Control of T H 17/T reg Balance by Hypoxia-Inducible Factor 1

Transcription

1 ontrol of T H 17/T reg alance by Hypoxia-Inducible Factor 1 Eric V. Dang, 1,9 Joseph arbi, 1,9 Huang-Yu Yang, 1,7,9 Dilini Jinasena, 1 Hong Yu, 1 Ying Zheng, 1 Zachary ordman, 1 Juan Fu, 2 Young Kim, 2 Hung-Rong Yen, 1,8 Weibo Luo, 3 Karen Zeller, 4 Larissa Shimoda, 5 Suzanne L. Topalian, 6 Gregg L. Semenza, 3 hi V. Dang, 4 Drew M. Pardoll, 1, * and Fan Pan 1, * 1 Immunology and Hematopoiesis Division, Department of Oncology and Medicine, Sidney Kimmel omprehensive ancer enter 2 Department of Otolaryngology-Head and Neck Surgery, Oncology 3 Vascular Program, Institute for ell Engineering; Departments of Pediatrics, Medicine, Oncology, Radiation Oncology, and iology hemistry; and McKusick-Nathans Institute of Genetic Medicine 4 Division of Hematology, Department of Medicine, Oncology 5 Division of Pulmonary and ritical are Medicine, Department of Medicine 6 Department of Surgery Johns Hopkins University School of Medicine, Maryland 21231, USA 7 Department of Nephrology 8 Department of Pediatrics, enter for Traditional hinese Medicine hang Gung Memorial Hospital, hang Gung University ollege of Medicine, Taoyuan 333, Taiwan 9 These authors contributed equally to this work *orrespondence: dmpardol@jhmi.edu (D.M.P.), fpan1@jhmi.edu (F.P.) DOI 1.116/j.cell SUMMARY T cell differentiation into distinct functional effector and inhibitory subsets is regulated, in part, by the cytokine environment present at the time of antigen recognition. Here, we show that hypoxia-inducible factor 1 (HIF-1), a key metabolic sensor, regulates the balance between regulatory T cell (T reg ) and T H 17 differentiation. HIF-1 enhances T H 17 development through direct transcriptional activation of RORgt and via tertiary complex formation with RORgt and p3 recruitment to the IL-17 promoter, thereby regulating T H 17 signature genes. oncurrently, HIF-1 attenuates T reg development by binding Foxp3 and targeting it for proteasomal degradation. Importantly, this regulation occurs under both normoxic and hypoxic conditions. Mice with HIF-1a-deficient T cells are resistant to induction of T H 17-dependent experimental autoimmune encephalitis associated with diminished T H 17 and increased T reg cells. These findings highlight the importance of metabolic cues in T cell fate determination and suggest that metabolic modulation could ameliorate certain T cellbased immune pathologies. INTRODUTION Host defense against microorganisms requires a complex network of specialized T cell populations that are responsible for triggering inflammation to eradicate the infection, resolving the inflammatory phase after elimination of the threat, and attenuating dysregulated or inappropriate immune responses. Despite their diverse functions, these T cell subsets largely differentiate from the same pool of precursor naive D4 + T cells upon stimulation by antigen in the presence of unique cytokine signals that are present in the microenvironment (Murphy and Reiner, 22; Zhu et al., 21). For instance, T H 1 cells are induced by type 1 interferons and propagated by IL-12; T H 2 cells require IL-4; and T H 17 cells are induced by IL-6 and TGF-b and propagated by IL-23 and IL-21. The activity of all of these effector T cells is attenuated by anti-inflammatory regulatory T cells (T regs ) that inhibit T cell proliferation and autoimmune responses (arnes and Powrie, 29; Sakaguchi et al., 28). T regs can be induced from naive T cells upon exposure to TGF-b and are propagated by IL-2 (hen et al., 23; Davidson et al., 27; Rajewsky and von oehmer, 28). T reg are commonly categorized into thymus-derived natural T reg (nt reg ) and induced T reg (it reg ). oth it reg and nt reg express Foxp3 as a core subset-specific transcription factor, which activates a large bank of genes that mediate the suppressive phenotype of T reg and also silences many effector T cell genes (Getnet et al., 21; Pan et al., 29; Sakaguchi et al., 28). TGF-b has been shown to maintain peripheral nt reg cells that develop in the thymus, and its deficiency leads to the development of early lethal autoimmunity. Moreover, TGF-b induces Foxp3 expression in peripheral naive T cells, leading to the differentiation of it regs, which exhibit a suppressive phenotype similar to that seen in ntreg cells (Sakaguchi et al., 28). T H 17 and T reg cells share a common requirement for TGF-b in their differentiation requirements despite expressing distinct transcriptional regulators (RORgt versus Foxp3, respectively) and demonstrating opposing functions (inflammatory versus anti-inflammatory) (ettelli et al., 26; Dong, 28; Littman and Rudensky, 21; Mangan et al., 26; O Quinn et al., 28; Veldhoen et al., 26). T H 17 cells are important in responses mounted against extracellular bacterial infections of 772 ell 146, , September 2, 211 ª211 Elsevier Inc.

2 the intestine and the airways (Korn et al., 29). Despite providing a benefit in these settings, T H 17 can play a pathologic role in the induction of several autoimmune diseases, including collagen-induced arthritis, experimental autoimmune encephalomyelitis (EAE), inflammatory bowel diseases (Fife et al., 29; Weaver et al., 27; Wu et al., 29), and inflammation-induced carcinogenesis (Wu et al., 29). In these models, T reg -mediated suppression of T H 17 responses often plays a protective role against pathology associated with the disease (Ahern et al., 21). In addition to TGF-b, initial T H 17 priming also requires the cooperative action of IL-6 signaling. Subsequently, IL-23 and IL-21 play a key role in the maintenance of T H 17 differentiation by enhancing the transcription of IL-17 and other T H 17 signature cytokines (Korn et al., 29; Littman and Rudensky, 21). Interestingly, IL-6, IL-21, and IL-23 all activate Stat3, which is critical for the effects of these cytokines on T H 17 cell differentiation (Zhou et al., 28; Harris et al., 27). Though it is now known that Stat3 induces RORgt gene expression (Harris et al., 27; Yang et al., 27) the key transcription factor required for T H 17 development (Ivanov et al., 26) the mechanism by which this is accomplished remains unclear. Interestingly, the defect in T H 17 differentiation seen in the absence of Stat3 can be only partially rescued by RORgt overexpression (Yang et al., 28b). Furthermore, although Stat3 binds most of the genes that have been shown to promote T H 17 cell fate determination, it also binds to many genes that are involved in T cell survival and proliferation (Durant et al., 21). Genes bound by RORgt, on the other hand, are fairly specific for the T H 17 differentiation program. learly, additional regulators contribute to the control of the T H 17 transcriptional program. Interestingly, accumulating evidence suggests that T H 17 and it reg arise from a common precursor. In response to TGF-b in vitro as well as in vivo, many T cells coexpress RORgt and Foxp3 (Veldhoen et al., 26; Yang et al., 28a). Ultimately, depending on the interplay between additional environmental cues, such as the relative amounts of IL-6 and TGF-b, one or the other subset emerges as the dominant phenotype. High TGF-b levels in the absence of IL-6 induce Foxp3 and repress IL-23R transcription. Foxp3 can, in turn, bind to the RORgt protein and antagonize its ability to bind DNA, thus pushing T cell differentiation away from the T H 17 transcriptional program and toward the T reg lineage. On the other hand, proinflammatory cytokines, such as IL-6 or IL-21 in the presence of low TGF-b, activate Stat3, which overcomes Foxp3 inhibition of RORgt transcriptional activity. This leads to the upregulation of the IL-23R, thus pushing T cell differentiation toward a T H 17 fate (Zhou et al., 28). Ultimate T H 17 differentiation is associated with Foxp3 downregulation and sustained, unopposed RORgt and Stat3 transcriptional activity, although the mechanisms underlying Foxp3 suppression during T H 17 lineage commitment are not fully understood. In this study, we explored the role of a key metabolic sensor and regulator in T cell fate determination. We report that hypoxia-inducible factor 1 (HIF-1), the transcription factor that mediates the metabolic switch from oxidative phosphorylation to aerobic glycolysis in response to hypoxia (Semenza, 27), in fact regulates the T H 17/T reg balance. Specifically, we have discovered that HIF-1 promotes T H 17 differentiation by directly inducing RORgt transcription and subsequently collaborates with RORgt to regulate downstream T H 17 genes. In addition, HIF-1 inhibits T reg differentiation through an active process that targets Foxp3 protein for degradation. Given the plasticity between T H 17 and T reg programs, our study sheds light on how this balance is subject to metabolic regulation and suggests new strategies to manipulate these cell lineage decisions in order to treat diseases associated with T H 17/T reg imbalance. RESULTS HIF-1a Expression Is Upregulated in a Stat3-Dependent Manner in T ells under T H 17-Skewing onditions HIF-1 is a heterodimeric transcription factor consisting of a highly regulated oxygen-sensitive HIF-1a subunit and a constitutively present b subunit (termed HIF-1b) (Semenza, 27). HIF-1a expression in T cells can be induced both by hypoxic and nonhypoxic stimuli, including TR-triggered and PI3K-mediated pathways, that result in mrna upregulation and protein stabilization (Lukashev et al., 26). Intrigued by a potential link between the requirement for Stat3 in the T H 17 program and our previous observation that HIF-1a is a target gene for activated Stat3 in tumor cells (Harris et al., 27; Xu et al., 25), we tested the hypothesis that HIF-1a plays a positive role in T H 17 development. We began by analyzing HIF-1a mrna expression in helper T cell subsets using qrt-pr and found that HIF-1a mrna was most highly expressed in T H 17 cells (Figure 1A). When naive T cells were activated in the presence of TGF-b + IL-6 (T H 17- skewing conditions), HIF-1a mrna increased, peaked after 48 hr of culture, and then began to decline (Figure 1) relative to consistently low background signal established using HIF-1a knockout (HIF-1a / ) T cells (generated by crossing D4 cre 3 HIF-1a flox/flox mice, termed T-HIF-1 / mice). T-HIF-1 / mice had normal development of D4 + and D8 + T cells, cells, and dendritic cells (Figure S1 available online). Whereas wildtype (WT) T cells activated under T H 17-skewing conditions accumulated IL-17A and RORgt transcripts over time, the induction of these T H 17-associated mrnas by HIF-1a deficient T cells was minimal (Figure 1 and 1D). ecause HIF-1a mrna was upregulated during in vitro T H 17 differentiation and RORgt and IL-17A mrna induction were significantly diminished in the absence of HIF-1a, we suspected that HIF-1a regulates important components of the T H 17 pathway. Furthermore, we measured HIF-1a and HIF-1b protein in T cells stimulated with differing levels of IL-6 and TGF-b. Whereas HIF-1b expression was constitutive as expected under all treatments, both TGF-b and IL-6 alone induced HIF-1a, and the combination of the two generated significantly higher levels of HIF-1a protein even under normoxic conditions (Figure 1E). As Stat3 is known to mediate T H 17 lineage commitment and has previously been shown to regulate HIF-1a in tumor cell lines, we hypothesized that HIF-1a induction may be diminished in Stat3-deficient T cells activated under T H 17-generating conditions. Indeed, after isolating D4 + D62L high D25 (naive) Stat3-deficient T cells from D4-cre 3 Stat3 flox/flox mice and culturing them under T H 17 skewing conditions for 4 days, very little HIF-1a protein was detected by western blot (Figure 1E). Furthermore, a chromatin immunoprecipitation (hip) experiment was carried out ell 146, , September 2, 211 ª211 Elsevier Inc. 773

