Role for MyD88, TLR2 and TLR9 but Not TLR1, TLR4 or TLR6 in Experimental Autoimmune Encephalomyelitis

This information is current as Socorro Miranda-Hernandez, Nicole Gerlach, Julie M. of September 23, 2021. Fletcher, Erik Biros, Matthias Mack, Heinrich Körner and Alan G. Baxter J Immunol published online 17 June 2011 http://www.jimmunol.org/content/early/2011/06/17/jimmun

ol.1001992 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2011/06/17/jimmunol.100199 Material 2.DC1 http://www.jimmunol.org/

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 by guest on September 23, 2021

*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 © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 17, 2011, doi:10.4049/jimmunol.1001992 The Journal of Immunology

Role for MyD88, TLR2 and TLR9 but Not TLR1, TLR4 or TLR6 in Experimental Autoimmune Encephalomyelitis

Socorro Miranda-Hernandez,* Nicole Gerlach,*,1 Julie M. Fletcher,*,1 Erik Biros,* Matthias Mack,† Heinrich Ko¨rner,* and Alan G. Baxter*

The potential roles of TLRs in the cause and pathogenesis of autoimmune CNS inflammation remain contentious. In this study, we examined the effects of targeted deletions of TLR1, TLR2, TLR4, TLR6, TLR9, and MyD88 on the induction of myelin oligoden-

drocyte glycoprotein 35–55 (MOG35–55) peptide/CFA/pertussis toxin-induced autoimmune encephalomyelitis. Although C57BL/6. Tlr12/2, C57BL/6.Tlr42/2 and C57BL/6.Tlr62/2 mice showed normal susceptibility to disease, signs were alleviated in female C57BL/6.Tlr22/2 and C57BL/6.Tlr92/2 mice and C57BL/6.Tlr2/92/2 mice of both sexes. C57BL/6.Myd882/2 mice were completely protected. Lower clinical scores were associated with reduced leukocyte infiltrates. These results were confirmed by passive 2 2 2 2 adoptive transfer of disease into female C57BL/6.Tlr2 / and C57BL/6.Tlr9 / mice, where protection in the absence of TLR2 Downloaded from was associated with fewer infiltrating CD4+ cells in the CNS, reduced prevalence of detectable circulating IL-6, and increased proportions of central (CD62L+) CD4+CD25+Foxp3+ regulatory T cells. These results provide a potential molecular mechanism for the observed effects of TLR signaling on the severity of autoimmune CNS inflammation. The Journal of Immunology, 2011, 187: 000–000. http://www.jimmunol.org/ ultiple sclerosis (MS) is a complex genetic trait positive skin-prick tests for house dust mite (3). Similarly, in resulting from the contributions of multiple genetic a population in Guinea-Bissau, measles was associated M and environmental factors (1). It is a chronic immune- with an 80% reduction in the risk for skin-prick test positivity to mediated inflammatory/demyelinating disease of the CNS that house dust mite (4). occurs in young adulthood and is more common in women. To Almost 10 y after Strachan first published his idea, Rook and date, there is no cure or preventative therapy for MS, and the Stanford (5, 6) extrapolated it from allergy to autoimmunity, in- mechanisms underlying the initiation and progression of MS are cluding MS, arguing that vaccination replaced an infectious sti- still unclear. Genetic and epidemiological studies have identified mulus that generated a predominately Th1 response with a much some reproducible associations, such as HLA alleles, UV weaker stimulus that generated a predominately Th2 response. The by guest on September 23, 2021 exposure, cigarette smoking, and microbial exposure. original proposed association with deviation did not sur- The hygiene hypothesis, originally proposed by David Strachan vive testing; regression analysis of the Third National Health and of the London School of Hygiene and Tropical Medicine (2), Nutrition Examination Survey demonstrated a strong positive as- postulated that the inverse association between the prevalence of sociation between a reported history of “Th2-mediated allergic hay fever and family size was causally related to infection in early disorder” (hay fever, , or both) and “Th1-mediated autoim- childhood transmitted by unhygienic contact with older siblings. mune disorders” (thyroid disorder, rheumatoid arthritis, or type 1 Evidence of causality was provided when a cohort of school diabetes treated with insulin and not tablets) (7). Our new un- children in the Gabon treated for Ascaris and Trichuris derstanding of the IL-17/IFN-g/IL-4 counterregulation paradigm, with anti-helminth drugs were subsequently twice as likely to have and the central role of IL-17 in experimental autoimmune en- cephalomyelitis (EAE; an animal model of MS-like autoimmune CNS inflammation), may give this aspect of the hypothesis new life. *Comparative Genomics Centre, James Cook University, Townsville, Queensland Despite this, key issues remain. Without knowing what microbial † 4811, Australia; and Regensburg University Medical Center, Regensburg 93053, species are protective, what expressed constituents are responsible Germany for this activity, or how their presence is recognized by the immune 1N.G. and J.M.F. contributed equally to this article. system, attempts to establish immunotherapies based on manipu- Received for publication June 17, 2010. Accepted for publication May 12, 2011. lating microbial exposure are fraught with danger. For example, the This work was supported by the Australian National Health and Medical Research proposed use of bacillus Calmette-Gue´rin to treat MS (8) and Council, Lions Clubs of Australia, Multiple Sclerosis Research Australia, and intra- mural funding from James Cook University and James Cook University Faculty of diabetes (9) was associated with the precipitation of a lupus-like Medicine, Health and Molecular Sciences. A.G.B. is supported by a Senior Research disease in mice (10–12). Fellowship from the Australian National Health and Medical Research Council. One potential mechanism by which microbial flora could reg- Address correspondence and reprint requests to Dr. Alan G. Baxter, Comparative ulate autoimmune responses is via the TLR system and associated Genomics Centre, Molecular Sciences Building 21, James Cook University, Towns- ville, QLD 4811, Australia. E-mail address: [email protected] pathways. The family of TLRs plays an essential role in innate The online version of this article contains supplemental material. and adaptive immune recognition in mammals (13). TLRs are Abbreviations used in this article: CDI, cumulative disease index; DC, ; type 1 transmembrane receptors and recognize distinct pathogen- EAE, experimental autoimmune encephalomyelitis; LFB, Luxol fast blue; MOG, associated molecular patterns through their leucine-rich repeats in myelin oligodendrocyte glycoprotein; MS, multiple sclerosis; PTX, pertussis toxin; the extracytoplasmic domain (14). Because many of these mo- SS, side scatter; TIR, Toll/IL-1R; Treg, regulatory T cell; WT, wild type. lecular patterns are shared with nonpathogens, they are now Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 commonly referred to as microbe-associated molecular patterns

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1001992 2 ROLE OF TLR IN EAE

(15). The cytoplasmic portions of the receptors include a con- treated with recombinant rat MOG (33). In the case of TLR2 served motif, termed the Toll/IL-1R (TIR) domain. The TIR deficiency, Prinz et al. (32) reported a normal pattern of suscep- domains of TLRs are homologous with the respective domain of tibility to MOG35–55 peptide-induced EAE. In contrast, Rag- the IL-1R and the cytoplasmic adaptor family. The TIR deficient hosts reconstituted with CD4+ T cells from TLR2- domains of the adaptor interact with the TIR domains deficient donors were less susceptible to the induction of EAE of the TLRs or IL-1R and trigger the activation of downstream with MOG peptide and adjuvant than recipients reconstituted with protein kinases and multiple transcription factors, including the wild type (WT) T cells (35). It is unclear whether these differences NF-kB family. All TLRs except TLR3 signal through the adaptor stem from variation in methods of induction or housing conditions. protein MyD88; TLR3 signals through a MyD88-independent, In summary, the role of TLR1 is unknown, and those of TLR2, -dependent pathway, and TLR4 uses both MyD88-depen- TLR4, and TLR9 are still controversial. In this article, we revisit dent and -independent pathways (16). this issue in a specific pathogen-free facility specifically con- To date, 10 human and 12 mouse TLRs have been identified, structed to minimize environmental variation. and each TLR recognizes a characteristic collection of microbial constituents. For example, TLR2 detects microbe-derived pepti- Materials and Methods doglycan and lipopeptides (17–19), TLR4 binds LPS (20), and Mice TLR5 binds bacterial flagellin (21). The range of microbial C57BL/6.Tlr12/2 and C57BL/6.Tlr62/2 mice, backcrossed to C57BL/6 products recognized by TLR is extended by the ability of some 2 2 2 2 mice six generations, and C57BL/6.Tlr2 / , C57BL/6.Tlr4 / , C57BL/ TLRs to heterodimerize. For example, the TLR1/TLR2 hetero- 6.Tlr92/2, and C57BL/6.Myd882/2 mice, backcrossed to C57BL/6 mice dimer binds triacylated lipopeptides, whereas the TLR2/TLR6 for 10 generations, were kindly provided by Dr. Shizuo Akira (Osaka Downloaded from heterodimer binds diacylated lipopeptides (22–24). TLR3, TLR7, University, Osaka, Japan) (19, 26, 36–39). 2/2 2/2 and TLR9 are expressed in endosomes and are ligated by microbial C57BL/6.Tlr1 and C57BL/6.Tlr6 mice were further backcrossed to C57BL/6 mice for a total of 10 generations before intercrossing to nucleic acids (25). TLR9, for example, recognizes bacterial CpG generate homozygous mutants. C57BL/6.Tlr22/2 and C57BL/6.Tlr92/2 DNA sequences (26). mice were intercrossed twice and homozygous breeders selected to obtain The role of TLR signaling in the pathogenesis and regulation of double-congenic TLR2/TLR9-deficient mice. All mice were bred and

CNS inflammation is far from clear. Eighteen patients with chronic maintained in the Immunogenetics Research Facility at James Cook http://www.jimmunol.org/ progressive MS were treated in an open phase I trial with the TLR3 University under specific pathogen-free conditions. The facility has strict environment-controlled mouse holding rooms that are HEPA-filter venti- ligand carboxymethylcellulose, polyinosinic-polycytidylic acid, and lated and maintained at 19 6 1.5˚C and 40–50% humidity. Food is sourced poly-L-lysine dsRNA (i.e., poly-ICLC). All patients produced sub- from a specialty composition-defined manufacturer (Specialty Feeds, Glen stantial amounts of IFN and experienced acute adverse effects, in- Forrest, Western Australia) that purchases their grain from a restricted cluding fever and transient worsening of neurologic symptoms. Of number of farms and subject it to gamma irradiation sterilization before delivery. A four-pass air filtration system, advanced drinking water ster- nine patients with rapid neurologic deterioration at the time of entry ilization and purification, as well as a unidirectional operator flow design into the study, only three had disease progression during treatment with shower-in access limits microbial infections of the mice. These (27). The possibility that TLR3 stimulation might be therapeutic is studies have been reviewed and approved by James Cook University An- supported by the findings of Touil et al. (28), who found that poly- imal Care and Ethics Committee. by guest on September 23, 2021 inosinic:polycytidylic acid treatment induced IFN-b secretion and Induction and scoring of active EAE suppressed EAE in the SJL/PLP139–151 model. The manipulation of TLR signaling in MS may carry significant risks, as the modulation Age-matched 7- to 13-wk-old mice were immunized s.c. at days 0 and 7 in inguinal regions and flanks with 200 mgMOG35–55 peptide (Auspep, of inflammatory in MS has previously resulted in un- Parkville, VIC, Australia) in 100 ml PBS emulsified in an equal volume (1:1) expected outcomes. For example, treatment of patients with rIFN-g of IFA (Sigma-Aldrich, Castle Hill, NSW, Australia) containing 0.5 mg exacerbated MS in 7 of 18 patients (29), and TNF neutralization was heat-killed M. tuberculosis H37RA (Difco; BD Diagnostic Systems, Sparks, associated with increased proportions of patients experiencing MD). Each animal received an i.p. injection of 250 ng PTX (List Biological Laboratories, Campbell, CA) at day of immunization and 2 d later [modified exacerbations in clinical trials (30) and was temporarily associated from Prinz et al. (32)]. with unexpected neurologic events suggestive of demyelination Mice were scored daily for clinical signs of EAE using the following during therapy for inflammatory arthritis (31). scale: 0 = no detectable signs of EAE; 0.5 = hunching; 1 = slight tail In studies of TLR-deficient, -targeted mutant mice, C57BL/ weakness; 1.5 = distal limp tail or hind-limb weakness; 2 = total tail pa- 6.Tlr6 mice were as susceptible as C57BL/6 mice to myelin oli- ralysis; 2.5 = hind leg totally paralyzed; 3 = both hind legs totally paralyzed; 3.5 = bilateral hind-limb paralysis in combination with unilateral fore-limb godendrocyte glycoprotein 35–55 (MOG35–55) peptide-induced paralysis or weakness of forelimbs; 4 = bilateral hind-limb and forelimb EAE and whole recombinant rat MOG-induced EAE (32, 33). paralysis; 5 = moribund state or death. In contrast, both Marta et al. (33) and Prinz et al. (32) found 2 2 Histopathology C57BL/6.Myd88 / mice to be resistant to EAE. The production of the inflammatory cytokines IL-6 and IL-23 by dendritic cells Brain and spinal cord (CNS) were collected on day 40 after EAE induction. (DCs), as well as the subsequent priming of Th17 cells, was Mice were euthanized by CO2 asphyxiation. The CNS was perfused with 24 ml ice-cold PBS to avoid leukocyte contamination from blood before MyD88 dependent (33). Conflicting results have been reported for being removed and fixed in 10% buffered formalin (Sigma-Aldrich). Tis- mice deficient in TLR2, TLR4, and TLR9. Whereas C57BL/6. sues were embedded in paraffin, and 4-mm sections were stained with Tlr4-deficient mice showed increased EAE susceptibility to H&E or Luxol fast blue (LFB) to assess the degree of infiltration and demyelination. Protocols were adapted from Humason (40) and IHC recombinant rat MOG-induced EAE (33), in the MOG35–55 peptide-induced model, the severity of disease was unaffected by World, Life Science Information Network (http://www.ihcworld.com/ _protocols/special_stains/fast_blue.htm). targeted deletion of Tlr4, unless a marginal dose of Mycobacte- Cell infiltration and defects in myelination of spinal cords were assessed rium tuberculosis was used in the inoculum, in which case the using the following semiquantitative scoring system (modified from Ref. severity of disease was sometimes reduced compared with C57BL/6 41): 0 = no lesions, no cell infiltration, and no reduction in LFB staining; in a mechanism that appeared to involve pertussis toxin (PTX) 1 = solitary lesions with cell infiltration of low cellular density with or 2/2 without mildly reduced LFB staining; 2 = two to three lesions per level signaling through TLR4 (34). Whereas C57BL/6.Tlr9 mice with moderate cell infiltration associated with reduced LFB staining; 3 = showed a decreased severity of disease after EAE induction with many lesions in almost all fields with extensive cell infiltration associated MOG35–55 (32), disease was exacerbated in TLR9-deficient mice with severe reduction in LFB staining. The Journal of Immunology 3

