Association of Kidney Transplant Failure and Antibodies Against MICA

Association of Kidney Transplant Failure and Antibodies Against MICA

Association of Kidney Transplant Failure and Antibodies Against MICA Kazuo Mizutani, Paul Terasaki, Jean Denis Bignon, Maryvonne Hourmant, Anne Cesbro...

166KB Sizes 2 Downloads 42 Views

Association of Kidney Transplant Failure and Antibodies Against MICA Kazuo Mizutani, Paul Terasaki, Jean Denis Bignon, Maryvonne Hourmant, Anne Cesbron-Gautier, Remi N.J. Shih, Rui Pei, Jarhow Lee, and Miyuki Ozawa ABSTRACT: Despite the progress in renal transplantation, acute rejection and graft failure still occur and chronic rejection continues to be the main problem in long-term allograft survival. Although kidney transplant rejection has been linked to anti-HLA antibodies, not all patients with failed kidney transplants have anti-HLA antibodies, indicating that other loci may be involved. Sera of 63 patients who experienced kidney rejection were compared against sera of 82 patients with functioning transplants. Sera were examined for IgG and IgM antiHLA Class I and II antibodies. They were also tested by cytotoxicity against panels of 26 endothelial cell lines, 8 MHC class I chain-related gene A (MICA) recombinant cell lines, and 28 B lymphoblast cell lines. Among patients whose transplants failed, 65% had anti-HLA antibodies compared with 45% of those with functioning kidneys (p ⬍ 0.05). Similarly, among those whose transplants failed, 41% had anti– endothelial cell antibodies in contrast ABBREVIATIONS HLA human leukocyte antigen MICA MHC class I chain-related gene A

INTRODUCTION We previously summarized the evidence indicating that anti– human leukocyte antigen (HLA) antibodies are the likely cause of chronic immunologic rejection [1, 2]. Although rejection can often be attributed to HLA, research has also shown that graft failure can occur in HLA identical transplants. Previous research has estab-

From the Terasaki Foundation Laboratory, Los Angeles, CA; Laboratoire d’Histocompatibilité. Etablissement Français du Sang, Nante, France; Service de Nephrologie et d’Immunologie Clinique, Nantes, France; and One Lambda, Inc., Canoga Park, CA Address reprint requests to: Paul I. Terasaki, Terasaki Foundation Laboratory, 11570 W. Olympic Blvd, Los Angeles, CA 90064; Tel: (310) 479-6101; Fax: (310) 445-3381;E-mail: [email protected] Received March 24, 2006; revised June 15, 2006; accepted June 21, 2006. Human Immunology 67, 683– 691 (2006) © American Society for Histocompatibility and Immunogenetics, 2006 Published by Elsevier Inc.

to 22% in functioning patients (p ⬍ 0.05). Among patients whose grafts failed, 52% had anti-MICA antibodies versus 21% of those with functioning grafts (p ⬍ 0.001). Eleven patients with failed grafts and 32 with functioning grafts were negative for all of the above. However, 6 of the former and 7 of the latter showed positive cytotoxicity against B lymphoblasts (p ⬍ 0.05). Taking all antibodies together, 92% of patients with graft failure had antibodies as opposed to 70% of patients with functioning grafts (p ⬍ 0.001). We postulate that antibodies against HLA, MICA, endothelial cells, and B lymphoblasts could be independently involved in the slow process of chronic graft failure. Human Immunology 67, 683– 691 (2006). © American Society for Histocompatibility and Immunogenetics, 2006. Published by Elsevier Inc. KEYWORDS: Kidney-transplant rejection; MICA; endothelial cell; B lymphoblast

PBS

phosphate-buffered saline

lished graft-versus-host reactions and graft failure among HLA-identical sibling donor kidney transplants as well as a 38% 10-year kidney graft failure in HLA-identical sibling transplants [3]. Further evidence that HLA alone is not responsible for graft failure is provided by a comparison of kidney transplant survival in HLA-identical sibling donors and unrelated living donors in which calculations suggested that as many as 38% kidney graft failures in deceased donor transplants are the result of other histocompatibility loci [3]. Finally, a recent study by Opelz indicated that anti–non-HLA antibodies contribute substantially to long-term kidney-transplant failure in identical sibling donor transplants, suggesting the need for characterizing non-HLA antigens that can lead to graft failure [4]. 0198-8859/06/$–see front matter doi:10.1016/j.humimm.2006.06.002

