Complement Genes Strongly Predict Recurrence and Graft Outcome in Adult Renal Transplant Recipients with Atypical Hemolytic and Uremic Syndrome

Authors


Corresponding authors: Moglie Le Quintrec

m.lequintrec@hopital-foch.org; and Véronique Frémeaux-Bacchi

veronique.fremeaux-bacchi@egp.aphp.fr

Abstract

Atypical hemolytic and uremic syndrome (aHUS) is a severe disease strongly associated with genetic abnormalities in the complement alternative pathway. In renal posttransplantation, few data are available on recurrence risk and graft outcome according to genetic background in aHUS patients. The aim of this study was to identify risk factors for recurrence and transplant outcome and, in particular, the role of complement gene abnormalities. We retrospectively studied 57 aHUS patients who had received 71 renal transplants. A mutation in complement gene was identified in 39 (68%), in factor H (CFH), factor I (CFI), membrane cofactor-protein (MCP), C3 and factor B (CFB). At 5 years, death-censored graft survival was 51%. Disease recurrence was associated with graft loss (p = 0.001). Mutations in complement genes were associated with higher risk of recurrence (p = 0.009). Patients with CFH or gain of function (C3, CFB) mutations had a highest risk of recurrence. M-TOR inhibitor was associated with significant risk of recurrence (p = 0.043) but not calcineurin inhibitor immunosuppressive treatment (p = 0.29). Preemptive plasmatherapy was associated with a trend to decrease recurrence (p = 0.07). Our study highlights that characterization of complement genetic abnormalities predicts the risk of recurrence-related graft loss and paves the way for future genetically based individualized prophylactic therapeutic strategies.

Abbreviations
aHUS

atypical Hemolityc and uremic syndrome

STEC

shiga toxine-producing Escherichia Coli

ESRF

end stage renal failure

CAP

complement alternative pathways

Introduction

Hemolytic uremic syndrome (HUS) is a microvascular occlusive disorder characterized by hemolytic anemia with fragmented erythrocytes, thrombocytopenia and acute renal failure [1]. Unlike typical HUS, atypical HUS (aHUS) is not related to Shiga toxin-producing Escherichia coli (STEC) infection and affects both children and adults. aHUS is associated with a poor functional outcome, as up to 50% of patients with aHUS develop end-stage renal disease (ESRD) with a subsequent high risk of recurrence [2-4] after renal transplantation. Over the last decade, a close link between aHUS and uncontrolled activation of the complement alternative pathway (CAP) has been well established [5, 6]: genetically determined or acquired dysregulation of CAP has been identified in roughly 70% of aHUS patients [3, 6]. Loss or gain of function mutations have been described in genes encoding for complement regulatory proteins complement factor H (CFH), factor I (CFI), membrane cofactor-protein (MCP) as in those encoding the two components of the C3 convertase, factor B (CFB) or C3 [7-16]. In addition, anti-CFH antibodies have been identified in 10% of aHUS patients [17-19]. Moreover, genetic variants, including single nucleotide polymorphisms, and haplotypes in the genes encoding for CFH and MCP may also contribute to aHUS development ([8], [12], [20-22]). In particular, CFH haplotypes defined with four exonic polymorphisms located in SCR1 (Val62), SCR7 (Tyr402), SCR11 (Gln672) and SCR16 (Asp 936), known as the at-risk CFH gtgt haplotype, is strongly associated with aHUS [22]. The risk of posttransplant aHUS recurrence, which has been reported as ranging from less than 20% to more than 90%, critically depends on whether the recipient carries a mutation in a gene encoding for a membrane protein membrane cofactor protein (MCP) [2-4] or a circulating factor (CFH, CFI, CFB, C3). Moreover, in posttransplant settings, environmental factors, including acute rejection, immunosuppressive drugs, viral infections and ischemia reperfusion lesions, may initiate endothelial damage [23-26] but their impact on aHUS recurrence has not been investigated. The aim of this retrospective study was to assess the risk of recurrence and graft loss in renal transplanted aHUS patients. Complement genetics and posttransplant characteristics, including acute rejection, immunosuppressive regimen, delayed graft function, aHUS recurrence and plasmatherapy treatment were considered in the analysis.

