Effects of Experimental Ocular Inflammation on Ocular Immune Privilege

Abstract

purpose. To determine whether the inflammation of endotoxin-induced uveitis (EIU) and experimental autoimmune uveoretinitis (EAU) alters key in vivo and in vitro parameters of ocular immune privilege.

methods. For EIU induction, C3H/HeN mice received 200 μg lipopolysaccharide (LPS). For EAU induction, B10.A mice were immunized with 50 μg interphotoreceptor retinoid-binding protein (IRBP) mixed with complete Freund’s adjuvant. Aqueous humor (AqH) was collected at periodic intervals and assayed for leukocyte content and the ability to suppress or enhance T-cell proliferation. Eyes with EAU were assessed for the capacity to support anterior chamber (AC)-associated immune deviation (ACAID) induction after injection of ovalbumin (OVA).

results. Inflammation within the anterior segment in EIU peaked at 12 to 24 hours and was detected from 10 days onward in EAU. In AqH of EIU, protein content rose within 4 hours, followed by infiltrating leukocytes. EIU AqH promptly lost its capacity to suppress T-cell proliferation and became mitogenic for T cells. In AqH of EAU, protein and leukocyte content rose at 11 days and continued to remain elevated thereafter. Whereas 11-day EAU AqH failed to suppress T-cell proliferation, AqH at later time points reacquired immunosuppressive properties. Injection of OVA into the AC of eyes of mice with EAU failed to induce ACAID.

conclusions. The intraocular inflammation of EIU and EAU disrupted important parameters of immune privilege, ranging from breakdown of the blood–ocular barrier, to loss of an immunosuppressive microenvironment, to abrogation of ACAID. Because AqH from inflamed EAU reacquired the ability to suppress T-cell proliferation, the authors conclude that the capacity to regulate immune expression and inflammation can be a property even of inflamed eyes.

Immune privilege is a constitutive feature of the anterior chamber (AC) of the normal eye. ¹ In its strictest definition, immune privilege refers to the fact that foreign tissue grafts placed in the AC survive for prolonged, often indefinite, intervals, whereas placement of such grafts at conventional body sites leads to acute, irreversible immune rejection. As our understanding of the mechanisms responsible for immune privilege have expanded over the past 3 decades, the definition of immune privilege has been relaxed and now embraces, on the one hand, the induction of AC-associated immune deviation (ACAID) after intracameral injection of antigenic materials, ^{2 3 4} and on the other hand, the capacity of the ocular microenvironment (especially aqueous humor [AqH]) to suppress immune effector responses and inflammation. ^{5 6} Existence of ocular immune privilege is believed to serve the purpose of limiting the extent to which innate and adaptive immunity can cause intraocular inflammation. By limiting intraocular inflammation, immune privilege preserves the integrity of the visual axis and thereby prevents blindness. 

Ocular inflammation, whether expressed within the cornea or within the uveal tract, is a frequent cause of visual impairment. A variety of experimental models have been developed in laboratory animals as a means of studying the pathogenesis of ocular inflammation. ^{7 8 9 10} Yet, virtually nothing is known about the extent to which ocular inflammation interferes with ocular immune privilege. This may not be an idle concern because inflammation at other body sites can lead, directly or indirectly, to the development of autoimmunity, and as a consequence, local tissue damage is further exaggerated. ¹¹ For this reason, and because we wanted to understand the critical factors that contribute to the existence of ocular immune privilege, we examined the so-called immune privileged status of eyes of mice in which intraocular inflammation had been induced experimentally. Using the model systems of endotoxin-induced uveitis (EIU) and experimental autoimmune uveoretinitis (EAU), we found that inflamed eyes displayed weakened or broken blood–ocular barriers and that aqueous humor (AqH) from these eyes either lost or displayed altered immunosuppressive properties. Moreover, EAU-affected eyes no longer supported the induction of ACAID. The relevance of ocular inflammation to retention and maintenance of immune privilege in the AC of the eye is discussed. 

Materials and Methods

Animals

C3H/HeN (Taconic Farms, Germantown, NY),C3H/HeJ, and B10.A (Jackson Laboratory, Bar Harbor, ME) mice were purchased at 6 to 8 weeks of age. Normal BALB/c mice were obtained from our domestic, inbred mouse-breeding colony. All animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Antigens

Interphotoreceptor retinoid-binding protein (IRBP) was isolated from bovine retinas as described previously. ¹² Ovalbumin (OVA) was purchased from Sigma (St. Louis, MO). 

