Intrinsic STAT4 Expression Controls Effector CD4 T Cell Migration and Th17 Pathogenicity

CD4 T cells are implicated in the etiology of a number of autoimmune diseases, including multiple sclerosis (MS), rheumatoid arthritis, and inflammatory bowel disease, yet the underlying mechanisms that potentiate T cell–mediated chronic inflammatory diseases remain ill-defined (1–7). Single-nucleotide polymorphisms in the human STAT4 locus are associated with susceptibility to multiple autoimmune disorders, including MS (8), and STAT4 is required for the induction of T cell–mediated autoimmunity in animal models, including the mouse model of MS, experimental autoimmune encephalomyelitis (EAE) (9–11). IL-12 signaling in CD4 T cells via STAT4 promotes IFN-γ production and Th1 differentiation (12–14); however, for the induction of EAE, neither IL-12 nor IFN-γ is obligatory (15–19), indicating that STAT4 functions outside the Th1 lineage to elicit autoimmune inflammation. Therefore, understanding which cell types use STAT4 to mediate autoimmunity and how STAT4 directs pathogenicity in these cells is vital for devising new therapeutic interventions for MS and other autoimmune disorders.

Th17 effector CD4 T cells differentiate in response to TGF-β1 and IL-6 signaling, are dependent on the transcription factors RORγt and STAT3, and produce IL-17A, IL-17F, and other inflammatory cytokines (20–27). Studies using the EAE model have established that Th17 cells are critical for the induction of CNS inflammatory disease. Importantly, in the absence of IL-23 signaling, Th17 cells are deemed nonpathogenic and do not induce autoimmune inflammation, whereas exposure of Th17 cells to IL-23 promotes the acquisition of a pathogenic phenotype as exhibited by the ability to induce EAE in mice (28–32). The gain of pathogenic properties is associated with a unique transcriptional signature; however, the exact mechanisms by which Th17 cells function to elicit autoimmune chronic inflammation are not fully resolved. Elucidating additional mechanisms that distinguish between pathogenic and nonpathogenic Th17 cells is vital to create therapeutic options for patients with autoimmune diseases.

Recent studies assessing pathogenic Th17 CD4 T cells during MS/EAE highlight an unexpected association with the transcription factor STAT4, which is a prototypic Th1 transcription factor (12, 13, 30, 33, 34). Notably, although IL-17A–producing Th17 CD4 T cells develop independently of STAT4, the contribution of STAT4 to Th17 pathogenicity has not been established (21, 26). Therefore, we interrogated the importance of STAT4 to CD4 T cell–mediated encephalogenicity, with a focus on the Th17 lineage. We find that STAT4 regulates multiple aspects of the CD4 T cell response during autoimmune inflammation, including the striking observation that STAT4 controls the migration of effector CD4 T cells to the inflamed CNS. Importantly, we show that IL-23–mediated CNS inflammation is dependent on CD4 T cell expression of STAT4, and that Th17 cells require STAT4 to induce EAE. Transcriptional profiling experiments revealed that STAT4 regulates the expression of numerous genes in Th17 cells that are associated with pathogenicity; however, we find that Il17a mRNA and IL-17A protein are not controlled by STAT4, highlighting the function of STAT4 in Th17 cells is independent of IL-17A. These data highlight multiple functions for STAT4 in mediating CD4 T cell autoimmune inflammation.

The following mice were purchased from Jackson Laboratory: C57BL/6J, B6.SJL-Ptprca Pep3^b/BoyJ (wild-type [WT] CD45.1), C57BL/6-Il17a^Tm1Bcgen/J (WT IL-17A-IRES-GFP-KI), Il17a^{tm1.1(icre)Stck}/J (Il17a^cre), B6.129X1-Gt(ROSA)26Sor^tm1(EYFP)Cos/J (ROSA26-YFP^fl/fl), and B6.Cg-Tg(Lck-icre)3779Nik/J (distal Lck [dLCK]). B6.STAT4^fl/fl mice were created by the Heflin Center for Genomic Sciences at The University of Alabama at Birmingham (see later). B6.STAT4^−/− (STAT4^−/−) mice were generously provided by Dr. M. Kaplan (13). B6.IFN-γ^−/− mice were generously provided by Dr. F. Lund at the University of Alabama at Birmingham. Mice were maintained and procedures approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham.

