Signal transducer and activator of transcription 4 limits the development of adaptive regulatory T cells

Abstract

T-cell responses to a cytokine milieu instruct the development of multiple effector phenotypes. While transforming growth factor-β₁ (TGF-β₁) inhibits the development of T helper type 1 (Th1) and Th2 cells, we demonstrate that like interleukin-6 (IL-6) and IL-4, IL-12 can inhibit the development of TGF-β₁-induced Foxp3-expressing adaptive T regulatory (aTreg) cells. Signal transducer and activator of transcription 4 (STAT4) is critical for the response to IL-12, although there is a parallel pathway involving T box expressed in T cells (T-bet), and cells from mice double-deficient in STAT4 and T-bet are refractory to the inhibition of aTreg-cell development by IL-12. While the ability of these cytokines to promote Th differentiation may contribute to this effect, we observe that culture with IL-12, or other instructive cytokines, results in an increase in repressive chromatin modifications at the Foxp3 locus that limit STAT5 binding to Foxp3, without observed effects on IL-2 signalling pathways. In a model of allergic lung inflammation there are increased percentages of Treg cells in the lungs of Stat4^−/− mice, compared with wild-type mice, and increases in Treg cells correlate with decreased allergic inflammation. Overall, these results suggest an important role for STAT4 in regulating Treg-cell development.

Introduction

Forkhead Box P3 (Foxp3) expression instructs CD4⁺ T cells to become inducible or adaptive T regulatory (aTreg) cells following stimulation with anti-T-cell receptor in the presence of transforming growth factor-β₁ (TGF-β₁). Adaptive Treg cells can suppress T-cell proliferation and effector function in vitro and in vivo.¹ However, the regulation of aTreg-cell development is not completely understood. Interleukin-2 (IL-2) enhances the development of TGF-β₁-primed Foxp3⁺ aTreg cells through a signal transducer and activator of transcription 5 (STAT5)-dependent pathway.^2–5 In contrast to the effects of IL-2 and STAT5, IL-6, through STAT3-dependent pathways, and IL-4, through STAT6-dependent pathways, have been shown to inhibit the development of aTreg cells and to promote altered cytokine secretion.^6–15 Previous reports have suggested that IL-12, through a STAT4-dependent pathway, also impairs aTreg-cell development.¹³ However, it was suggested that the IL-12 effect was mediated entirely through the STAT4-dependent induction of interferon-γ (IFN-γ), and was dependent upon STAT1 and T box expressed in T cells (T-bet).¹³ These effects of IL-12 have not been studied further.

In this report we show that IL-12, through STAT4 but not other STAT proteins, inhibits the development of aTreg cells in a pathway parallel to T-bet-dependent inhibition. Like other inflammatory cytokines, IL-12 promotes repressive chromatin modifications of Foxp3 and inhibits the binding of STAT5 to the Foxp3 promoter. Moreover, in a model of allergic airway inflammation, mice that lack STAT4 have increased percentages of Treg cells in the bronchoalveolar lavage (BAL) and increased Foxp3 messenger RNA in the lung that correlates with attenuated airway inflammation. Consequently, STAT4 is not only required for the promotion of the development of inflammatory subsets, but also limits the development of aTreg cells in vitro and in vivo.

Materials and methods

Mice

All animal studies in this report were approved by the Indiana University Institutional Animal Care and Use Committee. The generation of Stat4^−/−, Stat6^−/−, Tbx21^−/−, Stat4^−/−Tbx21^−/− and Stat3^fl/fl with a CD4-Cre (Stat3^CD4−/−) transgene were previously described.^16–20 C57BL/6 Stat4^−/− and BALB/c Stat6^−/− mice were used with matched wild-type (WT) mice (Harlan Sprague Dawley, Indianapolis, IN). Stat3^fl/fl mice are on a mixed 129-C57BL/6 genetic background and WT mice in experiments using Stat3^CD4−/− mice were Cre-negative littermates.

