STAT3/SOCS3 - Semantic Scholar

1 downloads 73 Views 926KB Size Report
Suppressor of cytokine signaling (SOCS) proteins are feedback inhibitors of the JAK/STAT pathway. SOCS3 has a crucial role in inhibiting STAT3 activation ...
Signal transducer and activator of transcription3/suppressor of cytokine signaling-3 (STAT3/SOCS3) axis in myeloid cells regulates neuroinflammation Hongwei Qina,1, Wen-I Yeha, Patrizia De Sarnob, Andrew T. Holdbrooksa, Yudong Liua, Michelle T. Muldowneya, Stephanie L. Reynoldsa, Lora L. Yanagisawaa, Thomas H. Fox IIIa, Keun Parka, Laurie E. Harringtona, Chander Ramanc, and Etty N. Benvenistea,1 Departments of aCell Biology, bPsychiatry and Behavioral Neurobiology, and cMedicine, University of Alabama at Birmingham, Birmingham, AL 35294 Edited by George R. Stark, Lerner Research Institute, Cleveland, OH, and approved February 14, 2012 (received for review October 20, 2011)

Suppressor of cytokine signaling (SOCS) proteins are feedback inhibitors of the JAK/STAT pathway. SOCS3 has a crucial role in inhibiting STAT3 activation, cytokine signaling, and inflammatory gene expression in macrophages/microglia. To determine the role of SOCS3 in myeloid cells in neuroinflammation, mice with conditional SOCS3 deletion in myeloid cells (LysMCre-SOCS3fl/fl) were tested for experimental autoimmune encephalomyelitis (EAE). The myeloid-specific SOCS3-deficient mice are vulnerable to myelin oligodendrocyte glycoprotein (MOG)-induced EAE, with a severe, nonresolving atypical form of disease. In vivo, enhanced infiltration of inflammatory cells and demyelination is prominent in the cerebellum of myeloid-specific SOCS3-deficient mice, as is enhanced STAT3 signaling and expression of inflammatory cytokines/chemokines and an immune response dominated by Th1 and Th17 cells. In vitro, SOCS3-deficient macrophages exhibit heightened STAT3 activation and are polarized toward the classical M1 phenotype. SOCS3-deficient M1 macrophages provide the microenvironment to polarize Th1 and Th17 cells and induce neuronal death. Furthermore, adoptive transfer of M2 macrophages into myeloid SOCS3-deficient mice leads to delayed onset and reduced severity of atypical EAE by decreasing STAT3 activation, Th1/Th17 cells, and proinflammatory mediators in the cerebellum. These findings indicate that myeloid cell SOCS3 provides protection from EAE through deactivation of neuroinflammatory responses. Th1 cells

| M1 polarization

M

ultiple sclerosis (MS) is an autoimmune neurodegenerative disease in which the immune system plays a pathogenic role (1, 2). Hallmarks include demyelination, gliosis, inflammatory lesions, and neurodegeneration (2). Activated macrophages and microglia in the CNS express aberrant levels of cytokines/chemokines, leading to neuronal damage and demyelination (3). Experimental autoimmune encephalomyelitis (EAE) is an animal model for MS (1). CD4+ Th1 and Th17 cells are primary effector T cells in EAE, whereas Th2 and Treg cells promote disease resolution (1, 4). Macrophages and microglia promote both injury and repair in the CNS (5–9). These divergent functions are dictated by the cellular microenvironment, which induces polarization of macrophage/microglial phenotypes. Polarization to the M1 phenotype is induced by IFN-γ, LPS, and GM-CSF and characterized by production of IL-12, IL-1, IL-23, IL-6, and inducible (i)NOS (10, 11). M1 cells participate in the induction of Th1 and Th17 responses (11). M2 cells are induced by IL-4, IL-10, and IL-13, identified by production of IL-10 and expression of arginase 1 and Ym1, and are implicated in resolution of inflammation and promotion of Th2 responses (12). M2 macrophages are protective in MS, EAE, and spinal cord injury (13–16). The JAK/STAT signaling pathway is used by cytokines and IFNs, and dysregulation of this pathway has pathological implications (17). Suppressors of cytokine signaling (SOCS) proteins inhibit JAK/STAT signaling by various mechanisms (18). SOCS proteins are induced by cytokines, creating a negative feedback loop to prevent excessive activation of the JAK/STAT pathway (18). The major function of SOCS3 is inhibition of signaling by the IL-6 5004–5009 | PNAS | March 27, 2012 | vol. 109 | no. 13

