Extrathymically derived regulatory T cells (iTregs) protect against autoimmunity to tissue-specific Ags. However, whether Ag-specific iTreg generation and function is limited to secondary lymphoid tissue or whether it can occur within the tissue-specific local environment of the cognate Ag remains unresolved. Mice expressing β-galactosidase (βgal) on a retina-specific promoter (βgal mice) in conjunction with mice expressing GFP and diphtheria toxin (DTx) receptor (DTR) under control of the Foxp3 promoter, and βgal-specific TCR transgenic (BG2) mice were used to examine this question. Local depletion (ocular DTx), but not systemic depletion (i.p. DTx), of βgal-specific iTregs enhanced experimental autoimmune uveoretinitis induced by activated βgal-specific effector T cells. Injections of small amounts of βgal into the anterior chamber of the eye produced similar numbers of βgal-specific iTregs in the retina whether the mouse was depleted of pre-existing, circulating Tregs. Taken together, these results suggest that protection from tissue-specific autoimmunity depends on the function of local Ag-specific iTregs and that the retina is capable of local, “on-demand” iTreg generation that is independent of circulating Tregs.

The control of effector T cell (Teff) generation and responses by other T cells capable of immune regulation is vital to immune homeostasis. Although the phenotypes of T cells capable of regulating the activity of other T cells are diverse (1, 2), the CD4+CD25+Foxp3+ T cell is considered the dominant, prototypical regulatory T cell (Treg). A majority of these Tregs (natural Tregs, nTregs) develop in the thymus, partially in response to aire promoted expression of self-Ags by medullary thymic epithelial cells (37), and are crucial for protection from autoimmunity.

Foxp3+ Tregs can also be generated extrathymically from peripheral naive CD4+ T cells (induced Tregs, iTregs) (812). By analysis of TCR repertoire it was estimated that 4–7% of circulating Tregs were generated peripherally (13). Although use of the transcription factor Helios as a marker of nTregs is unsettled (14, 15), a recent study analyzing Helios expression suggested that iTregs may comprise as much as 30% of the circulating Tregs (14). Other studies using mice deficient in the Foxp3 enhancer CNS1, which is essential for iTreg but not nTreg generation, demonstrated that iTregs could constitute up to 50% of the Tregs in certain secondary lymphoid tissues (16, 17). Once formed, there is phenotypically little to distinguish nTregs and iTregs (14, 18); however, there are studies suggesting that nTregs and iTregs are functionally different in terms of TCR repertoire (Ag specificity), conditions necessary for induction and regulatory action, stability, regulatory capacity, site of action, and ability to function in quiescent or inflamed tissue (13, 1923). In addition to acting in consort with nTregs to limit autoimmune inflammation, iTregs are important in controlling immune responses to microorganisms, allergens, and Ags encountered through the gut (18, 24). Several studies suggest that iTreg induction from peripheral naive T cells is especially efficient in the gut and that iTregs constitute a majority of the Tregs found in GALT (2527). This rapid induction of Tregs by the gut immune system provides an acutely sensitive way to regulate responses to the wide variety of new, non–self-Ags that it constantly encounters.

Although central tolerance (negative selection and nTreg generation) is the primary mechanism for establishing tolerance to self-Ags, it is not complete (28). Thus, the ability to generate iTregs that regulate adaptive immune responses to self-Ags could be a crucial mechanism in immune homeostasis. Initial studies did not examine iTreg induction to bona fide tissue-specific self-Ags but rather used Ag-transgenic (Tg) mice or transplantation models that mimicked conditions associated with peripheral self-Ag expression—prolonged exposure of T cells to low doses of Ag without inflammation—to demonstrate efficient iTreg induction (12, 2931). However, recent studies have shown that iTregs induced to tissue-specific self-Ag can regulate Teffs (32, 33). Similarly, we have used Tg mice expressing Escherichia coli β-galactosidase (βgal, arrβgal mice) in retina photoreceptor cells in conjunction with TCR-Tg mice specific for βgal to demonstrate that a consistent outcome of retinal βgal expression is a downregulation of systemic immune responses to βgal that is, in part, due to the induction of Tregs from naive, peripheral βgal-specific CD4+ cells (3437).

It is well established that iTregs can be generated in peripheral lymphoid tissue as a result of interaction between Ag-bearing dendritic cells (DC) and T cells, where they exert a regulatory effect by limiting the priming of Teffs (38). It is also known that Ag-specific iTregs—albeit artificially generated—can migrate to Ag-expressing tissue and modulate established immune responses associated with autoimmune gastritis (39), collagen-induced arthritis (21), and type 1 diabetes (23). Among organs harboring tissue-specific self-Ags, the retina is unique in that it is part of the CNS yet lacks the meninges and lymphatic drainage found in brain and spinal cord (40, 41). This raises questions as to how and where retinal Ag-specific iTregs are generated and whether they contribute effective regulatory activity within a quiescent retina. Recently, we established that the retina has a population of DC (42). When isolated from quiescent retina and mixed with Ag-specific naive T cells plus cognate Ag, retinal DC induced the production of Foxp3+ T cells (43). This led us to investigate whether iTregs could be produced locally within the retina and whether they act locally to maintain retinal immune privilege.

To determine the origin and function of retinal iTregs, we used βgal-specific, MHC class II–restricted (CD4+ T cell), TCR-Tg mice in conjunction with arrβgal mice, and mice expressing diphtheria toxin receptor (DTR) and/or GFP under control of the Foxp3 promoter. We found that the presence of the local retinal iTregs, but not systemic Tregs, was necessary to protect mice from T cell–mediated experimental autoimmune uveoretinitis (EAU). Furthermore, in mice systemically depleted of Tregs, local Ag challenge induced Ag-specific iTreg production within the retina. Taken together, the data suggest that iTregs specific to retinal Ags are made within the retina and can protect the retina from circulating autoreactive T cells.

The arrβgal mice (B10.A-arrβgal, MHC haplotype I-Ak) have been described in detail elsewhere (4446). Briefly, rod photoreceptor cell expression of βgal in arrβgal mice mimics that of endogenous arrestin, producing ∼150 ng βgal/retina and <0.5 ng βgal/pineal gland. No other sites of βgal expression, including the thymus, have been found. Arrβgal mice on the B6 background (B6-arrβgal mice, MHC haplotype I-Ab) mice were generated by crossing B10.A-arrβgal mice with normal C57BL/6J (B6) mice. F1 mice were crossed again to wild-type (wt) B6, and the F2 mice were screened for βgal expression and I-Ab. Mice positive for βgal and I-Ab were continually backcrossed to wt B6 mice for >10 generations. Presence and expression of the βgal gene were assayed by PCR or by o-nitrophenyl-β-galactopyranoside (Sigma-Aldrich) analysis of retinal homogenates (47). MHC haplotype analysis was done by flow cytometry of PBMCs. βgalTCR mice (B10.A) and BG2 mice (B6 background) mice carry MHC class II–restricted (CD4+) T cells specific for βgal protein, including epitopes YVVDEANIETHGMV (βgalTCR) or SVTLPAASHAI (BG2), and have been described elsewhere (36, 48). FG and FDG mice (both B6), which express GFP only or DTR and GFP under control of the Foxp3 promoter, respectively, have been described previously (49, 50). Breeding stock was provided by Dr. S. S. Way (University of Minnesota, Minneapolis, MN). All mice were handled in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and the University of Minnesota animal use and care guidelines. Mice were housed under specific pathogen-free conditions on lactose-free chow.

Pooled spleen and lymph node (LN) cell suspensions from the indicated mice were prepared by tissue homogenization, followed by filtration through a 70-μM cell strainer. Lymphocytes were also prepared from whole blood. RBCs were lysed using 0.17 M NH4Cl, and the remaining cells were washed twice in PBS and resuspended in X-Vivo 15 medium (Lonza) supplemented with sodium pyruvate (100 μg/ml), l-glutamine (784 μg/ml), penicillin (100 U/ml), streptomycin (0.1 mg/ml), 2-ME (50 μM), 1× MEM nonessential amino acids, and 2% FCS (Sigma-Aldrich). All cultures were maintained at 37°C in an atmosphere of 6% CO2. Thy1.2+ (CD90.2+) or CD4+ negative selection of BG2 lymphocytes was done by magnetic bead separation (Miltenyi Biotec) per manufacturer’s protocol.

Spleen, LNs, or PBMCs were prepared as described above, except that the final suspension was made in FACS buffer (PBS with 2% FCS and 0.02% sodium azide). A total of 0.25–2.0 μl/106 cells of the appropriate fluorescent-labeled Abs (BD Biosciences or eBioscience) were added to the cell suspension and incubated on ice for 30 min. The cells were washed, resuspended in FACS buffer, and analyzed on FACSCalibur or FACSCanto flow cytometers using CellQuest (BD Biosciences) or FlowJo (Tree Star) software. For analysis of retinal cells by flow cytometry, mice were euthanized and perfused, and the retinas were removed as described previously (42). Control experiments showed that the perfusion procedure effectively removed passenger T cells from the retinal vasculature so that their contribution was not significant (data not shown). The retinas were dissociated using a solution of 0.2 μg/ml Liberase/TM (Roche) and 0.05% DNase in PBS, stained with the indicated Abs, and analyzed as described previously (42). Analysis of all cells collected from a single retina comprised a single sample.

Cultures containing 5 × 105 unfractionated spleen/LN cells from the indicated mice were aliquoted in triplicate into 96-well plates in final volume of 200 μl with or without βgal protein (10 μg/ml). Supernatants were harvested 48 h poststimulation and assayed for cytokines by cytometric bead array (BD Biosciences) per manufacturer’s protocol.

For the transfer of activated T cells, purified BG2 lymphocytes were stimulated with cognate peptide (0.5 μM) and irradiated (3000 rad) B6 splenocytes at a 1:10 ratio. IL-2 (10 U/ml) was added 48 h poststimulation, and the cells were cultured an additional 6 h. The cultures were washed and resuspended in PBS to a concentration of 2 × 107 cell/ml. Mice were inoculated i.p. with 5–20 × 106 cells. βgal-specific T cells were also generated by s.c. immunization of mice with a single 50-μl dose containing 200 μg βgal emulsified in CFA containing 5 μg/ml Mycobacterium tuberculosis (H37Ra; Sigma-Aldrich), followed by 0.5 μg pertussis toxin (Sigma-Aldrich) per mouse given in 100 μl saline i.p. At the indicated times posttransfer or postimmunization, the eyes were harvested, fixed in 10% buffered formalin, paraffin embedded, sectioned (5 μM), and stained with H&E. The slides were examined in a masked fashion, and the induced EAU was scored from 0 (no disease) to 5 (complete loss of photoreceptor cells plus damage to the inner layers of the retina) based on histopathological changes in the retina (51).