3 H H H H D E F Figure 1. HIF-1a mrna Is Upregulated in T ells under T H 17-Skewing onditions in a Stat3-Dependent Manner (A) Naive (D4 + D25 D62L high ) T cells were cultured under TH-cell subset-inducing conditions, and HIF-1 mrna was detected by qrt-pr. ( D) Naive T cells from wild-type (WT, HIF-1a +/+ )ord4 re 3 HIF-1a flox/flox mice (ko, HIF-1a / ) were stimulated in the presence of TGFb and IL-6. RNA was isolated from these cells, and qrt-pr was performed at different time points during culture to measure HIF-1a (), IL-17 (), and RORgt (D) transcript levels. (E) WT or Stat3 / (obtained from D4 re 3 Stat3 flox/flox mice) naive D4 + T cells were isolated and cultured under the indicated conditions for 2 days, followed by SDS-PAGE and western blotting using antibodies against HIF-1a (top), HIF-1b (middle), and tubulin (bottom), respectively. (F) A hip assay was used to examine direct Stat3 binding to the HIF-1a promoter. (A) (D) and (F) depict the mean + SD of at least three experiments; (E) is a representative result. See also Figure S1. to test whether Stat3 directly induces HIF-1a gene expression in T H 17 cells. Stat3 was directly associated with the promoter region of HIF-1a in the WT T H 17-skewed T cells, but not at the promoter region of GAPDH (Figure 1F). These findings confirm that HIF-1a expression is induced in T cells during T H 17 differentiation in a Stat3-dependent manner. HIF-1 Positively Regulates T H 17 Development at Multiple Levels Further analysis of T H 17 differentiation by intracellular staining was performed on WT and HIF-1a / T cells. We observed significant induction of IL-17A protein by WT T cells stimulated with TGF-b and IL-6 (Figure 2A). Strikingly, the pattern of IL- 17A and Foxp3 protein expression in HIF-1a / T cells cultured under the same conditions was dramatically altered (Figure 2A, right). In agreement with the mrna results from Figure 1, these HIF-1-deficient T cells displayed a marked reduction in IL-17A + cells. Despite this notable disparity in IL-17 induction, no significant differences in IFN-g or IL-4 were seen between WT and HIF-1a / T cells under T H 1- or T H 2-skewing conditions (Figure S2A), and D4 + cells from both groups showed comparable proliferative capacity (Figure S2). These results suggest that HIF-1a deficiency specifically affects the T H 17 pathway, as opposed to creating a global defect in T cell proliferation or differentiation capacity. The mrna expression of other T H 17-related genes such as IL- 17F and IL-23R was also much lower in HIF-1a / T cells than in WT T cells, indicating that HIF-1 regulates multiple genes in the T H 17 program (Figure 2). Furthermore, we also noted a dramatic increase in the proportion of Foxp3 + cells in HIF-1a / T cells (Figure 2A). To directly assess the role of HIF-1a in T H 17 programming, we used retroviral transduction to overexpress HIF-1a in naive D4 + cells followed by detection of IL-17A and IFN-g by intracellular cytokine staining (IS). The ectopic expression of HIF-1a substantially increased the percentage of IL-17A + cells even in the absence of the T H 17-driving cytokines IL-6 and TGF-b (Figure 2). This induction of IL-17A was associated with the upregulation of both RORgt mrna and protein (Figure S2). The IL-17A + cell population was also significantly increased upon overexpression of HIF-1a under T H 17-skewing conditions (Figure 2, right). These results support the notion that HIF-1a might facilitate T H 17 development through the upregulation of RORgt. 774 ell 146, , September 2, 211 ª211 Elsevier Inc.

4 Relative mrna expression D Figure 2. HIF-1a Is Required for T H 17 Development In Vitro (A) Naive T cells isolated from wild-type (WT, HIF-1a +/+ )ord4 re 3 HIF-1a flox/flox (HIF-1a / ) mice were cultured under T H 17-skewing conditions with anti-d3/ D28 and TGFb, IL-6, and anti-ifng, IL-12, and IL-4 antibodies for 6 days. ells were then stained for IL-17 and Foxp3 (see Experimental Procedures). Numbers represent the percentage of D4 + cells positive for the indicated marker. () FAS-sorted naive D4 + T cells from WT or T-HIF-1a / mice were activated and cultured under T H 17 skewing for 4 days. Total RNA was isolated, and mrna expression of Il17, Il17f, and IL23r genes was assessed by qrt-pr. For each gene, expression level in HIF-1a / T cells was set to 1. The mean + SD of at least three trials is shown. () Naive D4 + T cells were activated with anti-d3/d28 under either neutral (anti-il-4 and anti-ifng antibodies) or T H 17-skewing conditions and were transduced with a bicistronic retrovirus expressing HIF-1a-GFP or GFP alone. Intracellular cytokines were stained and analyzed in GFP + cells. Numbers represent the percentage of gated GFP + cells. (D) Naive T cells isolated from WT or T-HIF-1a / mice were activated as described for (A) under normoxia or hypoxia (see Experimental Procedures). () and (D) represent at least two independent experiments. See also Figure S2. The fact that HIF-1a is a major sensor of metabolic cues, such as oxygen tension, suggests that this protein could be an important link between metabolism and T cell fate determination. To formally test this, we asked whether hypoxia, which stabilizes HIF-1a protein levels and subsequent transcriptional activity, could affect T H 17 differentiation. Indeed, after T H 17 skewing, we observed a higher proportion of IL-17A + cells in hypoxic compared to normoxic culture conditions. This increase was abrogated in HIF-1a / T cells (Figure 2D and Figure S2E), indicating that the hypoxia-induced enhancement of T H 17 differentiation is HIF-1a dependent. These findings support the idea that metabolic cues modulate T cell differentiation and that HIF-1 is an important mediator of this effect. ell 146, , September 2, 211 ª211 Elsevier Inc. 775

5 e Relative Luciferase Val mock PMA/Iono elative Luciferase Valu mock PMA/Iono Enrichment (fold) IgG hip HIF-1α hip RORγt-Luc wt RORγt-Luc Mut. HIF-1α Wt HIF1α-ΔDD Gmpr RORγt D vector HIF-1α +/+ HIF-1α -/- E RORγt HIF-1α +/+ HIF-1α -/- 7 IL- IL- 7 Foxp3 Foxp3 Figure 3. HIF-1a Transactivates RORgt Transcription and Is Necessary for RORgt-Driven T H 17 Differentiation In Vitro (A) Jurkat T cells were cotransfected with a luciferase reporter under the control of a wild-type RORgt promoter or one with a mutated HIF-1-binding site. At 24 hr posttransfection, cells were either treated with PMA and ionomycin or left untreated prior to analysis of luciferase activity, which was normalized to that of Renilla luciferase. () Jurkat T cells were transfected with a RORgt-luciferase reporter plasmid (incorporating the RORgt promoter) along with plasmid encoding either wild-type HIF-1a or a HIF-1a mutant (with a DNA-binding domain deletion). At 24 hr posttransfection, cells were stimulated and luciferase activity was assessed as described in (A). () A hip assay was used to measure direct HIF-1 binding to the RORgt promoter. Data are shown for three independent experiments (mean + SD). (D and E) Naive WT or HIF-1a / D4 + T cells were activated with anti-d3/anti-d28 under neutral (anti-il4 and anti-ifng antibodies) conditions and were transduced with either a bicistronic retrovirus expressing RORgt-GFP or the GFP containing empty vector. Intracellular IL-17 and Foxp3 were stained. The plots shown are gated on GFP + cells. These experiments were repeated at least twice with consistent results. See also Figure S3. HIF-1a Activates RORgt Transcription and ooperates with RORgt and p3 to Activate the IL-17A Gene during T H 17 Development Next, we sought to further dissect the molecular mechanism by which HIF-1a regulates T H 17 development. Given that RORgt is a key transcriptional regulator of T H 17 cells, as well as the dramatically blunted induction of RORgt observed in HIF-1a / T cells (Figure 1D), we determined whether HIF-1a directly regulates RORgt gene expression. We used a reporter assay with the luciferase gene under the control of the 1.1 kb RORgt promoter (Figures S3A and S3). Transfection of Jurkat cells with the RORgt promoter-driven reporter construct, together with increasing amounts of HIF-1a-encoding plasmid, indeed caused corresponding increases in RORgt promoter activity, particularly upon PMA and ionomycin stimulation (Figure S3). Analysis of the RORgt promoter sequence revealed that a hypoxia response element (HRE, a conserved HIF-1a-binding site) is located in the proximal region of the RORgt promoter (Figure S3A). To further assess the importance of HIF-1a in RORgt promoter activation, we tested the reporter activity of a promoter with mutated HIF- 1a-binding sites (Figure S3). The mutated constructs had much lower luciferase activity compared to that of a wild-type RORgt promoter-driven luciferase reporter (Figure 3A). Likewise, the expression of a Hif1a mutant gene lacking the DNA-binding domain failed to activate the wild-type RORgt promoter (Figure 3). ecause the human and mouse RORgt promoter sequences share a high degree of identity as well as a conserved HRE (Figure S3A), we also tested the importance of HIF-1a in activation of the human RORgt promoter. Luciferase assays using the human sequence demonstrated that HIF-1a activates this promoter in a similar fashion to that of the mouse (Figures S3D and S3E). These results suggest that HIF-1 can regulate RORgt gene expression. To further dissect this mechanism, a hip assay was employed to test whether HIF-1a directly binds to the promoter region of the RORgt gene in primary, in vitro-generated T H ell 146, , September 2, 211 ª211 Elsevier Inc.