STAT4 Flox CRISPR/sgRNA design and synthesis were as follows: CRISPR guides were designed using CRISPOR (http://crispor.tefor.net/) to introns 5 and 6 of the Stat4 locus (5′ G1, 5′-CTAATCACAACTCCACTTAC-3′; 3′ G1, 5′-TAGGGGTGCACATGTTGCAG-3′). Modified synthetic single-guide RNAs (Synthego) were allowed to complex with Alt-R S.p. Cas9 Nuclease V3 (IDT) at room temperature for 15 min before addition of the repair template (IDT) and dilution with microinjection buffer to a final concentration of 50 ng/μl each of guide 1, guide 2, Cas9, and repair template, respectively. Gonadotropins were female C57B6J embryo donors from 3 to 4 wk of age administered 5 IU of pregnant mare serum gonadotropin (Sigma, St. Louis, MO) on day −3 followed by 5 IU of human chorionic gonadotropin (Sigma) on day −1 to induce superovulation. Donor and recipient females were mated to stud and vasectomized males, respectively, on day −1. The collection of embryos was as follows: at day 0.5 postconception, superovulated donor females with copulatory plugs were humanely sacrificed using CO₂ followed by cervical dislocation. Oviducts were dissected with sterile medium and nicked to expose the cumulus masses containing fertilized embryos. Embryos were cultured in potassium-supplemented simplex optimised medium (Millipore, Darmstadt, Germany) under 5% blood gas before electroporation. Founder animals were identified by PCR using primers flanking the target loci that amplified an 851-bp fragment in WT animals (F1: 5′-ACTAGTACAGAGGGCAGCAGA-3′; R1: 5′-GCAATTACATGCACGTGCCA-3′). PCR samples were run on 6% polyacrylamide/Tris-borate-EDTA gels at 100 V for 45 min before staining with ethidium bromide. Positive samples were confirmed by modified Sanger sequencing.

EAE was induced and scored as previously described: classical EAE disease was scored as published previously (10), and atypical EAE disease was scored as published previously (35). Mice that reached a score of 4 were humanely euthanized. Scores were discontinued in the graph.

Littermate WT controls for Fig. 1 were either STAT4^WT/Flox mice or mice negative for the dLCKcre transgene. Similarly, WT littermate controls for Fig. 4 were either STAT4^WT/Flox or mice negative for the Il17a-cre transgene.

Ten days after EAE induction, spleen cells were cultured for 3 d under Th17-polarizing conditions as previously described (36). After culture, live CD4 T cells were isolated by centrifugation over Histopaque 1083 (Sigma) followed by Dynabeads FlowComp Mouse CD4 kit (Invitrogen). For bulk CD4 transfer, 2 × 10⁶ cells were injected i.p. into B6 (C57BL/6 or B6.SJL-Ptprca Pep3^b/BoyJ) mice. For IL-17aGFP⁺ cell transfer, after CD4 purification, GFP⁺ CD4 T cells were sorted using Aria II, and 5 × 10⁵ cells were injected i.p. into B6 (C57BL/6 or B6.SJL-Ptprca Pep3^b/BoyJ) recipient mice.

Mixed bone marrow chimeric mice were generated as previously described (10, 18). B6.Rag1^−/− mice were lethally irradiated and reconstituted with a 1:1 ratio of CD45.1 WT and CD45.2 WT or STAT4^−/− bone marrow (WT:WT and WT:STAT4^−/−, respectively). Ten weeks after reconstitution, mice were immunized for EAE.

Cell surface and intracellular staining was performed as described previously (10, 18). A viability dye (Aqua; Life Technologies) was applied to exclude dead cells. Intracellular staining was performed after restimulation with 50 ng/ml PMA (Sigma) and 750 ng/ml ionomycin (Calbiochem) and BFA for 4 h using either the eBioscience Fixation/Permeabilization Diluent (eBioscience) or the Cytofix/Cytoperm plus Fixation/Permeabilization kit (BD Biosciences) according to the manufacturer’s instructions.

