Abstract
ABSTRACT. Xenotransplantation is increasingly viewed as a promising way to alleviate the problem of patients who have alloreactive lymphocytotoxic antibodies and therefore tend to accumulate on the waiting list for renal transplantation. One barrier to xenotransplantation in these patients could be the hyperacute or acute vascular rejection as a result of preexisting anti-HLA antibodies that recognize swine leukocyte antigens. The cross-reactivity of sera from 98 patients with pig lymphocytes was studied by flow cytometry. After absorption of xenoreactive natural antibodies (XNA), isotype, class, and antibody specificity causing a positive cross-match (XM) were determined. For nonsensitized patients, all of the antibody binding to pig lymphocytes was due to XNA, which were removed by pig red blood cells absorption. In contrast, in sensitized patients, after removal of XNA, pig lymphocyte XM remained positive. There was no correlation between antibody binding to pig lymphocytes and Ig isotype (IgG or IgM) or HLA class-specific antibodies. For testing evidence that class II–specific antibodies were responsible for antibody binding to pig lymphocytes, HLA class I–specific antibodies were absorbed with pooled human platelets. It was confirmed that HLA class II–specific antibodies were responsible for the positive pig XM, but the strength of the positive XM was weaker than the strength caused by HLA class I–specific antibodies. Sera with multiple specificities (plurispecific sera) displayed a greater frequency of cross-reactivity with swine leukocyte antigens (P < 0.05). Seven of 11 highly immunized patients without cross-reactivity IgG with porcine lymphocytes showed positive XM before an IgM was used. The results demonstrate the cross-reactive nature of HLA antibodies and therefore point out the need to perform a prospective XM after absorption of XNA in presensitized individuals. E-mail: idv@udc.es
The shortage of human kidneys for transplantation is severe, resulting in a significant waiting time for any allograft. This shortage of human organs for allotransplantation has generated interest in using xenotransplantation as an alternative approach for patients with end-stage organ failure (1). Although for successful application of xenotransplantation there are significant hurdles, such as immunologic responses of the recipient against the graft, physiologic limitations of the graft, infection, and ethical considerations, this strategy could offer a real solution to the growing disparity between the demand for transplantable organs and the supply (2). The initial barrier to xenotransplantation, a process of hyperacute rejection, is now overcome by using organs from transgenic pigs expressing human regulators of complement (3), depleting xenoreactive natural antibodies (XNA) from the recipient’s circulation (4), or eradicating Gal(α)1-3Gal by knocking out (α)1,3 galactosyltransferase in cloned pigs (5). Acute vascular rejection and T lymphocyte–dependent xenograft rejection are now the major immunologic barriers that remain to be overcome (6).
In kidney transplantation, patients with high levels of anti-HLA antibodies are considered at high risk for an immediate antibody-mediated graft rejection and hence wait for prolonged periods for a suitable cross-match (XM)–negative allograft. This prolonged waiting may entitle them to become primary candidates for newer therapeutic approaches, such as xenogeneic organ transplantation. In this setting, the successful performance of xenotransplants assumes that anti-HLA antibodies do not cross-react with porcine major histocompatibility antigens (swine leukocyte antigens [SLA]).
SLA and HLA share considerable sequence homology, and preexisting antibodies that are cross-reactive with SLA may indicate humoral and cellular priming to porcine MHC (7,8⇓). Also, that human T cells can recognize SLA antigens via both direct and indirect antigen presentation pathways further strengthened this hypothesis (9,10⇓). Although some studies found little or no cross-reactivity between anti-HLA antibodies and SLA molecules (11,12⇓), such antibodies have been shown to bind to vascular endothelium of pig kidneys during ex vivo perfusion with human blood (13); other studies have found that these antibodies can recognize SLA molecules (14,15⇓).
To determine whether preformed anti-HLA antibodies, present in sensitized patients waiting for a renal transplant, will pose a threat to a potential xenograft, the cross-reactivity of these antibodies against SLA was tested. Anti-HLA antibodies devoid of XNA were tested against a panel of porcine lymphocytes for cross-reactivity to SLA by flow cytometry. The dependence of this cross-reactivity of antibody isotype was also tested. Furthermore, HLA class I–specific antibodies were absorbed, and sera devoid of such antibodies were tested for binding by HLA class II–specific antibodies. The HLA specificities found in sera of the renal transplant waiting list patients were also studied, and establishment of a profile of antibody reactivity with porcine cells was attempted. The results clearly showed that anti-HLA antibodies recognized SLA in addition to HLA. Thus, anti-HLA sensitization of recipients may pose problems for xenograft survival if xenografts are transplanted across sensitization barriers.
