In some parts of the world, shortages of human donor corneas

Transcription

1 CD8 T Cell Mediated Delayed Rejection of Orthotopic Guinea Pig Cornea Grafts in Mice Deficient in CD4 T Cells Ryotaro Higuchi and J. Wayne Streilein PURPOSE. To determine the immunopathogenesis of delayed orthotopic corneal xenograft rejection in mice deficient in the xenoreactive CD4 T cells that mediate acute rejection. METHODS. CB.17 SCID and BALB/c mice were used as recipients of orthotopic cornea grafts obtained from strain 13 guinea pigs. Before transplantation, SCID recipients, which do not normally reject guinea pig cornea grafts, were reconstituted with spleen cells (whole, CD4-depleted, CD4/CD8-depleted) or purified CD8 T cells from normal BALB/c donors. Graft survival was assessed by clinical examination, and median survival times (MST) were calculated. Lymphocytes from mice that rejected guinea pig cornea grafts were analyzed in vitro for their capacity to respond to guinea pig xenoantigens and to lyse guinea pig target cells. RESULTS. SCID mice reconstituted with whole spleen cells from BALB/c donors rejected guinea pig corneas with a vigor identical with that of normal BALB/c mice (MST 15 and 14 days, respectively), whereas SCID mice reconstituted with CD4-depleted BALB/c spleen cells rejected guinea pig corneas in a delayed fashion (MST 27 days), as did SCID mice reconstituted with purified CD8 T cells from BALB/c donors. Although CD8 T cells from rejector mice failed to lyse guinea pig target cells in vitro, the T cells proliferated and secreted IFN- in response to in vitro stimulation with guinea pig xenoantigens. CONCLUSIONS. Guinea pig cornea xenografts that avoid acute rejection in CD4 T cell-depleted mice are vulnerable to rejection by CD8 T cells. Effector CD8 T cells destroy corneal xenografts through release of proinflammatory mediators (IFN- ) rather than by cytotoxicity. (Invest Ophthalmol Vis Sci. 2003;44: ) DOI: /iovs In some parts of the world, shortages of human donor corneas lead to long waiting times for patients who need corneal transplants. 1 It is possible that corneas harvested from suitable xenogeneic sources could be a reasonable alternative to this problem. However, the barriers to success of solid-tissue xenografts significantly exceed the barriers to success of allografts. 2,3 The heightened barrier to successful xenografts results primarily from the presence in unimmunized recipients of so-called natural antibodies that bind to vascular endothelial From the Schepens Eye Research Institute and Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts. Supported by National Institutes of Health (NIH) Grant EY Submitted for publication January 15, 2002; revised April 25, 2002; accepted June 19, 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 solely to indicate this fact. Corresponding author: J. Wayne Streilein, Schepens Eye Research Institute, 20 Staniford St., Boston, MA 02114; waynes@vision.eri.harvard.edu. cells of xenografts, triggering complement activation, clot formation, and hyperacute rejection. In the case of the cornea, distinctive features exist which partially exempt this tissue from this type of pathogenesis. 4 First, the normal cornea is an avascular tissue. Second, the corneal endothelium constitutively expresses potent membrane-bound complement inhibitors. 5 7 Third, the cornea forms the anterior wall of the anterior chamber an immune-privileged site with a fluid (aqueous humor) that contains anticomplementary activity. 8,9 Fourth, the cornea itself possesses inherent immune privilege, 10 in part through the ability of its cells to secrete immunosuppressive factors 11,12 and express CD95 ligand. 13,14 Our laboratory has recently demonstrated that guinea pig corneas grafted heterotopically into the anterior chamber of eyes of normal mice whose sera naturally contain guinea pig reactive antibodies are not vulnerable to hyperacute, antibody-mediated rejection, 15,16 and that acute rejection of corneal xenografts occurs in mice genetically incapable of producing antibodies. 17 These findings suggest that naturally occurring antibodies may not be an important barrier to successful orthotopic cornea grafting in humans. However, the absence of antibody-mediated corneal xenograft rejection does not mean that mice are incapable of rejecting corneal xenografts. Tanaka et al. 17 have reported that BALB/c and C57BL/6 mice reject guinea pig corneas in an acute manner, and they have provided convincing evidence that acute rejection is mediated by CD4 xenoreactive T cells. Because the normal cornea lacks class II MHC-bearing antigenpresenting cells (APCs), it is not particularly surprising that the CD4 T cells mediating acute corneal xenograft rejection are of the indirect type, and that is what has been found. 18 That is, murine T cells appear to be incapable of recognizing and responding to MHC-encoded guinea pig xenoantigens, but they readily respond to guinea pig xenoantigens processed and presented on murine APCs. Thus, rejection of orthotopic guinea pig corneal xenografts by mice is accomplished almost exclusively by indirect xenoreactive T cells, a situation virtually identical with the manner in which mice reject orthotopic corneal allografts. 19,20 When CD4 T cells are eliminated from mice that receive orthotopic guinea pig cornea grafts, acute rejection is avoided, but the grafts eventually succumb to a delayed destructive inflammation that typically comes to completion by 30 days. 17 Because immune deficient SCID mice have been found to be incapable of rejecting guinea pig cornea grafts, 17 the delayed rejection observed in CD4 T cell deficient mice strongly suggests that adaptive immune effectors are responsible. In the present study, we tested the hypothesis that CD4-independent CD8 T cells are the mediators of delayed rejection of guinea pig corneal xenografts. The results support the validity of this hypothesis, and suggest that the effector CD8 T cells cause graft rejection by releasing proinflammatory cytokines, such as IFN-, that promote macrophage-mediated graft destruction. Investigative Ophthalmology & Visual Science, January 2003, Vol. 44, No. 1 Copyright Association for Research in Vision and Ophthalmology 175

