Immune Responses to Retinal Self-Antigens in CD25⁺CD4⁺ Regulatory T-Cell–Depleted Mice

Abstract

purpose. Prior work has shown that autoimmune uveoretinitis develops spontaneously in CD25⁺CD4⁺ regulatory T-cell–depleted mice (Tr-depleted mice). In this study, the generation of autoantibodies and autoreactive T-cells specific to retinal antigens was examined in Tr-depleted mice with uveoretinitis, and the pathogenic and immunogenic abilities of the autoreactive T cells were evaluated.

methods. Tr-depletion was achieved in (C57BL/6 x A/J) F1 (B6A) mice by thymectomy on day 3 of life followed by intraperitoneal injection of an anti-CD25 mAb. At 6 months of age, autoantibodies to the retina were evaluated by indirect immunofluorescence, and total IgG2a levels in sera were assessed by ELISA. The pathogenic abilities of the splenic T cells were examined by adoptive transfer to syngeneic nu/nu mice, and the proliferation responses and the secretion of granulocyte-macrophage–colony-stimulating factor (GM-CSF), IFN-γ, and IL-10 on stimulation by retinal self-antigens was also evaluated.

results. Autoantibodies to the retinal photoreceptor cell layer were detected in Tr-depleted mice, and the titers correlated well with the grades of inflammatory lesions. The splenic CD4⁺ T cells of Tr-depleted mice induced uveoretinitis in the recipients by adoptive transfer and exhibited proliferative responses and secretion of IFN-γ, but not of IL-10, by in vitro stimulation with S-Ag and interphotoreceptor retinoid-binding protein (IRBP). Moreover, the total IgG2a level in serum was markedly and significantly augmented in Tr-depleted mice.

conclusions. The results suggest that in Tr-depleted mice in which uveoretinitis develops, S-Ag and IRBP-specific T cells are spontaneously sensitized and shifted to a Th1-phenotype. These sensitized T cells may account for the development of autoimmune uveoretinitis in Tr-depleted mice.

Uveitis is a group of diverse intraocular inflammatory diseases that cause serious visual loss and morbidity and is characterized by inflammatory attack of the uvea and neuroretina. ¹ Uveitis can be classified into two groups: infectious and noninfectious. Noninfectious uveitis, including Behçet’s disease, Vogt-Koyanagi-Harada disease, and sympathetic ophthalmia, is thought to have an autoimmune origin and involves activated self-reactive T cells in the development. ¹ Experimental autoimmune uveoretinitis (EAU) is an animal model of human uveitis produced in susceptible animals by immunization with retinal self-antigens (Ag) such as S-antigen (S-Ag) and interphotoreceptor retinoid-binding protein (IRBP). ^{2 3 4} Although this model allows analysis of the ensuing immunogenic responses and processes of the disease, it does not permit investigation of how immunologic tolerance to ocular self-antigens is broken down. To understand the cause and the pathogenic mechanism of autoimmune uveoretinitis, it is necessary to elucidate how T cells reactive to ocular self-antigens are controlled normally in vivo, so as not to elicit harmful autoimmune responses, and what conditions are required for their activation and expansion to mediate autoimmune disease. 

Two major mechanisms have been demonstrated by which tolerance to self-antigens is maintained. One is central tolerance in the thymus by clonal deletion (negative selection) of self-reactive T cells ^{5 6} and another is peripheral tolerance. Egwuagu et al. ⁷ reported the detection of IRBP transcript and its protein in the thymus of mouse strains resistant to EAU, but not in the thymus of the susceptible strains. Furthermore, recent studies have found that the transcription factor, autoimmune regulator (AIRE), regulates autoimmune diseases including uveoretinitis, by promoting the ectopic expression of peripheral tissue-restricted antigens in medullary epithelial cells of the thymus. ^{8 9} Therefore, central tolerance by clonal deletion of S-Ag- and IRBP-reactive T cells in the thymus is an important mechanism in the prevention of autoimmune uveoretinitis. However, because immunization of susceptible mice with IRBP combined with a potent adjuvant results in development of uveoretinitis, ¹⁰ this evidence indicates that some of the IRBP-reactive T cells escape the thymus-negative selection and exist in the periphery. Although thymus-negative selection is critical, peripheral tolerance is also involved in maintaining immunogenic tolerance to ocular self-antigens. 