Materials and Methods
Study Population
Sera from all patients on the waiting list for cadaveric renal transplantation at Juan Canalejo Hospital (A Coruña, Spain) were selected for study. Sensitization was determined by screening sequential serum samples for lymphocytotoxic antibodies against a selected panel of 30 HLA-typed volunteer blood donors.
Of the 244 patients on the waiting list, 98 (40.16%) were included in the study and distributed in five groups according to the sensitization status. Of a total of 98 sera tested, 10 were obtained from nonsensitized patients (negative controls, 0% complement-dependent cytotoxicity [CDC] panel reactive antibodies [PRA]) and the remaining 88 were from sensitized patients with varying %CDC PRA (29 with 1 to 10%, 14 with 11 to 50%, 22 with 51 to 80%, and 23 with >80% CDC PRA).
Cell Preparation
Human and pig peripheral blood lymphocytes from seven unrelated large white pigs, for use as targets in flow cytometric XM assays, were purified by density gradient centrifugation. Peripheral blood lymphocytes prepared from anticoagulated blood (lithium heparin) were purified from threefold-diluted heparinized blood with PBS using density gradient centrifugation on Histopaque-1077 (Sigma, St. Louis, MO) at 1100 × g for 30 min at room temperature. Cells at the interface were harvested, and platelets were removed by differential centrifugation with three 10-min washes at 200 × g with PBS.
Pig red blood cells (P-RBC), used for absorption of xenoreactive natural antibodies, were prepared from packed whole blood cells subjected to three cycles of buffy coat removal followed by density gradient centrifugation as above. The P-RBC pellet was harvested, resuspended in PBS to the original diluted blood volume, and centrifuged a second time on Histopaque to deplete further the few remaining white blood cells.
Absorption of XNA
Volumes of sera (1 ml) from our patients were each divided into two equal aliquots. One aliquot from each pair remained untreated (unabsorbed control), whereas the second was absorbed twice with an equal volume of P-RBC. Serum (500 μl) was mixed with an equal volume of packed P-RBC and incubated overnight at 4°C. Tubes were then centrifuged at 14,000 × g for 5 min, and the serum was transferred to a second tube containing a new batch of packed P-RBC. These were mixed and incubated at 22°C for 2 h before centrifugation as above. Absorbed and unabsorbed sera were stored on ice until use.
Enzyme-Linked Immunosorbent Assay for Gal-(α)-1,3-Gal-(β)-1,4-GlcNAc-Specific Natural Antibodies
Levels of xenoreactive natural antibodies present before and after P-RBC absorption were assessed by enzyme-linked immunosorbent assay (ELISA) specific for Gal-α-1,3-Gal, using methods as described previously (15).
Flow Cytometry XM with Human and Pig Lymphocytes
Human and pig lymphocytes were suspended at a concentration of 5 × 106/ml in Terasaki Park medium containing 0.1% NaN3. Duplicate 10-μl volumes of absorbed and unabsorbed sera were incubated with 40 μl of lymphocyte suspension at 22°C for 30 min. Cells were washed twice in 1 ml of chilled PBS containing 0.1% BSA and 0.1% NaN3 (pH 7.2), followed by 5 μl of per-CP–conjugated mouse IgG1 heavy chain and κ light chain anti-human–CD3 (Becton Dickinson, Mountain View, CA) when human cells were used, and 50 μl of FITC-conjugated rabbit F(ab′)2 anti-human IgG (Dako, Glostrup, Denmark; diluted 1:20) or 50 μl of phycoerythrin-conjugated rabbit F(ab′)2 anti-human IgM (Dako) for both human and pig cells. Cells were mixed gently and incubated further for 30 min at 4°C, washed twice, and resuspended in 250 μl of CellFix (Becton Dickinson). Tubes were stored at 4°C in the dark until flow cytometric analysis on a FACScan (Becton Dickinson), collecting data for FITC on 256 channels using a logarithmic scale. Data were collected on 104 cells, and histograms were generated for cells in the lymphocyte gate for pig cells and lymphocyte gate and CD3-PE positive for human cells. The FACScan was calibrated daily using CaliBRITE Beads (Becton Dickinson) to ensure comparable readings over the period of the study.
