Opposing roles of STAT1 and STAT3 in IL-21 function in + CD4 T cells

Chi-Keung Wana,1, Allison B. Andraskia,1,2, Rosanne Spolskia, Peng Lia, Majid Kazemiana, Jangsuk Oha, Leigh Samselb, Phillip A. Swanson IIc, Dorian B. McGavernc, Elizabeth P. Sampaiod, Alexandra F. Freemand, Joshua D. Milnere, Steven M. Hollandd, and Warren J. Leonarda,3

aLaboratory of Molecular Immunology and Immunology Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892-1674; bFlow Cytometry Core, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892-1674; cViral Immunology and Intravital Imaging Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892-1674; dLaboratory of Clinical Infectious Diseases, National Institutes of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892-1674; and eLaboratory of Allergic Diseases, National Institutes of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892-1674

Contributed by Warren J. Leonard, June 18, 2015 (sent for review May 19, 2015; reviewed by Thomas R. Malek and Howard A. Young) + IL-21 is a type I essential for immune cell differentiation STAT1 to promote CD8 T-cell cytotoxicity and of and function. Although IL-21 can activate several STAT family mantle cell lymphoma (16, 17). We now have elucidated the transcription factors, previous studies focused mainly on the role roles of STAT1 in IL-21 signaling and identified an interplay of STAT3 in IL-21 signaling. Here, we investigated the role of STAT1 between STAT1 and STAT3 in mediating the actions of IL-21 + and show that STAT1 and STAT3 have at least partially opposing in CD4 T cells, and have also found increased IL-21–mediated rolesinIL-21signalinginCD4+ T cells. IL-21 induced STAT1 phos- + induction of STAT1 phosphorylation in cells from patients with phorylation, and this was augmented in Stat3-deficient CD4 T + autosomal dominant hyper-IgE syndrome (AD-HIES), and in cells. RNA-Seq analysis of CD4 T cells from Stat1- and Stat3-de- patients with a STAT1 gain-of-function (GOF) , which ficient mice revealed that both STAT1 and STAT3 are critical for IL- correlates with increased IFNG ( gamma) and TBX21 – 21 mediated regulation. Expression of some , including (T-box 21) expression after IL-21 stimulation. Tbx21 and Ifng, was differentially regulated by STAT1 and STAT3. Moreover, opposing actions of STAT1 and STAT3 on IFN-γ expression + Results in CD4 T cells were demonstrated in vivo during chronic lympho- IL-21 Induces Sustained STAT1 and STAT3 Activation in CD4+ T Cells. – cytic choriomeningitis infection. Finally, IL-21 mediated induction of IL-21 was previously shown to induce strong and sustained STAT3 STAT1 phosphorylation, as well as IFNG and TBX21 expression, were + phosphorylation (pSTAT3) but only weaker and more transient higher in CD4 T cells from patients with autosomal dominant hyper- STAT1 phosphorylation (pSTAT1) in total splenocytes, B cells, IgE syndrome, which is caused by STAT3 deficiency, as well as in cells + and CD8 T cells (18, 19). We first compared IL-21–induced from STAT1 gain-of-function patients. These data indicate an inter- + + play between STAT1 and STAT3 in fine-tuning IL-21 actions. pSTAT1 and pSTAT3 in preactivated CD4 and CD8 Tcells. IL-21 induced strong pSTAT1 and pSTAT3 at 30 min (Fig. 1 A IL-21 | T cells | STATs | IFN-γ | AD-HIES Significance nterleukin-21 (IL-21) is a type I cytokine that signals via a composed of IL-21R and the common cytokine re- IL-21 is a type I cytokine important for immune cell differenti- I ation and function. We found that transcription factors STAT1 ceptor γ-chain, γc (1). γc is also shared by the receptors for IL-2, IL-4, IL-7, IL-9, and IL-15 and is mutated in humans with and STAT3 play partially opposing roles in IL-21 function in + – X-linked severe combined immunodeficiency (XSCID), a dis- CD4 T cells. Both STAT1 and STAT3 control IL-21 mediated gene regulation, with some genes, including Ifng, Tbx21, and ease characterized by the absence of T and natural killer (NK) γ cells and the presence of nonfunctional B cells (2). IL-21 is Il21 reciprocally regulated by these STATs. IFN- production + was also differentially regulated by these STATs in vitro during primarily produced by CD4 T cells and natural killer T (NKT) + CD4 T-cell differentiation and in vivo during chronic lympho- cells, but it has pleiotropic actions on both adaptive and innate cytic choriomeningitis infection. Importantly, IL-21–induced immune cells, including T, B, NK, NKT, and dendritic cells (1). + IFNG and TBX21 expression was higher in CD4 T cells from In T cells, IL-21 can act as a comitogen and cooperates with IL-7 + patients with autosomal dominant hyper-IgE syndrome or with and IL-15 to expand CD8 T cells (3), promotes Th17 differ- – STAT1 gain-of-function , suggesting that dys-regu- entiation (4 6), and induces BCL6 expression (7) to promote T lated IL-21–STAT signaling partially explains the clinical mani- follicular helper cell development (8, 9). In B cells, IL-21 pro- festations of these patients. motes plasma cell differentiation (10, 11), and in combination with IL-4, drives IgG1 and IgG3 class switch (11, 12). Defective Author contributions: C.-K.W., A.B.A., P.L., and W.J.L. designed research; C.-K.W., A.B.A., R.S., signaling by IL-21 appears to substantially explain the B-cell P.L., M.K., J.O., L.S., P.A.S., E.P.S., and A.F.F. performed research; C.-K.W., A.B.A., R.S., P.L., defect observed in patients with XSCID (11, 13). Furthermore, M.K., D.B.M., and W.J.L. analyzed data; and C.-K.W., A.B.A., P.L., D.B.M., J.D.M., S.M.H., and W.J.L. wrote the paper. IL-21 can enhance the cytotoxic activity of NK and NKT cells (1) Reviewers: T.R.M., Miller School of Medicine, University of Miami; and H.A.Y., National and induce the apoptosis of conventional dendritic cells (14). Cancer Institute. IL-21 activates multiple signaling pathways, including the Conflict of interest statement: W.J.L. and R.S. are inventors on NIH patents related to IL-21. JAK-STAT, PI 3-kinase (PI3K), and MAPK pathways (15). Of Data deposition: The data reported in this paper have been deposited in the Gene Ex- these, the JAK-STAT pathway has been most extensively stud- pression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE63204). ied. IL-21 induces phosphorylation of JAK1 and JAK3, which 1C.-K.W., and A.B.A. contributed equally to this work. in turn leads to phosphorylation and nuclear translocation of 2Present address: Department of Nutrition, Harvard T.H. Chan School of Public Health, STAT3, which then binds to IFN-γ–activated sequence (GAS) Boston, MA 02115. – 3 motifs and modulates expression of IL-21 responsive genes. To whom correspondence should be addressed. Email: [email protected]. – IL-21 also activates STAT1, but the function of IL-21 activated This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. STAT1 is largely unknown, although IL-21 was suggested to use 1073/pnas.1511711112/-/DCSupplemental.