Analysis of T helper cell differentiation

Total CD4⁺ T cells were isolated from Stat4^−/−, Stat6^−/− or Stat3^CD4−/− and control spleens (magnetic antibody cell sorting isolation system; Miltenyi Biotec, Auburn, CA). T cells were activated with plate-bound anti-CD3 (4 μg/ml 145-2C11) and soluble anti-CD28 (1 μg/ml; BD Pharmingen, San Jose, CA) and were cultured under conditions that prime aTreg cells [TGF-β₁ (2 ng/ml; R&D Systems, Minneapolis, MN) and anti-IL-4 (10 μg/ml 11B11)], T helper type 17 [Th17; TGF-β₁ and IL-6 (100 ng/ml; Peprotech, Rocky Hill, NJ)], IL-12 + TGF-β₁ [aTreg-cell conditions + IL-12 (5 ng/ml; Peprotech)] or IL-4 + TGF-β₁ [anti-IFN-γ (10 μg/ml R46A2), TGF-β₁ and IL-4 (10 ng/ml; Peprotech)]. After 5 days in culture, cells were restimulated with plate-bound anti-CD3 (4 μg/ml) for 24 hr (or 96 hr for TGF-β₁) before cell-free supernatants (acid treated for TGF-β₁ analysis) were analysed for IFN-γ, IL-4 and TGF-β₁ using enzyme-linked immunosorbent assay (ELISA; reagents from BD Pharmingen or R&D Systems).²¹ Foxp3 intracellular staining was performed using the eBioscience fixation–permeabilization kit before staining with fluorescein isothiocyanate-conjugated Foxp3 (eBioscience, San Diego, CA) and analysis by flow cytometry. The % repression of Foxp3⁺ cells was calculated as (% Foxp3⁺ cells in cultures incubated with Th differentiative cytokine/% Foxp3⁺ cells in cultures with TGF-β₁ alone) × 100. Statistics were performed using an unpaired Student’s t-test. Phospho-Stat intracellular staining was performed using 1·5% paraformaldehyde fixing of cells before methanol permeabilization for 10 min at 4°. Cells were stained using pSTAT5 antibody (BD Pharmingen) for 30 min at room temperature and analysed by flow cytometry. RNA was isolated from Th cultures and RNA levels of the genes indicated were analysed by quantitative polymerase chain reaction (PCR) as described elsewhere.²¹ Chromatin immunoprecipitation analysis was performed as described previously²² using PCR primers that span the Foxp3 promoter and first intron.²³ Statistics were performed using an unpaired Student’s t-test.

Suppressor assay

CD4⁺ CD25⁻ cells (5 × 10⁴) isolated from WT mice were incubated with 2 μg/ml soluble anti-CD3 in the presence of irradiated, T-cell-depleted WT splenocytes (5 × 10⁴) and increasing numbers of CD4⁺ CD25⁺ cells from the indicated primary cultures, or from purified splenic natural Treg (nTreg) cells, isolated from control or gene-deficient mice. Cultures were pulsed with 0·8 μCi [³H]thymidine for the last 16–24 hr of a 72-hr incubation. Radioactivity incorporated was counted using a flatbed beta-counter (Wallac/PerkinElmer, Waltham, MA). % suppression was calculated as 1 – (proliferation of Treg + T effector/proliferation of T effector) × 100. Statistics were performed using an unpaired Student’s t-test.

