A transgenic model of central nervous system autoimmunity mediated by CD4⁺ and CD8⁺ T and B cells

Abstract

Experimental autoimmune encephalomyelitis (EAE) is a widely used model of multiple sclerosis (MS). In NOD mice, EAE develops as a relapsing-remitting disease that transitions to a chronic progressive disease, making the NOD model the only mouse model that recapitulates the full clinical disease course observed in most MS patients. We have generated a T cell receptor (TCR) transgenic mouse that expresses the alpha and beta chains of a myelin oligodendrocyte glycoprotein (MOG) 35–55-reactive TCR (1C6) on the NOD background. 1C6 TCR transgenic mice spontaneously generate both CD4⁺ and CD8⁺ T cells that recognize MOG and produce pro-inflammatory cytokines, allowing for the first time the simultaneous examination of myelin-reactive CD4⁺ and CD8⁺ T cells in the same host. 1C6 CD8⁺ T cells alone can induce optic neuritis and mild EAE with delayed onset; however, 1C6 CD4⁺ T cells alone induce severe EAE and predominate in driving disease when both cell types are present. When 1C6 mice are crossed with mice bearing an immunoglobulin heavy chain specific for MOG, the mice develop spontaneous EAE with high incidence but surprisingly the disease pattern does not resemble the Neuromyelitis optica (NMO)-like disease observed in mice bearing CD4⁺ T cells and B cells reactive to MOG on the C57BL/6 background. Collectively our data show that while myelin-reactive CD8⁺ T cells contribute to disease, disease is primarily driven by myelin-reactive CD4⁺ T cells and that the co-existence of myelin-reactive T and B cells does not necessarily result in a distinct pathological phenotype.

Introduction

Experimental autoimmune encephalomyelitis (EAE) is the most widely used model of multiple sclerosis (MS). In NOD mice, EAE induced by immunization with myelin oligodendrocyte glycoprotein (MOG) 35–55 follows a relapsing-remitting course after which the animals develop chronic progressive disease, thereby recapitulating the full clinical disease course observed in most MS patients. In addition, the NOD strain is attractive for the study of the genetics of autoimmunity because many disease-modifying loci are shared between type 1 diabetes and EAE (1–3); moreover, congenic strains in which disease-modifying loci have been introgressed from resistant strains onto the NOD strain have been generated. However, in the NOD strain the kinetics of disease development is delayed relative to that in C57BL/6 mice, i.e. NOD mice have to be observed for at least 3 months in order to observe the full clinical spectrum of disease. We, therefore, set out to generate a T cell receptor transgenic model of EAE on the NOD background in which we could study the full spectrum of clinical EAE from relapsing-remitting disease to chronic progressive disease and with which we could take advantage of the genetic tools available on the NOD background.

Several different T cell receptor transgenic models of EAE have been developed over the years. Most of these are based on CD4⁺ T cells that recognize myelin basic protein (MBP) (4–6), myelin proteolipid protein (PLP) (7) or MOG (8) on different genetic backgrounds. While these models have contributed greatly to our understanding of the role of CD4⁺ T cells in central nervous system (CNS) autoimmunity, increasing evidence suggests that CD8⁺ T cells are also important in the pathogenesis of MS. Indeed, CD8⁺ T cells outnumber CD4⁺ T cells in the brains of MS patients, and oligoclonal expansion of CD8⁺ T cells has been observed in the blood, cerebrospinal fluid (CSF) and brains of MS patients (9–11). MBP-reactive CD8⁺ T cells have also been isolated from MS patients (12).

In EAE, myelin antigen-reactive CD8⁺ T cells have been shown to induce disease (13–15). However, models in which the role of myelin-reactive CD8⁺ T cells can be studied in the presence of myelin-reactive CD4⁺ T cells have been lacking. Moreover, the recent success of antibody therapy targeted to B cells in MS has renewed interest in understanding the role of B cells in driving CNS autoimmunity. For many years, it has been known that immunoglobulins (Igs) are present in the CSF in a large proportion of MS patients and that B cells, plasma cells and myelin-specific antibodies are present in chronic MS plaques and in areas of active myelin breakdown (16, 17). Whether the Ig itself or the antigen-presenting cell function of putative myelin-reactive B cells are important in driving disease is not known. The generation of a knock-in mouse bearing the immunoglobulin heavy chain of a MOG-specific antibody (IgH^MOG) has facilitated the study of the role of myelin-specific B cells in EAE (18). When these mice are crossed with transgenic mice expressing CD4⁺ T cells that recognize MOG 35–55, the mice spontaneously develop a disease that resembles Neuromyelitis optica (NMO), a subtype of MS (19, 20). However, the disease phenotype that results when both CD4⁺ and CD8⁺ T cells as well as B cells reactive to myelin antigen are present in the same individual, a situation that likely best replicates the in vivo situation in MS patients, has never been examined. Here we describe a T cell receptor transgenic model of EAE on the NOD background that allows for the first time the examination of CNS autoimmunity when both CD4⁺ and CD8⁺ T cells as well as B cells that recognize the same myelin antigen are present.

