|Home | About | Journals | Submit | Contact Us | Français|
Stimulation of naïve mouse CD4+Foxp3− T cells in the presence of TGF-β results in the induction of Foxp3 expression and T suppressor function. However, Foxp3 expression in these induced T regulatory cells (iTreg) is unstable raising the possibility that iTreg would not be useful for treatment of autoimmune diseases. To analyze the factors that control the stability of Foxp3 expression in iTreg, we generated OVA-specific iTreg from OT-II Foxp3-GFP knock in mice. Following transfer to normal C57BL/6 mice, OT-II GFP+ cells maintained high levels of Foxp3 expression for 8 days. However, they rapidly lost Foxp3 expression upon stimulation with OVA in IFA in vivo. This unstable phenotype was associated with a strong methylation of the Treg-specific demethylated region (TSDR) within the Foxp3 locus. Administration of IL-2/anti-IL-2 complexes expanded the numbers of transferred Foxp3+ iTreg in the absence of antigen challenge. Notably, when the iTreg were stimulated with antigen, treatment with IL-2/anti-IL-2 complexes stabilized Foxp3 expression and resulted in enhanced demethylation of the TSDR. Conversely, neutralization of IL-2 or disruption of its signaling by deletion of Stat5 diminished the level of Foxp3 expression resulting in decreased suppressor function of the iTreg in vivo. Our data suggest that stimulation with TGF-β in vitro is not sufficient for imprinting T cells with stable expression of Foxp3. Administration of IL-2 in vivo results in stabilization of Foxp3 expression and may prove to be a valuable adjunct for the use of iTreg for the treatment of autoimmune diseases.
The transcription factor Foxp3 plays a major role in the development of T regulatory cells (Treg) in the thymus and is critical for their suppressive function in vivo and in vitro (1, 2). Although Foxp3+ T cells primarily develop in the thymus (so-called natural or nTreg), it has become clear that naïve CD4+ T cells selected as Foxp3− during development in the thymus have the potential to be converted into functional Foxp3+ Treg cells (so-called adaptive or induced iTreg) in peripheral lymphoid organs or in tissue culture (3–6). Transforming growth factor beta (TGF-β) plays a critical role in the induction of Foxp3 expression in vitro and in vivo. iTreg generated in vitro have been shown to efficiently suppress inflammation in Scurfy mice, and to inhibit effector cell function in several different animal models including autoimmune gastritis, asthma, EAE, and inflammatory bowel disease (5–10).
Epigenetic regulation of gene expression has been of the most active fields of research in immunology in the past years. Mechanisms such as histone modifications and DNA methylation contribute to a stable epigenetic imprint of gene expression programs and govern cell fate decisions in developing lymphocytes (11). DNA methylation is the most well established epigenetic mechanism involved in Treg development and is linked to stable gene expression patterns. Both mouse and human studies of the methylation of CpG motifs in the Foxp3 locus of nTreg revealed complete demethylation within an evolutionary conserved region upstream of exon 1, named TSDR (Treg-specific demethylated region) (12–15). In the mouse, this CpG island is approximately 5 kb upstream of the transcriptional start site (TSS). Demethylation of the TSDR region appears to be a specific marker of nTreg, as both CD4+Foxp3− cells and in vitro generated Foxp3+ iTreg display almost full methylation state of TSDR (16, 17). Surprisingly, induction of iTreg in vivo by targeting delivery of antigen to DC with the DEC-205 antibody in the absence of pro-inflammatory signals resulted in the induction of Foxp3+ iTreg with complete demethylation of the TSDR region and stable Foxp3 expression (18). Taken together these studies suggest that TSDR demethylation does not act as an on/off switch for Foxp3 expression, but instead determines the stability of Foxp3 expression. Control of TSDR demethylation is the key for maintaining stable Foxp3 expression and a fully functional Treg cell phenotype in iTreg.