2 176 Higuchi and Streilein IOVS, January 2003, Vol. 44, No. 1 MATERIALS AND METHODS Animals and Anesthesia Strain 13 inbred guinea pigs ( g) were purchased from Crest Caviary (Prunedale, CA). BALB/c and C.B-17SCID (C.B-17/IcrTac-scid) mice were obtained from our animal facility or purchased from Taconic Farm (Germantown, NY). All mice were males, 8 to 12 weeks of age. All animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. In all experiments, mice were used as recipients and guinea pigs were used as donors. Corneal grafts were prepared from eyes enucleated after the donor guinea pigs were killed. For experimental manipulations, mice were deeply anesthetized with an intraperitoneal injection of 3 mg ketamine and mg xylazine. Reconstitution of SCID Mice with Normal Lymphoid Cell Populations In some experiments, adult SCID mice were reconstituted with whole spleen cells (or fractions thereof) obtained from normal BALB/c. Subsequently, the reconstituted mice received orthotopic guinea pig cornea grafts. For reconstitution with whole spleen cells, BALB/c splenocytes, depleted of RBCs by lysis with Tris-NH 4 Cl, were suspended in Hanks balanced salt solution (HBSS) and injected intravenously into CB17-SCID mice (one donor equivalent, approximately cells/ donor). For reconstitution with splenocytes depleted of CD4 T cells or CD4 CD8 T cells, RBC-lysed splenocytes from BALB/c mice were incubated with anti-cd4 (GK1.5) alone, or in conjunction with anti-cd8 (2.43) antibodies (ascites fluid, a kind gift from Joan Stein Streilein, Schepens Eye Research Institute, Boston, MA) for 30 minutes on ice, washed twice with complete medium (CM) composed of RPMI 1640 medium, 10 mm HEPES, 0.1 mm nonessential amino acids, 1 mm sodium pyruvate, 100 U/mL penicillin, 100 g/ml streptomycin (all from BioWhitaker, Walkersville, MD), and 1 10 to 5 M 2-mercaptoethanol (ME; Sigma Chemical Co., St. Louis, MO), supplemented with heat-inactivated 10% fetal calf serum (Sigma). The cells were then incubated with infant rabbit complement (Pel-Freez Biologicals, Rogers, AZ) for 30 minutes at room temperature, washed once, and resuspended in HBSS and injected (one donor equivalent) intravenously into SCID mice. 21,22 On the same day, the mice received 100 L of anti-cd4 (GK1.5) antibodies (ascitic fluid diluted 1:2 with HBSS) injected intraperitoneally. Anti-CD4 antibodies were injected intraperitoneally at weekly intervals thereafter until the experiment came to completion. For reconstitution with purified CD8 T cells, cervical, inguinal, and mesenteric lymph nodes and spleens were obtained from naive BALB/c mice (two donors/recipient) and pressed through nylon mesh. After the RBCs were lysed, the suspension was passed through a T cell-enrichment column (Biotecx Laboratories, Inc, Houston, TX), followed by passage through a CD8 T cell-enrichment column (R&D Systems, Minneapolis, MN). The purity was 88% with no CD4 T cells ( 0.1%). Purified CD8 T cells (approximately two donor equivalents, cells/recipient) were injected intravenously into SCID mice. These recipient mice also received systemic treatment with anti CD4 antibody as described earlier. All reconstituted SCID mice received orthotopic guinea pig cornea transplants 1 day after reconstitution. Flow Cytometry Analysis of Peripheral Blood T Cells in Reconstituted SCID Mice On the day after reconstitution, peripheral blood cells were removed from SCID mice, depleted of red blood cells by lysis with Tris-NH 4 Cl, then triple-stained with Cy-Chrome- anti-cd3e (CD3), PE-anti-CD4 (L3T4) and FITC-anti-CD8a (Ly-2). Antibodies and isotype controls were purchased from PharMingen (San Diego, CA). The stained cells were analyzed by flow cytometry using a cell counter (Epics XL Analyzer; Coulter Inc., Hialeah, FL). Orthotopic Corneal Xenografting The detailed surgical procedure for performing orthotopic corneal xenotransplants in mice was previously reported. 17 Briefly, grafts were removed from donor guinea pig corneas by excision with a 2.0-mm diameter trephine. The corneal tissue was then placed in HBSS until grafting. The graft bed was prepared by excising with Vannas scissors a 1.5-mm site from the central cornea of the right eye. The xenograft was placed in the recipient bed and secured with 10 interrupted 11-0 nylon sutures (Sharpoint; Vanguard, Houston, TX) juxtaposed to the epithelial surface of the recipient cornea. Antibiotic ointment was applied to the corneal surface, and the lids were closed with an 8-0 nylon tarsorrhaphy. Tarsorrhaphy was maintained (except for clinical inspection purposes) until graft rejection was documented. Corneal sutures were removed on day 8. Assessment of Xenograft Survival Corneal xenografts were evaluated by slit-lamp biomicroscopy three times a week. The day of rejection was defined when graft transparency was lost (i.e., the iris margin and iris structures