Analysis of the delayed-type hypersensitive (DTH) response (ear swelling assay) was done by injection of βgal (50 μg in 10 μl) into the ear pinna as described previously (52).

Diphtheria toxin (DTx), βgal, and saline injections into the eye were done by transcorneal deposition into the anterior chamber (AC) as described previously (42). One-microliter doses containing saline or the indicated amount of DTx or βgal were given. DTx injections into the cheek were given s.c. using 25 ng DTx in 10 μl saline as indicated. Systemic depletion of Tregs was done by i.p. injections of DTx with dose and timing indicated.

Information in the online supplemental material includes flow cytometry gating strategy for analysis of retinal cells, sensitivity of DTR+ retinal cells to systemic DTx, and analysis of T cell activation (L-selectin levels) in retinal and circulating T cells following DTx treatment.

Previously, we described a TCR-Tg mouse specific for βgal (βgalTCR) on the B10.A background (35, 36). When backcrossed to mice that expressed βgal in the retina, the T cells from the double Tg mice (βgalTCR × B10.A-arrβgal) maintained a naive phenotype, did not induce spontaneous EAU, and did not mediate EAU when immunized with βgal. However, the double Tg mice had a reduced DTH response to βgal compared with βgalTCR mice. In total, our results suggested that the downregulation of immune responses associated with retinal βgal was due the generation of Ag-specific CD4+CD25+ Tregs from naive CD4+ precursors and was dependent on the presence of retinal βgal (3437). To promote analysis of the nature of retinal Ag-specific Tregs, the model was moved onto the B6 background. B6 mice expressing GFP (49) or DTR/GFP (50) from the Foxp3 promoter were acquired (designated FG and FDG mice, respectively) along with B6 TCR-Tg BG2 mice (48).

T cells from TCR-Tg BG2 mice express the Vα11 TCR on the majority of CD4+ T cells (Fig. 1A) and CD4+Vα11+ T cells from BG2 × B6-arrβgal double Tg mice (BG2-βgal) maintained a naive phenotype (Fig. 1B). In BG2-βgal mice, spontaneous EAU was not observed, and EAU following βgal immunization was rare (data not shown). Analysis of cell surface markers associated with Ag recognition from βgal-immunized BG2 mice showed increases typically associated with T cell activation (elevated CD44, CD69, and reduced CD45RB, L-selectin; Fig. 1B). Cytokine analysis showed that Ag-activated BG2 T cells produced primarily IL-2, IL-6, IL-10, TNF-α, and smaller amounts of IFN-γ and IL-17 in response to stimulation with βgal (Fig. 1C), consistent with the Th1 and Th17 phenotypes associated with T cells that can induce EAU (5356). Comparison of the DTH response to βgal in naive BG2-βgal versus naive BG2 mice showed significant reduction in the ear swelling of BG2-βgal mice (Fig. 1D). Transfer of activated BG2 T cells induced mild disease in ∼15% of B6-arrβgal recipients (Fig. 1E). The low incidence of EAU induced by BG2 T cells is consistent with the limited EAU susceptibility of B6 mice (57). The severity and histological features of BG2 T cell induced EAU was similar in all respects to the disease induced in arrßgal mice on the B10.A background by either adoptive transfer of activated T cells or βgal immunization (36, 46).

FIGURE 1.

Analysis of βgal-specific T cells and TCR-Tg mice. (A) Representative FACS analysis of CD3+ splenocytes for the clonotypic TCR (Vα11) from BG2 mice and control B6 mice. Percent of CD3+ cells positive for CD4 and Vα11 is given. (B) FACS analysis of BG2 T cells for surface molecules associated with activation. Lymphocytes were gated on CD4+Vα11+ cells and then analyzed for the indicated surface marker. Representative FACS plots shown with mean percentage (n = 4) of cells having a naive (N) or activated (A) phenotype indicated. For immunized BG2 mice, lymphocytes were obtained from draining inguinal LN 5 d postimmunization with βgal. (C) In vitro cytokine production. BG2 lymphocytes from naive mice were cultured with and without βgal stimulation. Results are given as mean ± SD, n = 3. (D) Inhibition of CD4+ T cell–induced DTH by retinal βgal. Ear swelling assays of BG2 versus BG2-βgal mice. Results are given as mean ± SD. (E) Representative photomicrographs of autoimmune pathology in the eyes of B6-arrβgal mice induced by the transfer of activated BG2 T cells. The peripheral edge (a) and optic nerve head (b) of control retinas and the same regions in diseased eyes (c, d) are shown. The frequency of disease and the average score of diseased eyes is indicated. Original magnification ×20. (F) Comparison of Tregs levels in BG2 and BG2-βgal mice. Splenocytes were gated on CD4+Vα11+ cells and analyzed for CD25 and Foxp3. Representative FACS plots shown with mean percentage of CD25+Foxp3+ cells indicated (n = 3). (G) Association of Foxp3+ T cells with the Vα11 phenotype in BG2 mice. Lymphocytes from blood were gated on CD3+CD4+ cells and then analyzed for Vα11 and GFP expression. Representative FACS plots shown with average percent of GFP+ cells within the Vα11 and Vα11+ T cells populations indicated (n = 20). Where indicated, p values were determined by t test.

FIGURE 1.

Analysis of βgal-specific T cells and TCR-Tg mice. (A) Representative FACS analysis of CD3+ splenocytes for the clonotypic TCR (Vα11) from BG2 mice and control B6 mice. Percent of CD3+ cells positive for CD4 and Vα11 is given. (B) FACS analysis of BG2 T cells for surface molecules associated with activation. Lymphocytes were gated on CD4+Vα11+ cells and then analyzed for the indicated surface marker. Representative FACS plots shown with mean percentage (n = 4) of cells having a naive (N) or activated (A) phenotype indicated. For immunized BG2 mice, lymphocytes were obtained from draining inguinal LN 5 d postimmunization with βgal. (C) In vitro cytokine production. BG2 lymphocytes from naive mice were cultured with and without βgal stimulation. Results are given as mean ± SD, n = 3. (D) Inhibition of CD4+ T cell–induced DTH by retinal βgal. Ear swelling assays of BG2 versus BG2-βgal mice. Results are given as mean ± SD. (E) Representative photomicrographs of autoimmune pathology in the eyes of B6-arrβgal mice induced by the transfer of activated BG2 T cells. The peripheral edge (a) and optic nerve head (b) of control retinas and the same regions in diseased eyes (c, d) are shown. The frequency of disease and the average score of diseased eyes is indicated. Original magnification ×20. (F) Comparison of Tregs levels in BG2 and BG2-βgal mice. Splenocytes were gated on CD4+Vα11+ cells and analyzed for CD25 and Foxp3. Representative FACS plots shown with mean percentage of CD25+Foxp3+ cells indicated (n = 3). (G) Association of Foxp3+ T cells with the Vα11 phenotype in BG2 mice. Lymphocytes from blood were gated on CD3+CD4+ cells and then analyzed for Vα11 and GFP expression. Representative FACS plots shown with average percent of GFP+ cells within the Vα11 and Vα11+ T cells populations indicated (n = 20). Where indicated, p values were determined by t test.

Close modal

The level of CD25+Foxp3+ T cells was also similar between naive BG2 and naive BG2-βgal mice (Fig. 1F). Although Vα11+ T cells predominant in mice carrying the BG2 transgene, a greater percentage of Vα11 T cells were GFP+ Tregs compared with the Vα11+ T cells (Fig. 1G), and Treg frequency in blood (shown) or peripheral lymphoid organs (data not shown) was not affected by retinal βgal expression (Fig. 1G, bottom panels).

We recently described a GFP+ DC population in the retinas of CD11c-DTR/GFP mice that could be specifically eliminated from the retinal parenchyma by AC injection of 1–5 ng DTx (42, 43). We also determined that DTx given systemically could efficiently access the retina and eliminate DTR/GFP+ cells (Supplemental Table I). Analysis of FDG mice showed that AC injections of DTx efficiently depleted GFP+ cells from the retinas (data not shown). Knowing that immune cells expressing DTR in the retina can be selectively and efficiently depleted by local DTx, we asked whether AC injection of DTx into FDG × B6-arrβgal double Tg mice (FDG-βgal) would alter the pathogenesis of EAU by elimination of DTR+ Tregs from the retina. Ag-activated, βgal-specific, DTx-sensitive T cells from BG2 × FDG double Tg mice were adoptively transferred into FDG-βgal mice by i.p. injections. AC injection of 25 ng DTx was done three times per week for 3 wk to locally deplete Foxp3+ Tregs, whether derived from the endogenous Tregs of the recipient, the donor T cell population, or newly generated in response to retinal βgal. Mice that did not receive DTx failed to develop EAU, whereas mice that received multiple, unilateral (right eye) AC injections of DTx developed severe EAU in their ipsilateral (right) eyes (nine of nine eyes, average score 2.8) but not in their contralateral (left) eyes (Fig. 2A). Analysis of blood during the course of the adoptive transfer experiments showed Treg frequency in mice receiving multiple AC DTx injections was similar to normal nontransferred FDG-βgal mice (Fig. 2B), suggesting that local DTx treatments did not affect circulating Treg levels. Mice that received multiple systemic doses of DTx failed to develop EAU in either eye (Fig. 2A), despite being substantially depleted of circulating Tregs (Fig. 2B). The observation of no EAU in the contralateral eyes of transferred mice that received AC DTx, in mice receiving systemic DTx, and no DTx control mice was consistent with the low frequency of EAU induced by adoptive transfer of BG2 T cells into B6-arrβgal mice (Fig. 1E) given the limited number of mice analyzed in each group. Furthermore, both the incidence (100 versus 15%) and severity (2.8 versus 1.3 in diseased eyes) of EAU was significantly greater in the transferred FDG-βgal mice receiving AC DTx (ipsilateral eyes) compared with BG2 T cell transferred B6-arrβgal mice. Control wt B6 mice (data not shown) and FDG-arrβgal mice that received only AC injections of DTx showed no retinal pathology (Fig. 2A), demonstrating a lack of nonspecific toxicity of DTx toward retinal cells.