Single-cell suspensions of pooled brain and spinal cord lymphocytes were incubated for 30 min in the presence of media only or 10 ng/ml recombinant mouse IL-23. Phosflow staining was performed using anti–p-STAT3 (Cell Signaling) and the BD Phosflow kit (BD Biosciences), according to the manufacturer’s instructions.

cDNA synthesis and real-time PCR were described previously (37). Relative gene expression was calculated according to the ΔΔ threshold cycle method by using β₂-microglobulin as a housekeeping gene.

Ten days after EAE induction, splenocytes were cultured for 3 d under Th17-polarizing conditions as previously described (10, 38). After culture, live cells were isolated by centrifugation over Histopaque 1083 (Sigma), and CD4 T cells were isolated using Dynabeads FlowComp Mouse CD4 kit (Invitrogen).

RNA isolation, sequencing, and alignment were completed as described previously (37). RNA was isolated from WT and STAT4^−/− CD4 cells using the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer’s protocols and submitted to GENEWIZ for sequencing. Total RNA was sequenced on illumine HiSeq 2500 (2 × 100 bp, single-end reads) and aligned using mouse mm10 reference genome. The GI tools software along with DESeq2 software was used for differential expression analysis. RNA sequencing data have been deposited in the Gene Expression Omnibus Repository (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE227049).

For gene set enrichment analysis (GSEA), RNA sequencing data were preranked according to an adjusted p value and the sign of differential expression. The normalized enrichment score, nominal p value, and false detection rate q value were assessed using GSEA software from Broad Institute by running in preranked list mode with 1000 permutations. To perform GSEA with pathogenic Th17 cell gene signature (TGF-β + IL-6 ± IL-23), we used the microarray data (GSE43955). GraphPad Prism 9 was used to generate a volcano plot.

In vitro Th1 and Th17 differentiation was completed as described previously (39). Naive CD4 T cells were isolated from spleen and lymph nodes with the EasySep Mouse Naive CD4+ T Cell Isolation Kit (StemCell Technologies) according to the manufacturer’s instructions. For Th1 conditions, cultures were supplemented with IL-12 (10 ng/ml; PeproTech) and anti–IL-4 (10 mg/ml; BioXCell). For Th17 conditions, cultures were supplemented with rmIL-6 (20 ng/ml; BioLegend), rhTGFb1 (2.5 ng/ml; PeproTech), anti–IL-4 (10 mg/ml; 11B11; BioXCell), and anti–IFN-γ (10 mg/ml; XMG1.2; BioXCell).

Supernatants were removed from undisturbed Th1 cell cultures on day 3. IFN-γ ELISA (eBioscience) was completed according to the manufacturer’s protocol.

STAT4 regulates effector CD4 T cell differentiation and function and is required for the induction of CNS demyelination in mice (9–11); nevertheless, the role of T cell–intrinsic STAT4 in the disease pathogenesis remains elusive. To determine whether T cell–intrinsic STAT4 is necessary for EAE, we established T cell–specific STAT4-deficient mice by breeding newly generated mice that harbor a floxed Stat4 allele with mice carrying the dLck cre recombinase transgene (STAT4^ΔdLck), which is known to delete in mature T cells (40). We verified the phenotype of newly developed STAT4^ΔdLck mice in vitro by comparing the expression and function of CD4 T cells derived from these mice with WT and global STAT4-deficient (STAT4^−/−) mice. First, naive CD4 T cells from the different strains of mice were analyzed directly ex vivo for Stat4 mRNA levels by real-time PCR, and we did not detect a significant difference in the Stat4 mRNA levels between the STAT4^−/− and STAT4^ΔdLck CD4 T cells (Fig. 1A). Because STAT4 is activated and necessary during Th1 differentiation (12–14), we stimulated naive CD4 T cells from the WT, STAT4^−/−, and STAT4^ΔdLck mice under Th1 conditions for 3 d. As expected, the WT Th1 CD4 T cells exhibited robust p-STAT4 staining, whereas the STAT4^−/− and STAT4^ΔdLck CD4 T cells displayed negligible p-STAT4 staining (Fig. 1B, 1C). In addition, the production of the STAT4-dependent cytokine IFN-γ by CD4 T cells was markedly reduced in the STAT4^ΔdLck CD4 T cells compared with the WT cells (Fig. 1D–G). It is worth noting that the STAT4^ΔdLck CD4 T cells did harbor a low level of p-STAT4 and IFN-γ production after Th1 differentiation, likely because of incomplete cre deletion by the dLck transgene.

