γδ T Cells in Anterior Chamber-Induced Tolerance in CD8⁺ CTL Responses

Abstract

purpose. Delivery of antigen to the anterior chamber (AC) of the eye induces a systemic form of tolerance, referred to as anterior chamber-associated immune deviation (ACAID). ACAID is characterized by decreases in delayed-type hypersensitivity responses and complement-fixing antibodies on subsequent challenge with an immunogenic form of the antigen. The current study was designed to test whether priming of antigen-specific CD8⁺ cytotoxic T-lymphocytes (CTLs) are inhibited by injection of soluble antigen into the AC and whether γδ T cells play a role in the inhibition of such responses.

methods. Antigen was administered through the AC to normal γδ T-cell-deficient or reconstituted γδ T-cell-deficient mice. Seven days after the AC injection, the mice were primed with antigen in adjuvant and 10 days later, their spleen cells were cultured for 5 to 7 days and the CTL responses measured.

results. CTL responses were inhibited by antigen delivered through the AC in normal but not γδ T-cell-deficient mice. Tolerance was reconstituted in δ-chain knockout mice by the adoptive transfer of γδ T cells from normal mice. Moreover, spleen cells and splenic γδ⁺ T cells, but not γδ⁻ T cells, from mice injected with antigen through the AC inhibited development of CTL responses when cultured together with primed effector T cells.

conclusions. These data show, for the first time, that administration of soluble antigen in the AC inhibits development of CD8⁺ cytotoxic T-cell responses and that γδ T cells play a critical role in inhibition of CTL responses.

To prevent excessive tissue damage during an infection that could lead to blindness, the eye has the means of minimizing inflammation. This is exemplified by the extended survival of foreign tissues placed in the AC, the vitreous cavity, and the subretinal space of the eye, compared with tissues grafted in conventional sites, such as the peritoneal cavity or the skin. ^{1 2} In addition, antigens introduced into the AC of the eye elicit a deviant form of systemic immunity, anterior chamber-associated immune deviation (ACAID), in which animals are refractory to the development of delayed-type hypersensitivity (DTH) when an immunogenic form of antigen is injected subcutaneously (SC). ³ In contrast to DTH responses, the overall antigen-specific antibody response is preserved. Nevertheless, the isotypes of IgG antibodies that fix complement, IgG2a, -2b, and -3, are inhibited, whereas IgG₁ is unaffected by intraocular antigen in some ⁴ but not all ⁵ strains of mice. These deficiencies suggest that Th1 functions of CD4⁺ T cells are inhibited in this form of tolerance. 

Although CD8⁺ T cells are the primary effector cells that eliminate viral infections from various host tissues including the eye, they also can be involved in ocular disease. For example, development of herpes stromal keratitis (HSK) after infection with herpes simplex virus (HSV)-1 involves a variety of immunologic mechanisms. ⁶ In certain strains of mice, HSV-1-induced keratitis is dependent on CD8⁺ T cells, but injection of HSV-1 into the AC suppresses development of HSV-specific cytotoxic T-lymphocytes (CTLs) and prevents keratitis. ⁷ Yet, the mechanisms regulating CD8⁺ CTL in the eye are poorly understood. 

One goal of the studies reported herein was to determine whether the injection of soluble proteins into the AC would inhibit the induction of CD8⁺ CTLs induced by priming with antigen in CFA. Activation of CD8⁺ αβ T cells requires presentation of endogenous peptides by the major histocompatibility complex (MHC) class I molecules. ⁸ Exogenous antigens, such as ovalbumin (OVA), are not processed for MHC class I presentation by most cells. ⁸ However, phagocytic cells have been shown to process exogenous proteins, load peptides onto MHC class I, and stimulate CD8⁺ T cells. ⁹ Moreover, injection of OVA emulsified in CFA primes precursors of CD8⁺, MHC class I-restricted, OVA-specific CTLs in H-2^b mice, and this process involves adjuvant and phagocytic cells. ¹⁰ Thus, phagocytic cells that have processed exogenous antigen can activate CD8⁺ T cells, raising the possibility that CTL activated by this route may be susceptible to ACAID. 

