Background
Multiple sclerosis (MS) is a demyelinating autoimmune disease characterized by the presence of CD4 T cells in inflammatory lesions within the central nervous system (CNS) [
1,
2]. Studies using the mouse model of MS, experimental autoimmune encephalomyelitis (EAE), have demonstrated that the Th1 and Th17 CD4 T cell subsets are associated with disease onset and that both subsets are capable of causing disease. Interestingly, while both Th1 and Th17 cells can initiate disease, the mechanisms by which these cells mediate inflammation and characteristics of the disease are different [
3‐
5]. For instance, Th1 cells preferentially migrate to the spinal cord and recruit macrophages to sites of inflammation, whereas Th17 cells primarily infiltrate the brain and recruit neutrophils. Nevertheless, it is possible that Th1 and Th17 CD4 T cells share properties that contribute to pathogenicity, and defining these potential commonalities may lead to new therapeutic targets.
A recent genome-wide association study (GWAS) identified a polymorphism in the signal transducer and activator of transcription 4 (STAT4) gene that is associated with MS susceptibility [
6]. STAT4 is a member of the STAT family of transcription factors and is a Th1 transcriptional regulator [
7,
8]. STAT4, when activated by IL-12, results in the development of Th1 cells and the production of the hallmark Th1 cytokine IFNγ. Paradoxically, neither IL-12 nor IFNγ is required for EAE, while STAT4 is essential. This highlights an important, unknown function for STAT4 during chronic CNS inflammation that is independent of the classic Th1 pathway [
9‐
15]. During Th1 differentiation, STAT4 is necessary to establish the genomic landscape, which then allows other transcription factors to bind Th1 lineage-associated genes [
7,
16,
17]. It remains unclear if STAT4 instructs the epigenetic landscape in effector CD4 T cells outside of the Th1 lineage. While Th17 differentiation is not contingent on STAT4, the role of this molecule in Th17 plasticity, and potentially Th17 gene expression, remains controversial [
18‐
20]. Hence, STAT4 may function in Th17 cells during EAE, possibly by shaping the accessibility and expression of encephalogenic genes.
In addition to IFNγ, the Th17 prototypic cytokine IL-17A is also dispensable for EAE, raising the question as to how these CD4 T cells mediate disease and if these effector subsets share an encephalogenic molecule [
15,
21]. Both Th1 and Th17 CD4 T cells produce GM-CSF, which has been demonstrated to be critical for EAE pathogenesis by both cell types [
22,
23]. GM-CSF functions to activate microglia within the CNS as well as recruit and stimulate peripheral macrophages and dendritic cells during EAE [
24,
25]. Recent studies show that STAT4 knockout (STAT4
−/−) T cells had diminished GM-CSF production [
26,
27]. These data suggest that STAT4 may regulate GM-CSF, which in turn drives the development of EAE. However, whether STAT4 acts in a CD4 T cell-intrinsic manner and if STAT4 regulates both Th1 and Th17-derived GM-CSF production during EAE were not determined.
In this study, we investigated the relationship between STAT4 and GM-CSF production during EAE. We demonstrate that STAT4 regulates GM-CSF expression by not only MOG-specific Th1 cells, but also Th17 and a population of GM-CSF+IFNγ−IL-17A− single-producing subset of cells during EAE. Coincident with a lineage-indiscriminate role for STAT4, we find that in vitro STAT4 functions to promote optimal GM-CSF production in CD4 T cells activated under non-polarizing, Th0 conditions. Using mixed bone marrow chimeric mice, we show that CD4 T cell-intrinsic STAT4 expression is important for GM-CSF production during EAE. Furthermore, STAT4 is able to directly bind to and regulate the Csf2 promoter in encephalogenic CD4 T cells. Overall, this study illustrates that STAT4 directly regulates the transcription of GM-CSF and highlights a previously unrecognized role for STAT4 in the function of Th17 cells.