Induction of Uveitis

To induce endotoxin-induced uveitis (EIU), C3H/HeN mice received a footpad injection of 200 μg lipopolysaccharide (LPS) from Salmonella typhimurium (Difco, Detroit, MI) in 100 μl phosphate-buffered saline (PBS) solution. For induction of experimental autoimmune uveoretinitis (EAU), B10.A mice were immunized subcutaneously with 50μ g IRBP in 0.2 ml emulsion mixed 1:1 with complete Freund’s adjuvant (CFA; Difco) that had been supplemented with Mycobacterium tuberculosis to a final concentration of 2.5 mg/ml. Simultaneously, the mice were injected intraperitoneally with 500 ng pertussis toxin (Sigma) as an additional adjuvant. 

Assessment of Uveitis

Fundus examination of eyes of mice immunized with IRBP was performed every other day after immunization. These examinations were performed in masked fashion in which the observer was unaware of the nature of the prior experimental manipulations. According to the ocular findings, an arbitrary 4-point score was devised to provide semiquantitative evaluation ¹³ of the extent of inflammation and damage. Histopathologic examination of uveitic eyes was performed on methacrylate-embedded sections of eyes enucleated at selected times after immunization with IRBP. These sections were stained with hematoxylin and eosin and evaluated according to criteria of ocular inflammation described elsewhere. ⁸  

Aqueous Humor Collection and Analysis

LPS-induced EIU generates an acute intraocular inflammation that reaches peak intensity within 24 hours ^{14 15 16} and is largely dissipated by 48 hours. By contrast, EAU generates intraocular inflammation that develops over a more protracted course: The clinical expression is not uniformly evident until 11 days after immunization, and the inflammation usually persists beyond 28 days. ^{17 18} Therefore, in our experiments the sampling times for AqH were different in the two model systems. Aqueous humor was obtained from eyes of C3H/HeN and B10.A mice for in vitro analysis at 0, 2, 4, 6, 8, 12, 24, and 48 hours after endotoxin injection in EIU-affected mice, and on days 0, 11, 17, and 28 after IRBP immunization in EAU-affected mice. Aqueous humor was obtained immediately after death from both eyes, using a 30-gauge needle and 10 μl micropipets (Fisher Scientific, Pittsburgh, PA) by capillary attraction and pooled into a siliconized microcentrifuge tube (Fisher Scientific). Aqueous humor samples from panels of at least five mice (10 eyes) were pooled and centrifuged at 3000 rpm for 3 minutes, and the cell-free supernatant was frozen immediately at −70°C. On average, 6 μl AqH was obtained from the two eyes of each mouse. Leukocytes that were present in the pellet of centrifuged AqH were resuspended in medium, stained with 0.4% trypan blue solution, and counted by phase-contrast microscopy. The total protein content in AqH samples was measured using a protein assay reagent kit (BCA; Pierce, Rockford, IL) in reference to a bovine albumin standard. Endotoxin content of AqH was assessed with a limulus assay kit (Limulus Amebocyte Lysate [LAL]; Bio-Whittaker, Walkersville, MD). Briefly, AqH from eyes of mice with EIU was mixed with LAL and incubated at 37°C for 10 minutes. A substrate solution was then mixed with the LAL sample and incubated at 37°C for 6 minutes. The reaction was stopped with sodium dodecyl sulfate solution, and absorbance was determined spectrophotometrically at 405 nm. Each experiment with pooled AqH was repeated at least twice. In the repeat experiments, AqH was collected again for new groups of mice. 