The development of EAE is associated with Th17 cells, and previous studies have shown that STAT4 is dispensable for the development of Th17 cells (10, 21, 26). Therefore, it was necessary to test the ability of STAT4^ΔdLck CD4 T cells to differentiate into Th17 cells. We isolated naive WT, STAT4^−/−, and STAT4^ΔdLck CD4 T cells and cultured these cells for 3 d under Th17 conditions. Consistent with published reports (10, 21, 26), there was no difference in the production of IL-17A between WT and STAT4^−/− CD4 T cells after Th17 differentiation, and the levels of IL-17A production by the STAT4^ΔdLck CD4 T cells were comparable (Fig. 1H, 1I). Together, these experiments demonstrate that the CD4 T cells in the STAT4^ΔdLck mice lack functional STAT4.

Confident that the STAT4^ΔdLck mice lack functional STAT4 in mature CD4 T cells, we set out to test the requirement of T cell–intrinsic STAT4 for the induction of EAE. To do this, we immunized WT and STAT4^ΔdLck mice with MOG_35–55 peptide to induce EAE. As expected, the WT mice exhibited EAE symptoms; however, the STAT4^ΔdLck mice did not show signs of disease (Fig. 2A). Although these data do not rule out a role for STAT4 in non-T cells during EAE, they demonstrate T cell expression of STAT4 is indispensable to mediate CNS demyelination.

To assess the impact of intrinsic STAT4 deficiency on effector CD4 T cells during EAE, we analyzed the functional capacity of the brain- and spinal cord–infiltrating cells. Strikingly, in the absence of STAT4, the frequency and number of CD4 T cells in the spinal cord were significantly decreased (Fig. 2B, 2C). Functional cytokine analysis showed that the percentage of CD4 T cells in the brain and spinal cord producing IFN-γ only or both IFN-γ and IL-17A was reliant on STAT4, whereas the requirement for IL-17A production was less stringent (Fig. 2D–G), and this is consistent with previous reports using global STAT4-deficient mice (9, 10). These data, combined with the diminished CD4 T cell recovery from the CNS, translated into significant reductions in the absolute numbers of IFN-γ⁺, IL-17A⁺, and IFN-γ⁺IL-17A⁺ CD4 T cells in the spinal cords of the STAT4^ΔdLck mice, and similar trends were also observed in the cells recovered from the brains (Fig. 2D–G). Taken together, these experiments reveal that STAT4 not only instructs the functional properties of CD4 T cells but potentially impacts the ability of these cells to migrate to the CNS.

Because differences in the inflammatory environment can influence the access of CD4 T cells to the CNS during EAE, we used mixed (50:50) bone marrow chimeric mice to alternatively test the T cell–intrinsic role of STAT4 in T cell trafficking to the brain and spinal cord in the context of disease. Notably, we have previously published that WT:STAT4^−/− mixed bone marrow chimeric mice develop EAE to the same extent as WT:WT chimeric mice (10), providing a system to study the role of CD4 T cell–intrinsic STAT4 expression during active CNS inflammation. In the control WT:WT group, at the peak of EAE disease severity (day 17), the frequency of CD45.2⁺ CD4 T cells recovered from the brain and spinal cord was similar to the frequency detected in the spleen (Fig. 3A, 3B). However, in the experimental cohort, there was a significant reduction in the frequency of CD45.2⁺STAT4^−/− CD4 T cells in both the brain and spinal cord compared with the spleen at this time point. To test whether the reduced accumulation of CD4 T cells in the CNS was due to a defect in cellular migration, we analyzed CD4 T cells in the brain and spinal cord at disease onset (day 11), a time point reflective of early T cell recruitment. There was a significant reduction in the migration efficiency of STAT4^−/− CD4 T cells in the brain and spinal cord compared with WT CD4 T cells when normalized to the spleen at the initiation of disease (Fig. 3B), which is consistent with a role for STAT4 in entry of CD4 T cells into the inflamed CNS. Together, these findings highlight a previously unrecognized function of STAT4 in directing the migration of effector CD4 T cells to the sites of inflammation during EAE.

