IL-6 Antagonizes TGF-β and Abolishes Immune Privilege in Eyes with Endotoxin-Induced Uveitis

Abstract

purpose. To determine the immunosuppressive status of aqueous humor (AqH) from mouse eyes afflicted with endotoxin-induced uveitis (EIU) and to identify the relevant cytokines responsible for immunomodulatory activity within EIU AqH.

methods. Bacterial lipopolysaccharide (LPS) was injected into hind footpads of C3H/HeN mice; and AqH, collected at 6, 12, 24, and 48 hours, was evaluated for content of transforming growth factor (TGF)-β, tumor necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6, and interferon (IFN)-γ and capacity to suppress anti-CD3–driven T-cell proliferation. Cytokine mRNA expression in iris–ciliary body (I/CB) was analyzed by RNase protection assays.

results. During 6 to 24 hours after LPS injection, total TGF-β levels in AqH increased even though the fluid lost its capacity to suppress T-cell activation. At this time, AqH contained high levels of IL-6, and I/CB contained high levels of IL-6 mRNA. When IL-6 was neutralized with specific antibodies, inflamed AqH reacquired its capacity to suppress T-cell activation, which correlated with high levels of TGF-β. Coinjection of IL-6 plus antigen into the anterior chamber of the eye of normal mice prevented antigen-specific anterior chamber–associated immune deviation (ACAID).

conclusions. LPS-induced intraocular inflammation is associated with local production of IL-6, which robs AqH of its immunosuppressive activity, perhaps by antagonizing TGF-β. The fact that IL-6 antagonized ACAID induction in normal eyes suggests that strategies to suppress the intraocular synthesis of IL-6 may reduce inflammation and restore ocular immune privilege.

The capacity of the ocular microenvironment (especially the aqueous humor, AqH) to suppress immune effector responses and inflammation ^{1 2 3} is believed to be an important dimension of ocular immune privilege. Normal AqH contains immunosuppressive and modulatory factors, such as transforming growth factor (TGF)-β2,α -melanocyte–stimulating hormone (α-MSH), vasoactive intestinal peptide (VIP), calcitonin gene–related peptide (CGRP), and macrophage migration inhibitory factor (MIF). ^{3 4 5 6 7 8} Normal AqH has been shown to inhibit T-cell activation, leading to proliferation and cytokine production. ^{2 4 5} Among the factors found in AqH, TGF-β2 has been thought to be the most important agent responsible for inhibiting T-cell responses in vitro. TGF-β is produced locally within the eye, and explants if iris–ciliary body (I/CB) has been found to produce immunosuppressive factors, including TGF-β. ^{9 10 11 12}  

Ocular immune privilege is believed to serve the purpose of limiting the extent to which innate and adaptive immune responses lead to 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. ^{13 14} Yet, virtually nothing is known about the extent to which ocular inflammation interferes with ocular immune privilege. For this reason, and because we wish 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. 

In this study, we used one popular murine model of ocular inflammation called endotoxin-induced uveitis (EIU). EIU is generated by injecting lipopolysaccharide (LPS) into the footpads of susceptible mice and rats. Shortly after injection, an acute inflammatory response emerges in the anterior segment of the eyes. ¹³ EIU is considered to represent a number of sight-threatening human inflammatory eye diseases, such as Behçet’s disease, Crohn’s disease, Reiter’s disease, and ulcerative colitis. A variety of inflammatory factors have been suspected of contributing to the development of EIU. Upregulation of tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6 mRNAs within ocular tissues during EIU has been reported, ^{15 16 17} and these cytokines have been detected directly in AqH of rat eyes displaying EIU. ^{18 19}  

Although de Boer et al. ²⁰ reported that TGF-β levels were decreased in AqH obtained from human eyes with uveitis, no consensus exists concerning changes in TGF-β levels in the AqH of rodent eyes with experimental uveitis. ^{15 21 22} More important, no data are available concerning the extent to which TGF-β in the AqH from eyes of mice with uveitis is active or latent. In normal AqH, the vast majority of TGF-β is latent. ^{4 5} Unless activated, latent TGF-β has little if any immunosuppressive properties. Thus, we are interested not only in whether the absolute levels of TGF-β in AqH change during acute intraocular inflammation but whether the TGF-β present is in its active or latent form. 

