Clinical significance of C4d deposition in renal tissues from patients with primary Sjögren’s syndrome—a preliminary study

Background To evaluate renal expression of C4d, a complement component in the classical/mannose binding lectin (MBL) pathway, in patients with primary Sjögren’s syndrome (pSS)-associated renal impairments. Methods We retrospectively reviewed the clinical and pathological data from 39 patients with pSS presenting with renal impairments. C4d was examined in paraffin-embedded biopsy tissues using immunohistochemistry. Glomerular C4d positive was defined when > 75% glomeruli were globally stained. Tubulointerstitial C4d (TI-C4d) were scored semi-quantitatively as 0 (absent), 1 (spotty or weak), 2 (patchy) and 3 (diffuse). A TI-C4d score ≥ 2 was considered TI-C4d positive and included in the TI-C4d+ group and vice versa. Peritubular capillary (PTC) C4d was scored as 0 (absent), 1 (0~10%, minimal), 2 (10%~ 50%, focal), and 3 (> 50%, diffuse). Results Glomerular C4d deposition was observed in all 8 patients with pSS-related membranous nephropathy (MN) without obvious C1q deposition. Two of 5 patients with mesangial proliferative glomerulonephritis and 1 of 2 patients with IgA nephropathy had mild mesangial C4d deposition. Sixteen patients (6 glomerular dominant and 10 tubulointerstitial dominant) presented TI-C4d score ≥ 2. Patients in the TI-C4d+ group exhibited a higher serum creatinine level at the time of renal biopsy (TI-C4d+ 132.5 [89.7, 165.5] vs. TI-C4d− 83.0 [70.7, 102.0] μmol/L, P = 0.008). PTC C4d was observed in 12 patients, with each of minimal, focal and diffuse staining being noted in 4 patients. Conclusions The MBL pathway of complement activation was potentially involved in pSS-related MN. Tubulointerstitial C4d might be a pathological marker of severe renal injury in patients with pSS-related renal impairments. Electronic supplementary material The online version of this article (10.1186/s12882-019-1341-y) contains supplementary material, which is available to authorized users.

Primary membranous nephropathy (PMN) is a kidneyspecific disease with increased proteinuria and pathological characteristics of sub-epithelial immune complex deposition (predominantly IgG4 and C3 deposition) [9]. PMN is characterized by the presence of serum autoantibodies against the podocyte component M-type phospholipase A2 receptor (PLA2R) in approximately 70% of patients [10] and anti-thrombospondin type 1 domain containing 7A (THSD7A) antibodies in approximately 10% of those with negative anti-PLA2R antibodies (nearly 3% of patients with intrinsic antibodies related MN) [11,12]. The mannosebinding lectin pathway of complement activation is involved in PMN, as evidenced by sub-epithelial deposition of hypogalactosylated IgG4 and C3 and 100% glomerular C4d deposition without C1q co-deposition [13].
Notably, pSS is one of the possible causes of secondary MN [3]. Patients with pSS-MN may present extra renal manifestations that do not exist in patients with PMN [6,14]. In patients with MN secondary to pSS, the complement activation pathway has not been studied. Patients with pSS-TIN are characterized by interstitial lymphocyte infiltration with rare tubular and interstitial immune complex deposition [15]. Autoantibodies against tubular components have been identified in patients with pSS-TIN by some authors [16][17][18], and the complement component C9 was also observed around TBM in some patients with pSS-TIN [4]. It has not extensively researched whether immune complex-mediated interstitial injury plays a role in pSS-TIN. C4d deposition in various compartments of renal tissues have been reported. C4d deposition in the glomerulus indicates local complement activation through the classical or lectin pathway [19]. It is used to differentiate C3 glomerulopathy and immune complex-mediated proliferative GN [20]. Mesangial C4d deposits have been observed in patients with IgA nephropathy (IgAN) [21], lupus nephritis (LN) [22], and thrombotic microangiopathy (TMA) [23]. Peritubular capillary (PTC) C4d correlates with antibody-mediated transplant rejection and inferior renal allograft outcomes [24]. Tubular C4d deposition is associated with an increased WHO grade of IgAN [25]. To our knowledge, neither glomerular C4d deposition nor tubulointerstitial C4d deposition have been investigated in patients with pSS-MN and in patients with pSS-TIN.
In the present study, we performed a retrospective examination of 39 patients with pSS with renal involvement (21 cases with TIN and 18 cases with GMN) from a single centre. C4d were examined in formaldehydefixed, paraffin-embedded biopsy tissues using immunohistochemistry. The objectives were to investigate the prevalence and localization of C4d deposits in renal biopsy tissues in patients with pSS-MN and those with pSS-TIN.

Patient selection
We retrospectively analysed the clinical and pathological data from 82 hospitalized patients with biopsy-proven renal involvements secondary to "Sjögren's syndrome" (ICD: M35.001) in Peking Union Medical College Hospital (PUMCH) from 1996 to 2011. After Sjögren's syndrome was redefined based on the 2002 American and European Classification Criteria (AE), we excluded 13 patients with systemic lupus nephritis (LN), 2 patients with primary biliary cirrhosis (PBC), 3 patients with rheumatoid arthritis (RA), 2 patients who were positive for anti-HCV antibodies, 2 patients with unidentified GMN due to glomerular sclerosis, and 3 patients for whom adequate kidney biopsy samples were unavailable. In addition, 18 patients who only qualified based on the Fox criteria which does not require the presence of anti-SSA/SSB antibodies were excluded. Thirty-nine patients were enrolled (Fig. 1). All inpatients received a consultation with a rheumatologist and had a confirmed diagnosis of primary Sjögren's syndrome. After patient selection, archived formalin-fixed and paraffin-embedded kidney tissues were retrieved and stained as described below. The study was performed in

Clinical data
Medical records including the demographic data, the duration of pSS, the duration of renal impairment from symptom onset to renal biopsy, medical comorbidities and treatment history were retrospectively reviewed. The immunological screening for antinuclear antibodies (ANA), anti-SSA/SSB antibodies, immunoglobulins (IgG, IgM and IgA), complement (CH50, C3 and C4), C-reactive protein (CRP), rheumatoid factor (RF) and cryoglobulins were obtained. Renal function was evaluated by the estimated glomerular filtration rate (eGFR) with the CKD Epidemiology Collaboration (CKD-EPI) equation. Renal tubular function screening included blood and urine pH values, electrolytes, blood and urine α1-macroglobulin and β2-microglobulin level, urine transferrin protein and N-acetyl-β-amino-glucosidase levels. Ammonium chloride load test and sodium hydrogen carbonate reabsorption rate were applied to verify the diagnosis of renal tubular acidosis. Urine and blood osmolality, fluid deprivation/vasopressin test and cranial images were evaluated to assess the renal tubular concentration function and to differentiate diabetes insipidus. Proteinuria based on 24-h urine protein excretion rate was classified as mild (< 1.5 g/day), moderate (1.5-3.5 g/day) and nephrotic (> 3.5 g/day). Urinary ultrasound results were reviewed for evidence of anomalies and renal calcification.

Immunohistochemical staining of C4d
Rabbit anti-human polyclonal C4d antibodies (1:40; Biomedica, Vienna, Austria) were applied on 2-μm formaldehyde-fixed kidney sections. Antigen retrieval was performed by incubating the sections with 0.4% pepsin (Zhongshan Golden Bridge Biotechnology, Beijing, China) for 30 min at 37°C. Sections were incubated with the primary antibody overnight at 4°C and with avidin-free horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulins using the two-step Envision kit (PV9001, Zhongshan Golden Bridge Biotechnology, Beijing, China) for 30 min at 37°C . Sections were treated with a freshly prepared 3-3-diaminobenzidine solution (DAB) (Zhongshan Golden Bridge Biotechnology, Beijing, China) for 3 min and counter-stained with haematoxylin for 30 s.
A positive control for C4d was obtained from dendritic lymphocytes in human tonsil tissues from a patient undergoing an operation for obstructive sleep apnoea hypopnea syndrome (Additional file 1: Figure S1) [26]. Renal tissues from patients with mild lesions who underwent a biopsy for microscopic haematuria were used as negative controls (Additional file 1: Figure S2).

Statistics
Continuous variables with a normal distribution were analysed using Student's t-test or the Mann-Whitney U-test for abnormally distributed data. Categorical data were compared using the Fisher's exact test. Correlations were analysed by calculating Pearson's correlation coefficients for continuous normally distributed variables, Spearman's Rho for nominal and ordinal variables, and performing a linear regression analysis for continuous variables. P < 0.05 was considered statistically significant. Calculations were performed using SPSS statistical software version 11.5 (SPSS, Chicago, IL, USA).