Material and Methods

Mice

Generation of MOG 35-55-specific TCR transgenic mice on NOD background

The TCR alpha (Vα5Jα22) and beta (Vβ7Dβ2Jβ2.7) regions of a CD4⁺ T cell clone (1C6) specific for MOG 35–55 (MEVGWYRSPFSRVVHLYRNGK) were amplified by PCR from genomic DNA and cloned into the PTα and PTβ genomic expression vectors (21). Linearized TCR containing constructs were injected directly into the pronuclei of fertilized NOD oocytes. Transgenic founders were identified by PCR using primers specific for the Vα and Vβ regions. Three independent transgenic founders were identified and bred with NOD mice (Taconic Farms). Animals were housed in a specific-pathogen free facility and maintained in accordance with the guidelines established by the committee on Animals of Harvard Medical School. IgH^MOG mice on the C57BL/6 background (18) were backcrossed onto the NOD background for five generations prior to inter-crossing with 1C6 TCR transgenic mice.

Flow Cytometry

Single cell suspensions from thymus, lymph node or spleen were prepared and stained with the indicated antibodies. All antibodies were purchased from eBiosciences, BD Biosciences and Biolegend. All flow cytometry data were collected on a BD FACS Calibur or LSRII (BD Biosciences) and analyzed with FlowJo Software (Tree Star).

Proliferation assays and Cytokine measurement

Splenocytes (4×10⁵/well) were cultured in triplicate in 96-well plates in the presence of the indicated concentration of antigen. CD4⁺ and CD8⁺ T cells (1×10⁵/well) were cultured in triplicate in 96-well plates in the presence of the indicated concentration of antigen and irradiated wild type NOD splenic APC (5×10⁵). At 48 h, supernatants were collected for the measurement of cytokines by either ELISA or cytometric bead array (CBA) (BD Biosciences) according to the manufacturer’s protocol. Plates were then pulsed with [³H]Thymidine and harvested 16 hours later. Thymidine incorporation was measured using a β–scintillation counter (1450 microbeta;PerkinElmer).

Degranulation assay

Mice were immunized with 10 μg of MOG 35–55 peptide emulsified in complete Freund’s adjuvant (CFA; Difco) supplemented with 4 mg/ml of Mycobacterium tuberculosis (H37RA; Difco). Each mouse also received 200 ng of pertussis toxin (List Biological Laboratories) intravenously on Days 0 and 2. Draining lymph nodes were harvested on Day 7 and cultured in medium at 2.5×10⁶/well in the presence of MOG 35–55 peptide (0–50μg/ml), PMA (Sigma) and ionomycin (Sigma) or medium as an unstimulated control. LAMP-1 (CD107a; Biolegend) and monensin (Golgi Stop, BD Bioscience) were added to cultures containing PMA and ionomycin at the beginning of the incubation for 4 h at 37° C. Cultures stimulated with MOG 35–55 were preincubated for 2 hrs before the addition of monensin and LAMP-1. Cells were then stained for cell surface expression of CD8 and intracellular expression of Granzyme B, IFN-γ, IL-2, and TNF. Flow cytometric data were collected on a BD LSRII (BD Biosciences) and analyzed with FlowJo Software (Tree Star).

Induction and assessment of EAE

1C6 Tg mice and wild-type littermate controls were immunized with 10 μg MOG 35–55 emulsified in complete Freund’s adjuvant (CFA; Difco) supplemented with 4 mg/ml of Mycobacterium tuberculosis (H37RA; Difco). Each mouse also received 200 ng of pertussis toxin (List Biological Laboratories) intravenously on Days 0 and 2. Clinical disease was assessed as follows: 0, no disease; 1, decreased tail tone; 2, hind limb paresis; 3, complete hind limb paralysis; 4, forelimb and hind limb paralysis; 5, moribund state.

Adoptive transfer experiments

1–1.75 × 10⁶ sorted 1C6 CD4⁺ or CD8⁺ T cells were injected intravenously into NOD.Scid mice. After 21 days, mice were immunized with 25–50 μg MOG 35–55 emulsified in CFA and given 200 ng Pertussis toxin intravenously on Days 0 and 2 post immunization. Clinical disease was assessed as described above.

