Participation of Pigment Epithelium of Iris and Ciliary Body in Ocular Immune Privilege. 2. Generation of TGF-β–Producing Regulatory T Cells

Abstract

purpose. To determine whether T cells exposed to cultured iris and ciliary body pigment epithelial (I/CB PE) cells acquire the capacity to modify the activation, differentiation, and effector functions of bystander T cells, and if so, to identify the mechanism.

methods. T cells from naive BALB/c mice were cultured with I/CB PE cells, x-irradiated, and used as regulators (a) of T-cell activation in vitro and (b) of delayed hypersensitivity expression in vivo. Neutralizing anti–TGF-β and –IL-10 antibodies were used to abolish regulatory function. T-cell activation was assessed for proliferation by[ ³H]thymidine incorporation and for IL-2, IFN-γ, IL-4, and IL-10 production by semi-quantitative RT-PCR for mRNA and by supernatant analysis by ELISA. I/CB PE–exposed T cells were evaluated for mRNA content of IFN-γ, IL-4, TNF-α, TGF-β1, TGF-β2, and IL-10, and their supernatants were analyzed for content of TGF-β.

results. T cells exposed to I/CB PE cells inhibited anti-CD3–driven activation of bystander naive T cells in vitro and suppressed the expression of delayed hypersensitivity in vivo. Bystander T cells cocultured with I/CB PE–exposed T cells failed to proliferate and secreted high levels of IL-4 and IL-10 but low amounts of IL-2 and IFN-γ. Regulation of bystander T-cell activation was mediated via enhanced secretion of TGF-β by I/CB PE–exposed T cells.

conclusions. T cells exposed to cultured I/CB PE cells were induced to secrete active and latent TGF-β, which conferred on the T cells the capacity to inhibit the differentiation as well as the effector function of Th1-type cells.

Tcells that enter intraocular compartments (anterior chamber, vitreous cavity, subretinal space) carry with them the threat of immunogenic inflammation. Should infiltrating T cells encounter an antigen that binds their receptor for antigen, the cells can become activated, leading to release of proinflammatory cytokines with the potential to disrupt the visual axis and cause blindness. To avoid this catastrophe, the eye utilizes a variety of anatomic, cellular, and molecular strategies to limit the consequences of T-cell activation and to avoid the expression of intraocular inflammation. ¹ In part, these strategies account for the existence of immunologic privilege. ^{2 3} One strategy that limits immunogenic inflammation and promotes immune privilege is the existence of an immunosuppressive intraocular microenvironment. Aqueous humor contains a remarkable array of cytokines, growth factors, and low-molecular-weight molecules that suppress T-cell activation and limit the effects of proinflammatory cytokines. ^{4 5} In addition, parenchymal cells that surround the intraocular compartments (pigment epithelium of iris and retina, secretory epithelium of ciliary body, corneal endothelium) express cell surface molecules such as CD95L that trigger apoptosis among CD95-bearing leukocytes that encounter these cells. ^{6 7} This property of the corneal endothelium accounts in part for the privilege that orthotopic corneal allografts experience. ^{8 9}  

Recently, our laboratory has studied immunomodulatory properties of pigment epithelial (PE) cells cultured from explants of murine iris and ciliary body (I/CB). ¹⁰ When murine T cells were stimulated with ligands for their T-cell receptor (Tcr) for antigen in the presence of cultured I/CB PE cells, the responding T cells failed to proliferate or secrete the typical cytokines usually evoked by these stimuli. Suppression of T-cell activation by cultured I/CB PE cells was found to be accomplished by a contact-dependent mechanism, rather than by a soluble product secreted by the PE cells. However, the suppression of T cells was not mediated by CD95/CD95L interactions, nor did exposure of T cells to I/CB PE cells trigger T-cell apoptosis. To the contrary, T cells that were stimulated with a Tcr ligand in the presence of cultured I/CB PE cells were protected from apoptosis. 

In the present study, we have investigated the immunomodulatory properties of T cells cultured in the presence of I/CB PE cells. Our experiments reveal that T cells exposed to I/CB PE cells acquired the capacity to alter the functional properties of bystander T cells in both in vitro and in vivo assays. Specifically, naive T cells that were stimulated by anti-CD3 antibodies in the presence of T cells previously exposed to cultured I/CB PE cells failed to proliferate. Moreover, these cells secreted enhanced amounts of IL-4, and IL-10, but reduced levels of IL-2 and IFN-γ. In a local adoptive transfer assay, T cells first exposed to I/CB PE cells suppressed the in vivo expression of ovalbumin-specific delayed hypersensitivity (DH) by specific effector T cells. The ability of I/CB PE–exposed T cells to regulate bystander T-cell functions was mediated, at least in part, by TGF-β. 