Interestingly, CD4 T cells that lack the ability to signal via the IL-23R complex (IL-23R^−/− or IL-12Rβ1^−/− CD4 T cells) also exhibit a defect in CNS accumulation (28, 41, 42). However, IL-23 is published to support CD4 T cell proliferation in the CNS, not migration to the site; the defect in CNS accumulation of IL-23R^−/− CD4 T cells was not present at the onset of disease but at the peak (42), which is distinct from the phenotype observed with STAT4^−/− CD4 T cells. Nevertheless, we further probed the potential link between the IL-23 pathway and STAT4 in CD4 T cells during EAE using mixed bone marrow chimeric mice. Unlike IL-23R^−/− CD4 T cells, CD4 T cells devoid of STAT4 expression did not exhibit reduced proliferation at the peak of disease as evidenced by Ki-67 staining (Fig. 3C, 3D). Moreover, IL-23 induction of STAT3 phosphorylation was unaffected by the lack of STAT4 in CD4 T cells (Fig. 3E, 3F). Together, these data indicate that notable downstream actions of IL-23 signaling are intact in the absence of STAT4, and that STAT4 modulates the accumulation of CD4 T cells in the inflamed CNS via a mechanism distinct from proliferation, potentially by promoting cell migration.

IL-23 is a key pathogenic cytokine during EAE and is critical for the induction of pathogenic Th17 cells (28, 30, 33, 42). Upon IL-23 signaling, activated STAT3 dimerizes and translocates to the nucleus, where it initiates transcription, and we demonstrate earlier that this signaling conduit is intact in the absence of STAT4. However, it has been illustrated that IL-23 can also induce the phosphorylation of STAT4 (34, 43), potentially linking the IL-23 and STAT4 pathways. Hence we hypothesized that STAT4 is critical for IL-23–mediated autoimmunity. Using a passive transfer model of EAE to test this, we find that transfer of IL-23–stimulated WT CD4 T cells elicited EAE, whereas STAT4^−/− CD4 T cells cultured in the presence of IL-23 did not induce disease (Fig. 4A). This is associated with a significant decrease in the number of donor STAT4^−/− cells recovered from both the brain and spinal cord of recipient mice (Fig. 4B–D). Importantly, the donor WT and STAT4^−/− CD4 T cells did not differ in proliferation during the in vitro culture (Supplemental Fig. 1A), and there were equivalent numbers of both donor WT and STAT4^−/− cells recovered in the spleen (Fig. 4C, 4D), indicating the migratory phenotype observed in the CNS is irrespective of cell proliferation and survival. Before transfer, the production of IL-17A or IFN-γ was similar between the two genotypes (Supplemental Fig. 1B, 1C); however, after transfer, significant differences in the number of recovered donor WT and STAT4^−/− cells producing the proinflammatory cytokines IL-17A and IFN-γ were detected (Fig. 4E–H). In the recipient mice, the WT CD4 T cells predominantly produced IFN-γ and minimal IL-17A, whereas STAT4^−/− CD4 T cells produced high amounts of IL-17A and less IFN-γ. Because production of IFN-γ was one of the key differences in the donor cells, we performed the same experiment with IL-23–stimulated IFN-γ^−/− CD4 T cells. These cells did mediate EAE, indicating that IFN-γ is dispensable for disease (Supplemental Fig. 1D–F) and revealing that STAT4 functions independently of IFN-γ to promote neuroinflammation.

