Various microbial infections evoke innate immune responses

Melanoma Differentiation-Associated Gene 5 Is Critical for Protection against Theiler s Virus-Induced Demyelinating Disease Young-Hee Jin, a Seung Jae Kim, a * Eui Young So, a * Liping Meng, a Marco Colonna, b and Byung S. Kim a Department of Microbiology-Immunology, Northwestern University Medical School, Chicago, Illinois, USA, a and Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, Missouri, USA b Infection of dendritic and glial cells with Theiler s murine encephalomyelitis virus (TMEV) induces various cytokines via Tolllike receptor- and melanoma differentiation-associated gene 5 (MDA5)-dependent pathways. However, the involvement and role of MDA5 in cytokine gene activation and the pathogenesis of TMEV-induced demyelinating disease are largely unknown. In this study, we demonstrate that MDA5 plays a critical role in the production of TMEV-induced alpha interferon (IFN- ) during early viral infection and in protection against the development of virus-induced demyelinating disease. Our results indicate that MDA5-deficient 129SvJ mice display significantly higher viral loads and apparent demyelinating lesions in the central nerve system (CNS) accompanied by clinical symptoms compared with wild-type 129SvJ mice. During acute viral infection, MDA5- deficient mice produced elevated levels of chemokines, consistent with increased cellular infiltration, but reduced levels of IFN-, known to control T cell responses and cellular infiltration. Additional studies with isolated CNS glial cells from these mice suggest that cells from MDA5-deficient mice are severely compromised in the production of IFN- upon viral infection, which results in increased cellular infiltration and viral loads in the CNS. Despite inadequate stimulation, the overall T cell responses to the viral determinants were significantly elevated in MDA5-deficient mice, reflecting the increased cellular infiltration. Therefore, the lack of MDA5-mediated IFN- production may facilitate a massive viral load and elevated cellular infiltration in the CNS during early viral infection, leading to the pathogenesis of demyelinating disease. Various microbial infections evoke innate immune responses that recognize foreign materials using pattern recognition receptors (PRRs). These receptors include Toll-like receptors (TLRs), nucleotide-binding oligomerization domain proteins (NODs), double-stranded RNA (dsrna)-activated protein kinase R (PKR) and caspase recruitment domain (CARD)- containing RNA helicases, including retinoic acid-induced protein I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5) (29). All of the PRRs can sense a distinct pathogenassociated molecular pattern and stimulate innate immune responses against the infectious agents, including the synthesis of multiple cytokines. In particular, nonself nucleic acids, such as dsrna and single-stranded RNA (ssrna) of viral genomes and/or replication intermediates, are mainly recognized by TLRs, PKR, and RNA helicases (21). MDA5 was first identified as a type I interferon (IFN)- responsive apoptosis-inducing gene that contains a CARD and an RNA helicase domain (17, 18). Both RIG-I and MDA5 bind synthetic poly(i C) and viral dsrna. These molecules are essential for type I IFN production in response to RNA viruses (47). RIG-I and MDA5 appear to recognize a specific subset of viruses, although some reoviruses and flaviviruses trigger innate immune responses in both a RIG-I- and MDA5-dependent manner (8, 27). MDA5 mediates type I IFN production in response to picornaviruses, such as encephalomyocarditis virus (EMCV) and Theiler s murine encephalomyelitis virus (TMEV), and MDA5-deficient mice are highly susceptible to viral infections (9, 20). After engagement of MDA5 to dsrna, the complex interacts with a specific adaptor, beta interferon (IFN- ) promoter stimulator 1 (IPS-1), to activate downstream protein kinases. IRF3/IRF7 and NF- B are subsequently activated, leading to the production of various cytokines (22, 24). TMEV contains an ssrna genome with positive polarity and belongs to the genus Cardiovirus of the family Picornaviridae (33, 36). TMEV infection leads to chronic viral persistence in the central nervous system (CNS) of susceptible mice, and infected mice develop an immune-mediated demyelinating disease similar to human multiple sclerosis (5). TMEV infects various glial cells, including astrocytes and microglia, as well as professional antigen-presenting cells (APCs), including macrophages and dendritic cells (DCs) (14, 32, 34, 41). Such viral infections activate the production of proinflammatory cytokines and chemokines, including type I IFNs, tumor necrosis factor alpha (TNF- ), interleukin 6 (IL-6), CCL2, and CCL5. These proinflammatory chemokines and cytokines are known to be involved in cellular infiltration and activation and consequently affect the development and progression of virus-induced demyelinating disease (23). TMEV-induced cytokine gene expression, as well as viral replication, is dependent on the NF- B pathway (19, 25, 35). NF- B activation following viral infection requires TLR3 and TLR2 (42, 43) and is partially dependent on PKR (3, 35). However, the induction of type I IFNs and IRF3/IRF7 activation, critical for type I IFN production, are not fully dependent on the presence of TLRs Received 3 October 2011 Accepted 4 November 2011 Published ahead of print 16 November 2011 Address correspondence to Byung S. Kim, bskim@northwestern.edu. * Present address: S. J. Kim, Department of Cancer Genetics, Roswell Park Cancer Institute, Buffalo, New York, USA; E. Y. So, Division of Rheumatology, University of Illinois at Chicago, Chicago, Illinois, USA. Y.-H. Jin, S. J. Kim, and E. Y. So contributed equally to this article. Copyright 2012, American Society for Microbiology. All Rights Reserved. doi:10.1128/jvi.06457-11 0022-538X/12/$12.00 Journal of Virology p. 1531 1543 jvi.asm.org 1531

