The influence of cathelicidin LL37 in human anti-neutrophils cytoplasmic antibody (ANCA)-associated vasculitis

Adult subjects with AAV, who were diagnosed according to the Chapel Hill definition [12], were recruited from the Department of Nephrology, Xinqiao Hospital, Third Military Medical University. The patient characteristics are presented in Table 1. Serum samples were collected from the patients and healthy controls (HCs). Among the 85 patients, 40 patients underwent renal biopsies concomitantly obtained with the serum samples, and all biopsies were reviewed and classified by an experienced nephropathologist according to the revised criteria for PiCGN. Criteria used in the study to define PiCGN by immunofluorescence analysis for frozen tissue sections presented deficiency of immunoglobulin and complements. PR3-ANCA and MPO-ANCA were evaluated using EUROBlot kits (DL-1200-6421-3G; Euroimmun;Lübeck, Germany) and indirect immunofluorescence (FA-1200-2010; Euroimmun; Lübeck, Germany) according to the protocol provided by the manufacturer. The levels of creatinine, blood urea nitrogen and complement factor C3 in the sera were determined using routine techniques. Crescentic GN was defined as crescents in >50% glomeruli. Frozen sections of renal biopsy specimens obtained from 40 patients, among of them, 15 patients with crescentic GN and 25 patients without crescentic GN were included in the present study. The study protocols were approved by the ethics committee board of Xinqiao Hospital, and all subjects gave written informed consent.

Fluorescein isothiocyanate (FITC) anti-human CD16 was purchased from Biolegend (San Diego, CA, USA). Anti-human LAMP-2 mouse monoclonal antibody (H4B4), anti-human LAMP-2 mouse monoclonal antibody-FITC, anti-human PR3 mouse monoclonal antibody-FITC (WGM2), anti-human MPO mouse monoclonal antibody-FITC (266.6 K2), anti-LL37 (pAbC14), anti-IFNα (pAbFL-189) and normal mouse IgG1 were purchased from Santa Cruz (Heidelberg, Germany). Anti-human histone H3 rabbit polyclonal antibody was purchased from Abcam (Cambridge, UK). All secondary antibodies conjugated with fluorescence were purchased from ZSBio (Beijing, China). Phorbol 12-myristate 13-acetate was purchased from Sigma-Aldrich (St Louis, MO, USA). HBSS (without Ca²⁺ and Mg²⁺), RPMI-1640 medium (phenol red-free) and penicillin/streptomycin solution were purchased from Invitrogen (San Diego, CA, USA). The 12-mm round glass cover slips were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Polymorphprep™ was purchased from Axis-Shield (Oslo, Norway). The red blood cell lysis buffer was purchased from Roche Diagnostics (Mannheim, Germany). The immunostaining fix solution, immunostaining blocking buffer, immunostaining primary antibody dilution buffer, immunofluorescence staining secondary antibody dilution buffer and anti-fade mounting medium were purchased from Beyotime (Shanghai, China). Anti-human BDCA2 mouse monoclonal antibody-FITC (130-090-510) and anti-human BDCA2 mouse monoclonal antibody-PE (130-090-511) were purchased from Miltenyi Biotec (Bergisch Gladbach, Germany). The ELISA kits for LL37 (HK321) and IFNα (3423-1A-20) were purchased from Hycult Biotech (Uden, Netherlands) and MabTech (Nacka Strand, Sweden), respectively.

Human neutrophils were isolated from HCs or patients with PiCGN by density centrifugation using Polymorphprep™ and red blood cell lysis buffer [13]. Briefly, 5 ml blood containing ethylenediamine tetraacetic acid was layered onto 5 ml Polymorphprep™. After 35 minutes of centrifugation at 500 g, the neutrophils were separated from the polymorphonuclear leukocyte-rich pellet. Residual erythrocytes were eliminated by red blood cell lysis. The neutrophil purity was routinely ~95%, as assessed by forward-scatter and side-scatter flow cytometric analyses [8]. Unless otherwise stated, the cells were resuspended in RPMI medium (phenol red-free) supplemented with 1% penicillin/streptomycin. Then, 5×10⁶ neutrophils/ml were seeded onto tissue culture plates for culture and 5×10⁵ neutrophils/ml were seeded onto glass coverslips for immunofluorescence staining. The cells were incubated at 37°C in the presence of 5% CO₂.

The levels of LL37 and IFNα from sera or cell culture supernatants were quantified by ELISA according to the protocols provided by the manufacturers. The sensitivity was 0.1 ng/ml for LL37 and 7.8 pg/ml for IFNα.

