Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 May 25.
Published in final edited form as: Immunity. 2012 May 10;36(5):717–730. doi: 10.1016/j.immuni.2012.03.020

Loss of epigenetic modification driven by the Foxp3 transcription factor leads to regulatory T cell insufficiency

Matthew L Bettini 1, Fan Pan 4, Maria Bettini 1, David Finkelstein 2, Jerold E Rehg 3, Stefan Floess 5, Bryan D Bell 6, Steven F Ziegler 6, Jochen Huehn 5, Drew M Pardoll 4, Dario AA Vignali 1,*
PMCID: PMC3361541  NIHMSID: NIHMS374204  PMID: 22579476

SUMMARY

Regulatory T (Treg) cells, driven by the Foxp3 transcription factor, are responsible for limiting autoimmunity and chronic inflammation. We showed that a well-characterized Foxp3gfp reporter mouse, which expresses an N-terminal GFP-Foxp3 fusion protein, is a hypomorph that causes profoundly accelerated autoimmune diabetes on a NOD background. Although natural Treg cell development and in vitro function are not markedly altered in Foxp3gfp NOD and C57BL/6 mice, Treg cell function in inflammatory environments was perturbed and TGFβ-induced Treg cell development was reduced. Foxp3gfp was unable to interact with the histone acetyltransferase Tip60, the histone deacetylase HDAC7, and the Ikaros family zinc finger 4, Eos, which led to reduced Foxp3 acetylation and enhanced K48-linked polyubiquitylation. Collectively this results in an altered transcriptional landscape and reduced Foxp3-mediated gene repression, notably at the hallmark IL-2 promoter. Loss of controlled Foxp3-driven epigenetic modification leads to Treg cell insufficiency that enables autoimmunity in susceptible environments.

INTRODUCTION

Optimal immune regulation is essential for dampening ongoing immune responses and preventing damaging autoimmune pathology. CD4+Foxp3+ regulatory T (Treg) cells are a major immunosuppressive cell population responsible for controlling inflammation and autoimmune responses, such as inflammatory bowel disease (IBD) and type 1 diabetes (T1D) (Coombes et al., 2005; Vignali et al., 2008). Mutations in the human FOXP3 locus result in the loss of Treg cells and development of immunodysregulation polyendocrinopathy enteropathy X-linked (IPEX), severe autoimmune and inflammatory syndrome (Campbell and Ziegler, 2007). Likewise, mice that carry the lethal Scurfy mutation or genetic deletion of Foxp3 also lack Treg cells leading to T cell dysregulation, widespread autoimmunity and eventual death by 3–4 weeks of age. Thus, Foxp3 is an essential transcription factor that is required for Treg cell development and function.

There are two types of CD4+Foxp3+ Treg cells. Natural Treg (nTreg) cells develop in the thymus during T cell selection, whereas induced (iTreg) cells develop following TGF-β cytokine exposure in the periphery from naïve CD4+ T cells. Whereas nTreg cells are generated in restricted niches in the thymus as a consequence of high affinity T cell receptor (TCR) stimulation (Bautista et al., 2009; Hsieh et al., 2004), iTreg cells can be generated in response to antigenic stimulation at mucosal sites, during chronic inflammation, or following induction of transplantation tolerance (Curotto de Lafaille and Lafaille, 2009). Even though autoimmune diabetes will eventually develop in NOD mice, a model for T1D in humans, Treg cells can limit disease progression as acute Foxp3+ Treg cell deletion can rapidly accelerate diabetes onset (Chen et al., 2005; Feuerer et al., 2009). The relative contribution of nTreg and iTreg cells in controlling disease progression is currently unknown. Adoptive transfer of islet-specific iTreg cells was sufficient to restore euglycemia in diabetic mice (Tarbell et al., 2007). Also, recent studies suggest that Foxp3+ iTreg cells can develop in the islets during the natural disease course and can contribute to the control of diabetes onset and disease progression (Bluestone and Tang, 2005; Thompson et al., 2011; Wan and Flavell, 2007).

Although considerable insight has been gained into the mechanism of Foxp3 function, important questions remain. Foxp3 is known to induce and repress gene expression either directly or in concert with a host of interacting proteins in a tightly regulated manner (Marson et al., 2007; Zheng and Rudensky, 2007). Foxp3 competes with Fos-Jun heterodimers for NFAT transcription factor binding, which allows for transcription of Treg cell-specific genes as well as the production of inhibitory cytokines, such as IL-10 (Bettini and Vignali, 2009; Campbell and Ziegler, 2007). A large number of genes are repressed by Foxp3 in Treg cells, a process that is dependent on its interaction with transcription factors like Ikaros family zinc-finger 4, Eos (Hill et al., 2007; Pan et al., 2009). In addition to modulating the expression of numerous genes, Foxp3 also regulates its own expression (Tone et al., 2008; Williams and Rudensky, 2007).

Foxp3 also associates with many proteins that epigenetically modulate transcriptional activity of target gene loci via altering DNA methylation, transcription factor and histone post-translational modifications, such as acetylation. These include the histone acetyltransferases (HAT), Tip60 and p300, and the histone deacetylase (HDAC), HDAC7 (Li et al., 2007; Tao et al., 2007). Thus Foxp3:HDAC and Foxp3:HAT complexes may modulate target gene expression via histone or Foxp3 acetylation or deacetylation. However, the physiological contribution and impact of these events in shaping Treg cell development and function, their contribution to nTreg and iTreg cell stability or function, and how this might impact diabetes progression remain unclear.

This project was initiated by a striking and serendipitous observation. Foxp3gfp mice (MGI: Foxp3tm2Ayr), a frequently used Treg cell reporter mouse line, were generated by fusing GFP to the amino-terminus of Foxp3 (Fontenot et al., 2005). Although Foxp3gfp C57BL/6 mice had no apparent disease manifestations or Treg cell abnormalities, Foxp3gfp NOD mice, which we generated to track and analyze Treg cells in this model of T1D, developed dramatically accelerated autoimmune diabetes with 100% penetrance. Subsequent analysis of Foxp3gfp NOD and C57BL/6 mice identified this knockin as a hypomorphic allele, which we used to gain insight into the mechanism and function of Foxp3. We showed that Foxp3gfp failed to associate with Tip60, HDAC7 and Eos, which lead to altered Foxp3-dependent transcription and epigenetic modification. This lead to nTreg cell insufficiency and impaired iTreg cell development.

RESULTS

Foxp3gfp NOD mice develop accelerated autoimmune diabetes

Foxp3gfp C57BL/6 mice were bred onto the NOD/LtJ background. As expected, 80% of the littermate NOD control female mice became diabetic by ~30 weeks of age (Figure 1A). Surprisingly, the Foxp3gfp NOD mice exhibited substantially increased incidence of diabetes with 100% penetrance by 12 weeks of age, when only 10% of wild-type (WT) female NOD mice had become diabetic (Figure 1A). As the Foxp3 gene is located on the X-chromosome, Foxp3gfp/wt NOD heterozygous (HET) female mice have half their Tregs expressing WT Foxp3 and half expressing the Foxp3gfp fusion protein due to random X-inactivation. The Foxp3gfp/wt NOD HET mice developed an intermediate incidence of diabetes and onset kinetics with 47% diabetic at 12 weeks (Figure 1A). These data suggest that the ~50% WT Tregs present in these mice were unable to prevent accelerated autoimmune diabetes.

Figure 1.

Figure 1

Open in a new tab

Foxp3gfp NOD mice develop accelerated autoimmune diabetes. (A) Female Foxp3gfp (n=29), Foxp3gfp/wt HET (n=19) and WT (n=25) NOD mice in the same colony were monitored weekly for diabetes onset (***p<0.001, **p<0.003; Kalpan-Meier with groups compared using the Log-rank test). (B) Insulitis scoring of Foxp3gfp and WT NOD mice was performed at 8 weeks of age. (C) Cellular infiltrate in the islets, PLN, and spleen of 8 week old pre-diabetic WT, Foxp3gfp and FoxP3GFP-hCre NOD mice were analyzed by flow cytometry (n=6–10 mice per group; *p<0.05, **p<0.01, ***p<0.001). Error bars represent SEM.