Using the model system of EIU, we previously reported that when the blood–ocular barrier was breached, AqH lost its immunosuppressive properties and acquired the novel capacity to stimulate T-cell proliferation in vitro. ²³ In the current experiment, breakdown of the blood–ocular barrier has been correlated with enhanced levels of TGF-β in AqH and, more important, with the presence of IL-6, which appeared to be produced locally. When this cytokine was neutralized with specific antibodies, the underlying immunosuppressive properties of AqH (probably due to active TGF-β) were revealed. Finally, coinjection of IL-6 with antigen into the anterior chamber of normal eyes prevented the induction of anterior chamber–associated immune deviation (ACAID), implying that IL-6 is a major threat to ocular immune privilege. 

Methods

Mice

Female C3H/HeN (Taconic, Germantown, NY) 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 on the Use of Animals in Ophthalmic and Vision Research. 

Induction of Uveitis

To generate EIU, C3H/HeN mice received a footpad injection of 200μ g of LPS from Salmonella typhimurium (Difco, Detroit, MI) in 100 μl of phosphate-buffered saline solution (PBS). 

Aqueous Humor Collection and Analysis

AqH was obtained from eyes of C3H/HeN mice for in vitro analysis at 0, 6, 12, 24, and 48 hours after LPS injection. AqH was obtained immediately from both eyes after rats were euthanatized, using a 30-gauge needle and 10-μl micropipets (Fisher Scientific, Pittsburgh, PA) by capillary attraction, and multiple samples were pooled into a siliconized microcentrifuge tube (Fisher Scientific). Pooled AqH samples from panels of at least five mice (10 eyes) were centrifuged at 3000 rpm for 3 minutes, and the cell-free supernatant was frozen immediately at −70°C. On average, a total of 6 μl of AqH was obtained from the two eyes of each mouse. Every experiment was repeated at least three times with similar results. 

Determination of Cytokine Production

TGF-β1 and -β2 levels in stocked AqH were assessed with a commercially available human enzyme-linked immunosorbent assay (ELISA) kit (Promega, Madison, WI). This immunoassay detects biologically active TGF-β2. Cross-reactivity with other TGF-β isoforms is less than or equal to 5%, according to manufacturer’s manual. Interferon (IFN)-γ and IL-2 levels in AqH were also measured using anti-mouse monoclonal antibody (mAb) pairs: Rat IgG1, 18181D and IgG1, 18112D and Rat IgG2a, 18161D and IgG2b, 18172D, respectively (PharMingen, San Diego, CA). IL- β, IL-6, and TNF-α were also estimated using an ELISA kit from R&D Systems (Minneapolis, MN) according to the manufacturer’s instructions. 

TGF-β Bioassay

To measure total TGF-β, AqH was added to Mv1Lu cells (CCL-64; American Type Culture Collection, Rockville, MD) as described previously. ²² In brief, 1 × 10⁵ Mv1Lu cells in 200 μl with AqH diluted with Eagle’s Minimum Essential Medium (EMEM; BioWhitaker, Walkersville, MD) were incubated for 20 hours at 37°C, 5% CO₂. To each well, 20 μl of 50 Ci/ml ³H-thymidine (New England Nuclear–DuPont) was added, and the plate was incubated for an additional 4 hours. After incubation, the media was discarded and 50 μl of 10× trypsin–EDTA (BioWhitaker) solution was added to each well, then the plate was incubated for 15 minutes at 37°C. The cells were recovered using a Harvester 96 (Tomtec, Orange, CT), and[ ³H]thymidine incorporation was measured in counts per minute (cpm), using a 1205 Betaplate Liquid Scintillation Counter (Wallac, Gaithersburg, MD). Cultures of known amounts of pure TGF-β1 (R&D Systems) were prepared in the same plates as the assayed samples. A standard curve of TGF-β concentration (20 ng/ml to 2 pg/ml) versus counts per minute was used to measure TGF-β in AqH from eyes with EIU. Each 15 μl of AqH was diluted to 100 μl with assay medium. To convert all TGF-β from the latent to active form, 1N HCl was added to these samples. After incubation for 1 hour at 4°C, the acid was neutralized with a 1:1 mixture of 1N NaOH/1 M HEPES. 

