Hyperacute rejection of a kidney allograft is the most dramatic manifestation of antibody-mediated graft failure, but donor-specific humoral presensitization, unmasked with the use of ultrasensitive crossmatch protocols, leading to antibody-mediated vasculitis and early graft failure, has been reported in as early as 1980 (1). The contributions of antibodies to rejection and allograft failure have reemerged forcefully with the improved detection of donor-specific anti–human leukocyte antigen (HLA) antibodies (DSAs) and the use of intragraft deposition of the degradation product of complement component 4 (C4d) as a histologic feature of antibody-mediated rejection (AMR) (2–4). The current Banff diagnostic criteria for acute AMR include (1) C4d deposits in peritubular capillaries; (2) circulating DSAs; and (3) morphologic evidence for acute tissue injury (5). Acute allograft dysfunction distinguishes clinical AMR from the subclinical form, and the presence of chronic tissue injury (e.g., glomerular double contours) in patients with circulating DSAs and intragraft C4d are the criteria for the diagnosis of chronic active AMR. Time from transplantation to clinical manifestation of AMR has been reported to be associated with responsiveness to antirejection therapy and graft outcome (6), and histologic features such as intragraft C4d deposition, neutrophilic glomerulitis, and monocyte-macrophage infiltration and immunologic characteristics such as the presence or persistence of DSAs are reported to be harbingers of graft failure after an episode of AMR (7–11).
The primary objective of the current investigation was to identify risk factors for graft failure in kidney graft recipients with clinically indicated biopsy results displaying both C4d and morphologic features of acute AMR. We examined whether concurrent histologic features such as acute T-cell mediated rejection (TCMR), chronic active AMR, and/or interstitial fibrosis–tubular atrophy (IF/TA) are associated with graft survival. We also determined whether time from transplantation to biopsy and allograft function (estimated glomerular filtration rate [eGFR] and proteinuria) are related to graft survival.
Characteristics of the Study Cohort at the Time of Kidney Transplantation
We reviewed the results of 1120 clinically indicated biopsies from 833 kidney allograft recipients and identified 87 biopsy specimens from 87 patients that were positive for C4d immunofluorescence staining and displayed morphologic evidence of acute tissue injury consistent with acute AMR. The 1120 biopsies were performed at our center between December 2003 and February 2011, and all biopsy specimens were stained for C4d.
The demographic, pretransplant, and transplant characteristics including donor type and induction and maintenance immunosuppression used in the 87 patients are summarized in Table 1.
Characteristics of the Study Cohort at the Time of Clinically Indicated Allograft Biopsy
The clinical indications for performing the biopsy were acute graft dysfunction in 69 (79%) patients, delayed graft function (DGF) in 7 (8%), and significant proteinuria (defined as >0.5 g/day) or chronic graft dysfunction (eGFR <40 mL/min/1.73 m2) in 11 (13%) patients. The median time from transplantation to biopsy was 3.7 months, and the time from transplantation to biopsy was longer than 1 year in 35 kidney transplant recipients. At the time of diagnostic biopsy, the mean (±SD) eGFR was 23.2 ± 15.1 mL/min per 1.73 m2, and proteinuria was more than 1 g per day in 48 of 87 patients.
All 87 biopsy specimens showed C4d+ acute AMR, and among them, 26 showed acute AMR only, 43 concurrent chronic active AMR (29 acute plus chronic active AMR and 14 acute and chronic active AMR and acute TCMR), and 32 concurrent acute TCMR (18 acute AMR and concurrent acute TCMR and 14 acute and chronic active AMR and acute TCMR). IF/TA was present in 37 of the 87 biopsy specimens. The histologic features of the allograft biopsies, classified and graded using the Banff ’09 schema, are summarized in Table 2.
Results of circulating DSAs were available in 66 of 87 patients; 60 of the 66 patients were classified as DSA-positive using mean fluorescence intensity (MFI) higher than 500 as the threshold, using microbeads coated with purified single class I or II HLA antigens.
