Post-Transplant Anti-HLA Class II Antibodies as Risk Factor for Late Kidney Allograft Failure

Authors


*Corresponding author: Maria Gerbase-DeLima, gerbase@igen.epm.br

Abstract

The purpose of this study was to prospectively analyze the relationship between the post-transplant anti-HLA class I and/or class II panel reactive antibodies and graft failure due to chronic allograft nephropathy (CAN). We studied 512 first kidney recipients transplanted at a single center, with a graft functioning for at least 3 years. A single blood sample was collected from each patient for antibody evaluation. The median posttransplant time after blood collection was 4.4 years and did not differ between patients with (n = 91) or without anti-HLA antibodies (n = 421). Female gender, pregnancies and blood transfusions were associated with the presence of anti-HLA class I antibodies. Graft function deterioration was associated with anti-HLA class II antibodies. Multivariate analysis showed independent association for creatinine levels (RR = 7.5), acute rejection (RR = 2.6), recipient male gender (RR = 3.6) and anti-HLA class II antibodies (RR = 2.9) and CAN-associated graft loss. In conclusion, the presence of anti-HLA class II antibodies conferred a risk for graft loss before a decline in renal function and increased the risk of graft failure in patients who already had a decline in graft function. Thus, anti-HLA class II antibody monitoring is a useful tool for the management of long-term kidney recipients.

Introduction

Chronic allograft nephropathy (CAN) and death of the patient with a functioning graft are the two leading causes of graft loss after the first year following transplantation (Tx), resulting in an annual graft loss rate of 3–5% (1). The pathogenesis of CAN is complex, comprising immunological and nonimmunological factors, whose importance may vary depending on the time interval after Tx (2,3). Among all the variables that may contribute to the development of CAN, the pivotal role of immunological factors is clearly demonstrated by the extremely low incidence of CAN in grafts from HLA-identical living-related donors (4). The mechanisms by which the immune system operates to destroy the graft are not yet clear. Classically, attention has been mainly focused on cell-mediated immunity but in the last years a possible important role of antibody-mediated mechanisms has come into evidence.

Retrospective or cross-sectional studies have shown association between the presence of anti-HLA antibodies in serum and impaired graft function, CAN or graft loss (5–7). In the majority of these studies the antibodies were assessed against a panel of HLA antigens (panel reactive antibodies, PRA). Interestingly, quite frequently the antibodies present in serum were apparently not directed against HLA antigens from the donor and this finding was probably related to the fact that donor-specific antibodies were not in the blood but were bound to the graft. The detection of C4d, a protein from the complement activation cascade that remains attached to the site of complement activation, in biopsies with CAN has also been claimed as an argument for the participation of antibodies in the pathogenesis of CAN (8,9). Although these studies do not allow a definitive conclusion concerning the role of anti-HLA antibodies in the pathogenesis, or even as predictive markers, of CAN, they definitely called attention to the possible role of antibodies in the pathogenesis of CAN.

The preliminary results of a multicenter study designed to prospectively evaluate the relationship of anti-HLA antibodies and kidney graft loss were published in 2004. Panel-reactive anti-HLA antibodies were found in the serum of 20.9% of the 2278 kidney graft recipients evaluated at least 6 months after Tx. After 1 year of follow-up, 6.6% of the antibody-positive recipients had lost their grafts, as opposed to 3.3% of the antibody-negative group (p = 0.0007). Although very important considering the large number of patients included, this study did not address questions concerning the different impact of class I or II anti-HLA antibodies, the cause of graft loss, nor did it attempt to evaluate other possible factors associated with graft loss (10).

The present study was designed to determine whether the presence of anti-HLA class I and/or class II panel-reactive antibodies in serum was predictive of graft loss due to CAN. This investigation differs from all other studies published thus far concerning anti-HLA antibodies and late kidney graft loss because it is prospective and based on a large number of patients from a single Tx center and under very similar immunosuppressive therapy. The occurrence of CAN was based on clinical and histological features, and the roles of anti-HLA class I and class II antibodies were separately evaluated.