Absorption of Serum to Remove Class I Antibodies
Platelets were obtained as units of platelet concentrate from the blood bank service and were processed as follows. Platelets were centrifuged at 200 × g for 40 min in 50-ml centrifuge tubes. The supernatant was removed, and platelets were centrifuged again at 2000 × g for 15 min. After the supernatant was removed, 20 ml of 0.8% ammonium chloride was added to lyse red blood cells, and the mixture was placed on a rotary mixer for 50 min. The platelets were washed twice with 1% Tris-buffered EDTA/saline at 2000 × g for 15 min and stored in this solution after the addition of 0.1% sodium azide for at least 1 mo before being used to absorb antibodies.
For the absorption of 0.25 ml of each sera, the platelets were centrifuged at 2000 × g for 20 min, the supernatant was removed, and platelets were washed twice with complement fixing buffer (Oxoid), centrifuging as before. A 50% volume of complement fixing buffer was added to packed platelets, and 1 ml of the above mixture was placed in a microcentrifuge tube and centrifuged at 10,000 × g for 5 min. Before the supernatant was removed, 0.25 ml of each serum was mixed and incubated at 22°C for 2 h. Before centrifuging at 10,000 × g for 5 min, the absorption procedure was repeated, increasing the incubation time to overnight at 22°C. Nonplatelet and platelet absorbed sera were stored at 4°C until use.
ELISA for Anti-HLA Antibodies
Presence of anti-HLA antibodies before and after platelet absorption was assessed by ELISA specific for HLA class I or class II antigens, Lambda Antigen Tray (LAT M) assay (One Lambda Inc., Canoga Park, CA).
CDC XM
The CDC XM was performed using unseparated peripheral blood lymphocytes using the micro-dye exclusion lymphocytotoxicity technique of the National Institutes of Health (16). IgG and IgM panel-reactive antibodies were identified using CDC after reduction of IgM with dithiothreitol by screening serum samples against a selected panel of 30 HLA-typed volunteer blood donors. When possible, the HLA antibody specificity was determined using 2 × 2 table analysis (Software One Lambda 4.02) using χ2 statistics to determine significant correlations between serum reactivity patterns and the presence of specific markers in the cell panel.
Statistical Analyses
Flow cytometry XM (FCXM) results expressed as the mean of the duplicate median channel fluorescence values were compared between unabsorbed and P-RBC–absorbed sera, for nonsensitized and sensitized patients against human and pig lymphocyte targets. A positive pig lymphocyte XM was defined as 2 SD above the mean of the control group after absorption of xenoreactive natural antibodies. All statistical analyses were performed by use of SPSS 10.0.6 software. P ≤ 0.05 was considered to be significant. Tests of association (χ2 and Fisher’s exact test) were used for comparison of categorical variables, whereas the Kruskal Wallis and Mann-Whitney U test were applied for comparison of continuous variables.
Results
Depletion of XNA from Sera
The removal of XNA from sera was accomplished by absorption with porcine erythrocytes, which express 107 α-Gal residues per cell (17). For confirming complete absorption of these antibodies, sera were tested for IgG and IgM Gal-α-1,3-Gal–specific antibodies by ELISA. Data were calculated with respect to a standard serum by integrating the areas under each titration curve using the software kaleidagraph (Synergy Software Technologies Inc., Canada) and expressed as relative titer with respect to the standard. The amount of XNA, after absorption with P-RBC, was reduced to 10% in almost all sera tested (data not shown).
Effect of Absorption with P-RBC on Anti–HLA-Specific Antibodies
P-RBC do not express MHC molecules, and therefore absorption of XNA will not deplete cross-reactive HLA-specific antibodies. For confirming this, FCXM was carried out between patient sera, before and after absorption with porcine erythrocytes, and a panel of human lymphocytes. Reactivity results, expressed as median channel fluorescence, for the sera showed no significant differences for any of the patient groups, before and after XNA absorption (Table 1). This result demonstrates the irrelevance of anti–Gal-α-1,3-Gal antibodies in the allogenic XM and the participation of anti-HLA antibodies, among other antibodies, in xenoreactivity.
Table 1. Human lymphocyte cross-match results
Recognition of Pig PBMC by Anti-HLA Antibodies
For determining the cross-reactivity of anti-HLA antibodies to porcine PBMC, unabsorbed and absorbed sera were tested against a panel of porcine PBMC by flow cytometry. The FCXM results from group 2 patients (PRA >0% and <10%) presented no significant differences when compared with patients from the control group. The patients of this group were excluded from later studies.
The reactivity of the samples after absorption with porcine erythrocytes decreased significantly in all patient groups; however, for groups 4 and 5 patients, this xenoreactivity was significantly greater than that obtained with sera from the control group. These differences were not statistically significant for any of the patient groups when compared with the xenoreactivity values using sera not absorbed with porcine erythrocytes (Table 2). The magnitude of xenoreactivity as a result of anti-HLA antibodies increased in parallel with level of anti-HLA sensibilization.
