- Email: [email protected]
Histocompatibility (Hla) Antigens and Keratoplasty
AMERICAN JOURNAL OF OPHTHALMOLOGY VOLUME 86
NUMBER 5
NOVEMBER, 1978
HISTOCOMPATIBILITY (HLA) ANTIGENS AND KERATOPLASTY W A L T E R J. STARK, M.D., H U G H R. T A Y L O R , M.D., W I L M A B. B I A S , P H . D . , AND A. E D W A R D M A U M E N E E , M.D. Baltimore,
Allograft rejection occurs in about 10% of keratoplasty cases if the recipient cor nea is avascular. 1 - 3 However, if the preoperative corneal bed is densely vascularized, rejection can lead to graft failure in u p to 65% of cases. 1 ' 4 The exact role of donor-recipient histocompatibility (HLA) antigens in rejection after kerato plasty is not yet certain. In these high-risk cases of vascularized corneas, some have found a significant correlation between the number of donor and recipient HLA antigens matched and the graft out come,4*5 whereas others have not. 6 The HLA antigen system is the major transplantation antigen system in man as indicated by analysis of familial renal transplantation results. 7,8 Because of the complexity of the HLA antigen system, the probability for complete matching of donor and recipient antigens for unrelat ed, cadaver transplants is low. In renal transplantation, preimmunization of the recipient to HLA antigens of the donor increases the risk of rejection. 9-13 Howev er, for keratoplasty, the effect of recipient preimmunization has not been fully eval uated. From the Wilmer Ophthalmological Institute (Drs. Stark, Taylor, and Maumenee), and the Immunogenetics Laboratory (Dr. Bias) of the Johns Hop kins Hospital, Baltimore, Maryland. This work was supported in part by Grant EY01302 from the Na tional Eye Institute (Dr. Stark). Reprint requests to Walter J. Stark, M.D., Wilmer Institute, Johns Hopkins Hospital, 600 N. Wolfe St., Baltimore, MD 21205.
Maryland
Patients can become immunized and develop lymphocytotoxic antibodies by exposure to foreign HLA antigens from previous transplants, including corneal transplants, 1 4 blood transfusions, 10 or pregnancy. 1 5 Preimmunization of the re cipient to antigens of the donor may be detected by preoperative cross-match testing between the recipient's serum and the donor's lymphocytes. 1 3 , 1 6 A negative cross-match test indicates that the recipi ent has no circulating lymphocytotoxic antibodies specific for the antigens of the donor, whereas a positive cross-match test indicates that the recipient has previously been sensitized and developed antibod ies to one or more antigens of the donor. A positive donor-recipient cross-match before renal transplantation almost al ways results in hyperacute allograft re jection. 13,16 We report herein our results of pene trating keratoplasty in 84 potentially preimmunized recipients with severe pre operative corneal vascularization. The donor corneas that were used for these patients were selected on the basis of obtaining a negative donor-recipient cross-match test before surgery. No at tempt was made to match donor and re cipient HLA antigens specifically. M A T E R I A L AND M E T H O D S
Eighty-four penetrating keratoplasties were performed between November 1973 and November 1977. Patients selected for
AMERICAN JOURNAL O F OPHTHALMOLOGY 86:595-604, 1978
595
596
AMERICAN JOURNAL OF OPHTHALMOLOGY
this study had significant vascularization of the corneal stroma in at least three quadrants, extending into the visual axis, Of the 84 patients studied, 64 (76%) had previously had at least one corneal trans plant failure in that same eye. These 64 patients had an average of 1.9 previous graft failures, with a range of from one to six previous graft failures in that eye. The remaining 20 patients had previously been exposed to foreign HLA antigens by keratoplasty in the other eye, or by blood transfusion or pregnancy. Of the 84 pa tients 50 were female and 34 were male. The age of recipients ranged from 13 to 80 years (mean, 60 years). The original diag nosis before the first graft in the eyes studied included: vascularized corneal scars, 31 eyes; corneal edema, 17 eyes; corneal dystrophies, ten eyes; alkali burns, eight eyes; and vascularized corne as of unknown cause, 14 eyes. Preoperative serum from each recipient was screened for lymphocytotoxic anti bodies against a standard panel of lym phocytes from 50 selected donors that, collectively, had all of the known HLA antigens. Each serum was tested against the 50 samples of lymphocytes in the stan dard donor panel by the microlymphocytotoxicity test, 17 which was performed as follows: crossmatch trays of sera from all patients awaiting corneal grafts were pre pared and stored at —80°C. This was done by placing ljxl of each serum in a well of a microtest tray that contained 1 (JLI of min eral oil. The trays also included positive and negative control sera. These prepared trays were tested against the standard panel of lymphocytes and against corne al donor lymphocytes by adding 1 JJLI of 1 X 10G cells/ml of lymphocyte suspen sion to each well. Cells and antisera were incubated for 30 minutes at room temper ature. After incubation, excess antibody was washed from the well by adding 10 u.1 of barbital buffer, allowing cells to settle, and "flicking" buffer from the
NOVEMBER, 1978
wells. Four microliters of previously titered rabbit complement was added to each well, and incubation was continued for one hour. Trypan blue was then added to stain cells that were killed in the cytotoxic reaction. Reactions were determined microscopically, and greater than 20% cell death was considered a positive reac tion. Cytotoxic reactivity of a recipient's serum against 2% or more of the 50 sam ples in the panel of donor lymphocytes was considered a significant antibody lev el. A patient's serum with less than 2% reactivity against the standard panel was determined as having no significant amount of detectable lymphocytotoxic antibody. Preoperatively, sera from all kerato plasty patients were cross-matched indi vidually against lymphocytes of all the potential corneal donors, as described for the microlymphocytotoxicity test. A spe cific donor cornea for an individual recip ient was selected on the basis of a nega tive cross-match reaction with the donor's lymphocytes. If a positive cross-match was obtained, the cornea was not used for that recipient. Several of the potentially preimmunized recipients, who had no significantly detectable lymphocytotoxic antibodies on the 50-sample screening tests against our standard panel, still showed a positive cross-match reaction against lymphocytes from some of the potential corneal donors. Therefore, the sera of all. of the keratoplasty patients were added to the crossmatch trays re gardless of whether they appeared to have antibodies by panel screening. Addition ally, HLA typing of both donor and recip ient was obtained, but no attempt was made to match donor and recipient anti gens specifically. Donor material for these 84 cases was stored in McCarey-Kaufman medium 1 8 until used for keratoplasty. Surgery was performed within 60 hours of the donor's death. All grafts were sutured with 10-0
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
597
ma, glaucoma or persistent epithelial de fects with exposure keratitis. RESULTS
Fig. 1 (Stark and associates). Allograft reaction: keratic precipitates on endothelium of donor cornea.