Past studies have demonstrated that self-reactive T cells that have escaped thymus-negative selection fail to be activated in the periphery because of seclusion from the target self-antigens, low avidities of T-cell receptors (TCRs), or lack of costimulation from antigen-presenting cells. ^{11 12 13} There is accumulating evidence that, besides these passive mechanisms of self-tolerance, regulatory T-cell–mediated dominant control of self-reactive T cells also contributes to the maintenance of immunotolerance. The term “regulatory T cells” is used to denote a variety of immunoregulatory cells that can be subdivided into several subsets based on the expression of cell surface molecules and the pattern of cytokine production. ^{14 15} These subsets of regulatory T cells have been characterized in experimental models using different assays; therefore, the interrelationship between the subsets is difficult to understand. One of the best characterized subsets of CD4⁺ regulatory T cells is defined by its constitutive expression of the IL-2R-γ chain (CD25⁺CD4⁺ T cells). ^{16 17} CD25⁺CD4⁺ regulatory T cells are thought to be generated in the thymus and migrate to the periphery after day 3 of life. In genetically susceptible mice thymectomized on day 3 of life, the development of organ-specific autoimmune diseases such as gastritis, oophoritis, and prostatitis can be prevented by reconstitution of the mice with CD25⁺CD4⁺, ^{18 19 20 21} but not CD25⁻ T cells by day 10 to 14 of life. ²² Furthermore, when the CD4⁺ T cell fraction isolated from peripheral lymphoid tissues of normal adult mice is depleted of CD25⁺CD4⁺ T cells and the remaining CD25⁻CD4⁺ T cells are injected into nu/nu recipients, organ-specific autoimmune disease develops spontaneously in these animals at a high incidence, without exogenous immunization with self-antigens. Again, cotransfer of cell populations enriched in CD25⁺CD4⁺ T cells inhibits the induction of disease by CD25⁻CD4⁺ T cells. ^{23 24} We hypothesized that if retinal self-antigen–specific T cells are also controlled by CD25⁺CD4⁺ regulatory T cells in the periphery, then uveoretinitis may develop in CD25⁺CD4⁺ regulatory T-cell–depleted mice. In fact, although autoimmune uveoretinitis did not develop in (C57BL/6 x A/J) F1 (B6A) mice thymectomized on day 3 of life, after injection of anti-CD25 mAbs into Tx-3 mice to further eliminate CD25⁺CD4⁺ T cells (Tr-depleted mice), uveoretinitis was induced spontaneously in the Tr-depleted mice. ²⁵  

In the present study, we characterized the cellular and humoral immune responses to the retina in Tr-depleted mice in which uveoretinitis developed and also evaluated the pathogenic and immunogenic abilities of the autoreactive T cells. Our results demonstrated that autoantibodies in the serum of Tr-depleted mice with uveoretinitis reacted with the retinal photoreceptor layer where S-Ag and IRBP are located and that the autoreactive T cells were able to induce uveoretinitis in the recipients by adoptive transfer and exhibit proliferation responses and secretion of IFN-γ, but not IL-10, on stimulation by S-Ag and IRBP in vitro. Moreover, the serum IgG2a level was markedly and significantly increased in Tr-depleted mice, suggesting that self-reactive T cells including S-Ag and IRBP-reactive T cells were sensitized and shifted to Th1 phenotype in vivo, which would induce multiple autoimmune diseases. 

Materials and Methods

Animals

Specific pathogen-free (C57BL/6 x A/J) F1 (B6A) mice were raised for this experiment. B6A nude (nu/nu) mice were raised as described previously. ²⁶ All mice were maintained in the animal facilities of Aichi Cancer Research Institute and were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Antigens

S-Ag and IRBP were purified from bovine retina by using methods described elsewhere. ^{4 27} S-Ag and IRBP preparations were aliquoted and stored at −80°C. BSA and porcine thyroglobulin were purchased from Sigma-Aldrich (St. Louis, MO). 

Culture Medium

Culture medium was composed of RPMI 1640 medium, 10 mM HEPES, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 μg/mL streptomycin (all from BioWhittaker, Walkersville, MD), and 1 × 10⁻⁵ M 2-mercaptoethanol (Sigma-Aldrich), and supplemented with 10% fetal bovine serum (Sigma-Aldrich). 

Production of CD25⁺CD4⁺ T-Cell–Depleted Mice

B6A mice were thymectomized on day 3 of life by a method described previously (Tx-3 mice). ²⁶ In addition, Tx-3 mice from 3 weeks of age were injected with 1 mg of rat mAbs (IgG₁) against mouse IL-2R (PC61) ²⁸ twice a week for 4 weeks (Tr-depleted mice). 

Histopathology and Uveoretinitis Grading

Eyes were carefully enucleated from 6-month-old naive, Tx-3, and Tr-depleted mice and fixed in 10% buffered formalin. Specimens were dehydrated through graded alcohol and embedded in methacrylate. Serial transverse sections through the papillary–optic nerve plane were cut and stained with hematoxylin and eosin. Five sections cut at different levels were examined for each eye in a masked fashion, and the presence and extent of lesions were determined. The severity of uveoretinitis was scored on a scale of 0 to 2. In brief: 0, normal retina; 1, partial retinal destruction with lymphocyte infiltration; and 2, complete loss of the photoreceptor layer from the layer of rods and cones to the outer nuclear layer. 

