Neutrophils from vasculitis patients exhibit an increased propensity for activation by anti-neutrophil cytoplasmic antibodies



Anti-neutrophil cytoplasmic antibodies (ANCA) are thought to be pathogenic in ANCA-associated vasculitis (AAV) by stimulating polymorphonuclear leucocytes (PMNs) to degranulate and produce reactive oxygen species (ROS). The aim of this study was to investigate if PMNs from AAV patients are stimulated more readily by ANCA compared with PMNs from healthy controls (HCs). Differences in ANCA characteristics that can account for different stimulation potential were also studied. PMNs from five AAV patients and five HCs were stimulated with 10 different immunoglobulins (Ig)Gs, purified from PR3–ANCA-positive patients, and ROS production, degranulation and neutrophil extracellular trap (NET) formation was measured. ANCA levels, affinity and clinical data of the AAV donors were recorded. The results show that PMNs from AAV patients produce more intracellular ROS (P = 0·019), but degranulate to a similar extent as PMNs from HCs. ROS production correlated with NET formation. Factors that may influence the ability of ANCA to activate PMNs include affinity and specificity for N-terminal epitopes. In conclusion, our results indicate that PMNs from AAV patients in remission behave quite similarly to HC PMNs, with the exception of a greater intracellular ROS production. This could contribute to more extensive NET formation and thus an increased exposure of the ANCA autoantigens to the immune system.


ANCA-associated vasculitides (AAV) is a group of autoimmune diseases comprising granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA). They are characterized by autoantibodies against the neutrophil granule proteins proteinase 3 (PR3) or myeloperoxidase (MPO). These anti-neutrophil cytoplasmic antibodies, or ANCAs, can activate primed neutrophils to degranulate and produce reactive oxygen species (ROS) [1], induce the formation of neutrophil extracellular traps (NETs) [2] and mediate the release of microparticles from polymorphonuclear leucocytes (PMNs) [3]. The exact mechanisms are not completely understood, but activation by ANCA seems to involve both the Fc portion of the antibody, as well as the Fab portion binding to its antigen on the neutrophil surface [4, 5]. However, there are also reports indicating that ANCA are internalized into the PMNs after binding, but the importance of this remains unknown [6]. Activation of neutrophils by ANCA is thought to play a major role in the pathogenesis and progress of these diseases. The pathogenicity of ANCA was first proved in vivo in a MPO–ANCA mouse model [7]. More recently, a PR3–ANCA mouse model has been described [8]. However, ANCA levels alone cannot predict relapses conclusively [9], and there is an unmet need for better methods to assess the degree of disease and risk of relapse. Several attempts have been made to map both MPO [10] and PR3 for particularly harmful epitopes, and several conformational epitopes have been found [11, 12]. This information will be of importance for understanding the underlying mechanisms of these diseases and could also lead to better methods for assessing the disease course. Furthermore, immunoglobulin (Ig)G subclass specificity also matters for ANCA-mediated activation of the neutrophil [13].

To our knowledge, there are no systematic studies of PMNs from AAV patients and their response to ANCA stimulation. In this paper we aimed to investigate if IgG from PR3–ANCA-positive patients activates neutrophils more or less easily from AAV patients compared with neutrophils from healthy donors. Furthermore, we characterize IgG purified from different patients to determine why some ANCA are more potent activators than others.

Material and methods

Blood samples and patients

PMNs were isolated from PR3–ANCA-positive AAV patients in stable clinical remission, as assessed by a Birmingham Vasculitis Activity Score (BVAS) of 0 [14]. Serum samples for IgG purification were obtained from AAV patients with PR3–ANCA specificity and varying activity. The PR3–ANCA levels in a standard direct enzyme-linked immunosorbemt assay (ELISA) (Wieslab, Malmö, Sweden) ranged between 6 and 991 IU/ml (Table 1). Healthy control (HC) IgG was used as negative control in some assays. Written informed consent was retrieved from all donors and these studies were conducted with permission from the Ethical Review Board, Lund, Sweden.