9394–9399 | PNAS | July 28, 2015 | vol. 112 | no. 30 www.pnas.org/cgi/doi/10.1073/pnas.1511711112 Downloaded by guest on September 29, 2021 and B). pSTAT1 as well as pSTAT3 was still detected at 4 and 8 h signaling (19). Moreover, we observed augmented IL-21– of stimulation (Fig. 1 A and B), and when we deleted Stat3 by induced expression of a number of genes in the absence of crossing Stat3-floxed mice to CD4-Cre transgenic mice (referred to STAT1 or STAT3, suggesting that these STAT also − − as Stat3 / herein), IL-21 induced more pSTAT1 at 4 and 8 h in directly or indirectly inhibit expression. Notably, nearly ∼50% − − + Stat3 / CD4 T cells than in WT cells (Fig. 1 A and C); this in- (84 of 173) of IL-21–regulated, STAT1-dependent genes were + crease was minimally observed in CD8 Tcells(Fig.1B and C). also STAT3-dependent (Fig. 2C and Dataset S3). Some genes Interestingly, we observed diminished IL-21–induced expression were differentially regulated by STAT1 and STAT3. For exam- − − of Socs3 and Socs1 in Stat3 / cells (Fig. 1D). Because SOCS3 and ple, the Th1 cell signature genes Tbx21 (Fig. 2D) and Ifng (Fig. SOCS1 can negatively regulate cytokine signaling by blocking 2E) were not induced by IL-21 in the absence of STAT1, but STAT phosphorylation (20), this might in part explain the their expression was elevated in the absence of STAT3, consis- increased pSTAT1 in the absence of STAT3. Using ImageStream tent with an inducing role for STAT1 and repressive role for + STAT3. This finding is also consistent with a report showing that to image fluorescently labeled pSTAT1 and pSTAT3 within CD4 + IL-21 induces Tbet expression in CD8 T cells via STAT1 (16). T cells, we confirmed that both STAT1 and STAT3 translocated – to the nucleus after IL-21 stimulation (Fig. 1E and Fig. S1). We hypothesized that IL-21 activated STAT1 and STAT3 di- rectly regulated the expression these genes and therefore used + Both STAT1 and STAT3 Contribute to IL-21–Mediated Gene Regulation ChIP-Seq to determine STAT1 and STAT3 binding in CD4 T − − – in CD4+ T Cells. We next used RNA-Seq in WT, Stat1 / , and cells after IL-21 stimulation. IL-21 induced STAT3 bound at − − + ′ Stat3 / CD4 T cells to elucidate the roles of STAT1 and sites 5 of the Tbx21 and Ifng loci (Fig. 2 F and G) containing STAT3 in IL-21–mediated gene regulation (Fig. 2A). Consistent GAS motifs (Table S1), and the level of STAT3 binding was lower in Stat1-deficient cells (Fig. 2 F and G). STAT1 binding with its activation by IL-21 (Fig. 1), STAT1 was responsible for was detected at the same sites and additionally at multiple sites the regulation of ∼10% (173 of 1,648) of the genes (fold- where STAT3 binding was minimal (Fig. 2 F and G). In contrast change > 1.5) after IL-21 stimulation (Fig. 2B and Dataset S1). to Tbx21 and Ifng, IL-21–induced expression of the Il21 gene was Although STAT3 is considered to be the major transcription ’ ∼ elevated in the absence of STAT1 but decreased in the absence factor responsible for IL-21 s effect, it only affected 40% (834 of STAT3 (Fig. 2H), and correspondingly, IL-21–mediated of 2,101) of the genes regulated by IL-21 (Fig. 2B and Dataset STAT3 binding was enhanced in Stat1-deficient cells (Fig. 2I), and S2), suggesting contributions of STAT-independent (e.g., MAPK STAT1 and STAT3 bound at distinct sites for regulating Il21 ex- – and PI3K) pathways, which are also involved in IL-21 mediated pression (Fig. 2I). These data collectively indicate key roles for + STAT1 for IL-21–induced in CD4 T cells and suggest that STAT1 and STAT3 can regulate gene expression through binding at the same or different sites at the Tbx21, Ifng, A CD4+ T cells B CD8+ T cells and Il21 loci. IL-21 4h 8h IL-21 4h 8h Stat3 +/+ -/- +/+ -/- +/+ -/- +/+ -/- Stat3 +/+ -/- +/+ -/- +/+ -/- +/+ -/- Opposing Effects of STAT1 and STAT3 on IL-21–Mediated IFN-γ pSTAT3 pSTAT3 Expression. By ChIP-Seq, we observed STAT1 binding in pre- Total Total + STAT3 activated CD4 T cells even without IL-21 stimulation (Fig. 2 F, INFLAMMATION

