T-Cell Mediated Responses in a Murine Model of Orthotopic Corneal Transplantation

Transcription

1 T-Cell Mediated Responses in a Murine Model of Orthotopic Corneal Transplantation Choun-Ki Joo*% Jay S. Pepose,*-\ and P. Michael Stuart*% Purpose. To evaluate the role that delayed-type hypersensitivity (DTH) and cytotoxic T lymphocyte responses play in a murine model of orthotopic corneal allograft transplantation. Methods. Corneal transplantation was performed by grafting C57BL/6 donor corneas into BALB/c corneal beds. After transplantation, the mice were observed by slit lamp biomicroscopy on a weekly basis and graded for signs of graft rejection and delayed-type hypersensitivity (DTH) and cytotoxic T lymphocyte (CTL) responses to donor alloantigens assessed at selected times after grafting. Results. It was determined that between 40% and 65% of BALB/c mice rejected C57BL/6 corneas by 8 weeks after engraftment. Mice with opacity scores >2 demonstrated significantly greater DTH responses than did mice with opacity scores <2 at 2, 3, and 4 weeks after engraftment. After 4 weeks, the DTH responses for all groups were essentially the same as for naive BALB/c mice. The DTH responses were specific for C57BL/6 alloantigens and are primarily directed against non-major histocompatibility complex (MHC) C57BL/6 alloantigens and are primarily directed against non-major histocompatibility complex C57BL/6 alloantigens, as evidenced by the ability of B10.D2 cells to elicit DTH responses whereas C.B10- H-2 b cells did not. However, although BALB/c mice engrafted with C57BL/6 tail skin demonstrated significantly greater CTL activity than naive BALB/c mice, there was no significant difference in CTL activity between BALB/c mice whose C57BL/6 corneal allografts displayed opacity scores greater than (rejected) or less than (accepted) 2. Conclusions. The mechanism whereby corneal allografts in this strain combination are rejected is best associated with die ability to generate strong DTH responses and not CTL activity. This DTH response also demonstrates alloantigen specificity and appears to be primarily directed against the non-mhc component of the corneal transplant. Invest Ophthalmol Vis Sci. 1995; 36: JM.ore than 43,000 corneal transplants are performed each year in the United States, making it second only to bone transplantation as the most common form of transplantation. 1 The outcome of corneal transplants placed into nonvascularized corneal beds differs From the Departments of * Ophthalmology and Visual Sciences, ^Pathology, and % Molecular Microbiology, Washington University School of Medicine, St. Louis, Missouri, and the ^Department of Ophthalmology, Catholic University Medical College, Seoul, Korea. Supported by United States Public Health grant P30 EY-02687, a grant from Korea Science and Engineering Foundation , and the Mid-America Transplant Association and by unrestricted grants from Research to Prevent Blindness. PMS is a recipient of an Arthritis Foundation Investigator Award and a Research to Prevent Blindness Career Development Award. Submitted for publication November 10, 1994; revised fanuary 11, 1995; accepted February 27, Proprietary interest category: N. Reprint requests: P. Michael Stuart, Department of Ophthalmology and Visual Sciences, Campus Box 8096, Washington University Medical Center, 660 South Euclid Avenue, St. l^ouis, MO greatly from those transplanted into hosts in whom the corneal bed is vascularized. When the corneal bed is nonvascularized, the 5-year success rate is particularly impressive and may be as high as 90%. 2 ~ 5 However, when corneal transplantation is performed on vascularized, inflamed eyes, the success rate falls dramatically and can be as low as 20% to 40%. 6 This high failure rate in the group at high risk is, in large part, caused by allograft rejection, which may occur despite the use of immunosuppressive agents, such as cyclosporin A and corticosteroid therapy. 7 ' 8 The unusually high success rate for most corneal transplants into avascular beds is likely the result of multiple factors. The cornea and the anterior chamber of the eye generally are considered immunologically privileged sites. 9 ' 10 This immune privilege status derives, in large part, from the unique regional micro Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 Copyright Association for Research in Vision and Ophthalmology

2 T-Cell Responses in Murine Corneal Transplantation 1531 environment of the anterior chamber and the lack of corneal vascularization, limiting both afferent and efferent limbs of the immune response. 9 "" Aqueous humor contains various immunosuppressive cytokines, such as TGF-/?, a-melanocyte stimulating hormone, and vasoactive intestinal peptide, known to downregulate cellular immune responses. 11 " 13 The altered immune response toward antigens introduced into the anterior chamber is characterized by depressed DTH with normal or elevated antibody responses and has been termed anterior chamber associated immune deviation. 14 " 17 The two primary T-cell dependent immune effector mechanisms that have been implicated in the rejection of solid tissue grafts are CTL and DTH responses. The role that these effector mechanisms play in corneal graft rejection remains controversial. Some investigators 1819 have suggested that the ability to generate strong CTL responses correlates with corneal graft rejection, whereas DTH responses play a minor or an insignificant role. In contrast, studies on the anterior chamber associated immune deviation-inducing capacity of corneal transplants imply that the critical effector mechanism in corneal allograft rejection is DTH. 20 ' 21 In the current study, we evaluated the role that DTH and CTL responses play in orthotopic corneal transplantation using mice mismatched at major and minor histocompatibility antigens. The results indicate that rejection of these allogeneic corneas best correlates with the ability of these mice to generate strong DTH responses directed against minor histocompatibility antigens, not with CTL responses. MATERIALS AND METHODS Mice Female C57BL/6 (H-2 b ) and BALB/c (H-2 d ) mice were purchased from the National Cancer Institute (Frederick, MD). The NIH mice were purchased from Harlan Olac UK (Bicester, UK). The C.B10-H-2 b (H- 2 b ) and B10.D2 (H-2 d ) mice were purchased from Jackson Labs (Bar Harbor, ME). These animals were maintained in our mouse facility and were used at 7 to 10 weeks of age for all experiments described. All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Orthotopic Corneal Grafting Corneal transplantation was performed as previously described with some modifications. 