monofilament nylon suture. 2 Of the 84 grafts, 70 were 7.5 mm in diameter, ten were 8.0 mm in diameter, and four were 7.0 mm in diameter. Forty-four grafts were performed on phakic eyes, and 19 of these were combined with lens removal. The remaining 40 grafts were performed on aphakic eyes. The diagnosis of immune graft rejec tion was made only if the graft remained clear for at least two weeks after surgery, and then developed edema or clouding associated with keratic precipitates on the graft endothelium or a rejection line (Figs. 1 and 2). Nonimmune causes of graft failure included grafts that clouded within two weeks of surgery, or graft clouding associated with infection, trau-
Fig. 2 (Stark and associates). Allograft rejection: rejection line on endothelial surface of donor cornea (arrows denote advancing edge of rejection line).
Negative donor-recipient cross-match pairs—Postoperatively 65 of the 84 corneal grafts (77%) were clear (Figs. 3-5) with an average follow-up of 14.3 months (3 to 45 months). Of the grafts, 15% (13 of 84) have failed from immune allograft rejec tion, and 7% have failed from causes other than rejection (Table 1). Of graft failures from rejection, 92% (12 of 13) occurred within six months of surgery, with the average time for occurrence of rejection being 3.6 months. Of the 84 patients, 22 (26%) had evidence of an immune event, either rejection or a re versible allograft reaction. Of the 22 im mune events, nine (41%) could be re versed with corticosteroids, and 13 (59%) of the cases progressed to graft clouding (Table 2). The incidence of allograft rejection or the reversibility of an immune event did not differ in those with positive preoperative lymphocytotoxic antibody reactivity of greater than or equal to 2% and those without detectable antibodies (Tables 1 and 2). The incidence of graft rejection is com pared with the number of previous grafts in the same eye in Table 3. A trend exists for an increased risk of rejection with a greater number of previous grafts. A simi lar trend is seen with the total number of immune events (Table 4), although there is no difference in reversibility of these immune events. In 84 keratoplasty cases with dense corneal vascularization, our incidence of graft rejection, which was 15%, is less than that reported in similar high-risk cases. Khodadoust 1 reported graft failure from rejection in 65% of cases with dense corneal vascularization, and Batchelor and co-workers 4 reported 60% of such cases with graft failure from rejection.
598
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978
Fig. 3 (Stark and associates). Left, Preoperative appearance of eye after 3 corneal transplant failures from rejection. Right, Postoperative appearance of same eye ten months after keratoplasty. Graft is clear with no evidence of rejection.
Fig. 4 (Stark and associates). Left, Preoperative appearance of eye with dense corneal vascularization after two previous graft failures. Right, Postoperative appearance of same eye 9 months after keratoplasty with no evidence of rejection.
IA Fig. 5 (Stark and associates). Left, Preoperative appearance of eye with 4 t corneal vascularization. Other eye was lost from complications after keratoplasty. Right, Postoperative appearance of same eye. Vessels are surrounding graft but there is no evidence of rejection 12 months after keratoplasty.
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
599
TABLE 1 O U T C O M E O F KERATOPLASTY IN P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION*
Recipient Preoperative Lymphocytotoxic Antibody Screening Level
Total
Allograft Rejection
Non-Immune Graft Failure
Total Clear
Mean Follow-up (range-mos)
&2% <2% Total
48 36 84
6(13%) 7(19%) 13(15%)
4(8%) 2(6%) 6(7%)
38(79%) 27(75%) 65(77%)
14.0(3-40) 14.9(3-45) 14.3(3-45)
Nu imber of Patients
'Negative donor-recipient cross-match.
Although there may have been differ ences between their series and ours, such as the quality of donor material or patient follow-up availability, these factors are unlikely to account for all of the large differences in incidence of rejection fail ures between those series and ours. A major cause of the lower incidence of rejection in our series is probably the use of negatively cross-matched donor mate rial. These findings for keratoplasty are in agreement with the results in other tissue and organ transplantation systems. 13,16 Retrospective donor-recipient HLA an tigen matching—The donor and recipient HLA typing results were known for the 84 patients who received negative crossmatched grafts. There were 19 additional patients for whom the donor and recipi ent HLA types were known. These addi
tional 19 patients also had dense corneal vascularization, but the cross-match test had not been performed. The donorrecipient HLA antigen matching was ana lyzed retrospectively for these 103 pa tients. The occurrence of graft rejection was compared with the number of matching donor and recipient antigens (Table 5). When the donor and recipient shared no matching antigens, rejection failures oc curred in 14% of the cases (7 of 51). When one matching antigen was shared, rejec tion developed in 18% of the cases (6 of 34). In 11 cases in which the donor and recipient shared two matching antigens, there were no graft failures from rejec tion. However, rejection accounted for failure in two of five cases (40%) in which three matching antigens were shared. In
TABLE 2 IMMUNE REACTIONS IN POTENTIALLY PRE-IMMUNIZED RECIPIENTS W I T H D E N S E CORNEAL VASCULARIZATION
Recipient Preoperative Lymphocytotoxic Antibody Screening Level
Total Number of Grafts
Total Number of Immune Events (Rejection or Reaction)
Allograft Reaction (Graft Clear)
Percentage Reversed
£2% <2% Total
48 36 84
9(19%) 13(36%) 22(26%)
3 6 9
33 46 41
600
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978
TABLE 3 C O M P A R I S O N O F N U M B E R O F PREVIOUS G R A F T S IN O P E R A T E D E Y E AND O U T C O M E O F KERATOPLASTY IN P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION*
Number of Previous Grafts
Total
AUograft Rejection
Non-Immune Graft Failure
Total Clear
0 1 2 3-6
20 24 28 12
1(5%) 3(13%) 5(18%) 4(25%)
2 2 2 0
17 19 21 8
Number of Patients
*Negative donor-recipient cross-match.
