated the differentiation potential of naïve CD4+ T cells, we stimulated lymphocytes with activating antibodies (anti-CD3/CD28) specific for the TCR complex component CD3 and the co-receptor CD28 in either complete medium or medium deficient in glucose or glutamine. Whereas the early TCR-stimulated increases in the production of the cytokines interleukin-2 (IL-2) and IL-17 were not inhibited, glutamine withdrawal resulted in an increase in the percentage of Foxp3+ cells (Fig. 1A). Furthermore, we also measured the increased abundance of foxp3 mRNA in cells cultured under glutamine-deprived, but not glucose-deprived, conditions (Fig. 1B). Thus, this increased gene expression did not represent a response to nutrient deprivation per se but rather was a specific consequence of glutamine starvation. To determine whether the increased Foxp3 expression in cells cultured under glutamine-deprived conditions was specifically a result of changes in glutamine catabolism, we blocked glutaminolysis with 6-diazo-5-oxo-l-norleucine (DON), a glutaminase inhibitor that functions only when the enzyme is active (37). There was an approximately 45-fold increase in foxp3 mRNA abundance in DON-treated cells compared to that in cells cultured under normal nutrient conditions (Fig. 1B). Thus, either limiting the available amount of extracellular glutamine or inhibiting intracellular glutaminolysis resulted in increased foxp3 expression in stimulated naïve CD4+ T cells.

Fig. 1 Glutamine deprivation promotes the conversion of naïve CD4+ T cells into Foxp3+ T cells.

(A) Naïve mouse CD4+ T cells were activated by immobilized anti-CD3 and anti-CD28 monoclonal antibodies (anti-CD3/CD28 activation) in nutrient-replete medium (Nutr+), glucose-free medium (Nutr–GLC), or glutamine-free medium (Nutr–GLN). After 96 hours, the percentages of CD4+ cells that were positive for IL-2, IFN-γ, IL-17A, or Foxp3 were determined by intracellular staining and flow cytometric analysis. Percentages are indicated in the dot plots, which are representative of four experiments. (B) Naïve CD4+ T cells cultured under the indicated conditions and activated for 96 hours with anti-CD3/CD28 were analyzed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) to determine the relative amounts of Foxp3 mRNA normalized to that of Hprt mRNA. Data are means ± SD of triplicate samples from one experiment, which is representative of four independent experiments. *P < 0.05 by analysis of variance (ANOVA). (C) Left: Freshly isolated CD4+ lymph node T cells loaded with Cell Trace Violet (CTV) and cultured under the indicated conditions were activated for 24 or 96 hours and then were analyzed by flow cytometry to determine the percentages of Foxp3+ cells. Right: At 96 hours, the proliferation profiles of Foxp3CD4+ T cells (dashed line) and Foxp3+CD4+ T cells (solid line) were monitored by flow cytometric analysis of the dilution of CTV. Dot plots and histograms are representative of three experiments. (D) Naïve conventional CD4+CD62L+CD44CD25GFP T cells were sorted by FACS (fluorescence-activated cell sorting) from Foxp3-GFP reporter mice, and their purity was monitored by flow cytometry (left panel; day 0). The cells were then activated under the indicated conditions, and the appearance of Foxp3+ cells was determined by flow cytometric analysis after 96 hours. The percentages of Foxp3+ cells are indicated in the dot plots. Data are representative of four independent experiments. SSC, side scatter.

The increased percentage of Foxp3+ T cells under glutamine-deprived conditions was detected 96 hours after TCR stimulation (Fig. 1C) and was not a result of their preferential proliferation as compared to that of CD4+Foxp3 T cells. Whereas CD4+ T cell proliferation was reduced in the absence of glutamine or in the presence of DON, the relative proliferation profiles of the Foxp3 and Foxp3+ cells were similar under all conditions (Fig. 1C, right panels). Moreover, although glucose deprivation also reduced the TCR-stimulated proliferation of CD4+ T cells, there was no accompanying increase in the percentage of Foxp3+ cells (Fig. 1C). Thus, the biased TCR-stimulated generation of Foxp3+ T cells was specific to conditions of glutamine deprivation. Furthermore, this biased generation of Foxp3+ T cells was reversed by the addition of low amounts of glutamine to the cell culture medium (fig. S1).

These data gave rise to two nonexclusive hypotheses. First, the survival of Foxp3+ cells relative to that of Foxp3 cells is modulated when glutamine abundance is limiting. Second, Foxp3CD4+ T cells are converted to Foxp3+ cells under conditions of glutamine deprivation. To test the second hypothesis, we purified CD4+CD62L+CD44CD25GFP conventional naïve T cells from reporter mice expressing green fluorescent protein (GFP) under the control of the foxp3 promoter (Foxp3-GFP) (38, 39) and found that deprivation of glutamine, but not glucose, led to an increase in the percentage of Foxp3+ cells that were generated in response to T cell stimulation (Fig. 1D, P < 0.0001). These data suggest that glutamine deprivation in the context of T cell stimulation favors the conversion of Foxp3 cells to Foxp3+ cells.