Material and Methods

Study design

This was a retrospective, multicenter study of renal transplant recipients with aHUS-related ESRD. Sixty-six patients included in the French aHUS registry from 21 French transplantation centers were considered for inclusion. Renal transplant centers were contacted to date and update clinical data regarding graft outcome. Criteria for inclusion in the study were: (1) Patients with adult-onset (=18 years) aHUS who had received a renal transplant between January 1995 and December 2009 and (2) Patients who had undergone a genetic investigation of the complement alternative pathway. Exclusion criteria were: (1) HUS with an identified cause such as STEC-induced HUS, Thrombotic Thrombopenic Purpura (TTP) associated with ADAMST 13 deficiency, malignancy or autoimmune-related HUS. (2) Patients without updated information. (3) Patients who were transplanted before 1995. The last follow-up time point was June 2010.

Complement assays and genetic screening

Complement assessment was performed on EDTA plasma samples at the immunology laboratory of the Georges Pompidou European Hospital. C3, C4 and CFB levels were measured by nephelometry (Dade Behring, Deerfield, IL, USA). CFH and CFI levels were measured by ELISA. Membrane expression of CD46 was analyzed using flow cytometry and anti-CFH antibodies were screened by ELISA as previously reported [27]. Complete exon sequencing of CFH, MCP CFI, C3, CFB, THMD genes was undertaken using direct sequencing analysis as described [27]. Screening for nonallelic homologous recombination (NAHR) between CFH and CFHR1 namely “hybrid gene” was performed using multiplex ligation-dependent probe amplification (MLPA; SALSA MLPA kit P236-A1 ARMD from MRC Holland and probes designed by the laboratory; Ref. [28]. Functional evaluation of FH was performed with hemolytic assay using sheep erythrocytes [27]. Complete CFHR1/CFHR3 deletion was identified by analysis of the SNP rs7542235 [29]. The aHUS at-risk CFH gtgt haplotype namely “at risk CFH haplotype” was tagged by genotyping the SNPs rs800292 (c.184G), rs1061170 (c.1204T); rs3753396 (c.2016G) and rs1065489 (c.2808T) as previously reported [22]. All patients had provided informed written consent for gene screening.

Recurrence and graft survival

Atypical HUS recurrence or acute rejections were diagnosed by renal graft biopsy. Diagnosis of posttransplant recurrence was assessed by histological criteria of thrombotic microangiopathy (TMA), including arterial and glomerular lesions [30]. Biopsy-proven acute rejections (BPAR) were graded using the Banff classification as noted at the time of renal biopsy. C4d staining was routinely performed on biopsies from 2005 to 2009. Graft survival was defined as the time between the date of transplantation and the date of return to dialysis. Death with functioning graft was used as the censor for the graft survival study.

The impact of curative plasmatherapy on graft survival was analyzed. Curative plasmatherapy for recurrence was defined by a minimum of five fresh frozen plasma infusions or by five plasma exchanges over 10 days. The impact of preemptive plasmatherapy on disease recurrence and graft survival was also assessed.

Clinical data

The following clinical data were obtained from medical records: pretransplant characteristics, the age at the time of aHUS onset, ESRD, renal transplantation and posttransplant recurrence. Meaningful independent variables for all kidney transplants were recorded: deceased or living donor kidney, donor age, number of HLA mismatches, cold ischemia time, induction therapy (anti-thymoglobulin [ATG] or interleukin 2 receptor alpha anti-IL2Ra]), calcineurin inhibitor (CNI) or mTOR-based maintenance regimen, plasmatherapy regimen, delayed graft function and number of acute rejection episodes.

Statistics

Continuous variables are expressed as mean values with standard deviations. Categorical variables are expressed as tallies and percentages. Statistical analysis was performed on 71 grafts. Kaplan–Meier graft survival estimates were calculated for death-censored graft survival. Death-censored graft survival curves were compared between patients with and without HUS recurrence. Death-censored graft survival without HUS recurrence curves were compared between grafts from patients with and without genetic mutation. Graft survival without HUS recurrence curves for patients with and without individual mutations were then compared. The Bonferronni correction for multiple comparisons was applied. Cox model univariate analysis was used to test the association between independent variables. Variables with p values of less than 0.20 in the univariate analysis were entered into multivariate analysis through a forward subset selection with no switching. Risk ratios (RRs) are shown with their upper and lower 95% confidence limits. For all statistical measures, p < 0.05 was considered significant. Statistical analysis was performed using the NCSS software (Hintze, J. 2012. NCSS 8. NCSS, LLC. Kaysville, UT, USA)

Results

Characteristics of the study population

Out of the initial 66 patients considered for inclusion, 57 patients were finally included in the study. Nine patients were not included as they had been transplanted before 1995. The 57 patients received a total of 71 renal transplants: 43 patients had one graft whereas 14 had two grafts. No patients were lost to follow-up. Forty-eight (85%) of the 57 patients were female. Nine (16%) and 48 (84%) patients had a familial or sporadic form of aHUS, respectively. Three grafts were from living related donors and the others were from deceased donors. There were no renal transplantations from donors after cardiac death. The median age at the time of aHUS onset, ESRD and kidney transplantation was 32, 34 and 38 years, respectively. Characteristics of the grafts and immunosuppressive therapies are summarized in Table 1.