684

To test for antibodies against non-HLA, we used recombinant MICA cell lines and a panel of endothelial cells, as endothelial cells are likely targets of alloresponse, and antibodies against endothelial cells have been found in renal transplant recipients before and after transplantation [5–10]. Antibodies to MHC class I chain-related gene A (MICA) antigens [11] are interesting given that the MICA locus is highly polymorphic and closely linked to the HLA-B locus [11]. Zwirner et al. identified MICA antigens which are expressed by endothelial cells and monocytes but not by lymphocytes [12]. In other research, MICA antibody was found in patients whose kidney transplants had been rejected [13, 14]. Our aim here was to conduct a systematic study of antibodies ageinst HLA, MICA, and endothelial cells in patients who had failed grafts and in those with functioning grafts. Finally, for those patients with none of the above antibodies, antibodies reactive against B-lymphoblast cells were sought by cytotoxicity. PATIENTS AND METHODS Patients We retrospectively examined the sera of 173 patients who had undergone either a first-kidney or combined first-kidney and pancreas transplantation performed in Nantes, France, between January 1998 and March 2003. The sera were taken until December 2003 as part of an annual check-up to test for immunologic response, which is the accepted procedure at this institution to monitor for rejection. One serum sample was available for each patient was available in this study and tested in December 2003. Patients were followed in Nantes until May 2005. A total of 28 patients who had graft failure before serum collection were used as references for the study (After-Failure group). The average number of days between the sample collection date and the failure date in the After-Failure group was 173 days (5–1087 days). Unfortunately, the exact cause of each failure could not be accurately determined. We assumed that not all failure were the result of chronic rejection. Because the sera had been collected over a long period of time, biopsy data, especially those from C4d staining, were not uniformly available. After testing was completed in December 2003, a total of 146 patients (excluding the After-Failure group) were followed; of those, 63 patients had experienced kidney rejection or had died before the end of the study (the Before-Failure group). The average number of days between the sample collection date and failure date in the Before-Failure group was 164 days (0 –1193 days). A total of 82 patients with functioning

K. Mizutani et al.

first transplants from the same center were used as control subjects. Most of the study and control patients had initial triple therapy (cyclosporine [CsA], steroids, and azathioprine or mycophenolate mofetil [MMF]), with an anti-IL2 receptor monoclonal antibody, Basiliximab, induction. Some recipients were also participating in experimental protocols and were given other monoclonal antibodies or related molecules (antiLFA1, anti-CD4 or CTLA4-Ig), Sirolimus, Everolimus, or FTY720 instead of the induction therapy described above. All patients with acute rejection episodes were treated with steroid boluses as first line therapy and with anti-lymphocyte globulins in case of cortico-resistance. The immunosuppressive treatment was modified according to clinical events [15]. “Return to dialysis” was considered to indicate graft loss. Detection of Anti-HLA Antibodies Testing for anti-HLA antibodies was performed by flow cytometry or Luminex methods, using FlowPRA®, or LABScreen® assays, respectively, according to the manufacturer’s specifications (One Lambda, Inc., Canoga Park, CA). IgG anti-HLA antibody screening was done using the antihuman reagent provided by the kits. IgM anti-HLA antibody screening was performed using RPhycoerythrin-conjugated AffiniPure F(ab=)2 Fragment Donkey Anti-Human IgM, Fc5␮ Fragment Specific (Jackson ImmunoResearch Laboratories, West Grove, PA). Antibodies to HLA DP (0101, 0201, 0301, 0401, 0501, and 1101) were tested with FlowPRA® single antigen HLA-DP beads. Antibodies were detected by the fluorescent signal which were measured on a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) or LABScan 100™(One Lambda Inc, Canoga Park, CA). The bead was considered positive when it exceeded the cut-off value set by the manufacturer. We confirmed negative levels of these tests with 20 normal healthy men. Cytotoxicity Tests With Recombinant MICA Cell Lines MICA-expressing cell lines were produced from the m-HMY2.CIR cell line, which expressed no HLA Class I and II antigens. We electrotransformed these host cells with vector plasmid, MICA cDNA (*001,*002,*004,*007,*008,*012,*017, and *018). cDNA’s of MICA were modified with the coding sequences (exon 2 to 4), signal peptides (exon 1), transmembrane regions (exon 5), and cytosolic tails (exon 6 and 7) with G418 resistance. All cell lines were cultured in RPMI Medium 1640 with 15% fetal bovine serum, 1% of L-glutamine penicillin/streptomycin solution, membrane filtered, and 600 ␮g/ml G418 sulfate, and incubated at 37°C in a 5% CO2 humidified

