Site-Specific Production of IL-6 in the Central Nervous System Retargets and Enhances the Inflammatory Response in Experimental Autoimmune This information is current as Encephalomyelitis of September 26, 2021. Albert Quintana, Marcus Müller, Ricardo F. Frausto, Raquel Ramos, Daniel R. Getts, Elisenda Sanz, Markus J. Hofer, Marius Krauthausen, Nicholas J. C. King, Juan Hidalgo and

Iain L. Campbell Downloaded from J Immunol 2009; 183:2079-2088; Prepublished online 13 July 2009; doi: 10.4049/jimmunol.0900242 http://www.jimmunol.org/content/183/3/2079 http://www.jimmunol.org/

Supplementary http://www.jimmunol.org/content/suppl/2009/07/14/jimmunol.090024 Material 2.DC1 References This article cites 50 articles, 15 of which you can access for free at: http://www.jimmunol.org/content/183/3/2079.full#ref-list-1 by guest on September 26, 2021

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Site-Specific Production of IL-6 in the Central Nervous System Retargets and Enhances the Inflammatory Response in Experimental Autoimmune Encephalomyelitis1

Albert Quintana,2,3*§ Marcus Mu¨ller,2,4* Ricardo F. Frausto,* Raquel Ramos,§ Daniel R. Getts,5† Elisenda Sanz,3*§ Markus J. Hofer,6* Marius Krauthausen,4* Nicholas J. C. King,†‡ Juan Hidalgo,§ and Iain L. Campbell7*‡

IL-6 is crucial for the induction of many murine models of autoimmunity including experimental autoimmune encephalomyelitis (EAE), an animal model for multiple sclerosis. To establish the role of site-specific production of IL-6 in autoimmunity, we examined myelin oligodendrocyte glycoprotein immunization-induced EAE in transgenic mice (GFAP-IL6) with IL-6 production restricted to the cerebellum. Myelin oligodendrocyte glycoprotein-immunized (Mi-) GFAP-IL6 mice developed severe ataxia but Downloaded from no physical signs of spinal cord involvement, which was in sharp contrast to Mi-wild type (WT) animals that developed classical EAE with ascending paralysis. Immune pathology and demyelination were nearly absent from the spinal cord, but significantly increased in the cerebellum of Mi-GFAP-IL6 mice. Tissue damage in the cerebellum in the Mi-GFAP-IL6 mice was accompanied by increased total numbers of infiltrating leukocytes and increased proportions of both neutrophils and B-cells. With the exception of IL-17 mRNA, which was elevated in both control immunized and Mi-GFAP-IL6 cerebellum, the level of other and mRNAs were comparable with Mi-WT cerebellum whereas significantly higher levels of IFN-␥ and TNF-␣ mRNA http://www.jimmunol.org/ were found in Mi-WT spinal cord. Thus, site-specific production of IL-6 in the cerebellum redirects trafficking away from the normally preferred antigenic site the spinal cord and acts as a leukocyte “sink” that markedly enhances the inflammatory cell accumulation and disease. The mechanisms underlying this process likely include the induction of specific , activation of microglia, and activation and loss of integrity of the blood-brain barrier present in the cerebellum of the GFAP-IL6 mice before the induction of EAE. The Journal of Immunology, 2009, 183: 2079–2088.

nterleukin-6 is a plurifunctional cytokine that is implicated in is crucial for the induction of immune pathology and clinical disease

the pathogenesis of numerous autoimmune disorders in humans, in collagen-induced (3) or spontaneous (4) arthritis and in experimen- by guest on September 26, 2021 I including rheumatoid arthritis (1) and multiple sclerosis (MS)8 tal autoimmune encephalomyelitis (EAE; Ref. 5, 6), which are murine (2). Experimental studies in IL-6 knockout (KO) mice show that IL-6 models for the human autoimmune disorders rheumatoid arthritis and MS, respectively. Recent studies have identified a subpopulation of CD4ϩ Th cells, namely Th17 T-cells, that are key Ag-specific effector *School of Molecular and Microbial Biosciences, †Department of Pathology and ‡ § T cells in these murine models (4, 7–9). The demonstration that the Bosch Institute, University of Sydney, Australia; and Institute of Neurosciences and ␤ Department of Cellular Biology, Physiology and Immunology, Autonomous Univer- formation of Th17 T cells is driven by IL-6 together with TGF- in sity Barcelona, Spain vitro and in vivo (7, 10–12) explains, in part, why the murine auto- Received for publication January 22, 2009. Accepted for publication June 1, 2009. immune models are critically dependent on IL-6. However, in the case The costs of publication of this article were defrayed in part by the payment of page of EAE, studies to pin down the precise role of IL-6 in the induction charges. This article must therefore be hereby marked advertisement in accordance of EAE have failed to identify a unifying mechanism. One study with 18 U.S.C. Section 1734 solely to indicate this fact. suggested that IL-6 is necessary for the induction of cerebrovascular 1 This work was supported by a New South Wales Spinal Cord Injury and Other Neu- adhesion molecules such as VCAM-1, which are crucial for leuko- rological Conditions Project Grant and a start up grant from the University of Sydney (to I.L.C.); and by SAF2008–00435, FP6 Integrated Project Exgenesis (LSHM-CT-2004– cyte trafficking to the CNS in EAE (13); a second study con- 005272), and RETICS (REEM, RD07/0060/0002) to J.H. D.G. was supported by an cluded that local, perivascular-produced IL-6 is needed for the Australian Postgraduate Award. M.J.H. and M.M. were postdoctoral fellows of the Deut- sche Forschungsgemeinschaft HO3298/1–1 and Mu17–07/3–1, respectively. M.M. was induction of EAE and not peripheral IL-6 (5), while a third also supported by the fund “Innovative Medical Research” of the University of Mu¨nster study proposed that IL-6 is required for autoantigen-reactive Medical School, Germany. A.Q. and E.S. were predoctoral fellows of the Ministerio de CD4ϩ-T cell proliferation during EAE (6). Finally, in a more Educacio´n y Ciencia, BES-2003–2427 and AP2003–0534, respectively. R.F. was sup- ported by a University of Sydney International Research Scholarship. R.R. was a pre- recent report IL-6 was shown to be a dominant inhibitor of the doctoral fellow of Generalitat de Catalunya 2006FI-01355. conversion of conventional T cells to FoxP3-positive T regulatory 2 A.Q. and M.M. contributed equally to this work. (reg) cells with the loss of IL-6 resulting in an overwhelming T reg 3 Current address: Department of Biochemistry (A.Q.) and Department of Pharma- cology (E.S.), University of Washington, Seattle, WA. 4 Current address: Department of Neurology, University of Bonn, Germany. 5 Current address: Department of Microbiology-Immunology, Northwestern Univer- 8 Abbreviations used in this paper: MS, multiple sclerosis; KO, knockout; EAE, sity, Chicago, IL. experimental autoimmune encephalomyelitis; T reg, regulatory T cell; MOG, my- 6 elin oligodendrocyte glycoprotein; BBB, blood-brain barrier; WT, wild type; Current address: Department of Neuropathology, University of Marburg, Germany. LFB, luxol fast blue; RPA, RNase protection assay; Ni, nonimmunized; Ci, BSA- 7 Address correspondence and reprint requests to Dr. Iain L. Campbell, School of CFA-immunized; Mi, MOG-CFA-immunized. Molecular and Microbial Biosciences G08, University of Sydney, New South Wales, 2006, Australia. E-mail address: [email protected] Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900242 2080 SITE-SPECIFIC PRODUCTION OF IL-6 TARGETS AUTOIMMUNITY