The conversion of naïve CD4+ T cells into Foxp3+ CD4+ T cells under glutamine-deprived conditions results from TGF-β–dependent signaling

Given the central role of transforming growth factor–β (TGF-β) in the conversion of naïve T cells to Foxp3+ Treg cells, we evaluated its potential involvement in the enhanced generation of these cells under glutamine-deprived conditions. Although we had not added exogenous TGF-β to the cell culture medium, we evaluated the possibility that either the low amount of TGF-β present in the fetal calf serum (FCS) or any TGF-β produced by the cells themselves was sufficient to drive the differentiation of naïve CD4+ T cells into Foxp3+ T cells under glutamine-deprived conditions. Indeed, a neutralizing anti–TGF-β antibody reduced the percentage of Foxp3+ T cells generated under glutamine-deprived conditions in a proliferation-independent manner (Fig. 2A). Furthermore, SB431542, an inhibitor of the TGF-β receptor 1 (TGFβR1) signaling pathway (40), statistically significantly decreased the generation of Foxp3+ cells under conditions in which glutamine abundance was limiting (Fig. 2B, P < 0.0001). Finally, we evaluated the Treg cell–specific demethylated region (TSDR), an enhancer element of the Foxp3 locus that is selectively demethylated in stable Foxp3+ Treg cells (4143), and found that the methylation status of the TSDR in TGF-β–induced Foxp3+ Treg (iTreg) cells was similar to that in Foxp3+ cells generated in the absence of glutamine (fig. S2).

Fig. 2 Glutamine deprivation promotes the TGF-β–dependent conversion of naïve CD4+ T cells into Foxp3+ Treg cells.

(A) Naïve CD4+ T cells loaded with CTV were activated for 4 days under the indicated nutrient conditions in the absence or presence of a neutralizing anti–TGF-β antibody (αTGF-β). Left: The percentages of Foxp3+ cells on day 4 of activation are indicated in the dot plots. Right: The relative proliferation profiles of Foxp3CD4+ and Foxp3+CD4+ T cells under each condition were monitored on day 4 of activation as a function of CTV fluorescence and are presented in histograms. Bottom: Quantification of the percentages of Foxp3+ T cells generated under glutamine-deficient conditions, in the presence or absence of TGF-β. Data are means ± SD of three independent experiments. ****P < 0.0001 by χ2 test. (B) Naïve CD4+ T cells were activated for 5 days under the indicated nutrient conditions in the presence or absence of the TGF-β receptor signaling inhibitor SB431542. The percentages of Foxp3+ T cells on day 5 of activation were assessed by flow cytometric analysis. Data are means ± SD of two experiments.

Foxp3+ T cells generated in the absence of glutamine show enhanced proliferation in vivo

Our earlier experiments demonstrated that glutamine depletion promotes the in vitro differentiation of naïve CD4+ T cells into Foxp3+ T cells. It was nonetheless not clear whether these Foxp3+ T cells would be maintained after their transfer into lymphopenic hosts and how their proliferation in vivo would compare with that of the Foxp3 T cells derived from the same cultures. To evaluate this question, we activated naïve CD4+ T cells from C57BL/6-Thy1.1 mice for 5 days in nutrient-replete or glutamine-deprived conditions or in the presence of DON (Fig. 3A). Under the glutamine-altered conditions, the percentages of Foxp3+ T cells increased between 15 to 35% (Fig. 3A). The cells were then labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) and adoptively transferred into sublethally irradiated C57BL/6-Thy1.2 recipient mice. The ratio of Foxp3+ to Foxp3 donor T cells from both the glutamine-depleted cultures and the DON-treated cultures remained substantially greater than that of cells derived from the nutrient-replete control culture on day 7 after adoptive transfer (means of 20 to 40% as compared to 10%; P < 0.005) and was associated with an increased in vivo proliferation of the Foxp3+ cells (Fig. 3B). Thus, the Foxp3+ T cells generated ex vivo by culturing under glutamine-deprived conditions were maintained after their adoptive transfer to recipient mice.

Fig. 3 Foxp3+ T cells generated in the absence of glutamine exhibit enhanced proliferation in vivo.

(A) Lymph node (LN) Thy1.1+CD4+ T cells were activated for 5 days under the indicated nutrient conditions before being analyzed by flow cytometry to determine the relative percentages of Foxp3+ cells, which are shown in the representative dot plots (right). The cells were then labeled with CTV and transferred into congenic Thy1.2+ C57BL/6 mice that were previously rendered lymphopenic by irradiation (5.5 Gy). The mice were sacrificed 7 days later, and the status of the adoptively transferred T cells was assessed by flow cytometry. FSC, forward scatter. (B) The engraftment of the transferred T cells was monitored as a function of Thy1.1 expression (dot blots, top left), and the presence of Foxp3 within the indicated transferred CD4+ T cell populations was also assessed (right panel). The percentages of Foxp3+ cells are indicated in each dot plot. Right: Quantification of the percentages of Foxp3+ donor T cells. Each point represents data from a single mouse. Means are indicated by horizontal lines. *P < 0.05, **P < 0.01, ***P < 0.001 by Student’s t test and ***P < 0.001 by ANOVA. The proliferation of Foxp3CD4+ T cells relative to that of Foxp3+CD4+ T cells was monitored in individual mice as a function of CTV fluorescence. Representative histograms for the indicated conditions are shown (bottom panels). Data are representative of two independent experiments and each experiment was performed with three or four mice per group.

Foxp3+ T cells generated under glutamine-deficient conditions show suppressive activity in vivo

To evaluate the in vivo suppressive potential of Foxp3+ T cells generated under glutamine-deprived conditions in vitro, we assessed their ability to prevent inflammatory bowel disease (IBD) in Rag2−/− mice injected with naïve CD4+ Teff cells. Briefly, naïve CD4+ T cells isolated from Foxp3-GFP reporter mice were stimulated under glutamine-deprived conditions for 5 days before being sorted on the basis of GFP abundance (fig. S3). Consistent with previous reports (4448), injection of Rag2−/− mice with Treg cells generated in the thymus of wild-type mice [which are referred to as thymic Treg or natural Treg (nTreg) cells] protected the Rag2−/− mice from IBD induced by CD4+CD45RBhi Teff cells during the entire 9-week follow-up period (Fig. 4, A and B). Injection with Foxp3+ cells generated in vitro under glutamine-deprived conditions also fully protected Rag2−/− mice co-injected with Teff cells from IBD during the entire follow-up period, an effect that was associated with reduced cell surface abundance of the effector molecule CD44 on the injected Teff cells (Fig. 4C). Moreover, the Foxp3+ cells generated in vitro under glutamine-deprived conditions exhibited enhanced persistence compared to that of nTreg cells (Fig. 4D). These data suggest that glutamine withdrawal promotes the suppressive activity and in vivo persistence of the generated Treg cells.