Kidney Transplant Failure and Antibodies

environment. MICA expression of each cell line was confirmed with MICA monoclonal antibody with a flow cytometer. The complement-dependent cytotoxicity method was used to detect specific antibody against MICA. The target cells were from 8 MICA antigen expressing cell lines: MICA *001, *002, *004, *007, *008, *012, *017, and *018. The m-HMY2.CIR cell lines were used as controls. Cytotoxicity tests were performed using 0.001 ml of sera in Terasaki trays, and 0.001 ml of a 3 ⫻ 106 cell/ml suspension. Class II complement (rabbit complement non-toxic to B lymphocytes, One Lambda, Inc.) were used, and test results were read by fluorescence. The cytotoxicity score was the estimate of percentage of cell death beyond that of the negative control as follows: score 1 (0 –10% dead cells), 2 (11– 20%), 4 (21–50%), 6 (51– 80%), and 8 (81–100%). We considered 6 and 8 as positive. An alloserum from a highly sensitized patient, previously shown to have the same levels of reactions as a monoclonal MICA antibody, was used as a positive control. Negative levels of these tests were confirmed on sera from 20 normal healthy men. T- and B-Cell Cytotoxicity Test Sera that were negative against HLA, endothelial, and MICA were tested for other antibodies by cytotoxicity against T and B cells. Lambda Cell TrayTM assays (One Lambda, Inc.) were used for this test, based upon the manufacturer’s instructions. Cytotoxicity tests with endothelial cell lines A total of 26 endothelial cell lines, which were kindly donated to us by Dr. Suciu-Foca, were cultured in EGM BulletKit® (Cambrex Corporation, East Rutherford, NJ), and incubated at 37°C in a 5% CO2 humidified environment. Endothelial cell lines were used between the second and fifth generations. To activate the antigen expression, endothelial cells were cultured with 10 U/ml recombinant human IFN-␥ (R&D Systems, Inc., Minneapolis, MN) for 48 –72 hours before testing. Endothelial cells were suspended in trypsinethylenediaminetetraacetic acid (EDTA; Invitrogen, Carlsbad, CA), washed with culture medium, and harvested. The complement-dependent cytotoxicity method was used to detect the specific antibody against endothelial cells. Cytotoxicity tests were performed using 0.001 ml of sera in Terasaki trays, and 0.001 ml of a 3 ⫻ 106 cell/ml suspension of 26 lines of endothelial cells. Class I rabbit complement (One Lambda, Inc.) was added, and test results were read by fluorescence. To rule out cytotoxicity by anti-HLA antibodies, the cells were typed for HLA Class I and Class II (HLA DR/DQ) with HLA-ABC First Tissue Typing Tray and