Sensitization and challenge

Mice were sensitized by two intraperitoneal injections (0·5 ml) of ovalbumin (OVA; grade V) alum [Al(OH)₃/Mg(OH)₂, both from Sigma-Aldrich, St Louis, MO; 20 μg/2 mg] on days 0 and 7, of the protocol. From days 14 to 18, mice were challenged intranasally with 50 μg OVA in 30 μl. Mice were killed by intraperitoneal injection of pentobarbital (5 mg/mouse in phosphate-buffered saline) 48 hr after the last intranasal challenge. Bronchoalveolar lavage was performed with 3 × 1 ml phosphate-buffered saline. The cells recovered in BAL fluid were counted using a haemocytometer. The cellular composition of BAL for the populations of eosinophils, neutrophils, T cells, B cells, dendritic cells and macrophages was measured as described previously, using flow cytometry.²⁴ Briefly, eosinophils, neutrophils, T cells, B cells and mononuclear cells were distinguished by cell size and the expression of CD3, B220, CCR3, CD11c and major histocompatibility complex class II. Cytokines in cell-free BAL fluid were analysed using ELISA as described above. For real-time PCR measurements, lung tissues were homogenized in a tissue lyser (Qiagen, Valencia, CA) and RNA isolated with an RNeasy kit (Qiagen) was used to synthesize complementary DNA and for subsequent quantitative PCR. Statistics were performed using an unpaired Student’s t-test.

Results

Repression of TGF-β₁-induced Foxp3 expression by IL-12 requires STAT4

The ability of IL-6, IL-21 and IL-4 to divert the differentiation of aTreg cells into cells with distinct phenotypes suggests that in an inflammatory cytokine environment the development of aTreg cells is inhibited^8,11–13,25. The ability of a Th1-promoting cytokine environment containing IL-12 to inhibit aTreg-cell development has not been clearly documented. To test this directly, we examined cells cultured in Th1 (IL-12 + anti-IL-4) conditions in the presence of TGF-β₁ for Foxp3 expression and suppressor activity compared with cells cultured in aTreg (TGF-β₁ + anti-IL-4), Th2 (IL-4 + anti-IFN-γ) or Th17 (TGF-β₁ + IL-6 + anti-IL-4 + anti-IFN-γ) conditions. The Th2 conditions repressed TGF-β₁-induced Foxp3 expression and suppressor activity as efficiently as Th17 culture conditions (Fig. 1a). Although Th1 conditions were not as efficient at repressing the aTreg phenotype as Th17 culture conditions, IL-12 was able to decrease Foxp3 expression and suppressor activity (Fig. 1a). The ability of cells in each culture to proliferate in response to anti-CD3 correlated with the percentage of Foxp3⁺ cells (Fig. 1a). The results of experiments with purified naïve (CD4⁺ CD62L⁺) cells for differentiation were similar (data not shown). These results suggest that the ability of instructive cytokines to inhibit Foxp3 expression also decreases their suppressive function.

Figure 1.

Interleukin-12/signal transducer and activator of transcription (IL-12/STAT4) represses Foxp3 expression and suppressive activity. (a) CD4⁺ CD25⁻ responder cells from wild-type (WT) C57BL/6 mice were stimulated in the presence of anti-CD3 and irradiated T-cell-depleted splenocytes in the presence or absence (CD4⁺ CD25⁻ cells alone) of WT cells cultured with transforming growth factor-β₁ (TGF-β₁) alone (adaptive regulstory T cells; aTreg), TGF-β₁ + IL-4, T helper tye 17 (Th17), or TGF-β₁ + IL-12 (4 : 1 ratio). Per cent suppression ± SD was calculated as described in the Materials and methods. The % Foxp3⁺ cells within each culture and the proliferation of each culture without effector cells are indicated to the right of the graph. The data are representative of two independent experiments with two mice in each experiment. *Significantly different (P<0·05) from cells cultured with TGF-β₁ alone. (b) Five-day cultured cells from the indicated T helper cell culture conditions were assessed for the percentage of Foxp3⁺ cells ± SD from Stat4^−/−, Stat6^−/− or Stat3^CD4−/− mice and WT strain-matched controls. Relative % repression of Foxp3⁺ T cells was calculated as described in the Materials and methods. Results are indicative of two to five independent experiments with two to four mice in each experiment. *Significantly different (P<0·05) from WT cells cultured under the same conditions. (c) CD4 T cells from WT, Stat4^−/−, and STAT4α and STAT4β transgenic, mice were cultured under aTreg or TGF-β₁ + IL-12 conditions for 5 days before the percentage of Foxp3⁺ cells ± SD was assessed by flow cytometry. *Significantly different (P<0·05) from Stat4^−/− cells cultured under the same conditions. (d) CD4 T cells from WT, Stat4^−/−, Tbx21^−/− and Stat4^−/−Tbx21^−/− mice were cultured under aTreg or TGF-β₁ + IL-12 conditions for 5 days before the percentage of Foxp3⁺ cells ± SD was assessed by flow cytometry. Results are expressed as percentage of the IL-12-induced inhibition of Foxp3 expression compared to aTreg-cell cultures and are the average ± SD of values from three to four mice in separate experiments. *Significantly different (P<0·05) from WT cells cultured under the same conditions; **significantly different (P<0·05) from Tbx21^−/− cells cultured under the same conditions.

