- Email: [email protected]
Detection of HLA Class I and II Antigens in Rejected Human Corneal Allografts
Detection of HLA Class I and II Antigens in Rejected Human Corneal Allografts JAY S. PEPOSE, MD, PhD, KATHRYN M. GARDNER, MD, MARK S. NESTOR, BA, ROBERT Y. FOOS, MD, THOMAS H. PETTIT, MD
Abstract: We compared the distribution of HLA-ABC (class I) and HLA-OR (class II) antigens on fresh human donor corneal tissue, donor corneas following a 72-hour storage in McCarey-Kaufman (M-K) medium, and corneal buttons from patients with allograft rejection and with chronic herpetic stromal keratitis. Incubation in M-K media had little or no effect on the distribution of HLA antigens as compared with fresh tissue. In contrast to control corneas, both HLA class I and II antigens were detected on corneal endothelial cells, cells in the stroma, and on basal epithelial cells in rejected allografts. Corneal endothelium in herpetic buttons did not express detectable HLA antigens. HLA-OR positive Langerhan's cells were demonstrated in the central corneal epithelium of rejected allografts, as well as in herpetic corneas, but not in control corneas except at the limbus. Based upon these observations, a theory of corneal allograft rejection in humans is proposed based upon the induction of class I HLA-ABC and class II HLA-OR antigens on cells in the donor button by a factor(s) associated with cellular inflammation. [Key words: corneal transplantation, HLA antigens, monoclonal antibodies, transplantation immunology.] Ophthalmology 92: 1480-1484, 1985
Corneal allograft rejection on an immunologic basis is the leading cause of graft failure, 1 especially in patients with vascularized corneas or a history of multiple allograft rejections. However, the specific mechanisms that underlie and initiate the rejection reaction remain unclear. The major transplantation (histocompatability) antigens in man are the human leukocyte antigens (HLA), which are clustered together in a small segment of chromosome six. 2 The HLA-A,B, and C antigens are termed From the Jules Stein Eye Institute and the Departments of Ophthalmology, Medicine and Pathology, UCLA School of Medicine, Los Angeles. Presented at the Eighty·ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15, 1984. Supported in part by grants from the National Society to Prevent Blindness .(Dr. Pepose) and from the Jules Stein Eye Institute (Dr. Pepose) and Research Grants EY 00331 and EY 00725 from the National Eye Institute and by a Research Manpower Award from Research to Prevent Blindness, Inc., New York, New York (Dr. Foos). Dr. Pepose was the recipient of a Damon Runyon-Walter Winchell Research Fellowship DRG-025. Reprint requests to Jay S. Pepose, MD, The Wilmer Institute B-23, The Johns Hopkins Hospital, 601 North Wolfe Street, Baltimore, MD 21205.
1480
class I and are serologically determined. The antigens coded by the genes at the HLA-D locus are referred to as the class II antigens and were originally defined by the mixed lymphocyte culture reaction, but can now also be typed with specific isoantisera. Recent studies have localized HLA antigens on specific corneal cells3- 7 and a serologic response to donor HLA antigens has been detected following human corneal transplantation. 8 The importance of the major transplantation antigens in a broad spectrum of experimental and human organ transplants 9 has prompted clinical trials of HLA matching in high risk corneal allograft patients. 10-17 These studies suggest that the HLA system may, in some circumstances, play a role in the rejection of corneal transplants and that a negative preoperative lymphocytoto xi city crossmatch may decrease the incidence of graft rejection in high risk individuals. l3,l7 Whereas the distribution of HLA antigens has been demonstrated on normal human corneal tissue and corneal cells examined shortly after enucleation or following tissue culture,3-7 we are unaware of any studies of HLA antigens on corneal tissue that has been processed following the standard protocol that precedes the typical corneal allograft or following graft rejection. For example, storage
PEPOSE,
et al • HLA ANTIGENS IN REJECTED CORNEAL ALLOGRAFTS
of donor corneas for up to 72 hours in McCarey-Kaufman (M-K) media 18 has become a routine procedure in preparation for penetrating keratoplasty. 19 To further our understanding of the functional role of HLA antigens in human corneal allograft rejection, we compared the distribution of HLA antigens before and after incubation in M-K medium and following corneal allograft rejection. For additional comparison, HLA antigens were also localized in corneal buttons harvested from patients with chronic herpetic stromal keratitis. We discuss the implications of our findings with respect to possible mechanisms of corneal allograft rejection and avenues for further clinical and experimental research.