STAT3 IMMUNOLOGY AND pSTAT1 G, and I), and indeed, the basal level of phosphorylated STAT1 pSTAT1 + Total Total was also higher in preactivated Stat3-deficient CD4 T cells (Fig. STAT1 STAT1 1A). We thus investigated if T-cell receptor (TCR) signaling in 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 + CD naïve CD4 T cells augmented STAT1 phosphorylation in these Stat3+/+ ** Stat3+/+ cells, and found that TCR signaling strongly induced STAT1 2 15 Stat3-/- 6 Stat3-/- ** phosphorylation at 24 and 48 h after stimulation and persisted 10 4 for at least 72 h (Fig. 3A, lanes 3–8 vs. 1 and 2), but augmented 1 + 5 2 STAT1 phosphorylation in Stat3-deficient CD4 T cells was only Socs1/Rpl7 Socs3/Rpl7 0 0 detected at 48 and 72 h (Fig. 3 A and B, lanes 6 and 8 vs. 5 and 0 4h 8h 4h 8h Ctrl IL-21 Ctrl IL-21 γ (relative to total STAT1) 7). Addition of an IFN- blocking antibody diminished STAT1 CD4+ T cells CD8+ T cells pSTAT1 in Stat3-/- / Stat3+/+ pSTAT1 phosphorylation induced by TCR signaling (Fig. 3A, lanes 9–14), E DAPI pSTAT1 Merge DAPI pSTAT3 Merge suggesting that TCR-mediated IFN-γ expression induced STAT1 Ctrl activation. Importantly, however, STAT1 phosphorylation was + still higher in Stat3-deficient CD4 T cells even in the presence IL-21 of anti–IFN-γ (Fig. 3A, lanes 12 vs. 11 and 14 vs. 13, and Fig. 3B), indicating that IFN-γ was not responsible for this difference and Fig. 1. IL-21–induced STAT1 phosphorylation is enhanced in the absence of + + + + + suggesting that other might contribute to the aug- STAT3 in CD4 T cells. (A and B) CD4 (A) and CD8 (B) T cells from Stat3 / − − mented STAT1 phosphorylation. Because IL-21–induced Ifng / μ + + and Stat3 mice were stimulated with 5 g/mL plate-bound anti-CD3 mRNA expression was enhanced in Stat3-deficient CD4 T cells 2 μg/mL soluble anti-CD28 for 2 d, rested 1 d, then not stimulated or stim- (Fig. 2E), it was possible that IFN-γ signaling contributed to ulated with IL-21 for 30 min, 4 h, and 8 h. Levels of pSTAT3, STAT3, pSTAT1, – and STAT1 were assessed by Western blotting. Data are representative of the IL-21 induced augmented STAT1 phosphorylation in these three to five experiments. (C) Quantification of Western blotting images by cells; however, an IFN-γ blocking antibody did not significantly + ImageJ. (D) IL-21–mediated Socs3 and Socs1 mRNA expression in CD4 T cells affect STAT1 phosphorylation induced by IL-21 (Fig. 3 C and + + + − − is STAT3-dependent. CD4 T cells from Stat3 / and Stat3 / mice were stim- D), indicating that IL-21 worked independently from IFN-γ ulated with IL-21 as in A for 4 h, mRNA expression of Socs3 (Left) and Socs1 signaling for inducing higher STAT1 activation in Stat3-deficient + (Right) was analyzed by RT-PCR. Shown is expression relative to Rpl7; data CD4 T cells. are representative of three experiments. (E) IL-21–activated STAT1 and + Above, we showed that IL-21 could induce Tbx21 and Ifng STAT3 translocate to the nucleus. CD4 T cells were stimulated with IL-21 as + expression in preactivated CD4 T cells after 4 h of stimulation, in A for 30 min. Samples were fixed and stained with DAPI (nuclear stain, with an enhanced effect in the absence of STAT3. However, it purple), and with antibodies to pSTAT1 (green) or pSTAT3 (yellow), and γ analyzed by ImageStream. Overlapped DAPI and pSTAT1 or pSTAT3 staining was previously reported that IL-21 suppresses IFN- expression (Merge) indicates nuclear translocation. Data are representative of three during T-cell differentiation (21). We therefore investigated the + independent experiments. effect of IL-21 on IFN-γ expression during CD4 T-cell

Wan et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9395 Downloaded by guest on September 29, 2021 A Stat1 Stat3 B STAT1-independent STAT3-independent D +/+ -/- +/+ -/- STAT1-dependent STAT3-dependent IL-21 - + - + - + - + Untreated 3 10 1475 1648 1267 2101 IL-21 Socs1 genes genes 2 Socs3 173 834 5 Osm 1 Tbx21/Rpl7 472 0 0 100 91 500 Stat1+/+ Stat1-/- Stat3+/+ Stat3-/- 82 362 Tbx21 50 250 Irf8 E H Untreated Gzmb 50 Untreated IL-21 0 0 10 6 0.6 Induced Repressed Induced Repressed IL-21 4 0.4 C STAT1- STAT3- 5 25

dependent dependent Il21/Rpl7 2 0.2 Ifng/Rpl7 (173) 89 84 750 (834) 0 0 0 0 Stat1+/+ Stat1-/- Stat3+/+ Stat3-/- Stat1+/+ Stat1-/- Stat3+/+ Stat3-/- 1 2 3 4 5 6 7 8

FGChIP-Seq: STAT3 ChIP-Seq: STAT1 ChIP-Seq: STAT3 ChIP-Seq: STAT1 I ChIP-Seq: STAT3 ChIP-Seq: STAT1 57 24 54 Stat1+/+, untreated 5.3 9.7 Stat1+/+, untreated Stat1+/+, untreated 22.4 57 24 54 Stat1+/+, IL-21 Stat1+/+, IL-21 57 24 Stat1+/+, IL-21 334 57 24 54 Stat1-/-, untreated Stat1-/-, untreated 5.2 7.9 Stat1-/-, untreated 44.9 57 24 54 Stat1-/-, IL-21 Stat1-/-, IL-21 17 18.3 Stat1-/-, IL-21 554 28 15 15 Stat3+/+, untreated Stat3+/+, untreated 18.1 2.8 3.6 Stat3+/+, untreated 9.6 00.8 15 15 28 Stat3+/+, IL-21 21.5 19.4 3.2 Stat3+/+, IL-21 14.7 Stat3+/+, IL-21 00.2 28 15 15 Stat3-/-, untreated 27.6 4.6 3.6 Stat3-/-, untreated 10.3 Stat3-/-, untreated 22.3 28 15 15 Stat3-/-, IL-21 26 13.3 14.4 Stat3-/-, IL-21 10.5 Stat3-/-, IL-21 11.2 Tbx21 Ifng Il21