family of cytokines, causing inhibition of STAT3 activation (18). IL-6 has a deleterious role in EAE by activation of STAT3, which is pivotal for induction of Th17 cells (19–21). In EAE models, SOCS3 transcripts increase at the height of disease, and then decline, coincident with disease remission (22). Injection of SOCS3-expressing dendritic cells at EAE induction reduces clinical severity by suppression of Th1 and Th17 cell differentiation (23). Furthermore, SOCS3 expression in T cells inhibits IL-23 signaling, which constrains Th17 cell differentiation (24). These results indicate that the STAT3/SOCS3 axis influences the T-cell repertoire in EAE, with SOCS3 providing protection against disease. SOCS3 is integral for limiting excessive STAT3 activation and cytokine signaling in macrophages/microglia (25). Mice with conditional knockout of SOCS3 in cells of the myeloid lineage, LysMCreSOCS3fl/fl, were generated to investigate the function of myeloid SOCS3 in neuroinflammation. These studies reveal a role for SOCS3 in myeloid cells, through inhibition of STAT3 activation, to provide protection from detrimental inflammatory responses in the CNS. Results SOCS3 Influences M1 Polarization in Macrophages. To test whether deletion of SOCS3 in macrophages/microglia influences their phenotype, mice with conditional knockout of SOCS3 in cells of the myeloid lineage were generated (LysMCre-SOCS3fl/fl). SOCS3 deletion was successful in bone marrow derived macrophages (BMDMs) [Fig. 1A, band corresponding to Cre-mediated deletion of the SOCS3 gene (△)]. Inducible SOCS3 mRNA expression was abolished in BMDMs (Fig. 1B), microglia, peritoneal macrophages, and dendritic cells (DCs) from LysMCreSOCS3fl/fl mice (Fig. S1 A–C). Ly-6Clo and Ly-6Chi monocytes were also SOCS3 deficient, as were Gr-1+ neutrophils (Fig. S1D). IL-6–induced STAT1 and STAT3 phosphorylation was enhanced in SOCS3-deficient BMDMs compared with SOCS3fl/fl BMDMs (Fig. 1C and Fig. S2), as was IL-6–induced IL-12, TNFα, iNOS, CCL2, and CXCL10 expression (Fig. 1 D and E). This gene expression pattern is reflective of the M1 proinflammatory phenotype. We next tested stimulation of these cells with LPS, a classical activator of M1 cells. SOCS3 deletion led to an enhancement of M1 genes such as iNOS, IL-6, IL-12, and IL-23 (Fig. 1F), as well as IRF5 protein expression (Fig. 1G) (11). Myeloid Ablation of SOCS3 Leads to Development of Atypical EAE. To determine the impact of myeloid SOCS3 on the disease course of EAE, immunization with MOG35-55 peptide was performed (26). Most strikingly, myeloid SOCS3-deficient mice developed early onset of a severe and nonresolving disease with features

Author contributions: H.Q., P.D.S., and E.N.B. designed research; H.Q., W.-I.Y., P.D.S., A.T.H., Y.L., M.T.M., S.L.R., L.L.Y., T.H.F., and K.P. performed research; H.Q., W.-I.Y., L.E.H., C.R., and E.N.B. analyzed data; and H.Q., P.D.S., L.E.H., C.R., and E.N.B. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

To whom correspondence may be addressed. E-mail: [email protected] or [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1117218109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1117218109