To decipher the mechanistic role of STAT4 on Th17 cell pathogenicity, we performed RNA sequencing on CD4 T cells extracted from EAE-immunized WT and STAT4^−/− mice that were subsequently cultured under Th17 conditions. We found that STAT4 significantly influenced the expression of >200 genes, and importantly, expressions of the Th17 lineage genes Rorc, Stat3, Il22, and Il17a were not negatively impacted by STAT4 deficiency (Fig. 5A). We scrutinized this RNA sequencing dataset for the expression of genes previously published to be associated with pathogenic and nonpathogenic Th17 cells by GSEA. As expected, we show that WT EAE-Th17 cells were positively associated with genes upregulated in pathogenic, IL-23–stimulated Th17 cells (Fig. 5B) (32). In striking contrast, we find that the STAT4^−/− EAE-Th17 cells were enriched for genes downregulated under these conditions (Fig. 5B). These data, together with the in vivo EAE experiments, indicate that STAT4 controls IL-23–mediated inflammation in part by regulating the balance between pathogenic and nonpathogenic Th17 cells.

IL-23 imprints pathogenic properties to Th17 cells, yet the earlier adoptive transfer experiments involved a heterogeneous population of IL-17A⁺ and IL-17A⁻ CD4 T cells. To test whether STAT4 is requisite in IL-17A⁺ CD4 T cells to elicit neuroinflammation, we used WT and STAT4^−/− IL-17A-GFP knockin reporter mice in conjunction with the IL-23–mediated passive transfer model of EAE. We performed the same IL-23–mediated passive transfer EAE model as earlier, but instead used FACS-sorted GFP⁺ (IL-17A⁺) CD4 T cells as the donor population. After transfer, the WT GFP⁺ (IL-17A⁺) CD4 T cells rapidly induced EAE symptoms in all mice (100%; 11/11), promoting classical EAE with ascending paralysis in the majority of the recipient mice (82%; 9/11), and evoking ataxia or atypical EAE in almost half of the mice (45%; 5/11) (Fig. 5C). In contrast, transfer of an equivalent number of STAT4^−/− GFP⁺ (IL-17A⁺) CD4 T cells did not result in either classical or atypical EAE in the recipient mice (0/10 mice) (Fig. 5C). The disparities in disease severity correlate with differences in the donor CD4 T cell populations after transfer. Consistent with our earlier data, lower numbers of donor STAT4^−/− CD4 T cells were recovered from the CNS sites compared with the WT donor cells (Fig. 5D), again reaffirming a role for STAT4 in Th17 migration. Even though all donor cells were IL-17A⁺ pretransfer, at the peak of disease the WT CD4 T cells had largely extinguished IL-17A production and exhibited robust manufacture of IFN-γ (Fig. 5E–H). This functional profile was not observed in the donor STAT4^−/− cells after transfer, which did downregulate IL-17A expression but did not increase IFN-γ production to the extent the WT CD4 T cells did (Fig. 5E–H). These data further support the observation that STAT4 influences Th17 and CD4 T cell migration, as well as the proinflammatory activities, and demonstrate that STAT4 is essential for IL-23–mediated Th17 disease.

In the earlier experiments, STAT4 is deleted in the CD4 T cells before TCR stimulation, Th17 differentiation, and the initiation of EAE disease. Therefore, we sought to determine whether STAT4 functions early in CD4 T cells to direct pathogenic potential or after commitment to the Th17 lineage. To test this, we created Th17-specific STAT4-deficient mice by crossing the STAT4^fl/fl mice and the IL-17A-Cre × ROSA26-YFP^fl/fl fate-mapping mice (STAT4^ΔIL-17A), and we immunized control and STAT4^ΔIL-17A mice to induce EAE. As expected, the control group of animals developed EAE symptoms; however, the STAT4^ΔIL-17A mice displayed significantly lower disease scores (Fig. 6A). This was associated with dramatically decreased numbers of CD4 T cells in the brain and spinal cord of the STAT4^ΔIL-17A mice compared with the control mice (Fig. 6B) at the peak of disease. Although these data do not rule out a role for STAT4 early in the programming of pathogenic CD4 T cells, it does show that STAT4 does function after the commitment to the Th17 lineage to promote disease.

