Delayed neutrophil apoptosis in granulomatosis with polyangiitis: dysregulation of neutrophil gene signature and circulating apoptosis-related proteins.

Objectives: Neutrophil apoptosis is mandatory for resolving inflammation and is regulated by expression of pro- and anti-apoptotic genes. We studied neutrophils isolated from patients with granulomatosis with polyangiitis (GPA) to investigate apoptosis alterations and to identify transcriptional and circulating factors affecting this process. Method: We enrolled 36 patients (18 in active stage, 18 in remission) and 18 healthy controls. Circulating levels of tumour necrosis factor-α (TNF-α), granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage migration inhibitory factor, plasminogen activator inhibitor-1, interferon-γ, vascular cell adhesion molecule-1, intercellular adhesion molecule-1, platelet endothelial cell adhesion molecule-1, soluble Fas (sFas), sFas ligand, survivin, and pentraxin-3 (PTX3) were evaluated by enzyme-linked immunosorbent assay/Luminex; circulating apoptotic neutrophils by flow cytometry; and apoptosis-related gene transcripts by real-time polymerase chain reaction. Results: Patients had decreased fractions of circulating apoptotic neutrophils and delayed neutrophil apoptosis was present in vitro. Circulating levels of TNF-α, GM-CSF, sFas, and PTX3 were higher in GPA. Delayed neutrophil apoptosis was accompanied by decreased mRNA of pro-apoptotic genes and transcription factors (DIABLO, PMAIP1, BAX, CASP3, CASP7, RUNX3, E2F1, TP53) and increased anti-apoptotic CFLAR and BCL2A1 mRNA. TNF-α and sFas levels correlated with circulating apoptotic neutrophils and expression of apoptosis genes. Stimulation with TNF-α of neutrophils from controls significantly down-regulated E2F1 and CASP3 expression. Conclusions: Circulating neutrophils in GPA have anti-apoptotic phenotype involving both intrinsic and extrinsic pathways of apoptosis. This is accompanied by increased levels of circulating pro-survival factors (GM-CSF, TNF-α, sFas), independent of disease activity. Anti-apoptotic phenotype of neutrophils in GPA is reproduced by exposure to low concentrations of TNF-α.

Granulomatosis with polyangiitis (GPA) is a rare systemic autoimmune disease affecting several organs including the respiratory tract and kidneys. GPA is characterized by the presence of circulating autoantibodies against the neutrophil serine protease proteinase-3 [immunoglobulin G (IgG) anti-PR3] (1,2). Activation of neutrophils by anti-PR3 IgG antibodies initiates the expression of several genes and incites pathways causing the release of cytokines, production of reactive oxygen species (ROS), or formation of neutrophil extracellular traps (3)(4)(5). The eventual fate of the neutrophil is apoptosis. Acceleration of this process is observed in some infectious or hereditary inflammatory diseases (6)(7)(8)(9), whereas delay is reported in asthma, idiopathic pulmonary fibrosis, and rheumatoid arthritis (7,10). Two apoptotic pathways are present in neutrophils: intrinsic and extrinsic. Mitochondrial outer membrane permeabilization (MOMP) is a key trigger for the intrinsic apoptotic pathway (11). MOMP is mediated by two pro-death proteins belonging to the Bcl2 family: BAX and Bak. Mitochondrial membrane depolarization leads to the release of cytochrome c and mitochondrial direct IAPbinding protein with low pI (DIABLO) into the cytosol. DIABLO inhibits pro-survival molecules, whereas cytochrome c and apoptotic protease-activating factor (APAF) form the caspase-9/caspase-3 activation platform (12). The extrinsic pathway is initiated by cross-linking of cellular death receptors such as the Fas receptor (FasR), tumour necrosis factor receptor-I and -II (TNFRI and TNFRII), and TRAILR. However, some variability in the downstream mechanism of this pathway has been reported. ROS are produced in apoptosis activated by tumour necrosis factor-α (TNF-α) and mediated by the effector caspase-3. In Fasmediated apoptosis, caspase-8 activation is followed by induction of the intrinsic mitochondrial death pathway (13).
Neutrophil apoptosis in systemic autoimmune diseases has been investigated in several studies but conflicting results have been reported in GPA. Most previous studies investigated spontaneous apoptosis of cultured neutrophils from GPA patients (14,15). In the current study, we focused on the transcriptional regulation of apoptosis in circulating neutrophils from patients with GPA and included measurements of several circulating proteins modulating the process.