+ + + + − − + + − − Fig. 2. IL-21–mediated gene regulation in CD4 T cells requires both STAT1 and STAT3. (A–C) CD4 T cells from Stat1 / , Stat1 / , Stat3 / , and Stat3 / mice were preactivated with plate-bound anti-CD3 + anti-CD28 for 3 d, rested overnight, and not stimulated or stimulated with IL-21 for 4 h. Gene expression was determined by RNA-Seq. For each sample, RNA was prepared from combined cells from three mice. (A) Heat map comparing the IL-21–mediated gene ex- + pression in CD4 T cells with the indicated genotypes. (B) Number of IL-21 regulated genes that are STAT1- or STAT3-dependent (pie charts, Upper) and the number of these genes that are induced or repressed (bar graphs). (C) Venn diagram showing numbers of genes that are independently or coregulated by STAT1 and STAT3. (D and E) IL-21–mediated mRNA expression of Tbx21 (D) and Ifng (E)inStat1+/+, Stat1−/−, Stat3+/+, and Stat3−/− CD4+ T cells was determined + + + − − + + by RT-PCR. Shown are combined results from three mice with expression relative to Rpl7.(F and G) Preactivated CD4 T cells from Stat1 / , Stat1 / , Stat3 / , − − and Stat3 / mice were stimulated with IL-21 for 1 h, and STAT1 and STAT3 binding at Tbx21 (F) and Ifng (G) loci was determined by ChIP-Seq analysis. Tag numbers for peaks are indicated. (H)AsinD and E, IL-21–mediated mRNA expression of the Il21 gene was determined by RT-PCR. Shown are combined results of three mice and expression relative to Rpl7.(I)AsinF and G, STAT1 and STAT3 binding at the Il21 was determined by ChIP-Seq analysis.

differentiation. After 24 h of stimulation, IL-21 induced Tbet IL-21, with higher induction in Stat3-deficient cells at the 24 + expression in CD4 T cells, and Tbet expression was enhanced and 72 h time points (Fig. 4A and Dataset S4), including Ifng, when Stat3 was deleted (Fig. 3E). IL-21 diminished Ifng ex- Tbx21, Ifit1,andIfit2 (Fig. 4 B–E). The addition of anti–IFN-γ pression in Stat3-sufficent cells during 4 d of differentiation (Fig. reduced IL-21–induced expression of these Th1 signature genes 3F), but it increased Ifng expression in Stat3-deficient cells (Fig. in both WT and Stat3-deficient cells (Fig. 4 and Dataset S4), 3F). These results are consistent with lower IFN-γ protein pro- suggesting that even though IFN-γ signaling did not affect the duction after 4 d of differentiation with IL-21 in WT cells but augmented STAT1 phosphorylation induced by IL-21, it was γ increased IFN- expression in similarly treated Stat3-deficient partly responsible for IL-21–mediated gene expression in Stat3- cells (Fig. 3G). Consistent with the fact that IL-21 acted in- deficient cells. dependently of IFN-γ signaling, addition of anti–IFN-γ sup- + pressed intracellular IFN-γ expression in WT CD4 T cells, but it Opposing Effects of STAT1 and STAT3 During Lymphocytic γ did not affect the increased IFN- expression in Stat3-deficient Choriomeningitis Infection. The opposing effects of STAT1 and γ + cells (Fig. 3G). STAT1 is absolutely required for IFN- expres- STAT3 on IFN-γ production in CD4 T cells were further ex- sion, as the basal expression of intracellular IFN-γ was abolished + amined in vivo using lymphocytic choriomeningitis (LCMV) in- − − − − in Stat1-deficient CD4 T cells (Fig. 3H). As a result, the IL- fection. We infected Stat1 / , Stat3 / , and control mice with – γ Stat3 21 mediated IFN- expression that we observed in - LCMV clone 13 and determined IFN-γ production in antigen- deficient cells could not be detected in Stat1-deficient cells. + specific CD4 T cells. Consistent with our in vitro results, the Although the IL-21–induced augmented STAT1 phosphory- numbers of antigen-specific (Fig. 5A) and IFN-γ–producing (Fig. lation was independent of IFN-γ signaling, expression of IFN-γ + − − 5B) CD4 T cells were decreased in Stat1 / mice infected with mRNA (Figs. 2E and 3F) and protein (Fig. 3G) was increased in −/− these cells. Therefore, we asked if the increased IFN-γ expres- LCMV compared with WT littermates, but increased in Stat3 sion affected IL-21–mediated gene expression, particularly for mice (Fig. 5 C and D). Interestingly, however, the numbers of γ– + Th1 signature genes (e.g., Ifng and Tbx21) whose induction by antigen-specific (Fig. 5E) and IFN- producing (Fig. 5F) CD8 −/− IL-21 was enhanced in Stat3-deficient cells. We performed a T cells were not significantly altered in Stat3 mice. These data + suggest that STAT1 and STAT3 play opposing roles in regulating kinetic analysis by stimulating WT and Stat3-deficient CD4 + T cells with IL-21, without or with anti–IFN-γ, and determined IFN-γ production in CD4 cells. IL-21 signaling is likely in- gene expression by RNA-Seq after 24 and 72 h of stimulation. volved, as previous reports showed that IL-21 is induced and By Ingenuity Upstream Regulator Analysis, we identified 53 plays a protective role in LCMV infection (22–24), although other STAT1-regulated genes whose expression was induced by STAT1- or STAT3-activating cytokines—including perhaps IL-6,