FIGURE 2.

Only local depletion of Tregs induces EAU. (A) Induction of EAU in FDG-βgal mice following transfer of activated BG2 × FDG T cells and treatment with DTx. Transferred and nontransferred control mice were given three times 25 ng DTx per week into the right AC starting the day of T cell transfer or 250 ng DTx i.p. on days 0, 2, 5, and 12 after transfer. Eyes were harvested 22 d posttransfer. (B) Analysis of circulating Treg levels in FDG-βgal mice transferred with activated BG2 × FDG T cells and given DTx into the right AC or systemic DTx. AC injected mice were analyzed at 1 wk (three times 25 ng DTx) or 2 wk (six times 25 ng DTx) posttransfer and were compared with normal FDG-βgal mice. Mice receiving systemic DTx were given 250 ng on days 0, 2, and 5 posttransfer and were analyzed day 6 posttransfer. Results are given as mean ± SD.

FIGURE 2.

Only local depletion of Tregs induces EAU. (A) Induction of EAU in FDG-βgal mice following transfer of activated BG2 × FDG T cells and treatment with DTx. Transferred and nontransferred control mice were given three times 25 ng DTx per week into the right AC starting the day of T cell transfer or 250 ng DTx i.p. on days 0, 2, 5, and 12 after transfer. Eyes were harvested 22 d posttransfer. (B) Analysis of circulating Treg levels in FDG-βgal mice transferred with activated BG2 × FDG T cells and given DTx into the right AC or systemic DTx. AC injected mice were analyzed at 1 wk (three times 25 ng DTx) or 2 wk (six times 25 ng DTx) posttransfer and were compared with normal FDG-βgal mice. Mice receiving systemic DTx were given 250 ng on days 0, 2, and 5 posttransfer and were analyzed day 6 posttransfer. Results are given as mean ± SD.

Close modal

To ascertain whether local Tregs could maintain their protective effect under conditions of ongoing autoreactive Teff generation, FDG-βgal mice were immunized with βgal in CFA and then given regular unilateral injections of DTx into the AC (25 ng, three times per week, right eye). The mice were analyzed at 7 d, 10 d, and 2 wk postimmunization, to ascertain whether local depletion of Tregs accelerated the development of EAU, as well as at 3 wk postimmunization, a time typically associated with the development of EAU from immunization in susceptible strains. No EAU was observed in mice analyzed 7 d postimmunization (data not shown), and minimal disease was observed in mice analyzed 10 d postimmunization (one of eight right eyes, score 0.5; Fig. 3A). However, at 2 and 3 wk postimmunization, FDG-βgal mice that received unilateral AC injections of DTx developed a similarly significant incidence and severity of EAU in the ipsilateral eyes (Fig. 3A) as compared with their contralateral eyes. Although mice analyzed at 3 wk had a slightly higher numerical disease score, likely reflective of the extra time for EAU to progress, it was not statistically significant (1.8 ± 1.9 versus 2.7 ± 2.1, p = 0.21; Fig. 3A). To minimize effects associated with repeated AC injections and locally high DTx concentrations and to determine whether a minimal depletion protocol could enhance EAU, βgal immunized FDG-βgal mice were given right AC injections of 5 ng DTx twice per week and analyzed for EAU at 3 wk. The incidence and severity of EAU (six of seven eyes, average score of all eyes 1.9 ± 1.5; Fig. 3A) was also significantly elevated compared with contralateral eyes. Combined analysis of the untreated contralateral eyes (left eyes from mice analyzed at 2 and 3 wk) showed only minimal incidence and severity of EAU (3 of 36 mice; Fig. 3A), typical of the incidence and severity of EAU observed in the minimally susceptible B6 strain. This modest background level of EAU was also observed in βgal-immunized FDG-βgal mice that did not receive DTx (2 of 28 eyes; Fig. 3B, right panel) and in immunized B6-βgal mice that received unilateral AC injections of DTx (two of seven ipsilateral eyes, zero of seven contralateral eyes; Fig. 3B, left panel), suggesting that DTx itself does not enhance EAU nonspecifically.

FIGURE 3.

Local Treg depletion and induction of EAU in FDG-βgal mice immunized with βgal. (A) Immunized FDG-βgal mice received right AC injections of DTx starting the day of immunization. DTx dose, frequency, and time to EAU analysis are indicated. Mean EAU score ± SD for all right (ipsilateral) eyes is indicated. For mice analyzed at 2 and 3 wk; p > 0.05 comparing groups of ipsilateral eyes and p < 0.001 comparing ipsilateral to contralateral eyes (Mann–Whitney U test). (B) Induction of EAU in immunized B6-βgal mice given DTx into the right AC and in immunized FDG-βgal mice not receiving DTx. DTx dose, frequency, and time to EAU analysis are indicated. All p values determined by Mann–Whitney U test. (C) Analysis of immunized FDG-βgal mice receiving right AC saline (1 μl, three times per week), right cheek DTx (25 ng, three times per week), systemic DTx (250 ng i.p. on days 0, 2, 5, and 12 d postimmunization for 2-wk mice and on days 0, 3, 6, 10, and 14 for 3-wk mice), and systemic DTx for 2 wk plus right AC saline injections. All eyes were analyzed for EAU at the indicated time.

FIGURE 3.

Local Treg depletion and induction of EAU in FDG-βgal mice immunized with βgal. (A) Immunized FDG-βgal mice received right AC injections of DTx starting the day of immunization. DTx dose, frequency, and time to EAU analysis are indicated. Mean EAU score ± SD for all right (ipsilateral) eyes is indicated. For mice analyzed at 2 and 3 wk; p > 0.05 comparing groups of ipsilateral eyes and p < 0.001 comparing ipsilateral to contralateral eyes (Mann–Whitney U test). (B) Induction of EAU in immunized B6-βgal mice given DTx into the right AC and in immunized FDG-βgal mice not receiving DTx. DTx dose, frequency, and time to EAU analysis are indicated. All p values determined by Mann–Whitney U test. (C) Analysis of immunized FDG-βgal mice receiving right AC saline (1 μl, three times per week), right cheek DTx (25 ng, three times per week), systemic DTx (250 ng i.p. on days 0, 2, 5, and 12 d postimmunization for 2-wk mice and on days 0, 3, 6, 10, and 14 for 3-wk mice), and systemic DTx for 2 wk plus right AC saline injections. All eyes were analyzed for EAU at the indicated time.

Close modal

βgal-immunized FDG-βgal mice receiving nonocular DTx and/or ocular saline were analyzed for EAU to further assess the specificity of EAU induction by DTx-induced local Treg depletion. Immunized FDG-βgal mice that received unilateral AC saline injections (incidence, 2 of 13 for ipsilateral eyes and 1 of 13 for contralateral eyes) failed to develop significant EAU (Fig. 3C). Because Tregs in LNs have been shown to inhibit T cell responses, we asked whether specific depletion of Tregs from LNs draining the eye would enhance EAU in βgal-immunized FDG-βgal mice. Serial injections of DTx into the right cheek of βgal-immunized FDG-βgal mice resulted in no EAU in either the ipsilateral or contralateral eyes (Fig. 3C). As with adoptively transferred mice, βgal-immunized FDG-βgal mice that received regular systemic DTx (250 ng) did not develop EAU when analyzed 2 wk (0 of 8 eyes; Fig. 3C) or 3 wk postimmunization (0 of 10 eyes; Fig. 3C), despite showing symptoms of systemic autoimmunity. Immunized FDG-βgal mice that received systemically a single large dose of DTx (1 μg at time of immunization) or small daily doses of DTx (50 ng/day for 3 wk) did not develop EAU (data not shown). Systemically depleted, βgal-immunized FDG-βgal mice that were also given unilateral AC saline injections did not develop EAU (Fig. 3C), even though systemic Treg depletion leads to a polyclonal expansion and activation of T cells (50) that increases both the number of circulating Ag-specific Teffs and their numbers within the retina (Supplemental Table II). This suggested that just an increase in Ag-specific Teffs in the retina is not sufficient to induce EAU and that the trauma associated with AC injection is not an instigating factor for EAU even when the periphery is full of primed Teffs unencumbered by Tregs. The ability to develop severe EAU specifically in eyes given a strong and sustained local Treg depletion but not in mice systemically depleted of Tregs, despite increased retinal Teffs, supports the idea of a very strong and active local component to EAU-protective Tregs. Our data suggest that βgal-specific, EAU-protective Tregs are generated within the retina and that they act locally to exert their regulatory effects in the retina.

To confirm that the induction of EAU following AC injection of DTx was due to the loss of locally acting, retinal Ag-specific Tregs, we determined the effects of AC DTx injections on Treg frequency in proximal and distal LN and on the ability of Tregs to modulate the DTH response to retinal Ags in a remote site (ear pinna). The frequency of Tregs was ∼13–15% of the CD3+CD4+ T cells in secondary lymphoid tissue of normal FDG-βgal mice (Fig. 4A, bottom). This was relatively unchanged in spleen, submandibular LN, and inguinal LN of FDG mice that were immunized with βgal and then given cheek or AC injections of DTx (Fig. 4A, top and middle), suggesting that the local Treg depletion protocols had no discernible effect on overall Treg levels. Analysis of the DTH response to βgal in FDG-βgal mice immunized with βgal showed no difference between right and left ears in mice that received right AC DTx injections and that the response was not different from immunized mice that received right AC saline injections (Fig. 4B, top). In a separate analysis, FDG mice that were βgal immunized and systemically depleted of Tregs had a much greater DTH response than mice that were immunized and not depleted of Tregs, showing that the DTH response was modulated by systemic changes in Tregs but not by local AC injection of DTx (Fig. 4B, bottom). Taken together, the data indicate that the AC DTx injections had little effect on the overall and local LN levels of Tregs or their ability to modulate DTH responses, suggesting that the effect of DTx injected into the AC is concentrated to the eye and that induction of EAU in response to local DTx treatment was not due to significant effects on Tregs outside of the retina.

FIGURE 4.

