C57BL/6 Mice Genetically Deficient in IL-12/IL-23 and IFN-γ Are Susceptible to Experimental Autoimmune Myasthenia Gravis, Suggesting a Pathogenic Role of Non-Th1 Cells¹

Cytokines influence the development and modulation of autoimmune responses. Most cytokines are synthesized by and act on a variety of cells. Moreover, they may influence the synthesis and/or activity of other cytokines, directly or indirectly. Because of the complex, pleiotropic, and redundant signaling through the cytokine network, cytokines have complex and at times even contrasting effects on both normal and autoimmune responses.

Many cytokines are growth and differentiation factors or effector molecules of CD4⁺ T cells. Differentiated CD4⁺ T cells can be classified into different subsets based on the cytokines they secrete (1, 2): The Th1 subset secretes proinflammatory cytokines, such as IFN-γ and IL-2, which activate APC and in mice stimulate the synthesis by B cells of IgG that fix complement. Th1 cells can be cytotoxic and are involved in cell-mediated immune responses. The Th2 subset secretes anti-inflammatory cytokines (e.g., IL-4, IL-5, IL-13), which also stimulate humoral immune responses. In mice, Th2 cytokines promote the secretion of IgE and IgG1, which do not fix complement. The Th3 subset secretes TGF-β and is involved in immunosuppressive mechanisms. Th17 cells secrete IL-17, a proinflammatory cytokine that is involved in the pathogenesis of several experimental autoimmune diseases (1, 2, 3). Other CD4⁺ T cell subsets have been identified: among them, the CD4⁺CD25⁺ T regulatory (Treg)⁴ cells are especially important in maintaining self-tolerance and inhibiting the development of autoimmunity (4).

Myasthenia gravis (MG) and its experimental model, experimental autoimmune myasthenia gravis (EAMG), are Ab-mediated, T cell-dependent autoimmune diseases. Their symptoms are caused by high-affinity IgG against the muscle acetylcholine receptor (AChR) at the NMJ (5, 6). EAMG can be induced in mice and other experimental animals by immunization with AChR (7, 8). In both MG and rodent EAMG, synthesis of pathogenic anti-AChR Ab requires CD4⁺ T cells (5, 6).

Th1 cells are necessary for EAMG development and IL-12, the major growth and differentiation factor for Th1 cells, has a crucial role for development of mouse EAMG. Wild-type (WT) B6 mice treated with exogenous IL-12 during TAChR immunization developed EAMG earlier, and synthesized more IFN-γ and anti-TAChR IgG2a than IL-12 untreated mice (9). Also, BALB/c mice genetically deficient in STAT4 transcription factor, which is crucial for IL-12 intracellular signaling, were EAMG resistant (10). Moreover, B6 mice carrying a disrupted gene for the common p40 subunit shared by both IL-12 and IL-23, and therefore were genetically deficient in both IL-12 and IL-23 (IL-12/IL-23^−/− mice), were resistant to EAMG after TAChR immunization (9, 11): those studies confirmed the importance of Th1 cells from EAMG development, although they did not dissociate possible distinct roles of IL-12 and IL-23 in this model disease.

The role in EAMG of the Th1 effector cytokine, IFN-γ, is unclear. In one study, B6 mice genetically deficient in IFN-γ (IFN-γ^−/− mice) were as susceptible to EAMG as WT B6 mice (11): IFN-γ^−/− and WT mice had comparable concentrations of anti-AChR Ab, and both had IgG and complement at the NMJ (11). Other studies found that mice genetically deficient in IFN-γ or IFN-γ receptor were less susceptible to EAMG than WT mice and produced less anti-AChR Ab (12, 13).

Th17 cells are important in the pathogenesis of several experimental autoimmune diseases. Development of experimental autoimmune encephalomyelitis (EAE) was significantly suppressed in IL-17^−/− mice (2); vaccination against IL-17 suppressed collagen-induced arthritis and EAE (14, 15); and in experimental autoimmune uveitis, neutralizing IL-17 inhibited disease in WT mice (D. Luger and R. R. Caspi, submitted for publication). A pathogenic role of Th17 cells could explain the susceptibility to autoimmunity in animal models that lack Th1 cells (2, 14, 16, 17).