We next analyzed the islet infiltrate of pre-diabetic mice. At 8 weeks of age, Foxp3gfp NOD mice exhibited substantially increased peri-insulitis and invasive insulitis compared with NOD WT controls (Figure 1B). This correlated with an increase in the number of CD4+ T cells within the islets of Foxp3gfp NOD mice compared to WT NOD mice and Foxp3GFP-hCre NOD BAC transgenic mice (MGI: Tg(Foxp3-EGFP/cre)1aJbs) (Zhou et al., 2009; Zhou et al., 2008), included as an additional control that had a comparable diabetes incidence to WT NOD mice (Figure 1C). Interestingly, there was a significant decrease in the percentage of Foxp3+ Tregs within the islets and peripheral secondary lymphoid organs of Foxp3gfp NOD mice compared to WT NOD or Foxp3GFP-hCre NOD control mice (Figure 1C).

To better assess the cell intrinsic defect of Foxp3gfp, we further analyzed Foxp3gfp/wt NOD HET female mice to compare the relative ratio and expression of the Foxp3 protein. Foxp3 expression (MFI) was significantly increased in Foxp3gfp Tregs in all lymphoid and non-lymphoid organs including the islets when compared with WT Foxp3+ Tregs (Figure S1A available online). The increased expression of Foxp3 did not appear to be a consequence of higher Foxp3 transcription as determined by real-time PCR analysis of Tregs isolated from Foxp3gfp NOD HET mice (Figure S1B). Whereas a ~50:50 ratio of Foxp3gfp versus WT Tregs was found in the thymus, spleen, inguinal lymph nodes (LN) and lung, a modest increase in the percentage of Foxp3gfp Tregs was observed in the liver. Also, thymic Treg development appeared largely intact on the NOD background. Interestingly, the mesenteric lymph nodes (MLN), pancreatic lymph nodes (PLN) and islets displayed a reduced Foxp3gfp:WT Treg ratio in Foxp3gfp NOD HET mice, suggesting that this may occur at or near draining sites of inflammation (Figure S1A). To determine if this was due to reduced proliferation, we assessed expression of the cell cycle protein Ki67 in Foxp3gfp and WT Tregs. There was a slight but significant decrease in the percentage Ki67+ Foxp3gfp Tregs in the islets and MLN, but not in the PLN (Figure S1A). Thus, while the reduced number of Foxp3gfp Tregs observed may be impacted by reduced proliferative potential, we cannot rule out the possibility that there may also be reduced survival or migration to these inflammatory sites. Nevertheless, the reduction observed is presumably a cell intrinsic defect in Foxp3gfp Tregs as this was not observed with the WT Tregs in the same Foxp3gfp NOD HET microenvironment.

We then performed a mixed reconstitution assay in NOD.scid mice to determine if the accelerated autoimmune diabetes in the Foxp3gfp mice was due to defects in nTregs, iTregs or both. Mice received the same number of CD4-depleted splenocytes plus different combination of WT or Foxp3gfp Tconv and WT or Foxp3gfp Tregs at a 10:1 ratio, and recipients were monitored for disease incidence over 30 weeks. This experimental set up allowed us to assess the ability of the two nTreg populations to limit disease and whether any iTregs that emerge from the two Tconv populations contribute to immune regulation. Interestingly, the two Foxp3gfp Treg recipient groups developed accelerated diabetes onset compared with their wild type counterparts (~4 weeks vs ~12 weeks, respectively), although it was not fully penetrant and the incidence at 30 weeks was the same for all four groups (Figure S1C). These data suggest that the Foxp3gfp Tregs may not be as effective as WT Tregs at delaying diabetes onset, and that iTregs (generated in vivo from the conventional T cells (Tconv)) are either not generated, or have no impact on diabetes development. Surprisingly the two Foxp3gfp, but not the WT, Treg recipient groups also developed colitis-like symptoms with severe diarrhea and rectal prolapse, suggesting that the former are unable to prevent intestinal inflammation in this setting (Figure S1D). Colitis onset and incidence were exacerbated in the group that also received Tconv cells from Foxp3gfp mice, which may have been due to limited or ineffective iTreg induction or function. While a detailed analysis of the differential impact and contribution of Foxp3gfp nTregs and iTregs would require more extensive analysis, taken together there appears to be substantial Foxp3gfp Treg insufficiency particularly in highly inflammatory and stressed sites in vivo (Figure 1 and Figure S1E).

The accelerated diabetes incidence observed coupled with the reduced number of Tregs detected in the islets and their diminished capacity to limit diabetes progression suggests that the Foxp3gfp knock-in mutant may function as a hypomorph that is particularly penetrant on an autoimmune background, such as NOD, which may lead to Treg insufficiency. This raised two questions: (i) Are there additional defects in Treg development, homeostasis and/or function caused by expression of the Foxp3gfp hypomorph? (ii) What is the molecular basis for the defects observed?

Foxp3gfp nTreg insufficiency in vivo

Although Foxp3gfp C57BL/6 mice do not manifest any noticeable autoimmune or inflammatory lesions, the knockin mutation is the same and thus may still impart some defect or insufficiency. Thus we included WT and Foxp3gfp mice (C57BL/6 background), as well as Foxp3DTR-GFP C57BL/6 mice (MGI: Foxp3tm3Ayr), a control GFP knock-in that does not exhibit any defects (Kim et al., 2007), in subsequent analysis. Although there was a reduced ratio of Foxp3gfp versus WT Tregs in Foxp3gfp C57BL/6 HET mice, this was comparable in the thymus and secondary lymphoid sites (LNs, spleen) suggesting subtle developmental defect (Figure S2A). It is possible that the Treg insufficiency seen in NOD mice, manifest as reduced Treg numbers and an inability to limit diabetes onset, is a consequence of the ‘challenges’ placed on Tregs in an autoimmune or inflammatory environment. To determine if this ratio is further perturbed in inflammatory sites, Foxp3gfp C57BL/6 HET mice were injected with B16 melanoma cells and analyzed 14 days later. Consistent with observations in the islets of Foxp3gfp NOD mice, there was a substantially reduced ratio of Foxp3gfp Tregs in the inflammatory tumor microenvironment in contrast to the spleen (Figure S2B).

Foxp3gfp Tregs on either a NOD or C57BL/6 background were functionally equivalent in an in vitro suppression assay (Figure S2C and S2D). Likewise, Foxp3gfp and WT C57BL/6 Tregs were equally capable of suppressing the homeostatic proliferation of Tconv cells in lymphopenic Rag1−/− mice (Figure 2A). To test the fitness and function of Foxp3gfp Tregs in a more stringent model, their capacity to prevent systemic autoimmunity was assessed in Foxp3−/− mice, which die ~3–5 weeks of age without adoptive transfer of exogenous Tregs. CD4+CD45RBloCD25+ Tregs from WT and Foxp3gfp C57BL/6 mice were injected intraperitoneally into 2–3-day-old Foxp3−/− mice. At day 35, mice were assessed by histological score and the number of CD4+ T cells in the spleen and LNs. Interestingly, Foxp3gfp Treg recipients exhibited a significant increase in cumulative histological score (liver, skin and lung) and the number of CD4+ T cells in the spleen and LNs of WT and Foxp3gfp Treg recipients (Figure 2B and 2C). Although this caused a significant reduction in the percentage of Foxp3gfp Tregs compared with WT Tregs, the number of Foxp3gfp and WT Tregs was the same (Figure 2C). Collectively, these data suggest that there may not be a reduction in the systemic ‘fitness’ of Foxp3gfp Tregs, although this may be compromised at sites of inflammation. Regardless, Foxp3gfp Tregs are unable to limit the immune pathology observed and there appears to be a deficiency in the capacity of Foxp3gfp Tregs to suppress the T helper 1 (Th1) cell-driven pathology in Foxp3−/− mice.

Figure 2.

