The expression of NOD2, NLRP3 and NLRC5 and renal injury in anti-neutrophil cytoplasmic antibody-associated vasculitis

Thirty-four patients with active AAV receiving renal biopsies before immunosuppressive therapy in Peking University First Hospital from 2016 to 2018 were enrolled in this study. All the patients met the criteria of the 2012 Chapel Hill Consensus Conference definition for AAV [1]. No patient had coexistence of other kidney diseases, secondary vasculitis, other autoimmune diseases or infections at the time of renal biopsy according to the clinical and laboratory data. Renal tissue from normal parts of nephrectomized kidneys from seventeen patients with solitary renal cell carcinoma was collected as normal controls. They were identified normal with regular light microscopy, immunofluorescence and electron microscopy. Renal specimens from nineteen patients with minimal change disease (MCD) and twenty patients with class IV lupus nephritis (LN) according to the International Society of Nephrology/Renal Pathology Society (ISN/RPS) [25] were collected as disease controls. Informed consent was signed by each participant. This research was in compliance with the Declaration of Helsinki and approved by the ethics committee of Peking University First Hospital.

Renal histopathology of AAV patients was evaluated according to the standardized protocols [26‐28] described previously. In brief, the lesions of glomeruli including fibrinoid necrosis, glomerulosclerosis and total crescents were calculated as affected percentage of the total number of glomeruli in biopsies. The lesions of tubulointerstitium were semi-quantitatively scored according to the affected area of tubulointerstitium: tubular atrophy, interstitial fibrosis (− for 0%, + for 0–50% and ++  for > 50%) and interstitial infiltration (− for 0%, + for 0–25%, ++  for 25–50% and +++ for > 50%). Furthermore, each biopsy was classified as focal, crescentic, mixed or sclerotic category, according to the pathologic classification system for ANCA-associated glomerulonephritis proposed by Berden et al. [29].

Formaldehyde-fixed renal slides of AAV patients, normal and disease controls were dewaxed in xylene ethanol at room temperature and rehydration through graded ethanol. Antigen retrieval was then performed by heating the slides in citrate buffer (0.01 M, pH 6.0) in a pressure cooker at the highest pressure for 2 min and 30 s. After cooled to room temperature and washed with PBS, the slides were immersed in freshly prepared 3% hydrogen peroxide for 10 min at room temperature to quench endogenous peroxidase activity. Non-specific staining was blocked by incubating specimens with 3% bovine serum albumin (BSA) in PBS at room temperature for 60 min. After the removal of blocking BSA without washing, primary antibodies for NOD2 (1:500, 20,980–1-AP; ProteinTech Group, Chicago, USA), NLRP3 (1:200, ab214185; Abcam, Cambridge, UK) or NLRC5 (1:500, ab105411; Abcam, Cambridge, UK) were added and incubated overnight at 4 °C. For negative controls, specimens were incubated with 3% BSA instead of primary antibodies. Secondary antibodies (PV9001, ZSGB-Bio, Beijing, China) were incubated with the specimens at 37 °C for 20 min. Sections were developed in fresh hydrogen peroxide plus 3–3′-diaminobenzidine tetra hydrochloride solution for 40 s. Nuclear staining was performed by incubating the specimens with haematoxylin for 8 min. Finally, the slides were dehydrated through ethanol and xylene and sealed with neutral gum. All the glomeruli in a section at × 400 and at least 10 fields of tubulointerstitial vision per kidney section at × 400 were observed blindly for quantitative assessments of immunohistochemical staining.

The staining results of NLRs were evaluated by the Image Pro Plus (version 6.0; Media Cybernetics, Dallas, TX, USA): The optical intensity threshold of pictures was corrected to 0–255. Hue-Saturation-Intensity model was used to select stained area and calculate the integrated optical density and relative area in pixel. The mean optical density = integrated optical density/relative area.

