ANTI-HLA ANTIBODIES AFTER SOLID ORGAN TRANSPLANTATION1

Abbreviations: BOS, bronchiolitis obliterans, CDC, complement-dependent cytotoxicity, PRA, panel-reactive antibody. One of the earliest controversies in the field of transplantation, after the demonstration by Brent that the homograft (allograft) reaction was immunologic, was whether grafts are rejected by humoral or cellular mechanisms (1).Currently the cellular theory of graft rejection dominates. However, despite many attempts to detect the pretransplantation state of immunization by testing T lymphocyte function, only the humoral anti-HLA panel-reactive antibody (PRA*) test (2) has been effective in measuring alloimmunization. The clinical relevance of the humoral response was clearly demonstrated in 1969 in a study that showed that HLA antibodies can instantly kill an entire kidney (3). Two excellent reviews were published recently, showing that patients with pretransplantation anti-HLA PRA have an increased risk of graft failure (4, 5). Thus it is now routine to measure anti-HLA antibodies before transplantation. However posttransplantation monitoring of anti-HLA antibodies is not performed routinely. Recent advances in the techniques used to detect anti-HLA antibodies, as well as a substantial number of studies showing an association of posttransplantation anti-HLA antibodies with adverse events, suggest that such testing may be clinically useful as well as give some insight into the mechanism of graft rejection. In this review, we will describe advances in HLA antibody testing and recent studies to determine the relevance of posttransplantation HLA antibody in solid organ transplantation. We summarize the evidence that HLA antibodies developed after transplantation are associated with acute as well as chronic rejection of allografts. On a historical note, the first posttransplantation antibody study was performed in 1968 on 24 patients who underwent transplantation in Richmond, Virginia (6). The presence of HLA antibodies detected by complement-dependent cytotoxicity (CDC) was associated with poor function of the kidney transplants. Early studies on the detection of HLA antibodies were based on the standard lymphocyte cytotoxicity test (7). Renewed interest in posttransplantation antibodies has been driven by enhancements in the CDC assay (8) and the recent development of more sensitive and specific ELISA (9, 10) and flow cytometry tests (11–14). Using these tests, the association of HLA antibodies with rejection is becoming clearer and stronger. Methods to Detect Anti-HLA Antibodies CDC. This is the original assay in use for more than 30 years to detect antibodies to HLA antigens (7). Prepared plates containing live lymphocytes from 60 to 96 different individuals can be tested at once against any unknown serum to determine its specificity and extent of reactivity (PRA level). Additionally, reactivity against only donor target cells (T or B lymphocytes) can also be tested for. The greatest drawback of the CDC test is that it also detects non-HLA autoantibodies. Attempts to circumvent this complication consisted of incubating the test at 37°C to avoid reactions of cold autoantibodies and the addition of dithiothreitol to the serum to inactivate IgM antibodies. Because some strong autoantibodies, present particularly in patients with systemic lupus erythematosus, are IgG (15), dithiothreitol does not completely solve the problem of excluding extraneous reactions. A possible reason that early attempts to monitor the development of donor-specific antibody failed may be due to the fact that the CDC assay requires complement fixation and lysis of target lymphocytes to measure antibody reactivity. Fuller et al. (8) described a phenomenon whereby antibody that did not react by CDC to donor antigens could be removed by absorption with donor platelets. The addition of an antihuman immunoglobulin reagent revealed donor reactivity and markedly increased the sensitivity of HLA antibody characterization by increasing the efficiency of complement C1q binding (16). This modification of the CDC assay is now used routinely in many HLA laboratories to detect anti-HLA antibodies. Flow cytometry. This assay uses a laser beam to detect antibodies using a similar second antihuman immunoglobulin antibody that is linked to a fluorescent marker. With this assay, antibodies can be detected with 10- to 250-fold greater sensitivity than the CDC test. T and B lymphocytes are generally used as targets to distinguish anti-HLA class I from class II reactivity. Complement fixation is unnecessary, and the isotype of the antibody can be determined by using fluorescent labeled anti-human globulin isotype-specific reagents (e.g., anti-human IgG or anti-human IgM). Finally, the intensity of the reaction, which is related to the ratio of antibody molecules per target antigen, can be determined quantitatively by median channel shift (17) or relative change in fluorescence ratio (18). One difficulty, if living cells are used as targets, is that strong IgG autoantibodies cannot be distinguished from HLA antibodies by flow cytometry. Another problem often encountered with flow cytometry is the high background of B