Materials and Methods

Animals

Normal male BALB/c and C57BL/6 mice at 6 to 8 week of age were obtained from our domestic breeding facility or purchased from Taconic Farms (Germantown, NY). 

Animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Culture Medium

T-cell cultures were carried out with serum-free medium, composed of RPMI 1640, 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-ME (Sigma Chemical, St. Louis, MO) and supplemented with 0.1% bovine serum albumin (Sigma Chemical) and 0.2% ITS + culture supplement (Collaborative Biochemical Products, Bedford, MA). 

Cultures of I/CB PE Cells

I/CB PE cells were cultivated as previously described. ¹⁰ Briefly, whole I/CB tissues were collected from 6- to 8-week-old male BALB/c mice, and single-cell suspensions were prepared by incubating the tissue in phosphate-buffered saline (PBS) containing 1 mg/ml of Dispase (Boehringer Mannheim, Mannheim, Germany) and 0.05 mg/ml of DNaseI (Boehringer Mannheim) at 37°C for 1 hour, followed by trituration through 21- and 23-gauge needles. Monodisperse I/CB cells were washed twice with RPMI complete medium and then cultured in 35-mm cell culture dishes at 37°C in 5% CO₂ atmosphere for 14 days. At this time point,> 95% of the cells in the culture stained positive with FITC-labeled anti-pan keratin antibody (Clone PCK-26; Sigma Immunochemicals, St. Louis MO), indicating that they were virtually all pigment epithelial cells. 

Activation of Naive T Cells with Anti-CD3 Antibodies

T cells were prepared as a single cell suspension from naive BALB/c mouse spleens and then purified through T-cell enrichment columns (R&D Systems, Minneapolis, MN). Purified T cells, 1 × 10⁵, were placed in culture wells with 1 μg/ml of anti-CD3 antibody (Clone 2C11; Pharmingen, San Diego, CA).[ ³H]thymidine (0.5 μCi) was added to these cultures during the terminal 8 hours of the 3-day culture interval, and then the cells were harvested onto glass filters using an automated cell harvester (Tomtec, Orange, CT). Radioactivity was assessed by liquid scintillation spectrometry, and the amount was expressed as counts per minute (cpm). 

Exposure of Naive T Cells to Cultured I/CB PE Cells

Naive BALB/c spleen T cells were purified as described above, placed in culture wells containing I/CB PE cells, BALB/c 3T3 cells (fibroblasts; ATCC, Rockville, MD) or no other cells. After 48 hours, the T cells were harvested by gentle pipetting from the culture and were used for further experiments after being washed twice with serum-free RPMI medium. The level of contamination of the harvested T cells by either I/CB PE cells or 3T3 cells was found to be <1% by flow cytometry. 

Assays of In Vitro Functional Properties of Naive T Cells Cultured in the Presence of T Cells Exposed to Cultured I/CB PE Cells

Proliferation.

T cells first cultured with I/CB PE cells (serially diluted from 1 × 10⁵/well; referred to as I/CB PE–exposed T cells) were harvested, x-irradiated (2000 R) and cocultured with naive BALB/c T cells (serially diluted from 1 × 10⁵/well) in the presence of 1 μg/ml of antiCD3 Ab. Supernatants were collected at 48 hours for cytokine assays, and proliferation was assessed after 72-hour incubations as described above. 

Cytokine Content.

To measure cytokine (IL-2, IFN-γ, IL-4, IL-10) content in supernatants of T-cell cultures, quantitative capture ELISAs were used. Culture supernatants were collected at 48 hours and immediately frozen and stored at −20°C until used. ELISAs were carried out according to manufacturer’s instructions (Pharmingen). Rat mAbs to mouse IFN-γ (Clone: R4-6A2), IL-2 (JES6-1A12), IL-4 (11.B11), or IL-10 (JES-2A5; PharMingen) were used as coating Abs. Biotinylated rat mAbs to mouse IFN-γ (XMG1.2), IL-2 (JES6-5H4), IL-4 (BVD6-24G2), and IL-10 (SXC-1; PharMingen) were used as detecting Abs. All recombinant cytokines for assay standards were purchased from Pharmingen. 

Bioassay for TGF-β.