Method
Patients and design of the study In this observational non-randomized study, we enrolled 36 patients with GPA (18 in the active stage of disease and 18 in remission). As a control group, 18 healthy volunteers were matched for age and gender (for details see Table 1). Disease activity was evaluated using the Birmingham Vasculitis Activity Score (BVAS, version 3). Patients with score = 0 were in remission and patients with score ≥ 1 were in the active stage of GPA. Basic laboratory tests (complete blood count, C-reactive protein, and anti-PR3 IgG level) were performed in all study participants at the time of collection of the peripheral blood for isolation of neutrophils, evaluation of serum or plasma levels of apoptosis-related proteins, and flow cytometry. All patients in the active stage of GPA had blood samples collected before high-dose corticosteroid or immunosuppressive therapy was initiated. Written informed consent was obtained from all participants in the study and the study protocol was accepted by Jagiellonian University Ethics Committee.

Neutrophil isolation and culture
Neutrophils were isolated using density gradient centrifugation (Histopaque-1077; Sigma-Aldrich Chemical Co., St Louis, MO, USA) followed by purification with a commercial magnetic negative selection kit, according to the manufacturer's protocol (EasySep Human Neutrophil Enrichment Kit; STEMCELL Technologies, Vancouver, BC, Canada). Purity of the neutrophil fraction (> 98%) was determined by flow cytometry and cell viability (> 95%) was verified by Trypan blue exclusion staining (for details see Supplementary figure S1). Immediately after isolation of neutrophils, total cellular RNA was isolated for gene expression studies or the neutrophil pellet was

Flow cytometry
The percentage of circulating apoptotic neutrophils was ascertained in the whole citrate anticoagulated blood using an Epic XL Beckman flow cytometer, as follows: 50 µL of blood was lysed with 1 mL of ammonium sulphate lysis buffer (10 min on ice), centrifuged (5 min at 500 g), and the cell pellet was suspended in the assay buffer. Staining was carried out for 15 min with annexin-V/7-AAD (Annexin-V Apoptosis Detection Kit, BD Biosciences, San Jose, CA, USA). Spontaneous apoptosis of isolated neutrophils was evaluated at 1, 4, and 24 h of cell culture. Cultured cells were collected by centrifugation, resuspended in the assay buffer, stained with annexin-V/7-AAD and analysed by flow cytometry (for gating strategy details, see Supplementary figure S2).
Total RNA isolation, reverse transcription, and gene expression in neutrophils Total RNA was isolated from neutrophils using RNAzol reagent according to the manufacturer's procedure (Sigma-Aldrich Chemical Co., USA). Reverse transcription was performed using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies, USA). Screening for 93 apoptosis-related transcripts in six randomly selected patients with active GPA and six healthy controls was carried out using quantitative realtime polymerase chain reaction (qRT-PCR) with a TaqMan Human Apoptosis Array (Life Technologies, USA; for a list of genes see Supplementary table S1). The mRNA abundance of candidate genes selected by the screening was measured in all study participants by qRT-PCR using TaqMan chemistry (7900HT Fast Real-Time PCR System; Life Technologies, USA). TaqMan Human Apoptosis Arrays were also used to evaluate the expression of apoptosis-related genes in neutrophils after 24 h of incubation. Quantification cycle data were normalized to ribosomal 18S rRNA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Results were calculated using the 2 -ΔCt (relative expression) or 2 -ΔΔCt (fold change) formula from the appropriate endogenous controls (16).