6 cells. Indeed, HIF-1a was found to bind to the HRE located at the RORgt promoter, but not the promoter of the non-hif-1a-regulated Gmpr gene (Figure 3 and Figures S4A S4D). These results collectively indicate that HIF-1a directly transactivates RORgt gene expression. ecause our results suggested that RORgt acts directly on the IL-17A promoter, we asked whether RORgt could rescue T H 17 development in HIF-1a / T cells. In HIF-1a +/+ T cells, retroviral RORgt expression induced a significantly higher proportion of IL-17A + cells (45.7%) under nonskewing conditions, compared to control retrovirus recipients (.72%) (Figures 3D and 3E, left). However, ectopic RORgt expression in HIF-1a / T cells induced a strikingly diminished population of IL-17A + cells (1.33%) (Figure 3E, right). These results suggest that HIF-1a not only directly mediates RORgt gene expression, but also is required for the full function of RORgt in IL-17A gene transcription. To gain further insight into how HIF-1a and RORgt coregulate IL-17A transcription, we performed a luciferase reporter assay to determine whether transient expression of HIF-1a and RORgt could activate the 1.3 kb IL-17 promoter. The expression of wild-type HIF-1a in the presence of RORgt resulted in significantly increased reporter gene activity over that induced by RORgt alone (Figure 4A). In striking contrast to our findings with the RORgt promoter, a HIF-1a mutant with a characterized DNA-binding domain deletion (Arany et al., 1996) showed no defect in inducing reporter gene activity, which implies that DNA binding is dispensable for HIF-1a-mediated IL-17A expression. ecause the DNA-binding domain of HIF-1a did not appear to be required for its function in facilitating RORgt activity, we speculated that HIF-1a might physically associate with RORgt, serving as a coactivator for RORgt without direct DNA binding. To test this hypothesis, a Flag-tagged HIF-1a plasmid was coexpressed with Myc-tagged RORgt in Jurkat T cells. HIF-1a was found to coimmunoprecipitate (co-ip) with RORgt (Figure S4E). The association between HIF-1a and RORgt was further confirmed by co-ip of endogenous HIF-1a and RORgt in T cells cultured under T H 17-skewing conditions (Figure 4). To verify that the interaction between HIF-1a and RORgt was direct, we employed affinity-purified recombinant GST fusion proteins to pull down Myc-RORgt from 293T cell lysate. GST-HIF-1a (amino acids1 8), rather than GST alone, was bound to RORgt (Figure S4F). Further testing this apparent direct association of RORgt and HIF-1a, we purified His-tagged RORgt from E. coli, and a GST pull-down assay was carried out using GST or GST-HIF-1a fragments as indicated. Only GST fusions of the HIF-1a N terminus (1 8 aa), but not GST alone or other HIF-1a fragments, successfully pulled down RORgt (Figure 4). Taken together, these results suggested a direct interaction between the two proteins in vitro. ecause it has been shown that the transcription factor p3 is critical for HIF-1a-mediated gene activation under hypoxic conditions, we hypothesized that HIF-1a might function in a complex with RORgt by recruiting p3 to activate target genes such as IL-17A. To test this hypothesis, a hip assay was employed to determine whether RORgt, HIF-1a, and p3 associate with the IL-17A promoter during T H 17 differentiation. Indeed, we detected localized binding of RORgt, HIF-1a, and p3 to the IL-17A promoter determined using primers flanking the RORgt-binding site located at the IL-17A promoter region (Figure 4D). This binding, as analyzed by hip, was specific because no significant signal could be detected at a non- RORgt-binding region within the IL-17A promoter. Furthermore, hip analysis revealed that HIF-1a was recruited to several other putative RORgt-binding regions, including NS2. This recruitment of HIF-1a was dependent on RORgt because it was abrogated in T cells from RORgt knockout mice (Figures S5A S5D). Additionally, we observed similar colocalized binding of RORgt, HIF-1a, and p3 at the promoters of the T H 17 signature genes IL-17F and IL-23R (Figure 4E). The role of the RORgt/HIF-1a/ p3 complex in the activation of the IL-17A promoter was further confirmed by reporter assays in Jurkat and 293T cells. Surprisingly, coexpression of RORgt, HIF-1a, and p3 led to a more than 1-fold increase in luciferase activity compared to that induced by combinations of any two of these proteins (Figures S5E and S5F). However, a HIF-1a mutant with a -terminal deletion of the p3-binding domain (Arany et al., 1996) failed to synergize with RORgt and p3 in the activation of the IL-17A promoter, implying that p3 is required for the full function of RORgt and HIF-1a in the regulation of IL-17A gene expression (Figure S5E). In addition, IL-17A production in Jurkat T cells cotransfected with plasmids encoding RORgt, HIF-1a, and p3 was further increased under hypoxic conditions (Figure 4F). These results suggest that, though active under normoxic conditions, HIF-1a s ability to promote IL-17A production is indeed modulated by oxygen tension. Of note, the gut, which is known to be relatively hypoxic under physiologic conditions (Koch, 22), contains a high proportion of IL-17A-producing cells. We found that, under steady-state conditions, T-HIF-1 / mice had reduced proportions of IL-17A + cells in the lamina propria (Figures S6A S6D), indicating a potential role for HIF-1a in promoting T H 17 development in the context of microbial influences. p3 possesses histone acetyltransferase (HAT) activity (Thompson et al., 24). In order to determine whether this HAT activity is required for the RORgt/HIF-1a /p3 complexmediated IL-17A gene expression, we performed cotransfection studies using a p3 mutant with a HAT domain deletion (Dp3) (Youn et al., 2). Wild-type p3, but not Dp3, could synergistically enhance IL-17A promoter activity in the presence of HIF-1a and RORgt (Figure S5E). Furthermore, a HIF-1a mutant with a deletion in its p3-binding domain (Arany et al., 1996) fails to activate the IL-17A promoter. These findings indicate that RORgt and HIF-1a, through recruitment of p3 to the IL-17A promoter, acetylate histones to open the chromatin structure and facilitate gene expression. In support of this notion, hip assays revealed that histones H3 and H4 around the IL-17A promoter region were highly acetylated in WT compared to HIF-1a / T cells under T H 17-inducing conditions (Figure 4G). Similarly, hyperacetylated histones H3 and H4 were associated with the promoter of T H 17 signature genes IL-17F and IL-23R in WT, but not in HIF-1a /,T H 17 cells (Figures S6E and S6F). Taken together, these results demonstrate that HIF-1 plays a dual role in regulating IL-17A transcriptional activity by directly activating RORgt transcription and then associating with RORgt at the IL-17A promoter to recruit p3, thus generating a permissive chromatin structure. ell 146, , September 2, 211 ª211 Elsevier Inc. 777

7 tivity Relative luciferase act mock PMA/Iono EV HIF-1α only HIF-1α- ΔDD RORγt only RORγt +HIF-1α RORγt +HIF-1α-ΔDD ontrol IgG IP anti-rorγt lysate W anti-hif-1α anti-rorγt t ontrol IgG o IP nti-hif-1α an sate lys W anti-rorγt anti-hif-1α anti-his HIF-1 (81-329aa) HIF-1 (1-8aa) GST only Input:His-RORγt W:anti-His D chment (fold) Enric Gmpr Putative RORγt inding region IgG hip RORγt hip HIF-1α hip p3 hip Non-RORγt inding region E Enrichment (fold) Putative RORγt inding region IL-17F loci Non-RORγt inding region Putative RORγt inding region IL-23R loci IgG hip RORγt hip HIF-1α hip p3 hip Non-RORγt inding region F ve Luciferase Value Relativ PMA/Iono Normoxia Hypoxia IL-17A loci G En nrichment (fold) IgG HIF-1α HIF-1α H3(K9) H4(K8) IgG H3(K9) H4(K8) IL-17A actin Figure 4. RORgt, HIF-1a, and p3 ind to the IL-17 Promoter to Regulate Its Gene Expression (A) A HIF-1a mutant (HIF-1a-DDD, with the DNA-binding domain deleted) retains the capacity to activate IL-17 promoter-driven luciferase activity in the presence of RORgt. Jurkat T cells were transfected with an IL-17 promoter-driven luciferase reporter along with the indicated plasmids followed by stimulation and assessment as described for Figure 3A. Data are representative of at least three experiments (mean and SD of triplicate transfections). () The interaction between HIF-1a and RORgt was examined with coimmunoprecipitation. FAS-sorted D4 + T cells were activated and cultured under T H 17- skewing conditions for 5 days. The whole-cell lysate was immunoprecipitated with either anti-rorgt (left) or anti-hif-1a antibodies (right), resolved by SDS-PAGE, and blotted with the indicated antibodies. () RORgt interacts with the N terminus of HIF-1a. His-tagged RORgt purified from E. coli was incubated with different fragments of GST-HIF-1a also purified from E. coli as indicated, followed by pull-down with GST beads, resolution by SDS-PAGE, and Western blotting with anti-rorgt (top and bottom) antibodies, or anti-gst (middle). (D and E) FAS-sorted D4 + T cells from WT or HIF-1a / mice were activated under T H 17-skewing conditions (as described for ) prior to harvest for hip assay utilizing anti-rorgt, anti-hif-1a, or anti-p3 antibodies. (F) Hypoxia enhances the activation of IL-17 promoter-driven luciferase activity by HIF-1, RORgt, and p3. Jurkat T cells were transfected with a IL-17 promoterdriven luciferase reporter along with the indicated plasmids under normoxia or hypoxia, followed by stimulation and assessment as described for Figure 3. (G) Histone hyperacetylation at the IL-17 promoter was detected by hip assay in WT and HIF-1a / T cells under T H 17-skewing conditions for 5 days. (D G) Mean + SD of at least three independent experiments. See also Figure S4, Figure S5, Figure S6, and Table S1. HIF-1a Mediates Foxp3 Protein Degradation during T H 17 Development Although we established a role for HIF-1 in IL-17A gene regulation, it was unclear how HIF-1 deficiency caused Foxp3 protein accumulation in T cells under in vitro T H 17-inducing conditions (Figure 2A). uriously, Foxp3 mrna levels in WT and HIF-1a / T cells stimulated under T H 17-skewing conditions were essentially identical (Figure 5A). This implies that HIF-1 might be 778 ell 146, , September 2, 211 ª211 Elsevier Inc.