Materials and Methods

Subjects

The subjects of this study were 512 recipients with primary kidney grafts functioning for at least 3 years. All transplants (200 with kidneys from deceased donors and 312 with kidneys from living non-HLA identical donors) were performed at the Federal University of São Paulo/Hospital do Rim between 1981 and 2000. All recipients were negative for pre-transplant antibodies against donor HLA class I antigens, as assessed by a complement-dependent cytotoxicity test. Pre-transplant anti-donor HLA class II antibodies were not regularly investigated and, when present, were not considered to be a contraindication for transplant. Among the recipients, 55.7% were males, 28.5% were Caucasian, and the median age at Tx was 34.5 years. The median number of pre-Tx blood transfusion was three. Most patients (81.4%) received a triple-therapy regimen consisting of prednisone, cyclosporine and azathioprine, with initial doses and tapering regimens that varied according to the year of Tx, and a maintenance dose adjusted to obtain appropriate levels or to avoid toxicity. The remaining patients (18.6%) received alternative regimens with the addition of tacrolimus, mycophenolate mofetil, rapamycin, everolimus or FTY. Of the 512 patients, only 11 (2.1 %) received induction therapy, which consisted of anti-CD3 monoclonal antibody (n = 5), anti-lymphocytic polyclonal antibodies (n = 1) or anti-IL2R monoclonal antibody (n = 5). A single blood sample was collected from each patient for anti-HLA class I and class II antibody evaluation. The protocol was approved by the Institutional Medical Ethics Committee and informed consent was obtained from each patient prior to enrollment in the study.

Anti-HLA antibodies determination

The presence of anti-HLA class I and II panel-reactive antibodies was assessed by an enzyme-linked immunosorbent assay test (LATM One Lambda Inc., Canoga Park, CA).

Data analysis and endpoints

Patients with anti-HLA class I, anti-HLA class II and anti-HLA class I and class II antibodies were compared with antibody-negative patients in terms of baseline characteristics, clinical events after Tx (cytomegalovirus infection, diagnosed by clinical symptoms associated with seroconversion or positivity in more than five cells in the antigenemia assay, acute tubular necrosis and acute cellular rejection), and renal function at the time of antibody evaluation. Renal function was estimated by the variation of serum creatinine levels: % creatinine delta =[(X − Y)/Y]× 100; where X is the creatinine level at the time of antibody testing, and Y is the lowest documented post-transplant creatinine level. Graft function deterioration was defined as a creatinine delta >100%. The primary endpoint was graft loss due to CAN after antibody evaluation. Another endpoint was death of the patient.

The diagnosis of CAN was based on laboratory tests (increase in serum creatinine levels and/or proteinuria) and histological criteria (presence of chronic interstitial fibrosis and tubular atrophy, with or without fibrointimal vascular thickening, as recommended in the Banff classification). Unfortunately C4d staining was not adopted into the biopsy histopathology protocol during the major part of the study period, and a retrospective analysis was not possible.

Statistical analysis

Baseline characteristics were compared between groups by the chi-squared test and Mann-Whitney nonparametric test for categorical and continuous variables, respectively. The chi-square test was used to estimate the association between the presence of anti-HLA antibodies and graft loss in the first year of follow-up after antibody evaluation. Graft loss due to CAN throughout the follow-up (12–23 months) was calculated by the Kaplan-Meier method. The comparison of graft survival curves, censored for death and graft loss due to other causes, was performed using the log-rank test. Multifactorial Cox regression analysis was performed to investigate the impact of other prognostic factors on the observed association of the presence of anti-HLA class II antibodies with CAN. Statistical analysis was performed using SPSS v 10.0 for Windows. All p values were two-sided, and a value of less than 0.05 was considered to indicate statistical significance.

Results

Proportion of patients with anti-HLA antibodies

Among the 512 recipients, 91 (17.8%) were positive for anti-HLA antibodies, with 55 (10.7%) being positive for anti-HLA class II, 20 (3.9%) for anti-HLA class I and 16 (3.1%) for anti-HLA class I and class II antibodies.

Baseline characteristics of the patients and variables associated with the presence of anti-HLA antibodies.

The time from transplant to blood collection (3–22 years, median 4.4 years) did not differ between patients with or without anti-HLA antibodies.

The baseline characteristics of the patients are shown in Table 1. The proportion of recipients under the classic triple immunosuppression scheme (cyclosporine/prednisone/azathioprine) did not differ among the four groups of patients. Female gender, pregnancies and blood transfusions were associated (p < 0.05) with the presence of anti-HLA class I antibodies. Recipient age or race, donor age or type (living or deceased), cold ischemia time, the occurrence of acute tubular necrosis, acute rejection or cytomegalovirus disease did not differ between groups. Graft function deterioration (creatinine delta >100%) was associated with class II anti-HLA antibody either isolated (p = 0.03) or accompanied by class I antibodies (p = 0.02). The proportion of patients with a creatinine delta >100% did not differ between patients with only anti-HLA class II antibodies and patients with anti-HLA class I and class II antibodies.