response that prevents the induction of Th1 and Th17 effector cells (3 mg/ml) and 100 ␮l CFA (Sigma-Aldrich) supplemented with 4 mg/ml in EAE (14). Mycobacterium tuberculosis H37RA (Difco). In addition, animals received During an autoimmune response, the production of IL-6 is an i.p. injection of 500 ng pertussis toxin (Sigma-Aldrich), which was repeated 2 days after immunization. As controls, another cohort of mice known to arise from a large number of different cell types that from each genotype (n ϭ 4) was immunized similarly with the exception include infiltrating leukocytes as well as tissue-resident cells (15). that MOG peptide was replaced with BSA. Adoptive transfer experiments in a spontaneous arthritis model (4) Immunized animals were examined daily and scored for EAE. Initially, as well as in EAE (5) suggest that the local cytokine milieu may we observed that GFAP-IL6 mice showed reduced classical clinical signs (which typically reflect spinal cord involvement), while displaying non- also play an important role in modulating the development and classical clinical signs of ataxia. To better characterize the clinical evolu- progression of disease. Due to the plurifunctional nature of IL-6, it tion of the disease in the two genotypes used, an elaborated general clinical is likely that in addition to its potential to control the activity of score that included the prototypical physical signs of EAE and an addi- infiltrating self-reactive Th17 T cells, that this cytokine may in- tional score for ataxia was used. The general clinical score was assessed for ϭ fluence the pathogenesis of autoimmunity at multiple levels, e.g., each animal according to the following criteria: 0 no signs of disease, 1 ϭ loss of tail tonus, 1.5 ϭ moderate hind limb paraparesis, 2 ϭ severe via the regulation of adhesion molecule expression on vascular hind limb paraparesis, 3 ϭ hind limb paralysis, 4 ϭ tetraplegia, and 5 ϭ endothelium (16). However, the degree to which localized, tissue- death. For each animal, we determined the time to disease onset (clinical specific production of IL-6 may influence the development of au- score Ն1 or severe ataxia Ն1), time to peak disease, peak-score, cumula- toimmune pathology within specialized tissues such as the CNS tive score (sum of all scores from disease onset to day 50), outcome (final score at day 50), and grade of remission (difference between peak score and and the mechanisms that are involved remain largely unknown. outcome). To score for ataxia, a previously reported protocol (22) was used To begin to address these outstanding issues, in the present study as follows: ataxic signs were scored using a cumulative scale of 4 points, we investigated myelin oligodendrocyte glycoprotein (MOG) immu- giving 1 point to each of these four physical signs: splayed legs, dragging Downloaded from nization-induced EAE in a unique transgenic mouse model with weight on the trunk rather than on the legs, wobbling, and falling from side CNS-restricted, astrocyte-targeted production of IL-6. The so- to side. called GFAP-IL6 transgenic mouse has been characterized exten- Histology sively (17–21) and shown to develop a progressive inflammatory At peak disease, mice were euthanized under halothane anesthesia and the encephalopathy with increased adhesion molecule expression by brain and spinal cord removed immediately and fixed in 4% paraformal-

the cerebrovascular endothelium, loss of integrity of the blood- dehyde. After embedding in paraffin, sections (5-␮m) were cut and depar- http://www.jimmunol.org/ brain barrier (BBB), reactive astrocytosis and microgliosis, and affinized in xylene and rehydrated through a series of graded ethanol so- induction of a number of acute phase response including lutions. For some processed sections, pretreatment was performed with ␮ alpha1-antichymotrypsin and complement C3. Thus, this model proteinase K (Sigma-Aldrich, 10 g/ml, 15 min at 37°C) to unmask Ag. Slides were then incubated for1hatroom temperature with primary Abs afforded us the opportunity to examine the induction and evolution (anti-CD3) or biotinylated tomato lectin (Sigma-Aldrich). After washing in of an autoimmune response to a target tissue in which the micro- PBS, a biotinylated secondary Ab (Vector Laboratories; 1/200, 45 min) and environment contained and was modified by IL-6. HRP-coupled streptavidin (Vector Laboratories; 1/200, 30 min) was used. Nova red (Vector Laboratories) was applied as the immunoperoxidase sub- strate according to the manufacturer’s instructions. Sections were counter- Materials and Methods stained with hematoxylin (Sigma-Aldrich). Animals Dual-label staining on paraffin sections was performed using an addi- by guest on September 26, 2021 tional primary Ab of a different species, which was detected with a sec- Construction and characterization of the GFAP-IL6 transgenic mice with ondary alkaline phosphatase-coupled Ab and Vector blue (Vector Labora- astrocyte-targeted production of IL-6 was described previously (19). The tories) as the second color substrate. Counterstaining was omitted for dual GFAP-IL6 transgenic and control nontransgenic littermates used in this label immunohistochemistry. For apoptotic cell counting a TUNEL assay study were produced by breeding of GFAP-IL6 (C57BL/6 strain) mice (Roche) was performed on paraffin slides according to the manufacturer’s hemizygous for the transgene with C57BL/6 wild type (WT) mice. All instructions. mice were kept under constant temperature and with free access to food and water under pathogen-free conditions in the Blackburn animal facility Analysis of demyelination and cell counting of the University of Sydney. Ethical approval for the use of all mice in this study was obtained from the University of Sydney Animal Care and Ethics Sections were stained either with H&E for routine histological analysis or Committee. with luxol fast blue (LFB) for myelin evaluation. Conventional stained sections were examined with a DM4000B bright field and fluorescence Determination of IL-6 levels in different brain regions microscope (Leica). Bright field images were acquired using a Leica DFC480 camera and Leica Firecam 1.7.1 software (Leica). To quantify GFAP-IL6 and WT littermate mice (n ϭ 4 for both genotype groups) at 12 infiltrated areas of longitudinal spinal cord sections, images of H&E stains wk of age were euthanized and the cerebrum, cerebellum, and spinal cord were examined with Image analysis software AnalySIS 3.0 (Softimaging). were dissected on ice and each lysed and homogenized in complete RIPA The total area of the spinal cord cross-section and infiltrated areas were lysis buffer (100 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% measured by an observer blinded to the sample identity. Histological anal- deoxycholic acid, 1% Triton X-100, 0.1% SDS, 2 mM PMSF, 50 mM yses were performed on three sections (50 ␮m apart) of each brain from NaF, 1 mM NaVanadate, Protease Inhibitor Cocktail Set III (Calbiochem- three mice per group. Merck), and Phosphatase Inhibitor Cocktail Set II (Calbiochem-Merck). The homogenates were centrifuged at 14,000 rpm for 20 min at 4°C. The RNase protection assay (RPA) in the cleared lysates was quantified using the BCA Protein Assay Kit (Thermo Scientific). The cerebellum and spinal cord were collected (as described above) and IL-6 protein was measured using a Ready-Set-Go IL-6 ELISA Kit frozen in liquid nitrogen pending isolation of the RNA. Total RNA was (eBioscience) performed as per the manufacturer’s instructions. The levels isolated using TRIzol (Sigma-Aldrich). For RPA (see below), 5 ␮g of total were corrected for the amount of total protein added (ϳ100 ␮g) to the RNA was used in each analysis. wells. The construction and characterization of the multiprobe sets used in the RPAs, namely for host-response (iNOS, Mac1, A20, ICAM 1, EB22/5, Induction of EAE and clinical evaluation GFAP), and for various and chemokines (TNF-␣, TGF-␤, CCL- 12/MCP-5, IFN-␥, IL-6, RANTES/CCL-5), has been described elsewhere For the induction of EAE (see below), 3-mo-old male (n ϭ 16 WT, n ϭ 10 (20, 23, 24). GFAP-IL6) and female (n ϭ 10 WT, n ϭ 17 GFAP-IL6) mice were used. GFAP-IL6 mice at this age did not show any physical signs of neurological IL-17 mRNA determination by quantitative real time-PCR disease. An additional cohort of WT and GFAP-IL6 (n ϭ 3 per group) were used as nonimmunized controls. EAE was induced by active immunization Total RNA isolated as described above was reverse-transcribed into cDNA

with MOG35–55 peptide. On day 0, GFAP-IL6 and WT animals were in- by using SuperScript II Reverse Transcriptase (Invitrogen). Real-time ␮ jected s.c. into the hind flanks with an emulsion of 100 l MOG35–55 quantitative PCR assays were performed using SYBRgreen. The composition The Journal of Immunology 2081

of the reaction mixture was as follows: 2 ␮l of cDNA corresponding to 300 ng of total RNA, 100 nM of each primer, 1 ϫ SYBR Green PCR Master Mix (Applied Biosystems) in a total volume of 25 ␮l. Samples were analyzed simultaneously for GAPDH mRNA as the internal control, and quantitative expression values were extrapolated from separate standard curves. Each sam- ple was assayed in triplicate and normalized to GAPDH. Primer sequences used to amplify the IL-17A and GAPDH cDNA were: IL-17 sense 5Ј-AAG GCAGCAGCGATCATCC-3Ј and IL-17 anti-sense 5Ј-GGAACGGTTGAG GTAGTCTGAG-3Ј; GAPDH sense 5Ј-TCACCAGGGCTGCCATTTGC-3Ј and GAPDH anti-sense 5Ј-GACTCCACGACATACTCAGC-3Ј. Flow cytometry analysis