22 " 24 Briefly, donor and recipient mice were anesthetized by intraperitoneal injection of Nembutal sodium solution (phenobarbital, 50 mg/kg). Donor corneal grafts were prepared by removing a central 2-mm corneal button from C57BL/6J mice (allogeneic graft) or from BALB/c mice (syngeneic graft) using an inner-edged trephine (Storz, St. Louis, MO) and Vannas scissors (Storz). The corneal button was placed in chilled Optisol solution (Chiron, Irvine, CA) until the recipient corneal graft bed was prepared. The recipient corneal graft bed was prepared by removing a central 2-mm corneal button from BALB/c recipient mice using an outer-edged trephine (Storz) and Vannas scissors. Once the graft bed was prepared, the donor corneal button was immediately applied to the graft bed and secured in place with 10 interrupted sutures (11-0 monofilament nylon, 70-//m diameter needle, Sharppoint; Vanguard, Houston, TX). After surgery, each recipient mouse was placed in a separate cage for 1 day to eliminate mouse-to-mouse contact, which might have damaged the engrafted cornea during anesthetic recovery and led to technical failure of the corneal grafts. Ocumycin ointment (Pharmafair Hauppauge, NY) was placed on the corneal surface for 8 days after surgery. All grafted eyes were examined with a vertically mounted slit lamp biomicroscope (Marco, Jacksonville, FL), and sutures were removed 7 days after surgery. At this time, recipient mice with signs of surgical complications such as hyphema, cataract, shallow or flat anterior chamber, or an opacity score of 4 were excluded from the study. Orthotopic Skin Grafting Skin grafting was performed using tail skin-to-tail skin grafts according to previously described techniques. 25 BALB/c mice were recipients of tail skin from C57BL/ 6 mice. These grafts were evaluated at 4 days for technical failures. Those mice whose grafts were determined to be technical failures were excluded from any subsequent studies. All C57BL/6 tail skin grafts were rejected by day 15 after transplantation. Evaluation and Scoring of Orthotopic Cornea and Skin Transplants After transplantation, grafts were examined using a vertically mounted slit lamp biomicroscope (slit beam angle 30, length 14) at weekly intervals. At each time point, the corneal grafts were scored by two masked observers for opacity and neovascularization as described by Sonoda and Streilein. 24 The semiquantitative grading scales used to evaluate corneal opacity and neovascularization for the corneal grafts are shown in Table 1. In agreement with previous studies, we considered that any corneal grafts with an opacity score >2 were undergoing an inflammatory reaction. 24 If the opacity score was >28 weeks after surgery, we defined those grafts as rejected. Histopathologic Evaluation of Corneal Grafts At various times throughout the study, engrafted eyes were enucleated, fixed, embedded in paraffin, sec-

3 1532 Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 TABLE l. Clinical Scoring System for Orthotopic Corneal Grafts Score Opacity Neovascularization Clinical Appearance Clear graft Minimal, superficial opacity not involving corneal stroma Minimal to moderate opacity including corneal stroma Moderate stromal opacity, where pupil is visible and iris is obscured Intense stromal opacity, where only outline of pupil is visible Severe stromal opacity, where pupil is not visible No neovascularization Vessels in <50% of recipient graft bed only Vessels in >50% of recipient graft bed only Vessels in <50% of recipient-graft border Vessels in >50% of recipient-graft border Vessels in <50% of peripheral graft stroma Vessels in >50% of peripheral graft stroma Vessels in <50% of central graft stroma Vessels in >50% of central graft stroma tioned, and stained with hematoxylin and eosin. These sections were then evaluated by light microscopy. Delayed-Type Hypersensitivity Assay Delayed-type hypersensitivity responses were determined by measuring footpad swelling after antigenic challenge. On selected BALB/c mice, 1 X 10 7 gammairradiated (3000 R) C57BL/6 spleen cells (30 fil) were injected subcutaneously into the right hind foot pad. As a negative control, irradiated BALB/c spleen cells were injected into the left hind foot pad. NIH (H-2 C ) spleen cells were used to determine the antigen specificity of the response. Both footpads were measured with an engineer's micrometer (Mitutoyo, Tokyo, Japan) before spleen cell injection to determine time 0 footpad thickness, and they were remeasufed at 24 hours and 48 hours after the injection. Footpad swelling was calculated as follows: Experimental swelling at 24- or 48-hours is corrected by the footpad thickness at 0-hour in the right footpad, and this value is further corrected by the value for control swelling, which is the swelling of the left footpad at 24- or 48-hours corrected by the 0-hour thickness of the left footpad. The corrected result is then expressed in 10" 3 inches. Cytotoxic T-Cell Lympholysis Assay Cytotoxic T-lymphocyte assays were performed as described. 26 Briefly, spleens were removed from BALB/ c mice, and single-cell suspensions were produced. Spleen cells were resuspended in Dulbecco's minimum essential medium supplemented with 10% fetal bovine serum, 2-mm L-glutamine, 5 X 10~ 5 M betamercaptoethanol, 10 mm HEPES buffer, 100 IU penicillin, 100 fig streptomycin/ml at a concentration of 25 X 10 6 cells/ml. These cells either were used direcdy in a CTL assay or were further stimulated with irradiated C57BL/6 spleen cells (12.5 X 10 6 cells/ml). C57BL/6 derived EL-4 (H-2 b ) cell line for allogeneic targets and DBA/2 derived P815 (H-2 d ) mastocytoma for syngeneic targets. These target cells were harvested and resuspended in complete medium at a concentration of 5 X 10 6 cells/ml and labeled with [ 51 Cr] sodium chromate (Amersham, Arlington Heights, IL) at a concentration of 10 7 cells/ml for 1 hour at 37 C. Effector cells and labeled target cells (2.5 X 10 4 cells/ determination) were plated on round-bottom microtiter plates in triplicate, and effectontarget ratios were 100:1, 50:1, 25:1, and 5:1. The plates were incubated at 37 C in 5% CO 2, 100% humidity, for 4 hours. At the end of the incubation period, 100 /A of supernatant was harvested by centrifugation (10 minutes at looog), and the released 51 Cr determined using a gamma counter (Gamma 5500B; Bechman, Fullerton, CA). For secondary CTL assay, BALB/c spleen cells (2.5 X 10 6 /ml) were stimulated with irradiated (2000R) C57BL/6 spleen cells (1.25 X,10 6 /ml) for 4 days at 37 C. At the end of this incubation period, activated cells were harvested by passing over lymphocyte separation media and tested in a CTL assay as described above. Specific cytotoxicity of each sample was determined by the following formula: % Specific 51 Cr release = (Experimental cpm Spontaneous release cpm)/(maximum release cpm Spontaneous cpm) X 100 ± SEM. Maximum 51 Cr release was determined by measuring the amount of 51 Cr released into the medium by 51 Cr-labeled targets incubated with 10% Triton X-100 (Sigma, St. Louis, MO) for 4 hours. Spontaneous 51 Cr release was determined by culturing 51 Cr-labeled target cells in a culture medium only for 4 hours and