Jin et al. (43). These data indicate that other PRRs are also involved, either cooperatively or independently, in mediating TMEV-induced innate cytokine responses, particularly the production of type I IFNs. In addition, TMEV-induced type I IFN production has been shown to be significantly impaired in MDA5-deficient murine embryonic fibroblasts (MEFs) and bone-marrow-derived DCs, indicating that MDA5 could also be responsible for type I IFN production after TMEV infection (20, 37). However, the role of MDA5 signaling in the production of type I IFNs and in protection against TMEV-induced demyelinating disease accompanied by viral persistence in the CNS is unknown. In this study, we utilized MDA5-deficient mice to examine the potential role of MDA5 signals in the development of TMEVinduced demyelinating disease. Our results indicate that MDA5- deficient mice, with a relatively resistant genetic background, display significantly higher viral loads and cellular infiltration in the CNS than 129/SvJ wild-type (WT) mice. Virus-infected MDA5- deficient (MDA5 knockout [KO]) mice developed demyelinating disease accompanied by axonal losses in the brain and spinal cord. The production of various inflammatory cytokines in the CNS was generally uncompromised in virus-infected MDA5 KO mice during an early stage of infection, except for the drastic reduction of IFN- levels. Despite the low proportions of virus-specific T cells, the overall levels of antiviral T cell responses in virus-infected MDA5 KO mice, including Th17 responses, were higher than those in WT mice due to elevated cellular infiltration. The levels of costimulatory molecules were lower, but the level of the inhibitory PDL-1 molecule was higher on APCs from virus-infected MDA5 KO mice. This finding probably reflects a deficiency in IFN- production that results in lower proportions of functional T cells. Therefore, these results collectively indicate that MDA5-mediated signals play a critical role via IFN- production in controlling viral loads and preventing pathogenic T cell responses, which lead to the development of demyelinating disease. MATERIALS AND METHODS Mice. MDA5 / (MDA5 KO) mice were obtained from Marco Colonna (Washington University School of Medicine, St. Louis, MO). Control 129X1/SvJ (129SvJ) mice were obtained from the Jackson Laboratories (Bar Harbor, ME). OT-II.2 TCR-transgenic (Tg) mice recognizing ovalbumin 323 to 339 (1) were obtained from William Karpus at Northwestern University. Experimental procedures with mice were approved by the Institutional Animal Care and Use Committee of Northwestern University, and the mice were used in accordance with the animal care guidelines of the National Institutes of Health (NIH). Preparation of macrophages. Macrophages were isolated from the peritoneal exudates of WT and MDA5 KO mice that had been injected with 3% thioglycolate solution 3 days prior to cell harvest. The cells were seeded onto a 10-cm plate, and nonadherent cells were removed after incubation for 1hat37 C. Cell cultures and reagents. Macrophages were cultured in Dulbecco s modified Eagle s medium (DMEM) (Mediatech, Inc., Herndon, VA) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals, Lawrenceville, GA), 2 mm L-glutamine, 10 mm sodium pyruvate, penicillin (100 U/ml)/streptomycin (100 g/ml), and 5 M -mercaptoethanol. Synthetic peptides. All synthetic peptides were purchased from Genmed Synthesis (San Francisco, CA). Stock peptides of the previously defined TMEV-specific H-2 b -restricted CD4 (VP2 203-220 and VP4 25-38 ) and CD8 (VP2 121-130 and VP3 110-120 ) T cell epitopes were prepared in 8% dimethyl sulfoxide in phosphate-buffered saline (PBS). Determination of viral titer and viral infection. The BeAn strain of TMEV was propagated in BHK-21 cells grown in DMEM supplemented with 7.5% donor calf serum. Viral titers were determined by plaque assays on BHK cell monolayers, followed by staining with 0.1% crystal violet as previously described (39). For in vitro viral infection, the cells were washed and subsequently incubated in infection medium (DMEM supplemented with 0.1% bovine serum albumin) with TMEV at 10 multiplicities of infection (MOIs). For viral infection of mice, approximately 30 l (1 10 7 PFU) of TMEV was injected into the right hemispheres of 5- to 7-week-old mice anesthetized with isoflurane. Immunohistopathological examinations. Mice at 30 days postinfection were perfused with 50 ml of PBS via intracardiac puncture. The brains and spinal cords from the MDA5 KO mice and the 129SvJ control mice were dissected and fixed in 4% formalin in PBS for 24 h and embedded in paraffin for sectioning and staining. The paraffin-processed brains and spinal cords were sectioned at 6 m. Sets of three adjacent sections from each animal were deparaffinized, rehydrated, and evaluated by hematoxylin and eosin (H&E) staining for inflammatory infiltrates, by Luxol Fast Blue (LFB) staining for axonal demyelination, and by Bielschowsky silver staining for axon loss and damage. For immunofluorescent examination, 6- m cross sections of the spinal cords were deparaffinized and subjected to antigen retrieval in a ph 6.0 citrate buffer in a microwave. For staining with rabbit polyclonal anti-tmev antibodies, Alexafluor 488-labeled goat anti-rabbit IgG (Invitrogen) was utilized as the secondary reagent following overnight incubation with anti-tmev antibodies. Nuclei were counterstained with DAPI (4=,6-diamidino-2- phenylindole) (Sigma) according to the manufacturer s instructions. The sections were observed and recorded using a Leica DMR fluorescence microscope, and images were captured using a AxioCam MRc camera and AxioVision imaging software. Isolation of CNS-infiltrating mononuclear cells. Brains and spinal cords were removed from the mice after perfusion with Hanks balanced salt solution through the left ventricle. The tissues were forced through a steel mesh to prepare single-cell suspensions and incubated at 37 C for 45 min in 250 g/ml collagenase type 4 (Worthington Biochemical Corp., Lakewood, NJ). A continuous 100% Percoll gradient (Pharmacia, Piscataway, NJ) was achieved after centrifugation at 27,000 g for 30 min to enrich CNS-infiltrating cells. Reverse transcription (RT)-PCR and real-time PCR. Total RNA was purified from the lysates of brain and spinal cord cells using TRIzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer s instructions. First-strand cdna was synthesized from 1 to 4 g total RNA using Moloney murine leukemia virus (MMLV) reverse transcriptase and oligo(dt) 18. Primers were obtained from Integrated DNA Technologies (Coralville, IA). The sense and antisense primer sequences were as follows: IL-6, 5=-AGTTGCCTTCTTGGGACTGA-3= and 5=-TCCACGATT TCCCAGAGAAC-3=; CXCL1, 5=-GCTGGGATTCACCTCAAGAA-3= and 5=-TGGGGACACCTTTTAGCATC-3=; CCL2, 5=-AGCAGGTGTCC CAAAGAAGCTGTA-3= and 5=-AGAAGTGCTTGAGGTGGTTGTGGA- 3=; CCL5, 5=-GTGCCCACGTCAAGGAGTAT-3= and 5=-GGGAAGCGT ATACAGGGTCA-3=; IFN- 4, 5=-AAGGCTCAAGCCATCCTTGTGCT A-3= and 5=-TTGCCAGCAAGTTGGTGAGGAAG-3=; IFN-, 5=-CCCT ATGGAGATGACGGAGA-3= and 5=-CTGTCTGCTGGTGGAGTTGA- 3=; IL-1, 5=-TCATGGGATGATAACCTGCT-3= and 5=-CCCATAC TTTAGGAAGACACGGAT-3=; IFN-, 5=-ACTGGCAAAAGGATGGTG AC-3 and 5-TGAGCTCATTGAATGCTTGG-3; IL-17A, 5=-CTCCAGAA GGCCCTCAGACTAC-3= and 5=-AGCTTTCCCTCCGCATTGACACA G-3=; TGF-,5=-CGTGGAAATCAACGGG-3= and 5=-CAGAAGTTGGC ATGGT-3=; GAPDH (glyceraldehyde-3-phosphate dehydrogenase), 5=-A ACTTTGGCATTGTGGAAGG-3= and 5=-ACACATTGGGGGTAGGAA CA-3=; and TMEV genome, 5=-CCCAGTCCTCAGGAAATGAAGG-3= and 5=-TCCAAAAGGAGAGGTGCCATAG-3=. For a quantitative analysis of gene expression, real-time PCR was performed using icycler SYBR green I master mix and an icycler real-time machine (Bio-Rad, Hercules, CA). GAPDH expression served as an internal reference. Each real-time PCR was performed in triplicate. 1532 jvi.asm.org Journal of Virology