were no longer visible clearly through the graft), and graft clarity never recovered subsequently. In Vitro T-Cell Activation Assays Cervical lymph nodes ipsilateral to the xenograft-containing eye were removed from BALB/c (n 6) and reconstituted SCID mice (n 6) 4 weeks after grafting that is, at a time when the grafts were judged clinically to have been rejected. Lymph node cells from naïve BALB/c mice were used as negative controls (n 6). Single-cell suspensions prepared from lymph nodes of individual animals were used as responders ( cells/well) and added to x-irradiated (2000 R) 13 guinea pig stimulator spleen cells ( cells/well) in a final volume of 200 L of culture medium. Triplicate cultures were prepared in 96-well flat-bottomed microculture plates (Corning, Corning, NY). The cultures were incubated at 37 C in an atmosphere of 5% CO 2 for 3 days. Eight hours before termination, the cultures were pulsed with 0.5 Ci [ 3 H]-thymidine, and then harvested onto glass filters using an automated cell harvester (Tomtec, Orange, CT). Radioactivity was assessed by liquid scintillation spectrometry, and the amount expressed as counts per minute (cpm). For anti-cd4 or CD8 blocking experiments, anti-cd4 or anti-cd8 antibody ascitic fluid was added to the cultures at inception. In companion experiments, similar cultures were established and supernatants were collected at 96 hours and IFN- levels were measured using enzyme-linked immunosorbent assay provided by a kit purchased from PharMingen. Coating antibody, detecting antibody and recombinant IFN- as positive control were purchased from PharMingen. These experiments were repeated twice and similar results were obtained. In Vitro Cytotoxicity Assay To assess direct lymphocyte-mediated cytotoxicity by putative effector T cells obtained from SCID mice reconstituted with CD8 T cells or from BALB/c mice that had rejected guinea pig corneas, 51 chromium (Cr)-release assays were performed on day 28 after corneal xenotransplantation. To create suitable target cells, 13 guinea pig spleen cells were co-cultured with 5 g/ml concanavalin A (ConA) for 3 days, then exposed to 51 Cr for 2 hours. After washing, these cells served as targets of the putative effector T cells. Effector T cells were prepared from draining cervical lymph nodes obtained from mice that had rejected orthotopic corneal xenografts. These cells were restimulated in vitro by coculturing them with x-irradiated (2000 R) 13 guinea pig spleen cells for 3 days. Thereafter, the T cells were harvested from these cultures and added to guinea pig Con A blast target cells in triplicate wells at ratios of effector-to-target cells of 6:1, 12:1, 25:1, 50:1, 75:1, and 100:1. Six wells containing medium and target cells were used to measure spontaneous release, and six wells containing 5

3 IOVS, January 2003, Vol. 44, No. 1 Delayed Graft Rejection 177 N HCl and target cells were used to measure maximal radioisotope release. After 4 hours incubation at 37 C in an atmosphere of 5% CO 2, 25 L of culture supernatant was removed from each well, and radioactivity was measured. 23 Statistical Methods Statistical analysis of graft survival, enabling comparison of median survival times (MST), was performed using the Mantel-Cox rank test. Comparison of proliferation assay values was made by the Student s t-test. P 0.05 was deemed significant. RESULTS The goal of these experiments was to create mice deficient in CD4 T cells to determine the extent to which delayed rejection of orthotopic corneal xenografts is mediated by CD8 T cells. Because SCID mice are incapable of rejecting guinea pig cornea grafts, our strategy was to reconstitute SCID mice with various splenic or lymph node cell populations and then to assess the capacity of these mice to reject orthotopic guinea pig cornea grafts. Fate of Guinea Cornea Grafts Placed in Eyes of SCID Mice Reconstituted with BALB/c Spleen Cells Adult SCID mice received intravenous infusions of cell suspensions prepared from spleens of normal BALB/c mice (one donor equivalent/recipient). One day later these mice received an orthotopic guinea pig cornea graft in the right eye. For comparison s sake, panels of normal BALB/c and SCID mice also received orthotopic guinea pig cornea grafts. One day later, peripheral blood was obtained from reconstituted SCID mice, as well as from normal SCID and BALB/c mice. Leukocytes in the blood samples were analyzed by flow cytometry for content of CD4 and CD8 T cells. As the results presented in Figure 1A reveal, peripheral blood from BALB/c mice contained easily identifiable populations of CD4 and CD8 T cells, whereas peripheral blood from 8-week-old SCID mice was devoid of these cells. Peripheral blood from SCID mice that received an infusion of BALB/c spleen cells 24 hours previously contained large numbers of both CD4 and CD8 T cells, indicating successful reconstitution of these adaptive immune cellular elements. The survival of guinea pig cornea grafts was assessed clinically in reconstituted SCID mice and the relevant controls. The day of rejection was found when clinical inspection of the grafted eye failed to reveal an identifiable pupil or detailed structures of the iris. The results of this experiment are displayed in Figure 2A. As expected, untreated SCID mice retained their cornea xenografts in perfect condition throughout the 8-week observation interval, MST 56 days. Cornea xenografts in eyes of BALB/c mice were rejected acutely, as were grafts placed in eyes of SCID mice reconstituted with BALB/c spleen cells. The tempos of these acute rejections were virtually identical (MST 15 and 14 days, respectively,) for reconstituted SCID and untreated BALB/c recipients). Thus, SCID mice recover completely their capacity to reject guinea pig cornea grafts if they are reconstituted with spleen cells from normal BALB/c donors. Fate of Guinea Cornea Grafts Placed in Eyes of SCID Mice Reconstituted with BALB/c Spleen Cells Depleted of CD4 T Cells It has previously been demonstrated that CD4 T cells are responsible for acute rejection of orthotopic guinea pig cornea grafts in normal BALB/c mice, and that mice depleted of CD4 T cells retain the capacity to reject guinea pig cornea grafts, but in a delayed fashion. 