The demonstration that in vitro induced iTreg have a fully methylated TSDR region has raised doubts about the potential therapeutic usefulness of iTreg for the treatment of autoimmune disease despite several studies demonstrating not only their effectiveness (5–10), but also their sustained expression of Foxp3 in vivo (7, 8). Here, we have re-explored the factors that regulate the expression of Foxp3 in iTreg in vivo. In contrast to previous studies that used T cell populations enriched in Foxp3+ iTreg, we generated antigen-specific iTreg from CD4+Foxp3− T cells from OT-II TCR transgenic mice bred to Foxp3-GFP KI mice (19). To avoid any possible confounding influence of small numbers of contaminating, activated Foxp3− T cells, GFP+ cells were sorted from the induction cultures prior to transfer to normal C57BL/6 mice. We demonstrate that Foxp3+ iTreg maintain high levels of Foxp3 expression in the absence of TCR stimulation for 8 days after transfer, but rapidly lose Foxp3 expression upon antigen stimulation. The simultaneous administration of IL-2 in the form of an anti-IL-2/IL-2 complex and TCR stimulation resulted in a stabilization of Foxp3 expression accompanied by enhanced demethylation of the TSDR. Furthermore, by neutralizing IL-2 in vivo, we found that local secretion of IL-2 by conventional T cells also augmented the stability of Foxp3 expression in antigen-specific iTreg, as well as potentiating their suppressive activity in vivo. Thus, in addition to controlling nTreg homeostasis and expansion, IL-2 also can contribute to the regulation of the methylation of TSDR and the stability of Foxp3 expression in iTreg in vivo. The implications of these results for the clinical use of iTreg will be discussed.
Foxp3gfp mice (IRES-GFP knockin into the Foxp3 locus) (19) on the C57BL/6 background were kindly provided by Dr. Vijay Kuchroo (Harvard Medical School, Boston, MA). OVA-specific TCR-transgenic OT-II mice were obtained from Taconic Farms and bred to Foxp3gfp mice to generate OT-II Foxp3-GFP KI mice. B6.SJL (CD45.1) and OT-II CD45.1 Rag −/− were obtained from Taconic Farms. Thy1.1 mice on the C57BL/6 background were purchased from The Jackson Laboratory. TCR-transgenic OT-II Stat5 fl/fl and OT-II Stat5 fl/flCD4-Cre mice were previously described (20). All animals used for this study were 4–8 wk of age. They were housed and handled according to National Institutes of Health institutional guidelines under an approved animal protocol.
For purification of DCs, mouse spleens were fragmented and digested for 30 min at 37° C in the presence of liberase blendzyme II (Roche) and DNase (2 μg/ml) (Roche) in complete medium (modified RPMI 1640 supplemented by 10% FBS (HyClone),50 μM 2-ME (Sigma-Aldrich), 1% sodium pyruvate, 1% nonessential amino acids, 1% HEPES, 100 U/ml penicillin and 100 μg/ml streptomycin, and 2 mM L-glutamine; all from Invitrogen). They were stained with anti-CD11C+ (HL3 clone) (BD Pharmingen) and purified by FACS sorting. The purity was higher than 95%.
T cells were obtained from pooled mouse spleens and lymph nodes and the CD4+ fraction was first purified on the AutoMACS Cell Separator using anti-mouse CD4 beads (Miltenyi Biotec). The enriched CD4+ fractions were then further separated into conventional T or Treg populations by FACS sorting for the CD4+GFP− or CD4+GFP+ fractions using either the FACS Vantage Diva or FACS Aria flow cytometers (BD Biosciences). For different experiments as indicated, CD4+GFP− fraction or Vα2+Vβ5+ OT-II transgenic CD4+CD25− T cells were sorted to induce GFP in GFP-Foxp3 KI cells or Foxp3 expression in OT-II TCR Tg Stat5fl/fl or Stat5fl/fl CD4-Cre cells, respectively. The postsort purity for each cell type was higher than 98% and the Foxp3 purity for Treg cells was higher than 95%.
Single-cell suspensions were stained using the following conjugated Abs according to the manufacturer’s protocol: anti-mouse Thy1.2 (53-2.1) and anti-Vβ5+ (MR 9-4) (BD Pharmingen), anti-mouse CD4 (L3T4), anti-mouse CD45.2 (104), anti-mouse CD45.1 (A20), anti-Vα2+ (B20.1), anti- Thy1.1 (HIS51), anti-mouse Foxp3 (FJK-16s), anti-IFN-γ (XMG1.2), anti-IL-2 (JES6-5H4) and anti-IL-17 (eBio17B7) (all from eBioscience). For staining of Foxp3 or intracellular cytokines, the cells were fixed and permeabilized using a fixation/permeabilization kit (eBioscience). Acquisition was done on an LSR II (BD Biosciences)equipped with a FACS Diva digital upgrade. Analysis was done using FlowJo software (Tree Star).