Local Treg depletion does not alter the level or function of peripheral Tregs. (A) Analysis of Treg levels in proximal and distal secondary lymphoid tissue following βgal immunization and DTx treatment of FDG-βgal mice. Dose and timing of DTx treatments were as described (Fig. 3). Mice were analyzed 2 wk postimmunization. (B) Analysis of DTH response following DTx treatment. βgal-immunized FDG-βgal mice were given saline or 25 ng DTx into the right AC (three times per week, 2 wk), challenged with βgal in both ears, and analyzed 24 h later (top). βgal-immunized FDG mice were given 250 ng DTx i.p. on days 0, 2, and 5 postimmunization, then ear tested with βgal on day 6 postimmunization, and analyzed 24 h later (bottom). Results are given as mean ± SD.

FIGURE 4.

Local Treg depletion does not alter the level or function of peripheral Tregs. (A) Analysis of Treg levels in proximal and distal secondary lymphoid tissue following βgal immunization and DTx treatment of FDG-βgal mice. Dose and timing of DTx treatments were as described (Fig. 3). Mice were analyzed 2 wk postimmunization. (B) Analysis of DTH response following DTx treatment. βgal-immunized FDG-βgal mice were given saline or 25 ng DTx into the right AC (three times per week, 2 wk), challenged with βgal in both ears, and analyzed 24 h later (top). βgal-immunized FDG mice were given 250 ng DTx i.p. on days 0, 2, and 5 postimmunization, then ear tested with βgal on day 6 postimmunization, and analyzed 24 h later (bottom). Results are given as mean ± SD.

Close modal

Because DTx depletion of circulating, systemic Tregs had little ability to promote EAU, and retinal Ag-specific Tregs appeared to act locally, we asked whether they were made locally. We previously reported that injection of soluble βgal into the AC of the eye led to the appearance of T cells in the retinal parenchyma of BG2 mice (43). Therefore, we used the DTx sensitivity of the DTR+ Tregs to eliminate existing circulating Tregs and then examined the number and phenotype of Tregs found in the retina in response to AC injections of Ag to determine whether the retina can locally generate Ag-specific Tregs.

The parenchyma of a normal retina contains ∼50 T cells of which only a few are Tregs (Fig. 5, gating strategy for retinal FACS analysis shown in Supplemental Fig. 1). Experiments to optimize the dose of βgal and analyze the timing of retinal T cell appearance showed that AC injection of 20 μg βgal was effective at increasing the number of Tregs (GFP+) and Teffs (GFP) within the retina of mice containing BG2 T cells (Fig. 5). The overall retinal T cell response, primarily GFP cells, peaked 2 d after βgal injection and was significantly higher than control (normal and saline injected) BG2 × FG mice and AC βgal/saline injected mice lacking βgal-specific T cells (FG mice) (Fig. 5A, 5B), which indicated Ag dependency to the recruitment of T cells to the retina. A majority of all T cells in βgal injected BG2 × FG mice were Vα11+, but Vα11 cells predominated in saline and βgal-injected FG mice, consistent with the low frequency of T cells having βgal specificity (Fig. 5A, 5B). Retinal Tregs (GFP+ T cells) generated in response to AC injection of βgal were significantly increased in BG2 × FG mice over controls (βgal-injected FG mice and saline-injected BG2 × FG mice) and were at a maximum 3 d after injection and were highly skewed toward the Vα11+ phenotype associated with βgal specificity (Fig. 5C). Retinal Treg levels returned to normal by day 5 post-βgal injection, and total T cell levels returned to normal by day 6. These experiments showed that T cells, including Tregs, could be readily detected in the retina in response to Ag, and that the appearance of Tregs in the retina is highly dependent on the interaction of Ag-specific T cells with cognate Ag.

FIGURE 5.

Analysis of the retinal T cell response to AC injections of Ag. FG and BG2 × FG mice were injected in the right AC with 1 μl saline or 20 μg βgal in final volume of 1 μl. The retinas from the right eyes were harvest at the indicated times and the retinal T cells were analyzed by FACS as shown in Supplemental Fig. 1. Rag−/− mice are shown as a control for the specificity of the T cell analysis (see Supplemental Fig. 1). (A) Total retinal T cells. (B) Retinal GFP T cells analyzed for Vα11. (C) Retinal Tregs (GFP+ T cells) analyzed for Vα11. Results given as mean ± SD for each group, n for all groups indicated, and p values determined by t test comparing GFP-Vα11+ and GFP+Vα11+ T cells numbers in saline versus βgal-injected eyes indicated.

FIGURE 5.

Analysis of the retinal T cell response to AC injections of Ag. FG and BG2 × FG mice were injected in the right AC with 1 μl saline or 20 μg βgal in final volume of 1 μl. The retinas from the right eyes were harvest at the indicated times and the retinal T cells were analyzed by FACS as shown in Supplemental Fig. 1. Rag−/− mice are shown as a control for the specificity of the T cell analysis (see Supplemental Fig. 1). (A) Total retinal T cells. (B) Retinal GFP T cells analyzed for Vα11. (C) Retinal Tregs (GFP+ T cells) analyzed for Vα11. Results given as mean ± SD for each group, n for all groups indicated, and p values determined by t test comparing GFP-Vα11+ and GFP+Vα11+ T cells numbers in saline versus βgal-injected eyes indicated.

Close modal

To determine whether retinas from BG2 × FDG mice given AC βgal generated significant numbers of Vα11+ and Vα11 Tregs, they were compared with retinas from control mice, mice given systemic DTx only (day 0 mice), and AC saline injected mice (Fig. 6A–C). If circulating Tregs were a source for the Tregs found in the retinal parenchyma, the large reduction in their number following systemic DTx treatment should affect their appearance in the retina. The finding on day 3 post-AC βgal injection that similar numbers of retinal Tregs developed in both normal and systemic Treg-depleted mice (Fig. 6C, top right panel) suggested that generation of retinal Tregs can occur without circulating Tregs and is independent from any pool of circulating Tregs. Systemic depletion of Tregs also resulted in a large increase in the number of Vα11GFP cells found in the retina (Fig. 6C, bottom panels). The increase was evident on day 0, highly enhanced by day 3, and likely due to the activation and expansion of polyclonal self-reactive Vα11 T cells no longer constrained by Tregs (50). However, the number of retinal GFPVα11+ T cells in the BG2 × FDG mice increased significantly only in response to Treg depletion plus deposition of cognate Ag within the eye (Fig. 6C, bottom panels).

FIGURE 6.

Local generation of Ag-specific Tregs. (A) BG2 × FDG mice were or were not treated with DTx (250 ng i.p., days −4 and −2) and then were given 20 μg βgal or saline (1 μl) into the AC (day 0). Control mice received no DTx and βgal, whereas day 0 mice received DTx only. Mice were analyzed for retinal and circulating T cells on day 0 (control and DTx only mice) and day 3 (AC βgal or saline). (B) Representative FACS plots of CD45+CD3+ gated retinal T cells (day 3) with number of retinal T cells having each phenotype indicated. (C) Composite analysis of number and Vα11 phenotype of retinal GFP+ and GFP T cells of control, day 0, and day 3 mice. (D) Analysis of circulating T cells for Vα11 and GFP (left panel) and comparison of the percent of Tregs that are Vα11+ in the retina vs. the blood on day 3 (right panel). For all analysis, the results are given as mean ± SD with p values determined by t test; *p < 0.05, **p < 0.01.

FIGURE 6.

Local generation of Ag-specific Tregs. (A) BG2 × FDG mice were or were not treated with DTx (250 ng i.p., days −4 and −2) and then were given 20 μg βgal or saline (1 μl) into the AC (day 0). Control mice received no DTx and βgal, whereas day 0 mice received DTx only. Mice were analyzed for retinal and circulating T cells on day 0 (control and DTx only mice) and day 3 (AC βgal or saline). (B) Representative FACS plots of CD45+CD3+ gated retinal T cells (day 3) with number of retinal T cells having each phenotype indicated. (C) Composite analysis of number and Vα11 phenotype of retinal GFP+ and GFP T cells of control, day 0, and day 3 mice. (D) Analysis of circulating T cells for Vα11 and GFP (left panel) and comparison of the percent of Tregs that are Vα11+ in the retina vs. the blood on day 3 (right panel). For all analysis, the results are given as mean ± SD with p values determined by t test; *p < 0.05, **p < 0.01.

Close modal

It is of interest to note that the majority of retinal Tregs made in response to AC βgal injections are Vα11+, even though both the peripheral and retinal T cell milieus are skewed toward Vα11 T cells. In normal BG2 × FDG mice, most of the Teffs are Vα11+, whereas the majority of Tregs are Vα11. However, upon Treg depletion, the portion of Vα11 Teffs increases reaching ∼50% at day 3 analysis (Fig. 6D, left panel). By day 0, DTx treatments have significantly reduced circulating Tregs, but their numbers have partially rebounded by day 3 (Fig. 6D, middle panel). However, in mice receiving AC βgal, most circulating Tregs are Vα11 (70%), whereas the majority of the retinal Tregs are Vα11+ (63%) (Fig. 6D, right panel). This skewing of retinal Treg toward Vα11+ occurs despite a large influx of Vα11 Teffs into the retina (Fig. 6C, bottom right panel).

As it has been demonstrated that a portion of material placed in the AC leaves the eye and is found in both blood and secondary lymphoid tissue (5860), it was an important control to ask whether the efflux of βgal from the eye into the circulation affected the frequency and phenotype of peripheral and retinal Tregs in normal and systemic Treg-depleted mice. BG2 × FDG mice were depleted of Tregs by systemic DTx injections, given 20 μg βgal i.v., and then analyzed for Tregs in the blood and the retina 3 d later (Fig. 7A). Although Vα11+ cells are >80% of all CD4+ T cells in mice carrying the BG2 TCR transgene, a majority of Tregs were Vα11 T cells (Fig. 7B, no DTx FACS panels, see also Fig. 1G). Overall, Treg depletion was routinely 90% or greater when assayed 2 d after the last DTx injection (day 0). Analysis of blood showed that 20 μg i.v. βgal did not specifically increase the number of Vα11+ Tregs in either normal or DTx-treated mice (Fig. 7B, bottom right panel). However, there was a slight increase in the number of circulating Vα11 Tregs from days 0 to 3, regardless if the mice received βgal (Fig 7B, bottom left panel). This increase is in response to increased numbers of circulating GFPVα11 T cells (Fig. 7B, lower left quadrant of FACS panels, DTx versus no DTx mice), which, as noted before, results from systemic depletion of Tregs leading to the polyclonal activation and expansion of T cells, especially those specific for endogenous self-Ag. There were very few Tregs to be found in the retinas of BG2 × FDG mice whether they were given DTx and/or i.v. βgal (Fig. 7C, bottom left panel). However, systemic Treg depletion led to a substantial increase in the number of Vα11GFP retinal T cells (Fig. 7C, bottom right panel). This again is the result of systemic depletion of Tregs leading to activation and expansion of self-reactive, polyclonal T cells, which have an enhanced ability to access the retina. Interestingly, there was not an increase in GFPVα11+ T cells in the retina after systemic Treg depletion (Fig. 7C, bottom right panel) likely reflecting their specificity for βgal, which is not present in retinas of BG2 × FDG mice. These results, together with the data from Fig. 5, showed while AC βgal-induced retinal Tregs, the same dose of βgal given i.v. did not induce an increase in either retinal or peripheral blood Tregs by 3 d after injection.

FIGURE 7.

Analysis of T cells response in blood and retina to the potential efflux of βgal from the AC. (A) BG2 × FDG mice were or were not given 250 ng DTx i.p. on days −4 and −2 and 20 μg βgal i.v. on day 0. (B) Analysis of circulating Tregs. T cells from blood were analyzed on days 0 and 3 by FACS for GFP and Vα11 expression. Representative FACS plots of CD3+CD4+ gated lymphocytes from day 0 analysis is shown with percentage of CD3+CD4+ lymphocytes being GFP+Vα11 or GFP+Vα11+ indicated (top panel). Composite analysis showing percent of CD4+ T cell being GFP+Vα11 or GFP+Vα11+ T cells at day 0 and day 3 (bottom panel). (C) Analysis of retinal T cells. Retinas were analyzed for T cells on day 3 post-βgal injection. Representative FACS plots with number and phenotype of retinal T cells indicated (top panel). Composite analysis of retinal T cells for number and phenotype (bottom panel). For all composite analysis, the results are given as mean ± SD with p values determined by t test; *p < 0.05, **p < 0.01.

FIGURE 7.

Analysis of T cells response in blood and retina to the potential efflux of βgal from the AC. (A) BG2 × FDG mice were or were not given 250 ng DTx i.p. on days −4 and −2 and 20 μg βgal i.v. on day 0. (B) Analysis of circulating Tregs. T cells from blood were analyzed on days 0 and 3 by FACS for GFP and Vα11 expression. Representative FACS plots of CD3+CD4+ gated lymphocytes from day 0 analysis is shown with percentage of CD3+CD4+ lymphocytes being GFP+Vα11 or GFP+Vα11+ indicated (top panel). Composite analysis showing percent of CD4+ T cell being GFP+Vα11 or GFP+Vα11+ T cells at day 0 and day 3 (bottom panel). (C) Analysis of retinal T cells. Retinas were analyzed for T cells on day 3 post-βgal injection. Representative FACS plots with number and phenotype of retinal T cells indicated (top panel). Composite analysis of retinal T cells for number and phenotype (bottom panel). For all composite analysis, the results are given as mean ± SD with p values determined by t test; *p < 0.05, **p < 0.01.

Close modal

Further evidence that the retina can efficiently and rapidly produce Tregs from circulating precursors came from analysis of Treg levels in mice given DTx after Treg induction. Retinal Treg numbers in BG2 × FDG mice given DTx daily up to the time of AC βgal injection were significantly reduced compared with mice not given DTx (Fig. 8, group 1 versus 2). However, even with DTx continued until the time of retinal harvest, there persisted a small number of Vα11+ Tregs but not Vα11 Tregs (Fig. 8, group 2 versus 3). The eventual loss of Vα11 Tregs from mice given continual DTx showed that systemic DTx does access the retina and eliminates GFP+ Tregs, but the presence of βgal provides a stimulus for the ongoing production of Ag-specific Tregs within the retina.

FIGURE 8.

Continual generation of Ag-specific Tregs in the retina. BG2 x FDG mice received no DTx or daily i.p. injections of 50 ng DTx through day 7 or 11 and 20 μg βgal into the AC on day 8. Retinas were analyzed by FACS for Tregs on day 11. For all analysis, results are given as mean ± SD with p values determined by t test.

FIGURE 8.

Continual generation of Ag-specific Tregs in the retina. BG2 x FDG mice received no DTx or daily i.p. injections of 50 ng DTx through day 7 or 11 and 20 μg βgal into the AC on day 8. Retinas were analyzed by FACS for Tregs on day 11. For all analysis, results are given as mean ± SD with p values determined by t test.

Close modal

Taken together, these results are consistent with the idea that the retina is capable of Ag-dependent, local production of Tregs. The data show that retinal Tregs are generated to cognate Ag from Ag-specific precursor T cells. This Treg generation is not dependent on, or reflective of, the pool of pre-existing Tregs, and it is not reflective of the overall phenotype of retinal and circulating precursor T cells.

Because the number of retinal Tregs is small even after AC injection of Ag, we sought evidence for their ability to regulate a pathogenic response. Because unilateral, AC injections of DTx in BG2 × FDG-βgal mice permitted EAU in the ipsilateral retina of βgal-immunized mice, a comparison of retinal T cells from susceptible ipsilateral retinas versus resistant contralateral retinas from the same mice could be informative. BG2 × FDG-βgal mice were immunized with βgal, given serial injections of DTx into the AC of the right eye, and analyzed for retinal T cells by FACS. As a control to ensure that EAU developed, some of the right eyes were examined histologically (six of nine positive for EAU; data not shown), whereas the remaining ipsilateral retinas (n = 10) were assayed by FACS and compared with the (n = 19) contralateral retinas. Although less sensitive than histology, funduscopic examination of the contralateral eyes just prior to harvest showed them to be free of retinal pathology, whereas 3 of 10 remaining ipsilateral retinas showed evidence of inflammation. FACS analysis of ipsilateral retinas showed a wide range in number of infiltrating lymphocytes reflecting the varied levels of EAU in these retinas. Overall, there was a significant increase in the number of T cells (both GFP+ Tregs and GFP effector T cells) compared with normal controls (Fig. 9). Although subjected to periodic Treg depletion, high numbers of Tregs were observed in several ipsilateral retinas 3 d after the last DTx treatment, likely in response to the large number of recruited and activated effector T cells infiltrating the retina during active inflammation. Analysis of contralateral retinas showed a smaller but still significant increase in GFP+ and GFP T cells. Although total T cell numbers varied greatly among ipsilateral, contralateral, and control retinas, the percentage of GFP+ T cells was similar (Fig. 9). The number of CD11b+ cells was also elevated in ipsilateral retinas compared with normal controls, consistent with inflammation (Fig. 9). Further analysis showed that most of the CD11b+ cells were polymorphonuclear granulocytes (data not shown). However, the number of CD11b+ cells in contralateral retinas was similar to normal controls, consistent with a lack of disease in these eyes. Because the circulation to the retina is the same for both eyes, immunization at a remote site would generate effector T cells having equal access to both retinas. However, only untreated retinas resisted EAU. This outcome is consistent with their uninterrupted local generation of Tregs as needed.

FIGURE 9.

T cell and myeloid cell analysis of ipsi- and contralateral retinas following EAU induction by immunization and AC DTx. BG2 × FDG-βgal mice were immunized with βgal (day 0) and given 25-ng doses of DTx into the right (ipsilateral) AC on days 0, 2, 5, 7, and 9. Ipsilateral and contralateral retinas from immunized mice were analyzed by FACS on day 12 along with control retinas from normal BG2 × FDG-βgal mice and Rag−/− mice. Normal Rag−/− mice were used as controls for setting flow cytometry gating. Results are given as mean ± SD with p values determined by t test; *p < 0.05. nd, None detected.

FIGURE 9.

T cell and myeloid cell analysis of ipsi- and contralateral retinas following EAU induction by immunization and AC DTx. BG2 × FDG-βgal mice were immunized with βgal (day 0) and given 25-ng doses of DTx into the right (ipsilateral) AC on days 0, 2, 5, 7, and 9. Ipsilateral and contralateral retinas from immunized mice were analyzed by FACS on day 12 along with control retinas from normal BG2 × FDG-βgal mice and Rag−/− mice. Normal Rag−/− mice were used as controls for setting flow cytometry gating. Results are given as mean ± SD with p values determined by t test; *p < 0.05. nd, None detected.

Close modal

Immune privilege of ocular tissue, including retina, was reported in the 1940s (61) and continues to be a paradigm for how critically sensitive tissues maintain protective immune responses while limiting tissue destructive immune responses. In several murine models used to study the interactions of the retina and the immune system, there is a robust resistance to retinal autoimmune disease that has been, in part, attributed to Tregs. In this study, we have proposed and presented evidence supporting the concept that retinal Ag-specific iTregs are generated locally and on demand within the retina and are essential in protecting the retina from autoimmunity.

The generation of iTregs in peripheral lymphoid tissue is well accepted (25, 62). However, the lack of lymphatic drainage from the retina as well as uncertainty about the destinations of the very small number of emigrating retinal DC led us to question where iTregs associated with retinal immune privilege were made and function. Our previous studies suggested a strong connection between the retina and Tregs. First, using B10.A-arrβgal mice, we found that retinal Ag expression made a substantial contribution to the generation of circulating iTregs that could inhibit CD4+ T cells in remote sites. Furthermore, enucleated mice failed to develop this ability, showing that the retina was the source of the Ag necessary for iTreg development (34). Second, we reported that retinal DC isolated from quiescent retina have the ability to promote Foxp3 expression in resting mature T cells if specific Ag is present (43). Third, we recently observed that the pathogenic capacity of βgal-specific CD8+ T cells to induce EAU in B6-arrβgal mice was modulated by local Foxp3+ Tregs (63). The dependence of iTreg development on TGF-β and retinoic acid (23, 25, 64, 65)—two factors found extensively in ocular tissue—further strengthens the retina/Treg connection. In a recent report, it was shown that aqueous humor, which contains TGF-β and retinoic acid, induced naive T cells to express Foxp3 and CD103 and gain suppressive activity in vitro. However, activated committed Teffs were resistant to gaining Treg function and maintained their effector function (66). This observation was extended in a study demonstrating the ocular production of Tregs from naive T cells after their injection into the eye (32). Taken together, these observations suggest the ocular environment is well suited to support the generation and function of iTregs.

Discrimination of iTregs from nTregs is currently a matter of great interest. Expression of the transcription factor Helios has been proposed to mark nTregs (14), but more recent studies (15, 67) have associated Helios expression with T cell activation, calling into question whether it can be used as a definitive marker for discriminating nTregs from iTregs. Neuropilin-1 (NRP-1) expression has also been associated with nTregs (68), but other reports suggest that its expression is also influenced by the local cytokine milieu, particularly TGF-β (69). Given that the eye is rich in TGF-β, it cannot be concluded that NRP-1+ Tregs from the retina are nTregs. Using RT-PCR, we found variable expression of Helios and NRP-1 between individual Vα11+GFP+ T cells isolated from retinas of BG2 × FDG mice given AC injections of βgal (data not shown), which supports the idea that neither Helios nor NRP-1 can reliably distinguish the source of Tregs. Two observations that would initially appear to diminish the role of Tregs in retinal immune privilege support the idea of local iTreg generation upon further consideration. First is the observation that efficient systemic depletion of Tregs in substantial groups of βgal-immunized and T cell adoptive transfer mice did not lead to retinal inflammation. If protective, retinal–Ag-specific iTregs were generated outside of the retina, their depletion via systemic DTx should have allowed EAU to develop. Although systemic DTx can deplete cells expressing DTR from the retina, the inability of systemic DTx to induce EAU suggests there is rapid, ongoing, local regeneration of iTregs on demand within the target tissue. We propose that although depleted of circulating Tregs, these mice are still replete with Ag-specific precursor T cells that are readily converted to iTregs in a consistent and ongoing basis within the favorable Ag-containing environment of the retina. In this model, nascent retinal iTregs individually contribute a transient protection, and collectively a continual protection, against autoreactive Teffs. Ultimately, retinal iTregs enter and persist in the circulation as evidenced by the downregulation of DTH responses to retinal-specific Ags and the ability to ablate this regulation with systemic DTx. Conversely, local Treg depletion allowed the T cell adoptive transfer and immunization models of EAU to produce a high incidence of severe EAU in only the ipsilateral retina. Although the amount of DTx used in the AC injections was small, repeated administration within the small volume of the eye likely resulted in a sustained local concentration sufficient to disrupt local Treg production. We estimated that retinal DTx concentrations achieved by AC injections were at least 100-fold greater than systemic injections. The amount of systemic DTx needed to achieve concentrations equivalent to AC DTx injections would likely result in rapid and fatal autoimmunity. Despite the presence of βgal-specific iTregs circulating within FDG-βgal mice, induction of EAU following local Treg depletion suggests that retinal Ag-specific Tregs need to be within the retina to protect against T cell–mediated autoimmunity.

The second observation relates to Treg numbers within retinas. Given the volume and extensive vasculature of the retina, it would appear that the small number of Tregs present in quiescent retina could not be a direct barrier to pathogenic T cells. However, our observation that contralateral retinas in the βgal immunization/adaptive transfer plus local Treg depletion experiments remained disease free, despite large numbers of Ag-specific pathogenic T cells circulating through both retinas, showed that a substantial residency of pre-existing retinal Tregs was not needed to protect against T cell–mediated autoimmunity. During the course of EAU in the ipsilateral retinas, there was an increase in the number of Tregs in the contralateral retinas. This increase in Tregs results from there being an uninterrupted, local, and on-demand production of Ag-specific iTregs within the contralateral retina in response to the threat associated with increased perusal of the retina by activated, retinal-Ag specific T cells. It is of interest to note that there were large numbers of Tregs in diseased ipsilateral retinas, suggesting that once disease is induced they may be relatively inefficient at protecting the retina.

Another factor that may reduce the need for large numbers of Tregs to maintain retinal homeostasis is the growing evidence for an important and dynamic relationship between Tregs and DC. Depletion of Tregs in FDG mice resulted in a several fold increase in the number of “activated” DC in secondary lymphoid organs (50), which led to the stimulation and expansion of Teffs, including those that were self-reactive and mediated autoimmunity (70). Conversely, “immature” DC are well known inducers of Treg activity (71) and that systemic loss of immature DC leads to the loss of tolerance (72). Furthermore, there is evidence that Tregs can mediate their regulatory effects through DC (73) and can suppress DC maturation (74) and that Treg/DC interaction can induce a switch in the DC phenotype from stimulatory to regulatory, thus favoring the production of more Tregs (75, 76). The dynamic Treg/DC relationship is evident in the retina. Within the EAU model, it was demonstrated that immature DC inhibited disease when given prior to adoptive transfer of T cells or were used during the in vitro activation of the T cells (77). Furthermore, we have shown that local conditions determined whether retinal DC favored Treg or Teff production (42, 43). A cooperative or synergistic interaction of Tregs and regulatory DC at the local level likely provides the regulatory functions necessary to maintain a healthy retina without the presence of large numbers of either cell type. Perturbation by local depletion of Tregs, as we have shown in this study, may be an example of the loss of this homeostatic mechanism, leading to EAU.

We have also demonstrated a preferential production of Ag-specific retinal iTregs in response to local Ag injection in mice depleted of pre-existing Tregs, suggesting again that iTregs can be produced locally and “on demand”. A concern in the interpretation of these results is that the AC βgal injections could induce AC-associated immune deviation (ACAID) (78). Ag deposited within the AC can be carried to spleen by eye-derived F4/80+ cells, where there is the formation of ACAID regulatory cells that modulate the immune response to subsequent challenge with the Ag (60). However, the timing for appearance of retinal Foxp3+ Tregs was not consistent with that of ACAID-induced regulatory cells. The course of ACAID is well known, requiring 7- to 8-d post-AC injection to develop functional, circulating ACAID regulatory cells. In contrast, we found that Foxp3+ Tregs were present in retina by 2 d post-βgal injection, peaked at 3 d postinjection, and were even declining by day 4. ACAID-induced regulatory cells would ultimately develop in response to AC βgal injections, but the timing of our analysis likely excludes them from having any effect on our assays. Regardless, it remained a formal possibility that ACAID-derived Tregs generated in the spleen could preferentially home to the retina before reaching levels in circulation require to modulate distal immune responses. However, experiments to address this question in which βgal was given direct access to the lymphoid organs via i.v. injection failed to alter retinal Treg numbers. Thus, within the parameters of our assays, retinal Tregs appear to be the result of local generation and not generation and migration from secondary lymphoid tissues.

We thank Drs. Mike Farrar and Sing Sing Way for helpful discussion and critique of this manuscript, Heidi Roehrich for histology, and Mark Pierson for technical assistance.

This work was supported by National Institutes of Health Research Grants R01-EY021996, R01-EY016376, and Core Facility Grant P30-EY011374. Additional support was provided by Research to Prevent Blindness and the Minnesota Lions Clubs.

The online version of this article contains supplemental material.

Abbreviations used in this article:

AC

anterior chamber

ACAID

anterior chamber–associated immune deviation

arrβgal

mice expressing β-galactosidase under control of the rod photoreceptor arrestin promoter

B6

C57BL/6J

DTH

delayed-type hypersensitivity

DTR

diphtheria toxin receptor

DTx

diphtheria toxin

EAU

experimental autoimmune uveoretinitis

FDG

mice expressing GFP and DTR under control of the Foxp3 promoter

FG

mice expressing GFP under control of the Foxp3 promoter

βgal

β-galactosidase

iTreg

extrathymically derived regulatory T cell

LN

lymph node

NRP-1

neuropilin-1

Teff

effector T cell

nTreg

natural regulatory T cell

Tg

transgenic

Treg

regulatory T cell

wt

wild-type.

1
Shevach
E. M.
2006
.
From vanilla to 28 flavors: multiple varieties of T regulatory cells.
Immunity
25
:
195
201
.
2
Jutel
M.
,
Akdis
C. A.
.
2008
.
T-cell regulatory mechanisms in specific immunotherapy.
Chem. Immunol. Allergy
94
:
158
177
.
3
Derbinski
J.
,
Schulte
A.
,
Kyewski
B.
,
Klein
L.
.
2001
.
Promiscuous gene expression in medullary thymic epithelial cells mirrors the peripheral self.
Nat. Immunol.
2
:
1032
1039
.
4
Anderson
M. S.
,
Venanzi
E. S.
,
Klein
L.
,
Chen
Z.
,
Berzins
S. P.
,
Turley
S. J.
,
von Boehmer
H.
,
Bronson
R.
,
Dierich
A.
,
Benoist
C.
,
Mathis
D.
.
2002
.
Projection of an immunological self shadow within the thymus by the aire protein.
Science
298
:
1395
1401
.
5
Liston
A.
,
Lesage
S.
,
Wilson
J.
,
Peltonen
L.
,
Goodnow
C. C.
.
2003
.
Aire regulates negative selection of organ-specific T cells.
Nat. Immunol.
4
:
350
354
.
6
Aschenbrenner
K.
,
D’Cruz
L. M.
,
Vollmann
E. H.
,
Hinterberger
M.
,
Emmerich
J.
,
Swee
L. K.
,
Rolink
A.
,
Klein
L.
.
2007
.
Selection of Foxp3+ regulatory T cells specific for self antigen expressed and presented by Aire+ medullary thymic epithelial cells.
Nat. Immunol.
8
:
351
358
.
7
Sakaguchi
S.
2011
.
Regulatory T cells: history and perspective.
Methods Mol. Biol.
707
:
3
17
.
8
Lohr
J.
,
Knoechel
B.
,
Abbas
A. K.
.
2006
.
Regulatory T cells in the periphery.
Immunol. Rev.
212
:
149
162
.
9
Apostolou
I.
,
Verginis
P.
,
Kretschmer
K.
,
Polansky
J.
,
Hühn
J.
,
von Boehmer
H.
.
2008
.
Peripherally induced Treg: mode, stability, and role in specific tolerance.
J. Clin. Immunol.
28
:
619
624
.
10
Thorstenson
K. M.
,
Khoruts
A.
.
2001
.
Generation of anergic and potentially immunoregulatory CD25+CD4 T cells in vivo after induction of peripheral tolerance with intravenous or oral antigen.
J. Immunol.
167
:
188
195
.
11
Curotto de Lafaille
M. A.
,
Lino
A. C.
,
Kutchukhidze
N.
,
Lafaille
J. J.
.
2004
.
CD25 T cells generate CD25+Foxp3+ regulatory T cells by peripheral expansion.
J. Immunol.
173
:
7259
7268
.
12
Kretschmer
K.
,
Apostolou
I.
,
Hawiger
D.
,
Khazaie
K.
,
Nussenzweig
M. C.
,
von Boehmer
H.
.
2005
.
Inducing and expanding regulatory T cell populations by foreign antigen.
Nat. Immunol.
6
:
1219
1227
.
13
Lathrop
S. K.
,
Santacruz
N. A.
,
Pham
D.
,
Luo
J.
,
Hsieh
C. S.
.
2008
.
Antigen-specific peripheral shaping of the natural regulatory T cell population.
J. Exp. Med.
205
:
3105
3117
.
14
Thornton
A. M.
,
Korty
P. E.
,
Tran
D. Q.
,
Wohlfert
E. A.
,
Murray
P. E.
,
Belkaid
Y.
,
Shevach
E. M.
.
2010
.
Expression of Helios, an Ikaros transcription factor family member, differentiates thymic-derived from peripherally induced Foxp3+ T regulatory cells.
J. Immunol.
184
:
3433
3441
.
15
Gottschalk
R. A.
,
Corse
E.
,
Allison
J. P.
.
2012
.
Expression of Helios in peripherally induced Foxp3+ regulatory T cells.
J. Immunol.
188
:
976
980
.
16
Josefowicz
S. Z.
,
Niec
R. E.
,
Kim
H. Y.
,
Treuting
P.
,
Chinen
T.
,
Zheng
Y.
,
Umetsu
D. T.
,
Rudensky
A. Y.
.
2012
.
Extrathymically generated regulatory T cells control mucosal TH2 inflammation.
Nature
482
:
395
399
.
17
Samstein
R. M.
,
Josefowicz
S. Z.
,
Arvey
A.
,
Treuting
P. M.
,
Rudensky
A. Y.
.
2012
.
Extrathymic generation of regulatory T cells in placental mammals mitigates maternal-fetal conflict.
Cell
150
:
29
38
.
18
Curotto de Lafaille
M. A.
,
Lafaille
J. J.
.
2009
.
Natural and adaptive foxp3+ regulatory T cells: more of the same or a division of labor?
Immunity
30
:
626
635
.
19
Horwitz
D. A.
,
Zheng
S. G.
,
Gray
J. D.
.
2008
.
Natural and TGF-β‑induced Foxp3+CD4+ CD25+ regulatory T cells are not mirror images of each other.
Trends Immunol.
29
:
429
435
.
20
Housley
W. J.
,
Adams
C. O.
,
Nichols
F. C.
,
Puddington
L.
,
Lingenheld
E. G.
,
Zhu
L.
,
Rajan
T. V.
,
Clark
R. B.
.
2011
.
Natural but not inducible regulatory T cells require TNF-α signaling for in vivo function.
J. Immunol.
186
:
6779
6787
.
21
Lan
Q.
,
Fan
H.
,
Quesniaux
V.
,
Ryffel
B.
,
Liu
Z.
,
Zheng
S. G.
.
2012
.
Induced Foxp3+ regulatory T cells: a potential new weapon to treat autoimmune and inflammatory diseases?
J Mol Cell Biol
4
:
22
28
.
22
Lee
H. M.
,
Bautista
J. L.
,
Hsieh
C. S.
.
2011
.
Thymic and peripheral differentiation of regulatory T cells.
Adv. Immunol.
112
:
25
71
.
23
Daniel
C.
,
von Boehmer
H.
.
2011
.
Extrathymic generation of regulatory T cells—chances and challenges for prevention of autoimmune disease.
Adv. Immunol.
112
:
177
213
.
24
Josefowicz
S. Z.
,
Rudensky
A.
.
2009
.
Control of regulatory T cell lineage commitment and maintenance.
Immunity
30
:
616
625
.
25
Coombes
J. L.
,
Siddiqui
K. R.
,
Arancibia-Cárcamo
C. V.
,
Hall
J.
,
Sun
C. M.
,
Belkaid
Y.
,
Powrie
F.
.
2007
.
A functionally specialized population of mucosal CD103+ DCs induces Foxp3+ regulatory T cells via a TGF-β and retinoic acid-dependent mechanism.
J. Exp. Med.
204
:
1757
1764
.
26
Atarashi
K.
,
Tanoue
T.
,
Shima
T.
,
Imaoka
A.
,
Kuwahara
T.
,
Momose
Y.
,
Cheng
G.
,
Yamasaki
S.
,
Saito
T.
,
Ohba
Y.
, et al
.
2011
.
Induction of colonic regulatory T cells by indigenous Clostridium species.
Science
331
:
337
341
.
27
Lathrop
S. K.
,
Bloom
S. M.
,
Rao
S. M.
,
Nutsch
K.
,
Lio
C. W.
,
Santacruz
N.
,
Peterson
D. A.
,
Stappenbeck
T. S.
,
Hsieh
C. S.
.
2011
.
Peripheral education of the immune system by colonic commensal microbiota.
Nature
478
:
250
254
.
28
Gallegos
A. M.
,
Bevan
M. J.
.
2006
.
Central tolerance: good but imperfect.
Immunol. Rev.
209
:
290
296
.
29
Verginis
P.
,
McLaughlin
K. A.
,
Wucherpfennig
K. W.
,
von Boehmer
H.
,
Apostolou
I.
.
2008
.
Induction of antigen-specific regulatory T cells in wild-type mice: visualization and targets of suppression.
Proc. Natl. Acad. Sci. USA
105
:
3479
3484
.
30
Knoechel
B.
,
Lohr
J.
,
Kahn
E.
,
Bluestone
J. A.
,
Abbas
A. K.
.
2005
.
Sequential development of interleukin 2-dependent effector and regulatory T cells in response to endogenous systemic antigen.
J. Exp. Med.
202
:
1375
1386
.
31
Apostolou
I.
,
von Boehmer
H.
.
2004
.
In vivo instruction of suppressor commitment in naive T cells.
J. Exp. Med.
199
:
1401
1408
.
32
Zhou
R.
,
Horai
R.
,
Silver
P. B.
,
Mattapallil
M. J.
,
Zárate-Bladés
C. R.
,
Chong
W. P.
,
Chen
J.
,
Rigden
R. C.
,
Villasmil
R.
,
Caspi
R. R.
.
2012
.
The living eye “disarms” uncommitted autoreactive T cells by converting them to Foxp3+ regulatory cells following local antigen recognition.
J. Immunol.
188
:
1742
1750
.
33
Thompson
L. J.
,
Valladao
A. C.
,
Ziegler
S. F.
.
2011
.
Cutting edge: de novo induction of functional Foxp3+ regulatory CD4 T cells in response to tissue-restricted self antigen.
J. Immunol.
186
:
4551
4555
.
34
Gregerson
D. S.
,
Heuss
N. D.
,
Lehmann
U.
,
McPherson
S. W.
.
2009
.
Peripheral induction of tolerance by retinal antigen expression.
J. Immunol.
183
:
814
822
.
35
Gregerson
D. S.
,
Heuss
N. D.
,
Lehmann
U.
,
McPherson
S. W.
.
2008
.
Evidence for extrathymic generation of regulatory T cells specific for a retinal antigen.
Ophthalmic Res.
40
:
154
159
.
36
McPherson
S. W.
,
Heuss
N. D.
,
Gregerson
D. S.
.
2009
.
Lymphopenia-induced proliferation is a potent activator for CD4+ T cell-mediated autoimmune disease in the retina.
J. Immunol.
182
:
969
979
.
37
McPherson
S. W.
,
Heuss
N. D.
,
Lehmann
U.
,
Gregerson
D. S.
.
2011
.
Generation of regulatory T cells to antigen expressed in the retina.
Curr. Immunol. Rev.
7
:
344
349
.
38
DiPaolo
R. J.
,
Brinster
C.
,
Davidson
T. S.
,
Andersson
J.
,
Glass
D.
,
Shevach
E. M.
.
2007
.
Autoantigen-specific TGFβ-induced Foxp3+ regulatory T cells prevent autoimmunity by inhibiting dendritic cells from activating autoreactive T cells.
J. Immunol.
179
:
4685
4693
.
39
Nguyen
T. L.
,
Sullivan
N. L.
,
Ebel
M.
,
Teague
R. M.
,
DiPaolo
R. J.
.
2011
.
Antigen-specific TGF-β‑induced regulatory T cells secrete chemokines, regulate T cell trafficking, and suppress ongoing autoimmunity.
J. Immunol.
187
:
1745
1753
.
40
Yamada
S.
,
DePasquale
M.
,
Patlak
C. S.
,
Cserr
H. F.
.
1991
.
Albumin outflow into deep cervical lymph from different regions of rabbit brain.
Am. J. Physiol.
261
:
H1197
H1204
.
41
Cserr
H. F.
,
Harling-Berg
C. J.
,
Knopf
P. M.
.
1992
.
Drainage of brain extracellular fluid into blood and deep cervical lymph and its immunological significance.
Brain Pathol.
2
:
269
276
.
42
Lehmann
U.
,
Heuss
N. D.
,
McPherson