Flow Cytometry

Spleen cells obtained from 3-month-old naive, Tx-3, and Tr-depleted mice were stained with FITC-conjugated anti-CD4 (GK1.5) and biotin-conjugated anti IL-2R (PC61). PE-conjugated avidin (Biomeda, Foster City, CA) was added as a second reagent. The spleen cells were assayed using a fluorescence-activated cell sorter (FACScan; BD Biosciences, Mountain View, CA). 

Indirect Immunofluorescence

Six-micrometer-thick cryostat tissue sections were prepared from the eyeball of a normal adult B6A mouse. The sections were fixed in acetone and incubated with test serum (1:40–1:2560 dilutions) for 40 minutes, washed in PBS, and incubated with FITC-labeled anti-mouse IgG (Cappel, Durham, NC) for 40 minutes. Blood samples were collected under anesthesia from tail veins of naive, Tx-3, and Tr-depleted mice aged 1 to 6 months. 

Splenic T-Cell Purification

Spleens were removed from 6-month-old naive, Tx-3, and Tr-depleted mice and pressed through a nylon mesh to produce a single-cell suspension. Red blood cells were lysed with Tris-NH₄Cl. Splenic T cells were subsequently purified to more than 95% by a T-cell enrichment column (R&D Systems, Minneapolis, MN). 

T-Cell Proliferation Assay

Purified splenic T cells (4 × 10⁴) were added to 96-well plates containing 4 × 10⁵ irradiated syngeneic spleen cells and cultured in triplicate with or without various concentrations of BSA, thyroglobulin, S-Ag, or IRBP. Cells were cultured for 72 hours at 37°C in an atmosphere of 5% CO₂, pulsed with 0.5 μCi [³H]thymidine for 12 hours before termination of culturing, and harvested onto glass filters, using an automated cell harvester (Tomtec, Orange, CT). Radioactivity was assessed by liquid scintillation spectrometry, and data are presented as the stimulation index (SI), which is the mean counts in cultures with stimulus/mean counts in control cultures without stimulus. 

GM-CSF, IL-10, and IFN-γ Assays

Purified splenic T cells (4 × 10⁴) were added to 96-well plates containing 4 × 10⁵ irradiated syngeneic spleen cells and cultured with or without various concentrations of BSA, thyroglobulin, S-Ag, or IRBP. Seventy-two hours later, supernatants were collected and assayed for granulocyte-macrophage–colony-stimulating factor (GM-CSF), IL-10, and IFN-γ by ELISA kits (Endogen, Cambridge, MA), according to the manufacturer’s instructions. 

IgG2a Production Assay

Serum samples were obtained from five each of naive mice, Tx-3 mice, and Tr-depleted mice at the time of death at 6 months of age. Serum levels of IgG2a were determined by mouse IgG2a ELISA quantitation kits (Bethil, Montgomery, TX). Briefly, goat anti-mouse IgG2a Abs were coated on 96-well flat plates, and then serial dilutions of the standards and samples were added, followed by horseradish-peroxidase–conjugated goat anti-mouse IgG2a Abs. The plates were developed with chromogenic substrate reagent (TMB) solution, and stopped with 2 M HCl. Absorbance at 450 nm was measured on a microtiter plate reader. 

Adoptive Transfer of Uveoretinitis

Purified splenic T cells obtained from 6-month-old naive, Tx-3, and Tr-depleted mice with autoantibodies against the retina were treated with complement together with PBS, anti-CD4 mAb (H129.19), or anti CD8 mAb (53-6.7; BD PharMingen, San Diego, CA). Viable T cells (2 × 10⁷) were injected into syngeneic naive nude mice. Four weeks after injection, the recipient mice were killed, and development of uveoretinitis was histologically evaluated. 

Statistics

The statistical significance of differences between means of each experimental group was determined by analysis of variance and the Fisher protected least-significant difference (PLSD) test. Mean differences were considered to be significant when P < 0.05. 