685

HLA Class II Tissue Typing Tray (One Lambda), and compared with anti-HLA antibody specificity. We confirmed negative levels of these tests on sera from 20 normal healthy men. Test results are summarized in Figure 1. Flow Cytometry With Endothelial Cell Lines To confirm the activation of endothelial cell lines, the degree of HLA antigen expression was measured by a flow cytometer. Endothelial cells were cultured with EGM BulletKit® media with either 10 units/ml or 100 units/ml of recombinant human IFN-␥, and examined for HLA expression. After a 48-hour incubation with IFN-␥, endothelial cell lines (1 ⫻ 105 cells/ sample) were harvested with trypsin-EDTA, washed with culture medium, and incubated at room temperature for 30 minutes with fluorescein isothiocyanate (FITC)-labeled anti–HLA class II antibody (FH0002, One Lambda, Inc.). The flow cytometry readings are given in Fig. 2. Statistical Analysis Significance between frequencies was determined by Chisquare analysis or Fisher’s exact test. Values of p ⬍ 0.05 were considered statistically significant. RESULTS Test results for antibodies are presented in Figure 1 and Table 1 and grouped into three categories: 1) serum samples from patients with functioning grafts, including those still functioning at the time of this writing (“Functioning,” n ⫽ 82); 2) serum samples from patients taken before failure of their grafts (“Before Failure,” n ⫽ 63); and 3) serum samples taken from patients after graft failure (“After Failure,” n ⫽ 28). Anti-HLA Antibody Among the Before-Failure group, 40% (n ⫽ 25) had anti-HLA class I IgG antibody and 27% (n ⫽ 17) IgM antibody (Table 1 and Figure 1). Of the Functioning group, 7% (n ⫽ 6) had anti-HLA class I IgG antibody and 28% (n ⫽ 23) IgM antibody. Of the Before-Failure group, 43% (n ⫽ 27) had anti-HLA DR/DQ IgG antibody and 33% (n ⫽ 21) had IgM antibody. These percentages were notably different for the Functioning group, in which 6% (n ⫽ 5) were positive for anti-HLA DR/DQ IgG antibody and 24% (n ⫽ 20) IgM antibody. In total, as many as 65% (n ⫽ 41) of patients in the Before-Failure group had both anti-HLA class I and/or class II antibodies (HLA A, B, DR, DQ: IgG and IgM, DP: IgG), compared with 45% (n ⫽ 37) of those with functioning grafts. (p ⬍ 0.05) (Table 3). However, IgG antibody was more clearly associated with failure than was anti-HLA class I IgM antibody.

686

K. Mizutani et al.

Failure Group HLA HLA HLA HLA DP class I DP/DQ DP/DQ IgG IgM IgG IgM

Endo Cat 1

Functioning Group

cyto

Cat 2

MICA Cat 3

HLA class I IgG

B cell cyto

HLA HLA HLA HLA DP class I DP/DQ DP/DQ IgG IgM IgM IgG

Endo Cat 1

cyto

Cat 2

MICA Cat 3

B cell cyto

Before-failure samples

HLA class I IgG

After-failure samples

63

28

82 *Endothelial cells cytotoxicity Cat 1: Endothelial cells reacted only to anti-HLA antibodies present in patient’s serum. Cat 2: Endothelial cells reacted to anti-HLA antibodies of the patients as well as other HLA types. Cat 3: Endothelial cells reactions could not be accounted for by the patient’s anti-HLA antibodies. B cell cyto: test with B lymphoblast Antibody positive

FIGURE 1 Anti– human leukocyte antigen (HLA), endothelial cell, MHC class I chain-related gene A (MICA), and B lymphoblast antibodies in patients who had transplants that failed (left side) and in patients who still had functioning grafts (right side). Each row shows results of one serum sample. Gray area indicates a positive reaction. Endothelial cell cytotoxicity results were divided into three categories: 1) Endothelial cells that reacted only to anti-HLA antibodies present in the patient’s serum; 2) endothelial cells that reacted to anti-HLA antibodies of the patients as well as other HLA types; 3) endothelial cell reactions that could not be accounted for by the patient’s anti-HLA antibodies.

There was no significant difference between the Before-Failure group and the After-Failure group in terms of the prevalence of HLA class I and class II antibodies. A total of 52% (n ⫽ 33) were positive for HLA class I and 51% (n ⫽ 32) were positive for HLA class II (DR, DQ, and DP) (Figure 1). In the After-Failure group, 50% (n ⫽ 14) were positive for HLA class I, whereas 71% (n ⫽ 20) were positive for HLA class II. Anti-MICA antibodies Anti-MICA antibodies were more frequent in the sera of patients who had failed rather than functioning grafts. In the Before-Failure group, 52% (n ⫽ 33) had anti-MICA antibody, whereas 21% (n ⫽ 16) of the Functioning group were positive for anti-MICA antibody (p ⬍ 0.001) (Table 1, Figure 1). Endothelial Cell Cytotoxicity and Anti-HLA Antibody Cultured resting endothelial cells constitutively expressed HLA class I but not HLA class II antigens. To