Table 1. Clinical and biochemical characteristics of immunoglobulin (IgG) preparations from anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) patients.
  1. IU/ml = international units/ml; BVAS = Birmingham Vasculitis Activity Score. Activation score represents a summarized value from reactive oxygen species (ROS), degranulation and neutrophil extracellular trap (NET) formation assays. See text for details. ELISA = enzyme-linked immunosorbent assay.
Age (years)45·656·938·667·071·458·965·855·875·159·1
Direct ELISA PR3–ANCA (IU/ml)99163612631935643411
Capture ELISA PR3–ANCA (IU/ml)745320179059020318832014296
Affinity (ml/μg)15·411·12·52·82·12·12·73·21·82·2
Activation score99·360·530·134·224·111·725·449·332·047·8

IgG purification

Serum or plasma samples were spun down for 5 min at 10 000 g and clear supernatants were added to NAb protein G spin columns (Thermo Scientific, Rockford, IL, USA) that were prepared according to the manufacturer's instructions. The samples were incubated at room temperature with rotation for 20 min. Columns were washed three times with binding buffer (20 mM sodium phosphate buffer, pH 7·0) and eluted with 0·1 M glycin-HCl, pH 2·7. Eluates were neutralized with 0·1 volumes of 1 M Na2HPO4, pH 8–9. Buffer was exchanged with PD Miditrap G-25 columns (GE Healthcare, Uppsala, Sweden) to phosphate-buffered saline (PBS) without Ca2+ and Mg2+ (Hyclone, South Logan, UT, USA). IgG were concentrated to 2 mg/ml using Vivaspin 4 ml columns, 10 000 MWCO (Sartorium stedim biotech, Goettingen, Germany). The concentration was measured using a Nanodrop®. IgG aliquots were frozen at −20°C and spun down at 12 000 g for 10 min before use to avoid IgG aggregates in the assays.

PMN purification

Heparinized blood was collected and sedimented with 2% dextran T500 (Pharmacosmos, Holbæk, Denmark) in 0·9% NaCl followed by density gradient centrifugation using Lymphoprep™ (Axis-Shield, Oslo, Norway) at 4°C. All subsequent steps were performed at this temperature. The remaining erythrocytes were lysed using water as hypotonic solution for a maximum of 30 s to leave PMNs intact. Ionic concentration was restored by adding 1 volume 1·8% NaCl.

Membrane PR3

Purified PMNs were stained with monoclonal antibodies against CD177 fluorescein isothiocyanate (FITC) (Serotech, Stockholm, Sweden) and PR3-Alexa647 (Eurodiagnostica, Malmö, Sweden) and analysed using a fluorescence activated cell sorter (FACS)Canto II (BD Biosciences, Stockholm, Sweden), as described previously [15]. Both the percentage of PR3-positive cells and mean fluorescence intensity (MFI) of the positive cells were recorded.

ROS assay

Purified PMNs were adjusted to 2 × 106 cells/ml in Hanks's balanced salt solution (HBSS++) (Hyclone) and primed with tumour necrosis factor (TNF)-α (R&D Systems, Abingdon, UK; 2 ng/ml) for 15 min at 37°C followed by Cytochalsin B (Sigma-Aldrich, Steinheim, Germany; 5 μg/ml) for 5 min at 37°C. PMNs were then stimulated with the 10 different purified AAV IgG (200 μg/ml), a monoclonal PR3 antibody (a kind gift from Dr Zhao, Beijing, 5 μg/ml), phorbol 12-myristate 13-acetate (PMA) (Glycotope Biotechnology, Heidelberg, Germany; 100 ng/ml) or PBS at 37°C. This was performed in duplicate samples in white 96-well polystyrene plates (Thermo Scientific) in the presence of luminol (Sigma-Aldrich, Fluka; 1 μg/ml) and the scavengers superoxide dismutase (SOD) (Sigma-Aldrich; 5 U/ml) and catalase (Sigma-Aldrich; 2400 U/ml) to detect intracellular ROS, or isoluminol (Sigma-Aldrich; 1 μg/ml) alone to detect extracellular ROS. PMNs were stimulated for a total of 70 min and luminescence, as a measure of ROS production, was monitored in a microplate reader (Tristar LB941; Berthold Technologies, Bad Wildbad, Germany) every 10 min. Each well was measured for 0·1 s with no filter. The software MicroWin 2000 was used for analysis. We chose to report the value after 20 min stimulation, as in most cases this time-point corresponds to the maximum stimulation (Fig. 1b).