4 T-Cell Responses in Murine Cornea! Transplantation 1533 FIGURE l. Morphologic appearance of C57BL/6 cornea! allografts placed onto BALB/c recipient mice, {top left} Photomicrograph of nonrejected C57BL/6 allograft at 6 weeks, {top right) Photomicrograph of rejected C57BL/6 allograft at 6 weeks, {bottom left) Histologic section of nonrejected, neovascularized graft at 6 weeks (original magnification, X250). {bottom right) Histologic section of rejected C57BL/6 comeal allograft at 3 weeks. Notice the dense mononuclear cell infiltrate in the comeal stroma (original magnification, X250). measuring the amount of 51Cr released into the medium in the absence of effector cells. Statistical Analysis Unpaired Student's Rests were used to determine the statistical significance of the differences between nongrafted or syngeneic corneal grafting (negative control), accepted corneal graft (opacity score < 2), rejected corneal graft (opacity score > 2), and tail skin grafting (positive control). RESULTS Clinical Observations of Syngeneic and Allogeneic Orthotopic Corneal Grafts and Allogeneic Skin Grafts We investigated the pattern of rejection seen in orthotopic corneal transplantation between C57BL/6 corneal donors and BALB/c recipients. It should be pointed out that some corneal allografts that initially demonstrated opacity scores > 2, possibly from edema, subsequently became clear and remained clear indefinitely. We and others have observed that the corneal opacity scores at 7 to 8 weeks after engraftment do not change in any appreciable manner when observed at later times. Representative of these observations are data in Figure 1, which displays representative eyes from BALB/c mice with an accepted C57BL/6 corneal allograft (Fig. 1A) and those with rejected C57BL/6 corneal allografts (Fig. IB). Also displayed are the results of histologic examination of the corneas from accepted (Fig. 1C) and rejected (Fig. ID) C57BL/6 corneas. Consequently, we considered transplanted corneas at 7 weeks or later with an opacity score < 2 to be accepted, whereas those that consistently displayed an opacity score > 2 were designated as rejected (Table 1). Using the criteria for corneal opacity and vascularization outlined in Table 1, we determined corneal acceptance-rejection status at 8 weeks after engraftment of C57BL/6 donor corneas into BALB/c mice. Once it was determined which mice had rejected and which had accepted the engrafted corneas, we compared the opacity scores during the first 8 weeks

5 1534 Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 Isograft -- Accepted -*- Rejected WEEKS POST ENGRAFTMENT FIGURE 2. Corneal opacity scores for BALB/c mice engrafted orthotopically with C57BL/6 donor corneas and compared with BALB/c recipients of syngeneic BALB/c donor cornea controls. Using the scoring system described in Table 1, corneal grafts were evaluated weekly, and the groups were determined by the opacity score of the corneal graft at 8 weeks (rejected: opacity score = >2; accepted: opacity score = = 2). The values displayed on the graph are the mean ± SEM for BALB/c mice receiving BALB/c corneas (syngeneic, n = 10), for BALB/c mice diat accepted C57BL/6 corneal allografts (accepted, n = 24), and for BALB/c mice that rejected C57BL/6 corneal allografts (rejected, n = 17). after engraftment for those mice that rejected C57BL/ 6 corneas (opacity score > 2 at 8 weeks) with those mice that accepted C57BL/6 corneas (opacity score < 2 at 8 weeks) and with BALB/c mice engrafted with syngeneic BALB/c corneas (Fig. 2). It should be noted that neovascularization scores were not particularly helpful in determining the fate of corneal grafts, because most mice, including those with syngeneic grafts, displayed neovascularization scores > 5 8 weeks after surgery (data not shown). These observations are consistent with previous reports that documented neovascularization in corneas with clear grafts and intact corneal endothelium and thus were not rejected. 24 The opacity scores for these three groups were the same at 1 week. After 1 week, BALB/c mice engrafted with allogeneic C57BL/6 corneas consistently displayed greater opacification than BALB/c mice engrafted with syngeneic BALB/c corneas (Fig. 2). When BALB/c mice engrafted with allogeneic corneas were separated into those that ultimately accepted or rejected the C57BL/6 corneas and compared statistically, significant differences in opacity scores were apparent by 4 weeks and continued until the conclusion of the experiment. The rate of rejection in this study was 42%, with 17 rejections out of 41 mice grafted. However, subsequent experiments consistently have demonstrated rejection rates between 55% and 65%. Consequently, the overall rejection rate for BALB/c mice engrafted with C57BL/6 corneas ranged between 40% and 65%. There were two general patterns of clinical responses of BALB/c mice toward C57BL/6 corneal allografts. For mice that ultimately accepted C57BL/6 corneal allografts, 29% showed no signs of corneal opacity during the 8 weeks after transplantation. On the other hand, most allogeneic grafts that were ultimately accepted (71%) developed transient opacity (opacity scores > 2) at least once during the 8-week period of the experiment, which typically did not last more than 2 weeks and did not compromise the final outcome of the corneal allograft. A proportion of BALB/c mice (38%) engrafted with syngeneic BALB/ c corneas also displayed a transient opacity, which disappeared within 2 weeks and was restricted to peripheral cornea near