IL-2 Receptor β-Dependent STAT5 Activation Is Required for the Development of Foxp3⁺ Regulatory T Cells

Mice

IL-2^−/−, IL-2Rα^−/−, IL-2Rβ^−/−, γ_c^−/−, and IL-7Rα^−/− mice were obtained from The Jackson Laboratory and maintained in our breeding colony. In some experiments, IL-2^−/− × rag-2^+/− and IL-2Rβ^−/− × rag-2^+/− mice were used; rag-2 heterozygosity had no effect on results since identical findings were obtained with IL-2^−/− and IL-2Rβ^−/− mice. IL-15^−/− and rag-2^−/− mice were obtained from Taconic Farms. IL-2^−/− × IL-15^−/− mice were generated by crossing IL-2^−/− × rag2^−/− with IL-15^−/− mice; STAT5b-CA × IL-2Rβ^−/− mice were generated by crossing STAT5b-CA × IL-2Rβ^+/− mice with IL-2Rβ^−/− × rag-2^−/− mice. STAT5a/b^FL/FL and CD4-Cre mice were provided by Dr. L. Hennighausen (National Institutes of Health, Bethesda, MD) and Dr. K. Hogquist (University of Minnesota, Minneapolis, MN), respectively. B6.C.H2bm12/KhEg (termed bm12) mice were purchased from The Jackson Laboratory. C57BL/6 mice were purchased from the National Institutes of Health or Taconic Farms. The foxp3 transgenic mice were generated as described (18) and provided by Dr. F. Ramsdell (Zymogenetics, Seattle, WA). All mouse strains used were backcrossed at least 10 generations onto the C57BL/6 genetic background except for STAT5a/b^FL/FL mice. Mice used for analysis were 4–6 wk old, unless otherwise noted. The University of Minnesota Institutional Animal Care and Use Committee approved all animal experiments.

Intracellular flow cytometry

Spleen, thymus, and lymph nodes were harvested and single-cell suspensions were generated by disruption with ground glass slides. Before enumeration, RBC were removed from spleen preparations by ammonium chloride lysis. Abs against the following cell surface markers were obtained from eBioscience: CD8 (53-6.7), CD25 (PC61.5), and glucocorticoid-induced TNFR family-related gene (DTA-1). In some experiments, potential contaminating cells were removed by gating out cells expressing B220, CD19, GR-1, CD11b, TER-119, NK1.1, and TCRγδ. Additional Abs directed at the cell surface molecule CD4 (RM4-5) were obtained from BD Pharmingen. Following surface staining, samples were fixed and permeabilized overnight using the PE anti-mouse/rat Foxp3 staining set obtained from eBioscience. Following fixation, cells were stained with either PE-labeled rat/anti-mouse Foxp3 (FJK-16s) or PE-labeled rat IgG2a,κ (P3) isotype control. Labeled cells were then collected and analyzed using a LSRII flow cytometer (BD Biosciences). Analysis of flow cytometry data was conducted using FlowJo software (Tree Star). Staining for intracellular phospho-STAT5 and phospho-Erk was done as previously described (19).

Proliferation assays

For proliferation assays involving IL-2Rβ^−/− × foxp3tg mice, CD4⁺CD25⁻ T cells were isolated from lymph nodes (LN) using a wire mesh to obtain a single-cell suspension. Cells were resuspended in medium consisting of DMEM (BioWhittaker), 10% FBS (HyClone), 50 μM 2-ME (Sigma-Aldrich), 10 mM HEPES buffer, 1 mM sodium pyruvate (Invitrogen Life Technologies), amino acid supplements (1.5 mM l-glutamine, l-arginine, and l-asparagine; Sigma-Aldrich), and antibiotics (100 U/ml penicillin and 100 mg/ml streptomycin; Sigma-Aldrich). Importantly, the percentage of Foxp3⁺ cells in the sorted CD4⁺ IL-2Rβ^−/− × foxp3tg cells, CD4⁺foxp3tg cells, and WT CD4⁺CD25⁺ cells was roughly equivalent (63–75% Foxp3⁺). LN cells were depleted of NK cells (hybridoma PK136) and CD8⁺ T cells (hybridoma 2.43) by incubation with mAb, followed by passage through a goat anti-mouse and goat anti-rat Ig-coated column (Cedarlane Laboratories). CD25⁺ T cells were depleted using anti-CD25 PE (hybridoma PC61; BD Pharmingen) and MACS anti-PE MicroBeads followed by LS MACS separation columns (Miltenyi Biotec). T cell- and NK cell-depleted stimulator cells were prepared from bm12 spleens, incubated with anti-Thy1.2 mAb (hybridoma 30H-12) and NK mAb (hybridoma PK136) followed by incubation with rabbit complement (Nieffenegger). Cells were irradiated at 30 Gy. CD4 T cells from WT, foxP3tg, IL-2Rβ^−/−, and IL-2Rβ^−/− × foxP3tg mice were enriched from the spleen using anti-CD4 PE (hybridoma RM4-5) followed by MACS separation. Cells were further purified using a FACSAria sorter to 99% purity. One hundred thousand CD4⁺CD25⁻ T cells were mixed with equal numbers of bm12 stimulators and indicated numbers of sorted CD4⁺ T cells from the various strains in 200 μl in 96-well plates. Cultures were incubated for 7 days in DMEM at 37°C and 10% CO₂ and pulsed with [³H]thymidine at 1 μCi/well (Amersham Biosciences) for the last 16–18 h of culture. Cells were harvested and counted in the absence of scintillation fluid using a beta plate reader (Packard Instrument) and analyzed in quadruplicates.