Results

Depletion of CD25⁺CD4⁺ T Cells by Thymectomy on Day 3 of Life and Injection of Anti-CD25 mAbs

To produce Tr-depleted mice, B6A mice were thymectomized on day 3 of life (Tx-3 mice). From the age of 3 weeks, the Tx-3 mice were injected intraperitoneally with 1 mg of anti-CD25 mAbs (PC61) twice a week for 4 weeks (Tr-depleted mice). As baseline data, we measured the population of CD25⁺CD4⁺ T cells in the spleen of naive, Tx-3, and Tr-depleted mice at the age of 3 months. Total splenocytes were collected from each group of mice, stained for CD4 and CD25 molecules, and analyzed by flow cytometry. Representative results are displayed in Figure 1 . Whereas 3.4% of total splenocytes in naive B6A mice were CD25⁺CD4⁺ T cells, the population of CD25⁺CD4⁺ T cells was reduced to 1.64% in Tx-3 mice, although the number of total CD4⁺ T cells also decreased. However, the fraction of CD25⁺CD4⁺ T cells was further reduced to 0.62% in Tr-depleted mice. It is uncertain whether this small population of CD25⁺CD4⁺ T cells in Tr-depleted mice are regulatory T cells or activated T cells, since some Tr-depleted mice had autoimmune diseases, such as dacryadenitis and prostatitis (data not shown), which developed as early as 3 months of age, although autoimmune uveoretinitis was not evident. 

Autoimmune Diseases Induced in Tr-Depleted Mice

Table 1 indicates the incidence of autoimmune diseases detected in Tx-3 mice and Tr-depleted mice. The histopathological observations in Table 1 were performed using mice at the age of 6 months. Although gastritis, dacryadenitis, and prostatitis were observed in both groups of mice, the incidence of these lesions was apparently higher in Tr-depleted mice than in Tx-3 mice. Furthermore, uveoretinitis was observed in Tr-depleted mice but not in Tx-3 mice, although the incidence of uveoretinitis was not high (20%). 

The ocular histopathology of a Tr-depleted mouse and the results of indirect immunofluorescence using the sera are shown in Figure 2 . Uveoretinitis characterized by massive lymphocyte infiltration of the retina together with disappearance of the photoreceptor cell layer and outer nuclear layer developed spontaneously in Tr-depleted mice (Fig. 2B) . In addition, indirect immunofluorescence technique demonstrated circulating autoantibodies to the layer of rods and cones of retina in the sera of Tr-depleted mice with uveoretinitis (Fig. 2F) . Different from experimental autoimmune uveoretinitis in mice induced by immunization with IRBP, cell infiltration into the choroid was not observed and the retinal lesions were partial with some areas remaining intact (data not shown). The correlation between individual uveoretinitis scores and titers of circulating autoantibodies to the retinal photoreceptor cell layer in Tx-3 and Tr-depleted mice is indicated in Figure 3 . Corresponding to the results of histopathological examination, autoantibodies to the retina were not observed in Tx-3 mice, whereas the autoantibodies were consistently detected in Tr-depleted mice with uveoretinitis, but not in the mice without uveoretinitis. Furthermore, the titers of autoantibodies correlated well with the grades of inflammatory lesions: Tr-depleted mice with higher titer of the autoantibodies exhibited more severe uveoretinitis. Because the detection of autoantibodies to the retinal photoreceptor cell layer in Tr-depleted mice reflected the occurrence of uveoretinitis, Tr-depleted mice with the autoantibodies were used in the subsequent in vitro experiments. 

Adoptive Transfer of T Cells Obtained from Tr-Depleted Mice

We next performed adoptive transfer experiments to examine whether splenic T cells obtained from Tr-depleted mice with the autoantibodies are able to induce uveoretinitis and which subset of T cells is responsible. As shown in Table 2 , adoptive transfer of T cells from either naive or Tx-3 mice did not induce uveoretinitis in any recipient mouse, although T cells from Tx-3 mice induced dacryadenitis and prostatitis in the recipients (data not shown). However, adoptive transfer of splenic T cells from Tr-depleted mice with autoantibodies induced uveoretinitis in three of four recipient B6A nu/nu mice, at a severity level of grade 2 (Fig. 2C , Table 2 ). The splenic T cells depleted of the CD4⁺ subset failed to adoptively transfer uveoretinitis, whereas the splenic T cells depleted of the CD8⁺ subset preserved the ability to induce uveoretinitis. These results indicate that autoreactive T cells are sensitized in Tr-depleted mice and that some of the T cells, probably mainly CD4⁺ T cells, acquire the property to induce uveoretinitis. 

Self-Antigens for Spontaneously Sensitized T Cells in Tr-Depleted Mice

S-Ag and IRBP are retina-specific antigens localized in the retinal photoreceptor cell layer, and they were targeted by autoantibodies in Tr-depleted mice with uveoretinitis. As mentioned in the introduction, since immunization with these antigens induces autoimmune uveoretinitis, S-Ag- and IRBP-reactive T cells are present normally in the body but are inactivated under normal conditions. Therefore, we examined whether S-Ag or IRBP is a target antigen for the spontaneously sensitized T cells in Tr-depleted mice. Splenic T cells were collected from 6-month-old naive, Tx-3, and Tr-depleted mice and stimulated with BSA, thyroglobulin (used to induce experimental autoimmune thyroiditis), S-Ag, or IRBP in the presence of irradiated syngeneic spleen cells serving as antigen-presenting cells. Representative results of antigen-stimulated splenic T cell proliferation responses are depicted in Figure 4 . None of the three groups showed T cell proliferation responses to stimulation with BSA or thyroglobulin. However, T cells from both Tx-3 mice and Tr-depleted mice exhibited proliferation responses to S-Ag and IRBP challenge in a dose-dependent manner. If the S-Ag or IRBP-specific T cells were naive, they would not proliferate in vitro after primary stimulation with the antigen. These results indicate that the T cells specific for S-Ag and IRBP preexist and are sensitized in Tx-3 and Tr-depleted mice. 