Figure 2

Open in a new tab

Foxp3gfp nTreg insufficiency in vivo. (A) Thy1.1+ Tconv (2×106 CD4+CD45RB+CD25− from C57BL/6 mice; >99% purity post-sort) were injected into Rag1−/− mice or mixed with Thy1.2+ Tregs (0.5×106) from either WT or Foxp3gfp C57BL/6 mice (CD4+CD45RBloCD25+; >95% purity post-sort, confirmed by Foxp3 intracellular stain). The number of CD4+Thy1.1+ splenic T cells was determined 7 days later. 11–20 mice per group from 4 separate experiments are shown. Statistical analysis was one-way ANOVA analysis of variance (** p<0.01 and ***p<0.001). (B–C) Foxp3−/− mice (2–3 day old) were adoptively transferred with 106 sorted Tregs (CD4+CD45RBloCD25+) from either WT or Foxp3gfp C57BL/6 mice (>95% purity by Foxp3 intracellular stain). After 35 days, mice were sacrificed and CD4+ T cell numbers and Foxp3+ percentages in the spleens and peripheral LNs quantified. Three separate experiments with 5–9 mice per group are shown (*p<0.05, ***p<0.001). Error bars represent SEM. (D) Foxp3−/− mice treated as in (C) were evaluated over 18 weeks for clinical signs of autoimmunity. When mice reached a clinical score of 5 (out of six), they were euthanized. n=10 mice/group (**p<0.01, ***p<0.001; Kalpan-Meier with groups compared using the Log-rank test).

Differences in histological score and CD4+ T cell numbers were far greater at 35 days versus 23 days post-transfer, suggesting a progressive failure of Foxp3gfp Tregs over time (Figure 2B, 2C, and data not shown). Thus, we asked if there were differences in the length of time Foxp3gfp or WT Tregs could control the onset of autoimmunity that develops in Foxp3−/− mice. Strikingly, Foxp3−/− mice reconstituted with Foxp3gfp Tregs developed autoimmune symptoms significantly faster than WT Treg recipients (Figure 2D). Taken together, these data suggest that while Foxp3gfp Tregs appear to retain normal in vitro suppressive activity, they exhibit substantial functional insufficiency in vivo, particularly in highly stressed environments which may impact their capacity to limit chronic autoimmune or inflammatory diseases.

Impaired development of Foxp3gfp iTregs

TGFβ-induced Tregs (iTregs) are generated at inflammatory sites and can contribute to immune homeostasis (Curotto de Lafaille et al., 2008; Geuking et al., 2011; Thompson et al., 2011) We sought to determine whether there was a defect in iTreg development from Foxp3gfp Tconv cells. Consistently diminished expression of Foxp3 was observed with CD4+CD45RBhiCD25−GFP− Foxp3gfp NOD and C57BL/6 Tconv cells stimulated with anti-CD3 and anti-CD28-coated beads, TGFβ and IL-2 compared with WT and Foxp3GFP-hCre/Foxp3DTR-GFP controls (Figure S3A and S3B). We next utilized Foxp3gfp/wt female HET NOD and C57BL/6 mice to determine if Tconv cells that could express the Foxp3gfp fusion protein had any competitive disadvantage over cells that could express WT Foxp3. There was a ~50% reduction in the efficiency of Foxp3+ iTreg conversion with cells expressing Foxp3gfp compared with WT Foxp3 (Figure 3A, 3B and Figure S3C). The lack of conversion observed did not appear to be due to an increase in cell death by apoptosis or altered proliferative capacity (Figure S4D). However, the cells that did convert to iTregs appear to have comparable in vitro suppressive capacity (Figure S4A and S4B). Taken together, these data indicated a substantial deficiency in the ability of Foxp3gfp Tconv cells to convert into Foxp3+ iTregs.

Figure 3.

Figure 3

Open in a new tab

Impaired Foxp3gfp iTregs conversion in vitro. (A) CD4+CD45RBhiCD25− Tconv were purified from Foxp3gfp/wt NOD HET mice and stimulated with anti-CD3/CD28 coated beads, 5ng/ml TGFβ and 100U/ml IL-2. After 5 days, cells were stained with anti-Foxp3. Cells were gated on GFP+Foxp3+ and GFP−Foxp3+. A representative flow plot and an average of 5 experiments are shown (**p<0.01 by 2-way analysis paired t-test). (B) CD4+CD45RBhiCD25− Tconv were sorted from Foxp3gfp and Foxp3DTR-GFP C57BL/6 HET mice, stimulated and analyzed as in (A). A representative of 6 separate experiments is shown (*** p<0.001 by 2-way analysis paired t-test).

We then assessed the capacity of Foxp3gfp Tconv cells to convert into Foxp3+ iTregs in two in vivo models. First, CD4+CD45RBhiCD25− Tconv cells from WT and Foxp3gfp NOD mice were injected intravenously into NOD.scid mice and the number and percentage of Foxp3+ iTregs in the spleen and MLN determined 21 days later. Remarkably, substantially reduced conversion of Foxp3gpf Tconv (Foxp3−) into Foxp3+ iTregs cells was observed in both lymphoid sites when compared with WT controls (Figure 4A and 4B). Despite the lack of autoimmune or inflammatory disease in Foxp3gfp C57BL/6 mice, an essentially identical defect in iTreg generation after transfer of CD4+CD45RBhiCD25− Tconv cells from Foxp3gfp C57BL/6 mice into Rag1−/− mice was observed. Furthermore, the ratio of converted Foxp3+ cells was skewed toward cells expressing WT Foxp3 following transfer of Tconv cells from Foxp3gfp/wt C57BL/6 HET mice injected into Rag1−/− mice (Figure 4C).

Figure 4.

Figure 4

Open in a new tab

Impaired development of Foxp3gfp iTregs in vivo. CD4+CD45RBhiCD25− Tconv were sorted (>98% purity) from Foxp3gfp and WT C57BL/6 or NOD mice and 2×106 cells injected i.v. into Rag1−/− or NOD.SCID mice, respectively. Mice were sacrificed 21 days post-injection; spleens and LNs harvested and stained for CD4, CD25 and intracellular Foxp3. (A) Representative flow plots. (B) Percentage and total number of CD4+Foxp3+ cells in spleens and MLN (n=9 mice per group, ***p<0.001, *p<0.03 (C) CD4+CD45RBhiCD25− Tconv (C57BL/6 Foxp3gfp/wt HET mice, >99% purity; 2×106) were injected i.v. into Rag1−/− mice and analysed as in (A) (n=9 mice; *p<0.05 by 2-way analysis paired t test). (D) CD45.1+ C57BL/6 mice were injected i.v. with a 50:50 mixture of OT-II CD4+CD45RBhiCD25− Tconv (>99.7% Foxp3−) from either Thy1.1+CD45.2+ WT OT-II mice or Thy1.2+CD45.2+ Foxp3gfp OT-II mice, and placed on drinking water with OVA antigen. After 6 days, the percentage of Foxp3+ iTregs in PP, MLN, spleen and peripheral LN was determined (n=15; **p<0.005, *p<0.05 by 2-way analysis paired t-test).

Second, CD45.1+ congenic mice were injected intravenously with a 50:50 ratio of Thy1.1+CD45.2+CD4+CD45RBhiCD25− Tconv cells from WT OT-II TCR transgenic mice and from Foxp3gfp OT-II mice and given OVA antigen in the drinking water. After 6 days, the percentage of Foxp3+ iTregs in the spleen and peripheral LNs was determined. Consistent with the observations above, the ratio of Foxp3+ iTregs was highly skewed toward WT Foxp3+ versus Foxp3gfp iTregs, which was particularly evident in the Peyer’s patches (PP) and the MLN, (Figure 4D). Given that the proliferative capacity of WT and Foxp3gfp iTregs was comparable (Figure S4C), the lack of Foxp3gfp iTreg development is likely due to diminished rate of stable Foxp3gfp expression.

Taken together, these data suggest a defect in the capacity of Foxp3gfp Tconv to induce Foxp3 expression and convert into iTregs. Expression of the Foxp3gfp hypomorph limits iTreg development and causes nTreg insufficiency resulting in a substantial impact on their capacity to control autoimmune and inflammatory disease.

Foxp3gfp exhibits selective impaired molecular interactions

Certain epigenetic modifications of the Foxp3 locus and molecular associations between Foxp3 and a variety of transcription factors and modifying proteins have been collectively shown to modulate Treg stability, homeostasis, migration and function. Thus, we assessed whether the Foxp3 N-terminal fusion of GFP had altered any specific associations or epigenetic modifications which might underlie the phenotypic defects observed. The methylation status of the CpG-rich Treg-specific demethylated region (TSDR) of the Foxp3 locus has been shown to correlate with Treg stability (Baron et al., 2007; Floess et al., 2007; Nagar et al., 2008). However, no differences in the methylation status of the Foxp3 TSDR were observed in Foxp3gfp nTregs or iTregs compared with their respective controls (Figure S5A).