Isolation of CNS leukocytes and FcR block (rat anti-mouse CD16/26 [FcIII/II , clone 93]; Bio- Legend) and, in some cases, 50% mouse serum and 10% rat serum were Mice were euthanized by CO2 asphyxiation. The CNS was perfused with used to prevent nonspecific binding of Abs. If possible, a total of 1 million 24 ml ice-cold PBS to avoid leukocyte contamination from blood, before cells was analyzed per sample; a forward scatter area against forward scatter being removed and minced through a 180-mm wire stainless steel mesh. height gate was used to exclude doublets, and propidium iodide was used to 3 The dissociated material was centrifuged at 400 g for 15 min at 4˚C. gate out the dead cells from analyses. Samples were acquired using a CyAn Enzymatic digestion was performed with Collagenase Type II (0.5 ml/mg) ADP Flow Cytometer and analyzed with Summit software (Beckman and DNase I (500 U/ml) at 37˚C for 60 min. The resulting digest was Coulter, Gladesville, NSW, Australia). pelleted and resuspended in 30% isotonic Percoll (Amersham Biosciences, Rydalmere, NSW, Australia) underlaid with 70% isotonic Percoll and subsequently overlaid with PBS. The tube was centrifuged at 600 3 g for 25 min at 25˚C. Cells were collected between the 70 and 30% density layers [modified from Ford et al. (42)]. Cell suspensions Splenocyte and cell suspensions were prepared by gently grinding organs between two frosted microscope slides in FACS buffer (PBS containing 1 mM EDTA (Amresco, Solon, OH) and 10% (v/v) bovine serum (Invitrogen Life Technologies, Mulgrave, VIC, Australia) and 0.02% sodium azide (BDH, Poole, U.K.). Splenocytes were treated with RBC lysis buffer (Sigma-Aldrich).

DC culture Downloaded from DCs were derived from bone marrow. In brief, bone marrow cells were isolated from the tibias and femurs of mice by flushing with RPMI 1640 containing 5% (v/v) FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ ml streptomycin sulfate (Invitrogen Life Technologies), and 15% GM- CSF–containing supernatant of hybridoma cell line X63-AG8. After one wash, cells were seeded in a 24-well plate at a concentration of 1 3 106 cells in 500 ml and cultured at 37˚C for 7 d. http://www.jimmunol.org/ In vitro stimulation of splenocytes Isolated splenocytes were resuspended and plated at 5 3 106/ml in RPMI 1640 supplemented with 10% (v/v) FCS (PAA Laboratories, Morningside, QLD, Australia), 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin sulfate (Invitrogen Life Technologies). Cells were treated for 24 h with TLR1/2 synthetic tripalmi- toylated lipopeptide (Pam3CSK4), TLR4 agonist LPS, TLR9 agonist ODN1826, and/or TLR6/2 agonist FSL-1 (Invivogen Life Technologies) dissolved at a concentration of 100 mg/ml in endotoxin-free water. by guest on September 23, 2021 Splenocytes were stimulated with a range of agonist concentrations and an optimal concentration for each agonist was chosen: 500 ng/ml for Pam3CSK4, FLS-1, and LPS, and 5 mg/ml for CpG1826. The cells were collected and analyzed for cell surface markers by flow cytometry. Flow cytometry Directly labeled fluorescent Abs anti–CD11b-FITC, –PE-Cy7, or –allophycocyanin-Cy7 (clone M1/70), anti–TCRb-allophycocyanin or -FITC (clone H57-597), anti–CD4-Pacific blue or -V450 (clone RM4-5), anti-CD3 allophycocyanin (clone 145-2c11), anti–CD8-allophycocyanin- Cy7 (clone 53-6.7), anti-CD11c PE-Cy7 (clone HL3), anti–CD19-PE (clone 1D3), anti–CD45/B220-PerCP (clone RA3-B2), or anti–CD69- PE-Cy7 (clone H1.2F3) and anti–Ly6C-PerCP-Cy5.5 (clone AL-21) were obtained from BD Biosciences (San Diego, CA). Anti–CD86-allophycocyanin (clone GL1), anti–CD62L-PE-Cy7 (clone MEL-14), anti–TLR1-Alexa Fluor (clone TR23), anti–TLR2-FITC (clone 6C2), and anti–TLR4-PE (clone UT41) were obtained from eBioscience (San Diego, CA), and anti–CD11c-PE (clone N418) was obtained from BioLegend (San Diego, CA). CCR2 staining with mAbs MC-21 (43) was detected by donkey anti-rat Dylight 649 (Jackson Immunoresearch, West Grove, PA). CD45 staining with biotin-conjugated anti-mouse CD45 (clone 30-F11; BD Biosciences) and CD25 staining with biotin-conjugated anti-mouse CD25 (clone IHC; BD Biosciences) was detected by streptavidin-Pacific orange (Invitrogen Life Technologies). TLR6 staining with rabbit polyclonal anti-TLR6 Ab (Abnova, Taipei, Taiwan) was detected by goat anti–rabbit-FITC human and mouse absorbed (Southern Biotech, Birmingham, AL). TLR9 staining with biotin-conjugated anti-

TLR9 (clone m9.D6; eBioscience) was detected with streptavidin-PE FIGURE 1. Clinical course of active MOG35–55/CFA + PTX -induced (BD Biosciences). EAE in TLR-deficient C57BL/6 mice and C57BL/6 WT control mice. For intracellular staining of Foxp3, the cells were stained with surface Each data point represents the mean 6 SEM of at least five animals per markers, then fixed, permeabilized, and stained with anti-Foxp3 FITC (clone group. Statistical differences in the clinical score between the controls and FJK-16S; eBioscience). MyD88 expression was detected by intracellular the TLR-deficient C57BL/6 mice are indicated (*p , 0.05, uncorrected staining using the affinity-purified goat anti-mouse/rat MyD88 Ab (R&D Systems, Minneapolis, MN) followed by donkey anti-goat IgG-FITC Mann–Whitney U test; n = 5–10). Representative experiments are illus- mouse/human absorbed Ab (clone SC-2024; Santa Cruz Biotechnology, trated; data showing significant differences in clinical scores have been 2 2 2 2 Santa Cruz, CA). Cells were fixed and permeabilized in Cytofix/Cytoperm repeated at least once for female C57BL/6.Tlr2 / , C57BL/6.Tlr9 / , buffers (BioLegend) following the manufacturer’s protocol. C57BL/6.Tlr2/92/2 and C57BL/6.MyD882/2 mice (see Table I). 4 ROLE OF TLR IN EAE

Cytokine assays Passive EAE

Blood was collected by retro-orbital venipuncture, and plasma separated Donor mice were immunized with MOG35–55/CFA + PTX, as described and stored at 280˚C until assayed. Quantitative detection of plasma GM- earlier. The inguinal and axillary lymph nodes and were collected CSF, IFN-g, IL-1a, IL-2, IL-4, IL-5, IL-6, IL-10, IL-17, IL-18, IL-23, 11 d after immunization. Cells were cultured for 3 d in RPMI 1640 medium TNF, CXCL1/KC, and MIP-1b were performed by cytometric bead ar- (Invitrogen Life Technologies) containing 10% (v/v) FCS, 2 mM L-gluta- ray using Bender Medsystems (Vienna, Austria) mouse Th1/Th2 10plex, mine, 100 U/ml penicillin, 100 mg/ml streptomycin sulfate, 1 mM sodium Mouse IL-23, and mouse CXCL/KC, mouse IL-18, and mouse MIP-1 kits. pyruvate, and 2 mM 2-ME (Invitrogen Life Technologies), supplemented Samples were analyzed following the manufacturer’s protocol. The de- with 50 mg/ml MOG35–55 peptide (Auspep), 10 ng/ml IL-23 (eBioscience), tection limits were 10.9 pg/ml for GM-CSF, 6.5 pg/ml for IFN-g, 15.7 pg/ 10 ng/ml rIL-6 (Sigma-Aldrich or eBioscience), and 5 ng/ml recombinant ml for IL-1a, 8.8 pg/ml for IL-2, 0.7 pg/ml for IL-4, 4.0 pg/ml for IL-5, 22 human TGF-b1 (Jomar Bioscience, Kensington, South Australia). After 3 d, pg/ml for IL-6, 5.5 pg/ml for IL-10, 2.4 pg/ml for IL-17, 2.1 pg/ml for cells were harvested, washed, and dead cells were removed using Histo- TNF, 10.0 pg/ml for IL-18, 14.5 pg/ml for IL-23, 3.0 pg/ml for CXCL1/ paque 1083 (Sigma-Aldrich). Recipient mice were preconditioned with 3 Gy KC, and 14.9 pg/ml MIP-1b. gamma irradiation, and 4 h later, 5 3 107 cells were injected i.p. to induce Data were acquired on either a CyAn ADP (Beckman Counter) or an LSR passive EAE. Each animal received an i.p. injection of 350 ng PTX (List Fortessa (BD Biosciences) Flow Cytometer and analyzed with a multipa- Biological Laboratories) on the day of cell transfer and 2 d later. rameter logistic nonlinear regression model using Flow Cytomix Pro 2.2.1 Statistical analyses (Bender Medsystems) or GraphPad Prism version 5.00 (GraphPad for Mac Software; GraphPad, San Diego, CA; http://www.graphpad.com). Statistical differences in quantitative traits were compared by uncorrected Plasma samples were assayed in duplicates using an ELISA specific for Mann–Whitney U test or Kruskal–Wallis test with Dunn’s multiple- mouse IFN-b (PBL Biomedical Laboratories, Piscataway, NJ), to quanti- comparison posttest. Qualitative traits were compared by Fisher’s exact tate the amounts of IFN protein. The ELISA was performed in accordance test. Statistical calculations were performed using InStat 3.0b (1992–2003) with the manufacturer’s protocol and analyzed at an absorbance of 450 nm. and GraphPad Prism 4.0c (San Diego, CA) software. Differences were Downloaded from The limit of detection of IFN-b was 15.6 pg/ml. considered statistically significant when the p values were ,0.05.