17 Our next experiments were designed to determine whether CD8 T cells are responsible for delayed rejection of guinea pig cornea grafts. Adult SCID mice were reconstituted with BALB/c spleen cells from which CD4 T cells were removed by treatment with anti-cd4 antibodies plus complement. One day later peripheral blood was removed from reconstituted mice and, along with peripheral blood from untreated SCID mice, was analyzed for content of CD4 and CD8 T cells. As displayed in Figure 1B (compare the histogram at left with the histogram in middle), peripheral blood samples of SCID mice reconstituted with CD4 T cell depleted spleen cells contained no CD4 cells, but numerous CD8 cells. SCID mice reconstituted in this fashion received orthotopic guinea pig cornea grafts 24 hours later, and the fate of these grafts was assessed clinically. The results presented in Figure 2B indicate that SCID mice reconstituted with CD4 T cell-depleted spleen cells acquired the capacity to reject guinea pig cornea grafts, and that the tempo of rejection was prolonged compared to that of SCID mice reconstituted with whole spleen cells from BALB/c donors (MST 27 vs. 15 days, P 0.04). These results indicate that effectors present in normal BALB/c spleen cells depleted of CD4 T cells possess the capacity to reject guinea pig cornea grafts in the absence of CD4 T cell help. The pattern of rejection observed in these mice is strikingly similar to that reported for CD4 knockout mice that received orthotopic guinea pig cornea grafts (MST 27 days). 17 Fate of Guinea Cornea Grafts Placed in Eyes of SCID Mice Reconstituted with BALB/c Spleen Cells Depleted of CD4 and CD8 T Cells The previous result supports the hypothesis that CD8 T cells, in the absence of CD4 T cells, are capable of causing rejection of guinea pig cornea grafts, albeit in a delayed fashion. If the hypothesis is valid, it should be possible to eliminate the rejection of cornea xenografts in mice reconstituted with CD4 T-cell depleted spleen cells by also eliminating CD8 T cells from the reconstituting inoculum. Accordingly, adult SCID mice received suspensions of splenocytes that had been treated with both anti CD4 and anti CD8 antibodies plus complement. One day later the mice received orthotopic guinea pig cornea grafts, and immediately before the procedure peripheral blood was collected from these mice and analyzed for content of CD4 and CD8 cells. As displayed in Figure 1B (compare histogram at left with the histogram at right), peripheral blood samples from these reconstituted mice contained no detectable CD4 or CD8 cells. Moreover, when the survival of guinea pig cornea grafts was assessed clinically (Fig. 2C), the most grafts were found to survive indefinitely (MST 49 days). This result strongly suggests that the effector modality primarily responsible for delayed rejection of guinea pig grafts is a CD8 T cell. However, this conclusion must be tempered by the observation that a minority (25%) of guinea pig cornea grafts were rejected in SCID mice reconstituted with BALB/c spleen cells depleted of both CD4 and CD8 cells. Fate of Guinea Cornea Grafts Placed in Eyes of SCID Mice Reconstituted with CD8 T Cells Obtained from BALB/c Donors To confirm that CD8 T cells on their own are capable of mounting rejection reactions that destroy orthotopic guinea pig cornea grafts, purified suspensions of CD8 T cells were prepared from lymph nodes and spleens of normal BALB/c mice. As revealed in Figure 1C, peripheral blood of adult SCID

4 178 Higuchi and Streilein IOVS, January 2003, Vol. 44, No. 1 FIGURE 1. Flow cytometric evidence of reconstitution of SCID mice with CD4 and/or CD8 T cells. Peripheral blood leukocytes obtained from SCID mice that received one day previously BALB/c (A) unfractionated spleen cells; (B) splenocytes depleted of CD4 or CD4 /CD8 T cells; and (C) purified lymph node and splenic CD8 T cells. The leukocytes were stained with antibodies directed at CD4 and CD8 molecules and then analyzed by flow cytometry. For positive and negative controls, peripheral blood leukocytes were obtained from normal BALB/c and untreated SCID mice. Histograms of individual assays are presented with CD4 cells along the y- axis and CD8 cells along the x-axis. mice that received purified CD8 T cells one day previously contained large numbers of CD8 cells as detected by flow cytometry. Moreover, mice reconstituted in this fashion proved capable of rejecting orthotopic guinea pig cornea grafts (Fig. 2D). However, a minority of grafts was not rejected at all. On the one hand, these results permit us to conclude that CD8 T cells in the absence of CD4 T cell help are capable in their own right of developing xenograft immunity that leads to graft rejection. On the other hand, these results indicate that the efficiency of rejection mediated solely by CD8 T cells is less than that achieved with spleen cell suspensions depleted of CD4 T cells. This leads to the suspicion that yet another immune effector (non-cd4, non-cd8 ) lurks among spleen cells of normal BALB/c mice cells capable of promoting re-