The idea that soluble antigen delivered through the AC may inhibit the CTL response is also supported by the observation that antigen administered by the oral route inhibits activation of CD8⁺ CTLs, as well as CD4⁺ T cells. ¹¹ Moreover, mice that are deficient in γδ T cells, TCR δ-chain knockout (δ^−/−) mice, or wild-type mice treated with anti-δ-chain antibody are resistant to tolerance induced by oral OVA, as measured by antibody, cytokine production, and priming of CTL precursors. ¹¹ These results suggest that γδ T cells play an essential role in oral tolerance. ¹² Therefore, we tested the hypothesis that delivery of soluble antigen in the AC of the eye would inhibit the induction of CD8⁺ CTL in a γδ T-cell-dependent fashion. 

Materials and Methods

Experimental Animals

Female C57BL/6 (B6) (H-2^b) mice, 7 to 8 weeks of age, were purchased from the National Cancer Institute (Frederick Cancer Research and Development Center, Frederick, MD). Male or female C57BL/6J-Tcrd^{tm1 Mom} (δ^−/−) (H-2^b) mice ¹³ were bred in the animal facilities at Emory University and used at 7 to 8 weeks of age. All procedures involving animals were conducted according to the principles in the guidelines of the Committee on Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Research Council and in adherence to the provisions of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Antigens

Chicken egg albumin (OVA, grade VI) and keyhole limpet hemocyanin (KLH) were purchased from Sigma Chemical Co. (St. Louis, MO). CFA containing Mycobacterium tuberculosis strain H37Ra and incomplete Freund’s adjuvant (IFA) were purchased from Difco Laboratories (Detroit, MI). Emulsions of OVA in CFA (2 mg/mL) or OVA in IFA (0.5 mg/mL) were prepared by mixing equal volumes of aqueous antigen and adjuvant. 

Target Cell Lines

Two H-2^b tumor cell lines were used for these studies: EL4, an MHC class II-negative T cell lymphoma, and E.G7-OVA, which is an EL4 cell line transfected with the chicken OVA gene ¹⁴ (kindly provided by Michael J. Bevan, University of Washington, Seattle, WA). All cells were cultured in standard growth medium (RPMI 1640 medium supplemented with 5% fetal bovine serum, 1 mM l-glutamine, 1 mM sodium pyruvate, 50 μM 2-mercaptoethanol, and antibiotics) at 37°C in 6% CO₂ in air. E.G7 was periodically cultured with 250 μg/mL neomycin to maintain expression of the OVA gene. All cell lines were free of mycoplasma. 

Inoculations

Mice were anesthetized by intramuscular injection of 100 μL of a mixture containing 10 mg/mL ketamine (Sigma Chemical Co., St. Louis, MO) and 2 mg/mL xylazine (Bayer Corp., Shawnee Mission, KS). One drop of proparacaine HCl (Alcon Inc., Humacao, Puerto Rico) was applied topically on the eye before injection. Under a dissecting microscope, 50 μg OVA or KLH in 2 μL phosphate-buffered saline (PBS) was injected into the AC of one eye with a microliter syringe and a 33-gauge needle (Hamilton, Reno, NV). 

Cytotoxicity Assay

Mice were primed for OVA-CTL responses by injection of 200 μg OVA in CFA into the rear footpad. Ten days later, their spleens were harvested and single-cell suspensions prepared. Mononuclear cells (30 × 10⁶ cells/10 mL culture) were incubated with irradiated (20,000 rad) E.G7-OVA at a ratio of 10:1 in standard growth medium for 5 to 7 days. In the coculture system, effector cells (15 × 10⁶) from mice primed with an SC injection of 200 μg OVA in CFA 10 days earlier were cultured with regulatory cells (15 × 10⁶) from naïve mice or mice primed with antigen injected into the AC 10 days earlier and incubated for 5 days in the presence of irradiated EG7-OVA. 