Table I. Active EAE in TLR-deficient C57BL/6 mice and C57BL/6 WT control mice http://www.jimmunol.org/

Experiment No. Strain n Incidence (%) Onset (d) Mean Maximum Score CDI Female 1 C57BL/6 5 100 16.0 6 0.0 3.1 6 0.1 61.1 6 3.8 C57BL/6.Tlr12/2 6 100 15.2 6 0.2** 2.7 6 0.3 53.6 6 9.7 C57BL/6.Tlr22/2 6 100 15.7 6 0.8 1.6 6 0.3** 14.2 6 8.8* C57BL/6.Tlr42/2 6 100 15.7 6 1.6 3.2 6 0.4 64.0 6 19.2 C57BL/6.Tlr62/2 6 100 15.5 6 1.6 3.1 6 0.5 65.1 6 19.8 C57BL/6.Tlr92/2 6 33.3 17.0 6 1.2* 0.8 6 0.5** 8.5 6 7.8** 6 6 6 2 C57BL/6 6 100 13.7 0.3 2.9 0.3 57.7 10.3 by guest on September 23, 2021 C57BL/6.Tlr92/2 6 100 17.0 6 0.0** 2.2 6 0.2* 32.0 6 5.5* C57BL/6.Myd882/2 5 40 22.0 6 2.6** 1.1 6 0.4** 8.8 6 7.3** 3 C57BL/6 5 100 15.8 6 0.6 2.6 6 0.2 48.4 6 2.8 C57BL/6.Tlr22/2 5 100 22.8 6 3.1 1.4 6 0.3* 14.0 6 9.0 4 C57BL/6 6 100 16.2 6 0.3 2.4 6 0.2 46.9 6 4.3 C57BL/6.Tlr22/2 5 100 19.6 6 2.5 1.5 6 0.4 16.2 6 10.0 5 C57BL/6 8 100 15.9 6 0.5 2.9 6 0.1 53.3 6 3.7 C57BL/6.Tlr22/2 8 100 15.4 6 0.2 2.6 6 0.2 44.9 6 3.5 C57BL/6.Myd882/2 8 0 na 0.0 6 0.0*** 0.0 6 0.0*** 6 C57BL/6 6 100 17.3 6 0.8 2.7 6 0.2 45.5 6 5.5 C57BL/6.Tlr42/2 6 100 17.0 6 1.5 2.4 6 0.3 47.7 6 7.6 7 C57BL/6 6 100 17.0 6 0.8 2.6 6 0.3 48.0 6 8.0 C57BL/6.Tlr62/2 5 100 14.8 6 0.8 2.9 6 0.2 54.8 6 4.2 8 C57BL/6 8 100 15.4 6 0.2 2.9 6 0.1 56.0 6 5.0 C57BL/6.Tlr92/2 8 100 21.6 6 2.5*** 1.6 6 0.2*** 19.9 6 6.7*** C57BL/6.Tlr2/92/2 8 100 16.2 6 0.8 1.6 6 0.2*** 19.1 6 4.9*** 9 C57BL/6 5 100 14.8 6 1.1 2.7 6 0.3 53.1 6 9.9 C57BL/6.Tlr2/92/2 8 100 15.4 6 0.6 2.3 6 0.2 44.8 6 7.3 Male 1 C57BL/6 5 100 14.2 6 0.2 3.1 6 0.1 71.8 6 2.9 C57BL/6.Tlr12/2 6 100 14.7 6 0.2 3.7 6 0.3 77.3 6 8.7 C57BL/6.Tlr22/2 6 100 14.5 6 0.2 3.8 6 0.4 80.2 6 12.1 C57BL/6.Tlr92/2 6 100 14.5 6 0.4 3.2 6 0.2 60.2 6 4.3 2 C57BL/6 5 100 13.2 6 0.2 2.9 6 0.2 62.1 6 8.3 C57BL/6.Tlr42/2 6 100 13.2 6 0.2 3.5 6 0.0* 76.7 6 2.0 C57BL/6.Tlr62/2 6 100 13.0 6 0.0 3.3 6 0.2 67.3 6 4.3 3 C57BL/6 6 100 15.2 6 0.5 3.0 6 0.0 64.9 6 5.2 C57BL/6.Tlr92/2 6 100 15.3 6 0.2 2.8 6 0.2 60.5 6 3.7 4 C57BL/6 7 100 13.4 6 0.2 3.4 6 0.1 80.9 6 7.0 C57BL/6.Tlr22/2 8 100 13.8 6 0.3 3.8 6 0.2 74.5 6 8.3 C57BL/6.Tlr2/92/2 10 100 13.8 6 0.2 2.8 6 0.1** 49.4 6 5.4*** 5 C57BL/6 7 100 14.3 6 0.4 3.0 6 0.0 70.3 6 2.9 C57BL/6.Myd882/2 8 0 na 0.0 6 0.0*** 0.0 6 0.0***

TLR-deficient C57BL/6 and C57BL/6 WT control mice were immunized with MOG35–55 emulsified in IFA and M. tuberculosis. Statistical differences between the TLR-deficient C57BL/6 and C57BL/6 WT control mice are indicated. *p , 0.05, **p , 0.01, ***p , 0.001, uncorrected Mann–Whitney U test. na, not appropriate. The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 2. Histopathology of the spinal cords of male and female TLR-deficient C57BL/6 mice and C57BL/6 WT control mice after induction of active

MOG35–55/CFA + PTX-induced EAE. Mice were culled 40 d after immunization; the spinal cords were harvested, fixed, sectioned, and stained with H&E or LFB. Representative spinal cord sections are shown. A, Whole spinal cord sections representative of the four gradings of inflammation applied in the histopathological assessment using a modification of the scoring system of Hempel et al. (41). Rectangular boxes indicate the general area from which the high-power sections in B were sampled. B, Representative sections from male (left two columns) or female (right two columns) mice from groups (in rows listed from top to bottom) of PBS-treated C57BL/6 mice; C57BL/6 mice treated with PBS/CFA + PTX; and then MOG35–55/CFA + PTX-treated mice from the following strains: C57BL/6 (positive control), C57BL/6.Tlr12/2, C57BL/6.Tlr22/2, C57BL/6.Tlr42/2, C57BL/6.Tlr62/2, C57BL/6.Tlr92/2, C57BL/6. Tlr2/92/2, and C57BL/6.MyD882/2. Original magnification 3100. 6 ROLE OF TLR IN EAE

Results p , 0.001, Mann–Whitney U test); in the second experiment, two Generation and characterization of C57BL/6 TLR-deficient of five mice developed some weakness, but with delayed onset (p , mice 0.01, Mann–Whitney U test) and a very much reduced severity of disease (mean maximum disease score: p , 0.01; CDI: p , 0.01, To study the potential role of TLR ligation by bacterial products in Mann–Whitney U test) (Table I). These results confirm those the regulation of autoimmune CNS inflammation, we rederived previously published by Marta et al. (33) and Prinz et al. (32). C57BL/6 mice deficient for TLR1, TLR2, TLR4, TLR6, TLR9 and 2 2 Less severe EAE was also observed for C57BL/6.Tlr2 / mice MyD88 into an environmentally controlled “better than specific with a mean maximum score of 1.6 6 0.3 and a CDI of 14.2 6 8.8 pathogen-free” facility. These lines were originally generated on compared with WT C57BL/6 mice with a mean maximum score the 129 background by Takeda and Akira and subsequently of 3.1 6 0.1 (p , 0.01) and CDI of 61.1 6 3.8 (p , 0.05, Mann– backcrossed to C57BL/6 mice (19, 26, 36–39). Mice deficient for Whitney U test). Similar results were obtained in two of three TLR1 and TLR6 were supplied at the BC6 generation and were additional independent experiments, with a comparable trend seen backcrossed to BC10 before being intercrossed to produce 2/2 2/2 in the third (exact p value for CDI was 0.051) (Table I, experiment C57BL/6.Tlr1 and C57BL/6.Tlr6 breeder pairs. Those 2 2 4). A pooled analysis of the four cohorts of C57BL/6.Tlr2 / bearing mutations of Tlr2, Tlr4, Tlr9 and Myd88 were supplied at 2 2 2 2 mice and their respective control C57BL/6 mice showed highly BC10. C57BL/6.Tlr2 / and C57BL/6.Tlr9 / mice were inter- significant differences in mean maximum score (p , 0.0005) and crossed twice and the F2 progeny selected to produce a C57BL/6. 2/2 2/2 CDI (p , 0.0001, Mann–Whitney U test). Similarly, C57BL/6. Tlr2 .Tlr9 double-mutant line. 2 2 Tlr9 / mice exhibited a milder form of EAE compared with WT To confirm the lack of the corresponding TLR in the TLR- C57BL/6 mice, with a mean maximal score of 0.8 6 0.5 and Downloaded from deficient C57BL/6 mice, we compared their constitutive expres- a CDI of 8.5 6 7.8 (p , 0.01 for both comparisons, Mann– sion of TLR1, TLR2, TLR4, TLR6, TLR9 and MyD88 with that in Whitney U test), again with similar results obtained in two addi- WT C57BL/6 mice by flow cytometry. As expected, C57BL/6. 2 2 2 2 2 2 2 2 tional experiments (Fig. 1, Table I). Tlr1 / , C57BL/6.Tlr2 / , C57BL/6.Tlr4 / , C57BL/6.Tlr6 / , 2 2 2 2 Because active MOG-induced EAE appears to be at least partially C57BL/6.Tlr9 / and C57BL/6.Myd88 / mice did not express dependent on signaling through both TLR2 and TLR9, the severity the corresponding TLR in comparison with the C57BL/6 WT mice 2/2 of disease was also examined in C57BL/6.Tlr2/9 doubly de- http://www.jimmunol.org/ (Supplemental Fig. 1). 2 2 ficient mice. Compared with either C57BL/6.Tlr2 / or C57BL/6. In addition, we compared the responses to TLR ligands of 2 2 Tlr9 / mice, the mean onset of disease (16.2 6 0.8), the mean B cells from TLR-deficient C57BL/6 mice with those of C57BL/6 maximal score (1.6 6 0.2), and the CDI (19.1 6 4.9) were not WT mice (44). WT C57BL/6 splenic B cells showed an increased statistically significantly different (Fig. 1, Table I). Given the strong expression of the costimulatory molecules CD69 and CD86 after 2 2 suppression of EAE in C57BL/6.Myd88 / mice, these results in- stimulation with TLR2/1 ligand Pam3CSK4,TLR2/6ligandFSL-1, dicate that MyD88 plays a role in the pathogenesis of EAE in ad- TLR4 ligand LPS, TLR9 ligand CpG1826, a combination of dition to its role as the adaptor molecule for TLR2 and TLR9. FSL-1 and CpG1826, or a combination of all four TLR ligands. In Evidence of sexual dimorphism was sought by analyzing

contrast, TLR2/1 ligand Pam3SCK4 failed to induce expression by guest on September 23, 2021 MOG35–55/CFA-induced EAE in male C57BL/6 WT, C57BL/6. of either CD69 or CD86 in TLR1- or TLR2/9-deficient B cells; 2 2 2 2 2 2 2 2 Tlr1 / , C57BL/6.Tlr2 / , C57BL/6.Tlr4 / , C57BL/6.Tlr6 / , TLR2/6 ligand FSL-1 failed to induce expression in TLR2, TLR6, or TLR2/9-deficient B cells; TLR4 ligand LPS failed to induce expression in TLR4-deficient B cells; and CpG1826 failed to in- duce expression in TLR9- or TLR2/9-deficient splenic B cells (Supplementary Fig. 2). MyD88-deficient splenic B cells failed to respond to Pam3CSK4, FSL1, LPS, or CpG1826 (Supplemental Fig. 2). Taken together, these data confirmed that the TLR- deficient mice used in these studies do not express the corre- sponding TLR and do not respond to their corresponding ligands.