Figure 1.

Reactive oxygen species (ROS) activation of polymorphonuclear leucocytes (PMNs). (a) PMNs from anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) patients (n = 5) and healthy controls (HCs) (n = 5) stimulated with 10 AAV immunoglobulin (Ig)G at 200 μg/ml. The bars represent mean values of 10 stimulations read after 20 min incubation. Phosphate-buffered saline (PBS) is the negative control. AAV PMNs gave rise to more intracellular ROS production (measured with luminol + scavengers) than PMNs from healthy controls (HCs) (P = 0·019). Extracellular ROS was measured with isoluminol. (b) Representative graph showing ROS production over time in PMNs (from an AAV patient, filled circles, from a HC open circles) stimulated with one representative AAV IgG. Filled lines represent extracellular ROS and dotted lines intracellular ROS. (c,d) The characteristic of each IgG is shown with a mean value of five experiments in each donor group (HC and AAV PMNs). PMA is excluded from extracellular ROS graphs for clarity, as these values were much higher than for IgG stimulations. There were no significant differences between HC and AAV PMNs in any individual IgG. RLU = relative luminescence units. Statistics were calculated using the Mann–Whitney U-test. Error bars represent standard deviations.

Degranulation assay

Purified PMNs were adjusted to 6 × 106 cells/ml in KRG buffer (Krebs–Ringer buffer with glucose: 130 mM NaCl, 5 mM KCl, 1·27 mM MgSO4, 0·95 mM CaCl2, 10 mM NaH2PO4/Na2HPO4, pH 7·4, 5 mM glucose). PMNs were primed with Cytochalasin B (5 μg/ml) for 5 min followed by TNF-α (2 ng/ml) for 15 min. The cells were then stimulated with IgG (200 μg/ml) for 15 min. All stimulations were performed at 37°C, with gentle shaking. As controls, cells stimulated with PMA (4 μg/ml) or buffer only were used. Stimulations were stopped by adding 1 volume ice-cold KRG buffer with 0·5 mM phenylmethylsulphonyl fluoride (PMSF) (Sigma-Aldrich) and spun down at 250 g for 6 min at 4°C. Supernatants were collected. Pellets were resuspended in 1 volume ice-cold KRG buffer with 0·5 mM PMSF. One volume of ELISA dilution buffer [EDB: 0·5 mM NaCl, 3 mM KCl, 8 mM Na2HPO4/KH2PO4, 1% (w/v) bovine serum albumin (BSA), 1% (v/v) Triton-X100, 0·05% NaN3, pH 7·2] with 0·5 mM PMSF were added to both supernatant and pellet before freezing at −20°C. Release values are calculated as amount of protein in supernatant compared to total amount of protein (in supernatant + pellet) and reported as percentages.

Degranulation ELISA

Five different ELISAs, with supernatants and pellets from the degranulation assay, were performed to detect markers of different granula, as has been described previously [16]. Albumin was used to detect secretory vesicles, gelatinase for gelatinase granula, lactoferrin for specific granula, MPO for azurophilic granula and PR3, mainly present in azurophilic granula, but also others.