Additional measurements
Evaluation of neutrophil surface CD11b expression, serum myeloperoxidase (MPO), neutrophil elastase (NE), and DNA-MPO complexes has been described in detail and published elsewhere (17). In brief, neutrophils' CD11b expression was analysed in whole blood samples by flow cytometry. Serum levels of DNA-MPO complexes were measured by ELISA, and MPO and NE were assessed by a Luminex assay.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA, USA). All comparisons were made using one-way analysis of variance (ANOVA) with Tukey's post-hoc test or the Kruskal-Wallis test for non-parametrically distributed variables with Dunn's post-hoc test. Descriptive statistics are presented as mean ± standard deviation (sd) or median ± interquartile range (IQR). Correlations between analysed factors were calculated by Spearman's rank method. mRNA expression level analysis was conducted using DataAssist Software (Thermo Fisher Scientific, Waltham, MA, USA). Type I statistical error p < 0.05 was considered significant. Bioinformatics pathway analyses were performed with the use of DAVID Bioinformatics Resources 6.7 and EVEX databases (18,19).

Clinical characteristics of study participants
In this study, we analysed neutrophil apoptosis and levels of circulating apoptosis-related proteins in 36 GPA patients and 18 healthy controls (Table 1). All GPA patients were anti-PR3 IgG positive. Eighteen patients with GPA were in the active stage of the disease (BVAS median = 14) and 18 were in remission from GPA (BVAS = 0). In the group of patients with active GPA, 10 patients were newly diagnosed while remaining ones had disease exacerbation. However, none of patients with active GPA received intensive immunosuppressive treatment (rituximab, cyclophosphamide, or high-dose corticosteroids) before collection of blood samples.
Clinical laboratory results showed elevated peripheral blood mononuclear cell (PBMC) and neutrophil counts in the group of patients with active GPA (compared to healthy controls, p < 0.05), while the platelet number was elevated in both active GPA patients and those in remission.

Circulating neutrophil apoptosis and expression of apoptosis-related genes
First, we analysed apoptosis of circulating neutrophils in GPA patients and healthy controls. In patients with GPA, the percentage of circulating apoptotic neutrophils was lower in comparison with healthy controls (in active GPA 2.9 ± 3.1%, in remission 2.47 ± 4.9% vs healthy controls 6.9 ± 6.1%; median ± IQR, p = 0.001) ( Figure 1A).
To explain the low percentage of circulating apoptotic neutrophils in GPA patients at the molecular level, we analysed the expression of 93 genes involved in the regulation of apoptotic pathways. mRNA was detectable for all 93 analysed genes. After the initial screening on neutrophils from GPA patients (n = 6) and healthy volunteers (n = 6), 13 candidate transcripts were selected for measurements in all participants in the study. Eleven were down-regulated (PMAIP1, DIABLO, CASP3, CASP7, CASP8, BAX, CHUK, CARD9, BCL3, BBC3, and CASP8AP2), and two were up-regulated (CFLAR and BCL2A1) (Supplementary figure S3). We also included quantification of mRNA from three transcription factor (RUNX3, E2F1, and TP53), selected after bioinformatics analysis.
All patients with GPA had eight transcripts significantly down-regulated in neutrophils (PMAIP1, DIABLO, CASP3, CASP7, BAX, RUNX3, E2F1, and TP53), whereas two mRNAs were up-regulated (CFLAR and BCL2A1) in comparison with healthy controls ( Figure 1B, Supplementary table S2). There were no differences in the percentage of circulating apoptotic neutrophils or expression of the analysed mRNAs between treatment-naïve patients who did not receive glucocorticoids and the remaining GPA group (Supplementary figure S4).
The activation of neutrophils observed in GPA patients leads to the release of cytokines and granule proteins and the formation of neutrophil extracellular traps. In our previous research, we showed elevated neutrophil CD11b expression, serum MPO, NE, and DNA-MPO complexes in patients with GPA (17). The current study included 48 participants (18 in the active stage of GPA, 12 in remission of GPA, and 18 healthy controls) examined in the previous study. To analyse whether delayed neutrophil apoptosis could be linked to the mentioned parameters, we correlated them to the percentage of circulating apoptotic neutrophils and neutrophil expression of apoptosis-related genes. We observed that a low level of circulating apoptotic neutrophils was accompanied by high levels of neutrophil CD11b expression, circulating DNA-MPO complexes, MPO, and NE (Supplementary figure S5A-D). Moreover, levels of these parameters were positively correlated with upregulated neutrophil anti-apoptotic genes (BCL2A1 and CFLAR), and negatively correlated with down-regulated neutrophil pro-apoptotic genes (CASP3, CASP7, DIABLO, and PMAIP1) and transcription factors (RUNX3, E2F1, and TP53) (Supplementary figure S5E).