lymphocytes due to adherence of the second antiglobulin antibody to membrane Fc receptors. Determining the specificity of HLA antibodies is generally difficult, for several mixtures of lymphocytes must be used (12). With the advent of purified class I and class II molecules coated on microparticles, flow cytometry has become more useful (13, 14). Reactions with autoantigens are excluded with these beads. Moreover, mixes of beads having HLA antigens of different specificities enable the determination of antibody specificity, as well as differentiation of antibodies to HLA class I and class II molecules in a single test (14). ELISA. Use of the ELISA method in detection of HLA antibodies was introduced in 1993 to detect IgG (9) and IgA antibodies to HLA (10). Purified HLA antigen is adhered to a microplate, and antibody reactivity is measured using a sandwich assay. The ELISA assays seem to detect HLA antibodies with greater sensitivity than CDC tests (19, 20). Although the conventional colorimetric substrates of ELISA are probably less sensitive than flow cytometry, fluorescent substrates may be as sensitive (our unpublished observations). The main advantage of ELISA is that ELISA readers are readily available and less expensive than flow cytometers. Moreover, many reactions can be examined in a 96-well plate, thus permitting determination of the specificity of HLA antibodies. As with flow cytometry, complement fixation is unnecessary, and the isotype of the antibody can be readily resolved. A current disadvantage of the ELISA test is that it does not provide crossmatch capability between recipient serum and donor cells to test for donor-specific reactivity. After transplantation, however, the donor cells are often not available for the determination of donor-specific activity. Therefore, most ELISA studies have expediently detected HLA antibodies reactive against a panel of antigens. HLA Antibodies After Kidney Transplantation. In this review, we focused our attention on studies that defined the percentage of transplant recipients who developed antibody reactive to either donor lymphocytes or to a panel of HLA antigen targets. In addition, where the study was informative, we identified the clinical relevance (acute or chronic rejection or graft survival) of the immunoglobulin isotype, antibody specificity (donor or otherwise) and, in studies that used donor lymphocytes, the donor target antigen (HLA class I versus class II) (Table 1). Donor lymphocytes were not readily available for all of the 12 renal studies shown in Table 1. Anti-HLA antibodies were detected in 12–60% of recipients after transplantation (11, 21–31). In studies utilizing a donor-specific crossmatch, antibody was detected in 20–30% of transplants where the CDC method was used (21–23, 25) compared with 12–60% with flow cytometry (11, 24, 27, 29–31). Studies utilizing cell panels or pooled antigen in CDC or ELISA assays detected HLA antibody in 25–50% (21–23) and 19–31% (26, 28, 29) of the respective sets of patients. Therefore the percentage of patients with antibody was not technique dependent. Table 1: Summary of current studies of anti-HLA antibodies after transplantation and association with rejection and graft survival in kidney transplantsIn the 7 studies that correlated acute rejection with the presence of anti-HLA antibodies, patients with antibodies had a 2- to 10-fold higher incidence of acute rejection (11, 23, 24, 26, 29–31). In transplant recipients without HLA antibodies, chronic rejection was less frequent and graft survival higher than in patients with anti-HLA antibodies (27, 28). In three major studies, despite the difference in techniques used to detect HLA antibodies, de novo antibody formation correlated with the poorest graft outcome. In the Manchester series (21), the 5-year graft survival rate of patients with antibodies detected by CDC after transplantation was 12%, compared with 76% for those without antibodies. The group from Columbia reported 4-year survival rates of 45% for those that developed CDC antibodies and 71% for patients without antibodies (22). In a recent report from Maastricht, the 5-year survival rate was 34% in patients with anti-HLA antibodies detected by flow cytometry compared with 76% in those without antibodies (31). The finding that 14–45% of patients with de novo antibody had functioning grafts after 5 years indicates that antibody does not necessarily cause immediate rejection of grafts and that many patients can have well-functioning grafts despite the presence of anti-HLA antibodies. Long-term follow-up of these patients will be needed to see whether those patients who subsequently reject their grafts come from the antibody-positive group. The implication that antibody is strongly associated with chronic rejection was seen in a recent study from Tokyo (27). Antibody reactive to donor B cells was detected in 25 of 29 patients diagnosed with chronic rejection compared with only 5 of 33 patients with stable renal function. HLA Antibodies After Heart Transplantation Unlike kidneys, there is no biochemical readout for cardiac function, so antibody monitoring may fill a need for a noninvasive alternative to protocol