Fig. 4 Treg cells generated under glutamine-deficient conditions inhibit effector T cell–mediated autoimmune colitis.

(A to D) Rag2-deficient mice were injected with sorted CD4+CD45RBhi (CD45.1+) Teff cells alone or together with purified nTreg cells or glutamine-deprived in vitro–generated Foxp3+CD4+ T cells (TregNutr-GLN). (A) Top: The injected Rag2-deficient mice were examined for weight loss on a weekly basis, and the mean weights in each group are presented as a function of the initial weight. Data are means ± SD of five mice per group. Bottom: The survival rate in each group is presented at the indicated times as means ± SD. The log-rank (Mantel-Cox) test and test for trend indicate P values of 0.0089 and 0.0087, respectively. (B) Colons from mice injected with the indicated T cells were collected, fixed, and stained with hematoxylin and eosin (H&E). Representative sections from each group of mice are shown. The magnifications in the upper and lower panels are ×5 and ×10, respectively. (C) CD45.1+CD4+ Teff cells isolated from Rag2-deficient mice injected with the indicated T cells were analyzed by flow cytometry to determine the cell surface abundance of the activation marker CD44. Representative histograms are shown. The vertical dotted line delineates the mean fluorescence intensity (MFI) of CD44 staining on CD45.1+CD4+ T cells isolated from Rag2-deficient mice injected with Teff cells alone. Right: Mean MFIs of CD44 staining in each group. Data are means ± SD of five mice in each group. (D) Left: The relative recovery of injected Teff cells (CD45.1+) and Treg cells (CD45.2+) in the different groups of mice was assessed 9 weeks after adoptive transfer. Representative dot plots discriminating between CD45.1+ Teff cells and CD45.2+ Treg cells in the mesenteric lymph node (mLN) of the mice are presented for each condition. Right: The proportion of CD45.2+ Treg cells recovered from the mLN of individual injected Rag2-deficient mice from each group is quantified, and each point represents data from a single mouse. Data are means from five mice per group. Means are indicated by horizontal lines. ***P < 0.001 by Student’s t test.

Glutamine deprivation inhibits the differentiation of naïve CD4+ T cells into TH1 cells and increases the generation of Foxp3+ cells

Next, we investigated whether glutamine deprivation affected the generation of Foxp3-expressing cells after the activation of naïve T cells under cell-polarizing conditions. As expected, most naïve CD4+ T cells cultured under iTreg-polarizing conditions became Foxp3+ cells, and this percentage was not modulated by glutamine availability (Fig. 5A). These data suggest that the potential of a naïve CD4+ T cell to differentiate into an iTreg cell occurs in a glutamine-independent manner.

Fig. 5 Glutamine deprivation inhibits the generation of TH1 cells but promotes the generation of Treg cells.

(A) Naïve CD4+CD62L+CD44CD25 T cells were activated under iTreg-inducing conditions in the presence or absence of glutamine. After 6 days of activation, the percentages of Foxp3+ cells were determined by intracellular staining and flow cytometric analysis. Dot plots are representative of three experiments. (B) Naïve CD4+CD62L+CD44CD25 T cells cultured under the indicated polarizing and nutrient conditions were analyzed after 30 hours by qRT-PCR to determine the relative amounts of Tbet (top) and Gata3 (bottom) mRNAs normalized to that of Hprt mRNA. Data are means ± SD of triplicate samples from one experiment and are representative of two independent experiments. *P < 0.05 by Student’s t test. (C) Naïve CD4+CD62L+CD44CD25 T cells were activated for 6 days under TH0-, TH1-, or TH2-polarizing conditions in complete or glutamine-deficient medium, as indicated. Left: The percentages of CD4+ T cells that were positive for IFN-γ were determined by intracellular staining and flow cytometric analysis. Dot plots are representative of eight experiments. Middle: Mean percentages ± SD of IFN-γ–secreting cells (n = 8 experiments). Statistical difference was determined by χ2 (****P < 0.0001) and Mann-Whitney tests (P < 0.001). Right: IL-4 production by the indicated cells was assessed by cytometric bead array. Data are means from one experiment and are representative of two independent experiments. (D) Naïve CD4+CD62L+CD44CD25 T cells were activated under TH0-, TH1-, or TH2-polarizing conditions in nutrient-replete or glutamine-depleted medium. Left: After 6 days of activation, the percentages of Foxp3+ cells were determined by flow cytometric analysis. Dot plots show the percentages of Foxp3+ cells. Middle: Quantification of the percentages of Foxp3+ T cells. Data are presented as means ± SD of eight experiments. P values were determined by χ2 (****P < 0.0001) and Mann-Whitney tests (P < 0.05). Right: The indicated cells were analyzed by qRT-PCR to determine the relative amounts of Foxp3 mRNA normalized to that of Hprt mRNA. Data are means ± SD of triplicate samples from a single experiment and are representative of two independent experiments.