Table 2. Porcine lymphocyte cross-match results
Determination of Anti-HLA Class I and Class II Antibodies
Before the xenoreactivity study, the anti-HLA antibody class was determined in sera of 59 patients in groups 3, 4, and 5. These results appear in Table 3. Analysis of the data showed no statistically significant relation between cross-reactivity and the presence in the sera of anti-HLA class I antibodies, mixture of class I and class II anti-HLA antibodies, or non-HLA antibodies.
Table 3. IgG porcine lymphocyte crossmatch results and antibody class
Because most sera presented antibodies against class I and class II HLA antigens, HLA class I–specific antibodies were absorbed to differentiate whether anti-HLA class I or anti-HLA class II antibodies were responsible for the positive porcine lymphocyte crossmatch. Six sera from group 3, 11 sera from group 4, and 12 sera from group 5 had at least one positive porcine lymphocyte XM with any of the seven pigs tested and therefore were further absorbed with pooled human platelets. For confirming that absorption of class I anti-HLA antibodies had been complete, an ELISA specific for class I and class II anti-HLA antibodies was assessed. Unabsorbed sera, absorbed sera with P-RBC, and absorbed sera with P-RBC and platelets were tested against the same panel of pig lymphocytes. After the elimination of anti-HLA class I antibodies, 82% of the XM were postabsorption negative. The remaining 18% were positive. Xenoreactivity values caused by class II anti-HLA antibodies showed that the strength of the positive XM was weaker than the strength caused by HLA class I–specific antibodies (data not shown).
Determination of Antibody Specificity
The specificities found for patients on the waiting list were similar to those detected for patients studied by other groups. Of a total of 59 patients, 17 presented plurispecific anti-HLA activity, 31 patients presented monospecific anti-HLA activity, and in sera from 31 patients specificity could not be determined. As expected, most patients (13 of 17 patients) from group 5 presented plurispecific activity, and in the group of patients with intermediate, PRA we found more patients with monospecific activity (24 of 31 patients).
The variability in the number of positive porcine lymphocyte FCXM within the sera of sensitized patients is likely to reflect both SLA heterogeneity within the porcine lymphocyte panel and the different antibody specificities. There was considerable variability in the number of positive XM observed, ranging from one to seven, so a correlation with any single antibody specificity could not be established. The positive XM appearance was dependent on specificity only with plurispecific sera (P < 0.001).
Determination of HLA Isotypes
The reactivity of the samples with porcine lymphocytes was studied as a function of the Ig isotype detected for each patient. The anti-HLA antibodies of IgG isotype were responsible for most of the positive XM with porcine lymphocytes (Table 4). The statistical analysis of the data showed no statistically significant differences, and therefore it was not possible to attribute a particular pattern as having a greater or lesser tendency to react with porcine lymphocytes.
Table 4. IgG porcine lymphocyte crossmatch results and immunoglobulin isotype
In sera showing a PRA >80% and negative XM with the panel of porcine lymphocytes (n = 11), the possibility of any of them presenting autoantibodies or anti-HLA antibodies of IgM isotype was studied. For the study of autoantibodies, microlymphocytotoxicity techniques were used with and without dithiothreitol and autologous lymphocytes. For the study of IgM isotype anti-HLA reactivity, flow cytometry techniques were applied along with a panel of human lymphocytes (n = 15) using an IgM isotype PE-Ig.
The behavior of the sera with porcine and human lymphocytes was different. Seven sera from 11 patients with negative IgG xenoreactivity gave a positive result in the XM when using an IgM. Of these seven sera, four presented only IgG isotype anti-HLA antibodies. Two sera samples had anti-HLA antibodies of both isotypes (IgG and IgM), and one serum had anti-HLA antibodies of IgM isotype.
Discussion
This study confirms that cross-reactivity exists between sera, absorbed and nonabsorbed with porcine erythrocytes, of patients awaiting renal transplant and porcine lymphocyte antigens. This cross-reactivity is attributed, in the first case, to anti-HLA antibodies and, in the second case, to anti-Gal-α-1,3-Gal antibodies.
The cross-reactivity between anti-HLA antibodies and SLA had been studied before (11,13–15⇓⇓⇓). All of the authors agree on the contribution of XNA in xenoreactivity and in rejection of xenografts. However, there are differences in the interpretation of the role played by anti-HLA antibodies in xenoreactivity, and its possible impact on the prognosis of a hypothetical xenotransplant is unknown.