9396 | www.pnas.org/cgi/doi/10.1073/pnas.1511711112 Wan et al. Downloaded by guest on September 29, 2021 anti-IFN- negative mutations in STAT3. In most cases, mutations occur in ABTCR 24h 48h 72h 24h 48h 72h control the SH2 or DNA binding domains, leading to the partial loss of Stat3 +/+ -/- +/+ -/- +/+ -/- +/+ -/- +/+ -/- +/+ -/- +/+ -/- 2 anti-IFN- STAT3 activity (29), and Th17 cell differentiation is typically pSTAT1 defective in these patients (30). We stimulated preactivated + 1 CD4 T cells purified from AD-HIES patients (Table S2) with Total IL-21, and compared pSTAT1 levels and IFNG and TBX21 STAT1 0 24 48 72 mRNA expression with those seen with healthy controls. In Time (h) + -Actin (relative to total STAT1) normal human donors, IL-21 induced pSTAT1 in CD4 T cells pSTAT1 in Stat3-/- / Stat3+/+ pSTAT1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 by 30 min and then declined, but was still detected after 4 h (Fig. + CDanti-IFN- 6A). In CD4 T cells from patients with AD-HIES, pSTAT1 was TCR 4h 8h 4h 8h control anti-IFN- higher than in cells from normal donors, particularly at 4 h after Stat3 +/+ -/- +/+ -/- +/+ -/- +/+ -/- +/+ -/- + 3 IL-21 treatment (Fig. 6 A and B). Interestingly, CD4 T cells pSTAT1 2 from some AD-HIES patients also expressed higher levels of Total 1 total STAT1 (Fig. 6A). Moreover, IL-21 induced IFNG and + STAT1 0 TBX21 mRNA expression in human CD4 T cells, and particu- 4 8 + Time (h) -Actin (relative to total STAT1) larly for TBX21 the fold-induction was higher in AD-HIES CD4

pSTAT1 in Stat3-/- / Stat3+/+ pSTAT1 T (Fig. 6 C and D and Fig. S2), consistent with the suppressive 1 2 3 4 5 6 7 8 9 10 EFIsotype 2.5 Stat3+/+ ** Untreated Stat3-/- Stat3+/+ IL-21 1.5 ** ** anti-IFN- ** ** A Stat3 +/+ -/- +/+ -/- 0.5 Day 01 3 0 1 3 0 1 301 3 Stat3-/- Ifng/Rpl7 0.1 ** ** Igtp Irgm1 0 ** Tgtp2 2 Gbp10 IL-21 -+-+-+-+-+-+-+-+ Stat2 Tbet Day 1 Day 2 Day 3 Day 4 Gm14446 1 Irf7 anti-IFN- Oas2 G H Ctrl IL-21 Ifit1 0 Ctrl IL-21 Ctrl IL-21 Fam26f 35 832 19 4 Usp18 Fcer1g 1 Stat3+/+ Stat1+/+ Cxcr3 Prf1 0.6 0.4 Egr1 2 0.7 0.3 0.4 0.2 Slfn8 0.9 0.6

IFN- Ifng IFN- 38 53 727 Ccl3 Stat1-/- Ccl4 Stat3-/- Dpp4 I830012O16Rik 1.1 1.2 0.8 0.9 0.5 0.5 Ifit2 IL-17 Herc6 IL-17 Mx2 Casp1 Cmpk2 – γ + AI607873 Fig. 3. STAT3 suppresses IL-21 mediated Tbet and IFN- expression in CD4 Rsad2 INFLAMMATION + + + − − T cells. (A) Naïve CD4 T cells from Stat3 / and Stat3 / mice were stimu- Oasl1 IMMUNOLOGY AND H2 Eb1 lated with 5 μg/mL plate-bound anti-CD3 + 2 μg/mL soluble anti-CD28 Hif1a Fgl2 without or with 10 μg/mL anti-IFN-γ for 24–72 h. Levels of pSTAT1 and STAT1 Tbx21 Neurl3 were determined by Western blotting. Data are representative of three Gata3 Runx1 experiments. (B) Quantification of Western blotting images by ImageJ. Fasl −/− +/+ Ccrl2 Numbers indicate the ratios of pSTAT1 levels in Stat3 cells vs. Stat3 cells Irf5 −/− Gbp5 and dividing this value by total STAT1 levels in Stat3 cells relative to Psmb9 +/+ + +/+ −/− Casp3 Stat3 cells. (C) CD4 T cells from Stat3 and Stat3 mice were stimu- Gzmb lated with 5 μg/mL plate-bound anti-CD3 + 2 μg/mL soluble anti-CD28 for Klrk1 Isg15 3 d, rested 1 d, and then not stimulated or stimulated with IL-21 for 4 h, and Serpina3f Clic5 8 h, without or with the addition of 10 μg/mL anti–IFN-γ. Levels of pSTAT1 Tap1 Ccl5 and STAT1 were determined by Western blotting. Data are representative of Csf2 Eif2ak2 two experiments. (D) Quantification of Western blotting images by ImageJ Ccnd2 + +/+ −/− Cd86 S as in B.(E) Naïve CD4 T cells from Stat3 and Stat3 mice were stimu- W W W S S W W W S S S lated with 5 μg/mL plate-bound anti-CD3 + 2 μg/mL soluble anti-CD28 Stat3+/+ Stat3+/+ + anti-IFN- Stat3-/- Stat3-/- + anti-IFN- without or with 100 ng/mL IL-21 for 24 h, and Tbet expression was analyzed BC10 2 by flow cytometry. Data are representative of three experiments. (F) Naïve + + + − − CD4 T cells from Stat3 / and Stat3 / mice were treated as in E, and Ifng 5 1

expression was determined at different points relative to Rpl7; data are Ifng / Rpl7

+ Tbx21 / Rpl7 representative of three experiments. (G and H) Naïve CD4 T cells from 0 0 +/+ −/− +/+ −/− Stat3 and Stat3 (G), or Stat1 and Stat1 (H), mice were treated as in DE5 3 E for 4 d, without or with the addition of 10 μg/mL anti–IFN-γ. Cells were 2 / Rpl7

washed and restimulated with PMA + ionomycin and golgiplug for 4 h, and / Rpl7 2.5 1 intracellular IFN-γ and IL-17 were analyzed by flow cytometry. Data are 0 0 representative of three experiments. Day 031 013 013 013 Day 031 013 013 013

Fig. 4. IFN-γ signaling partially controls IL-21–mediated gene expression. + + + − − IL-10, or type I IFN, which are involved in LCMV responses (25– (A)CD4 TcellsfromStat3 / and Stat3 / mice were stimulated with 5 μg/mL 28)—might also contribute. plate-bound anti-CD3 + 2 μg/mL soluble anti-CD28 for 3 d, rested 1 d, then not stimulated or stimulated with IL-21 for 24 h or 72 h, without or with 10 μg/mL – γ Higher IL-21–Induced STAT1 Phosphorylation and IFNG and TBX21 anti IFN- . Gene expression was determined by RNA-Seq. Heat map showing a + list of STAT1-regulated genes (identified by Ingenuity Upstream Regulator Expression in CD4 T Cells from Patients with AD-HIES and STAT1 Analysis) whose expression is induced by IL-21, and the induction was further GOF Mutation. To clarify the roles of STAT1 and STAT3 in the – + increased in Stat3-deficient cells. (B E)ExpressionofIfng (B), Tbx21 (C), Ifit1 response to IL-21 in humans, we used CD4 T cells from pa- (D), and Ifit2 (E), which are in the list of genes shown in A, was confirmed by tients with AD-HIES, a genetic disorder caused by dominant- RT-PCR. Shown are representative of two experiments.