In this study, we have used dKO B6 mice to investigate the effect of simultaneous absence of both IL-12/IL-23 and IFN-γ on EAMG development.

Eight- to 10-wk-old female B6 mice were purchased from The Jackson Laboratory. Mice genetically deficient in IL-12/IL-23 and IFN-γ (dKO mice) were generated as follows: IFN-γ KO mice, originally made by Dalton et al. (18) and IL-12p40 KO mice originally made by Magram et al. (19) were purchased from Jackson ImmunoResearch Laboratories. The strains (both on C57BL/6 background) were crossed to generate heterzygotes, who were then crossed again to generate double knockouts. Genomic DNA was extracted from tails of the pups using a Qiagen kit (DNeasy kit), and PCR was used to determine the genotype of individual mice using published primers (18, 19). The individual mouse was considered double positive if it was positive for the neomycin gene but negative for both IFN-γ and IL-12p40 genes. Homozygous mice of the desired genotype were used as founders for the double knockout colony. The mice were maintained under specific pathogen-free conditions and bred at the animal facility of the University of Minnesota. For all of the experiments, we used only female mice. All of the studies were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Minnesota.

We immunized each mouse with AChR purified from the electric organ of Torpedo californica as we described previously (20). Briefly, we emulsified 25 μg of TAChR in 100 μl of PBS in an equal volume of CFA and used it for multiple s.c. injections along the back of the mouse. We boosted the mice twice, at 4-wk intervals, with the same amount of TAChR in IFA.

We measured the strength of the mice sensitized with pancuronium bromide (0.03 mg/kg) by forcing them to hang from an inverted grid. We described this test in detail previously (11, 21, 22). “Holding time” is the time it took for the mouse to fall three times from the grid. We considered sick mice with holding times shorter than 8.3 min (11, 22). The myasthenic nature of the muscular weakness was verified by injecting the mice with edrophonium chloride (0.1 mg/kg, i.p.), a cholinesterase inhibitor that immediately increases the holding time. We tested the mice every 2 wk.

Twelve weeks after the first immunization, we sacrificed the mice, collected their spleen, and depleted the splenocytes of CD8⁺ T cells using rat anti-mouse CD8 Ab (BD Biosciences Pharmingen) and paramagnetic beads coated with goat anti-rat IgG (Qiagen), as we described previously (22). We cultured the cells for 3 days in the presence of the following Ag: TAChR (2 μg/ml); individual synthetic peptides (∼20 residues long and overlapped by ∼5 residues) spanning the TAChR or mouse AChR α subunit (10 μg/ml), as described in Refs. 10 and 22 . We used Con A (5 μg/ml; Sigma-Aldrich) as positive control. We determined the basal rate of cell proliferation from cell cultured without any Ag or with a synthetic peptide of a random sequence unrelated to AChR. We determined the rate of cell proliferation from the incorporation of [³H]thymidine, measured by liquid scintillation, and expressed as cpm. When an Ag induced a statistically significant increase in the cell proliferation, we calculated the stimulation index (SI: the ratio between the proliferation of cells cultured in the presence of the Ag and in the presence of the unrelated synthetic peptide). We disregarded SI lower than 2.

We immunized WT and dKO mice with TAChR/CFA and after 2 wk boosted them with TAChR/IFA. Within 1 wk after the boost, we collected the mouse splenocytes and used a CD4⁺CD25⁺ regulatory T cell isolation kit (Miltenyi Biotec) to purify the following cell populations: CD4⁺CD25⁺ cells, CD4⁺CD25⁻ cells, and the residual cell population (non-CD4⁺ cells). The non-CD4⁺ cells were further depleted of CD8⁺ cells using rat anti-mouse CD8 Ab and paramagnetic beads coated with goat anti-rat IgG, irradiated (10,000 rad) and used as APC. We cultured 50,000 CD4⁺CD25⁻ cells in U-bottom 96-well plates with 50,000 APC, 1 μg/ml anti-CD3 mAb (2C11; BD, Pharmingen) as stimulant, and with or without 40,000 CD4⁺CD25⁺ cells, for 72 h at 37°C in 5% CO₂. We pulsed the cultures with 1 μCi/well of [³H]thymidine for the last 18 h, harvested the cells, and counted the [³H]thymidine that had been incorporated by the cells. We represent the suppressive activity of the CD4⁺CD25⁺ T cells as percentage of inhibition, which is calculated as follows: 100 × (cpm of cultures without CD4⁺CD25⁺ cells − cpm of cultures with CD4⁺CD25⁺ cells)/cpm of cultures without CD4⁺CD25⁺ cells.

