Relevance of Posttransplant Flow Cytometric T- and B-cell Crossmatches in Tacrolimus-Treated Renal Transplant Patients

Abstract

Background. 

De novo development of anti–human leukocyte antigen (HLA) antibodies after transplantation is associated with increased rejection and decreased graft survival. In this study, the effect of posttransplant HLA antibodies on clinical outcome was evaluated in patients treated with tacrolimus by means of flow cytometric crossmatches (FCXm).

Methods. 

T- and B-cell FCXm were performed retrospectively on posttransplant sera of patients who received a graft between 1997 and 1999. Ninety-four kidney-only recipients were tested and all FCXm positive sera were investigated for the presence of HLA class I and II antibodies by Flow panel reactive antibodies.

Results. 

From 94 patients with a negative pretransplant complement-dependent cytotoxicity crossmatch, seven (7%) showed a positive pretransplant FCXm. After transplantation the FCXm became positive in five patients (6%). The predictive value of a positive FCXm after transplantation, and the log-transformed relative change in fluorescence ratio between pretransplant and posttransplant serum, were not significant to rejection within six months, nor to graft survival censored for death.

Conclusions. 

The presence of HLA antibodies before rejection or graft failure could only be shown in a minority of patients; most antibodies were detected after graft failure, especially after transplantectomy. Monitoring through antibody testing after transplantation on the basis of our results has no added value with tacrolimus-based immunosuppression.

De novo development of anti–human leukocyte antigen (HLA) antibodies after transplantation is associated with increased acute and chronic rejection and decreased graft survival (^1–4). Antibodies to both HLA class I and class II antigens seem to be detrimental, but some studies found HLA class I antibodies to be associated with early graft failure and class II antibodies with steady deterioration in graft function (^{5, 6}). Terasaki et al. (^{7, 8}) showed that posttransplantation HLA antibody testing is useful in predicting subsequent graft failure. HLA antibodies are detected before graft failure (^{5, 6, 9}) and even before rejection (²).

An elegant way to show the presence of HLA-antibodies directed against the potential donor is the performance of crossmatches (Xm) (¹⁰). Historically complement-dependent cytotoxicity (CDC) was used, but the flow cytometric (FC) method has become increasingly popular. FC is more sensitive in the detection of antidonor antibodies (^{1, 11–13}) and a positive flow cytometric crossmatch (FCXm) before transplantation is considered a risk factor for early rejection, chronic rejection, and graft failure (^{12, 14, 15}). Performing FCXm after transplantation may be a tool for clinical monitoring of patients after transplantation (^16–19).

In contrast with these findings, we demonstrated in a previous study that patients can be safely transplanted over a positive pretransplant T-cell FCXm when they are carefully selected, screened, and followed regarding immunological events during their time on the waiting list (²⁰). This was also found by Scornik et al. (¹²), who in al large cohort of patients demonstrated that transplantation should not be denied to FCXm-positive patients who have been subject to careful selection. However, when posttransplant FCXm results were expressed as relative change in fluorescence ratio (RCFR), they proved to be a significant risk factor for graft survival in Cox regression analysis (²¹).

The immunosuppressive therapy used is an important factor in posttransplant analysis. In several studies, a correlation between a positive FCXm after transplantation and graft survival was proven under cyclosporine-based immunosuppression (^{16, 17, 19, 21}). Superiority of tacrolimus over cyclosporine with respect to rejection-free survival was demonstrated in trials in Europe and the United States (^22–24). Five-year graft survival results of the FK506 Kidney Transplant Study Group showed a lower number of graft failures (²⁵).

The influence of a positive B-cell FCXm on renal transplantation remains unclear. Several centers found a positive pretransplant B-cell FCXm to be associated with detrimental graft survival (^26–28), while others did not (²⁹). A positive B-cell FCXm that is not caused by anti-HLA antibodies results in the same graft survival as in negative patients; however, when positivity is caused by anti-HLA class II antibodies, a clear effect on graft survival is shown (²⁹).