Wan et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9397 Downloaded by guest on September 29, 2021 CD4+ gated Discussion A Stat1+/+ Stat1-/- B ) ** 5 4 The role of STAT3-independent pathways in IL-21 signaling is not 4 0.3 20 ** Stat1+/+ Stat1-/- well understood. Here, we investigated the possible role of STAT1 3 0.4 + (%) + 19 15 in regulating IL-21 function in CD4 T cells, and we now show that I-Ab gp66 2 10 STAT1, as well as STAT3, is a mediator of the actions of IL-21 in 0.01 0.03 IFN- 20 16 these cells. Our RNA-Seq analysis revealed that 10% of the IL-21– CD4+ IFN- CD44 0 + 24 16 0 Stat1+/+ Stat1-/- regulated genes in CD4 T cells were STAT1-dependent and that I-Ab gp66+ (x10 Ctrl-tetramer Stat1+/+ Stat1-/- – CD44 almost half of these genes were coregulated by both IL-21 activated STAT1 and STAT3, suggesting that STAT1 not only exerts its own CDCD4+ gated Stat3+/+ Stat3-/- * actions but that it also interplays with STAT3 for fine-tuning IL-21– ) 20 ** 5 Stat3+/+ Stat3-/- mediated gene regulation. Conceivably, some of these genes may be 4 8 6 + (%) 2 7 regulated by STAT1-STAT3 heterodimers, although it is possible 38 52 4 10 I-Ab gp66 that some require both STAT1 and STAT3 homodimers instead of IFN- 0.3 0.4 15 40 heterodimers. Previously, it was shown that there is enhanced IFN- 2 CD4+ IFN- CD44 0 Stat3+/+ Stat3-/- γ–mediated STAT3 activation in Stat1-deficient mouse embryonic 40 61 0

I-Ab gp66+ (x10 fibroblasts (31) and enhanced IL-6 mediated STAT1 activation in

Ctrl-tetramer Stat3+/+ Stat3-/- CD44 Stat3-deficient mouse embryonic fibroblasts (32), revealing CD8+ gated that cytokines can activate more of one STAT when another is ) Stat3+/+ Stat3-/- )

E 5 5 NS NS absent. The balance of STAT1 and STAT3 is intriguing given that 8930 20

H-2Db gp33 77 81 15 5 7 10 A IL-21 4h IL-21 4h HIESND HIESNDND HIES ND HIES HIESND HIESND 68 70 0 0 pSTAT1 pSTAT1 Stat3+/+ Stat3-/- Stat3+/+ Stat3-/- H-2Db gp33+ (x10 H-2Db gp276

CD44 H-2Db gp276+ (x10 Total STAT1 Total STAT1 CD8+ gated F Stat3+/+ Stat3-/- -Actin -Actin NS 14 11 NS gp33-41 + (%) +20 (%) 10 peptide IL-21 4h IL-21 4h 59 72 HIESND ND HIES HIESND HIESND HIESND HIESND

IFN- 9 10 10 5 gp276-286 pSTAT1 pSTAT1 CD8+ IFN- peptide CD8+ IFN- 0 0 Total STAT1 Total STAT1 61 73 Stat3+/+ Stat3-/- Stat3+/+ Stat3-/- gp33-41 gp276-286 CD44 -Actin -Actin

Fig. 5. STAT1 and STAT3 have opposing effects on regulation of IFN-γ 4 expression in CD4+ T cells during LCMV infection. (A and B) Stat1+/+ and IL-21 4h B − − Stat1 / mice were infected with 2 × 105 pfu of LCMV clone 13, and total ND HIES HIESND HIESND 3 spleen cells were isolated at day 6 after infection. (A) Percentages (Left) pSTAT1 + 2 and numbers (Right) of LCMV-specific CD4 T cells were identified by I-Ab Total STAT1 gp66 tetramer staining. (B) Total spleen cells were stimulated with LCMV- (HIES/ND) 1

+ + + Relative pSTAT1 specific gp66-77 peptide for 5 h, percentages of CD4 CD44 IFN-γ cells -Actin 0 were determined. (Left) Individual mice; (Right) summary of data from 4h Time multiple animals. In A and B, data are from two experiments with a total + + − − of six to seven mice per group. (C–F) Stat3 / and Stat3 / mice were CDNormal donor Normal donor infected with 2 × 106 pfu of LCMV clone 13 and total spleen cells were AD-HIES AD-HIES 1.8 2.4 isolated at day 9 after infection. (C) Percentages (Left) and numbers P = 0.049 (n=5) P = 0.043 (n=5) + b (Right) of LCMV-specific CD4 T cells were identified by I-A gp66 tetra- 1.6 2.0 mer staining. (D) Total spleen cells were stimulated as in B; percentages of / ACTB 1.4 CD4+ CD44+ IFN-γ+ cells were determined. (Left) Individual mice; (Right) IFNG TBX21 / ACTB 1.6 summary of data from multiple animals. (E) Percentage (Left) and num- 1.2 + bers (Right) of LCMV-specific CD8 T cells were identified by H-2Db gp33 1.2 b 1