8 D pression Foxp3 mrna ex % of Max HIF-1α HIF-1α Foxp3 E ontrol IgG IP % of Max anti-hif-1α FL1-H lysate W ontrol IgG IP 64.9±5.2% 36.8±4.1% lysate F HIF-1α HIF-1α IL-6 (ng/ml) anti-tubulin 49.2±3.2% anti-hif-1α W ±5.4% anti-flag FL1-H anti-hif-1α anti-tubulin GFP-Foxp3 Foxp3+EV Foxp3+wt-ubi HIF1α IP: anti-ubi G H I J A5-HIF1 HIF-1 Mut (P42/564A) Wt-HIF-1 Wt-Ubi A5-HIF _ IP: HIF-1 Mut 1 2 (P42/564A) Wt-HIF IP: _ Wt-Ubi Foxp Foxp anti-igg MG132 normoxia hypoxia _ _ + _ + IP: normoxia hypoxia anti-ubi _ anti-igg + MG132 + _ + _ + anti-ubi anti-ubi WL anti-tubulin WL anti-tubulin WL anti-tubulin WL anti-tubulin Figure 5. HIF-1a-Mediated Foxp3 Degradation through Proteasomal Degradation Pathways (A) Naive T cells from either WT or T-HIF-1a / mice were cultured and stimulated in the presence of TGFb and IL-6, and qrt-pr was performed to measure Foxp3 mrna at different times. () HIF-1a / T cells displayed enhanced Foxp3 accumulation during in vitro Treg differentiation. Naive T cells from HIF-1a +/+ and T-HIF-1a / were isolated by FAS and activated under Treg skewing (5 ng/ml TGFb, 1 U IL-2). ells were stained for Foxp3 after 72 hr. Shown are representative histograms for HIF-1a +/+ (red line) and HIF-1a / (blue line) cells from three experiments. An isotype control is shown in green. Numbers represent the mean percentage of Foxp3 + cells. () Foxp3 protein is lost upon culture of WT T cells with IL-6 but remains unchanged in HIF-1a / Foxp3 + T cells. Naive T cells were activated under Treg-skewing conditions (TGFb, 5 ng/ml) with the indicated doses of IL-6 for 4 days. Western blotting was used to measure Foxp3 protein level. (D) Hypoxia reduced expression of Foxp3 by naive T cells during in vitro differentiation. Naive T cells isolated from Foxp3-GFP reporter mice (D4 + GFP D62L high ) were cultured under Treg-skewing conditions (see ) in either a hypoxic chamber or under normoxia (blue and red lines, respectively). Numbers represent the mean percentage of D4 + cells expressing GFP (Foxp3 + ) from three experiments. (E) HIF-1a interacts with Foxp3 in itreg cells. FAS-sorted D4 + T cells were activated and cultured under itreg-skewing conditions for 4 days. ell lysates were immunoprecipitated with anti-hif-1a (left) or antibodies (right), followed by SDS-PAGE and western blotting. (F H) HIF-1a mediates Foxp3 degradation. 293T cells were cotransfected with plasmids encoding Foxp3 +/ ubiquitin and increasing amounts of WT HIF-1a (F) or mutant HIF-1a (p42a and p564a) (G) or a deleted ODD domain (A5-HIF-1a) (H). oip and Western blots were probed as indicated. (I and J) Naive T cells were cultured under Treg-skewing conditions in normoxia (N) or hypoxia (H) for 4 days. ells were harvested and lysed, followed by immunoprecipitation with or control IgG antibodies. The pulled down protein along with an input control were resolved by SDS-PAGE, followed by western blot. Depicted are typical findings from three independent experiments. Data shown are representative of at least three independent experiments (mean + SD). See also Figure S7. involved in Foxp3 protein modulation in addition to its role in the transcriptional activation of the T H 17 pathway. Further evidence supporting this hypothesis comes from analysis of Foxp3 protein levels in T cells cultured under T reg -skewing conditions (TGF-b plus IL-2). In concordance with our findings in T cells cultured under T H 17 skewing conditions, HIF-1a / T cells cultured under T reg conditions also expressed substantially more Foxp3 than did WT T cells (Figure 5 and 5). These findings suggest that, under both T reg - and T H 17-skewing conditions, HIF-1a downregulates Foxp3 protein levels despite demonstrating no effect on Foxp3 mrna levels (Figure 5A). ased on these findings, we sought to determine whether Foxp3 levels were sensitive to hypoxia. T cells were thus cultured with TGF-b and IL-2 under hypoxic conditions. Foxp3 + cells were significantly decreased ell 146, , September 2, 211 ª211 Elsevier Inc. 779

9 under these hypoxic conditions relative to normoxia (Figure 5D). Thus, hypoxic culture conditions have opposite effects on IL-17 and Foxp3 induction. ompatible with this in vitro finding, steady-state levels of Foxp3 + cells in the lamina propria of the gut, which, as indicated above (Koch, 22), is relatively hypoxic, are significantly increased in T-HIF-1 / mice (Figures S6A S6D). We next asked whether a physical interaction between Foxp3 and HIF-1a existed. Immunoprecipitation and western blot analysis of 293 T cells cotransfected with plasmids encoding HIF-1a and Foxp3, in fact, demonstrated a physical interaction between these two molecules (Figure S7A). This interaction between HIF-1a and Foxp3 was further confirmed by endogenous co-ip in it reg cells (Figure 5E). The HIF-1a-binding domain was further mapped to the terminus of Foxp3 (Figure S7), and the Foxp3-binding domain in HIF-1a was found to be within the N-terminal 8 amino acids by GST pull-down experiments (Figure S7). ased on these results, we sought to determine whether HIF-1a could regulate Foxp3 protein levels through a mechanism completely distinct from its classic role as a transcriptional regulator. In particular, we hypothesized that HIF-1 might target Foxp3 for degradation. Indeed, upon cotransfection of 293T cells with Foxp3 and increasing amounts of Flag-tagged HIF-1a, Foxp3 protein levels progressively decreased (Figure 5F). On the other hand, RORgt protein levels remained unchanged upon cotransfection with HIF-1a (Figure S7D). It has been well established that ubiquitin-dependent degradation of HIF-1a itself occurs via proline hydroxylation at amino acid positions 42 and 564 by prolyl hydroxylases (PHDs). Hydroxylated HIF-1a is subsequently bound by the von Hippel-Lindau protein (VHL), which recruits the Elongin--Elongin--ullin-2-E3- ubiquitin ligase complex, thereby targeting HIF-1a for degradation by the 26S proteasome (Semenza, 27). We wondered whether HIF-1a s ability to induce Foxp3 protein degradation was dependent on this very mechanism. To address this, we cotransfected 293T cells with Foxp3 and a HIF-1a mutant in which prolines 42 (p42a) and 564 (p564a) were mutated to alanines, making HIF-1a resistant to both hydroxylation by PHDs and subsequent proteasomal degradation. The expression of P42A and P564A HIF-1a failed to induce Foxp3 degradation (Figures 5G and 5H). Similar cotransfection assays using a HIF-1a plasmid containing an oxygen dependent domain (ODD) deletion (A5- HIF-1a) further demonstrated that a nonproline hydroxylatable HIF-1a was incapable of mediating Foxp3 degradation (Figures 5G and 5H). Knockdown of PHD2 with sirna also eliminated the ability of wild-type HIF-1a to mediate the degradation of Foxp3 (Figure S8A). HIF-1a-mediated Foxp3 degradation was completely inhibited in the presence of a proteasome inhibitor (Figure S8). Furthermore, we assessed whether Foxp3 itself could be ubiquitinated under hypoxic conditions in primary it regs. Hypoxic culture significantly enhanced both Foxp3 ubiquitination and degradation in these cells (Figures 5I and 5J). Furthermore, this ubiquitin-mediated Foxp3 degradation process could be prevented by addition of the proteasomal inhibitor MG132 to it reg cells under hypoxia. ollectively, these results are compatible with a scenario in which HIF-1 mediates Foxp3 degradation via PHD-VHL-ubiquitin-mediated proteasomal degradation, although future studies are needed to delineate precisely how HIF-1a protein turnover dynamics in T cells modulate Foxp3 levels. ollectively, these in vitro differentiation studies, alongside our transcriptional and biochemical analyses, suggest a model in which HIF-1a is transcriptionally activated via Stat3 signaling in differentiating T cells, resulting in the enhancement of the T H 17 genetic program via a RORgt/p3-dependent mechanism and repression of the T reg transcriptional program through induced degradation of Foxp3. Mice with HIF-1a-Deficient T ells Fail to Mount a Strong T H 17 Response, Have Increased T reg Numbers, and Are Resistant to EAE Our in vitro findings that HIF-1a regulates the T H 17/T reg balance prompted us to test its role in vivo in the setting of a pathologic T H 17-dependent autoimmune disease. T H 17 cells are the major pathogenic population in experimental autoimmune encephalomyelitis (EAE), a mouse model of multiple sclerosis, although factors besides IL-17A and IL-17F can contribute to the disease. Immunization of 6 mice with myelin oligodendrocyte glycoprotein peptide (MOG ) in complete Freund s adjuvant (FA) induces a T H 17-dependent response that induces encephalitis. After a peak of encephalitis, clinically discernible by various gradable neurologic signs (including tail paralysis, progressing to hind and front limb paralysis), the disease remits, concomitant with expansion of T regs that can silence the T H 17 encephalitic process. Strikingly, WT mice immunized with MOG developed EAE, whereas T-HIF-1a / mice were highly resistant (Figure 6A). At peak disease, lymph node-, spleen-, and central nervous system (NS)-infiltrating D4 + T cells from WT mice contained higher proportions of IL-17 + cells compared to those of T-HIF-1a / mice (Figures 6 and 6). In contrast, T-HIF-1 / mice showed a significant increase in the percentage of D4 + Foxp3 + cells during the recovery phase of disease (21 days postinjection), concomitant with a decrease in the proportion of IL-17 + D4 + cells in T-HIF-1a / animals (Figures 6D and 6E). This altered balance between T reg cells and T H 17 cells exactly parallels our in vitro culture studies and suggests that HIF-1a indeed plays a role in modulating the T H 17/T reg balance. Taken together, these results support the notion that HIF-1 plays a role in setting the T H 17/T reg balance in a T H 17-dependent disease model and that the mechanisms defined in the in vitro differentiation assays are also operative in vivo (Figure 7). DISUSSION Here, we identify HIF-1 as a major player in the development of T H 17 cells and further demonstrate its role in modulating the T H 17:T reg balance. Specifically, HIF-1 promotes T H 17 development through: (1) direct transcriptional activation of RORgt, the major T H 17 transcription factor, and (2) direct collaboration with RORgt to activate T H 17 signature genes, such as IL-17A, through mechanisms involving p3 recruitment and histone acetylation. Other transcription factors have been identified to contribute to the induction of IL-17A in T H 17- polarized cells: Runx1 (Zhang et al., 28), ATF (Schraml et al., 29), Stat3, c-maf (auquet et al., 29), AHR (Veldhoen et al., 28), and RORa (Yang et al., 28b). Interestingly, like most of these 78 ell 146, , September 2, 211 ª211 Elsevier Inc.