Clear 96-well polystyrene plates (Maxisorp; Nunc, Roskilde, Denmark) were coated with rabbit anti-albumin IgG (Dako, Glostrup, Denmark; 1:2500) in coating buffer (0·0106 M Na2CO3, 0·0393 M NaHCO3, 0·02% NaN3) overnight at 4°C. ELISA plates were blocked in incubation buffer (0·137 M NaCl, 8 mM Na2HPO4 × 2H2O, 2·7 mM KCl, 1·5 mM KH2PO4, 0·02% NaN3, 0·05% Tween 20) for 1 h, with gentle shaking at room temperature; all subsequent incubations were performed under these conditions. Plates were washed three times with ELISA wash (0·9% NaCl, 0·05% Tween 20) between each incubation step. Supernatants and pellets were diluted in incubation buffer to 1:10 and 1:20, respectively. A standard curve with twofold serial dilutions of albumin (Kabivitrum, 250–3·9 ng/ml) was included in each plate. After sample incubation a detection antibody was added: rabbit anti-albumin IgG (Dako; 1:2000), biotinylated with NHS-D-Biotin (Sigma-Aldrich) followed by secondary reagent ImmunoPure® avidin-alkaline phosphatase (AP) (Thermo Scientific; 1:1000). A coloured product was developed using substrate phosphatase tablets (Sigma-Aldrich) diluted in substrate solution (1 M dietanolamine pH 9·8, 0·5 mM MgCl × 6H2O, 0·02% NaN3). Absorbance was measured at 405 nm in a plate reader (Epoch; Biotek Instruments, Bad Friedrichshall, Germany). All other ELISAs were performed in the same manner, except that blocking and dilutions were performed in EDB unless stated otherwise. Supernatants and pellets were diluted 1:500 and 1:2000, respectively, for all following ELISAs. Other specific differences are indicated as follows.


Antibody for coating was mouse anti-MMP-9 IgG1, MAB936 (R&D Systems), clone 36020, 2 μg/ml. For the standard curve, gelatinase (Enzo, Plymouth Meeting, PA, USA; 5–0·08 ng/ml) was used. The detection antibody was polyclonal goat anti-MMP9, biotinylated (R&D Systems, 1:2000) and secondary reagent ExtrAvidin® peroxidase (Sigma-Aldrich; 1:20 000, diluted in PBS). Substrate was 3,3′,5,5′-tetramethyl-benzidine liquid substrate (TMB) (Sigma-Aldrich) and absorbance was measured at 370 nm.


Antibody for coating was polyclonal sheep anti-lactoferrin (Biogenesis, Poole, UK, 2 μg/ml). Lactoferrin (Wieslab; 200–3·1 ng/ml) was used for the standard curve. The detection antibody used was rabbit anti-lactoferrin (Sigma-Aldrich, 1:4000); the secondary reagent was goat anti-rabbit IgG-AP (Sigma-Aldrich; 1:15 000).


Antibody for coating was 2B11 (MPO-specific monoclonal antibody (mAb) purified from a hybridoma, 1 μg/ml). MPO (a kind gift from Dr I. Olsson, Haematology Department, Lund University, Sweden; 100–1·56 ng/ml) was used for the standard curve. The detection antibody was polyclonal rabbit anti-MPO (Dako; 1:2000); the secondary reagent was swine anti-rabbit IgG-AP (Dako; 1:1000).


Antibodies for coating were 4A5 and 4A3 (PR3-specific mAbs purified from hybridomas), 1·5 μg/ml each. PR3 (Wieslab, 200–3·1 ng/ml) was used for the standard curve. The detection antibody used was rabbit anti-PR3, affinity-purified from serum (Wieslab; 1:1200); the secondary reagent was swine anti-rabbit IgG-AP (Dako; 1:1000).