Demographic characteristics of patients with pSS presenting renal involvement
Among the 39 patients included in this study, 21 were diagnosed with TIN, and 18 were diagnosed with GMN. Eight of 18 patients with GMN also had TIN. The male:female ratio was 9:30. The average age of the patients was 42 ± 14.1 (9-72) years at the time of renal biopsy. ANA was positive in 37/39 patients, with different titres: 8 patients at a 1:1280 titre, 8 patients at 1:640, 9 patients at 1:320, 9 patients at 1:160, 2 patients at 1:80, and 1 patient at 1:40.
Hyper-γ-globulinemia was present in 11/36 (28.2%) patients. Elevated serum Ig levels and hypocomplementemia were also observed ( Patients with GMN exhibited more prominent microscopic haematuria (TIN 5/21 vs GMN 13/ 18, P = 0.004) and proteinuria (TIN 0.8 ± 0.6 g/day vs. GMN 3.9 ± 4.9 g/day, P = 0.033). Renal function at kidney biopsy did not differ between the TIN and GMN groups (Table 1). Specifically, 2 patients were diagnosed with acute interstitial nephritis, and the other 19 were diagnosed with chronic interstitial nephritis. All 8 patients with MN exhibited C4d deposition along the glomerular capillary and/ or in mesangium (Fig. 2A1, A2). IF staining revealed IgG deposition in 7/8 of patients with MN and C3 deposition in 2/8 of patients with MN. Only 1 patient displayed weak C1q deposition using IF staining. Sub-epithelial electrondense deposits (EDDs) and occasional mesangial EDDs were also observed in patients with MN using EM, including the patient with negative IgG deposition, confirming the diagnosis of MN ( Table 2). Six of the patients with MN were diagnosed with stage II-III disease and exhibited different degrees of mesangial proliferation. Four cases had segmental thickening of GBM. Three out of five patients with MePGN and 1/2 patients with IgAN had mild mesangial deposition of C4d, but the staining area did not meet the criteria for glomerular C4d positive (Fig. 2A3).
PTC C4d deposition was observed in 12 patients, including 4 patients with minimal deposition (Fig.  2C1), 4 patients with local deposition (Fig. 2C2) and 4 patients with diffuse deposition (Fig. 2C3). The PTC C4d score was positively correlated with the ANA titre (Spearman's Rho = 0.458, P = 0.003). All of 4 patients with diffuse deposition had elevation of serum IgG levels, and 3 of them had concomitant elevation of serum IgA levels and ANA titre at 1:1280, the serum levels of IgM were within normal range for all 4 patients.