The purpose of the present study was to evaluate the effect of posttransplant FCXm (expressed as positive FCXm and RCFR) on rejection and graft survival in patients treated with tacrolimus. Both T- and B-cell crossmatches were analyzed.

MATERIALS AND METHODS

Patients

From January 1997 until January 1999, 124 transplantations were performed at the University Hospital of Maastricht. Excluded from analysis were patients with CsA-based therapy (n=17), combined kidney-pancreas transplantation (n=5), and patients lacking donor material (n=8). Thus 94 consecutive kidney transplantations with tacrolimus-based therapy were analyzed. Clinical follow-up was continued through August 1, 2005. Pretransplant characteristics are given in Table 1.

TABLE 1: Pretransplant patient characteristics (n = 94)

Immunosuppression

The immunosuppressive regimen was tacrolimus-based (TAC) in all recipients. In addition, all recipients received prednisone (PRED), mycophenolate mofetil (MMF) was added in 30, and azathioprine (AZA) in three. TAC trough-levels were measured in whole blood by lmx (Abbott) or HPLC-MS/MS; target levels were 15–20 ng/ml for week 1+2 and 10–15 ng/ml for week 3+4, thereafter tapering to 5–7 ng/ml.

Clinical Outcome Parameters

Acute rejection was defined as any rejection treatment within six months after grafting. Rejection episodes were proven by needle core biopsy (Banff ≥1A) except for three cases in which biopsy material was inadequate. Rejection treatment consisted of three doses of methylprednisolone (0.5–1.0 g/dose). For steroid-resistant rejections, second-line therapy consisted of antithymocyte globulin (ATG). Graft failure was defined as return to dialysis or transplant nephrectomy but did not include death of recipient with functioning graft.

Center Policy of Tissue Typing

All patients were typed for class I by serology and for class II by molecular methods (polymerase chain reaction [PCR]-SSP) at the time of transplant. HLA antigens were considered unacceptable for a patient if antibodies against the specificity had ever been demonstrated. Mismatches from previous transplants were excluded, as were the paternally inherited antigens of children in female patients. The laboratory is accredited for the American Society for Histocompatibility and Immunogenetics (ASHI) and the European Federation for Immunogenetics (EFI). Crossmatches before transplantation included different CDC crossmatches: the standard National Institutes of Health (NIH) crossmatch with and without dithiothreitol (DTT) and the two-color fluorescence crossmatch. Sera used were the pretransplant serum and all relevant positive historical samples (peak sera). A negative class I CDC crossmatch was mandatory for transplantation. Screening was performed in ELISA and CDC every three months according to Eurotransplant protocol and after every immunizing event. Enzyme-linked immunosorbent assay (ELISA) screening (LAT-M, One Lambda Inc., Canoga Park, CA) was performed to identify the presence of HLA class I and/or II antibodies. CDC-screening was performed with and without DTT using a selected 60 cell panel. Reading was performed using a Leitz Patimed automated microfluorometer microscope, positivity was defined as >10% of cell death in the patient serum over the negative controls.

Serum Samples

Serum samples were collected after transplantation: weekly during hospitalization, monthly until six months, and yearly thereafter. After transplantectomy sera were collected at week zero, one, two, three and four, and every three months thereafter if the patient returned to the waiting list. Retrospective FCXms were performed on pretransplant and posttransplant sera. Sera tested for acute rejection were samples of rejection dates −3 and +3 weeks. Sera of day eight, 20, and 32 posttransplant were tested for recipients without rejection episodes. In case of nephrectomy, sera at −3 and +3 weeks were tested if available. On average, eight sera per patient were tested. Sera drawn at the time of ATG-treatment were excluded from analysis (³⁰).