and recipients shared increasingly more matched antigens. Our work does not confirm this in similar high-risk cases (Fig. 6). However, most of our patients received corneas from negatively crossmatched donors. Batchelor and coworkers 4 did not report having used cross-match testing for donor selection, although the probability of negative cross-matches will be higher with donorrecipient pairs that are well-matched for specific antigens. Vannas and co-workers 5 prospectively matched donor and recipi ent antigens and obtained a negative donor-recipient cross-match test in 18 keratoplasty cases with all degrees of cor neal vascularization. Their cases showed a lower incidence of rejection as the num-
one case the donor and recipient shared four matching antigens, and this graft was clear. The incidence of graft failure from re jection does not correlate with the num ber of matching HLA antigens shared between the donor and recipient. Also, the incidence of immune events and the reversibility of allograft reaction did not correlate with the number of antigens shared (Table 6). Batchelor and co-workers 4 have the largest series published to date regarding donor-recipient HLA antigen matching and the graft outcome. Seventy-three keratoplasty recipients with dense corneal vascularization showed a reduction in graft failure from rejection as the donor
TABLE 4 C O M P A R I S O N O F NUMBER O F PREVIOUS G R A F T S IN O P E R A T E D E Y E AND I M M U N E REACTIONS I N P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION
Number of Previous Grafts
Total Number of Grafts
Total Number of Immune Events (Rejection or Reaction)
Allograft Reaction (Graft Clear)
Percentage Reversed
0 1 2 3-6
20 24 28 12
2 5 9 6
1 2 4 2
50 40 44 33
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
601
TABLE 5 DONOR-RECIPIENT H L A ANTIGEN MATCHES AND GRAFT OUTCOME
Number of HLA Antigens Matched
Number of Patients
Allograft Rejection
Nonimmune Graft Failure
Clear Graft
0 1 2 3 4 Total
51 35 11 5 1 103
7(14%) 6(17%) 0 2(40%) 0 15(15%)
5(10%) 1(3%) 0 0 0 6(6%)
39(76%) 28(80%) 11(100%) 3(60%) 1(100%) 82(80%)
ber of matched donor-recipient antigens increased (Fig. 7). However, Vannas and co-workers' series included cases with any degree of corneal vascularization, and their recipients may not have had as high a risk of graft rejection as those included in our series and those reported by Batchelor and co-workers. 4 DISCUSSION
The role of humoral immunity in cor neal graft rejection is unknown. Previous work suggests that corneal graft rejection is mediated by cellular immune re sponses. 1 9 However, the HLA serologic reactions can be used as indicators of recipient preimmunization, because the major gene loci controlling cell-mediated responses also map within the HLA com plex. Thus far, about ten loci, all control ling some aspect of immune function,
have been assigned to the HLA region of chromosome 6 in man (Fig. 8). 20 The D locus (or subregion) controls mixed lym phocyte responsiveness, and is an in vitro indicator of cellular immune responsive ness. The entire HLA complex occupies a small segment of the chromosomal length, allowing identification of a given haplotype (the gene complex on one chro mosome of the pair) by the serologic specificities it contains. Our results indicated a lower incidence of corneal graft failure from rejection when cross-match testing was used for donor selection than did other series in volving similar high-risk cases with dense corneal vascularization. 1,4 We believe this difference is due mainly to our use of preoperative donor-recipient cross-match testing as a means for donor selection. Thus, we are able to avoid giving a donor
TABLE 6 DONOR-RECIPIENT H L A ANTIGEN MATCH AND GRAFT OUTCOME
Number of HLA Antigens Matched
Number of Patients
Total Number of Immune Events (Reaction or Rejection)
Allograft Reaction (Graft Clear)
0 1 2 3 4 Total
51 35 11 5 1 103
14(27%) 9(26%) 1(9%) 2(40%) 0 26(25%)
7 3 1 0 0 11
Percent Reversed 50 33 100 0 42
602
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978 HLA Region
I STARK ] BATCHELOU
-C4,Ch,Rg-
PGM,
GLO
-i-£i—\-i
L
3
D
Bf
B C
I
I
1I
MLR" A
M-
Map distances in centimorgans
cornea to a recipient who is presensitized to that cornea's antigens. 21 Although we found no consistent corre lation between the number of matching HLA antigens shared between the donor and recipient and graft outcome, the role of antigen matching deserves further evaluation. However, if further attempts at donor-recipient HLA antigen matching are made, they must be combined with preoperative cross-match testing, because even if all but one of the known donorrecipient antigens are favorably matched,
I
1-
Fig. 8 (Stark and associates). Map of HLA region of chromosome 6.
Fig. 6 (Stark and associates). The incidence of rejection after keratoplasty in recipients who have shared zero, one, two, three or four HLA antigens with their donor. Comparison between this study and that of Batchelor and co-workers. 4
LLi I
C2
the recipient may still be preimmunized to the one unmatched donor antigen or to unrecognized donor HLA antigens. Al though perfect donor-recipient HLA matching for corneal transplantation may be beneficial, it is not currently practical because of the complexity of the HLA system (Table 7). 22 Our study indicates that the use of cross-matching to avoid donors with transplantation antigens to which the recipient is immunized is feasi ble and seems beneficial. We do not have positive donorrecipient cross-match controls, and we therefore consider this to be a prelimiary report. A well-controlled clinical trial is necessary to evaluate fully the results of preoperative donor-recipient cross-match testing for donor selection in keratoplas ty, and such a study is currently under way. If controlled studies continue to show a significant beneficial effect from using cross-match testing in high-risk cases for keratoplasty, then this procedure could easily be coordinated with kidney transplant projects without the need for additional expensive equipment and peronnel.