10 D E Figure 6. Mice Lacking HIF-1a in D4 + T ells Are Deficient in IL-17 Production, Have Increased Numbers of Foxp3 Treg, and Are More Resistant to EAE (A) EAE was induced in HIF-1a +/+ and T-HIF-1a / mice by injection of MOG in FA and Pertussis Toxin. Disease severity was monitored and scored daily. T-HIF-1a / mice failed to develop the severe disease seen in HIF-1a +/+ mice. Mean scores for HIF-1a +/+ and T-HIF-1a / mice over time are represented (± SEM; *p <.5). Shown is a representative of four experiments (n = 7 1 per group). ( and ) During peak disease (day12), draining lymph node, splenic, and NS-infiltrating T cells were recovered and stained for IL-17 and IFNg. The mean percentage of NS D4 + cells during peak disease (days 12 14) that were positive for IL-17 was determined (). (D and E) Similarly, during the recovery phase (day 21), tissue-infiltrating cells were isolated, and the percentage of D4 + that were Foxp3 + was found (D and E). () and (E) present the mean (± SEM; *p <.5) of at least three trials, and dot plots are representative analyses. factors, HIF-1a not only regulates RORgt expression at the mrna level, but also cooperates with RORgt protein to regulate IL-17A-related genes during T H 17 development. HIF-1 may promote the T H 17 response in additional ways. Indeed, studies by Miossec and colleagues indicate that the hypoxia-induced pathway is itself activated by IL-17A and IL-17F (Hot and Miossec, 211), suggesting that IL-17 signaling can sustain the HIF-1 pool and thereby perpetuate an existing T H 17response. oordinate with these transcriptional mechanisms for the promotion of T H 17 development, HIF-1 inhibits differentiation toward the T reg lineage through a distinct, nontranscriptional mechanism. We show that HIF-1 targets Foxp3 for ubiquitination and proteasomal degradation, using the same ubiquitin ligase system that is responsible for degradation of HIF-1a itself (Figure 7). Our findings help to clarify the molecular events involved in lineage differentiation from the precursors of T H 17 and T reg cells into the distinct T cell subsets. In addition, because HIF-1 responds to metabolic cues, our findings link metabolism to T cell differentiation and provide a basis for understanding how the metabolic environment of a T cell can modulate its fate decisions. Our studies on HIF-1 presented here and those of Delgoffe et al. (Delgoffe et al., 29, 211; Delgoffe and Powell, 29) and Shi et al. (Shi et al., 211) on the mtor pathway ell 146, , September 2, 211 ª211 Elsevier Inc. 781

11 Figure 7. A Model for the Multifactorial Role of HIF-1 in Modulating the T H 17/T reg alance Stat3 activation by factors such as IL-6 transcriptionally activates HIF-1. HIF-1 levels are further regulated by oxygen tension and other metabolites, representing a key molecular link between metabolic cues and T cell lineage commitment. HIF-1 directly activates Rorc gene transcription and, furthermore, recruits p3 to RORgt transcription complexes on the promoters of T H 17 genes (i.e., Il17). These activities promote T H 17 differentiation. oncomitantly, HIF-1 induces Foxp3 protein degradation via targeting for ubiquitination and proteasomal degradation. further link metabolic sensing and cytokine signaling in T cell fate determination. HIF-1 is a major sensor of oxygen tension, and indeed, we found that T H 17 differentiation is enhanced under hypoxia in a HIF-1a-dependent manner. Interestingly, a number of studies have demonstrated that inflammatory environments are relatively hypoxic. It is therefore likely that HIF-1a activity represents a major mechanism by which the hypoxic conditions associated with inflammation can promote T H 17 differentiation. It is quite clear, however, that HIF-1a protein is readily detectable in T H 17 cells under normoxia, and the experiments utilizing HIF-1a / T cells confirm that, under such conditions, T H 17 differentiation is significantly impaired without HIF-1a. Even though HIF-1 was originally discovered as a hypoxia sensor, it is now well appreciated that HIF-1a levels can be significantly affected by many other important metabolites such as reactive oxygen species and succinate (Pouysségur and Mechta- Grigoriou, 26). This regulation appears to be at the level of PHD enzyme activity, which in normoxia hydroxylates specific prolines on HIF-1a, targeting it for VHL-dependent ubiquitination and proteosomal degradation. Our findings regarding the role of HIF-1 in targeting Foxp3 for degradation help to explain the unresolved question of how dual T H 17/T reg precursors that express both RORgt and Foxp3 eventually eliminate Foxp3 when they commit to T H 17 differentiation in a STAT3-dependent fashion. Rudensky and colleagues have recently proposed that conserved noncoding sequence 2 (NS2) in the Foxp3 gene represents a site for positive transcriptional autoregulation of Foxp3, providing a mechanism for the protein to maintain its own expression levels (Zheng et al., 21). Thus, degradation of Foxp3 in T H 17/T reg progenitors mediated by HIF-1 would be expected to eventually diminish Foxp3 gene transcriptional activity, thereby antagonizing stable Foxp3 autoregulation and consequent T reg lineage commitment. Sitkovsky and colleagues have suggested that HIF-1 activity in T cells is a general negative regulator of T cell activation (Lukashev et al., 26). These experiments did not evaluate specific T cell differentiation decisions, however. We found that neither IFN-g nor IL-4 production was significantly affected by the absence of HIF-1 in T cells cultured under Th1 or Th2 conditions, respectively. ecause T H 17 cytokines can inhibit Th1/IFN-g development in some circumstances (Luger et al., 28), it is possible that the enhanced IFN-g production seen in HIF-1a / T cells under the nonskewing in vitro conditions used by Lukashev et al. is, in fact, a reflection of diminished T H 17 development. Our in vivo findings demonstrate that mice with HIF-1a / T cells are resistant to induction of EAE, a T H 17-dependent disease whose remission is dependent upon T reg cells (Schraml et al., 29). These experiments validate the role of HIF-1a identified by our in vitro studies and further suggest the possibility of metabolically modulating autoimmune diseases. Indeed, there is great interest in developing HIF-1a inhibitors based on its role in cancer. It would be worthwhile to determine whether any of 782 ell 146, , September 2, 211 ª211 Elsevier Inc.

12 these inhibitors can alter the T H 17:T reg balance in vivo and ameliorate T H 17-dependent autoimmune diseases. EXPERIMENTAL PROEDURES Mice All animal experiments were performed in specific pathogen-free facilities in the Johns Hopkins Animal Resource enter following national, state, and institutional guidelines. Animal protocols were approved by the Johns Hopkins Animal are and Use ommittee. Foxp3-GFP mice (Fontenot et al., 25) were kindly provided by A. Rudensky. 57L/6 HIF-1a fl/fl mice were produced in Dr. Semenza s lab. Rorc / mice were purchased from the Jackson Laboratory. T ell Differentiation Naive T cells were purified using a FAS Aria sorter prior to stimulation with anti-d3 /D28 antibodies in a 24-well plate (1 and 4 mg/well, respectively; iolegend) for 3 7 days. T H 17-skewing conditions consisted of IMDM media supplemented with 5% FS, 2 ng/ml IL-6, 2.5 ng/ml TGFb (Peprotech), and 1 ug/ml neutralizing antibodies against IFNg, IL-4, and IL-12 (iolegend). For hypoxia experiments, cells were cultured in a GasPak Plus anaerobic chamber (1% O 2 ) for 2 hr cycles interrupted by normoxic rest. Quantitative RT-PR RNA was isolated by a minirna extraction kit (QIAGEN). The cdna archival kit (Applied iosystem) was used as per the manufacturer s instruction. Triplicate reactions were run using an AI Prism 75. mrna levels were determined by comparative T method and normalized to b-actin or 18 s rrna expression. EAE Induction Six- to eight-week-old, sex-matched WT and T-HIF-1a / littermates were injected s.c in the rear flank with 1 mg MOG peptide (2HN- MEVGWYRSPFSRVVHLYRNGK-OOH) in complete Freund s adjuvant (Sigma), and 25 ng pertussis toxin (List iological) was injected i.p. Mice were monitored daily, and disease severity was scored. Immunoprecipitation and Western lotting Immunoprecipitation and western blotting were performed as elsewhere (Pan et al., 25). Immunoprecipitations were done using a Pierce rosslink IP kit and clean-lot IP detection system (Thermo Scientific). Reporter Gene Assays The luciferase reporter gene assay reagents were obtained from Promega, and the assay was performed per manufacturer s instructions. hromatin Immunoprecipitation Assays hip analysis was carried out according to the manufacturer s Millipore, Massachusetts, USA). riefly, cells (1-/Millipore, Massachusetts, USA). The amount of immunoprecipitated DNA was quantified by real-time PR with the AI PRISM 75 Sequence Detection System (Applied iosystems) using SYR Green. All primers used for hip assays are listed in Table S1. GST Pull-Down ell lysate or purified protein was incubated with 3 mg of affinity-purified GST protein (with a GSTrap column, Amersham) in the presence of.2% SA in 1 uffer on a rotator overnight at 4. Proteins were pulled down using GST beads, followed by washing, glutathione elution, and resolution by SDS-PAGE. Statistical Analysis An unpaired Student s t test was used to determine significance (*p <.5). SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, eight figures, and one table and can be found with this article online at doi:1.116/ j.cell AKNOWLEDGMENTS We thank the members of the D. Pardoll, F. Pan, J. Powell, and. Drake laboratories for helpful discussions. We are grateful to Drs. J. Liu (Department of Pharmacology at Johns Hopkins), A. Rudensky (Memorial Sloan Kettering ancer enter), and D. Littman (NYU) for helpful suggestions and/or reagent contribution. We thank Dr. H. Wei (Dr. Semenza s laboratory) for the HIF1 fl/fl mice and genotyping primer sequences. This work was supported by grants from NIH and the Melanoma Research Alliance, the Janey Fund and Seraph Foundation, and gifts from ill and etty Topecer and Dorothy Needle. F.P. is a recipient of the Stewart Trust Scholar Award. Received: September 13, 21 Revised: June 22, 211 Accepted: July 25, 211 Published online: August 25, 211 REFERENES Ahern, P.P., Schiering,., uonocore, S., McGeachy, M.J., ua, D.J., Maloy, K.J., and Powrie, F. (21). Interleukin-23 drives intestinal inflammation through direct activity on T cells. Immunity 33, Arany, Z., Huang, L.E., Eckner, R., hattacharya, S., Jiang,., Goldberg, M.A., unn, H.F., and Livingston, D.M. (1996). An essential role for p3/p in the cellular response to hypoxia. Proc. Natl. Acad. Sci. USA 93, arnes, M.J., and Powrie, F. (29). Regulatory T cells reinforce intestinal homeostasis. Immunity 31, auquet, A.T., Jin, H., Paterson, A.M., Mitsdoerffer, M., Ho, I.., Sharpe, A.H., and Kuchroo, V.K. (29). The costimulatory molecule IOS regulates the expression of c-maf and IL-21 in the development of follicular T helper cells and TH-17 cells. Nat. Immunol. 1, ettelli, E., arrier, Y., Gao, W., Korn, T., Strom, T.., Oukka, M., Weiner, H.L., and Kuchroo, V.K. (26). Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, hen, W., Jin, W., Hardegen, N., Lei, K.J., Li, L., Marinos, N., McGrady, G., and Wahl, S.M. (23). onversion of peripheral D4+D25- naive T cells to D4+D25+ regulatory T cells by TGF-beta induction of transcription factor Foxp3. J. Exp. Med. 198, Davidson, T.S., DiPaolo, R.J., Andersson, J., and Shevach, E.M. (27). utting Edge: IL-2 is essential for TGF-beta-mediated induction of Foxp3+ T regulatory cells. J. Immunol. 178, Delgoffe, G.M., and Powell, J.D. (29). mtor: taking cues from the immune microenvironment. Immunology 127, Delgoffe, G.M., Kole, T.P., Zheng, Y., Zarek, P.E., Matthews, K.L., Xiao,., Worley, P.F., Kozma, S.., and Powell, J.D. (29). The mtor kinase differentially regulates effector and regulatory T cell lineage commitment. Immunity 3, Delgoffe, G.M., Pollizzi, K.N., Waickman, A.T., Heikamp, E., Meyers, D.J., Horton, M.R., Xiao,., Worley, P.F., and Powell, J.D. (211). The kinase mtor regulates the differentiation of helper T cells through the selective activation of signaling by mtor1 and mtor2. Nat. Immunol. 12, Dong,. (28). TH17 cells in development: an updated view of their molecular identity and genetic programming. Nat. Rev. Immunol. 8, Durant, L., Watford, W.T., Ramos, H.L., Laurence, A., Vahedi, G., Wei, L., Takahashi, H., Sun, H.W., Kanno, Y., Powrie, F., and O Shea, J.J. (21). Diverse targets of the transcription factor STAT3 contribute to T cell pathogenicity and homeostasis. Immunity 32, Fife,.T., Pauken, K.E., Eagar, T.N., Obu, T., Wu, J., Tang, Q., Azuma, M., Krummel, M.F., and luestone, J.A. (29). Interactions between PD-1 and PD-L1 promote tolerance by blocking the TR-induced stop signal. Nat. Immunol. 1, Fontenot, J.D., Rasmussen, J.P., Williams, L.M., Dooley, J.L., Farr, A.G., and Rudensky, A.Y. (25). Regulatory T cell lineage specification by the forkhead transcription factor foxp3. Immunity 22, ell 146, , September 2, 211 ª211 Elsevier Inc. 783