After examining the T-cell proliferation response, we measured the amounts of GM-CSF in these cultures, which is an alternative method for evaluation of the level of T-cell activation. Figure 5 shows GM-CSF production by splenic T cells from naive, Tx-3, and Tr-depleted mice when stimulated with BSA, thyroglobulin, S-Ag, or IRBP. Different from the results of T-cell proliferation responses, an obvious dose-dependent GM-CSF production in response to S-Ag or IRBP stimulation was observed in Tr-depleted mice but not in Tx-3 mice. These results suggest that S-Ag- and IRBP-specific T cells were more sensitized in Tr-depleted mice than in Tx-3 mice. In addition, a greater amount of GM-CSF was produced in response to IRBP stimulation than S-Ag stimulation, suggesting that IRBP was a more potent self-antigen than S-Ag in Tr-depleted mice. 

Preferential Production of Th1-Type Cytokine in Tr-Depleted Mice

After determining that T cells specific for S-Ag and IRBP are sensitized in Tr-depleted mice, we subsequently examined what types of cytokines are produced by the sensitized T cells. Similar to the earlier experiment, splenic T cells from 6-month-old naive, Tx-3, and Tr-depleted mice were cultured with irradiated syngeneic spleen cells and S-Ag, or IRBP, and the amounts of IFN-γ and IL-10 in the supernatants were measured by ELISA. Figure 6 displays representative results of IFN-γ and IL-10 production. T cells from Tr-depleted mice produced IFN-γ in response to stimulation by both S-Ag and IRBP, although the amount of IFN-γ induced by S-Ag stimulation was significantly less than that by IRBP stimulation. T cells from Tx-3 mice produced IFN-γ only in response to IRBP stimulation, and the magnitude was much lower than that in Tr-depleted mice. In contrast, T cells from Tx-3 mice and Tr-depleted mice exhibited no antigen-induced IL-10 production when they were stimulated with S-Ag or IRBP. These results demonstrated that retinal self-antigen–specific T cells, especially IRBP-specific T cells, had already shifted to Th1 phenotype in Tr-depleted mice manifesting autoimmune uveoretinitis. 

IgG2a Levels in Sera of Tr-Depleted Mice

If the S-Ag– or IRBP-specific T cells had shifted to Th1-type cells in vivo, other pathogenic T cells should also have shifted to the Th1-phenotype. Because Th1-cell–mediated immune responses promote class switching of immunoglobulin to IgG2a, we examined IgG2a production in sera of naive, Tx-3, and Tr-depleted mice. The results are displayed in Figure 7 . Although IgG2a levels were significantly augmented in sera of both Tx-3 and Tr-depleted mice compared with that of naive mice, the IgG2a level in Tr-depleted mice was significantly higher than that in Tx-3 mice. It is conceivable that from these results, Th1-cell–mediated immune responses are more augmented in Tr-depleted mice than in Tx-3 mice, which would account for the differences in incidence of autoimmune diseases in multiple organs. 

Discussion

In the present study that Tr-depleted mice with uveoretinitis produced autoantibodies to the retinal photoreceptor cell layer, and their T cells, which induced uveoretinitis in naive nu/nu recipients by adoptive transfer, were sensitized by S-Ag or IRBP and had already shifted to a Th1 phenotype in vivo. 

As shown in Table 1 , Tx-3 mice have various organ-specific autoimmune diseases, such as dacryadenitis and prostatitis, but uveoretinitis was not induced in these mice. The additional process of injecting anti-CD25 mAbs into Tx-3 mice to further deplete CD25⁺CD4⁺ T cells increased the incidence of the other autoimmune diseases and also induced uveoretinitis and thyroiditis (Table 1) . Therefore, the small number of CD25⁺CD4⁺ T cells observed in the spleen of Tx-3 mice (Fig. 1) were considered to be regulatory T cells. However, it is uncertain whether the CD25⁺CD4⁺ T cells detected in the spleen of Tr-depleted mice were regulatory T cells for the following reason. Flow cytometry was conducted in cells from mice aged 3 months, at which age some Tr-depleted mice were already exhibiting dacryadenitis and prostatitis (data not shown), and the pathogenic T cells responsible for these autoimmune diseases would have been activated and also would express the CD25 molecule. 