Supernatants of T cells exposed for 48 hours to cultured I/CB PE cells were collected, and then biologically active TGF-β was measured using Mv1Lu cells (ATCC). To detect mature TGF-β, the supernatants were diluted 1:4 with Eagle’s minimum essential medium (EMEM; Biowhitaker), which consisted of 2 mM l-glutamine, 10 mM HEPES, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.5% fetal calf serum. Diluted supernatants (100 μl) were added to 96-well flat-bottom plates. To measure total TGF-β, supernatants were pretreated with 1 N HCl (1:10) for 1 hour and then neutralized with a mixture of 1 N NaOH:1 M HEPES (1:5). These acidified supernatants were diluted 1:10 with complete EMEM containing 0.5% fetal calf serum, and then 100 μl was added to 96-well flat-bottom plates. Mv1Lu cells (1 × 10⁵/100 μl) were added to each well and cultured for 24 hours at 37°C in 5% CO₂. Cultures were pulsed with 1 μCi [³H]thymidine 6 hours before termination and assayed as described above. Half-maximal inhibition was determined by polynomial regression on log-log transformation of standard curves and experimental samples. The results are expressed as picograms per milliliter. 

Semiquantitative Reverse Transcriptase Polymerase Chain Reaction (RT-PCR).

RT-PCR was carried out as previously described. ¹¹ Briefly, total RNA was extracted from T cells cultured for 48 hours with or without I/CB PE cells by using Rneasy RNA extraction kit (Qiagen, Chatsworth, CA). cDNA was synthesized with AMV reverse transcriptase (Promega, Madison, WI) according to manufacturer’s instructions. The following primers were used from 5′ to 3′: GAPDH sense, GGTGAAGGTCGGTGTGAACGGA; GAPDH antisense, TGTTAGTGGGGTCTCGCTCCTG; TGF-β1 sense, CAAGGAGACGGGAATACAGGGCT; TGF-β1 antisense, CGCACACAGCAGTTCTTCTCTGT; TGF-β2 sense, CACCACAAAGACAGGAACCTGG; TGF-β2 antisense, GCGAAGGCAGCAATTATCCTGCAC; IL-4 sense, ACGGAGATGGATGTGCCAAACGTC; IL-4 antisense, ATGGTACTCCAGAAGACCAGAGG; IL-10 sense, GGTTGCCAAGCCTTATCGGAAATG; IL-10 antisense, TTGTAGACACCTTGGTCTTGGAGC; TNF-α sense, GAAAGCATGATCCGCGACGTGGA; TNF-α antisense, TACGACGTGGGCTACAGGCTTG; IFN-γ sense, CCTCATGGCTGTTTCTGGCTGTTA; IFN-γ antisense, CATTGAATGCTTGGCGCTGGACC. 

All primers were designed by Pascale Alard (University of Virginia, Charlottesville, VA), and their specificities were confirmed in preliminary experiments. Samples containing mRNA for PCR analysis were treated with an excess of RNase-free DNA before DNA synthesis. PCR was carried out under the following conditions: denaturation: 94°C, 30 seconds; annealing: 55°C, 30 seconds; extension: 72°C, 60 seconds. The amplification cycles used were determined by preliminary experiments to obtain the best results. The cycles were as follows: GAPDH, 20; TGF-β1, 25; TGF-β2, 25; IL-10, 30; IL-4, 40; IFN-γ, 25; and TNF-α, 30. In addition, we carried out the PCR using 1:2 or 1:3 serially diluted template cDNAs, which showed a linearity of amplification through at least 30 cycles of amplification in our system. After 20 to 40 cycles of amplification, PCR products were electrophoresed in 2% agarose gel and visualized by ethidium bromide staining. TGF-β2 expression was compared by semiquantitative PCR by the method of Jaffe et al. with slight modification. PCR was carried out with serially diluted cDNA. Then PCR products were electrophoresed in 2% agarose gel and stained with 0.5 μg/ml ethidium bromide for 20 minutes. Photographs of the gel were taken with a high-resolution camera, and the density of the band of negative image was analyzed by NIH Image software. The expression level of mRNA was standardized by the expression of GAPDH as an internal control. The predicted sizes of PCR product for relevant cytokines and growth factors are TGF-β1, 260 bp; TGF-β2, 327 bp; IL-10, 233 bp; and GAPDH, 245 bp. The PCR product for each cytokine was subjected to restriction enzyme digestion to confirm its presence. In the case of TGF-β2, the proper restriction patterns were observed when the product was digested with XhoI and with MboI. 

Neutralizing Antibodies.

Anti-pan TGF-β antibody and anti–TGF-β2 antibody were purchased from R&D systems. Neutralizing anti–IL-10 antibody was purified from culture supernatant of hybridoma JES5-2A5. Anti-pan TGF-β2 antibody was added from the beginning of culture at a concentration of 5μ g/ml. Similarly, anti–IL-10 antibody (100 ng/ml) was used for neutralization. Anti-pan TGF-β was used for neutralizing TGF-β contained in the conditioned medium of T cells treated with I/CB PE cells. 