biopsies. Interestingly, HLA antibodies occur at about the same frequency following heart transplantation as they do after kidney transplantation (Table 2). Using CDC PRA tests, 42–88% of patients had antibody after transplantation (32–35). Using the more specific CDC crossmatch, 15–40% had donor-specific antibody (33, 35). In one study using flow cytometry, anti-donor B-cell reactivity was found at twice the rate of T-cell reactivity (50% vs. 25%;36). Whether flow cytometry or CDC assays were used, the incidence of acute rejection was significantly higher in recipients with HLA antibody (3–4X) than in recipients where no antibody was detected (33, 36). Patients with CDC-detected antibodies had a 5-year survival rate of 78% as compared with 91% in patients without antibodies (34). Similar 2-year survival results were observed when using flow cytometry (75% in patients with T-cell antibody and 90% for those without antibodies;36). Thus, heart transplant recipients have better graft survival with anti-HLA antibodies than do recipients of kidney transplants who develop anti-HLA antibodies. Table 2: Summary of current studies of anti-HLA antibodies after transplantation and association with rejection and graft survival of heart transplantsAnti-HLA Antibodies After Transplantation and Rejection and Graft Survival of Lung, Liver, and Corneal Transplants Lung. Using CDC-based assays, 11–50% of lung transplant recipients were found to have anti-HLA antibody after transplantation (Table 3) (37–39). Bronchiolitis obliterans (BOS) is the most common cause of mortality in lung transplants and is generally considered to represent chronic allograft rejection. Patients who made antibody were two (38) to four (37) times more likely to have BOS than patients who did not make antibody. When the antibody specificity could be identified, it was found to be directed to donor-mismatched class I antigens (38, 39). Interestingly, in one study anti-HLA antibody detected by ELISA was present up to 20 months before the diagnosis of BOS (39). Table 3: Summary of current studies of anti-HLA antibodies after transplantation and association with rejection and graft survival of lung and corneal transplantsLiver. In liver transplantation, in contrast to other solid organs, the majority of patients are unaffected by the presence of preformed antibodies against donor antigens (40) and, therefore, the liver seems to be partially resistant to antibody-mediated graft rejection. However, in a recent systematic flow cytometry study of antibodies made after transplantation against the specific donor, as many as 21% of 58 live donor liver transplant recipients formed posttransplantation donor-specific antibody (41). All 12 of the patients diagnosed with early acute rejection episodes developed posttransplantation antibody compared with only 17% of antibody-negative patients. Cornea. Two large studies from Canada and the United States found between 12% and 45% of corneal transplant recipients made anti-HLA antibodies detected by CDC after transplantation (Table 3) (42, 43). Both studies found that patients who made antibodies after transplantation were more likely to have graft reactions. In one study that examined the time of development of antibody relative to graft reaction, 82% of patients had an increased PRA at the time of or preceding the diagnosis of a graft reaction (42). A higher failure rate was noted with donor-specific rather than nonspecific antibodies (43). Pathologic Features of Antibody-Associated Rejection Few studies have correlated the presence of systemic anti-HLA antibodies with the histological findings in graft biopsy specimens to address the question of whether there is a distinct pathology (and therefore potentially distinct mechanism of rejection) associated with anti-HLA antibodies. Humoral rejection is usually considered to have a strong vascular component. There are two distinctive vascular lesions seen in renal allograft biopsy specimens during acute rejection: endothelialitis and necrotizing vasculitis. Endothelialitis is usually associated with a good response to increased immunosuppression, whereas necrotizing arteritis often presages graft loss. Although these two terms are descriptive only, they are taken to denote etiology, endothelialitis being equated with cellular rejection and necrotizing arteritis with antibody-mediated rejection (44). It is unlikely that morphology equates so neatly with the two arms of the immune system, however. Acute vascular rejection or necrotizing arteritis where there is fibrinoid necrosis and immunoglobulin deposition but no mononuclear cell infiltrate in renal allografts is thought to antibody mediated (44). However, this is a very infrequent histological finding (less than 1% of allograft biopsy specimens;44) and certainly is much lower than the frequency of anti-HLA antibodies associated with acute graft rejection reported in the studies cited in the present review. Immunohistologic evidence of humoral rejection has been difficult to demonstrate in allografts, perhaps because immunoglobulin is rapidly shed or capped from endothelial cells (5). Recent studies by Feucht et al. (5, 45) suggest that the presence of immunoglobulin