MATERIALS AND METHODS PATIENT SELECfION
Corneal buttons of seven patients who had undergone corneal allograft rejection and were being regrafted at the Jules Stein Eye Institue or affiliated UCLA hospitals were examined. Criteria for the diagnosis of immunologic allograft rejection included the onset of graft edema and keratic precipitates with or without a rejection line beginning several weeks or later after a technically successful, clear graft. The specific clinical characteristics of the seven patients with allograft rejection have been described elsewhere, along with an analysis of the mononuclear cell infiltrates in the rejected corneal buttons. 2o No attempt was made to histocompatability match the donor and recipient prior to the original grafting. In. addition, nine corneal buttons from patients with chronic herpetic stromal keratitis (with the last recurrence of active disease at least one year prior to penetrating keratoplasty) were identically processed for HLA class I and II antigens along with the rejected allografts. Finally, three donor corneas from the Jules Stein Eye Bank were hemisected; one half of each was snap frozen in liquid nitrogen and processed immediately for HLA antigens; the other portion was processed identically following a 72-hour incubation in McCarey-Kaufman (M-K) medium at 4 0 C. IMMUNOCYTOLOGIC STUDIES
Eacn rejected or herpetic corneal button was mounted in Tissue-Tek, immersed in liquid nitrogen and transversely hemi-sectioned at 6 J.Lm intervals using a Slee cryostat. The remaining half of each corneal button was fixed in 10% neutral buffered formalin for routine staining. Frozen sections were fixed in chilled acetone for 5 seconds and then stored dessicated at -70 0 C. Every eleventh slide was fixed in formalin, stained with hematoxylin and eosin and studied by light microscopy to assess the histopathologic features. Selected slides were refixed in chilled acetone for 15 minutes and air dried. Following rehydration in phosphate buffered saline (PBS), the slides were incubated with a 1/ 200 dilution of purified murine monoclonal anti-HLA-
DR antibody (Beckton-Dickinson) or 1/100 dilution of murine monoclonal anti HLA-A,B,C antigen (shared determinant; clone W6/32; Sera-lab, Inc.) for 40 minutes. Following a wash in PBS, slides were incubated with a 1/30 dilution of affinity purified peroxidase conjugated goat anti-mouse IgG (Boehringer Mannheim) for 25 minutes at room temperature in a moist chamber. After another wash step, slides were incubated in 3-amino-9-ethyl carbazole developer (Sigma) and hydrogen peroxide for 10 minutes. After counter-staining with hematoxylin, the slides were examined with a light microscope. As positive reaction product was indicated by a reddish-brown discoloration on cellular membranes. As reported elsewhere,2° the mononuclear cellular infiltrates (i.e. T-cells, macrophages, Langerhan's cells, B-cells, natural killer cells) in these corneal samples were analyzed using an identical immunoperoxidase method and the leu series (Beckton Dickinson) of monoclonal antibodies.
RESULTS DISTRIBUTION OF HLA ANTIGENS IN NORMAL CORNEA
In freshly obtained control corneas, HLA-ABC (class I) antigens were localized on corneal epithelium (Fig 1) and, at lower levels, on stromal keratocytes (Fig 2). The corneal endothelium was devoid of detectable HLA-ABC antigens (Fig 3). HLA-DR (class II) antigens were not detected on central corneal epithelium, stroma or endothelium. In contrast, HLA-DR antigens were readily demonstrated on cells at the corneal limbus (Fig 4), including vascular endothelium, leu-6 positive Langerhan's cells and leu-M3 positive macrophages. The distribution of HLA antigens on normal corneal tissue is schematically diagramed in Figure 5. EFFECfS OF CORNEAL INCUBATION IN MCCAREY-KAUFMAN (M-K) MEDIA
Normal donor corneas were hemi-sectioned, with half of each cornea stored at 4 0 C for 72 hours in M-K medium prior to staining and the other half processed immediately for HLA antigens without storage. There was little or no difference in the expression of HLA antigens following storage as compared to fresh cornea. HLA ANTIGENS IN REJECfED CORNEAL ALLOGRAFTS
In rejected corneal buttons, HLA-DR and HLA-ABC antigens (Fig 6) were detected on cells in the stroma. Cells in the basal epithelium (Fig 7) were also HLA-DR positive. Both HLA-ABC (class I) antigens (Fig 8) and HLA-DR (class II) antigens (Fig 9) were localized on corneal endothelial cells in rejected corneal buttons where endothelial cells were still present at the time of keratoplasty. Leu6 positive, HLA-ABC- and HLA-DR-bearing Langerhan's cells were seen in central and pericentral epithelium in the rejected allografts. The distribution of HLA antigens 1481
OPHTHALMOLOGY •
NOVEMBER 1985 •
VOLUME 92 •
NUMBER 11
Fig. 1. Top left. HLA-ABC antigens (red reaction product) are demonstrated on control corneal epithelium (arrow) using the indirect immunoperoxidase method (hematoxylin, X285). Fig. 2. Top center. HLA-ABC positive keratocytes (arrow) are seen (red reaction product) in a normal cornea (indirect immunoperoxidase; hematoxylin, X285). Fig. 3. Top right, the control corneal endothelium is devoid of detectable HLA-ABC antigens. Note the absence of a red reaction product (indirect immunoperoxidase method; hematoxylin, X285). Fig. 4. Second row left, HLA-DR antigens (red reaction product) are demonstrated on cells at the corneal limbus, including vascular endothelium (asterisk), Langerhan's cells (arrow) and macrophages (arrowheads) (indirect immunoperoxidase technique; hematoxylin, X285). Fig. 5. Second row center, the distribution of HLA-ABC (yellow) and HLA-DR (blue) antigens in the normal cornea and limbus, chronic stromal herpes keratitis and in the rejected corneal allograft is diagrammed schematically. See text for detailed description. Fig. 6. Second row right, HLA-ABC antigens (red reaction product) are localized on cells in the stroma of a rejected corneal allograft (indirect immunoperoxidase; hematoxylin, X285). Fig. 7. Bottom left. a rejected corneal graft with HLA-DR positive cells (red reaction product) seen in the basal epithelium (arrow) and anterior stroma (indirect immunoperoxidase; hematoxylin, X285). Fig. 8. Bottom center. HLA-ABC antigens are readily detected on the corneal endothelium (arrows) in a rejected corneal allograft button (indirect immunoperoxidase; hematoxylin, X285). Fig. 9. Bottom right, HLA-DR antigens (red reaction product) are demonstrated on an attenuated endothelial cell (arrow) in a rejected corneal allograft (indirect immunoperoxidase; hematoxylin, X285).