WT or OT II CD4+GFP− cells were stimulated with plate-bound anti-mouse CD3 (2 μg/well) and anti-CD28 (1 μg/well) in the presence of rhTGF-β1 (5 ng/ml, PeproTech) in complete RPMI 1640 medium (Invitrogen) containing 100 U/ml rhIL-2 supplemented by 10% FBS (HyClone),50 μM 2-ME (Sigma-Aldrich), 1% sodium pyruvate, 1% nonessential amino acids, 1% HEPES, 100 U/ml penicillin and 100 μg/ml streptomycin, and 2 mM L-glutamine (all from Invitrogen). The cells were incubated at 37°C in 5% CO2 for 4 to 5 days before sorting the GFP+ fractions. For the in vitro differentiation of CD4+ T cells from OT-II Stat5 fl/fl or Stat5 fl/flCD4-Cre mice, the differentiation cultures were supplemented with all-trans retinoic acid (ATRA, Sigma-Aldrich) at a final concentration of 1μM. Cultures containing ATRA were protected from light during the culture.
Purified GFP+ cells were washed once in 0.1% BSA in PBS, and labeled with 1 μl of 10 mM CFSE (Invitrogen) at a density of 2 × 106 cells/ml in PBS containing 0.1% BSA for 8 min at 37°C in the dark. The reaction was stopped with RPMI 1640 with 10% FCS and washed twice. For in vivo studies, unless stated otherwise, 1 × 106 cells were injected into WT mice via the tail vein. For analysis, GFP fluorescence was quenched by a fixation and permeabilization procedure using the fixation/permeabilization kit (eBioscience) permitting analysis of CFSE-dilution in the absence of GFP fluorescence.
6 hours after cell transfer, mice were immunized with OVA323–339 (50 μg, synthesized by the peptide synthesis facility, NIAID) in IFA in each hind flank. Inguinal lymph nodes were collected on day 5 after immunization, unless stated otherwise. The rat anti-IL-2 mAb (IgG2a, clone S4B6) was purified from as cites using protein G columns (Amersham Biosciences). S4B6 (0.5 mg) was injected i.p. on day 0 and day 1 after adoptive transfer. IL-2 mAb (IgG2a, clone JES6-1-A12) were purchased from Bioxcell. To prepare IL-2-anti-IL-2 mAb complexes, rmIL-2 (1 μg, PeproTech) was mixed with JES6-1 (5 μg) at the optimal 1:2 molar ratio and incubated at 37°C for 10 min. Mice were injected i.p. on day 0, 1 and day 2 after adoptive transfer.
OTII transgenic CD4 T cells (CD45.1+Thy1.2+, 0.5 × 106) were CFSE-labeled and co-transferred with either OTII transgenic activated conventional CD4+ T cells (3 × 106) or GFP+ iTreg (CD45.2+Thy1.2+, 3 × 106) into WT Thy1.1 congenic mice via the tail vein. 6h after cell transfer, mice were immunized with OVA323–339 peptide in IFA in each hind flank. Inguinal lymph nodes were collected on day 5 after immunization.
Foxp3+ or Foxp3− cells from male mice were used for the methylation analysis. Sorted OT-II iTreg (GFP+CD45.2+) (3 × 106) were transferred into CD45.1+ mice (3 mice per group). On day 0, recipients were left untreated or immunized with OVA/IFA. IL-2/anti-IL-2 complexes were injected from day 0 i.p. daily for 3 days. 5 days later, LNs or DLNs were harvested and donor cells were sorted into different population based on the expression of GFP and CD45.2 and subjected to methylation analysis. The sorting purity was higher than 95%. As controls, GFP+ nTreg and GFP− conventional CD4+ T cells that have been activated by plate-bound anti-CD3/anti-CD28 were adoptively transferred into recipients and left untreated. 5 days later, the GFP+ or GFP− donor cells were isolated. Genomic DNA was purified from isolated cells with DNasy Blood & Tissue kit (Qiagen) and subjected to DNA methylation analysis. Bisulfide modification and TSDR-specific CpG methylation analysis using Pyrosequencing was performed by EpigenDx (Worcester, MA). In this study, 9 CpGs (ADS568, X; 7161111–7161273) were read and analyzed among 14 CpGs of the TSDR.