N" 13
SUMMARY 3-4
Fig. 7 (Stark and associates). The incidence of rejection after keratoplasty and the number of HLA antigens shared by recipient and donor. Comparison between this study and that of Vannas and coworkers. 5
In 84 cases of keratoplasty, all patients had densely vascularized corneas and a history of prior exposure to HLA antigens by pregnancy, blood transfusion, or pre vious transplantation. Donors were se lected on the basis of a negative cross-
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
603
TABLE 7 L I S T I N G O F R E C O G N I Z E D HLA S P E C I F I C I T I E S ( 1 9 7 8 )
A Locus
B Locus
C Locus
D Locus
HLA—Al HLA—A2 HLA—A3 HLA—A9 HLA—A10 HLA—All HLA—A28 HLA—A29 HLA—AW23 HLA—AW24 HLA—AW25 HLA—AW26 HLA—AW30 HLA—AW31 HLA—AW32 HLA—AW33 HLA—AW34 HLA—AW36 HLA—AW43 Blank
HLA—B5 HLA—B7 HLA—B8 HLA—B12 HLA—B13 HLA—B14 HLA—B15 HLA—B17 HLA—B18 HLA—B27 HLA—B37 HLA—BW16 HLA—BW21 HLA—BW22 HLA—BW35 HLA—BW38 HLA—BW39 HLA—BW40 HLA—BW41 HLA—BW42 HLA—BW44 HLA—BW45 HLA—BW46 HLA—BW47 HLA—BW48 HLA—BW49 HLA—BW50 HLA—BW51 HLA—BW52 HLA—BW53 HLA—BW54 Blank
HLA- -CW1 HLA- -CW2 HLA--CW3 HLA--CW4 HLA--CW5 HLA- -CW6 Blank
HLA—DW1 HLA—DW2 HLA—DW3 HLA—DW4 HLA—DW5 HLA—DW6 HLA—DW7 HLA—DW8 HLA—DW9 HLA—DW10 HLA—DW11 Blank
match to avoid donor antigens to which the recipient was preimmunized. T h e overall rate of graft failure from rejection was 15%. A retrospective analysis of donor-recipient HLA matching in 103 high-risk cases showed no consistent cor relation between the n u m b e r of antigens shared and graft outcome. These findings indicated negative cross-match to be im portant in donor selection for keratoplasty in high-risk cases.
REFERENCES 1. Khodadoust, A. A.: The allograft rejection reac tion. T h e leading cause of late failure of clinical corneal grafts. In Corneal Graft Failure, Ciba Foun dation Symposium. Amsterdam, Elsevier, 1973, p .
2. Stark, W. J., Paton, D., Maumenee, A. E., and Michelson, P. E.: T h e results of 102 penetrating keratoplasties using 10-0 monofilament nylon su ture. Ophthalmic Surg. 3:11, 1972. 3. Chandler, J. W., and Kaufman, H. E.: Graft reactions after keratoplasty for keratoconus. Am. J. Ophthalmol. 77:543, 1974. 4. Batchelor, J. R., Casey, T. A., Gibbs, D. C , Lloyd, D. F., Werb, A., Prasad, S. S., and James, A.: HLA matching and corneal grafting. Lancet 1:551, 1976. 5. Vannas, S., Karjalainen, K., Ruusuvaara, P., and Tiilikainen, A.: HLA-compatible donor cornea for prevention of allograft reaction. Albrecht von Graefes Arch. Klin. Ophthalmol. 198:217, 1976. 6. Allansmith, M. R., Fine, M., and Payne, R.: Histocompatibility typing and corneal transplanta tion. Trans. Am. Acad. Ophthalmol. Otolaryngol. 78:445, 1974. 7. Opelz, G., Mickey, M. R., and Terasaki, P. I.: Calculations on long-term graft survival in human kidney transplantation. Transplant. Proc. 9:27, 1977. 8. Mickey, R. M., Kreisler, M., Albert, E. D.,
604
AMERICAN JOURNAL OF OPHTHALMOLOGY
Tanaka, N., and Terasaki, P. I.: Analysis of HL-A incompatibility in human renal transplants. Tissue Antigens 2:57, 1971. 9. Terasaki, P. I., Kreisler, M., and Mickey, R. M.: Presensitization and Kidney transplant failures. Postgrad. Med. J. 47:89, 1971. 10. Opelz, G., Terasaki, P. I.: Histocompatibility matching utilizing responsiveness as a new dimen sion. Transplant. Proc. 4:433, 1971. 11. Opelz, G., Sengar, D. P. S., Mickey, R. M., and Terasaki, P. I.: Effect of blood transfusions on subsequent kidney transplants. Transplant. Proc. 5:253, 1973. 12. Opelz, G., Mickey, R. M., and Terasaki, P. I.: Identification of unresponsive kidney-transplant re cipients. Lancet 1:868, 1972. 13. Terasaki, P. I., Thrasher, D. L., Hauber, T. H.: Serotyping for homotransplantation. 13. Im mediate kidney transplant rejection and associated preformed antibodies. In Dausset, J., Hamburger, J., and Mathe, G. (eds.): Advance in Transplantation. Baltimore, Williams and Wilkins, 1968, p . 225. 14. Stark, W. J., Opelz, G., Newsome, D., Brown, R., Yankee, R., and Terasaki, P. I.: Sensitization to human lymphocyte antigens by corneal transplanta tion. Invest. Ophthalmol. 12:639, 1973. 15. Dausset, J.: Leuco-agglutinins. 4. Leuco-
NOVEMBER, 1978
agglutinins and blood transfusion. Vox Sang. 4:190, 1954. 16. Kissmeyer-Nielsen, F., Olsen, S., PosborgPetersen, V., and Fjeldborg, O.: Hyperacute rejec tion of kidney allografts associated with pre-existing humoral antibodies against donor cells. Lancet 2: 662, 1966. 17. Amos, D. B., Bashir, H., Boyle, W., MacQueen, M., and Tiilikainen, A.: A simple microcytotoxicity test. Transplantation 7:220, 1970. 18. McCarey, B. E., and Kaufman, H. E.: Im proved corneal storage. Invest'. Ophthalmol. 13:165, 1974. 19. Polack, F. M., and Gonzales, C. E.: The re sponse of the lymphoid tissue to corneal heterografts. Arch. Ophthalmol. 80:321, 1968. 20. Bias, W. B., and Hsu, S. H.: HLA, the major histocompatibility system in man. An overview. Acta Anthropogenetica. 1:15, 1977. 21. Newsome, D. A., Takasugi, M., Kenyon, K. R., Stark, W. J., and Opelz, G.: Human corneal cells in vitro: morphology and histocompatibility (HL-A) antigens of pure cell populations. Invest. Ophthalmol. 13:23, 1974. 22. World Health Organization Committee: No menclature for factors of the HLA system 1977. Tissue Antigens 11:81, 1978.