suture sites (data not shown). Similarly, BALB/c mice that rejected C57BL/6 allogeneic grafts demonstrated two patterns of reactivity. Most of the BALB/c mice that rejected C57BL/6 corneal allografts (82%) did so within 4 weeks of transplantation, whereas the remaining mice in this group (18%) did not display any signs of rejection until 4 weeks after surgery, at which time they were rejected rapidly. These results indicate that most C57BL/6 corneal allografts undergo some form of inflammatory response that may be representative of a rejection reaction. BALB/c mice transplanted with C57BL/6 allogeneic tail skin rejected 100% of these skin grafts by the middle of the third week after surgery and had a mean graft survival of between 12 and 14 days (data not shown). Induction of Alloantigen-Specific Delayed-Type Hypersensitivity in Mice With Orthotopic Allogeneic Cornea Transplantation To determine whether the corneal allograft rejection in our model was associated with the activity of CTLs or with DTH-mediating T cells, we performed the following experiments. Female BALB/c mice were engrafted orthotopically with corneas from female C57BL/6 (allogeneic grafts) or female BALB/c (syngeneic grafts) donors, and the opacity scores were determined on a weekly basis. At weekly intervals, five mice whose corneal grafts displayed opacity scores > 2 and five mice whose opacity scores were <2 were tested for their ability to generate DTH and CTL responses. Negative controls for these experiments initially consisted of testing DTH and CTL responses in two BALB/c mice engrafted with syngeneic BALB/c corneas and three naive BALB/c mice not grafted. Results from these assays indicated that both BALB/ c mice engrafted with syngeneic corneas and naive BALB/c mice displayed indistinguishable DTH and CTL responses (data not shown). Therefore, all subsequent experiments used naive mice as negative con-

6 T-Cell Responses in Murine Cornea! Transplantation 1535 u (/> o < a. O u. FIGURE 3. Comparison of delayed-type hypersensitivity (DTH) responses from mice with corneal allografts. At weekly intervals, BALB/c mice engrafted with C57BL/6 tail skin (skin graft, n = 4 to 5 mice/time point) engrafted with C57BL/6 corneas. Those with opacity scores of >2 (rejected, n = 4 to 7 mice/time point) or of <2 (accepted, n = 4 to 7 mice/time point) and ungrafted mice (naive, n = 4 mice for all time points) were monitored for their ability to generate DTH responses toward C57BL/6 alloantigens. Results shown are the mean ± SEM footpad swelling measured at 24 hours after footpad challenge. trols. The positive controls for these experiments consisted of five BALB/c mice transplanted with full-thickness C57BL/6 tail skin. Delayed-type hypersensitivity responses were measured by injecting the footpads of these mice with 10 7 spleen cells from either C57BL/6 or BALB/c, and the swelling response was measured 24 and 48 hours later. BALB/c mice engrafted with C57BL/6 corneas that demonstrated opacity scores > 2, and thus were considered to be undergoing a graft rejection episode, generated significantly greater DTH responses (P < 0.05) than did BALB/c mice that did not demonstrate signs of rejecting their C57BL/6 corneal allografts or negative control BALB/c mice at 2, 3, and 4 weeks after surgery (Fig. 2). However, by the fifth week after engraftment, there were no significant differences in DTH responses between any of the groups tested (Fig. 3). Maximal DTH responses for BALB/c mice engrafted with C57BL/6 corneas occurred at approximately 3 weeks, whereas the peak DTH response for BALB/c mice engrafted with C57BL/6 tail skin grafts occurred at 2 weeks after engraftment (Fig. 3), corresponding to the tempo of allograft rejection at these disparate sites and tissues. To define better these DTH responses, we determined whether they were specific for the alloantigens expressed by C57BL/6 cells. Consequently, we engrafted BALB/c mice with C57BL/6 corneas, and at 3 and 4 weeks after surgery, we measured DTH responses after footpad injection with C57BL/6 spleen cells (specific alloantigen), BALB/c spleen cells (syngeneic, negative control), and NIH spleen cells (nonspecific alloantigen, third party). These time points were chosen because maximal DTH for BALB/c mice that reject C57BL/6 corneas occurs between 3 and 4 weeks after engraftment (Fig. 3). Three types of injection protocols were used in these studies. One group of mice was injected into the right hind footpad with C57BL/6 cells and the left hind footpad with BALB/ c cells (B6 versus BALB). A second group was injected with NIH cells into the right hind footpad and BALB/ c cells into the left hind footpad (NIH versus BALB). In the final group, C57BL/6 cells were injected into the right hind footpad and NIH into the left hind footpad (B6 versus NIH). By using this injection scheme, we were able to determine whether NIH cells elicited a swelling response on their own and whether the difference between the swelling response elicited by C57BL/6 and NIH cells was essentially the same as the difference between the swelling response elicited by C57BL/6 and BALB/c cells. Results from these studies indicate that neither ungrafted BALB/c mice nor BALB/c mice that accepted C57BL/6 corneas displayed significant swelling responses toward either NIH or C57BL/6 cells (Fig. 4). In contrast, BALB/c mice that reject either C57BL/6 corneal allografts or Ungrafted Accepted Rejected GROUPS Skin Graft FIGURE 4. Delayed-type hypersensitivity responses of BALB/ c mice toward C57BL/6 are antigen specific. BALB/c mice engrafted with C57BL/6 corneas were challenged 3 weeks after corneal engraftment with irradiated spleen cells from BALB/c (syngeneic), C57BL/6 (specific alloantigen), or NIH (nonspecific alloantigen) mice and footpad swelling measured 24 hours after injection. Three types of injection schemes were used: One group of mice was injected in the right hind footpad with C57BL/6 cells and the left hind footpad with BALB/c cells (B6 versus BALB). A second group was injected with NIH cells into the right hind footpad and BALB/c cells in the left hind footpad (NIH versus BALB). The final group consisted of injecting C57BL/6 cells in the right hind footpad and NIH in the left hind footpad (B6 versus NIH). Values shown represent the mean of the differences in footpad swelling of the indicated injected cell types (right hind footpad swelling left hind footpad swelling) ±SEM of at least five mice per group.