We performed double immunofluorescence of NLRs and specific markers of various cell types, including renal intrinsic and inflammatory cells in AAV patients, to determine concrete locations of NLRs in the kidney from AAV patients. Glomerular endothelial cells, mesangial cells and podocytes were identified using immunofluorescence staining with primary antibodies against CD31 (1:50, sc53411; Santa Cruz Biotechnology, CA), integrin-α8 (1:50, sc365798; Santa Cruz Biotechnology, CA) and synaptopodin (1:50, MAB8977; R&D Systems, Minneapolis, MN, USA) respectively. Infiltrating cells (monocytes/macrophages) were identified using immunofluorescence staining with antibodies against CD68 (1:100, ab31630; Abcam, Cambridge, UK). After dewaxed, antigen retrieval and non-specific staining blocked, the renal specimens were incubated with mixed primary antibodies overnight at 4 °C and secondary antibody of Alexa Fluor (AF488)-labeled donkey anti-rabbit IgG (1:200; Jackson ImmunoResearch, West Grove, PA, USA) and cyanin 3 (Cy3)-labeled donkey anti-mouse IgG (1:200; Jackson ImmunoResearch) at 37 °C for 1 h. After washed with PBS, the specimens were stained with 4,6-diamidino-2-phenylindole (DAPI) and eventually mounted with Mowiol. For negative controls, the mixture of primary antibodies was replaced with 3% BSA. Confocal images were captured with a Zeiss LSM 710 confocal microscope (Zeiss, Jena, Germany). Images were exported from the ZEN 2012 (blue edition) microscopy software.

The immortalized human podocyte cell line was a gift from Prof. Saleem MA. Immortalized human podocytes were grown at 33 °C in RPMI 1640 media supplemented with 10% fetal bovine serum (FBS, Gibco, Thermo Fisher Scientific, MA, USA) and insulin, transferrin, selenium, ethanolamine solution (ITS, 41400045, Thermo Fisher Scientific, MA, USA). To allow the cells to differentiate, the cells were grown to 70–80% confluence, trypsinized and transferred at a dilution of 1:4 in growth media and cultured at the non-permissive temperature of 37 °C (in 5% CO₂). After 14 days, the differentiated podocytes were starved for 12 h followed by stimulation with 10 ng/ml human tumor necrosis factor α (TNF-α, R&D Systems, Minneapolis, MA, USA) for 12 h to imitate the inflammatory condition in AAV [30, 31].

To determine messenger RNA (mRNA) expression of three NLRs, total RNA was extracted from differentiated podocytes using Trizol. After reverse transcription of RNA, the mRNA level of each target gene was quantified on ViiATM7 Dx Real-Time PCR (Applied Biosystems, Foster City, CA, USA) using SYBR Green Universal PCR Master Mix (Applied Biosystems) according to the manufacturer’s instructions. The fold change of expression levels were calculated using the comparative ΔCt method. Primers used for PCR: NOD2 forward ACCTTTGATGGCTTTGACG and reverse CACCTTGCGGGCATTCTT, NLRP3 forward TGAAGAAAGATTACCGTAAGAAGTACAGA and reverse GCGTTTGTTGAGGCTCACACT, NLRC5 forward TGGGAAGACACTCAGGCTAA and reverse ATCATCGTCCTCACAGAGGTT, beta-actin (housekeeping gene) forward ACCACACCTTCTACAATGAGC and reverse CAGCCTGGATAGCAACGTAC.

Experimental statistics are shown as mean ± standard error for variables with Gaussian distribution and median (interquartile range, IQR) for the others. Differences of quantitative parameters among groups were assessed with t-test or one-way ANOVA for two or more independent samples as appropriated. Correlations among parametric variables were performed using Pearson's test while the correlations among one or more non-parametric variables were performed using Spearman’s test. Results were considered significant at P < 0.05. Data analysis was performed with SPSS version 20.0 (SPSS, Chicago, IL, USA).