Eighty-one (93%) of the 87 patients received standard antirejection therapy consisting of methylprednisolone, intravenous immunoglobulin (IVIg), or plasmapheresis (PP), and six patients did not receive antirejection treatment for reasons such as extensive IF/TA in their biopsy results. In the 81 patients treated, the components of the antirejection therapy reflected evolution over time and included the addition of antithymocyte globulin (n=39 patients), rituximab (n=25 patients), and bortezomib (n=15 patients) to the standard therapy (Table 2). Reversal of a rejection episode, as defined by the return of the eGFR to within 15% of the baseline within 8 weeks after the initiation of antirejection treatment, occurred in 41 (47%) patients. The median follow-up after biopsy was 28 months at the time of preparation of this report. Thirty-seven (43%) patients developed allograft failure, defined as a persistent decline in the eGFR to less than 15 mL/min per 1.73 m2 or return to dialysis, during the study period; the median time from the biopsy to allograft failure was 4 months.
Kaplan-Meier Graft Survival Analysis
The impact of concurrent biopsy findings, the time from transplantation to biopsy (<12 vs. >12 months), the presence or absence of proteinuria greater than 1 g per day, and the type of antirejection therapy were examined with the use of the Kaplan-Meier curve analysis.
Figure 1 illustrates that concurrent acute TCMR (P=0.001, log-rank test) and concurrent chronic active AMR (P=0.037), but not concurrent IF/TA (P=0.46), are associated with graft survival time in patients with biopsies showing C4d+ acute AMR. Figure 1 also illuminates that late rejection (biopsy >12 months) is associated with a shorter graft survival time (P=0.04) compared with early rejection (biopsy <12 months). The presence or absence of proteinuria (P=0.09) and the addition of antithymocyte globulin (P=0.27), rituximab (P=0.70), or bortezomib (P=0.96) to the standard antirejection therapy were not associated with allograft survival time. We did not analyze the impact of the presence or absence of DSAs, because 60 of 66 patients in whom DSA information was available were positive for DSAs.
Twenty-five grafts failed during the 12 months after the for-cause biopsy, and the 1-year actual graft survival postbiopsy was 70% in the 87 patients with C4d+ biopsy results. Among the 25 graft failures, nine failed in the 55 patients without concurrent acute TCMR, and the actual 1-year graft survival in this subgroup was 83%. Sixteen grafts failed in the 32 patients with concurrent acute TCMR, and the 1-year actual graft survival in this subset was 46%.
Because concurrent acute TCMR, concurrent chronic active AMR, and time to biopsy were associated with graft survival, we examined whether the renal allograft biopsy results were related to the time from transplantation to clinically indicated biopsies. This bivariate analysis showed that the proportion of biopsy results with acute TCMR were not associated with time to biopsy (P=0.37, the Fisher exact test) (Table 3). As expected, chronic active AMR (P<0.0001) and IF/TA (P<0.0001) were overrepresented in late (>12 months) compared with early (<12 months) biopsies.
Cox Proportional Hazards Multivariable Analysis
We used the Cox proportional hazards multivariable analysis, a statistical tool for determining the relative contributions of different causes to a single event, to identify independent predictors for graft failure. To adjust for confounding, we decided to use all of the histologic and treatment variables at the time of diagnosis, a priori, in the Cox model, irrespective of their significance level by bivariate analyses. The six recipients who did not receive antirejection therapy owing to clinical and histologic considerations were excluded from the prediction model, and the remaining 81 recipients were included in the analysis.
The multivariable analysis using the Cox proportional hazards regression model showed that concurrent acute TCMR (hazard ratio [HR], 2.59; 95% confidence interval [CI], 1.21–5.55; P=0.01) remained an independent predictor of graft failure (Table 4). The analysis identified that the eGFR at the time of rejection (HR, 0.65 [95% CI, 0.48–0.88]; P=0.01) is an independent risk factor and that neither concurrent chronic active AMR nor time to biopsy is significantly associated with graft survival.
HLA-specific DSA information was available in 66 of 87 patients, and 60 of the 66 recipients had circulating DSAs. We did not test for non-HLA antibodies in our study subjects. Of the 60 recipients who had positive circulating DSAs, 55 had complete data such as the number of donor HLA antigens recognized by DSA, the sum of MFI of all DSAs detected, and the MFI of immunodominant DSAs. We examined the effect of each of the three DSA characteristics in the multivariable analysis and found that none of the DSA characteristics are independently associated with graft failure. In the analysis restricted to the same 55 patients, concurrent acute TCMR (P=0.03, 0.02, and 0.04, respectively) and the eGFR (P=0.002, 0.002, and 0.002, respectively) remained significantly associated with graft failure.