characteristic of atypical EAE (Fig. 2A and Movies S1 and S2). The mean day of onset of clinical signs was 8 d, and mean clinical score at peak of disease was 4.3. Because of disease severity, experiments were terminated at day 13 (Fig. 2A). No signs of classical EAE were noted in LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice, which developed classical EAE at day 10, with peak of disease at days 16–20 (Fig. 2B). These results indicate that SOCS3 ablation in myeloid cells leads to a shift from classical to atypical EAE. To assess changes in the inflammatory response between SOCS3fl/fl and LysMCre-SOCS3fl/fl mice, groups of mice were killed at day 12 (when LysMCre-SOCS3fl/fl mice had peak atypical EAE) and mononuclear cells isolated from the brain. The absolute number of mononuclear cells in the brain was increased approximately fivefold in LysMCre-SOCS3fl/fl mice (Fig. 2C). The proportion and absolute numbers of CD11b+/Gr-1+ neutrophils, Ly-6Clo and Ly-6Chi monocytes, CD11c+ DCs, B220+ B cells, and CD4+ T cells were also increased in LysMCre-SOCS3fl/fl mice (Fig. 2C and Fig. S3). No differences were observed in absolute numbers of mononuclear cells, T cells, B cells, and neutrophils from the spinal cord between SOCS3fl/fl and LysMCre-SOCS3fl/fl mice (Fig. S4). Enhanced demyelination was more prevalent in the cerebellum of LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice, which was accompanied by enhanced infiltration of CD4+ T cells, activated neutrophils, and enhanced macrophage/microglia activation (Fig. 2D). Qin et al.

Fig. 2. Ablation of SOCS3 in myeloid cells causes atypical EAE. (A) SOCS3fl/fl (n = 13) and LysMCre-SOCS3fl/fl (n = 14) mice were immunized with MOG35-55 peptide. Mean ± SD of atypical EAE scores. (B) Mean ± SD of classical EAE scores. (C) CNS-infiltrating mononuclear cells isolated from the cerebellum of LysMCre-SOCS3fl/fl mice (day 12) and SOCS3fl/fl mice (day 14) after MOG immunization. Cells were stained with trypan blue and counted. Cells stained with Abs to CD4, CD11b, Gr-1, CD11c, Ly-6C, and B220, and the percentage of CD11b+/Gr-1+ neutrophils, CD11b+/Ly-6Chi, and CD11b+/Ly-6Clo monocytes, CD11c+ dendritic cells, B220+ B cells, and CD4+ T cells were gated. The absolute number of cells is shown. **P < 0.001 and *P < 0.05. (D) Sections from the cerebellum of SOCS3fl/fl and LysMCre-SOCS3fl/fl mice (day 13) after MOG35-55 immunization were stained with H&E, LFB, GS-I-B4 for activated macrophages/ microglia, CD4 Ab for CD4+ T cells, and myeloperoxidase (MPO) Ab for activated neutrophils. Arrows indicate inflammatory infiltrates and demyelination.

STAT Signaling Pathway Is Aberrantly Activated in LysMCre-SOCS3fl/fl fl/fl Mice. We examined the status of STAT activation in SOCS3

and LysMCre-SOCS3fl/fl MOG-immunized mice, which revealed that LysMCre-SOCS3fl/fl mice have exaggerated STAT3 activation in the cerebellum compared with SOCS3fl/fl mice (Fig. 3A). Furthermore, STAT4 activation was evident, with STAT1 activation being modest (Fig. 3A). We also examined NF-κB p65, ERK, and p38 MAPK activation; no changes were observed between LysMCre-SOCS3fl/fl and SOCS3fl/fl mice. Levels of STAT3 activation in spinal cord were comparable in LysMCreSOCS3fl/fl and SOCS3fl/fl mice (Fig. S5A). Significantly elevated levels of CCL2, IL-6, and CXCL10 protein were observed in the brain of LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. 3B). CD11b+ cells expressing IL-6 were more numerous in the CNS of LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. S5B). Thus, the absence of SOCS3 in myeloid cells causes enhanced STAT3 and STAT4 activation and elevated cytokine/chemokine expression in the brain. Ablation of SOCS3 in Myeloid Cells Leads to Expanded Numbers of Th1 and Th17 Cells in the CNS. Staining for IFN-γ and IL-17A

demonstrated that the percentage of Th1, Th17, and IFN-γ+/IL17A+ cells in the brain is significantly higher in LysMCreSOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. 3C). In the spinal cord, there was a modest increase in Th1 and IFN-γ/IL17A double positive cells in LysMCre-SOCS3fl/fl mice, whereas Th17 cells were not detectable (Fig. S5C). In the inflammatory PNAS | March 27, 2012 | vol. 109 | no. 13 | 5005