Histopathology

When animals were moribund or at the end of the experiment, they were sacrificed and brains and spinal cords were fixed in 10% formalin, processed for histologic analysis, and evaluated as described (22). Specifically, inflammatory foci (> 10 perivascular mononuclear cells) were counted in leptomeninges and CNS parenchyma in Luxol fast blue-H & E-stained paraffin sections of the entire CNS of each mouse by a neuropathologist who was blinded to the clinical status and genotypes. Each sample counted contained representative 8 μm thick sections of 5 levels of brain tissue and 15–20 sections of spinal cord tissue. Thus, both the sampling and total areas of CNS tissue analyzed were uniform for each mouse. In some analyses, the localization of parenchymal inflammatory foci in cerebrum, brain stem, cerebellum and spinal cord were separately determined. Optic neuritis was scored as either present or absent for each mouse.

Isolation and analysis of cells from CNS

Following perfusion, CNS tissue (brain and spinal cord) was digested with collagenase (2.5 mg/ml, Sigma) prior to centrifugation on a discontinuous percoll gradient (GE Healthcare). Cells were then stimulated with PMA (50 ng/ml) and ionomycin (1 μg/ml) in the presence of a golgi transport inhibitor for 3 hours prior to staining with 7AAD and antibodies against CD4 and CD8. Following fixation, cells were permeabilized and stained with antibodies against IFNγ and IL-17. Flow cytometry data were collected on an LSRII (BD Biosciences).

Results

Generation of 1C6 TCR transgenic mice and selection of 1C6 TCR transgenic T cells

We isolated a CD4⁺ T cell clone from a NOD mouse immunized with MOG 35–55. This clone called 1C6 expressed a TCR composed of Vα5 and Vβ7. The rearranged TCRα and TCRβ chains from the 1C6 clone were isolated and cloned into expression cassettes and micro-injected into NOD oocytes. We then screened progeny for the presence of the transgenic Vα and Vβ and successfully identified three independent founders (data not shown).

We have analyzed all three 1C6 founder lines and have not observed any gross differences in thymic cellularity in 1C6 transgenic (Tg) mice relative to wild type littermate controls (data not shown). When we analyzed CD4, CD8 and Vβ7 expression in thymocytes, we were surprised to find that there was not a strong bias towards CD4 single positive (SP) T cells in 1C6 Tg mice relative to wild type littermate controls and that there was prominent expression of Vβ7 on both CD4SP and CD8SP thymocytes in 1C6 Tg mice (Figure 1). Accordingly, we observed similar CD4/CD8 ratios in the spleen and lymph nodes of 1C6 Tg mice relative to wild type littermate controls as well as expression of Vβ7 on both CD4⁺ and CD8⁺ T cells in the spleen and lymph nodes of 1C6 TCR Tg mice (Figure 1 and data not shown). Indeed, we have observed the presence of Vβ7 positive CD8⁺ T cells in three independent transgenic lines (data not shown), indicating that the presence of CD8⁺Vβ7⁺ cells is not due an integration artifact. We next addressed whether the CD8⁺ T cells from 1C6 Tg mice expressed a rearranged Vα5 gene. Indeed, we found that 1C6 CD8⁺ T cells express a rearranged Vα5 gene and thus should have specificity for MOG 35–55 (supplementary data 1). We further examined the extent of allelic exclusion at the TCR alpha locus and found that 1C6 CD8⁺Vb7⁺ T cells exhibit rearrangement of some endogenous alpha chains but not others (supplementary data 1). Lastly, we examined whether the frequency of regulatory T cells was altered by the presence of the 1C6 transgene and found that the frequency of CD4⁺FoxP3⁺ cells is the same in wild type and 1C6 Tg mice (supplementary data 2).

Figure 1.

Flow cytometric analysis of thymocytes and peripheral T cells in 1C6 TCR transgenic mice.

Thymocytes (A) and splenocytes (B) from 7 week old female 1C6 transgenic (line#1191) and wild type littermate were stained with antibodies against CD4, CD8, and Vβ7. Data shown are representative of 8 independent analyses of 6–13 week old male and female mice across three independent founder lines.

Responses to MOG in 1C6 TCR transgenic mice

Next, we tested responsiveness of spleen cells from 1C6 Tg mice to MOG 35–55 peptide and recombinant MOG (rMOG) protein (consisting of the extracellular domain). Splenocytes from 1C6 Tg mice showed a strong proliferative response to both MOG 35–55 peptide and rMOG (Figure 2A and B). In terms of cytokine, the response was dominated by production of IFN-γ and IL-17. Other cytokines such as TNF, IL-6 and IL- 10 were variable and were generally produced in lower quantities. IL-4, IL-9, and IL-21 were not detected.