NET formation assay

PMNs were isolated from a HC using Percoll density gradient separation (GE Healthcare, Uppsala, Sweden). Cell concentration was adjusted to 0·5 × 106 cells/ml in RPMI-1640 medium (Gibco, Paisley, UK) containing 2% fetal bovine serum (FBS; Pasching, Austria) and seeded onto a 0·001% poly-l-lysine (Sigma-Aldrich) pretreated 96-well plate (Sarstedt, Nümbrecht, Germany). Cells were primed with TNF-α (Sigma-Aldrich; 2 ng/ml) for 15 min at 37°C. They were then stimulated with IgG, 200 μg/ml for up to 7 h, and extracellular DNA were measured with Sytox Green (Invitrogen, Paisley, UK; 2·5 μg/ml) using a FL600 Microplate Fluorescence Reader (Bio-Tek Instruments, Winooski, VT, USA). As a negative control, healthy control IgG (Berigloblin; CSL Behring, Marburg, Germany, 200 μg/ml) was used. PMA (Sigma-Aldrich, 20 nM) was used as a positive control. Each IgG was measured in one or two experiments with the same PMN donor.

Activation score

An activation score was calculated for each IgG using the values for intracellular and extracellular ROS, release of gelatinase and MPO and NET formation. They were normalized, setting the highest value in each assay to 100% and the lowest to 0%, and mean values (five HC and five AAV PMNs) for all five assays were calculated and regarded as the activation score for each IgG.

Epitope mapping ELISA

Clear 96-well polystyrene plates (Maxisorp; Nunc) were coated overnight with 0·5 μg/ml of the different recombinant mouse/human chimeric PR3, as described previously [17, 18]. The eight different chimeric proteins consisted of mouse (m) and human (H) sequences in one-third, one-half or two-thirds parts, and named thus thereafter. For example, the chimera HHm consists of two-thirds human parts in the N-terminal region and one-third mouse PR3 in the C-terminal region. After washing plates twice in ELISA wash, 50 μg/ml IgG, diluted in incubation buffer with BSA (0·05% Tween 20, 0·02% NaN3, 0·2% BSA in PBS, pH 7·3–7·4), were added to the plates. All incubations were performed for 1 h at room temperature and thereafter washed twice in ELISA wash. A goat anti-human Fc-specific AP-conjugated antibody (Sigma-Aldrich, 1:7000) was used to detect bound IgG. Substrate phosphatase tablets (Sigma-Aldrich) diluted in substrate solution were used to detect colour change at 405 nm. To be able to compare different chimeras, a background for each chimera, consisting of the median value of each particular chimera, was subtracted from the ELISA values.

PR3 affinity ELISA

All 10 AAV IgG were diluted to give the same absorbance in PR3 ELISA after 1 h development time. The IgG were mixed and preincubated overnight at 4°C with increasing concentrations of PR3 (Wieslab; 0 to 2 μg/ml). Clear 96-well polystyrene plates (Maxisorp; Nunc) were coated with 1·5 μg/ml each of the 4A5 and 4A3 mAbs in coating buffer overnight at 4°C. After blocking the ELISA plates with incubation buffer with BSA, PR3 (Wieslab; 0·5 μg/ml) was added to the plates for 1 h at room temperature and then washed three times in ELISA wash (the same conditions were employed for all subsequent incubations). The IgG samples with PR3 were then added, followed by a goat anti-human Fc-specific AP-conjugated antibody (Sigma-Aldrich; 1:15 000). Substrate phosphatase tablets (Sigma-Aldrich) diluted in substrate solution were used to detect colour change at 405 nm. Absorbance values for each IgG were normalized to the same scale, setting the first value (incubated with 0 μg/ml PR3) to 100% and the last value (incubated with 2 μg/ml PR3) to 0%. The PR3 concentration of the 50% absorbance value for each IgG was read and the inverse value of this was interpreted as the affinity for PR3.

Statistical analysis

All statistical calculations were performed using GraphPad Prism, version 6.0d. To compare groups, the Mann–Whitney U-test was used. For correlations, Spearman's correlation coefficients were calculated.


ANCA stimulation of PMNs

To determine if PMNs from AAV patients are activated more readily by ANCA, ROS assays, measuring intracellular and extracellular ROS, and a degranulation assay were performed in parallel. PMN from each donor (five AAV and five HC) were stimulated with 10 IgGs purified from PR3–ANCA-positive AAV patients.