in rejected corneal allografts is shown schematically in Figure 5. HLA ANTIGENS IN CHRONIC HERPETIC STROMAL KERATITIS
In buttons obtained from patients with chronic herpetic stromal keratitis, HLA-DR positive cells were observed in the basal epithelium and in cells in the central stroma, as in rejected buttons. The corneal endothelium remained negative for both HLA-ABC and HLA-DR antigens. Leu6 positive, HLA class I and II bearing Langerhan's cells were seen in the central and pericentral corneal epithelium. The distribution ofHLA antigens in normal cornea, and in rejected and herpetic buttons is shown schematically in Figure 5. 1482
DISCUSSION Our results indicate that HLA class I and II antigens are present on the corneal endothelium and on stromal keratocytes in rejected allografts. Class II antigens were also detected on cells in the basal epithelium and on Langerhan's cells in the epithelium. Whereas herpetic corneas also contained HLA-DR positive Langerhan's cells in the epithelium and class II positive basal epithelial cells, we could not demonstrate either HLA class I or II antigens on the endothelium. Cells in the stroma of herpetic buttons bore HLA class I and class II antigens, but the large number of infiltrating macrophages precluded differentiating at the light microscopic level whether specific cells
PEPOSE, et al •
HLA ANTIGENS IN REJECTED CORNEAL ALLOGRAFTS
represented keratocytes or macrophages. Thus, further studies are indicated to determine whether keratocytes express class II antigens in these buttons. It is of interest that the cellular immune response in the herpetic corneas was localized in the anterior and mid-stroma, but did not involve the posterior stroma as in the rejected allografts. The possible significance of this finding in relation to the failure to detect HLA antigens on the corneal endothelium in the herpetic buttons is discussed below. Normal human corneas were shown to express HLAABC (class I) antigens on epithelial cells and at lower levels on stromal keratocytes. Class I antigens were also readily detected on cells at the limbus (ie. macrophages, Langerhan's cells, and vascular endothelium), but were notably not detected on corneal endothelium. HLA-DR (class II) antigens were restricted to cells at the limbus (ie. macrophages, Langerhan's cells and vascular endothelium) and were not detected on corneal epithelium, stroma, or endothelium. These data are in agreement with the recent study by Whitsett and Stulting. 7 Similar results were obtained by Fujikawa et al,6 although HLA class I antigens were not detected on stromal keratocytes in their study. Following the original observation by Newsome et al,4 several investigators6,7 have confirmed the localization of HLA class I antigens on corneal epithelium, stroma and endothelium in tissue culture. In addition, fresh corneal endothelium, which is devoid of detectable HLA-ABC antigens, can be induced to express class I antigens by tissue culture conditions.6 Therefore, we thought it important to determine the effect of storage of the donor cornea in M-K medium, since any derepression of HLA expression may be undesirable for allografting. We found no change following 72-hour storage ofthe donor cornea in M-K medium on the expression of HLA class I or II antigens. The detection of Langerhan's cells in central and pericentral cornea in both rejected allografts and herpetic corneas are consistent with the results of the study by Gillette et a1. 21 However, in contrast to their findings, we did not detect Langerhan's cells in transverse sections of control cornea (by either HLA-DR or leu-6 staining) except at the limbus. Whether Langerhan's cells are present within the normal corneal donor button is a controversial 21 ,22 and important point, since these HLA-DR positive cells could theoretically contribute to the initiation of allograft rejection. Because we utilized a monoclonal antibody directed against the non polymorphic determinants of HLADR, it was not possible to determine whether these cells in rejected buttons expressed class II antigens of host or donor origin. However, our failure to detect Langerhan's cells in the non-limbal regions of control corneas and the limited life span of dendritic cells in other organs,23 suggest that the Langerhan's cells seen in the rejected corneal allografts are most likely of host origin. Our results are consistent with the notion that HLAclass I and II antigens are induced on corneal endothelial cells and class II antigens are induced on stromal keratocytes and basal epithelial cells in rejecting allografts. Studies of cadaveric renal transplantation in man have stressed the importance of HLA-DR (class II) compata-
bility between donor and recipient, and note a beneficial effect of HLA-A,B matching only in HLA-DR mismatched combination. 24 The aberrant localized expression of class II antigens is also becoming recognized in several autoimmune disorders such as autoimmune thyroiditis,2S in addition to skin allograft rejection 26 and graft vs. host disease. 27 A lymphokine, gamma interferon,28 has been shown to induce class II antigens and enhance the expression of class I antigens on human umbilical vein endothelial cells and dermal fibroblasts. Recent studies by Young et al 29 have demonstrated that HLA-DR antigens can be induced on cultured human corneal endothelial and stromal cells by gamma interferon. Therefore a lymphokine may similarly induce the expression of class II antigens on rejected corneal endothelium. The failure to detect class II antigens on the herpetic endothelium may reflect the restricted localization of the cellular inflammatory response limited to the anterior stroma with limited diffusion oflymphokines. One possible pathogenic scheme for corneal allograft rejection is as follows. In the majority of low risk corneal transplants, HLA class I incompatability (on corneal epithelium and stromal keratocytes) may not represent a strong enough transplantation barrier to initiate graft rejection. However, in transplants with localized nonspecific inflammation (e.g. loose sutures,30 vascularized recipient beds, subclinical infection) the production oflymphokines such as gamma interferon may result in the local induction of class I and II antigens on the corneal endothelium, keratocytes and basal epithelium. This may elicit further production of lymphokines, more inflammation and the subsequent induction of HLA-DR across the cornea in a vicious cycle. It is of interest that in addition to its proposed beneficial lymphocytotoxic effect in corneal allograft rejection,31 it has been suggested that glucocorticosteroids may partially inhibit the expression of HLA antigens in renal allograft rejection,32 and theoretically could playa similar role in corneal graft rejection. The results of this study indicate that class II antigens can be detected on human corneal transplants that are rejected. Our data provide further impetus to ongoing double blind clinical trials of lymphocytoxicity antibody screening and cross-match testing in high risk patients33 to determine the overall influence, as well as the significance of matching at any particular histocompatability locus.