The cerebellum from animals with peak disease was excised and placed in FIGURE 1. IL-6 production in GFAP-IL6 mice was restricted to the ␮ ice-cold PBS buffer solution before extrusion through a coarse 100- m cerebellum. The IL-6 content in lysates from different CNS regions, metal cell strainer. The subsequent homogenate was digested for 60 min in namely cerebrum, cerebellum, and spinal cord, of WT and GFAP-IL6 mice PBS with deoxyribonuclease (0.005 g/ml; Sigma-Aldrich) and collagenase IV (0.05 g/ml Sigma-Aldrich). Digested samples were incubated at 37°C in was analyzed by ELISA as described in the Materials and Methods section. Significant levels of IL-6 were detected only in cerebellum of GFAP-IL6 a humidified atmosphere of 5% CO2 and were resuspended every 15 min. .(p Ͻ 0.001 vs WT ,ء ,Digestion was stopped with 10% FCS and the homogenate was passed mice (for statistical significance through a 70-␮m cell strainer (BD Biosciences). The pellet obtained after 10 min centrifugation at 340 ϫ g was dissolved in 30% Percoll (GE Healthcare). Subsequently, the 30% Percoll homogenate mix was layered mice showed a mild ataxia that developed late (ϳ45 days) fol- Downloaded from over 80% Percoll. Leukocytes were collected from the 30%/80% interface lowing immunization (Fig. 2B). As expected, MOG-CFA-im- ϫ after centrifugation (1140 g) for 25 min at room temperature. The col- munized (Mi)-WT mice showed a typical EAE clinical course, lected cells were washed in PBS and blocked with CD16/CD32 (BD Bio- sciences) Ab. Viable cells were counted using trypan blue exclusion. Iso- with an onset of initial signs at 8.5 days, a peak of clinical signs lated leukocytes were incubated with fluorochrome-conjugated Abs to around 15 days postimmunization, followed by a gradual re- detect CD8, CD4, Ly6G/GR1, CD11b, CD11c, CD45, B220, and the ap- duction in clinical severity with incomplete remission at 50 propriate isotype controls (all from BD Biosciences). Samples were run on days postimmunization (Fig. 2, A and C). However, and in con- a FACSAria (BD Biosciences) and analyzed with Flow Jo (Tree Star). http://www.jimmunol.org/ trast to Mi-WT mice, Mi-GFAP-IL6 mice showed significantly Statistical analysis reduced signs of classical EAE over the entire period of study, For all clinical evaluations, a nonparametric U test was performed fol- but showed an earlier onset of disease in the form of ataxia at lowing the recommendations described elsewhere (25). For histologi- 7.1 days (Fig. 2C). The Mi-GFAP-IL6 mice did not develop hind cal, IL-6 protein, RNA levels and FACS analyses, a two-way ANOVA limb paralysis but, rather, exhibited signs of severe ataxia that was performed. In all cases, a statistical analysis software package (Sta- prompted us to use a different clinical scoring system that allowed p Ͻ tistical Program for Social Sciences v 13.0; SPSS) was used. 0.05 more accurate quantification of the degree of physical impairment was considered significant throughout all analyses. in these animals. Based on this assessment compared with Mi-WT

Results controls, Mi-GFAP-IL6 mice showed significantly increased by guest on September 26, 2021 IL-6 is present in the cerebellum but not detectable in the ataxia over the entire period of study (Fig. 2B). cerebrum or spinal cord from GFAP-IL6 mice Previous characterization of the histopathological alterations in the CNS of the GFAP-IL6 transgenic mice used in the current study revealed that changes including loss of BBB integrity and gliosis were present at 3 mo age and were localized predominantly in the cerebellum compared with other brain regions (18, 19). Because the presence of IL-6 in different brain regions of these mice has not been studied, it was necessary to determine the levels of IL-6 in lysates of the cerebrum, cerebellum, and spinal cord from WT or GFAP-IL6 mice. In lysates of all three CNS tissues from WT mice, the level of IL-6 was close to the detection limit of the assay (Fig. 1). Similarly, IL-6 was barely detectable in lysates of cerebrum and spinal cord from GFAP-IL6 mice. However, in the cerebellum of GFAP-IL6 mice, IL-6 was detectable at significantly higher levels than the other brain regions. These findings indicated that, in FIGURE 2. GFAP-IL6 mice developed an atypical course of disease the GFAP-IL6 mice, the transgene-encoded IL-6 production is characterized by severe ataxia without signs of limb paralysis. GFAP-IL6 largely restricted to the cerebellum and thus correlates with the (n ϭ 23) and WT (n ϭ 22) mice were immunized with 300 ␮gofMOG previously reported neuropathology that is predominant in this re- peptide emulsified in CFA (MOG-immunized, Mi-) and injected i.p. with gion of the CNS (18, 19). 500 ng of pertussis toxin at days 0 and 2. A control cohort of WT and GFAP-IL6 mice (n ϭ 4 for each group) received 300 ␮g of BSA (CFA GFAP-IL6 mice develop an altered clinical course following immunized, Ci-). A group of naive (nonimmunized, Ni-) mice (n ϭ 3 per MOG immunization genotype) was also used for comparisons. A, Clinical EAE evaluation in To determine whether or not astrocyte-targeted production of MOG immunized WT (Mi-WT) and Mi-GFAP-IL6 mice. B, Ataxia scores (assessed as described in the Materials and Methods section) in nonim- IL-6 influenced the clinical course of EAE, GFAP-IL6 and WT munized (Ni), CFA immunized (Ci), and MOG-immunized (Mi-) WT and littermate mice were immunized with MOG35–55 peptide and GFAP-IL6 mice. Note that ataxia was unable to be scored in Mi-WT at observed over 50 days (Fig. 2). Nonimmunized (Ni)-WT and peak disease due to bilateral hind limb paralysis. C, Disease incidence, Ni-GFAP-IL6 mice were clinically indistinguishable (Fig. 2B). mortality, and average disease onset (either EAE score Ն1 or ataxia) scores p Ͻ 0.05 ,ء ,BSA-CFA-immunized (Ci) WT mice showed no physical in Mi-WT and Mi-GFAP-IL6 mice. For statistical significance changes compared with control animals while Ci-GFAP-IL6 vs WT. 2082 SITE-SPECIFIC PRODUCTION OF IL-6 TARGETS AUTOIMMUNITY Downloaded from

FIGURE 3. Reduced leukocyte infiltration and demyelination in the spinal cord of GFAP-IL6 mice. Paraffin-embedded sections of spinal cord were prepared and processed for routine histological staining as described in the Materials and Methods section. Although no leukocytic infiltration (assessed by H&E staining) nor demyelination (assessed by LFB staining) can be observed in paraffin-embedded spinal cord sections of Ci-WT animals (A and B) or Ci-GFAP-IL6 mice (E and F), after MOG immunization (Mi-) multiple perivascular infiltrates were observed at peak disease in Mi-WT mice (C, arrows) which corresponded with areas of demyelination (D, arrows). In contrast, minimal infiltration (G, arrow) together with no signs of demyelination (H) were http://www.jimmunol.org/ observed in Mi-GFAP-IL6 mice at peak disease. Lesion area relative to the white matter area (I) was quantified in longitudinal sections of spinal cord from p Ͻ 0.05 ,ء ,Mi-WT and Mi-GFAP-IL6 (n ϭ 3 for each group) and showed a significant reduction in the Mi-GFAP-IL6 mice. For statistical significance vs WT. Scale bar: 100 ␮m(A, C, E, G)or50␮m(B, D, F, H).