Table 1. Characteristics of the grafts and immunosuppressive therapies
GraftsN = 71
% of deceased donor (n/total)95 (68/71)
Rank of renal transplantation >1 (%), (n/total)25 (18/71)
Median donor age (range) [years]44 (15–73)
Mean HLA mismatch number ± SD3.2 ± 1.2
Mean HLA DR mismatch number ± SD1.2 ± 0.8
Median cold ischemia time (range) [hours]20 (4–39)
% of induction therapy (n/total)98 (66/67)
 % Anti-thymoglobulin antibodies, (n/total)75 (50/66)
 % Anti-IL2Ra monoclonal antibody, (n/total)25 (16/66)
Maintenance immunosuppressive therapy 
% CNI-based regimen (n/total)84 (57/68)
 % of tacrolimus-based regimen (n/total)77 (44/57)
 % of cyclosporine-based regimen (n/total)23 (13/57)
 % of mTOR inhibitor-based regimen (n/total)16 (11/68)
 % of delayed graft function (n/total)14 (9/66)
% of preemptive plasmatherapy (n/total)13 (9/70)

Complement investigations

Eighteen patients (33%) displayed a low C3 plasma level (<660 mg/L). Thorough genetic and C3 level assessments were done either before (n = 29) or after (n = 28) renal transplantation. A mutation in complement factors was identified in 39 of the 57 patients (68%; Table 2, Table S1, Figure S1). Forty-six percent (18/39) of these patients carried a CFH mutation identified by direct sequencing analysis of Factor H gene (12/18) or nonallelic homologous recombination (NAHR) (6/18) identified by MLPA. In seven of the 18 patients (39%) CFH/NAHR mutations were associated with low-plasma CFH levels (type I mutation), and 11 (61%) with normal FH levels (type II mutation). All mutations were heterozygous. The frequency of CFI, MCP, C3 and CFB mutations was 23% (n = 9), 8% (n = 3), 10% (n = 4) and 3% (n = 1), respectively. Four of the nine patients with CFI mutations had homozygous CFHR1 deletion. Four patients (7%) had combined mutations. In 18 patients, no mutation was identified in any of the screened genes; six of them carried two at risk CFH haplotypes.

Table 2. Frequency of complement abnormalities in the cohort
 Patients (n = 57)Grafts (n = 71)
Genetic/acquired abnormalityn%n
  1. THBD = Thrombomodulin; w = with; w/o = without.

  2. Combined mutations were (1) in 2 familial cases: CFH/MCP mutations in one and an additional CFI in the other, (2) in 2 sporadic cases: CFI/CFB in one and MCP/CFI in the other.

Mutation396848
CFH1831.622
CFI916.011
MCP35.53
C347.05
CFB11.82
THBD000
Combined mutation47.05
No mutation183223
w/o.at-risk CFH haplotype1221.116
w two at-risk CFH haplotype610.57
Anti-CFH antibodies000

Patient and graft survival

Four patients (7%), aged from 21 to 50 years, died following cardio-vascular events (acute cardiac failure [n = 1], myocardial infarction [n = 2] and extensive stroke [n = 1]. Three of them died at 17 months, 2 and 3 years posttransplant with a functioning graft and one patient died of myocardial infarction 5 years after he had returned to dialysis.

One and five-year post transplantation death-censored graft survival were 76% and 51%, respectively. The median graft survival time (lower and upper 95% CL) was 61 (32–87) months (Figure 1). Graft survival and recurrence were not different for first and second grafts (Tables 3 and 4). Death-censored graft survival was significantly worse in patients with recurrent aHUS than in those without, in univariate (RR = 3.79 [1.74–8.28]; p = 0.0001) and multivariate analysis RR = 4.86 (1.30–13.81); p = 0.001 (Figure 2, Table 3). The mean time between recurrence and graft lost was 185 days (58–656). Graft survival was 44 ± 7% and 74 ± 5% at 1 year and 36 ± 7% and 70 ± 8% at 5 years in patients with or without recurrence, respectively. The disease recurrence occurred during the early posttransplant course (Figure 3).