Role of TLRs in active EAE To determine whether signaling through TLRs plays a role in the induction and/or effector phase of EAE, we immunized female C57BL/6 WT, C57BL/6.Tlr12/2, C57BL/6.Tlr22/2, C57BL/6. FIGURE 3. Histopathological assessment of the severity of in- Tlr42/2 Tlr62/2 Tlr92/2 Tlr2/92/2 , C57BL/6. , C57BL/6. , C57BL/6. flammation in the spinal cords of male and female TLR-deficient C57BL/6 and C57BL/6.Myd882/2 mice (5–8 mice/group/experiment) with mice and C57BL/6 WT control mice after induction of active MOG35–55/ MOG35–55 peptide in IFA supplemented with M. tuberculosis CFA + PTX -induced EAE applying a modification of the scoring system (CFA) accompanied by i.p. PTX injection. The clinical signs of of Hempel et al. (41). The mean scores of individual mice are represented EAE induced in female C57BL/6 WT mice appeared within 13–19 by the data points and the mean value for each group by the horizontal bar. d, and all female C57BL/6 WT mice experienced development of The first and second groups of each panel are negative control C57BL/6 EAE. The maximum severity of disease was observed at day 21, mice treated with PBS or PBS/CFA + PTX, respectively, followed by followed by a plateau phase until the end of the experiment. No MOG35–55/CFA + PTX-treated C57BL/6 (positive control; closed circles), 2/2 2/2 significant differences were observed in the maximum score of the C57BL/6.Tlr1 (open circles), C57BL/6.Tlr2 (closed squares), C57BL/6.Tlr42/2 (open downward triangles), C57BL/6.Tlr62/2 (open disease or cumulative disease index (CDI) between C57BL/6 WT 2/2 2/2 2/2 upward triangles), C57BL/6.Tlr9 (closed downward triangles), C57BL/ mice and C57BL/6.Tlr1 , C57BL/6.Tlr4 and C57BL/6. 2/2 2/2 2/2 6.Tlr2/9 (closed upward triangles), and C57BL/6.MyD88 mice Tlr6 mice, ruling out a role for these receptors in EAE (Fig. 1, (closed diamonds). Statistical differences between the C57BL/6 control Table I). mice and the TLR-deficient mice are indicated (*p , 0.05, **p , 0.01, In contrast, MyD88-deficient mice were completely resistant to ***p , 0.001, Mann–Whitney U test for C57BL/6 versus TLR1, TLR2, EAE in one experiment (mean maximum score: p , 0.001; CDI: TLR4, TLR6, TLR9, TLR2/9, MyD88; n = 5–14). The Journal of Immunology 7

C57BL/6.Tlr92/2, C57BL/6.Tlr2/92/2, and C57BL/6.Myd882/2 WT mice developed significantly more severe clinical signs of mice (5–10 mice/group/experiment). After immunization with EAE than female C57BL/6 WT mice (mean maximum score: p , MOG35–55 in CFA accompanied by PTX injection, male C57BL/6 0.01; CDI: p , 0.0001; n = 30–55, Mann–Whitney U test; Fig. 1). Male C57BL/6 WT, C57BL/6.Tlr12/2, C57BL/6.Tlr42/2, and C57BL/6.Tlr62/2 mice developed EAE with the same onset of disease (13–17 d), maximum score (3.1 6 0.1, 3.7 6 0.3, 3.5 6 0.0, and 3.3 6 0.2, respectively), and CDI (71.8 6 2.9, 77.3 6 8.7, 76.7 6 2.0, and 67.3 6 4.3, respectively) as each other, as ob- served for the female groups of mice. Similarly, like their female counterparts, male C57BL/6.Myd882/2 mice showed no clinical signs of EAE. In contrast with C57BL/6.Tlr22/2 and C57BL/6. Tlr92/2 female mice, which showed less severe EAE, disease Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 5. Absolute numbers of CNS leukocyte subsets from CNS of male (left column) and female (right column) C57BL/6.Tlr22/2 (closed squares), C57BL/6.Tlr92/2 (closed downward triangles), C57BL/6.Tlr2/ 92/2 (closed upward triangles), C57BL/6.MyD882/2 (open diamonds), and FIGURE 4. Representative flow cytometry showing the proportions of WT C57BL/6 control mice (closed circles) 21 d after MOG/CFA immu- leukocytes from CNS of female C57BL/6.Tlr22/2, C57BL/6.Tlr92/2, nization. The cell numbers of individual mice are represented by the data C57BL/6.Tlr2/92/2, C57BL/6.MyD882/2 and WT C57BL/6 control mice points and the mean value for each group by the horizontal bar, with 4–8 21 d after MOG/CFA immunization. CD45hi and bTCR in combination with mice/group, with the exception of the PBS-treated control (n = 1; closed CD4 or CD8 were used to detect CD4 or CD8 T cells, respectively. Other diamonds). Statistical differences between the C57BL/6 control group and cell populations were identified based on the differential expression of the TLR-deficient C57BL/6 mice are indicated (*p , 0.05, **p , 0.01, CD45int, CD11c, and CD11b. The values in plots indicate mean 6 SD of ***p , 0.001, Kruskal–Wallis test with Dunn’s multiple-comparison four to six mice, with the exception of the PBS-treated control (n =1). posttest). 8 ROLE OF TLR IN EAE in male C57BL/6.Tlr22/2 and C57BL/6.Tlr92/2 mice was not Similarly, like their female counterparts, male C57BL/6. alleviated and showed the same pattern as that in male C57BL/6 Myd882/2 mice lacked inflammatory lesions in their spinal cords WT mice. Examination of the individual scores shows a very (p , 0.0001, Mann–Whitney U test) (Fig. 3). Significantly re- substantial overlap in clinical scores between the two sexes of WT duced infiltration was also found in male C57BL/6.Tlr2/92/2 mice mice. If a threshold in disease severity were responsible for the (2.15 6 0.23; p , 0.01, Mann–Whitney U test) compared with failure of Tlr2 or Tlr9 deletion to modulate EAE in male mice, then C57BL/6 WT mice (2.8 6 0.05) (Fig. 3). a reduction in the average scores of TLR-deficient mice should still Altogether, these results show that female C57BL/6.Tlr22/2 and be seen (reflecting reduced scores in a proportion of mice in the C57BL/6.Tlr92/2 mice, as well as female and male C57BL/6. group). Such a situation may be occurring in TLR9-deficient mice, Tlr2/92/2 mice, show less severe EAE signs, whereas female and which trend toward lower average scores, but not in TLR2-deficient male C57BL/6.Myd882/2 mice are resistant to EAE. mice, which do not show this trend. Phenotypic analysis of cells infiltrating the CNS of Although neither male C57BL/6.Tlr22/2 nor male C57BL/6. TLR-deficient mice in active EAE Tlr92/2 mice were protected, alleviation of EAE was observed in 2/2 male C57BL/6.Tlr2/92/2 double-mutant mice, with a maximum Potential explanations for the relative resistance of C57BL/6.Tlr2 , 2/2 2/2 2/2 EAE score of 2.8 6 0.1 and a CDI of 49.4 6 5.4 compared with C57BL/6.Tlr9 , C57BL/6.Tlr2/9 and C57BL/6.Myd88 WT male C57BL/6 mice (mean maximum score: 3.4 6 0.1; CDI: mice to active EAE induced with MOG35–55/CFA + PTX include 80.9 6 7.0; p , 0.01 and p , 0.001, respectively, Mann–Whitney effects on recruitment of inflammatory cells to the CNS and dif- U test). ferences in subsets recruited. To determine the nature of cells

infiltrating the CNS of TLR-deficient C57BL/6 mice, CNS in- Downloaded from Histological analysis of the role of TLRs in active EAE filtrating leukocytes of female and male C57BL/6, C57BL/6. C57BL/6 WT, C57BL/6.Tlr12/2, C57BL/6.Tlr22/2, C57BL/6. Tlr22/2, C57BL/6.Tlr92/2, C57BL/6.Tlr2/92/2 and C57BL/6. Tlr42/2, C57BL/6.Tlr62/2,C57BL/6.Tlr92/2, C57BL/6.Tlr2/92/2 Myd882/2 mice were compared by flow cytometric analysis 2/2 and C57BL/6.Myd88 mice (5–14 mice/group) were treated 21 d after immunization with MOG35–55/CFA + PTX (Fig. 4). We with MOG/CFA + PTX and killed ∼40 d postimmunization for observed a significant decrease in the total number of CD4+ and + 2/2

histological analysis, semiquantitative scoring, and statistical anal- CD8 T cells infiltrating the CNS of female C57BL/6.Myd88 http://www.jimmunol.org/ ysis. Representative histological sections of the spinal cord are mice compared with C57BL/6 mice (p , 0.001 and p , 0.01, shown in Fig. 2. Histological evaluation of spinal cord sections respectively, Kruskal–Wallis test with Dunn’s multiple-com- showed multiple inflammatory foci in female C57BL/6 WT mice, parison posttest; Fig. 5) and a significant reduction in CD8+ and the histopathological scores observed in spinal cords averaged T cells in female C57BL/6.Tlr92/2 mice (p , 0.05). Because 2.39 6 0.15. In contrast, female C57BL/6.Myd882/2 mice had no TLR2- and TLR9-deficient mice showed a trend to lower T cell inflammation in the spinal cord (histopathological score: 0; p , numbers, an additional cohort of C57BL/6, C57BL/6.Tlr22/2 and 0.0005, Mann–Whitney U test) (Fig. 3). Female C57BL/6.Tlr22/2 C57BL/6.Tlr92/2 mice was assessed and the data combined. (histopathological score: 0.93 6 0.52), C57BL/6.Tlr92/2 (histo- Numbers of CD4+ T cells infiltrating the CNS of female C57BL/6. 2/2 2/2 2/2 pathological score: 0.53 6 0.23), and C57BL/6.Tlr2/9 double- Tlr2 and C57BL/6.Tlr9 mice of the combined experiment by guest on September 23, 2021 mutant mice (histopathological score: 1.35 6 0.46) also had sig- were significantly decreased (p , 0.05 and p , 0.01, respectively, nificantly reduced scores (p , 0.05, p , 0.001, and p , 0.05, Kruskal–Wallis test with Dunn’s multiple comparison posttest; n = respectively, uncorrected Mann–Whitney U test) (Fig. 3). These 11–13), as were the numbers of CD8+ T cells infiltrating the CNS results confirmed the clinical assessments. of C57BL/6.Tlr92/2 mice (p , 0.01).

Table II. Passive EAE in TLR-deficient C57BL/6 mice and C57BL/6 WT control mice

Experiment No. Strain n Incidence (%) Onset (d) Mean Maximum Score CDI Female 1 C57BL/6 5 100 17.0 6 2.3 2.8 6 0.2 34.4 6 4.7 C57BL/6.Tlr22/2 3 0 na 0.0 6 0.0* 0.0 6 0.0* C57BL/6.Tlr92/2 5 80 13.8 6 1.8 1.6 6 0.4* 25.1 6 9.3 2 C57BL/6 6 100 23.8 6 0.5 2.7 6 0.2 21.8 6 2.1 C57BL/6.Tlr22/2 4 0 na 0.0 6 0.0** 0.0 6 0.0** C57BL/6.Tlr92/2 4 100 23.3 6 1.3 1.5 6 0.2** 4.4 6 1.5** 3 C57BL/6 4 100 19.5 6 1.8 2.5 6 0.5 33.4 6 7.9 C57BL/6.Tlr22/2 6 0 na 0.0 6 0.0** 0.0 6 0.0** C57BL/6.Tlr92/2 5 100 19.4 6 2.8 2.1 6 0.5 26.0 6 11.1 Combined C57BL/6 15 100 20.4 6 1.2 2.7 6 0.2 29.1 6 3.0 C57BL/6.Tlr22/2 13 0 na 0.0 6 0.0*** 0.0 6 0.0*** C57BL/6.Tlr92/2 14 92.8 18.8 6 0.5 1.8 6 0.2** 19.5 6 5.5 Male 1 C57BL/6 4 100 23 6 1.0 2.6 6 0.4 16.9 6 3.2 C57BL/6.Tlr22/2 3 0 na 0.0 6 0.0 0.0 6 0.0 2 C57BL/6 6 100 21 6 0.2 3.0 6 0.0 22.8 6 2.4 C57BL/6.Tlr22/2 6 0 na 0.0 6 0.0** 0.0 6 0.0** Combined C57BL/6 10 100 21.8 6 0.5 2.8 6 0.1 18.9 6 1.9 C57BL/6.Tlr22/2 9 0 na 0.0 6 0.0*** 0.0 6 0.0*** Passive EAE was induced by adoptive transfer of in vitro-activated lymph node and splenic T cells from MOG/CFA- immunized C57BL/6 mice into C57BL/6.Tlr22/2, C57BL/6.Tlr92/2, and C57BL/6 control WT mice. Statistical differences between recipients of cells transferred from TLR-deficient C57BL/6 and C57BL/6 WT control mice are indicated. *p , 0.05, **p , 0.01, ***p , 0.001, uncorrected Mann–Whitney U test. na, not appropriate. The Journal of Immunology 9