Purified CD4⁺ T cells (10⁶/well) and APC from WT and dKO mice immunized twice with TAChR were incubated with TAChR (50 μg/ml) or Con A or with medium alone in 24-well plates. CD4⁺ T cells were purified by positive selection using CD4 (L3T4) MicroBeads (Miltenyi Biotec) and were cultured in the presence of 4 × 10⁵ irradiated APC (CD8⁺-depleted non-CD4⁺ cells). After 48 h of incubation at 37°C in 5%CO₂, the supernatants were tested for IL-17 using an IL-17 ELISA kit (eBioscience). The samples were tested in a series of dilutions to generate a linear curve. The results were calculated by using the slope of samples divided by the slope obtained in the standard curve expressed in nanograms per milliliter.

The following anti-mouse Ab were used to characterize the phenotype of splenocytes from naive dKO and WT mice: CD3, CD4, CD8, CD30, NK1.1, CD11c, CD19, IgM (all from BD Pharmingen). The Ab were directly conjugated with FITC or PE. Briefly, the staining was done for 30 min on ice in PBS/0.1% NaN₃/1% BSA. The cells were then washed three times, resuspended in PBS, and analyzed. We examined the presence of CD4⁺CD25⁺ regulatory T cells among the mouse spleen CD4⁺ T cells using PerCP-CD4, PE-CD25, and FITC-forkhead box P3 (FoxP3) triple staining. Specifically, intracellular specific marker Foxp3 was detected using the FITC anti-mouse/rat Foxp3 Staining Set (eBioscience) according to the manufacturer’s instruction. Flow cytometry analysis was conducted in a FACScan flow cytometer with CellQuest software (BD Biosciences).

As a source of mouse AChR, we used carcasses of WT B6 mice solubilized with Triton X-100 as we described previously (22) and labeled with¹²⁵I-α-bungarotoxin (¹²⁵I-αBTX). We incubated increasing dilutions of mouse serum with labeled muscle extracts. The complexes of ¹²⁵I-αBTX mouse AChR and IgG were precipitated using Zysorbin (Zymed Laboratories) and were counted in a 5500 gamma counter (Beckman Coulter). We express the Ab concentration as nanomolar of precipitated ¹²⁵I-α BTX binding sites.

We used an ELISA that we described previously (10, 11) to measure the anti-TAChR total IgG and IgG subclasses (IgG1, IgG2b, and IgG2c) in mouse sera collected 12 wk after the first immunization. Briefly, sera were diluted up to 1/4000 and added to ELISA plates (Nalge Nunc) previously coated with 5 μg/ml TAChR in coating buffer (10 mM NaHCO₃, pH 9.6). As a detecting system, we used HRP-conjugated goat anti-mouse IgG (Sigma-Aldrich) or the appropriate biotinylated anti-mouse IgG subclass Ab (BD Biosciences Pharmingen) followed by avidin-HRP (Sigma-Aldrich). The developing solution was made with equal volumes of 2,2′-azino-di (3-ethylbenzthiazoline-6-sulfonate)-peroxidase substrate and H₂O₂ (Kirkegaard & Perry Laboratories). We measured the absorbance at 405 nm with an ELx800 automated microplate reader (Bio-Tek Instruments). For each sample, we calculated the serum concentration of IgG, IgG1, IgG2b, and IgG2c (expressed as micrograms per milliliter) by comparing the slope of the curve obtained for the sample and the slope obtained for the standard curve. Because purified anti-TAChR IgG or IgG subclasses are not available, we constructed standard curves using purified mouse IgG, IgG1, IgG2a, and IgG2b. The standard curves are not ideal, yet they allowed the comparison of results obtained in different experiments.