We next examined the stability of protein:protein interactions known to occur between Foxp3 and other transcription factors or epigenetic modifiers which collectively shape the Treg transcriptional landscape. There was a complete loss of Tip60 HAT association and partial reduction of p300 HAT association with Foxp3gfp but not WT Foxp3 (Figure 5A and 5B). There was also a partial reduction in Foxp3:Tip60 association in Foxp3gfp iTregs compared with their Foxp3GFP-hCre counterparts (Figure S5B). There was also a substantially reduced association between HDAC7 and Foxp3gfp but not WT Foxp3 (Figure 5C). Taken together, these data suggested that epigenetic modifications driven by the association between Foxp3 and Tip60 and/or HDAC7 might be deficient.

Figure 5.

Figure 5

Open in a new tab

Foxp3gfp exhibits selective impaired molecular interactions. CD4+CD45RBloCD25+ WT and Foxp3gfp NOD nTregs (sorted to >93% purity) were analysed by IP/western blot using the Abs indicated. Representative blots of at least three separate experiments are shown. (G) Arrows highlight ubiquitylated Foxp3.

Gene repression by Foxp3 is mediated by Eos, an Ikaros family zinc-finger transcription factor (Pan et al., 2009). Interestingly, Foxp3gfp exhibited substantially reduced association with Eos compared with WT Foxp3 in nTregs and iTregs (Figure 5D and Figure S5C). Given that Eos, Tip60 and HDAC7 are known to associate with the N-terminal region Foxp3 (Xiao et al., 2010), we examined association with NFAT and AML-1 (Runx-1), two known Foxp3 C-terminal associated proteins. No differences were observed in the association of NFAT or AML-1 with Foxp3gfp compared with WT Foxp3 (Figure 5E and data not shown). Furthermore, luciferase reporter assays demonstrated a comparable ability of Foxp3gfp and WT Foxp3 to repress NFAT and RORα promoter activity (Figure S5D). Similar results were observed following co-transfection of HEK 293T cells with plasmids encoding Foxp3gfp or WT Foxp3 and NFAT, Eos or Tip60, demonstrating that the differential associations observed in Tregs were inherent properties of the Foxp3gfp fusion protein rather than an indirect consequence of Treg dysregulation (Figure S5E–S5G). These data suggest that Foxp3gfp possesses a selective defect which is limited to interactions utilizing the Foxp3 N-terminus.

Tip60 has been shown to be responsible for the acetylation of Foxp3 while HDAC7 is involved in the deacetylation and inactivation of Foxp3 (Xiao et al., 2010). The consequence of losing both associations and how this might affect Foxp3 acetylation is unclear. Interestingly, total acetyl-lysine was reduced on Foxp3gfp compared with WT Foxp3 in nTregs and iTregs (Figure 5F and Figure S5H), suggesting that the loss of Tip60:Foxp3 interaction may result in reduced Foxp3gfp stability and/or function (Li et al., 2007). Interestingly, there was a marked increase in total ubiquitylation and K48-linked polyubiquitylation of Foxp3gfp compared with WT Foxp3 (Figure 5G), suggesting that there may be increased Foxp3gfp proteasomal degradation and/or altered transcriptional activity. Furthermore, Foxp3gfp NOD nTregs and iTregs showed a small but significant reduction in Foxp3 protein half-life compared with WT Foxp3 NOD Tregs (Figure S6A and S6B). These data raise the possibility that this subtle increase in Foxp3gfp instability could accumulate over time and/or have a greater impact at sites of inflammation in vivo.

In order to provide further insight into the global transcriptional consequences of these otherwise selective defects in Foxp3gfp protein associations and epigenetic modifications, we isolated mRNA from highly purified CD4+CD3+GFP+ NOD Foxp3gfp and Foxp3GFP-hCre splenic Tregs, and GFP− Tconv controls, from 6 week old pre-diabetic mice and performed Affymetrix GeneChip microarray analysis with four independent replicates. Principal component analysis (PCA) demonstrated that the Foxp3gfp and Foxp3GFP-hCre GFP− Tconv replicate datasets were similar as they clustered in close proximity while the Foxp3GFP-hCre GFP+ Treg replicates were tightly clustered but distinct from the Tconv datasets (Figure 6A). While the Foxp3gfp Treg replicates were in close proximity to the Foxp3GFP-hCre Treg, they were dispersed relative to each other suggesting some degree of variability between individual mice within this population. Global analysis of the differentially expressed genes in Foxp3gfp and Foxp3GFP-hCre Tregs compared with their Tconv counterparts revealed that they possessed a predominantly intact Treg signature with many of the genes associated with previously described Treg signatures differentially expressed in both Treg populations (Figure 6B and 6C, and Figure S6C). Nevertheless, differences were evident between the most differentially expressed genes amongst the two Treg populations (Figure S6D and S6E). This is highlighted by several known Treg genes being significantly upregulated in Foxp3gfp Tregs (Figure 6D). Surprisingly, around half the most differentially expressed genes between Foxp3gfp and Foxp3GFP-hCre Tregs were not Treg signature genes (Figure 6D). For instance, several genes that are expressed in activated, migratory cells (Ccr2, Ntn1, Itgae, Lamc1, Anxa2, S100a6), and upregulated (Ikzf3, Bcl2l15, Lgals1, Egln3, Matk) or downregulated (Arhgap5) in active but dying cells are differentially expressed in Foxp3gfp Tregs compared with their Foxp3GFP-hCre Treg controls, which collectively may cause deregulated migration and/or altered functional sufficiency in inflammatory environments. Taken together, this analysis indicates a perturbation in the expression of some but not all genes that may in turn lead to the Treg insufficiency observed.

Figure 6.

Figure 6

Open in a new tab

Altered transcriptional landscape in Foxp3gfp Tregs. FoxP3GFP-hCre or Foxp3gfp NOD Treg and Tconv purified by FACS (>99% purity), mRNA isolated and subjected to Affymetrix analysis (4 independent samples per group). (A) PCA analysis. (B–C) Volcano plot comparing FoxP3GFP-hCre (B) or Foxp3gfp (C) Tregs with their respective Tconv controls. Highest modulated genes are marked. (D) Most differentially expressed genes in Foxp3gfp Treg compared to wild type Tregs and Tconv controls are depicted in a heat map. Treg signature genes are in red (Supplemental Figure 14 and Pillai et al., 2011) (E) FoxP3GFP-hCre/wt or Foxp3gfp/wt HET Tregs were stained for Itgae (CD103). Data are representative of 5–6 mice (**p<0.01 by 2 way analysis paired t test).

Given the altered transcriptional landscape observed, we questioned whether Foxp3gfp Tregs exhibited an altered phenotype which could impact their capacity to suppress the predominantly Th1 cell-driven disease observed in NOD and Foxp3−/− mice. Expression of the inflammatory cytokines IFNγ and IL-17, and the transcription factor T-bet, which is required for selective repression of Th-1 driven disease (Koch et al., 2009), was not altered in Foxp3gfp NOD Tregs (Figure S7A, S7B and data not shown). However the percentage of Tregs expressing CXCR3, a T-bet downstream target, was reduced in Foxp3gfp NOD mice (Figure S7C). Consistent with the CD103 (Itgae) upregulation observed in the Affymetrix analysis, an increased percentage of the CD103+ Treg subset was detected in Foxp3gfp versus WT mice (Figure 6E). Although one might predict that upregulation of two genes that contribute to cell migration might increase migration into inflammatory sites, it is possible that this in fact leads to an altered migratory pattern. Indeed, there appeared to be a selective reduction of Tregs at inflammatory sites in Foxp3gfp NOD mice and B16 tumor-baring Foxp3gfp C57BL/6 mice but an increase in Foxp3gfp Tregs in the liver (Figure 1C, S1A and S2B).