Retroviral bone marrow chimeras

The murine WT IL-2Rβ cDNA was used as a template to generate both IL-2Rβ-STAT5 and IL-2Rβ-ΔAH constructs. In brief, the IL-2Rβ-ΔAH construct encodes for a receptor that is truncated after serine-326. The IL-2Rβ-STAT5 construct was made by overlap extension PCR and encodes a chimeric receptor consisting of the first 326 aa of the murine WT IL-2Rβ construct fused to a sequence encoding the distal aa 494–505. WT-IL-2Rβ, IL-2Rβ-STAT5, and IL-2Rβ-ΔAH cDNAs were cloned into a modified pMIGR retroviral vector and retroviral particles were produced using 293T cells and a replication-deficient helper virus. Four-week-old IL-2Rβ^−/− donor mice were treated with 5-fluorouracil to enrich for long-term hemopoietic stem cells, and 4 days posttreatment bone marrow was harvested from donor mice. High titer viral supernatants were used to infect donor bone marrow with either WT-IL-2Rβ, IL-2Rβ-STAT5, or IL-2Rβ-ΔAH constructs as previously described (20). Following retroviral infection, donor bone marrow was introduced into sublethally irradiated rag2^−/− recipient mice (650 rad). Eight weeks after reconstitution, thymii and spleens were harvested from recipient mice and analyzed by flow cytometry.

Chromatin immunoprecipitation (ChiP) assays

Spleen and LN were isolated from C57BL/6 mice and single-cell suspensions were obtained by disrupting organs with ground glass slides. Cells were resuspended in 1× PBS containing 2% FBS (Atlas Biologicals). Isolated cells were then enumerated and stained with biotin-labeled CD8 (53-6.7), CD19 (MB19-1), CD11b (M1/70), CD45R (RA3-6B2), DX5 (DX5), Ly6C (RB6-8C5), I-A^b (M5/114.15.2), IgM (eB121-15F9), and Ter-119 obtained from eBioscience and γδ TCR (GL3) obtained from BD Pharmingen. Following Ab staining, the cells were incubated with streptavidin-labeled MicroBeads and fractionated with an LS column (Miltenyi Biotec). The enriched CD4 lymphocytes were then stained with allophycocyanin-labeled CD4 (L3T4) and PE-labeled CD25 (PC61), and sorted into CD4⁺CD25⁻ and CD4⁺CD25⁺ populations to a purity of >95% using a FACSAria (BD Biosciences). Alternatively, the enriched CD4 lymphocytes were stained with PE-labeled CD25 and fractionated into CD4⁺CD25⁻ and CD4⁺CD25⁺ populations using LS columns to a purity of >80%. Following fractionation, CD4⁺CD25⁻ and CD4⁺CD25⁺ populations were fixed with a 1× PBS solution containing 27% formaldehyde, and the mixture was then sonicated to generate DNA fragments of <1 kb. The size of the DNA fragments after sonication was confirmed by running aliquots of sheared DNA on ethidium bromide-stained agarose gels; DNA fragments typically averaged between 250 and 500 bp with no detectable fragments >1 kb. ChiP assays were performed with anti-STAT5 (sc-835; Santa Cruz Biotechnology) or anti-Mek1 (sc-219; Santa Cruz Biotechnology) Abs as previously described (21) using a kit from Upstate Biotechnology. Real-time PCR was performed using the Brilliant QPCR core reagent kit (Stratagene). All PCR were conducted for one cycle at 95°C for 10 min, followed by 45 cycles at 95°C (15 s) and 60°C (90 s). Serial dilutions of genomic DNA were used to obtain a standard curve used to quantify levels of DNA in ChIP assays. Oligonucleotides and TaqMan probes were obtained from Integrated DNA Technologies and are listed as follows: 5′ promoter primer, 5′-CTCACTCAGAGACTCGCAGCA-3′; 3′ promoter primer, 5′-GCAAGCATGCATATGATCACC-3′; TaqMan promoter probe, FAM-CCAGCCATTCTGAGACTCTCTGATTCTGTG-BHQ1; 5′ intron 1 primer, 5′-CAGAAAAATCTGGCCAAGTT-3′; 3′-intron 1 primer, 5′-AGGACCTGAATTGGATATGGT-3′; TaqMan intron 1 probe, FAM-ACCTCCATACAGCTTCTAAGAAACAGTCAAACA-BHQ1; 5′ intron 11, 5′-CTGTGAGTTGGAAGCCAG-3′; 3′ intron 11 primer, 5′-CTGAGCTATCTGTCCAGTTCC-3′; and TaqMan intron 11 probe, FAM-AACTGTAAATCTGTAATC CCAGCATGTGTACABHQ1.