induce expression of HLA class II by IFN-␥, 10 units/ml of IFN-␥ was used, as this was found to be as effective as 100 units/ml (Figure 2). These stimulated cells were used to perform HLA typing and antibody tests. Each endothelial cell was typed for HLA Class I and Class II (HLA DR/DQ). To exclude the reactions resulting from anti-HLA antibodies, we divided the results into three categories: 1) endothelial cells that reacted only to anti-HLA antibodies present in the patient sera; 2) endothelial cells that reacted to antiHLA antibodies of the patients as well as other HLA types; and 3) endothelial cell reactions that could not be accounted for by the patient’s anti-HLA antibodies. Among the patients in the Before-Failure group, 41% (n ⫽ 26) had anti– endothelial cell–specific antibodies compared with 22% (n ⫽ 18) of the patients with functioning grafts (p ⬍ 0.05) (Table 1 and Figure 1). The reactions of anti– endothelial cell–specific antibody (category 3, Fig. 1) with anti-MICA antibody did not have an association (data not shown). This lack of association was true for all patients regardless of

Kidney Transplant Failure and Antibodies

687

E19

E20

Control (FITC-mIgG) IFN gamma (10 unit/ml) IFN gamma (100 unit/ml) FIGURE 2 Example of anti– human leukocyte antigen (HLA) class II (DR/DQ) antibody reaction in endothelial cells, stimulated by interferon-␥ (IFN-␥). Endothelial cells incubated with media alone is compared with those incubated with 10 units/ml and 100 units/ml of recombinant human IFN-␥ for 48 hours. Control was FITC-mouse IgG (FITC-mIgG).

whether their grafts were functional. However, as these endothelial cell lines were not specifically tested for MICA antigens, we cannot rule out a relationship between anti– endothelial antibodies and anti–MICA antibodies. Anti-B Lymphoblast Line Antibody The sera of 43 patients were negative for all the previously tested antigens. These 43 included 11 whose

grafts had failed and 32 whose grafts were still functional. These sera were then tested for cytotoxicity against a panel of 28 B lymphoblasts composed mostly of CLL cells. As shown in Figure 1 and Table 2, six of 11 patients whose grafts failed had antibodies against B lymphoblasts, whereas seven of 32 patients with functioning grafts were positive for these same antibodies (p ⬍ 0.05). These antibodies were clearly not anti-HLA, as these sera did not react with HLA antigen– coated

TABLE 1 Frequency of antibodies by type Failure Before-failure samples

After-failure samples

n ⫽ 63 (100%)

n ⫽ 28 (100%)

Type of antibodies

Functioning n ⫽ 82 (100%) Positive

Positive

p Value

Positive

p Value

HLA Class I IgG Class I IgM DR/DQ IgG DR/DQ IgM DP IgG Endothelial cell cytotoxicitya MICA cytotoxicity

6 (7%) 23 (28%) 5 (6%) 20 (24%) 0 (0%) 18 (22%) 16 (21%)

25 (40%) 17 (27%) 27 (43%) 21 (33%) 6 (10%) 26 (41%) 33 (52%)

⬍0.0001 NS ⬍0.0001 NS ⬍0.01 ⬍0.05 ⬍0.001

10 (36%) 8 (29%) 13 (46%) 16 (57%) 0 (0%) 10 (36%) 9 (32%)

⬍0.001 NS ⬍0.0001 ⬍0.01 NS NS NS

Abbreviations as in text. a Endothelial cell cytotoxicity (included category 1,2, and 3).

688

beads. Furthermore, the antibodies did not react against a panel of 28 B lymphoblasts, showing that the non-HLA antigen against which they reacted was specific to B lymphoblasts. Overall Antibody Of the Before-Failure group, 92% (n ⫽ 58) had at least one antibody against HLA IgG, HLA IgM, endothelial cells, MICA, or B lymphoblasts, as opposed to 70% (n ⫽ 57) of the patients with functioning grafts (p ⬍ 0.001) (Table 3 and Figure 1). There were 63 first, 16 second, and 5 third transplants among patients whose grafts were rejected. Within this group, patients who had rejected first transplants had lower positive percentages of each antibody. Of the patients with rejected first transplants, 92% (n ⫽ 58) had at least one antibody against HLA IgG, HLA IgM, endothelial cell, MICA, or B lymphoblast. This percentage rose to 94% (n ⫽ 15) in the rejected second transplants and 100% (n ⫽ 5) in the rejected third transplants. The co-appearance of the antibodies is shown in Table 4. The co-appearances of IgG anti-HLA class I/II and anti-MICA antibodies were more strongly associated with failure than each antibody alone. DISCUSSION In a recent serial serum study of patients followed over a 10-year period, we found a good association between graft rejection and both anti-HLA and anti-MICA antibodies [16]. These antibodies were present in a significantly higher proportion of patients who rejected their grafts than in control patients with functioning grafts for a similar period. The current study was undertaken to determine whether a similar association would be found in another set of patients performed at another center. In the present study, only one serum sample taken at a single time was examined from a set of patients with functioning grafts and a set whose grafts had failed. This new study has an advantage over the previous study [16] in that we have a larger number of patients overall in the two comparison groups, thus giving the current study more power to demonstrate an association between graft rejection and both anti-HLA and anti-MICA antibodies. The frequency of IgG antiHLA antibodies in the Before-Failure group (Class I 40%; Class II 43%) was similar to that of the previous serial serum rejected group (Class I, 26%; Class II, 38%) [16]. Among the functioning patients, the frequencies in the current study (Class I, 7%; Class II, 6%) were similar to those of the previous serial serum study (Class I, 15%; Class II, 12%). Also, as in the prior studies [16], we were surprised to see IgM anti-