We next tested the requirement for STAT proteins in Th1-mediated repression of TGF-β₁-induced Foxp3. While IL-12 activates STAT3 and STAT4, only STAT4 was required for the ability of IL-12 to inhibit aTreg-cell development (Fig. 1b). Similar to published results, IL-4 only required STAT6 and IL-6 only required STAT3 for inhibition of Foxp3 expression (Fig. 1b).

We have recently described isoforms of STAT4 that have overlapping but distinct functions.^26,27 To determine if there was a differential ability of the isoforms, we isolated T cells from transgenic mice expressing either STAT4α or STAT4β on a Stat4^−/− background and cultured them in aTreg conditions, in the presence or absence of cytokines and antibodies that promote Th1 development. Both STAT4α and STAT4β were capable of repressing the development of Foxp3⁺ cells (Fig. 1c).

Previous data on the repression of aTreg-cell development by IL-12 suggested that it was mechanistically occurring through a STAT4/IFN-γ/STAT1/T-bet-dependent pathway.¹³ We hypothesized that if this were the case then mice deficient in both STAT4 and T-bet would have a similar phenotype to each of the single gene-deficient mice. To test this we generated Stat4^−/−Tbx21^−/− mice by mating single gene-deficient mice and compared the ability of IL-12 to repress Foxp3 expression in these mice. Interleukin-12 had a similarly decreased ability to repress Foxp3 in either Stat4^−/− or Tbx21^−/− cultures compared with repression in WT cells (Fig. 1d). However, IL-12 had less of an effect on repressing Foxp3 expression in Stat4^−/−Tbx21^−/− cultures than in Tbx21^−/− cultures (Fig. 1d). These data suggest that STAT4 and T-bet are not in a linear pathway in the repression of Foxp3 and that STAT4 has effects in repressing Foxp3 that are independent of T-bet.

STAT4 deficiency does not affect nTreg-cell development or function

Since STAT4 inhibited the development of aTreg cells, we tested whether there were any defects in nTreg-cell development in the absence of STAT4. Thymocytes, splenocytes and mesenteric lymph node cells were isolated from WT and Stat4^−/− mice and stained for CD4, CD25 and intracellular Foxp3. As shown in Fig. 2(a), there were no significant differences in the percentages of CD4⁺ CD25⁺ Foxp3⁺ T cells present in WT and Stat4^−/− mice in any of these lymphoid organs. To assess the function of these Treg cells, CD4⁺ CD25⁺ splenic T cells from WT and Stat4^−/− mice were purified and tested for suppressor activity using a [³H]thymidine incorporation assay. There was no significant difference between WT and Stat4^−/− Treg cells in their ability to suppress the proliferation of CD4⁺ CD25⁻ T cells (Fig. 2b). Consequently, STAT4 does not significantly affect nTreg-cell development or function.

Figure 2.