IL-2Rβ signaling is critical for Treg development

To investigate the role of IL-2 and IL-2R in Treg development, we compared littermate control mice (LMC) with IL-2^−/−, IL-2Rβ^−/−, and γ_c^−/− mice for the presence of CD4⁺Foxp3⁺ cells. As recently reported (7, 8), we observed comparable populations of CD4⁺Foxp3⁺ cells in LMC and IL-2^−/− mice (Fig. 1⇓A). Importantly, the absolute number of CD4⁺Foxp3⁺ cells in IL-2^−/− mice was not significantly different from the number of CD4⁺Foxp3⁺ cells observed in LMC mice (Fig. 1⇓B). Similar results were found in IL-2Rα^−/− mice (data not shown). In contrast, we found that IL-2Rβ^−/− mice have a significant decrease in both the percentage and absolute cell numbers of CD4⁺Foxp3⁺ cells in the thymus and in the periphery (Fig. 1⇓, C and D). Specifically, IL-2Rβ^−/− mice show a 4-fold decrease in numbers of thymic Tregs relative to LMC; the decrease is even more dramatic in the spleen and LN (27- and 6-fold decrease, respectively, relative to LMC mice). Analysis of γ_c-deficient mice revealed an even more striking defect with a virtual complete absence of Tregs in both the thymus and spleen (Fig. 1⇓, E and F). A potential concern with this latter experiment is that γ_c^−/− mice have dramatically reduced numbers of all lymphocyte subsets due to a requirement for γ_c in IL-7R-dependent early B and T cell development (22, 23). Hence, the abrogation of Treg development observed in γ_c^−/− mice may be an indirect effect of the paucity of effector T cells in these mice. This is particularly relevant because Sakaguchi and colleagues (6) have provided evidence that IL-2 produced by activated effector T cells may be involved in the maintenance of Tregs. To address this issue, we analyzed Treg development in IL-7Rα^−/− mice. Like γ_c^−/− animals, IL-7Rα^−/− mice have an identical block in thymocyte development (at the double-negative (DN)2 to DN3 transition) and exhibit a similar paucity of mature T cells (24). However, as shown in Fig. 1⇓E, IL-7Rα^−/− mice contain a substantial population of CD4⁺Foxp3⁺ cells in both the thymus and the periphery that are significantly greater in total cell numbers than the few cells observed in γ_c^−/− mice (albeit still reduced in numbers relative to WT mice, Fig. 1⇓F). Importantly, the ratio of CD4⁺Foxp3⁺:CD4⁺Foxp3⁻ T cells in IL-7Rα^−/− mice was greater than or equal to that seen in WT mice, demonstrating that Treg development was not impaired. These results suggest that one or more cytokines that bind both IL-2Rβ and γ_c are required for the development of Tregs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

