Relationship between complement deposition and the Oxford classification score and their combined effects on renal outcome in immunoglobulin A nephropathy

Abstract

Background

Complement activation has been highlighted in immunoglobulin (Ig) A nephropathy pathogenesis. However, whether the complement system can affect the downstream phenotype of IgA nephropathy remains unknown. Herein, we investigated the association of mesangial C3 deposition with the Oxford classification and their joint effects on worsening kidney function.

Methods

We investigated 453 patients with biopsy-proven IgA nephropathy. C3 deposition was defined as an immunofluorescence intensity of C3 ≥2+ within the mesangium. The subjects were classified according to the combination of C3 deposition and Oxford classification lesions. The primary endpoint was a composite of ≥30% decline in the estimated glomerular filtration rate or an increase in proteinuria ≥3.5 g/g during follow-up.

Results

Among the Oxford classification lesions, mesangial hypercellularity (M1), segmental glomerulosclerosis (S1) and tubulointerstitial fibrosis (T1–2) and crescentic lesion significantly correlated with C3 deposition. During a median follow-up of 33.0 months, the primary endpoint occurred more in patients with M1, S1, T1–2 and mesangial C3 deposition than in those without. In individual multivariable-adjusted Cox analyses, the presence of M1, S1, T1–2 and C3 deposition was significantly associated with higher risk of reaching primary endpoint. In the combined analyses of C3 deposition and the Oxford classification lesions, the hazard ratios for the composite outcome were significantly higher in the presence of C3/M1, C3/S1 and C3/crescent than in the presence of each lesion alone.

Conclusions

Complement deposition can strengthen the significance of the Oxford classification, and the presence of both components portends a poorer prognosis in IgA nephropathy.

complement, IgA nephropathy, Oxford classification

INTRODUCTION

Immunoglobulin (Ig) A nephropathy is the most frequently occurring type of primary glomerulonephritis worldwide, and leads to end-stage renal disease in up to 40% of patients within a few decades after diagnosis [1, 2]. Since Dr Jean Berger first described the unique feature of IgA nephropathy as intercapillary deposition of IgA–IgG within the glomeruli in 1968 [3], many studies have been performed to delineate the pathogenesis of IgA nephropathy. Although circulating galactose-deficient IgA1 is the important initiating factor, this abnormal IgA1 alone does not cause the disease. IgA1 binds with an anti-glycated antibody to form pathogenic IgA1-containing immune complexes; these immune complexes will then eventually deposit in the mesangium and trigger sequential downstream cascades to cause kidney injury [4].

Among the events within the glomeruli upon the deposition, complement activation has been considered to play a key role in the glomerular inflammation associated with IgA nephropathy; the mesangial co-deposition of C3, the product of the alternative pathway, with IgA1 is frequently observed in biopsy specimens via immunofluorescence during the initial phase of IgA nephropathy [5, 6]. In fact, complement activation by IgA1-containing immune complexes is thought to intensify kidney injury [7–9]. In addition, our group and others have shown that high-degree C3 deposition within the mesangium can predict adverse renal outcomes [10–13].

Recently, the Working Group of the International IgA Nephropathy Network and the Renal Pathology Society proposed the Oxford classification to have five well-described key pathologic phenotypes of IgA nephropathy: mesangial hypercellularity (M1), endocapillary proliferation (E1), segmental glomerulosclerosis/adhesion (S1), tubular atrophy or interstitial fibrosis (T1 or T2) and cellular or fibrocellular crescent (C1 or C2) [14]. The clinical usefulness of this classification has been tested and verified in many studies conducted to date [15–17]. These Oxford-MEST-C lesions are the consequences of the pathogenic process of IgA nephropathy. Given the potential role of complement activation in IgA nephropathy, it would be interesting to explore how the morphologic features are linked to this pathogenic step. Herein, we studied the association between C3 deposits and individual components of the Oxford classification, and their joint effects on the progression of IgA nephropathy.

MATERIALS AND METHODS

Ethics statement

This study was performed in accordance with the Declaration of Helsinki principles and approved by the institutional review board (IRB) of Yonsei University Health System (YUHS) Clinical Trial Center. Although all patients in this study were informed regarding the description of the investigations, it was conducted as a medical record-based retrospective analysis, and the included subjects were anonymized. Therefore, the IRB approved the exemption from obtaining written consent.