Discussion
This study investigated the clinical significance of renal C4d deposition in patients with pSS-related renal lesions which has not been reported before. Patients with pSS-MN exhibited glomerular C4d deposition without obvious concomitant C1q deposition, indicating the involvement of the lectin pathway of complement activation in pSS-MN. Tubulointerstitial and PTC C4d deposition were also observed. The tubulointerstitial-C4d score was correlated with kidney function at the time of kidney biopsy, indicating a possible link between renal interstitial injury and autoantibody-mediated complement activation in patients with pSS. GMN was more prevalent in our study (46.2%) than in previous studies using serum and urine biochemical examinations (13.9-29%) [15,29]. In studies employing pathological investigations, the proportion of GMN ranged from 37.1 to 48.6% [7,14,30]. In the GMN group, there were 8 patients with the pathological diagnosis of MN. The PLA2R/THSD7A antibodies in combination with IgG4 is a useful tool to diagnose intrinsic antibodies related MN [31].As the retrospective nature of our study, there were not enough biopsy tissues for us to test the expression of PLA2R/THSD7A antibodies as well as IgG subtypes. However, 6 of the 8 patients with MN presented with mesangial proliferation and occasional mesangial EDDs were observed in 5 patients, 4 of them had segmental thickening of GBM, which are more likely features of secondary MN [32]. One patient with negative IF was confirmed by detecting sub-epithelial EDDs under the EM. The IF negativity was possibly due to the early disease stage and lack of abundant immune complex. No obvious clinical and pathological features of lupus nephritis were observed both at the time of kidney biopsy and during follow up, we excluded the possibility of MN secondary to SLE. MN was the most common type of glomerular involvement in our study. In previous studies, MPGN was the most common type of GMN [7,33] and was typically accompanied by cryoglobulinemia. The reason for the increased proportion of patients with MN in our study may be due to differences in environmental and genetic factors, because the Chinese population exhibits an increased prevalence of MN and less prevalence of MPGN [34].
The contribution of complement to MN (primary and secondary) was documented elsewhere [35]. Besides C3, the component of the alternative pathway, glomerular deposition of C4d and MBL were detected in PMN, indicating involvement of the lectin pathway [11,13]. The lectin pathway of complement may be possibly triggered by the in situ formed immune complexes comprising the IgG4 type of anti-PLA2R antibodies [9]. In patients with SLE-MN, circulating immune complexes deposition and the subsequent activation of the complement system resulted in the full house deposition of immune complexes and C1q, and all the three pathways of complement activation were involved, namely the classical, lectin [22] and alternative pathway [36]. All patients with MN in our study exhibited glomerular C4d deposition without obvious C1q deposition, indicating local complement activation by the lectin pathway, which is similar to PMN [37] and different from SLE-MN [38], the underlying mechanism remains to be investigated.
The significance of tubulointerstitial C4d deposition is not clear, and few studies are available. Tubular C4d deposition was associated with an increased grade of WHO classification of IgAN [25]. Patients with TBM C4d deposition exhibited a significantly higher score for the viral cytopathic effect on BK nephropathy [39]. In patients with lupus nephritis, no practical significance of TBM-C4d was identified [40]. We identified spread C4d deposition along TBM and in the renal interstitial compartment with different degree. The semi-quantified TI-C4d score was correlated with worse renal function at kidney biopsy and more severe interstitial fibrosis, suggesting potential contribution of complement activation to tubulointerstitial injuries and renal impairment [41]. For most patients with pSS-related TIN, the IF staining results were negative and cellular immunity is considered a main pathological mechanism. The occasional TBM deposition of complement in images of IF staining is typically considered nonspecific [4]. Technically, immunohistochemistry for TBM C4d in paraffinembedded renal allograft tissues is more specific than IF in frozen tissues (4% by IHC vs. 48% by IF) [42]. Thus, we considered that the tubulointerstitial C4d deposits identified by IHC in our study were more inclining to suggest its potential role in the pathoetiological aspect. Due to the complexity of the complement system in diseases [43], the cause-effect relationship of complement activation and renal lesions in patients with pSS can not be demonstrated in this study, which certainly is an area needing to be further investigated.
As a marker for antibody-mediated renal allograft rejection [24], PTC C4d has been less frequently reported in glomerulonephritis, and different pathological mechanisms may be involved. In patients with LN, local immune complexes may be responsible for PTC C4d deposition as EDDs along the basement membrane of PTC were identified in 24/31 (77.4%) of patients with LN presenting PTC C4d deposition, whereas none of the patients with acute antibody-mediated rejection exhibited deposition [44]. In addition to autoantibodies, local shear stress on endothelial cells may play a role in the renal crisis in patients with scleroderma [45,46], and PTC C4d correlated with an increased risk of renal failure [47]. In our study, PTC C4d deposition was identified in 12 patients. The PTC C4d score was positively correlated with the ANA titre. Thus, PTC C4d deposition in pSS-related lesions was possibly mediated by the effects of autoantibodies, which is in accordance with the detection of anti-carbonic anhydrase II antibodies [16] and antibodies against intercalated cells [18] in pSS-related renal tubular lesions. As few patients exhibited diffuse PTC C4d deposition, the significance of PTC C4d deposits in pSS-related renal injury remains to be further clarified.
In consistence with Kidder et al' s study showing similar renal survival rate between the GMN and the TIN pathological entities [7], our study also showed that renal function at kidney biopsy in patients with GMN did not differ from those with TIN. PTC C4d deposition was a prognostic marker in renal allograft [48]. Although 10 patients in our study exhibited improvement of renal function, two of the 4 patients with PTC C4d diffuse deposition did not show significant improvement of renal function during follow-up, indicating poor response to treatment. The other two patients with PTC C4d diffuse deposition were lost follow up shortly after discharge, which resulted in the small sample size, the impact of C4d on renal function in patients with pSS related renal impairments remains to be explored in a prospective study with large sample size.
The retrospective nature of this study may lead to possible bias, as more patients with GMN were listed in the kidney biopsy database. Other limitations of our study were that we did not examine the serum anti-PLA2R antibody titre, and an IgG sub-type analysis was not performed in renal tissues. The inadequacy of biopsy tissues hindered the testing of other complement components (such as lectin and C5b-9).

Conclusions
This preliminary study investigated C4d deposition in patients with pSS with renal involvement. Glomerular C4d positivity was detected in 100% of patients with pSS-MN, in the absence of C1q, indicating possible complement activation through the lectin pathway. The tubulointerstitial C4d deposition observed in our patients with pSS indicated possible involvement of the MBL pathway of complement activation in pSS-renal lesions. Tubulointerstitial C4d deposits might be a candidate of pathological markers of severe renal injury in patients with pSS-related renal impairments, which remains to be further investigated.

Take Home Messages
The MBL pathway of complement activation may be involved in pSS related MN and possibly mediated pSS related tubulointerstitial injury.