Table 1.  Recipient, donor and transplant (Tx) variables in relation to positivity for anti-HLA class I and/or class II antibodies (Ab)
Variable1Without anti-HLA AbAnti-HLA class I AbAnti-HLA class II AbAnti-HLA class I and II Ab
  1. 1Variables not represented by proportions are shown as median values.

  2. 20.01 < p < 0.05, compared to the group without antibodies.

  3. 30.001 < p < 0.01, compared to the group without antibodies.

  4. 4p = 0.07, compared to the group without antibodies.

  5. 5Creatinine delta: percent variation of serum creatinine at the time of Ab evaluation in relation to the lowest post-Tx value.

N421205516
Years after Tx, at Ab evaluation4.43.75.24.6
Recipient age at Tx (years)34.241.531.941.2
Female sex (%)44.260.030.975.02
Women with previous pregnancies (%)61.6100.0346.675.0
Non-white recipients (%)69.185.074.556.2
Number of pre-Tx blood transfusions3.07.523.04.0
Months on dialysis before Tx24.031.026.038.5
Tx with deceased donors (%)39.730.034.550.0
Donor age (years)38.037.031.032.5
Hours of cold ischemia time24.016.0226.019.5
Cyclosporine/prednisone/azathioprine (%)81.275.081.8100.0
Acute tubular necrosis (%)20.210.018.243.74
Acute cellular rejection (%)43.555.060.050.0
Cytomegalovirus disease (%)7.115.014.56.2
Creatinine delta ≥100% (%)519.215.032.7243.72
Creatinine values at Ab evaluation (mg/dL)1.701.652.001.95

Anti-HLA antibodies and CAN-associated graft loss after antibody testing

During the first year of follow-up, 12.1% of antibody-positive recipients lost their grafts, while this occurred in only 5.5% of antibody-negative patients (p = 0.03, RR = 2.34 (CI 95% 1.11–5.07). When the different groups of antibody-positive patients were separately analyzed, graft loss was associated with the presence of anti-HLA class II (p = 0.04) or anti-HLA class II plus class I (p = 0.006) antibodies, but not with the presence of only anti-HLA class I antibodies. There was no association between the presence of antibodies and patient death.

During the entire follow-up period of 12–23 (median 17) months after antibody evaluation, 18 patients (3.5%) died with a functioning graft, 32 (8.2%) lost the graft (returned to dialysis or received a second transplant) and 4 patients (0.8%) were lost to follow-up. Among the 32 graft losses, 27 were due to CAN, and the other 5 due to urinary tract infection or use of nonsteroidal anti-inflammatory drugs. As shown in Table 2, the following variables were independently (Cox regression analysis) associated with CAN-associated graft loss: anti-HLA class II antibodies (p = 0.03), recipient gender (male) (p = 0.01), previous occurrence of acute rejection (p = 0.05) and a creatinine delta >100% (p < 0.0001). Patients with a creatinine delta ≥100% did not differ from the patients with a delta <100% in terms of immunosuppression, occurrence of acute tubular necrosis or acute rejection.

Table 2.  Factors associated with kidney graft loss due to chronic allograft nephropathy, after the third year of transplantation
 Univariate analysis (p-value)Multivariate analysis (p-value)Relative risk (95% confidence interval)
Recipient gender (male)0.050.013.58 (1.31–9.82)
Acute cellular rejection0.020.052.62 (1.01–6.83)
Creatinine delta ≥100%<0.0001<0.00017.52 (3.11–18.18)
Anti-HLA class II antibodies0.0010.032.89 (1.13–7.38)

The survival curves (Kaplan-Meier) in Figure 1 show the lower CAN-free graft survival in recipients with anti-HLA class II antibodies and the curves shown in Figure 2 illustrate the independent association of creatinine levels and positivity for anti-HLA class II antibodies with the occurrence of CAN-associated graft loss.

Figure 1.

Graft loss due to chronic allograft nephropathy (CAN) according to the presence of anti-HLA class I or II antibodies (Ab). The curves, censored for death and graft loss due to other causes, were constructed according to the method of Kaplan-Meier and compared by the log-rank test.