Study population

We conducted an observational study in 528 patients with biopsy-proven primary IgA nephropathy between October 2009 and December 2016 from the YUHS and National Health Insurance Service Ilsan Hospital. A flow diagram of the selection of the participants is presented in Figure 1. We excluded patients who met the following criteria: (i) age of <18 or ≥75 years; (ii) missing data during follow-up; (iii) history of kidney transplantation; or (iv) Henoch–Schonlein purpura nephritis. Finally, a total of 453 patients with IgA nephropathy were included in this analysis (Figure 1).

FIGURE 1

Flow diagram of the study.

Clinical, biochemical and histologic data collection

Using the database in the Glomerulonephritis Registry of the YUHS, the demographic, clinical and biochemical data of the patients with IgA nephropathy at the time of renal biopsy, including age, sex, blood pressure and body mass index, were retrieved and considered baseline data. Body mass index was calculated as weight/height² (kg/m²). The following biochemical laboratory data were also collected: levels of blood urea nitrogen, serum creatinine, random urine protein-to-creatinine ratio (UPCR), urine red blood cell, serum IgA, hemoglobin, serum albumin, total cholesterol, triglyceride, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) and high-sensitivity C-reactive protein (hs-CRP). The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation [18]. Microscopic hematuria was defined as a number of red blood cells in the urinary sediments of ≥5 per high-power microscopic field. The serum total cholesterol, HDL-C, LDL-C and triglyceride levels were measured through enzymatic colorimetry using an autoanalyzer (Hitachi 7150; Hitachi Ltd, Tokyo, Japan), while the hs-CRP levels were determined through a latex-enhanced immunonephelometric method using a BNII analyzer (Dade Behring, Newark, DE, USA). Follow-up data, such as blood pressure, UPCR and eGFR, were recorded at 3-month interval visits. All renal biopsy specimens were re-assessed by one pathologist blinded to the patients’ clinical data according to the Oxford classification [16]. Crescentic lesions were defined as C0, C1 or C2 according to the new Oxford-MEST-C classification [19]. However, because only 17 patients had C2 lesions, we simply designated crescentic (+) lesions to encompass both C1 and C2 lesions. Similarly, there were only seven patients with T2, and thus T1 and T2 together were incorporated into T1–2. We generally followed the glucocorticoid protocol suggested by Pozzi et al. [20]. Glucocorticoid users were defined as patients who received glucocorticoids for ≥1 month.

Follow-up and endpoints

The primary endpoint was a composite of the onset of a ≥30% decline in the eGFR or an increase in proteinuria ≥3.5 g/g during follow-up. Because of relatively short-term observation of our study and preserved kidney function in the study subjects, we used an outcome of a ≥30% decline in the eGFR as previously suggested [21–23]. In addition, we followed the definition of proteinuria outcome as suggested by Coppo et al. [24]. This endpoint was defined as a sustained decrease in the eGFR of ≥30% or persistently elevated proteinuria ≥3.5 g/g for at least two consecutive measurements. The first of these consecutive measurements was retrospectively designated as the study endpoint.

Statistical analysis

Continuous variables were presented as means and standard deviations for data with a normal distribution or medians with interquartile ranges for data with a skewed distribution. Categorical variables were expressed as counts and percentages. The normality of distribution was ascertained using the Kolmogorov–Smirnov test and Shapiro–Wilk test, and skewed continuous parameters were logarithmically transformed before use in the parametric procedures. P-value for trend was calculated by Jonckheere–Terpstra trend test or Cochran–Armitage trend test, as appropriate. To identify the correlation between mesangial C3 deposition and the Oxford-MEST-C classification, logistic regression analysis was used. Based on the findings of these analyses, the study subjects were classified according to the combination of C3 deposits and each lesion in the Oxford classification. To compare differences among the categories of mesangial C3 deposits and the Oxford classification lesions, analysis of variance with Bonferroni correction or the Kruskal–Wallis test was used for continuous variables and Chi-square test or Fisher’s exact test for categorical variables. Cox proportional hazard models were constructed to determine the association between the Oxford-MEST-C classification with or without C3 deposits and the development of a 30% decline in the eGFR. Firstly, a univariable analysis was performed, and significant variables identified in this analysis were then incrementally included in the multivariable analysis. Model 1 included age, sex and history of hypertension. Model 2 additionally adjusted the mean arterial pressure, proteinuria and eGFR in addition to those in Model 1. Finally, Model 3 was created after further adjustment for glucocorticoid treatment. The adjusted cumulative renal survival curves were plotted after the same level adjustments included in the Cox models. Survival time was defined as the interval between the time of biopsy and the onset of the endpoint or last follow-up. All results were expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). Statistical analyses were performed using Stata version 15 (StataCorp., College Station, TX, USA) and R version 3.5.2 (www.r-project.org; R Foundation for Statistical Computing, Vienna, Austria). P <0.05 were considered significant.