IMMUNOLOGY

Fig. 1. SOCS3 influences M1 polarization of macrophages. (A) SOCS3 floxed mice were cross-bred with B6.129P2-Lyz2tm1(cre)Ifo/J mice to generate deletion of SOCS3 in the myeloid cell lineage. Genomic DNA of BMDMs was distinguished as floxP (fl) and cre-excised alleles (Δ). (B) BMDMs were treated with IL-6 (10 ng/mL) plus sIL-6R (25 ng/mL) for up to 4 h, and mRNA was analyzed by RT-PCR. (C) BMDMs were incubated with IL-6/sIL-6R for up to 4 h, then lysates immunoblotted with the indicated antibodies. (D) BMDMs were incubated with IL-6/sIL-6R for 4 h, and mRNA was analyzed by qRT-PCR. (E) BMDMs were treated with IL-6/sIL-6R for 24 h, cytokine and chemokine protein expression in the supernatants was determined by Multiplex. (F) BMDMs were treated with LPS (10 ng/mL) for 4 h, and mRNA was analyzed by RT-PCR. (G) BMDMs were incubated with LPS for 24 h, then lysates immunoblotted with IRF5, IRF3, and GAPDH Abs. *P < 0.05. Three independent experiments are represented.

the presence of M1 macrophages from LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. S8). In vitro T-cell polarization was next examined. M1 macrophages from LysMCreSOCS3fl/fl mice promoted a greater extent of Th1 and Th17 cell differentiation as reflected by IFN-γ/Tbx21 and IL-17A/Rorc expression, respectively, compared with M1 macrophages from SOCS3fl/fl mice (Fig. 4 B and C). SOCS3-deficient DCs also had a stronger capacity to induce Th1 and Th17 cell differentiation than DCs from SOCS3fl/fl mice (Fig. S9A), and also expressed higher levels of IL-6, IL-12, IL-23, and iNOS (Fig. S9B). We evaluated the influence of supernatants from M1/M2 polarized macrophages on neuronal viability. Supernatants from M1-polarized macrophages from LysMCre-SOCS3fl/fl mice induced significantly more neuronal toxicity than M1 macrophages from SOCS3fl/fl mice (Fig. 4D). Interestingly, M2 macrophages provided a neuroprotective effect. These results indicate that SOCS3-deficient macrophages, polarized to the M1 phenotype,

Fig. 3. STAT3/4 signaling pathways are aberrantly activated in LysMCre-SOCS3fl/fl mice. (A) Protein extracts from brains of unimmunized (day 0) or MOG35-55–immunized (day 13) SOCS3fl/fl and LysMCre-SOCS3fl/fl mice (n = 5) were immunoblotted with indicated antibodies. (B) Supernatants from brains of MOGimmunized mice (n = 5; day 13) were analyzed for protein expression of CCL2, IL-6, or CXCL10 by Multiplex. *P < 0.05. (C) CNS-infiltrating mononuclear cells from brain at day 13 after MOG immunization. Cells were stimulated with PMA/ionomycin/GolgiStop for 4 h and stained for the surface marker CD4 and intracellular flow for IFN-γ and IL-17A. *P < 0.05, **P < 0.001. (D) Cells were stained for surface markers CD11b and CD45. *P < 0.05. (E) CCL2 and CXCL10 mRNA expression was analyzed by qRT-PCR in CD45int/CD11b+ microglia (Mi) or CD45hi/CD11b+ macrophages/activated microglia (Ma) obtained from the cerebellum. *P < 0.05.

EAE brain, there was an increase in the percentage of CD45hi/ CD11b+ macrophages/activated microglia (34.9 vs. 16.7%) and the percentage of CD45int/CD11b+ microglia (19.0 vs. 13.8%) in the cerebellum of LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. 3D). These infiltrating macrophages/activated microglia (Ma) and microglia (Mi) expressed higher levels of CCL2 and CXCL10 mRNA (Fig. 3E). There were fewer CD45hi/CD11b+ cells in the spinal cord of LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. S5C). These data indicate the absence of SOCS3 in myeloid cells promotes myeloid cell recruitment and activation in the brain, which is associated with enhancement of Th1 and Th17 cells. We next examined the influence of myeloid cell SOCS3 on the priming phase of EAE. CD4+ T-cell proliferation to MOG antigen was enhanced in LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. S6). Furthermore, MOG-induced Th1 and Th17 cell differentiation in lymph nodes was enhanced in LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice (Fig. S7 A–E). In contrast, the number of MOG-specific Th2 cells was lower in LysMCre-SOCS3fl/fl mice, as were levels of IL-4 and IL13 (Fig. S7 C and D). These findings suggest that the atypical EAE observed in LysMCre-SOCS3fl/fl mice may be initiated by enhanced Th1/Th17 cell priming in the periphery. Myeloid Cell SOCS3 Expression Influences Th1 and Th17 Cell Polarization and Neuronal Toxicity. The absence of SOCS3 in BMDMs led to