AMR generally carries a poorer prognosis compared with acute TCMR, and it has been reported that kidney graft survival is shorter in those with C4d+ AMR compared with C4d− AMR (7). We have identified that concurrent acute TCMR is an independent risk factor for graft failure in kidney graft recipients with biopsies positive for C4d and acute AMR.
Among the 1120 biopsies from 833 kidney transplant recipients, 87 biopsies from 87 patients were positive for C4d and acute AMR changes. The biopsy results showed a 7.8% incidence of C4d+ acute AMR in our study. The incidence of C4d+ with acute AMR and concurrent acute TCMR among all 1120 biopsies was 2.9%, and 37% of the 87 C4d+ biopsy results with identified acute AMR features displayed features of concurrent acute TCMR. Importantly, the majority of graft failures, 16 (64%) of the 25 failures in the first year postbiopsy, occurred in the subset with concurrent acute TCMR. An interesting biologic question is the sequence of events in patients manifesting both humoral immunity and T-cell immunity and whether acute TCMR preceded C4d+ AMR. The biologic question of what came first (C4d+ acute AMR vs. acute TCMR) cannot be resolved with this investigation, because we did not analyze results from sequential biopsies. As T-cell help is needed for IgG antibody production, a cascade of events with T-cell priming preceding a C4d+ acute AMR is biologically plausible. However, the majority (63%) of the biopsy specimens showing C4d+ AMR did not display concurrent TCMR. We therefore speculate that acute TCMR is not an essential prerequisite for the development of C4d+ AMR.
Kidney Allograft Biopsy Findings and Graft Survival Time
Feucht et al. made the seminal observations that vascular deposition of complement fragments C4d and C3d is a histologic feature of kidney graft biopsy specimens displaying cell-mediated rejection and that capillary deposition of C4d is associated with early graft dysfunction and early graft loss after deceased-donor kidney transplantation (3, 7). In the study by Feucht and colleagues (7), the 1-year graft survival was 57% to 63% in those with C4d+ biopsy results and 90% in those without C4 staining of capillaries. Gaston et al. (8) also reported inferior graft survival rates in 40 C4d+DSA+ patients compared with 74 C4d−DSA− patients (HR, 4.58; 95% CI, 1.75–12.00; P=0.002). Neutrophilic glomerulitis, peritubular capillary dilatation with neutrophilic infiltrates, and interstitial edema at the time of first biopsy have also been reported as determinants of graft failure in AMR patients (9, 11). In our study, concurrent chronic active AMR was related to graft survival by the bivariate Kaplan-Meier curve analysis but was not significant by multivariable analysis. The presence of IF/TA was also not associated with graft survival time. The median follow-up in our study, however, was 2 years, and this short duration might have been insufficient to witness the adverse impact previously observed with time to biopsy or chronic changes (6, 12, 13).
Potential Mechanisms for the Differential Kidney Graft Outcomes in Those With Acute AMR Versus Acute AMR and Concurrent Acute TCMR
In our study, 25 grafts failed within 1 year of biopsy, and the overall 1-year actual graft survival postbiopsy was 70% in the 87 patients with C4d+ biopsy results that displayed acute AMR features. The 1-year actual graft survival was 83% in the 55 patients without concurrent acute TCMR and 46% in the 32 patients with concurrent acute TCMR. The plausible mechanisms for the heightened risk associated with concurrent acute TCMR include (1) the participation of memory T cells that are relatively resistant to conventional anti–T-cell therapy; (2) the contribution of multiple cell types including macrophages-monocytes to the “cellular” component of cell-mediated rejection; (3) the possibility of the antibody type (e.g., IgG subtype or complement-fixing capacity) being different with active T-cell help compared to when T cells are absent; and (4) the possible contribution of an antibody-dependent cell-mediated cytotoxicity effector mechanism to the graft injury process (because Fc receptor–bearing cells and antibodies are both present in the graft). The contribution of concurrent chronic changes to the adverse impact of concurrent acute TCMR cannot be excluded completely but appears unlikely in view of the data shown in Table 4. Although the precise biologic basis for our novel observation remains unresolved, the findings advance the idea that anti–T-cell therapy is an integral component of patients with biopsy specimens displaying both humoral (AMR) and cellular (TCMR) immunity. In this regard, currently prescribed immunosuppressive drugs such as calcineurin inhibitors are highly active against T-cell–mediated immune responses. The occurrence of concurrent TCMR in our patients may reflect underimmunosuppression, owing to clinical circumstances (e.g., drug toxicity) or nonadherence, or both. In view of the lack of robust parameters to define the immune status of the organ graft recipients, we are unable to address the important issue of whether the occurrence of TCMR in our patients reflected the potentially remediable problem of underimmunosuppression.