Table 3. Variables associated with significant graft loss risk by univariate and multivariate Cox analysis (n = 71 grafts)
 Univariate analysisMultivariate analysis
Independent variableRR (CI)pRR (CI)p
  1. ATG = anti-thymoglobulin antibodies; BPAR = biopsy proven acute rejection; CNI = calcineurin inhibitors; MM = mismatch; mTOR = mTor inhibitors. Rank of Tx = rank of transplantation.

  2. 1Since patients are eliminated listwise when one missing value is present, the MM HLA DR, which had a significant number of missing values was dropped from the analysis.

Recipient age0.99 (0.97–1.03)0.776  
Donor age1.00 (0.98–1.02)0.718  
Living donor0.40 (0.05–2.91)0.365  
Cold ischemia time0.97 (0.93–1.02)0.286  
HLA MM1.00 (0.75–1.35)0.983  
HLADR MM10.70 (0.42–1.15)0.158  
Induction0.39 (0.05–2.90)0.358  
ATG1.46 (0.67–3.20)0.344  
CNI0.81 (0.36–1.85)0.624  
mTOR1.19 (0.53–2.69)0.676  
DGF2.05 (0.85–4.95)0.1114.77 (1.80–12.67)0.002
BPAR1.55 (0.81–2.94)0.1831.70 (0.83–3.49)0.149
Time on dialysis1.00 (0.99–1.00)0.648  
Rank of Tx0.88 (0.39–2.00)0.766  
HUS recurrence3.79 (1.74–8.28)0.00014.86 (1.30–13.81)0.001
Curative PE2.51 (1.34–4.68)0.0041.17 (0.49–2.83)0.724
Preemptive PE0.19 (0.03–1.40)0.1030.11 (0.01–0.85)0.035
Table 4. Variables associated with significant recurrence risk by univariate and multivariate Cox analysis (n = 71 grafts)
 Univariate analysisMultivariate analysis
Independent variableRR (CI)pRR (CI)p
  1. ATG = anti-thymoglobulin antibodies; BPAR = biopsy proven acute rejection; CNI = calcineurin inhibitors; MM = mismatch; mTOR = mTor inhibitors; PE = plasma exchange; Rank of Tx = rank of transplantation.

  2. 1Since patients are eliminated listwise when one missing value is present, the MM HLA DR, which had a significant number of missing values was dropped from the analysis.

Recipient age1.022 (0.99–1.05)0.1491.04 (1.00–1.07)0.031
Donor age1.01 (0.99–1.03)0.233  
Living donor0.85 (0.21–3.50)0.817  
HLA MM0.92 (0.70–1.20)0.532  
HLADR MM10.73 (0.47–1.14)0.156  
Induction0.65 (0.26–1.66)0.361  
ATG0.80 (0.48–1.580.525  
CNI0.67 (0.33–1.39)0.294  
mTOR2.00 (0.98–4.17)0.0652.21 (1.03–4.74)0.043
DGF1.382 (0.58–3.29)0.464  
BPAR1.15 (0.61–2.18)0.656  
Time on dialysis1.00 (0.99–1.06)0.948  
Rank of Tx1.20 (0.59–2.44)0.612  
Previous recurrence1.26 (0.58–2.72)0.551  
Low C31.99 (1.05–3.48)0.0351.61 (0.85–3.04)0.144
Mutation3.10 (1.43–6.68)0.0042.88 (1.30–6.37)0.009
Preemptive PE0.16 (0.02–1.15)0.0680.34 (0.10–1.13)0.078
Figure 1.

Death-censored graft survival. Kaplan–Meier analysis of overall death-censored graft survival in atypical HUS patients. The median graft survival time (lower and upper 95% CL) was 61 (2.4–87) months.

Figure 2.

Death-censored graft survival after renal transplantation according to the occurrence of recurrence. Kaplan–Meier analysis (p = 0.0001, log rank).

Figure 3.

Recurrence hazard rate over posttransplant time. The hazard rate is the derivative (slope) of the cumulative hazard function.

In univariate aHUS was the only factor significantly associated with an increased risk of graft loss (RR = 3.79 [1.7–8.27]; p = 0.0001) whereas in multivariate analysis, delayed graft function (RR = 4.77[1.81–12.67]; p = 0.002) and aHUS recurrence RR = 4.86 (1.30–13.81); p = 0.001 were factors independently associated with an increased risk of graft loss. In multivariate analysis, preemptive plasmatherapy reduced graft loss (RR = 0.11 [0.01–0.85]; p = 0.035; Table 3).

Acute rejection occurred in 30% of the transplanted kidneys (22/71) including 17 cellular rejections (ACR) and five antibody-mediated rejections (AMR). However, they were not associated with a significant increased risk of graft failure (RR = 1.55 [0.86–2.94]; p = 0.183; Table 3). Sixty per cent (14/22) of acute rejections coincided with aHUS recurrence. Statistical analyses performed per graft and per patient gave similar results (data not shown).