RESULTS

Baseline characteristics according to mesangial C3 deposition

Table 1 presents the baseline characteristics of 453 patients according to mesangial C3 deposition. The mean age of the patients was 40.3 years, and 45.3% of them were men. The mean eGFR was 89.5 mL/min/1.73 m², the median UPCR was 0.77 g/g Cr. There were 87 (19.2%), 168 (37.1%) and 198 (43.7%) patients having C3 staining of negative, trace or 1+ and ≥2+, respectively. The patients with C3 deposition were significantly younger and had a higher prevalence of microscopic hematuria, lower serum C3 level and higher proportion of M1, S1 and crescentic (+) lesions than those without. However, there were no significant differences in the prevalence of hypertension, mean arterial blood pressure (MAP), kidney function, prevalence of proteinuria and proportion of E1 lesions. Because C3 deposition of ≥2+ was associated with poor renal outcome in our previous study [25] and this association was consistent in this study (Supplementary data, Table S1), we defined mesangial C3 deposition as immunofluorescence intensity of C3 ≥2+ within the mesangium.

Table 1

Baseline characteristics according to mesangial C3 deposition

Variables	Total (n = 453)	Mesangial C3 deposition			P-value for trend
		C3 negative (n = 87)	C3 trace/1+ (n = 168)	C3 2+ or higher (n = 198)
Age (years)	40.3 ± 13.3	43.2 ± 15.2	40.6 ± 13.1	38.6 ± 12.2	0.013
Men (%)	205 (45.3)	40 (46.0)	79 (47.0)	86 (43.4)	0.596
Body mass index (kg/m2)	23.1 ± 3.4	23.2 ± 3.2	23.5 ± 3.8	22.8 ± 3.1	0.126
Hypertension (%)	297 (65.6)	55 (63.2)	111 (66.1)	131 (66.2)	0.673
MAP (mmHg)	96.3 ± 12.8	96.6 ± 11.5	96.2 ± 13.8	96.2 ± 12.5	0.678
UPCR (g/g Cr)	0.77 (0.38–1.51)	0.67 (0.22–1.37)	0.76 (0.38–1.58)	0.82 (0.43–1.51)	0.104
Microscopic hematuria (%)	351 (77.5)	56 (64.4)	128 (76.2)	167 (84.3)	<0.001
eGFR (mL/min/1.73 m2)	89.5 ± 28.5	91.4 ± 27.4	86.8 ± 28.1	91.0 ± 29.2	0.569
BUN (mg/dL)	15.5 ± 7.2	15.5 ± 8.0	15.8 ± 7.3	15.4 ± 6.9	0.671
Hemoglobin (g/dL)	13.1 ± 1.7	13.3 ± 1.5	13.0 ± 1.7	13.0 ± 1.7	0.190
Albumin (g/dL)	4.0 ± 0.5	4.0 ± 0.6	4.0 ± 0.5	4.0 ± 0.4	0.120
Total cholesterol (mg/dL)	187.0 (165.0–215.0)	185.0 (152.0–218.0)	187.0 (165.5–211.5)	188.0 (168.0–215.0)	0.267
Triglyceride (mg/dL)	109.0 (76.0–158.5)	121.0 (82.0–165.0)	117.0 (74.0–155.0)	102.0 (78.0–147.0)	0.186
C3 (mg/dL)	110.8 ± 20.3	113.9 ± 23.1	113.6 ± 21.4	107.1 ± 17.2	0.012
C4 (mg/dL)	27.3 ± 8.8	27.5 ± 1.7	27.1 ± 8.4	27.4 ± 7.7	0.423
hs-CRP (mg/L)	1.3 (0.6–3.6)	1.2 (0.8–2.9)	1.6 (0.8–4.8)	1.1 (0.5–3.2)	0.426
Serum IgA (mg/dL)	310.0 (247.0–390.0)	293.0 (241.0–380.0)	315.0 (253.0–398.0)	314.0 (246.0–383.0)	0.552
Oxford classification
M1 (%)	86 (19.0)	6 (6.9)	31 (18.5)	49 (24.7)	<0.001
E1 (%)	128 (28.3)	18 (20.7)	49 (29.2)	61 (30.8)	0.108
S1 (%)	307 (67.8)	43 (49.4)	115 (68.5)	149 (75.3)	<0.001
T1–2 (%)	40 (8.8)	5 (5.7)	14 9 (8.3)	21 (10.6)	0.174
Crescent (+) (%)	97 (21.4)	12 (13.8)	34 (20.2)	51 (25.8)	0.021
Glucocorticoid treatment (%)	84 (18.6)	16 (18.4)	29 (17.4)	39 (19.7)	0.704