Figure 2.

T cell responses in 1C6 TCR transgenic mice.

Splenocytes from 1C6 transgenic (line#1183) and wild type littermates were stimulated with MOG 35–55 as indicated (A) and recombinant MOG (B) at 50 μg/ml. (C) Sorted CD4⁺ and CD8⁺ T cells from 1C6 mice were stimulated with MOG 35–55 (100 μg/ml) and rMOG (50 μg/ml). Proliferation was determined by [³H]Thymidine incorporation of triplicate wells. At 48 h, supernatants were harvested for measurement of production of the indicated cytokines by cytometric bead array (CBA). Data shown are representative of 2–3 independent experiments for each founder line. Error bars indicate s.e.m.

Since we had observed the presence of both CD4⁺ and CD8⁺ T cells expressing Vβ7 in 1C6 Tg mice, we next examined the responsiveness of CD4⁺ and CD8⁺ T cells from 1C6 Tg mice to both MOG 35–55 peptide and rMOG. We found that both CD4⁺ and CD8⁺ T cells from 1C6 Tg mice respond to MOG 35–55 although the response of CD8⁺ T cells is lower in magnitude (Figure 2C). We also addressed the MHC restriction of 1C6 CD8⁺ T cells and found that they can recognize MOG 35–55 in the context of both MHC Class I and Class II (Supplementary data 3). In terms of cytokine, the response was again dominated by IFN-γ and IL-17. Collectively, these data demonstrate that 1C6 CD4⁺ and CD8⁺ T cells are able to respond to both MOG peptide and MOG protein and that they are predisposed to develop into Th1/Tc1 and/or Th17 phenotype.

Properties of 1C6 CD8⁺ T cells

Given our observation that nearly all the CD8⁺ T cells in 1C6 Tg mice express Vβ7 and that they respond to MOG, produce proinflammatory cytokines and can recognize MOG in both the context of MHC Class I and II, we addressed whether 1C6 CD8⁺ T cells exhibited features of classical CD8⁺ T cells. To do this, we immunized 1C6 Tg or wild type control littermates with MOG 35–55 in CFA. Indeed, immunization with MOG 35–55/CFA has been shown previously to recruit MOG 35–55 reactive CD8⁺ cells (14, 15). We isolated draining lymph nodes and examined the ability of CD8⁺ T cells to respond to re-stimulation with MOG. CD8⁺ T cells from 1C6 Tg but not wild type littermates degranulated as shown by surface expression of CD107α (LAMP-1, lysosomal-associated membrane protein-1) (Figure 3). As classically seen in CD8⁺ T cells, degranulated cells also expressed intracellular IFN-γ and granzyme B in response to stimulation with MOG 35–55. Similar results were obtained for CD107α/TNF and CD107α/IL-2 (supplementary data 4).

Figure 3.

Function of CD8⁺ T cells in 1C6 TCR transgenic mice.

1C6 transgenic (line#1183) and wild type littermate were immunized with 10 μg of MOG 35–55 peptide emulsified in complete Freund’s adjuvant supplemented with 4 mg/ml of Mycobacterium tuberculosis (H37RA). Each mouse also received 200 ng of pertussis toxin intravenously on Days 0 and 2. On Day 7, cells were harvested from draining lymph nodes and stimulated with MOG 35–55 peptide as indicated for 4 h. Cells were stained with antibodies against CD8, CD107a (LAMP-1) and Granzyme B or IFN-γ. Expression of CD107a and Granzyme B or IFN-γ is shown on gated CD8⁺VB7⁺ (Tg) and total CD8⁺ (WT) T cells. Data shown are representative of five independent experiments.

Disease in 1C6 Tg mice

MOG 35–55 TCR transgenic mice on the C57BL/6 background develop spontaneous optic neuritis as well as EAE, albeit at very different frequencies (>30% vs. 4%, respectively) (8). We therefore followed 1C6 Tg mice for the development of spontaneous disease. We have observed over 400 mice and found only 5 mice with spontaneous optic neuritis and 4 with spontaneous EAE. These cases were confirmed by histopathology and occurred in both male and female transgenic mice (data not shown). Thus, while 1C6 Tg mice can develop spontaneous disease, they do not do so at an appreciable frequency.