K. Mizutani et al.

bodies in the late period. There were roughly comparable frequencies of anti-HLA antibodies in the two groups. As is apparent in Figure 1, the isolated appearance of IgM antibody was more frequent among patients with functioning grafts, suggesting that in the failed patients whose grafts failed, IgG antibody had replaced IgM antibody in patients who previously produced IgM antibodies to HLA. Another indication of this subclass-switch process is that IgM antibodies were not significantly associated with failure for anti-HLA class I antibodies, and only marginally in combined antiHLA class I and II antibodies (Table 1). In contrast, the association with failure was stronger for IgG antibodies, suggesting that the production of IgM antibodies may have been the earlier response to the transplants. Anti-MICA antibody frequencies in sera that did not have IgG/M anti-HLA antibodies were similar in our two studies for functioning patients (20% in the current study vs. 15% in the previous study) [16]. However, for patients whose grafts were rejected, this frequency was 36% in the current study compared with 80% in the previous study. This significantly higher percentage in the previous study likely results from the sequential sampling of the earlier study as opposed to single serum sampling of the current study. MICA is determined by a genetic locus closely linked to the HLA-B locus [17], which consists of more than 55 alleles [18]. A series of publications by Zwirner et al. [11] showed that antibodies to this highly polymorphic locus have been detected in patients who had transplants. Moreover, these antibodies have been implicated in graft failure by Sumitran-Holgersson et al. [13]. In this study, we show that 52% of recipients whose kidney transplants failed had anti-MICA antibodies compared with 21% of those with functioning grafts. (p ⬍ 0.001). When the results for anti-MICA antibodies are combined with those for anti-HLA antibodies, an even higher significance was obtained (Table 4). New to the current study were tests on a panel of 26 endothelial cell lines that could be examined in relation to the anti-HLA antibodies. The results of these tests indicated that antibodies specific to anti-endothelial cells were not very frequent. One difficulty in working with endothelial cell lines is that because they have HLA antigens, antibodies that might react with HLA on the lines needed to be discounted first. Sera that contained antibodies to both the HLA and endothelial antigens were difficult to identify. Antibodies that did not contain anti-HLA antibodies and reacted with endothelial cells alone were not very frequent (category 3 in Figure 1), but were detected in 14% of patients with graft failure and 10% of patients with functioning grafts. The endothelial cell reactions noted in this

Kidney Transplant Failure and Antibodies

689

TABLE 2 Frequency of antibodies negative to HLA, endothelial cell, and MICA Failure

B cell cytotoxicity

Before-failure samples

After-failure samples

n ⫽ 11

n⫽4

Functioning n ⫽ 32 Positive

Positive

p Value

Positive

p Value

7

6

p ⬍ 0.05

1

NS

Abbreviations as in text.