Materials and methods
Mice
C57BL/6J, B6.SJL-
Ptprca Pep3b/BoyJ (WT CD45.1), and B6.129S7-
Rag1tm1Mom/J (Rag1
−/−) were purchased from the Jackson Laboratory. B6.STAT4
−/− (STAT4
−/−) mice were generously provided by Dr. Mark Kaplan [
28]. B6.
Ifng/Thy1.1 knock-in mice were described previously [
29]. Both C57BL/6J and B6.
Ifng/Thy1.1 knock-in mice were used as wild-type (WT) controls. All animals were bred and maintained under specific pathogen-free conditions at the University of Alabama at Birmingham according to Institutional Animal Care and Use Committee regulations.
Mixed bone marrow chimeric mice
Mixed bone marrow chimeric mice were generated as previously described [
30]. Rag1
−/− mice were irradiated with a split dose of 1000 rad and reconstituted with CD5-depleted bone marrow by intravenous injection. The transferred bone marrow cells were a mixture of 50 % CD45.1 WT bone marrow and 50 % CD45.2 WT bone marrow (WT:WT) or 50 % CD45.1 WT bone marrow and 50 % CD45.2 STAT4
−/− bone marrow (WT:STAT4
−/−). Recipient mice were maintained on antibiotic water for 6 weeks. Mice were immunized for EAE 10 weeks following reconstitution.
EAE induction and clinical scoring
Age and sex matched mice between 8 and 12 weeks of age were induced for EAE by subcutaneous immunization with 50 μg MOG35−55 peptide (Biosynthesis) emulsified in CFA (150 μg Mycobacterium tuberculosis; Difco) and intraperitoneal (i.p.) administration of pertussis toxin (200 ng; List Biological Laboratories) on days 0 and 2. Disease was monitored daily by the following criteria: 0, no disease; 1, tail paralysis; 2.0, hind limb paresis; 3.0, complete hind limb paralysis; 4.0, forelimbs paralysis; and 5, moribund.
Ex vivo stimulation
Single-cell suspensions of spinal cord, spleen, and inguinal LNs were prepared as previously described [
31]. The following antibodies were used: anti-CD4 PerCP-Cy5.5/APC/eFluor 450/PE-Cy7 (eBioscience, clone RM4-5), anti-CD45.1 FITC (eBioscience, clone A20), anti-CD45.2 PerCP-Cy5.5/APC (eBioscience, clone 104), and anti-CD44 FITC (eBioscience, clone IM7). For intracellular staining, cells were reactivated with culture media (negative control) or 5 μM MOG
35−55 peptide for 7 h with GolgiPlug (BD Biosciences) added for the final 4 h. The following intracellular antibodies were used in accordance with the manufacturer’s protocols: anti-IFNγ eFluor 450 (eBioscience, clone XMG1.2), anti-IL17A Alexa Fluor 647 (eBioscience, clone eBio17B7), anti-GM-CSF PE (BD Biosciences, clone MP1-22E9). A viability dye (Aqua, Life Technologies) was applied to exclude dead cells. Samples were acquired by using an LSRII flow cytometer (BD Biosciences) followed by data analysis using FlowJo version 9.x (Tree Star).
Naïve CD4 T cell polarization and activation
Naïve CD4+CD25−CD45RBhi (anti-CD25 PerCP-Cy5.5 (eBioscience, clone PC61.5); anti-CD45RB FITC (eBioscience, clone C363.16A)) T cells were sorted from WT and STAT4−/− mice using a FACSAria cell sorter (BD Biosciences) and cultured in the presence of irradiated WT feeders in R10 containing 2.5 μg/ml anti-CD3 (clone 145-11) for 5–6 days. Conditions also contained the following: Non-polarizing (Th0) with anti-IL-12p40—10 μg/ml anti-IL-12p40 (clone C17.8); Th1—10 ng/ml rmIL-12 and 10 μg/ml anti-IL-4 (clone 11B11); Th17 (TGFβ1)—10 ng/ml rmIL-23, 20 ng/ml rmIL-6, 5 ng/ml rhTGFβ1, 10 μg/ml anti-IL-4, and 10 μg/ml anti-IFNγ (clone XMG1.2); Th17 (IL-1β)—10 ng/ml rmIL-23, 20 ng/ml rmIL-6, 5 ng/ml rmIL-1β, 10 μg/ml anti-IL-4, and 10 μg/ml anti-IFNγ. For intracellular staining, cells were stimulated with R10 only (negative control) or platebound anti-CD3 (10 μg/ml)/soluble anti-CD28 (1 μg/ml) for 7 h with GolgiPlug (BD Biosciences) added for the final 4 h. Neutralizing antibodies were obtained from UAB hybridoma facility.