To determine whether glutamine availability conditions the potential of a cell to differentiate into a specific Teff cell type, we stimulated naïve CD4+ T cells under TH1- or TH2-polarizing conditions and monitored their differentiation as a function of the abundances of the mRNAs for the transcription factors Tbet and GATA3, respectively. Glutamine deprivation blocked the expression of Tbet under TH1-polarizing conditions but had no effect on the expression of GATA3 under TH2-polarizing conditions (Fig. 5B). Furthermore, this change in Tbet expression in glutamine-deprived cells was associated with an almost complete absence of IFN-γ secretion by CD4+ T cells exposed to TH1-polarizing cytokines (Fig. 5C, P < 0.0001). In marked contrast, glutamine deprivation resulted in enhanced IL-4 production by cells cultured under TH2-polarizing conditions (Fig. 5C). Note that the differentiation of naïve cells into TH1 cells was not inhibited by glucose starvation (fig. S4), suggesting that there are substantial differences in the capacity of energetic fuel sources to support Teff cell differentiation.

Given that glutamine-deprived naïve CD4+ T cells were unable to undergo differentiation into TH1 cells under TH1-polarizing conditions, we next monitored Foxp3 expression in these cells. We found that the percentage of Foxp3+ cells was markedly increased under TH1-polarizing, but not TH2-polarizing, conditions in the context of glutamine deprivation (Fig. 5D, P < 0.0001). Furthermore, Foxp3 mRNA abundance was increased fivefold in glutamine-deprived cells compared to that in cells cultured with complete nutrients (Fig. 5D). These data suggest that the nutrient milieu affects the expression of Foxp3, even in cells cultured in the presence of cytokines that would be expected to bias a naïve CD4+ T cell to differentiate into a specific Teff cell type.

On the basis of these data, it was of interest to determine whether endogenous TGF-β was also responsible for the emergence of Foxp3+ T cells under TH1-polarizing conditions. Indeed, anti–TGF-β antibodies inhibited the conversion of naïve CD4+ T cells to Foxp3+ cells under TH1-polarizing conditions (Fig. 6A). However, even though the generation of Foxp3+ cells was dependent on TGF-β, neutralizing this cytokine was not sufficient to promote TH1 cell polarization under glutamine-deprived conditions. Specifically, Tbet abundance was not restored in cells treated with the anti–TGF-β antibody and cultured under TH1-polarizing conditions (Fig. 6B), and cells cultured under these conditions did not produce sufficient amounts of IFN-γ (Fig. 6C). In contrast, the differentiation of naïve CD4+ T cells into TH2 cells was not sensitive to the inhibition of TGF-β (Fig. 6B). Together, these results indicate that, under limiting glutamine conditions, the conversion of Foxp3 cells into Foxp3+ cells under TH0- or TH1-polarizing conditions was dependent on TGF-β. Nevertheless, inhibiting TGF-β signaling was not sufficient to enable TH1 cell generation under these conditions of glutamine deprivation.

Fig. 6 Inhibition of TGF-β signaling abrogates the conversion of glutamine-deprived CD4+ T cells to Treg cells but cannot support TH1 cell generation.

(A to C) Naïve CD4+ T cells were stimulated under TH0-, TH1-, or TH2-polarizing conditions in nutrient-replete or glutamine-deprived medium in the presence or absence of an anti–TGF-β neutralizing antibody. (A) After 96 hours of activation, the CD4+ T cells were analyzed by flow cytometry to determine the percentages of Foxp3+ cells. Left: Dot plots show the percentages of Foxp3+ cells. Right: Quantification of the percentages of Foxp3+ T cells. Data are means ± SD of three independent experiments. ****P < 0.0001 by χ2 test. (B) After 96 hours of stimulation, the CD4+ T cells were analyzed by flow cytometry to determine the relative abundances of the transcription factors Tbet (top) and GATA3 (bottom). Histograms are representative of four independent experiments. (C) On day 6 of activation, the CD4+ T cells were analyzed by flow cytometry to determine the percentages of IFN-γ–producing cells. Dot plots show the percentages of IFN-γ+ cells and are representative of three independent experiments.

Generation of the glutamine-derived metabolite αKG is the rate-limiting step in the differentiation of naïve cells into TH1 cells

To determine how glutamine deprivation affected the intracellular metabolism of the cell, we first assessed glutamine uptake. Whereas T cell stimulation increased glutamine uptake by >25-fold compared to that of unstimulated cells, T cell stimulation of glutamine-deprived or DON-treated cells statistically significantly decreased their glutamine uptake (Fig. 7A, P < 0.05). Furthermore, glutamine deprivation attenuated the T cell–stimulated increase in the abundance of the mRNA encoding GLS2 (Fig. 7B), a key enzyme in the conversion of glutamine to glutamate. Because the catabolism of glutamine contributes to the metabolism of T cells, these data suggest that the metabolic phenotypes of CD4+ T cells activated under nutrient-replete or glutamine-deprived conditions are distinct.

Fig. 7 The cell-permeable αKG analog DMK rescues the generation of TH1 cells from naïve CD4+ T cells under glutamine-deprived conditions.