Our earlier data show that STAT4 is critical for Th1 differentiation and IFN-γ production by CD4 T cells but does not impact IL-17A production by Th17 cells, hence it was necessary to assess how STAT4 deletion in Th17 cells specifically influences functionality. The availability of the ROSA26-YFP fate-mapping allele provided us with the opportunity to track IL-17A⁺ CD4 T cells during disease; in both the brains and spinal cords of STAT4^ΔIL-17A mice, the frequency and number of YFP⁺ cells were significantly decreased (Fig. 6C–E, Supplemental Fig. 2A–C). Furthermore, intracellular staining for IFN-γ and IL-17A in the CNS-infiltrating CD4 T cells showed marked reductions in the numbers of IFN-γ⁺, IL-17A⁺, and IFN-γ⁺IL-17A⁺ CD4 T cells in the STAT4^ΔIL-17A mice (Fig. 6F–I). Taken together, these experiments complement the results obtained with the global and T cell–specific STAT4-deficient mice and further corroborate a definitive role for STAT4 in Th17 pathogenesis.

In this article, we demonstrate that STAT4 functions in IL-23–mediated autoimmunity by influencing migration and the pathogenic potential of Th17 cells. Using a strain of T cell–specific STAT4-deficient mice, we show that T cell–intrinsic expression of STAT4 is required to elicit autoimmune CNS inflammation, and this is in part by regulating the migration of CD4 T cells to the CNS. In addition, we employed multiple animal models, including STAT4-deficient IL-17A reporter mice and Th17-specific STAT4-deficient mice, to illustrate an indispensable role for STAT4 in Th17-mediated disease. Importantly, STAT4 does not control the differentiation of Th17 cells or the production of IL-17A, but instead governs the expression of genes associated with pathogenicity.

One of the major findings from this study is that STAT4 impacts the migration of effector CD4 T cells, including Th17 cells, to the inflamed CNS during EAE. The importance of T cell trafficking during MS and EAE is exemplified by the successful clinical use of natalizumab, a mAb specific for the integrin α4, and Fingolimod, a sphingosine 1-phosphate receptor agonist, to treat MS patients (44–47). Interestingly, using Th1- versus Th17-mediated adoptive transfer EAE models, Rothhammer et al. (48) demonstrated that integrin α4 is necessary for Th1 cells to infiltrate the CNS and mediate disease, but that Th17 cells use CD11a (Itgal), not integrin α4, to induce EAE. We previously published that LFA-1 expression by CD4 T cells is not regulated by STAT4 after acute viral infection (49), suggesting that STAT4 controls T cell migration via an alternative mechanism. Chemokine:chemokine receptor interactions are also known to be important for the recruitment of T cells to inflammatory sites. The chemokine receptor CXCR3 is preferentially expressed by Th1 cells, and notably, Ghoreschi et al. (33) published that Cxcr3 mRNA levels are elevated in IL-23–stimulated, pathogenic Th17 cells that induce EAE. Nevertheless, the CXCR3-deficient mice do not phenocopy the STAT4^−/− mice in regards to the EAE disease course (50, 51), thus indicating that this is not the likely pathway by which STAT4 regulates CD4 T cell–dependent CNS demyelination. Overall, we find that STAT4 influences the expression of numerous molecules in Th17 cells, hence we favor the hypothesis that STAT4 controls CD4 T cell encephalitogenicity in part by regulating a constellation of molecules that function cooperatively to facilitate entrance to the CNS during inflammation and demyelinating disease.

STAT4 is an integral component of the Th1 differentiation pathway; IL-12–induced STAT4 phosphorylation functions to promote the epigenetic remodeling of Th1-associated genes and acts to directly activate Ifng gene transcription (13, 52). In the EAE model of CNS inflammation, STAT4 is required for disease induction, whereas both IL-12 and IFN-γ signaling are dispensable (9, 15, 28, 53), highlighting a role for STAT4 independent of the classic Th1 lineage. Although the role of Tbet, the Th1 master transcription factor, in the development of EAE is debatable, our laboratory does not find an essential role for Tbet in EAE (38, 54–56). In addition, we published that Tbet is not important for the migration of CD4 T cells to the inflamed CNS during EAE, using the same mixed bone marrow chimera approach employed in this study (38). Together, these findings demonstrate that STAT4 functions to promote effector CD4 T cell encephalitogenicity, independent of its function in Th1 differentiation and IFN-γ production.