Data are presented as mean ± SD or median (interquartile range) unless otherwise indicated.

BUN, blood urea nitrogen.

Clinical outcomes according to mesangial C3 deposition

The primary kidney outcome events occurred in 5 (5.8%), 18 (10.7%) and 27 (13.6%) patients having C3 staining of negative, trace or 1+ and ≥2+, respectively. In multivariable Cox model, C3 deposition ≥2+ was associated with significantly higher risk of adverse kidney outcome (Supplementary data, Table S2). However, the hazard ratio (HR) did not differ between C3 negative and C3 trace or 1+. This finding was consistent with our previous study [25]. Thus, we used binary classification of C3 staining intensity as <2+ versus ≥2+ in the following analyses.

Correlation between mesangial C3 deposition and the Oxford classification score

Next, we examined the correlation between mesangial C3 deposition and the Oxford-MEST-C classification using multivariable logistic regression analysis. The results showed that the odds ratios (95% CI; P-value) for C3 deposition in the presence of M1, E1, S1, T1–2 and crescent (+) were 1.84 (1.11–3.04; P = 0.018), 1.24 (0.81–1.91; P = 0.322), 1.80 (1.17–2.76; P = 0.007), 2.03 (1.01–4.13; P = 0.048) and 1.58 (1.00–2.47; P = 0.048), respectively (Table 2). This relationship is collectively presented in Supplementary data, Figure S1, showing the histopathologic features of a 38-year-old male patient with mesangial C3 deposition and M1, S1 and crescentic lesions.

Table 2

Multivariable-adjusted logistic regression analysis for mesangial C3 deposition according to the Oxford classification

The Oxford classification lesions	Odds ratio (95% CI)	P-value
M1	1.84 (1.11–3.04)	0.018
E1	1.24 (0.81–1.91)	0.322
S1	1.80 (1.17–2.76)	0.007
T1–2	2.03 (1.01–4.13)	0.048
Crescent (+)	1.58 (1.00–2.47)	0.048

Baseline characteristics according to mesangial C3 deposition combined with the Oxford classification score

Based on the findings above, the study subjects were classified according to the combination of C3 deposition and each lesion in the Oxford classification, and their baseline characteristics are summarized in Supplementary data, Tables S3–S6. The patients with both mesangial C3 deposition and M1 lesions were younger and had a higher prevalence of hematuria and more components of the other Oxford scores than those with either mesangial C3 deposition or M1 lesions alone. Similar findings were observed in patients with C3 deposition and S1, T1–2 or crescentic (+) lesions.

Individual association of mesangial C3 deposition or the Oxford classification with kidney function decline in IgA nephropathy

During a median follow-up of 33.0 months, 50 (11.0%) patients reached the composite outcome. There were 46 and 17 patients who developed a ≥30% reduction of the eGFR and an increase in proteinuria ≥3.5 g/g, respectively (Supplementary data, Table S1). The composite primary endpoint occurred more in patients with M1 (24.4% versus 7.9%), S1 (13.7% versus 5.5%), T1–2 (35.0% versus 8.7%), crescentic (+) (13.4% versus 10.4%) and mesangial C3 deposition (13.6% versus 9.0%) than in those without each corresponding lesion. The HRs for the M1, S1, T1–2 and mesangial C3 deposition were 2.10 (95% CI 1.12–3.94; P = 0.020), 4.48 (95% CI 1.71–11.71; P = 0.002), 3.83 (95% CI 1.97–7.43; P < 0.001) and 2.70 (95% CI 1.41–5.17; P = 0.003), respectively, in multivariable Cox model after adjusting for age, sex, history of hypertension, proteinuria, eGFR, MAP and glucocorticoid treatment (Table 3).