A hallmark gene known to be potently repressed by Foxp3 is IL-2 (Chen et al., 2006; Crispin and Tsokos, 2009). Thus, IL-2 promoter activity and histone acetylation status can be used as a surrogate for the general functionality of Foxp3 in Tregs. Foxp3gfp expressing Jurkat cells exhibited a significant increase in IL-2 promoter activity compared to cells expressing WT Foxp3 (Figure 7A). Furthermore, there was substantially diminished Foxp3 binding to the IL-2 promoter in Foxp3gfp nTregs and iTregs compared with WT controls (Figure 7B and data not shown). Consistent with the reduced Foxp3gfp:HDAC7 association observed (Figure 5C), histone H3 and H4 lysine acetylation (H3Ac and H4Ac) at the IL-2 promoter was substantially increased in Foxp3gfp nTreg and iTregs compared with controls suggesting incomplete repression of the IL-2 promoter (Figure 7C and S7D). There was a small but significant increase in the percentage of IL-2+ Foxp3gfp Tregs compared with WT Tregs, which was approximately a third of the IL-2+ Tconv generated under the same stimulation conditions (5.3% vs 16.1%; Figure S7E). Furthermore, there was a significant increase in the IL-2 MFI of Foxp3gfp versus WT Tregs. This ultimately led to significant amounts of IL-2 produced by Foxp3gfp Tregs in contrast with WT Tregs (Figure 7D). Taken together, these data show that Foxp3gfp fails to productively engage with key mediators of its transcription program, specifically, HDAC7, Tip60 and Eos. This leads to a partial loss of gene repression, altered Foxp3 stability and limited histone deacetylation at target promoters, such as IL-2, suggesting that this loss of Foxp3-driven epigenetic modification may lead to the Treg insufficiency and enhanced diabetes observed in Foxp3gfp NOD mice.

Figure 7.

Figure 7

Open in a new tab

Enhanced IL-2 promoter activity and IL-2 production by Foxp3gfp Tregs. (A) Jurkat T cells were transfected with an IL-2-luciferase reporter plasmid along with indicated Foxp3-encoded plasmids. Empty vector pMIY was used to maintain equal DNA content for all transfections. Cells were either treated with PMA plus ionomycin or left untreated 12hr post-transfection, prior to analysis of luciferase activity, which was normalized to the Renilla luciferase activity. Data shown are representative of at least 3 independent experiments (mean and SD of triplicate transfections are shown; *p<0.005 by t test). (B) ChIP assay was performed with anti-Foxp3 Ab or IgG control followed by PCR using primers flanking the Foxp3-binding region of the IL-2 promoter. Shown are the mean (±SEM) of at least 3 independent experiments (**p<0.001 by t test). (C) Histone acetylation at the IL-2 promoter was determined by ChiP in the cell subsets indicated. Shown are the mean (±SEM) of at least 3 independent experiments (*p<0.005 by t test). Naive is defined as rested, while Tconv is defined as previously activated by anti-CD3/CD28. (D) CD4+CD45RBloCD25+ Tregs from WT, Foxp3gfp and FoxP3GFP-hCre NOD mice were sorted (>95% purity), stimulated (0.5×105) with plate-bound anti-CD3 and soluble anti-CD28 for 72h, and IL-2 concentration in supernatants determined by ELISA. Shown are the mean (±SEM) of at least 3 independent experiments (*p<0.05 by 2-way paired t test).

DISCUSSION

Foxp3 acts as a master transcription factor that is essential for Treg development and function, inducing and repressing a wide variety of genes that collectively constitute the Treg signature (Hill et al., 2007; Marson et al., 2007; Wan and Flavell, 2007; Zheng et al., 2007). These activities are mediated by its interaction with many transcription factors, such as Eos, AML1 (Runx-1), NFAT and RORα, or epigenetic regulators, such as Tip60 and HDAC7 (Campbell and Ziegler, 2007; Wang et al., 2009; Xiao et al., 2010). For instance, Foxp3 association with Eos and HDAC7 mediates the gene repression profile component of the Treg signature (Li et al., 2007; Pan et al., 2009), while Foxp3 association with Tip60 is thought to enhance Foxp3 stability (Tao et al., 2007). Although many of these interactions have been documented and linked to specific functions, their global physiological contribution to Treg development, homeostasis and function is unclear. In general, these proteins interact with many different partners and thus perform numerous diverse functions. Consequently, targeted deletion or inhibition of these proteins can produce results which are difficult to interpret.

Our serendipitous observation that Foxp3gfp was a previously unrealized hypomorph provided a unique opportunity to probe the physiological importance of these interactions on Treg development, homeostasis and function. Indeed, Foxp3gfp expression on the autoimmune-prone NOD background led to dramatically accelerated diabetes, that was reminiscent of the exacerbated disease observed following acute Treg depletion (Chen et al., 2005; Feuerer et al., 2009). The molecular defects described in this study are selective and restricted to associations with the N-terminus of Foxp3, likely due to steric hindrance caused the attachment of GFP without a flexible linker. Thus interactions with Eos, Tip60 and HDAC7 (this paper) or Hif1α (Darce et al., 2012) are substantially reduced or lost, while more distal interactions with NFAT, AML1/Runx-1 and RORα are maintained, or even enhanced in the case of Irf4 (Darce et al., 2012). This presented a unique opportunity to determine the selective physiological impact of these altered interactions on Treg development and function.

A significantly reduced number of Tregs was observed in the islets of Foxp3gfp NOD mice, although the severity of the disease suggested that factors other than reduced number also contributed to the precipitous disease onset. Although there are subtle defects in thymic nTreg generation on the C57BL/6 background (Fontenot et al., 2005), our data clearly show that Foxp3gfp expression results in nTreg insufficiency in competitive and/or inflammatory environments and limits iTreg generation. Indeed, a substantially reduced number of Foxp3gfp versus WT Tregs was observed in B16 tumors. The contribution, if any, of iTregs in limiting the rate of autoimmune diabetes in NOD mice has not been fully determined, so it is possible iTregs may play a role as their reduced production may underlie the accelerated rate of disease onset observed. However, our transfer experiments suggest that there may not be a significant role for iTregs in limiting autoimmune diabetes, although they appeared to limit colitis.

A curious, and at first glance counterintuitive, observation was the increased Foxp3 levels in Foxp3gfp NOD Tregs. Treg activation is known to induce the upregulation of Foxp3, so it is possible that the enhanced inflammatory environment in the islets of Foxp3gfp NOD mice causes increased Treg activation, which may in turn induce the increased Foxp3 expression observed. While steady state analysis did not suggest that there is substantial Foxp3gfp instability, the data were significant and in an inflammatory environment Foxp3gfp instability could be more pronounced. Thus, the increased Foxp3 expression observed may also be a physiological response to Foxp3 instability and an attempt to compensate for the defects inherent in Foxp3gfp Tregs, which may in turn be induced by the accelerated disease progression. We have shown that Foxp3gfp is unable to effectively associate with Tip60 and HDAC7, enzymes that control acetylation. Given that we see reduced acetylation, it is possible that the loss of Tip60 causes to Foxp3gfp instability and/or functional insufficiency as loss of Foxp3:Tip60 association has also been shown to lead to increased Foxp3 expression in 293T cells following Tip60 knock-down (Li et al., 2007; Tao et al., 2007). This would also be consistent with the increased K48-linked polyubiquitylation of Foxp3gfp observed, which typically leads to proteasomal degradation. Interestingly, it has been suggested that K48-linked polyubiquitylation can also lead to transcriptional repression without protein degradation, a possibility that warrants further investigation (Flick et al., 2004; Flick et al., 2006; Ouni et al., 2011). However, HDAC7 has also been suggested to deacetylate the ε-acetyllysine residues found on Foxp3 suggesting a level of complexity that has yet to be fully understood (Johnstone, 2002; Saouaf et al., 2009). Nevertheless, this reduced Foxp3 stabilization may underlie the reduced iTreg conversion and nTreg insufficiency observed.