Intracellular ROS production was measured by stimulating primed PMNs with IgG in the presence of luminol together with the scavengers SOD and catalase. The results showed significantly higher intracellular ROS production in AAV patients compared with HC (P = 0·019, Mann–Whitney U-test, Fig. 1a). Extracellular ROS was measured using isoluminol instead, as it is not able to cross membranes. The extracellular ROS production did not differ between the two groups. A representative experiment is shown in Fig. 1b. There was a substantial variation in the amount of intracellular and extracellular ROS obtained with different stimulating IgGs; however, there was no significant difference when comparing to what degree the individual IgG stimulated HC and AAV PMNs (Fig. 1c,d).

Degranulation was measured as percentage protein released from activated PMNs in relation to the total amount of protein detected. No significant differences were seen for any of the five measured granular proteins when comparing PMNs from AAV patients to HCs (Fig. 2a). AAV IgG stimulated HC PMNs to a greater extent than IgG purified from HC sera (Fig. 2b). The variation in the individual 10 AAV IgG preparations is shown in Fig. 3a–e. There were no significant differences between AAV and HC PMNs when studying individual AAV IgG stimulations.

Figure 2.

Degranulation. (a) Degranulation of five different granula markers after stimulation of polymorphonuclear leucocytes (PMNs) from anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) patients (n = 5) and healthy controls (ctrl) [HCs (n = 5)] with 10 AAV immunoglobulin (Ig)G. No significant differences were detected; 37 ctrl represents PMNs from both HC and AAV PMN donors that were stimulated with buffer only at 37°C. Release is the amount of protein in supernatant compared with total amount of protein. (b). AAV IgG compared to HC IgG in ability to stimulate HC PMNs to degranulate. The same 10 AAV IgG as in (a) were tested on five different HC PMNs (median of five is reported) compared to seven HC IgG (for albumin) or five HC IgG [for gelatinase, lactoferrin, proteinase 3 (PR3) and myeloperoxidase (MPO)]. Median of several experiments with the same IgG is reported in cases where the same IgG had been tested on several HC PMNs. All different granular markers were released to a higher degree when stimulated with AAV IgG compared to HC IgG: albumin (P = 0·0097), gelatinase (P = 0·0027), lactoferrin (P = 0·0007), PR3 (P = 0·0047), MPO (P = 0·0007). Statistics were calculated using the Mann–Whitney U-test. Error bars represent standard deviations.

Figure 3.

(a–e) Degranulation characteristics of the 10 different anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) immunoglobulin (Ig)G preparations. Each bar represents a mean value of five experiments in each donor group [HC and AAV polymorphonuclear leucocytes (PMNs)]. No significant difference was seen between HC and AAV PMNs for any IgG. PMA was used as a positive control (ctrl); 37 ctrl represents PMNs that were stimulated with buffer only at 37°C. Statistics were calculated using the Mann–Whitney U-test. Error bars represent standard deviations.

The ability of the purified AAV IgG to induce NET formation in HC PMNs was determined by measuring the release of extracellular DNA using a fluorometric assay. The ability to induce DNA release differed considerably between individual IgG preparations (Fig. 4a). NET formation correlated with ROS production (intracellular ROS: rs = 0·733, P = 0·031 and extracellular ROS: rs = 0·750, P = 0·026) (Fig. 4b,c).

Figure 4.

Neutrophil extracellular trap (NET) formation. (a) NET formation measured as the degree of extracellular DNA after stimulation with anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) immunoglobulin (Ig)G in a fluorometric assay. One or two experiments per IgG were performed and in cases of two, a mean value of these is shown. IgG number 4 is missing from NET formation experiments. (b,c) Intracellular ROS (luminol + scavengers) and extracellular reactive oxygen species (ROS) (isoluminol) correlates with NET formation; rs = 0·733, P = 0·031 and rs = 0·750, P = 0·026, respectively. Spearman's correlation coefficient rs was calculated.