Freshly purified neutrophils were allowed to adhere to glass coverslips in RPMI-1640 medium. After incubation for 30 minutes at 37°C in 5% CO₂, the cells were stimulated with 20 μg/ml LAMP-2 antibody or left untreated. The neutrophils were then fixed in 2.5% glutaraldehyde for 2 hours at 4°C. After washing with physiological saline three times, the samples were dehydrated through a graded ethanol series. The cover slips were then transferred into a critical point dryer and dried. The surface of the specimen was coated with a 5 nm platin/carbon layer using a thin layer evaporator. The samples were then viewed with a scanning electron microscope.

Immunofluorescence staining was performed on frozen sections of kidney needle biopsies from patients with PiCGN as described previously [8]. Briefly, after fixation in paraformaldehyde, the specimens were incubated with anti-LL37 (pAbC14), anti-histone H3 (pAbab8284), anti-IFNα (pAbFL-189) and anti-BDCA-2 (mAbCD303) antibodies or isotype control, followed by the appropriate secondary antibodies. DNA was stained with 4′,6-diamidino-2-phenylindole. The neutrophils were seeded onto lysinated glass slides in a 24-well cell culture plate and incubated for 1 hour in a CO₂ incubator at 37°C. The cells were left unstimulated or were stimulated with either 100 nM phorbol 12-myristate 13-acetate or 20 μg/ml LAMP-2 antibody for up to 180 minutes at 37°C in 5% CO₂. Subsequently, the cells were fixed and permeabilised. After rehydration with phosphate-buffered saline at room temperature, the cells were incubated with blocking buffer overnight at 4°C, and then the specimens were incubated with fluorescently labelled antibodies. The chromatin was stained with an anti-histone H3 rabbit antibody. The images were acquired as projections of a confocal stack.

The experiments were performed in triplicate at least three separate times. Data are presented as the mean ± standard error of the mean and were analysed using GraphPad Prism software 5 (GraphPad Software, San Diego, CA, USA). Where appropriate, either two-tailed Student’s t tests or the Kruskal–Wallis and Mann–Whitney U tests were used. Differences were considered significant at P <0.05.

To analyse the correlation between the serum levels of LL37 and AAV, we first performed ELISAs to determine the LL37 levels in 136 serum samples collected from 85 unrelated patients with documented AAV (50 females, 35 males; female/male ratio 10:7) (Table 1) and 51 HCs. The AAV sera contained higher levels of LL37, with a mean concentration of 100.3 ng/ml, as compared with the HC sera, which showed a mean concentration of 28.53 ng/ml (U = 829.00, P <0.01; Figure 1A). Serum levels of LL37 >100.3 ng/ml were arbitrarily considered to represent high levels, and those <100.3 ng/ml were considered to represent low levels. Among the 40 patients with renal biopsy, as shown in Figure 1B, the renal tissue of patients with crescentic GN showed higher levels of LL37 as compared with those from patients with noncrescentic GN (P <0.01) and HCs (P <0.001). The serum levels of LL37 in patients without crescentic GN were also significantly higher than those in HCs (P <0.01). Forty patients underwent renal biopsy and 22 of them expressed high serum levels of LL37 (>100.3 ng/ml), and 13 of these high-expression patients had crescentic GN. In contrast, only two of the 18 patients expressing low serum levels of LL37 (<100.3 ng/ml) had crescentic GN (χ² = 9.724, P = 0.002; Figure 1C). These results indicated that patients with high LL37 expression had a greater risk of having crescentic GN as compared with patients with low LL37 expression.

We also performed ELISAs to analyse the serum IFNα levels in 136 serum samples collected from the aforementioned 85 patients. We found that the AAV sera contained higher levels of IFNα, with a mean concentration of 958.5 pg/ml, compared with the HC sera, which showed a mean concentration of 252.1 pg/ml (U = 1438.00, P = 0.000; Figure 2A).

Serum levels of IFNα >958.5 pg/ml were arbitrarily considered to represent high levels, whereas those <958.5 pg/ml were considered to represent low levels. Among the 40 patients with renal biopsy, as shown in Figure 2B, 15 crescentic GN patients showed high-level IFNα expression as compared with 25 of the noncrescentic GN patients (P <0.001) and the HCs (P <0.001). The IFNα serum level of the patients without crescentic GN were also significantly higher than those in HCs (P <0.05). Of the 40 patients who provided a renal biopsy, 22 showed high levels of IFNα (>958.5 pg/ml), and 13 of these 22 patients had crescentic GN. In contrast, only two of the 18 patients expressing low serum levels of IFNα (<958.5 pg/ml) had crescentic GN (χ² = 16.835, P = 0.000; Figure 2C). These results indicated that patients expressing high IFNα levels had a greater risk of crescentic formation in the glomerulus in comparison with patients expressing low levels of IFNα.