S. W.
,
Roehrich
H.
,
Gregerson
D. S.
.
2010
.
Dendritic cells are early responders to retinal injury.
Neurobiol. Dis.
40
:
177
184
.
43
Heuss
N. D.
,
Lehmann
U.
,
Norbury
C. C.
,
McPherson
S. W.
,
Gregerson
D. S.
.
2012
.
Local activation of dendritic cells alters the pathogenesis of autoimmune disease in the retina.
J. Immunol.
188
:
1191
1200
.
44
Kikuchi
T.
,
Raju
K.
,
Breitman
M. L.
,
Shinohara
T.
.
1993
.
The proximal promoter of the mouse arrestin gene directs gene expression in photoreceptor cells and contains an evolutionarily conserved retinal factor-binding site.
Mol. Cell. Biol.
13
:
4400
4408
.
45
Kimura
A.
,
Singh
D.
,
Wawrousek
E. F.
,
Kikuchi
M.
,
Nakamura
M.
,
Shinohara
T.
.
2000
.
Both PCE-1/RX and OTX/CRX interactions are necessary for photoreceptor-specific gene expression.
J. Biol. Chem.
275
:
1152
1160
.
46
Gregerson
D. S.
,
Torseth
J. W.
,
McPherson
S. W.
,
Roberts
J. P.
,
Shinohara
T.
,
Zack
D. J.
.
1999
.
Retinal expression of a neo-self antigen, β-galactosidase, is not tolerogenic and creates a target for autoimmune uveoretinitis.
J. Immunol.
163
:
1073
1080
.
47
Zack
D. J.
,
Bennett
J.
,
Wang
Y.
,
Davenport
C.
,
Klaunberg
B.
,
Gearhart
J.
,
Nathans
J.
.
1991
.
Unusual topography of bovine rhodopsin promoter-lacZ fusion gene expression in transgenic mouse retinas.
Neuron
6
:
187
199
.
48
Tewalt
E. F.
,
Grant
J. M.
,
Granger
E. L.
,
Palmer
D. C.
,
Heuss
N. D.
,
Gregerson
D. S.
,
Restifo
N. P.
,
Norbury
C. C.
.
2009
.
Viral sequestration of antigen subverts cross presentation to CD8+ T cells.
PLoS Pathog.
5
:
e1000457
.
49
Fontenot
J. D.
,
Rasmussen
J. P.
,
Williams
L. M.
,
Dooley
J. L.
,
Farr
A. G.
,
Rudensky
A. Y.
.
2005
.
Regulatory T cell lineage specification by the forkhead transcription factor foxp3.
Immunity
22
:
329
341
.
50
Kim
J. M.
,
Rasmussen
J. P.
,
Rudensky
A. Y.
.
2007
.
Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice.
Nat. Immunol.
8
:
191
197
.
51
Gregerson
D. S.
,
Obritsch
W. F.
,
Donoso
L. A.
.
1993
.
Oral tolerance in experimental autoimmune uveoretinitis: distinct mechanisms of resistance are induced by low dose vs high dose feeding protocols.
J. Immunol.
151
:
5751
5761
.
52
Gregerson
D. S.
,
Dou
C.
.
2002
.
Spontaneous induction of immunoregulation by an endogenous retinal antigen.
Invest. Ophthalmol. Vis. Sci.
43
:
2984
2991
.
53
Caspi
R. R.
2002
.
Th1 and Th2 responses in pathogenesis and regulation of experimental autoimmune uveoretinitis.
Int. Rev. Immunol.
21
:
197
208
.
54
McPherson
S. W.
,
Roberts
J. P.
,
Gregerson
D. S.
.
2005
.
Peripheral expression of rod photoreceptor arrestin induces an epitope-specific, protective response against experimental autoimmune uveoretinitis.
Curr. Eye Res.
30
:
491
502
.
55
Tang
J.
,
Zhu
W.
,
Silver
P. B.
,
Su
S. B.
,
Chan
C. C.
,
Caspi
R. R.
.
2007
.
Autoimmune uveitis elicited with antigen-pulsed dendritic cells has a distinct clinical signature and is driven by unique effector mechanisms: initial encounter with autoantigen defines disease phenotype.
J. Immunol.
178
:
5578
5587
.
56
Peng
Y.
,
Han
G.
,
Shao
H.
,
Wang
Y.
,
Kaplan
H. J.
,
Sun
D.
.
2007
.
Characterization of IL-17+ interphotoreceptor retinoid-binding protein-specific T cells in experimental autoimmune uveitis.
Invest. Ophthalmol. Vis. Sci.
48
:
4153
4161
.
57
Egwuagu
C. E.
,
Charukamnoetkanok
P.
,
Gery
I.
.
1997
.
Thymic expression of autoantigens correlates with resistance to autoimmune disease.
J. Immunol.
159
:
3109
3112
.
58
Camelo
S.
,
Kezic
J.
,
Shanley
A.
,
Rigby
P.
,
McMenamin
P. G.
.
2006
.
Antigen from the anterior chamber of the eye travels in a soluble form to secondary lymphoid organs via lymphatic and vascular routes.
Invest. Ophthalmol. Vis. Sci.
47
:
1039
1046
.
59
Camelo
S.
,
Shanley
A.
,
Voon
A. S.
,
McMenamin
P. G.
.
2004
.
The distribution of antigen in lymphoid tissues following its injection into the anterior chamber of the rat eye.
J. Immunol.
172
:
5388
5395
.
60
Wilbanks
G. A.
,
Mammolenti
M.
,
Streilein
J. W.
.
1991
.
Studies on the induction of anterior chamber-associated immune deviation (ACAID). II. Eye-derived cells participate in generating blood-borne signals that induce ACAID.
J. Immunol.
146
:
3018
3024
.
61
Medawar
P. B.
1948
.
Immunity to homologous grafted skin; the fate of skin homografts transplanted to the brain, to subcutaneous tissue, and to the anterior chamber of the eye.
Br. J. Exp. Pathol.
29
:
58
69
.
62
Guilliams
M.
,
Crozat
K.
,
Henri
S.
,
Tamoutounour
S.
,
Grenot
P.
,
Devilard
E.
,
de Bovis
B.
,
Alexopoulou
L.
,
Dalod
M.
,
Malissen
B.
.
2010
.
Skin-draining lymph nodes contain dermis-derived CD103 dendritic cells that constitutively produce retinoic acid and induce Foxp3+ regulatory T cells.
Blood
115
:
1958
1968
.
63
McPherson
S. W.
,
Heuss
N. D.
,
Gregerson
D. S.
.
2012
.
Regulation of CD8+ T cell responses to retinal antigen by local FoxP3+ regulatory T cells.
Front. Immunol.
3
:
166
.
64
Mucida
D.
,
Pino-Lagos
K.
,
Kim
G.
,
Nowak
E.
,
Benson
M. J.
,
Kronenberg
M.
,
Noelle
R. J.
,
Cheroutre
H.
.
2009
.
Retinoic acid can directly promote TGF-β‑mediated Foxp3+ Treg cell conversion of naive T cells.
Immunity
30
:
471
472, author reply 472–473
.
65
Sela
U.
,
Olds
P.
,
Park
A.
,
Schlesinger
S. J.
,
Steinman
R. M.
.
2011
.
Dendritic cells induce antigen-specific regulatory T cells that prevent graft versus host disease and persist in mice.
J. Exp. Med.
208
:
2489
2496
.
66
Zhou
R.
,
Horai
R.
,
Mattapallil
M. J.
,
Caspi
R. R.
.
2011
.
A new look at immune privilege of the eye: dual role for the vision-related molecule retinoic acid.
J. Immunol.
187
:
4170
4177
.
67
Akimova
T.
,
Beier
U. H.
,
Wang
L.
,
Levine
M. H.
,
Hancock
W. W.
.
2011
.
Helios expression is a marker of T cell activation and proliferation.
PLoS One
6
:
e24226
.
68
Yadav
M.
,
Louvet
C.
,
Davini
D.
,
Gardner
J. M.
,
Martinez-Llordella
M.
,
Bailey-Bucktrout
S.
,
Anthony
B. A.
,
Sverdrup
F. M.
,
Head
R.
,
Kuster
D. J.
, et al
.
2012
.
Neuropilin-1 distinguishes natural and inducible regulatory T cells among regulatory T cell subsets in vivo.
J. Exp. Med.
209
:
1713
1722, S1711‑1719
.
69
Weiss
J. M.
,
Bilate
A. M.
,
Gobert
M.
,
Ding
Y.
,
Curotto de Lafaille
M. A.
,
Parkhurst
C. N.
,
Xiong
H.
,
Dolpady
J.
,
Frey
A. B.
,
Ruocco
M. G.
, et al
.
2012
.
Neuropilin 1 is expressed on thymus-derived natural regulatory T cells, but not mucosa-generated induced Foxp3+ T reg cells.
J. Exp. Med.
209
:
1723
1742, S1721
.
70
Banchereau
J.
,
Steinman
R. M.
.
1998
.
Dendritic cells and the control of immunity.
Nature
392
:
245
252
.
71
Steinman
R. M.
,
Hawiger
D.
,
Liu
K.
,
Bonifaz
L.
,
Bonnyay
D.
,
Mahnke
K.
,
Iyoda
T.
,
Ravetch
J.
,
Dhodapkar
M.
,
Inaba
K.
,
Nussenzweig
M.
.
2003
.
Dendritic cell function in vivo during the steady state: a role in peripheral tolerance.
Ann. N. Y. Acad. Sci.
987
:
15
25
.
72
Ohnmacht
C.
,
Pullner
A.
,
King
S. B.
,
Drexler
I.
,
Meier
S.
,
Brocker
T.
,
Voehringer
D.
.
2009
.
Constitutive ablation of dendritic cells breaks self-tolerance of CD4 T cells and results in spontaneous fatal autoimmunity.
J. Exp. Med.
206
:
549
559
.
73
Tang
Q.
,
Adams
J. Y.
,
Tooley
A. J.
,
Bi
M.
,
Fife
B. T.
,
Serra
P.
,
Santamaria
P.
,
Locksley
R. M.
,
Krummel
M. F.
,
Bluestone
J. A.
.
2006
.
Visualizing regulatory T cell control of autoimmune responses in nonobese diabetic mice.
Nat. Immunol.
7
:
83
92
.
74
Serra
P.
,
Amrani
A.
,
Yamanouchi
J.
,
Han
B.
,
Thiessen
S.
,
Utsugi
T.
,
Verdaguer
J.
,
Santamaria
P.
.
2003
.
CD40 ligation releases immature dendritic cells from the control of regulatory CD4+CD25+ T cells.
Immunity
19
:
877
889
.
75
Schwarz
A.
,
Schwarz
T.
.
2010
.
UVR-induced regulatory T cells switch antigen-presenting cells from a stimulatory to a regulatory phenotype.
J. Invest. Dermatol.
130
:
1914
1921
.
76
Ullrich
S. E.
2010
.
Two-way traffic on the bridge from innate to adaptive immunity.
J. Invest. Dermatol.
130
:
1773
1775
.
77
Jiang
H. R.
,
Muckersie
E.
,
Robertson
M.
,
Forrester
J. V.
.
2003
.
Antigen-specific inhibition of experimental autoimmune uveoretinitis by bone marrow-derived immature dendritic cells.
Invest. Ophthalmol. Vis. Sci.
44
:
1598
1607
.
78
Streilein
J. W.
1987
.
Immune regulation and the eye: a dangerous compromise.
FASEB J.
1
:
199
208
.

The authors have no financial conflicts of interest.