in allograft biopsy specimens may best be revealed by assessing the capillary composition of the complement component C4d, which, in contrast to transiently bound immunoglobulin, is a stable and prominent marker of antibody reactivity against endothelial cells. The presence of this marker in renal allograft biopsy specimens was associated with cell-mediated rejection, early renal graft loss, and circulating anti-HLA antibodies (45, 46). These patients tended to have more DR mismatches (5), suggesting either a role for de novo class II antibody in rejection or a failure to detect donor-specific class II antibody in the pretransplantation crossmatch procedure. In one of the few studies to correlate biopsy findings with systemic anti-HLA antibodies, Halloran et al. (23) reported a delayed form of rejection in the first 3 months after transplantation that included endothelial injury in the microcirculation, neutrophils in the glomeruli and/or peritubular capillaries, and fibrin deposition in glomeruli or blood vessels, which was associated with the presence of antibodies to donor T cells. Two other studies, one in heart (36) and one in renal transplants (11) reported that the presence of anti-HLA antibodies was associated with acute vascular rejection, although detailed pathological features of the biopsy specimens were not given. Which Immunoglobulin Isotypes Are Detrimental? The initial method developed to detect HLA antibodies (CDC) uses a functional test, i.e., complement fixation of target cells, and it has been generally assumed that HLA antibodies that fix complement are detrimental to the graft recipient. IgM and the IgG1 and IgG3 subclasses of IgG are complement-fixing antibodies. Most studies of HLA antibodies have used methods that detect both IgG and IgM and, at best, distinguish between these isotypes. More recently, however, using flow cytometry and ELISA techniques, it has been possible to determine the different isotypes and subclasses of immunoglobulin that make up the alloantibody response. Most (47, 48) but not all (49) pretransplantation studies show that only IgG anti-HLA, and not IgM antibodies, are associated with a poor graft outcome. The posttransplantation significance of IgM HLA antibodies is unclear. In two studies of renal transplant recipients using flow cytometry, only IgG and not IgM anti-HLA antibody was associated with acute (11) and chronic (27) rejection. Another study reported that IgG antibody was associated with more severe rejection than IgM antibody (24). In heart transplant recipients, one study suggested that both IgM and IgG antibodies were associated with acute rejection (33), whereas another study suggested that only IgG and not IgM anti-HLA antibody was associated with poor graft survival (35). Acute graft rejection of liver transplants was associated with both IgG and IgM antibodies (41). There is evidence that some immunoglobulin isotypes may be protective. Renal patients with pretransplantation IgA anti-HLA antibodies detected by ELISA had improved graft outcome. The authors suggested that IgA antibodies may bind donor HLA antigens and block interaction with tissue-damaging antibodies (10, 20). Another possibility is that the isotype of the anti-HLA antibody is a reflection of the immune activation pathway involved in the recognition of the HLA antigens. IgA class switching is regulated by interleukin 5 and interleukin 10, two cytokines that are considered to be representative of a type 2 response (50). Thus, this response to donor HLA antigens may be less detrimental than a type 1 response, which is more likely to produce IgG1 antibodies. Clinical Relevance of Antibodies to Different HLA Antigens The major targets of alloantibodies that have been detected in transplantation are HLA class I and II antigens. There have been some reports of alloantibodies to non-HLA antigens with the latter mostly being associated with endothelial cells and not detectable with T or B lymphocytes (51). These non-HLA alloantibodies will not be considered further in this review because the methods used to detect the alloantibodies after transplantation in the series of studies cited here were designed to detect only anti-HLA antibodies. Both HLA class I and class II antigens were the targets of the anti-HLA antibodies associated with acute graft rejection of renal (24, 29) and heart (33, 36) transplants and patients with this antibody had poor graft outcome (21, 32). Interestingly, antibodies to B cells (presumably class II) were very strongly associated with chronic rejection in recipients of living donor kidneys (27), whereas chronic rejection in lung transplant recipients was strongly associated with antibodies to HLA class I antigens (39). These differences in antibody specificities may reflect differences in antigen expression on the different transplanted organs but will need to be confirmed in other studies. In studies that looked at total and donor-specific anti-HLA antibody, i.e., measured PRA, only donor-specific antibodies were associated with acute rejection (33) and graft loss (35) in heart transplant recipients. Similar findings were seen for corneal transplant recipients (43). In renal transplant