E.G7-OVA and EL4 cells were labeled with Na₂ ⁵¹CrO₄ (DuPont, Boston, MA) at 37° for 60 minutes, washed three times, and added to effector cells in 96-well plates at different E:T cell ratios. Cytolytic activity was quantified by a 4-hour ⁵¹Cr release assay. Supernatants were collected after 4 hours’ incubation at 37°C and radioactivity was detected in a gamma counter (Wallac, Turku, Finland). The percentage of specific lysis was calculated as 100 × [(release by CTL − spontaneous release)/(maximum release − spontaneous release)]. Maximum release was determined by addition of 1% Triton X-100 (EM Science, Gibbstown, NJ). Spontaneous release in the absence of CTLs was generally less than 15% of maximum release. It is important to note that development of OVA-specific CTLs absolutely depends on priming. Spleen cells from mice injected with OVA or CFA do not induce development of CTLs when stimulated with E.G7-OVA, whereas mice primed with both show vigorous CTL responses. ¹⁰  

The frequency of cytotoxic T-cell precursors (pTc) was determined by the classic limiting-dilution assay as described by Kruisbeek. ¹⁵ Briefly, B6 mice were injected with 50 μg of soluble KLH or OVA in the AC. After 10 days, the mice were immunized with SC injection of 200 μg OVA in CFA. Ten days after immunization, spleen cells from three mice per group were pooled and serially diluted in 96-well plates, starting at 3 × 10⁵ cells per well, 24 wells per dilution. Irradiated (20,000 rad) E.G7-OVA (1 × 10⁴ cells/well) were added, and all wells were restimulated 7 days later with irradiated (2,000 rad) splenocytes as filler cells (1 × 10⁵ cells/well), irradiated E.G7-OVA (1 × 10⁴ cells/well), and 20 U/mL recombinant human (rh)IL-2 (Hoffmann-La Roche, Nutley, NJ). Seven days later, the lytic activity of the cultured cells was tested by adding ⁵¹Cr-labeled E.G7-OVA targets (1 × 10⁴ cells/well) to the wells and incubating for 4 hours. Wells that had fewer counts per minute (cpm) than spontaneous release plus 3 SD were counted as negative. ¹⁵ The log of the percentage of negative wells was plotted against the number of spleen cells per well, and the precursor frequency was calculated at 37% negative wells. ¹⁵  

DTH Assay

To determine DTH response, mice were immunized with 100 μg OVA in 50 μL CFA injected SC at the base of the tail. Seven days later, mice were challenged with SC injection of 25 μL IFA containing 12.5 μg OVA in one hind footpad. The same volume of PBS in IFA was used as a negative control in the other hind footpad. The thickness of the footpad was measured 24 hours after challenge by using a micrometer (Mitutoyo 227-101; MTI Corp., Paramus, NJ). The swelling in response to OVA was calculated by the following formula: antigen-specific swelling (in millimeters) = the thickness of the footpad with injection of antigen-IFA (in millimeters) - the thickness of the footpad with injection of IFA-PBS (in millimeters). Footpads were used for DTH responses, because in our hands, larger, more reliable responses were obtained than in measurements of the ear-swelling response. Data from three experiments were pooled. Data were subjected to statistical analysis with Student’s t-test. 

T-Cell Purification

Spleen cells from B6 mice that were primed with 50 μg soluble OVA injected into the AC 10 days earlier were pooled and purified by positive selection with antibody-coated MicroBeads and separation through MACS columns (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s instructions. B cells were first depleted from the splenocyte suspension using anti-mouse CD19-coated beads. For γδ T-cell isolation, the B-cell-depleted cells were treated with Fc block (PharMingen, San Diego, CA) for 15 minutes followed by the incubation with biotin-conjugated anti-mouse TCR δ chain antibody (GL3; Pharmingen) on ice for 30 minutes. GL3-treated cells were washed, treated with streptavidin-loaded MicroBeads, and separated (MACS columns; Miltenyi Biotech). In our hands, approximately 0.5% to 1.0% of lymphocytes in the spleens of B6 mice were TCR γδ⁺ (GL3⁺) cells. Using the magnetic separation method, 0.4% to 0.6% of spleen cells were recovered in the GL3⁺ population, which was routinely 85% pure or more. Positive selection was chosen, because the purity of the cells is higher than in negatively selected cells, particularly when the cells of interest are a minority such as splenic γδ T cells. MACS super-paramagnetic MicroBeads are extremely small, approximately 50 nm in diameter, which is comparable to the size of a virus. Their size and composition (iron oxide and polysaccharide) make the MicroBeads biodegradable, so that labeled cells retain their physiological function. Positive selection with MicroBeads generally does not activate the cells. ^{16 17 18}  