Foxp3 associates with both HDAC7 and Tip60, which are collectively required for transcriptional repression (Li et al., 2007; Xiao et al., 2003). Likewise, Foxp3 association with Eos has been shown to be required for the gene repression component of the Treg signature (Li et al., 2007; Pan et al., 2009). Our observation that Foxp3gfp fails to efficiently associate with Eos, Tip60 and HDAC7, while other known associations are normal, implies a selective defect in Foxp3-mediated gene repression. Indeed, several genes that are not associated with Tregs were upregulated. Analysis of IL-2, a hallmark Foxp3-repressed gene, revealed reduced Foxp3 binding to the IL-2 promoter, concomitant increase in histone acetylation associated with the IL-2 promoter, increased promoter activity and increased IL-2 production. Collectively, these data suggest that loss of Foxp3gfp association with Eos, Tip60 and HDAC7 limits nTreg function at inflammatory sites and iTreg induction, leading to Treg insufficiency and accelerated autoimmunity in prone environments. While we observed substantial Treg insufficiency in two predominantly Th1-driven disease models, the NOD model of autoimmune diabetes and following adoptive transfer into Foxp3−/− mice, it is possible that distinct outcomes would be observed in other disease models. Indeed, enhanced Foxp3gfp:Irf4 association may underlie the enhanced Treg-mediated suppression of Th2 cell responses and protection from arthritis seen in Foxp3gfp.K/BxN mice (Darce et al., 2012). The Foxp3gfp reporter mouse has been a critical tool in advancing our understanding of Treg development, homeostasis and function (Liston et al., 2008; Zheng and Rudensky, 2007). Our serendipitous observation that this broadly utilized Foxp3gfp reporter is a hypomorph that causes profoundly accelerated autoimmune diabetes on an NOD background raises questions and opportunities. Although the numerous studies that have used the Foxp3gfp reporter are internally controlled (i.e. mutant and wild-type gene of interest both expressing the Foxp3gfp reporter), the presence of this hypomorph could potentially exaggerate or mitigate any enhanced or repressed disease observed, respectively, as a consequence of introducing unlinked compound mutations. At the very least, our observations suggest that a review of published observations using this reporter is warranted. Our observations also send a note of caution to those generating knockin mutants as seemingly innocuous alterations can lead to very different consequences in autoimmune-prone or disease-induced environments.

Despite these concerns, our findings also present a unique opportunity as the Foxp3gfp hypomorph represents a new tool that may facilitate further dissection of the complex processes orchestrated by Foxp3 in mediating Treg development, homeostasis and function. Our findings highlight the physiological contribution of Foxp3 association with Tip60, HDAC7 and Eos that warrant further study. Furthermore, the extent of Foxp3 acetylation and ubiquitylation may contribute a greater role in Treg stability than previously realized. Our data also imply that loss of controlled Foxp3-driven epigenetic modification may lead to Treg insufficiency that ultimately underlies the accelerated autoimmunity observed in certain prone environments.

METHODS

Mice

Foxp3gfp (MGI: Foxp3tm2Ayr) and Foxp3DTR-GFP (MGI: Foxp3tm3Ayr) mice (both 100% C57BL/6 by microsatellite analysis) were provided by Sasha Rudensky (Fontenot et al., 2005; Kim et al., 2007). Foxp3gfp mice were bred 10 times onto the NOD background to achieve 100% NOD genetic background based on microsatellite analysis. Foxp3GFP-hCre NOD (NOD/ShiLT-Tg(Foxp3-EGFP/cre)1Jbs/J; MGI: Tg(Foxp3-EGFP/cre)1aJbs) BAC transgenic mice were obtained from The Jackson Laboratory (Zhou et al., 2008). WT and Foxp3gfp OT-II mice were provided by Hongbo Chi (St. Jude). NOD/LtJ, NOD.SCID, Rag1−/−, C57BL/6, and B6.PL-Thy1a/CyJ (Thy1.1 congenic) mice were purchased from The Jackson Laboratory. NODmice were monitored weekly for diabetes onset, mice were considered diabetic when tested positive for glucose in urine by Clinistix® concomitant with blood glucose reading above 400mg/dL. All animal experiments were performed in American Association for the Accreditation of Laboratory Animal Care-accredited, specific pathogen- free, helicobacter-free facilities in the St. Jude Animal Resource Center following national, state, and institutional guidelines. Animal protocols were approved by the St. Jude Animal Care and Use Committee.

Measurement of insulitis

Insulitis was assessed as described previously (Burton et al.). Briefly, pancreata of NOD mice embedded in paraffin was cut at 4 μm–thick sections at 150-μm step-sections and stained with hematoxylin and eosin at the St. Jude Histology Core Facility. An average of 90–100 islets per mouse were scored in a blinded manner. See online for details of insulitis measurement.

Flow cytometric analysis and cell sorting

Tconv (CD4+ CD45RBhi CD25−) and Treg (CD4+ CD45RBlo CD25+)(Biolegend) cells were positively sorted by FACS on a Reflection (i-Cyt) from spleens and lymph nodes of WT or Foxp3gfp age-matched mice. Spleens, LN and pancreatic islets were removed, processed, counted, stained, and then analyzed on a LSRII flow cytometer (BD PharMingen, San Diego, CA). Pancreatic islets were isolated as previously reported by flushing the pancreas with Collagenase 4, then digesting the pancreas, and picking the isolated islets (Burton et al., 2008). Cell surface molecules including CD4, CD25, CXCL3 and CD103 (Biolegend) were stained with fluorescently conjugated mAbs. Flow cytometric analysis was performed using a LSR II and data analyzed using FlowJo 8.8.2 software.

Foxp3−/− rescue model

The Foxp3−/− rescue model was performed as described previously (Collison et al., 2010; Workman et al., 2011). Briefly, the Treg populations indicated were purified by FACS and injected (106) i.p. into 2–3 day old Foxp3−/− mice. On day 35, disease was assessed as a clinical score, and mice euthanized to determine the number of CD4+ Tconv in the spleen and LN by flow cytometry. In addition, lung, liver and ear pinna were prepared for H&E analysis and the severity of inflammation was assessed and scored in a blinded manner by an experienced veterinary pathologist (J.E.R.).

Induced Treg assay

Briefly, CD4+ Tconv (CD45RBhiCD25−) cells were sorted to greater than 99%. Purity was confirmed by intracellular Foxp3 staining. Cells were plated at 0.2×106 per well in a 96 well U-bottom plate and stimulated with anti-CD3/anti-CD28-coated beads at 1:1 ratio beads to cells, plus 5ng/ml TGFβ and 100U/ml IL-2. After 5 days, cells were stained with an anti-Foxp3 antibody and the percent conversion was calculated.

In vivo induced Treg conversion assay

Briefly, CD4+ Tconv (CD45RBhiCD25−) cells were sorted to greater than 98%. Purity was confirmed by intracellular Foxp3 staining. 2×106 Tconv cells were then injected into Rag1−/− mice. After 21 days, spleens and mesenteric LNs were harvested and stained with anti-CD4 and anti-Foxp3 to determine Foxp3 induction. For OVA fed in vivo iTregs, CD45.1+ congenic mice were injected intravenously with a 50:50 ratio of Thy1.1+CD45.2+CD4+CD45RBhiCD25− Tconv (sorted to greater than 99.7% purity for Foxp3−Tconv) cells from WT OT-II mice or Thy1.2+ CD45.2+CD4+CD45RBhiCD25− Tconv cells from Foxp3gfp OT-II mice. After 6 days of receiving water containing OVA antigen (Sigma, 2g/100ml), the PP, MLN, spleen and peripheral LNs were harvested. Donor cells were determined by first gating on CD45.2 and CD4, then gating on either Thy1.1 or Thy1.2.

Immunoprecipitation and western blot analysis

Immunoprecipitation and western blotting were performed as described previously (Pan et al., 2007). Immunoprecipitation experiments were carried out by using Pierce crosslink IP kit and clean-Blot IP detection system (Thermo Scientific). The following antibodies were used in the current study: anti-Foxp3 (eBioscience); anti-Myc, anti-GST, anti-p300, anti-NFAT and anti-HDAC7 (Santa Cruz); anti-ubiquitin, anti-Eos and anti-Flag (Sigma); anti-Tip60 and anti-acetyllysine (Cell Signaling); anti-lysine 48 (K48) (Invitrogen).

Affymetrix array and analysis

Wild type (Foxp3GFP-hCre) or Foxp3gfp NOD Tconv (CD3+CD4+GFP−) and Treg (CD3+CD4+GFP+) splenocytes were sorted twice to greater than 99.0% purity from 6 week old mice by FACS and mRNA isolated using TRIZOL (Life Technologies). Additional methods available online.