7 " * 1536 Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 o \D SWEL.LING x 1 F Hi C57BL/6 Hii B10.D2 control T accepted rejected skin graft FIGURE 5. Delayed-type hypersensitivity responses of BALB/ c mice toward C57BL/6 are directed primarily against non- MHC alloantigens. Three weeks after transplantation of BALB/c mice with C57BL/6 corneal allografts, they were challenged with irradiated spleen cells from BALB/c (syngeneic), C57BL/6 (alloantigen), B10.D2 (non-mhc-only-specific alloantigen), or C.BlO-H-2 b (MHC-only-specific alloantigen) mice and footpad swelling measured 24 hours after injection. Values shown represent the mean swelling response ± SEM for at least five mice per group. skin allografts demonstrate significant DTH responses toward C57BL/6 cells but not toward NIH spleen cells (Fig. 4). The next series of experiments was designed to determine against which general class of alloantigens these DTH responses were directed. Because earlier studies 24 have suggested that corneal allograft rejection is greatest when donor and recipient are mismatched for non- MHC gene products, we decided to test whether the DTH response elicited by C57BL/6 corneas was primarily directed against MHC alloantigens (H-2 b ) or non- MHC alloantigens. In these experiments, we engrafted BALB/c mice with C57BL/6 corneas, C57BL/6 tail skin, or nothing. Three weeks after engraftment, these mice were challenged by footpad injection with irradiated spleen cells from C57BL/6 (whole strain difference), C.BlO-H-2 b (MHC only difference), B10.D2 (non-mhc only difference), or BALB/c (syngeneic) mice. Footpad swelling was then measured 24 and 48 hours after footpad injection. The DTH responses for BALB/c mice engrafted with C57BL/6 tail skin toward C57BL/6, C.BlO-H-2 b, and B10.D2 spleen cells did not demonstrate any significant differences between these groups (Fig. 5), indicating that engrafting BALB/c mice with C57BL/6 skin grafts stimulated MHC-specific and non- MHC-specific T cells capable of generating significant DTH responses. In contrast, BALB/c mice in the process of rejecting C57BL/6 corneas were only able to mount significant DTH responses against C57BL/6 and B10.D2 spleen cells (Fig. 5). These data suggest that the dominant type of T cell capable of mediating antigen-specific DTH response in BALB/c mice engrafted with C57BL/ 6 corneas are specific for non-mhc antigens. We concluded that a significant DTH response was detected in animals with rejected allogeneic corneal graft (antigen discrepancy at major and minor histocompatibility loci) and that the response was an alloantigen specific event. Secondary Cytotoxic T Lymphocyte Response Induced in Orthotopic Corneal and Skin Allografts Cytotoxic T lymphocyte responses also were measured at weekly intervals after corneal or skin engraftment by isolating single cell suspensions from the spleens of all the same groups of mice described for DTH experiments. Initial studies determined the ability of these spleen cells to lyse 5l Cr-labeled EL-4 (C57BL/ 6) or P-815 (BALB/c) target cells. Results indicated that few mice, even from those engrafted with C57BL/ 6 tail skin, were able to lyse EL-4 targets at an effector:target (E:T) ratio of 100:1 (data not shown). Consequently, we set up in vitro stimulation cultures in which spleen cells from individual mice representing naive BALB/c mice (control), BALB/c mice engrafted with C57BL/6 corneas whose corneal opacity was :s2 (accepted), those whose corneal opacity was >2 (rejected), and BALB/c mice engrafted with C57BL/6 tail skin (skin graft) were incubated with irradiated spleen cells from C57BL/6 mice for 4 days. The purpose of these cultures was to expand and generate additional T cells capable of lysing C57BL/6 targets. After this in vitro stimulation, viable cells were harvested and tested for their ability to lyse 5l Cr-labeled EL-4 and P-815 target cells. Initially, we tested CTL responses in BALB/c mice that were engrafted with syngeneic BALB/c corneas. Because these responses were indistinguishable, at all E:T ratios used, from naive BALB/c mice (data not shown), we used naive BALB/c mice as controls for all subsequent experiments. Results from these studies indicated that there were no significant differences in the percent specific lysis at a E:T ratio of 100:1 between those BALB/c mice that rejected C57BL/6 corneal allografts and those BALB/c mice that accepted such grafts at any of the time points measured (Fig. 6). Interestingly, the CTL activity for BALB/c mice that rejected C57BL/6 corneal allografts were significantly greater than CTL activity from naive mice when measured at 3, 4, 5, and 6 weeks after engraftment. In contrast, the CTL activity for BALB/c mice that accepted C57BL/6 corneas, which were intermediate between rejected and naive mice, was not significantly greater than that seen in naive control mice for all time points, with the exception of 4 weeks after engraftment (Fig. 6). The CTL responses for BALB/c mice engrafted with C57BL/6 tail skin were significantly greater than all other groups tested, with the exception of those responses measured at 4 weeks