Assay of T-Cell Proliferation

Spleens were removed from naive BALB/c 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) according to manufacturer’s directions. The enriched, naive T cells (>95% Thy 1⁺ cells as 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 BioWhitaker) and 1 × 10^{− 5} M 2-ME (Sigma, St. Louis, MO) and supplemented with 0.1% bovine serum albumin (Sigma), insulin, transferrin, selenium (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. ^{7 24} To individual wells of a 96-well V-shaped bottom plate (Corning, Corning, NY) we added 2.0 to 2.5 × 10⁴ enriched T cells, hamster anti-mouse CD3e IgG (2C11; final concentration is 1.0μ g/ml; PharMingen, San Diego, CA), and 5 μl of AqH or PBS as 20% vol/vol. Total reaction volume was kept constant at 25 μl. The cells were pulsed with 2.5 μl of 20 μCi/ml[ ³H]thymidine for the final 8 hours of the 48-hour incubation (37°C, 5% CO₂/95% humidified air mixture). Then, the cells were recovered and[ ³H]thymidine incorporation was measured in counts per minute. Each sample was cultured in triplicate. In some assays, samples of AqH were neutralized with Ab against TGF-β2 (R&D Systems) or control polyclonal IgG (ICN Biochemicals, Lisle, IL), or anti-murine IL-6 (PharMingen) or control monoclonal IgG Ab (PharMingen). To study the potential interacting effects of TGF-β and proinflammatory cytokines, serially diluted cytokine recombinant porcine TGF-β2, murine IL-6, murine IL-1β, murine TNF-α, and murine IFN-γ (R&D Systems) were added to the T-cell proliferation assay instead of AqH. All proliferation experiments were performed at least three times with similar results. 

RNA Preparation and RNase Protection Assay

Total RNA was extracted by the single-step method using RNA-STAT-60 (Tel-Test, Friendswood, TX). I/CBs were dissected from eyes, homogenized, and centrifuged to remove cellular debris. The RNA pellet obtained from 20 eyes was resuspended in nuclease-free water and processed together as a group. Detection and quantification of murine cytokine mRNAs were accomplished with a multiprobe RNase protection assay system (PharMingen) as recommended by the supplier. Briefly, a mixture of [(-32P] UTP-labeled antisense riboprobes was generated from the mCK-3b Multi-Probe Template Set (PharMingen). This set contains anti-sense RNA probes that can hybridize with target mouse mRNAs encoding TNF-β, lymphotoxin (LT)-β, TNF-α, IL-6, IFN-γ, IFN-β, TGF-β1, TGF-β2, TGF-β3, and MIF as well as two housekeeping gene products, L32 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Five μg of total RNA was used in each sample. Total RNA was hybridized overnight at 56°C with 300 pg of the ³²P–anti-sense riboprobe mixture. Nuclease-protected RNA fragments were purified by ethanol precipitation. After purification, the samples were resolved on 5% polyacrylamide sequencing gels. The gels were dried and subjected to autoradiography. Protected bands were observed after exposure of gels to x-ray film. Specific bands were identified on the basis of their individual migration patterns in comparison with the undigested probes. The bands were quantitated by densitometric analysis (NIH Image) and were normalized to GAPDH. 

Induction and Assay of Delayed Hypersensitivity and ACAID

To induce delayed hypersensitivity, mice were immunized subcutaneously with 100 μg of ovalbumin (OVA) emulsified 1:1 in complete Freund’s adjuvant (CFA) in a total volume of 100 μl. To induce ACAID, OVA was injected (50 μg/3 μl PBS) into the AC of one eye of recipient mice. To examine the effects of IL-6 in ACAID, OVA was mixed with 20 ng/ml of recombinant murine IL-6 and then injected intracamerally. One week later these mice were immunized with OVA/CFA into the nape of neck. After 7 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; MTI Corporation, 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

Data were subjected to analysis by ANOVA and the Scheffé test. A value of P < 0.05 was deemed to be significant. 