Mouse AChR was extracted from mice carcasses as described previously (23). Briefly, 1 g of muscle taken from each individual carcass was homogenized and solubilized with 1.5% Triton X-100 in 0.05 M Tris buffer. The AChR content was determined using the ¹²⁵I-αBTX-binding assay we described previously (23, 24). We expressed the concentration of muscle AChR as picomoles of ¹²⁵I-αBTX binding sites per gram of muscle.

Mouse hind limb muscle tissue, frozen in liquid nitrogen and stored at −70°C, was embedded in Tissue-Tek OCT compound (Miles Laboratories) and sectioned in the transverse direction into 10-μm sections using a Jung Frigout 2800E Kryostat (Leica). For each mouse, we analyzed 4–10 muscle sections. We incubated the sections at room temperature in PBS for 10 min and then for 1 h with biotin-conjugated goat anti-mouse IgG polyclonal Ab (Sigma-Aldrich) in PBS containing 2% BSA. We washed the sections three times with PBS and stained them for 1 h at room temperature with Texas Red-labeled αBTX (Molecular Probes), FITC-labeled goat anti-mouse C3 Ab (Nordic Immunological Laboratories), and AMCA-S-labeled streptavidin (Molecular Probes). We washed the sections three times with PBS and viewed them in a fluorescence microscope (Nikon Eclipse E800; Nikon). We collected digital images using the program Image Pro Plus (Media Cybernetics,).

We used repeated measure ANOVA (StatView 5.0.1; Abacus Concepts) and a two-tailed Student’s t test (Excel; Microsoft). We considered a difference significant when p < 0.05.

In three independent experiments, we immunized dKO mice (n = 10, 10, and 13, respectively) and WT mice (n = 9, 6,and 11, respectively) with TAChR/CFA and boosted them twice with TAChR/IFA 4 wk apart. We tested the mouse muscle strength every 2 wk. Fig. 1,A illustrates the frequency of EAMG weakness (i.e., holding times of 8.3 min or less) among the WT and dKO mice used in the three experiments, about 2 wk after the first and second TAChR/IFA boost. In agreement with the results of previous studies (5, 10), the frequency of EAMG weakness differed among experiments: yet, in each experiment, it was the same in the dKO and the WT group. Fig. 1 B depicts the holding times of all of the individual mice used in the three experiments: 38% of the WT mice and 42% of the dKO mice had EAMG symptoms after the first boost and almost 69% of the mice of both groups had EAMG weakness after the second boost.

We sacrificed the mice after the second boost (12 wk after the first TAChR immunization) and we measured their muscle AChR content. In all experiments, the WT and the dKO mice had similar concentrations of muscle AChR, which in most mice were reduced as compared with those of eight naive WT mice that we used as controls. Fig. 2 shows the concentration of muscle AChR (expressed as picomoles per gram of tissue) of the individual WT and dKO mice used in the three experiments and of eight control naive WT mice. Both WT and dKO mice had significantly less muscle AChR than the naive WT mice (p = 0.042 and p = 0.0052, respectively).

TAChR-immunized IL-12/IL-23^−/− mice had IgG, but not complement at the NMJ, suggesting that ineffective synthesis of complement-fixing anti-AChR Ab contributes to their EAMG resistance (11). We examined whether simultaneous absence of IL-12/IL-23 and IFN-γ restored the mouse ability to synthesize anti-muscle AChR Ab that fix complement. We determined the presence of IgG and complement at the NMJ of TAChR-immunized mice using specific fluorescent Ab. We analyzed muscle sections from four WT mice and four dKO mice (two mice for each group from those used in experiments 1 and 2), sacrificed after the second boost, 12 wk after the first TAChR injection. Naive WT mice served as negative controls. For each mouse, we analyzed 20–28 sections. The results we obtained were consistent in all sections, for all of the mice we examined: WT and dKO mice had both IgG and the C3 complement component at the NMJ, identified by the presence of AChR. Fig. 3 shows the results of representative sets of sections.

We collected sera 12 wk after the first TAChR immunization from the mice used in experiments 1 and 2. We used ELISA to measure the concentration of anti-TAChR IgG and IgG subclasses. We obtained similar results for both experiments (reported together in Fig. 4). dKO mice had significantly less anti-TAChR IgG and all IgG subclasses than WT mice. They had virtually no anti-TAChR IgG2c and little anti-TAChR IgG1 and IgG2b.