Results

Effect of Soluble Antigen Injected into the AC on Priming of CTLs

To determine whether CD8⁺ CTL development is subject to ocular tolerance, OVA-specific CTL responses were measured in B6 mice that were untreated or injected with 50 μg OVA or KLH in the AC 7 days before SC immunization with 200 μg OVA in CFA. Ten days after immunization, spleen cells were harvested and cocultured with irradiated E.G7-OVA. As previously reported, OVA-specific cytolytic activity, was generated in control B6 mice that received no AC treatment (Fig. 1A) . OVA-specific lytic activity was mediated by CD8⁺ T cells because anti-CD8 mAb (53-6.72) blocked CTL activity when added to the in vitro activation culture or to the 4-hour lytic assay (data not shown). Soluble OVA injected into the AC inhibited the generation of CD8⁺ CTL to a level comparable to the lysis of the negative control when EL4 targets were used. By contrast, priming for OVA-specific CTLs in mice injected with the irrelevant antigen KLH was similar to that of control mice without AC treatment. Thus, administration of soluble antigen into the AC before immunization inhibited the priming of CD8⁺ CTL precursors in an antigen-specific manner. Mice that received OVA through the AC also exhibited reduced DTH responses when they were challenged with OVA in the footpad (Fig. 1B) , which verifies the ACAID phenomenon in our experiments. Mice pretreated with the irrelevant protein KLH mounted a strong DTH response to OVA that was comparable to that in the untreated mice. Thus, delivery of antigen to the AC induces tolerance in both the CD4⁺ and CD8⁺ T-cell compartments. 

γδ T cells have been shown to be necessary for inhibition of DTH responses in ACAID, ^{5 19} and therefore we tested whether they are also involved inhibition of CTL responses. Thus, the susceptibility of δ^−/− B6 mice to inhibition of CTL responses by OVA in the AC was compared with that of B6 mice. OVA-specific CTLs were primed in both B6 and δ^−/− mice, as displayed by equal lysis in these cultures (Fig. 2) . However, delivery of OVA through the AC did not inhibit the priming of OVA-specific CTLs in δ^−/− mice, as it did in B6 mice, suggesting that γδ T cells are also needed to inhibit CTL responses. 

One caveat to this experiment is that disruption of the δ-chain gene during the generation of δ^−/− mice may have inadvertently altered the expression of other genes. To address the possibility that the abrogation of ACAID in δ^−/− mice may be the result of something other than the deletion of γδ T lymphocytes, δ^−/− mice were reconstituted with B6 lymphocytes before the induction of ACAID. B-cell-depleted spleen cells from B6 mice were separated into γδ⁺ and γδ⁻ subsets. δ^−/− mice received no B6 T cells, 0.25 × 10⁶ γδ⁺ T cells, or 25 × 10⁶ γδ⁻ T cells before injection of OVA into the AC and subsequent measurement of CTL responses. The δ^−/− mice that received no transplanted cells but had OVA injected into the AC exhibited a CTL response that was equivalent to that in the untreated control animals, verifying that no tolerance was induced in δ^−/− mice (Fig. 3) . The transfer of 0.25 × 10⁶ γδ⁺ T cells, which is approximately equal to 50% of total γδ T cells in one spleen, reconstituted the sensitivity of δ^−/− mice to ACAID. By contrast, priming for OVA-CTL was not inhibited in the δ^−/− mice that were reconstituted with γδ⁻ T cells from B6 mice, confirming that γδ T cells play an important role in generating the tolerance of CTL responses in mice injected with soluble antigen in the AC of the eye. 

Effect of Antigen Injected into the AC on Regulatory Cells That Inhibit Development of CTLs

There are several potential mechanisms by which CTL responses may be inhibited in ACAID. The inhibition may be attributable to the immune deviation of CD4⁺ helper T cells. ²⁰ However, CTL responses induced by OVA in CFA are independent of CD4⁺ T cells, ¹⁰ presumably because the mycobacteria activate antigen-presenting cells (APCs) directly. ²¹ Induction of anergy, immune deviation of CD8⁺ T cells, or activation of regulatory cells are other mechanisms that could account for inhibition of CTL activity. 