The results presented here confirm that XNA from sera of patients on a waiting list to receive a transplant react with the antigenic epitope Gal-α-1,3-Gal expressed in porcine lymphocytes, as reported by other authors (11), and that xenoreactivity mediated by XNA and by anti-HLA antibodies is independent of each other. Some authors suggest that anti-HLA antibodies could contribute to the rejection of a possible xenograft through a complement-mediated reaction or because anti-HLA antibodies are indicative of immunologic memory for donor HLA antigen; individuals with anti-HLA antibodies possess primed T cells directed at these same alloantigens.
The present work suggests that there is a PRA threshold (PRA >50%) above which the probability of obtaining a positive XM with porcine lymphocytes is statistically significant. These results agree with those reported by Oostingh et al. (18), who found a greater probability of obtaining positive XM with patients who present a PRA >64%.
The class of anti-HLA antibody is important in the induction of rejection in the allogenic transplant (19,20⇓). In most studies, only anecdotal cases have been cited in which HLA class II antibodies were implicated in graft rejection (21,22⇓). Recent studies on allotransplants have demonstrated that presensitization against either HLA class I or class II is of no clinical consequence, whereas sensitization against both HLA class I and class II produces increased rejection of HLA-mismatched grafts (23). The differential contribution of class I and class II anti-HLA antibodies had not been studied before in xenotransplants. Therefore, the aim here was to investigate the relationship between preformed anti-HLA class I and class II antibodies and pig cross-reactivity, but no significant differences were found in the cross-reactivity between the sera with anti-HLA class I antibodies and those that presented a mixture of class I and class II anti-HLA antibodies.
In xenotransplants, other authors have demonstrated cross-reactivity between anti-HLA class I antibodies and porcine lymphocytes (13–15⇓⇓). The possible cross-reactivity between anti-HLA class II antibodies and SLA antigens was studied and revealed that 18% of FCXM with porcine lymphocytes presented xenoreactivity after absorbing anti-HLA class I antibodies, insinuating the xenoreactivity between anti-HLA class II antibodies and porcine leukocyte antigens. The importance of this finding in an eventual xenotransplant remains to be explored because the role of anti-HLA class II antibodies in graft outcome is still debatable. The results presented here reveal that class II reactivity suggests that a porcine B cell XM would be advisable before pig organ transplant to a human, at least for the moment, until the role of class II anti-HLA antibodies is clarified.
It is possible that some anti-HLA specificity may be more reactive; however, this aspect has not been studied sufficiently. The findings of the present work suggest that anti-HLA specificity can be important. Sera from patients who are on this waiting list and presented plurispecific activity or antibodies directed to public epitopes were the most reactive with porcine lymphocytes, and this is concordant with the proposal of Oostingh et al. (18) that conserved MHC class I public epitopes are shared between HLA and SLA.
The IgM isotype anti-HLA antibodies do not play a relevant role in the rejection of allografts (24–27⇓⇓⇓). The xenoreactivity studies performed up to now do not take into account the Ig isotype. The work presented here hints at IgM isotype xenoreactivity, but its role in xenotransplantation remains to be explored.
The IgM xenoreactivity found in four patients who were hyperimmunized with anti-HLA antibodies could be due to the presence of other antibodies of IgM isotype that are potentially relevant in the xenotransplant in the same way as that occurring in the rejection of allografts induced by IgM non-HLA antibodies directed to endothelial cells (28) and therefore not detected when lymphocytes are used as target cells. Immunologic rejection of kidney allograft is usually attributed to presensitization to HLA antigens. However, data on HLA identical transplant rejections indicate that non-HLA antigens may also be involved, and it has been suggested that vascular endothelium represents the main target cell (29). This target determinant is polymorphic (30), and therefore pig lymphocytes can express antigens that are cross-reactive with these antibodies. Further studies are required to establish the nature of these antibodies, and molecular characterization is required to elucidate the nature of the target antigen to which the antibodies detected here are directed.
The present work shows that a significant proportion of patients on renal waiting lists possess antibodies that are detectable by FCXM and are cross-reactive with SLA. These findings highlight the importance of accurately determining sensitization and also drawing attention to the necessity of using PRA screening techniques that are capable of detecting antibodies that may play a role in graft rejection and escape detection in conventional XM tests. Finally, these results point out the need for performing a prospective XM in sensitized individuals even in circumstances in which the kidney is being derived from a Gal-knockout pig to avoid antibody-mediated rejection of xenografts.
- © 2003 American Society of Nephrology
References