Cultured neutrophil apoptosis and expression of apoptosis-related genes
To analyse whether disturbance of neutrophil apoptosis may be linked to environmental factors, and to establish cell death kinetics, we conducted experiments in which isolated neutrophils were cultured for up to 24 h and apoptosis was evaluated by annexin-V/7-AAD staining. During neutrophil culture, a decreased ratio of spontaneous apoptosis was present in GPA (active and remission) only after 1 h and 4 h (1 h: active GPA 3.6 ± 1.6% and remission 3.2 ± 2.5% vs healthy controls 6.8 ± 7.2%, p = 0.01; 4 h: active GPA 6.9 ± 3.7% and remission 6.6 ± 5.6% vs healthy controls: 12.5 ± 5.6%, median ± IQR, p = 0.01) ( Figure 4A). There were no differences in the percentage of apoptotic neutrophils after 24 h of incubation. Moreover, comparison of the expression of 93 apoptosis-related genes between patients with active GPA, remission of GPA, and healthy controls after 24 h of incubation, found no significant differences ( Figure 4B, C).
To confirm that the mRNA expression pattern observed in circulating neutrophils of GPA patients is associated with environmental circulating factors such as TNF-α, we performed an experiment in which neutrophils isolated from healthy donors (n = 5) were stimulated with a concentration of TNF-α observed in patients with GPA (25 pg/mL). After   24 h of stimulation, the percentage of viable neutrophils (ANX5 -/7-AAD -) was higher when cells were stimulated with TNF-α, and it was followed by significant downregulation of E2F1, CASP7, and PMAIP1, and upregulation of CFLAR ( Figure 5).