To investigate this question, the frequency of cytotoxic pTcs was measured in the spleens of mice that were inoculated in the AC with 50 μg KLH or OVA 10 days before receiving SC injection of 200 μg OVA in CFA. The pTc frequency ranged from 0.677 to 4.000 per 10⁶ splenocytes in mice pretreated with KLH in the AC and 0.186 to 0.500 per 10⁶ splenocytes in mice pretreated OVA (Table 1) . The ratio of pTc frequencies (KLH-treated to OVA-treated) showed that 3 to 10 times more pTcs were detected in mice pretreated with the irrelevant antigen KLH than in mice treated with OVA in the AC. These data verify that introducing soluble antigen into the AC before immunization inhibited the activation or expansion of CTL precursors. 

The observation that CTLs were detected in the precursor frequency assays but not in the bulk cultures raised the possibility that an active regulatory population, which may have been present in the bulk cultures, failed to survive or were diluted out in the pTc assay. To directly test whether ACAID cells inhibit CTL responses, the spleen cells from mice primed with SC injection of 200 μg OVA in CFA (effector cells) were cocultured with the cells from mice injected with 50 μg soluble OVA in the AC (regulatory cells) at a 1:1 ratio in the presence of irradiated E.G7-OVA. Naïve B6 spleen cells were used as filler cells in the control cultures, which contained only effector or regulatory cells, to keep the total cell number constant in all cultures. As previously shown, OVA-specific CTL activity was induced by priming with SC injection of OVA in CFA, but not in mice that received only soluble OVA in the AC (Fig. 4) . Moreover, OVA-specific lytic activity of effector cells was reduced when they were cocultured with an equal number of ACAID regulatory cells. The inhibition of the CTL activation in vitro by cells from ACAID mice suggests that regulatory T cells were activated in the spleen by inoculating with antigen through the AC. 

To determine whether the regulatory cells were specifically induced by antigen, spleen cells from mice with OVA or KLH injected into the AC (regulatory cells) were cultured with spleen cells from mice primed with OVA in CFA (effector cells) in the presence of EG7-OVA. Five days later, lysis of E.G7-OVA targets was measured. In contrast to the dramatic decrease of OVA-specific cytolytic activity induced by regulatory cells from mice with OVA injected into the AC, cytolytic activity was inhibited only marginally by regulatory cells from mice with KLH injected into the AC (Fig. 5) . Thus, regulatory cells appear to be induced in an antigen-specific fashion. 

γδ T cells have been shown to exert a negative regulatory role in several forms of immune tolerance. ^{12 22 23 24 25} To test the possibility that γδ T cells were responsible for the regulatory activity, effector cells from mice primed with OVA in CFA were cocultured at a 1:1 ratio with regulatory spleen cells from naïve mice or unfractionated spleen cells, splenic γδ⁺, or γδ⁻ T cells from mice primed with soluble OVA in the AC. A portion of positively selected γδ T cells equal to that of γδ cells in the unfractionated population was added to cultures of effector cells. The total number of cells in each culture was equalized by the addition of naïve B6 splenocytes. OVA-specific killing by the effector cells was dramatically inhibited by the splenic γδ⁺ T subset from mice primed with soluble OVA in the AC (Fig. 6) . In addition, the degree of inhibition was greater than when unfractionated whole spleen cells from these mice were added. These data suggest that γδ T cells from the spleens of mice that received with OVA in the AC were necessary and sufficient to suppress the CTL response by effector cells. 

Discussion

These studies show for the first time that the administration of soluble antigen by injection into the AC of the eye reduces CD8⁺ T-cell-mediated cytotoxic responses to proteins such as OVA under the same conditions that inhibit development of DTH responses, which are mediated by CD4⁺ T cells. OVA-specific CTL responses were inhibited in an antigen-specific fashion. Thus, cell-mediated immune responses by both CD4⁺ and CD8⁺ T cells are susceptible to tolerance induced by delivery of antigen into the AC. Inhibition of CTL responses did not occur in δ-chain knockout mice, which extends recent reports that they are necessary for inhibition of DTH responses. ^{5 19} Furthermore, B6 γδ T cells reconstituted ACAID in δ-chain knockout mice when adoptively transferred before AC injection. 