Recent studies have suggested that murine Foxp3+ iTreg induced by short-term in vitro stimulation in the presence of TGF-β rapidly lose Foxp3 expression following re-stimulation in cultures without exogenous TGF-β (12, 18, 21). One problem in the interpretation of these studies is the possibility of outgrowth of non-Foxp3+ T cells during restimulation. To exclude this possibility, we repeated these studies using Foxp3-GFP KI mice and sorted the induced Foxp3+ T cells after the induction culture. The iTreg population was then either re-stimulated with plate-bound anti-CD3/anti-CD28 antibodies in the absence or presence of exogenous TGF-β, or cultured with IL-2 alone (Fig. 1A), and cell proliferation was measured by CFSE dilution. In agreement with previous studies (12, 18, 21), re-stimulation for 3 days resulted in loss of Foxp3 expression by the majority of the cells with only 12% remaining Foxp3+ on day 3. In contrast, almost 95% of cells maintained Foxp3 expression when re-stimulated in the presence of TGF-β. Although culture of the iTreg with IL-2 alone induced a vigorous proliferative response as measured by CFSE dilution, almost all of the iTreg retained Foxp3 expression. We next examined whether the iTreg that had lost Foxp3 converted into effector CD4+ T cells able to produce inflammatory cytokines including IL-2, IL-4, IL-17 or IFN-γ. Upon PMA/ionomycin stimulation for 4 h, very few cells produced IL-2, IL-4 and IL-17 (data not shown). A small number (15.2%) of IFN-γ producing cells was found in the Foxp3− population, but this number was only slightly higher than the percentage of IFN-γ producers in the Foxp3+ T cells stimulated in the presence of TGF-β (10.5%) or with IL-2 alone (7.2%) (Fig. 1A, lower panel). These studies demonstrate that the loss of Foxp3 expression in iTreg following re-stimulation did not results into their conversion into effector T cells. .
To further analyze the role of TCR stimulation on the maintenance of Foxp3 expression in iTreg, we performed a more detailed kinetic analysis and compared the effects of TCR stimulation on iTreg with nTreg (Fig. 1B). Foxp3− T cells were cultured for 5 days in the presence of TGF-β, washed, and then Foxp3+ T cells were isolated by sorting. By day 3 of the induction culture, more than 50% of the Foxp3− cells had been induced to express Foxp3 and the percentage of Foxp3+ T cells increased to ~90% after 5 days. nTreg expanded in parallel maintained high levels of Foxp3 expression for 5 days. When the sorted Foxp3+ iTreg were stimulated with anti-CD3, they rapidly lost Foxp3 expression and by day 4 of the re-stimulation culture only low percentages of Foxp3+ cells could be detected. In contrast, TCR stimulated nTreg maintained high levels of Foxp3, as did iTreg maintained in IL-2 alone. Taken together, these studies demonstrate that TCR, but not IL-2, stimulation of iTreg results in rapid down-regulation of Foxp3 expression in vitro, but the loss of Foxp3 is not accompanied by conversion to T effector cells, particularly Th17 cells.
To examine whether TCR stimulation also induced down-regulation of Foxp3 expression in iTreg in vivo, we bred the Foxp3-GFP mice to OT-II TCR transgenic mice. We initially sorted CD45.2+ nTreg cells from these mice, labeled them with CFSE, and transferred them to CD45.1 wild type (WT) mice (Fig. 2A). Five days following immunization with OVA in IFA, most of the transferred nTreg in draining lymph nodes (DLN) maintained Foxp3 expression (87%, Fig. 2A, middle). 34% of Foxp3+ cells recovered from DLN had not divided wheras 66% of cells had undergone at least one round of CFSE dilution (Fig. 2A, right).