13 Getnet, D., Grosso, J.F., Goldberg, M.V., Harris, T.J., Yen, H.R., runo, T.., Durham, N.M., Hipkiss, E.L., Pyle, K.J., Wada, S., et al. (21). A role for the transcription factor Helios in human D4(+)D25(+) regulatory T cells. Mol. Immunol. 47, Harris, T.J., Grosso, J.F., Yen, H.R., Xin, H., Kortylewski, M., Albesiano, E., Hipkiss, E.L., Getnet, D., Goldberg, M.V., Maris,.H., et al. (27). utting edge: An in vivo requirement for STAT3 signaling in TH17 development and TH17-dependent autoimmunity. J. Immunol. 179, Hot, A., and Miossec, P. (211). Effects of interleukin (IL)-17A and IL-17F in human rheumatoid arthritis synoviocytes. Ann. Rheum. Dis. 7, Ivanov, I.I., McKenzie,.S., Zhou, L., Tadokoro,.E., Lepelley, A., Lafaille, J.J., ua, D.J., and Littman, D.R. (26). The orphan nuclear receptor RORgammat directs the differentiation program of proinflammatory IL-17+ T helper cells. ell 126, Koch,.J. (22). Measurement of absolute oxygen levels in cells and tissues using oxygen sensors and 2-nitroimidazole EF5. Methods Enzymol. 352, Korn, T., ettelli, E., Oukka, M., and Kuchroo, V.K. (29). IL-17 and Th17 ells. Annu. Rev. Immunol. 27, Littman, D.R., and Rudensky, A.Y. (21). Th17 and regulatory T cells in mediating and restraining inflammation. ell 14, Luger, D., Silver, P.., Tang, J., ua, D., hen, Z., Iwakura, Y., owman, E.P., Sgambellone, N.M., han,.., and aspi, R.R. (28). Either a Th17 or a Th1 effector response can drive autoimmunity: conditions of disease induction affect dominant effector category. J. Exp. Med. 25, Lukashev, D., Klebanov,., Kojima, H., Grinberg, A., Ohta, A., erenfeld, L., Wenger, R.H., Ohta, A., and Sitkovsky, M. (26). utting edge: hypoxiainducible factor 1alpha and its activation-inducible short isoform I.1 negatively regulate functions of D4+ and D8+ T lymphocytes. J. Immunol. 177, Mangan, P.R., Harrington, L.E., O Quinn, D.., Helms, W.S., ullard, D.., Elson,.O., Hatton, R.D., Wahl, S.M., Schoeb, T.R., and Weaver,.T. (26). Transforming growth factor-beta induces development of the T(H)17 lineage. Nature 441, Murphy, K.M., and Reiner, S.L. (22). The lineage decisions of helper T cells. Nat. Rev. Immunol. 2, O Quinn, D.., Palmer, M.T., Lee, Y.K., and Weaver,.T. (28). Emergence of the Th17 pathway and its role in host defense. Adv. Immunol. 99, Pan, F., Means, A.R., and Liu, J.O. (25). almodulin-dependent protein kinase IV regulates nuclear export of abin1 during T-cell activation. EMO J. 24, Pan, F., Yu, H., Dang, E.V., arbi, J., Pan, X., Grosso, J.F., Jinasena, D., Sharma, S.M., Mcadden, E.M., Getnet, D., et al. (29). Eos mediates Foxp3-dependent gene silencing in D4+ regulatory T cells. Science 325, Pouysségur, J., and Mechta-Grigoriou, F. (26). Redox regulation of the hypoxia-inducible factor. iol. hem. 387, Rajewsky, K., and von oehmer, H. (28). Lymphocyte development: overview. urr. Opin. Immunol. 2, Sakaguchi, S., Yamaguchi, T., Nomura, T., and Ono, M. (28). Regulatory T cells and immune tolerance. ell 133, Schraml,.U., Hildner, K., Ise, W., Lee, W.L., Smith, W.A., Solomon,., Sahota, G., Sim, J., Mukasa, R., emerski, S., et al. (29). The AP-1 transcription factor atf controls T(H)17 differentiation. Nature 46, Semenza, G.L. (27). Hypoxia-inducible factor 1 (HIF-1) pathway. Sci. STKE 27, cm8. Shi, L.Z., Wang, R., Huang, G., Vogel, P., Neale, G., Green, D.R., and hi, H. (211). HIF1alpha-dependent glycolytic pathway orchestrates a metabolic checkpoint for the differentiation of TH17 and Treg cells. J. Exp. Med. 28, Thompson, P.R., Wang, D., Wang, L., Fulco, M., Pediconi, N., Zhang, D., An, W., Ge, Q., Roeder, R.G., Wong, J., et al. (24). Regulation of the p3 HAT domain via a novel activation loop. Nat. Struct. Mol. iol. 11, Veldhoen, M., Hocking, R.J., Atkins,.J., Locksley, R.M., and Stockinger,. (26). TGFbeta in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 24, Veldhoen, M., Hirota, K., Westendorf, A.M., uer, J., Dumoutier, L., Renauld, J.., and Stockinger,. (28). The aryl hydrocarbon receptor links TH17- cell-mediated autoimmunity to environmental toxins. Nature 453, Weaver,.T., Hatton, R.D., Mangan, P.R., and Harrington, L.E. (27). IL-17 family cytokines and the expanding diversity of effector T cell lineages. Annu. Rev. Immunol. 25, Wu, S., Rhee, K.J., Albesiano, E., Rabizadeh, S., Wu, X., Yen, H.R., Huso, D.L., rancati, F.L., Wick, E., McAllister, F., et al. (29). A human colonic commensal promotes colon tumorigenesis via activation of T helper type 17 T cell responses. Nat. Med. 15, Xu, Q., riggs, J., Park, S., Niu, G., Kortylewski, M., Zhang, S., Gritsko, T., Turkson, J., Kay, H., Semenza, G.L., et al. (25). Targeting Stat3 blocks both HIF-1 and VEGF expression induced by multiple oncogenic growth signaling pathways. Oncogene 24, Yang, X.O., Panopoulos, A.D., Nurieva, R., hang, S.H., Wang, D., Watowich, S.S., and Dong,. (27). STAT3 regulates cytokine-mediated generation of inflammatory helper T cells. J. iol. hem. 282, Yang, X.O., Nurieva, R., Martinez, G.J., Kang, H.S., hung, Y., Pappu,.P., Shah,., hang, S.H., Schluns, K.S., Watowich, S.S., et al. (28a). Molecular antagonism and plasticity of regulatory and inflammatory T cell programs. Immunity 29, Yang, X.O., Pappu,.P., Nurieva, R., Akimzhanov, A., Kang, H.S., hung, Y., Ma, L., Shah,., Panopoulos, A.D., Schluns, K.S., et al. (28b). T helper 17 lineage differentiation is programmed by orphan nuclear receptors ROR alpha and ROR gamma. Immunity 28, Youn, H.D., hatila, T.A., and Liu, J.O. (2). Integration of calcineurin and MEF2 signals by the coactivator p3 during T-cell apoptosis. EMO J. 19, Zhang, H., Qian, D.Z., Tan, Y.S., Lee, K., Gao, P., Ren, Y.R., Rey, S., Hammers, H., hang, D., Pili, R., et al. (28). Digoxin and other cardiac glycosides inhibit HIF-1alpha synthesis and block tumor growth. Proc. Natl. Acad. Sci. USA 15, Zheng, Y., Josefowicz, S., haudhry, A., Peng, X.P., Forbush, K., and Rudensky, A.Y. (21). Role of conserved non-coding DNA elements in the Foxp3 gene in regulatory T-cell fate. Nature 463, Zhou, L., Lopes, J.E., hong, M.M., Ivanov, I.I., Min, R., Victora, G.D., Shen, Y., Du, J., Rubtsov, Y.P., Rudensky, A.Y., et al. (28). TGF-beta-induced Foxp3 inhibits T(H)17 cell differentiation by antagonizing RORgammat function. Nature 453, Zhu, J., Yamane, H., and Paul, W.E. (21). Differentiation of effector D4 T cell populations (*). Annu. Rev. Immunol. 28, ell 146, , September 2, 211 ª211 Elsevier Inc.