Risk factors for posttransplantation aHUS recurrence

To determine risk factors associated with aHUS recurrence after renal transplantation, we analyzed clinical and genetic variables (Table 4). Donor age, acute rejection and CNI-based immunosuppressive regimen did not represent significant risk factors for aHUS recurrence. In contrast, a low-C3 level (RR = 1.99 [1.05–3.48]; p = 0.035) and the presence of a mutation in genes encoding complement proteins in univariate analysis (RR = 3.10 [1.43–6.68]; p = 0.004) were associated with a greater risk of recurrence (Figure 4, Table 4). In multivariate analysis, the presence of mutation (RR = 2.88 [1.30–6.37]; p = 0.009), m-Tor inhibitor regimen (RR = 2.21 [1.03–4.74]; p = 0.043) and recipient age (RR = 1.04[1.00–1.07]; p = 0.031) were associated independently with an increased rate of aHUS recurrence. Of the 14 patients who received two grafts, 11 (78%) had a recurrence on the first graft and 11 (78%) on the second. Nine of the 11 patients (81%) who had recurrence on the first graft had a recurrence on the second graft.

Figure 4.

Graft survival without HUS recurrence according to presence or not of a mutation in complement genes (p = 0.035, log rank).

To further assess the risk of aHUS recurrence on graft, we compared graft survival (n = 71) without HUS recurrence curves according to the presence of mutations or two at risk CFH haplotypes (Figure 5, Table 5). Grafts in patients with neither mutations nor the two at-risk CFH haplotypes were considered as a control group (n = 16 grafts). Grafts in C3 or CFB carriers were pooled to analyze gain of function mutations (n = 7). Results of univariate cox are given after Bonferroni correction. Mutations in CFH/NAHR were associated with a significantly greater risk of recurrence as compared with the control group (RR = 5.6 [1.7– 9.6]; p = 0.0024), the seven grafts from the six patients who carried a Hybrid gene mutation experienced a recurrence in the first year after transplantation. Mutations in C3/CFB had a trend towards significance (RR = 4.80 [1.3–20.9]; p = 0.04; Figure 5, Table 5). No statistically significant difference for recurrence risk was found between carriers of CFI mutations, MCP mutations, combined mutations or two the at-risk CFH haplotypes without mutation compared to the control group (Table 5). Graft recurrence was not different between patients with CFI mutation associated with CFHR1 deletion and those with CFI mutation without CFHR1 deletion (p = 0.99; data not shown). Posttransplant recurrence occurred in two of the three patients with an isolated MCP mutation. Both of them had the two at-risk CFH haplotypes for aHUS. Two years after aHUS recurrence, 75%, 71%, 72% and 50% of grafts from carriers of CFH, CFI, C3 /B and MCP were lost.

Table 5. Individual mutations associated with significant recurrence risk by univariate Cox analysis
 RR95% CIp-Value
  1. C = control, no mutation, no polymorphism.

  2. At risk CFH: two at risk CFH haplotypes.

  3. All group were then compared to a control group (grafts from patients without mutations or two at-risk CFH haplotypes).

  4. Two at-risk CFH haplotypes without mutation, CFI, gain function mutation (C3/B) and CFH mutations, MCP and Combined mutations were compared to a Control group.

  5. Bonferroni correction for multiple comparisons was applied.

CFH vs. C5.62.4–13.00.0024
C3/CFB vs. C4.81.12–20.60.042
CFI vs. C2.90.9–9.80.444
At risk CFH vs. C2.80.6–13.20.810
MCP vs. C3.40.3–370.804
Combined mutations vs. C3.20.5–20.80.666
Figure 5.

Graft survival without HUS recurrence according to the recipient genetic background. Fractions of graft without aHUS recurrence at any time according the presence or not of mutations (CFH, C3/CFB, CFI, MCP, combined mutations or without mutation but with 2 at risk CFH haplotype). All group were then compared to a control group (grafts from patients with neither mutation nor homozygous at-risk CFH haplotype). The results are shown in Table 5.

Influence of plasmatherapy on graft survival or recurrence

Curative plasmatherapy, which consisted either of fresh-frozen plasma infusions (n = 3) or plasma exchanges (n = 30), was performed in 33 of the 44 patients (75%) with aHUS recurrence. Overall, graft outcome was poor in patients who received curative plasmatherapy for recurrence in univariate analysis RR = 2.51 [1.34–4.68] p = 0.004). After adjustment for recurrence, multivariate analysis showed that curative plasmatherapy did not improve graft survival (RR = 1.17 [0.49–2.83] p = 0.7; Table 3, Figure 6).