Role of MDA5 in Resistance to TMEV Enzyme-linked immunosorbent assay (ELISA). The cytokine levels produced by T cells stimulated with isolated APCs for 36 h in the presence of PBS or epitope peptides were determined. IFN- (Opteia kit; BD Pharmingen, San Diego, CA) and IL-17 (R&D Systems, Minneapolis, MN) levels were assessed. The plates were read using a Spectra Max 190 microplate reader (Molecular Devices, Sunnyvale, CA) at a 450-nm wavelength. Intracellular-cytokine staining. CNS mononuclear cells from 129SvJ and MDA5 KO mice were cultured in plates coated either with anti-cd3 plus anti-cd28 antibodies or with viral epitope peptides in the presence of a Golgi-Plug (BD Pharmingen) for 6 h at 37 C. The cells were then incubated in 50 l of 2.4G2 hybridoma (American Type Culture Collection) supernatant for 30 min at 4 C to block the Fc receptors. The cells were incubated for an additional 30 min at 4 C in the presence of allophycocyanin-conjugated anti-cd8 (clone 53-6.7) or anti-cd4 (GK1.5) antibody diluted in 50 l of Fc-block (2.4G2 supernatant). After two washes, intracellular IFN- staining was performed according to the manufacturer s instructions (BD Pharmingen) using phycoerythrinlabeled rat monoclonal anti-ifn- antibody (XMG1.2). Flow cytometry. CNS cell types and their expression of molecules associated with antigen presentation were assessed using flow cytometry. The cells were stained with PE-conjugated anti-cd11b (clone M1/70) and allophycocyanin-conjugated anti-cd45 (clone Ly-5) for resident microglia cells and infiltrated macrophages and lymphocytes. Allophycocyaninconjugated anti-cd8 (clone 53-6.7) and phycoerythrin-conjugated anti- CD4 (GK1.5) and anti-pdl-1 (clone MIH5) antibodies were used. In addition, fluorescein isothiocyanate (FITC)-conjugated antibodies to CD80 (clone 16-10A1), CD86 (clone GL-1), CD40 (clone HM40-3), H-2D b (clone 28-14-8), and I-A b (clone AF6-120.1) were used. All antibodies used for flow cytometry were purchased from BD Pharmingen (San Diego, CA). The cells were analyzed on a Becton Dickinson FACSCalibur flow cytometer. Statistical analyses. The significance of the differences (two-tailed P value) between the experimental groups with various treatments and the control was analyzed with an unpaired Student s t test using the InStat Program (GraphPad Software, San Diego, CA). P values of 0.05 were considered significant. RESULTS MDA5 KO mice developed a neurological disease after infection with TMEV. To determine the role of MDA5 in TMEV-induced demyelinating disease, we monitored clinical disease after intracerebral inoculation of the BeAn strain of TMEV. When 1 10 6 PFU of TMEV was inoculated, which consistently induces clinical disease in susceptible SJL mice, both WT and MDA5 KO 129SvJ mice did not develop clinical disease (data not shown). However, inoculation with a 10-fold increase in virus (1 10 7 PFU) induced clinical symptoms starting at 10 days postinfection (Fig. 1A). Interestingly, the severity of disease was significantly higher in MDA5 KO mice than in WT mice at 14 and 21 days postinfection (P 0.01). The incidence of disease was also significantly higher in MDA5 KO mice (80%) than in WT mice (20%). However, it was unexpected to observe that the severity of disease in both WT and MDA5 KO mice peaked at 14 days postinfection, and the majority of the mice recovered after 40 days postinfection (Fig. 1A). Recovery of mice from TMEV-induced demyelinating disease is very rare, and it was previously seen only in C57L/J mice (6). Therefore, the recovery from clinical disease may reflect the 129SvJ strain background. MDA5 is required for effective viral clearance from the CNS of mice infected with TMEV. MDA5 recognizes the dsrna intermediates of picornaviruses, such as TMEV, EMCV, and Mengo virus, to trigger antiviral innate immunity (9, 20). To test the potential role of MDA5 in TMEV clearance in the CNS of infected animals, we inoculated MDA5 KO and WT (129SvJ) mice with TMEV (1 10 7 PFU). Viral-load levels in the CNS were then measured by quantitative RT-PCR and plaque assays at 8 and 28 days postinfection (Fig. 1B and C). The levels of VP1 RNA and the TMEV loads in both the brain and spinal cord were significantly higher in MDA5 KO mice than in WT mice. Approximately 10- to 700-fold higher numbers of TMEV plaques were found in MDA5 KO mice at 8 and 28 days postinfection. In addition, the levels of viral RNA measured with VP1 RNA were consistently higher in the brains and spinal cords of MDA5 KO mice at all time points (8 and 28 days postinfection) than in WT mice. Additional immunohistological examination of the spinal cords from mice at 30 days postinfection using anti-tmev antibodies confirmed the elevated viral loads in MDA5 KO mice (Fig. 1D). The distribution of viral proteins in MDA5 KO mice was concentrated around the perivascular and white-matter regions. In contrast, the distribution of viral proteins in WT mice was very limited and sparse. Taken together, these results indicate that MDA5 plays a critical role in viral clearance in the CNS of TMEV-infected mice. TMEV-infected MDA5 KO mice display demyelinating lesions in the CNS. Consistent with the severity of clinical disease, histopathological examinations clearly showed more severe encephalomyelitis and demyelination in TMEV-infected MDA5 KO mice than in WT mice at 30 days postinfection (Fig. 2). We chose to present the representative histopathology results at the time when the clinical symptoms in the WT mice completely disappeared while they were maintained in MDA5 KO mice. The levels of focal and perivascular mononuclear cell infiltration in the cerebral cortex, corpus callosum, hippocampus, meninges, and periventricular region, which are characterized by a lesion in an early-phase TMEV-infected brain, were higher in MDA5 KO mice than in WT mice at 10 days postinfection (data not shown). As expected, WT 129SvJ mice of the resistant H-2 b strain infected with TMEV showed minimal to mild pathological lesions at 30 days postinfection (Fig. 2). In contrast, the lack of MDA5 expression in the same mouse strain resulted in moderate to severe demyelination in the white matter of the CNS during this period (Fig. 2d). MDA5 KO mice displayed multiple demyelination foci combined with mononuclear cell infiltration in the white matter of the brain stem, such as the pons and medulla oblongata, at 30 days postinfection (Fig. 2d). These results clearly indicate that the loss of MDA5 expression alters the outcome of TMEV infection. In the spinal cords, cellular infiltration, vacuole formation, and myelin damage were more extensive in MDA5 KO mice than in WT mice at 10 days postinfection (data not shown), consistent with their levels of clinical symptoms during this early phase of TMEV infection (Fig. 1A). As expected in a mouse strain resistant to TMEV-induced demyelinating disease, the lumbar spinal cords of WT 129SvJ mice displayed minimal pathological appearance at 30 days postinfection (Fig. 3A, a, c, and e, and B, a). Although few lymphocytes that had infiltrated into the CNS were found by H&E staining (Fig. 3A, e), the myelin sheaths and axons appeared to be normal based on the Luxol Fast Blue (Fig. 3A, a) and Bielschowsky silver (Fig. 3A, c) stains. In contrast, MDA5-deficient mice with the same genetic background displayed moderate to severe demyelination in the white matter of the CNS during TMEV infection (Fig. 3A, b, d, and f, and B, b). Abnormal vacuoles and axon damage in the white matter of the spinal cords were also apparent February 2012 Volume 86 Number 3 jvi.asm.org 1533