a more pronounced M1 phenotype; thus, we assessed antigen presenting cell functions by these cells. There was no difference in MHC class II expression, but stimulated LysMCre-SOCS3fl/fl BMDMs expressed significantly higher levels of CD40, CD80, and CD86 compared with SOCS3fl/fl BMDMs (Fig. 4A). MOG35-55– stimulated proliferation of T cells was significantly enhanced in 5006 | www.pnas.org/cgi/doi/10.1073/pnas.1117218109

Fig. 4. Myeloid cell SOCS3 expression influences Th1 and Th17 differentiation and neuronal viability. (A) BMDMs were incubated with LPS (10 ng/mL) plus IFN-γ (10 ng/mL) for 48 h, then expression of MHC class II, CD40, CD80, or CD86 was examined. *P < 0.05. (B) M1 macrophages and 2D2 CD4+ T cells were cultured at a 1:5 ratio for anti-CD3 or MOG-specific T-cell differentiation. Th1 cells were differentiated with IL-12 (10 ng/mL) and anti–IL-4 (10 μg/mL), and Th17 cells were differentiated with TGF-β (5 ng/mL), IL-6 (20 ng/mL), IL-23 (10 ng/mL), anti–IFN-γ (10 μg/mL), and anti–IL-4 (10 μg/mL). At 4 d, mRNA was extracted from cells and analyzed by qRT-PCR. *P < 0.05. (C) At day 4, cells were analyzed with intracellular flow for IFN-γ or IL-17A. *P < 0.05. (D) BMDMs were polarized to M1 with LPS plus IFN-γ or to M2 with IL-4 (10 ng/mL) for 48 h. Supernatant (400 μL) from each condition was added to primary neuron cultures and incubated for 24 h. Neuronal viability was assessed using the lactate dehydrogenase (LDH) release assay. Average percentage of cytotoxicity ± SD of three independent experiments are shown. *P < 0.05 comparing WT-M1 vs. WT-UN; &P < 0.05 comparing SOCS3−/−M1 vs. SOCS3−/−UN; and #P < 0.05 comparing SOCS3−/−M1 vs. WT-M1.

Qin et al.

Protective Effect of M2 Macrophages in Atypical EAE. M2 macrophages are protective in classical EAE (16). We examined whether M2 macrophages from SOCS3fl/fl mice could influence atypical EAE. BMDMs from SOCS3fl/fl mice were incubated with M-CSF, which results in polarization to an M2 phenotype (Fig. S10A), transferred by i.v. injection into LysMCre-SOCS3fl/fl mice 1 d before MOG35-55 immunization and 3 d after EAE induction. Group I LysMCreSOCS3fl/fl mice developed severe atypical EAE and were killed at day 14 (Fig. 5A). LysMCre-SOCS3fl/fl mice that received M2 macrophages (group II) developed less severe EAE, reaching an average atypical clinical score of 3.1 at day 14. From day 14 to day 23 atypical disease stabilized in these mice, and at days 24–27 there was marked recovery, after which time the mice relapsed (Fig. 5A). Transfer of M2 macrophages resulted in lower expression of numerous cytokines and