We next sorted Foxp3− T cells from the OT-II-GFP mice and induced expression of Foxp3 by stimulation with plate-bound anti-CD3/CD28, IL-2 and TGF-β for 5 d. The induced GFP+Foxp3+ iTreg were sorted, CFSE-labeled, and then adoptively transferred into CD45.1+ mice. Controls recipients received equivalent numbers of cells, but were not immunized. Following immunization with OVA in IFA, the frequency and absolute numbers of transferred CD45.2+ cells was increased in DLN by day 3 after transfer, reached a maximum by day 5, and decreased by day 8 (Fig. 2B, 2D). By day 3 after transfer, 37% of the donor cells had lost Foxp3 and 69% of the cells had lost Foxp3 on day 5 (Fig. 2C, 2D). In contrast, in the absence of immunization, ~90% of the cells remained Foxp3+. The iTreg that had lost Foxp3 expression did not produce IL-2, IFN-γ or IL-17 (Fig. 2E). These results are consistent with our in vitro experiments (Fig. 1) and demonstrate that TCR stimulation induces down-regulation of Foxp3 expression, but that the cells that have lost Foxp3 did not convert to an effector phenotype.
Approximately 30% of transferred cells remained Foxp3+ on day 5 after OVA stimulation in vivo (Fig. 2). To determine whether these cells represented a potential subset induced by TGF-β with a stable phenotype, we sorted the Foxp3+ cells (CD45.2+GFP+, Fig. 3A) from the recipient mice on day 5 after immunization, and transferred them to a second normal (CD45.1+) recipient. Half of the recipient mice were immunized with OVA/IFA. The results clearly indicate that Foxp3 expression was also unstable after the second round of antigen stimulation, as 64% of the transferred cells lost Foxp3 (Fig. 3B). Interestingly, all the transferred cells retained Foxp3 without stimulation.
To further analyze impact of the strength of TCR signals on Foxp3 expression, we performed co-transfer studies in which antigen-specific iTreg were stimulated with a fixed number of DC (DC: iTreg 1:20) that had been pulsed with varying amounts of antigen. Simulation of the iTreg with low dose (0.05 mM OVA) pulsed DC induced a modest decrease of Foxp3 expression (73% Foxp3+), whereas a 10-fold increase of peptide concentration (0.5 mM) induced a much greater loss of Foxp3 expressing cells (35% Foxp3+) (Fig. 4, upper panel). Similarly, when we co-transferred DC pulsed with a fixed amount of peptide, but varied the DC: iTreg ratio, the stronger TCR stimulus induced by the higher number of DC resulted in a marked decrease in Foxp3 expression (Fig. 4, lower panel).
Our initial in vitro studies demonstrated that stimulation of iTreg by IL-2 alone induced their expansion by preserved Foxp3 expression. As these studies raised the possibility of an effect of IL-2 on the stability of Foxp3 expression, we took advantage of a recently described (22, 23) approach to greatly enhance the biological activity of IL-2 in vivo by administration of IL-2 in the form of IL-2/anti-IL-2 complexes. These studies demonstrated that administration of IL-2/anti-IL-2 complexes markedly expanded the numbers of thymic-derived CD25+ Foxp3+ nTreg in WT animals. We first tested the effects of complex administration on iTreg in vivo in the absence of TCR stimulation. CFSE-labeled OT-II iTreg were transferred to normal recipients and the mice were treated with three daily i.p. injections of IL-2/anti-IL-2 complexes. Stimulation with IL-2 induced cell cycling in iTreg without loss of Foxp3 expression (Fig. 5A, left panels). More importantly, when the transferred iTreg were stimulated with antigen, administration of the IL-2/anti-IL-2 complexes significantly prevented loss of Foxp3 expression while at the same time enhancing the proliferation of the Foxp3+ iTreg (Fig. 5A, right panels). As a consequence, the increase in the total number of transferred cells seen in the IL-2 treated mice mostly resulted from expansion of Foxp3+ iTreg (Fig. 5B).