GM-CSF ELISA
Single-cell suspensions from draining inguinal lymph nodes were prepared and stimulated with either R10 only or 5 μg MOG35−55 peptide for 16 h. Supernatants were then collected and assessed for GM-CSF production by ELISA (eBioscience).
RNA purification, cDNA synthesis, and real-time PCR
Positively selected CD4 T cells were isolated following stimulation. RNA collection, cDNA synthesis, and real-time PCR analysis were performed as described previously [
31]. Primers used for indicated genes are as follows:
Csf2 forward: 5′-TGGAAGCATGTAGAGGCCATCA-3′; and
Csf2 reverse: 5′-GCGCCCTTGAGTTTGGTGAAAT-3′.
Chromatin-immunoprecipitation PCR
ChIP assays were adapted from previously described methods [
32]. Single-cell suspensions from pooled spleen and dLN were prepared and reactivated with either R10 or 5 μM MOG
35−55 peptide for 5 h. CD4 T cells were purified, fixed, lysed with T cell lysis buffer (20 mM HEPES, pH 7.4), 150 mM NaCl, 1.5 mM MgCl
2, 2 mM EGTA, 1 % Triton X-100, 12.5 mM β-glycerophosphate, 10 mM NaF, 1 mM Na
3VO
4), and then sonicated. Equal amounts of lysate were pre-cleared with BSA and SS-DNA-blocked protein A beads. Afterwards, 1/10th volume was removed and saved as “Input.” The remainder was immunoprecipitated with 4 μg of either STAT4 (Cell Signaling, clone C46B10) or Ser-2-Pol II CTD (Covance, clone H5) antibodies, and the immune complexes were absorbed with BSA and SS-DNA-blocked protein A beads (Upstate Cell Signaling Solutions, Charlottesville, VA). Immunoprecipitated DNA was analyzed by qRT-PCR using Sybr Green reagents. Primers used for indicated promoter regions are as follows:
Csf2 forward: 5′-GGTCTCCTCAGTGGGAGTCTGT-3′;
Csf2 reverse: 5′-GGGGTTTGGGAGATACTGAGTG-3′;
Ifng forward: 5′-TTTCTGGGCACGTTGACCCT-3′; and
Ifng reverse: 5′-ACAGCACAGGGAGCCTTTGT-3′. Reactions for each sample were performed in triplicate using an ABI StepOnePlus Detection System (Applied Biosystems, Foster City, CA) and a PCR protocol comprising an initial 10-min incubation at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at 60–65 °C. The raw data were analyzed using StepOnePlus software (Applied Biosystems), and ∆∆Ct values for each gene in each sample were determined.
Statistical analysis
Unpaired Student’s t test and one-way ANOVA were utilized as indicated and generated by GraphPad Prism 6 (version 6.0e).