Detection of Antidonor Antibodies by Flow Cytometric Crossmatching

The FCXm technique was performed as described previously (⁷) with small modifications. In short, 3×10⁵ lymphocytes were incubated with 50 μl of recipient serum for 30 min at 20°C. After washing three times with 4 ml of buffer, the supernatant was removed and 20 μl of FITC (fluorescein isothiocyanate)-conjugated F(ab′)2 goat anti-human IgG (Tago, Burlingame, CA), diluted 1:16 were added, as well as 5 μl of PE (phycoerythrin)-conjugated monoclonal mouse anti-human T-cell (CD3) and 5 μl of R-PE-cyanine5-conjugated monoclonal mouse anti-human B-cell (CD19) (Dakopatts a/s, Glostrup, Denmark) and incubated for 30 min at 4°C. After washing, the cells were resuspended in 250 μl of cold PBS and stored in the dark at 4°C until use.

Donor lymphocytes were incubated with all relevant recipient sera, a positive and a negative control serum. The positive control was an alloantiserum with 100% panel reactive antibody (PRA), the negative one a commercial serum pool obtained from healthy male donors. All tests were performed in duplicate.

Lymphocytes were analyzed by logarithmic amplification on a 1024-channel resolution with a FACS Calibur system (Becton Dickinson, San Jose, CA, USA). Quality control beads (Calibrite III, Becton Dickinson) were run before testing to set standards for photomultiplier tubes, fluorescence compensation and sensitivity. In the CD3 and CD19 positive, viable lymphocyte gate a minimum of 5000 cells were counted and analyzed for each sample. The linear median value of fluorescein isothiocyanate (FITC)-fluorescence was collected for CD3- and CD19-positive cells.

External proficiency testing for FCXm are performed five times yearly (UK NEQAS) and proven adequate.

A FCXm was considered positive if the ratio of median value of recipient serum to negative control was >1.97 (mean+3SD) for T-cells and >2.99 for B-cells, provided that the ratio positive control to negative control was > than the cut-off value. Background staining in B-cell FCXm was diminished by inactivating the sera at 56°C for 30 min.

Presence of donor-specific antibodies was also defined by RCFR (²¹). RCFR is defined as the median channel fluorescence of posttransplant serum minus pretransplant serum divided by pretransplant serum times 100%. All patients including pretransplant FCXm positive patients were analyzed. RCFR was calculated at the time of acute rejection and >3 weeks thereafter for T- and B-cell FCXm values.

Detection of Class I and II Antibodies by Flow Cytometric Screening

To detect the presence of HLA class I and II antibodies the Flow PRA Screening Test (One Lambda) was used according to the manufacturer's recommendations. Microbeads are coated with purified HLA antigens, consisting of a pool of 30 class I and 30 class II coated beads representing the most common HLA antigens. Five microliter class I and 5 μl class II beads were incubated with 20 μl of recipient serum for 30 min at 20°C. After two washes, 100 μl of goat FITC-conjugated F(ab′)2 anti-human IgG Fc(One Lambda), 1:100 were added and incubated for 30 min at 20°C. Beads were gated based on different FL2 fluorescence emission. Antibodies on the bead surface were identified by comparing FL1 fluorescence intensity in patient sera to positive and negative control sera. Class I and II antibodies were distinguished in a single test. Based on results using 20 negative serum samples, a serum was considered positive if the PRA percentage exceeded 10%.

Statistical Analysis

All statistical analyses were performed by using Statistical Package for the Social Sciences software (SPSS, version 11.5 for Windows). Pearson Chi-square test was performed when indicated. P values <0.05 were considered statistically significant.

The proportional hazards regression analysis (Cox regression) was performed by both stepwise forward and backward selection techniques for all suspected risk factors. In case acute rejection within six months after transplantation was the outcome parameter, the suspected risk factors tested for included the result of the FC technique (FCXm positivity or RCFR), recipient age, gender, donor age, transplant number, pretransplant PRA, HLA-AB mismatches, HLA-DR mismatches, and the use of MMF as additional immunosuppression. Pretransplant PRA was classified as: nonimmunized (0–5%), immunized (6–50%), and highly immunized (51–100%). With graft survival as outcome parameter, acute rejection within six months was included additionally as a risk factor in the model.