Statistical Analysis

Unless otherwise noted in figure legends, statistical significance was determined using Mann-Whitney unpaired t-test. All statistical significance is noted in figures. In select figures (where noted in legends) statistical analysis was determined using Wilcoxon matched paired t-test. All statistics were determined with GraphPad Prism software (San Diego, CA).

Supplementary Material

01

HIGHLIGHTS.

  • Foxp3gfp is a hypomorph that causes accelerated autoimmune diabetes in NOD mice

  • Foxp3gfp causes nTreg cell insufficiency and impaired iTreg cell development

  • Foxp3gfp has defective interaction with Tip60, HDAC7 and Eos

  • Foxp3gfp causes altered Foxp3-dependent transcription and epigenetic modification

Acknowledgments

We thank Sasha Rudensky and Hongbo Chi for mice; Meghan Turnis for performing B16 injections, Kate Vignali for technical assistance; Matt Smeltzer for biostatistical support; Karen Forbes, Ashley Castellaw and Amy McKenna for maintenance, breeding and genotyping of mouse colonies; Richard Cross, Greig Lennon and Stephanie Morgan for FACS; the people from Epiontis for the DNA methylation analysis; the staff of the Shared Animal Resource Center at St Jude for the animal husbandry; the Veterinary Pathology Core Laboratory at St. Jude for histology and immunohistochemistry support; and the Hartwell Center for Biotechnology and Bioinformatics at St Jude for real-time PCR primer/probe synthesis. Supported by the National Institutes of Health (AI039480, AI091977, DK089125; to D.A.A.V.), a Hartwell Postdoctoral Fellowship (to M.L.B.), JDRF Postdoctoral Fellowship (3-2009-594 to M.B.), the St Jude National Cancer Institute Center (CA-21765; to D.A.A.V. and J.E.R.), the American Lebanese Syrian Associated Charities (to D.A.A.V. and J.E.R.) and the German Research Foundation (DFG, SFB621 and SFB738; to J.H.).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Baron U, Floess S, Wieczorek G, Baumann K, Grutzkau A, Dong J, Thiel A, Boeld TJ, Hoffmann P, Edinger M, et al. DNA demethylation in the human FOXP3 locus discriminates regulatory T cells from activated FOXP3(+) conventional T cells. Eur J Immunol. 2007;37:2378–2389. doi: 10.1002/eji.200737594. [DOI] [PubMed] [Google Scholar]
  2. Bautista JL, Lio CW, Lathrop SK, Forbush K, Liang Y, Luo J, Rudensky AY, Hsieh CS. Intraclonal competition limits the fate determination of regulatory T cells in the thymus. Nat Immunol. 2009;10:610–617. doi: 10.1038/ni.1739. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bettini M, Vignali DAA. Regulatory T cells and inhibitory cytokines in autoimmunity. Current Opinion in Immunology. 2009;21:612–618. doi: 10.1016/j.coi.2009.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bluestone JA, Tang Q. How do CD4+CD25+ regulatory T cells control autoimmunity? Curr Opin Immunol. 2005;17:638–642. doi: 10.1016/j.coi.2005.09.002. [DOI] [PubMed] [Google Scholar]
  5. Burton AR, Vincent E, Arnold PY, Lennon GP, Smeltzer M, Li CS, Haskins K, Hutton J, Tisch RM, Sercarz EE, et al. On the pathogenicity of autoantigen-specific T-cell receptors. Diabetes. 2008;57:1321–1330. doi: 10.2337/db07-1129. [DOI] [PubMed] [Google Scholar]
  6. Campbell DJ, Ziegler SF. FOXP3 modifies the phenotypic and functional properties of regulatory T cells. Nat Rev Immunol. 2007;7:305–310. doi: 10.1038/nri2061. [DOI] [PubMed] [Google Scholar]
  7. Chen C, Rowell EA, Thomas RM, Hancock WW, Wells AD. Transcriptional regulation by Foxp3 is associated with direct promoter occupancy and modulation of histone acetylation. J Biol Chem. 2006;281:36828–36834. doi: 10.1074/jbc.M608848200. [DOI] [PubMed] [Google Scholar]
  8. Chen ZB, Herman AE, Matos M, Mathis D, Benoist C. Where CD4(+) CD25(+) T reg cells impinge on autoimmune diabetes. Journal of Experimental Medicine. 2005;202:1387–1397. doi: 10.1084/jem.20051409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Collison LW, Chaturvedi V, Henderson AL, Giacomin PR, Guy C, Bankoti J, Finkelstein D, Forbes K, Workman CJ, Brown SA, et al. IL-35-mediated induction of a potent regulatory T cell population. Nat Immunol. 2010;11:1093–1101. doi: 10.1038/ni.1952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Coombes JL, Robinson NJ, Maloy KJ, Uhlig HH, Powrie F. Regulatory T cells and intestinal homeostasis. Immunol Rev. 2005;204:184–194. doi: 10.1111/j.0105-2896.2005.00250.x. [DOI] [PubMed] [Google Scholar]
  11. Crispin JC, Tsokos GC. Transcriptional regulation of IL-2 in health and autoimmunity. Autoimmun Rev. 2009;8:190–195. doi: 10.1016/j.autrev.2008.07.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Curotto de Lafaille MA, Kutchukhidze N, Shen S, Ding Y, Yee H, Lafaille JJ. Adaptive Foxp3+ regulatory T cell-dependent and -independent control of allergic inflammation. Immunity. 2008;29:114–126. doi: 10.1016/j.immuni.2008.05.010. [DOI] [PubMed] [Google Scholar]
  13. Curotto de Lafaille MA, Lafaille JJ. Natural and adaptive foxp3+ regulatory T cells: more of the same or a division of labor? Immunity. 2009;30:626–635. doi: 10.1016/j.immuni.2009.05.002. [DOI] [PubMed] [Google Scholar]
  14. Darce J, Rudra D, Li L, Nishio J, Cipolletta D, Rudensky AY, Mathis D, Benoist C. Divergent immunoregulation by a mutant FoxP3 protein: better arthritis, worse diabetes. Immunity. 2012 doi: 10.1016/j.immuni.2012.04.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Feuerer M, Shen YL, Littman DR, Benoist C, Mathis D. How Punctual Ablation of Regulatory T Cells Unleashes an Autoimmune Lesion within the Pancreatic Islets. Immunity. 2009;31:654–664. doi: 10.1016/j.immuni.2009.08.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Flick K, Ouni I, Wohlschlegel JA, Capati C, McDonald WH, Yates JR, Kaiser P. Proteolysis-independent regulation of the transcription factor Met4 by a single Lys 48-linked ubiquitin chain. Nat Cell Biol. 2004;6:634–641. doi: 10.1038/ncb1143. [DOI] [PubMed] [Google Scholar]
  17. Flick K, Raasi S, Zhang H, Yen JL, Kaiser P. A ubiquitin-interacting motif protects polyubiquitinated Met4 from degradation by the 26S proteasome. Nat Cell Biol. 2006;8:509–515. doi: 10.1038/ncb1402. [DOI] [PubMed] [Google Scholar]
  18. Floess S, Freyer J, Siewert C, Baron U, Olek S, Polansky J, Schlawe K, Chang HD, Bopp T, Schmitt E, et al. Epigenetic control of the foxp3 locus in regulatory T cells. PLoS Biol. 2007;5:e38. doi: 10.1371/journal.pbio.0050038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fontenot JD, Rasmussen JP, Williams LM, Dooley JL, Farr AG, Rudensky AY. Regulatory T cell lineage specification by the forkhead transcription factor foxp3. Immunity. 2005;22:329–341. doi: 10.1016/j.immuni.2005.01.016. [DOI] [PubMed] [Google Scholar]
  20. Geuking MB, Cahenzli J, Lawson MA, Ng DC, Slack E, Hapfelmeier S, McCoy KD, Macpherson AJ. Intestinal bacterial colonization induces mutualistic regulatory T cell responses. Immunity. 2011;34:794–806. doi: 10.1016/j.immuni.2011.03.021. [DOI] [PubMed] [Google Scholar]
  21. Hill JA, Feuerer M, Tash K, Haxhinasto S, Perez J, Melamed R, Mathis D, Benoist C. Foxp3 transcription-factor-dependent and -independent regulation of the regulatory T cell transcriptional signature. Immunity. 2007;27:786–800. doi: 10.1016/j.immuni.2007.09.010. [DOI] [PubMed] [Google Scholar]