8 T-Cell Responses in Murine Corneal Transplantation 1537 FIGURE 6. Cytotoxic T-lymphocyte responses in BALB/c engrafted with C57BL/6 corneal allografts. Mice from the same groups described in Figure 2 were tested for their ability to generate C57BL/6-specific CTL activity. Spleen cells from the indicated groups of mice were tested for their ability to lyse 51 Cr-labeled EL-4 (H-2 b ) using effector:target ratios of 100:1. Results reflect the mean percent specific lysis ± SEM of between 6 and 12 individual mice per determination. There was no significant amount of lysis of 5l Cr-labeled P-815 syngeneic targets (data not shown). after engraftment (Fig. 6). As is shown in Figure 7, similar patterns of CTL responses were observed when other E:T ratios were used (Fig. 7). Figure 7 also presents representative data demonstrating that there was no significant lysis of the syngeneic P-815 by BALB/c mice engrafted with C57BL/6 tail skin (group E). We conclude that because there were no significant differences in CTL activity between mice that accepted C57BL/6 corneal grafts and mice that rejected such grafts, measurements of systemic CTL activity do not correlate well with the ultimate fate of corneal allografts when using this mouse strain combination. DISCUSSION The striking success of human corneal transplantation, which can be as high as 90%, probably is caused by multiple factors. These include the lack of significant corneal blood vessels and lymphatics, the absence of MHC class II + cells in central cornea that could serve as antigen-presenting cells, and the presence of various immunosuppressive factors found in aqueous humor. 9 " 13 All these factors work in concert to form a unique microenvironment that tends to select against the development of strong cellular immune responses. When this microenvironment is disrupted, such as when corneal transplantation is performed in vascularized corneal beds, the success of the corneal graft is significantly reduced as evidenced by published rejection rates of between 60% and 80%.'" 8 To understand better the mechanisms involved in rejection of "high-risk" corneal transplantation, we used a murine model of corneal transplantation in which the rate of graft failure ranges from 40% to 65%. The results of these studies indicate that most BALB/c mice engrafted with C57BL/6 corneas (83%) display measurable corneal opacity after transplantation. In contrast, only 38% of BALB/c mice engrafted with syngeneic BALB/c corneas demonstrated transient corneal clouding. Although it is not possible to determine the exact reason for this cornea opacity, the pattern of focal opacification in syngeneic corneal grafts is restricted to peripheral areas of the cornea, near suture sites, and is probably caused by a nonspecific inflammatory response initiated by the sutures or surgical trauma. The transient opacity seen in BALB/ c mice that ultimately accept C57BL/6 corneal allografts also begins near suture sites. However, unlike syngeneic corneal grafts, this clouding of the cornea tends to persist longer, is more severe, and, in some cases, progresses to involve the central cornea. Because this pattern of corneal opacity is clinically distinct from that seen in syngeneic corneal grafts, it suggests that some form of allorecognition takes place. Thus, the transient opacity observed in BALB/c mice that ultimately accepted C57BL/6 corneal allografts probably represented a mild rejection reaction that did not lead to failure of the corneal graft. Although the mechanism responsible for corneal O u. O UJ a. (0 A B -* - C D --a-- E 1:25 1:50 EFFECTOR:TARGET RATIO 1:100 FIGURE 7. Cytotoxic T-lymphocyte (CTL) responses in BALB/c (H-2 d ) mice 3 weeks after corneal transplantation with C57BL/6 (H-2 b ) corneal allografts. Spleen cells from BALB/c mice were tested for their ability to lyse 5l Cr-labeled EL-4 (H-2 b ) [A-D] or P-815 (H-2 d ) [E] targets using effectontarget ratios of 5:1, 25:1, 50:1, and 100:1. Mice from the groups described in Figure 2 were tested for their ability to generate C57BL/6-specific CTL activity. These include control BALB/c mice (A), BALB/c mice with C57BL/6 and mice whose opacity scores were ^2 (B). BALB/c mice with C57BL/6 corneal grafts whose opacity scores are >2 (C). BALB/c mice engrafted with C57BL/6 tail skin (D, E). Results reflect the mean percent specific lysis ± SEM of 6 to 13 individual mice per determination.

9 1538 Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 allograft rejection is unknown, our results indicate that corneal graft rejection in BALB/c recipients of C57BL/6 corneas is best associated with the ability to generate strong DTH responses and not the presence of significant CTL activity. The CTL responses for BALB/c mice that have rejected C57BL/6 corneal allografts, though significantly greater than that measured for naive BALB/c mice for all time points beyond 2 weeks, was not greater than CTL responses for BALB/c mice that did not display clinical signs of rejecting C57BL/6 corneal allografts. Further supporting the notion that CTL activity is not the primary means of corneal allograft rejection is the fact that the CTL activity of BALB/c mice that do not display signs of corneal allograft rejection are significantly greater than that measured in naive BALB/c mice 4 weeks after engraftment. This suggests that even though cells that mediate CTL activity in these animals have been sensitized toward the alloantigens of the C57BL/6 corneal allograft, these cells are not able to inflict measurable damage on the corneal graft itself. These findings are not in agreement with earlier studies by Niederkorn and coworkers, 18 ' 19 who reported that rejection of heterotopic corneal allografts in mice and orthotopic corneal allografts in rats was associated the development of allospecific CTLs in the apparent absence of detectable DTH responses. Although they used a mixture of spleen and regional lymph node cells, 1819 and we measured CTL activity in lymphocytes isolated from spleens alone, it seems unlikely that this alone could be the reason for the discrepancy in results. Furthermore, the studies we report here clearly indicate that BALB/c mice that rejected C57BL/6 corneal allografts generated significant allospecific DTH responses against the alloantigens expressed by the C57BL/6 corneal allografts, whereas BALB/c mice that accept such grafts do not. It is interesting to note that He et al 23 reported in 1991 that treatment of mice with anti-cd4 antibodies, not with anti-cd8 antibodies, prevented the rejection of allogeneic corneas. These results indicated that CD4 + T cells are required for C57BL/6 rejection of BALB/c corneas and that CD8 + cells are neither sufficient