In summary, GFAP-IL6 mice developed an atypical EAE dis- (Fig. 4A), there was increased cellularity of the gray and white ease after MOG immunization that suggested primary involvement matter regions as well as the presence of perivascular leukocyte of the cerebellum rather than the spinal cord. accumulations particularly in the cerebellar sulci (Fig. 4E; arrows). Examination of the myelin integrity also revealed alterations by guest on September 26, 2021 The pathological lesions induced by MOG immunization in present in cerebellum from the Ci-GFAP-IL6 mice (Fig. 4F) com- GFAP-IL6 mice are largely restricted to the cerebellum pared with Ci-WT animals (Fig. 4B) with the myelin appearing To determine the basis for the atypical physical phenotype pre- less well-organized and focal areas of pallor visible. In Mi-WT sented by the Mi-GFAP-IL6 mice, we examined the brain and the mice, the structural integrity of the cerebellum remained largely spinal cord by routine histology at disease peak (day 20 postim- intact (Fig. 4, C and D). Leukocyte infiltrates were observed, but munization). No significant difference in the histological appear- these were mostly found to be associated with blood vessels (Fig. ance of the spinal cord was observed in Ci-WT mice (Fig. 3, A and 4C; arrow), while LFB staining revealed limited disruption of the B) compared with Ci-GFAP-IL6 mice (Fig. 3, E and F). By con- myelin proximal to the perivascular infiltrates (Fig. 4D; arrow). trast, in Mi-WT mice in the white matter of the spinal cord, mul- Compared with Mi-WT mice, cerebellum from Mi-GFAP-IL6 tifocal perivascular infiltrates were observed by H&E staining mice had extensive hypercellular lesions corresponding to the ac- (Fig. 3C, arrows) and corresponded with extensive areas of demy- cumulation of large numbers of inflammatory cells which were elination as evidenced by the absence of LFB staining (Fig. 3D, diffusely located throughout the white matter (Fig. 4G) and asso- arrows). However, in Mi-GFAP-IL6, mice far fewer infiltrates ciated with widespread disintegration of the myelin tracts (Fig. 4H; were observed (Fig. 3G) and correspondingly little, if any, demy- arrowheads). elination was observed (Fig. 3H). Quantitative analyses of the le- In summary, these findings indicated that MOG immunization sion area as a proportion of the total white matter showed a sig- induced much more severe and extensive inflammation and demy- nificant reduction in the Mi-GFAP-IL6 mice compared with elination in the cerebellum of GFAP-IL6 mice compared with WT Mi-WT mice. In addition, in Mi-GFAP-IL6 mice the number of littermates. This pathological presentation in the Mi-GFAP-IL6 lesions was reduced significantly (74.56 Ϯ 9.22 vs 33.00 Ϯ 5.48 mice correlated with the clinical signs of severe ataxia noted in for Mi-WT vs Mi-GFAP-IL6, respectively) as was the average size these animals. of the inflammatory lesions (744.51 Ϯ 141.21 ␮m2 vs 203.74 Ϯ 47.03 ␮m2 for Mi-WT vs Mi-GFAP-IL6, respectively). In all, The distribution, numbers, and phenotype of leukocytes are these findings showed that the extent of the inflammatory lesion altered markedly in the CNS of Mi-GFAP-IL6 mice formation and tissue destruction due to immunization with MOG The results presented above indicated that the localized production peptide was reduced significantly in spinal cord from GFAP-IL6 of IL-6 in the cerebellum of GFAP-IL6 mice redirects the EAE mice compared with WT littermate controls. inflammatory response to MOG immunization and consequent tis- As noted above, the severe ataxic phenotype observed in Mi- sue destruction away from the spinal cord to the cerebellum. We GFAP-IL6 mice suggested the involvement of the cerebellum. next investigated how the localized production of IL-6 in the cer- Routine H&E examination of the cerebellum of Ci-GFAP-IL6 ebellum influenced the distribution, numbers and phenotype of the mice (Fig. 4E) revealed that in comparison with Ci-WT animals infiltrating leukocyte populations. The Journal of Immunology 2083

FIGURE 4. Increased mononuclear cell infiltration and demyelination in the cerebellum of Mi-GFAP-IL6 mice. Paraffin embedded sections of brain were prepared and processed for routine histological staining as described in the Materials and Methods section. No abnormal pathological changes were seen in H&E (A) or LFB (B) stained sections from Ci-WT mice. Occasional perivascular infiltrates were observed in Mi-WT mice at peak disease (C and D, arrows) as well as adjacent areas of demyelination (D, arrowhead). In contrast, increased cellularity was seen in H&E stained sections from Ci-GFAP-IL6 mice (E), particularly in the cerebellar sulci (arrows) as well as disorganization of the white matter (F). After MOG immunization, cerebellum from GFAP-IL6 mice showed severe, extensive and diffuse parenchymal accumulation of leukocytes at peak disease (G) with widespread disintegration of the white matter (H, arrowheads). Scale bar: 25 ␮m. Downloaded from

To better characterize the leukocyte populations in situ, immu- tiple lesions containing macrophage and T cells were observed in nohistochemistry was performed to identify T cells and microglia/ the spinal cord of Mi-WT mice (Fig. 5C); however, in clear con- macrophages (Fig. 5). In the spinal cord, lectin-positive cells with trast, there was a paucity of these mononuclear cell infiltrates a similar and number and distribution corresponding to vascular present in the spinal cord of Mi-GFAP-IL6 mice (Fig. 5D). In the endothelial cells and microglia were observed in both Ci-WT (Fig. cerebellum, clear evidence of microglial activation (Fig. 5F) and http://www.jimmunol.org/ 5A) or Ci-GFAP-IL6 (Fig. 5B) mice. After EAE induction, mul- modest T cell accumulation (Fig. 5H) was evident in Ni-GFAP-IL6 by guest on September 26, 2021

FIGURE 5. Altered distribution and increased accumulation of macrophage/microglia and T cells in the cerebellum of Mi-GFAP-IL6 mice. Paraffin- embedded sections of spinal cord (A–D) or cerebellum (E–P) were incubated with tomato lectin (to identify macrophage/microglia; A–F, I, J, M–P)or anti-CD3 (C, D, G, H, K, L, O, P) and stained as described in the Materials and Methods section. No microglial activation was observed in the spinal cords of either Ci-WT (A) or Ci-GFAP-IL6 (B) mice. Induction of EAE increased the number of perivascular infiltrates formed by CD3ϩ cells surrounded by lectin positive microglia/macrophages in spinal cords of Mi-WT mice (C). However, almost no infiltrates where observed in the spinal cords of Mi- GFAP-IL6 (D). Conversely, in the cerebellum there were increased numbers of lectin positive macrophage/microglia in Ni-GFAP-IL6 (F) and Ci-GFAP- IL6 mice (J), which was further increased markedly in Mi-GFAP-IL6 (N) whereas only discrete foci of lectin-macrophage/microlgia were observed in Mi-WT mice (M) compared with Ni-WT (E) or Ci-WT (I) animals. No CD3-positive cells were observed in either Ni-WT (G) or Ci-WT (K) mice. In contrast, scattered CD3-positive T cells were found in the cerebellum of Ni-GFAP-IL6 (H) mice, and the number of these cells further increased after CFA treatment, in particular in the cerebellar sulci of Ci-GFAP-IL6 mice (L). At peak EAE, cerebellum from both Mi-WT and Mi-GFAP-IL6 mice showed the presence of CD3-positive T cells (O and P). At higher-magnification the CD3-positive T cells were seen to be localized in a perivascular fashion in the inflammatory foci of Mi-WT mice (O). In contrast, CD3-positive cells were found scattered throughout the parenchyma of Mi-GFAP-IL6 mice (P). Scale bar: 50 ␮m(A–D), 25 ␮m(E–N), or 8 ␮m(O and P). 2084 SITE-SPECIFIC PRODUCTION OF IL-6 TARGETS AUTOIMMUNITY