We used radioimmunoprecipitation assays to measure the anti-mouse AChR Ab in those sera. In both experiments, several mice from both the dKO and the WT groups had measurable anti-muscle AChR Ab in the serum (Fig. 5). The WT mice had serum anti-muscle AChR Ab somewhat more frequently than dKO mice, but the difference was not significant.

We used CD8⁺-depleted splenocytes from mice sacrificed 12 wk after the first TAChR immunization to test their proliferative response to the TAChR and to overlapping peptides spanning the TAChR α subunit sequence. We conducted two independent proliferation assays using cells from mice of the first experiment and one using cells from mice of the second experiment and obtained consistent results. The TAChR-induced proliferative responses of dKO mice, although significant, were consistently weaker that those of WT mice. CD8⁺-depleted splenocytes of WT mice recognized several peptide sequences of the TAChR α subunit (α55–74, α106–126, α146–169, α181–200, α215–249, α346–378, and α406–423), which included sequences previously identified as immunodominant epitopes in B6 mice (α111–126, α146–169, α180–200, α346–368, see Refs. 20 , 22 , 25 , and 26). In contrast, CD8⁺-depleted splenocytes from dKO mice responded to TAChR peptides poorly and recognized only the immunodominant peptides α111–126, α146–161, and α346–364. Fig. 6 shows representative results.

We used CD8⁺-depleted splenocytes from mice sacrificed 12 wk after the first TAChR immunization to test their proliferative response to synthetic peptides spanning the sequence of the mouse muscle AChR α subunit (22). We conducted two independent experiments: Fig. 7 shows their results. We observed small yet significant responses to a few peptides for both WT and dKO mice. The intensity of the peptide-elicited responses was of the same order of magnitude for cells from WT and dKO mice. Both strains recognized peptides within the sequence regions α191–235 and α414–433. WT mice, but not dKO mice, recognized also peptide α276–295.

We used cultures of purified spleen CD4⁺ T cells plus APC from TAChR-immunized dKO and WT mice to determine their ability to secrete IL-17 when restimulated with TAChR in vitro. We found that the TAChR-induced secretion of IL-17 was slightly higher in WT mice, but the difference was not significant (p > 0.05). In contrast, the nonspecific IL-17 production induced in vitro by stimulation with Con A was significantly higher for the CD4⁺ T cells from WT mice than from the dKO mice. Fig. 8 shows the IL-17 concentrations, in nanograms per milliliter, in supernatants of cell cultures from individual dKO or WT mice.

To determine whether genetic absence of both IL-12/IL-23 and IFN-γ grossly affected the development of CD4⁺ or CD8⁺ T cells, or B cells, or dendritic cells, or NK and NK T cells, or of the potentially regulatory CD4⁺ CD25⁺ T cell subset (27, 28), we analyzed splenocytes from naive dKO and WT mice for the presence of the following surface markers, individually or in the appropriate combinations as indicated: CD3, CD4; CD3, CD8; CD3, CD30 (the latter as an indicator of activation of T cells); CD3, NK1.l; NK1.1, CD11.c (a marker of dendritic cells); CD4, CD25; and CD19, IgM. To determine the presence of CD4⁺CD25⁺ Treg cells, we used a triple fluorescence staining of Foxp3 (a specific marker of CD4⁺CD25⁺ regulatory T cells) coupled with surface staining for CD4 and CD25: A of Fig. 9 shows the results of an individual representative experiment done with this approach. This is a more sensitive and specific approach for detection of CD4⁺CD25⁺ regulatory T cells than assessment of the CD25 expression (29). Fig. 9 B depicts a summary of the results of those experiments: each column represents the average of results obtained in several individual mice (n = 3–12). Spleen of naive dKO and WT mice contained comparable concentrations of CD4⁺ T cells, CD8⁺ T cells, activated T cells, B cells, and dendritic cells. In contrast, dKO mice had significantly fewer NK (2.68%), NKT cells (1.06%), and CD4⁺CD25⁺ FoxP3 Treg cells (15.99%) than WT mice (4.49% for NK, 3.3% for NKT, and 17.97% for CD4⁺CD25⁺ FoxP3 Treg cells). We obtained similar results when testing the frequency of CD4⁺CD25⁺ FoxP3 Treg cells in dKO mice and WT mice after three TAChR immunizations (data not shown).