Among the thirty-four patients with active ANCA-associated vasculitis, twelve were males and twenty-two were females with an age of 61.7 ± 11.9 years at renal biopsy. Thirty patients were MPO-ANCA positive and four patients were PR3-ANCA positive. The median level of serum creatinine at renal biopsy was 266.9 (IQR: 196.3–410.0) μmol/L. The Birmingham Vasculitis Activity Score (BVAS) [32] was 18.6 ± 5.2. The majority of histopathologic categorization was crescentic class (24/34, 70.6%), according to the pathologic classification system by Berden et al. [29]. Detailed data are listed in Table 1.

In AAV patients, immunohistochemical staining showed that NOD2, NLRP3 and NLRC5 were widely expressed in glomeruli and tubulointerstitium (Fig. 1).

The mean optical densities of NOD2, NLRP3 and NLRC5 in glomeruli of AAV patients were significantly higher than those of normal controls (0.157 ± 0.004 vs. 0.090 ± 0.004, P < 0.001; 0.064 ± 0.002 vs. 0.050 ± 0.002, P < 0.001; 0.131 ± 0.004 vs. 0.093 ± 0.006, P < 0.001, respectively). Consistently, the mean optical densities of NOD2, NLRP3 and NLRC5 in tubulointerstitium were also significantly higher in AAV patients than those of normal controls (0.166 ± 0.003 vs. 0.109 ± 0.005, P < 0.001; 0.087 ± 0.002 vs. 0.071 ± 0.003, P < 0.001; 0.150 ± 0.004 vs. 0.096 ± 0.006, P < 0.001, respectively). Compared with disease controls, the expression of three NLRs in AAV patients was significantly higher than those in MCD or class IV LN. There were also close correlations between the expression in glomeruli and tubulointerstitium in AAV patients (r = 0.644, P < 0.001 for NOD2; r = 0.762, P < 0.001 for NLRP3; r = 0.844, P < 0.001 for NLRC5) (Fig. 2).

The immunohistochemical assay showed that three NLRs were universally expressed in glomeruli and tubulointerstitium in AAV patients. To further investigate locations of NLRs, we performed double immunofluorescence staining for NLRs and markers of various cell types (Figs. 3, 4, 5, 6). Among the renal intrinsic cells, NOD2, NLRP3 and NLRC5 were predominantly expressed in podocytes and sparsely expressed in glomerular endothelial or mesangial cells. The staining of three NLRs was also detected in infiltrating monocytes/macrophages specifically marked with CD68. Collectively, three NLRs co-localized with podocytes and infiltrating inflammatory cells.

In order to validate the expression results of immunohistochemical staining, we employed quantitative polymerase chain reaction to determine the expression of three NLRs on mRNA levels. Given the formation of crescents in AAV was associated with podocyte injury [33] and NLRs mainly co-localized with podocyte in glomeruli, we employed human podocytes for the functional study in vitro. Consistent with the results of immunohistochemical staining, the expression of NOD2, NLRP3 and NLRC5 were upregulated upon TNF-α stimulation (Fig. 7).

Then, we made correlation analysis between the expression of NLRs and clinicopathological parameters in AAV patients (Fig. 8). For NOD2, the mean optical density in glomeruli correlated with proteinuria level and serum creatinine at renal biopsy (r = 0.506, P = 0.002; r = 0.416, P = 0.014). For NLRP3, there was no significant correlation between the expression and clinicopathological data. For NLRC5, the mean optical density in glomeruli correlated with proteinuria level, BVAS and the proportion of crescents in renal specimen (r = 0.433, P = 0.011; r = 0.365, P = 0.034; r = 0.471, P = 0.005, respectively). As for the classification scheme proposed by Berden et al., the mean optical density of NOD2 and NLRC5 in glomeruli was significantly higher in crescentic class than that in non-crescentic class (0.163 ± 0.005 vs. 0.143 ± 0.004, P = 0.021; 0.136 ± 0.004 vs. 0.119 ± 0.007, P = 0.041) (Fig. 8).