Signal transducer and activator of transcription (STAT4) is not required for natural regulatory T (nTreg) cell development or function. (a) Splenocytes, thymocytes and mesenteric lymph nodes were isolated from wild-type (WT) and Stat4^−/− mice and stained for CD4, CD25 and intracellular Foxp3. Dot plots indicate surface CD25 and intracellular Foxp3 staining for the gated CD4⁺ population. Numbers indicate the percentage of cells positive ± SD for cells from three mice. (b) CD4⁺ CD25⁺ Treg cells isolated by magnetic antibody cell sorting selection from WT and Stat4^−/− spleens were assayed for suppressor function as described in Fig. 1(a).

IL-12 stimulation inhibits STAT5 binding to the Foxp3 gene

The ability of IL-2 to promote Foxp3 expression is dependent upon STAT5.⁴ We hypothesized that a potential mechanism for the STAT4-dependent inhibition of Foxp3 was altering the IL-2/STAT5-dependent effects on Foxp3 expression. The addition of IL-12 did not greatly affect IL-2 production and other Th-promoting cytokines increased IL-2 production (Fig. 3a). Moreover, the addition of IL-2 did not alter the repression of Foxp3 in the presence of IL-12 or other Th-promoting cytokines (Fig. 3b). Therefore, the STAT4-dependent inhibition of Foxp3 was not the result of a lack of IL-2 in the culture system.

Figure 3.

Normal interleukin-2 (IL-2) signalling and phosphorylated signal transducer and activator of transcription 5 (pSTAT5) activation in cultures stimulated with T helper instructive cytokines. (a) Cell-free supernatants were collected 48 hr after CD4⁺ T cells were plated in the indicated culture conditions. Interleukin-2 production was tested using enzyme-linked immunosorbent assay. Data are represented as mean ± SD of two independent replicates. Results are representative of at least two independent experiments. (b) CD4⁺ T cells were cultured in the indicated conditions ± 100 U/ml hIL-2 for 3 days. After 3 days, cells were collected, washed and intracellularly stained for Foxp3 expression. Data are represented as mean ± SD of two replicate samples. Results are representative of two independent experiments. (c) CD4⁺ T cells were cultured in the indicated conditions for 72 hr and stained with anti-CD25 and anti-Foxp3 for analysis by flow cytometry. Numbers represent percentage of cells present in the upper left and upper right quadrant. Results are representative of two independent experiments. (d) CD4⁺ T cells cultured in the indicated conditions for 4, 24 or 48 hr were collected for intracellular staining with anti-pSTAT5 and anti-Foxp3. Numbers represent the percentage of total pSTAT5⁺ cells. Results are representative of two to three independent experiments.

We next wanted to examine if IL-2 signalling was affected by the presence of IL-12 in these cultures by assessing IL-2 receptor α chain expression and STAT5 phosphorylation during the first 2 days of culture, a time when the expression of Foxp3 is not different between aTreg and TGF-β₁ + instructive cytokine-stimulated cultures (Fig. 4a and data not shown). Expression of the IL-2 receptor α (CD25), was not compromised in the presence of IL-12, and was actually increased by the addition of IL-4 and IL-6 to cultures on both Foxp3⁺ and Foxp3⁻ cells (Fig. 3c). Interleukin-12-stimulated cells, or cells cultured with other instructive cytokines, did not have diminished pSTAT5 after 4, 24 or 48 hr of culture, compared with aTreg-cell cultures (Fig. 3d). Hence, IL-2 signalling was intact in aTreg-cell cultures in the presence of IL-12.

Figure 4.