Results

Evidence for Breakdown of Blood–Ocular Barrier during Inflammation

According to previous reports ^{16 19} and our previous studies, ²³ we selected 6, 12, 24, and 48 hours after LPS administration as sampling times for AqH. As an indirect assessment of breakdown of the blood–ocular barrier, the numbers of infiltrating leukocytes and the protein levels within AqH samples collected at various time points were assessed. In brief, protein first appeared at 6 hours postinjection, reaching peak levels at 12 to 24 hours. A temporally commensurate appearance and rise of leukocytes were also observed. The peaks of protein concentration were 12 or 24 hours after LPS administration. Elevated protein concentrations, along with small numbers of leukocytes, were still evident at 48 hours. These data suggest that the blood–ocular barrier is breached between 0 and 6 hours and begins to reform after 24 hours. 

Eyes of mice afflicted with EIU were also enucleated periodically during the course of their disease and subjected to histologic analysis. 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 attached to the corneal endothelium, and to the epithelial surfaces of I/CB. Vasodilatation was particularly intense at 24 hours post-LPS injection, and leukocytes were found to be adjacent to dilated vessels in I/CB. A small number of leukocytes was also found in the posterior vitreous body. Otherwise, the posterior compartment of the eyes of endotoxin-treated C3H/HeN mice was unchanged from that of untreated C3H/HeN mice. 

TGF-β2 Levels in AqH from Inflamed Eyes and Its Influence on Capacity to Suppress T-Cell Activation

Samples from normal and inflamed AqHs were collected as described above, acid-activated, and assayed for total TGF-β (latent plus active forms) concentration using the mink lung epithelial cell bioassay. This assay measures all isoforms of TGF-β1, -β2, and -β3. As revealed in Figure 1 , total TGF-β levels increased by 6 hours after LPS injection and peaked at 12 and 24 hours. Thus, development of intraocular inflammation in EIU correlated with a rise, not a fall, in total TGF-β levels in the AqH. 

TGF-β2 is the predominant isoform in normal AqH, whereas TGF-β1 is the dominant isoform in blood plasma. In an effort to determine to what extent the rise in total TGF-β levels (latent plus active forms) in inflamed AqH was due to intraocular production as opposed to leakage of plasma proteins into the AqH after breakage of the blood–ocular barrier, we assayed AqH samples by ELISA for content of TGF-β1 and TGF-β2 separately. The results of this experiment, in which peak inflammation in the AC was reached at 24 rather than 12 hours post–LPS injection, are presented in Figure 2 . As expected, normal AqH displayed little or no immunoreactive TGF-β1 because the level of active TGF-β in this fluid is measured in picograms (which is below the sensitivity of this ELISA assay). However, samples collected at all times after footpad injection of LPS did contain significant levels of active TGF-β1 (Fig. 2) . Moreover, levels of active TGF-β2 also rose after LPS injection. Thus, the rise in total TGF-β levels in AqH of EIU reflects the accumulation of both TGF-β1 (presumably blood origin) and TGF-β2 (presumably generated in oculi). However, a rise in active TGF-β levels cannot explain the ability of AqH from eyes inflamed by systemic injection of LPS to promote T-cell activation in vitro. 

Evidence for the Presence of Proinflammatory Cytokines in AqH from Inflamed Eyes

Proinflammatory cytokines can promote T-cell proliferation, and several cytokines of this type have been detected in rodent eyes suffering from EIU. ^{18 19} Our next experiments were designed to detect by sandwich ELISA the presence of TNF-α, IL-1β, and IL-6 as proinflammatory cytokines, and IL-2 and IFN-γ as activated lymphocyte-derived inflammatory cytokines. AqH samples were collected as above and subjected to ELISA assay. The results of these experiments are summarized in Figure 3 . Within 6 hours of LPS injection, AqH contained high levels of TNF-α, IL-6, and IFN-γ and low levels of IL-1β. No evidence of IL-2 was found. The absolute levels of IL-6 were particularly high (>10 ng/ml), compared with the other cytokines (in the range of 1.5 ng/ml or less). Moreover, IL-6 continued to be present in AqH throughout the experiment; at 48 hours the amount of IL-6 was in the 100 pg/ml range. 