(Ly6Gϩ/CD11cϪ) and B220ϩ/CD11cϪ cells (B-cells) isolated from Mi-GFAP-IL6 transgenic animals than from Mi-WT controls (Fig. 6B). Inflammation-related expression in GFAP-IL6 and WT mice during EAE To delineate further the nature of the immune response in the spi- nal cord and cerebellum of Mi-WT vs Mi-GFAP-IL6 mice, the level of mRNA transcripts for a number of key cytokine and che- mokine genes was analyzed by RPA. In the spinal cord of Ci-WT mice, the levels of all the examined cytokine and chemokine transcripts were very low or undetectable (Fig. 7, A and B). A similar pattern of expression was seen in the Ci-GFAP-IL6 mice with the exception of a small but significant FIGURE 6. Markedly increased accumulation of various leukocyte sub- increase in both the IL-6 and CCL12 mRNAs. In general, this sets in the cerebellum of Mi-GFAP-IL6 mice. Tissue leukocytes were iso- pattern of expression was maintained in the preclinical phase of lated and analyzed by flow cytometry as described in the Materials and disease in both Mi-WT and Mi-GFAP-IL6 mice. However, at the Methods. Cerebellar cell counts (A) and relative ratios (B)ofCD4ϩ, CD8ϩ, peak phase of disease, the expression of the TNF, TGF-␤, CCL12, ϩ ϩ ϩ ϩ ϩ Ϫ ϩ CD45 CD11b , CD45 CD11c , Ly6G CD11c and B220 cell popu- IL-6, CCL5, and IFN-␥ mRNAs was increased markedly in Downloaded from lations in Mi-WT and Mi-GFAP-IL6 mice. For B, quantification of each Mi-WT mice and were significantly higher than were found in the specific cell population is shown as a percentage of the total number of Mi-GFAP-IL6 mice (Fig. 7B). These differences were particularly leukocytes gated according to side- and forward-light scatter. A near 10- striking in the case of TNF, CCL5, and IFN-␥. Finally, during the fold increase of all leukocyte subsets was found in Mi-GFAP-IL6 mice chronic phase of disease, the levels of the TNF, CCL5, and IFN-␥ compared with Mi-WT animals (A). However, only the relative proportion ϩ Ϫ ϩ but not the TGF-␤, CCL12, and IL-6 mRNA transcripts decreased of Ly6G /CD11c granulocytes and B220 B cells was found to be in- creased in Mi-GFAP-IL6 compared with Mi-WT mice (B). For statistical significantly in Mi-WT mice and showed a similar change in the http://www.jimmunol.org/ p Ͻ 0.05 vs WT. Mi-GFAP-IL6 animals. In all, these findings showed that the levels ,ء ,significance of expression of a number of cytokine and chemokine genes was significantly reduced in the spinal cord of Mi-GFAP-IL6 mice, mice compared with Ni-WT (Fig. 5, E and G) animals. After CFA consistent with the reduced numbers of leukocytes and tissue immunization there was negligible microglial activation (Fig. 5I) destruction. or T cell infiltration (Fig. 5K) in WT mice. However, while Ci- The results for the cerebellum (Fig. 7, C and D) were similar to GFAP-IL6 animals showed a similar degree of microglial activa- those of the spinal cord in CFA-treated mice. Thus, low levels of tion and morphology as compared with Ni-GFAP-IL6 (Fig. 5J) all the examined cytokine and chemokine mRNA transcripts were mice, increased numbers of T cells were observed in Ci-GFAP-IL6 observed in Ci-WT mice, whereas IL-6 and CCL12 mRNA tran- by guest on September 26, 2021 mice, in particular in the cerebellar sulci (Fig. 5L). This latter scripts were increased in the Ci-GFAP-IL6 mice. However, in con- finding is consistent with the routine histological findings noted trast to the spinal cord, TNF and TGF-␤ were also increased in the above indicating that the cerebellum of the GFAP-IL6 mice favors cerebellum of the Ci-GFAP-IL6 mice. Essentially the same pattern the accumulation and activation of inflammatory cells following of expression was observed in both genotypes in the preclinical peripheral immune activation. In Mi-WT mice with EAE, perivas- phase of disease. In contrast, the expression of all the cytokine and cular infiltrates consisting of T cells surrounded by activated mi- chemokine mRNAs studied was significantly increased in Mi-WT croglia/macrophages were located in the white matter of the cer- mice at the peak phase of disease, while only TNF and CCL5 ebellum (Fig. 5, M and O). In contrast to Mi-WT mice, the mRNAs were increased clearly in the Mi-GFAP-IL6 mice com- cerebellum of Mi-GFAP-IL6 mice showed extensive accumulation pared with the preclinical phase. Finally, during the chronic phase of large numbers of lectin-positive macrophage/microglia, which of disease, the levels of all transcripts were significantly reduced were distributed throughout the white matter (Fig. 5N). These mac- from their peak values. rophages had a more rounded morphology and were more densely As noted above, it is now clear that CD4ϩ Th17 cells are crucial stained compared with the more ramified morphology of the mac- effector T cells in the development of MOG-EAE. Therefore, we rophage/microglia in the Mi-WT mice and were accompanied by compared the levels of IL-17 mRNA in the cerebellum of Mi-WT extensive T cell infiltration into the white matter (Fig. 5P). Overall, and Mi-GFAP-IL6 mice. Initially, RPA was used for this purpose these findings revealed that while the Mi-GFAP-IL6 mice dis- however, due to extremely low levels of IL-17 mRNA, we could played reduced disease scores and exhibited little spinal cord pa- barely detect this transcript. As an alternative, we used the more thology, there was a marked increase in the distribution and se- sensitive real time-PCR technique for this analysis. As shown in verity of the EAE inflammatory response in the cerebellum of Fig. 7E the level of IL-17 mRNA was significantly higher in the these mice. cerebellum of Ci-GFAP-IL6 mice compared with Ci-WT litter- To aid further in the phenotypical characterization of the infil- mates. During the peak disease phase, there was a significant in- trating cell populations, we isolated cells from Mi-GFAP-IL6 and crease of IL-17 mRNA in Mi-WT mice compared with Ci-WT, Mi-WT (Fig. 6A) cerebellum at peak disease and performed flow while the level of IL-17 mRNA in Mi-GFAP-IL6 mice was in- cytofluorometric analysis. In support of the immunohistochemical creased to similar levels as seen in the Ci-GFAP-IL6 controls. results, Mi-GFAP-IL6 animals displayed a 10-fold increase in all To determine further the nature of the inflammatory response in leukocyte subsets examined (Fig. 6A). Although, the relative pro- the cerebellum of the Mi-GFAP-IL6 mice, an analysis of the ex- portion of T cell subsets (CD4ϩ and CD8ϩ), dendritic cells, pression of several host response-related genes was performed (CD45ϩ/CD11cϩ) and macrophages (CD45ϩ/CD11bϩ) remained (Fig. 8, A and B). The levels of the ICAM-1, MAC1, EB22/5, and similar between the two strains of immunized mice, there was an GFAP mRNA transcripts were all significantly increased in Ci- approximate 2-fold increase in the percentage of neutrophils GFAP-IL6 mice compared with Ci-WT littermates (Fig. 8B). This The Journal of Immunology 2085 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 7. Expression of selected inflammatory cytokine and chemokine genes is decreased in the spinal cord and increased in the cerebellum of Mi-GFAP-IL6 mice with EAE. Total RNA was isolated from the cerebellum and spinal cord of mice at the times shown and 5 ␮g was used for analysis by RPA as specified in the Materials and Methods section. RPA data for selected chemokine and cytokine RNA transcripts in spinal cord samples of WT and GFAP-IL6 mice (A) were quantified by densitometry and normalized to the L32 loading control (B) and showed an overall decrease in Mi-GFAP-IL6 compared with Mi-WT mice. The opposite was observed in cerebellum (C and D), with an increased expression selected chemokine and cytokine RNA transcripts in Ci-GFAP-IL6 mice. After EAE immunization, Mi-WT, and to a lesser extent Mi-GFAP-IL6 mice, showed an increase in the expression of all genes studied. E, qRT-PCR was performed to determine the level of IL17A mRNA in the cerebellum of Ci- and Mi-WT (n ϭ 6 and 5, respectively) and Ci and Mi-GFAP-IL6 (n ϭ 3 and 4, respectively) mice. Compared with Ci-WT, the Ci-GFAP-IL6 p Ͻ ,ء ,mice had significantly higher levels of IL-17A expression that remained similarly elevated in Mi-GFAP-IL6 mice. For statistical significance 0.05 vs WT or ࡗ, p Ͻ 0.05 vs Ci mice. pattern of expression of these host response-related mRNA Discussion transcripts was maintained in Mi-GFAP-IL6 and Ci-WT mice IL-6 has emerged as a crucial cytokine in the pathogenesis of a during the preclinical phase of disease. At peak disease in Mi- number of experimental and human autoimmune diseases (15). GFAP-IL6 mice, with the exception of EB22/5 mRNA, which IL-6 is not only multifunctional, but during the course of an au- was increased further, the ICAM-1, MAC1, and GFAP mRNAs toimmune response, the production of this cytokine arises from remained elevated to similar levels while the levels of all these transcripts were significantly increased in the Mi-WT mice. Ad- diverse cellular sources both in the periphery and the target tissue. Although it is now known that IL-6 is crucially involved in the ditionally, during the peak phase of disease, the levels of the ϩ iNOS and A20 mRNAs were increased significantly in both development of autoreactive CD4 Th17 effector T cells in vari- Mi-WT and Mi-GFAP-IL6 mice, however the level of iNOS ous experimental autoimmune disease models (7, 11, 12), the rel- mRNA lower in Mi-GFAP-IL6 mice when compared with ative contribution of IL-6 production in the target tissue to devel- Mi-WT animals, although this did not reach statistical signifi- opment of autoimmunity remains uncertain. In EAE (26) and in cance. During the chronic phase of disease, the levels of all MS (27), IL-6 or IL-6 transcripts are elevated in the CNS, which these mRNAs were reduced to almost control levels in the likely reflects, in part, production by intrinsic cells to the CNS Mi-WT mice while ICAM-1, MAC1, EB22/5, and GFAP re- such as the astrocytes and microglia (28). In this study, we mained elevated in the Mi-GFAP-IL6 animals. assessed the role of such tissue-localized production of IL-6 in 2086 SITE-SPECIFIC PRODUCTION OF IL-6 TARGETS AUTOIMMUNITY

ations affecting the cerebellum, the BBB integrity is compromised in the cerebellum but not in the spinal cord or in the cerebral hemispheres (18). Evidence of early inflammatory changes were also present in the cerebellum but not spinal cord of the GFAP-IL6 mice and included astrocytosis and microgliosis as well as in- creased expression of some proinflammatory cytokine genes. The intensity of this inflammatory process was increased after CFA immunization in these mice which produced mild clinical signs. These findings together with the grossly exaggerated accumulation of inflammatory cells following MOG-immunization indicate that the local milieu of the tissue conditioned by IL-6 can dramatically influence the development of both autoantigen-specific as well as nonspecific immune responses and associated pathology. It is well documented that changes in the CNS microenviron- ment, for example, following induction of nerve injury (29–31) or resulting from cortical cryolesion in rats (32) targets EAE inflam- mation to the lesion site. Similar to the GFAP-IL6 mice, focal cortical cryolesion is associated with localized breakdown of the