Previous studies suggested that stable iTreg that expressed enhanced TSDR demethylation could only be generated in vivo by targeting antigen to DC under non-inflammatory conditions (6). Given that administration of IL-2/anti-IL-2 induced strong proliferation of in vitro converted iTreg, we asked whether the treatment could also modify the demethlyation status of the TSDR. The TSDR methylation status of the foxp3 locus was analyzed by sorting both the Foxp3+ and Foxp3− transferred cells (Fig. 5A). Before adoptive transfer, the TSDR of the sorted Foxp3+ iTreg was approximately 89% methylated. Five days after transfer, the level of methylation (~83.7%) of the TSDR of Foxp3+ cells from mice treated IL-2/anti-IL-2 alone was comparable to level of the Foxp3+ T cells from the non-treated mice one (~88.7%, Fig. 5C). However, a striking difference in the level of demethylation of the TSDR was observed in the Foxp3+ T cells obtained from mice that had only been TCR stimulated and mice that had received both TCR and IL-2 stimulation. Antigen stimulation alone resulted in modest demethylation (65.7%) of the TSDR, while stimulation with the antigen and IL-2 resulted in highly efficient level of demethylation (~21.1%) that approximated the level of demethylation (~4.5%) seen in expanded nTreg. The results demonstrated that both TCR and IL-2 signals were required to induce epigenetic modification of regulatory regions in the foxp3 locus leading to in vitro converted iTreg with stable Foxp3 expression in vivo.
As administration of IL-2 enhanced the stability of Foxp3 expression in iTreg in vivo, we next examined whether endogenous IL-2 plays a role in maintaining Foxp3 expression in iTreg in vivo. Neutralization of IL-2 by treatment with a mAb (S4B6) that had been previously shown to be effective in vivo (24) had no significant effect in lowering Foxp3 expression in unstimulated iTreg. Interestingly, we observed that the MFI of Foxp3 in iTreg treated with S4B6 was decreased (1,662) compared with the MFI (2,567) in Foxp3+ cells from untreated mice. This finding supports the critical role of IL-2 in the maintenance of Foxp3 expression in vivo. Furthermore, in the presence of S4B6, a markedly enhanced loss of Foxp3 was observed from antigen-stimulated cells (Fig. 6A). As IL-2 primarily signals via Stat5, we examined the stability of Foxp3 expression in Stat5-deficient (−/−) iTreg generated in vitro from OT-II Stat5 fl/flCD4-Cre mice. We used a protocol (20) which had been previously been shown to induce high levels of Foxp3 in Stat5−/− CD4+ T cells by combing ATRA with TGF-β in the induction cultures. Following induction in vitro, approximately 99% of the Stat5 fl/fl CD4 cells and 97% of the Stat5 fl/fl CD4-cre cells expressed Foxp3 (Fig. 6B). They retained Foxp3 (> 95%) expression upon transfer to non-stimulated recipients (Fig. 6C). In contrast, antigen stimulation resulted in a greater loss of Foxp3 expression from Stat5−/− T cells compared to Stat5 fl/fl Foxp3+ CD4+ cells generated with ATRA and TGF-β (Fig. 6C lower panel). These findings are again consistent with a requirement for IL-2 signaling for the maintenance of Foxp3 expression.
The studies described above demonstrating the rapid loss of Foxp3 expression from iTreg upon TCR stimulation in vivo are in conflict with multiple studies that demonstrated that in vitro generated iTreg are effective suppressors in different autoimmune and inflammatory disease models (5–10, 25). Importantly, in several of these studies, iTreg were shown to maintain high levels of Foxp3 expression for prolonged periods of time in an inflammatory microenvironment in vivo (7, 8). To directly compare our present results with these earlier studies, we co-transferred GFP+ iTreg generated from OT-II mice with naïve OT-II CD4+Foxp3− T cells to normal recipients followed by immunization with OVA in IFA. As a control, we co-transferred activated Foxp3− OT-II cells with naïve Foxp3− OT-II cells (Fig. 7A). In the presence of the antigen-specific iTreg marked suppression of the proliferation (Fig. 7B, top panel) of the effector T cells (CD45.1+) was observed accompanied by a marked decrease in the absolute number of recovered T effectors (Fig. 7C). In contrast, no suppression of proliferation was seen when activated Foxp3− T cells were co-transferred (Fig. 7B, bottom panel, Fig. 7C). In contrast to the studies when iTreg were transferred alone, a high percentage of iTreg co-transferred with T effector cells maintained Foxp3 expression (69%). To determine whether factors in the microenvironment, such as IL-2 secreted by effectors, could contribute to the maintenance of Foxp3 stability in iTreg, we neutralized IL-2 by treating the recipient animals with anti-IL-2. Interestingly, anti-IL-2 augmented the proliferation of the effector cells when the activated Foxp3− cells were co-transferred (Fig. 7B). It is likely that this modest augmentation results from neutralization of the competition for IL-2 by the pre-activated effector T cells. However, neutralization of IL-2 resulted in down-regulation of Foxp3 expression (26%) and significantly reversed iTreg-mediated suppression of proliferation as measured by CFSE dilution and in the absolute numbers of T effectors recovered (Fig. 7B, 7C). Thus, IL-2 is critically required for required for homeostatic maintenance of iTreg, for maintenance of Foxp3 expression, and for initiation of their suppressive function.