Local Adoptive Transfer Assay of DH.

To assess suppression of DH expression, a local adoptive transfer as previously described was used. ¹² Normal female BALB/c mice at 6 to 8 weeks of age were used as recipients. Putative regulatory T cells were harvested from cultures in which naive T cells were exposed to cultured I/CB PE cells for 48 hours. Responder cells (DH-mediating effectors) were obtained from spleens of BALB/c mice immunized SC 10 days previously with ovalbumin (OVA, 200 μg; Sigma Chemical) plus complete Freunds’ adjuvant (CFA). Responder cells (1 × 10⁶) were mixed with regulator cells (5 × 10⁵/mouse) and OVA (100 μg/ml) and injected (10μ l/injection) into ear pinnae of naive BALB/c mice. Ear swelling responses were measured with an engineer’s micrometer (Mitsutoyo; MTI Corporation, Paramus, NJ) 24 and 48 hours later. Results were expressed as specific ear swelling = (24-hour measurement − 0-hour measurement [experimental ear]) − (24-hour measurement − 0-hour measurement [negative control ear]) × 10⁻³ mm. 

Statistical Analysis

Statistical significance was analyzed by Student’s t-test. Differences were considered significant at P < 0.05. Each experiment was repeated at least twice with similar results. 

Results

Effect of T Cells Exposed to Cultured I/CB PE Cells on Activation of Bystander T Cells

Naive BALB/c T cells were cultured for 48 hours in the presence of cultured I/CB PE cells as well as in the absence of any ligand for the Tcr for antigen. The T cells were then harvested, exposed to x-irradiation (2000 R), and added (in a dose-dependent manner) as“ regulators” to secondary cultures that contained naive BALB/c T cells plus monoclonal anti-CD3 antibodies (1 μg/ml). The secondary cultures were incubated for 72 hours, with[ ³H]thymidine added during the terminal 8 hours. The results of a representative experiment (of 3) are presented in Figure 1 . Although anti-CD3 antibodies induced intense proliferation in positive control cultures (to which no regulators had been added), the presence of regulator T cells that previously encountered cultured I/CB PE cells inhibited radioisotope incorporation by the responding T cells (Fig. 1A) . In separate experiments, when x-irradiated BALB/c T cells, not previously exposed to PE cells, were added to naive T cells stimulated with anti-CD3 antibodies, no evidence of suppression of proliferation was observed (data not shown). 

Similar experiments were conducted in which naive T cells were first exposed to I/CB PE cells in the presence of anti-CD3 antibodies. T cells harvested from these cultures displayed regulatory properties in secondary cultures, similar to those describe above (data not shown). Thus, the capacity of I/CB PE cells to confer regulatory functions on T cells is independent of whether the T cells are stimulated through the Tcr or not at the time of coculture. 

To determine whether the alteration of T cells by I/CB PE was unique, T cells were also exposed to cultured fibroblasts (3T3 cells). Naive BALB/c T cells were cultured for 48 hours in the presence of syngeneic 3T3 fibroblasts or I/CB PE cells. The T cells were subsequently harvested, x-irradiated, and added as regulators to cultures in which naive BALB/c T cells were stimulated with anti-CD3 antibodies. The results of a representative experiment are displayed in Figure 1B . T cells first exposed to 3T3 cells displayed no capacity to inhibit bystander T-cell activation in secondary cultures stimulated with anti-CD3 antibodies. These findings indicate that T cells first exposed to cultured I/CB PE cells (but not fibroblasts) acquire the novel capacity to suppress activation of bystander T cells stimulated in vitro with anti-CD3 antibodies. 

Functional Phenotype of Bystander T Cells Cultured with I/CB PE-Exposed T Cells

We next wanted to evaluate the functional properties of bystander T cells stimulated with anti-CD3 antibodies in the presence (or absence) of T cells previously exposed to I/CB PE cells. I/CB PE–exposed T cells (or T cells cultured in the absence of I/CB PE) were prepared as before, x-irradiated, and added to cultures of naive T cells stimulated with anti-CD3. After 48 hours, the supernatants of these cultures were harvested and assayed by ELISA for content of IL-2, IL-4, IL-10 and IFN-γ. As the results displayed in Figure 2 reveal, naive T cells stimulated with anti-CD3 in the presence of x-irradiated control T cells secreted large amounts of IL-2, IL-4, and IFN-γ and small amounts of IL-10. However, in the presence of T cells previously exposed to I/CB PE, anti-CD3–stimulated T cells secreted reduced amounts of IL-2 and IFN-γ, slightly enhanced amounts of IL-4, and greatly enhanced amounts of IL-10. These results indicate that in the presence of I/CB PE–exposed T cells, the cytokine profiles of naive T cells stimulated by anti-CD3 antibodies were biased away from the Th1 phenotype. IL-4 and IL-10 synthesis was promoted at the expense of IL-2 and IFN-γ. 