Whereas male C57BL/6.Myd882/2 mice showed a statistically significant decrease in only the total number of infiltrating CD4+ T cells (p , 0.01, Kruskal–Wallis test with Dunn’s multiple- comparison posttest; n = 7–8/group; Fig. 5), neither the C57BL/ 6.Tlr22/2 nor C57BL/6.Tlr92/2 male mice showed any significant difference in CNS infiltration, consistent with the clinical data. As the C57BL/6.Tlr2/92/2 double-mutant and C57BL/6.Myd882/2 male mice showed trends from the control mice in terms of some infiltrating cell subsets, an additional cohort of C57BL/6, C57BL/ 6.Tlr2/92/2, and C57BL/6.Myd882/2 male mice was assessed and the data combined. Numbers of CD4+ and CD8+ T cells in- filtrating the CNS of male C57BL/6.Myd882/2 mice of the com- bined experiment were significantly decreased (p , 0.001 and p , 0.05, respectively, Kruskal–Wallis test with Dunn’s multiple comparison posttest; n = 13/group). CD45 staining in combination with CD11b and/or CD11c was used to distinguish between DCs (CD45intSShiCD11c+), / (CD45hiSShiCD11b+), and inflammatory monocytes (CD45hiSShiCD11c+CD11b+) (Fig. 4). In contrast with Downloaded from female C57BL/6 mice, which developed severe EAE with high numbers of infiltrating cells, female C57BL/6.Myd882/2 mice had statistically significantly fewer infiltrating DCs, macrophages, and inflammatory monocytes (p , 0.05, p , 0.01, and p , 0.001, respectively, Kruskal–Wallis test with Dunn’s multiple compari-

son posttest; n = 4–6/group). Analysis of the combined experiment http://www.jimmunol.org/ also found a significant reduction in CNS infiltrating macrophages in both C57BL/6.Tlr22/2 and C57BL/6.Tlr92/2 female mice (p , 0.05 in both cases) and a significant reduction in inflammatory monocytes in female C57BL/6.Tlr22/2 mice (p , 0.01). As for the T cells, male mutant mice tended to a phenotype more similar to that of WT mice in terms of / infiltration. Only male C57BL/6.Myd882/2 mice showed statisti- cally significant decreases in numbers of CNS-infiltrating DCs

FIGURE 6. Clinical course of passive EAE induced by adoptive transfer by guest on September 23, 2021 and macrophages (p , 0.05 and p , 0.01, respectively, Kruskal– of in vitro-activated lymph node and splenic T cells from MOG35–55/CFA + Wallis test with Dunn’s multiple comparison posttest; n = 7–8/ PTX immunized C57BL/6 mice into C57BL/6.Tlr22/2, C57BL/6.Tlr92/2, group) (Fig. 5). Analysis of the combined experiment confirmed and C57BL/6 control WT mice. A, Data pooled from three experiments reduction of DCs, macrophages, and inflammatory monocytes in using female donors and recipients, with three to six replicates/group/ex- 2 2 only the male C57BL/6.Myd88 / mice (p , 0.05, p , 0.01, and periment, as indicated in Table II. Each data point represents the mean 6 p , 0.05, respectively, Kruskal–Wallis test with Dunn’s multiple SEM of $13 animals/group. B, The effect of direction of adoptive transfer 2 2 comparison posttest; n = 13/group). between C57BL/6.Tlr2 / and C57BL/6 control WT mice. Each data point Thus, the milder clinical disease and resistance to EAE seen in represents the mean 6 SEM of 6–10 animals/group. C, Adoptive transfer 2 2 2 2 2 2 / / / of in vitro-activated lymph node and splenic T cells from MOG35–55/CFA + female C57BL/6.Tlr2 , C57BL/6.Tlr9 , C57BL/6.Tlr2/9 2/2 and C57BL/6.Myd882/2 mice and male C57BL/6.Myd882/2 mice PTX immunized male C57BL/6 mice into male C57BL/6.Tlr2 and C57BL/6 control WT mice (n = 6). Statistical differences in the clinical is associated with decreased numbers of CNS-infiltrating cells score from the C57BL/6 positive control group are indicated in each panel compared with C57BL/6 WT mice. (*p , 0.05, uncorrected Mann–Whitney U test). Role of TLRs in passive EAE Because the induction of EAE in this model involves the use of killed diation triggers the release of intrinsic TLR ligands (51), as well as M. tuberculosis as a component of the adjuvant, and M. tuberculosis the translocation of commensal gut microflora into mesenteric possesses for TLR2 (45–48), TLR4 (47, 49), and TLR9 lymph nodes and elevation of circulating LPS (52). We therefore (50), one potential explanation for the inhibitory effect of targeted examined the dependence on irradiation, PTX, and TLR4 of passive deletion of TLR2 and TLR9 is that suppression of TLR signaling EAE induced by adoptive transfer of in vitro activated leukocytes via these receptors has inhibited the adjuvanting action of CFA. harvested from MOG35–55 peptide/CFA + PTX immunized mice. Although this finding would be of significant interest, its relevance Four groups of six C57BL/6 mice were compared; all received to human health is unclear. The effects of targeted deletion of TLR2 in vitro-activated donor leukocytes. The group that was precondi- and TLR9 were therefore examined in the passive model of EAE. tioned by irradiation and received PTX injections (i.e., the standard Passive EAE is a model in which disease is adoptively transferred protocol) showed the most severe clinical signs. Although the other from an actively immunized donor. A number of refinements have three groups also developed EAE, the group that lacked PTX been proposed to enhance the model, including mild (3 Gy) gamma injections had the mildest disease. Although clinical signs were also 2 2 irradiation and pretreatment with PTX of the recipient. Kerfoot et al. alleviated in the C57BL/6.Tlr4 / mice, the inhibition of disease (34) proposed that the action of PTX in active EAE is mediated by was not as great as in the mice that lacked PTX, indicating that, if TLR4. Although we found no significant role for TLR4 in active PTX does act through TLR4, it is also likely to have activities EAE, it remained possible that it did contribute to the activity of mediated by other mechanisms in this model. The group that re- PTX in a passive model. Similarly, it has been proposed that irra- ceived PTX but not 3 Gy irradiation also showed some alleviation 10 ROLE OF TLR IN EAE of disease severity, similar in intensity to that seen in C57BL/6. of TLR2 in passive EAE, we performed adoptive transfers in which Tlr42/2 mice (Supplemental Fig. 3). TLR2 was deleted from either the donor cells, the recipient, both, or Passive EAE was induced in 3 Gy irradiated, PTX-treated female neither. Again, adoptive transfer of WT into C57BL/ C57BL/6.Tlr22/2, C57BL/6.Tlr92/2 and C57BL/6 recipients by 6.Tlr22/2 mice resulted in no disease, in stark contrast with the i.p. injection of in vitro-activated lymph node cells and splenocytes transfer of aliquots of the same pool of cells into WT mice. The harvested from C57BL/6 CFA/MOG + PTX immunized mice adoptive transfer of C57BL/6.Tlr22/2 cells into either WT or 11 d postimmunization. Data were pooled from 3 independent ex- C57BL/6.Tlr22/2 recipients resulted in alleviated disease (Fig. periments with 3–6 replicates/group/experiment (total n = 13–15/ 6B). These results are consistent with previously published com- group) (Table II). In none of the experiments did lymphocytes from parisons of passive EAE induced by adoptive transfer of WT or C57BL/6 mice adoptively transfer EAE into C57BL/6.Tlr22/2 TLR2-deficient T cells (35) and confirm that TLR2 plays a role in mice, resulting in a highly significant and highly reproducible exacerbating the severity of EAE. In this context, the prevention of decrease in mean maximum score and CDI compared with WT disease in TLR2-deficient recipients of WT cells suggests that the recipients (p , 0.001 for both, uncorrected Mann–Whitney U test) presence of the receptor at induction of disease creates a de- (Fig. 6A, Table II). Suppression of the mean maximum score of pendence on TLR2 signaling in the effector phase (Fig. 6C). C57BL/6.Tlr92/2 recipients occurred in two of the three trials and Adoptive transfer of in vitro-activated cells from male C57BL/6 was confirmed in the pooled analysis (p , 0.01, uncorrected mice into male C57BL/6.Tlr22/2 recipients also failed to induce Mann–Whitney U test) (Fig. 6A, Table II). Although the onset of disease (Fig. 6C), consistent with this hypothesis. disease in C57BL/6.Tlr92/2 mice consistently appeared earlier Phenotypic analysis of cells infiltrating the CNS of than that in C57BL/6 control mice in each trial, this difference did Downloaded from TLR-deficient mice in passive EAE not reach significance in any trial or in the analysis of pooled data. The greater effect of TLR2 deficiency on passive EAE than on Flow cytometric analysis of CNS-infiltrating leukocytes from 2/2 2/2 active disease was unexpected. To gain further insight into the role female C57BL/6.Tlr2 , C57BL/6.Tlr9 and C57BL/6 WT http://www.jimmunol.org/

FIGURE 7. Representative flow cytom- etry of brain and spinal cord leukocytes 2/2 from female C57BL/6.Tlr2 , C57BL/6. by guest on September 23, 2021 Tlr92/2 and C57BL/6 WT control mice 34 d after adoptive transfer of C57BL/6 MOG- reactive leukocytes. A, CD45 and SS area were used to distinguish between lym- phocytes (high CD45 expression and low SS area) and monocytes/macrophages/ DCs (high to intermediate CD45 and high SS area). Lymphocytes were stained with bTCR, CD4, and CD8 to determine major T cell subsets, whereas the APC population was stained with Ly6C, CD11b, and CD11c. B, Total numbers of CD4, CD8 T cells and inflammatory macrophages and mean fluo- rescence intensity (MFI) of CD11c on my- eloid DCs in the CNS of female C57BL/6. Tlr22/2 (closed squares), C57BL/6.Tlr92/2 (closed triangles), and C57BL/6 WT control mice (closed circles). The values in plots indicate mean 6 SD of four to six mice, with the exception of the PBS control (n = 1; closed diamonds). Statistical differences between the C57BL/6 control group and C57BL/6.Tlr22/2 or C57BL/6.Tlr92/2 mice are indicated (*p , 0.05, **p , 0.01, Mann–Whitney U test; n = 4–6) (B). The Journal of Immunology 11 control mice was performed 34 d after adoptive transfer of MOG- of activation and expansion of autoreactive T cells via the activity reactive leukocytes from female C57BL/6 donors. CD45 and side of regulatory T cells (Tregs). Because prior flow cytometric scatter (SS area) were used to distinguish between lymphocytes characterization of NKT cell numbers and subsets in TLR- (high CD45 expression and low SS area) and monocytes/ deficient C57BL/6 mice had failed to identify any significant macrophages/DCs (high to intermediate CD45 and high SS area) differences (data not shown), we examined the numbers of splenic (Fig. 7A). Coexpression of CD45hi and bTCR in combination with CD4+CD25+Foxp3+ Tregs of female C57BL/6.Tlr22/2, C57BL/6. CD4 or CD8 were used to detect CD4+ or CD8+ T cells (Fig. 7A). Tlr92/2 and C57BL/6 recipients of MOG-reactive leukocytes, Consistent with the clinical data, C57BL/6.Tlr22/2 recipients had 34 d after adoptive transfer. Because expression of CD62L has fewer CNS-infiltrating CD4+ T cells (p , 0.01, pairwise Mann– been reported to be associated with suppressive activity of Tregs Whitney U test, n = 4–6/group) (Fig. 7B) and reduced activation in vivo in three models of autoimmune/inflammatory disease, in- (CD11c expression levels) of myeloid DCs (*p , 0.05). Numbers cluding MOG/CFA + PTX-induced EAE (56–59), we specifically of inflammatory (CCR2+Ly6C+CD11b+) macrophages did not examined the CD62L-expressing subset of splenic Tregs. Whereas differ significantly (Fig. 7B). CD4+CD25+Foxp3+ Tregs were inconsistently increased in At least two mechanisms could account for the reduced re- spleens of C57BL/6.Tlr22/2 recipients of WT cells, absolute cruitment of CD4 T cells to the brains of C57BL/6.Tlr22/2 mice in numbers of CD62L-expressing Tregs were significantly increased this system. Recruitment of leukocytes to the CNS is at least partly in C57BL/6.Tlr22/2, but not C57BL/6.Tlr92/2 recipients of dependent on the expression of CCR2 (53, 54). CCR2 levels were MOG-reactive leukocytes compared with WT recipients (p , examined on the infiltrating CD4 T cells of female C57BL/6.Tlr22/2, 0.01, Mann–Whitney U test, n = 4–6/group) (Fig. 8). This pattern 2 2 C57BL/6.Tlr9 / and C57BL/6 recipients of MOG-reactive leu- was also seen in the male transfers; numbers of CD62L-expressing Downloaded from kocytes, but virtually no expression was found (data not shown). Tregs were significantly increased in C57BL/6.Tlr22/2 mice (p , This was not due to technical difficulties because CCR2 expres- 0.05; Mann–Whitney U test, n = 6/group; data not shown). sion on inflammatory macrophages was easily detected (Fig. 7). In an attempt to identify a molecular mechanism for the in- Because CCR2 expression on T cells may be downregulated after creased numbers of central (CD62L+) Tregs in C57BL/6.Tlr22/2 recruitment to the CNS (55), its levels of expression were com- passive EAE recipients, we examined plasma cytokine levels of