TGF-β mediated conversion of CD4+Foxp3− cells has been used in various models to generate Foxp3+ iTreg with suppressive capacity (5–10, 25). However, a number of studies have now demonstrated that Foxp3 expression in iTreg is unstable upon re-stimulation in the absence of exogenous TGF-β (18, 21, 26). In this report, we have re-evaluated the factors that control the stability of Foxp3 expression in iTreg both in vivo and in vitro. We have mainly used “pure” iTreg generated from Foxp3 GFP-KI mice to avoid any complicating factors produced by contaminating Foxp3− T cells. We have confirmed the observations of other groups (18, 21) that Foxp3 expression is rapidly lost when iTreg are re-stimulated in vitro in the absence of TGF-β. In contrast with a previous study indicating that both IL-2 and TGF-β were required to sustain expression of Foxp3 (27), we now demonstrate that in the absence of stimulation via the TCR, Foxp3 expression is remarkably stable in vitro for at least 2 weeks of culture even when the iTreg proliferate in response to the addition of exogenous IL-2. TCR stimulated iTreg that had lost Foxp3 expression did not become cytokine producing Th1 or Th17 cells although a low levels of IFN-γ producing cells could be detected in both Foxp3+ and Foxp3− populations. Nevertheless, as reported by Polansky et al (18) the TSDR of the induced Foxp3+ T cells remained fully methylated.
The major focus of this report was the fate of iTreg that had been transferred to normal recipients. Some studies (21) have reported that iTreg in vivo rapidly lost Foxp3 expression and suppressor function, but these observations run counter to the studies that have demonstrated the remarkable therapeutic efficacy of iTreg in several autoimmunity models as well as their capacity to rescue Scurfy mice when administered shortly after birth. The results of our in vivo studies parallel the in vitro experiments. In the absence of TCR stimulation, iTreg isolated from DLN or spleen maintain high levels of Foxp3 expression for long periods of time (up to one month, data not shown) and exhibit minimal proliferative responses. This observation differs from the studies of Selveraj and Geiger (21) who demonstrated that iTreg rapidly lost Foxp3 expression in most sites in the absence of TCR stimulation and only a small number of Foxp3+ T cells could be detected in LN or bone marrow. One possible difference between the studies is that they transferred polyclonal Treg, while we transferred iTreg generated from OT-II mice on a RAG−/− background that expressed a restricted TCR repertoire. The iTreg that had retained Foxp3 expression in the absence of stimulation retained a methylated phenotype of their TSDR. Upon stimulation with antigen by immunization with peptide in IFA or antigen-pulsed spleen cells, the OVA-specific iTreg rapidly lost Foxp3 expression. The cells that had lost Foxp3 proliferated more rapidly than those that retained Foxp3 expression, but did not differentiate to Th1 or Th17 cells. The fate of the iTreg in vivo was markedly dependent on the strength of the TCR signal, as weak stimulation with lower doses of peptide or lower numbers of DC that resulted in a decreased proliferative response induced less loss of Foxp3. We could find no evidence for a stable subpopulation of Foxp3+ cells as suggested by Selveraj and Geiger (21), as the minor population of Foxp3+ cells that remained after stimulation with antigen in vivo, when transferred to a new host, lost Foxp3 expression upon a second exposure to antigen in vivo.