Given that 1C6 Tg mice harbor both CD4⁺ and CD8⁺ T cells that respond to MOG, we sought to address the relative contribution of 1C6 CD4⁺ and CD8⁺ T cells to disease. We therefore isolated highly purified populations of CD4⁺ and CD8⁺ 1C6 T cells from unimmunized mice and transferred them either alone or together into NOD.Scid recipients. We allowed 21 days for reconstitution of the peripheral lymphoid compartment. None of the mice developed spontaneous disease during this time. Therefore, we immunized mice with MOG 35–55/CFA. Mice that received 1C6 CD4⁺ T cells alone developed typical EAE with almost 100% incidence and good severity (Figure 4 and Table I). Mice that received both 1C6 CD4⁺ and CD8⁺ T cells developed disease with nearly identical clinical courses as mice that received CD4⁺ T cells alone. In contrast, mice that received 1C6 CD8⁺ T cells alone developed mild disease with a much lower clinical incidence. Histopathologic analysis revealed predominance of spinal cord lesions in all three groups (data not shown). Interestingly, there was a large discordance between incidence of clinical disease and histological disease, (48% vs. 70%, respectively) in recipients of 1C6 CD8⁺ T cells alone (Table I). Also noteworthy is that there was a trend towards a higher incidence of optic neuritis in mice that received both 1C6 CD4⁺ and CD8⁺ T cells relative to mice that received either 1C6 CD4⁺ T cells or 1C6 CD8⁺ T cells alone; however, this did not reach statistical significance.

Figure 4.

EAE induced by 1C6 CD4⁺ and CD8⁺ T cells.

CD4⁺ and CD8⁺ T cells from 1C6 mice were transferred either alone or together into NOD.Scid recipients. On Day 21 post transfer mice were immunized for EAE with MOG 35-55/CFA and administered Pertussis toxin.

Table I.

Disease induced by 1C6 CD4⁺ and CD8⁺ T cells

	Clinical Disease			Histological Disease Inflammatory Foci: Mean ± sem
Group	Incidence	Mean Onset (Days)	Mean Max Score	Incidence	Meninges	Parenchyma	Total	Optic Neuritis
CD4	23/24 (96%)	11.5±1.3	3.02±0.21	23/24 (96%)	48.8±11.9	87.3±4.4	138±15.7	16/24 (67%)
CD8	11/23 (48%)	12.3±2.0	1.59±0.4	16/23 (70%)	24.3±13	56.5±24.8	80.3±38.1	11/23 (48%)
CD4+CD8	16/19 (84%)	12.4±1.2	3.06±0.35	17/19 (89%)	51.3±10.5	79.8±17.6	131±27.6	17/19 (89%)

Data are pooled from six independent experiments.

We next examined active induction of disease in 1C6 Tg mice. To do this, we immunized 1C6 Tg mice with a low dose of MOG 35–55. With this dose of MOG 35–55, 1C6 Tg mice developed disease by day 9 while none of the wild-type littermates developed disease. Disease consisted of an initial wave of mild severity that resolved completely. This was then followed by a relapse of greater severity that did not resolve completely. After the second wave of disease, 1C6 Tg mice developed chronic disease (Figure 5A). Thus, 1C6 Tg mice exhibit the progression of relapsing-remitting to chronic EAE typically observed in NOD mice. Examination of the CNS of 1C6 Tg mice with EAE showed the presence of both CD4⁺ and CD8⁺ T cells, although CD4⁺ T cells outnumbered CD8⁺ T cells by approximately 30 fold (Figure 5B). The CD4⁺ T cells in the CNS of 1C6 Tg mice consisted of cells producing IFN-γ, IL-17 and IFN-γ plus IL-17 (Figure 5C). IL-17-producing cells appeared to be the most abundant population of CD4⁺ T cells; however, this was only significant when compared to cells producing both IFN-γ and IL-17. Although CD8⁺ T cells were low in frequency, a large fraction of these cells were producing IFN-γ (Figure 5C). Lastly, histopathologic analysis showed the presence of typical EAE lesions (data not shown). Collectively, our adoptive transfer and active disease induction data indicate that myelin-reactive CD8⁺ T cells can induce optic neuritis and mild clinical disease on their own and can contribute to disease after immunization for EAE; however, myelin-reactive CD4⁺ cells appear to have the dominant role in driving disease.

Figure 5.

Active disease induction in 1C6 Tg mice.

10 week old female 1C6 Tg (n=4) and wild-type littermates (n=4) were immunized with 10 μg of MOG 35-55/CFA and administered 200 ng pertussis toxin intravenously on Days 0 and 2. Immunized mice were monitored for the development of EAE. A) The mean clinical disease score is shown. Error bars indicate SEM. (B and C) Mononuclear cells were isolated from the CNS of 1C6 Tg mice with EAE (score 2–2.5). Cells were stimulated with PMA/ionomycin prior to staining with antibodies against CD4 and CD8 and the examination of intracellular cytokine production by flow cytometry. Data shown are the mean of three independent experiments. Error bars indicate s.e.m. * denotes p<0.05, one-way ANOVA, Tukey’s multiple comparison test.