RESULTS

Pretransplant FCXm

Pretransplant FC crossmatches were performed retrospectively for all patients. From 94 patients with negative pretransplant and historic CDC crossmatches, seven (7%) showed a positive FC crossmatch (2T, 3B, 2TB). From the two T-cell positive recipients, one had an uneventful course, the other underwent graft failure after six years. Two of the three B-cell positive patients underwent a steroid-sensitive acute rejection-episode 18 and 47 days after transplantation. The third one lost his graft due to thrombosis after six days. One TB-cell positive patient underwent a steroid-resistant acute rejection-episode six days after transplantation, was treated with ATG and plasmapheresis, but the graft was lost two years later. The other TB positive patient has a functioning graft seven years after transplantation.

Graft outcome and acute rejection rate were compared between patients with a negative (n=87) and positive (n=7) pretransplant FCXm. In Cox regression analysis, the FCXm result as independent risk factor did not contribute to acute rejection within six months (P=0.30; CI 95% 0.57–6.35) graft survival nor to graft survival censored for death (P=0.40; CI 95% 0.50–5.60). Of the independent parameters gender, MMF, recipient age, donor age, HLA-AB mismatch, HLA-DR mismatch, pretransplant PRA, and transplantation number, none contributed to acute rejection within six months or to graft survival censored for death.

Clinical Outcome

Acute rejection episodes within six months were shown in 23/87 of the pretransplant FC negative patients (26%). For three recipients, no adequate biopsy material was available; in the other 20 a Banff grade ≥1A was found. In total, 13 Banff 1A, four 1B and three 2A were detected. All primary acute rejections, except one (ATG), were treated with methylprednisolone. Three patients were steroid-resistant and treated with ATG.

Grafts failed in 23/87 recipients (26%). Causes of graft failure were acute rejection (7), primary nonfunction (4), vascular problems (3), thrombosis (3), septicemia (2), and other (4).

Posttransplant FCXm before Graft Failure

The sera of the pretransplant FC negative recipients (n=87), who had their grafts in situ, were studied for at least seven years posttransplant. FCXms became positive in five patients (6%), whereas 82 stayed negative (94%). From the positive recipients, two were T- and three were B-cell positive (Fig. 1). Time between transplantation and antibody detection was 19, 22, 26, 1,083, and 1,777 days.

FIGURE 1.: Patients with positive posttransplant FCXm. Period of positive FCXm ▪, negative FCXm , not tested □. ↓, serum tested; R, rejection episode; F, graft failure; †, deceased.

Acute rejections were found in one B-cell positive patient (graft loss at six months) and 22 in negative recipients (27%, Chi-square P=0.26). Two patients turned positive after acute rejection treatment (1T, 1B), both still have a functioning graft after seven years.

Graft failure occurred in two (40%) positive (1B, 1T) and 21 negative patients (25%, Chi-square P=0.48). The first patient (Bpos) became positive at day 22 after transplantation but 3 days before acute rejection, with graft failure at day 205, the second one (Tpos) at day 1,083, with graft failure at day 1,170. The predictive value of a positive FCXm after transplantation is not significant to acute rejection within six months or to graft survival censored for death.

Relative Change in Fluorescence Ratio

The difference between pretransplant and posttransplant FCXm value was expressed as RCFR. By definition patients with a pretransplant positive FCXm were included. Serum samples tested were drawn at the time of acute rejection (median day 20) or around day 20 for patients without acute rejection. RCFR was calculated separately for T- and B-cell Xm.