Capacity of IL-6 to Antagonize the Capacity of TGF-β to Inhibit T-Cell Proliferation In Vitro

To this point, our results indicate that AqH from eyes suffering from EIU acquired, on the one hand, increased levels of TGF-β, an immunosuppressive cytokine, and, on the other hand, high levels of IL-6, a proinflammatory cytokine. We next wished to determine which of these cytokines prevailed in our AqH samples. To approach this complex matter, anti-CD3–stimulated T-cell cultures were first established in medium containing 5 μl (final 20% vol/vol) of TGF-β2 at 0.5 ng/ml, comparable to the level of active TGF-β reported by Cousins et al. ⁵ to be present in normal AqH, and at 5.0 ng/ml (comparable to levels of TGF-β we detected in inflamed AqH, Fig. 1 ). Five microliters (final 20% vol/vol) of IL-6 was then added to these cultures at concentrations of 0.1, 1.0, and 10 ng/ml (comparable to levels in inflamed AqH at 6 and 12 hours). Proliferation was assessed in these cultures, and the results are summarized in Figure 4 . In the presence of no TGF-β, T cells proliferated vigorously (approximately 7500 cpm). Addition of increasing concentrations of IL-6 had no effect on T-cell proliferation. However, in the presence of TGF-β at 0.5 or 5.0 ng/ml (and in the absence of IL-6), T-cell proliferation was dramatically reduced (approximately 2500 cpm). Addition of increasing concentrations of IL-6 to these cultures resulted in a stepwise restoration of T-cell proliferative capacity such that at 10 ng/ml of IL-6, T-cell proliferation in the presence of TGF-β was virtually comparable to T-cell proliferation in the absence of TGF-β. These results indicate, first, that IL-6 is not by itself mitogenic for T cells; second, that TGF-β is a potent inhibitor of T-cell proliferation; and, third, that IL-6, especially at concentrations found in inflamed AqH, antagonizes almost completely the immunosuppressive effects of TGF-β. 

Effects of Anti–IL-6 Antibodies on Capacity of Inflamed AqH to Suppress T-Cell Proliferation

The loss of immunosuppressive activity by inflamed AqH during EIU could be explained by the presence of IL-6. Because inflamed AqH also contains significantly increased amounts of TGF-β, and because the results above indicate that IL-6 antagonizes TGF-β in suppressing T-cell mitogenesis, we wondered whether IL-6 was masking any immunosuppression that might be present. To test this idea, neutralizing anti–IL-6 antibodies were added to cultures containing normal or inflamed AqH, T cells, and anti-CD3. Anti–IL-6 Ab (20μ g/ml), which is capable of neutralizing 100 ng/ml of murine IL-6 according to manufacturer’s protocol, was added. When proliferation was assessed in these cultures (see Fig. 5 ), AqH collected from inflamed eyes at 6, 12, and 24 hours inhibited T-cell activation. These findings suggest that IL-6 emerged in AqH of inflamed eyes and masked the underlying capacity of the fluid to suppress T-cell activation. Because the fluid contains high levels of TGF-β at this time, we speculate that IL-6 antagonizes TGF-β, preventing this cytokine from carrying out immunosuppressive functions in the ocular microenvironment. 

Evidence of Cytokine Gene Expression in Tissues of Inflamed Eyes

Having found increased levels of TGF-β1 and -β2, IL-6, TNF-α, and IFN-γ in inflamed AqH, we next inquired whether these factors were produced locally by parenchymal ocular cells or were elevated because the blood–ocular barrier had been breached. I/CBs were harvested from normal eyes and from eyes with EIU at the stipulated times post–LPS injection. These tissues were analyzed for mRNA content using a RPA. The results are presented in Figure 6 . An autoradiograph of I/CB mRNA samples displayed in Figure 6A reveals that IL-6 mRNA is virtually nondetectable in normal tissues but is massively upregulated at 6 (especially), 12, and 24 hours after LPS injection. Similar, though quantitatively much fewer, changes were observed for TNF-α, IFN-γ, and TGF-β1 mRNAs. No significant increase was observed in mRNA for TGF-β2. Densitometric evaluations of this autoradiograph are presented in Figures 6B and 6C . The rise in IL-6 mRNA from normal to 6-hour samples is approximately 30-fold. An approximate threefold rise for TGF-β1 mRNA was observed in the same interval. These results suggest that the remarkable rise in intraocular IL-6 early in EIU is due to production locally. Because this rise was so much greater than the elevations observed with TNF-α and IFN-γ, we infer that the production arises from ocular parenchymal cells rather than infiltrating leukocytes. 