14 Supplemental Information EXTENDED EXPERIMENTAL PROEDURES Mice The HIF-1a fl/fl mouse line was bred to 57L/6 mice carrying a D4 cre transgene. In the resulting offspring a region encompassing exon 2 was excised in D4 + cells. Genotyping on tail DNA was performed as previously described (Harris et al., 27). An amplification of 25 bp by primers DP11 (5 -GAGTTAAGAGATAGTTG) and DP12 (5 -GGAGTATTTTAGA) indicated the presence of a floxed HIF-1 allele. An amplification of 2 bp indicated a WT HIF-1a allele. PR was used to genotype tail DNA for the presence of D4 cre (forward 5 -GATGAAGAGTGATGAGG, reverse 5 GATAAAGTGAAAAG). T ell Differentiation D4 + D62L high D25 (naive) T cells were purified using a FAS Aria high speed sorter. These cells were stimulated for three-toseven days in 24 well plates containing immobilized anti-d3e and soluble anti-d28 antibodies (1 and 4 mg per well, respectively; iolegend). T H 17 skewing conditions consisted of IMDM culture media supplemented with 5% FS, 2ng/ml IL-6, 2.5ng/ml TGFb (Peprotech), and 1ug/ml cytokine neutralizing antibodies directed against IFNg, IL-4, and IL-12 (all from iolegend).at indicated time periods, cells were removed from culture, restimulated with PMA/ionomycin in the presence of Golgi-stop for 5hrs and intracellular cytokine staining (IS) analysis was performed to assay production of IL-17 or expression of Foxp3. For some experiments, Ovaspecific T cells (obtained from an OTII transgenic donor) were stimulated under T H 17 skewing in vitro culture conditions by coincubation with splenocyte APs pulsed with their cognate antigen Ova. Hypoxic hamber ell Growth Studies For hypoxia experiments, 2ml of FAS-sorted T cells (1x1 6 cells/ml) were plated on 6-well culture dishes (ecton Dickinson, NJ) under T H 17-skewing conditions. The culture medium was changed immediately before placing the dishes in a GasPak Plus anaerobic culture chamber containing hydrogen and a palladium catalyst (GasPak Plus Hydrogen O2 generators, ecton Dickinson) to remove all traces of oxygen. ells were incubated under normoxic conditions in between approximately 2hr-long hypoxic culture cycles (1% O 2, 99% N 2 ) each day during the course of T reg /T H 17-skewing in vitro. EAE Induction Six-eight week old, sex matched wild-type and HIF-1a / littermates were injected s.c in the rear flank with 1mg MOG peptide (2HN-MEVGWYRSPFSRVVHLYRNGK-OOH) in complete Freund s Adjuvant (Sigma). On days and 2 post-injection, mice also received 25ng Pertussis Toxin (List iological) injected i.p. in 2ul PS. Mice were monitored daily and disease severity was scored as follows: = no disease, 1 = flaccid tail, 2 = hind limb weakness, 3 = hind limb paralysis, 4 = forelimb weakness, 4.5 = hind and forelimb paralysis, 5 = dead or moribund (Korn et al., 28). At specified time points during disease progression, four mice per group were sacrificed and cells of the injection site draining lymph nodes and spleen were recovered for flow cytometric analysis along with those of the NS which were isolated by enzymatic digestion and Percoll gradient centrifugation as described elsewhere (Wu et al., 29). For IS, recovered cells were treated as described above prior to staining of surface D4 and D3, cellular fixation, permeabilization and eventual staining with flourochrome labeled anti-cytokine or isotype control IgG. Quantitative Real-Time PR Total RNA was isolated from whole cells using the QIAGEN minirna extraction kit following the manufacturer s instructions. RNA was quantified spectrophotometrically, and complementary DNA was reverse-transcribed using the cdna archival kit (Applied iosystem) following the manufacturer s instruction. The cdna samples were distributed on plates at 2ng/well and run in triplicate. PR reactions were set up in 25-ml volumes using TaqMan Universal PR Master Mix (Applied iosystem) on an AI Prism 75 Sequence Detection System. Quantification of relative mrna expression was determined by the comparative T (critical threshold) method where the amount of target mrna, normalized to endogenous b-actin or 18 s rrna expression, is determined by the formula 2-DT. Primers used to quantify HIF-1a mrna were 5 -GGGAAGAAAGAGTTG-3 (forward) and 5 -ATAATGATGGT GAGTATAA-3 (reverse). Reporter Gene Assays The luciferase reporter gene assay reagents were obtained from Promega, and the assay was performed per manufacturer s instructions. riefly, transient transfections were performed using electroporation. Jurkat cells ( ) were harvested, washed once in RPMI medium, and mixed with 2 mg of luciferase reporter plasmid and different amount of expression plasmids as indicated in 3 ml of RPMI in a sterile cuvette. An electric pulse (25 V, 96 mf) was applied (io-rad Gene Pulser II). Where appropriate, the transfection efficiency was measured by cotransfecting 1 mg of Renilla luciferase expression vector and measuring enzymatic Renilla luciferase activity. hromatin Immunoprecipitation Assays hip analysis was carried out according to the manufacturer s protocol (Upstate/Millipore, Massachusetts, USA). riefly, cells (1-5x1 6 ) were fixed with 1% formaldehyde, followed by micrococcal nuclease digestions, and chromatin was precleared by ell 146, , September 2, 211 ª211 Elsevier Inc. S1

15 incubation with Protein A/G agarose/salmon sperm DNA (Upstate/Millipore). Subsequently, chromatin was immunoprecipitated by overnight incubation at 4 with 4 mg antibodies (rabbit isotype, anti-rorgt and anti-p3, Santa ruz; anti-hif-1a, Novus iologicals; anti-acetyl-histone H3, H4 Upstate/Millipore, USA) followed by incubation with protein A/G agarose/salmon sperm DNA for 1h. Precipitates were defixed and DNA was purified. The amount of immunoprecipitated DNA was quantified by real-time PR with the AI PRISM 75 Sequence Detection System (Applied iosystems) using SYR Green. Retroviral Transduction D4 + naive T cells were isolated from mouse spleens and lymph nodes, stimulated with anti-d3 and D28 for 24 hr, and spin-infected with retrovirus containing supernatant from Phoenix packaging cells transfected with retroviral expression plasmids (KMV IRES-GFP, empty or encoding RORgt or HIF-1a) (Pan et al., 25; Schraml et al., 29). Immunoprecipitation and Western lotting Immunoprecipitation and Western blotting were performed as described previously (Pan et al., 25). Immunoprecipitation experiments were carried out by using a Pierce rosslink IP kit and clean-lot IP detection system (Thermo Scientific). GST Pull-Down RORgt or Foxp3 expressing Jurkat T or 293T cell lysate (or purified protein from E.coli) was incubated with 3 mg of the affinity purified GST protein (with a GSTrap column, Amersham) as indicated in the presence of.2%sa in 1 (2mM Tris ph7.9, 1mM Nal, 1% glycerol,.2mm EDTA,.1% Triton X-1) on a rotator overnight at 4. The proteins were pulled down using GST beads, followed by five washes with 1 before elution with 5 ml of 1 plus 2mM reduced glutathione for 1 min with gentle rotation. Eluted materials were resolved by SDS-PAGE. SUPPLEMENTAL REFERENES Harris, T.J., Grosso, J.F., Yen, H.R., Xin, H., Kortylewski, M., Albesiano, E., Hipkiss, E.L., Getnet, D., Goldberg, M.V., Maris,.H., et al. (27). utting edge: An in vivo requirement for STAT3 signaling in TH17 development and TH17-dependent autoimmunity. J. Immunol. 179, Korn, T., Mitsdoerffer, M., roxford, A.L., Awasthi, A., Dardalhon, V.A., Galileos, G., Vollmar, P., Stritesky, G.L., Kaplan, M.H., Waisman, A., et al. (28). IL-6 controls Th17 immunity in vivo by inhibiting the conversion of conventional T cells into Foxp3+ regulatory T cells. Proc. Natl. Acad. Sci. USA 15, Pan, F., Means, A.R., and Liu, J.O. (25). almodulin-dependent protein kinase IV regulates nuclear export of abin1 during T-cell activation. EMO J. 24, Schraml,.U., Hildner, K., Ise, W., Lee, W.L., Smith, W.A., Solomon,., Sahota, G., Sim, J., Mukasa, R., emerski, S., et al. (29). The AP-1 transcription factor atf controls T(H)17 differentiation. Nature 46, Wu, S., Rhee, K.J., Albesiano, E., Rabizadeh, S., Wu, X., Yen, H.R., Huso, D.L., rancati, F.L., Wick, E., McAllister, F., et al. (29). A human colonic commensal promotes colon tumorigenesis via activation of T helper type 17 T cell responses. Nat. Med. 15, S2 ell 146, , September 2, 211 ª211 Elsevier Inc.

16 HIF-1α +/+ HIF-1α -/- A LN SPL THY LN SPL THY FL3-H D8 FL3-H FL3-H FL3-H FL3-H FL3-H D HIF-1α +/+ HIF-1α -/- LN SPL THY LN SPL THY FL2-H FL2-H FL2-H FL2-H FL2-H Foxp3 FL2-H 63.3 D HIF-1α +/ HIF-1α-/- / FL1-H H: 22 FIT LN : D11 D11c FL1- -H: 22 FIT SPL : D11 L1-H: 22 FIT FL LN : D11 FL L1-H: 22 FIT SPL : D FL2-H: F4/8 FL2-H: F4/8 F4/8 FL2-H: F4/8 FL2-H: F4/ : D : D : D : D11.91 D11c Figure S1. The Development of T,, and Dentritic ells in T-HIF-1a / Mice Is Normal, Related to Figure 1 WT and T-HIF-1a / mice aged 5-8 weeks were sacrificed and the indicated tissues were harvested and the leukocytes isolated from them were stained for various immune cell population markers. (A and ) oth WT and T-HIF-1a / mice have similar percentages of D4+ and D8+ T cells in the lymph node (LN), spleen (SPL) and thymus (THY) tissues. () HIF-1a / mice possess WT levels of 22+ cells (top row) and D11c+ dendritic cells (bottom row). All panels depict representative findings from two-to-three independent experiments utilizing at least three organs per group. Numbers represent the percentage of viable cells. ell 146, , September 2, 211 ª211 Elsevier Inc. S3

17 T H 1 T H 2 HIF-1α +/+ HIF-1α -/- HIF-1α +/+ HIF-1α -/ HIF-1α +/+ HIF-1α -/ # ells 79.1 # ells IFN-γ IL FL1-H: FSE FL1-H: FSE ROR Rγt mrna expression D hrs anti-rorγt E IL-17+ cells % HIF-1α +/+ HIF-1α -/- hrs anti-tubulin 5 Normoxia Hypoxia Figure S2. Further haracterization of Wild-Type and T-HIF-1a / T ells, Related to Figure 2 (A) Naive D4+D62L hi T cells isolated from WT or T-HIF-1a / mice were activated and cultured under Th1 and Th2 skewing conditions for 4 days, followed by re-stimulation with PMA and ionomycin for 5 hr with GolgiStop, followed by IS staining. Numbers represent the percentage of D4+ cells positive for the indicated marker. () Analysis of T cell proliferation in wild-type and T-HIF-1a / mice. Sorted D4+D25-D62L high naive T cells were incubated with 2.5 um FSE for 7 min at 37, and excess dye was washed away. 5x 1 4 FSE-labeled naive T cells were then mixed with 5x 1 5 irradiated APs and were cocultured for 6 days with anti- D3 stimulation (2 mg/ml). T cell proliferation was measured by flow cytometry and the percentage of FSE low T cells is shown. Shown are representative histograms from at least two independent experiments. ( and D) Naive D4+ T cells were activated with anti-d3 plus anti-d28 under either neutral (anti-il-4 plus anti-ifng) or Th17-skewing conditions before being transduced with either a bicistronic retrovirus expressing HIF-1a-GFP or GFP alone. GFP+ cells were sorted out at the indicated time points and were subjected to qrt-pr () or Western blotting with the indicated antibodies (D). (E) Naive D4+D62L high T cells isolated from WT or T-HIF-1a / mice were activated with anti-d3/d28 in combination with TGFb, IL-6 and anti-ifng, anti-il- 12 and IL-4 antibodies for 6 days under normoxia or hypoxia as described in the Methods section. ells were re-stimulated with PMA and ionomycin for 5 hr with GolgiStop, followed by IS staining. () and (E) depict the mean + SD of three different experiments. S4 ell 146, , September 2, 211 ª211 Elsevier Inc.