5 IOVS, January 2003, Vol. 44, No. 1 Delayed Graft Rejection 179 FIGURE 2. Survival patterns of strain 13 guinea pig corneas grafted orthotopically to eyes of SCID mice reconstituted with diverse BALB/c lymphoid cells. Guinea pig corneas were grafted orthotopically to SCID mice that were reconstituted one day previously with BALB/c (A) unfractionated spleen cells; (B) spleen cells depleted of CD4 T cells, (C) spleen cells depleted of CD4 T cells alone, or both CD4 and CD8 T cells; and (D) CD8 T cells purified from lymph node and spleen cell suspensions. Control panels of mice consisted of normal BALB/c and untreated SCID mice. Median survival times of xenografts (A) on BALB/c mice (14 d) and SCID mice reconstituted with whole splenocytes (15 d) are not significantly different; (B) on SCID mice reconstituted with unfractionated BALB/c splenocytes (15 d) versus CD4 T cell-depleted spleen cells (27 d) are significantly different (P 0.004); (C) on SCID mice reconstituted with CD4 T cell depleted splenocytes (27 d) versus CD4 /CD8 T cell depleted splenocytes (49 d) are significantly different (P 0.05); (D) on SCID mice reconstituted with purified BALB/c CD8 T cells are significantly different from grafts placed on untreated SCID mice. jection of corneal xenografts in the absence of either CD4 or CD8 T cells. Cytotoxic Capacity of CD8 T Cells Obtained from Mice that Rejected Guinea Pig Cornea Grafts A primary functional property of CD8 T cells is to recognize immunogenic peptides loaded onto class I major histocompatibility complex (MHC) molecules expressed on target cells as a prelude to delivery of a lytic signal that results in target cell death. We tested to determine whether cell-mediated cytotoxicity is the method by which xenoreactive CD8 T cells bring about cornea xenograft rejection in this system. Ipsilateral cervical lymph node cells were obtained from cervical lymph nodes draining eyes (of normal BALB/c recipients as well as SCID mice reconstituted with CD8 T cells) bearing rejected strain 13 guinea pig cornea grafts that had been in place for 28 days. The CD8 T cells were stimulated in vitro for 3 days by exposure to irradiated strain 13 guinea pig spleen cells. Thereafter, the responding T cells were placed in cytotoxicity assays in which 51 Cr-labeled strain 13 guinea pig spleen cell Con A blasts were used as targets. Irrespective of the source of CD8 T cells, no evidence of cytotoxicity was observed (data not shown). These results suggest that direct cytotoxicity by effector CD8 T cells is not the mechanism by which cornea xenografts are rejected. Capacity of Guinea Pig Cells to Activate In Vitro CD8 T Cells Obtained from Mice that Have Rejected Guinea Pig Cornea Grafts In orthotopic corneal allografts in mice, rejection is achieved largely by CD4 T cells that recognize donor-derived alloantigens by the indirect pathway that is, when processed and presented on recipient APCs that infiltrate the graft. 19 This mechanism has been found to be responsible also for acute rejection of guinea pig cornea xenografts in mice. 18 We next tested whether CD8 T cells from mice that had rejected guinea pig cornea grafts in the absence of CD4 T cells were activated in vitro by guinea pig spleen cells. The results are presented in Figures 3A and 3B. Spleen cells from mice that rejected strain 13 guinea pig cornea grafts proliferated when stimulated with strain 13 guinea pig antigens in vitro. The extent of proliferation by cells removed from CD8 T cellreconstituted SCID mice was considerably less than that of normal BALB c mice, but nonetheless significantly higher than negative control. Moreover, spleen cell suspensions from both types of rejector mice produced IFN- when stimulated with guinea pig spleen cells in vitro. These results indicate that lymphoid cells obtained from mice that reject guinea pig corneas respond to guinea pig xenoantigens in vitro by proliferating and by secreting the proinflammatory cytokine IFN-.

6 180 Higuchi and Streilein IOVS, January 2003, Vol. 44, No. 1 In the final set of experiments, we inquired into the CD4/ CD8 phenotype of the T cells responding in the restimulation cultures. Cultures similar to those described were established and then either anticd4 or anticd8 antibodies were added. After 3 days, the extent of proliferation was once again examined. The results presented in Figure 3C indicate that the responding T cells in suspensions obtained from BALB/c mice that have rejected guinea pig corneas were almost exclusively CD4. By contrast, the responding T cells in suspensions obtained from SCID mice reconstituted with CD8 T cells that rejected guinea pig corneas were almost exclusively CD8. Thus, in the absence of CD4 T cells, xenoreactive CD8 T cells acquire the capacity to become primed to guinea pig xenoantigens and to produce IFN-, the presumed mediator of graft rejection. DISCUSSION FIGURE 3. Capacity of strain 13 guinea pig lymphoid cells to activate T cells obtained from mice that have rejected orthotopic guinea pig cornea grafts. In (A) and (C) [ 3 H]-thymidine was added 8 hr before culture termination at 3 days, after which radioisotope uptake (cpm) was assessed (y-axis). In (B), supernatants were collected and assayed by ELISA for content of IFN- (ng/ml, y-axis). In (C), anti CD4 or anti CD8 antibodies were added at the onset