Assay of T-Cell Proliferation

Spleens were removed from naive BALB/c or LPS-resistant C3H/HeJ mice and pressed through nylon mesh to produce single-cell suspensions. Red blood cells were lysed with Tris-NH₄Cl. T cells were subsequently purified by passage through a T-cell enrichment column (R&D Systems, Minneapolis, MN) according to the manufacturer’s directions. The enriched, naive T cells (>95% Thy+ cells, measured by flow cytometry) were suspended in serum-free medium. Serum-free medium was composed of RPMI 1640 medium, 10 mM HEPES, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin (all from Bio-Whittaker) and 1× 10⁻⁵ M 2-ME (Sigma) and supplemented with 0.1% bovine serum albumin (Sigma), ITS+ culture supplement (1 μg/ml iron-free transferrin, 10 ng/ml linoleic acid, 0.3 ng/ml Na₂Se, and 0.2 μg/ml Fe[NO₃]₃; Collaborative Biomedical Products, Bedford, MA). The proliferation assay used was a modification of one described previously. ¹⁹ To individual wells of a 96-well V-shape bottomed plate (Corning, Corning, NY), we added 10 μl 2.5 × 10⁴ enriched T cells, 10 μl hamster anti-mouse CD3e IgG (2C11, 2.5 μg/ml; PharMingen, San Diego, CA), or 10 μl serum-free medium, and 5 μl AqH or PBS. Total reaction volume was kept constant at 25 μl. The cells were pulsed with 2.5 μl 20 μCi/ml [³H]thymidine for the final 8 hours of the 72-hour incubation (37°C; 5% CO₂–95% humidified air mixture). On day 3, the cells were recovered using a cell harvester (model 96; Tomtec, Orange, CT), and [³H]thymidine incorporation was measured in counts per minute, using a liquid scintillation counter (Betaplate 1205; Wallac, Gaithersburg, MD). 

Induction and Assay of Delayed Hypersensitivity and ACAID

To induce delayed hypersensitivity, mice were immunized subcutaneously with 100 μg OVA emulsified 1:1 in CFA in a total volume 100 μl. To induce ACAID, OVA was injected (50 μg/3 μl PBS) into the AC of one eye of recipient mice. One week later these mice were immunized with OVA-CFA injected into the nape of neck. After seven days, OVA (200 μg/10 μl) was injected into the right ear pinna, and ear swelling responses were assessed 24 and 48 hours later using an engineer’s micrometer (Mitutoyo 227-101; MTI, Paramus, NJ). Ear swelling was expressed as follows: specific ear swelling = (24-hour measurement of right ear − 0-hour measurement of right ear) − (24-hour measurement of left ear − 0-hour measurement of left ear) × 10⁻³ mm. Ear swelling responses of groups of mice are presented as mean ± SEM. 

Statistical Evaluations

Significance of differences between mean values of ear swelling responses was evaluated using Student’s t-test. P < 0.05 was deemed significant. 

Results

Our experimental plan was to induce EIU and EAU in C3H/HeN and B10.A mice, respectively, and to examine the immunosuppressive properties of AqH and the capacity of inflamed eyes to support induction of ACAID. At selected times after the initiating injection, AqH was collected from eyes of panels of mice (at least five animals per time point), pooled, and assayed for protein content, content of leukocytes, and capacity to inhibit anti-CD3–driven T-cell proliferation. Because the ocular inflammation induced in mice by footpad injection of endotoxin and by IRBP immunization varies somewhat from laboratory to laboratory, it was necessary for us to describe the clinical and histologic features of EIU and EAU in the animals from which we collected AqH and in which we attempted to induce ACAID. 

Clinical Assessment of Ocular Inflammation in Experimental Mice

Few signs of anterior segment inflammation were observed in eyes of C3H/HeN mice that received endotoxin by footpad, although the corneal surface was sometimes covered with a discharge. In B10.A mice with EAU, the first sign of ocular inflammation was hyperemia of limbal vessels, which was detected at day 11. Evidence of inflammation in the anterior ocular segment was minimal. In a few instances, posterior synechiae were observed between days 14 and 20. However, marked aqueous flare or definite evidence of leukocytes in the AC were not observed. EAU is primarily a posterior uveitis. In immunized mice, the frequency of clinically detectable posterior inflammation approached 80% (Fig. 1) . After day 11, signs of retinal vasculitis developed in increasing numbers of eyes. In the more severe cases, retinal or subretinal exudates, retinal hemorrhages, and/or disc edema were observed. Retinal detachments were often severe and were detected between days 14 and 22 in severely affected eyes. 