Jin et al. FIG 1 Development of clinical disease accompanies viral load in the CNS of MDA5 KO mice. (A) The development of clinical symptoms was monitored after intracerebral TMEV inoculation (1 10 7 PFU) into age- and sex-matched WT 129X1/SvJ (n 10) and MDA5 KO (n 10) mice. Clinical symptoms were assessed weekly according to the following grading scale: grade 0, no clinical signs; grade 1, mild waddling gait or flaccid tail; grade 2, severe waddling gait; grade 3, moderate hind limb paresis; and grade 4, severe hind limb paralysis. % affected represents the percentage of mice displaying clinical symptoms of TMEV-induced demyelinating disease. The significance of the differences in clinical scores was analyzed by a paired t test. (B and C) The viral loads of the brains and spinal cords from WT and MDA5 KO mice were determined at 8 and 28 days postinfection (DPI). (B) The levels of infectious TMEV in whole brains and spinal cords from 3 to 5 mice per group were determined by plaque assays in BHK cells. (C) The relative expression of VP1 RNA was determined by quantitative RT-PCR normalized with GAPDH. The error bars indicate the standard deviations (SD) of triplicate samples from 3 mice per group. The levels of viral load in the CNS of MDA5 KO mice were significantly higher than those in WT mice (P 0.01 at all time points, based on a paired Student t test). (D) Immunofluorescent staining of TMEV-infected cells in the spinal cords of MDA5 KO and WT mice at 30 days postinfection. TMEV-infected cells were stained with rabbit anti-tmev antibodies (green) and counterstained with DAPI (blue). The scale bar represents 50 m., P 0.001;, P 0.01;, P 0.05 (Student s t test). with Bielschowsky silver staining (Fig. 3A, d). High magnification of demyelinated lesions in the ventral spinal cords of MDA5 KO mice showed clustered mononuclear infiltrations in the meninges and demyelinating areas (Fig. 3A, f). Glial cells in the white matter, including microglia and astrocytes, appear to be highly activated based on their morphological changes (Fig. 3A, f). These results indicate that the loss of MDA5 expression renders the resistant mouse strain susceptible to the development of TMEV-induced demyelinating disease. Virus-infected MDA5 KO mice display higher cellular infiltration and viral load in the CNS, accompanied by lower IFN- production. To further understand the cellular responses that 1534 jvi.asm.org Journal of Virology

Role of MDA5 in Resistance to TMEV FIG 2 Axonal damage in the brains of virus-infected MDA5 KO mice. Histopathology of the brains of TMEV-infected WT and MDA5 KO mice at 30 days postinfection was analyzed by hematoxylin and eosin staining (a and b), Luxol Fast Blue stain for myelin (c and d), and Bielschowsky silver stain for axons (e and f). In contrast to normal-appearing medulla oblongata (a; the inset presents an enlarged section [boxed area]) with intact myelination (c) and integrity of axons (e) in the brains of WT control mice, the brains of MDA5 KO mice displayed chronic inflammation in the white matter of the medulla oblongata, with elevated parenchymal and perivascular infiltration of mononuclear cells (b and inset) and focal demyelination (d), as well as axonal damage (f). lead to differential susceptibility to TMEV-induced demyelinating disease, the levels of cellular infiltration into the CNS of virusinfected MDA5 KO and WT control mice were compared (Fig. 4A and B). The CNS of MDA5 KO mice displayed higher levels of cellular infiltration, including both CD4 and CD8 T cells, as well as CD11b CD45hi macrophages. In contrast, there was no difference in the percentages of resident microglia (CD11b CD45int) in the CNS between MDA5 KO and WT mice. The expression of chemokines (CCL2, CCL5, and CXCL1) that are known to be involved in cellular infiltration into the CNS (46) also increased in MDA5 KO mice when assessed with quantitative realtime PCR at 8 days postinfection (Fig. 4C). However, the levels of type I IFNs were very low in virus-infected MDA5 KO mice, accompanying high viral message levels (Fig. 4D). Because type I IFNs are known to inhibit cellular infiltration and viral replication (2, 16), these results, together with increased chemokine levels, are consistent with increased cellular infiltration into the CNS. Interestingly, the level of IFN- 4 was significantly elevated in WT mice at 21 days, whereas there was no induction in MDA5 KO mice. February 2012 Volume 86 Number 3 These results imply that robust viral replication and elevated cellular infiltration are associated with a lack of IFN- induction in MDA5 KO mice. Therefore, MDA5 appears to play a role in TMEV-induced type I IFN production, controlling viral spreading and cellular infiltration into the CNS. MDA5 is required to control the levels and subtypes of antiviral CD4ⴙ T cells. To determine whether the levels of antiviral immune responses are affected in MDA5 KO mice, the levels of CD4 and CD8 T cells, as well as their associated cytokines, in the CNS and periphery of virus-infected WT and MDA5 KO mice were compared at 8 and 21 days postinfection (Fig. 5). Because IL-17-producing T cells may play a critical role in pathogenesis (13), we first compared IFN- and IL-17 mrna expression levels in the CNS at 8 and 21 days postinfection (Fig. 5A). The overall levels of both IFN- and IL-17 gene expression in the CNS were significantly higher, particularly at 8 days postinfection. The proportions of IFN- -producing antiviral CD4 T cells in the CNS of infected mice were similar in WT and MDA5 KO mice at 8 days postinfection, which was the peak of the immune responses, al- jvi.asm.org 1535