The observation that CTL responses to OVA were inhibited by injection of OVA into the AC appears to contradict a previous report that P815 tumor cells from DBA/2 (H-2^d) mice delivered to the AC of BALB/c (H-2^d) mice inhibited DTH responses but stimulated CTL responses equal to those induced by SC injection of P815. ²⁶ These differences in sensitivity of CTLs to ACAID are not yet understood, but there are several potential explanations. For example, the antigen presented to BALB/c mice by the P815 tumor is a minor histocompatibility antigen, which is an endogenous peptide that is presented by MHC class I directly to H-2^d-compatible host T cells without the need for processing by APCs. Soluble antigen injected into the AC of the eye is thought to be processed locally by APCs, under the influence of TGFβ and other mediators. 

Antigen-bearing APCs are believed to migrate from the AC into the bloodstream and take up residence in the spleen where they induce ACAID. ²⁷ If the P815 tumor cells exited the AC, they could present antigen through MHC class I directly to peripheral T cells, in the absence of an ACAID-inducing signal, which could promote a CTL response. Alternatively, direct antigen presentation may fail to induce the γδ suppressor cells that appear to be necessary for inhibition of CTL priming in spleen, whereas soluble antigen, which can be processed and presented through the MHC class I and II pathway by APCs, ²⁸ may generate the appropriate epitopes to activate both αβ and γδ T cells. 

In the current findings, it was also shown that injecting antigen into the AC specifically induced regulatory T cells that inhibited in vitro activation of OVA-specific CTL precursors. The γδ T-cell-enriched fraction from ACAID spleens inhibited the lytic function of CTL effector cells, whereas the γδ T-cell-depleted fraction, containing both CD4⁺ and CD8⁺ αβ T cells, did not. These results suggest that γδ T cells are required for regulatory cells that inhibit CTL responses. These results extend previous reports that γδ T cells serve as negative regulators in immune responses to pathogens, ^{29 30} tumors, ²⁵ allografts, ^{31 32} and autoimmunity. ^{22 33}  

The observation that splenic γδ⁺ cells, but not γδ⁻ cells that contained CD4⁺ and CD8⁺ αβ T cells, inhibited OVA-specific CTL responses seems to contradict the previous report that CD4⁺ and CD8⁺ T-cell-regulatory cells inhibit DTH responses. ³⁴ One explanation for the discrepancy is that the in vitro CTL response may be subject to different forms of regulation than the DTH responses. On the other hand, these two observations are not necessarily mutually exclusive. The experiments of Wilbanks et al., ³⁴ demonstrating that AC injection induces splenic CD4⁺ afferent and CD8⁺ efferent suppressor T cells, were performed by depletion, by the use of antibody and complement. This method would have left some gd⁺ T cells in both treated populations, because some splenic γδ T cells are CD8⁺, whereas others are CD8⁻. ³⁵ If the γδ T cells transmitted the ACAID-inducing signal to the remaining CD4⁺ and CD8⁺ T cells in each subset, they could have developed into afferent and efferent suppressors, respectively. 

The observation that administration of KLH did not induce suppressor cells that inhibited an OVA-induced CTL response raises the possibility that the γδ T cells in the spleen may be antigen specific. γδ T cells have also been reported to inhibit IgE-specific antibody responses to inhaled soluble antigens. ²⁴ In addition, they have been identified in the spleens of tolerant, αβ T-cell receptor (TCR)-deficient mice. ³⁶ In the latter study, γδ T-cell-mediated regulatory activity was MHC unrestricted, indicating that recognition by γδ T cells may not involve antigen presentation by MHC, as αβ T cells do, or that antigen was presented by a nonpolymorphic MHC molecule. 