Donor-Specific Anti-HLA Antibodies and Graft Outcome
The presence of DSAs and their persistence after treatment have both been associated with graft loss after AMR (6, 9–11, 14–17). HLA-specific DSA information was available in 66 of 87 patients, and 60 of the 66 recipients had circulating DSAs. We did not test for non-HLA antibodies in our study subjects. The almost universal presence of DSAs in our study cohort precluded an analysis regarding the significance of the presence or absence of DSAs on graft outcome, and we restricted our analysis to investigating the impact of DSA characteristics (number of donor HLA antigens recognized, immunodominant DSAs, and MFI) rather than the presence or absence on graft outcome. Our focused analysis found no relationship between DSA characteristics and graft survival time. A limitation of our study is that DSA data were available only in 66 of 87 patients studied, and we did not have longitudinal information. These shortcomings may have contributed to underappreciation of the role of DSA characteristics in those with C4d+ acute AMR.
Time From Transplantation to Biopsy Diagnosis of C4d+ Acute Antibody-Mediated Rejection
Time of onset of acute rejection has been reported to be a determinant of graft outcome with late acute rejection being more recalcitrant to antirejection therapy compared with early-onset acute rejection (6, 9, 18,19). In the present study, time from transplantation to biopsy was indeed significantly associated with graft survival by bivariate analysis but lost its significance after multivariable analysis. As expected, our study did find that chronic changes are more common in late compared with early biopsies (Table 3). In this regard, the time-related late features such as chronic active AMR and IF/TA were not independent predictors of graft survival time in our study, whereas acute TCMR that was not related to time to biopsy heightened the risk of graft failure.
Treatment of Antibody-Mediated Rejection
Anti-CD20 monoclonal antibodies target B cells; plasma cells are targets of the proteasome inhibitor bortezomib, and antithymocyte globulin can target multiple cell types including T, B, and plasma cells. All three have been used for the treatment of acute AMR, particularly in refractory cases (11, 20–22), and emerging data suggest an improved treatment response with bortezomib compared with rituximab (21, 23). Contrary to our expectations, we did not find an improvement in graft survival with the adjunct use of rituximab, bortezomib, or antithymocyte globulin to the antirejection therapy with corticosteroids, plasmapheresis, or IVIg. Our retrospective study, however, was neither designed nor powered to fully evaluate the rapidly evolving therapeutic options for acute AMR.
In summary, after adjusting for potential confounders in the 87 kidney transplant recipients with biopsy specimens that displayed C4d+ acute AMR, we show that concurrent acute TCMR and reduced eGFR at the time of biopsy are associated with inferior graft survival in this high-risk population.
MATERIALS AND METHODS
Kidney Transplant Patients and Biopsies
We reviewed 1120 clinically indicated biopsy results from 833 kidney allograft recipients and identified 87 biopsy specimens from 87 patients that were positive for C4d immunofluorescence staining and displayed morphologic evidence of acute tissue injury consistent with acute AMR. The biopsies were performed at our center between December 2003 and February 2011. ABO-incompatible kidney transplants were excluded.