We determined the ability of CD4⁺CD25⁺ Treg cells from naive and TAChR-immunized dKO or WT mice to inhibit the anti-CD3 Ab-induced proliferation of CD4⁺CD25⁻ T cells T cells from the same mice. We conducted five independent experiments using CD4⁺CD25⁺ Treg cells from TAChR-immunized WT and dKO mice and four experiments using CD4⁺CD25⁺ Treg cells from naive WT and dKO mice. In each experiment, we set up four to six cultures for each of the conditions tested. When stimulated with anti-CD3 mAb in the absence of CD4⁺CD25⁺ T cells, cultures of CD4⁺CD25⁻ T cells (50,000 cells plus 50,000 APC) from WT mouse cells proliferated more strong than the cells from dKO mice. The range of [³H]thymidine incorporation (expressed as cpm per well) varied in the different experiments: it was 5,290 cpm to 148,693 cpm in the different experiments that used WT cells and 3,700 cpm to 11,538 cpm in the different experiments that used dKO cells. After TAChR immunization, CD4⁺CD25⁺ Treg cells from dKO mice inhibited the proliferation of CD4⁺ T cells significantly less (p = 0.02) than the CD4⁺CD25⁺ T cells from WT mice. In contrast, the suppressive ability of CD4⁺CD25⁺ Treg cells of naive dKO and WT mice was comparable. Fig. 10 illustrates the average of the results obtained in all of the individual experiments we conducted.

Several studies have shown that Th1 cells and Th1-driven, complement-fixing anti-AChR Ab are important in the development of mouse EAMG (e.g., see Refs. 8 , 14 , and 22). However, the role of IFN-γ is still unclear (11, 12, 13). In this study, we show that B6 mice are susceptible to EAMG in the absence of IFN-γ, IL-12, and IL-23, which are the main differentiation and effector cytokines of Th1 cells (9, 11). dKO mice immunized with TAChR/CFA-IFA developed modest and much lower Ab and CD4⁺ T cell responses to the xenogeneic TAChR than WT mice (Figs. 4 and 6). Yet, dKO mice presented EAMG symptoms (Fig. 1) and developed an autoimmune response against mouse AChR that involved both CD4⁺ T cells and Ab-producing B cells: 1) they had a reduced muscle AChR content, comparable to that of WT mice (Fig. 2); 2) their sera contained anti-mouse AChR Ab in concentrations comparable to those of WT mice (Fig. 5); and 3) the CD4⁺ T cell responses against peptide sequences of mouse AChR α subunit were similar in dKO and WT mice (Fig. 7). Both dKO and WT mice recognized peptide 215–235, which is immunodominant in B6 mice (22) most strongly and consistently. In one of the two experiments we conducted, mice of both strains recognized also peptide 203–218, which overlaps the immunodominant region 215–235, and peptide 414–433, which is frequently recognized by mice of the B6 background after TAChR immunization. WT mice recognized also peptide 191–207, which overlaps peptide 203–218, recognized by both strains, and peptide 276–296, which was not recognized by dKO mice. Recognition of both of these peptides by CD4⁺ T cells of B6 mice with EAMG has been described previously (22). The present finding that CD4⁺ T cells of dKO mice recognized a more limited set of AChR peptides than WT mice is well explained by a less efficient spreading (22) of the repertoire of the autoimmune anti-mouse AChR CD4⁺ T cells in the combined absence of IL-12/IL-23 and IFN-γ, because of the resulting lesser efficiency of APC in dKO mice than in WT mice.

The modest intensity of the anti-TAChR response of dKO mice even after multiple TAChR immunizations in CFA-IFA, and the seemingly paradoxical ability of such a low response to trigger a pathogenic autoimmune response to mouse AChR, suggests that absence of IL-12/IL-23 and IFN-γ might have impaired the sensitization/differentiation and/or function of both the CD4⁺ T and B cells that cooperate in the xenogeneic anti-TAChR response and perhaps that of CD4⁺CD25⁺ Treg cells that curb the progression of the anti-TAChR response to involve the mouse AChR. Without an efficient activation of Treg cells, the modest sensitization of pathogenic CD4⁺ T cells and the IL-12-independent synthesis of complement-fixing anti-AChR Ab that occur in IL-12/IL-23-deficient mice (11) might trigger a pathogenic, self-sustaining anti-muscle AChR response and result in the appearance of EAMG symptoms.