Discussion
Neutrophil apoptosis is mandatory for the resolution of inflammation. Although increased apoptosis may lead to an incompetent immune response, a delayed response would contribute to the persistence of inflammation. An early study by Harper et al demonstrated that neutrophils from the peripheral blood of patients with anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) had enhanced apoptosis in vitro. ANCAs themselves accelerated apoptosis of neutrophils (14). In contrast, Abdgawad et al reported decreased apoptosis of cultured neutrophils from GPA patients (15). Our results are in line with the latter observation. In our study, spontaneous apoptosis of neutrophils from GPA patients was delayed. When analysed at different time points ex vivo, this delay was noticeable only within the first 4 h of incubation. We suggest that the decreased rate of spontaneous apoptosis in GPA patients could be explained by a persistent stimulation of neutrophils with pro-survival proteins present in the circulation. Apoptotic cells are distinguished by their surface expression of certain membrane molecules (20). Physiologically, the clearance of apoptotic cells by macrophages coincides with the suppression of pro-inflammatory cytokines (TNF-α) and increased production of anti-inflammatory mediators such as transforming growth factor-β and prostaglandin E 2 . Thus, the resolution activity of macrophages is not accompanied by the release of pro-inflammatory cytokines (20). This process was reported to be dysregulated in systemic lupus or chronic granulomatous disease (21)(22)(23). To test the hypothesis that clearance of apoptotic neutrophils is altered in GPA patients, we analysed the level of circulating PTX3. This protein is produced locally at the site of inflammation under the control of pro-inflammatory signals, usually by endothelial cells and neutrophils (24). It has been demonstrated that PTX3 can bind to the late apoptotic neutrophils, inhibiting their phagocytosis by macrophages (25,26). Levels of PTX3 were increased in the active phase of GPA but did not correlate with the percentage of apoptotic neutrophils in the circulation. A high concentration of PTX3, especially in GPA patients with active disease, seems to be a biomarker of neutrophil activation and extracellular trap formation (27,28) rather than the modifier of apoptosis. The positive correlation between the plasma level of PTX3 and circulating DNA-MPO complexes reported in this study seems to confirm this observation.
TNF-α, GM-CSF, and Fas-signalling proteins are probably the most studied among various factors modifying the apoptosis of neutrophils. GM-CSF is a pro-survival Cross-correlations between measured circulating proteins and circulating apoptotic neutrophils or their expression of selected genes in all study participants. Significant correlations coefficients are marked with bold font. (C) Cross-correlations between mRNA expression of analysed apoptosis-related genes in all study participants. Significant correlations coefficients are marked with bold font. The association between analysed parameters was tested using Spearman rank correlation. TNF-α, tumour necrosis factor-α; GM-CSF, granulocytemacrophage colony-stimulating factor; sFas, soluble Fas; PTX3, pentraxin-3. haemopoietin delaying neutrophil apoptosis, mainly by the induction of anti-apoptotic BCL2A1 and MCl-1 (29). Chiewchengchol et al demonstrated that GM-CSF stimulation of neutrophils also leads to down-regulation of proapoptotic BAX (29). Increased levels of circulating GM-CSF were previously reported in GPA patients (15,29) and the current study supports this observation.
Fas-mediated neutrophil apoptosis is caspase dependent. Binding to the cellular Fas receptor (FasR) leads to activation of caspase-8, increased expression of caspase-3, and activation of mitochondria-dependent death machinery, which includes mitochondrial apoptotic proteins Bid, BAX, and DIABLO (11). In addition to the cellular FasR, a splicing variant of this molecule circulates in the blood as a soluble form (sFas). sFas can inhibit binding between Fas ligand (FasL) and FasR (30). Elevated levels of circulating sFas have been reported in several autoimmune diseases, including AAV (15,31,32). In the current study, we measured circulating levels of sFas and its soluble ligand (sFasL). We observed no differences in sFasL but the sFas level was significantly higher in GPA patients. This is in line with other observations suggesting a defect of Fas-mediated death signalling in GPA. Moreover, the lowered abundance of transcripts encoded by pro-apoptotic BAX, DIABLO, and PMAIP1, all participating in the mitochondrial apoptosis pathway, provides another argument that Fas signalling fails in neutrophil apoptosis in GPA patients.