Histopathologic Changes in Mice with Experimental Intraocular Inflammation

Eyes of mice afflicted with EIU and EAU were enucleated periodically during the course of these diseases and subjected to histologic analysis. Concerning EIU, within 6 hours of footpad injection of endotoxin, the AC of recipient eyes contained an eosin-staining amorphous substance. Rare leukocytes were detected on the surface of ciliary body epithelium and iris at this time, and vessels within the stromae of ciliary body and iris were dilated. By 12 hours, these changes were exaggerated, and significant numbers of leukocytes were now attached to the corneal endothelium and to the epithelial surfaces of iris and ciliary body. Vasodilatation was particularly intense at 24 hours after LPS injection, and leukocytic cells were found to be adjacent to dilated vessels in iris and ciliary body. A small number of leukocytes were also found in the posterior vitreous body. Otherwise, the posterior compartments of the eyes of endotoxin-treated C3H/HeN mice was unchanged from those of untreated C3H/HeN mice. 

Regarding EAU, by approximately 11 days after immunization with IRBP, eyes removed from B10.A mice displayed dilatation of the choroidal and retinal vessels. By 17 days after immunization, numerous leukocytes were observed in the vitreous, within the neuronal retina, in the subretinal space, and even in the AC. An intense accumulation of leukocytes was often observed in the angle between the iris and the corneal endothelium. At this time, more than 40% of the eyes displayed moderate to severe retinal detachments. Intraocular evidence of intense inflammation persisted in eyes removed for study 28 days after immunization. 

Protein and Leukocyte Content of Aqueous Humor from Eyes with EIU

Aqueous humor was removed from eyes of C3H/HeN mice with EIU at selected times and separated by centrifugation into a cellular fraction (pellet) and a soluble fraction (supernatant). The samples were evaluated for protein concentration and content of leukocytes (Fig. 2) . As anticipated, no leukocytes were detected in control AqH samples. Moreover, no leukocytes were found in samples harvested at 2, 4, or 6 hours after LPS injection. At 8 hours and thereafter, leukocytes (chiefly polymorphonuclear neutrophils) were detected, and peak concentration of these cells was reached at 12 hours (70.3 ± 5.5 cells/μl). Considerably reduced numbers of these cells were detected in AqH removed at 24 and 48 hours after endotoxin injection. The protein concentration in control AqH was low (1.8 ± 0.1 mg/ml). It remained at this low level in samples of AqH obtained at 2 hours after LPS footpad injection, but a significant increase in protein concentration was detected at 4 hours, and the level continued to increase at 6, 8, 12, and 24 hours. At the latter time point, protein concentration in AqH was 5.8 mg/ml. Although the peak protein concentration differed among experiments, it occurred at either 12 or 24 hours. Protein concentration in AqH returned almost to baseline by 48 hours. If elevated protein levels in AqH is taken as evidence of breakdown of the blood–ocular barrier, these results indicate that endotoxin injection into the footpad induced a leak through this barrier within 2 to 4 hours. These results indicate that endotoxin-dependent breakdown of the blood–ocular barrier, measured by protein concentration, precedes intrusion of leukocytes into the AC. This result could mean that leukocyte infiltration into the AC is a secondary consequence of the action of LPS, and/or it could mean that LPS (directly or indirectly) alters ocular microvessels, rendering them permissive for leukocyte immigration. 

Protein and Leukocyte Content of Aqueous Humor from Eyes with EAU

Aqueous humor samples were collected at selected times from eyes of mice that received a uveitogenic regimen of IRBP and were analyzed in a manner similar to that described for AqH from EIU-affected eyes. At the earliest time (11 days) that AqH was harvested from eyes of immunized B10.A mice, protein concentration was found to be elevated. The levels of protein in AqH gradually decreased over the next 17 days, although the concentration on day 28 was still significantly higher than that of control AqH. We interpret these findings to mean that the blood–ocular barrier was broken within 11 days of immunization and that partial restoration of the barrier was eventually achieved during the remainder of the observation period. Although the antigenic target of the autoimmune attack in eyes with EAU was in the retina, leukocytes penetrated into the AqH within 11 days of immunization with IRBP (Fig. 3) . Unlike the changes in protein concentration, the density of leukocytes in AqH continued to rise through days 17 and 28 after immunization. These findings indicate that protein concentration and leukocyte density are not identical markers of a broken blood–ocular barrier. Moreover, as we expected, the findings show that an intense inflammatory reaction that is focused in the posterior compartment of the eye (i.e., the retina) spills over into the anterior segment, modifying the AqH content of protein and leukocytes. 