chemokines (Fig. 5 B and C), which correlated with less severe EAE (score of 2.0) (Fig. 5A). In addition, mRNA expression for IL-17A and IFN-γ was lower in the cerebellum of mice with transferred M2 macrophages, whereas expression of Foxp3, reflective of Treg cells, was higher (Fig. 5D). In a separate experiment, all mice were killed at day 18 to allow comparisons at the same time point. Group I had an atypical score of 4.0, and group II an atypical score of 2.9. Phosphorylated STAT3 was detected in the brain of group I mice, whereas reduced STAT3 activation was observed in group II mice with less severe atypical EAE (Fig. 5E). Examination of mRNA from the brain revealed reduced expression of Th1 markers (IFN-γ and Tbx21), Th17 markers (IL-17A and Rorc), and enhanced expression of GATA3 mRNA (marks Th2 cells) in group II (Fig. 5F). LysMCre-SOCS3fl/fl mice receiving M2 macrophages also had reduced levels of IL-1, iNOS, TNF-α, CCL2, or CXCL10 mRNA (Fig. 5G). Furthermore, the percentage of CD45hi/CD11b+

Fig. 5. Protective effect of M2 macrophages in atypical EAE. (A) EAE was induced with MOG as described in LysMCre-SOCS3fl/fl mice. BMDMs from SOCS3fl/fl (WT) mice were cultured with 10 ng/mL of MCSF for 5 d. SOCS3fl/fl M2 macrophages (5 × 106) were transferred by i.v. injection into LysMCre-SOCS3fl/fl mice. Group I (◆), no cells, PBS; group II (☐), cell transfer 1 d before MOG and 3 d after MOG injection (black arrows). Development of atypical EAE was compared between the two groups (n = 3). Protein expression levels of cytokines (B) and chemokines (C) was determined by Multiplex from brains of group I or group II mice (days 13 and 28), respectively. *P < 0.05. (D) mRNA from group 1 or group II brains (days 13 and 28), respectively, was analyzed by qRT-PCR. *P < 0.05. (E) In a separate experiment, brain homogenates from group I (n = 5) and group II (n = 5) (day 18) were immunoblotted with the indicated antibodies (F and G). mRNA from group 1 or group II brain (day 18) was analyzed by qRT-PCR. *P < 0.05. (H) CNS-infiltrating mononuclear cells from brains of group I and group II mice 18 d after MOG immunization and CD11b and CD45 expression determined by flow cytometry. (I) Sections from LysMCre-SOCS3fl/fl mice stained by H&E (day 18).

Qin et al.

PNAS | March 27, 2012 | vol. 109 | no. 13 | 5007

IMMUNOLOGY

provide signals to induce Th1/Th17 responses and neuronal demise in a more potent manner than WT M1 cells.