Decreased signal transducer and activator of transcription 5 (STAT5) binding to Foxp3 following incubation with cytokines instructive in T helper development. (a) Wild-type (WT) CD4 T cells were activated with anti-CD3/anti-CD28 and cultured with transforming growth factor-β₁ (TGF-β₁) alone or TGF-β₁ + IL-12 and analysed for Foxp3 expression by intracellular staining. *Significantly different (P<0·05) from cells cultured in the presence of IL-12. Results are representative of at least three independent experiments. (b) Wild-type CD4 T cells were cultured with TGF-β₁ alone (adaptive regulatory T cells; aTreg), TGF-β₁ + IL-4, TGF-β₁ + IL-6 (Th17) or TGF-β₁ + IL-12 for 48–72 hr, collected and fixed; then sonicated chromatin was precipitated with antibodies to tri-methyl-H3K9. Results are expressed as fold-enrichment over chromatin immunoprecipitation with control immunoglobulin G and are the average ± SD of two experiments. *Significantly different (P<0·05) from cells cultured in the presence of IL-4, IL-6 or IL-12. (c) Chromatin from cultures generated as in (b) was immunoprecipitated with anti-STAT5. The amount of STAT5 bound to the indicated region in the aTreg-cell culture conditions was set at 100% and the amount of STAT5 bound to the indicated regions in the other culture conditions was compared to the aTreg conditions. Data are represented as mean ± SD and are averages of three independent experiments. *Significantly different (P<0·05) from cells cultured in the presence of IL-4 or IL-6 (top), or IL-4, IL-6 and IL-12 (bottom).

To further address a mechanism for how STAT4 inhibited the development of aTreg cells, we first wanted to determine when, during the differentiation period, Foxp3 was regulated. We examined Foxp3 induction using CD4⁺ T cells stimulated under aTreg or TGF-β₁ + Th1 conditions over a 5-day differentiation assay. In the first 48 hr of culture, Foxp3 was similarly induced in aTreg and TGF-β₁ + Th1 cultures (Fig. 4a). After 48 hr, the presence of IL-12 limited further induction of Foxp3 protein levels while aTreg cultures increased Foxp3 expression through the remainder of the culture period. We therefore focused our analysis on the 48–72 hr time-frame when Foxp3 was being actively regulated.

To determine if chromatin at the Foxp3 locus was being altered by cytokine exposure during this time-frame where the percentages of Foxp3 in culture are not different, we performed chromatin immunoprecipitation (ChIP) for the H3K9me3 modification associated with gene repression.^28–30 The addition of IL-12, as well as IL-6 and IL-4 to TGF-β₁-induced aTreg-cell cultures resulted in a fourfold increase in this modification (Fig. 4b), suggesting that these cytokines promote a repressive chromatin environment and might limit the association of Foxp3-inducing transcription factors.

STAT5 has been shown to bind the Foxp3 promoter and intron.^4,23 To test if IL-12 stimulation alters STAT5 binding, we used chromatin immunoprecipitation to assess the level of STAT5 bound to the promoter and first intron in aTreg TGF-β₁ + IL-4, Th17, or TGF-β₁ + IL-12 cultured cells at the 72-hr time-point where there is active repression of Foxp3 (Fig. 4a). While IL-4 and IL-6 significantly decreased STAT5 binding to the Foxp3 promoter and intron compared to CD4⁺ T cells cultured under aTreg conditions, IL-12 induced a significant decrease of STAT5 binding only to the intron site (Fig. 4c). This suggests that IL-12, by promoting a repressive chromatin environment at the Foxp3 locus, is able to limit STAT5 binding.