pared on splenic CD4 T cells of TLR-deficient mutants, but again IL-1, IL-2, IL-4, IL-5, IL-6, IL-10, IL-17, IFN-g, TNF, and GM- http://www.jimmunol.org/ no difference was found (data not shown). CSF at 10 or 34 d after adoptive transfer in two independent experiments. As a generalization, most of the cytokines were undetectable in the plasma of most mice. A significant exception Role of central regulatory T cells in modulating passive EAE in 2/2 to this was IL-6, which was detected in 4 of 6 C57BL/6 recipients C57BL/6.Tlr2 mice at 10 d and 3 of 5 at 34 d, but was undetectable in the plasma of An alternative explanation for the reduced recruitment of CD4+ any C57BL/6.Tlr22/2 mice (7/11 versus 0/9; p , 0.005, Fisher’s T cells to the brains of C57BL/6.Tlr22/2 mice is the suppression exact test). by guest on September 23, 2021

FIGURE 8. Representative flow cytometry of splenic Treg cells from female C57BL/6.Tlr22/2, C57BL/6.Tlr92/2 and C57BL/6 WT control mice 34 d after adoptive transfer of MOG-reactive leukocytes. A, Lymphocytes were gated by forward scatter and SS and cell clusters excluded by forward scatter height versus forward scatter area gating. Dead cells were excluded by propidium iodide staining. Viable cells were gated on CD3+ and CD4+, and Tregs were identified by CD25 and Foxp3 expression. The histograms show the expression of CD62L on Treg cells. B, Total numbers of CD4 T cells, Treg cells and CD62L-expressing Tregs in the of C57BL/6.Tlr22/2, C57BL/6.Tlr92/2 and C57BL/6 WT control mice. At least four mice per group were analyzed and the mean value of each group is indicated as a bar. Statistical differences between the C57BL/6 control group and C57BL/6.Tlr22/2 or C57BL/6.Tlr92/2 mice are indicated (**p , 0.01, Mann–Whitney U test; n = 4–6) (B). 12 ROLE OF TLR IN EAE

Discussion more readily affected by environmental change (77). Although Although encephalomyelitis can be experimentally induced in these data do not contribute to understanding the greater suscep- laboratory animals in the absence of adjuvant, the severity and tibility of women to autoimmunity, they do provide a potential incidence of disease is greatly enhanced by its addition (reviewed in mechanism by which environmental factors could play a greater Ref. 60). Although the conventional model applies CFA as an role in the prevalence of disease in women than in men. adjuvant, zymosan (a ligand for TLR2 and TLR6) (61), CpG- Although it could be postulated that in active EAE the role of ODN (TLR9) (62), and/or LPS (TLR4) (35, 61, 63) have been TLR2 and TLR9 signaling was merely to mediate the adjuvanting successfully (but inconsistently) (61, 63, 64) substituted for killed activity of CFA (45–48, 50), the demonstration of inhibition of M. tuberculosis in the induction of EAE, indicating the capacity of passive (adoptively transferred) EAE in targeted deletional TLR signaling to promote the disease. Paradoxically, LPS can also mutants for TLR2 and TLR9 not only confirms the importance of inhibit the induction of conventional CFA-induced EAE (65, 66). these TLRs in the pathogenesis of EAE, but also implicates these In this article, we have shown that, although targeted deletion of pathways in ongoing immune control of the effector phase of the TLR1, TLR4, or TLR6 did not affect the incidence or severity of autoimmune response. Although Prinz et al. (32) also reported a partial dependency on TLR9 expression in the adoptive transfer MOG35–55 peptide-induced EAE, signaling via MyD88 is neces- sary, and TLR2 and TLR9 contribute to the severity of the disease. recipient, our finding of complete dependency on TLR2 expres- The dependence of EAE induction on MyD88 signaling has been sion in recipients of WT cells is novel but consistent with Rey- reported previously (32, 33, 67), as has the lack of an EAE phe- nolds et al.’s report (35) that bone marrow chimeras reconstituted notype in TLR6-deficient mice (33); the role of TLR1 in EAE has with TLR2-deficient marrow expressed ameliorated disease after not been previously reported. TLR2 can signal as a heterodimer the induction of active EAE. Downloaded from with TLR1 or TLR6; nevertheless, functional activity of TLR2 The dependency of passive induction of EAE on TLR2 ex- homodimers acting in the absence of either TLR1 or TLR6 has pression in the recipient indicates the presence of tonic signaling been reported (68) and may be responsible for the dependence of through the receptor. This could occur via endogenous or exoge- EAE on TLR2 expression in our hands. Alternately, functional re- nous ligands. Although many putative endogenous ligands for dundancy between the effects of ligands for TLR2/1 and TLR2/6 TLR2 have been proposed, none of them meets the reasonable are possible but difficult to test, because Tlr1 and Tlr6,thegenetic criteria of purity and validation required to demonstrate their ability http://www.jimmunol.org/ loci encoding these two TLR, are in strong linkage disequilibrium. to directly bind and signal through TLR2 (78). In contrast, mi- Although we found that TLR4 deficiency did not affect the crobial products from intestinal microflora have been detected in severity of active EAE, Marta et al. (33) reported mild exacerbation the blood of healthy humans (79) and mice (80). Specifically, the and Kerfoot et al. (34) reported inconsistent patterns of disease mucosal translocation and systemic circulation of enterically de- (protection, amelioration, or normal severity) in TLR4-deficient rived peptidoglycan was confirmed using an NF-kB reporter cell mutants. This variability is in keeping with the inconsistent action line transfected with Nod1 and by comparing the phagocytic of LPS when used as an adjuvant in the induction of disease, as priming of bone marrow in WT and Nod1-deficient targeted mutant mice (80). Extracellular recognition of high m.w. well as its variable ability to prevent EAE (65), and may reflect the polymeric peptidoglycan is mediated by CD14 and TLR2 (17–19, by guest on September 23, 2021 effects of natural LPS exposure early in life (69), possibly a con- 81, 82). sequence of the induction of tolerance to LPS induced by previous To our knowledge, we show for the first time, in the context of exposure (70). Similarly, although we found a partial dependence CNS autoimmune inflammatory disease, that TLR2 signaling is of both TLR2 and TLR9 signaling for the full expression of active associated with detectable levels of circulating IL-6, reduced EAE, Marta et al. (33) reported slight exacerbation of disease in numbers of central (CD62L-expressing) Tregs, and increased re- TLR9-deficient mice, whereas Prinz et al. (32) reported mild al- cruitment of activated CD4 T cells to the brain. EAE is suppressed leviation and found no effect of TLR2 deficiency. Nevertheless, by the activity of Tregs (83–86), including central Tregs (54, 87), the results published in this study were highly reproducible and + via the action of IL-10 (84, 86, 88). Tregs are induced by CD4 are consistent with the known roles of TLR2 and TLR9 in me- T cell activation in the presence of TGF-b (86), and their in- diating many of the adjuvant-like effects of killed M. tuberculosis. duction is inhibited by IL-6 via trans signaling of soluble IL-6R The observation that although MyD88-deficient mice were into T cells causing SMAD7 suppression of TGF-b signaling (89). completely protected from the induction of active EAE, TLR2/9 IL-6 is a proinflammatory growth and survival factor whose ex- doubly deficient mice showed no more resistance than TLR2 or pression is induced by TLR2 ligation (19) and is, in part, con- TLR9 singly deficient mice, indicating that MyD88 plays a role in the trolled by the IL-6–NF-kB transcription factor binding site 59 to pathogenesis of EAE in addition to its role as the adaptor molecule the gene’s transcription initiation site (90). IL-6–deficient targeted for TLR2 and TLR9. This result could either be explained by a degree mutant mice are completely resistant to the induction of EAE by of redundancy between TLR2/TLR9 and TLR4 signaling or by MOG/CFA + PTX (91). dependence of EAE induction on IL-18 and/or IL-1 signaling, which These data therefore provide molecular insight into the differ- is also MyD88 dependent, as has been shown for the analogous ential environment-dependent susceptibility of females to CNS animal model experimental autoimmune uveitis (71). autoimmune inflammatory disease: female C57BL/6 mice were Sexual dimorphism was observed in this study in the active EAE dependent on TLR2 signaling driving the production of IL-6, which experiments, particularly in TLR2-deficient mice, but also to inhibited the induction of central Tregs, thereby enhancing the a lesser extent in TLR9-deficient and TLR2/9 doubly deficient proliferation and/or inhibiting the cell death of encephalitogenic mice. Although sexual dimorphism of TLR effects on disease has CD4+ T cells. not been previously reported, postpubertal mice show sex differ- ences in cytokine responses to TLR ligands, with females pro- Acknowledgments ducing less IL-1 in response to the TLR2 ligand We thank Dr. Shahead Chowdhury for contributing to the backcrossing and (72), and it is a typical feature of both EAE (73–76) and MS genotyping of the lines studied and Roby Jose with some of the cell sus- [reviewed in Orton et al. (77)]. Paradoxically, although the in- pensions. We are grateful for corrections by Margaret Jordan and Benjamin cidence and severity of MS is generally higher in women, it is also Crowley, the advice of Drs. Clett Erridge and Ashley Mansell, the help and The Journal of Immunology 13 suggestions of Phil Fromm, and the invaluable assistance of the animal 25. Kawai, T., and S. Akira. 2005. Pathogen recognition with Toll-like receptors. attendants of the Comparative Genomics Centre: Nicole Fraser, Kylie Rob- Curr. Opin. Immunol. 17: 338–344. 26. Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, ertson, Jo-Anne Diaz, and Kirilly O’Reilly. K. Hoshino, H. Wagner, K. Takeda, and S. Akira. 2000. A Toll-like receptor recognizes bacterial DNA. Nature 408: 740–745. [Published erratum appears in 2001 Nature 409: 646.] Disclosures 27. Bever, C. T., Jr., A. M. Salazar, E. Neely, B. E. Ferraraccio, J. W. Rose, The authors have no financial conflicts of interest. H. F. McFarland, H. B. Levy, and D. E. McFarlin. 1986. Preliminary trial of poly ICLC in chronic progressive multiple sclerosis. Neurology 36: 494–498. 28. Touil, T., D. Fitzgerald, G. X. Zhang, A. Rostami, and B. Gran. 2006. Cutting Edge: TLR3 stimulation suppresses experimental autoimmune encephalomy- References elitis by inducing endogenous IFN-beta. J. Immunol. 177: 7505–7509. 1. Sawcer, S. 2008. The complex genetics of multiple sclerosis: pitfalls and pros- 29. Panitch, H. S., R. L. Hirsch, A. S. Haley, and K. P. Johnson. 1987. Exacerbations of pects. Brain 131: 3118–3131. multiple sclerosis in patients treated with gamma interferon. Lancet 329: 893–895. 2. Strachan, D. P. 1989. Hay fever, hygiene, and household size. BMJ 299: 1259– 30. The Lenercept Multiple Sclerosis Study Group and The University of British 1260. Columbia MS/MRI Analysis Group. 1999. TNF neutralization in MS: results of 3. van den Biggelaar, A. H., L. C. Rodrigues, R. van Ree, J. S. van der Zee, a randomized, placebo-controlled multicenter study. Neurology 53: 457–465. Y. C. Hoeksma-Kruize, J. H. Souverijn, M. A. Missinou, S. Borrmann, 31. Mohan, N., E. T. Edwards, T. R. Cupps, P. J. Oliverio, G. Sandberg, H. Crayton, P. G. Kremsner, and M. Yazdanbakhsh. 2004. Long-term treatment of intestinal J. R. Richert, and J. N. Siegel. 2001. Demyelination occurring during anti-tumor helminths increases mite skin-test reactivity in Gabonese schoolchildren. J. In- necrosis factor alpha therapy for inflammatory arthritides. Arthritis Rheum. 44: fect. Dis. 189: 892–900. 2862–2869. 4. Shaheen, S. O., P. Aaby, A. J. Hall, D. J. Barker, C. B. Heyes, A. W. Shiell, and 32. Prinz, M., F. Garbe, H. Schmidt, A. Mildner, I. Gutcher, K. Wolter, M. Piesche, A. Goudiaby. 1996. Measles and atopy in Guinea-Bissau. Lancet 347: 1792– R. Schroers, E. Weiss, C. J. Kirschning, et al. 2006. Innate immunity mediated 1796. by TLR9 modulates pathogenicity in an animal model of multiple sclerosis. J. 5. Rook, G. A., and J. L. Stanford. 1998. Give us this day our daily germs. Clin. Invest. 116: 456–464. Immunol. Today 19: 113–116. 33. Marta, M., A. Andersson, M. Isaksson, O. Ka¨mpe, and A. Lobell. 2008. Un- Downloaded from 6. Rook, G. A., G. Ristori, M. Salvetti, G. Giovannoni, E. J. Thompson, and expected regulatory roles of TLR4 and TLR9 in experimental autoimmune en- J. L. Stanford. 2000. Bacterial vaccines for the treatment of multiple sclerosis cephalomyelitis. Eur. J. Immunol. 38: 565–575. and other autoimmune disorders. Immunol. Today 21: 503–508. 34. Kerfoot, S. M., E. M. Long, M. J. Hickey, G. Andonegui, B. M. Lapointe, R. C. 7. Sheikh, A., L. Smeeth, and R. Hubbard. 2003. There is no evidence of an inverse O. Zanardo, C. Bonder, W. G. James, S. M. Robbins, and P. Kubes. 2004. TLR4 relationship between TH2-mediated atopy and TH1-mediated autoimmune dis- contributes to disease-inducing mechanisms resulting in central nervous system orders: lack of support for the hygiene hypothesis. J. Allergy Clin. Immunol. 111: autoimmune disease. J. Immunol. 173: 7070–7077. 131–135. 35. Reynolds, J. M., B. P. Pappu, J. Peng, G. J. Martinez, Y. Zhang, Y. Chung, L. Ma,