NUMBER 5
NOVEMBER, 1978
HISTOCOMPATIBILITY (HLA) ANTIGENS AND KERATOPLASTY W A L T E R J. STARK, M.D., H U G H R. T A Y L O R , M.D., W I L M A B. B I A S , P H . D . , AND A. E D W A R D M A U M E N E E , M.D. Baltimore,
Allograft rejection occurs in about 10% of keratoplasty cases if the recipient cor nea is avascular. 1 - 3 However, if the preoperative corneal bed is densely vascularized, rejection can lead to graft failure in u p to 65% of cases. 1 ' 4 The exact role of donor-recipient histocompatibility (HLA) antigens in rejection after kerato plasty is not yet certain. In these high-risk cases of vascularized corneas, some have found a significant correlation between the number of donor and recipient HLA antigens matched and the graft out come,4*5 whereas others have not. 6 The HLA antigen system is the major transplantation antigen system in man as indicated by analysis of familial renal transplantation results. 7,8 Because of the complexity of the HLA antigen system, the probability for complete matching of donor and recipient antigens for unrelat ed, cadaver transplants is low. In renal transplantation, preimmunization of the recipient to HLA antigens of the donor increases the risk of rejection. 9-13 Howev er, for keratoplasty, the effect of recipient preimmunization has not been fully eval uated. From the Wilmer Ophthalmological Institute (Drs. Stark, Taylor, and Maumenee), and the Immunogenetics Laboratory (Dr. Bias) of the Johns Hop kins Hospital, Baltimore, Maryland. This work was supported in part by Grant EY01302 from the Na tional Eye Institute (Dr. Stark). Reprint requests to Walter J. Stark, M.D., Wilmer Institute, Johns Hopkins Hospital, 600 N. Wolfe St., Baltimore, MD 21205.
Maryland
Patients can become immunized and develop lymphocytotoxic antibodies by exposure to foreign HLA antigens from previous transplants, including corneal transplants, 1 4 blood transfusions, 10 or pregnancy. 1 5 Preimmunization of the re cipient to antigens of the donor may be detected by preoperative cross-match testing between the recipient's serum and the donor's lymphocytes. 1 3 , 1 6 A negative cross-match test indicates that the recipi ent has no circulating lymphocytotoxic antibodies specific for the antigens of the donor, whereas a positive cross-match test indicates that the recipient has previously been sensitized and developed antibod ies to one or more antigens of the donor. A positive donor-recipient cross-match before renal transplantation almost al ways results in hyperacute allograft re jection. 13,16 We report herein our results of pene trating keratoplasty in 84 potentially preimmunized recipients with severe pre operative corneal vascularization. The donor corneas that were used for these patients were selected on the basis of obtaining a negative donor-recipient cross-match test before surgery. No at tempt was made to match donor and re cipient HLA antigens specifically. M A T E R I A L AND M E T H O D S
Eighty-four penetrating keratoplasties were performed between November 1973 and November 1977. Patients selected for
AMERICAN JOURNAL O F OPHTHALMOLOGY 86:595-604, 1978
595
596
AMERICAN JOURNAL OF OPHTHALMOLOGY
this study had significant vascularization of the corneal stroma in at least three quadrants, extending into the visual axis, Of the 84 patients studied, 64 (76%) had previously had at least one corneal trans plant failure in that same eye. These 64 patients had an average of 1.9 previous graft failures, with a range of from one to six previous graft failures in that eye. The remaining 20 patients had previously been exposed to foreign HLA antigens by keratoplasty in the other eye, or by blood transfusion or pregnancy. Of the 84 pa tients 50 were female and 34 were male. The age of recipients ranged from 13 to 80 years (mean, 60 years). The original diag nosis before the first graft in the eyes studied included: vascularized corneal scars, 31 eyes; corneal edema, 17 eyes; corneal dystrophies, ten eyes; alkali burns, eight eyes; and vascularized corne as of unknown cause, 14 eyes. Preoperative serum from each recipient was screened for lymphocytotoxic anti bodies against a standard panel of lym phocytes from 50 selected donors that, collectively, had all of the known HLA antigens. Each serum was tested against the 50 samples of lymphocytes in the stan dard donor panel by the microlymphocytotoxicity test, 17 which was performed as follows: crossmatch trays of sera from all patients awaiting corneal grafts were pre pared and stored at —80°C. This was done by placing ljxl of each serum in a well of a microtest tray that contained 1 (JLI of min eral oil. The trays also included positive and negative control sera. These prepared trays were tested against the standard panel of lymphocytes and against corne al donor lymphocytes by adding 1 JJLI of 1 X 10G cells/ml of lymphocyte suspen sion to each well. Cells and antisera were incubated for 30 minutes at room temper ature. After incubation, excess antibody was washed from the well by adding 10 u.1 of barbital buffer, allowing cells to settle, and "flicking" buffer from the
NOVEMBER, 1978
wells. Four microliters of previously titered rabbit complement was added to each well, and incubation was continued for one hour. Trypan blue was then added to stain cells that were killed in the cytotoxic reaction. Reactions were determined microscopically, and greater than 20% cell death was considered a positive reac tion. Cytotoxic reactivity of a recipient's serum against 2% or more of the 50 sam ples in the panel of donor lymphocytes was considered a significant antibody lev el. A patient's serum with less than 2% reactivity against the standard panel was determined as having no significant amount of detectable lymphocytotoxic antibody. Preoperatively, sera from all kerato plasty patients were cross-matched indi vidually against lymphocytes of all the potential corneal donors, as described for the microlymphocytotoxicity test. A spe cific donor cornea for an individual recip ient was selected on the basis of a nega tive cross-match reaction with the donor's lymphocytes. If a positive cross-match was obtained, the cornea was not used for that recipient. Several of the potentially preimmunized recipients, who had no significantly detectable lymphocytotoxic antibodies on the 50-sample screening tests against our standard panel, still showed a positive cross-match reaction against lymphocytes from some of the potential corneal donors. Therefore, the sera of all. of the keratoplasty patients were added to the crossmatch trays re gardless of whether they appeared to have antibodies by panel screening. Addition ally, HLA typing of both donor and recip ient was obtained, but no attempt was made to match donor and recipient anti gens specifically. Donor material for these 84 cases was stored in McCarey-Kaufman medium 1 8 until used for keratoplasty. Surgery was performed within 60 hours of the donor's death. All grafts were sutured with 10-0
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
597
ma, glaucoma or persistent epithelial de fects with exposure keratitis. RESULTS
Fig. 1 (Stark and associates). Allograft reaction: keratic precipitates on endothelium of donor cornea.