Table 3

HRs for a composite of a ≥ 30% decline in eGFR from baseline values or an increase in proteinuria ≥3.5 g/g according to C3 deposition or the Oxford classification

The Oxford classification lesions	Events/total (%)	Unadjusted		Model 1		Model 2		Model 3
		HR (95% CI)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value
M1	21/86 (24.4)	2.83 (1.59–5.04)	<0.001	2.88 (1.59–5.21)	<0.001	2.10 (1.12–3.93)	0.021	2.10 (1.12–3.94)	0.020
S1	42/307 (13.7)	3.35 (1.50–7.48)	0.003	3.09 (1.36–7.00)	0.007	4.48 (1.72–11.70)	0.002	4.48 (1.71–11.71)	0.002
T1–2	14/40 (35.0)	5.41 (2.89–10.10)	<0.001	4.61 (2.44–8.73)	<0.001	3.87 (1.99–7.50)	<0.001	3.83 (1.97–7.43)	<0.001
Crescent (+)	13/97 (13.4)	1.30 (0.68–2.51)	0.429	1.26 (0.65–2.44)	0.497	1.22 (0.62–2.41)	0.556	1.21 (0.62–2.39)	0.576
C3 deposition	27/198 (13.6)	2.52 (1.41–4.52)	0.002	2.98 (1.64–5.45)	<0.001	2.71 (1.42–5.18)	0.003	2.70 (1.41–5.17)	0.003

Model 1: adjusted for age, sex and history of hypertension.

Model 2: Model 1 + mean arterial blood pressure, proteinuria and eGFR.

Model 3: Model 2 + glucocorticoid treatment.

Combined effects of mesangial C3 deposition and the Oxford classification on kidney function decline in IgA nephropathy

Next, we evaluated whether the combination of mesangial C3 deposition and the Oxford scoring system could adversely affect the deterioration of kidney function more in IgA nephropathy. The primary endpoint occurred more in patients with double positivity for C3 deposition and M1, S1, T1–2 or crescentic (+) lesions in the Oxford classification than in those without these lesions and those with either C3 deposition or each lesion alone (Supplementary data, Table S7). Multivariable Cox models showed that the risk of adverse renal outcomes was significantly the highest in these double-positive lesions. The HRs were 4.22 (95% CI 1.80–9.89; P = 0.001) for C3 deposition and M1, 2.55 (95% CI 1.11–5.90; P = 0.028) for C3 deposition only and 1.93 (95% CI 0.72–5.21; P = 0.194) for M1 lesions only as compared with the absence of C3 deposition and M0 lesions. This association was consistently observed in the other combinations of C3 deposition with S1 or crescentic (+) lesions and the risk of kidney function decline further increased when both lesions were present. The corresponding adjusted HRs were 7.17 (95% CI 2.33–22.06; P = 0.001) and 2.96 (95% CI 1.20–7.29; P = 0.019) (Figure 2; Supplementary data, Table S7). An adjusted cumulative renal survival curve also showed that renal events were more frequent in the combination of C3 deposition with M1, S1, T1–2 or crescentic (+) lesions than in single positive lesions of C3 deposition or the respective Oxford classification (Figure 3). For the combined C3 deposition and T1–2 lesion, the HR was 3.57 (95% CI 1.34–9.53; P = 0.011), but the presence of both C3 deposition and T1–2 did not have higher risk of the composite outcome than t-score alone and the HRs were similar between these two groups.

FIGURE 2

HRs for a composite of a ≥30% decline in eGFR from baseline values or an increase in proteinuria ≥3.5 g/g according to the combination of mesangial C3 deposition and the Oxford classification. The Cox proportional hazards model was adjusted for age, sex, history of hypertension, log-transformed UPCR, eGFR, MAP and glucocorticoid treatment. (A) C3 deposition and M-score, (B) C3 deposition and S-score, (C) C3 deposition and T-score and (D) C3 deposition and C-score; *P < 0.05.