of the cultures. In (A) and (B), [ 3 H]-thymidine incorporation and IFN- content of supernatants, respectively, in cultures containing CD8 T cell reconstituted SCID lymphoid cells is significantly greater than naïve control. (Error bars represent standard deviations.) The pathogenesis of orthotopic corneal xenografts is not completely understood. Early studies in which guinea pig corneas were transplanted orthotopically to rat eyes revealed a tempo of rejection that was not hyperacute, 24 even though the investigators assumed that antibodies were responsible. More recently, Larkin et al. 25 performed similar experiments and found that corneal xenografts were rejected in rat eyes within 2 to 3 days, strongly suggesting the importance of preformed antibodies in graft rejection. By contrast, our experiments over the past several years have come to a different conclusion, using a model in which guinea pig corneas are transplanted into eyes of mice, a species that is discordant that is, the murine serum contains preformed anti-guinea pig antibodies. Our experiments have yielded two significant observations that have encouraged us to continue to study the immunopathogenesis of rejection of orthotopic corneal xenografts. First, guinea pig corneas grafted into the eyes of SCID mice survive indefinitely and maintain their clarity. 17 This indicates that no important nonadaptive immune mechanism exists to limit the success of xenogeneic cornea transplants. Second, guinea pig corneal xenografts placed in eyes of mice with naturally occurring, preformed antibodies against guinea pig xenoantigens suffer no evidence of hyperacute or delayed acute rejections of the antibody-mediated type. 17 As mentioned previously, features that confer immune privilege on the cornea itself, as well as on the anterior chamber into which the graft is placed, undoubtedly account for the relative invulnerability of orthotopic corneal xenografts to antibody mediated destruction. Encouraging as these observations are, they cannot obscure the important finding that guinea pig corneas placed in eyes of immunocompetent mice suffer acute rejection (within 8 20 days) that is mediated almost exclusively by CD4 T cells that recognize guinea pig xenoantigens through the so-called indirect pathway. 6 There is little evidence that immune privilege mitigates in any way the vigor and intensity of CD4 T cell-mediated rejection of orthotopic guinea pig cornea grafts. Nor is there evidence that acute rejection of this type can be controlled by local or systemic immunosuppression with corticosteroids or cyclosporine A, agents that primarily target CD4 T-cell dependent immune rejection. As studies of this type proceed, it is important to determine whether immune effector modalities in addition to CD4 T cells can also cause rejection of orthotopic guinea pig corneas in mice, and if so, what the nature of these effectors might be. The results reported here identify CD8 T cells as capable of effecting destruction of guinea pig corneas grafted into the eyes of mice. Whereas SCID mice reconstituted with BALB/c spleen cells rejected guinea pig corneas acutely (within 20 days) in a manner similar to intact BALB/c recipients, SCID mice reconstituted with BALB/c spleen cells depleted of CD4 T cells or with purified CD8 T cells from BALB/c donors rejected guinea pig corneas in a delayed fashion (between 20 and 40

7 IOVS, January 2003, Vol. 44, No. 1 Delayed Graft Rejection 181 days). A similar delayed rejection pattern was reported recently by Tanaka et al. 17 who studied rejection of guinea pig cornea grafts in C57BL/6 mice in which the CD4 gene had been disrupted. Together these results permit the conclusion that CD8 T cells are sufficient, in the absence of CD4 T cells, to promote rejection of orthotopic guinea pig cornea grafts in mice. One of the unexplained curiosities of immune rejection of orthotopic allogeneic cornea grafts is their relative resistance to rejection by direct alloreactive CD4 and CD8 T cells 20 that is, the effector T cells that reject corneal allografts recognize donor alloantigens chiefly when processed and presented on APCs of recipient origin that infiltrate the graft. The same appears to be true for cornea xenografts. Tanaka et al. 18 have recently reported that CD4 T cells mediated acute rejection of guinea pig cornea grafts exclusively by indirect xenoreactive T cells. Our current results extend this interpretation to the CD8 T cells that effect delayed rejection of guinea pig cornea grafts. We were unable to detect any guinea pig-specific cytolytic T cells in mice that rejected guinea pig cornea grafts. However, we did detect the presence of both CD4 and CD8 T cells that proliferated when stimulated in vitro with guinea pig spleen cells. Moreover, the responding T cells secreted significant amounts of IFN-. These results lead us to speculate that rejection of orthotopic guinea pig cornea grafts in mice is triggered by proinflammatory cytokines (such as IFN- ) that are released from primed guinea pig xenoantigen-specific T cells that respond to guinea pig antigens displayed on recipient APCs that infiltrate the graft. A similar mechanism operates in the rejection of orthotopic corneal allografts in mice, and in this instance infiltrating recipient APCs have been well-documented to arrive in and on the graft shortly before and during the interval when the graft is rejected To that end, Fox et al. 29 showed that, once activated, macrophages can act as direct effectors and the responses of these cells is stronger against xenoantigens than alloantigens. Although our experimental results are the first concerning the contribution of CD8 T cells to corneal xenograft rejection, other investigators examining other types of solid tissue xenografts have found a pathogenic role for CD8 T cells: porcine islet cell clusters to monkeys, 30 porcine or rat pancreas to mice, 31 or hamster heart to rats. 32 Most of these reports showed that CD8 T cells display cytotoxic activity for xenogeneic target cells, via perforin-mediated or Fas-FasL interactions. 