Jin et al. FIG 3 Focal demyelinated lesions and axonal damage in the spinal cords of virus-infected MDA5 KO mice. Adjacent sections of the lumbar regions of the spinal cords from virus-infected WT and MDA5 KO mice were compared at 30 days postinfection. (A, a and b) LFB staining. In contrast to no loss of myelin in the ventral white matter of the spinal cords of WT mice (a), multiple severe demyelinating lesions were observed in the same areas of the spinal cords from MDA5 KO mice (b). (c and d) Bielschowsky silver staining. The presence of axonal damage and loss in the demyelinating areas of the spinal cords from MDA5 KO mice (d) was observed, in contrast to WT mice (c). (e and f) H&E staining of boxed areas in a and b, respectively. Few infiltrating cells and activated glial cells were found in the spinal cords from WT mice (e). However, elevated infiltrations of inflammatory cells in the meninges and demyelinating areas of the spinal cords from MDA5 KO mice and highly activated glial cells in the demyelinating lesions were observed (f). (B) Severe demyelination was also detected in the dorsal columns of the lumbar spinal cords from MDA5 KO mice, as determined by LFB staining (b). However, no demyelination was seen in the spinal cords of control mice (a). Scale bars, 100 m (A, a, b, c, and d, and B, a and b) and 20 m (A, e and f). though they were somewhat lower in MDA5 KO mice at 21 days postinfection (Fig. 5B). In contrast, the proportion of IFN- producing antiviral CD8 T cells was significantly lower in MDA5 KO mice at the peak of the immune responses (Fig. 5C). Despite the lower proportions of antiviral T cell responses, the overall antiviral T cell levels were higher in the CNS of MDA5 KO mice due to the elevated T cell infiltration (Fig. 5D). These results collectively suggest that both Th1 and Th17 response levels are significantly higher in the CNS of virus-infected MDA5 KO mice in the early stage of viral infection, and the increase in the CD4 T cell levels may lead to the development of demyelinating disease. 1536 jvi.asm.org To further determine whether T cell responses are uniquely high in the CNS, we also assessed the levels of proliferation and cytokine production by virus-specific CD4 and CD8 T cells in the spleens of virus-infected mice (Fig. 5E). The levels of both IFN- and IL-17 production upon specific stimulation by both splenic CD4 and CD8 T cells from MDA5 KO mice were greater than by WT mice at 8 and 21 days postinfection. These results suggest that T cell levels in the CNS are regulated similarly to those in the periphery during early stages of infection. However, the ability of T cells in the CNS to produce cytokines wanes as viral persistence continues, which is in contrast to the steady cytokine Journal of Virology

Role of MDA5 in Resistance to TMEV FIG 4 Elevated cellular infiltration and expression of viral and cytokine genes in the CNS of MDA5 KO mice. CNS mononuclear cells from WT (129SvJ) and MDA5 KO mice (3 or 4 mice per group) were analyzed at 8 days postinfection by flow cytometry. The cells were gated on forward and side scatter plots. (A) Infiltrating lymphocytes were stained for the presence of CD4 and CD8 T cells. The numbers in the cytometric plots represent the percentages of CD4 and CD8 T cells. (B) The percentages of CD45 int CD11b, CD45 hi CD11b, and CD45 hi CD11b populations were analyzed. The gated populations represent microglial cells (CD45 int CD11b ), macrophages (CD45 hi CD11b ), and leukocytes (CD45 hi CD11b ). The data are representative of two independent experiments. (C) Total RNA isolated from the mixture of brains and spinal cords at day 8 postinfection was used to determine chemokine RNA levels. (D) Total RNAs isolated separately from the brains (BR) and spinal cords (SC) at days 8 and 21 postinfection were used to assess TMEV, IFN- 4, IFN-, and IL-1 message levels. Gene expression levels were analyzed by real-time quantitative PCR (qpcr) using specific primers. The GAPDH level was used as an internal control. The data shown are representative of two independent experiments. The error bars indicate standard deviations., P 0.001;, P 0.01;, P 0.05 (Student s t test). production by peripheral T cells. These differences may reflect a higher viral load in the CNS than in the periphery, where the virus is cleared. A localized high viral load in the CNS may lead to downregulation of T cell responses. Antigen-presenting cells in the CNS of MDA5 KO mice express lower levels of costimulatory molecules but higher levels of inhibitory molecules. Because the T cell responses in the CNS of virus-infected WT and MDA5 KO mice were different, we further assessed the expression of costimulatory molecules on CNS CD11b antigen-presenting cell populations containing microglia and macrophages (Fig. 5F). It is interesting that the expression of CD80, CD86, and CD40, representing important costimulatory molecules in T cell stimulation, was lower in MDA5 KO cells, although the expression of major histocompatibility complex (MHC) molecules was not compromised. These results are consistent with the lower proportions of functional T cells in the CNS of MDA5 KO mice (Fig. 5A and B). In contrast, the expression of PDL-1, which is known to downregulate T cell functions, was markedly elevated in the CNS CD11b cells from virus-infected MDA5 KO mice. These results may also explain the poor efficiency of CD8 T cell stimulation, particularly in MDA5 KO mice (Fig. 5C). TMEV-infected macrophages from MDA5 KO and WT mice similarly stimulate T cells. To determine the possibility that virus-infected MDA5 KO macrophages stimulate CD4 T cells more efficiently than WT macrophages, T cells from naïve OTII February 2012 Volume 86 Number 3 jvi.asm.org 1537

Jin et al. 1538 jvi.asm.org Journal of Virology

Role of MDA5 in Resistance to TMEV TCR-Tg (Fig. 6A), virus-infected WT (Fig. 6B), and MDA5 KO (Fig. 6C) mice were stimulated with WT or MDA5 KO macrophages infected for 24 h in vitro with TMEV. After stimulating IFN- -producing CD4 T cells from naïve OTII T cells with the cognate OVA 323-339 peptide and primed CD4 T cells from virusinfected WT and MDA5 KO mice with viral peptides, no differences were detected between virus-infected WT and MDA5 KO macrophages. However, MDA5 KO macrophages stimulated naïve CD4 T cells from OTII TCR-Tg mice and CD4 T cells from virus-infected MDA5 KO mice to produce a higher level of IL-17 than WT macrophages, whereas they failed to stimulate CD4 T cells from virus-infected WT mice (Fig. 6A to C). These results suggest that MDA5 signals may inhibit the development of Th17 responses from naïve CD4 T cells but may not be able to convert the precommitted CD4 Th1 cell subtype from virus-infected WT mice; therefore, virus-infected WT mice display a low-level Th17 response, but MDA5 KO mice mount a high-level response (Fig. 5E). To further understand the mechanisms underlying the discrepancies between T cells in the CNS of virus-infected mice and T cells stimulated by in vitro virus-infected antigen-presenting cells, the expression of costimulatory molecules on peritoneal