FIGURE 3

Adjusted-cumulative survival curve for a composite of a ≥30% decline in the eGFR from baseline values or an increase in proteinuria ≥3.5 g/g according to the categorization of mesangial C3 deposition and the Oxford classification. (A) C3 deposition and M-score, (B) C3 deposition and S-score, (C) C3 deposition and T-score and (D) C3 deposition and C-score.

DISCUSSION

In this study, we found that glomerular complement deposition significantly correlated with M1, S1, T1–2 and crescentic lesion of the Oxford classification. In addition, we also showed that the presence of C3 deposition together with these lesions was significantly associated with an increased risk of kidney function decline. Our findings underscore the importance of complement deposition in conjunction with the Oxford classification in light of the clinical utility in predicting future adverse kidney outcomes.

The Oxford scoring system is a histologic classification designed to improve the predictability of the prognosis of patients with IgA nephropathy even after considering clinical factors, such as proteinuria, baseline renal function and blood pressure [26]. Neither the original nor the revised Oxford classification system [19, 26] covered biopsy-based complement immunostaining, although there were a number of studies suggesting that complement activation has a crucial role in the clinical outcomes of IgA nephropathy [10–13] and complement activation alone can induce ruinous glomerular lesions [27]. Our study is meaningful because it is the first study to show that complement deposition has a greater impact on the prognosis of IgA nephropathy by a synergistic action with the downstream phenotype of IgA nephropathy than that by complement deposition or individual lesions in the Oxford classification alone.

The underlying mechanisms responsible for the cooperative effects of complement deposition and M1 lesions can be inferred from previous experimental studies. In general, complement activation can cause tissue injury through formation of the C5b-9 membrane attack complex. C5b-9 can induce mesangial stress, leading to the elevated production of fibronectin, transforming growth factor-β (TGF-β) and interleukin-6 [28]. In addition, mesangial cells start to lose the contractile phenotype and exert proliferative features in response to exogenous administration of C3a [29]. Furthermore, these cells produce C3 under an inflammatory environment and upon binding with galactose-deficient IgA1 [29, 30]. Thus, this mechanistic link can explain the increased risk of disease progression in double-positive lesions of C3 deposition and M1 in our study.

Another interesting finding of our study is the association between C3 deposition and S1 lesions. In IgA nephropathy, synechiae is considered as a consequence of podocyte injury [31]. The role of the complement system in podocyte injury has been demonstrated in other forms of glomerulopathy. C5b-9, the ultimate product of complement activation, is inserted to podocyte cell membranes to secrete various proteases, oxidants, cytokines and components of the extracellular matrix [32]. This eventually impairs the function of the glomerular basement membrane and causes apoptosis and glomerular scarring [33, 34]. Previous experimental studies by Lai et al. have suggested a potential cross-talk between mesangial cells and podocytes in IgA nephropathy [35, 36]. They demonstrated that polymeric IgA did not bind to podocytes. However, in response to IgA-conditioned medium prepared from patients with IgA nephropathy, mesangial cells secreted tumor necrosis factor-α and TGF-β, which induced podocyte de-differentiation and loss [37]. These detrimental effects were attenuated by a neutralizing antibody against tumor necrosis factor-α and TGF-β. Given the role of the complement system in mesangial cells and potential interactions between different cells in IgA nephropathy, the significant association of C3 deposition with S1 lesions can support the plausible mechanistic link between complement-derived activation of mesangial cells and podocyte injury.

There has been much controversy on the role of crescentic lesions as a prognostic factor in IgA nephropathy. Nevertheless, the new Oxford classification system has adopted crescentic lesions as an independent predictor of adverse outcomes because a large proportion of these lesions, particularly when they are present in ≥25% of the glomeruli, were significantly associated with a faster decline in kidney function [19, 38]. In our study, C1 or C2 lesions alone did not predict adverse renal outcomes. Notably, the presence of crescentic lesions was significantly correlated with C3 deposition and an increased risk of progression of kidney failure when C3 deposition is present. Although the mechanisms for this association are unclear, our finding is supported by those of previous studies, suggesting that complement can play an important role in pauci-immune necrotizing and crescentic glomerulonephritis [39, 40]. To our knowledge, no studies have examined the association between complement and crescent formation in IgA nephropathy. Further studies should explore the mechanistic link responsible for this association.