REFERENCES 1. Khodadoust AA. The allograft rejection reaction: the leading cause of late failure of clinical corneal grafts. In: Corneal Graft Failure; Ciba Foundation Symposium 15 (new series). Amsterdam: Associated Scientific Publishers, 1973; 151-67. 2. McDevitt HO. Regulation of the immune response by the major histocompatability system. N Engl J Med 1980; 303:1514-7. 3. Ehlers N, Ahrons S. Corneal transplantation and histocompatability. Acta Ophthalmol1971; 49:513-27. 4. Newsome DA, Takasugi M, Kenyon KR, et al. Human corneal cells in vitro: morphology and histocompatability (HL-A) antigens of pure cell populations. Invest Ophthalmol1974; 13:23-32.
1483
OPHTHALMOLOGY
•
NOVEMBER 1985 •
5. Klareskog L, Forsum U, Tjemlund UM, et al. Expression of la antigenlike molecules on cells in the comeal epithelium. Invest Ophthalmol Vis Sci 1979; 18:310-3. 6. Fujikawa LS, Colvin RB, Bhan AK, et al. Expression of HLA-AfBfC and -DR locus antigens on epithelial, stromal, and endothelial cells of the human com ea. Cornea 1982; 1:213-22. 7. Whitsett CF, Stulting RD. The distribution of HLA antigens on human comeal tissue. Invest Ophthalmol Vis Sci 1984; 25:519-24. 8. Stark WJ, Opelz G, Newsome 0, et al. Sensitization to human lymphocyte antigens by corneal transplantation. Invest Ophthalmol1973; 12:639-45. 9. Morris PJ, Batchelor JR, Festenstein H. Matching for HLA in transplantation. Sr Med Bull 1978; 34:259-62. 10. Batchelor JR, Casey TA, Gibbs DC, et al. HLA matching and corneal grafting. Lancet 1976; 1:551-4. 11. Allansmith MR, Fine M, Payne R. Histocompatability typing and corneal transplantation. Trans Am Acad Ophthalmol Otolarnygol 1974; 78: OP445-60. 12. Kissmeyer-Nielson F, Ehlers N. Comeal transplantation and matching for HLA-A and HLA-B. In: Ferrone S, Curtoni ES, Gorini S, eds. HLA Antigens in Clinical Medicine and Biology. New York:Garland, 1979; 200-3. 13. Stark WJ, Taylor HR, Bias WS, Maumenee AE. Histocompatability (HLA) antigens and keratoplasty. Am J Ophthalmol 1978; 86:595604. 14. Vannas S, Karjalainen K. Ruusuvaara P, Tiilikainen A. HLA- compatible donor comea for prevention of allograft reaction. Albrecht von Graefes Arch Klin Exp Ophthalmol1976; 198:217-22. 15. Fronterre A, Trimarchi F, Bo G. HLA antigens and selection of donors in corneal transplants. Curr Ther Res 1980; 27:749-56. 16. Volker-Dieben HJM, Kok-Van Alphen CC, Kruit PJ. Advances and disappointments, indications and restrictions regarding HLA-matched corneal grafts in high risk cases. Doc Ophthalmol1979; 46:219-26. 17. Foulks GN, Sanfilippo FP, Locascio JA III, et al. Histocompatability testing for keratoplasty in high-risk patients. Ophthalmology 1983; 90: 239-43. 18. McCarey BE, Kaufman HE. Improved corneal storage. Invest Ophthalmol 1974; 13:165-73. 19. Stark WJ, Maumenee AE, Kenyon KR.lntermediate-term corneal stor-
1484
20.
21. 22. 23. 24.
25.
26.
27. 28.
29.
30. 31.
32.
33.
VOLUME 92
•
NUMBER 11
age for penetrating keratoplasty. Am J Ophthalmol 1975; 79:795802. Pepose JS, Nestor MS, Gardner KM, et al. Composition of cellular infiltrates in rejected human corneal allografts. Graefes Arch Clin Exp Ophthalmol 1985; 222: 128-33. Gillette TE, Chandler JW, Greiner JV. Langerhans cells of the ocular surface. Ophthalmology 1982; 89:700-10. Streilein JW, Toews GB, Bergstresser PRo Corneal allografts fail to express la antigens. Nature 1979; 282:326-7. Steinman RM, Nussenzweig MC. Dendritic cells: features and functions. Immunol Rev 1980; 53:127-47. Moen T, Albrechtsen 0, Flatmark A, et al. Importance of HLA-DR matching in cadaveric renal transplantation: a prospective one-center study of 170 transplants. N Engl J Med 1980; 303:850-4. Bottazzo GF, Pujol-Borrell R, Hanafusa T, Feldmann M. Role of aberrant HLA-DR expression and antigen presentation in induction of endocrine autoimmunity. Lancet 1983; 2: 1115-9. de Waal RMW, Bogman MJJ, Maass CN, et al. Variable expression of la antigens on the vascular endothelium of mouse skin allografts. Nature 1983; 303:426-9. Barclay AN, Mason Ow. Graft rejection and la antigens--paradox resolved? Nature 1983; 303:382-3. Pober JS, Collins T, Gimbrone MA Jr, et al. Lymphocytes recognize human vascular endothelial and dermal fibroblast la antigens induced by recombinant immune interferon. Nature 1983; 305:726-9. Young E, Stark WJ, Prendergast RA. Immunology of corneal allograft rejection: HLA-DR antigens on human comeal cells. Invest Ophthalmol Vis Sci 1985; 26:571-4. Paque J, Poirier RH. Corneal allograft reaction and its relationship to suture site neovascularization. Ophthalmic Surg 1977; 8:71-4. Polack FM. Lymphocyte destruction during corneal homograft reaction; a scanning electron microscopic study. Arch Ophthalmol 1973; 89: 413-6. Hayry P, von Willebrand E, Ahonen J, Eklund B. Glucocorticosteroids in renal transplantation. I. Impact of high- versus low-dose post-operative methylprednisolone administration on the first episode(s) of rejection. Scand J Immunol1982; 16:39-49. Stark WJ. Transplantation immunology of penetrating keratoplasty. Trans Am Ophthalmol Soc 1980; 78: 1079-117.