study were not related to the MICA antigens, as the two reactions were not correlated. It is possible that, in the present study, the endothelial lines used did not express MICA antigens. We can also tentatively conclude that even if all antibodies reactive to endothelial cells were counted, these antibodies would not be sufficient to account for all of the rejections. This conclusion must be viewed with some caution, however, as stimulation by IFN-␥ may not have been sufficient, and the cells in culture could have lost some of their antigens. We tested the expression of HLA class I and II antigens on endothelial cells using standard serologic typing trays. HLA class II expression was examined with FACS, as shown in Figure 2. Further studies with endothelial cells would be of interest to determine whether these cells have unique antigens influencing graft rejection. Many studies have identified antibodies in the sera of patients whose transplants are functioning [1, 19, 20]. Studies of serial serum samples over as many as 10 years [16] have clearly shown that antibodies can exist in patients with functioning transplants as well as in patients with rejected grafts. In fact, patients can survive for many years while having circulating antibodies. It has been postulated that the mechanism of the

effect of antibody in the post-transplantation period is much like cholesterol, in that it causes small incremental damage that accumulates over many years [2, 21]. In this case, the correlation of antibodies with failure can be seen only as a higher incidence of antibody in patients with rejected grafts compared with those with functioning grafts. The total antibody frequency in this study for HLA, MICA, and endothelial cells was indeed higher in the group of patients with failed grafts (92%), compared with those with functioning grafts (70%) (p ⬍ 0.001). In addition, if many different types of antibodies are involved, and the effect is cumulative, it is necessary to measure each different type of antibody separately. We have attempted to do this in the current study, looking at anti-MICA antibodies and anti– endothelial cell antibodies. The correlation with failure was significant individually for all three types of antibodies (Table 1). To test the presence of anti–non-HLA, anti–B-lymphoblast–specific antibodies, we examined sera that were negative to soluble HLA antigen-coated beads by cytotoxicity against a panel of 28 B lymphoblasts. Again, as shown in Figure 1, the presence of antibodies was more frequent in the patients with failed grafts than in those with functioning transplants (p ⬍ 0.05),

TABLE 3 Frequency of antibodies among Patients who had at least one of the following antibodies Failure

HLAa HLAa, Endob, or MICAc HLAa, Endob, MICAc, or B celld

Before-failure samples

After-failure samples

n ⫽ 63 (100%)

n ⫽ 28 (100%)

Functioning n ⫽ 82 (100%) Positive

Positive

p Value

Positive

p Value

37 (45%) 50 (61%) 57 (70%)

41 (65%) 52 (83%) 58 (92%)

⬍0.05 ⬍0.01 ⬍0.001

22 (79%) 24 (86%) 25 (89%)

⬍0.01 ⬍0.05 ⬍0.05

Abbreviations as in text. a HLA anti-HLA (IgG:A,B,DR,, DQ,DP, IgM: A,B,DR,DQ) antibody. b Endo anti-endothelail cell antibody (category 1, 2, and 3). c MICA anti-MICA antibody d B cell anti-B cell antibody

690

K. Mizutani et al.

TABLE 4 Frequency of antibodies among Patients who had at least two of the following antibodies Failure

IgG IgG IgG IgG

HLA HLA HLA HLA

class class class class

I and class II Abs I and MICA Abs II and MICA Abs I, class II, and MICA Abs

Before-failure samples

After-failure samples

n ⫽ 63 (100%)

n ⫽ 28 (100%)

Functioning n ⫽ 82 (100%) Positive

Positive

3 (4%) 1 (1%) 1 (1%) 0 (0%)

16 (25%) 16 (25%) 19 (30%) 13 (21%)

p p p p

p Value

Positive

p Value

⬍ ⬍ ⬍ ⬍

5 (18%) 3 (11%) 4 (14%) 2 (7%)

NS ⫽0.05 ⬍0.05 NS

0.05 0.0001 0.0001 0.0001

Abbreviations as in text. Abs⫽antibodies; NS: p ⬎ 0.05 compared with functioning transplants. HLA class II: DR, DQ, and DP;