IL-2Rβ and γ_c chains are required for the development of CD4⁺Foxp3⁺ cells. A, Shown are CD4⁺CD8⁻-gated thymocytes (THY), total splenocytes (SPL), or LN cells stained with Abs to CD4 and Foxp3. Numbers to the right of the outlined areas represent the percentage of cells in the designated gate. B, Shown are average total Foxp3⁺ cell numbers from C57BL/6 (▪) and IL-2^−/− (□) mice. C, Thymii, spleens, and LN were harvested from 4-wk-old C57BL/6 or IL-2Rβ^−/− mice and stained with Abs to CD4 and Foxp3. Flow cytometry plots shown depict CD4⁺CD8⁻-gated thymocytes, lymphocyte-gated splenocytes, and LN cells . Numbers indicate the percentage of cells in the gate. D, Total CD4⁺Foxp3⁺-gated cells from both C57BL/6 (▪) and IL-2Rβ^−/− (□) mice. E, Thymii and spleens were harvested from 6- to 8-wk-old IL-7Rα^−/− and γ_c^−/− mice and stained with Abs to CD4, CD25, and Foxp3. Numbers indicate the percentage of CD4⁺CD8⁻ thymocytes or total splenocytes inside the depicted gate. F, Total numbers of CD4⁺Foxp3⁺ cells from IL-7Rα^−/− (▪) and γ_c^−/− (□). Error bars, SEM. Data in each case are representative of three or more independent experiments using between two and five mice per group. Mice used were between 4 and 5 wk of age, except as otherwise noted. The p values for all figures were determined by a two-tailed Student’s t test.

Other than IL-2, the only cytokine known to signal through an IL-2Rβ/γ_c receptor complex is IL-15. IL-15^−/− mice do not develop autoimmunity, and we have observed that they have virtually identical numbers of Tregs as WT mice. In contrast, IL-2^−/− × IL-15^−/− mice show a comparable decrease in Treg numbers as that observed in IL-2Rβ^−/− mice (6-fold ↓ in the thymus and 12-fold ↓ in the spleen; Fig. 2⇓). Thus, IL-2 and IL-15 are redundant cytokines, either of which can signal through an IL-2Rβ/γ_c receptor complex to promote Treg development.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

IL-2 and IL-15 act as redundant factors to promote the development of Foxp3⁺ cells. A, Shown are CD4⁺CD8⁻-gated thymocytes (THY) or total splenocytes (SPL) stained with Abs to CD4 and Foxp3. Numbers to the right of the outlined areas represent the percentage of cells in the designated gate. B, Shown are average total CD4⁺Foxp3⁺ cell numbers from C57BL/6 (▪), IL-15^−/− (□), and IL-2^−/− × IL-15^−/− (▩) mice. Error bars, SEM. Mice used in these experiments were between 4 and 5 wk of age. Data are representative of nine WT C57BL/6 and IL-15^−/− mice and four IL-2^−/− × IL-15^−/− mice.