Effect of I/CB PE–Exposed T Cells on In Vivo Expression of Delayed Hypersensitivity

T cells were harvested from cultures containing I/CB PE cells, x-irradiated, and used as “regulators” in local adoptive transfer assays. Control regulators were naive BALB/c splenic T cells that had been x-irradiated. Responder T cells were prepared from spleens of BALB/c mice immunized 10 days previously with OVA plus CFA. Control responder T cells were obtained form naive BALB/c mice. Responder cells (1 × 10⁶) were mixed with regulator cells (5 × 10⁵) plus OVA (100 μg/ml) and injected (10 μl) into the ear pinnae of naive BALB/c mice. Ear swelling responses were assessed 24 hours later and are presented in Figure 3 . Responder T cells coinjected with control regulator cells evoked ear swelling responses significantly greater than negative controls. By contrast, responder T cells coinjected with T cells previously exposed to I/CB PE cells elicited ear swelling responses not significantly different from negative controls. These results (a) reveal that T cells first exposed to cultured I/CB PE cells in vitro acquire the capacity to suppress expression of delayed hypersensitivity in vivo, and (b) support the previous results, which indicated that I/CB PE-regulator T cells bias bystander T cells away from Th1 responses. 

Importance of Cell Contact on Ability of I/CB PE–Exposed T Cells to Suppress Bystander T-Cell Activation In Vitro

When cultured I/CB PE cells alter the functional properties of T cells (converting them into regulators), direct cell-to-cell contact has been found to be necessary. ¹⁰ Our next experiments examined whether a similar contact-dependent restriction applied to the inhibitory activities of regulatory I/CB PE–exposed T cells. T cells were first cultured with I/CB PE cells, x-irradiated, and then added to transwell cultures in which (a) both regulatory cells and responding naive T cells (plus anti-CD3) were in the same compartment, or (b) regulatory T cells were separated from responding naive T cells (plus anti-CD3) by a transwell membrane. The cultures were allowed to incubate for 72 hours, with [³H]thymidine being added during the terminal 8 hours. As the results displayed in Figure 4 reveal, bystander T cells proliferated poorly in the presence of I/CB PE regulatory T cells, whether the latter cells were admixed with the responder cells or separated from the responders by a membrane. This result suggests that I/CB PE—exposed T cells secrete soluble factor(s) that inhibit anti-CD3–driven proliferation among bystander T cells. 

Identification of TGF-β as Mediator of Suppression by I/CB PE–Exposed T Cells

Because the previous experiments suggested that T cells first exposed to cultured I/CB PE cells secreted immunosuppressive factors, we next examined mRNA expression of various soluble cytokines and growth factors in T cells of this type. Naive T cells were cultured for 48 hours in the presence or absence of cultured I/CB PE cells. At the conclusion of the culture, the T cells were harvested, and mRNA was collected and subjected to semi-quantitative RT-PCR analysis for the following gene products: IFN-γ, IL-4, TNF-α, TGF-β1, TGF-β2, and IL-10. The results of a representative experiment are displayed in Figure 5 . Several similar experiments were performed and the results were virtually identical. These results are summarized in Table 1 . In analyzing the results, valid comparisons can be made only between a PCR product for one cytokine obtained from T cells not exposed to I/CB PE and PCR product for the same cytokine obtained from T cells exposed to I/CB PE. Thus, T cells cultured in the absence of I/CB PE cells contained significant amounts of mRNA for TGF-β1 and IFN-γ, small amounts of mRNA for TGF-β2 and TNF-α, and no mRNA for IL-10 or IL-4. By contrast, T cells exposed to I/CB PE cells contained significantly enhanced amounts of mRNA for TGF-β2, IL-10, and small amounts of IL-4 mRNA. The significant increases in mRNA levels for TGF-β2 and IL-10 (cytokines with known immunosuppressive properties) suggested that either or both of these factors might be responsible for the capacity of I/CB PE–exposed T cells to suppress bystander T-cell activation. 