recipients, some studies found both donor-specific and nonspecific antibodies were associated with acute and chronic rejection (26, 28), whereas others found a higher rate of acute rejection and graft loss with donor-specific antibody (22, 29). There could be a number of reasons for the conflicting results. The studies that found that both donor-specific and nonspecific antibodies were clinically relevant used an ELISA technique to detect anti-HLA antibodies, whereas the studies that found that only donor-specific antibodies were relevant used a CDC technique. Thus, differences in the sensitivities of the two assays may explain the different findings. Nondonor-specific antibodies could be associated with rejection because their presence may reflect an inadequately immunosuppressed recipient. Further studies using sensitive techniques will be required to address the issue of the clinical relevance of nondonor-specific antibodies definitively. The Humoral Hypothesis of Graft Rejection The strong association between anti-HLA antibodies and acute and chronic rejection, as well as lower graft survival, noted in the studies reviewed, indicates that HLA antibody provides a readout of patient alloreactivity. It is highly unlikely that this association has occurred by chance, for in each of the 23 studies, a positive association was found. It could be argued that this association shows that antibodies appear secondarily to other key cellular events that are the real factors responsible for the rejection. However, we consider here the alternative hypothesis that antibodies can initiate rejection by directly binding to the allotransplantation antigens (HLA). Early humoral rejection is associated with poorer graft outcome than cellular rejection (23, 52). Whether the antibody associated with acute humoral rejection involves a de novo or amnesic response still needs to be determined. The correlation of HLA antibody with chronic rejection may prove to have the most profound clinical significance because early predictors of long-term outcome are needed for drug efficacy studies (53). Renal patients with HLA antibody were five to six times more likely to develop chronic rejection (27, 28). The development of BOS in lung allograft recipients was two to four times greater in patients with antibody (37–39). Detection of antibody preceded the diagnosis of rejection by an average of 17 months (39). A possible mechanism for antibody-mediated chronic renal and cardiac allograft rejection has been developed in animal and in vitro human models. This theory suggests that arterial thickening associated with chronic rejection is a result of a repair response to damage caused by donor-specific antibody (54). Vascular lesions developed only in mouse strain combinations that produced humoral antibodies or by passive transfer of hyperimmune serum (55). Intimal thickening of carotid artery allografts occurred in mice deficient in CD8 T cells or natural killer cells but did not occur in mice deficient in antibodies or CD4 T cells (56). In vitro studies have shown that human endothelial and smooth muscle cells can be stimulated to proliferate by HLA antibody (57). One dilemma currently facing clinicians is whether graft dysfunction is due to immunologic or nonimmunologic causes. In the early period after transplantation, the decision to increase immunosuppressive dosage may further impair kidney function, while tapering the dosage may exacerbate the rejection episode. Likewise, it is difficult to assess whether impairment of long-functioning grafts is due to senescence or an immune response. Every transplantation center measures pretransplantation HLA antibody. There is no reason that this test could not be performed in the posttransplantation setting. Summary We have cited more than 23 studies showing that de novo development of anti-HLA antibodies is associated with increased acute and chronic rejection and decreased graft survival in kidney, heart, lung, liver, and corneal transplants. Antibodies to both HLA class I and class II antigens seem to be detrimental. Antibodies of the IgG isotype and possibly the IgM isotype were clinically relevant. Most studies showed that donor-specific antibodies were associated with rejection and graft loss. Therefore, HLA antibodies provide a clinical readout for patient alloreactivity that may have the ability to distinguish graft dysfunction due to immunologic and nonimmunologic causes. Antibody may act as a critical trigger for rejection of allografts and may serve as an early indicator of a slowly smoldering chronic rejection that is not manifested at a given time by biochemical measures such as serum creatinine levels. The effectiveness of various drugs on chronic rejection should be evaluable by their effects on HLA antibody production. We predict that recently developed ELISA and flow cytometry techniques using purified HLA antigen will increase the clinical relevance of posttransplantation HLA antibody monitoring by (1) allowing the detection of low levels of donor antibody; (2) easily distinguishing the isotype and target (HLA class I or class II) of the antibodies; and (3) correlating the antibody with specific graft pathology.