Capacity of IL-6 to Inhibit the Induction of ACAID

The ability of an eye to support ACAID induction is thought to be an important dimension to ocular immune privilege. The results of the experiments described above suggest that IL-6 is a potent antagonist of TGF-β. Our initial plan was to determine whether ACAID could be induced by injecting soluble antigen into the inflamed eyes of mice with EIU. However, mice that receive an injection of LPS are rendered incapable of displaying delayed hypersensitivity responses; thus, we could not examine this issue directly. The discovery that IL-6 is present in inflamed AqH in concentrations that effectively eliminate local immunosuppression led us to examine whether exogenous IL-6 could inhibit ACAID induction in normal mice. Panels of BALB/c mice received intracameral injections of OVA alone, or mixed with IL-6. One week later these mice were immunized subcutaneously with OVA plus CFA. Seven days later their ears were challenged with OVA, and the ear swelling responses measured at 24 and 48 hours. As revealed in Figure 7 , coinjection of IL-6 with OVA prevented the induction of ACAID. Recipients of these injections mounted OVA-specific delayed hypersensitivity responses comparable to positive controls. Thus, IL-6 not only antagonizes the immunosuppressive properties of TGF-β with respect to T cells but it also prevents the ocular environment from promoting ACAID. 

Discussion

Potent immune regulatory forces are operative in the normal eye. ^{2 3 25 26} The existence of immune privilege in the eye offers experimental verification that these forces exist. The eye creates a microenvironment that ensures 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. ^{27 28 29} In a complementary fashion, the ocular microenvironment is also inhospitable to the expression of those forms of immunity that use nonspecific inflammation in carrying out their effector function. ^{2 3} 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“ relative” 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. In the laboratory, intraocular inflammation can be evoked in informative model systems in mice. ³⁰  

Using the murine EIU model, we recently reported that, on the one hand, AqH from inflamed eyes transiently loses its immunosuppressive properties (capacity to suppress T-cell activation in vitro), and that, on the other hand, inflamed AqH displays the capacity to activate T cells (in the absence of a cognate ligand for the T-cell receptor for antigen). The results of experiments reported here provide a molecular explanation for these attributes of AqH from EIU-inflamed eyes. TGF-β has been known to be one of the important immunosuppressive factors in normal AqH. ^{4 5} However, in normal AqH virtually all the TGF-β is present in the latent, rather than active (i.e., immunosuppressive) form. Previous reports of AqH displaying T-cell inhibitory activity in vitro have been performed after the AqH had been acid activated. ⁵ This procedure not only activates latent TGF-β but it also destroys the numerous immunomodulatory neuropeptides that are normally present in this ocular fluid. By contrast, when fresh (not acid-activated) AqH is examined for its immunosuppressive properties, TGF-β contributes very little. Instead, other factors such as VIP and α-MSH are the major immunosuppressive agents. Our results showing only small amounts of active TGF-β in AqH removed directly from eyes of normal mice are consistent with this concept. Our experimental results indicate that total TGF-β levels (both 1 and 2 isoforms) were increased as early as 6 hours after EIU induction and remained elevated through 24 hours (i.e., during the peak interval of inflammation). Unlike in normal AqH, much of this TGF-β was “active.” By RPA of I/CB from inflamed eyes, we found that TGF-β1 mRNA levels rose modestly, whereas there was little if any change in mRNA for TGF-β2. Similar results were reported for EIU in rats. ^{15 17} We infer that, because the rise in intraocular TGF-β levels correlated with a rise in AqH protein levels (reflecting a breakdown in the blood–ocular barrier), increased TGF-β in AqH was probably delivered from blood plasma rather than the eye itself. The paradox is that AqH displayed increased levels of a potent and active immunosuppressive factor (TGF-β) at the same time the eye was experiencing the development of inflammation in the anterior segment. 