Effect of Aqueous Humor from EIU-Inflamed Eyes on T-cell Proliferation

Normal T cells can be profoundly activated by exposure to the monoclonal antibody 2C11, which recognizes an epitope on the CD3 complex associated with the T-cell receptor for antigen. T cells activated by anti-CD3 antibodies are induced to undergo proliferation. Because it has been shown that AqH from normal mouse eyes inhibits T-cell proliferation induced through the T-cell receptor for antigen, we next examined the potential T-cell–inhibiting properties of AqH from inflamed mouse eyes. In these experiments, BALB/c T cells were cultured with anti-CD3 antibodies in the presence of medium containing 20% AqH. In control cultures, PBS was substituted for AqH. After 72 hours’ incubation, T-cell proliferation was assessed by measuring the uptake of [³H]thymidine. Aqueous humor samples from different times after induction of EIU and EAU were tested in this assay. The results of one set of representative experiments are presented in Figure 4 . Aqueous humor from eyes of normal C3H/HeN mice (0 hours) suppressed T-cell proliferation completely (112.7 ± 61.1 cpm) when compared with radioisotope incorporation by T cells in the absence of AqH (3384.4 ± 588.8 cpm). Using AqH samples from eyes of mice with EIU, the capacity to suppress T-cell proliferation was found to be reduced within 2 hours of LPS injection and was virtually gone by 4 hours. Loss of capacity to suppress T-cell proliferation also characterized AqH samples collected at 6, 8, 12, and 24 hours after LPS injection. Only at 48 hours did AqH from eyes of LPS-injected mice recover the ability to inhibit T-cell activation. These results indicate that EIU not only causes a break in the blood–ocular barrier, but it alters the intraocular microenvironment by eliminating, at least transiently, its ability to suppress T-cell activation. 

T-cell proliferation in cultures to which AqH collected 6 hours after LPS injection had been added actually exceeded that of positive controls. Although this difference was small, it suggests the possibility that AqH from EIU-inflamed eyes contain T-cell mitogenic factors. Because EIU is triggered by an injection of LPS, and because LPS is mitogenic for lymphocytes, the possibility exists that the putative mitogenic activity of EIU-AqH is caused by LPS. To examine this possibility, we tested AqH samples with the limulus assay for the presence of LPS. All AqH samples, beginning 2 hours after LPS injection, contained detectable levels of LPS (Fig. 5) . In two of three experiments, peak levels of LPS were sustained in AqH harvested 8, 12, and 24 hours after endotoxin injection into the footpad. To determine whether the mitogenic activity in EIU-AqH samples was caused by LPS, T cells were harvested from C3H/HeJ mice. These cells were selected because this strain of mice is genetically resistant to LPS, including the mitogenic effect of LPS on lymphocytes. Cultures were established with C3H/HeJ T cells to which anti-CD3 antibodies were or were not added. Aqueous humor samples from eyes of mice with EIU were then added and T-cell proliferation assessed after 72 hours. The results of this experiment are presented in Figure 6 . In the absence of anti-CD3 antibodies, C3H/HeJ T cells failed to proliferate, in either the presence or the absence of normal (0-hour) AqH. However, AqH obtained 4 hours after LPS injection into the footpad stimulated C3H/HeJ T-cell proliferation in the absence of anti-CD3 antibodies. Mitogenic activity of this type was also detected with AqH samples collected at 6, 8, 12, 24, and 48 hours after LPS injection into the footpad. Because C3H/HeJ T cells are genetically unable to respond to LPS, the mitogenic activity found in AqH of eyes from mice with EIU must arise from unknown factors generated by LPS, but not from the action of LPS itself. 

Effect of Aqueous Humor from EAU-Inflamed Eyes on T-Cell Proliferation

Similar studies were conducted with AqH samples from eyes of B10.A mice with EAU. The results of this experiment are presented in Figure 7 . As anticipated, AqH from normal B10.A mice suppressed T-cell proliferation induced by anti-CD3 antibodies (234.4 ± 105.4 cpm versus 1870.6 ± 194.2 cpm). By contrast, AqH obtained from eyes of mice immunized with IRBP 11 days previously completely failed to inhibit T-cell proliferation. T-cell–inhibiting activity returned to AqH samples collected from mice immunized 17 days previously, and AqH harvested at 28 days continued to suppress T-cell activation. These findings are surprising, because eyes with EAU at 17 and 28 days display evidence of intense inflammation and destruction. Aqueous humor samples from mice with EAU were also tested for their mitogenic properties, using BALB/c T cells as responders. As the results presented in Figure 8 indicate, only AqH samples collected 11 days after immunization induced T-cell proliferation in the absence of anti-CD3 antibodies. In all other AqH samples from eyes with EAU, no mitogenic activity was detected. Thus, EAU caused a transient loss of immunosuppressive activity from AqH that was restored even though inflammation continued to damage the posterior uveal tract. 