Foxp3 expression restores a functional population of Tregs in IL-2Rβ^−/− mice

One potential mechanism by which IL-2Rβ-dependent signals could govern Treg development would be by influencing foxp3 expression. To test this hypothesis, we made use of foxp3 transgenic mice. The foxp3 transgene (made by Ramsdell and colleagues, 18) was derived from a 30.8-kb cosmid encompassing the entire coding region of the foxp3 gene as well as a significant portion of the upstream regulatory elements (18 kb). However, expression of the transgene does not mimic that of the endogenous foxp3 gene, as it is expressed in most CD4⁺ (75%) and CD8⁺ (44%) T cells. We intercrossed IL-2Rβ^−/− mice with foxp3 transgenic mice (foxp3tg) and observed that the foxp3 transgene effectively restores a large population of CD4⁺Foxp3⁺ T cells in IL-2Rβ^−/− × foxp3tg mice (Fig. 3⇓A). Notably, expression of the foxp3 transgene in IL-2Rβ^−/− mice prevented activation of the remaining CD4⁺Foxp3⁻ T cells typically seen in IL-2Rβ^−/− mice. For example, CD62L expression was restored, and CD69 levels were reduced, to levels seen on WT T cells (Fig. 3⇓B). Furthermore, the lymphoproliferative disorder seen in IL-2Rβ^−/− mice was completely inhibited in IL-2Rβ^−/− × foxp3tg mice as demonstrated by the reversal of splenomegaly (Fig. 3⇓C) and the decreased total lymphocyte counts observed in IL-2Rβ^−/− × foxp3tg mice relative to IL-2Rβ^−/− controls (data not shown). Because peripheral T cells in IL-2Rβ^−/− × foxp3tg mice express high levels of Foxp3, but are completely incapable of responding to IL-2Rβ-dependent signals, we used them to examine whether IL-2R signaling is required to induce Treg suppressor function in vitro. For these experiments, we isolated WT CD4⁺CD25⁺ T cells, CD4⁺ T cells from foxp3tg mice, and CD4⁺ T cells from IL-2Rβ^−/− × foxp3tg mice and assayed their ability to inhibit proliferation of WT CD4⁺CD25⁻ responder cells that had been stimulated with allogeneic APCs from bm12 mice. As shown in Fig. 3⇓D, the CD4⁺ T cells from IL-2Rβ^−/− × foxp3tg mice suppressed responder cell proliferation to an equivalent degree as that achieved when using WT CD4⁺CD25⁺ T cells. CD4 T cells from foxp3tg mice also inhibited proliferation of WT effector cells, whereas CD4⁺ T cells from IL-2Rβ^−/− mice did not (data not shown). These results demonstrate that Foxp3 expression is sufficient, in the absence of other IL-2Rβ-dependent signals, to effectively inhibit effector T cell proliferation in vitro.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Foxp3 expression rescues Treg development in IL-2Rβ^−/− mice. A, Spleens were harvested from 5- to 8-wk-old LMC, IL-2Rβ^−/−, foxp3tg, and IL-2Rβ^−/− × foxp3tg mice and stained with Abs to CD4 and Foxp3. B, CD4⁺-gated T cells were stained with Abs to CD69 and CD62L. C, Shown are spleens from 5-wk-old LMC, foxp3tg, IL-2Rβ^−/−, and IL-2Rβ^−/− × foxp3tg mice. Data are representative of five mice. D, One hundred thousand CD4⁺CD25⁻ responder T cells from C57BL/6 mice were mixed with equal numbers of allogeneic bm12 splenocytes to induce proliferation. FACS-sorted CD4⁺ cells from foxp3 transgenic mice (□), IL-2Rβ^−/− × foxp3tg mice (▦), or CD4⁺CD25⁺ T cells from C57BL/6 mice (▪) were cultured at the indicated Treg to responder cell ratio. Data are presented as mean [³H]thymidine incorporation of quadruplicate cultures and error bars represent SEM. Data presented are representative of three independent experiments.

STAT5 signaling is required for Treg development

The IL-2R activates a number of downstream signaling pathways including those involving Erk, PI3K, and STAT5. Having previously demonstrated that STAT5 is preferentially activated in Tregs (25), we hypothesized that STAT5 signaling may play a critical role in Treg development. In fact, we and others have previously demonstrated that mice in which the first exon of STAT5a and STAT5b is disrupted (referred to as STAT5a/b^Ex1/Ex1 mice) have a decreased number of cells with regulatory activity (26, 27). However, because these studies predated the availability of methods to stain for intracellular foxp3, we reexamined this issue. We found that although STAT5a/b^Ex1/Ex1 mice do have a reduced percentage of Tregs, substantial numbers of these cells could still be detected. However, we also observed that lymphocytes from these mice express significant amounts of truncated STAT5a/b protein (data not shown). Importantly, this truncated STAT5 protein has recently been shown to exhibit functional activity (28). To circumvent this issue, we made use of mice generated by Hennighausen and colleagues (28), in which the entire STAT5a and STAT5b genes are flanked by loxP sites (referred to as STAT5a/b^FL/FL mice), allowing for Cre-mediated deletion of both STAT5a and STAT5b genes. We crossed these animals to CD4-Cre transgenic mice in which Cre protein is expressed in CD4⁺CD8⁺ and CD4⁺ thymocytes, as well as in peripheral CD4⁺ T cells (29). Consistent with recent reports (30), we observed a significant decrease in CD8⁺ T cells in the thymus and the periphery of CD4-cre × STAT5a/b^FL/FL mice. Notably, CD4⁺Foxp3⁺ T cells were also clearly reduced in both the thymus and spleen of CD4-cre × STAT5a/b^FL/FL mice when compared with STAT5a/b^FL/FL or CD4-cre⁻ × STAT5a/b^FL/+ controls (Fig. 4⇓, A–C). Consistent with the reduction in Treg numbers, we observed a significant increase in the percentage of activated CD4⁺ T cells in these mice based on decreased CD62L and increased CD69 expression (Fig. 4⇓D). However, it was also clear that not all CD4⁺Foxp3⁺ Tregs were eliminated in CD4-Cre × STAT5a/b^FL/FL mice. One explanation for the remaining Tregs is that STAT5 may not have been efficiently eliminated in that cell lineage. To examine this issue, we stimulated T cells from CD4-Cre × STAT5a/b^FL/FL mice with IL-2 and IL-7, which results in STAT5 phosphorylation that can be detected by flow cytometry. Because the intracellular staining protocols for Foxp3 and phospho-STAT5 require different fixation conditions that are not compatible, we used CD25 as a surrogate marker to identify Tregs. Phospho-STAT5 was clearly detected in CD4⁺CD25⁻, CD4⁺CD25⁺, and CD8⁺ T cells from LMC (Fig. 4⇓, E and F). However, in CD4-Cre × STAT5a/b^FL/FL mice, the majority of CD4⁺CD25⁻ and CD8⁺ T cells had deleted STAT5a/b protein (95 and 70%, respectively) as demonstrated by the failure to detect phospho-STAT5 protein following IL-2/IL-7 stimulation. Importantly, our results with loss of STAT5 protein are in perfect agreement with previous findings by Yao et al. (30), who reported that >90% of double-positve thymocytes in CD4-Cre × STAT5a/b^FL/FL mice have deleted the STAT5a and STAT5b genes. In contrast, CD4⁺CD25⁺ T cells from CD4-Cre × STAT5a/b^FL/FL mice exhibited minimal to no deletion of STAT5a/b as demonstrated by their robust ability to induce phospho-STAT5 upon IL-2/IL-7 stimulation in both the thymus and the spleen (Fig. 4⇓, E and F). Similar results were obtained when using glucocorticoid-induced TNFR family-related gene as a cell surface marker to identify Tregs (data not shown). These results indicate that the only remaining Tregs are those that had not deleted STAT5 and thus strongly support a role for STAT5 in the development of Tregs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