T-cell RCFR

An increase in RCFR was shown in 24/94 patients. The RCFR ranged from −93% to 232% with a median of −7% (Table 2). Because of the skewed distribution, log transformation of the data was performed. In Cox-regression analysis, the log-transformed RCFR did not contribute significantly to acute rejection within six months (P=0.68) nor to graft survival censored for death (P=0.87).

TABLE 2: RCFR of T- and B-cell FCXm at the time of rejection

B-cell RCFR

An increase in RCFR was found in 38/94 patients. The RCFR ranged from −97% to 393% with a median of −5% (Table 2). Cox-regression analysis showed no significant contribution to acute rejection within six months (P=0.82) nor to graft survival censored for death (P=0.18).

Posttransplant FCXm after Graft Failure

The number of pretransplant FCXm negative patients having graft failure was 23. In two of them, FCXm antibodies were found before failure. In eight of the remaining patients (38%), a positive FC crossmatch was shown after graft failure (1T, 7 TB). Median time between graft failure and antibody detection was 161 days (n=8; range 19–424 days). Causes of graft failure in these patients were acute rejection (2), primary nonfunction (2), and vascular problems (4).

Presence of Class I or II Antibodies

All FCXm positive sera were tested for HLA antibodies by Flow PRA in the pre- and posttransplant serum. In the four pretransplant T-cell positive patients, class I antibodies were detected in the sera of two; in the five B-cell positive patients, four showed the presence of class II antibodies, whereas one was negative. From 10 posttransplant T-cell positive patients, eight presented with class I antibodies, whereas two did not. Class II antibodies were detected in eight out of 10 B-cell positive recipients.

Five patients showed a positive FCXm posttransplant at the time of functioning graft. Two of them showed HLA- antibodies: both became positive after acute rejection and still have functioning grafts after seven years of follow-up. In three patients with positive FCXm, no HLA-specificities were detected.

DISCUSSION

Posttransplant HLA antibody testing is considered to be useful for monitoring recipients because HLA antibodies are claimed to be found prior to rejection and graft failure. The purpose of the present study was to determine the prevalence of posttransplant positive FCXm in relation to acute rejection and graft survival in tacrolimus-treated patients.

From 94 patients with a negative pretransplant and historical CDC crossmatch, 7% showed a positive pretransplant FCXm. Although the influence of a positive pretransplant FCXm on clinical outcome has been reported in several studies (^{13, 15, 31–33}), Cox regression analysis of the present data revealed no advantage of pretransplant FCXm in predicting graft survival. Although the number of positive patients is small, it is consistent with our earlier findings and the findings of Scornik et al., who demonstrated the same phenomenon in a large group of patients (^{12, 20}). The explanation most probably is that the knowledge about pretransplant immunologic history of the patients differs between centers. All pregnancies for female patients, transfusions, and prior transplants are registered. Antibody screening is performed on all sera that are drawn according to existing pre- and posttransplant protocols. Antibodies are defined in different techniques, the antigens against which antibodies have been detected are excluded from a future transplant. Paternally inherited antigens of the children in female patients are also excluded as are the mismatched antigens from earlier transplantations. Negative CDC crossmatch results are mandatory with pretransplant and historical recipient sera.

The importance of a well-documented history is illustrated in one of the patients. Although this knowledge is indispensable, sometimes it is lacking and no historical sera are available. A 30-year-old female patient was referred to us from elsewhere. The immunizing antigens of her two pregnancies were known and excluded, but no historical sera were available nor was the transfusion history documented. She was transplanted with a kidney from her sister, with only one HLA-A mismatch, and experienced six days later a severe tubulointerstitial and vascular acute rejection. Treatment consisted of ATG and plasmapheresis but the graft failed after two years. In the FCXm-positive sera, HLA class I and II antibodies were detected. This illustrates that in situations where little or no information is available on the history of a transplant patient, the relevance of FCXm is clearly demonstrated.