Capacity of Inflamed Eyes to Support ACAID Induction

The ability of an eye to support ACAID induction is thought to be an important dimension to ocular immune privilege. Therefore, we attempted to assess the capacity of OVA injected into eyes affected with EIU and EAU to induce ACAID. To determine that ACAID has been induced, positive control mice must display intense, antigen-specific delayed hypersensitivity. Unfortunately, mice that received a footpad injection of LPS were rendered refractory to delayed hypersensitivity induction, an observation reported by others. ²⁰ For this reason, we were unable to determine whether OVA could induce ACAID in eyes of mice with EIU. We were, however, able to assess this response in B10.A mice with EAU. We selected day 14 after IRBP immunization to assess ACAID, primarily because our previous results had indicated that this time point corresponds to the interval during which the blood–ocular barrier is broken, and EAU is rapidly developing within the retina. Accordingly, one panel of B10.A mice that first received 50μ g of IRBP in CFA with 500 ng pertussis toxin received an injection of 50 μg OVA into the AC of one eye 14 days later. A second panel similarly immunized with IRBP received no AC injection of OVA, and these mice served as the EAU-positive control. An ACAID control panel of normal B10.A mice received 50 μg OVA in the AC of one eye. One week later, these panels, along with normal B10.A mice (positive controls) received a subcutaneous immunization with OVA (100 μg) plus CFA. The ear pinnae of all mice were challenged with 200 μg OVA 1 week later, and the ear swelling responses assessed 24 and 48 hours later. The results are presented in Figure 9 . IRBP-immune B10.A mice that were immunized with OVA/CFA displayed intense delayed hypersensitivity responses, greater in fact than the responses of the positive controls. IRBP-immune B10.A mice that received an AC injection of OVA before immunization with OVA-CFA also showed intense ear swelling responses that were even greater than those of either of the positive controls. By contrast, B10.A mice that were not immunized with IRBP but received an AC injection of OVA before immunization with OVA-CFA mounted feeble DH responses, indicating that they had ACAID. These results show that eyes undergoing the inflammatory insult of EAU lose the ability to support ACAID, a further indication of their loss of immune privilege. 

Discussion

Potent immune regulatory forces are operative in the normal eye. ^{5 6 21 22} The existence of immune privilege in the eye offers experimental verification that these forces exist. The ability of the eye to regulate immunity is twofold. On the one hand, the eye creates a microenvironment that insures that antigens that are injected into, or arise within, the AC induce a deviant systemic immune response termed ACAID, in which T cells and antibodies that evoke immunogenic inflammation are suppressed. ^{2 3 4} On the other hand, the ocular microenvironment is inhospitable to the expression of those forms of immunity that use nonspecific inflammation in performing their effector function. ^{5 6} By shaping the immune response to eye-derived antigens at both the induction and expression stages, the eye is relatively protected from the vision-damaging effects of intraocular inflammation. The term “relatively” is important in the preceding sentence because the eye can, and does, experience inflammation. In clinical ophthalmology, acute and chronic uveitis are common afflictions that all too often lead to visual impairment and blindness, and in the laboratory intraocular inflammation can be evoked in informative model systems. ^{9 10} The experiments reported here have taken advantage of two different model systems with which to explore the extent to which intraocular immune regulatory mechanisms are compromised in inflamed eyes. 