To be able to compare the activation potential of the different AAV IgG preparations, an activation score was constructed based on the levels of intracellular and extracellular ROS, gelatinase and MPO release and NET formation (see Materials and methods and Table 1 for details). The activation score did not correlate with disease activity as measured by the BVAS.

The membrane expression of PR3 (mPR3) did not differ significantly between HCs and AAV PMNs (63 and 67%, respectively). No significant associations were found between the percentages of mPR3-positive PMNs with either ROS production or degranulation (data not shown). Moreover, there was no correlation with MFI of mPR3-positive cells and activation of the PMNs.

Importance of ANCA levels, epitope specificity and affinity

A possible source of the variations in PMN activation capacity between different IgG preparations could be differences in the proportion of IgG being ANCA. However, when comparing results from the different functional assays performed above with ANCA levels, no significant correlations were found. This was true regardless of whether ANCA was measured with a direct ELISA (data not shown) or with a capture ELISA (Fig. 5a).

Figure 5.

(a) Activation score do not correlate with capture proteinase 3 (PR3)-anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) levels. (b,c) Gelatinase and lactoferrin degranulation correlates with the human–mouse chimeric PR3 proteins HHm and Hm added together; rs = 0·649, P = 0·049 and rs = 0·661, P = 0·044, respectively. (d) Activation score correlates with PR3 affinity: rs = 0·758, P = 0·015. Spearman's correlation coefficient rs was calculated.

Commercially available PR3 ELISAs are designed to measure all epitope specificities, and previous studies indicate that certain epitope specificities of PR3–ANCA are more pathogenic than others [11, 12]. To investigate this, we used eight different mouse–human chimeric PR3 proteins developed previously by our group [17, 18], and tested all IgGs for reactivity towards the different parts of the PR3 molecule. Special emphasis was put upon two of the chimeras: HHm (two-thirds human sequence starting in the N-terminal part and one-third mouse sequence) and Hm (one-half human sequence, one-half mouse sequence), as they include an epitope defined by Silva et al. (designated by them as ‘epitope 1’). This same epitope corresponds to an epitope proposed by Kuhl et al. to be pathogenic, as it seems to interfere with α1-anti-trypsin binding and perhaps PR3's enzymatic activity [11, 12]. It includes the amino acids at positions 35–39 and 72–78 in the PR3 molecule. When combining the epitope ELISA results from HHm and Hm, we found significant correlations between reactivity to these chimeras and gelatinase release (rs = 0·649, P = 0·049), as well as lactoferrin release (rs = 0·661, P = 0·044, Fig. 5b,c). No other significant correlations, including the activation score, were found. The epitope assay did not correlate with ROS and NET assays. No correlations were found when any of the eight individual constructs were correlated with any of the activation assays.

Another independent factor possibly influencing the activation of PMNs is affinity, which was measured using a competition ELISA. The IgGs were incubated with increasing concentrations of PR3 overnight before being tested in a PR3 ELISA. The results show that PR3 affinity correlated with the activation score (rs = 0·758, P = 0·015, Fig. 5d) as well as with total ROS production (luminol assay, rs = 0·721, P = 0·023, data not shown). Extracellular ROS alone, gelatinase and lactoferrin release had a tendency to correlate with affinity, but did not reach statistical significance. There was no correlation between affinity and ANCA levels (as measured by standard commercial PR3 ELISAs), emphasizing the importance of affinity for ANCA-mediated activation of neutrophils.