Discussion
STAT4 is a transcription factor necessary for the differentiation of the Th1 lineage of effector CD4 T cells [
7,
8]. However, while STAT4 is critical for EAE, neither the upstream STAT4 activating cytokine IL-12 nor the downstream Th1 effector cytokine IFNγ are needed for disease induction [
9‐
15]. This implicates a role for STAT4 outside of the traditional Th1-associated IL-12 signaling pathway. We find that, when activated in vitro under non-polarizing conditions, the absence of STAT4 results in the decreased ability of CD4 T cells to produce GM-CSF, whereas blocking IL-12 has minimal effect on GM-CSF production by CD4 T cells. This is consistent with published data, as well as the data presented herein, showing CD4 T cell production of GM-CSF during EAE is independent of IL-12 signaling but dependent on STAT4 expression. Our findings do not negate the ability of IL-12 to induce GM-CSF in a STAT4-dependent manner, but do indicate that additional molecules signal via STAT4 to promote GM-CSF expression. Taken together, these data demonstrate STAT4 can function to regulate GM-CSF separately from the classic Th1 pathway, particularly during EAE. In keeping with these data, our study shows a marked decrease in Th1-like IFNγ+GM-CSF+, Th17-like IL-17A+GM-CSF+, and GM-CSF+ single-producing CD4 T cells, indicating that STAT4 may be a central regulator of GM-CSF expression in a lineage-indiscriminant manner.
CD4 T cells co-producing IFNγ and IL-17A are present during EAE and are postulated to be highly pathogenic [
18,
41]. Our interrogation of GM-CSF expression by effector CD4 T cells revealed that the majority of the IFNγ+IL-17A+ double-producing cells in WT mice during EAE concurrently express GM-CSF, and this may explain the pathogenic propensity of this cell population. The IFNγ+IL-17A+ double-producing cells have been shown to arise from Th17 CD4 T cells [
41], inferring that the triple cytokine-producing cells are also derived from plastic Th17 cells. The role of STAT4 in Th17 plasticity is debatable; the emergence of IFNγ+IL-17A+ cells from Th17 cells after repeated IL-23 stimulation in vitro has been shown to be both dependent on, as well as independent of, STAT4 [
18,
19]. Interestingly, we show that STAT4
−/− CD4 T cells lack the IFNγ+IL-17A+GM-CSF+ triple cytokine-producing cell population; thus, the requirement of STAT4 to induce CNS inflammation may be linked to Th17 plasticity and the development of this particular effector cell subset.
The production of IL-17A by Th17 cells has been shown to be independent of STAT4 and, in fact, we demonstrate that during EAE, the frequency of IL-17A+ CD4 T cells is actually increased in the absence of STAT4. One potential explanation for this observation is the role of STAT4 in modulating expression of the Th1 master transcription factor, Tbet. In the absence of STAT4, Tbet is induced, but not to the same levels as in WT CD4 T cells (McWilliams, data not shown). Tbet can repress IL-17A production [
42,
43]; hence, the effect of STAT4 may be a consequence of lower Tbet levels. Similarly, the Th1 cytokine IFNγ is able to suppress IL-17A secretion by CD4 T cells [
31,
33] and STAT4 is necessary for optimal expression of IFNγ [
44]; therefore, the increase in IL-17A+ CD4 T cells may be the result of diminished autocrine IFNγ signaling. Another possible reason for the augmented IL-17A frequencies in the absence of STAT4 is the inability of these cells to undergo Th17 plasticity [
19]. Th17 cells that are unable to convert into IL-17A+IFNγ+, and potentially then IFNγ+ CD4 T cells, may yield higher percentages of these cells during EAE.
Both GM-CSF single and triple cytokine-producing CD4 T cells were recently described in MS patients [
40], but the ontogeny of these effector cell populations and how these cells contribute to MS and EAE pathology remain unclear. In this study, we show that there is a marked reduction not only in IFNγ+IL-17A+GM-CSF+ triple-producing CD4 T cells in the absence of STAT4 during EAE, but in all GM-CSF producing effector CD4 T cells. This includes GM-CSF single-producing cells as well as IL-17A+GM-CSF+ double-producing cells, which would presumably be Th17 cells. Therefore, STAT4 regulates GM-CSF production in various subsets of effector CD4 T cells, and the decreased production of GM-CSF by STAT4
−/− CD4 T cells is not solely the result of impaired Th17 plasticity. It is unclear if the same molecule mediates the development of these different GM-CSF-producing CD4 T cell populations. One cytokine shown to regulate GM-CSF production in mice is IL-23 [
22,
23,
45,
46], and it is interesting to speculate that the impaired GM-CSF production by STAT4
−/− CD4 T cells is linked to defective IL-23 signaling. Importantly, data from our lab demonstrates that IL-23-induced STAT3 activation occurs independent of STAT4 expression during EAE (McWilliams, manuscript submitted), indicating that the predominant IL-23 signaling pathway is intact in the absence of STAT4. Together, these data suggest that the function of STAT4 in the formation of GM-CSF single, double, and triple cytokine-producing CD4 T cells may be separate from the role of IL-23 in the expression of this cytokine during EAE and that a potentially novel STAT4 ligand drives CNS inflammation.