By harvesting AqH at periodic intervals after induction of EIU and EAU in mice we have been able to identify the times at which the blood–ocular barrier is broken and to examine the intervals during which the ability of this fluid to inhibit T-cell activation is abrogated. Our results indicate that leakage of plasma proteins into AqH occurred early in the course of both EIU and EAU—in fact, before the first clinical evidence of disease was apparent. Moreover, in both types of ocular inflammation, the AqH promptly lost its capacity to suppress T-cell proliferation. Four factors could contribute to the loss of immunosuppressive capacity of AqH from inflamed eyes. First, plasma proteins display protease activity, and many of the immunosuppressive factors in AqH (e.g., α-MSH, VIP, CGRP) ^{19 23 24 25} are neuropeptides with exquisite vulnerability to enzymatic degradation. Thus, the mere entry of plasma proteins into AqH may cause the depletion of factors that suppress T-cell proliferation. Second, shortly after plasma proteins leaked into AqH, leukocytes also penetrated into the intraocular microenvironment. It is possible that activated neutrophils or macrophages could take up and/or destroy factors in AqH that confer on the fluid its immunosuppressive properties. Third, the rapid turnover of AqH in the eye applies to its content of immunosuppressive factors. It is possible that endotoxin in EIU, or the autoimmune attack directed at IRBP in EAU, halts the intraocular production of immunosuppressive factors (e.g., TGFβ-2) ^{26 27} and that the new AqH that is formed in inflamed eyes is deficient in these immunosuppressive factors. Fourth, AqH obtained from eyes of mice with EIU and EAU (to a more limited extent) was found to stimulate T-cell proliferation in vitro, in the absence of the addition to the culture of a T-cell mitogen. Limulus assay showed that AqH from eyes of mice with EIU contained significant amounts of LPS, yet this did not appear to be responsible for the mitogenic activity. By using T cells from LPS-resistant C3H/HeJ mice, we determined that AqH from EIU eyes still induced proliferation. Therefore, the abnormal microenvironment associated with inflamed eyes may actually promote, rather than inhibit, T-cell activation. At present, we have no information concerning the nature of the proliferation-inducing property discovered in AqH obtained during intraocular inflammation. 

In AqH obtained at periodic intervals from eyes with EIU, a relatively good correlation was observed when the amount and timing of protein concentration were compared with the AqH’s loss of immunosuppressive activity. Maximal loss of AqH ability to inhibit T-cell proliferation (4–24 hours, with peak at 6 hours after LPS footpad injection) corresponded reasonably well to maximum protein concentration (4–24 hours, with peak at 12–24 hours). A similarly good correlation was not observed in periodic AqH samples obtained from eyes of mice with EAU. High levels of protein were observed in AqH from EAU-affected eyes at days 11 and 17, but only AqH from day 11 displayed a significant loss of immunosuppressive activity. Aqueous humor removed from eyes with EAU on day 17 inhibited T-cell proliferation in vitro as profoundly as did normal AqH. No better correlation could be made between loss of immunosuppressive capacity and the number of leukocytes found in AqH of mice with EAU. Although AqH in EAU lost its T-cell proliferation–inhibiting capacity coincident with the breakdown of the blood–ocular barrier, continued loss of barrier function did not prevent these eyes from restoring their immunosuppressive microenvironment, at least in the AC. We are eager to identify the mechanism responsible for reestablishment of immunosuppression in these inflamed eyes. 

In the aggregate, the in vitro studies of AqH from EIU- and EAU-inflamed eyes confirm that experimentally induced intraocular inflammation abrogates (partially to completely) one important dimension of immune privilege—that is, the ability to suppress T-cell–dependent immunity within the eye. In addition, our experiments attempting to induce ACAID by injecting OVA into the AC of eyes of mice with EAU indicate that intraocular inflammation also interferes with another important dimension of the privileged state—that is, the capacity of the eye to shape the nature of systemic immune responses to eye-derived antigens. B10.A mice that received a uveitogenic IRBP regimen 14 days previously and that showed development of intense posterior uveitis, did not acquire OVA-specific ACAID when OVA was injected into the AC. This finding implies that intraocular inflammation, even if focused on the posterior segment, robs the AC of its capacity to support ACAID induction. More important, the loss of the ability to promote ACAID extends to antigens unrelated to the autoantigens that are the targets of the autoimmune disease. Whereas in our experiments the unrelated exogenous antigen was OVA , a nonpathogenic molecule, the threat posed by intraocular inflammation is that ACAID may not develop when it is needed—for example, when a corneal allograft is placed orthotopically. Experimental evidence in mice indicates that ACAID contributes to the long-term success of keratoplasties. ^{28 29 30} We suspect that if we can learn how to restore the capacity of an inflamed eye to promote ACAID, we may generate therapeutic strategies that will ensure engraftment of allogeneic corneas and perhaps mitigate the destructive potential of intraocular infection with herpes viruses and other pathogens. 