STAT5 signaling is necessary for the development of Tregs. A and B, Left panels, show thymocytes (A) or splenocytes (B) from LMC (CD4-Cre⁻ × STAT5a/b^FL/+ or STAT5a/b^FL/FL) and CD4-Cre⁺ × STAT5a/b^FL/FL mice stained with Abs to CD4 and CD8. Center and middle panels, CD4-gated thymocytes (A) or splenocytes (B) stained with Abs to CD25 or Foxp3. C, Shown are total numbers of CD4⁺Foxp3⁺ T cells from either LMC (□) or CD4-Cre × STAT5a/b^FL/FL mice. D, Shown are histograms of CD4-gated T cells from LMC or CD4-Cre × STAT5a/b^FL/FL mice stained with Abs to CD62L or CD69. E and F, Thymocytes (E) or splenocytes (F) were stimulated with 1000 U/ml IL-2 and 30 ng/ml IL-7 for 30 min and then stained with Abs to CD4, CD8, and CD25. Cells were then fixed, permeabilized, and stained for phopsho-STAT5. Shown is a representative example of three or more experiments.

STAT5 activation restores CD4⁺Foxp3⁺ Tregs in IL-2Rβ^−/− mice

Having demonstrated that STAT5 is required for Treg development, we decided to examine whether STAT5 activation was sufficient to rescue Treg development in IL-2Rβ^−/− mice. For these studies, we made use of transgenic mice that express a constitutively active form of STAT5b (STAT5b-CA) (21, 26, 31). Expression of STAT5b-CA is regulated by a compound promoter element consisting of the lck proximal promoter with the Eμ enhancer inserted into it. We have previously demonstrated that this transgene is expressed early during thymocyte development (at the early T cell progenitor stage) and that it continues to be expressed in both mature CD4⁺ and CD8⁺ T cells (21, 26). As shown in Fig. 5⇓, expression of the STAT5b-CA transgene restored a substantial population of CD4⁺CD25⁺Foxp3⁺ Tregs in the thymus of IL-2Rβ^−/− mice. Likewise, CD4⁺Foxp3⁺ Tregs could still be detected in the spleen of STAT5b-CA × IL-2Rβ^−/− mice. However, these Tregs no longer expressed CD25 (Fig. 5⇓B) and could not prevent T cell lymphoproliferation in vivo or animal death (data not shown). Thus, in many respects the CD4⁺CD25⁻Foxp3⁺ Tregs observed in the spleen and lymph nodes of STAT5b-CA × IL-2Rβ^−/− mice are very similar to the Tregs observed in IL-2^−/− mice, which also fail to express CD25 and do not prevent autoimmunity (8).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

STAT5b activation restores Treg development in the thymus of IL-2Rβ^−/− mice. Shown are CD4⁺CD8⁻-gated thymocytes (A) or splenocytes (B) stained with Abs to CD4, CD25, and Foxp3 from WT controls; STAT5b-CA transgenic mice; IL-2Rβ^−/− mice; and STAT5b-CA × IL-2Rβ^−/− mice. C, Shown are total numbers of CD4⁺Foxp3⁺ T cells in WT controls, STAT5b-CA transgenic mice, IL-2Rβ^−/− mice, and STAT5b-CA × IL-2Rβ^−/− mice. Error bars, SEM. Data are representative of three separate experiments.