macrophages/activated microglia was reduced in group II (Fig. 5H), as was leukocyte infiltration (Fig. 5I). Discussion Our study demonstrates that SOCS3 has a critical role in limiting activation of macrophages/microglia, indicating that SOCS3 promotes and/or maintains quiescence of these cells. In vitro, the absence of SOCS3 led to a pronounced polarization to the M1 phenotype, which was associated with expression of M1 genes including iNOS, IL-1β, IL-12p40, IL-23p19, IL-6, CCL2, CXCL10, and IRF5. SOCS3-deficient macrophages promoted Th1/Th17 responses, another characteristic of M1 cells. Enhanced M1 polarization in the absence of SOCS3 had functional consequences for neurons; supernatants from these cells induced significant neuronal toxicity. This is relevant because neuronal toxicity with synaptic dysfunction during EAE is associated with activated microglia (27, 28), and axonal transection in MS lesions is related to the abundance of activated macrophages (29). Our findings also demonstrate that myeloid cell expression of SOCS3 is crucial in limiting CNS autoimmunity. Ablation of SOCS3 in myeloid cells caused a shift from classical EAE to atypical EAE, associated with infiltration of macrophages, Ly-6Clo, and Ly-6Chi monocytes, DCs, neutrophils, B cells, and CD4+ T cells into the CNS, demyelination, heightened STAT3/STAT4 signaling and cytokine/ chemokine expression, and a shift in T-cell profile exclusively in the cerebellum, but not the spinal cord. This change in leukocyte migration pattern suggests that SOCS3 in macrophages is essential to regulate leukocyte infiltration into the brain. A critical role has been established for monocytes/macrophages/microglia in immune modulation of the CNS (9, 30, 31). MicroRNA-124 (miR-124) was shown to promote microglial quiescence and skew polarization from an M1 toward an M2 phenotype (8). In vivo administration of miR-124 suppressed EAE by deactivation of macrophages and reduced activation of T cells. In our study, SOCS3 shapes the innate immune response of myeloid cells with respect to M1 polarization. In vitro, we demonstrated that T-cell proliferation was greater in the presence of M1 cells from LysMCre-SOCS3fl/fl mice compared with SOCS3fl/fl mice. Polarization to Th1 and Th17 phenotypes was also enhanced, likely due to SOCS3-deficient M1 secretion of IL12, and IL-6, IL-23, and/or IL-1, respectively. Th1 cells in turn produce IFN-γ and GM-CSF, whereas Th17 cells produce GMCSF (32, 33), which will sustain the M1 phenotype, establishing a feed-forward regulatory loop for M1, Th1, and Th17 cells. Atypical EAE is characterized by involvement of brainstem/ cerebellum, vertigo, and ataxia (19, 34, 35). Stromnes et al. (34), demonstrated that T-cell infiltration and inflammation in the brainstem vs. spinal cord was observed when Th17 cells outnumbered Th1 cells. Another feature of atypical EAE is enhanced infiltration of the cerebellum by neutrophils and B cells (19). The atypical EAE seen in LysMCre-SOCS3fl/fl mice share some of these features: pronounced neutrophil and B-cell infiltration, elevated levels of Th17 cells, and increased production of IL-6 and IL-17. We also observed increased Th1 cells and levels of IFN-γ in the cerebellum of LysMCre-SOCS3fl/fl mice, as well as 100% mortality between days 13–18, which differs from other atypical models. Our data thus far suggest that heightened STAT3/STAT4 activation in the cerebellum is a major contributor to atypical EAE, which can be tested in the future by use of JAK/STAT inhibitors. The primary function of M2 macrophages is resolution of inflammation and promotion of Th2 responses (12). Glatiramer acetate treatment induces antiinflammatory type II monocytes, characterized by secretion of IL-10 and TGF-β. Upon adoptive transfer, these cells inhibited classical EAE, suppressed Th17 development, and promoted Th2 and Treg cells in recipient mice (16). In our study, M2 macrophages delayed the onset and reduced the severity of atypical EAE, which correlated with a decrease in STAT3 activation in the cerebellum, decreased Th1, Th17, and M1 gene expression, reduced macrophage/microglial activation, and less leukocyte infiltration. Th2 and Treg cell genes such as GATA3 and Foxp3, respectively, were increased. Preliminary results indicate that M2 products IL-10 and TGF-β inhibit LPSinduced M1 polarization (Fig. S10B). Thus, in LysMCre-SOCS3fl/fl 5008 | www.pnas.org/cgi/doi/10.1073/pnas.1117218109

