C5a and its receptors in human anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis

Twenty-four consecutive patients with AAV in the active phase of initial onset before initiation of immunosuppressive therapy and 19 consecutive patients with AAV in the remission phase after immunosuppressive therapy, diagnosed at Peking University First Hospital from 2008 to 2009, were included. All these patients had a positive test for perinuclear ANCA (P-ANCA) by indirect immunofluorescence and MPO-ANCA by antigen-specific enzyme-linked immunosorbent assay (ELISA). All these patients met the Chapel Hill Consensus Conference (CHCC) definition of AAV [8] and had complete clinical data. All the previously mentioned 24 patients with AAV in the active phase received renal biopsy at diagnosis and before immunosuppressive therapy. "Remission" was defined as "absence of disease activity attributable to active disease qualified by the need for ongoing stable maintenance immunosuppressive therapy" (complete remission), or "50% reduction of disease activity score and absence of new manifestations" (partial remission), as described previously [9]. Patients with secondary vasculitis or with any other coexisting renal disease were excluded.

Twenty-five age- and gender-matched healthy blood donors and 20 patients with biopsy-proven lupus nephritis (LN), diagnosed in the same period in our center, were enrolled as normal control and disease control, respectively. All the patients with lupus nephritis fulfilled the 1997 American College of Rheumatology revised criteria for systemic lupus erythematosus (SLE) [10]. Of the 20 patients with lupus nephritis, six were classified as type III, and 14 were classified as type IV, according to the abbreviated version of the ISN/RPS classification [11].

Estimated glomerular filtration rate (eGFR) was calculated by using the modification of diet in renal disease (MDRD) equations [12].

The research was in compliance of the Declaration of Helsinki and approved by the ethics committee of our hospital. Written informed consent was obtained from each participant.

Neutrophils were isolated from heparinized venous blood of healthy donors by density gradient centrifugation on Lymphoprep (Nycomed, Oslo, Norway). Erythrocytes were lysed with ice-cold ammonium chloride buffer, and neutrophils were washed in Hanks balanced salt solution without Ca²⁺/Mg²⁺ (HBSS^-/-; Chemical Reagents, Beijing, China). Neutrophils were then suspended in HBSS with Ca²⁺/Mg²⁺(HBSS^+/+; Chemical Reagents) to a concentration of 2.5 × 10⁶ cells/ml. Neutrophils were activated by 2 ng/ml tumor necrosis factor (TNF)-α. The specimens were fixed with 70% ethanol for 5 minutes, and permeabilized with 0.5% Triton X-100 for 5 minutes at 4°C.

Urine and plasma concentrations of human fragment C5a were determined with enzyme-linked immunoassays (Quidel Corporation, San Diego, CA, USA). Optimal dilutions of urine and plasma samples were obtained according to the ELISA kit instructions. The urinary creatinine also was measured. The urinary level of C5a was calculated by the ratio of urinary concentration of C5a to the urinary creatinine.

The renal histology of patients with AAV in the active phase was evaluated according to the previous standardized protocol [13‐15]. The presence of glomerular lesions, including fibrinoid necrosis, crescents, and glomerulosclerosis, was calculated as the percentage of the total number of glomeruli in a biopsy. Interstitial and tubular lesions were scored semiquantitatively on the basis of the percentage of the tubulointerstitial compartment that was affected: interstitial infiltrate ("-"for 0; "+" for 0 to 20%; "++" for 20% to 50%; and "+++" for > 50%), interstitial fibrosis ("-" for 0; "+" for 0 to 50%; and "++"for > 50%) and tubular atrophy ("-" for 0; "+" for 0 to 50%; and "++" for > 50%).

In renal sections stained with hematoxylin/eosin (HE), the numbers of neutrophils were counted in the glomeruli of AAV by an experienced pathologist.

Five renal tissues were obtained from the normal part of nephrectomized (because of renal carcinoma) kidneys and were used as normal controls. Their renal tissues were considered normal with light microscopy, immunofluorescence, and electron microscopy.

Tissue sections were deparaffinized in xylene and rehydrated through grading alcohols. Slides were immersed in citric acid buffer (0.01 M, pH 6.0) and were treated in an 800 W microwave oven for 2 minutes and then at 200 W for 8 minutes, for antigen retrieval. Slides were rinsed in phosphate-buffered saline (0.01% PBS). Endogenous peroxidase was blocked with 0.3% H₂O₂ in methanol for 30 minutes at room temperature. After a 30-minute blocking step with 3% normal goat serum diluted in PBS, mouse anti-human CD88 (1:50 dilution; Abcam, Cambridge, UK; catalog number, ab11867), rabbit anti-human C5L2 (1:500 dilution; Abcam, catalog number, ab77982), and mouse anti-human CD66b (1:200 dilution; Abnova, Taiwan; catalog number, MAB7198) were applied and incubated overnight at 4°C in a moisture chamber. After multiple PBS rinses, the slides were exposed to secondary antibodies. The detection system used, Dako EnVision HRP (Dako A/S, Copenhagen, Denmark), was an avidin-free two-step indirect method with goat anti-rabbit and goat anti-mouse immunoglobulins conjugated with horseradish peroxidase (HRP) as secondary antibodies. The secondary antibodies were incubated for 30 minutes at 37°C, and sections were developed in fresh hydrogen peroxide plus 3-3-diaminobenzidine tetrahydrochloride solution for 1 minute. As negative controls, primary antibodies were replaced by normal rabbit IgG or normal mouse IgG. Finally, the sections were incubated with hematoxylin and dehydrated through alcohols and xylene. The sections were examined with light microscopy.