nor required for corneal graft rejection when using that strain combination. Because DTH responses typically are mediated by CD4 + T cells of the Thl phenotype and because CTL responses are predominantly the result of direct killing by CD8 + cells T cells, 27 ' 28 their data could be interpreted as indicating that murine corneal allograft rejection is primarily the result of allospecific DTH responses. Further evidence, supporting the ability of mice engrafted with allogeneic corneas to generate DTH responses, comes from a report by Niederkorn and Mayhew, 29 in which C57BL/6 mice given BALB/c corneal heterotopic allografts generated measurable and significant DTH responses to BALB/c alloantigens. Furthermore, studies from Streilein and coworkers have reported that BALB/c mice that accepted allogeneic C57BL/6 corneas, are not capable of generating significant DTH responses after subcutaneous injection of C57BL/6 lymphoid cells, whereas BALB/c mice that rejected C57BL/6 corneas developed significant DTH responses after such injections. 20 ' 21 They concluded that suppression of alloantigen-specific DTH responses plays a critical role in the success of grafted corneas. 20 ' 21 Finally, a recent report by He et al, 30 describing the effects of anti-lfa-1 and anti-icam-1 treatment on the ability to CB6F1 to reject C3H corneal allografts, demonstrated that anti-lfa-1 not anti- ICAM-1 treatment significandy prolonged graft survival and led to a higher percentage of permanently accepted allografts. Although both treatments prevented the induction of CTL and DTH responses toward donor alloantigens, the authors interpreted these data to mean that anti-lfa-1 acted by interfering with alloantigen presentation to potential effector T cells. 30 This interpretation is supported by a report by Gorczynski and Wojcik, 31 who used these antibodies in a murine model of skin transplantation. Their data indicated that anti-icam-1/lfa-1 treatment prolonged allogeneic skin graft survival by blocking the signals required for Thl activation, but they did allow for the development of Th2 cells. 31 Taken together, these data, though not conclusive, best support the concept that CD4 + Thl cells play a central role in the rejection of primary murine corneal allografts. Furthermore, the mechanism they use for graft destruction is most likely one that involves an antigen-specific DTH response. This does not, however, mean that CTL responses are irrelevant to this process, only that it is unlikely they play a determinative role. As expected, the DTH response engendered by C57BL/6 tail skin graft was directed equally to MHC and non-mhc alloantigens as determined by the ability of spleen cells from C.B10-H-2 b (MHC difference) and B10.D2 (non-mhc difference) to elicit a DTH response. However, when BALB/c mice that were rejecting C57BL/6 corneal allografts were tested for their responses against these same cells, the C.B10-H- 2 b cells did not elicit significant DTH responses in these animals, whereas spleen cells from B10.D2 did elicit a significant DTH response. We expected that we would observe a pattern similar to that seen for BALB/c mice engrafted with C57BL/6 tail skin, in which similar DTH responses would be elicited by these spleen cells. This surprising result suggests that the T cells mediating DTH responses during allograft rejection of whole-strain, mismatched corneas are primarily recognizing non-mhc antigens. These data are particularly significant because C57BL/6 and C57BL/ 10 mice do not share all the same minor histocompati-

10 T-Cell Responses in Murine Corneal Transplantation 1539 bility antigens, as indicated by the lack of skin graft acceptance between these two strains of mice (Stuart PM, unpublished observations, 1984). This indicates that the DTH response measured against B10.D2 mice is directed against some, but not all, minor histocompatibility antigens expressed by C57BL/6 mice. This observation, however, is consistent with the studies reported by Sonada and Streilein 24 concerning the genetic pattern for corneal allograft rejection. They reported that the genetic disparity that gives rise to the highest rate of allograft rejection is mismatch at non-mhc loci. The rate of rejection for non-mhconly mismatch was 41%, essentially the same as MHGand non-mhc mismatch (44%) and considerably greater that MHC class I only (30%) or MHC class II only (11%) mismatch. 24 In addition to these results involving MHC and non-mhc mismatched corneal transplants in mice, a similar pattern has been described for orthotopic corneal allografts involving inbred strains of rats. 32 " 34 The observation that DTH responses generated after corneal allograft transplantation in mice is directed primarily toward non-mhc-encoded alloantigens may provide an explanation for the findings of the Collaborative Corneal Transplantation Studies, sponsored by the National Eye Institute 35 and designed to evaluate the effectiveness of donor-recipient matching in patients at high risk in the United States using a postoperative immunosuppressive regimen considered the standard of care. This multicenter study addressed the question of whether matching HLA-A, HLA-B, and/or HLA-DR antigens between donor and recipient had any impact on the ultimate success or failure of corneal allografts to survive 3 years after surgery. Results indicated that matching for these MHC-encoded alloantigens had no effect on overall graft survival, the incidence of irreversible rejection, or the incidence of rejection episodes. 35 One possible explanation for the Collaborative Corneal Transplantation Studies findings suggested by our data and that of Sonada and Streilein, 24 is that MHC matching is not particularly useful in an outbred population in which multiple non-mhc mismatches are invariably present and may be overriding. Key Words corneal transplantation, delayed-type hypersensitivity, immunoregulation, major histocompatibility complex, minor histocompatibility antigens Acknowledgments The authors thank Dr. Keith Laycock for his observations of mouse corneas, and they thank Kelly Hook and Elena Jimenez for fine technical assistance. References 1. Brady SE, RapuanoJC, ArentsenJJ, Cohen EJ, Laibson PR. Clinical indications and procedures associated with penetrating keratoplasty. Am J Ophthalmoi. 