The antigenic epitopes recognized by γδ T cells are much less well-defined than those recognized by αβ T cells, but they include nonpeptide epitopes that can be presented by nonclassic MHC molecules. ^{37 38} γδ T cells from tumor-bearing mice have been reported to recognize directly the tumor cell-associated antigen, Q5, which is a nonclassic class I antigen. ²⁵ These activated γδ T cells are suppressors and inhibit CTL responses, which facilitates escape of tumor cells from immune surveillance. Anti δ-chain mAb (GL3) treatment of mice downregulates TCR γδ in vivo and inhibits γδ T-cell functions that are essential for the induction of oral tolerance, ¹² suggesting that the normal functional activity of γδ T cells involves an interaction of the γδ TCR with a ligand. Moreover, the potential diversity of the γδ TCR is greater than that of any other antigen receptor. ^{38 39} However, the question of whether the γδ T cells that play a role in ACAID recognize the nominal antigen administered through the AC, peptides presented by conventional APCs, the effector T cells with which they were mixed, or some other cells has not yet been investigated. Conceivably, γδ TCR interacts with their ligands directly or in the context of other cell surface molecules. To determine whether these regulatory cells require the encounter of antigen in vitro to be activated and become inhibitory, stimulators expressing a second protein antigen, such as KLH, are needed. Until a reciprocal system is developed, it is not possible to determine whether the antigenic specificity is a function of the γδ T cells or the αβ T cells that are subject to regulation. 

The mechanism involved in the inhibition by γδ T cells of development of CTLs in vivo and in vitro is not yet clear. Alteration of αβ T cell development has been reported as one mechanism of γδ T cells’ regulating immune responses. ^{33 40} γδ T cells are capable of producing a broad spectrum of cytokines, including regulatory cytokines such as TGFβ, IL-4, and IL-10. ^{23 31 41 42} CD8⁺ OVA-induced CTLs primed by OVA administered in CFA are Tc1 cells that make type 1 cytokines, such as IFNγ and TNFα. ⁴³ The IFNγ production by cells from spleens or lymph nodes of ACAID mice has been reported to be reduced. ⁴⁴ Therefore, it is possible that cytokines produced by γδ T cells divert the differentiation of Tc1 cells into a nonlytic pathway. There are also reports that γδ T cells are cytotoxic. ^{45 46} Thus, γδ T cells could regulate OVA-induced CTLs by eliminating APCs. Preliminary data suggest that direct cell contact between γδ T cells and OVA-primed spleen cells is required for inhibition of CTL responses. 

Splenic B cells ⁴⁷ and CD1-reactive natural killer T (NKT) cells ⁴⁸ have also been reported to be necessary in the generation of ACAID. Moreover, NKT cells accumulate in the spleen after the induction of ACAID and form clusters with APCs and T cells in the splenic marginal zone. ⁴⁹ We have observed that there is also an increase in the percentage of γδ T cells in the spleens of ACAID mice, and CD44 is upregulated in this population. ⁵ γδ T cells have been shown to be present primarily in the sinusoids of the avian spleen, ⁵⁰ whereas they have been localized to the red pulp in human spleens. ⁵¹ Studies are under way to determine where γδ T cells reside in the murine spleen and whether they physically interact with APCs, B cells, or NKT cells in generating the signals for ACAID. 

²  Present affiliation: Center for Neurological Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts.

 Supported by a research grant and a center grant from the Foundation Fighting Blindness, National Eye Institute Core Grant P30 EYO0636 and Grant EY13459, and a gift from Malcolm and Musette Powell. JAK is the recipient of the Research to Prevent Blindness Jules and Doris Stein Professorship in Ophthalmology.

 Submitted for publication July 2, 2002; accepted July 20, 2002.

 Commercial relationships policy: N.

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

 

Figure 1.

 

Intraocularly delivered antigen inhibited priming of CTLs and the DTH responses. B6 mice received no antigen, 50 μg OVA, or 50 μg KLH in the AC of the eye 10 days before SC injection of 200 μg OVA in CFA. Spleen cells of some mice (n = 3 per group) were harvested 10 days after the last exposure to antigen and cultured individually with irradiated E.G7-OVA stimulator cells for 5 or 6 days. OVA-specific cytolytic activity was measured using ⁵¹Cr-labeled E.G7-OVA or EL4 targets at various E-to-T cell ratios (A). Data are the mean ± SD of results in three individual mice from a representative experiment repeated three times with similar results. DTH responses to OVA were detected by challenging the hind footpad with SC injection of 12.5 μg OVA in IFA 7 days after immunization (B). Footpad swelling was measured 24 hours after the challenge. Each symbol represents an individual mouse, and the data are the accumulation of results in three experiments with the same protocol. The bar represents the average swelling. Probabilities were calculated with Student’s t-test to compare the control group that received no intraocular treatment with those that received either OVA or KLH.