monofilament nylon suture. 2 Of the 84 grafts, 70 were 7.5 mm in diameter, ten were 8.0 mm in diameter, and four were 7.0 mm in diameter. Forty-four grafts were performed on phakic eyes, and 19 of these were combined with lens removal. The remaining 40 grafts were performed on aphakic eyes. The diagnosis of immune graft rejec tion was made only if the graft remained clear for at least two weeks after surgery, and then developed edema or clouding associated with keratic precipitates on the graft endothelium or a rejection line (Figs. 1 and 2). Nonimmune causes of graft failure included grafts that clouded within two weeks of surgery, or graft clouding associated with infection, trau-
Fig. 2 (Stark and associates). Allograft rejection: rejection line on endothelial surface of donor cornea (arrows denote advancing edge of rejection line).
Negative donor-recipient cross-match pairs—Postoperatively 65 of the 84 corneal grafts (77%) were clear (Figs. 3-5) with an average follow-up of 14.3 months (3 to 45 months). Of the grafts, 15% (13 of 84) have failed from immune allograft rejec tion, and 7% have failed from causes other than rejection (Table 1). Of graft failures from rejection, 92% (12 of 13) occurred within six months of surgery, with the average time for occurrence of rejection being 3.6 months. Of the 84 patients, 22 (26%) had evidence of an immune event, either rejection or a re versible allograft reaction. Of the 22 im mune events, nine (41%) could be re versed with corticosteroids, and 13 (59%) of the cases progressed to graft clouding (Table 2). The incidence of allograft rejection or the reversibility of an immune event did not differ in those with positive preoperative lymphocytotoxic antibody reactivity of greater than or equal to 2% and those without detectable antibodies (Tables 1 and 2). The incidence of graft rejection is com pared with the number of previous grafts in the same eye in Table 3. A trend exists for an increased risk of rejection with a greater number of previous grafts. A simi lar trend is seen with the total number of immune events (Table 4), although there is no difference in reversibility of these immune events. In 84 keratoplasty cases with dense corneal vascularization, our incidence of graft rejection, which was 15%, is less than that reported in similar high-risk cases. Khodadoust 1 reported graft failure from rejection in 65% of cases with dense corneal vascularization, and Batchelor and co-workers 4 reported 60% of such cases with graft failure from rejection.
598
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978
Fig. 3 (Stark and associates). Left, Preoperative appearance of eye after 3 corneal transplant failures from rejection. Right, Postoperative appearance of same eye ten months after keratoplasty. Graft is clear with no evidence of rejection.
Fig. 4 (Stark and associates). Left, Preoperative appearance of eye with dense corneal vascularization after two previous graft failures. Right, Postoperative appearance of same eye 9 months after keratoplasty with no evidence of rejection.
IA Fig. 5 (Stark and associates). Left, Preoperative appearance of eye with 4 t corneal vascularization. Other eye was lost from complications after keratoplasty. Right, Postoperative appearance of same eye. Vessels are surrounding graft but there is no evidence of rejection 12 months after keratoplasty.
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
599
TABLE 1 O U T C O M E O F KERATOPLASTY IN P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION*
Recipient Preoperative Lymphocytotoxic Antibody Screening Level
Total
Allograft Rejection
Non-Immune Graft Failure
Total Clear
Mean Follow-up (range-mos)
&2% <2% Total
48 36 84
6(13%) 7(19%) 13(15%)
4(8%) 2(6%) 6(7%)
38(79%) 27(75%) 65(77%)
14.0(3-40) 14.9(3-45) 14.3(3-45)
Nu imber of Patients
'Negative donor-recipient cross-match.
Although there may have been differ ences between their series and ours, such as the quality of donor material or patient follow-up availability, these factors are unlikely to account for all of the large differences in incidence of rejection fail ures between those series and ours. A major cause of the lower incidence of rejection in our series is probably the use of negatively cross-matched donor mate rial. These findings for keratoplasty are in agreement with the results in other tissue and organ transplantation systems. 13,16 Retrospective donor-recipient HLA an tigen matching—The donor and recipient HLA typing results were known for the 84 patients who received negative crossmatched grafts. There were 19 additional patients for whom the donor and recipi ent HLA types were known. These addi
tional 19 patients also had dense corneal vascularization, but the cross-match test had not been performed. The donorrecipient HLA antigen matching was ana lyzed retrospectively for these 103 pa tients. The occurrence of graft rejection was compared with the number of matching donor and recipient antigens (Table 5). When the donor and recipient shared no matching antigens, rejection failures oc curred in 14% of the cases (7 of 51). When one matching antigen was shared, rejec tion developed in 18% of the cases (6 of 34). In 11 cases in which the donor and recipient shared two matching antigens, there were no graft failures from rejec tion. However, rejection accounted for failure in two of five cases (40%) in which three matching antigens were shared. In
TABLE 2 IMMUNE REACTIONS IN POTENTIALLY PRE-IMMUNIZED RECIPIENTS W I T H D E N S E CORNEAL VASCULARIZATION
Recipient Preoperative Lymphocytotoxic Antibody Screening Level
Total Number of Grafts
Total Number of Immune Events (Rejection or Reaction)
Allograft Reaction (Graft Clear)
Percentage Reversed
£2% <2% Total
48 36 84
9(19%) 13(36%) 22(26%)
3 6 9
33 46 41
600
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978
TABLE 3 C O M P A R I S O N O F N U M B E R O F PREVIOUS G R A F T S IN O P E R A T E D E Y E AND O U T C O M E O F KERATOPLASTY IN P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION*
Number of Previous Grafts
Total
AUograft Rejection
Non-Immune Graft Failure
Total Clear
0 1 2 3-6
20 24 28 12
1(5%) 3(13%) 5(18%) 4(25%)
2 2 2 0
17 19 21 8
Number of Patients
*Negative donor-recipient cross-match.