8. Ristori, G., M. G. Buzzi, U. Sabatini, E. Giugni, S. Bastianello, F. Viselli, X. O. Yang, R. I. Nurieva, Q. Tian, and C. Dong. 2010. Toll-like receptor 2 http://www.jimmunol.org/ C. Buttinelli, S. Ruggieri, C. Colonnese, C. Pozzilli, and M. Salvetti. 1999. Use signaling in CD4(+) T lymphocytes promotes T helper 17 responses and reg- of Bacille Calmette-Gue`rin (BCG) in multiple sclerosis. Neurology 53: 1588– ulates the pathogenesis of autoimmune disease. Immunity 32: 692–702. 1589. 36. Adachi, O., T. Kawai, K. Takeda, M. Matsumoto, H. Tsutsui, M. Sakagami, 9. Shehadeh, N., F. Calcinaro, B. J. Bradley, I. Bruchim, P. Vardi, and K. J. Lafferty. K. Nakanishi, and S. Akira. 1998. Targeted disruption of the MyD88 gene results 1994. Effect of adjuvant therapy on development of diabetes in mouse and man. in loss of IL-1- and IL-18-mediated function. Immunity 9: 143–150. [Published erratum appears in 2006 Lancet 367: 1240.] Lancet 343: 706–707. 37. Hoshino, K., O. Takeuchi, T. Kawai, H. Sanjo, T. Ogawa, Y. Takeda, K. Takeda, 10. Baxter, A. G., A. C. Horsfall, D. Healey, P. Ozegbe, S. Day, D. G. Williams, and and S. Akira. 1999. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice are A. Cooke. 1994. Mycobacteria precipitate an SLE-like syndrome in diabetes- hyporesponsive to : evidence for TLR4 as the Lps gene prone NOD mice. Immunology 83: 227–231. product. J. Immunol. 162: 3749–3752. 11. Jordan, M. A., P. A. Silveira, D. P. Shepherd, C. Chu, S. J. Kinder, J. Chen, 38. Takeuchi, O., T. Kawai, P. F. Mu¨hlradt, M. Morr, J. D. Radolf, A. Zychlinsky, L. J. Palmisano, L. D. Poulton, and A. G. Baxter. 2000. Linkage analysis of K. Takeda, and S. Akira. 2001. Discrimination of bacterial lipoproteins by Toll-

systemic lupus erythematosus induced in diabetes-prone nonobese diabetic mice like receptor 6. Int. Immunol. 13: 933–940. by guest on September 23, 2021 by Mycobacterium bovis. J. Immunol. 165: 1673–1684. 39. Takeuchi, O., S. Sato, T. Horiuchi, K. Hoshino, K. Takeda, Z. Dong, 12. Silveira, P. A., and A. G. Baxter. 2001. The NOD mouse as a model of SLE. R. L. Modlin, and S. Akira. 2002. Cutting edge: role of Toll-like receptor 1 in Autoimmunity 34: 53–64. mediating immune response to microbial lipoproteins. J. Immunol. 169: 10–14. 13. Janeway, C. A., Jr., and R. Medzhitov. 2002. Innate immune recognition. Annu. 40. Humason, G. H. 1972. Animal Tissue Techniques. W.H. Freeman & Company, Rev. Immunol. 20: 197–216. San Franscisco, p. 238–239. 14. Medzhitov, R., P. Preston-Hurlburt, and C. A. Janeway Jr. 1997. A human ho- 41. Hempel, K., A. Freitag, B. Freitag, B. Endres, B. Mai, and G. Liebaldt. 1985. mologue of the Toll protein signals activation of adaptive immunity. Unresponsiveness to experimental allergic encephalomyelitis in Lewis rats Nature 388: 394–397. pretreated with complete Freund’s adjuvant. Int. Arch. Allergy Appl. Immunol. 15. Ausubel, F. M. 2005. Are innate immune signaling pathways in plants and 76: 193–199. animals conserved? Nat. Immunol. 6: 973–979. 42. Ford, A. L., A. L. Goodsall, W. F. Hickey, and J. D. Sedgwick. 1995. Normal 16. Kumar, H., T. Kawai, and S. Akira. 2009. Toll-like receptors and innate im- adult ramified microglia separated from other central nervous system macro- munity. Biochem. Biophys. Res. Commun. 388: 621–625. phages by flow cytometric sorting. Phenotypic differences defined and direct ex 17. Yoshimura, A., E. Lien, R. R. Ingalls, E. Tuomanen, R. Dziarski, and vivo antigen presentation to myelin basic protein-reactive CD4+ T cells com- D. Golenbock. 1999. Cutting edge: recognition of Gram-positive bacterial cell pared. J. Immunol. 154: 4309–4321. wall components by the occurs via Toll-like receptor 2. J. 43. Mack, M., J. Cihak, C. Simonis, B. Luckow, A. E. Proudfoot, J. Plachy´, Immunol. 163: 1–5. H. Bru¨hl, M. Frink, H. J. Anders, V. Vielhauer, et al. 2001. Expression and 18. Schwandner, R., R. Dziarski, H. Wesche, M. Rothe, and C. J. Kirschning. 1999. characterization of the chemokine receptors CCR2 and CCR5 in mice. J. Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll- Immunol. 166: 4697–4704. like receptor 2. J. Biol. Chem. 274: 17406–17409. 44. Gururajan, M., J. Jacob, and B. Pulendran. 2007. Toll-like receptor expression 19. Takeuchi, O., K. Hoshino, T. Kawai, H. Sanjo, H. Takada, T. Ogawa, K. Takeda, and responsiveness of distinct murine splenic and mucosal B-cell subsets. PLoS and S. Akira. 1999. Differential roles of TLR2 and TLR4 in recognition of gram- ONE 2: e863. negative and gram-positive bacterial cell wall components. Immunity 11: 443– 45. Brightbill, H. D., D. H. Libraty, S. R. Krutzik, R. B. Yang, J. T. Belisle, 451. J. R. Bleharski, M. Maitland, M. V. Norgard, S. E. Plevy, S. T. Smale, et al. 1999. 20. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. Van Huffel, X. Du, D. Birdwell, Host defense mechanisms triggered by microbial lipoproteins through toll-like E. Alejos, M. Silva, C. Galanos, et al. 1998. Defective LPS signaling in C3H/HeJ receptors. Science 285: 732–736. and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282: 2085–2088. 46. Jones, B. W., T. K. Means, K. A. Heldwein, M. A. Keen, P. J. Hill, J. T. Belisle, 21. Hayashi, F., K. D. Smith, A. Ozinsky, T. R. Hawn, E. C. Yi, D. R. Goodlett, and M. J. Fenton. 2001. Different Toll-like receptor agonists induce distinct J. K. Eng, S. Akira, D. M. Underhill, and A. Aderem. 2001. The innate immune macrophage responses. J. Leukoc. Biol. 69: 1036–1044. response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410: 47. Means, T. K., S. Wang, E. Lien, A. Yoshimura, D. T. Golenbock, and 1099–1103. M. J. Fenton. 1999. Human toll-like receptors mediate cellular activation by 22. Ozinsky, A., D. M. Underhill, J. D. Fontenot, A. M. Hajjar, K. D. Smith, Mycobacterium tuberculosis. J. Immunol. 163: 3920–3927. C. B. Wilson, L. Schroeder, and A. Aderem. 2000. The repertoire for pattern 48. Thoma-Uszynski, S., S. M. Kiertscher, M. T. Ochoa, D. A. Bouis, M. V. Norgard, recognition of pathogens by the innate immune system is defined by cooperation K. Miyake, P. J. Godowski, M. D. Roth, and R. L. Modlin. 2000. Activation of between toll-like receptors. Proc. Natl. Acad. Sci. USA 97: 13766–13771. toll-like receptor 2 on human dendritic cells triggers induction of IL-12, but not 23. Wyllie, D. H., E. Kiss-Toth, A. Visintin, S. C. Smith, S. Boussouf, D. M. Segal, IL-10. J. Immunol. 165: 3804–3810. G. W. Duff, and S. K. Dower. 2000. Evidence for an accessory protein function 49. Abel, B., N. Thieblemont, V.J. Quesniaux, N. Brown, J. Mpagi, K. Miyake, F. Bihl, for Toll-like receptor 1 in anti-bacterial responses. J. Immunol. 165: 7125–7132. and B. Ryffel. 2002. Toll-like receptor 4 expression is required to control chronic 24. Hajjar, A. M., D. S. O’Mahony, A. Ozinsky, D. M. Underhill, A. Aderem, Mycobacterium tuberculosis infection in mice. J. Immunol. 169: 3155–3162. S. J. Klebanoff, and C. B. Wilson. 2001. Cutting edge: functional interactions 50. Bafica, A., C. A. Scanga, C. G. Feng, C. Leifer, A. Cheever, and A. Sher. 2005. between toll-like receptor (TLR) 2 and TLR1 or TLR6 in response to phenol- TLR9 regulates Th1 responses and cooperates with TLR2 in mediating optimal soluble modulin. J. Immunol. 166: 15–19. resistance to Mycobacterium tuberculosis. J. Exp. Med. 202: 1715–1724. 14 ROLE OF TLR IN EAE