macrophages infected in vitro with TMEV for 24 h was assessed (Fig. 6D). No differences in the expression levels of costimulatory and MHC molecules were detected. This observation is consistent with the T cell responses, which were not different when stimulated by either virus-infected WT or MDA5-deficient macrophages. However, the expression patterns of these molecules on virus-infected macrophages were different from those on CD11b antigenpresenting cells in the CNS of virus-infected mice, showing reduced levels (Fig. 5F). Therefore, the altered expression of costimulatory molecules on the CNS cells from virus-infected MDA5-deficient mice may exert accumulated effects of viral infection on T cell stimulation compared with the effects of in vitro short-term viral infection. MDA5 is required for the induction of IFN- expression in TMEV-infected cells. Because our previous study demonstrated that proinflammatory cytokines and type I IFN are induced following TMEV infection in glial cells, we further investigated the potential contribution of MDA5 to the induction of cytokines in virus-infected microglia and macrophages (Fig. 7). For these studies, the cytokine expression levels in the microglia and macrophages from the CNS of virus-infected mice at 8 days postinfection (Fig. 7A), as well as the cytokine levels in peritoneal macrophages (Fig. 7B) and microglia from uninfected mice upon TMEV infection in vitro (Fig. 7C), were determined. Microglia from infected MDA5 KO mice showed elevated viral and reduced IL-6 message levels. No differences in IFN- and IFN- levels were observed in the microglia from MDA5 KO mice compared to the microglia from control WT mice. However, CNS-infiltrating CD45 hi CD11b macrophages from MDA5 KO mice displayed a significantly higher level of viral messages and a dramatically lower level of IFN- messages than the counterpart controls (Fig. 7A). These results indicate that the cytokine profiles of microglia and macrophages in the CNS of infected mice are distinct. Peritoneal macrophages from MDA5 KO mice infected in vitro with TMEV also showed increased viral RNA with lower IFN- message levels. In addition, the IFN- message level was low in MDA5 KO macrophages, similar to the IFN- level (Fig. 7B), indicating the differences in the cytokine profiles between the macrophages in the CNS of virus-infected mice and the peripheral macrophages infected in vitro. MDA5 KO microglia infected with TMEV in vitro displayed lower IFN- and IFN- levels, similar to the macrophages from the mice (Fig. 7C). Interestingly, uninfected microglia from MDA5 KO mice uniquely showed a significantly high level of transforming growth factor beta (TGF- ), although the level was reduced to the background, similar to WT microglia when infected with virus. Primary astrocytes and oligodendrocytes derived from new neonatal brains of MDA5 KO and WT mice were also infected with TMEV or stimulated with TLR ligands (data not shown). The expression of IFN- 4 in response to TMEV infection was completely abrogated in MDA5 KO cells compared with WT cells, similar to virus-infected microglia and peritoneal macrophages (Fig. 7B and C). In addition, we observed a reduction of IFN- expression in both the astrocytes and oligodendrocytes from MDA5 KO mice, suggesting that an MDA5 deficiency may also affect the TMEV-induced expression of IFN-. However, stimulation with LTA (TLR2) and poly(i C) (TLR3) induced nearly the same levels of cytokine expression in both MDA5 KO and WT cells, indicating that MDA5 KO cells exhibit uncompromised responses to these TLR ligands. These results confirm that MDA5 plays a critical role in the production of TMEV-induced IFN- 4 and IFN- but minor roles in the production of other cytokines in many different types of cells. DISCUSSION Induction of type I IFNs is a hallmark of viral infection. Type I IFNs mediate antiviral function in the infected host, including activation of other antiviral genes, such as interferon-stimulated genes, and upregulation of MHC and costimulatory molecules associated with T cell stimulation (2, 15, 26, 38). RNA viruses, including TMEV, generate dsrna in the host cytoplasm during their replication cycle. Newly made dsrna following TMEV infection is recognized by MDA5 and TLR3, and this interaction triggers innate immune responses, including the production of type I IFNs (20, 42). We have recently shown that 129S2/SvPasCrl mice lacking type I interferon receptors develop rapid fatal en- FIG 5 Regulation of antiviral T cell response levels by MDA5 signals. (A to E) CD4 and CD8 T cell responses in the CNS (A to D) and periphery (E) of WT and MDA5 KO mice (3 mice per group) were compared at 8 and 21 days postinfection. (A) Real-time PCR analysis of IFN- and IL-17A messages in the brains and spinal cords of virus-infected WT and MDA5 KO mice. (B) Proportions of IFN- -producing CNS CD4 T cells upon stimulation with a mixture of viral CD4 T cell epitopes or anti-cd3/cd28 antibodies. (C) Proportion of IFN- -producing CNS CD8 T cells upon stimulation with viral CD8 T cell epitopes or anti-cd3/cd28 antibodies. (D) Numbers of overall virus-specific IFN- -producing CD4 and CD8 T cells in the CNS of virus-infected control WT and MDA5 KO mice at 8 and 21 days postinfection. A representative of three separate results is shown. (E) T cell proliferative responses and cytokine production by the periphery (spleens) of virus-infected WT and MDA5 KO mice upon stimulation with viral epitopes. (F) Expression levels of costimulatory and MHC molecules by CNS CD11b APCs from infected WT and MDA5 KO mice at 8 days postinfection. The numbers on the cytometric plots represent the percentages of CD11b cells that express the indicated molecules out of the total number of CNS CD11b cells. The error bars indicate standard deviations., P 0.001;, P 0.01;, P 0.05 (Student s t test). February 2012 Volume 86 Number 3 jvi.asm.org 1539