and recipients shared increasingly more matched antigens. Our work does not confirm this in similar high-risk cases (Fig. 6). However, most of our patients received corneas from negatively crossmatched donors. Batchelor and coworkers 4 did not report having used cross-match testing for donor selection, although the probability of negative cross-matches will be higher with donorrecipient pairs that are well-matched for specific antigens. Vannas and co-workers 5 prospectively matched donor and recipi ent antigens and obtained a negative donor-recipient cross-match test in 18 keratoplasty cases with all degrees of cor neal vascularization. Their cases showed a lower incidence of rejection as the num-
one case the donor and recipient shared four matching antigens, and this graft was clear. The incidence of graft failure from re jection does not correlate with the num ber of matching HLA antigens shared between the donor and recipient. Also, the incidence of immune events and the reversibility of allograft reaction did not correlate with the number of antigens shared (Table 6). Batchelor and co-workers 4 have the largest series published to date regarding donor-recipient HLA antigen matching and the graft outcome. Seventy-three keratoplasty recipients with dense corneal vascularization showed a reduction in graft failure from rejection as the donor
TABLE 4 C O M P A R I S O N O F NUMBER O F PREVIOUS G R A F T S IN O P E R A T E D E Y E AND I M M U N E REACTIONS I N P O T E N T I A L L Y P R E - I M M U N I Z E D R E C I P I E N T S W I T H D E N S E CORNEAL VASCULARIZATION
Number of Previous Grafts
Total Number of Grafts
Total Number of Immune Events (Rejection or Reaction)
Allograft Reaction (Graft Clear)
Percentage Reversed
0 1 2 3-6
20 24 28 12
2 5 9 6
1 2 4 2
50 40 44 33
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
601
TABLE 5 DONOR-RECIPIENT H L A ANTIGEN MATCHES AND GRAFT OUTCOME
Number of HLA Antigens Matched
Number of Patients
Allograft Rejection
Nonimmune Graft Failure
Clear Graft
0 1 2 3 4 Total
51 35 11 5 1 103
7(14%) 6(17%) 0 2(40%) 0 15(15%)
5(10%) 1(3%) 0 0 0 6(6%)
39(76%) 28(80%) 11(100%) 3(60%) 1(100%) 82(80%)
ber of matched donor-recipient antigens increased (Fig. 7). However, Vannas and co-workers' series included cases with any degree of corneal vascularization, and their recipients may not have had as high a risk of graft rejection as those included in our series and those reported by Batchelor and co-workers. 4 DISCUSSION
The role of humoral immunity in cor neal graft rejection is unknown. Previous work suggests that corneal graft rejection is mediated by cellular immune re sponses. 1 9 However, the HLA serologic reactions can be used as indicators of recipient preimmunization, because the major gene loci controlling cell-mediated responses also map within the HLA com plex. Thus far, about ten loci, all control ling some aspect of immune function,
have been assigned to the HLA region of chromosome 6 in man (Fig. 8). 20 The D locus (or subregion) controls mixed lym phocyte responsiveness, and is an in vitro indicator of cellular immune responsive ness. The entire HLA complex occupies a small segment of the chromosomal length, allowing identification of a given haplotype (the gene complex on one chro mosome of the pair) by the serologic specificities it contains. Our results indicated a lower incidence of corneal graft failure from rejection when cross-match testing was used for donor selection than did other series in volving similar high-risk cases with dense corneal vascularization. 1,4 We believe this difference is due mainly to our use of preoperative donor-recipient cross-match testing as a means for donor selection. Thus, we are able to avoid giving a donor
TABLE 6 DONOR-RECIPIENT H L A ANTIGEN MATCH AND GRAFT OUTCOME
Number of HLA Antigens Matched
Number of Patients
Total Number of Immune Events (Reaction or Rejection)
Allograft Reaction (Graft Clear)
0 1 2 3 4 Total
51 35 11 5 1 103
14(27%) 9(26%) 1(9%) 2(40%) 0 26(25%)
7 3 1 0 0 11
Percent Reversed 50 33 100 0 42
602
AMERICAN JOURNAL OF OPHTHALMOLOGY
NOVEMBER, 1978 HLA Region
I STARK ] BATCHELOU
-C4,Ch,Rg-
PGM,
GLO
-i-£i—\-i
L
3
D
Bf
B C
I
I
1I
MLR" A
M-
Map distances in centimorgans
cornea to a recipient who is presensitized to that cornea's antigens. 21 Although we found no consistent corre lation between the number of matching HLA antigens shared between the donor and recipient and graft outcome, the role of antigen matching deserves further evaluation. However, if further attempts at donor-recipient HLA antigen matching are made, they must be combined with preoperative cross-match testing, because even if all but one of the known donorrecipient antigens are favorably matched,
I
1-
Fig. 8 (Stark and associates). Map of HLA region of chromosome 6.
Fig. 6 (Stark and associates). The incidence of rejection after keratoplasty in recipients who have shared zero, one, two, three or four HLA antigens with their donor. Comparison between this study and that of Batchelor and co-workers. 4
LLi I
C2
the recipient may still be preimmunized to the one unmatched donor antigen or to unrecognized donor HLA antigens. Al though perfect donor-recipient HLA matching for corneal transplantation may be beneficial, it is not currently practical because of the complexity of the HLA system (Table 7). 22 Our study indicates that the use of cross-matching to avoid donors with transplantation antigens to which the recipient is immunized is feasi ble and seems beneficial. We do not have positive donorrecipient cross-match controls, and we therefore consider this to be a prelimiary report. A well-controlled clinical trial is necessary to evaluate fully the results of preoperative donor-recipient cross-match testing for donor selection in keratoplas ty, and such a study is currently under way. If controlled studies continue to show a significant beneficial effect from using cross-match testing in high-risk cases for keratoplasty, then this procedure could easily be coordinated with kidney transplant projects without the need for additional expensive equipment and peronnel.