1989;108: Vail A, Gore SM, Bradley BA, Easty DL, Rogers CA. Corneal transplant follow-up study. ARVO Abstracts. Invest Ophthalmoi Vis Sri. 1993; 34: Williams KA, Roder D, Esterman A, Muehlberg SM, Coster DJ. Factors predictive of corneal graft survival: Report from the Australian Corneal Graft Registry. Ophthalmology. 1992;99: Sugar A. A analysis of corneal endothelial and graft survival in pseudophakic bullous keratopathy. Trans Am Ophthalmoi Soc. 1989;87: Boisjoly HM, Taourigny R, Bazin R, et al. Risk factors of corneal graft failure. Ophthalmology. 1993; 100: Council on Scientific Affairs. Report of the Organ Transplant Panel: Corneal Transplantation. JAMA. 1988;259:7l Aldredge OC, Krachmer JH. Clinical corneal transplant rejection: their manifestations, frequency, preoperative correlates and treatment. Arch Ophthalmoi. 1981;99: Hill JC. The use of cyclosporine in high-risk keratoplasty. Am J Ophthalmoi. 1989; 107: Streilein JW. Immune regulation and the eye: A dangerous compromise. FASEBJ. 1987; 1: NiederdornJY. Immune privilege and immune regulation in the eye. Adv Immunol. 1990;48: Granstein RD, Staszewski R, Knisely TL, et al. Aqueous humor contains transforming growth factor-/? and a small (< 3500 daltons) inhibitor of thymocyte proliferation. J Immunol. 1990; 144: Cousins SW, McCabe MM, Danielpour R, Streilein JW. Identification of transforming growth factor-/? as an immunosuppressive factor in aqueous humor. Invest Ophthalmoi Vis Sri. 1991; 32: Streilein JW, Wilbanks GA, Taylor A, Cousins S. Eyederived cytokines and the immunosuppressive intraocular microenvironment: A review. Curr Eye Res. 1992; 11: Kaplan HJ, Streilein JW. Immune response to immunization via the anterior chamber of the eye: I: Fl lymphocyte-induced immune deviation. J Immunol. 1977; 118: Ferguson TA, Waldrep JC, Kaplan HJ. The immune response and the eye: II: The nature of T suppressorcell induction in anterior chamber-associated immune deviation (ACAID). / Immunol. 1987; 139: Wilbanks GA, Streilein JW. Characterization of suppressor cells in anterior chamber-associated immune deviation (ACAID) induced by soluble antigen: Evidence of two functionally and phenotypically distinct T-suppressor cell populations. Immunology. 1990; 71: Streilein JW, Niederkorn JY. Characterization of the suppressor cell(s) responsible for anterior chamber-

11 1540 Investigative Ophthalmology & Visual Science, July 1995, Vol. 36, No. 8 associated immune deviation (ACAID) induced in BALB/c mice by P815. J Immunol. 1985; 134: Peeler J, Niederkorn J, Matoba A. Corneal allografts induce cytotoxic T cell but not delayed hypersensitivity responses in mice. Invest Ophthalmol Vis Sci. 1985; 26: Callanan D, Peeler J, Niederkorn JY. Characteristics of rejection of orthotopic corneal allografts in the rat. Transplantation. 1988; 45: Sonoda Y, Ksander B, Streilein JW. Impaired cell mediated immunity (ACAID) induced in mice that accept orthotopic corneal allografts indefinitely. ARVO Abstracts. Invest Ophthalmol Vis Sci. 1992; 33: Sonoda Y, Streilein JW. Impaired cell-mediated immunity in mice bearing healthy orthotopic corneal allografts. J Immunol. 1993;150:l727-l She S-C, Steahly LP, Moticka EJ. A method for performing full-thickness, orthotopic, penetrating keratoplasty in the mouse. Ophthalmol Surg. 1990; 21: He Y, Ross J, Niederkorn JY. Promotion of murine orthotopic corneal allograft survival by systemic administration of anti-cd4 monoclonal antibody. Invest Ophthalmol Vis Sci. 1991; 32: Sonoda Y, Streilein JW. Orthotopic corneal transplantation in mice: Evidence that the immunogenetic rules of rejection do not apply. Transplantation. 1992; 54: Stuart PM, Beck-Maier B, Melvold RW. Provocation of skin graft rejection across murine class II differences by non-bone-marrow-derived cells. Transplantation. 1984; 37: Stuart PM, Iannaccone PM, Miller SD, Jenkins MK, Del Muro FA, Melvold RW. In vitro and in vivo correlates of F hybrid resistance. J Natl Can Inst. 1987;78: Cher DI, Mosmann TR. Two types of murine helper T cell clones: II: Delayed-type hypersensitivity is mediated by Thl clones. J Immunol. 1987; 138: Rosenberg AS, Singer A. Cellular basis of skin allograft rejection: an in vivo model of immune-mediated tissue destruction. Ann Rev Immunol. 1992; 10: Niederkorn JY, Mayhew E. UVB irradiation renders corneal allografts tolerogenic for allospecific delayed hypersensitivity responses. Immunology. 1993; 79: He YG, Mellon J, Apte R, Niederkorn JY. Effect of LFA- 1 and ICAM-1 antibody treatment on murine corneal allograft survival. Invest Ophthalmol Vis Sci. 1994; 35: Gorczynski RM, Wojcik D. A role for nonspecific (cyclosporin A) or specific (monoclonal antibodies to ICAM-1, LFA-1, and IL-10) immunomodulation in the prolongation of skin allografts after antigen-specific pretransplant immunization or transfusion. J Immunol. 1994; 152: Katami M, Madden PW, White DJG, Watson PG, Kamada N. The extent of immunological privilege of orthotopic corneal grafts in the inbred eye. Transplantation. 1989; 48: RossJ, Callanan D, Kunz H, Niederkorn JY. Evidence that the fate of class II-disparate corneal grafts is determined by the timing of class II expression. Transplantation. 1991; 51: Niederkorn JY, RossJR, He YG. Effect of donor Langerhans cells on corneal graft rejection. / Invest Derm. 1992;99:104S-106S. 35. The collaborative corneal transplantation studies research group. The collaborative corneal transplantation studies (CCTS): Effectiveness of histocompatibility matching in high-risk corneal transplantation. Arch Ophthalmol. 1992; 110: