Glomerular C4d in Post-Transplant IgA Nephropathy is associated with decreased allograft survival

Background Glomerulonephritis (GN), including post-transplant IgAN (post-Tx IgAN) is an important contributor to decreased long-term allograft survival. The immunopathological detection of the complement degradation product C4d in glomeruli (C4dG) has been recently described as a risk factor in native kidney IgAN, however little is known about C4dG deposition in post-Tx IgAN. We hypothesized that glomerular C4d may indicate a more aggressive disease course and worse allograft survival in patients with post-Tx IgAN. Methods In this retrospective study we assessed the presence and clinical relevance of C4dG in patients with post-transplant IgAN. We analyzed 885 renal allograft recipients, including 84 patients with post-transplant GN. All patients were transplanted between January 1999 and April 2006 and underwent at least one biopsy for differnt causes. The primary endpoint was death-censored graft survival, with a median follow-up of 9.6 (IQR 3.8–13.2) years. Results The prevalence of post-Tx GN was 9.5%. Twenty-seven patients with post-Tx IgAN were included. C4dG positive patients (N = 18, 66.7%) had significantly worse allograft survival compared to C4dG negative post-Tx IgAN patients and patients without post-Tx IgAN [C4dG positive: 27.8% vs. 55.6% and 66.0%; log-rank: p = 0.01]. C4dG remained a significant risk factor (HR 2.22, 95% CI 1.27–3.87) for allograft loss even after adjustment for T cell mediated rejection (TCMR) and antibody mediated rejection. Conclusion Glomerular C4d deposition is an independent risk factor for worse graft-survival in patients with post-Tx IgAN, even after adjusting for other risk factors such as antibody mediated rejection. Assessment of glomerular C4d deposition may provide a valuable prognostic risk assessment tool to identify high risk patients in post-Tx IgAN. Graphic abstract Electronic supplementary material The online version of this article (10.1007/s40620-020-00914-x) contains supplementary material, which is available to authorized users.


Introduction
Chronic glomerulonephritis (GN) is one of the leading causes for the development of end-stage renal disease (ESRD) [1]. IgA nephropathy (IgAN) represents the most common primary glomerular disease and its clinical course varies from asymptomatic microhematuria to progressive chronic kidney disease (CKD) and ESRD [2][3][4][5]. Aberrant glycosylation of IgA1, circulating antibodies against the galactose-deficient IgA1 and the formation of IgA1-IgG immune complexes, are suggested mechanisms in the so called "multi-hit" hypothesis of IgAN [6][7][8][9]. In addition, emerging evidence supports an important role of complement pathways in IgAN [10][11][12][13][14][15][16][17][18][19]. Several retrospective studies suggested that glomerular C4d staining (C4dG) may have a prognostic value in native glomerular kidney disease [13,[20][21][22]. In 2012 Espinosa et al. [21] performed a retrospective multicenter analysis of 283 patients with native kidney IgAN. Around two thirds of their included patients had C4d deposits in glomerular capillaries. C4dG was associated with significantly worse 20-year renal survival (28% vs. 85%) and represented an independent risk factor for progression of CKD in patients with native IgAN [21]. Similar findings were observed in two subsequent analyses, independently from baseline estimated glomerular filtration rate (eGFR) [22,23]. In renal transplantation, immunopathological C4d staining in peritubular capillaries (C4dPTC) is an accepted marker for the diagnosis of antibody-mediated rejection (ABMR) [24]. Linear C4d deposits in PTC are considered to be induced by HLA-antibodies with subsequent activation of the classical pathway and are associated with a higher risk of allograft loss [25][26][27]. In native and post-transplant IgA Nephropathy (post-Tx IgAN) similar pathomechanisms have been suggested [28][29][30], but in contrast to native kidney IgAN, the role of C4dG in post-transplant IgAN is yet to be determined. Further, evidence regarding the frequency and clinical significance of C4dG depositions in post-transplant GN is scarce [31]. We hypothesize that similar to reports in native kidney IgAN [32], C4dG deposits in post-Tx IgAN may indicate worse clinical outcome. Aim of our study was therefore to analyze 1) the frequency of C4dG deposition in allograft recipients with recurrent or de-novo IgAN, 2) the long-term death-censored graft survival in C4dG positive post-Tx IgAN patients and 3) their clinical course compared to C4dG negative patients.

Study design and patients
The present study was designed as a retrospective singlecenter analysis. The objective was to assess the presence and clinical relevance of C4dG staining in patients with de-novo or recurrent post-transplant IgA nephropathy. The primary outcome was death-censored graft-survival, defined as date of initiation of any renal replacement therapy. We analyzed all patients with consecutive renal transplantation between January 1st 1999 and April 1st 2006 (N = 1248) and available biopsy results as well as clinical follow-up until January 1st 2017. All biopsies were for cause biopsies, performed upon unexplainable graft dysfunction and/or proteinuria. Study approval was obtained from the local institutional ethics committee (EK-number 1490/2017). All procedures were conducted according to the Declarations of Helsinki and Istanbul.

Baseline characteristics
Eight hundred and eighty-five patients underwent allograft biopsy for various causes and were included into the study. Baseline characteristics were analyzed from digital patient records and included relevant demographic and transplantassociated findings as well as the occurrence of early (within the first six months after transplant) rejections including T cell mediated rejection (TCMR), ABMR and Borderline lesions and eGFR and proteinuria after one year. The glomerular filtration rate was estimated using the Mayo Clinic formula [33].

Analysis of biopsies and C4d staining
Of the 885 included patients, a diagnosis of post-transplant glomerulonephritis was made in 84 (9.5%) of them.
Individuals with suspected post-Tx GN underwent detailed immunohistochemistry and ultrastructural analysis with electron microscopy (if material was available). Details on staining procedures and electron microscopy are provided as a supplemental section. The diagnosis of post-Tx IgAN was defined by the presence of glomerular IgA-dominant or codominant immune deposits assessed by immunohistochemistry or immunofluorescence, as described before [34]. In the literature the definition of C4dG in native kidney IgAN is variable, depending on the staining technique (IF versus IHC) and includes the mesangium and glomerular capillary walls [14,16,17,19]. Therefore, C4dG deposition was defined as an IHC-based detection of granular C4d deposits in > 5% of peripheral glomerular capillaries or C4d deposits in the mesangium (Fig. 1). All included biopsies with post-Tx GN were retrospectively re-analyzed by two nephropathologists (N.K., J.K.) blinded to clinical data. Biopsies were reassessed according to the 2017 Banff scheme [35]. C4d positive ABMR was considered present if linear, circumferential C4d deposits in PTC were above 0% [36]. Medullary vasa recta were also assessed if available. The diagnosis of C4d negative ABMR had not been established during the observation period for most patients. To accommodate the potential effect of C4d negative ABMR on graft survival, and due to the lack of post-transplant HLA antibody results (which were not standard of care at our center during the study period), we chose a broader definition of the histomorphology compatible for C4d negative ABMR (= suspicion of C4d negative ABMR), which consisted of the presence of inflammatory lesions compatible with ABMR in at least two different renal compartments including a g score > 0, a ptc score > 0, post-transplant thrombotic microangiopathy (TMA), V3 lesions and a cg score > 0. Additionally, following the BANFF guidelines, a ptc score > 0 was not considered sufficient for ABMR diagnosis in the presence of tubulointerstitial infiltrates. Second, g and cg lesions were not scored in the presence of glomerulonephritis. The Oxford criteria (MEST-C score) were applied according to the 2017 published consensus reports [37].

Treatment
The treatment policy for post-Tx IgAN during the respective time period was watchful waiting and the introduction of angiotensin converting enzyme inhibitor (ACEi)/angiotensin receptor blocker (ARB) (if not already established) in the majority of cases. Only in cases with rapid deterioration of graft function intensified immunosuppressive regimens were applied. Those regimens were: steroid bolus therapy in three patients, switch to calcineurin inhibitors and/or steroid bolus therapy in four patients and one patient received oral cyclophosphamide (2 mg per kilogram). All rejection episodes were treated as follows: patients with TCMR Banff ≥ IA-IIA or Banff Borderline lesions with concomitant allograft dysfunction received a steroid pulse. Steroidrefractory or Banff IIB rejections were treated with antithymocyte globulin (ATG) over a course of ten to fourteen days. C4d positive ABMR with allograft dysfunction was treated in the majority of cases as previously described [38].

Statistical analysis
Continuous variables were expressed as mean and standard deviation or median and interquartile range, whichever appropriate. Categorical variables were expressed as absolute and relative frequencies. Unpaired Student´s t or Mann-Whitney-U was used for comparison of continuous data; group comparisons of categorical variables were analyzed by Fisher´s exact or Chi Squared (χ 2 ) test. Kaplan-Meier analysis and Mantel-Cox log-rank test were used for comparison of graft survival. Graft survival was compared between 1) patients with and without post-Tx GN and 2) between C4dG positive and negative post-Tx IgAN patients and patients without post-Tx GN (control group). For Cox regression analysis, variables differing between the groups with GN and without GN or significantly related to the primary endpoint were included. Variables with p < 0.05 in the univariate model were included into the multivariate regression analysis. p-values < 0.05 were considered statistically significant. Statistical analysis and graph design were performed using commercially available software systems

Study cohort
In total 885 patients with indication biopsy and a median follow-up of 9.6/3.8-13.2 years (median/IQR) were included. Most patients were male (63.5%), mean donor and recipient age were 48.7 ± 15.0 and 50.9 ± 13.7 years, respectively. One hundred and two (11.5%) patients received transplants from living donors. Most patients (N = 656/74.3%) were on Cyclosporin A-based baseline immunosuppression. Further baseline parameters are shown in Table 1.

Immunohistochemistry in post-Tx IgAN patients
C4dG stainings were available in 27 patients with post-Tx IgAN. Of those, 18 (66.7%) showed positive C4dG staining. In most cases (N = 16), granular C4dG was found in peripheral capillary loops. Four patients had mesangial C4d deposits (two of them had mesangial and peripheral capillary C4d deposits). Glomerular basal membrane splitting was found in five out of eleven patients with available silver staining and peripheral capillary C4d deposits. In three biopsies concomitant C4d positive ABMR (linear C4d positive in PTC) was diagnosed. Those patients were diagnosed with chronic antibody-mediated rejection. Simultaneous tubulointerstitial infiltrates were observed in fourteen patients (BANFF ≥ 1 N = 7, BANFF Borderline N = 7). Neither the Oxford MEST-C single scores nor the sum score differed significantly between C4dG positive-and negative patients (data not shown). Crescents were found in two patients with post-Tx IgAN. As shown in Table 3, neither baseline variables nor laboratory findings at the time of index biopsy, including graft function and proteinuria, differed significantly between C4dG positive and negative post-Tx IgAN patients. Moreover, intensified immunosuppression as a treatment of post-Tx IgAN was rare and did not differ between both groups (data not shown). C4dG positivity was a constant finding in follow-up biopsies (available in six C4dG positive patients). All patients had positive C4dG stainings in follow-up biopsies (three with mesangial and three patients with peripheral capillary C4d deposits).

C4dG is associated with worse death-censored graft survival
Kaplan Meier analysis revealed a significantly worse deathcensored allograft survival of C4dG positive post-Tx IgAN patients [death-censored graft-survival: C4dG 27.8% vs. 55.6% in C4dG negative post-Tx IgAN patients and 66.0% in patients without GN; log-rank: p = 0.01] (Fig. 3). Exclusion of the three patients with concomitant C4d positive ABMR Table 1 Baseline findings of included patients ABMR antibody-mediated rejection, AZA azathioprine, DGF delayed graft function, GN glomerulonephritis, IQR interquartile range; MMF mycophenolate mofetil; N number, Neg. negative, SD standard deviation, TCMR T cell mediated rejection, Tx transplantation, Pos. positive. Mayo Clinic estimation was used to calculate the eGFR. Early rejection was classified as occurrence within the first six months after transplantation; *Suspicion of C4d negative ABMR, which consisted of the presence of at least two lesions being compatible with ABMR including a g score > 0, a ptc score > 0, post-transplant thrombotic microangiopathy (TMA), V3 lesions and a cg score > 0  (Table 4).

Discussion
In this analysis of a representative and well characterized cohort of renal transplant recipients, we demonstrated the clinical relevance of glomerular capillary C4d staining as a profound prognostic marker in patients with post-Tx IgA nephropathy. To the best of our knowledge, this is the first study showing that post-Tx IgAN patients with C4dG deposits had significantly worse death-censored allograft survival. The substantially increased risk for allograft loss remained significant even after adjustment for other known risk factors such as HLA-mismatch, donor age and the occurrence of biopsy-proven rejection. The prognostic impact of C4dG in post-Tx IgAN is in line with descriptions of patients with native kidney IgAN, where C4dG depositions have been associated with an overall worse renal prognosis [15,21,22,39]. However, in renal transplant cohorts with posttransplant glomerulonephritis the interpretation of immunopathological C4dG may be hampered due to HLA-antibody-triggered C4dPTC and/or C4dG. In our cohort only three subjects were diagnosed with concomitant ABMR, we can therefore safely exclude ABMR as a driving force of increased risk for graft loss in post-Tx IgAN patients. Nevertheless, the underlying pathomechanisms are still not fully understood. It was hypothesized that more progressive forms of IgAN are associated with endocapillary hypercellularity and vascular injuries, leading to complement activation and deposits of complement degradation products in various glomerular compartments-in particular in mesangial cells and glomerular capillary walls [40]. Recent studies in native kidney IgAN however have suggested additional clinical relevance of C4d staining in other renal compartments including tubules, interlobular arteries and arterioles [14,16,17,19]. Graft loss occurred in two thirds of our group of C4dG positive post-Tx IgAN patients, within a median of two years after the index biopsy. At the time of index biopsy, neither graft function nor proteinuria differed between C4d positive-and negative IgAN patients. Based on our findings, we therefore hypothesize that C4dG may help to identify a progressive course of disease in patients with post-Tx IgAN, even before a severe deterioration of graft function is observed. This is in contrast to studies in native kidney IgAN [21,39]. In 2014, Espinosa et al. [21] reported on 283 patients with IgAN (38.5% C4d positive). In their study, patients with C4dG positive biopsies already had significantly more proteinuria and worse eGFR at the time of index biopsy. We believe that the difference in our findings may be due to regular follow-ups of transplant recipients and thus earlier diagnosis of post-Tx IgAN compared to native kidney IgAN patients with very heterogeneous disease courses. Nevertheless, in both studies the association of C4dG with ESRD remained statistically significant even after correction for baseline eGFR. In contrast, in our study C4dG negative patients had death-censored graft survival that was comparable to the control group, suggesting an acceptable allograft prognosis even after post-Tx IgAN diagnosis. Most of our patients were not subjected to intensified immunosuppression after the diagnosis of post-Tx IgAN. Moreover, concomitant TCMR and Banff Borderline lesions were routinely treated with steroid boli, and their prevalence was not different between the two groups (data not shown). We can therefore safely exclude a relevant treatment bias between C4dG positive and negative groups.
Another finding that needs to be emphasized is that posttransplant glomerulonephritis was a major contributor for decreased long-term graft survival. Not only post-Tx IgAN patients but all patients with a diagnosed post-Tx GN had significantly worse death-censored graft-survival compared to the control group. This is in line with previously published findings from Chailimpamontree et al. [41].
One major strength of this study is the long followup period. We believe this is essential because post-Tx GNs often manifest late and show a heterogeneous disease course. Consequently, insufficient follow-up periods may lead to an underestimation of the decreased long-term survival of post-Tx GNs. Further, multiple relevant risk factors for decreased allograft survival, including donorassociated risk factors, were available and were included into the multivariate analysis.
Our study has limitations that need to be discussed. First, though the study design was retrospective, we have included a large sample of post-transplant biopsies with long-term follow-up and corrected our findings for multiple covariates including HLA-mismatch, and early-and concomitant rejections. Moreover, the follow-up period of included biopsies was comparable between C4dG positive and negative patients, arguing against a relevant time selection bias. We found predominantly peripheral capillary-rather than mesangial C4d positivity, a deposition pattern that may seem counterintuitive at first glance. Nevertheless, our findings are in concordance with previous reports from patients with native IgAN, where C4d was described in various glomerular compartments: Rath et al. reported that 40% of his patients with IgAN showed C4d depositions in the glomerular capillary walls [17]. These findings were in line with a report from Meng et al. [16] where C4d was found in the mesangium as well as in capillary loops. Differences in IgA and C4d staining patterns may be related to possible alterations of non-endothelial C4d deposits resulting from formalin-fixation or paraffinembedding [42]. Moreover, despite equally distributed mesangial-and peripheral C1q deposition, a potential role of MBL pathway activation cannot be excluded in certain patients. Another important limitation is the lack of systematic post-transplant assessment of anti HLA-antibodies. We have corrected this by using a broad definition of histological features compatible with C4d negative ABMR, and by the inclusion of both C4d positive and C4d negative ABMR in multivariate models. However, we cannot completely rule out the possibility that some biopsies were not attributed to ABMR.
Further, it is important to know that isolated microhematuria after transplantation was not considered a biopsy indication in the study period.
In conclusion, this study shows a prognostic value of C4dG deposition in patients with post-transplant IgAN. C4dG positivity was significantly associated with graft loss and remained an independent risk factor even after correction for covariates in the multivariate analysis. These findings are of clinical importance providing new evidence and suggesting that C4dG assement may be useful for a riskbased management of post-Tx IgAN patients.  ABMR antibody-mediated rejection, ACEi angiotensin converting enzyme inhibitor, ARB Angiotensin receptor blocker, eGFR estimated glomerular filtration rate, HLA human leukocyte antigen, IQR interquartile range, N number, neg. negative, pos. positive, Post-Tx IgAN post-transplant IgA nephropathy, TCMR T-cell mediated rejection. Early rejection was classified as occurrence within the first 6 months after transplantation; the antihypertensive treatment was known in 23 patients; Mayo Clinic estimation was used to calculate the eGFR; *: Suspicion of C4d negative ABMR, which consisted of the presence of at least two lesions being compatible with ABMR including a g score > 0, a ptc score > 0, post-transplant thrombotic microangiopathy (TMA), V3 lesions and a cg score > 0