One potential reason for the failure to prevent autoimmunity in STAT5b-CA × IL-2Rβ^−/− mice is that the lck proximal promoter, which drives expression of the STAT5b-CA transgene, is known to become inactive in mature T cells and typically results in decreased expression of transgenes in mature T cells (32). Indeed, the expression of the STAT5b-CA transgene is quite low in peripheral T cells (26). Moreover, expression of the STAT5b-CA transgene in CD8⁺ and CD4⁺Foxp3⁻ T cells renders them more resistant to Treg-mediated suppression (data not shown). Finally, Rudensky and colleagues (8) have shown that Tregs develop in IL-2^−/− mice (Fig. 1⇑A) but that they fail to prevent autoimmunity in these mice because they exhibit reduced fitness compared with IL-2^−/− effector T cells. We propose that a parallel situation exists in STAT5b-CA × IL-2Rβ^−/− mice in which STAT5b-CA × IL-2Rβ^−/− Tregs exhibit reduced fitness relative to STAT5b-CA × IL-2Rβ^−/− effector T cells. Supporting this proposal, we have recently observed that expression of the STAT5b-CA transgene in γ_c^−/− mice restores peripheral T cell numbers including CD4⁺CD25⁺Foxp3⁺ Tregs; importantly, these mice do not exhibit T cell activation, as determined by CD69 and CD62L expression, nor do they develop autoimmune disease (data not shown). Thus, in a situation where all T cells are equally affected (γ_c^−/−) STAT5b-CA expression is sufficient to induce a population of functional Tregs that are capable of preventing autoimmune disease.

To circumvent these possible problems, we expressed a chimeric IL-2Rβ construct that selectively activates STAT5 in bone marrow cells from IL-2Rβ^−/− mice (Fig. 6⇓A). Reconstitution of rag2^−/− mice with these transduced bone marrow cells results in the generation of T cells that should induce STAT5 phosphorylation at physiological levels when exposed to endogenously produced IL-2. Control constructs encoding either the full-length WT IL-2Rβ chain or a signaling defective receptor in which the cytoplasmic tail contained just the S region were also generated and tested using this approach (referred to as WT IL-2Rβ and IL-2Rβ-ΔAH, respectively). We first examined the signaling properties of control and chimeric receptors by introducing them into the cytokine-dependent BAF cell line. IL-2 stimulation of the cells expressing the WT IL-2Rβ chain led to activation of the ERK and STAT5 signaling pathways. In contrast, the chimeric IL-2Rβ-STAT5 receptor only activated STAT5, whereas the signaling defective IL-2Rβ-ΔAH receptor did not activate any of these pathways (data not shown). Thus, the chimeric IL-2Rβ-STAT5 receptor selectively leads to STAT5 activation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

STAT5 signaling is sufficient to mediate IL-2Rβ-dependent Treg development. A, Schematic representation of IL-2Rβ constructs used for retroviral reconstitution experiments. B, Spleens and thymii from recipient mice were harvested at 9 wk postreconstitution and stained with Abs to IL-2Rβ (open histograms). Shaded histograms represent Ab staining on IL-2Rβ-deficient mice. C and D, Foxp3 and CD62L levels of expression were analyzed on CD4⁺-gated cells. Numbers above the outlined areas represent the percentage of cells in the designated gate. Data are representative of three independent recipient mice per group.