Our suspicion that this paradox might be resolved by determining the factor responsible for inflamed AqH’s mitogenic activity proved to be correct. Although we assayed inflamed AqH for several cytokines known to be associated with inflammation, IL-6 turned out to be the factor responsible for enabling T-cell proliferation in vitro. Not only was IL-6 present in inflamed eyes at extremely high levels, but IL-6 mRNA levels in I/CB removed from inflamed eyes was 30-fold higher than IL-6 mRNA levels in similar tissues from untreated eyes. Moreover, inflamed AqH treated with anti–IL-6 antibodies was able to express a potent capacity to suppress anti-CD3–derived T-cell activation in vitro. Our evidence suggests that IL-6 is a potent proinflammatory factor in inflamed AqH. In fact, IL-6 may be a major mediator of uveitis. Although there is general agreement that normal AqH is devoid of this cytokine, IL-6 has been detected in ocular fluids of patients with uveitis ^{31 32} and in rodents with EIU. ¹⁸ However, IL-6 is not necessary for the development of uveitis subsequent to intravitreal injection of endotoxin in mice. ³³  

To the best of our knowledge, there is no prior evidence to suggest that IL-6, on its own, is mitogenic for T cells. However, evidence does exist to implicate IL-6 in the induction of IL-2 receptor expression, and in differentiation, and proliferation of T cells after stimulation through the T-cell receptor for antigen. ³⁴ Indeed, IL-6 is more active in this regard than either IL-1 or TNF-α. ³⁵ It is pertinent to this consideration that Kogiso et al. ³⁶ have reported that CD4⁺ T lymphocytes are required for the expression of EIU in mice, even though it is not usually thought of as a T cell–mediated disorder. With regard to the simultaneous presence of IL-6 and elevated levels of TGF-β in inflamed AqH, Reinhold et al. ³⁷ have reported that IL-6 and IL-2 are capable of abolishing the effect of TGF-β1 on DNA synthesis by T cells. In light of this varied evidence, we conclude that IL-6 is a potent antagonist of TGF-β, and we infer that the presence of IL-6 in inflamed AqH abolishes the potential immunosuppression that would otherwise be mediated by the elevated levels of active TGF-β. 

Our studies suggest, but do not prove, that the IL-6 found in inflamed AqH is derived from ocular parenchymal cells. Many cells have the potential to secrete IL-6, and macrophages that infiltrate inflammation sites, such as the anterior segment in EIU, clearly have the capacity to secrete IL-6. Typically, the synthesis and secretion of IL-6 by macrophages is coordinated with synthesis and secretion of IL-1 and TNF-α. We have reasoned that if the bulk of the IL-6 found in inflamed AqH was macrophage-derived, then similarly high levels of IL-1 and TNF-α should have also been found. Instead, IL-1 and TNF-α levels were only modestly elevated. Moreover, RPA revealed that upregulation of IL-6 mRNA in I/CB tissues removed from EIU eyes was 10-fold higher (or greater). Therefore, our current view is that the extremely high levels of IL-6 in inflamed AqH in EIU are produced by ocular parenchymal cells. Although the stimulus for ocular production of IL-6 in EIU is unknown, LPS itself can be detected in AqH samples from mice that receive footpad injections of this substance. ²³  

In our experiments, coinjection of soluble antigen injection with IL-6 into the AC of BALB/c mice impaired ACAID induction. Several other proinflammatory cytokines, such as IL-2, ³⁸ IL-1, ³⁹ TNF-α, and IFN-γ ⁴⁰ have been shown similarly to prevent ACAID induction. We suspect, but have no direct evidence, that these cytokines act by altering the functional properties of indigenous antigen-presenting cells that, under the influence of TGF-β, are responsible for preparing and delivering an ACAID-inducing signal to the spleen. The ability of IL-6 to interfere with ACAID induction implies that eyes experiencing LPS-induced anterior uveitis have lost immune privilege. It will be important to determine whether loss of immune privilege renders the transiently inflamed eye vulnerable to other immunopathogenic disorders that may have the potential to produce chronic inflammation, and therefore blindness. 

 Supported by National Institutes of Health (NIH) Grant EY05678. JWS is a recipient of a Research to Prevent Blindness Senior Scientific Investigator Award.