Given that the LL37 and IFNα serum levels were increased in AAV patients, particularly in those with crescentic GN, we further investigated whether the serum levels of LL37 were correlated with the IFNα levels and whether the levels of LL37 and IFNα were correlated with the serum levels of creatinine and complement C3. We found a positive correlation between the serum LL37 and IFNα levels (R = 0.577, P = 0.000; Figure 3A) as well as significant positive correlations between the serum levels of LL37 and creatinine (R = 0.437, P = 0.000; Figure 3B) and between the serum levels of IFNα and creatinine (R = 0.337, P = 0.004; Figure 3C). However, neither LL37 nor IFNα showed any correlation with complement C3 (LL37: R = 0.020, P = 0.871; IFNα: R = -0.12, P = 0.922) (Figure 3D,E). To further investigate whether LL37 and IFNα reflect systemic inflammation, we analysed the sera levels of LL-37 and IFNα before and after treatment (n = 7). We found that both LL-37 and IFNα levels had degraded after immunosuppressive therapy (Additional file 1).

To assess the roles of LL37 and IFNα in renal inflammation, we evaluated the presence and localisation of LL37 and IFNα in renal biopsies from AAV patients with crescentic GN. We found that the kidney tissues from AAV patients showed strong expression of LL37, especially in Bowman’s capsule. We also found that LL37 expression co-localised with neutrophil and pDC (stained for BDCA2) infiltrates in the affected glomeruli and the interstitium (Figure 4A,B), suggesting that LL37 expression occurs predominantly during crescentic GN. As autoinflammatory conditions such as psoriasis can be driven by pDCs, which produce large amounts of IFNα, we next examined the expression of IFNα in the kidney tissues. Immunostaining revealed the co-localisation of IFNα and LL37 in the kidney tissues from AAV patients (Figure 4C). However, there were only mild immunofluorescence signals stained with IFNα and LL37 in the renal tissues from patients without crescents (Figure 4D), suggesting that LL37 may mediate pDC activation to produce IFNα in renal local inflammatory in AAV.

ANCA plays a critical role in the vascular damage associated with AAV [14]. Kessenbrock and colleagues reported strong NET formation in patients with AAV [8], and the presence of anti-LAMP-2 antibodies has been suggested to represent a new subtype of ANCA, with a high prevalence in PiCGN [2, 4, 15]. Therefore, we asked whether targeted auto-antigens and antimicrobial peptides would be present in NETs stimulated by anti-LAMP-2-IgG. As expected, the immunofluorescent analysis of NETs revealed that LAMP-2 co-localised with extracellular chromatin fibres (Figure 5A). We also observed that PR3 and MPO were expressed within the NETs (Figure 5B,C), as reported previously [8], and we found that LL37 expression was enriched in parts of the extracellular chromatin fibres (Figure 5D). After culturing for 180 minutes, neutrophils isolated from PiCGN patients and HCs showed the typical features of NETosis, including web-like structures, as visualised by scanning electron microscopy (Figure 5E,F). To investigate the release of LL37 in the NET supernatant, neutrophils isolated from AAV patients or HCs were treated with anti-LAMP-2-IgG (AAV + H4B4, n = 5; HC + H4B4, n = 5), an isotype control (HC + ISO, n = 5) or phorbol 12-myristate 13-acetate as a positive control (n = 5) [8]. After 180 minutes, the supernatant was collected and measured by ELISA, and the results indicated that the HC + H4B4 group supernatant showed higher levels of LL37 as compared with the HC or HC + ISO groups (P <0.001). Moreover, the AAV + H4B4 supernatant showed higher levels of LL37 compared with the HC + H4B4 supernatant (P <0.001) (Figure 5G). No IFNα was detected in the supernatant by ELISA. We observed MPO, a marker of neutrophils, showing co-localisation with LL37, IFNα and BDCA-2 in AAV patients with crescentic GN (Additional file 2). To further clarify the source IFNα, pDCs were isolated by magnetically activated cell sorting using the BDCA-4 dendritic cell isolation kit (Miltenyi Biotec). pDCs (5 × 10⁴/ml) were incubated with supernatants from neutrophils treated as above mentioned, with 3 μg/ml ODN 2216 pDCs as positive control for 24 hours. Supernatants were harvested to detect the level of IFNα and LL-37 by ELISA. We found that IFNα was released by pDC incubation with supernatants of anti-LAMP-2 antibody-treated neutrophils and ODN2216 for 24 hours (Additional file 3). No LL-37 was released from pDC.