N" 13
SUMMARY 3-4
Fig. 7 (Stark and associates). The incidence of rejection after keratoplasty and the number of HLA antigens shared by recipient and donor. Comparison between this study and that of Vannas and coworkers. 5
In 84 cases of keratoplasty, all patients had densely vascularized corneas and a history of prior exposure to HLA antigens by pregnancy, blood transfusion, or pre vious transplantation. Donors were se lected on the basis of a negative cross-
VOL. 86, NO. 5
HLA ANTIGENS AND KERATOPLASTY
603
TABLE 7 L I S T I N G O F R E C O G N I Z E D HLA S P E C I F I C I T I E S ( 1 9 7 8 )
A Locus
B Locus
C Locus
D Locus
HLA—Al HLA—A2 HLA—A3 HLA—A9 HLA—A10 HLA—All HLA—A28 HLA—A29 HLA—AW23 HLA—AW24 HLA—AW25 HLA—AW26 HLA—AW30 HLA—AW31 HLA—AW32 HLA—AW33 HLA—AW34 HLA—AW36 HLA—AW43 Blank
HLA—B5 HLA—B7 HLA—B8 HLA—B12 HLA—B13 HLA—B14 HLA—B15 HLA—B17 HLA—B18 HLA—B27 HLA—B37 HLA—BW16 HLA—BW21 HLA—BW22 HLA—BW35 HLA—BW38 HLA—BW39 HLA—BW40 HLA—BW41 HLA—BW42 HLA—BW44 HLA—BW45 HLA—BW46 HLA—BW47 HLA—BW48 HLA—BW49 HLA—BW50 HLA—BW51 HLA—BW52 HLA—BW53 HLA—BW54 Blank
HLA- -CW1 HLA- -CW2 HLA--CW3 HLA--CW4 HLA--CW5 HLA- -CW6 Blank
HLA—DW1 HLA—DW2 HLA—DW3 HLA—DW4 HLA—DW5 HLA—DW6 HLA—DW7 HLA—DW8 HLA—DW9 HLA—DW10 HLA—DW11 Blank
match to avoid donor antigens to which the recipient was preimmunized. T h e overall rate of graft failure from rejection was 15%. A retrospective analysis of donor-recipient HLA matching in 103 high-risk cases showed no consistent cor relation between the n u m b e r of antigens shared and graft outcome. These findings indicated negative cross-match to be im portant in donor selection for keratoplasty in high-risk cases.
REFERENCES 1. Khodadoust, A. A.: The allograft rejection reac tion. T h e leading cause of late failure of clinical corneal grafts. In Corneal Graft Failure, Ciba Foun dation Symposium. Amsterdam, Elsevier, 1973, p .
2. Stark, W. J., Paton, D., Maumenee, A. E., and Michelson, P. E.: T h e results of 102 penetrating keratoplasties using 10-0 monofilament nylon su ture. Ophthalmic Surg. 3:11, 1972. 3. Chandler, J. W., and Kaufman, H. E.: Graft reactions after keratoplasty for keratoconus. Am. J. Ophthalmol. 77:543, 1974. 4. Batchelor, J. R., Casey, T. A., Gibbs, D. C , Lloyd, D. F., Werb, A., Prasad, S. S., and James, A.: HLA matching and corneal grafting. Lancet 1:551, 1976. 5. Vannas, S., Karjalainen, K., Ruusuvaara, P., and Tiilikainen, A.: HLA-compatible donor cornea for prevention of allograft reaction. Albrecht von Graefes Arch. Klin. Ophthalmol. 198:217, 1976. 6. Allansmith, M. R., Fine, M., and Payne, R.: Histocompatibility typing and corneal transplanta tion. Trans. Am. Acad. Ophthalmol. Otolaryngol. 78:445, 1974. 7. Opelz, G., Mickey, M. R., and Terasaki, P. I.: Calculations on long-term graft survival in human kidney transplantation. Transplant. Proc. 9:27, 1977. 8. Mickey, R. M., Kreisler, M., Albert, E. D.,
604
AMERICAN JOURNAL OF OPHTHALMOLOGY
Tanaka, N., and Terasaki, P. I.: Analysis of HL-A incompatibility in human renal transplants. Tissue Antigens 2:57, 1971. 9. Terasaki, P. I., Kreisler, M., and Mickey, R. M.: Presensitization and Kidney transplant failures. Postgrad. Med. J. 47:89, 1971. 10. Opelz, G., Terasaki, P. I.: Histocompatibility matching utilizing responsiveness as a new dimen sion. Transplant. Proc. 4:433, 1971. 11. Opelz, G., Sengar, D. P. S., Mickey, R. M., and Terasaki, P. I.: Effect of blood transfusions on subsequent kidney transplants. Transplant. Proc. 5:253, 1973. 12. Opelz, G., Mickey, R. M., and Terasaki, P. I.: Identification of unresponsive kidney-transplant re cipients. Lancet 1:868, 1972. 13. Terasaki, P. I., Thrasher, D. L., Hauber, T. H.: Serotyping for homotransplantation. 13. Im mediate kidney transplant rejection and associated preformed antibodies. In Dausset, J., Hamburger, J., and Mathe, G. (eds.): Advance in Transplantation. Baltimore, Williams and Wilkins, 1968, p . 225. 14. Stark, W. J., Opelz, G., Newsome, D., Brown, R., Yankee, R., and Terasaki, P. I.: Sensitization to human lymphocyte antigens by corneal transplanta tion. Invest. Ophthalmol. 12:639, 1973. 15. Dausset, J.: Leuco-agglutinins. 4. Leuco-
NOVEMBER, 1978
agglutinins and blood transfusion. Vox Sang. 4:190, 1954. 16. Kissmeyer-Nielsen, F., Olsen, S., PosborgPetersen, V., and Fjeldborg, O.: Hyperacute rejec tion of kidney allografts associated with pre-existing humoral antibodies against donor cells. Lancet 2: 662, 1966. 17. Amos, D. B., Bashir, H., Boyle, W., MacQueen, M., and Tiilikainen, A.: A simple microcytotoxicity test. Transplantation 7:220, 1970. 18. McCarey, B. E., and Kaufman, H. E.: Im proved corneal storage. Invest'. Ophthalmol. 13:165, 1974. 19. Polack, F. M., and Gonzales, C. E.: The re sponse of the lymphoid tissue to corneal heterografts. Arch. Ophthalmol. 80:321, 1968. 20. Bias, W. B., and Hsu, S. H.: HLA, the major histocompatibility system in man. An overview. Acta Anthropogenetica. 1:15, 1977. 21. Newsome, D. A., Takasugi, M., Kenyon, K. R., Stark, W. J., and Opelz, G.: Human corneal cells in vitro: morphology and histocompatibility (HL-A) antigens of pure cell populations. Invest. Ophthalmol. 13:23, 1974. 22. World Health Organization Committee: No menclature for factors of the HLA system 1977. Tissue Antigens 11:81, 1978.