Jin et al. FIG 6 There is no difference between WT and MDA5 KO macrophages infected with virus in vitro for CD4 T cell stimulation, except IL-17 production. Isolated CD4 T cells were cultured with peritoneal macrophages from either virus-infected WT or MDA5 KO mice for 4 days in the presence of PBS or CD4 T cell epitope peptides. T cell proliferative responses and cytokine production (IFN- and IL-17) of the cultures were analyzed using [ 3 H]thymidine (TdR) uptake and specific ELISAs, respectively. (A) Isolated CD4 T cells (1.3 10 5 ) from naïve OTII TCR-Tg mice were incubated with isolated peritoneal macrophages (1.3 10 4 ) from virus-infected WT 129SvJ or MDA5 KO mice at 7 days postinfection in the presence of PBS, OVA peptide, or OVA protein. (B) Isolated CD4 T cells (2.5 10 5 ) from virus-infected WT mice at 7 days postinfection were incubated with isolated peritoneal macrophages (1.3 10 4 ) from virus-infected WT or MDA5 mice at 7 days postinfection in the presence of PBS or CD4 T cell epitope peptides. (C) Isolated CD4 T cells (2.5 10 5 ) from virus-infected MDA5 KO mice at 7 days postinfection were incubated with isolated peritoneal macrophages (1.3 10 4 ) from virus-infected WT or MDA5 mice at 7 days postinfection in the presence of PBS or CD4 T cell epitope peptides. (D) Expression levels of costimulatory and MHC molecules on peritoneal macrophages from naïve WT and MDA5 KO mice infected in vitro with TMEV (MOI 10) for 24 h were determined using flow cytometry. The numbers on the cytometric plots represent the percentages of CD11b cells that were positive for the indicated molecules out of the total number of CD11b cells in the peritoneal exudate cell preparation. The error bars indicate standard deviations., P 0.001;, P 0.01;, P 0.05 (Student s t test). cephalitis accompanied by high viral loads and elevated cellular infiltration into the CNS but a low efficiency for T cell activation (16). Therefore, it would be interesting to assess the effects of the innate cytokines associated with MDA5 signals on the development of demyelinating disease and the stimulation of various antiviral immune responses during viral infection. In this study, we investigated the potential role of MDA5- mediated signals in the pathogenesis of TMEV-induced demyeli- 1540 jvi.asm.org Journal of Virology

Role of MDA5 in Resistance to TMEV FIG 7 MDA5 signals are critical for TMEV-induced type I IFNs in microglia and macrophages. (A) Microglia (MG) (CD45 int CD11b ) and infiltrating macrophages (MP) (CD45 hi CD11b ) were isolated from TMEV-infected WT and MDA5 KO mice at 8 days postinfection using MoFlo, and the message levels were analyzed by qpcr. (B) Peritoneal macrophages from WT and MDA5 KO mice were infected with TMEV (MOI 10) for 24 h. The message levels were then analyzed by qpcr. (C) Microglial populations isolated from naïve WT and MDA5 KO mice using a Percoll gradient (7) were incubated with PBS or TMEV (MOI 10) for 24 h. The message levels of virus and cytokines were similarly analyzed using qpcr. The GAPDH level was used as an internal control. The data shown are representative of two independent experiments., P 0.001;, P 0.01;, P 0.05 (Student s t test). The error bars indicate standard deviations. nating disease using MDA5 KO mice on the resistant 129X1/SvJ background. Our results indicate that MDA5 is critical for the production of type I IFNs in the CNS, IFN- in particular (Fig. 4), and plays a protective role in the development of TMEV-induced demyelinating disease (Fig. 1A). These results strongly suggest that an MDA5 deficiency accompanied by suboptimal levels of type I IFN production compromises viral clearance (Fig. 1) and regulation of immune responses (Fig. 5) and elevates cellular infiltration into the CNS (Fig. 4). These combined effects may contribute to the development of demyelinating disease in MDA5 KO mice. It was recently reported that murine norovirus type 1 replication is higher in MDA5 KO mice and in MAD5-deficient bone marrow-derived dendritic cells, indicating that MDA5 signals are involved in limiting viral replication (28). It is interesting that IFN- production following TMEV infection is particularly compromised in MDA5-deficient mice (Fig. 4 and 7), although type I IFN production is also induced via TLR3-mediated signals (42), which are intact in these mice. These results, therefore, suggest that an MDA5-mediated signal is the primary activator for type I IFN production in mice following infection with TMEV. A recent study showed that MDA5 KO mice infected with coxsackie B virus, another picornavirus, display impaired type I IFN production and early mortality (45). The reduced type I IFN production is consistent with our results with TMEV infection. However, we did not observe a similar fatal disease after TMEV infection, implying that the effects of MDA5 deficiency may be different depending on the virus. It is well known that IRF3 and IRF7 are essential for the induction of type 1 IFNs after viral infection, although they appear to play distinct roles (11). IRF3 appears to be involved in the initial production of mainly IFN-, which is required for further transcription of IRF7 and large amounts of type I IFNs. The presence of IRF7 is essential for IFN- gene expression, because IRF7- deficient cells completely failed to express it, but the expression of IFN- mrna was only partially compromised. Interestingly, IRF3/IRF7 doubly deficient cells completely failed to induce the gene expression of IFN- (12). Furthermore, it has recently been shown with West Nile virus (WNV) infection that IFN- production is mainly controlled by IRF7 (4). Thus, it appears that IFN- is controlled by IRF3/IRF7 cooperatively but IFN- is mainly induced by IRF7. Our previous data showed that TMEV infection induces IRF3-interferon-stimulated response element binding activity through a TLR3-mediated signaling pathway (42), suggesting that gene expression of IFN- following TMEV infection is partially regulated by TLR3, possibly in cooperation with MDA5. In contrast, IFN- 4 expression seems to be activated mainly by MDA5-IRF7-mediated signaling, as MDA5 KO mice produced drastically reduced levels of IFN- in many different cell types February 2012 Volume 86 Number 3 jvi.asm.org 1541

Jin et al. (Fig. 4D and 7). Additionally, virus-induced IRF7 gene transcription and activation are significantly lower in MDA5 KO cells (data not shown). These results collectively suggest that the viral load in the CNS of infected MDA5 KO mice increases due to the initial low levels of type I IFNs, in particular IFN-. The elevated viral load may subsequently activate the production of IFN- via the TLR3 and/or PKR pathway to a level similar to the levels of control WT mice. Many of the observations in virus-infected MDA5-deficient mice appear to reflect insufficient production of type I IFNs, yet enough production to prevent the fatal encephalitis seen in type I IFN receptor (IFNR)-deficient mice (16). For example, higher cellular infiltration into the CNS (Fig. 4) and reduced expression of costimulatory molecules in virus-infected MDA5 KO mice (Fig. 5F) may be attributable to insufficient levels of type I IFNs (10, 30, 31, 40, 44). Although the efficiency of T cell stimulation appears to be somewhat compromised, the reduction of costimulatory molecule expression does not affect the T cell response levels. However, the overall levels of T cells were increased due to the increased cellular infiltration into the CNS, reflecting low levels of type I IFNs that control brain blood barrier permeability. The expression of costimulatory molecules on MDA5-deficient macrophages infected with TMEV in vitro was largely intact (Fig. 6D). Therefore, the T cell-activating function of APCs from MDA5- deficient mice would not be compromised. These results suggest that the levels of type I IFN signals induced by other receptors may be sufficient to activate T cell responses in the absence of MDA5 signals. Relatively intact IFN- levels via TLR3 and PKR in MDA5-deficient mice (Fig. 4) may be sufficient for such T cell stimulation machinery. Taken together, the results from our current study strongly suggest that an overall optimal level of type I IFNs induced following viral infection is critically important in controlling the initial levels of viral load, cellular infiltration, and T cell subtype development, which ultimately affect the establishment of chronic viral persistence and the development of viral immune-mediated inflammatory disease. REFERENCES 1. Barnden MJ, Allison J, Heath WR, Carbone FR. 1998. Defective TCR expression in transgenic mice constructed using cdna-based alpha- and beta-chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76:34 40. 