Abstract: We compared the distribution of HLA-ABC (class I) and HLA-OR (class II) antigens on fresh human donor corneal tissue, donor corneas following a 72-hour storage in McCarey-Kaufman (M-K) medium, and corneal buttons from patients with allograft rejection and with chronic herpetic stromal keratitis. Incubation in M-K media had little or no effect on the distribution of HLA antigens as compared with fresh tissue. In contrast to control corneas, both HLA class I and II antigens were detected on corneal endothelial cells, cells in the stroma, and on basal epithelial cells in rejected allografts. Corneal endothelium in herpetic buttons did not express detectable HLA antigens. HLA-OR positive Langerhan's cells were demonstrated in the central corneal epithelium of rejected allografts, as well as in herpetic corneas, but not in control corneas except at the limbus. Based upon these observations, a theory of corneal allograft rejection in humans is proposed based upon the induction of class I HLA-ABC and class II HLA-OR antigens on cells in the donor button by a factor(s) associated with cellular inflammation. [Key words: corneal transplantation, HLA antigens, monoclonal antibodies, transplantation immunology.] Ophthalmology 92: 1480-1484, 1985
Corneal allograft rejection on an immunologic basis is the leading cause of graft failure, 1 especially in patients with vascularized corneas or a history of multiple allograft rejections. However, the specific mechanisms that underlie and initiate the rejection reaction remain unclear. The major transplantation (histocompatability) antigens in man are the human leukocyte antigens (HLA), which are clustered together in a small segment of chromosome six. 2 The HLA-A,B, and C antigens are termed From the Jules Stein Eye Institute and the Departments of Ophthalmology, Medicine and Pathology, UCLA School of Medicine, Los Angeles. Presented at the Eighty·ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15, 1984. Supported in part by grants from the National Society to Prevent Blindness .(Dr. Pepose) and from the Jules Stein Eye Institute (Dr. Pepose) and Research Grants EY 00331 and EY 00725 from the National Eye Institute and by a Research Manpower Award from Research to Prevent Blindness, Inc., New York, New York (Dr. Foos). Dr. Pepose was the recipient of a Damon Runyon-Walter Winchell Research Fellowship DRG-025. Reprint requests to Jay S. Pepose, MD, The Wilmer Institute B-23, The Johns Hopkins Hospital, 601 North Wolfe Street, Baltimore, MD 21205.
1480
class I and are serologically determined. The antigens coded by the genes at the HLA-D locus are referred to as the class II antigens and were originally defined by the mixed lymphocyte culture reaction, but can now also be typed with specific isoantisera. Recent studies have localized HLA antigens on specific corneal cells3- 7 and a serologic response to donor HLA antigens has been detected following human corneal transplantation. 8 The importance of the major transplantation antigens in a broad spectrum of experimental and human organ transplants 9 has prompted clinical trials of HLA matching in high risk corneal allograft patients. 10-17 These studies suggest that the HLA system may, in some circumstances, play a role in the rejection of corneal transplants and that a negative preoperative lymphocytoto xi city crossmatch may decrease the incidence of graft rejection in high risk individuals. l3,l7 Whereas the distribution of HLA antigens has been demonstrated on normal human corneal tissue and corneal cells examined shortly after enucleation or following tissue culture,3-7 we are unaware of any studies of HLA antigens on corneal tissue that has been processed following the standard protocol that precedes the typical corneal allograft or following graft rejection. For example, storage
PEPOSE,
et al • HLA ANTIGENS IN REJECTED CORNEAL ALLOGRAFTS
of donor corneas for up to 72 hours in McCarey-Kaufman (M-K) media 18 has become a routine procedure in preparation for penetrating keratoplasty. 19 To further our understanding of the functional role of HLA antigens in human corneal allograft rejection, we compared the distribution of HLA antigens before and after incubation in M-K medium and following corneal allograft rejection. For additional comparison, HLA antigens were also localized in corneal buttons harvested from patients with chronic herpetic stromal keratitis. We discuss the implications of our findings with respect to possible mechanisms of corneal allograft rejection and avenues for further clinical and experimental research.
MATERIALS AND METHODS PATIENT SELECfION
Corneal buttons of seven patients who had undergone corneal allograft rejection and were being regrafted at the Jules Stein Eye Institue or affiliated UCLA hospitals were examined. Criteria for the diagnosis of immunologic allograft rejection included the onset of graft edema and keratic precipitates with or without a rejection line beginning several weeks or later after a technically successful, clear graft. The specific clinical characteristics of the seven patients with allograft rejection have been described elsewhere, along with an analysis of the mononuclear cell infiltrates in the rejected corneal buttons. 2o No attempt was made to histocompatability match the donor and recipient prior to the original grafting. In. addition, nine corneal buttons from patients with chronic herpetic stromal keratitis (with the last recurrence of active disease at least one year prior to penetrating keratoplasty) were identically processed for HLA class I and II antigens along with the rejected allografts. Finally, three donor corneas from the Jules Stein Eye Bank were hemisected; one half of each was snap frozen in liquid nitrogen and processed immediately for HLA antigens; the other portion was processed identically following a 72-hour incubation in McCarey-Kaufman (M-K) medium at 4 0 C. IMMUNOCYTOLOGIC STUDIES
Eacn rejected or herpetic corneal button was mounted in Tissue-Tek, immersed in liquid nitrogen and transversely hemi-sectioned at 6 J.Lm intervals using a Slee cryostat. The remaining half of each corneal button was fixed in 10% neutral buffered formalin for routine staining. Frozen sections were fixed in chilled acetone for 5 seconds and then stored dessicated at -70 0 C. Every eleventh slide was fixed in formalin, stained with hematoxylin and eosin and studied by light microscopy to assess the histopathologic features. Selected slides were refixed in chilled acetone for 15 minutes and air dried. Following rehydration in phosphate buffered saline (PBS), the slides were incubated with a 1/ 200 dilution of purified murine monoclonal anti-HLA-
DR antibody (Beckton-Dickinson) or 1/100 dilution of murine monoclonal anti HLA-A,B,C antigen (shared determinant; clone W6/32; Sera-lab, Inc.) for 40 minutes. Following a wash in PBS, slides were incubated with a 1/30 dilution of affinity purified peroxidase conjugated goat anti-mouse IgG (Boehringer Mannheim) for 25 minutes at room temperature in a moist chamber. After another wash step, slides were incubated in 3-amino-9-ethyl carbazole developer (Sigma) and hydrogen peroxide for 10 minutes. After counter-staining with hematoxylin, the slides were examined with a light microscope. As positive reaction product was indicated by a reddish-brown discoloration on cellular membranes. As reported elsewhere,2° the mononuclear cellular infiltrates (i.e. T-cells, macrophages, Langerhan's cells, B-cells, natural killer cells) in these corneal samples were analyzed using an identical immunoperoxidase method and the leu series (Beckton Dickinson) of monoclonal antibodies.