Effect of Neutralizing Anti–TGF-β, but Not IL-10, Antibodies on Suppressive Properties of I/CB PE–Exposed T Cells

Using neutralizing anti–TGF-β and anti–IL-10 antibodies, we next examined whether TGF-β2 and/or IL-10 was responsible for the suppression mediated by I/CB PE–exposed T cells. Regulatory T cells were prepared by culturing naive BALB/c T cells with cultured I/CB PE cells for 48 hours. Immediately thereafter the T cells were harvested, x-irradiated, and used as regulators in secondary cultures containing naive BALB/c T cells, anti-CD3 antibodies, and anti-TGF-β2 or anti–IL-10 antibodies. [³H]thymidine incorporation was assessed after 72 hours of culture. The results of representative experiments are presented in Figure 6 . As revealed in Figure 6A , anti–TGF-β2 antibodies enabled responder T cells to proliferate vigorously in the presence of regulatory T cells. Alternatively, anti–IL-10 antibodies failed to restore T-cell proliferation in comparable cultures (Fig. 6B) . These results suggest that T cells exposed to cultured I/CB PE cells suppress bystander T-cell activation by a TGF-β–dependent mechanism. 

Because T cells exposed to cultured I/CB PE cells accumulated high levels of TGF-β2 mRNA, it was important to document that these cells were actually capable of producing enhanced amounts of this cytokine. Accordingly, T cells were cultured alone or in the presence of I/CB PE cells for 48 hours The T cells were then harvested, washed, and recultured in serum-free medium for 24 hours. The supernatants of these secondary cultures were collected and subjected to a bioassay that detects active TGF-β. In some assays, neutralizing anti–TGF-β was added to confirm that the activity being measured was due to TGF-β. The results of one such experiment are presented in Figure 7 . Supernatants of T cells cultured alone contained small amounts of active and somewhat larger amounts of total TGF-β. In comparison, supernatants of T cells cultured with I/CB PE cells contained significantly larger amounts of both active and total TGF-β. In the bioassay, neutralizing anti–TGF-β antibodies abolished virtually all activity. We conclude that T cells exposed to cultured I/CB PE cells upregulate TGF-β mRNA and secrete enhanced amounts of both active and total TGF-β. 

Discussion

Pigment epithelial cells of the iris, ciliary body, and retina provide the eye with a pervasive light sink that absorbs virtually all light energy passing through the cornea, except that which falls directly on and is absorbed by the photoreceptor cells of the retina. The cytoplasm of these epithelial cells is filled with melanin granules that accomplish this task. In addition, RPE cells provide metabolic and cellular support to photoreceptors, recycling critical molecules in the visual pigment and phagocytizing effete outer segments of rods and cones. PE cells of the iris and retina ^{13 14} also provide important barrier functions by virtue of extensive tight junctions that unite these cells and prevent the wanton passage of blood-derived molecules and cells into the anterior chamber and subretinal space, respectively. PE cells of the ciliary body are not joined by tight junctions as the superficial layer of secretory ciliary epithelium to which they are apposed takes responsibility for creating a blood–tissue barrier. The barrier properties of PE cells are not the only contribution these cells make to immune privilege. Numerous studies have demonstrated that RPE can secrete immunosuppressive factors, ^{15 16} and several recent publications document that the constitutive expression of CD95 ligand (CD95L) by RPE enables these cells to promote programmed cell death among CD95⁺ T cells, ¹⁷ although a soluble factor rather than CD95 itself appears to be the proximate cause of apoptosis in this instance. ¹⁸  

Pigment epithelial cells of the iris and ciliary body also display properties that suggest that they contribute to ocular immune privilege. Almost a decade ago our laboratory and others reported that cultured explants of murine iris and ciliary body suppressed mixed lymphocyte reactions in vitro. ^{19 20 21 22 23} In those reports, soluble factors found in the supernatants of these cultures accounted for the majority of the immunosuppressive activity, although evidence of a cell-associated inhibitory mechanism was also found. Recently, we have reported that PE cells alone cultured from murine iris and ciliary body profoundly inhibit the in vitro activation of T cells triggered through the Tcr via direct cell-to-cell contact. ¹⁰  

However, “inhibition” is not the most suitable term to use, as the results of the experiments on which this article is based emphasize. T cells exposed to cultured I/CB PE cells may have failed to proliferate when stimulated in vitro, but the T cells were nonetheless“ activated.” Using semi-quantitative RT-PCR analysis, we found that anti-CD3–stimulated T cells cultured in the presence of I/CB PE cells upregulated the expression of genes encoding TGF-β2, IL-10, and IL-4. In addition, the supernatants of T cells previously cultured in the presence of I/CB PE cells contained markedly enhanced (compared with controls) amounts of both active and latent TGF-β. Thus, T cells stimulated with anti-CD3 in the presence of cultured I/CB PE cells appear to be activated in a novel way. Although these T cells did not proliferate, they displayed enhanced production of cytokines that are often associated with immune suppression rather than immunogenic inflammation. 