(A) Naïve CD4+ T cells (5 × 105) were cultured for 3 days in the absence of any activating agents (quiescent) or were activated under the indicated nutrient conditions. Glutamine uptake under each condition was then assessed by incubating cells with l-[3,4-3H (N)]glutamine (0.5 μCi) for 5 min at room temperature, and incorporation per cell was measured as counts per minute (CPM). Data are means ± SD of triplicate samples from one experiment and are representative of three experiments. *P < 0.05 by ANOVA. (B) Cells cultured under the indicated conditions were subjected to qRT-PCR analysis after 2 days of activation to determine the relative abundance of Gls2 mRNA normalized to that of Hprt mRNA. Data are means ± SD of triplicate samples from one experiment and are representative of two independent experiments. *P < 0.05 by ANOVA. (C) Cells cultured under the indicated conditions were subjected to the measurement of cellular respiration and glycolysis as a function of the OCR and ECAR, respectively, with the Seahorse XF24 analyzer. Basal respiration, SRC, and ECAR are presented as means (pmol/min per 1 × 106 cells or mpH/min) ± SD of triplicate samples from three independent experiments. ***P < 0.001 by Student’s t test. ATP amounts were assessed as a function of oligomycin-sensitive oxygen consumption. **P < 0.01 by Student’s t test. (D) Schematic representation of the metabolism of glutamine to αKG and its introduction into the TCA cycle. GS, glutamine synthetase; GLS, glutaminase; GDH, glutamine dehydrogenase. (E) Naïve CD4+ T cells were activated under TH1-polarizing conditions in nutrient-replete or glutamine-depleted conditions in the absence or presence of DMK. After 6 days of stimulation, the cells were analyzed by flow cytometry to determine the percentages of Tbet+, IFN-γ+, and Foxp3+ cells. Plots are representative of data from five independent experiments. The relative abundance of pS6 ribosomal protein was analyzed by flow cytometry on day 1 after activation. The difference in the MFI of pS6 staining (black lines) relative to control staining (shaded) is indicated in each histogram. Data are representative of five independent experiments.

In cells cultured under nutrient-replete conditions, TCR stimulation resulted in a >10-fold increase in the basal oxygen consumption rate (OCR), an indicator of oxidative phosphorylation (OXPHOS); however, the OCR was statistically significantly decreased in glutamine-deprived cells (Fig. 7C, P < 0.001). These data are concordant with a study that showed that Treg cells have a lower level of mitochondrial OCR than Teff cells (49). Furthermore, Treg cells exhibit reduced glycolysis compared to Teff cells, as determined by measurement of the extracellular acidification rate (ECAR) (49). Indeed, we found that stimulated glutamine-deprived T cells showed a 10-fold decrease in basal ECAR compared to stimulated control T cells (Fig. 7C, P < 0.001). On the other hand, Treg cells have a high spare respiratory capacity (SRC) (49), which potentially enables them to respond to starvation conditions in which glucose is the only fuel, and this was indeed the case for glutamine-deprived CD4+ T cells (Fig. 7C). As expected from the decreased OXPHOS and ECAR in glutamine-deprived CD4+ T cells, they had a statistically significantly decreased adenosine triphosphate (ATP) concentration compared to that of cells cultured with complete nutrients (Fig. 7C, P < 0.01), revealing an important role for glutaminolysis in the energy homeostasis of an activated CD4+ T cell.

Glutamine is catabolized to generate αKG, which supports energy production through TCA cycle anaplerosis (Fig. 7D). We therefore assessed whether glutamine-derived αKG was critical for the commitment of activated CD4+ T lymphocytes to become TH1 cells. To this end, glutamine-deprived naïve CD4+ T cells activated under TH1-polarizing conditions were supplemented with a cell-permeable αKG derivative, dimethyl αKG (DMK). When supplemented with DMK, glutamine-deprived cells showed an increased abundance of Tbet, the transcription factor required for the generation of TH1 cells, which was comparable to that in control TH1-polarized cells (Fig. 7E). Furthermore, DMK-supplemented cultures exhibited a substantial increase in the percentage of IFN-γ+ cells (Fig. 7E). Supplementing glutamine-deprived CD4+ T cells with DMK also decreased the generation of Foxp3+ cells (Fig. 7E). Thus, the reduced abundance of αKG correlates with the increased production of Foxp3.

The differentiation of naïve CD4+ T cells into TH1 cells depends on the activation of the mTORC1 pathway (17, 22, 50), and glutamine transport and the subsequent catabolism of glutamine are required to stimulate mTOR signaling pathways (9, 5154). Furthermore, the extent of mTOR signaling inversely correlates with the generation of Treg cells (2227). Indeed, we found that the phosphorylation of ribosomal protein S6, a readout of mTORC1 activity, was reduced by more than fourfold in CD4+ T cells undergoing TH1 polarization in glutamine-deprived as compared to nutrient-replete conditions (Fig. 7E). Because αKG directly activates mTORC1 signaling in multiple primary and transformed cell lines (51), we next assessed whether DMK affected mTORC1 signaling in glutamine-deprived cells. We found that phosphorylated S6 (pS6) was increased two- to threefold in abundance in DMK-treated cells compared to that in glutamine-deprived cells (Fig. 7E, P < 0.05). Together, these data suggest that increasing the intracellular concentration of αKG enhances mTORC1 signaling, which supports the differentiation of naïve cells into TH1 cells.

DISCUSSION

Our data suggest that glutamine availability is a key determinant of the differentiation of naïve CD4+ T cells. Low glutamine availability promoted the conversion of naïve CD4+ T cells into Foxp3+ Treg cells, which displayed highly effective regulatory function. The marked increase in the generation of Foxp3+ T cells under these conditions was dependent on signaling by endogenous TGF-β. The Foxp3+ T cells that were generated under glutamine-deprived conditions exhibited robust in vivo proliferative potential as well as suppressor activity, controlling autoimmune colitis in a mouse model of T cell adoptive transfer. Moreover, the skewing of glutamine-deprived naïve CD4+ T cells toward Foxp3+ cells occurred even under TH1-polarizing conditions, blocking their differentiation into TH1 cells. Our data identify a metabolic program that conditions TH1 lineage specification. We found that the differentiation of naïve cells into TH1 cells required that glutamine be catabolized to αKG, replenishing the pools of metabolic intermediates in the TCA cycle and substantially increasing mTORC1 signaling. Under conditions in which glutamine use was limiting, the addition of an αKG ester restored the capacity of a cell to generate Tbet and adopt a TH1 cell phenotype.