and H-2D gp276 tetramer staining. (F) Total spleen cells were stimulated Relative Relative with LCMV-specific gp33-41 or gp276-286 peptides for 5 h, and the 0.8 0.8 Control IL-21 Control IL-21 percentage of CD8+ CD44+ IFN-γ+ cells was determined. (Left)Arepre- + sentative analysis and the right two panels summarize data from mul- Fig. 6. IL-21–induced STAT1 phosphorylation is enhanced in CD4 T cells from + tiple animals. In C–F, data are from three experiments with 10–11 mice patients with AD-HIES. CD4 T cells from normal donors (ND) and AD-HIES per group. patients (HIES) were stimulated with 2 μg/mL plate-bound anti-CD3 and 1 μg/mL soluble anti-CD28 for 3 d, rested 1 d, then not stimulated or stimulated with IL-21for30minand4h.LevelsofpSTAT1andSTAT1weredeterminedby effect of STAT3 on the expression of these genes. We also ex- Western blotting; β-actin was used as a loading control. (A) Data are shown for + amined the effect of IL-21 on CD4 T cells from two patients five patients and five normal controls. (B) Quantification of Western blotting with STAT1 GOF mutation (Table S2). Similar to the effect of images in A by ImageJ. Numbers indicate the ratios comparing pSTAT1 levels in + IL-21 in CD4 T cells from patients with AD-HIES, IL-21– cells from patients relative to cells from normal donors. (C and D)IL-21–induced IFNG and TBX21 mRNA expression is enhanced in CD4+ T cells from patients induced STAT1 phosphorylation was enhanced in STAT1 GOF + + – with AD-HIES. CD4 T cells from normal donors and AD-HIES patients were CD4 T cells (Fig. S3 A and B), and the IL-21 induced IFNG activated as described in panel A, then not stimulated or stimulated with IL-21 and TBX21 expression was also augmented (Fig. S3 C and D). for 4 h. mRNA expression of IFNG (C)andTBX21 (D) was determined by RT-PCR. These data indicate that STAT1/STAT3 interplay influences Shown is fold-induction by IL-21 relative to ACTB, data from five experiments, IL-21–mediated gene regulation. with the level for untreated samples set to 1.

9398 | www.pnas.org/cgi/doi/10.1073/pnas.1511711112 Wan et al. Downloaded by guest on September 29, 2021 STAT1 is classically activated by , mediating an anti- of opposing effects of STAT1 and STAT3 and that a balance of viral and tumor-suppressor response, whereas STAT3 is an on- these STAT proteins is critical for normal host defense. cogene (33). We now have shown that STAT1 and STAT3 have + opposing roles in regulating IFN-γ expression in CD4 Tcells, Materials of Methods both in vitro when stimulated with IL-21 and in vivo when the Mice. All mice were used under animal protocols approved by the National + mice were infected with LCMV. Moreover, we found that IL-21– Heart, Lung, and Blood Institute Animal Care and Use Committee. CD4 − − mediated activation of STAT1 suppressed STAT3-dependent Il21 T cells lacking STAT3 were described previously (14). Stat1 / mice were expression, which presumably normally serves to auto-amplify backcrossed to the C57BL/6 background for six generations. See SI Materials IL-21 signaling. Thus, STAT1 activation may be a mechanism for and Methods for more information. limiting IL-21–mediated immune responses. Isolation and Activation of T Cells. Murine total and naïve T cells were isolated Several possible mechanisms may account for the interplay + – from spleen using magnetic beads (Miltenyi Biotec). Human CD4 T cells of STAT1 and STAT3 induced by IL-21. First, we found that IL-21 from peripheral blood mononuclear cells were isolated using Dynabeads mediated induction of the Socs1 and Socs3 genes was diminished in (Invitrogen). Cells were preactivated as reported (14). See SI Materials and the absence of STAT3, suggesting that the prolonged STAT1 acti- Methods for more information. vation may be a result of the attenuation of a negative regulatory signal. Second, the absence of either STAT1 or STAT3 would pre- Flow Cytometry. Surface and intracellular staining were performed per the vent the formation of STAT1/STAT3 heterodimers, presumable fa- standard protocol. For nuclear translocation of pSTAT proteins, cells were stained voring formation of STAT3 or STAT1 homodimers, respectively, as described previously (14), and analyzed on an ImageStream system. See SI with potential effects on gene expression. Third, our ChIP-Seq Materials and Methods for more information analysis revealed altered STAT3 binding to DNA in the absence of STAT1,afindingthatmayhelptoexplainalteredexpressionof LCMV Clone 13 Infection. For LCMV clone 13 infection, see SI Materials and Methods. STAT3-dependent genes in the absence of STAT1. Fourth, the opposing effects of STAT1 and STAT3 may also result from dif- Immunoblot Analysis. Immunoblotting were performed per the standard ferential effects on other molecules. For example, we found that protocol. See SI Materials and Methods for more information. STAT1 and STAT3 reciprocally regulated expression of both IFN-γ and Tbet (which is a critical mediator of IFN-γ)(34)afterstimulation RNA-Seq and ChIP-Seq. RNA-Seq and ChIP-Seq were performed as reported (14, 36). See SI Materials and Methods for more information. with IL-21 in vitro. Thus, multiple mechanisms might contribute to the opposing effects of STAT1 and STAT3 upon IL-21 stimulation. Statistical Analysis. Statistical analysis was performed by Student’s t test. See Corresponding to our data in the mouse, we also demonstrated SI Materials and Methods for more information. increased IL-21–induced STAT1 activation and higher IFNG and + TBX21 expression in CD4 T cells from patients with AD-HIES. ACKNOWLEDGMENTS. We thank Dr. Jian-Xin Lin (National Heart, Lung, and AD-HIES patients have recurrent infections, including chronic Blood Institute) for critical comments. This work was supported by the Divisions mucocutaneous candidiasis and impaired Th17 cell differentiation of Intramural Research, National Heart, Lung, and Blood Institute, National Institute of Neurological Disorders and Stroke, and National Institute of Allergy (30). Interestingly, patients with GOF mutation of STAT1 also and Infectious Diseases, National Institutes of Health. RNA-Seq and ChIP-Seq suffer from chronic mucocutaneous candidiasis as a result of de- were performed in the National Heart, Lung, and Blood Institute DNA Sequencing

fective Th17 cell differentiation (35), further supporting the idea and Genomics Core. INFLAMMATION IMMUNOLOGY AND