McHugh and Shevach ²⁹ reported that autoimmune gastritis develops in BALB/c mice thymectomized on day 3 of life, whereas the disease was rarely induced in mice depleted of CD25⁺CD4⁺ T cells by treatment with anti-CD25 mAbs (PC61). However, because adoptive transfer of splenocytes from mice treated with anti-CD25 mAbs induces gastritis in nu/nu mice, they proposed that nonspecific lymphopenia-induced proliferation as a second signal is necessary for CD25⁻CD4⁺ T cells to induce autoimmune diseases. In our study, thymectomy was performed before administration of anti-CD25 mAbs to produce Tr-depleted mice, so that nonspecific lymphopenia-induced proliferation of CD25⁻CD4⁺ T cells should have been induced in both Tr-depleted mice and Tx-3 mice. Moreover, our model rules out the possibility that CD25⁺CD4⁺ T cells regenerated from the thymus after treatment with anti CD25 mAbs suppress the development of autoimmune diseases. Our present data showing that further depletion of CD25⁺CD4⁺ T cells by the additional injection of anti-CD25 mAbs into Tx-3 mice increases the incidence of autoimmune diseases do not contradict the results in their report. 

The presence of autoantibodies against the retinal photoreceptor cell layer in Tr-depleted mice reflected the occurrence of uveoretinitis, and we also demonstrated that Tr-depleted mice with higher titers of the autoantibodies developed more severe uveoretinitis. However, because organ-specific autoimmune diseases including uveoretinitis are T-cell–mediated disease, it is unlikely that the autoantibodies produced in Tr-depleted mice directly induce uveoretinitis. We speculate that retinal self-antigens, which are secluded from the body’s defense system under normal conditions, are exposed to systemic immune mechanisms as a result of ocular inflammation in Tr-depleted mice, and autoantibodies against the photoreceptor cell layer are produced as a result. Moreover, if the magnitude of the ocular inflammation is more severe, a greater amount of the retinal self-antigens are exposed. Therefore, the antibody titers correlate well with the grades of inflammatory lesions, as shown in Figure 3 . 

Because T cell proliferation responses to S-Ag and IRBP challenges were elicited in both Tx-3 mice and Tr-depleted mice at similar magnitudes, the T cells specific for S-Ag or IRBP were adequately sensitized in Tx-3 mice. However, GM-CSF production by T cells stimulated by S-Ag or IRBP was observed only in Tr-depleted mice but not in Tx-3 mice. In addition, T cells of Tx-3 mice produced no IFN-γ after S-Ag stimulation, and only scant IFN-γ, even with IRBP stimulation. From these results, we speculate that S-Ag- and IRBP-specific T cells are spontaneously sensitized in Tx-3 mice as well as in Tr-depleted mice, but they fail to differentiate into fully effector cells that are required for the induction of uveoretinitis. 

However, a question arises as to why pathogenic T cells in Tx-3 mice are able to cause other autoimmune diseases including dacryadenitis and prostatitis and why the frequencies of dacryadenitis and prostatitis are markedly higher than that of uveoretinitis in Tr-depleted mice. One of the reasons may be the nature of the eye, which is an immune-privileged site. Ocular immune privilege is generated and maintained by active suppressive features. ^{30 31} Aqueous humor in the eye prevents activation of T cells to proliferate, produce cytokines, and differentiate into effector cells. In contrast, ocular parenchymal cells constitutively express Fas ligand (CD95L), which induces activated Fas (CD95)-expressing T cells to undergo apoptosis. ³² Moreover, injection of Ag into the anterior chamber of the eye results in systemic impairment of Ag-specific delayed hypersensitivity, termed anterior chamber-associated immune deviation (ACAID), and is one of several central mechanisms that sustain immune privilege in the eye. ³³ Recent studies have indicated that Ag delivered into the anterior chamber induces generation of CD4⁺ and CD8⁺ regulatory T cells that downregulate delayed hypersensitivity. ³¹ However, Ohta et al. ³⁴ have reported that ACAID cannot be induced in eyes with EAU. ³⁴ Because our present data indicate that uveoretinitis is spontaneously induced in mice depleted of CD25⁺CD4⁺ T cells, CD25⁺CD4⁺ T cells may be involved in the establishment of ACAID. We are undertaking studies to examine the nature of immune privilege in the eye of Tr-depleted mice. 