In total, 20 fields of vision per kidney section (three to four sections per kidney) at ×200 were observed blindly as a semiquantitive assessment of immunohistochemical staining. The renal CD88 and C5L2 stainings were evaluated with the Image Pro Plus analysis software 6.0. The positive signals were quantified as the mean optical density (integrated option density/area).

Elastase is the marker for neutrophils and monocytes [16], and CD68 is the marker for macrophages [17]. Colocalization of CD88 and C5L2 with elastase and CD68 was judged by double-labeling immunofluorescence in formalin-fixed paraffin-embedded tissue. The deparaffinage and antigen-retrieval step were consistent with the previous immunohistochemistry described earlier. After a 30-minute blocking step with 3% normal goat serum diluted in PBS, mouse anti-human CD88 (1:5 dilution; Abcam) combined with rabbit anti-human neutrophil elastase (1:50 dilution; Abcam) or with rabbit anti-human CD68 (1:50 dilution; Abbiotec, San Diego, CA, USA), and rabbit anti-human C5L2 (1:50 dilution; Abcam) combined with mouse anti-human neutrophil elastase (1:5 dilution; Genway Biotech, San Diego, CA, USA) or with mouse anti-human CD68 (1:5 dilution; Abcam) were added and incubated overnight at 4°C in a moisture chamber, 3 × 5-minute washes in PBS. The secondary antibodies FITC-labeled goat anti-mouse IgG and TRITC-labeled goat anti-rabbit IgG (both 1:60; Southern Biotech) were incubated for 30 minutes at 37°C in a moisture chamber, 3 × 5-minute washes in PBS. Nuclei were stained with DAPI (ZSGB, Beijing, China) and sections mounted with Citifluor. In negative controls, primary antibodies were replaced by normal rabbit IgG or normal mouse IgG.

The spleen and neutrophils were used as positive controls. Neutrophils were isolated from peripheral blood of healthy blood donors and fixed with 70% ethanol. The spleen tissues were obtained from a patient subjected to splenectomy for traumatic splenic rupture. The spleen tissues were fixed in 10% buffered formalin, processed through graded alcohols and xylene, and embedded in paraffin. For the primary antibody, mouse anti-human CD88 (1:20 dilution for neutrophil-coated slides and 1:5 dilution for spleen tissues; Abcam) combined with rabbit anti-human neutrophil elastase (1:50 dilution for neutrophils and spleen; Abcam) were used. The other steps were consistent with aforementioned method.

The stained sections were visualized with fluorescence microscopy (80i; Nikon, Tokyo, Japan). Images were exported from the Nikon fluorescence microscopy software.

To confirm whether CD88 was expressed on renal infiltrated neutrophils of AAV patients, the renal sections of AAV were counterstained with HE on the basis of immmunohistochemical staining of CD88 (1:50 dilution; Abcam; catalog number, ab11867). The sections were examined with light microscopy.

Differences of quantitative parameters between groups were assessed by using the t test (for normally distributed data) or nonparametric test (for non-normally distributed data). Spearman correlation was used to measure the relation between two non-normally distributed variables or between one non-normally distributed variable and one normally distributed variable. Pearson correlation was used to measure the relation between two normally distributed variables. The difference was considered significant if the P value was < 0.05. Analysis was performed with the SPSS statistical software package (version 13; Chicago, IL, USA).

Among the 24 patients with AAV in the acute phase, 11 were male and 13 were female patients, with an age of 55.4 ± 15.7 years (range, 14 to 82 years) at diagnosis. Of the 24 patients, the level of initial serum creatinine was 297.0 ± 232.8 (range, 59 to 1,007) μM; the level of urinary protein excretion was 2.56 ± 1.72 g/24 hours (range, 0.23 to 7.35 g/24 hours). Twenty patients had renal insufficiency at diagnosis. The organ involvements are listed in Table 1.

Among the 19 patients with AAV in the remission phase, nine were male and 10 were female patients, with an age of 62.1 ± 13.9 years (range, 30 to 81 years) at diagnosis. The level of serum creatinine was 165.3 ± 99.3 (range, 58 to 441) μM. The level of Birmingham Vasculitis Activity Scores [18] (BVAS) was 0 and 2 in 18 and one patients, respectively.

Among the renal biopsies of the 24 patients with the active phase of AAV, all had little or no staining for IgG, IgA, and IgM by immunofluorescence microscopy (≤ 1+ on a scale of 0 to 4+). No or little electron-dense deposit was detected with electron microscopy. The glomerular and tubulointerstitial lesions are listed in Table 2.

We found 2.56 ± 0.14 neutrophils per glomeruli of AAV in HE staining sections, whereas no neutrophils were present in glomeruli of normal controls (see Additional file 1, Figure S1).

Plasma samples were obtained from 24 patients with AAV in the active phase, 19 patients with AAV in the remission phase, 20 patients with LN, and 25 healthy blood donors. Urinary levels of C5a were measured in 24 patients with AAV in the active phase, 19 patients with AAV in remission, 10 patients with LN, and 25 healthy blood donors.

The level of plasma C5a was significantly higher in patients with AAV in the active phase than that in patients with AAV in remission, patients with LN, and normal controls (75.67 ± 11.93 ng/ml versus 8.78 ± 1.12 ng/ml; P < 0.001; 75.67 ± 11.93 ng/ml versus 30.42 ± 4.07 ng/ml; P < 0.001; 75.67 ± 11.93 ng/ml versus 7.37 ± 0.90 ng/ml; P < 0.001, respectively). No significant differences in C5a levels were found between normal control and patients with AAV in remission (8.78 ± 1.12 ng/ml versus 7.37 ± 0.90 ng/ml; P = 0.887) (Figure 1).

The urinary C5a level was significantly higher in patients with AAV in the active phase than that in patients with AAV in remission and normal controls (3.00 ± 0.78 ng/ml versus 0.18 ± 0.044 ng/ml, P < 0.001; 3.00 ± 0.78 ng/ml versus 0.13 ± 0.02 ng/ml, P < 0.001, respectively). This was also the case after adjustment for urinary creatinine (14.13 ± 3.54 ng/mg Cr versus 0.023 ± 0.0058 ng/mg Cr; P < 0.001; 14.13 ± 3.54 ng/mg Cr versus 0.0246 ± 0.0037 ng/mg Cr; P < 0.001, respectively). No significant difference was noted in urinary C5a levels between patients with AAV in the active phase and patients with LN (Figure 1). No significant correlation was found between the levels of plasma C5a and urinary C5a in patients with AAV.

Studies on renal histopathology were performed in patients with AAV in the active phase. In renal specimens of patients with AAV, immunohistochemical examination revealed prominent expression of CD88 in proximal tubules, distal tubules, and collecting ducts. Expression of CD88 protein was scanty in glomeruli and vasculature of the normal controls and in the patients with AAV (Figure 2).

In contrast, expression of C5L2 was detected in the glomeruli and vasculature. The positive staining of C5L2 was also seen in proximal tubules, distal tubules, collecting ducts, and interstitium (Figure 2).

Immunohistochemical examination showed the prominent expression of CD66b, which is an ideal marker for neutrophils, in glomeruli of AAV patients. Expression of CD66b was scanty in the glomeruli of normal controls (see Additional file 2, Figure S2). It indicated that a number of neutrophils were present in the glomeruli of AAV patients, but no or few neutrophils in the glomeruli of normal controls.

CD88 and C5L2 staining in renal specimens were evaluated by the Image Pro Plus analysis software (version 6.0; Dallas, TX, USA). The mean optical density of CD88 in patients with AAV in the tubulointerstitium was significantly lower than that in normal controls (0.0052 ± 0.0011 versus 0.029 ± 0.0042; P = 0.005). The mean optical density of C5L2 in glomeruli was significantly higher in patients with AAV than that in normal controls (0.013 ± 0.0027 versus 0.0032 ± 0.0006; P < 0.001). However, no significant difference was seen for the mean optical density of C5L2 in the tubulointerstitium between patients with AAV and normal controls (0.034 ± 0.005 versus 0.044 ± 0.006; P = 0.227).

Among the patients with AAV, correlation analysis showed that the mean optical density of CD88 correlated with initial eGFR (r = 0.835; P < 0.001), and inversely correlated with initial serum creatinine (r = -0.628; P < 0.001). The mean optical density of CD88 in patients with interstitial infiltrate greater than 50% was significantly lower than that in those with interstitial infiltrate ≤ 50% (0.0035 ± 0.0008 versus 0.0070 ± 0.0023; P = 0.034).

The mean density of glomerular C5L2 staining correlated with the ratio of glomerular cell number to glomerular area (r = 0.561; P = 0.004).

By using immunofluorescence double-labeling with neutrophil elastase or CD68, we determined whether CD88 and C5L2 was expressed on neutrophils (or monocytes) or macrophages. It was shown that little or no CD88 was expressed on neutrophils, monocytes, or macrophages in renal specimens of patients with AAV, and this result was proved by two other anti-CD88 monoclonal antibodies (Santa Cruz Biotechnology, catalog number sc-70812, and Abcam, catalog number ab11884, (see Additional files 3, 4, 5, and 6, Figures S3, S4, S5, and S6). However, CD88 could colocalize with elastase in the spleen- and neutrophil-coated slides. C5L2 expression partly colocalized with neutrophils (or monocytes) and macrophages (Figures 3 and 4).

On the basis of immmunohistochemical staining of CD88, the renal sections of AAV were counterstained with HE. Little or no CD88 staining was noted in renal infiltrated neutrophils (see Additional file 7, Figure S7). This was consistent with the results of immunofluorescence double-labeling of CD88 and elastase.