Our previous studies (8) on the treatment of Scurfy mice with iTreg strongly suggested that factors present in the inflammatory environment of the Scurfy host might potentiate Foxp3 expression in iTreg, as iTreg that had been transferred to Scurfy mice expanded and retained expression of Foxp3, while iTreg transferred to normal neonates lost Foxp3 expression. As culture of iTreg with IL-2 resulted in stable expression of Foxp3, we first evaluated whether treatment of mice with IL-2/anti-IL-2 complexes resulted in stabilization of Foxp3 expression. Treatment of mice with the complexes alone enhanced the proliferation of the transferred iTreg that remained Foxp3+ in the absence of stimulation. Surprisingly, treatment of mice that had received iTreg and were subsequently immunized markedly enhanced Foxp3 expression. Moreover, analysis of the methylation status of the TSDR indicated that treatment of mice with both antigen stimulation and IL-2, but not IL-2 alone, resulted in significant demethylation of the TSDR. Thus, IL-2 alone is not sufficient to induce epigenetic imprinting of Foxp3 expression. Furthermore, our studies demonstrate that culture of iTreg either with IL-2 alone or with plate-bound anti-CD3 and IL-2 did not lead to significant demethylation of the TSDR (89.5±5.1%, 76.1±6.5%, methylated). It will be interesting in future studies to clarify the differences between the in vitro and in vivo effects of antigen and IL-2 treatments on the methylation status of the TSDR.
We used several additional approaches to confirm that IL-2 played an important role in the maintenance of Foxp3 expression in vivo. First, following TCR stimulation, neutralization of IL-2 potentiated the loss of Foxp3 expression. Secondly, iTreg generated from Stat5−/− mice demonstrated a greater loss of Foxp3 expression than iTreg from WT mice following antigen stimulation in vivo. Lastly, we used a co-transfer model in which antigen-specific iTreg inhibit the expansion of naïve antigen-specific T cells following antigen stimulation. In contrast to the stimulation of iTreg alone, stimulation of iTreg in the presence of effector cells resulted in significant maintenance of Foxp3. Both the maintenance of Foxp3 expression and T suppressor function in vivo were reduced when endogenous IL-2 was neutralized with anti-IL-2. Because optimal TCR and CD28 engagement can induce IL-2-independent cell cycle progression in vivo (27), depletion of endogenous IL-2 did not inhibit effector T cell proliferation. These results demonstrate that IL-2 is not only essential for the development and homeostasis of nTreg (28, 29), but that IL-2 produced by T effector cells is also critical for the function of iTreg in vivo.
It is clear from our studies and those of other groups that the conditions for induction of TSDR demethylation are lacking in TGF-β induction cultures in vitro. Both the biologic and molecular signals leading to TSDR demethylation in nTreg and iTreg remain elusive. It is unlikely that the induction of CREB/ATF by TCR stimulation or Stat5 activation by IL-2 stimulation are directly responsible for inducing demethylation as both CREB/ATF and Stat5 only bind to demethylated TSDR (13). Foxp3 itself is also unlikely to regulate demethylation as Foxp3 in the form of Foxp3-Runx-1-Cbf-β complexes also only bind to the TSDR after demethylation (30). It thus remains to be determined how the synergistic interaction of IL-2 signals with TCR signals promotes in the induction of TSDR demethylation.
One major issue that is not resolved by these studies is the development of an optimal protocol for the in vitro generation of iTreg that stably express Foxp3 and concomitantly have a demethylated TSDR. This issue is of greater significance for the generation of human iTreg in vitro as it has been shown that TGF-β stimulation of human cells, while inducing Foxp3 expression, fails to induce functional iTreg (31, 32). One criticism that has been raised about the in vitro methods used by most investigators is that the result in the rapid induction of Foxp3 expression in the presence of CD28-mediated co-stimulation and proliferation is detrimental to the generation of stable iTreg and that the addition of agents that inhibit the mTOR pathway increases the stability of Foxp3 expression in iTreg generated in vitro (25). However, it remains to be demonstrated that inhibition of the mTOR pathway results in demethylation of the TSDR. Previous studies have strongly suggested that IL-2 is absolutely required for the generation of iTreg in vitro (32, 33) and the present studies demonstrate that the administration of IL-2 in vivo can stabilize Foxp3 expression and promote demethylation of the TSDR. It is likely that the success of previous studies (5–10) that demonstrated the therapeutic effects of iTreg in several distinct models of autoimmunity was secondary to the production of IL-2 by the autoreactive effector cells. Future studies should address whether IL-2 can function as an adjuvant in vivo for the prevention and treatment of autoimmune disease, graft rejection, or graft versus host disease.
We thank the NIAID Flow Cytometry Section, particularly Bishop Hague, Carol Henry and Elina Stregevsky for cell sorting.
This work was supported by funds from the Intramural Program of the National Institute of Allergy and Infectious Diseases.