These data raise the question of why glutamine-derived αKG is required for the differentiation of naïve cells into TH1-type Teff cells, but not into anti-inflammatory Treg cells. The entry of metabolic intermediates into the TCA cycle is central to energy metabolism, whereas their exit fosters the synthesis of biological molecules that are required for cell growth and division. In cancer cells, the heightened need for biosynthetic intermediates results in a disproportionate dependency on glutamine, which undergoes anaplerotic reactions to form αKG (36, 55, 56). By analogy, this would suggest that TH1 cells, but not Treg cells, present a metabolic state that requires the support of a high level of anabolism. Indeed, Teff cells show high rates of glycolysis, whereas suppressive Treg cells exhibit a higher dependence on fatty acid oxidation (13, 14). Furthermore, although Treg cells can take up exogenous fatty acids, TH17 cells depend on the de novo fatty acid synthesis, which is costly in terms of ATP (16). Together, these data suggest that Teff cells have higher metabolic requirements than Treg cells. In this regard, it is interesting to note that the addition of αKG to cultures of glutamine-deprived naïve cells not only stimulated their differentiation into TH1 cells but also inhibited their differentiation into Treg cells. Thus, our findings suggest that altering the concentrations of intracellular metabolic intermediates conditions the balance between TH1 cell effector functions and Treg cell suppressor functions.

However, it is also important to note that TCA cycle intermediates can regulate the epigenetic state of an activated T cell. Specifically, Tet2 (ten-eleven translocation 2) is an αKG-dependent enzyme that alters DNA methylation by conversion of 5-methylcytosine to 5-hydroxymethylcytosine. Deficiency in Tet2 inhibits the differentiation of naïve CD4+ T cells into TH1 cells, but not TH2 cells, whereas their conversion into Treg cells is enhanced (57). Because these changes in differentiation directly parallel those that we observed under conditions of glutamine deficiency, the metabolism of this amino acid may be critical for Tet2-mediated demethylation and regulation of an epigenetic state that is required for the generation of TH1 cells, but not TH2 cells.

iTreg cells can be generated de novo by the stimulation of naïve CD4+ T cells through the TCR in the presence of exogenous TGF-β. We found that this conversion process was independent of glutamine abundance; large numbers of Treg cells were generated irrespective of the glutamine concentration. However, it was only under glutamine-deficient conditions that endogenous TGF-β, or the low abundance of TGF-β in FCS, was sufficient to promote the differentiation of naïve cells into Treg cells. Furthermore, the TSDR methylation status detected in TGF-β–induced Treg cells was similar to that of Foxp3+ T cells generated under glutamine-deficient conditions. These glutamine-deprived Foxp3+ T cells were also highly proliferative in vivo and were recovered at higher percentages than nTreg cells when cotransferred with Teff cells into Rag2-deficient mice. This conversion, which we detected under either TH0- or TH1-type stimulations, was abrogated under conditions in which TGF-β signaling was inhibited. These data are consistent with previous elegant work that demonstrated that the low abundance of endogenous TGF-β is sufficient to generate in vitro iTreg cells in an mTOR-deficient background (22). Indeed, decreased glutaminolysis resulted in decreased mTOR signaling in CD4+ T cells stimulated ex vivo, whereas supplementation of DMK activated mTORC1 and inhibited Foxp3 mRNA and protein expression.

Although it is not yet known how the intricate coordination of nutrient processing regulates the formation of specific metabolic intermediates in T cell subsets, our data point to the extracellular nutrient environment as a key factor in this equation. This may have substantial physiological consequences in microenvironments, such as those that occur because of tumors or infections, wherein the nutrient milieu into which a T cell enters can be altered (29, 30, 58). Indeed, glutamine abundance is reduced in patients with hepatocellular, colon, and stomach tumors (29, 30). This finding may explain the data that show that Treg cells preferentially accumulate in and around murine tumors, especially as the tumors progress (59). Furthermore, Foxp3+ T cells are often recruited to a tumor before Teff cells are recruited, which prevents the T cell–mediated eradication of the tumor cells (60). Thus, the glutamine-deficient microenvironment of a tumor, potentially caused by the “addiction” of tumor cells to glutamine (61, 62) or as an undesirable result of chemotherapy (3234), appears to be a critical factor in enforcing a Treg cell phenotype even under conditions in which the CD4+ T cells recruited into the tumor are exposed to conditions that promote their differentiation into effectors. Manipulating the metabolic state of T cells may modulate the balance between effector and suppressor functions, potentially opening new avenues for the development of immunotherapies. From an evolutionary perspective, it is tempting to speculate that it is in the interest of the organism to evade an energetically costly immune response under conditions of amino acid starvation.

MATERIALS AND METHODS

Mice

C57BL/6 mice and C57BL/6-Thy1.1 mice were purchased from Charles River laboratories, whereas Foxp3-GFP reporter mice have been previously described (38, 39). The mice were housed in conventional, pathogen-free facilities at the Institut de Génétique Moléculaire de Montpellier and at the National Cancer Institute at Frederick. Animal care and experiments were performed in accordance with the National Institutes of Health (NIH) and French national guidelines.