1. Spolski R, Leonard WJ (2008) -21: Basic biology and implications for cancer 20. Krebs DL, Hilton DJ (2001) SOCS proteins: Negative regulators of cytokine signaling. and autoimmunity. Annu Rev Immunol 26:57–79. Stem Cells 19(5):378–387. 2. Noguchi M, et al. (1993) Interleukin-2 receptor gamma chain mutation results in 21. Suto A, Wurster AL, Reiner SL, Grusby MJ (2006) IL-21 inhibits IFN-gamma production in – X-linked severe combined immunodeficiency in humans. Cell 73(1):147 157. developing Th1 cells through the repression of expression. JImmunol + 3. Zeng R, et al. (2005) Synergy of IL-21 and IL-15 in regulating CD8 expansion 177(6):3721–3727. – and function. J Exp Med 201(1):139 148. 22. Yi JS, Du M, Zajac AJ (2009) A vital role for interleukin-21 in the control of a chronic 4. Korn T, et al. (2007) IL-21 initiates an alternative pathway to induce proinflammatory viral infection. Science 324(5934):1572–1576. T(H)17 cells. Nature 448(7152):484–487. 23. Elsaesser H, Sauer K, Brooks DG (2009) IL-21 is required to control chronic viral in- 5. Nurieva R, et al. (2007) Essential autocrine regulation by IL-21 in the generation of fection. Science 324(5934):1569–1572. inflammatory T cells. Nature 448(7152):480–483. 24. Fröhlich A, et al. (2009) IL-21R on T cells is critical for sustained functionality and 6. Zhou L, et al. (2007) IL-6 programs T(H)-17 cell differentiation by promoting se- – quential engagement of the IL-21 and IL-23 pathways. Nat Immunol 8(9):967–974. control of chronic viral infection. Science 324(5934):1576 1580. 7. Ozaki K, et al. (2004) Regulation of differentiation and plasma cell generation 25. Harker JA, Lewis GM, Mack L, Zuniga EI (2011) Late interleukin-6 escalates T follicular – by IL-21, a novel inducer of Blimp-1 and Bcl-6. J Immunol 173(9):5361–5371. helper cell responses and controls a chronic viral infection. Science 334(6057):825 829. 8. Nurieva RI, et al. (2008) Generation of T follicular helper cells is mediated by interleukin- 26.CuiW,LiuY,WeinsteinJS,CraftJ,KaechSM(2011) An interleukin-21-interleukin-10-STAT3 21 but independent of T helper 1, 2, or 17 cell lineages. Immunity 29(1):138–149. pathway is critical for functional maturation of memory CD8+ Tcells.Immunity 35(5):792–805. 9. Vogelzang A, et al. (2008) A fundamental role for interleukin-21 in the generation of 27. Ray JP, et al. (2014) STAT3 and type I interferons are corepressive in- T follicular helper cells. Immunity 29(1):127–137. sulators for differentiation of follicular helper and T helper 1 cells. Immunity 40(3):367–377. 10. Kwon H, et al. (2009) Analysis of interleukin-21-induced Prdm1 gene regulation reveals 28. Havenar-Daughton C, Kolumam GA, Murali-Krishna K (2006) Cutting edge: The direct functional cooperation of STAT3 and IRF4 transcription factors. Immunity 31(6):941–952. action of type I IFN on CD4 T cells is critical for sustaining clonal expansion in response 11. Ozaki K, et al. (2002) A critical role for IL-21 in regulating immunoglobulin pro- to a viral but not a bacterial infection. J Immunol 176(6):3315–3319. duction. Science 298(5598):1630–1634. 29. Sowerwine KJ, Holland SM, Freeman AF (2012) Hyper-IgE syndrome update. Ann N Y 12. Pène J, et al. (2004) Cutting edge: IL-21 is a switch factor for the production of IgG1 Acad Sci 1250:25–32. – and IgG3 by human B cells. J Immunol 172(9):5154 5157. 30. Milner JD, et al. (2008) Impaired T(H)17 cell differentiation in subjects with autosomal γ 13. Recher M, et al. (2011) IL-21 is the primary common chain-binding cytokine required dominant hyper-IgE syndrome. Nature 452(7188):773–776. – for human B-cell differentiation in vivo. Blood 118(26):6824 6835. 31. Qing Y, Stark GR (2004) Alternative activation of STAT1 and STAT3 in response to 14. Wan CK, et al. (2013) The cytokines IL-21 and GM-CSF have opposing regulatory roles interferon-gamma. J Biol Chem 279(40):41679–41685. in the apoptosis of conventional dendritic cells. Immunity 38(3):514–527. 32. Costa-Pereira AP, et al. (2002) Mutational switch of an IL-6 response to an interferon- 15. Leonard WJ, Zeng R, Spolski R (2008) Interleukin 21: A cytokine/ gamma-like response. Proc Natl Acad Sci USA 99(12):8043–8047. system that has come of age. J Leukoc Biol 84(2):348–356. 33. Bromberg J, Darnell JE, Jr (2000) The role of STATs in transcriptional control and their 16. Sutherland AP, et al. (2013) IL-21 promotes CD8+ CTL activity via the transcription – factor T-bet. J Immunol 190(8):3977–3984. impact on cellular function. Oncogene 19(21):2468 2473. 17. Gelebart P, et al. (2009) Interleukin-21 effectively induces apoptosis in mantle cell 34. Szabo SJ, et al. (2000) A novel transcription factor, T-bet, directs Th1 lineage com- – lymphoma through a STAT1-dependent mechanism. Leukemia 23(10):1836–1846. mitment. Cell 100(6):655 669. 18. Konforte D, Paige CJ (2006) Identification of cellular intermediates and molecular 35. Liu L, et al. (2011) Gain-of-function human STAT1 mutations impair IL-17 immunity pathways induced by IL-21 in human B cells. J Immunol 177(12):8381–8392. and underlie chronic mucocutaneous candidiasis. J Exp Med 208(8):1635–1648. 19. Zeng R, et al. (2007) The molecular basis of IL-21-mediated proliferation. Blood 36. Lin JX, et al. (2012) Critical role of STAT5 transcription factor tetramerization for 109(10):4135–4142. cytokine responses and normal immune function. Immunity 36(4):586–599.

Wan et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9399 Downloaded by guest on September 29, 2021