1C6 x IgH^MOG Transgenic mice

We have next crossed IgH^MOG knock-in mice (18) onto the NOD background so that in conjunction with 1C6 Tg mice we could to study for the first time the immune response and development of disease in hosts that have CD4⁺ and CD8⁺ T cells as well as B cells that recognize the same myelin antigen, MOG. We first examined responses to MOG 35–55 and rMOG in 1C6 x IgH^MOG mice. Spleen cells from 1C6 x IgH^MOG mice proliferate better in response to stimulation with both MOG 35–55 and rMOG than spleen cells from 1C6 Tg mice (Figure 6A and B). Again we observed that the cytokine response was dominated by production of IFN-γ and IL-17 with spleen cells from 1C6 x IgH^MOG mice producing more IL-17 than spleen cells from 1C6 Tg mice in response to rMOG (Figure 6C).

Figure 6.

Responses to MOG 35-55 and rMOG in 1C6 and 1C6 x IgH^MOG mice.

Splenocytes from 1C6 transgenic (line#1183) and 1C6 x IgH^MOG were stimulated with MOG 35-55 (A) and recombinant MOG (B) as indicated. Proliferation (A and B) was determined by [³H]Thymidine incorporation of triplicate wells. Data shown are representative of 3 independent experiments. At 48 hours supernatants were harvested for measurement of cytokine production by cytometric bead array (CBA) and ELISA (C). Data shown are the mean of three independent experiments. *p=0.027. Error bars indicate s.e.m.

Next, we examined both the clinical and pathological phenotype of disease in 1C6 x IgH^MOG mice. We were particularly interested in the disease phenotype of these mice given that when MOG 35–55-reactive CD4⁺ T cells (from 2D2 TCR Tg) and MOG-reactive B cells (from IgH^MOG knock-in) co-exist on the C57BL/6 background, the mice develop spontaneous disease with 60% incidence and exhibit a distinct pathological pattern in which lesions are localized mostly in the spinal cord and optic nerve, thus resembling Neuromyelitis Optica (NMO), a subtype of MS most commonly observed in Asian populations (19, 20). Moreover, those mice exhibit development of ectopic lymphoid follicle-like structures in the spinal meninges (19).

In contrast to 1C6 Tg mice in which we observed only rare cases of spontaneous disease, 1C6 x IgH^MOG mice developed spontaneous disease with high incidence: 45% among males and 79% among females with an average age of onset of 15 weeks (Table II). Examination of the CNS of 1C6 x IgH^MOG mice with spontaneous disease revealed the presence of both CD4⁺ and CD8⁺ T cells. In comparison to 1C6 Tg mice with EAE, 1C6 x IgH^MOG exhibited more CD4⁺ and CD8⁺ T cells in the CNS and CD4⁺ T cells only outnumbered CD8⁺ T cells by approximately 7 fold (Figure 7A). Similar to 1C6 Tg mice with EAE, CD4⁺ T cells in 1C6 x IgH^MOG CNS also consisted of IFN-γ–producing, IL-17-producing, and IFN-γ plus IL-17-producing cells. While in 1C6 Tg mice there was a trend towards IL-17 producing cells being the most abundant population of CD4⁺ T cells, in 1C6 x IgH^MOG mice there was a trend towards IFN-γ-producing cells being the most abundant population of CD4⁺ T cells in the CNS. However, this was only significant when compared to IFN-γ plus IL-17-secreting cells (Figure 7B). As observed in 1C6 Tg mice, the CD8⁺ T cells in the CNS of 1C6 x IgH^MOG predominantly produce IFN-γ (Figure 7B). Collectively, these data indicate that relative to 1C6 Tg mice with EAE, 1C6 x IgH^MOG mice accumulate more CD4⁺ (3 fold) and CD8⁺ (14 fold) T cells in the CNS and that the majority of these cells produce IFN-γ. Given the disproportionate increase in CD8⁺ T cells in the CNS of 1C6 x IgH^MOG mice it is interesting to speculate whether the increased involvement of CD8⁺ T cells, at least in part, underlies the high frequency of spontaneous disease observed in these mice relative to 1C6 Tg.

Table II.

