Relevance of glomerular C4d deposition in pediatric patients with Henoch-Schönlein Purpura compared to IgA nephritis

Background IgA nephropathy (IgAN) is the most common primary glomerulonephritis (GN) in western countries and Henoch-Schönlein purpura nephritis (HSPN) is the most common form of vasculitis in childhood. Renal biopsy ndings in both nephropathies are often similar and are characterized by mesangioproliferative GN with mesangial or mesangiocapillary IgA and C3c deposits. Aim of this study was to investigate the signicances of glomerular C4d-deposition as discriminating factor between pediatric HSPN and IgAN Methods We retrospectively analysed patient records and renal biopsies from 53 pediatric patients from one single center with a median age of 10.5 years [range 2.3–18 years]. Twenty-two patients suffered from IgAN and 31 from HSPN. Work-up of all renal biopsies was performed using standard protocols including immunohistochemistry for C4d.

rash is de ning. Accordingly, when a patient shows biopsy proven IgAN in the presence of purpura or additional clinical ndings such as other organ involvement, the clinical diagnosis is HSPN instead of IgAN. Of note, in the recent Chapel Hill classi cation HSPN was o cially renamed IgA-vasculitis [2] due to its potential presentation with microangiopathy.
Complement factor 4d (C4d) is a degradation product of the complement cascade which is generated during activation of the classical and lectin pathways. In renal transplantation pathology C4d is routinely used as a marker for active antibody mediated rejection (ABMR). Speci cally, C4d-positivity of peritubular capillaries is a strong indicator of ABMR in AB0-compatible renal transplantation. Moreover, speci c glomerular C4d-deposition was identi ed as a prognostic marker in adult and pediatric patients with IgAN but not with HSPN. Positive C4d-staining was seen in 38-68% of adult patients with IgAN mostly in a mesangial pattern [3,4,5]. In these studies patients with glomerular C4d-positivity were older, more frequently presented with hypertension, lower gross hematuria, more proteinuria, more glomerulosclerosis and interstitial brosis and lower renal survival than C4d-negative patients pointing to an inferior outcome of C4d-positive IgAN [3]. Furthermore, glomerular C4d-positivity was shown to be an independent risk factor for development of end stage renal disease (ESRD) in adult IgAN patients [4]. In pediatric IgAN, C4d-positivity was detected in 21% of cases and was also identi ed as an independent predictor of worse renal outcome [6].
Most of the above studies, however, excluded patients with HSPN from the analyses. Only Espinosa et al. described C4d-positivity in 2/8 adult HSPN patients [3] but there is no information on its correlation to morphology or prognosis. Apart from this report no other studies on glomerular C4d in HSPN in general or in pediatric HSPN have been published. Thus, there is currently no information available whether glomerular C4d-deposition differs between HSPN and IgAN in pediatric patients and might therefore be used as a discriminating factor on renal tissue or whether it might be useful for prognostic or therapeutic purposes. Therefore, it was the aim of our study to retrospectively analyse renal biopsies of pediatric patients with HSPN and IgAN from one single centre with (i) respect to glomerular C4d-staining and (ii) other morphological and clinical ndings. Additionally, we were interested in whether glomerular C4dstaining correlates with any of the other ndings and could be used to subclassify the cases or to predict prognosis.
Methods: -Patient selection: All consecutive renal biopsies during a period of 16 years (2002-2018) of patients with biopsy-proven IgAN diagnosed at one single pediatric nephrology unit were analysed. Based on their clinical presentation these patients were divided into 2 groups, i.e. IgAN and HSPN, with the characteristical purpura before or at onset of nephritis as criteria for HSPN. Indication for renal biopsy included persistent (>4 weeks) or recurrent (>3 relapses) gross hematuria, nephrotic range proteinuria or renal insu ciency. Standard diagnostic criteria for histological IgAN or HSPN were used i.e. mesangioproliferation, the presence of predominant granular mesangial (or mesangiocapillary) deposition of IgA and less pronounced C3c in immunohistology, and predominant mesangial (and subendothelial) osmiophilic deposits in electron microscopy. We excluded only one patient with complex juvenile polyarthritis who was treated for several years with different immunosuppressive agents before IgAN was diagnosed. Furthermore, repeat renal biopsies performed in a subgroup of patients with severe renal injury after completion of cyclophosphamide treatment were excluded. The following parameters were documented at the time of renal biopsy and at last follow-up: age [ All renal biopsies were re-evaluated by a renal pathologist (K.A.) in a blinded manner. The following biopsy characteristics were documented: number of glomeruli, mesangial hypercellularity (score 0-2 with 0: absent, 1: mild and 2: moderate), mesangial matrix expansion (yes/no), segmental glomerulosclerosis (yes/no), endocapillary hypercellularity (yes/no), crescents (yes/no), interstitial brosis and tubular atrophy (IFTA, %), arterial wall thickening (yes/no) and Oxford classi cation for IgAN [7,8,9]. In addition, patient records were reviewed for immunosuppressive (steroids, cyclophosphamide and others) and ACEinhibitor (ACE-i) treatment. The study was conducted according to the general guidelines for studies using human material. The use of archival biopsy material was approved by the local ethics committee (re.-no.4415).
-De nitions: Nephrotic-range proteinuria was de ned as urinary total protein >2.0 g/g creatinine in spot urine or >1.0 g/m 2 /day in 24-h urine. Non-nephrotic-range proteinuria was de ned as urinary loss of protein between 0.2 and 2.0 g/g creatinine or between 0.15 and 1.0 g/m 2 /day in 24-h urine, respectively.
Hypertension was de ned as systolic blood pressure >95 th percentile [11].
-Statistical analysis: First, statistical differences between IgAN and HSPN patients were assessed. In case of binary variables, Fisher's exact test (two-tailed) was employed to assess the statistical signi cance of the association between the IgAN/HSPN diagnosis and the measured binary variable; for all other variables Wilcoxon Rank Sum test (with normal approximation) was employed to compare both groups. Afterwards, differences between C4d-positive and C4d-negative patients were assessed employing an analogous procedure. All statistical analyses were performed in the R (v. 3 Indication for renal biopsy in IgAN patients was recurrent or persistent gross hematuria in 64%, proteinuria in 41% and renal insu ciency in 23%. In contrast, indication for renal biopsy in HSPN patients was proteinuria (81%) in the majority of cases and less frequently recurrent or persistent gross hematuria (26%) or renal insu ciency (3%). Of course, in both groups more than one biopsy indication was observed in some individuals. At the time of renal biopsy, 23% of IgAN and 32% of HSPN patients had arterial hypertension (n.s.). Serum creatinine levels were signi cantly higher in the IgAN compared to the HSPN group with more IgAN patients showing renal insu ciency as indication for renal biopsy. In line with these results, glomerular ltration rate (GFR) was signi cantly lower in IgAN compared to HSPN patients (101.5 ± 41.8 ml/min/1.73 m 2 and 120.4 ± 33.0 ml/min/1.73 m 2 ; p < 0.05). Proteinuria was extremely variable in both groups and ranged from no to massive nephrotic range proteinuria. However, proteinuria (24-hour urine as well as spot urine) was signi cantly higher in HSPN compared to IgAN patients (Table 1). Of interest, in HSPN most patients (58%) showed nephrotic range proteinuria and only 13% had no proteinuria. In contrast, in IgAN only 23% presented with nephrotic range proteinuria and 36% had no proteinuria at the time of renal biopsy (p < 0.05). In parallel, serum protein and albumin values were signi cantly lower (p < 0.05 and p < 0.005, respectively) in HSPN compared to IgAN (Table 1). Interestingly, mean serum C3 levels were in the normal range in both groups but showed signi cantly higher values in HSPN compared to IgAN patients (1.28 g/l ± 0.17 g/l vs 1.15 g/l ± 0.16 g/l, p < 0.05).
On renal histology, we found a slightly higher rate of global glomerulosclerosis in C4d-positive vs. C4dnegative patients (29% vs 13%, n.s.) and a tendency to a higher crescent formation (57% vs 38%, n.s.). The rate of segmental glomerulosclerosis or endocapillary hypercellularity was comparable between both groups. Oxford classi cation showed no differences between C4d-positive and C4d-negative patients. In the further follow-up, 75% of C4d-positive and 48% of C4d-negative patients received at least one immunosuppressive drug. Of interest, cyclophosphamide treatment was signi cantly more frequent in C4d-positive than C4d-negative patients (35% vs 10%, p < 0.05) whereas use of ACE-i was comparable (95% vs 81%, n.s.). At last follow-up, which was at a median of 47 months (C4d-positive patients) and 41 months (C4d-negative patients) more C4d-positive (31%) than C4d-negative patients (12%) presented with reduced GFR < 90 ml/min/1.73 m 2 . Due to the small sample size this difference, however, was not statistically signi cant. Of note, the rate of proteinuria at follow up was similar in both groups (13% vs 15%, n.s.).

Discussion:
The aim of our study was to compare renal biopsies of pediatric patients with HSPN and IgAN with respect to glomerular C4d-deposition and other morphological ndings. As C4d is a routinely used antibody in renal transplant pathology that additionally seems to have some prognostic impact in some immunecomplex glomerulonephritis (GN), we were interested in whether mesangial C4d-deposition may differ between pediatric HSPN and IgAN. For IgAN, C4d-positivity was described in a variable percentage (21-68%) of patients [3][4][5][6]. Of interest, for HSPN this information is lacking as only very few studies included HSPN patients. Espinosa et al. analysed a total number of 8 HSPN patients of which 2 were positive for C4d (25%) resulting in a similar rate of C4d-positivity as in IgAN patients (32%) [3]. All other studies analyzing glomerular C4d-deposition sofar excluded HSPN patients. Therefore, our analysis is the rst to analyse glomerular C4d-positivity in pediatric HSP patients and also the largest study population for HSPN patients with respect to glomerular C4d-staining so far. Interestingly, we found a comparable rate of glomerular C4d-positivity in pediatric HSPN (42%) and IgAN patients (36%) indicating that C4ddeposition is not helpful in distinguishing between renal manifestation of HSPN and IgAN. In addition, we analyzed the differences between our HSPN and IgAN patients with regard to other histological and some clinical parameters. Of note, we found a higher rate of gross hematuria in IgAN compared to HSPN patients. In contrast, HSPN patients showed a higher rate and signi cant higher level of proteinuria and signi cantly more endocapillary hypercellularity. Recently, the CureGN study, a large register study, investigated a subgroup of 285 pediatric IgAN and HSPN patients [1]. Consistent with our results, they found a higher age in pediatric IgAN compared to HSPN patients with the mean ages of both groups being very similar to our monocentric patient cohort.
The second question of our study was to compare C4d-positive versus C4d-negative patients. Here, we found a signi cantly higher rate of gross hematuria in C4d-positive patients as well as slightly more proteinuria, hypertension and immunosuppressive treatment. C4d-positive patients more often suffered from lower grade of renal insu ciency at last follow-up, i.e. 4 years after renal biopsy. As in other studies in IgAN [4,6,12] glomerular C4d-positivity was also associated with poor renal outcome. This might be of practical relevance as staining for C4d is a standard method in renal pathology and thus widely used [13,14,15]. C4d, a degradation product of complement cascade, is generated by activation of classical and lectin pathways [16,17] and present in glomerular diseases [16]. Activation of the complement system plays a pathophysiological role in IgAN [17,18,19]. Apart from activation of the alternative pathway (as evidenced by C3c deposition) [20] about 25% of IgAN patients showed glomerular deposition of mannose binding lectin (MBL) and C4d pointing to additional activation of the lectin pathway. Of note, the C4ddeposits were not found to be in coincident with the MBL positivity, but the subgroup of C4d-positive IgAN patients presented more severe renal damage in histology and pronounced proteinuria [20]. In general, renal C4d-staining is predominantly seen in the mesangium (43%), but was also reported in tubules (11%) and interlobular arteries (22%); in contrast to ABMR no C4d-positive peritubular capillaries were reported in IgAN [5]. In 2009, Espinosa et al. were the rst to systematically analyse mesangial C4d-staining in patients with histological proven IgAN and found an association of positive C4d-staining with older age, hypertension, lower gross hematuria, more glomerulosclerosis, interstitial brosis, lower GFR and higher proteinuria, respectively [3]. When analysing long-term renal outcome they found a signi cantly higher 10 years renal survival in C4d-negative (91%) vs C4d-positive (44%) IgAN patients [3]. Only a small number of C4d-negative patients (7.5%) developed ESRD in the mean follow-up time of 12.7 years. These results were con rmed in a large multicentre Spanish study of 283 patients identifying glomerular C4dpositivity as an independent risk factor for development of ESRD in IgAN [4]. Also in pediatric IgAN patients with less chronic renal changes and a lower rate of hypertension, 21% showed positive glomerular C4d-staining which was an independent predictor of loss of renal function with renal survival of 8.6 years in C4d-positive vs. 15.1 years in C4d-negative patients [6]. In this analysis, C4d-positive patients showed higher levels of proteinuria, a tendency that was also seen in our study. Also in our follow-up analyses we found more patients with at least mild renal insu ciency (GFR < 90 ml/min/1.73 m 2 ) in the C4d-positive group (31% vs 12%). Interestingly, the rate of immunosuppressive treatment was slightly higher in C4d-positive compared to C4d-negative patients and this difference even reached statistical signi cance when analysing the use of cyclophosphamide separately. Both results indicate a more severe course of HSPN and IgAN when glomerular C4d is present and are in line with the report of Segarra et al [12] who investigated C4d in a subgroup of IgAN patients with GFR > 80 ml/min and found a lower amount of only 20% C4d-positivity compared to other studies [3][4][5]. However, C4dpositivity was associated with higher overall proteinuria. In the follow-up C4d-positive patients suffered from more ares, received more immunosuppressive treatment and had higher proteinuria compared to C4d-negative patients. In a multivariate analysis C4d was found to be an independent predictor of a decline in GFR slope and of ESRD [12]. A recent study [21] investigated renal microangiopathic lesions in adult and pediatric IgAN and HSPN patients using vascular C4d and C5b-9 positivity as markers of complement activation. Of note, they found renal microangiopathy associated with higher vascular C4d and C5b-9 deposits. Interestingly, patients with microangiopathy and vascular C4d-positivity had a signi cantly worse outcome with poorer renal survival, lower GFR and increased chronic lesions on histology [21]. The main analysis of this study, however, did not distinguish between pediatric and adult patients nor IgAN and HSPN patients The results of our monocentric retrospective study are limited by the relatively low number of pediatric patients and some dropout in the follow up. On the other hand, our single centre patients represent a relatively homogenous group in terms of diagnosis and treatment. We also have to acknowledge that C4d-staining might be variable as shown for IgAN patients and therefore one has to be careful with respect to the conclusions drawn from these data. We tried to take care of this problem, however, by staining all biopsies at one time point and by multiple separate analyses of the staining.

Conclusions:
In conclusion, glomerular C4d-staining does not differ between HSPN and IgAN as the rate of C4dpositivity was comparable in both groups. The ndings in our single centre pediatric HSPN patients, however, are in line with earlier studies in IgAN showing glomerular C4d-positivity to be an adverse histological marker associated with poor renal outcome.
Abbreviations ABMR active antibody mediated rejection C4d Complement factor 4d drafting the article or revising it: KB, KA, CD, MB. In providing intellectual content of critical importance to the work described: FF, EV. Final approval of the version to be published was performed by all authors.