The biopsy specimens were classified using the Banff ’09 update of the Banff ’97 classification (5); the Banff-specified morphologic features and the grades are summarized in Table 2. The biopsy results were read by a single pathologist (S.V.S). All biopsy specimens were stained for C4d, and C4d staining was performed on cryosections using a monoclonal anti-C4d antibody (Quidel, Santa Clara, CA). C4d was considered positive if peritubular capillary staining involved more than 50% (diffuse) or 10% to 50% (focal) of the cortical or medullary area (5).
We collected demographic, clinical, biochemical, pathology, treatment, and follow-up data. Delayed graft function was defined as the need for dialysis treatment within the first week of transplantation. We used the Modification of Diet in Renal Disease (MDRD) equation to estimate the GFR (eGFR) (24). The baseline eGFR was the average of the highest three measures of eGFR within 3 months before the diagnosis of acute AMR.
The Institutional Review Board of the Weill Cornell Medical College approved this study.
Pretransplant panel-reactive antibodies (PRA) were determined using complement-dependent cytotoxicity assay. Anti-HLA antibodies were detected using microparticles with individual purified HLA antigens covalently bound as targets (One Lambda Inc, Canoga Park, CA) on the Luminex platform (Luminex Corp, Austin, TX). We analyzed anti-HLA antibodies against donor HLA-A, HLA-B, HLA-DR, and HLA-DQ antigens. Transplant recipients with at least one DSA with MFI higher than 500 were considered positive for circulating DSA. Among the DSA-positive recipients, immunodominant DSA was the antibody with the highest MFI. Circulating antibodies to non-HLA antigens were not tested. Information on DSA was available in 66 of 87 transplant recipients studied.
Treatment and Outcome
The patients received standard antirejection therapy composed of methylprednisolone, IVIg, or PP. In addition to the standard therapy, subsets of patients were also treated with various combinations of antithymocyte globulin, rituximab, and bortezomib. Acute AMR was classified as resistant if the eGFR did not return to within 15% of the baseline within 8 weeks after the initiation of antirejection treatment. The primary outcome was graft loss defined as a persistent decline in the eGFR to less than 15 mL/min per 1.73 m2 or return to dialysis. Patients who were lost to follow-up, died with a functioning graft, or did not reach the outcome were censored at their last follow-up.
We summarized the continuous variables as mean (SD) for variables with Gaussian distribution and median (range) for variables with non-Gaussian distribution and the categorical variables as proportions.
We generated Kaplan-Meier survival curves to estimate the time to allograft loss. Survival curves were compared by Mantel-Cox log-rank test. We used the Cox proportional hazards regression model to test for the independent contribution of each of the variable at the time of diagnosis of C4d+ acute AMR in predicting subsequent graft loss. To adjust for confounders, we decided to use the following variables at the time of diagnosis of rejection, a priori, in the Cox model, irrespective of their significance level by bivariate analyses: (1) early versus late rejection; (2) eGFR; (3) proteinuria; (4) presence of the following concurrent features on biopsy: acute TCMR, chronic active AMR, and IF/TA; and (5) use of antithymocyte globulin, rituximab, or bortezomib in addition to the standard treatment of steroids, IVIg, or PP. We expressed the results of Cox regression as relative hazard with 95% CIs. We used Stata/IC 10.1 for Windows (StataCorp, College Station, TX) and GraphPad Prism 5.02 for Windows (GraphPad Software, La Jolla, CA) for all statistical analyses. A two-sided P<0.05 was considered statistically significant.
The authors thank their colleagues at the Division of Transplant Surgery, New York Presbyterian Hospital–Weill Cornell Medical Center and at The Rogosin Institute for their kind help in conducting this study. The research reported in this article is in partial fulfillment of Dr. Thangamani Muthukumar’s Clinical and Translational Science Center’s Graduate Program (K30) in Clinical & Translational Investigation.
1. Gailiunas P, Suthanthiran M, Busch GJ, et al.. Role of humoral presensitization in human renal transplant rejection. Kidney Int
1980; 17: 638.
2. Archdeacon P, Chan M, Neuland C, et al.. Summary of FDA antibody-mediated rejection workshop. Am J Transplant
2011; 11: 896.
3. Feucht HE, Felbert F, Gokel MJ, et al.. Vascular deposition of complement-split products in kidney allografts with cell-mediated rejection. Clin Exp Immunol
1991; 86: 464.