Spontaneous classical EAE in 1C6 x IgH^MOG mice

	Clinical Disease			Histological Disease Inflammatory Foci: mean ± sem
Sex	Incidence	Mean Onset (Days)	Mean Max Score	Meninges	Parenchyma	Total (n)	Optic Neuritis
Male	5/11 (45%)	105±21.1	2.3±0.2	18.3±5.8	32±13.1	50.3±15.5 (3)	2/3 (67%)
Female	11/14 (79%)	105.7±11.4	2.55±0.18	54.9±23.6	53.3±27.6	108.1±51.0 (7)	2/8 (25%)
	16/25 (64%)	105.6±9.9	2.47±0.14				4/11 (36%)

Mice were observed for a period of 22 weeks for the development of spontaneous clinical signs of EAE.

Figure 7.

Disease in 1C6 x IgH^MOG mice.

(A and B) Mononuclear cells were isolated from the CNS of 1C6 and IC6 x IgH^MOG mice with EAE (score 2–2.5). Cells were stimulated with PMA/ionomycin prior to staining with antibodies against CD4 and CD8 and the examination of intracellular cytokine production by flow cytometry. Data shown are the mean of three independent experiments. Error bars indicate s.e.m. * denotes p=0.0002, unpaired t-test, **p=0.014, unpaired t-test, ***p<0.05, one-way ANOVA, Tukey’s multiple comparison test. Data shown are representative of 2–3 independent experiments for each founder line. Error bars indicate s.e.m.

In sharp contrast to 2D2 x IgH^MOG mice, 1C6 x IgH^MOG mice do not exhibit NMO-like lesion distribution. Although the majority of lesions were restricted to the spinal cord, less than half of the mice (40%) had optic nerve lesions (Tables II and III); there was also no evidence of ectopic follicle-like structures in the CNS. These data indicate that the coexistence of myelin-reactive T and B cells does not necessarily result in an NMO-like disease phenotype. Rather, the genetic background in conjunction with T and B cell reactivity appear to influence the lesion distribution in the CNS and the development of ectopic follicle-like structures in this disease.

Table III.

Lesion localization and disease phenotype in 1C6 x IgH^MOG mice

		Lesion Localization
Clinical Disease	No. Mice with lesions	Cord	Cerebellum	Cerebrum	Optic Nerve
Classical EAE	10/10 (100%)	9/10 (90%)	0/4	1/10 (10%)	4/10 (40%)
None	6/11 (55%)1	0/4	3/4 (75%)2	1/4 (25%)	0/11

¹Lesion localization could not be assessed in two mice due to low number of lesions.

²All mice >20 weeks of age.

In addition to the lack of NMO-like disease, we also made a surprising observation in 1C6 x IgH^MOG mice that did not develop signs of classical EAE. These mice were sacrificed for use as controls for experiments or because they had reached greater than 20 weeks of age and did not exhibit any overt signs of neurologic disease at the time of sacrifice. Upon histopathological examination of the CNS, we observed the presence of CNS lesions in over half of these mice (Table III). Out of these, 75 percent exhibited lesions localized exclusively to the cerebellum, indicating the presence of a pathologically distinct neurologic disease.

Discussion

While EAE shares many clinical and histological features with MS, it has been argued that EAE is not an adequate model for MS. In most EAE models, disease is induced by immunization with MHC class II-restricted myelin antigens. Consequently, disease is driven largely by CD4⁺ T cells. In MS, however, there is considerable evidence that other lymphocytes such as CD8⁺ T cells and B cells likely play important pathogenic roles. In the case of CD8⁺ T cells, it is known that CD8⁺ T cells outnumber CD4⁺ T cells in MS lesions and exhibit oligoclonal expansion (9–11). While the presence of CD8⁺ T cells in CNS lesions is well established, it has been debated whether the CD8⁺ T cells are playing a pathogenic role or protective role. Myelin antigen-specific CD8⁺ T cells have been shown to induce EAE (13–15), but studies of EAE in CD8-deficient mice suggest that CD8⁺ T cells may also have a protective role (23, 24).

Recently some TCR transgenic models of CD8⁺ T cell-driven EAE have been generated. Two of these rely on expression of CD8⁺ transgenic TCRs specific for MBP. These models have been useful for the study of how MBP-specific CD8⁺ T cells undergo tolerance (25, 26). A unique feature of one of these models is the fact that disease can be induced by both control and MBP-expressing vaccinia virus. The disease induced by the two different viruses is indistinguishable and is characterized by weight loss, ataxia and tail weakness (27). The induction of disease by control vaccinia virus is due to the activation of transgenic T cells through endogenously rearranged TCR chains, thus supporting a role for dual-TCR-expressing cells in disease induction. A third model of CD8⁺ T cell induced CNS autoimmunity consists of a human PLP-reactive TCR expressed in an HLA-A3 transgenic mouse (28). In these mice, there is little spontaneous disease but after immunization, about 70% of mice develop mild weakness and 25% go on to develop more severe disease that includes hind limb paralysis. Interestingly, the development of hind limb paralysis was found to be dependent on the recruitment of endogenous CD4⁺ T cells reactive to other myelin epitopes. This observation is reminiscent of our data showing that while CD8⁺ T cells can induce mild disease and optic neuritis on their own, CD4⁺ T cells are the main drivers of disease.