Figure 6.

Graft survival following aHUS recurrence in patients treated or not with curative plasma therapy (p = 0.9, log rank).

Nine patients received preemptive plasmatherapy in an attempt to prevent posttransplant recurrence of aHUS but none received the anti-C5 treatment eculizumab as a preemptive therapy (Table 6). In multivariate analysis, preemptive plasmatherapy decreased graft loss RR = 0.11 [0.01–0.84] p = 0.035). There was also a nonsignificant decrease in disease recurrence RR = 0.34 [0.10–1.13] p = 0.078) in patients who received preemptive plasmatherapy (Table 4, Figure 7). Four of these patients experienced an event-free successful renal transplantation. Recurrence occurred in three other patients (33%; including the one with CFH mutation and one with combined mutations [CFB and CFI]) with severe concomitant graft impairment rescued by eculizumab, a complete reversal of aHUS activity was obtained in all of them. No patients treated by eculizumab suffered graft loss but renal function sequelae persisted in all of them at 1 year posttransplantation. Two other patients lost their graft due to AMR.

Table 6. Outcome of renal transplantation after preemptive plasmatherapy
PatientsMutationRecurrence on previous graft (Y/N)Perioperative plasmatherapy regimenMaintenance regimenDuration of plasmatherapy treatmentRecurrence/ rejection (delay)1 OutcomeCreatinine μmol/L (Follow-up)
  1. PE = plama exchanges; PI = plasma infusion; yrs = years; mth = month.

  2. No = no mutation, no 2 at risk CFH haplotype.

  3. At risk CFH: two at risk CFH haplotypes.

  4. 1Delay between renal transplantation and the event (recurrence or acute rejection).

  5. 2Mutations in CFB and CFI.

1CFH (SCR1)PI over 4 daysNo4 daysNo/No80 (2 yrs)
2CFH (SCR4)Daily PI over 7 daysMonthly PI4 yrsNo/No80 (4 yrs)
3CFH (SCR10)YesDaily PI over 7 daysPI every 2 weeks3 monthsYes/No (3 months)Anti-C5 at 3 mth 175 (12 mths)
4Combined2PE over 10 daysPE every 1 week1 monthYes/No (1month)Anti-C5 1 mth 180 (2 yrs)
5CFIDaily PE over 5 days following tapering PIPI every 3 weeks18 monthsNo/Yes (AMR at 12 months)Graft lost (18 mths)
6CFIYesPI J0 et J1PI every 2 weeks2 yrsYes /No (2 yrs)Anti-C5 at 2 yrs 250 (3 yrs)
7NOPE 3 times per 7 days over 1 monthPI every week2 yrsNo/Yes (AMR at 6 weeks)Graft lost (2 yrs)
8At risk CFHYesPI over 9 daysPE every 1 months2yrsNo/No110 (4 yrs)
9NOYesDaily PIMonthly PI4 yrsNo/No100 (4 yrs)
Figure 7.

Graft survival without recurrence according to prophylactic preemptive plasmatherapy (p = 0.03, log rank).

Discussion

We report the outcome of renal transplant in a large cohort of aHUS patients screened for mutation in all known aHUS related genes and for at-risk CFH haplotype. The present study highlights that, overall, poor graft survival is largely due to aHUS recurrence. We demonstrated that complement genetic abnormalities strongly impact the rate of early recurrence and thus the graft outcome.

In our study, the overall outcome of kidney transplantation in adults with aHUS was poor with 7% of deaths and 50% of graft failure at 5 years posttransplant. Four patients without major cardiovascular risk factors, aged from 21 to 50 years, died from cardiovascular events. This finding is reminiscent of aHUS cases with arterial involvement, including stenosis of cerebral arteries and myocardial ischemia, suggesting that permanent complement activation may also harm the macrovascular vessels [31-33].

The high frequency of graft loss in aHUS recipients was first documented based on case reports with or without genetics available [2, 34] and from an International registry [3].

Our study highlights that, overall, poor graft survival is largely due to aHUS recurrence which occurred in 68% of the patients. and to a lesser extent to delayed graft function. We showed that the recurrence risk is major during the first year (accounting for 70% of recurrence) and decreases markedly after 2 years.