The conclusion that I/CB PE cells bias T cells (stimulated with anti-CD3 or not) toward immunosuppressive/immunoregulatory properties is buttressed by our in vitro and in vivo studies of the effects of these T cells on bystander T cells. When x-irradiated T cells exposed to I/CB PE cells were cocultured with naive T cells stimulated with anti-CD3 antibodies, the responding T cells produced significantly reduced amounts of IL-2 and IFN-γ, cytokines that are produced in large amounts by T cells stimulated with anti-CD3 in the absence of I/CB PE cells. In fact, the pattern of cytokines produced by naive T cells activated in vitro under the influence of I/CB PE–exposed T cells suggests that the responders were biased away from the proinflammatory phenotype of Th1 cells. This suggestion was strengthened by our finding that OVA-specific effector T cells were inhibited in vivo from evoking DH responses in a local adoptive transfer assay if I/CB PE–exposed T cells were included in the inoculum. It is pertinent that neutralizing anti-TGF-β antibodies, but not anti–IL-10 antibodies, abolished the immunosuppressive activities of I/CB PE–exposed T cells in the in vitro assays. Together, the results of these studies indicate that T cells activated through their Tcr in the presence of cultured I/CB PE cells fail to acquire the capacity to mediate immunogenic inflammation. Instead, because they secrete large amounts of TGF-β2, they similarly inhibit bystander T cells from developing into mediators of immunogenic inflammation. 

The influence of the ocular microenvironment on bone marrow–derived cells that enter intraocular compartments is both profound and pervasive. In the current experiments, I/CB PE cells cultured from normal eyes were found to convert naive T cells into regulatory cells that alter the fate and function of bystander T cells. We have not yet tested whether I/CB PE cells would have a similar effect on previously primed T cells, especially cells of the Th1 phenotype. This is an important issue because many immunopathogenic eye diseases are believed to be mediated by Th1 cells that are specific for viral or autoantigens expressed intraocularly. 

It is of interest that I/CB PE cells induced T cells to secrete large amounts of active and latent TGF-β. T cells activated in the presence of aqueous humor have also been shown to secrete this immunosuppressive cytokine. ²⁴ Similarly, peritoneal exudate cells exposed to aqueous humor or TGF-β2 have been found to display enhanced secretion of TGF-β. ²⁵ When antigen-presenting cells (APC) of this type are pulsed with antigen and injected intravenously into naive mice, they induce systemic immune deviation similar to ACAID. ²⁶ These various results suggest that an underlying theme of the influence of the ocular microenvironment on bone marrow–derived cells is to promote the production of TGF-β, especially the active form. An immediate effect of this activity would be to enrich the intraocular concentration of TGF-β; however, because T cells and APCs are mobile and can migrate from one tissue to another, the possibility exists that the T cells and APCs that secrete large amounts of TGF-β upon exposure to the ocular microenvironment may carry this property with them if they migrate out of the eye and enter distant tissues. It has already been proposed that the eye-derived APCs that carry to the spleen the antigenic signals necessary for ACAID create a TGF-β–rich microenvironment at the site where they come to rest in splenic parenchyma. ²⁷ Eye-derived T cells may have a similar capacity. Creation in this manner of “metastatic” sites enriched for TGF-β content in extraocular organs may have systemic consequences. That is, TGF-β–producing T cells or APCs that leave the eye may modify immune responses systemically by altering APC-dependent T-cell activation in secondary lymphoid tissues or perhaps even in other somatic tissues. Experiments to test these interesting possibilities are currently underway. 

 Supported in part by National Institutes of Health Grant EY 05678. JWS is a recipient of a Research to Prevent Blindness Senior Scientific Investigator Award.

 Submitted for publication November 8, 1999; revised February 1, 2000 and July 13, 2000; accepted July 19, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Influence of I/CB PE–exposed T cells on anti-CD3–driven activation of naïve T cells. Naive BALB/c T cells (1 × 10⁵) were cultured with I/CB PE cells (cultured for 14 days from I/CB of BALB/c eyes, 1 × 10⁵/well) (A) or 3T3 cells (1 × 10⁵/well) (B) for 48 hours. The T cells were then harvested, x-irradiated (2000 R), and added as regulators (A, in a dose-dependent fashion) to cultures containing naive BALB/c T cells (1 × 10⁵/well) and anti-CD3 antibodies (1 μg/ml). After 72 hours, [³H]thymidine was added to assess proliferation. Results are presented as mean cpm ± SEM for triplicate samples.

Figure 2.