The WT IL-2Rβ, IL-2Rβ-STAT5, or IL-2Rβ-ΔAH constructs were then introduced into IL-2Rβ^−/− bone marrow using a modified version of the pMIGR retroviral vector (20). Infected IL-2Rβ^−/− bone marrow was then transferred into rag2^−/− recipients; 9 wk following reconstitution, recipient mice were analyzed for the presence of CD4⁺Foxp3⁺ and CD4⁺CD25⁺ Treg populations. Cell surface staining for the IL-2Rβ chain demonstrated that equal percentages of splenic CD4⁺ T cells in these bone marrow chimeric mice expressed the retrovirally driven IL-2Rβ chain (Fig. 6⇑B). Mice reconstituted with the WT IL-2Rβ construct developed a significant population of CD4⁺Foxp3⁺ and CD4⁺ CD25⁺ cells, indicating that reintroduction of the WT receptor could restore Tregs in IL-2Rβ^−/− mice (Fig. 6⇑C and data not shown). Similarly, the IL-2Rβ-STAT5 construct also restored a comparable population of CD4⁺Foxp3⁺ Tregs. Importantly, most Tregs in the spleen of these mice expressed high levels of CD25, indicating that they resembled Tregs that are normally observed in WT mice. In contrast, the signaling defective IL-2Rβ-ΔAH construct did not restore a population of CD4⁺Foxp3⁺ Tregs.

We next examined T cell activation in these mice. Virtually all T cells (95%) in mice reconstituted with the signaling defective IL-2Rβ-ΔAH construct expressed low levels of CD62L consistent with an activated phenotype (Fig. 6⇑D, right panel). In contrast, T cell activation was greatly diminished in mice reconstituted with either the WT or IL-2Rβ-STAT5 receptor as demonstrated by the increase in the percentage of T cells expressing high levels of CD62L (Fig. 6⇑D, left and center panels). Identical conclusions were reached when using expression of CD69 as a marker of T cell activation (data not shown). Thus, Tregs generated following reconstitution with an IL-2Rβ construct that can only activate STAT5 were capable of preventing T cell activation in vivo (although we cannot rigorously rule out the possibility that these Tregs may exhibit somewhat reduced suppressor function). These results demonstrate that STAT5 activation is sufficient to promote IL-2Rβ-dependent development of CD4⁺CD25⁻Foxp3⁺ and CD4⁺CD25⁺Foxp3⁺ Tregs in both the thymus and peripheral lymphoid organs.

STAT5 binds to the foxp3 gene promoter

The findings that foxp3 expression or STAT5 activation are both sufficient to restore Treg development in IL-2Rβ^−/− mice suggested that STAT5 may act by directly regulating foxp3 expression. To examine this issue, we first scanned the murine foxp3 gene for potential STAT5 consensus binding sites. As shown in Fig. 7⇓A, the murine foxp3 gene contains 11 potential STAT5 binding sites. Because functional binding sites should be evolutionarily conserved, we compared the foxp3 gene sequences from murine, human, monkey, and dog species and observed that six of these sites were conserved (Fig. 7⇓B); these include three closely spaced sequences in the foxp3 promoter and three sites located between murine exon-2a and exon-1 (Fig. 7⇓A). To determine whether STAT5 actually binds to these conserved sites we conducted ChIP assays. As seen in Fig. 7⇓C, immunoprecipitation with an anti-STAT5 Ab revealed STAT5 binding to the conserved sites in the promoter of the foxp3 gene in murine CD4⁺CD25⁺ T cells. In contrast, no binding could be seen to the three conserved sites in intron 2 or the distal site located in intron 11. No PCR product was observed when ChIP was done using an anti-Mek-1 control Ab. Finally, significant STAT5 binding to the foxp3 gene was not observed in CD4⁺CD25⁻ T cells, a cell population which is mostly comprised of CD4⁺Foxp3⁻ T cells (Fig. 7⇓D). Thus, these experiments provide evidence that STAT5 can play a role in regulating foxp3 gene expression.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

STAT5 binds to the foxp3 promoter in Tregs. A, Schematic representation of the foxp3 gene. Arrows, Locations of PCR primers used for quantification of ChIP assays. Asterisks, The conserved STAT5 binding sites shown in B and analyzed by ChIP assays in C. B, Alignment of selected regions of foxp3 gene among various species. C, Real-time PCR quantification of genomic DNA obtained from CD4⁺CD25⁺ T cells immunoprecipitated with either anti-STAT5 (▪) or anti-MEK-1. Left panel, The amount of PCR product normalized to input control of the foxp3 promoter region containing the STAT5 binding sites noted in A. Middle panel, The amount of PCR product normalized to input control of the foxp3 intron 1 region containing the STAT5 binding sites noted in A. Right panel, the amount of PCR product normalized to input control of the foxp3 intron 11 region noted in A. Input samples amplified DNA from all regions tested. Error bars, SEM. Data are representative of two independent experiments. D, Comparison of STAT5 binding to the foxp3 promoter sites in CD4⁺CD25⁺ vs CD4⁺CD25⁻ T cells. Error bars, SEM. Data are representative of two independent experiments.