18 Alignment of partial human and mouse RORγt promoter sequence Luc hrorγt Luc wt :HIF-1 response element Luc Luc hrorγt Luc Mut : Mutated HIF-1 response element mrorγt Luc wt :HIF-1 response element D Relative Lu uciferase value Luc mrorγt Luc Mut : Mutated HIF-1 response element mock PMA/Iono EV (μg) Wt-HIF-1 (μg) 7 15 HIF-1-mut ΔDD (μg) 7 15 HIF-1onsensus-binding Site: RGTG ative Luciferase Value Rela 5 45 mock prk5(μg) Flag-HIF-1α(μg) PMA/Iono E Re elative Luciferase value 1 emock 9 8 PMA/Iono Wt-HIF-1 (μg) 1 15 RORγt-Luc Wt 1 15 RORγt-Luc Mut Figure S3. HIF-1 inds and Transactivates the RORgt Promoter in oth Murine and Human ontexts, Related to Figure 3 (A) A sequence alignment of the partial mouse (mrorgt) and human (hrorgt) promoter region reveals two human HIF-1-consensus binding sites (highlighted in yellow) and one murine HIF-1 binding site (highlighted in green). The conserved site is underlined. () This schematic depicts a series of reporter constructs containing the HIF-1 response element(s) at the human and mouse RORgt promoter. () Jurkat T cells were transfected with an RORgt-luciferase reporter plasmid (incorporating the 1.1kb RORgt promoter) along with increasing amounts of prk5- HIF-1. Empty vector prk5 was used to maintain equal DNA content for all transfections. 12h post-transfection, cells were either treated with PMA plus ionomycin or were left untreated prior to analysis of luciferase activity, which was normalized to the Rellina luciferase activity. (D) Jurkat T cells were transfected with an hrorgt-luciferase reporter plasmid (incorporating the 1.1kb RORgt promoter) along with varying amounts of WT HIF-1 or a HIF-1 mutant with a deleted DD domain. A luciferase assay was carried out as in (). (E) Jurkat T cells were transfected with the indicated amount of WT HIF-1 together with a luciferase reporter under the control of either a wild-type hrorgt promoter or one with a mutated HIF-1-binding site. A luciferase assay was carried out as described above. Data are representative of at least three independent experiments (mean and s.d. of triplicate transfections). ell 146, , September 2, 211 ª211 Elsevier Inc. S5

19 Non-HIF-1 binding region to to -72 Human vs. Mouse -2kb -4kb -6kb -8kb 2.5 RORγt promoter to -428 RORγt promoter to D RORγt promoter non-hif-1 binding region ive binding Relat ontrol IgG Anti-HIF-1 elative binding Re ontrol IgG Anti-HIF-1 elative binding Re ontrol IgG Anti-HIF-1 HIF1+/+ HIF1+/+ HIF1+/+ HIF1-/- HIF1-/- HIF1-/- E Flag-HIF-1α Myc-RoRgt F GST RORγt IP: Flag W: RORgt GST-HIF-1α W: Flag GST Input:RORγt Figure S4. HIF-1 Is Recruited to the RORgt Promoter Region in T H 17 ells, Related to Figure 4 (A) Shown is a Vista blot depicting the sequence conservation of the human and mouse RORgt loci. The locations of primers used for hip analysis are indicated. ( D) Naive D4+ T cells were sorted from HIF-1 / or T-HIF-1 / mice and were activated with anti-d3/d28 under Th17 conditions (TGF-b/IL-6) for 3 days before being subjected to hip analysis using anti-hif-1 polyclonal antibody. A PR using primers flanking the indicated region was carried out. Data are presented as relative binding normalized to unprecipitated input DNA (shown is the mean+s.d. of 3 independent experiments). (E) Jurkat T cells were transfected with Flag-tagged HIF-1 along with either empty or myc-tagged RORgt plasmid. 48h post-transfection, cells were harvested and lysed in RIPA buffer. The cleared lysate was immunoprecipitated with anti-flag antibody, followed by Western blotting using anti-rorgt antibodies to detect immunoprecipitated protein. (F) RORgt interacts with the N-terminus of HIF-1. RORgt-expressing cell lysate was incubated with different fragments of GST-HIF-1 purified from E.coli as indicated, followed by pull-down with GST beads, resolved by SDS-PAGE and Westen blotting with anti-rorgt (top and bottom) antibodies, or anti-gst (middle). Depicted are representative panels from three independent experiments. S6 ell 146, , September 2, 211 ª211 Elsevier Inc.

20 NS to non-rorγt binding to -614 Human vs Mouse -2kb -4kb -6kb -8kb Relative binding IL17a promoter to ontrol IgG Anti-HIF-1 Relative binding NS to ontrol IgG Anti-HIF-1 D Relative binding Outside NS2 non-rorγt binding region ontrol IgG Anti-HIF-1 Th Th17 Th Th17 Th Th17 Th Th17 Th Th17 Th Th17 Th Th17 Th Th17 Th Th17 WT Rorc-/- HIF-1-/- WT Rorc-/- HIF-1-/- WT Rorc-/- HIF-1-/- E erase activity Relative lucife mock PMA/Iono F se activity Relative luciferas Mock PMA/Iono Figure S5. Recruitment of HIF-1 to the IL17A Promoter Region in T H 17 ells Occurs in a RORgt-Dependent Manner, Related to Figure 4 (A) Shown is a Vista blot depicting the sequence conservation of the human and mouse RORgt loci. The locations of primers used for hip analysis are indicated. ( D) Naive D4+ T cells were sorted from WT, Rorc / or T-HIF-1 / mice and were activated with anti-d3/d28 under Th17 conditions (TGF-b/IL-6) for 3 days prior to hip analysis using anti-hif-1 polyclonal antibody, and PR using primers flanking the region as indicated. Data are presented as relative binding based on unprecipitated input DNA (mean+s.d. of 3 independent experiments). (E and F) Jurkat T cells (E) or 293T cells (F) were transfected with a IL-17 promoter-driven luciferase reporter along with the indicated plasmids, followed by stimulation and assessment as described in Figure 3. HIF-1Mut refers to a HIF-1 mutant gene with a deletion in the p3 binding domain and p3dhat refers to a p3 deletion mutant in which histone acetyl transferase activity is eliminated. Data are representative of at least three independent experiments (mean and s.d. of triplicate transfections). ell 146, , September 2, 211 ª211 Elsevier Inc. S7

21 FL2-H LP HIF-1α +/+ HIF-1α -/- FL2-H % of D4+ cells recover red LP Foxp3 HIF1+/+ HIF1-/ Foxp FL3-H FL3-H D4 LP D FL2-H 13 HIF-1α +/+ HIF-1α -/ IL IFNγ FL2-H covered % of D4+ cells rec LP IL-17 HIF1+/+ HIF1-/- E Enrichment (fold) IgG HIF-1α +/+ HIF-1α -/- H3(K9) H4(K8) IgG H3(K9) H4(K8) IL-17F actin F Enrichme ent (fold) IgG HIF-1α +/+ HIF-1α -/- H3(K9) H4(K8) IgG H3(K9) H4(K8) IL-23R actin Figure S6. HIF-1a Deficiency Results in an Imbalance of T reg and T H 17 ells in the Steady-State Lamina Propria, Related to Figure 4 (A D)The colons from sex matched WT and T-HIF-1 / littermates were excised and the lamina propria (LP) residing leukocytes were isolated by Liberase digestion and percoll centrifugation. ells were stained for surface D3/D4 and intracellular Foxp3 (A and ). or (following restimulation) intracellular IL-17 and IFNg ( and D). Shown are representative density plots of three to five independent experiments (n = 3-5 per group). ar graphs in () and (D) depict the mean percentage of recovered D4+ T cells expressing either Foxp3 or IL-17 +/ SEM. Differences were found significant by an unpaired student s t test (P values =.33 and.52, respectively). (E F) Histone hyperacetylation at the IL-17F (E) and IL-23R (F) promoter was detected by hip assay. WT and HIF-1 / T cells were cultured under Th17 skewing conditions for 4 days prior to being subjected to the hip assay. Shown are the mean (+ SD) of at least three independent experiments. S8 ell 146, , September 2, 211 ª211 Elsevier Inc.

22 IP: anti-flag IP: anti-flag GST GST-HIF-1α Flag-HIF1α Foxp3 + _ + _ EV + _ Myc-Foxp3 Flag-HIF-1α FL N H H Foxp3 Foxp3 L W: W: anti-myc anti-gst D anti-rorγt Input W: Foxp3 anti-hif-1α anti-tubulin RORγt+Wt-Ubi HIF-1α Figure S7. HIF-1 Mediates the Degradation of Foxp3, Related to Figure 5 (A and ) terminus of Foxp3 interacts with HIF-1a. 293T cells were transfected with plasmids as indicated, lysates were immunoprecipitated with anti-flag (linked to H1F-1a) and Western blots were probed with (A) and anti-myc (linked to Foxp3 in ). () Foxp3 interacts with the N-terminus of HIF-1a. Foxp3-expressing cell lysate was incubated with different fragments of GST-HIF-1 purified from E.coli as indicated, followed by pull-down with GST beads, resolved by SDS-PAGE and Westen blotting with (top and bottom) antibodies, or anti-gst (middle). (D) RORgt protein stability is not regulated by HIF-1a. 293T cells were co-transfected with different doses of HIF-1a expression plasmid along with a constant amount of RORgt plasmid. 48hrs post-transfection, cells were harvested and subjected to SDS-PAGE, followed by Western blot with the indicated antibodies. ell 146, , September 2, 211 ª211 Elsevier Inc. S9

23 ) sirl Si-PHD2 ) Wt-HIF1α(μg) Wt-Ubi Foxp _ HiF-1α(μg) Foxp3 mock treatment Proteasome inhibitor W: anti-tubulin W: anti-tubulin Figure S8. PHD2 Is Required for HIF-1-Mediated Foxp3 Degradation, Related to Figure 5 (A) 293T cells were co-transfected with Foxp3, with or without wild-type ubiquitin and increasing amounts of wild-type HIF-1a in the presence or absence of si-phd2. Western blots were probed with Ab. Depicted are representative findings of at least 3 independent experiments. () Proteasomal inhibitor blocks HIF-1-mediated Foxp3 degradation. 293T cells were cotransfected with Foxp3 and increasing amounts of wild-type HIF-1a. After overnight recovery, cells were incubated with or without 1 mm of the proteasomal inhibitor MG132 for 8h before being harvested for the assay. Western blots were probed as indicated. S1 ell 146, , September 2, 211 ª211 Elsevier Inc.