RESULTS DISTRIBUTION OF HLA ANTIGENS IN NORMAL CORNEA
In freshly obtained control corneas, HLA-ABC (class I) antigens were localized on corneal epithelium (Fig 1) and, at lower levels, on stromal keratocytes (Fig 2). The corneal endothelium was devoid of detectable HLA-ABC antigens (Fig 3). HLA-DR (class II) antigens were not detected on central corneal epithelium, stroma or endothelium. In contrast, HLA-DR antigens were readily demonstrated on cells at the corneal limbus (Fig 4), including vascular endothelium, leu-6 positive Langerhan's cells and leu-M3 positive macrophages. The distribution of HLA antigens on normal corneal tissue is schematically diagramed in Figure 5. EFFECfS OF CORNEAL INCUBATION IN MCCAREY-KAUFMAN (M-K) MEDIA
Normal donor corneas were hemi-sectioned, with half of each cornea stored at 4 0 C for 72 hours in M-K medium prior to staining and the other half processed immediately for HLA antigens without storage. There was little or no difference in the expression of HLA antigens following storage as compared to fresh cornea. HLA ANTIGENS IN REJECfED CORNEAL ALLOGRAFTS
In rejected corneal buttons, HLA-DR and HLA-ABC antigens (Fig 6) were detected on cells in the stroma. Cells in the basal epithelium (Fig 7) were also HLA-DR positive. Both HLA-ABC (class I) antigens (Fig 8) and HLA-DR (class II) antigens (Fig 9) were localized on corneal endothelial cells in rejected corneal buttons where endothelial cells were still present at the time of keratoplasty. Leu6 positive, HLA-ABC- and HLA-DR-bearing Langerhan's cells were seen in central and pericentral epithelium in the rejected allografts. The distribution of HLA antigens 1481
OPHTHALMOLOGY •
NOVEMBER 1985 •
VOLUME 92 •
NUMBER 11
Fig. 1. Top left. HLA-ABC antigens (red reaction product) are demonstrated on control corneal epithelium (arrow) using the indirect immunoperoxidase method (hematoxylin, X285). Fig. 2. Top center. HLA-ABC positive keratocytes (arrow) are seen (red reaction product) in a normal cornea (indirect immunoperoxidase; hematoxylin, X285). Fig. 3. Top right, the control corneal endothelium is devoid of detectable HLA-ABC antigens. Note the absence of a red reaction product (indirect immunoperoxidase method; hematoxylin, X285). Fig. 4. Second row left, HLA-DR antigens (red reaction product) are demonstrated on cells at the corneal limbus, including vascular endothelium (asterisk), Langerhan's cells (arrow) and macrophages (arrowheads) (indirect immunoperoxidase technique; hematoxylin, X285). Fig. 5. Second row center, the distribution of HLA-ABC (yellow) and HLA-DR (blue) antigens in the normal cornea and limbus, chronic stromal herpes keratitis and in the rejected corneal allograft is diagrammed schematically. See text for detailed description. Fig. 6. Second row right, HLA-ABC antigens (red reaction product) are localized on cells in the stroma of a rejected corneal allograft (indirect immunoperoxidase; hematoxylin, X285). Fig. 7. Bottom left. a rejected corneal graft with HLA-DR positive cells (red reaction product) seen in the basal epithelium (arrow) and anterior stroma (indirect immunoperoxidase; hematoxylin, X285). Fig. 8. Bottom center. HLA-ABC antigens are readily detected on the corneal endothelium (arrows) in a rejected corneal allograft button (indirect immunoperoxidase; hematoxylin, X285). Fig. 9. Bottom right, HLA-DR antigens (red reaction product) are demonstrated on an attenuated endothelial cell (arrow) in a rejected corneal allograft (indirect immunoperoxidase; hematoxylin, X285).