51. Apetoh, L., F. Ghiringhelli, A. Tesniere, M. Obeid, C. Ortiz, A. Criollo, 71. Su, S. B., P. B. Silver, R. S. Grajewski, R. K. Agarwal, J. Tang, C. C. Chan, and G. Mignot, M. C. Maiuri, E. Ullrich, P. Saulnier, et al. 2007. Toll-like receptor 4- R. R. Caspi. 2005. Essential role of the MyD88 pathway, but nonessential roles dependent contribution of the immune system to anticancer and of TLRs 2, 4, and 9, in the adjuvant effect promoting Th1-mediated autoim- radiotherapy. Nat. Med. 13: 1050–1059. munity. J. Immunol. 175: 6303–6310. 52. Paulos, C. M., C. Wrzesinski, A. Kaiser, C. S. Hinrichs, M. Chieppa, L. Cassard, 72. Lamason, R., P. Zhao, R. Rawat, A. Davis, J. C. Hall, J. J. Chae, R. Agarwal, D. C. Palmer, A. Boni, P. Muranski, Z. Yu, et al. 2007. Microbial translocation P. Cohen, A. Rosen, E. P. Hoffman, and K. Nagaraju. 2006. Sexual dimorphism augments the function of adoptively transferred self/tumor-specific CD8+ T cells in immune response as a function of puberty. BMC Immunol. 7: 2. via TLR4 signaling. J. Clin. Invest. 117: 2197–2204. 73. Keith, A. B. 1978. Sex difference in Lewis rats in the incidence of recurrent 53. Fife, B. T., G. B. Huffnagle, W. A. Kuziel, and W. J. Karpus. 2000. CC che- experimental allergic encephalomyelitis. Nature 272: 824–825. mokine receptor 2 is critical for induction of experimental autoimmune en- 74. Bebo, B. F., Jr., A. A. Vandenbark, and H. Offner. 1996. Male SJL mice do not cephalomyelitis. J. Exp. Med. 192: 899–905. relapse after induction of EAE with PLP 139-151. J. Neurosci. Res. 45: 680–689. 54. Izikson, L., R. S. Klein, I. F. Charo, H. L. Weiner, and A. D. Luster. 2000. 75. Trooster, W. J., A. W. Teelken, P. O. Gerrits, T. H. Lijnema, J. G. Loof, Resistance to experimental autoimmune encephalomyelitis in mice lacking the J. M. Minderhoud, and P. Nieuwenhuis. 1996. The effect of gonadectomy on the CC chemokine receptor (CCR)2. J. Exp. Med. 192: 1075–1080. clinical course of chronic experimental allergic encephalomyelitis. Clin. Neurol. 55. Mahad, D., M. K. Callahan, K. A. Williams, E. E. Ubogu, P. Kivisa¨kk, B. Tucky, Neurosurg. 98: 222–226. G. Kidd, G. A. Kingsbury, A. Chang, R. J. Fox, et al. 2006. Modulating CCR2 76. Voskuhl, R. R., H. Pitchekian-Halabi, A. MacKenzie-Graham, H. F. McFarland, and CCL2 at the blood-brain barrier: relevance for multiple sclerosis patho- and C. S. Raine. 1996. Gender differences in autoimmune demyelination in the genesis. Brain 129: 212–223. mouse: implications for multiple sclerosis. Ann. Neurol. 39: 724–733. 56. Kohm, A. P., P. A. Carpentier, H. A. Anger, and S. D. Miller. 2002. Cutting edge: 77. Orton, S. M., S. V. Ramagopalan, D. Brocklebank, B. M. Herrera, D. A. Dyment, CD4+CD25+ regulatory T cells suppress antigen-specific autoreactive immune I. M. Yee, A. D. Sadovnick, and G. C. Ebers. 2010. Effect of immigration on responses and central nervous system inflammation during active experimental multiple sclerosis sex ratio in Canada: the Canadian Collaborative Study. J. autoimmune encephalomyelitis. J. Immunol. 169: 4712–4716. Neurol. Neurosurg. Psychiatry 81: 31–36. 57. Szanya, V., J. Ermann, C. Taylor, C. Holness, and C. G. Fathman. 2002. The 78. Erridge, C. 2010. Endogenous ligands of TLR2 and TLR4: agonists or assis- subpopulation of CD4+CD25+ splenocytes that delays adoptive transfer of di- tants? J. Leukoc. Biol. 87: 989–999. abetes expresses L- and high levels of CCR7. J. Immunol. 169: 2461– 79. Erridge, C., T. Attina, C. M. Spickett, and D. J. Webb. 2007. A high-fat meal 2465. induces low-grade endotoxemia: evidence of a novel mechanism of postprandial Downloaded from 58. Taylor, P. A., A. Panoskaltsis-Mortari, J. M. Swedin, P. J. Lucas, R. E. Gress, inflammation. Am. J. Clin. Nutr. 86: 1286–1292. B. L. Levine, C. H. June, J. S. Serody, and B. R. Blazar. 2004. L-Selectin(hi) but 80. Clarke, T. B., K. M. Davis, E. S. Lysenko, A. Y. Zhou, Y. Yu, and J. N. Weiser. not the L-selectin(lo) CD4+25+ T-regulatory cells are potent inhibitors of GVHD 2010. Recognition of peptidoglycan from the microbiota by Nod1 enhances and BM graft rejection. Blood 104: 3804–3812. systemic innate immunity. Nat. Med. 16: 228–231. 59. Ermann, J., P. Hoffmann, M. Edinger, S. Dutt, F. G. Blankenberg, J. P. Higgins, 81. Iwaki, D., H. Mitsuzawa, S. Murakami, H. Sano, M. Konishi, T. Akino, and R. S. Negrin, C. G. Fathman, and S. Strober. 2005. Only the CD62L+ sub- Y. Kuroki. 2002. The extracellular toll-like receptor 2 domain directly binds population of CD4+CD25+ regulatory T cells protects from lethal acute GVHD. peptidoglycan derived from Staphylococcus aureus. J. Biol. Chem. 277: 24315–

Blood 105: 2220–2226. 24320. http://www.jimmunol.org/ 60. Baxter, A. G. 2007. The origin and application of experimental autoimmune 82. Dziarski, R., and D. Gupta. 2005. Staphylococcus aureus peptidoglycan is a toll- encephalomyelitis. Nat. Rev. Immunol. 7: 904–912. like receptor 2 activator: a reevaluation. Infect. Immun. 73: 5212–5216. 61. Hansen, B. S., R. Z. Hussain, A. E. Lovett-Racke, J. A. Thomas, and 83. Kumar, V., K. Stellrecht, and E. Sercarz. 1996. Inactivation of T cell receptor M. K. Racke. 2006. Multiple toll-like receptor agonists act as potent adjuvants in peptide-specific CD4 regulatory T cells induces chronic experimental autoim- the induction of autoimmunity. J. Neuroimmunol. 172: 94–103. mune encephalomyelitis (EAE). J. Exp. Med. 184: 1609–1617. 62. Segal, B. M., J. T. Chang, and E. M. Shevach. 2000. CpG oligonucleotides are 84. McGeachy, M. J., L. A. Stephens, and S. M. Anderton. 2005. Natural recovery potent adjuvants for the activation of autoreactive encephalitogenic T cells and protection from autoimmune encephalomyelitis: contribution of CD4+CD25+ in vivo. J. Immunol. 164: 5683–5688. regulatory cells within the central nervous system. J. Immunol. 175: 3025–3032. 63. Wolf, N. A., T. K. Amouzegar, and R. H. Swanborg. 2007. Synergistic interaction 85. Yu, P., R. K. Gregg, J. J. Bell, J. S. Ellis, R. Divekar, H. H. Lee, R. Jain, between Toll-like receptor agonists is required for induction of experimental au- H. Waldner, J. C. Hardaway, M. Collins, et al. 2005. Specific T regulatory cells toimmune encephalomyelitis in Lewis rats. J. Neuroimmunol. 185: 115–122. display broad suppressive functions against experimental allergic encephalo-

64. Tigno-Aranjuez, J. T., R. Jaini, V. K. Tuohy, P. V. Lehmann, and M. Tary- myelitis upon activation with cognate antigen. J. Immunol. 174: 6772–6780. by guest on September 23, 2021 Lehmann. 2009. Encephalitogenicity of complete Freund’s adjuvant relative to 86. Selvaraj, R. K., and T. L. Geiger. 2008. Mitigation of experimental allergic CpG is linked to induction of Th17 cells. J. Immunol. 183: 5654–5661. encephalomyelitis by TGF-beta induced Foxp3+ regulatory T lymphocytes 65. Nagasawa, K., R. Mori, S. Kondo, and K. Hisatsune. 1979. Correlation between through the induction of anergy and infectious tolerance. J. Immunol. 180: 2830– the capacity of bacterial lipopolysaccharide to supress experimental allergic 2838. encephalomyelitis and its mitogenic activity for lymph node cells in guinea pigs. 87. Stephens, L. A., K. H. Malpass, and S. M. Anderton. 2009. Curing CNS auto- J. Neurol. Sci. 42: 417–427. immune disease with myelin-reactive Foxp3+ Treg. Eur. J. Immunol. 39: 1108– 66. Buenafe, A. C., and D. N. Bourdette. 2007. Lipopolysaccharide pretreatment 1117. modulates the disease course in experimental autoimmune encephalomyelitis. J. 88. Zhang, X., D. N. Koldzic, L. Izikson, J. Reddy, R. F. Nazareno, S. Sakaguchi, Neuroimmunol. 182: 32–40. V. K. Kuchroo, and H. L. Weiner. 2004. IL-10 is involved in the suppression of 67. Lampropoulou, V., K. Hoehlig, T. Roch, P. Neves, E. Caldero´nGo´mez, experimental autoimmune encephalomyelitis by CD25+CD4+ regulatory T cells. C. H. Sweenie, Y. Hao, A. A. Freitas, U. Steinhoff, S. M. Anderton, and Int. Immunol. 16: 249–256. S. Fillatreau. 2008. TLR-activated B cells suppress T cell-mediated autoimmu- 89. Dominitzki, S., M. C. Fantini, C. Neufert, A. Nikolaev, P. R. Galle, J. Scheller, nity. J. Immunol. 180: 4763–4773. G. Monteleone, S. Rose-John, M. F. Neurath, and C. Becker. 2007. Cutting edge: 68. Buwitt-Beckmann, U., H. Heine, K. H. Wiesmu¨ller, G. Jung, R. Brock, S. Akira, trans-signaling via the soluble IL-6R abrogates the induction of FoxP3 in naive and A. J. Ulmer. 2006. TLR1- and TLR6-independent recognition of bacterial CD4+CD25 T cells. J. Immunol. 179: 2041–2045. lipopeptides. J. Biol. Chem. 281: 9049–9057. 90. Xiao, W., D. R. Hodge, L. Wang, X. Yang, X. Zhang, and W. L. Farrar. 2004. Co- 69. Ellestad, K. K., S. Tsutsui, F. Noorbakhsh, K. G. Warren, V. W. Yong, operative functions between nuclear factors NFkappaB and CCAT/enhancer- Q. J. Pittman, and C. Power. 2009. Early life exposure to lipopolysaccharide binding protein-beta (C/EBP-beta) regulate the IL-6 promoter in autocrine hu- suppresses experimental autoimmune encephalomyelitis by promoting tolero- man cells. Prostate 61: 354–370. genic dendritic cells and regulatory T cells. J. Immunol. 183: 298–309. 91. Samoilova, E. B., J. L. Horton, B. Hilliard, T. S. Liu, and Y. Chen. 1998. IL-6- 70. Greisman, S. E., and W. E. Woodward. 1965. Mechanisms of endotoxin toler- deficient mice are resistant to experimental autoimmune encephalomyelitis: roles ance. 3. The refractory state during continuous intravenous infusions of endo- of IL-6 in the activation and differentiation of autoreactive T cells. J. Immunol. toxin. J. Exp. Med. 121: 911–933. 161: 6480–6486.