Recurrence is strongly determined by genetic background. Patients with mutations in complement factors had a threefold increase in posttransplant aHUS recurrence compared to aHUS patients without mutations. More importantly, this study provides new and useful information about the level of the risk of aHUS recurrence. The risk of recurrence is four times higher for patients with mutation in CFH gene or Hybrid gene between CFH/CFHR1 and CFB/C3 mutations as compared to patients without these genetic abnormalities. It is well documented that kidney transplantation in patients with CFH mutations is associated with a high recurrence rate. Of 42 published cases of transplanted patients (children or adult) with CFH genetic abnormalities, 32 had recurrence and 86% of the recurrences induced graft loss [3, 35-37]. We confirmed the high frequency of graft loss after recurrence in patients with CFH. Interestingly, in our study, the six patients who carried a Hybrid gene between CFH/CFHR1 in our study lost their renal allograft by early recurrence. This very high-risk subset of patients was not detected with the sequencing of the exonic sequence of the CFH but with MLPA technology. This finding is in line with the critical role of the C-terminus of CFH for binding to and protecting the endothelium [22, 36]. Data are emerging that patients with C3 and CFB mutations have a high risk of graft recurrence and graft loss. In our study, the risk of recurrence may be high for C3/B carriers, it is probably due to the small number of patient. Two patients who carried a CFB mutation lost their graft, one of whom has previously been reported [16]. Two other cases of CFB carriers who lost their grafts have been reported in the literature [14]. Recurrence occurred in four of the five grafts in four patients who carried C3 mutations. In contrast, Noris et al. reported only two recurrences in seven renal transplantations [3]. The main challenge remains to determine the functional consequences of the gene mutations to define the risk of aHUS disease recurrence [13].

Several previous reports suggest that patients with CFI mutation have a similar renal posttransplantation recurrence risk to those with a CFH mutation [9, 11, 38, 39]. We provide evidence that the risk of recurrence in grafts from patients with CFI mutations is lower compared with grafts from patients with CFH mutations. We have previously reported that complete deletion of the CFHR-1 gene, which is a common polymorphism, could adversely impact on the severity of the HUS disease in patients with Factor I mutation [7]. Our study failed to demonstrate that homozygous CFHR1 deletion increased recurrence in CFI carriers. Furthermore, we failed to identify the influence of at-risk CFH haplotype as our study included a relatively small number of patients with two at risk CFH haplotypes and without mutation. It is difficult to conclude about the specific impact of the at-risk CFH haplotype on the recurrence rate after kidney transplantation. Additional studies may help to determine the influence of these various polymorphisms on the risk of recurrence.

However, we determined that aHUS patients without gene mutations or two at-risk CFH haplotypes had a low-recurrence risk.

In our cohort, out of the three patients with an MCP mutation, only one transplantation was successful. MCP is the transmembrane protein which acts as a co-factor for CFI to inactivate C3b on the endothelial cell surface [40-43]. After renal transplantation, MCP production is driven by the endothelium cells from the donor and MCP carriers are therefore expected to have a low-recurrence risk [9]. To date, only two patients with MCP mutations have been reported as having experienced recurrence, one of whom was reported in this study [10]. The mechanism evoked to explain unexpected recurrence includes chimerism [10] or the presence of another mutation [42]. Interestingly, both carriers of an MCP mutation in the present cohort, including one already published [10], harbored two at-risk CFH haplotypes and the mutation in MCP was the only mutation identified.

Plasmatherapy failed to rescue graft function in most patients as previously reported in a review [44]. In this respect, we must nonetheless acknowledge that this retrospective study contains several caveats including variable times in plasmatherapy initiation and lack of homogeneity in frequency and duration of plasma exchange sessions. Nevertheless, the negative impact of recurrence on graft survival persisted independently whether patients received plasmatherapy or not. This reinforces the fact that patients should be administered prophylactic treatment.

We confirmed data previously reported that preemptive plasmatherapy may be effective in the prevention of disease recurrence and may reduce graft loss [36]–[49-51]. However three of the nine patients experienced disease recurrence. When switched to eculizumab renal dysfunction persisted but renal function was maintained. This study, as well as a recent publication of the outcome of 22 aHUS renal transplanted recipients treated by eculizumab, suggests that patients at very high risk of recurrence (mutations in CFH/Hybrid, or CFB/C3) should be administered preemptive therapy by eculizumab [45, 47-49] and that those with recurrence under plasmatherapy be rapidly switched to eculizumab [4-49]. Favorable renal outcome is rare in the CFH careers [50-52]. Patients at moderate risk, could be administered either preemptive plasmatherapy or eculizumab on a case-per-case basis. Finally, preemptive plasmatherapy may be sufficiently effective in patients at low risk (i.e. patients without mutations).

Although the duration of preemptive treatment is debatable, our study demonstrates that the rate of recurrence decreases markedly 2 years posttransplantation. These data may help clinicians to decide when to stop the treatment.

The estimated incidence of TMA due to CNI-based immunosuppressive therapy in patients without aHUS as initial nephropathy is 0.5% to 1% [53]. Our study failed to demonstrate a significant relationship between CNI therapy and recurrence of aHUS, in contrast to previous studies in patients with de novo posttransplant HUS [24, 53, 54]. At the opposite, we previously found a 30% prevalence of CFH and CFI mutations in patients with de novo posttransplant HUS [55]. Moreover, m-TOR inhibitors can induce TMA [25], and we found a significant risk of recurrence in m-TOR inhibitor-treated patients.

These results would suggest that patients should not be put on a CNI-free regimen in this setting as it exposes them to an increased risk of acute rejection [56, 57] whereas the recurrence risk is limited. Furthermore, switching to an mTOR inhibitor probably exposes them to a higher risk of recurrence.

Although the rate of acute rejection was high in our study, in line with a previous study which reports a rate of 30% [56], it was not significantly associated with an increased risk of graft loss. Moreover, the humoral rejection rate may be underestimated because the study covered a large period of time, without systematic screening for donor specific antibodies during the posttransplant course or systematic C4d staining on all renal biopsies. So far, it remains unclear whether the high incidence of acute rejection results from low exposure to CNI, fueled by the wish to limit CNI-related endothelial toxicity, or from a complement-dependent enhanced alloimmune response [57, 58].

Certain considerations need to be taken into account when interpreting the results. The first is that the patients were retrospectively selected from the French registry of aHUS. A second limitation is that the patients were enrolled over a large period of time and the inclusion criterion for previous genetic testing could have introduced selection bias. Such bias may be minimal as we failed to identify a significant difference in the outcome of grafts performed between 1995 and 2002 (before the knowledge of complement associated aHUS) and between 2002 and 2009 (data not shown). Third, therapeutic impact should be analyzed with caution. After adjustment for relevant covariates (DGF, BPAR and plasmatherapy) which could impact on graft survival, disease recurrence remains a significant factor associated with graft loss. We thus believe that our analysis represents an appropriate overview of the impact of complement genes on graft outcome.

In conclusion, our study provides strong evidence supporting previously published findings that aHUS disease recurrence impacts the graft outcome in renal transplant recipients. Recurrence is determined by the presence and the type of complement genetic abnormalities. We purpose to classify the risk of recurrence according to the genetic background. Patients with Factor H, C3 and CFB mutations or with CFH/CFHR1 hybrid gene carry a high risk of recurrence after renal transplantation. Patients with CFI, MCP, combined mutations or with two at risk CFH haplotype carry a moderate risk, whereas those with neither mutations nor CFH at risk haplotype have a low risk of recurrence.

Our study paves the way for future genetically based individualized prophylactic therapeutic strategies. In our opinion, preemptive eculizumab therapy should be administered to patients with high-risk mutations (CFH/Hybrid, CFB, C3) whereas preemptive plasmatherapy should be proposed to others patients with moderate and low risk of recurrence. Finally, unlike for mTOR inhibitors, we found no evidence supporting that CNI increased risk of recurrence. CNI-based regimens should thus be the immunosuppressive therapy of choice for aHUS transplanted patients to limit the risk of rejection.

Acknowledgments

We are grateful to the clinicians who referred their patients for complement investigations and accepted that patients were included in the study and particularly to Bruno Hurault De Ligny (Caen), Christophe Legendre (Necker), Benoit Barrou and Nadia Arzouk (Pitié Salpétrière), Antoine Durrbach (Bicêtre), Maryvonne Hourmant (Nantes), Pierre François Weestel (Amiens), Patrick Le Pogamp and Cécile Vigneau (Rennes), Marie Essig (Limoges), Olivier Toupance (Reims), Christophe Mariat and Nicolas Maillard (Saint Etienne), Denis Glotz (St Louis), Mathias Buchler (Tours), Philippe Lang (Mondor), Stéphane Burtey and Noemie Jourde (Marseille), Isabelle Etienne (Rouen).

We thank Stephanie Ngo, Delphine Beury, Nelly Poulain, Christine Hautreux and Jacques Blouin who provided expert technical support and Lubka Roumenina for the discussion about the functional consequences of complement mutations.

Disclosure

The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. J. Zuber, V. Frémeaux-Bacchi, C. Loirat received fees from Alexion Pharmaceuticals for invited lectures and are members of an expert board supported by Alexion Pharmaceuticals.

Ancillary