in rejected corneal allografts is shown schematically in Figure 5. HLA ANTIGENS IN CHRONIC HERPETIC STROMAL KERATITIS
In buttons obtained from patients with chronic herpetic stromal keratitis, HLA-DR positive cells were observed in the basal epithelium and in cells in the central stroma, as in rejected buttons. The corneal endothelium remained negative for both HLA-ABC and HLA-DR antigens. Leu6 positive, HLA class I and II bearing Langerhan's cells were seen in the central and pericentral corneal epithelium. The distribution ofHLA antigens in normal cornea, and in rejected and herpetic buttons is shown schematically in Figure 5. 1482
DISCUSSION Our results indicate that HLA class I and II antigens are present on the corneal endothelium and on stromal keratocytes in rejected allografts. Class II antigens were also detected on cells in the basal epithelium and on Langerhan's cells in the epithelium. Whereas herpetic corneas also contained HLA-DR positive Langerhan's cells in the epithelium and class II positive basal epithelial cells, we could not demonstrate either HLA class I or II antigens on the endothelium. Cells in the stroma of herpetic buttons bore HLA class I and class II antigens, but the large number of infiltrating macrophages precluded differentiating at the light microscopic level whether specific cells
PEPOSE, et al •
HLA ANTIGENS IN REJECTED CORNEAL ALLOGRAFTS
represented keratocytes or macrophages. Thus, further studies are indicated to determine whether keratocytes express class II antigens in these buttons. It is of interest that the cellular immune response in the herpetic corneas was localized in the anterior and mid-stroma, but did not involve the posterior stroma as in the rejected allografts. The possible significance of this finding in relation to the failure to detect HLA antigens on the corneal endothelium in the herpetic buttons is discussed below. Normal human corneas were shown to express HLAABC (class I) antigens on epithelial cells and at lower levels on stromal keratocytes. Class I antigens were also readily detected on cells at the limbus (ie. macrophages, Langerhan's cells, and vascular endothelium), but were notably not detected on corneal endothelium. HLA-DR (class II) antigens were restricted to cells at the limbus (ie. macrophages, Langerhan's cells and vascular endothelium) and were not detected on corneal epithelium, stroma, or endothelium. These data are in agreement with the recent study by Whitsett and Stulting. 7 Similar results were obtained by Fujikawa et al,6 although HLA class I antigens were not detected on stromal keratocytes in their study. Following the original observation by Newsome et al,4 several investigators6,7 have confirmed the localization of HLA class I antigens on corneal epithelium, stroma and endothelium in tissue culture. In addition, fresh corneal endothelium, which is devoid of detectable HLA-ABC antigens, can be induced to express class I antigens by tissue culture conditions.6 Therefore, we thought it important to determine the effect of storage of the donor cornea in M-K medium, since any derepression of HLA expression may be undesirable for allografting. We found no change following 72-hour storage ofthe donor cornea in M-K medium on the expression of HLA class I or II antigens. The detection of Langerhan's cells in central and pericentral cornea in both rejected allografts and herpetic corneas are consistent with the results of the study by Gillette et a1. 21 However, in contrast to their findings, we did not detect Langerhan's cells in transverse sections of control cornea (by either HLA-DR or leu-6 staining) except at the limbus. Whether Langerhan's cells are present within the normal corneal donor button is a controversial 21 ,22 and important point, since these HLA-DR positive cells could theoretically contribute to the initiation of allograft rejection. Because we utilized a monoclonal antibody directed against the non polymorphic determinants of HLADR, it was not possible to determine whether these cells in rejected buttons expressed class II antigens of host or donor origin. However, our failure to detect Langerhan's cells in the non-limbal regions of control corneas and the limited life span of dendritic cells in other organs,23 suggest that the Langerhan's cells seen in the rejected corneal allografts are most likely of host origin. Our results are consistent with the notion that HLAclass I and II antigens are induced on corneal endothelial cells and class II antigens are induced on stromal keratocytes and basal epithelial cells in rejecting allografts. Studies of cadaveric renal transplantation in man have stressed the importance of HLA-DR (class II) compata-
bility between donor and recipient, and note a beneficial effect of HLA-A,B matching only in HLA-DR mismatched combination. 24 The aberrant localized expression of class II antigens is also becoming recognized in several autoimmune disorders such as autoimmune thyroiditis,2S in addition to skin allograft rejection 26 and graft vs. host disease. 27 A lymphokine, gamma interferon,28 has been shown to induce class II antigens and enhance the expression of class I antigens on human umbilical vein endothelial cells and dermal fibroblasts. Recent studies by Young et al 29 have demonstrated that HLA-DR antigens can be induced on cultured human corneal endothelial and stromal cells by gamma interferon. Therefore a lymphokine may similarly induce the expression of class II antigens on rejected corneal endothelium. The failure to detect class II antigens on the herpetic endothelium may reflect the restricted localization of the cellular inflammatory response limited to the anterior stroma with limited diffusion oflymphokines. One possible pathogenic scheme for corneal allograft rejection is as follows. In the majority of low risk corneal transplants, HLA class I incompatability (on corneal epithelium and stromal keratocytes) may not represent a strong enough transplantation barrier to initiate graft rejection. However, in transplants with localized nonspecific inflammation (e.g. loose sutures,30 vascularized recipient beds, subclinical infection) the production oflymphokines such as gamma interferon may result in the local induction of class I and II antigens on the corneal endothelium, keratocytes and basal epithelium. This may elicit further production of lymphokines, more inflammation and the subsequent induction of HLA-DR across the cornea in a vicious cycle. It is of interest that in addition to its proposed beneficial lymphocytotoxic effect in corneal allograft rejection,31 it has been suggested that glucocorticosteroids may partially inhibit the expression of HLA antigens in renal allograft rejection,32 and theoretically could playa similar role in corneal graft rejection. The results of this study indicate that class II antigens can be detected on human corneal transplants that are rejected. Our data provide further impetus to ongoing double blind clinical trials of lymphocytoxicity antibody screening and cross-match testing in high risk patients33 to determine the overall influence, as well as the significance of matching at any particular histocompatability locus.
REFERENCES 1. Khodadoust AA. The allograft rejection reaction: the leading cause of late failure of clinical corneal grafts. In: Corneal Graft Failure; Ciba Foundation Symposium 15 (new series). Amsterdam: Associated Scientific Publishers, 1973; 151-67. 2. McDevitt HO. Regulation of the immune response by the major histocompatability system. N Engl J Med 1980; 303:1514-7. 3. Ehlers N, Ahrons S. Corneal transplantation and histocompatability. Acta Ophthalmol1971; 49:513-27. 4. Newsome DA, Takasugi M, Kenyon KR, et al. Human corneal cells in vitro: morphology and histocompatability (HL-A) antigens of pure cell populations. Invest Ophthalmol1974; 13:23-32.