mice, introduction of WT M2 cells may provide the microenvironment to inhibit development of M1 macrophages, contributing to the lessened atypical EAE phenotype. Aberrant activation of the JAK/STAT pathway has pathological implications for autoimmune and neurodegenerative diseases (20, 36). Dysregulation of the STAT3 pathway is involved in MS due to the importance of STAT3 in myeloid cell activation, T-cell differentiation, and cytokine/chemokine induction. Loss of STAT3 in CD4+ T cells prevents the development of EAE (20). In addition, STAT3 has been identified as an MS susceptibility gene (37). Monocytes from patients with relapsing–remitting MS express increased levels of activated STAT3 and less SOCS3 during relapse than do cells from patients in remission (38), suggesting an association among decreased SOCS3 expression, increased STAT3 activation, and MS relapse. We propose that the STAT3/SOCS3 axis in myeloid cells influences M1 polarization, the T-cell repertoire, and neuronal viability in EAE/MS, with SOCS3 expression providing protection against neuroinflammation and neurodegeneration. Furthermore, SOCS3-deficient DCs are also aberrantly activated and promote Th1/Th17 cell differentiation, indicating that SOCS3 serves to dampen DC activation. The LysMCre-SOCS3fl/fl model of atypical EAE provides an important model for MS patients with high incidence of cerebellar dysfunction, given that disease affecting the cerebellum has a poor prognosis and rapid progression (39–41). Understanding the role of STAT3/SOCS3 in neuroinflammation will lead to development of rational therapeutics for MS patients, including JAK/STAT inhibitors currently in clinical trials (42). Materials and Methods Mice. C57BL/6 and MOG–T-cell receptor transgenic 2D2 mice (43) were bred in the animal facility at the University of Alabama, Birmingham (UAB). SOCS3 floxed transgenic (SOCS3fl/fl) mice (44) were the generous gift of Warren Alexander (Walter and Eliza Hall Institute of Medical Research, Victoria, Australia), and were bred at UAB. SOCS3 conditional knockout (LysMCre-SOCS3fl/fl) mice were generated by serial breeding of SOCS3fl/fl mice with mice expressing Cre recombinase under the control of the LysM promoter. All experiments were reviewed and approved by the institutional animal care and use committee of UAB. Peptides, Antibodies, and Cytokines. MOG35-55 peptide was synthesized by New England Peptide. LPS was from Sigma-Aldrich. neutralizing antibodies (Abs) to IL-4 and IFN-γ, and Abs against CD3, B220, Gr-1, CD4, IFN-γ, IL-6, and IL-17A used for flow cytometry are from eBioscience. Mouse IL-6, IFN-γ, IL-12, IL-23, MCSF, and IL-4, and human TGF-β1, are from R&D Systems. Abs against phosphoSTAT1, phospho-STAT3, phospho-p65, phospho-p38, phospho-ERK1/2, STAT1, STAT3, p65, p38, ERK1/2, IRF3, and IRF5 are from Cell Signaling Technology. Ab against GAPDH is from Abcam, and Abs against Ly-6C, CD11c, MHC class II, CD40, CD80, CD86, CD11b, and CD45 are from BD Pharmingen. EAE Induction and Assessment. Active EAE was induced as previously described (45). Eight- to twelve-week-old mice were immunized s.c. with 200 μg of MOG35-55 emulsified in CFA (supplemented with 2 mg/mL of Mycobacterium tuberculosis) and injected i.p. on days 0 and 2 with 500 ng pertussis toxin. Assessment of classical EAE was as follows: 0, no disease; 1, decreased tail tone; 2, hind limb weakness or partial paralysis; 3, complete hind limb paralysis; 4, front and hind limb paralysis; and 5, moribund state (26, 45). Assessment of atypical EAE was done using the scoring system described by Stromnes et al. (34), with slight modifications: 1, hunched appearance, slight head tilt; 2, ataxia, scruffy coat; 3, severe head tilt, slight axial rotation, staggered walking; 4, severe axial rotation, spinning; and 5, moribund. Cell Preparation. Bone marrow cells were cultured with RPMI medium 1640 containing 10% (vol/vol) FBS and 20 ng/mL of mouse M-CSF for 7 d to obtain BMDMs. Primary microglia were prepared from 1-d-old neonatal mice. Primary murine neurons were prepared from embryonic day 17 C57BL/6 brains. Adoptive Transfer of Macrophages. Five × 106 BMDMs from SOCS3fl/fl mice were cultured for 5 d in M-CSF, analyzed for M2 markers, and then were injected i.v. into LysMCre-SOCS3fl/fl mice at −1 and +3 d after MOG immunization. Isolation of Mononuclear Cells. Spleen and draining cervical lymph nodes were excised and cell suspensions prepared as described (26). To isolate mononuclear cells from spinal cord and brain, tissues were separated on

Qin et al.

a 30/70% Percoll gradient. Cells at the 30/70% interface were collected for further analysis.

subjected to qRT-PCR. The data were analyzed using the comparative Ct method to obtain relative quantitation values.

Intracellular Cytokine Staining. Cells were stimulated for 4 h with PMA/ionomycin (25 ng/mL and 1 μg/mL) plus GolgiStop (BD Pharmingen). Cells were analyzed for the intracellular production of cytokines by staining with anticytokine Abs and subsequent flow cytometry, as described (46).

Immunoblotting. Thirty micrograms of cell lysate or brain homogenate was separated by electrophoresis and probed with antibodies as described previously (48). ELISA. Supernatants were assayed with the Millipore Mouse Cytokine/ Chemokine Panel (MPXMCYTO-70K). Expression levels were normalized to total protein levels. Statistical Analysis. Levels of significance for comparison between samples were determined by the Student’s t test distribution. A value of P