BBB and the induction of a proinflammatory state with increased Downloaded from expression of a number of chemokine genes including CCL5 and CCL12 (32). However, as we have shown (supplementary Fig. 1),9 in contrast to Mi-GFAP-IL6 mice, cryolesion-EAE is accompanied by classical ascending spinal cord inflammatory lesions with de- myelination and paralysis. Thus, a compound brain lesion caused

by cryoinjury induced as described (33), while it can target EAE http://www.jimmunol.org/ inflammation, does not cause the spinal cord sparing effects that FIGURE 8. Expression of selected inflammatory host response-related were observed in the Mi-GFAP-IL6 mice and thus argues for a genes is increased in the cerebellum of GFAP-IL6 mice. Total RNA was specific effect of IL-6 in modifying the cerebellar microenviron- isolated from the cerebellum of mice at the times shown and 5 ␮g used for ment. This idea that IL-6 induces a unique inflammatory environ- analysis by RPA as specified in the Materials and Methods section. RPA for selected host response-related mRNA transcripts in cerebellum from ment in the cerebellum of the GFAP-IL6 mice is further reinforced WT and GFAP-IL6 mice (A) were quantified by densitometry and normal- by our finding that the transgenic production of the proinflamma- ized to the L32 loading control (B). GFAP-IL6 animals consistently pre- tory cytokine IL-12 in the cerebellum, while causing an earlier sented with an increased expression of many of the genes studied that was onset of disease, did not significantly alter the development of already evident in CFA immunized animals. After EAE immunization, EAE (supplementary Fig. 1 and Ref. 34). by guest on September 26, 2021 Mi-WT, and to a lesser extent Mi-GFAP-IL6 mice, showed an increase in Mi-GFAP-IL6 mice showed a large increase in the number of the expression of genes studied. However, overall transcript levels were mononuclear cells accumulating in the cerebellum, which led to found to be higher in Mi-GFAP-IL6 mice compared with Mi-WT. For more severe demyelination when compared with Mi-WT animals. Ͻ ࡗ Ͻ ء statistical significance, , p 0.05 vs WT or , p 0.05 vs Ci mice. It is likely that the increased numbers of leukocytes in the cere- bellum of the Mi-GFAP-IL6 arose from increased recruitment from the circulation and/or proliferation locally. IL-6 increases the the induction and evolution of MOG-EAE, using transgenic levels of cellular adhesion molecules such as ICAM-1 on endo- mice with astrocyte-targeted and largely cerebellum-restricted thelial cells, which is crucial for tissue leukocyte recruitment me- production of IL-6. The results revealed a grossly exaggerated in- diated by this cytokine (16). Vascular ICAM-1 and VCAM-1 lev- volvement of the cerebellum in Mi-GFAP-IL6 mice which devel- els are increased in EAE and have been shown to mediate adhesion oped severe ataxia concomitant with increased immune pathology of to the inflamed cerebral vessels in the spinal cord and demyelination in the cerebellum. As is well known, in WT (35, 36). The loss of this interaction prevents the development of mice, MOG-EAE is characterized by an ascending paralysis re- EAE and is associated with a marked reduction in the numbers of flecting the primary targeting of the autoimmune response to the leukocytes in the CNS (37). The microenvironment of the cere- spinal cord and to a lesser degree the cerebellum. However, and bellum in the GFAP-IL6 mice is characterized by chronic activa- somewhat unexpectedly, there was only very mild inflammation in tion of the vascular endothelium. In particular, the levels of the the spinal cord of Mi-GFAP-IL6 mice with EAE. The few inflam- cellular adhesion molecules ICAM-1 and VCAM-1 as well as a matory lesions present were small in size and accompanied by number of integrins is increased significantly on vascular endothe- scant demyelination. Based on these findings, we conclude that the lium (19, 38). This pre-existing activation state of the vascular tissue-localized transgenic production of IL-6 acts as a potent stim- endothelium in the cerebellum of the GFAP-IL6 mice would be ulus to target the autoimmune response and may create a “sink” expected to favor the increased adhesion and subsequent transen- that retards leukocyte trafficking to normally preferred antigenic dothelial migration of leukocytes following MOG-immunization. sites. The BBB is a specialized and unique barrier to the CNS that The mechanisms underlying the almost exclusive targeting and regulates the flow of macromolecules and leukocytes into this tis- expansion of the EAE lesion development to the cerebellum in the sue. In adult GFAP-IL6 mice the BBB is impaired in the cerebel- Mi-GFAP-IL6 mice are likely to be complex. Astrocyte-targeted lum and is leaky for the life of the animal (18, 19). Therefore, it production of IL-6 in the cerebellum establishes a proinflammatory could be argued that this loss of BBB integrity would allow for milieu that progresses with the age of the animals (19, 21). Al- though at the age used for the current studies, the GFAP-IL6 mice appeared physically normal and had modest pathological alter- 9 The online version of this article contains supplemental material. The Journal of Immunology 2087 more indiscriminate entry of immunocompetent T cells and other proportions of CD4ϩ T cells and CD11bϩ microglia/macrophages leukocytes into the brain of the Mi-GFAP-IL6 mice. Indeed, our were similar between Mi-WT and Mi-GFAP-IL6 mice. The find- results showed a clear increase in the number of T cells present in ings suggest that the major impact of site-specific tissue production the cerebellum of Ci-GFAP-IL6 mice but not Ci-WT mice. How- of IL-6 is in the targeting of the inflammatory response and not the ever, the role of the BBB in this context is not so clear. It has been subsequent modulation of effector T cell function in the lesion. shown that activated T cells are capable of freely entering the CNS In this study, we have shown that localized transgenic produc- after migrating across the healthy intact BBB (39, 40). In EAE, tion of IL-6 in the cerebellum targets and enhances the inflamma- disease severity was reported to correlate closely with the degree tory response directed against a CNS-specific autoantigen. More- of BBB permeability suggesting that loss of BBB integrity is in- over, this local IL-6 milieu created a leukocyte “sink” that diverted tegral to the development of EAE (41). However, the relationship trafficking from and inflammation in the normally dominant anti- between BBB integrity and inflammatory cell accumulation in the genic site of the spinal cord. The mechanisms underlying these CNS in EAE is not straightforward. Areas of the CNS of naive responses to IL-6 in the cerebellum likely include the induction of mice that have constitutively greater BBB permeability do not nec- specific cytokines and chemokines as well as increased vascular essarily overlap with lesion distribution in EAE suggesting that activation and loss of integrity of the BBB. These findings estab- BBB integrity in of itself does not determine lesion establishment lish for the first time that intrinsic production of IL-6 in a target (42). Moreover, accumulation of inflammatory cells in the brain tissue can dramatically target and augment the development of an parenchyma in EAE can occur before detectable loss of BBB in- autoimmune response. tegrity (42).

Another plausible mechanism that could contribute to the tar- Acknowledgments Downloaded from geted recruitment of inflammatory cells to the cerebellum in the We thank Jane Radford and Barbara Hernandez (Department of Pathology, Mi-GFAP-IL6 mice may involve chemokines signaling. In this University of Sydney) for expert technical assistance with tissue processing regard, we identified CCL12 (MCP-5) mRNA as being markedly and routine histology and Drs. Bernardo Castellano, Berta Gonzalez, and increased in the cerebellum of the GFAP-IL6 mice compared with Bea Almolda for helpful discussions on this work. WT. In addition, the expression of some other chemokine genes Disclosures including CXCL10, CCL2, and CCL5 is elevated in the cerebel- http://www.jimmunol.org/ lum of the GFAP-IL6 mice (43) however, in comparison with The authors have no financial conflict of interest. CCL12, the level of expression of these other chemokine tran- References scripts is considerably lower (I.L. Campbell, unpublished obser- 1. Brennan, F. M., and I. B. McInnes. 2008. Evidence that cytokines play a role in vation). Interestingly, CCL12 gene expression was also increased rheumatoid arthritis. J. Clin. Invest. 118: 3537–3545. in the cerebellum and spinal cord of WT mice before onset of EAE 2. Steinman, L. 2008. Nuanced roles of cytokines in three major human brain dis- orders. J. Clin. Invest. 118: 3557–3563. and increased further during peak disease suggesting there may be 3. Sasai, M., Y. Saeki, S. Ohshima, K. Nishioka, T. Mima, T. Tanaka, Y. Katada, a role for this chemokine in leukocyte trafficking to the CNS in WT K. Yoshizaki, M. Suemura, and T. Kishimoto. 1999. Delayed onset and reduced mice with EAE. Studies in vitro have shown that CCL12 is a severity of collagen-induced arthritis in -6-deficient mice. Arthritis

Rheum. 42: 1635–1643. by guest on September 26, 2021 potent chemoattractant for /macrophages (44, 45). This 4. Hirota, K., M. Hashimoto, H. Yoshitomi, S. Tanaka, T. Nomura, T. Yamaguchi, function of CCL12 appears to be mediated primarily by the recep- Y. Iwakura, N. Sakaguchi, and S. Sakaguchi. 2007. T cell self-reactivity forms a ϩ tor CCR2 (45). Studies using CCR2 KO mice have failed to pro- cytokine milieu for spontaneous development of IL-17 Th cells that cause au- toimmune arthritis. J. Exp. Med. 204: 41–47. duce a unifying conclusion as to the function of CCR2 signaling in 5. Mendel, I., A. Katz, N. Kozak, A. Ben-Nun, and M. Revel. 1998. Interleukin-6 MOG-EAE. Initial reports concluded that CCR2 KO mice were functions in autoimmune encephalomyelitis: a study in gene-targeted mice. Eur. J. Immunol. 28: 1727–1737. resistant to EAE and exhibited a marked disease reduction (46, 6. Samoilova, E. B., J. L. Horton, B. Hilliard, T. S. Liu, and Y. Chen. 1998. IL-6- 47), however this was challenged in a later study (48). Given the deficient mice are resistant to experimental autoimmune encephalomyelitis: roles well-known promiscuity of this family (49, of IL-6 in the activation and differentiation of autoreactive T cells. J. Immunol. 161: 6480–6486. 50), it is conceivable the function of CCL12 in the pathogenesis of 7. Bettelli, E., Y. Carrier, W. Gao, T. Korn, T. B. Strom, M. Oukka, H. L. Weiner, EAE could be mediated via alternative CCRs but this, as well as and V. K. Kuchroo. 2006. Reciprocal developmental pathways for the generation the functions of CCL12 in EAE, remain to be determined. of pathogenic effector TH17 and regulatory T cells. Nature 441: 235–238. 8. Nakae, S., A. Nambu, K. Sudo, and Y. Iwakura. 2003. Suppression of immune Adoptive transfer studies in IL-6 KO mice suggest that local induction of collagen-induced arthritis in IL-17-deficient mice. J. Immunol. 171: production of IL-6 in the CNS is required for the optimal devel- 6173–6177. 9. Park, H., Z. Li, X. O. Yang, S. H. Chang, R. Nurieva, Y. H. Wang, Y. Wang, opment of EAE through the local priming of vascular function e.g., L. Hood, Z. Zhu, Q. Tian, and C. Dong. 2005. A distinct lineage of CD4 T cells by up-regulation of adhesion molecules (13). It is now clear that regulates tissue inflammation by producing . Nat. Immunol. 6: IL-6 together with TGF-␤ can drive the formation of effector Th17 1133–1141. 10. Mangan, P. R., L. E. Harrington, D. B. O’Quinn, W. S. Helms, D. C. Bullard, cells (7, 10, 12) and these cytokines are essential for the induction C. O. Elson, R. D. Hatton, S. M. Wahl, T. R. Schoeb, and C. T. Weaver. 2006. of EAE. Interestingly, in addition to IL-6, we observed that the Transforming -␤ induces development of the TH17 lineage. Nature level of TGF-␤ gene expression was also increased in the cerebel- 441: 231–234. 11. McGeachy, M. J., K. S. Bak-Jensen, Y. Chen, C. M. Tato, W. Blumenschein, lum of the Mi-GFAP-IL6 mice. This raised the question as to T. McClanahan, and D. J. Cua. 2007. TGF-␤ and IL-6 drive the production of whether the local production of these cytokines might also mod- IL-17 and IL-10 by T cells and restrain TH-17 cell-mediated pathology. Nat. Immunol. 8: 1390–1397. ulate effector T cell function during an autoimmune response. The 12. Veldhoen, M., R. J. Hocking, C. J. Atkins, R. M. Locksley, and B. Stockinger. spontaneous development of arthritogenic Th17 cells and augmen- 2006. TGF␤ in the context of an inflammatory cytokine milieu supports de novo tation of disease by zymosan in SKG mice was shown to be de- differentiation of IL-17-producing T cells. Immunity 24: 179–189. 13. Eugster, H. P., K. Frei, M. Kopf, H. Lassmann, and A. Fontana. 1998. IL-6- pendent on the synergistic interaction between T cell-derived and deficient mice resist myelin oligodendrocyte glycoprotein-induced autoimmune non-T cell-derived IL-6 (4). However, our findings here showed encephalomyelitis. Eur. J. Immunol. 28: 2178–2187. that, while elevated, the CNS levels of IL-17 mRNA were little 14. Korn, T., M. Mitsdoerffer, A. L. Croxford, A. Awasthi, V. A. Dardalhon, G. Galileos, P. Vollmar, G. L. Stritesky, M. H. Kaplan, A. Waisman, et al. 2008. altered between Ci-GFAP-IL6 and Mi-GFAP-IL6 mice while in- IL-6 controls Th17 immunity in vivo by inhibiting the conversion of conventional creased IFN-␥ mRNA levels at peak disease were similar for T cells into Foxp3ϩ regulatory T cells. Proc. Natl. Acad. Sci. USA 105: 18460–18465. Mi-WT and Mi-GFAP-IL6 mice. Furthermore, although increased 15. Ishihara, K., and T. Hirano. 2002. IL-6 in autoimmune disease and chronic in- significantly in number in the Mi-GFAP-IL6 mice, the relative flammatory proliferative disease. Cytokine Growth Factor Rev. 13: 357–368. 2088 SITE-SPECIFIC PRODUCTION OF IL-6 TARGETS AUTOIMMUNITY

16. Vardam, T. D., L. Zhou, M. M. Appenheimer, Q. Chen, W. C. Wang, 33. Penkowa, M., T. Moos, J. Carrasco, H. Hadberg, A. Molinero, H. Bluethmann, H. Baumann, and S. S. Evans. 2007. Regulation of a -endothelial-IL-6 and J. Hidalgo. 1999. Strongly compromised inflammatory response to brain trans-signaling axis by fever-range thermal stress: hot spot of immune surveil- injury in interleukin-6-deficient mice. Glia 25: 343–357. lance. Cytokine 39: 84–96. 34. Pagenstecher, A., S. Lassmann, M. J. Carson, C. L. Kincaid, A. K. Stalder, and 17. Barnum, S. R., J. L. Jones, U. Muller-Ladner, A. Samimi, and I. L. Campbell. I. L. Campbell. 2000. Astrocyte-targeted expression of IL-12 induces active cel- 1996. Chronic complement C3 gene expression in the CNS of transgenic mice lular immune responses in the central nervous system and modulates experimen- with astrocyte-targeted interleukin-6 expression. Glia 18: 107–117. tal allergic encephalomyelitis. J. Immunol. 164: 4481–4492. 18. Brett, F. M., A. P. Mizisin, H. C. Powell, and I. L. Campbell. 1995. Evolution of 35. Dopp, J. M., S. M. Breneman, and J. A. Olschowka. 1994. Expression of neuropathologic abnormalities associated with blood-brain barrier breakdown in ICAM-1, VCAM-1, L-selectin, and leukosialin in the mouse central nervous transgenic mice expressing interleukin-6 in astrocytes. J. Neuropathol. Exp. Neu- system during the induction and remission stages of experimental allergic en- rol. 54: 766–775. cephalomyelitis. J. Neuroimmunol. 54: 129–144. 19. Campbell, I. L., C. R. Abraham, E. Masliah, P. Kemper, J. D. Inglis, 36. Matsuda, M., N. Tsukada, C. S. Koh, T. Iwahashi, K. Shimada, and M. B. A. Oldstone, and L. Mucke. 1993. Neurologic disease induced in trans- N. Yanagisawa. 1994. Expression of intercellular adhesion molecule-1 and lym- genic mice by the cerebral overexpression of . Proc. Natl. Acad. Sci. phocyte function-associated antigen-1 in the spinal cord of rats during acute USA 90: 10061–10065. experimental allergic encephalomyelitis. Autoimmunity 19: 15–22. 37. Bullard, D. C., X. Hu, T. R. Schoeb, R. G. Collins, A. L. Beaudet, and 20. Chiang, C.-S., A. Stalder, A. Samimi, and I. L. Campbell. 1994. Reactive gliosis S. R. Barnum. 2007. Intercellular adhesion molecule-1 expression is required on as a consequence of interleukin-6 expression in the brain: studies in transgenic multiple cell types for the development of experimental autoimmune encephalo- mice. Dev. Neurosci. 16: 212–221. myelitis. J. Immunol. 178: 851–857. 21. Heyser, C. J., E. Masliah, A. Samimi, I. L. Campbell, and L. H. Gold. 1997. 38. Milner, R., and I. L. Campbell. 2006. Increased expression of the ␤4 and ␣5 Progressive decline in avoidance learning paralleled by inflammatory neurode- integrin subunits in cerebral blood vessels of transgenic mice chronically pro- generation in transgenic mice expressing interleukin 6 in the brain. Proc. Natl. ducing the pro-inflammatory cytokines IL-6 or IFN-␣ in the central nervous Acad. Sci. USA 94: 1500–1505. system. Mol. Cell. Neurosci. 33: 429–440. 22. Metten, P., K. L. Best, A. J. Cameron, A. B. Saultz, J. M. Zuraw, C. H. Yu, 39. Hickey, W. F. 1991. Migration of hematogenous cells through the blood brain D. Wahlsten, and J. C. Crabbe. 2004. Observer-rated ataxia: rating scales for barrier and the initiation of CNS inflammation. Brain Pathol. 1: 97–105. assessment of genetic differences in ethanol-induced intoxication in mice. J. Appl. 40. Hickey, W. F., B. L. Hsu, and H. Kimura. 1991. T-lymphocyte entry into the Downloaded from Physiol. 97: 360–368. central nervous system. J. Neurosci. Res. 28: 254–260. 23. Asensio, V. C., and I. L. Campbell. 1997. Chemokine gene expression in the 41. Fabis, M. J., G. S. Scott, R. B. Kean, H. Koprowski, and D. C. Hooper. 2007. brain of mice with lymphocytic choriomeningitis. J. Virol. 71: 7832–7840. Loss of blood-brain barrier integrity in the spinal cord is common to experimental 24. Hobbs, M. V., W. O. Weigle, D. J. Noonan, B. E. Torbett, R. J. McEvilly, allergic encephalomyelitis in knockout mouse models. Proc. Natl. Acad. Sci. USA R. J. Koch, G. J. Cardenas, and D. N. Ernst. 1993. Patterns of cytokine gene 104: 5656–5661. expression by CD4ϩ T cells from young and old mice. J. Immunol. 150: 42. Muller, D. M., M. P. Pender, and J. M. Greer. 2005. Blood-brain barrier disrup- 3602–3614. tion and lesion localisation in experimental autoimmune encephalomyelitis with

25. Fleming, K. K., J. A. Bovaird, M. C. Mosier, M. R. Emerson, S. M. LeVine, and predominant cerebellar and brainstem involvement. J. Neuroimmunol. 160: http://www.jimmunol.org/ J. G. Marquis. 2005. Statistical analysis of data from studies on experimental 162–169. autoimmune encephalomyelitis. J. Neuroimmunol. 170: 71–84. 43. Asensio, V. C., S. Lassmann, A. Pagenstecher, S. C. Steffensen, S. J. Henriksen, 26. Gijbels, K., J. Van Damme, P. Proost, W. Put, H. Carton, and A. Billiau. 1990. and I. L. Campbell. 1999. C10 is a novel chemokine expressed in experimental Interleukin 6 production in the central nervous system during experimental au- inflammatory demyelinating disorders that promotes the recruitment of macro- toimmune encephalomyelitis. Eur. J. Immunol. 20: 233–235. phages to the central nervous system. Am. J. Pathol. 154: 1181–1191. 27. Loch, C., G. Hermans, R. Pedotti, A. Brendolan, E. Schadt, H. Garren, 44. Jia, G. Q., J. A. Gonzalo, C. Lloyd, L. Kremer, L. Lu, A. C. Martinez, A. Langer-Gould, S. Strober, B. Cannella, J. Allard, et al. 2002. Gene-microarray B. K. Wershil, and J. C. Gutierrez-Ramos. 1996. Distinct expression and function analysis of multiple sclerosis lesions yields new targets validated in autoimmune of the novel mouse chemokine monocyte chemotactic protein-5 in lung allergic encephalomyelitis. Nat. Med. 8: 500–508. inflammation. J. Exp. Med. 184: 1939–1951. 45. Sarafi, M. N., E. A. Garcia-Zepeda, J. A. MacLean, I. E. Charo, and A. D. Luster. 28. Frei, K., U. V. Malipiero, T. P. Leist, R. M. Zinkernagel, M. E. Schwab, and 1997. Murine monocyte chemoattractant protein (MCP)-5: a novel CC chemo- A. Fontana. 1989. On the cellular source and function of interleukin 6 produced kine that is a structural and functional homologue of human MCP-1. J. Exp. Med. by guest on September 26, 2021 in the central nervous system in viral disease. Eur. J. Immunol. 19: 689–694. 185: 99–109. 29. Konno, H., T. Yamamato, H. Suzuki, H. Yamamoto, Y. Iwasaki, Y. Ohara, 46. Huang, D. R., J. Wang, P. Kivisakk, B. J. Rollins, and R. M. Ransohoff. 2001. H. Terunuma, and N. Harata. 1990. Targeting of adoptively transferred experi- Absence of monocyte chemoattractant protein 1 in mice leads to decreased local mental allergic encephalitis lesion at sites of Wallerian degeneration. Acta Neu- macrophage recruitment and antigen-specific T helper cell type 1 immune re- ropathol. 80: 521–526. sponse in experimental autoimmune encephalomyelitis. J. Exp. Med. 193: 30. Molleston, M. C., M. L. Thomas, and W. F. Hickey. 1993. Novel major histo- 713–726. compatibility complex expression by microglia and site-specific experimental 47. Izikson, L., R. S. Klein, I. F. Charo, H. L. Weiner, and A. D. Luster. 2000. allergic encephalomyelitis lesions in the rat central nervous system after optic Resistance to experimental autoimmune encephalomyelitis in mice lacking the nerve transection. Adv. Neurol. 59: 337–348. CC chemokine receptor (CCR)2. J. Exp. Med. 192: 1075–1080. 31. Tsunoda, I., T. Tanaka, Y. Saijoh, and R. S. Fujinami. 2007. Targeting inflam- 48. Gaupp, S., D. Pitt, W. A. Kuziel, B. Cannella, and C. S. Raine. 2003. Experi- matory demyelinating lesions to sites of wallerian degeneration. Am. J. Pathol. mental autoimmune encephalomyelitis (EAE) in CCR2Ϫ/Ϫ mice: susceptibility in 171: 1563–1575. multiple strains. Am. J. Pathol. 162: 139–150. 32. Sun, D., M. Tani, T. A. Newman, K. Krivacic, M. Phillips, A. Chernosky, P. Gill, 49. Horuk, R. 1994. Molecular properties of the chemokine receptor family. Trends T. Wei, K. J. Griswold, R. M. Ransohoff, and R. O. Weller. 2000. Role of Pharmacol. Sci. 15: 159–165. chemokines, neuronal projections, and the blood-brain barrier in the enhancement 50. Rot, A., and U. H. von Andrian. 2004. Chemokines in innate and adaptive host of cerebral EAE following focal brain damage. J. Neuropathol. Exp. Neurol. 59: defense: basic chemokinese grammar for immune cells. Annu. Rev. Immunol. 22: 1031–1043. 891–928.