1483
OPHTHALMOLOGY
•
NOVEMBER 1985 •
5. Klareskog L, Forsum U, Tjemlund UM, et al. Expression of la antigenlike molecules on cells in the comeal epithelium. Invest Ophthalmol Vis Sci 1979; 18:310-3. 6. Fujikawa LS, Colvin RB, Bhan AK, et al. Expression of HLA-AfBfC and -DR locus antigens on epithelial, stromal, and endothelial cells of the human com ea. Cornea 1982; 1:213-22. 7. Whitsett CF, Stulting RD. The distribution of HLA antigens on human comeal tissue. Invest Ophthalmol Vis Sci 1984; 25:519-24. 8. Stark WJ, Opelz G, Newsome 0, et al. Sensitization to human lymphocyte antigens by corneal transplantation. Invest Ophthalmol1973; 12:639-45. 9. Morris PJ, Batchelor JR, Festenstein H. Matching for HLA in transplantation. Sr Med Bull 1978; 34:259-62. 10. Batchelor JR, Casey TA, Gibbs DC, et al. HLA matching and corneal grafting. Lancet 1976; 1:551-4. 11. Allansmith MR, Fine M, Payne R. Histocompatability typing and corneal transplantation. Trans Am Acad Ophthalmol Otolarnygol 1974; 78: OP445-60. 12. Kissmeyer-Nielson F, Ehlers N. Comeal transplantation and matching for HLA-A and HLA-B. In: Ferrone S, Curtoni ES, Gorini S, eds. HLA Antigens in Clinical Medicine and Biology. New York:Garland, 1979; 200-3. 13. Stark WJ, Taylor HR, Bias WS, Maumenee AE. Histocompatability (HLA) antigens and keratoplasty. Am J Ophthalmol 1978; 86:595604. 14. Vannas S, Karjalainen K. Ruusuvaara P, Tiilikainen A. HLA- compatible donor comea for prevention of allograft reaction. Albrecht von Graefes Arch Klin Exp Ophthalmol1976; 198:217-22. 15. Fronterre A, Trimarchi F, Bo G. HLA antigens and selection of donors in corneal transplants. Curr Ther Res 1980; 27:749-56. 16. Volker-Dieben HJM, Kok-Van Alphen CC, Kruit PJ. Advances and disappointments, indications and restrictions regarding HLA-matched corneal grafts in high risk cases. Doc Ophthalmol1979; 46:219-26. 17. Foulks GN, Sanfilippo FP, Locascio JA III, et al. Histocompatability testing for keratoplasty in high-risk patients. Ophthalmology 1983; 90: 239-43. 18. McCarey BE, Kaufman HE. Improved corneal storage. Invest Ophthalmol 1974; 13:165-73. 19. Stark WJ, Maumenee AE, Kenyon KR.lntermediate-term corneal stor-
1484
20.
21. 22. 23. 24.
25.
26.
27. 28.
29.
30. 31.
32.
33.
VOLUME 92
•
NUMBER 11
age for penetrating keratoplasty. Am J Ophthalmol 1975; 79:795802. Pepose JS, Nestor MS, Gardner KM, et al. Composition of cellular infiltrates in rejected human corneal allografts. Graefes Arch Clin Exp Ophthalmol 1985; 222: 128-33. Gillette TE, Chandler JW, Greiner JV. Langerhans cells of the ocular surface. Ophthalmology 1982; 89:700-10. Streilein JW, Toews GB, Bergstresser PRo Corneal allografts fail to express la antigens. Nature 1979; 282:326-7. Steinman RM, Nussenzweig MC. Dendritic cells: features and functions. Immunol Rev 1980; 53:127-47. Moen T, Albrechtsen 0, Flatmark A, et al. Importance of HLA-DR matching in cadaveric renal transplantation: a prospective one-center study of 170 transplants. N Engl J Med 1980; 303:850-4. Bottazzo GF, Pujol-Borrell R, Hanafusa T, Feldmann M. Role of aberrant HLA-DR expression and antigen presentation in induction of endocrine autoimmunity. Lancet 1983; 2: 1115-9. de Waal RMW, Bogman MJJ, Maass CN, et al. Variable expression of la antigens on the vascular endothelium of mouse skin allografts. Nature 1983; 303:426-9. Barclay AN, Mason Ow. Graft rejection and la antigens--paradox resolved? Nature 1983; 303:382-3. Pober JS, Collins T, Gimbrone MA Jr, et al. Lymphocytes recognize human vascular endothelial and dermal fibroblast la antigens induced by recombinant immune interferon. Nature 1983; 305:726-9. Young E, Stark WJ, Prendergast RA. Immunology of corneal allograft rejection: HLA-DR antigens on human comeal cells. Invest Ophthalmol Vis Sci 1985; 26:571-4. Paque J, Poirier RH. Corneal allograft reaction and its relationship to suture site neovascularization. Ophthalmic Surg 1977; 8:71-4. Polack FM. Lymphocyte destruction during corneal homograft reaction; a scanning electron microscopic study. Arch Ophthalmol 1973; 89: 413-6. Hayry P, von Willebrand E, Ahonen J, Eklund B. Glucocorticosteroids in renal transplantation. I. Impact of high- versus low-dose post-operative methylprednisolone administration on the first episode(s) of rejection. Scand J Immunol1982; 16:39-49. Stark WJ. Transplantation immunology of penetrating keratoplasty. Trans Am Ophthalmol Soc 1980; 78: 1079-117.