Figure 2.

 

Reduced cytotoxicity was not observed in δ^−/− mice pretreated with soluble OVA in the AC before the priming for OVA-induced CTLs. B6 or δ^−/− mice that received either no treatment (▴ or •) or 50 μg OVA in the AC (▵ or ○) on day 0 were immunized with OVA in CFA on day 7 and then challenged with OVA in IFA on day 14. On day 21, spleen cells were harvested, pooled, stimulated with E.G7-OVA, and tested for OVA-specific lytic activity to EG7-OVA targets. The results of a representative experiment, which was repeated three times, are displayed.

Figure 3.

 

B6 γδ T cells reconstituted tolerance of priming for cytolytic T-cell responses in δ knockout mice with induced ACAID. δ^−/− mice without transplanted cells or OVA through the AC (□) or with OVA in the AC (⋄) were compared with δ^−/− mice that received 0.25 × 10⁶ B6 γδ T cells and AC injection of OVA (▵) or 25 × 10⁶ B6 γδ T cells before AC injection of OVA ( ). Seven days after AC treatment, all mice were primed with 200 μg OVA in CFA and OVA-specific cytolytic responses to E.G7-OVA targets. Data displayed are representative results ± SD in one of three similar experiments, in which mice were assayed individually.

Table 1.

 

OVA-Specific CTL Precursor Frequencies in the Spleens of Mice with Antigen Injected into the AC

Experiment	Precursor Frequency (N/1 × 106)			Ratio KLH:OVA
		KLH	OVA
1	0.677	0.207	3.26:1
2	4.000	0.500	8.00:1
3	1.802	0.450	4.00:1
4	0.980	0.186	5.28:1
5	3.817	0.356	10.73:1

 B6 mice (n = 3/group) were injected with 50 μg KLH or OVA into the AC and were immunized with SC injection of 200 μg OVA in CFA10 days later. After another 10 days, the mice were killed, the spleen cells from each group were pooled, and the CTL precursor frequencies were determined in five separate experiments.

Figure 4.

 

Soluble OVA delivered to the AC induced regulatory cells. Effector cells came from mice primed with SC injection of 200 μg OVA in CFA 10 days earlier, whereas regulatory cells came from mice primed with 50 μg soluble OVA injected into the AC 10 days earlier. Effector cells and regulatory cells were cultured separately with naïve spleen cells to make up the final cell number or together at a 1:1 ratio with irradiated E.G7-OVA. CTLs were measured with ⁵¹Cr-labeled E.G7-OVA or EL4 target cells. The means ± SD of results in three mice per group are shown, typical of results obtained in three experiments.

Figure 5.

 

Induction of suppressor cells in ACAID-affected mice was antigen specific. Equal numbers of splenic effector cells and regulatory cells from naïve mice (♦), mice inoculated in the AC with OVA (▴) or KLH (•) were cocultured with E.G7-OVA, and CTLs were measured with ⁵¹Cr-labeled E.G7-OVA. The means ± SD of results in three mice per group are shown, typical of results obtained in three experiments.

Figure 6.

 

Splenic γδ T cells from ACAID mice inhibited the development of cytolytic CD8⁺ T cells in vitro. Pooled effector cells (15 × 10⁶) from mice primed with SC injection of OVA in CFA 10 days earlier were cocultured in the presence of stimulator EG7-OVA with pooled regulatory cells consisting of 15 × 10⁶ normal B6 spleen cells (▪), 15 × 10⁶ unfractionated ACAID spleen cells (▴), 0.3 × 10⁶ γδ⁺ T cells from ACAID spleens plus 15 × 10⁶ normal B6 spleen cells (○), or 15 × 10⁶ γδ⁻ T cells (⋄) from ACAID spleens. OVA-specific CTLs were measured 5 days later. Shown is a representative result of three independent experiments.

The authors thank Wayne Streilein for generously providing time to train them in the techniques used for induction of ACAID, Kyle McKenna for critical review of the manuscript, and Jing Wen for technical assistance with cell separations.