Tubulointerstitial fibrosis is known as the strongest predictor of kidney function decline. In line with this notion, T-score well predicted adverse outcome in this study. Notably, the presence of both C3 deposition and T-score did not have higher risk of the composite outcome than T-score alone. This finding suggests that predictive ability of T-score is too strong and thus can weaken the association of other lesions with kidney outcome. To date, no studies have found the superiority of certain biomarkers over tubulointerstitial fibrosis in predicting kidney disease progression. However, this study included only 40 patients with T-score, partly because kidney biopsy is not commonly recommended in advanced stages of IgA nephropathy. This might result in lack of statistical power to detect the difference between groups, thus our findings need to be viewed with caution.

Our study showed that endocapillary hypercellularity poorly correlated with mesangial C3 deposition. This finding is consistent with our previous observation [12]. We surmise that the lack of association between these two lesions is possible because of the different location and timeframe of the individual pathogenic process. Endothelial cells reside in the vascular compartment, while mesangial cells are located within the mesangium, in which the complement system can be activated in response to various stimuli. In addition, endothelial proliferation can initially occur and then subside at an earlier time point before complement activation. However, two previous studies reported contradicting results showing a significant association between C3 deposition and endothelial proliferation [13, 41]. Notably, there were many clinical differences among studies. In particular, E1 lesions were less prevalent in our cohort than in those two studies; E1 lesions were observed in 42% of the subjects in the study by Katafuchi et al. [41], which is two times higher than that in our cohort. In addition, there were more patients who received steroid treatment in their study. These different clinical features can partly explain the discrepant findings among the studies. Nevertheless, glomerular cells can interact with each other, and thus, propagate injury as aforementioned. This complex relationship warrants further investigation.

Our study has several limitations. Firstly, our findings are limited by the observational design, uncertain causality, relatively small number of patients and possible influence of confounding factors that could not be captured. Therefore, the findings should be interpreted with caution. Secondly, only mesangial C3 deposition was investigated, and there was no additional staining for the other complement components, including C3a, C5a or C5b-9 complex, that are known as active players in tissue injury. In addition, growing attention has been paid to the mannose-binding lectin (MBL) pathway as an emerging prognostic marker of IgA nephropathy. We did not evaluate the relevance of the MBL pathway with kidney disease progression in this study. However, C3 deposition is most frequently found among the complement components in IgA nephropathy [42], and whether the MBL pathway can be a better prognostic factor than the alternative pathway is unknown. Detailed analyses on this were beyond the scope of this study, and this interesting issue is currently under investigation by our group. Thirdly, the inclusion of only Korean patients may limit the generalizability of our findings to other ethnic populations. Finally, we could not evaluate the effect of corticosteroids on the combined lesions with C3 deposition and the Oxford classification score, because only 7.4% of the patients were treated with such. However, recent randomized controlled trials have questioned the role of immunosuppressive treatment [43, 44], and it is unknown whether this treatment is particularly effective for specific lesions in the Oxford classification. Future clinical trials should explore this issue in depth and hopefully will be able to offer a helpful basis for decision making on appropriate indication of immunosuppressive treatment.

In conclusion, we showed that mesangial C3 deposition was significantly correlated with M1, S1 and crescentic lesions in the patients with IgA nephropathy. In addition, the combination of mesangial C3 deposition with each lesion in the Oxford classification was more predictive of kidney function decline than single lesions alone. These findings suggest that mesangial complement deposition can strengthen the significance of the Oxford classification, and the presence of both components portends a poorer prognosis in IgA nephropathy.

FUNDING

This research was supported by faculty research grant of Yonsei University College of Medicine for 2015, Seoul, Korea (4-2015-1084) and Research of Korea Centers for Disease Control and Prevention (2019ER690100).

AUTHORS’ CONTRIBUTIONS

S.H.H. and S.P. contributed to research idea and study design; S.-W.K., T.-H.Y., S.H.H., H.J.C., D.-R.R., E.W.K., T.I.C., J.T.P. and S.P. were involved in data acquisition; S.H.H., S.P. and H.W.K. were responsible for data analysis/interpretation; S.H.H., S.P. and H.W.K. contributed to statistical analysis; S.-W.K., H.J.J., T.-H.Y., S.H.H., B.J.L. and J.T.P. were responsible for data analysis/interpretation supervision or mentorship. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare and no financial disclosures to make with respect to this work. The results presented in this paper have not been published previously in whole or part, except in abstract format.

Author notes

Seohyun Park and Hyung Woo Kim contributed equally to this work.