EAU can be induced in susceptible animals by immunization with S-Ag or IRBP. ^{2 3 4} The S-Ag- or IRBP-specific effector T cells in EAU have been found to possess a Th1-like phenotype (high IFN-γ, low IL-4), ^{8 35} and this evidence is compatible with that in other animal models of autoimmune diseases, such as experimental autoimmune encephalomyelitis (EAE) and adjuvant-induced arthritis. ³⁶ Our present study also indicated that splenic T cells obtained from Tr-depleted mice, which were able to induce uveoretinitis by adoptive transfer, were Th1-type cells producing IFN-γ but not IL-10 on S-Ag or IRBP stimulation. Moreover, because total IgG2a levels in serum were elevated in Tr-depleted mice and also in Tx-3 mice, it is conceivable that not only the uveitogenic T cells but also the pathogenic T cells that are responsible for the development of the other autoimmune diseases were sensitized and shifted into the Th1 phenotype. 

Our results do not indicate whether S-Ag or IRBP is the first target antigen for uveitogenic T cells in Tr-depleted mice. It is difficult to elucidate the primary uveitogenic antigen in this model in which autoimmune uveoretinitis develops spontaneously without immunization with self-antigens. However, because T cells specific for both S-Ag and IRBP are activated in Tr-depleted mice, epitope spreading between these retinal self-antigens must have occurred. Furthermore, T cells from Tr-depleted mice were more reactive to IRBP than to S-Ag, and some of their T cells responded only to IRBP and not to S-Ag (data not shown), suggesting that IRBP is the primary target antigen and S-Ag is the secondary target. However, the possibility that retinal self-antigens other than IRBP or S-Ag participate in the development of uveoretinitis cannot be excluded. 

Polyclonal TCR stimulation (for example, with anti CD3 Ab) as well as antigen-specific stimulation can activate CD25⁺CD4⁺ T cells to mediate suppression, whereas irrelevant antigens incapable of activating CD25⁺CD4⁺ T cells fail to evoke suppression. ^{37 38} Once CD25⁺CD4⁺ regulatory T cells are stimulated, the suppression they mediate is not antigen specific. They suppress the proliferation of not only T cells with the same antigen specificity but also that of other T cells specific for other antigens. ³⁹ This implies that some CD25⁺CD4⁺ T cells may recognize self-antigens and be stimulated by them to regulate self-reactive pathogenic T cells in the normal internal environment. In other words, CD25⁺CD4⁺ regulatory T cells reactive for organ-specific self-antigens such as S-Ag and IRBP may also be present in the periphery of normal mice. We have reported that some self-antigens induce activation of tissue-specific suppressor T cells that control self-tolerance. ²⁷ These suppressor T cells may be identical with the CD25⁺CD4⁺ regulatory T cells, because they express CD25 molecule by stimulation with the self-antigens in vivo. 

In a recent study, CD4⁺ T cells expressing high levels of CD25 and exhibiting in vitro characteristics identical with those of CD25⁺CD4⁺ regulatory T cells in mice were detected in human peripheral blood. ⁴⁰ With TCR cross-linking, CD25⁺CD4⁺ T cells did not proliferate, but instead totally inhibited CD4⁺-activated T cells from proliferation and secretion of cytokines in a contact-dependent manner. Because spontaneous autoimmune diseases including uveoretinitis develop in CD25⁺CD4⁺ regulatory T-cell–depleted mice, it is possible that alternations of CD25⁺CD4⁺ regulatory T cells are involved in the induction of human autoimmune disorders. 

 Supported by the Japanese Society for the Promotion of Science.

 Submitted for publication October 7, 2002; revised January 28 and July 18, 2003, and January 9, 2004; accepted January 12, 2004.

 Disclosure: M. Takeuchi, None; H. Keino, None; T. Kezuka, None; M. Usui, None; O. Taguchi, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Flow cytometric analysis of spleen cells from 3-month-old naive, Tx-3, and Tr-depleted mice. The spleen cells were stained with FITC-conjugated anti-CD4 (GK1.5) and biotin-conjugated anti-IL-2R (PC61) antibodies. Numbers indicate percentages within quadrants. The population of CD25⁺CD4⁺ T cells in the spleen of Tr-depleted mice is decreased in Tx-3 mice compared with naive mice and is further decreased markedly in Tr-depleted mice.

Table 1.

 

Incidence of Autoimmune Diseases in Tx-3 and Tr-Depleted Mice

Mice (n)*	Gastritis	Dacryadenitis	Thyroiditis	Uveoretinitis	Prostatitis
Tx-3 (25)	2 (8.0)	21 (84.0)	0	0	18 (72.0)
Tr-depleted (30)	20 (66.7)	30 (100)	4 (13.3)	6 (20.0)	30 (100)

 Data are the number of mice with autoimmune diseases (%). Autoimmune disease was defined as the presence of at least one inflammatory lesion in the corresponding organ together with related autoantibodies.

*  All mice were killed at 6 months of age.

Figure 2.