Other models using adoptive transfer of MOG 35-55-reactive CD8⁺ T cells or MBP-reactive T cell clones into lymphopenic hosts also support a pathogenic role for CD8⁺ T cells in disease (13–15); however, the study by York et al. found that adoptive transfer of MOG 35-55-reactive CD8⁺ T cells alone does not induce clinical disease (29). Unfortunately, these recipient mice were not examined histopathologically for the presence of subclinical disease and/or optic neuritis as we have observed in the majority of mice receiving CD8⁺ T cells alone in our study. In contrast, the study by York et al. shows that adoptive transfer of MOG 35-55-reactive CD8⁺ T cells into mice at the peak of disease or into mice subsequently immunized for EAE can ameliorate disease (29). These findings are in line with the studies in CD8-deficient mice that suggest a protective/regulatory role for CD8⁺ T cells in EAE (23, 24). Our data showing the presence IFN-γ-secreting CD8⁺ T cells in the target organ of 1C6 Tg mice with both active disease induction and spontaneous disease further support a pathogenic role for CD8⁺ T cells in EAE. Thus, the majority of studies all employing different methodologies (adoptive transfer of in vitro activated CD8⁺ T cells, immunization of mice that receive CD8⁺ T cells alone, examination of CD8⁺ T cells in situ during induced and spontaneous disease) support a pathogenic role for CD8⁺ T cells in EAE. While these data do not exclude a protective/regulatory role for CD8⁺ T cells in CNS autoimmunity, it seems that such a role may be more the exception than the rule and may only manifest in specific circumstances.

As in MS, genetics have major effects on susceptibility and manifestations of EAE. Accordingly, genetic studies of EAE have identified a number of disease-modifying loci. Interestingly, many of these overlap with loci that have been identified in studies of type 1 diabetes in the NOD mouse, giving rise to the concept that there are ‘common autoimmune genes’ (1–3). In the NOD mouse, many congenic strains in which type 1 diabetes susceptibility loci (Idd) have been introgressed from resistant strains onto the NOD strain have been generated. The 1C6 TCR transgenic mouse now provides a tool that can be used in conjunction with NOD.Idd congenic mice to study the effects of Idd loci on both the selection and effector function of myelin-reactive CD4⁺ and CD8⁺ T cells.

The examination of 1C6 x IgH^MOG mice reveals that there is a gender bias in the development of spontaneous disease. This was not observed in the 2D2 x IgH^MOG mice on the C57BL/6 background. Moreover, on the NOD background the mice do not develop NMO-like disease nor are any ectopic follicle-like structures seen in the CNS. Whether these differences are due to genetic differences in the NOD background or the presence of MOG-reactive CD8⁺ T cells in 1C6 x IgH^MOG mice remains to be determined. Collectively, these data indicate that the simple co-existence of myelin-reactive CD4⁺ T cells and B cells with the same antigen specificity does not result in an NMO-like disease phenotype.

Why some 1C6 x IgH^MOG developed classical EAE and others developed exclusive cerebellar inflammation is unclear. Whether differential involvement of the CD4⁺ T, CD8⁺ T and B cell response is responsible for these two different disease pathologies requires further examination. To accomplish this, it would be necessary to identify mice with cerebellar inflammation antemortem, e.g. using additional tests of cerebellar function, such as the rotarod and treadmill/gait tests.

In summary, we have developed a novel TCR transgenic model of EAE that capitalizes on the unique features of disease and the tools available on the NOD background. Serendipitously, our model spontaneously generates both CD4⁺ and CD8⁺ T cells that express the transgenic TCR and recognize MOG 35–55, allowing for the first time the examination of the relative roles in disease of these two important cell types both in isolation and together. We have found that while CD8⁺ T cells alone can induce mild CNS disease, CD4⁺ T cells are the main drivers of CNS autoimmunity. Additionally, we have found that the simple presence of CD4⁺ T cells and B cells with the same myelin antigen specificity does not necessarily result in the induction of NMO-like disease.

Abbreviations

EAE

Experimental autoimmune encephalomyelitis

MS

Multiple sclerosis

MOG

Myelin oligodendrocyte glycoprotein

NMO

Neuromyelitis optica