2. Biron CA. 2001. Interferons alpha and beta as immune regulators a new look. Immunity 14:661 664. 3. Carpentier PA, Williams BR, Miller SD. 2007. Distinct roles of protein kinase R and toll-like receptor 3 in the activation of astrocytes by viral stimuli. Glia 55:239 252. 4. Daffis S, Samuel MA, Suthar MS, Gale M Jr, Diamond MS. 2008. Toll-like receptor 3 has a protective role against West Nile virus infection. J. Virol. 82:10349 10358. 5. Dal Canto MC, Kim BS, Miller SD, Melvold RW. 1996. Theiler s murine encephalomyelitis virus (TMEV)-induced demyelination: a model for human multiple sclerosis. Methods 10:453 461. 6. Dal Canto MC, Melvold RW, Kim BS. 1995. A hybrid between a resistant and a susceptible strain of mouse alters the pattern of Theiler s murine encephalomyelitis virus-induced white matter disease and favors oligodendrocyte-mediated remyelination. Mult. Scler. 1:95 103. 7. Ford AL, Goodsall AL, Hickey WF, Sedgwick JD. 1995. Normal adult ramified microglia separated from other central nervous system macrophages by flow cytometric sorting. Phenotypic differences defined and direct ex vivo antigen presentation to myelin basic protein-reactive CD4 T cells compared. J. Immunol. 154:4309 4321. 8. Fredericksen BL, Gale M Jr. 2006. West Nile virus evades activation of interferon regulatory factor 3 through RIG-I-dependent and -independent pathways without antagonizing host defense signaling. J. Virol. 80:2913 2923. 9. Gitlin L, et al. 2006. Essential role of mda-5 in type I IFN responses to polyriboinosinic:polyribocytidylic acid and encephalomyocarditis picornavirus. Proc. Natl. Acad. Sci. U. S. A. 103:8459 8464. 10. Grayson MH, et al. 2007. Controls for lung dendritic cell maturation and migration during respiratory viral infection. J. Immunol. 179:1438 1448. 11. Hiscott J. 2007. Triggering the innate antiviral response through IRF-3 activation. J. Biol. Chem. 282:15325 15329. 12. Honda K, Taniguchi T. 2006. IRFs: master regulators of signalling by Toll-like receptors and cytosolic pattern-recognition receptors. Nat. Rev. Immunol. 6:644 658. 13. Hou W, Kang HS, Kim BS. 2009. Th17 cells enhance viral persistence and inhibit T cell cytotoxicity in a model of chronic virus infection. J. Exp. Med. 206:313 328. 14. Hou W, So EY, Kim BS. 2007. Role of dendritic cells in differential susceptibility to viral demyelinating disease. PLoS Pathog. 3:e124. 15. Huber JP, Farrar JD. 2011. Regulation of effector and memory T-cell functions by type I interferon. Immunology 132:466 474. 16. Jin YH, et al. 2010. Type I interferon signals control Theiler s virus infection site, cellular infiltration and T cell stimulation in the CNS. J. Neuroimmunol. 226:27 37. 17. Kang DC, et al. 2004. Expression analysis and genomic characterization of human melanoma differentiation associated gene-5, mda-5: a novel type I interferon-responsive apoptosis-inducing gene. Oncogene 23: 1789 1800. 18. Kang DC, et al. 2002. mda-5: an interferon-inducible putative RNA helicase with double-stranded RNA-dependent ATPase activity and melanoma growth-suppressive properties. Proc. Natl. Acad. Sci. U. S. A. 99: 637 642. 19. Kang MH, So EY, Park H, Kim BS. 2008. Replication of Theiler s virus requires NF-kappaB-activation: higher viral replication and spreading in astrocytes from susceptible mice. Glia 56:942 953. 20. Kato H, et al. 2006. Differential roles of MDA5 and RIG-I helicases in the recognition of RNA viruses. Nature 441:101 105. 21. Kawai T, Akira S. 2007. Antiviral signaling through pattern recognition receptors. J. Biochem. 141:137 145. 22. Kawai T, et al. 2005. IPS-1, an adaptor triggering RIG-I- and Mda5- mediated type I interferon induction. Nat. Immunol. 6:981 988. 23. Kim BS, Palma JP, Kwon D, Fuller AC. 2005. Innate immune response induced by Theiler s murine encephalomyelitis virus infection. Immunol. Res. 31:1 12. 24. Kumar H, et al. 2006. Essential role of IPS-1 in innate immune responses against RNA viruses. J. Exp. Med. 203:1795 1803. 25. Kwon D, Fuller AC, Palma JP, Choi IH, Kim BS. 2004. Induction of chemokines in human astrocytes by picornavirus infection requires activation of both AP-1 and NF-kappa B. Glia 45:287 296. 26. Liu SY, Sanchez DJ, Cheng G. 2011. New developments in the induction and antiviral effectors of type I interferon. Curr. Opin. Immunol. 23:57 64. 27. Loo YM, et al. 2008. Distinct RIG-I and MDA5 signaling by RNA viruses in innate immunity. J. Virol. 82:335 345. 28. McCartney SA, et al. 2008. MDA-5 recognition of a murine norovirus. PLoS Pathog. 4:e1000108. 29. Mogensen TH, Paludan SR. 2005. Reading the viral signature by Toll-like receptors and other pattern recognition receptors. J. Mol. Med. 83: 180 192. 30. Montoya M, et al. 2002. Type I interferons produced by dendritic cells promote their phenotypic and functional activation. Blood 99:3263 3271. 31. Niemela J, et al. 2008. IFN-beta regulates CD73 and adenosine expression at the blood-brain barrier. Eur. J. Immunol. 38:2718 2726. 32. Olson JK, Girvin AM, Miller SD. 2001. Direct activation of innate and antigen-presenting functions of microglia following infection with Theiler s virus. J. Virol. 75:9780 9789. 33. Ozden S, Tangy F, Chamorro M, Brahic M. 1986. Theiler s virus genome is closely related to that of encephalomyocarditis virus, the prototype cardiovirus. J. Virol. 60:1163 1165. 34. Palma JP, Kim BS. 2001. Induction of selected chemokines in glial cells infected with Theiler s virus. J. Neuroimmunol. 117:166 170. 35. Palma JP, Kwon D, Clipstone NA, Kim BS. 2003. Infection with Theiler s murine encephalomyelitis virus directly induces proinflammatory cytokines in primary astrocytes via NF-kappaB activation: potential role for the initiation of demyelinating disease. J. Virol. 77:6322 6331. 36. Pevear DC, Calenoff M, Rozhon E, Lipton HL. 1987. Analysis of the 1542 jvi.asm.org Journal of Virology