Published studies have documented an essential role for the cytokine GM-CSF in the development of EAE (57–59). Both Th1 and Th17 effector CD4 T cell lineages produce GM-CSF, and the encephalitogenicity of both cell types is dependent on GM-CSF. El-Behi et al. (58) showed that adoptive transfer of Csf2^−/− Th1 cells or Th17 cells was not able to induce EAE in recipient mice, even though production of the lineage signature cytokines IFN-γ and IL-17A, respectively, was not corrupted. Moreover, Codarri et al. (57) published that adoptive transfer of lymphocytes from Ifng^−/−Il17a^−/− mice induced EAE in WT recipient mice, and this correlated with GM-CSF production by CD4 T cells. Interestingly, IL-23 stimulation was important for the upregulation of GM-CSF in CD4 T cells (57, 58), and in this study we show that STAT4 is required for IL-23–mediated EAE. Importantly, we previously reported that GM-CSF production by effector CD4 T cells during EAE, including Th17 cells, is dependent on STAT4; STAT4-deficient CD4 T cells secreted lower levels of GM-CSF and STAT4 directly bound to the Csf2 promoter (10). Taken together, these data are consistent with the data presented in this article that STAT4-deficient Th17 cells are able to produce ample amounts of IL-17A, yet not cause demyelination in IL-23–mediated CNS inflammation.

Consistent with a Th1-independent function for STAT4, we demonstrate in this study a vital role for STAT4 in IL-23–dependent, Th17-mediated EAE. IL-23 signaling is essential for the development of EAE in mice, and members of the IL-23 pathway are associated with MS susceptibility (28, 29, 42). Early findings documenting the discovery of IL-23 indicated that this cytokine not only induced the phosphorylation of STAT3, but that it also activated STAT4 (60). Moreover, a recent study by Lee et al. (34) detailed the induction of STAT3/STAT4 heterodimers by IL-23 in CD4 T cells if the cells were previously primed in the presence of certain inflammatory cytokines or APC populations. This link between IL-23 and STAT4 is further supported by the data in this study, as well as others, which show that IL-23R–deficient and STAT4-deficient CD4 T cells both exhibit a defect in accumulation in the inflamed CNS (42). However, we observed that STAT4 was essential for the early recruitment of CD4 T cells to the CNS, and it did not impact proliferation or survival, whereas the impaired accumulation of IL-23R–deficient CD4 T cells in the inflamed CNS is the result of reduced cellular proliferation. These data beg to question whether there are differences in IL-23 signals dependent on STAT3:STAT3 homodimers versus STAT3:STAT4 heterodimers. To this end, Poholek et al. (61) published that deletion of STAT3 in Th17 cells abrogated EAE disease, and many of the IL-23–dependent cell-cycle genes in CD4 T cells were similarly affected. Strikingly, though, certain genes previously identified to be regulated by IL-23 were not affected by STAT3 deficiency, highlighting the possibility that distinct IL-23 functions are mediated by distinct signaling modalities, potentially involving STAT4.

In conclusion, in this study, we show that STAT4 functions in IL-23–mediated autoimmune disease. We demonstrate that STAT4 directs the migration of effector CD4 T cells into the inflamed CNS during EAE. We also show that Th17-intrinsic expression of STAT4 is required for the encephalitogenic capacity of the cells during CNS demyelination, in part by controlling the expression of the Th17 pathogenic gene signature. These data, together with previous work from others, as well as our group, highlight that STAT4 operates at multiple levels in effector CD4 T cells to regulate autoimmune disease (10–12, 30, 34, 43, 62). Further understanding of the mechanisms by which STAT4 coordinates pathogenic gene expression in Th17 cells and other effector CD4 T cell populations will yield therapeutic options for individuals with chronic inflammation and autoimmunity, including MS, rheumatoid arthritis, and inflammatory bowel disease.

The authors have no financial conflicts of interest.

We thank the other members of the Harrington and Zajac laboratories for the thoughtful critiques and helpful input in the preparation of this manuscript. We also thank Dr. Mark Kaplan for providing the B6.STAT4^−/− mice, Dr. David Crossman for assistance with RNA sequencing data analysis, and the UAB Heflin Genomics Core Facility for the generation of mice with a floxed Stat4 allele.