33,34 Our results differ, in that we could find no evidence of direct cytolytic activity by in vivo primed, guinea pig xenoantigen-reactive CD8 T cells. A similar result was reported by Zhan et al., 35 who showed that murine CD8 T cells could not lyse porcine splenocytes. Whether this deficiency represents a special property of murine CD8 T cells remains to be determined. The precise mechanisms by which corneal xenografts in particular, and corneal allografts in general, are rejected remain elusive. Graft failure corresponds temporally with deterioration of function and eventual loss of corneal endothelium the posterior layer of cells that is responsible for deturgescing the corneal stroma, thereby maintaining its clarity. Yet direct lysis of allogeneic or xenogeneic corneal endothelial cells by cytotoxic T cells does not appear to take place. In the case of cornea xenografts, we found no evidence of direct xenoreactive cytolytic CD8 T cells. In the case of cornea allografts, direct alloreactive CD8 T cells with cytolytic potential are formed, but appear to play no role in rejection. 23 In the absence of direct cytotoxicity, immune rejection of allogeneic cornea grafts appears to be mediated by proinflammatory cytokines, especially IFN-, that are released by CD4 and/or CD8 T cells stimulated at the graft site by xenoantigenspresenting recipient APCs. Because activated macrophages and dendritic cells accumulate within the stroma and on the apical surface of corneal endothelium of rejecting grafts, 36,37 the nonspecific destructive capacities of activated macrophages, perhaps through release of nitric oxide (NO) and reactive oxygen intermediates (ROI), 38 are thought to be critical. Corneal endothelial cells are especially vulnerable to the deleterious effects of NO and ROI, 39 and this may be why these grafts eventually fail. Several findings reported here deserve special comment. First, CD8 T cells, on their own and without the aid of CD4 T cell help, are capable of causing the rejection of orthotopic guinea pig corneas in mice. This probably reflects the oftenoverlooked capacity of primed CD8 T cells to mediate reactions of the delayed hypersensitivity type. 40 Second, even in the absence of CD4 and CD8 T cells, a small number of orthotopic corneal xenografts undergo rejection. Because untreated SCID mice never reject guinea pig cornea grafts, 17 there must be another adaptive immune effector that can cause rejection on its own. Although we have no knowledge of the nature of this putative effector, we suspect (1) double negative T cells (perhaps cells using the / Tcr, or a subpopulation of NK T cells), and (2) B cells with an antibody product that could arm Fc receptor bearing leukocytes that mediate antibodydependent cell mediated cytotoxicity. Our report 17 that heavy-chain deficient mice still reject guinea pig cornea grafts acutely merely reveals that antibodies are not involved in acute rejection. However, they could be involved in a more desultory rejection process such as that observed in mice deficient in both CD4 and CD8 T cells. Finally, we would point out that the discovery of a potent, cornea xenograft-rejecting role for CD8 T cells is important on at least two accounts. On the first, CD8 T cells have been found to play virtually no role in rejection of orthotopic corneal allografts in mice. 23 Thus, something is fundamentally different about corneal xenografts that renders these grafts, but not their allogeneic counterparts, vulnerable to rejection by CD8 T cells. On the second, the most immunosuppressive regimens now used to prevent rejection of solid tissue allografts in the clinic are directed at CD4 T cells, with little or no capacity to suppress CD8 T cell activity. Our current results indicate that immunosuppressive regimens that would be developed to suppress rejection of corneal xenografts must be effective at inhibiting both CD4 and CD8 T cells, or graft rejection will be the unfortunate outcome. Acknowledgments The authors thank Jacqueline M. Doherty and Sharmila Masli for their kind help in the experiments, Patrick Michael Stuart for the kind gift of CD4 KO mice, and Marie Ortega and her staff for excellent care of the experimental animals. References 1. Coster DJ, Williams KA. Donor cornea procurement: some special problems in Asia. Asia Pacific J Ophthalmol. 1992;4: Auchincloss H Jr. Xenogeneic transplantation a review. Transplantation. 1998;46: Zhang Z, Bedard E, Luo Y et al. Animal models in xenotransplantation. Expert Opin Investig Drugs. 2000;9: Streilein JW. Immunobiology and immunopathology of corneal transplantation. In: Streilein JW, ed. Immune Response and the Eye. Basel, Switzerland: Karger; 1999: Bora NS, Gobleman CL, Atkinson JP, Pepose JS, Kaplan HJ. Differential expression of the complement regulatory proteins