Cell isolation and activation

CD4+ T cells were purified with the MACS CD4+ T cell isolation kit (Miltenyi Biotec). For experiments with CD4+ T cells depleted of Foxp3+ cells, enriched CD4+ T cells from Foxp3-GFP reporter mice were sorted on the basis of a CD4+CD62LhiCD44GFP expression profile on a FACSAria flow cytometer (BD Biosciences). For TH1-, TH2-, and iTreg-polarizing conditions, IL-12 (10 ng/ml) and anti–IL-4 antibody (5 μg/ml), IL-4 (10 ng/ml) and anti–IFN-γ antibody (10 μg/ml), or human TGF-β (3 ng/ml), respectively, were added to the cultures. TGF-β or the TGF-β signaling pathway were neutralized by the addition of an anti–TGF-β monoclonal antibody (clone 1D11, 10 μg/ml) or the inhibitor SB431542 (5 μM, Sigma), respectively. For rescue experiments, cultures were supplemented with DMK (3.5 mM, Sigma-Aldrich). Cell activations were performed with plate-bound anti-CD3 (clone 17A2 or 2C11, 1 μg/ml) and anti-CD28 (clone PV-1 or 37.5, 1 μg/ml) monoclonal antibodies in RPMI 1640 medium (Life Technologies) supplemented with 10% FCS and IL-2 (100 U/ml) in the presence or absence of 2 mM glutamine or 11 mM glucose. In some experiments, glutamine was added at the concentrations indicated in the figure legends. Exogenous IL-2 (100 U/ml) was added every other day starting on day 2 after stimulation. To block glutaminolysis, CD4+ T cells were activated in the presence of 3 μM l-DON (Sigma-Aldrich). Cell proliferation was monitored by labeling the cells with 2.5 μM CFSE (Life Technologies) or 5 μM CTV (Life Technologies) at 37°C for 3 or 8 to 10 min, respectively, for in vitro and in vivo manipulations.

Gene expression analysis

RNA was isolated from purified CD4+ T cells with the RNeasy Micro Kit (Qiagen) and then was reverse-transcribed into cDNA by oligonucleotide priming with the QuantiTect Reverse Transcription Kit (Qiagen). qRT-PCR analysis was performed with the LightCycler 480 SYBR Green I Master kit (Roche) and the following specific primers: Tbet sense, 5′-TCCCCCAAGCAGTTGACAGT-3′; Tbet antisense, 5′-CAACAACCCCTTTGCCAAAG-3′; Gata3 sense, 5′-AGTTCGCGCAGGATGTCC-3′; Gata3 antisense, 5′-AGAACCGGCCCCTTATCAA-3′; Foxp3 sense, 5′-CCCAGGAAAGACAGCAACCTT-3′; Foxp3 antisense, 5′-TTCTCACAACCAGGCCACTTG-3′; Gls2 sense, 5′-AGCGTATCCCTATCCACAAGTTCA-3′; Gls2 antisense, 5′- GCAGTCCAGTGGCCTTCAGAG-3′; Hprt sense, 5′-CTGGTGAAAAGGACCTCTCG-3′; Hprt antisense, 5′-TGAAGTACTCATTATAGTCAAGGGCA-3′.

Immunophenotyping and flow cytometric analysis

Immunophenotyping of cells was performed with fluorochrome-conjugated antibodies, and intracellular staining was performed after the fixation and permeabilization of the cells (intracellular staining kit, eBioscience). Phosphorylation of S6 was assessed after fixation in 4% paraformaldehyde and permeabilization by staining with an anti-pS6 antibody (clone 91B2, Cell Signaling) and revealed with a secondary anti-rabbit immunoglobulin G antibody conjugated to Alexa Fluor 647 (Life Technologies). IL-4 production was assessed on day 6 of polarization with a Cytometric Bead Array (CBA) Kit (BD Biosciences). Before being subjected to staining for intracellular cytokines, cells were activated with phorbol myristate acetate (100 ng/ml, Sigma-Aldrich) and ionomycin (1 μg/ml Sigma-Aldrich) in nutrient-replete medium in the presence of brefeldin A (10 μg/ml, Sigma-Aldrich) for 3.5 to 4 hours at 37°C. Cells were analyzed with a FACSCanto flow cytometer (BD Biosciences) or were sorted on a BD FACSAria flow cytometer. Data analysis was performed with FlowJo Mac version 8.8.7 software (Tree Star) and FCAP Array Software (CBA analysis).

Bisulfite pyrosequencing

For each analyzed cell type, cells (1 × 106) were subjected to FACS to >95% purity. Genomic DNA was isolated from the sorted cells with the DNeasy kit (Qiagen) and bisulfite-converted with the EZ DNA Methylation Kit (Zymo Research) according to the manufacturer’s instructions. The murine TSDR was amplified by PCR in a reaction containing 20 ng of bisulfite-converted genomic DNA, 25 μl of ZymoTaq PreMix (Zymo Research), and forward and reverse primers (0.4 μM each) in a final volume of 50 μl. PCR conditions were as follows: 95°C for 10 min; 45 cycles of 94°C for 30 s, 60°C for 1 min, and 72°C for 1 min; 72°C for 7 min; and 4°C for > 4 min. The PCR products were analyzed by gel electrophoresis. The PCR product (40 μl), PyroMark Gold Q96 reagent (Qiagen), PyroMark buffer (Qiagen), streptavidin Sepharose (GE Healthcare), and the sequencing primer were used for pyrosequencing on a PSQ96MA instrument (Qiagen) according to the manufacturer’s protocol. The primers mTSDR-amp forward (TAAGGGGGTTTTAATATTTATGAGGTTT), which was biotinylated at the 5′ end, and mTSDR-amp reverse (CCTAAACTTAACCAAATTTTTCTACCA) were used for TSDR amplification, whereas the primers mTSDR-seq1 (CCATACAAAACCCAAATTC), mTSDR-seq2 (ACCCAAATAAAATAATATAAATACT), mTSDR-seq3 (ATCTACCCCACAAATTT), and mTSDR-seq4 (AACCAAATTTTTCTACCATT) were used for pyrosequencing. Male mice were used to analyze DNA methylation status to avoid artificial recalculation because of X chromosome inactivation in female mice.