 

Uveoretinitis and circulating autoantibodies to the retinal photoreceptor cell layer induced in Tr-depleted mice. (A–C) Histopathology of the mouse retina (hematoxylin-eosin staining. The retina were obtained from (A) a 6-month-old Tx-3 mouse, (B) a 6-month-old Tr-depleted mouse with autoantibodies to the retina, and (C) a B6A nu/nu recipient mouse at 4 weeks after adoptive transfer of spleen cells (2 × 10⁷) from Tr-depleted mice with autoantibodies to the retina. Note massive inflammatory cells in the retina and complete destruction of the outer nuclear layer and photoreceptor cell layer in (B) and (C). (D–E) Indirect immunofluorescence staining of normal mouse retina. Frozen section of the retina from a normal adult B6A mouse was stained with (D) the secondary antibody alone, (E) serum (1:160 dilution) from a naive mouse, and (F) serum (1:160 dilution) from a Tr-depleted mouse with uveoretinitis. The layer of rods and cones of retina is stained in (F). G, ganglion cell layer; I, inner nuclear cell layer; O, outer nuclear layer; P, photoreceptor cell layer; *, retinal pigmented epithelium and choroid; V, vitreous chamber. Magnification, ×200.

Figure 3.

 

Correlation between individual uveoretinitis scores and titers of autoantibodies to the photoreceptor cell layer in Tx-3 and Tr-depleted mice. Tx-3 and Tr-depleted mice were killed at the age of 6 months, and serum samples and eyes were obtained. Antibody titers to the photoreceptor cell layer were measured by an indirect immunofluorescence method, and the frequency and severity of uveoretinitis were determined histopathologically. Circle, normal eye; square, partial retinal destruction with lymphocyte infiltration; triangle, complete loss of the photoreceptor layer; open symbols: individual mouse; closed symbols: mean of five mice.

Table 2.

 

Uveoretinitis Adoptively Transferred by T Cells Obtained from Tr-Depleted Mice with Autoantibodies to the Retina

Splenic T Cell Treatment	Naive	Tx-3	Tr-Depleted
PBS	0/5	0/5	3/4
Anti-CD4	ND	ND	0/3
Anti-CD8	ND	ND	4/4

 Splenic T cells were obtained from naive, Tx-3, or Tr-depleted mice with autoantibodies to the retinal photoreceptor cell layer and cultured with complement together with PBS, anti-CD4 mAb, or anti-CD8 mAb. Viable spleen cells (2 × 10⁷) were then injected into syngeneic nude mice (B6A nu/nu). Four weeks after injection, the recipient mice were killed and examined histopathologically. Data are number of mice showing uveoretinitis to number of mice tested. ND, not done.

Figure 4.

 

Splenic T-cell proliferation responses to S-Ag and IRBP challenge. Spleens were obtained from naive (□), Tx-3 (○), and Tr-depleted mice (▪). Each data point represents the result obtained from the pooled spleen T-cell sample of each group of five mice. Splenic T cells were stimulated with indicated concentrations of BSA (A), thyroglobulin (B), S-Ag (C), or IRBP (D) for 96 hours. The results are expressed as the SI. Background counts ranged from 188 to 454 cpm. The experiment was repeated three times with similar patterns of results.

Figure 5.

 

Antigen-specific GM-CSF production by splenic T cells. As in the T-cell proliferation experiment in Figure 4 , pooled splenic T cells obtained from naive (A), Tx-3 (B), and Tr-depleted (C) mice were stimulated with indicated concentrations of BSA (□), thyroglobulin (○), S-Ag (▪), or IRBP (•). Supernatants were collected after 72 hours, and GM-CSF production was assayed by ELISA. Each data point represents the mean of duplicate cultures. The experiment was repeated twice with similar patterns of results.

Figure 6.

 

Antigen-specific IFN-γ and IL-10 production by splenic T cells. Spleens were obtained from naive (□), Tx-3 (○), and Tr-depleted (▪) mice. Each point represents the result obtained from the pooled spleen T-cell sample of each group of five mice. Splenic T cells were stimulated with indicated concentrations of S-Ag (A, C), or IRBP (B, D). Supernatants were collected after 72 hours and IFN-γ (A, B) and IL-10 (C, D) contents were assayed by ELISA. Each data point represents the mean of duplicate cultures. The experiment was repeated twice with similar patterns of results.

Figure 7.

 

IgG2a levels in sera of naive, Tx-3, and Tr-depleted mice analyzed by ELISA. Serum samples were obtained from naive, Tx-3, and Tr-depleted mice at the time of death at 6 months of age. Serum levels of IgG2a were determined by ELISA. Values are mean ± SD of IgG2a concentrations in the serum of five mice per group. Significant differences between means of each group: *P < 0.05, **P < 0.005.