  22. Hsieh CS, Liang Y, Tyznik AJ, Self SG, Liggitt D, Rudensky AY. Recognition of the peripheral self by naturally arising CD25+ CD4+ T cell receptors. Immunity. 2004;21:267–277. doi: 10.1016/j.immuni.2004.07.009. [DOI] [PubMed] [Google Scholar]
  23. Johnstone RW. Histone-deacetylase inhibitors: novel drugs for the treatment of cancer. Nat Rev Drug Discov. 2002;1:287–299. doi: 10.1038/nrd772. [DOI] [PubMed] [Google Scholar]
  24. Kim JM, Rasmussen JP, Rudensky AY. Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice. Nat Immunol. 2007;8:191–197. doi: 10.1038/ni1428. [DOI] [PubMed] [Google Scholar]
  25. Koch MA, Tucker-Heard G, Perdue NR, Killebrew JR, Urdahl KB, Campbell DJ. The transcription factor T-bet controls regulatory T cell homeostasis and function during type 1 inflammation. Nat Immunol. 2009;10:595–602. doi: 10.1038/ni.1731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li B, Samanta A, Song X, Iacono KT, Bembas K, Tao R, Basu S, Riley JL, Hancock WW, Shen Y, et al. FOXP3 interactions with histone acetyltransferase and class II histone deacetylases are required for repression. Proc Natl Acad Sci U S A. 2007;104:4571–4576. doi: 10.1073/pnas.0700298104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liston A, Nutsch KM, Farr AG, Lund JM, Rasmussen JP, Koni PA, Rudensky AY. Differentiation of regulatory Foxp3+ T cells in the thymic cortex. Proc Natl Acad Sci U S A. 2008;105:11903–11908. doi: 10.1073/pnas.0801506105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Marson A, Kretschmer K, Frampton GM, Jacobsen ES, Polansky JK, MacIsaac KD, Levine SS, Fraenkel E, von Boehmer H, Young RA. Foxp3 occupancy and regulation of key target genes during T-cell stimulation. Nature. 2007;445:931–935. doi: 10.1038/nature05478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Nagar M, Vernitsky H, Cohen Y, Dominissini D, Berkun Y, Rechavi G, Amariglio N, Goldstein I. Epigenetic inheritance of DNA methylation limits activation-induced expression of FOXP3 in conventional human CD25-CD4+ T cells. Int Immunol. 2008;20:1041–1055. doi: 10.1093/intimm/dxn062. [DOI] [PubMed] [Google Scholar]
  30. Ouni I, Flick K, Kaiser P. Ubiquitin and transcription: The SCF/Met4 pathway, a (protein-) complex issue. Transcription. 2011;2:135–139. doi: 10.4161/trns.2.3.15903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pan F, Sun L, Kardian DB, Whartenby KA, Pardoll DM, Liu JO. Feedback inhibition of calcineurin and Ras by a dual inhibitory protein Carabin. Nature. 2007;445:433–436. doi: 10.1038/nature05476. [DOI] [PubMed] [Google Scholar]
  32. Pan F, Yu H, Dang EV, Barbi J, Pan X, Grosso JF, Jinasena D, Sharma SM, McCadden EM, Getnet D, et al. Eos mediates Foxp3-dependent gene silencing in CD4+ regulatory T cells. Science. 2009;325:1142–1146. doi: 10.1126/science.1176077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pillai MR, Collison LW, Wang X, Finkelstein D, Rehg JE, Boyd K, Szymczak-Workman AL, Doggett T, Griffith TS, Ferguson TA, Vignali DA. The plasticity of regulatory T cell function. J Immunol. 2011;187:4987–4997. doi: 10.4049/jimmunol.1102173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Saouaf SJ, Li B, Zhang G, Shen Y, Furuuchi N, Hancock WW, Greene MI. Deacetylase inhibition increases regulatory T cell function and decreases incidence and severity of collagen-induced arthritis. Exp Mol Pathol. 2009;87:99–104. doi: 10.1016/j.yexmp.2009.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Tao R, de Zoeten EF, Ozkaynak E, Chen C, Wang L, Porrett PM, Li B, Turka LA, Olson EN, Greene MI, et al. Deacetylase inhibition promotes the generation and function of regulatory T cells. Nat Med. 2007;13:1299–1307. doi: 10.1038/nm1652. [DOI] [PubMed] [Google Scholar]
  36. Tarbell KV, Petit L, Zuo X, Toy P, Luo X, Mqadmi A, Yang H, Suthanthiran M, Mojsov S, Steinman RM. Dendritic cell-expanded, islet-specific CD4+ CD25+ CD62L+ regulatory T cells restore normoglycemia in diabetic NOD mice. J Exp Med. 2007;204:191–201. doi: 10.1084/jem.20061631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Thompson LJ, Valladao AC, Ziegler SF. Cutting edge: De novo induction of functional Foxp3+ regulatory CD4 T cells in response to tissue-restricted self antigen. J Immunol. 2011;186:4551–4555. doi: 10.4049/jimmunol.1003573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Tone Y, Furuuchi K, Kojima Y, Tykocinski ML, Greene MI, Tone M. Smad3 and NFAT cooperate to induce Foxp3 expression through its enhancer. Nat Immunol. 2008;9:194–202. doi: 10.1038/ni1549. [DOI] [PubMed] [Google Scholar]
  39. Vignali DAA, Collison LW, Workman CJ. How regulatory T cells work. Nature Reviews Immunology. 2008;8:523–532. doi: 10.1038/nri2343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Wan YY, Flavell RA. Regulatory T-cell functions are subverted and converted owing to attenuated Foxp3 expression. Nature. 2007;445:766–770. doi: 10.1038/nature05479. [DOI] [PubMed] [Google Scholar]
  41. Wang L, Tao R, Hancock WW. Using histone deacetylase inhibitors to enhance Foxp3(+) regulatory T-cell function and induce allograft tolerance. Immunol Cell Biol. 2009;87:195–202. doi: 10.1038/icb.2008.106. [DOI] [PubMed] [Google Scholar]
  42. Williams LM, Rudensky AY. Maintenance of the Foxp3-dependent developmental program in mature regulatory T cells requires continued expression of Foxp3. Nat Immunol. 2007;8:277–284. doi: 10.1038/ni1437. [DOI] [PubMed] [Google Scholar]
  43. Workman CJ, Collison LW, Bettini M, Pillai MR, Rehg JE, Vignali DA. In vivo treg suppression assays. Methods Mol Biol. 2011;707:119–156. doi: 10.1007/978-1-61737-979-6_9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Xiao H, Chung J, Kao HY, Yang YC. Tip60 is a co-repressor for STAT3. J Biol Chem. 2003;278:11197–11204. doi: 10.1074/jbc.M210816200. [DOI] [PubMed] [Google Scholar]
  45. Xiao Y, Li B, Zhou Z, Hancock WW, Zhang H, Greene MI. Histone acetyltransferase mediated regulation of FOXP3 acetylation and Treg function. Curr Opin Immunol. 2010;22:583–591. doi: 10.1016/j.coi.2010.08.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Zheng Y, Josefowicz SZ, Kas A, Chu TT, Gavin MA, Rudensky AY. Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cells. Nature. 2007;445:936–940. doi: 10.1038/nature05563. [DOI] [PubMed] [Google Scholar]
  47. Zheng Y, Rudensky AY. Foxp3 in control of the regulatory T cell lineage. Nat Immunol. 2007;8:457–462. doi: 10.1038/ni1455. [DOI] [PubMed] [Google Scholar]
  48. Zhou X, Bailey-Bucktrout SL, Jeker LT, Penaranda C, Martinez-Llordella M, Ashby M, Nakayama M, Rosenthal W, Bluestone JA. Instability of the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo. Nat Immunol. 2009;10:1000–1007. doi: 10.1038/ni.1774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Zhou X, Jeker LT, Fife BT, Zhu S, Anderson MS, McManus MT, Bluestone JA. Selective miRNA disruption in T reg cells leads to uncontrolled autoimmunity. J Exp Med. 2008;205:1983–1991. doi: 10.1084/jem.20080707. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS374204-supplement-01.pdf (2.1MB, pdf)

RESOURCES