STAT4 limits Treg cells in allergic airway inflammation

Although STAT4 is required for Th1-mediated inflammation, it was also found to be important for the development of allergic inflammation.^31,32 The mechanism of this defect is still unclear but appears to involve decreased IL-17 and chemokine production. We speculated that these phenotypes might arise from altered Treg-cell generation. To test this we sensitized WT and Stat4^−/− mice to OVA and challenged mice with OVA intranasally. As shown before, Stat4^−/− mice have decreased, though not absent, pulmonary inflammation. There were specific decreases in lymphocyte and eosinophil populations from the BAL (Fig. 5a). We also examined the presence of Treg cells in the BAL and noted a significant increase in CD4⁺ CD25⁺ Foxp3⁺ cells in Stat4^−/− mice, compared to WT mice (P < 0·009) (Fig. 5b). While the absolute cell number of Foxp3⁺ and Foxp3⁻ CD4⁺ cells was decreased in the BAL of Stat4^−/− mice compared with WT mice, the ratio of Foxp3⁺ CD25⁺ CD4⁺ cells to total CD4⁺ cells was higher in Stat4^−/− than WT populations (Fig. 5c). Foxp3 messenger RNA in lung tissue was also increased in tissue from Stat4^−/− mice compared to WT mice (Fig. 5d). In contrast to the increases in Treg cells present in the lungs, there were decreases in the production of the inflammatory cytokines IL-5, IL-13 and IL-17 in the BAL fluid (Fig. 5e). These results suggest that STAT4 limits Treg cells during allergic inflammatory responses.

Figure 5.

Increased regulatory T (Treg) cells in lungs of signal transducer and activator of transcription 4 (STAT4) -deficient mice with allergic inflammation. (a) Wild-type (WT) and Stat4^−/− mice were sensitized with ovalbumin (OVA)/alum and challenged with OVA intranasally. Forty-eight hours after the last intranasal challenge bronchoalveolar lavage (BAL) was performed and cells were analysed by flow cytometry. Results are the average of four to five mice and are representative of two experiments. (b) Flow cytometric analysis of CD4⁺ CD25⁺ Foxp3⁺ cells in BAL from WT and Stat4^−/− mice challenged as in (a) with data presented as CD25 and Foxp3 expression of the CD4⁺ gated population. Percentage of Treg cells from Stat4^−/− BAL is significantly increased compared to WT cells (P < 0·009 by unpaired Student’s t-test). (c) The ratio of CD4⁺ CD25⁺ Foxp3⁺ cell number to total CD4⁺ cell number is indicated for BAL cells from WT and Stat4^−/− lungs. Data are presented as mean ± SD of the value from each mouse described in (a). (d) Lung tissue from WT and Stat4^−/− mice sensitized and challenged as in (a) was used to isolate RNA for quantitative polymerase chain reaction analysis of Foxp3. (e) Cytokine levels from BAL fluid obtained from mice described in (a) were determined using enzyme-linked immunosorbent assay.

Discussion

The development of adaptive Treg cells is regulated by a variety of cytokines. Although the ability of IL-6/STAT3, IL-4/STAT6 and IL-12/STAT4 to mediate repression of Foxp3 expression in developing aTreg cells is appreciated,^11,13 previous reports have left some questions regarding the precise mechanism of these functions. In this report we demonstrate that IL-12 inhibits the development of aTreg cells through the activation of STAT4, but not STAT3, which works in a parallel pathway with T-bet. Interleukin-12, like other Th instructive cytokines, promotes a repressive chromatin configuration at the Foxp3 locus and limits STAT5 binding to the locus. Moreover, in a model of allergic lung inflammation, Stat4^−/− mice have increased levels of Treg cells in the lung that correlate with decreased inflammation. These results suggest that STAT4 limits aTreg-cell development in vitro and Treg-cell numbers in a model of allergic inflammation.

The Treg cells in the allergic lung inflammation model could be either aTreg or nTreg, though it is unclear how STAT4 might contribute to nTreg-cell function. Previous reports have suggested that STAT4 may play a role in limiting the antigen-specific expansion of Treg cells though it is not clear from those studies if nTreg or aTreg cells are affected.³³ There are normal numbers of functional nTreg cells in Stat4^−/− mice (Fig. 2), and it seems unlikely that STAT4 contributes to nTreg-cell migration because this is not likely to be an IL-12-dependent response. Moreover, incubation of nTreg cells with IL-12 in vitro did not significantly alter the percentage of Foxp3⁺ cells, suggesting that IL-12 does not alter an already established Treg phenoytpe (data not shown). Based on our results demonstrating that STAT4 functions to limit aTreg-cell development in vitro, and in the absence of an IL-12 effect on nTreg cells, we speculate that the increases in Treg cells observed in the allergic lung inflammation model are increased aTreg cells.