TNF-α can modulate apoptosis of neutrophils in a nonlinear way, depending on the concentration of the cytokine. At high concentrations (> 10 ng/mL), TNF-α increases apoptosis via an ROS-dependent mechanism, whereas at low concentrations (< 10 ng/mL), neutrophil survival is prolonged. In a previous study, TNF-α was undetectable in the plasma of GPA patients (15). Using our immunofluorescence measurements, we found higher concentrations of serum TNF-α in GPA patients than in healthy volunteers. This discrepancy may result from the assay sensitivity. A detailed molecular background to TNF-α-regulated neutrophil apoptosis was described by Chiewchengchol et al (29). Stimulation of neutrophils with a low concentration of TNF-α protected cells against apoptosis and led to activation of anti-apoptotic genes (CLFAR, BCLA2A, and TNFAIP3) and inhibition of pro-apoptotic genes (CASP8, FADD, TNFRSF1A, and TNFRSF1B). Our results are in line with this report. The percentage of circulating apoptotic neutrophils in GPA patients was lower, while the expression of anti-apoptotic CFLAR was positively correlated with the serum level of TNF-α. Moreover, our experiments showed that stimulation of neutrophils with a much lower concentration of TNF-α (25 pg/mL) inhibited the expression of pro-apoptotic CASP7, probably by regulation of E2F1 transcription factor.
Neutrophils, owing to the fragility of the cells in vitro, can be investigated for genetic expression only by a limited number of methods. For this reason, only correlation analysis of transcripts was feasible, which is an evident limitation of the study.
RUNX3 transcription factor has been proposed as a master controller for the expression of PR3 and MPO in GPA (33). However, this protein can also participate in apoptosis. In cancerous cells, RUNX3 can control the expression of anti-apoptotic CFLAR and pro-apoptotic CASP3 and BAX (34)(35)(36). Our results seem consistent with this view. In neutrophils from GPA patients, expression of CASP3 and BAX was down-regulated, whereas expression of CFLAR was up-regulated. There was a positive correlation between RUNX3 mRNA and CASP3 or BAX, and negative correlation with CLFAR. RUNX3 mRNA correlated positively with the percentage of circulating apoptotic neutrophils. Moreover, RUNX3 can also regulate the function of other transcription factors related to apoptosis, e.g. p53. Zhai et al showed that RUNX3 can act as a co-activator of this transcription factor and has an impact on p53 phosphorylation (36). Since p53 responsive elements were identified in apoptosis-related genes such as PMAIP1 or BAX (37-39), some   activity downstream of p53 can occur. In our study, decreased expression of PMAIP1 or BAX was accompanied by decreased TP53 mRNA transcripts in neutrophils from GPA patients. Our mRNA quantification also included another transcription factor participating in apoptosis, namely E2F1. E2F1 can affect apoptosis directly by a p53-independent mechanism or in concert with p53. The p53-dependent pathway of E2F1 regulates the expression of apoptosisrelated genes by modulating p53 expression. This is achieved by a negative effect on the p53 inhibitor (MDM2) or by enhanced transcription of p53 cofactors (ASPP1, ASPP2, or JMY). Independent E2F1 activity induces pro-apoptotic genes, including CASP3, CASP7, PMAIP1, and DIABLO, by directly binding to their promoters (40,41). Decreased expression of E2F1 target genes and their correlation with E2F1 mRNA suggest the importance of this transcription factor in mechanisms regulating the neutrophil apoptosis pathway in GPA.
All of the analysed transactivators of expression had lower levels of transcripts in neutrophils from GPA patients compared to controls. Therefore, it is difficult to identify the dominant one in the current study. Hypermethylation of RUNX3 promoter was described in neutrophils from GPA patients (33), but no studies have investigated the expression of E2F1 or TP53. It would be useful to characterize the regulation of E2F1 expression because in our study this transcription factor was down-regulated by TNF-α.

Conclusion
This study presents novel findings on neutrophil pathobiology in GPA patients. An anti-apoptotic phenotype of circulating neutrophils from GPA patients is associated with high levels of neutrophil activation markers such as CD11b or the production of neutrophil extracellular traps. Circulating neutrophils in GPA have an anti-apoptotic umbrella characterized by low levels of pro-apoptotic transcripts and high levels of anti-apoptotic transcripts. This neutrophil property is readily removed following 4 h of cell culture, when the cells enter apoptosis in the same way as the control cells. The anti-apoptotic molecular phenotype of neutrophils from GPA patients involves both intrinsic and extrinsic pathways of programmed cell death and is probably imposed by increased levels of circulating pro-survival factors such as GM-CSF, TNF-α, or sFas. However, an important limitation of our research is the small size of the studied groups. To improve our understanding of the mechanism of neutrophil activation and apoptosis in GPA patients, further studies are required.