We show that during EAE, STAT4 directly interacts with the
Csf2 gene locus to regulate optimal GM-CSF expression. It is well documented that STAT4 controls the accessibility of multiple Th1-associated genes; however, these studies did not highlight GM-CSF as a STAT4 regulated target [
7,
16,
17]. The discrepancies in these data may reflect differences in the CD4 T cell populations examined or the manner in which the cells were activated; previous reports have studied IL-12-induced STAT4 gene regulation in differentiating Th1 cells, whereas we examined the role of STAT4 in GM-CSF production by bulk effector CD4 T cells during EAE, which is independent of IL-12. These data imply there exists a set of unidentified STAT4-dependent genes present in vivo during autoimmunity and possibly other inflammatory conditions. Deciphering what these genes are and which molecules operate via STAT4 to promote gene expression during disease will be important for dissecting the underlying causes of autoimmune inflammation.
Our study identifies STAT4 as a potent regulator of GM-CSF production by effector CD4 T cells of various lineages. Nevertheless, while GM-CSF levels are significantly decreased in STAT4
−/− CD4 T cells, this cytokine is still detectable suggesting that other transcription factors must mediate expression. Recent publications have identified both IL-2 and IL-7, signaling via STAT5, as additional regulators of T cell-derived GM-CSF [
40,
47]. This raises an interesting question as to the respective contributions of STAT4 and STAT5 to GM-CSF expression and if these molecules act cooperatively to promote optimal GM-CSF production. One proposed function of the STAT transcription factors is to modulate the accessibility of lineage-specific genes [
7,
16,
17]. Additionally, STAT4 alters the enhancer landscape around the
Csf2 locus [
17]. Therefore, a plausible explanation as to how STAT4 and STAT5 may both regulate GM-CSF is that STAT4 is controlling the accessibility or optimal transcription conditions of the
Csf2 locus for STAT5 involvement. Conversely, STAT5 may be necessary to open the
Csf2 locus in order for STAT4 to then interact, as we have shown that STAT4 is able to directly bind to the
Csf2 promoter after MOG stimulation. Future studies will be necessary to address this as well as to determine if STAT4 expression is critical for the ability of other transcription factors to bind and promote GM-CSF production in effector CD4 T cells.
Acknowledgements
We wish to thank the other members of the Harrington laboratory as well as the Zajac laboratory for the helpful discussions and critical reading of this manuscript. We also wish to thank Dr. Mark Kaplan for providing the B6.STAT4−/− mice, Marion Spell of the UAB Center for AIDS Research Flow Cytometry Core for the cell sorting, and the UAB Hybridoma Core Facility for the generation of neutralizing antibodies.
This study was supported by the National Institutes of Health Grants R01 DK084082, AI113007 (to L.E.H.), and T32 AI07051 (to I.L.M.) and funding from the National Multiple Sclerosis Society RG-5116-A-3 (to L.E.H.). This study was also supported in part by R01 NS57563 (to E.N.B.) and a Collaborative MS Research Center Award from the National Multiple Sclerosis Society CA 1059-A-13 (to E.N.B.).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ILM designed, executed, and analyzed the experiments and drafted the manuscript. LEH conceived the study and participated in the experimental design and manuscript preparation. RR, SN, and EB participated in the design and running of the STAT4 and Pol II ChIP protocols and subsequent analysis of data. All authors read and approved the final manuscript.