 Submitted for publication September 27, 1999; revised December 20, 1999 and March 14, 2000; accepted March 17, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Levels of total TGF-β in AqH from eyes of mice with EIU. AqH was collected and pooled from both eyes of panels of C3H/HeN mice (5 mice, 10 eyes per panel) at periodic intervals after LPS (200 μg) was injected into the hind footpads. The ocular fluid was acid-activated and then subjected to the mink lung epithelial cell bioassay for TGF-β. Inhibition of mink lung cell proliferation is expressed as nanograms per milliliter of total TGF-β. Results presented are representative of experiments repeated three times with similar results.

Figure 2.

 

Levels of TGF-β1 and TGF-β2 in AqH from eyes of mice with EIU. AqH was collected and assayed without acid activation for content of TGF-β1 and TGF-β2 using a sandwich ELISA.

Figure 3.

 

Levels of proinflammatory cytokines in AqH from eyes of mice with EIU. AqH was collected and assayed without acid activation by sandwich ELISA for content of IL-1β (A), IL-6 (C), TNF-α (B), and IFN-γ (D). Mean values of duplicate samples are presented ± SEM. Results presented are representative of experiments repeated twice with similar results. Lower limits of ELISA sensitivity were IL-1β, 15.6 pg/ml; IL-6, 51.5 pg/ml; TNF-α, 51.5 pg/ml; and IFN-γ, 500 pg/ml.

Figure 4.

 

Interactions of IL-6 and TGF-β in T-cell proliferation cultures driven by anti-CD3. BALB/c T cells were cultured with anti-CD3 antibodies in 20 μl medium to which was added 5 μl of active TGF-β2 (0.5 or 5.0 ng/ml) and/or 5 μl of IL-6 (0.1–10 ng/ml).[ ³H]thymidine was added during the terminal 8 hours of 48-hour incubations. Mean values of triplicate samples of radioisotope incorporation are presented ± SEM.

Figure 5.

 

Effect of anti–IL-6 antibodies on T-cell proliferation cultures in the presence of EIU AqH. BALB/c T cells were cultured with anti-CD3 antibodies in medium containing 20% AqH collected at selected intervals from eyes of mice with EIU. These samples were not acid-activated. Anti–IL-6 antibodies (or immunoglobulin isotope control) were added to these cultures. Radioisotope incorporation at 48 hours was assessed, and the results are presented as described in the legend to Figure 4 . * Indicates mean values significantly lower than cultures containing control isotype immunoglobulin (P < 0.01). Results presented are representative of experiments repeated three times with similar results.

Figure 6.

 

Detection of cytokine mRNA in I/CB tissues removed from eyes of mice with EIU. I/CB were excised from eyes of mice with EIU at selected times, and mRNA expression of TNF-α, IL-6, IFN-γ, TGF-β1, TGF-β-2, and GAPDH (internal housekeeping control) was determined by RNase protection assay. (A) Autoradiograph of RPA gel. Each band was identified by probes provided by the manufacturer’s kit and by positive control samples. Densitometric determination of quantitative recovery of mRNA of each cytokine compared with GAPDH mRNA is presented in (B) for IL-6 and TNF-α and in (C) for TGF-β1 and TGF-β2. Arbitrary units are expressed as a ratio of amount of cytokine mRNA/amount of GADPH mRNA, where the value of both in normal iris and ciliary body = 1.

Figure 7.

 

Effects of intracameral injection of IL-6 on capacity of intracameral injection of OVA to induce ACAID in normal mice. BALB/c mice received an intracameral injection of OVA (50 μg) plus IL-6 (20 ng/ml), or medium alone (ACAID control). Seven days later these mice, as well as naive positive controls, were immunized subcutaneously with OVA (100μ g/ml) plus CFA (A). One week later, the ear pinnae of these mice, plus naive negative controls, were challenged by an intradermal injection of OVA (200 μg/10 μl). Ear swelling responses were measured 24 and 48 hours later. Mean 24-hour ear swelling values ± SEM of each group of five mice are presented. * Indicates mean values significantly lower than positive control (P < 0.05). ** Indicates mean value significantly greater than ACAID control and insignificantly different from positive control (P > 0.05). Results presented are representative of experiments performed three times with similar results.

The authors thank Jacqueline M. Doherty, PhD, for her invaluable help during the preparation of this manuscript and Munenori Yoshida, MD, and Takeshi Kezuka, MD, for establishment of these experiments.