In this study we aimed to decipher whether PMNs from AAV patients behave differently in response to ANCA stimulation compared with PMNs from HCs. We could see an increased intracellular ROS production in comparison with HC PMNs, whereas extracellular ROS production and degranulation were similar. The reason for increased ROS production is unknown, but we speculate that AAV patients' PMNs, even though in clinical remission, are exposed to a low-grade activation or priming. It has been shown previously that the cytokine profile of patients in the active phase as well as the remission phase is different from healthy donors [19]. There might be other explanations for this, including intrinsic defects in the PMNs in AAV patients. AAV patients have an increased membrane expression of PR3 on their neutrophils and this has been connected to their higher degree of stimulation [20, 21]. However, PMNs from AAV patients in this study did not have significantly more PR3 on the surface than the HC PMNs. All patients received low-dose glucocorticoids, but that this has any stimulatory effect on ROS production is unlikely, as glucocorticoids have been shown to reduce ROS production in PMNs [22].

Why only intracellular and not extracellular ROS production differed remains elusive, and it must be emphasized that our data are based on a small number of observations. The major ROS-producing complex in phagocytes is the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase that resides in the plasma membrane (5%) and in granule membranes (95%) in neutrophils. Activation of NADPH oxidase in the plasma membrane leads to production of ROS that is released to the extracellular space. NADPH oxidase activation in the phagosome, by fusion of NADPH oxidase-containing granules to the phagosome, is also a well-established defence mechanism; however, there also exists intracellular non-phagosomal ROS production that seems to play an important role for cell signalling and immune regulation [23]. It is also known that the assembly is somewhat different for extracellular and intracellular ROS production. For example, the cytosolic component p40phox is needed for intracellular, but not extracellular, NADPH oxidase assembly [24]. There is also evidence for the involvement of different isoforms of PI3K and PLC in intracellular and extracellular ROS production, suggesting that different stimuli distinguish between these pathways [24, 25]. Moreover, ROS produced in one compartment can stimulate ROS production in another: so-called ROS-induced ROS release. If the balance between the anti-oxidant and oxidant system is disturbed, excessive ROS could possibly initiate a self-perpetuating cycle, leading to more damage [26].

NET formation has been described previously to be dependent upon ROS production [27]. A role for NET formation in AAV has also been proposed, as purified AAV IgG was found to induce more NET formation compared with HC IgG. In the same study it was also shown that AAV patients had more circulating NET remnants, in terms of MPO–DNA complexes [2]. Our data support this finding, as NET formation correlates with ROS release. Increased NET formation may be of importance for the development of ANCA, as NETs give rise to an increased antigen exposure for the immune system. In a proinflammatory milieu, this could supposedly break tolerance towards PR3 or MPO.

Our 10 different AAV IgGs exhibited different activation potential on PMNs. It is highly unlikely that this is due to differences in cytokine profiles [19] or different treatment regimens, as the IgG was purified and soluble factors would be eliminated. Moreover, the activity of the patients, as measured by the BVAS, from whom the IgG was purified, did not influence the results in any of our assays. More surprisingly, ANCA levels assessed by standard direct PR3–ANCA or capture PR3–ANCA ELISAs did not correlate with any of our data. Conversely, the antibody affinity for PR3 correlated with the activation score, and this indicates that ANCA affinity is more important than ANCA levels for PMN activation. A previously identified epitope region in the PR3 molecule has been suggested to be of importance for AAV [11, 12]. We could confirm this in this study, as degranulation correlated with reactivity to our chimeras that included this epitope. These results suggest that ANCA with certain epitope specificities are more potent PMN activators than others, and that these epitopes are situated in the N-terminal region of PR3.

In conclusion, PMNs from AAV patients show an increased intracellular ROS production, but otherwise behave in a similar way as PMNs from HCs, opening the door for further studies within the field of cell signalling and the involvement of ROS pathophysiology in these diseases.


The authors thank the nurses at the Nephrology Clinic at Skåne University Hospital for their assistance in collecting blood samples. This work was funded by the Crafoord Foundation, the Greta and Kock Foundation, Swedish Rheumatism Association and the Swedish Research Council 65x-15152.


The authors declare no conflicts of interest.

Author contributions

S. M. O. planned, performed experiments and wrote the article. S. O., M. S and T. H. planned experiments and wrote the article. D. S., L. G and Å. P. performed experiments.