CD4⁺CD25⁺ Treg cells play a key role in maintaining peripheral tolerance to self-Ag (30). Their activity reduced in several experimental autoimmune diseases and human autoimmune diseases (30, 31). IFN-γ is important for activation and differentiation of CD4⁺CD25⁺ Treg cells (32, 33, 34). Mice genetically deficient in STAT1, a mediator of intracellular IFN-γ signaling, were more susceptible to EAE than WT mice and had a reduced number and functional impairment of CD4⁺CD25⁺ Treg cells (32). After immunization with collagen, mice genetically deficient in IFN-γ were more susceptible to collagen-induced arthritis: they had fewer CD4⁺CD25⁺ Treg cells than WT mice and their CD4⁺CD25⁺ Treg cells had impaired suppressive activity (33). MG patients may have a functional impairment of thymic CD4⁺CD25⁺ Treg cells (35). Moreover, the number of circulating CD4⁺CD25⁺ Treg cells in MG patients increased after thymectomy, and the increase correlated with the symptom improvement (36).

Both naive and TAChR-immunized dKO mice had slightly but significantly fewer Foxp3⁺C25⁺CD4⁺ Treg cells than WT mice (Fig. 9 and data not shown). Moreover, after TAChR immunization, CD4⁺CD25⁺ Treg cells from dKO mice had a significantly less suppressive ability than CD4⁺CD25⁺ Treg cells from WT mice (Fig. 10), suggesting that in our system the decreased suppressive activity of CD4⁺CD25⁺ Treg in dKO mice is not due to the absence of IFN- γ only, but it is probably an effect of a combination of factors.

Other mechanisms could be involved in the EAMG susceptibility of dKO mice. Naive dKO mice had significantly fewer NK and NKT cells in their spleen than WT mice (Fig. 9). NK cells facilitate EAMG induction (37): their reduced number is not likely related to the EAMG susceptibility of dKO mice. In contrast, NKT cells may have a role in immunoregulation and in the maintenance of self-tolerance (38, 39). Thus, their reduction in dKO mice might contribute to the development of an autoimmune response to mouse AChR after TAChR immunization. In addition, a recent study demonstrated that α-galactosylceramide-activated NKT cells control the function of CD4⁺CD25⁺ Treg cells: mice treated with α-galactosylceramide had an increased number and suppressive function of CD4⁺CD25⁺ Treg cells (40). The decreased number of NKT cells in dKO mice might contribute to the reduced activity of their CD4⁺CD25⁺ Treg cells.

The recently discovered Th17 CD4⁺ T cell subset, an independent lineage from the Th1 and the Th2 cells, seems to be responsible for the autoimmune responses that occur in several experimental models of autoimmunity, such as collagen-induces arthritis, EAE (2, 14) and experimental autoimmune encephalomyelitis (D. Luger and R. R. Caspi, submitted for publication). The present observations, that TAChR-specific CD4⁺ T cells obtained from TAChR-immunized dKO and WT mice secrete similar levels of IL-17 when stimulated in vitro with TAChR (Fig. 8), indicate that an effective anti-TAChR sensitization of Th17 cells occurs in dKO mice. IL-23, which is also absent in dKO mice, has a role in maintaining the established Th17 responses (41), but it is not needed for the sensitization and differentiation of Th17 cells. This could explain how Th17 cells can differentiate, and could play a pathogenic role in EAMG of dKO mice, even if this observation is inconsistent with other findings (3, 42, 43).

The present results indicate that in EAMG, an Ab-mediated experimental autoimmune disease, Th1 cytokines, although important for development of autoimmune diseases, are not the exclusive players. Other mechanisms, such as those mediated by Th17 could have a role in both T cell- and Ab-mediated autoimmunity: they might become more obvious in situations when the Th1 cell function is impaired, as in the dKO mice we used here.

The authors have no financial conflict of interest.