The role for STAT4 in regulating a Th2 inflammatory disease is somewhat surprising considering that it is more associated with Th1 immunity. Previous reports have correlated decreased allergic inflammation in Stat4^−/− mice with decreases in IL-17 production and local chemokine production.^31,32 Indeed, these two observations may be linked because IL-17 is a potent inducer of chemokine production.³⁴ It is possible that Treg cells normally contribute to this aspect by regulating IL-17 production from Th17 or other proinflammatory cells. Recently, Treg cells have been shown to sculpt chemokine gradients during a viral infection, and while IL-17 production was not examined, it could be part of a regulatory circuit.³⁵ Our data, together with other reports, suggest that STAT4 may play a negative role in aTreg-cell development (this report) and a positive role in Th17 development,^31,36 both contributing to altered inflammation.

How STAT4 limits Foxp3 expression is still not entirely clear, although we provide a mechanism that is distinct from those previously described, specifically in a pathway parallel to T-bet function. We have shown that in the presence of the instructive cytokines IL-4, IL-6 and IL-12, there are increased levels of repressive chromatin modifications and decreased STAT5 binding to Foxp3, despite normal IL-2 signalling (Figs 3 and 4). It is interesting that IL-12 only reduced STAT5 binding to the Foxp3 intron and not to the promoter, and this may be linked to the decreased ability of IL-12 to repress Foxp3 compared with IL-4 or IL-6. Other reports have suggested that these cytokines repress Foxp3 by inducing their respective lineage determining factors.^11,13 However, while STAT6 induces GATA-3 and STAT3 induces RORγt, STAT4 is not an efficient inducer of T-bet,^9,36–40 and while transduced T-bet can partially rescue STAT4 deficiency it does not recapitulate WT levels of cytokine production or histone modification.^41–43 Moreover, while other reports have linked the effects of IL-12 and STAT4 to IFN-γ,¹³ we observed only modest effects of neutralizing IFN-γ or of STAT1 deficiency on the repression of Foxp3. The previous finding that IL-12-induced repression of Foxp3 was dependent on IFN-γ receptors could be explained if Ifngr1^−/− cells were not fully responsive to IL-12. Our results suggest that STAT4 and T-bet are not in a linear pathway, but rather function in parallel pathways to both direct Th1 development²⁰ and inhibit the generation of aTreg cells. STAT4 may bind to the Foxp3 locus to mediate gene repression, although there was not increased binding of STAT4 to the Foxp3 promoter or intron in IL-12-cultured cells (data not shown) suggesting that decreased STAT5 binding is not a simple displacement of one factor for the other. A recent report has suggested that STAT6 binds to another regulatory element in the Foxp3 gene to repress expression though whether STAT4 binds to this element is not clear.¹² Indeed, STAT4 may bind other regions or activate the expression of intermediate factors that repress Foxp3 expression. Further studies will explore these issues.

Although considerable literature has focused on the ability of TGF-β₁ to inhibit Th1 and Th2 differentiation,⁴⁴ our results support a role for IL-4 and IL-12 in decreasing the development of Foxp3⁺ aTreg cells. The STAT-dependent effects of Th1-, Th2- and Th17-promoting cytokines suggest that aTreg cells only efficiently develop in the absence of proinflammatory cytokines that mediate pathogen immunity, atopy and autoimmunity highlighting the delicate balance that exists between the development of proinflammatory and anti-inflammatory immunity.

Glossary

Abbreviations:

aTreg

adaptive T regulatory cell

BAL

bronchoalveolar lavage

nTreg

natural Treg

OVA

ovalbumin

Disclosures

The authors have no competing financial interests.