4. Terasaki PI. A personal perspective: 100-year history of the humoral theory of transplantation. Transplantation
2012; 93: 751.
5. Sis B, Mengel M, Haas M, et al.. Banff ’09 meeting report: antibody-mediated graft deterioration and implementation of Banff working groups. Am J Transplant
2010; 10: 464.
6. Walsh RC, Brailey P, Girnita A, et al.. Early and late acute antibody-mediated rejection
differ immunologically and in response to proteasome inhibition. Transplantation
2011; 91: 1218.
7. Feucht HE, Schneeberger H, Hillebrand G, et al.. Capillary deposition of C4d complement fragment and early renal graft loss. Kidney Int
1993; 43: 1333.
8. Gaston RS, Cecka JM, Kasiske BL, et al.. Evidence for antibody-mediated injury as a major determinant of late kidney allograft failure. Transplantation
2010; 90: 68.
9. Lefaucheur C, Nochy D, Hill GS, et al.. Determinants of poor graft outcome in patients with antibody-mediated acute rejection. Am J Transplant
2007; 7: 832.
10. Everly MJ, Everly JJ, Arend LJ, et al.. Reducing de novo donor-specific antibody levels during acute rejection diminishes renal allograft loss. Am J Transplant
2009; 9: 1063.
11. Lefaucheur C, Nochy D, Andrade J, et al.. Comparison of combination plasmapheresis/IVIg/anti-CD20 versus high-dose IVIg in the treatment of antibody-mediated rejection. Am J Transplant
2009; 9: 1099.
12. Nankivell BJ, Fenton-Lee CA, Kuypers DR, et al.. Effect of histological damage on long-term kidney transplant outcome. Transplantation
2001; 71: 515.
13. Rush DN, Cockfield SM, Nickerson PW, et al.. Factors associated with progression of interstitial fibrosis in renal transplant patients receiving tacrolimus and mycophenolate mofetil. Transplantation
2009; 88: 897.
14. Lefaucheur C, Loupy A, Hill GS, et al.. Preexisting donor-specific HLA antibodies predict outcome in kidney transplantation. J Am Soc Nephrol
2010; 21: 1398.
15. Everly MJ, Rebellato LM, Ozawa M, et al.. Beyond histology: lowering human leukocyte antigen antibody to improve renal allograft survival in acute rejection. Transplantation
2010; 89: 962.
16. Hidalgo LG, Campbell PM, Sis B, et al.. De novo donor-specific antibody at the time of kidney transplant biopsy associates with microvascular pathology and late graft failure. Am J Transplant
2009; 9: 2532.
17. Cooper JE, Gralla J, Cagle L, et al.. Inferior kidney allograft outcomes in patients with de novo donor-specific antibodies are due to acute rejection episodes. Transplantation
2011; 91: 1103.
18. Tejani AH, Stablein DM, Sullivan EK, et al.. The impact of donor source, recipient age, pre-operative immunotherapy and induction therapy on early and late acute rejections in children: a report of the North American Pediatric Renal Transplant Cooperative Study (NAPRTCS). Pediatr Transplant
1998; 2: 318.
19. Sijpkens YW, Doxiadis II, Mallat MJ, et al.. Early versus late acute rejection episodes in renal transplantation
2003; 75: 204.
20. Faguer S, Kamar N, Guilbeaud-Frugier C, et al.. Rituximab therapy for acute humoral rejection after kidney transplantation. Transplantation
2007; 83: 1277.
21. Everly MJ, Everly JJ, Susskind B, et al.. Bortezomib provides effective therapy for antibody- and cell-mediated acute rejection. Transplantation
2008; 86: 1754.
22. Walsh RC, Everly JJ, Brailey P, et al.. Proteasome inhibitor–based primary therapy for antibody-mediated renal allograft rejection. Transplantation
2010; 89: 277.
23. Waiser J, Budde K, Schütz M, et al.. Comparison between bortezomib and rituximab in the treatment of antibody-mediated renal allograft rejection. Nephrol Dial Transplant
2012; 27: 1246.
24. Levey AS, Stevens LA, Schmid CH, et al.. A new equation to estimate glomerular filtration rate. Ann Intern Med
2009; 150: 604.