Glutamine uptake assay

Before being analyzed for glutamine uptake, CD4+ T cells (0.5 × 106) were starved in glutamine-free RPMI for 30 min at 37°C. Glutamine uptake assays were initiated by the addition of l-2,3,4-[3H]glutamine (0.5 μCi, PerkinElmer) for 5 min at room temperature. Cells were solubilized in 500 μl of 0.1% SDS, and radioactivity was measured by liquid scintillation.

Metabolic flux analysis

OCR and ECAR were measured with the XF24 Extracellular Flux Analyzer (Seahorse Bioscience). Cells (1 × 106) were placed in XF medium (nonbuffered Dulbecco’s modified Eagle’s medium containing 2.5 mM glucose, 2 mM l-glutamine, and 1 mM sodium pyruvate) and monitored under basal conditions and in response to 1 μM oligomycin, 1 μM FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone), 100 nM rotenone, and 1 μM antimycin A (Sigma). The basal respiration rate was calculated as the difference between the OCR under basal conditions and the OCR after the inhibition of mitochondrial complexes 1 and 3 with rotenone and antimycin A, respectively. SRC was determined as the difference between the readings obtained under basal conditions and after injection of the ionophore FCCP. Mitochondrial ATP synthesis was estimated from the decrease in oligomycin-sensitive oxygen consumption using a phosphate/oxygen ratio of 2.3 as described previously (63).

Adoptive T cell transfer and induction of colitis

To induce colitis, Rag2−/− C57BL/6 mice were injected with sorted CD45.1+CD4+CD45RBhi T cells (2.5 × 105) alone or in combination with sorted CD4+Foxp3/GFP+ T cells (2.5 × 105) from CD45.2+Foxp3-GFP mice. The mice were injected with either freshly isolated Treg cells or Foxp3+GFP+ T cells obtained after 5 days of stimulation with anti-CD3 and anti-CD28 antibodies under glutamine-depleted conditions (fig. S3). After they were injected with the T cells, the mice were weighed weekly for 9 weeks or were sacrificed if their weight loss exceeded 20% of their total body weight. Lymphoid tissues were harvested, and colons were fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with H&E.

Statistical analyses

P values were determined with an unpaired Student’s t test, ANOVA, Mann-Whitney with a two-tailed distribution, or a χ2 test when indicated. Survival differences during the development of colitis were analyzed with a Mantel-Cox test (GraphPad Software).

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/8/396/ra97/DC1

Fig. S1. Reduced amounts of glutamine in the culture medium inhibit the conversion of naïve CD4+ T cells into Foxp3+ T cells.

Fig. S2. The methylation status of the TSDR is similar in iTreg and Treg cells generated in the absence of glutamine.

Fig. S3. Strategy for the purification of Foxp3+CD4+ T cells by flow cytometry.

Fig. S4. Naïve CD4+ T cells secrete IFN-γ after stimulation under TH1-polarizing conditions in glucose-depleted medium.

REFERENCES AND NOTES

Acknowledgments: We thank all members of our laboratories for discussions, scientific critique, and continual support. We are indebted to A. Singer for his important input and insights, as well as for his critical reading and correcting of the manuscript. This study was substantially fostered by extensive interactions and discussions with members of the ATTACK (Adoptive engineered T cell Targeting to Activate Cancer Killing) Consortium and specifically with R. Debets, D. Gilham, A. Mondino, H. Stauss, P. Velica, and M. Zech, and we are very appreciative to all of them for having provided this opportunity. We are grateful to M. Boyer and S. Gailhac of Montpellier Rio Imaging for support in cytometry experiments and to the RAM (Réseau des animaleries de Montpellier) animal facility of the Institut de Génétique Moléculaire de Montpellier and T. Gostan of the SERANAD platform for data analyses. Funding: D.K. was supported by a European Marie Curie fellowship (ATTACK); P.R. was supported by a Research Ministry fellowship administered by the University of Montpellier II; L.O. and G.C. were supported by fellowships from the Ligue Contre le Cancer; L.O. is presently supported by the Association de la Recherche contre le Cancer (ARC); M.C. was supported by a fellowship from the Portuguese Foundation for Science and Technology; V.F. was supported by a fellowship from the Association Française contre les Myopathies (AFM); C.Y. was supported by a fellowship from Cancer Therapeutics CRC and an Australian Postgraduate Award; X.T. was supported by the NIH intramural program; C.M., V.Z., S.K., and V.D. were supported by the CNRS; J.M. and N.T. were supported by INSERM. This work was supported by generous funding from the European Community [contracts LSHC-CT-2005-018914 (ATTACK) and PIRG5-GA-2009-249227 (T cell homeostasis)], a CNRS-NIH International Laboratory grant from the CNRS (LIA-BAGEL), ARC, French national research grants ANR (PolarATTACK, GlutStem, NutriDiff, and ANR-10-LABX-61), INCa, German Research Foundation (KFO250), and the French laboratory consortiums (Labex) EpiGenMed and GR-EX. Author contributions: D.K., V.D., and N.T. designed the study; D.K., X.T., P.R., M.C., G.C., L.O., C.M., S.F., V.F., M.I.M., C.Y., N.S., J.M., V.Z., S.K., and V.D. performed the experiments; D.K., X.T., P.R., M.C., G.C., L.O., C.M., S.F., V.F., M.I.M., C.Y., N.S., J.M., J.H., V.Z., S.K., V.D., and N.T. analyzed the data; and D.K., V.D., and N.T. wrote the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article