Figure 2.

Graft loss due to chronic allograft nephropathy (CAN) according to the presence of anti-HLA class II antibodies (Ab) and graft function. Graft function is expressed as percentual variation (delta) of serum creatinine at the time of antibody evaluation in relation to the lowest post-transplant value. The curves, censored for death and graft loss due to other causes, were constructed according to the method of Kaplan-Meier and compared by the log-rank test. The 16 patients that were positive for anti-HLA I and class I antibodies were not included in this analysis.

Discussion

The identification of the factor/s that are responsible for or major contributors to the development of CAN in a particular patient is of great importance in the clinical setting, since it may provide the basis for the establishment of interventions that may arrest injury progression. The present study was designed to evaluate the role of anti-HLA antibodies as risk factors for CAN-associated graft loss.

Among the 512 graft recipients evaluated at least 3 years after Tx, 18% presented anti-HLA class I and/or class II antibodies in serum, a frequency similar to those reported by other authors (10,11).

The unexpected low percentage of recipients with only anti-HLA class I antibodies (3.3%), or with anti-HLA class I and II antibodies (1.3%), may result from a negative selection of anti-HLA class I antibody producers, since these antibodies are associated with early graft losses, as shown previously (12,13). A surprising finding, however, was that no association was detected between the presence of HLA class I antibodies and late graft functional deterioration or loss. Corroborating this finding, it is interesting to note that the presence of anti-HLA class I antibodies was associated with female gender, previous pregnancies and blood transfusion before Tx and none of these factors were associated with CAN-associated graft loss.

Within a follow-up period of 12–23 months after blood collection for antibody evaluation, 5.3% of the patients lost their grafts due to biopsy-proven CAN and four independent factors were found to be associated with this endpoint: male gender (RR = 3.6), previous acute rejection (RR = 2.6), graft dysfunction at the time of antibody evaluation (RR = 7.5) and the presence of anti-HLA class II antibodies (RR = 2.9).

The association of CAN with male gender has been previously described but it is always a matter of debate whether this association is specifically related to the gender of the recipient or reflects size differences between donor and recipient (female donors to male recipients would frequently be a poor match concerning size) (14,15). The finding of acute rejection as a risk factor for CAN agrees with several other studies (2,16). Since an increase in serum creatinine levels is frequently a consequence of ongoing CAN, it is not surprising that this variable represents a major risk factor for CAN-associated graft loss (14).

The association of anti-HLA class II antibodies with graft loss due to CAN represents an original finding derived from a prospective study, and it is very important to point out that the presence of these antibodies, in patients with a graft functioning for at least 3 years, conferred a higher risk of CAN-associated graft loss, even in patients that, at the time of antibody testing did not present an important decrease in renal function, i.e. presented serum creatinine levels less than twice the lowest posttransplant value. In a recent publication concerning updated results of a multicentric study, Terasaki and Ozawa concluded that the presence of anti-HLA antibodies (this study did not discriminate anti-class I and anti-class II), in patients with functioning kidney grafts for at least 6 months, is predictive of subsequent graft failure, and that this predictive value is greater among patients with higher serum creatinine values (17).

On the other hand, it is important to note that some patients with anti-HLA antibodies did not lose the graft or show dysfunction within the follow-up period. We wonder if these patients could represent cases of accommodation, defined as the absence of graft dysfunction in the presence of anti-donor antibodies in the circulation, a phenomenon best characterized in studies of ABO-incompatible and xeno-transplants. The mechanisms underlying accommodation are unknown, and one interesting finding is that grafts might survive undamaged in spite of antibodies and even complement activation, represented by C4d deposits (18–21).

In conclusion, our findings lend support to the notion that post-transplant monitoring of anti-HLA class II antibodies represents an important tool for clinical management of kidney transplant recipients. The presence of these antibodies conferred a risk for graft loss before a decline in renal function, increased the risk of graft failure in patients who already had a decline in graft function, and therefore may indicate which patients could benefit from a change in immunosuppressant therapy toward the use of drugs that inhibit not only the T-cell function but also the antibody production by B cells. It would be of high clinical relevance to prospectively test the concept of anti-HLA class II antibodies as a modifiable, i.e. amenable to treatment, risk factor for CAN-associated graft loss.

Ancillary