indicating that these antibodies were also involved in transplant rejection. Possible immunogenic antigens determined by non-HLA loci were postulated to be the cause of as many as 38% of failures in transplants from deceased donors [3]. These antibodies were also suggested as the possible cause for failure of HLA-identical sibling donor kidney transplants into highly sensitized patients [4]. In conclusion, HLA, endothelial cell, MICA, and B lymphoblast cytotoxic antibodies were present independently on a more frequent basis in patients with failed grafts than those with functioning grafts. Whether these antibodies might act independently towards rejection of a graft remains to be determined. REFERENCES 1. Terasaki PI: Humoral theory of transplantation. Am J Transplant 3:665, 2003. 2. Terasaki PI, Cai J: Humoral theory of transplantation: further evidence. Curr Opin Immunol 17:541, 2005. 3. Terasaki PI. Deduction of the fraction of immunologic and non-immunologic failure in cadaver donor transplants. Exclusively in: Cecka JM, Terasaki PI (eds): Clinical Transplants. Los Angeles, CA: University of California–Los Angeles Immunogenetics Center 449 –552, 2003. 4. Opelz G: Non-HLA-transplantation immunity revealed by lymphocytotoxic antibodies. Lancet 365:1570, 2005. 5. Perrey C, Brenchley PE, Johnson RW, Martin S: An association between antibodies specific for endothelial cells and renal transplant failure. Transpl Immunol 6:101,1998. 6. Ball B, Mousson C, Ratignier C, Guignier F, Glotz D, Rifle G: Antibodies to vascular endothelial cells in chronic rejection of renal allografts. Transplant Proc 32:353, 2000. 7. Ferry BL, Welsh KI, Dunn MJ, et al: Anti-cell surface endothelial antibodies in sera from cardiac and kidney

8.

9.

10.

11.

12.

13.

14.

15.

16.

transplant recipients: association with chronic rejection. Transpl Immunol 5:17, 1997. Cerilli J, Brasile L, Galouzis T, Lempert N, Clarke J: The vascular endothelial cell antigen system. Transplantation 39:286, 1985. Sumitran-Karuppan S, Tyden G, Reinholt F, Berg U, Moller E: Hyperacute rejections of two consecutive renal allografts and early loss of the third transplant caused by non-HLA antibodies specific for endothelial cells. Transpl Immunol 5: 321, 1997. Colovai AI, Vasilescu ER, Foca-Rodi A, et al: Acute and hyperacute humoral rejection in kidney allograft recipients treated with anti-human thymocyte antibodies. Hum Immunol 66: 501, 2005. Zwirner NW, Marcos CY, Mirbaha F, Zou Y, Stastny P: Identification of MICA as a new polymorphic alloantigen recognized by antibodies in sera of organ transplant recipients. Hum Immunol 61:917, 2000. Bahram S, Bresnahan M, Geraghty DE, Spies T: A second lineage of mammalian major histocompatibility complex class I genes. Proc Natl Acad Sci U S A 91:6259, 1994. Zwirner NW, Fernandez-Vina MA, Stastny P: MICA, a new polymorphic HLA-related antigen, is expressed mainly by keratinocytes, endothelial cells, and monocytes. Immunogenetics 47:139, 1998. Sumitran-Holgersson S, Wilczek HE, Holgersson J, Soderstrom K: Identification of the nonclassical HLA molecules, mica, as targets for humoral immunity associated with irreversible rejection of kidney allografts. Transplantation 74:268, 2002. Hankey KG, Drachenberg CB, Papadimitriou JC, et al: MIC expression in renal and pancreatic allografts. Transplantation 73:304, 2002. Maryvonne Hourmant, Jean-Denis Bignon, Jean-Paul Soulillou, et al: Frequency and clinical implications of the development of donor-specific and non donor-specific

Kidney Transplant Failure and Antibodies

HLA antibodies after kidney transplantation. J Am Soc Nephrol 16:2804, 2005. 17. Mizutani K, Terasaki P, Rosen A, et al: Serial ten-year follow-up of HLA and MICA antibody production prior to kidney graft failure. Am J Transplant 5:2265, 2005. 18. Zhang Y, Lazaro AM, Lavingia B, Stastny P: Typing for all known MICA alleles by group-specific PCR and SSOP. Hum Immunol 62:620, 2001. 19. Lee PC, Terasaki PI, Takemoto SK, Lee PH, Hung CJ, Chen YL, Tsai A, Lei HY: All chronic rejection failures of

691

kidney transplants were preceded by the development of HLA antibodies. Transplant 74:294, 2002. 20. Worthington JE, Martin S, Al-Husseini DM, Dyer PA, Johnson RW: Posttransplantation production of donor HLA-specific antibodies as a predictor of renal transplant outcome. Transplant 75:1034, 2003. 21. Terasaki PI, Ozawa M: Predictive value of HLA antibodies and serum creatinine in chronic rejection: results of a 2-year prospective trial. Transplant 80:1194, 2005.