 Reprint requests: J. Wayne Streilein, Schepens Eye Research Institute, 20 Staniford Street, Boston, MA 02114.

 Supported by Grant EY05678 from the National Institutes of Health.

 Submitted for publication November 30, 1998; revised February 2, 1999; accepted April 9, 1999.

 Proprietary interest category: N.

 

Figure 1.

 

Clinical course of EAU. In our semiquantitative evaluation by funduscopy, one point each was awarded for retinal vascular dilatation, retinal vasculitis, retinal or subretinal exudates or bleeding, disc edema, or retinal detachment. Incidence (left) and uveitis score (right) are representative data from four separate experiments. Uveitis score was expressed as the mean of five mice.

Figure 2.

 

Protein concentration (A) and cell infiltration (B) in AqH from EIU mice after LPS injection. Values represent mean ± SE. Results are representative of data from four separate experiments.

Figure 3.

 

Protein concentration (A) and cell infiltration (B) in AqH from mice with EAU after IRBP immunization. Values represent mean ± SE. Results are representative data from three separate experiments.

Figure 4.

 

Effect of AqH from eyes with EIU on anti CD3-dependent T-cell proliferation. Enriched naive BALB/c T cells (2.5 × 10⁴ cells/well) were stimulated with anti-CD3 antibodies in the medium alone (M) or medium containing 20% PBS or 20% AqH obtained at each time point. After a 72-hour incubation, T-cell proliferation was assessed by measuring the uptake of [³H]thymidine. Each data point represents the mean ± SE of triplicate cultures. Similar results were obtained in five additional experiments.

Figure 5.

 

LPS concentration in AqH of EIU-affected mice. The data are representative of three separate experiments.

Figure 6.

 

Effect of AqH from eyes of mice with EIU on LPS-resistant T-cell proliferation. Enriched naive C3H/HeJ T cells (2.5 × 10⁴ cells/well) were stimulated with or without anti-CD3 antibodies in the presence or absence of 20% AqH (M). After a 72-hour incubation, T-cell proliferation was assessed by measuring the uptake of [³H]thymidine. Each data point represents the mean ± SE of triplicate cultures. Similar results were obtained in two additional experiments.

Figure 7.

 

Effect of AqH from eyes with EAU on anti-CD3–dependent T-cell proliferation. Enriched naive BALB/c T cells (2.5 × 10⁴ cells/well) were stimulated with anti-CD3 antibodies in the medium alone (M) or in medium containing 20% PBS or 20% AqH obtained at each time point. After a 72-hour incubation, T-cell proliferation was assessed by measuring the uptake of[ ³H]thymidine. Each data point represents the mean ± SE of triplicate cultures. Similar results were obtained in three additional experiments.

Figure 8.

 

Effect of AqH from eyes with EAU on T-cell mitogenesis. Enriched naive BALB/c T cells (2.5 × 10⁴ cells/well) were cultured in the absence of anti-CD3 antibodies with medium alone (M) or with medium containing 20% PBS or 20% AqH obtained at each time point. After a 72-hour incubation, T-cell proliferation was assessed by measuring the uptake of [³H]thymidine. Each data point represents the mean ± SE of triplicate cultures. Similar results were obtained in two additional experiments.

Figure 9.

 

ACAID in eyes from EAU-affected mice. Delayed-type hypersensitivity measurement after intraocular injection of OVA with or without preimmunization with IRBP. Ear-swelling analysis was performed 24 hours (data not shown) and at 48 hours after ear challenge with OVA in mice immunized subcutaneously with OVA in CFA 7 days earlier. One week before this immunization, OVA was injected into the AC. Two groups of animals had received IRBP immunization 14 days earlier. In 80% of these mice, clinical EAU developed after the AC inoculation. Negative control mice (Negative) were not immunized. Both groups of positive control mice (Positive) were immunized only, without AC inoculation.* Mean values significantly less than positive control group without IRBP immunization.

The authors thank Jacqueline M. Doherty for her valuable help during the preparation of the manuscript.