Posttransplant FCXm became positive in 6% of the patients (n=5) who had their grafts in situ, two were T-cell and three B-cell positive. The discrepancy between T-cell positivity and the reactivity in the corresponding B-cells of two recipients is the result of a higher cutoff point for positive B-cells. In fact, the B-cell FCXms of these two patients were borderline negative (T cell values 2.09 and 4.07, B-cell values 2.28 and 2.57, respectively). Of 23 patients with acute rejection episodes, one patient became positive before acute rejection, two turned positive after acute rejection without graft failure. Only two patients became FCXm positive before graft failure, the other three still have a functioning graft. A lower number of positive posttransplant FCXm were found compared to our previous study (6% versus 12%) (²¹). In the present study, no relation between FCXm positivity and clinical outcome was shown. Most posttransplant antibodies were detected after graft failure or transplantectomy (³⁴).

Comparison of the RCFR results from this study with the study from Christiaans et al. (²¹) showed striking differences in the lower number of patients with a positive RCFR (26% vs. 52%) and the lower maximum positivity (232% vs. 1000%). No B-cell RCFR values were available in the previous study, the present study showed an increase of 44% at the time of acute rejection. Maybe due to the lower RCFR, the significant correlation between RCFR and clinical outcome proven by Christiaans et al. (²¹) could not be confirmed in the present study. This could be the result of: (a) lesser sensitivity of the technique used, (b) different patient characteristics, or (c) the change in immunosuppression. (a) The technique used was the same as previously described, although with slight modifications; an additional washing step, a different negative control and duplicate testing of samples. The modified procedure was evaluated by external proficiency testing (NEQAS). Testing of the dilutions of identical positive and negative controls, showed no significant differences in sensitivity compared to the old protocol. Therefore, the different results in this study in our opinion are not accounted for by the modifications. (b) Concerning patient characteristics, the present study included both living and deceased donors, whereas the previous one had no living donations. In the present study the age of patients and donors was 10 years higher, the average number of retransplantations was lower (20% vs. 29%) and mean HLA-DR mismatch was higher (0.78 vs. 0.44). Mean HLA-AB mismatch (1.83 vs. 1.60) and mean CIT (24.6 vs. 29.6 hr) were comparable. (c) The most obvious difference between both studies was the change in immunosuppression from cyclosporine A to tacrolimus. Cyclosporine A has been shown to act on naive T cells, whereas tacrolimus also acts on memory cells (^{35, 36}) and a lower incidence of antibody formation has been shown in tacrolimus treated patients compared to CsA (³⁷). In addition, 32% of the patients in the pretransplant study were treated with MMF. MMF is known to reduce HLA class I and II antibody production in kidney transplant recipients (^{38, 39}). This may very well account for the lower number of patients with a positive RCFR and the lower level of positivity in the present study.

All positive patients were tested for the presence of class I and II antibodies in the relevant sera by flow cytometric methods (^40–42). Although anti-HLA antibodies were demonstrated in most positive sera, some samples were negative for HLA class I or II. This might be the result of the presence of non-HLA antibodies, which have been described (^{29, 43–45}). We did not study the cause of this positivity but non-HLA immunity might play an important role in graft survival (⁴⁶). In a recent study, Terasaki (⁸) found the CDC screening technique to be better correlated with graft outcome than techniques based on the use of HLA coated beads, probably because CDC also detects non-HLA antigens.

In this large group of patients followed for a long time, the presence of HLA antibodies before acute rejection or graft failure could only be shown in a minority of patients, most antibodies were detected after graft failure, which is clearly in contrast with the findings by Terasaki and others (^2–8). We were not able to prove an advantage of pretransplant FCXm over CDCXm, in patients whose immunological history is well-known, whose sera are screened regularly in different techniques, and whose immunizing results are taken into account at the time of transplant decision. Posttransplant T- and B-cell RCFR did not predict acute rejection nor graft survival in tacrolimus-treated renal recipients. Clinical monitoring through antibody testing after transplantation has on the basis of our results no additional value, most likely due to the use of tacrolimus and MMF.