 

Cytokine content of supernatants from T cells stimulated by anti-CD3 antibodies in the presence of I/CB PE–exposed T cells. X-irradiated I/CB PE–exposed T cells or T cells cultured in the absence of I/CB PE cells were added as regulators (1 × 10⁵/well) to wells containing naive BALB/c T cells (1 × 10⁵) plus anti-CD3 antibodies (1 μg/ml). In control cultures, no regulatory cells were added. After 48 hours, the supernatants were removed and assayed by ELISA for content of (A) IFN-γ, (B) IL-2, (C) IL-4, and (D) IL-10.

Figure 3.

 

Influence of I/CB PE–exposed T cells on expression of delayed hypersensitivity in vivo. BALB/c mice were immunized SC with OVA (200μ g) in CFA. After 10 days, purified T cells were prepared from their spleens as “responders” for local adoptive transfer assays. BALB/c T cells cultured for 48 hours in the presence of I/CB PE cells were used as “regulators.” Naive BALB/c T cells were used alternatively as control responders or as control regulators. Responder cells (1 × 10⁶) were mixed with OVA (100 μg) and regulator cells (I/CB PE–exposed T cells, 5 × 10⁵) and injected (10μ l) into the ear pinnae of naive BALB/c mice (5 mice per group). Ear swelling responses, assessed by engineer’s micrometer 24 hours later, are presented as mean swelling responses ± SEM.

Figure 4.

 

Determination whether regulation by I/CB PE–exposed T cells is mediated by soluble factor or by contact-dependent mechanism. (A) I/CB PE–exposed T cells (1 × 10⁵) were added to naive BALB/c T cells (1 × 10⁵) plus anti-CD3 in the same well (▪) or were separated from naive BALB/c T cells plus anti-CD3 antibodies by a transwell membrane ( ). (B) I/CB PE–exposed T cells (or naive T cells; 1 × 10⁵) were cultured for 24 hours in serum-free medium. Supernatants were collected and added to cultures containing naive BALB/c T cells plus anti-CD3 antibodies. After 72 hours, [³H]thymidine was added to assess proliferation. Results are presented as mean cpm ± SEM for triplicate samples.

Figure 5.

 

mRNA expression of selected cytokine genes in I/CB PE–exposed T cells. BALB/c T cells were cultured for 48 hours in the presence of I/CB PE cells. The T cells were harvested, and mRNA was extracted and subjected to semi-quantitative RT-PCR analysis. Amplified products were electrophoresed on 2% agarose gel and stained with ethidium bromide. M-pGEM molecular weight marker. Lanes 2, 4, 6, and 8: T cells exposed to I/CB PE cells; lanes 1, 3, 5, and 7: T cells not exposed to I/CB PE cells. Lanes 1 and 2: TGF-β1; lanes 3 and 4: TGF-β2; lanes 5 and 6: IL-10; lanes 7 and 8: GAPDH.

Table 1.

 

Expression of mRNA of Selected Cytokine Genes in T Cells Exposed to Cultured I/CB PE Cells

Cytokine	T Cells Cultured
		In Absence of I/CB PE	In Presence of I/CB PE
;l>TGF-β1	+	+
TGF-β2	+	++
IL-10	−	+
IL-4	−	+
IFN-γ	+	+
TNF-α	+	+

 Relative amounts of cytokine mRNA in each gel, analyzed according to the following rules: −, no mRNA was detected; +, mRNA for that cytokine was detected; ++, the intensity of the band for a particular cytokine (from T cells cultured in the presence of I/CB PE) is visibly greater than the intensity of the band for that same cytokine (from T cells cultured in the absence of I/CB PE).

Figure 6.

 

Inhibition of regulation by I/CB PE–exposed T cells with antibodies to TGF-β2 and IL-10. X-irradiated I/CB PE–exposed T cells (1 × 10⁵) were added to cultures of BALB/c T cells (1 × 10⁵) plus anti-CD3 antibodies in the presence of (A) anti–TGF-β2 (5 μg/ml) or (B) anti–IL-10 (100 ng/ml) antibodies, or their appropriate isotype controls. After 72 hours, [³H]thymidine was added to assess proliferation. Results are presented as mean cpm ± SEM for triplicate samples.

Figure 7.

 

Bioassay of TGF-β in supernatants of I/CB–exposed T cells. BALB/c T cells were exposed to cultured I/CB PE cells for 48 hours. The T cells were then harvested and recultured (1 × 10⁵/well) in serum-free medium for 24 hours. Supernatants were collected and added directly (active) or acid-activated before being added (total) to a mink lung cell assay for TGF-β. In some assays, neutralizing anti–TGF-β antibodies were added. Results of triplicate samples are presented as mean ± SEM.

The authors thank Pascale Alard and Jacqueline Doherty for assistance and advice in conducting these experiments and Peter Mallen for assistance with the figures.