8 182 Higuchi and Streilein IOVS, January 2003, Vol. 44, No. 1 in the human eye. Invest Ophthalmol Vis Sci. 1993;34: Liu L, Sun TK, Xi YP, Maffei A, Reed A, Harries P, Suciu-Foca N. Contribution of direct and indirect recognition pathways to T cell alloreactivity. J Exp Med 1993;177: Lee RS, Grusby MJ, Glimcher LH, Winn HJ, Auchincloss H. Indirect recognition by helper cells can induce donor-specific cytotoxic T lymphocytes in vivo. J Exp Med 1994;179: Streilein JW. Regional immunity and ocular immune privilege. In: Streilein JW, ed. Immune Response and the Eye. Basel, Switzerland: Karger; 1999: Goslings WRO, Prodeus AP, Streilein JW, et al. A small molecular weight factor in aqueous humor acts on C1q to prevent antibodydependent complement activation. Invest Ophthalmol Vis Sci. 1998;39: Hori J, Joyce, N, Streilein JW. Corneal allografts placed beneath the kidney capsule display inherent immune privilege. Invest Ophthalmol Vis Sci. 2000;41: Kawashima H, Prasad SA, Gregerson DS. Corneal endothelial cells inhibit T cell proliferation by blocking IL-2 production. J Immunol. 1994;153: Jager MJ, Bradley D, Streilein JW. Immunosuppressive properties of cultured human cornea and ciliary body in normal and pathological conditions. Transpl Immunol. 1995;2: Stuart PM, Griffith TS, Usui N, Pepose J, Yu X, Ferguson TA. CD95 ligand (FasL)-induced apoptosis is necessary for corneal allograft survival. J Clin Invest. 1997;99: Yamagami S, Kawashima H, Tsuro T, et al. Role of Fas-Fas ligand interactions in the immunorejection of allogeneic mouse corneal transplants. Transplantation 1997;64: Tanaka K, Yamada J, Joyce N, et al. Immunobiology of xenogeneic cornea grafts in mouse eyes. 1: fate of xenogeneic cornea tissue grafts implanted in anterior chamber of mouse eyes. Transplantation. 2000;69: Tanaka K, Yamada J, Joyce N, et al. Immunobiology of xenogeneic cornea grafts in mouse eyes. 2: immunogenicity of xenogeneic cornea tissue grafts implanted in anterior chamber of mouse eyes. Transplantation. 2000;69: Tanaka K, Yamada J, Streilein JW. Xenoreactive CD4 T cells and acute rejection of orthotopic guinea pig cornea in mice. Invest Ophthalmol Vis Sci. 2000;41: Tanaka K, Sonoda K, Streilein JW. Acute rejection of orthotopic corneal xenografts in mice depends on CD4 T cells and selfantigen-presenting cells. Invest Ophthalmol Vis Sci. 2001;42: Sano Y, Ksander BR, Streilein JW. Murine orthotopic corneal transplantation in high-risk eyes. Rejection is dictated primarily by weak rather than strong alloantigens. Invest Ophthalmol Vis Sci. 1997;38: Sano Y, Streilein JW, Ksander BR. Detection of minor alloantigenspecific cytotoxic T cells after rejection of murine orthotopic corneal allografts: evidence that graft antigens are recognized exclusively via the indirect pathway. Transplantation. 1999;68: Hu H, Stein-Streilein J. Hapten-immune pulmonary interstitial fibrosis (HIPIF) in mice requires both CD4 and CD8 T lymphocytes. J Leukoc Biol. 1993;54: Asea A, Stein-Streilein J. Signaling through NK1.1 triggers NK cells to die but induces NK T cells to produce interleukin-4. Immunology. 1998;93: Yamada J, Ksander BR, Streilein JW. Cytotoxic T cells play no essential role in acute rejection of orthotopic corneal allografts in mice. Invest Ophthalmol Vis Sci. 2001;42: Ross JR, Howell DN, Sanfilippo FP. Characteristics of corneal xenograft rejection in a discordant species combination. Invest Ophthalmol Vis Sci. 1993;34: Larkin DFP, Takano T, Standfield SD et al. Experimental orthotopic corneal xenotransplantation in the rat. Transplantation. 1995;60: Sano Y, Ksander BR, Streilein JW. Langerhans cells, orthotopic corneal allografts, and direct and indirect pathways of T cell allorecognition. Invest Ophthalmol Vis Sci. 2000;41: Dana MR, Yamada J, Streilein JW. Topical interleukin-1 receptor antagonistic promotes corneal transplant survival. Transplantation. 1997;63: Ross J, He Y, Mellon J, Niederkorn JY. The differential effects of donor versus host Langerhans cells in the rejection of MHCmatched corneal allografts. Transplantation. 1991;52: Fox A, Mountford J, Braakhuis A, et al. Innate and adaptive immune responses to nonvascular xenografts: evidence that macrophages are direct effectors of xenograft rejection. J Immunol. 2001;166: Soderlund J, Wennberg L, Castanos-Velez E, et al. Fetal porcine islet-like cell clusters transplanted to cynomolgus monkeys. Transplantation. 1999;67: Yi S, Feng X, Hawthorne W, et al. CD8 T cells are capable of rejecting pancreatic islet xenografts. Transplantation. 2000;70: Lin Y, Soares M P, Sato K, et al. Rejection of cardiac xenografts by CD4 or CD8 T cells. J Immunol. 1999;162: Smyth MJ, Sutton VR, Kershaw MH, Trapani JA. Xenospecific cytotoxic T lymphocytes use perforin- and Fas-mediated lytic pathways. Transplantation. 1996;62: Yi S, Feng X, Wang Y, et al. CD4 cells play a major role in xenogeneic human anti-pig cytotoxicity through the FAS/FAS ligand lytic pathway. Transplantation. 1999;67: Zhan Y, Brady JL, Sutherland RM, et al. Without CD4 help, CD8 rejection of pig xenografts requires CD28 costimulation but not perforin killing. J Immunol. 2001;167: Papulnock DM. Macrophage activation by T cells. Curr Opin Immunol. 1992;4: Junko H,Streilein JW. Dynamics of donor cell persistence and recipient cell replacement in orthotopic corneal allografts in mice. Invest Ophthalmol Vis Sci. 2001;42: Janeway CA. Macrophage activation by armed CD4 TH1 cells. Immunobiology 5th ed. New York: Garland Publishing; 2001: Yanagiya N, Akiba J, Kado M, Hikichi T, Yoshida A. Effects of peroxynitrite on rabbit cornea. Graefes Arch Clin Exp Ophthalmol. 2000;238: Rosenberg AS, Singer A. Cellular basis of skin allograft rejection: an in vivo model of immune-mediated tissue destruction. Annu Rev Immunol. 1992;10: