Transgenic Mice Expressing Human Proteinase 3 Exhibit Sustained Neutrophil-Associated Peritonitis

Neutrophil serine proteases play an important role in host defense and inflammation: not only do they act as powerful antimicrobial agents, they are also able to activate various receptors that are important in promoting inflammation and modulating the levels and activity of cytokines and chemokines (1). Indeed, inhibition of neutrophil serine proteases has been shown to reduce neutrophil infiltration and damage mediated by neutrophils in various animal models of inflammation, including noninfectious inflammation, ischemia and reperfusion injury, and collagen-induced arthritis (1–4). Proteinase 3 (PR3) belongs to the family of neutrophil serine proteases that also includes elastase (NE), cathepsin G (CG), and the enzymatically inactive azurocidin (5). Although PR3 is primarily stored within azurophilic granules of mature neutrophils, expression has also been observed at the plasma membrane and within secretory vesicles (6–8). During neutrophil activation and apoptosis, membrane expression of PR3 increases, and soluble PR3 can be released into the extracellular environment through degranulation (7, 9, 10).

PR3 is a multifunctional protein with a number of well-characterized proinflammatory properties. For example, PR3 is involved in the induction and processing of proinflammatory cytokines and chemokines, including TNF-α, IL-8, and IL-1β, which serve to increase their potency and amplify the inflammatory process (11). Additionally, PR3 can cleave and inactivate anti-inflammatory molecules, such as progranulin and annexin A1 (AnxA1) (4, 12, 13). It has recently been demonstrated that PR3 can disrupt the resolution of inflammation by disturbing macrophage activation (14, 15). The clearance of apoptotic neutrophils is an essential step in the resolution of inflammation, because dying neutrophils cause excessive tissue damage through the uncontrolled release of their proinflammatory cytotoxic contents. PR3 externalized during neutrophil apoptosis interferes with the ability of macrophages to efferocytose and clear dying cells (14). Furthermore, following efferocytosis, PR3 expressed on apoptotic neutrophils perpetuates inflammation by increasing the production of proinflammatory cytokines and chemokines, including TNF-α, CCL2, and IL-6 (15). Lastly, PR3 plays a role in cell survival and proliferation through its ability to cleave the cyclin-dependent kinase inhibitor p21/waf1/CIP1 and is involved in apoptosis via its interaction with procaspase-3 (8, 16, 17). PR3 is believed to play a role in a number of inflammatory diseases, including chronic inflammatory lung diseases, rheumatoid arthritis, and autoimmune vasculitis (18, 19). In particular, PR3 is the main target of anti-neutrophil cytoplasmic Abs (ANCAs) in granulomatosis with polyangiitis, a systemic autoimmune disease characterized by granulomatous inflammation and necrotizing vasculitis (20). It is worth noting that none of the other serine protease homologs are a specific target for ANCAs; this suggests that PR3 may have some specific biochemical or structural properties that promote the generation of ANCAs and contribute to the pathophysiology of the disease (21, 22).

To further investigate the function of PR3 in inflammation, we generated a novel human PR3 transgenic mouse (hPR3Tg) model. hPR3 was expressed because the human ortholog possesses unique structural and biochemical features that may amplify its pathogenic activity (23). Using these transgenic mice, this study examined the consequences of hPR3 expression on the initiation and resolution of inflammation, as well as its effects on neutrophil function and survival.

Animal studies were performed in accordance with European Community Guidelines. All protocols were approved by the Université Paris Descartes Ethics Committee for Animal Research or the Cantonal Veterinary Office of Bern. Mice had access to food and water ad libitum and were housed under a 12-h light/dark cycle. To generate hPR3Tg, hPR3 cDNA was cloned into a pWHERE-rEF1 plasmid (Invitrogen) and placed under control of the rEF1 promoter using NcoI and NheI restriction enzyme sites. The plasmid was purified and linearized using PacI, and several copies of target construct DNA were microinjected into the male pronucleus of fertilized mouse oocytes (CBA/J × C57BL/6 genetic background) (Service des Animaux Transgéniques, CNRS, Villejuif, France). Two independent hPR3Tg lines were established, and RT-PCR demonstrated that both lines expressed the transgene. The resulting offspring were backcrossed for at least six generations with C57BL/6J mice, resulting in an average 98.37% C57BL/6J background, and the transgenic line was maintained in a hemizygous state. All wild-type mice (WT) used throughout this study were littermate controls. Routine genotyping was carried out by PCR on genomic DNA from tail or ear biopsies using the following primers: 5′-GTTTCCAGGTGTTGTGAAAGCCACCGC-3′ and 5′-CCACTGATTAAGAGTGGGGTGGCAGG-3′ to yield an amplification product of 791 bp.

Two tibias and two femurs from each mouse were flushed to collect bone marrow (BM) cells, using HBSS–20 mM HEPES–0.5% FCS. Cells were resuspended in 5 ml of 0.2% NaCl for 40 s to lyse RBCs, and 5 ml of 1.6% NaCl was added. Separation of neutrophils and their precursors was performed, as previously described, using a two-layer Percoll gradient with densities 1.065 and 1.080 g/ml (24). Briefly, 2 × 10⁸ cells from the BM aspirate were placed on a two-layer Percoll gradient; following a 20-min centrifugation at 1000 × g at 4°C, neutrophils were collected from the pellet. Total RNA was extracted using TRIzol (Invitrogen), and sscDNA was synthesized from 5 μg of total RNA with random hexamer primers (Applied Biosystems) and Superscript II (Invitrogen). PCR was then used to detect hPR3 mRNA using the following primers: PR3 5′-CCGCAAGGCCGGCATCTGCT-3′ and 5′-GATAATGCTAGCTCAGGGGCGG-3′ and 18S 5′-GTAACCCGTTGAACCCCATT-3′ and 5′-CCATCCAATCGGTAGTAGCG-3′.

Sterile peritonitis was induced by i.p. injection of 1 mg of zymosan A from Saccharomyces cerevisiae (Sigma-Aldrich). After 4, 24, 48, and 72 h, the peritoneal cavity was washed with 1 ml of PBS, and lavage fluid was collected. Cells within the lavage fluid were collected by centrifugation for 5 min at 1500 rpm, counted, and subsequently used for flow cytometry analysis, Western blot, or to measure PR3 activity. Peritoneal fluid was collected and frozen at −20°C before measuring the secretion of proinflammatory cytokines and chemokines, as well as the levels of AnxA1. Protein concentrations were quantified using a BCA Protein Assay Kit (Bio-Rad), according to the manufacturer’s instructions.

Peritoneal lavage fluid was collected in 1 ml of PBS, and 15 μl of the sample was used for Western blot analysis. Cells within the lavage fluid were collected by centrifugation for 5 min at 1500 rpm and incubated on ice for 15 min in lysis buffer containing 10 mM HEPES, 0.3 mM DTT, 400 μM leupeptin, 400 μM pepstatin, 4 mM PMSF, 1 mM orthovanadate, 1 mM EDTA, 1 mM EGTA, and 1% Triton detergent. Fluid samples or cell lysates were separated on 10% SDS-PAGE gel (15 μl per lane), transferred to a polyvinylidene difluoride membrane (Bio-Rad), blocked for 1 h with 5% skim milk powder in TBST, and immunoblotted overnight at 4°C with rabbit anti-hPR3 Ab (1:200; gift from Dr. T. Hellmark, Lund University, Lund Sweden) or rabbit anti-mouse AnxA1 Ab (1:500; Invitrogen). Membranes were washed three times with TBST and incubated with HRP-conjugated goat anti-rabbit secondary Ab (Jackson ImmunoResearch). Blots were washed and developed using Super Signal West Pico Chemiluminescent Substrate (Thermo Scientific). Quantification of Western blots was performed using ImageJ software (National Institutes of Health); PR3 was normalized to actin and AnxA1 was normalized to total protein content within the lavage fluid.

Cells collected from the peritoneal lavage 4 h after i.p. injection of zymosan were lysed in 100 μl of lysis buffer containing 1% Triton in PBS. Protein concentrations were measured using a BCA Assay Kit, and PR3 activity was measured using the fluorescent substrate ABZ Tyr-Tyr-Abu-Asn-Glu-Pro-Tyr(3-NO2)-NH2; ABZ was excited at 320 nm, and its emission was followed at 450 nm, as previously described (25). Each reaction contained 20 μl of ABZ substrate (300 μg/ml), 210 μl of assay buffer containing 100 mM Tris–HCl/500 mM NaCl (pH 7.5), and one of the following: 20 μl of human neutrophil elastase or PR3 (final concentration 1 μg/ml), 0.01 mg of total protein from human polymorphonuclear cells (PMNs), or 0.5 mg of protein from each sample. Fluorescence was measured on a microplate spectrophotometer over 30 min, PR3 activity was determined as the slope of the reaction, and the results were expressed as the percentage increase relative to WT controls performed on the same day.

Peritonitis was induced by cecum ligation and puncture (CLP) in male mice aged between 12 and 16 wk, as previously described (26). All surgeries were performed in <15 min per mouse, with a maximum of 10 mice per procedure. After each mouse was anesthetized, the abdomen was shaved, and the cecum was exposed through a 1-cm midline incision. The cecum was ligated at three-quarters of its length with a number 5 silk suture (Ethicon) and punctured once with a 21-gauge needle. The cecum was replaced, and 1 ml of sterile saline (0.9% NaCl) was injected into the peritoneal cavity before suture. For experiments designed to assess survival, mice were monitored every 8 h within the first 3 d and then every 12 h until day 7. Animal well-being and overall health were scored using seven criteria, including activity, posture, movement, coat, breathing, alertness, and physiology of the eyes/nose at 24 and 48 h postsurgery. Animal well-being was not recorded at the later time points because of the death observed after 48 h. Cell recruitment into the peritoneal cavity was determined 6 h after the induction of sepsis. In these experiments, the peritoneal cavity was washed with 1 ml of PBS, and the lavage fluid was collected. Cells within the lavage fluid were collected by centrifugation, counted, and subsequently used for flow cytometry analysis. The number of bacteria within the peritoneal lavage was also estimated; the peritoneal cavity was washed with 1 ml of sterile PBS, and samples from these lavages were serially diluted on agar plates and grown at 37°C. The number of colonies was counted after 24 h, and CFU per milliliter was calculated. The remaining peritoneal fluid was collected and frozen at −20°C before measuring the secretion of proinflammatory cytokines and chemokines, as well as the levels of AnxA1. Cytokines released into the peritoneal cavity were analyzed, after i.p. injection of 1 ml of PBS, using a Cytometric Bead Array (CBA) Kit (BD Biosciences) or a membrane-based Ab array (Mouse Inflammation Array C1; RayBiotech), according to the manufacturers’ instructions.

P. aeruginosa PAO1 (ATCC-BAA-47; strain HER-1018) was grown on tryptic soy broth agar, and inoculum was prepared as described previously (27). For each experiment, bacterial inoculum was freshly prepared from a frozen glycerol stock, and the actual CFU in the inoculum was quantified by plating serial dilutions on tryptic soy agar plates. Groups of WT and hPR3Tg were sedated with 100 mg/kg ketamine and 10 mg/kg xylazine and inoculated intranasally with 8 × 10⁶ CFU per mouse by applying 10 μl of inoculum in each nostril. Mice were euthanized 24 h after inoculation. The dissected lungs were homogenized, serially diluted in 1% peptone, plated on tryptic soy agar plates, and incubated overnight at 37°C for CFU counting.

Blood samples were collected by cardiac puncture into tubes containing 1% 0.5 M EDTA and analyzed within 60 min using an MS9-5 Hematology Analyzer (Melet Schloesing).

Differential analysis of BM leukocytes was performed on 6–12-wk-old mice, as previously described (28). To identify hematopoietic stem and progenitor cells (HSPCs), BM was stained with fluorochrome-conjugated Abs against mouse Sca1 (E13-161.7), CD117/cKit (2B8), CD16/32 (E192222) (all from BioLegend), and CD34 (RAM34; eBioscience), and leukocyte subsets were identified using Abs against mouse CD11b (M1/70), B220/CD45R (RA3-6B2), CD19 (6D5), Ly6G (1A8), and CD115 (AFS98) (all from BioLegend). Peritoneal lavage cells were stained with fluorochrome-conjugated Abs against mouse CD11b (M1/70), Ly6C (NK1.4), Ly6G (1A8), CD45 (30-F11), CD3 (17-A2), CD19 (1D3), CD4 (RM4-5), CD8 (53-6.7), I-A/I-E (M5/114.15.2) (all from BD), and F4/80 (CI:A3-1) (Bio-Rad). After blocking Fc receptors with a combination of 2.4G2 supernatant, rat IgG, and hamster IgG, 5 × 10⁶ cells were stained with the Abs listed, diluted in PBS with 5% FCS for 30 min at 4°C. Cells were gated as follows: after gating out doublets, CD45⁺ cells were identified and then CD19⁺ B cells and CD3⁺ T cells (either CD4⁺ or CD8⁺) were gated. Macrophages (CD11b⁺F4/80⁺Ly6G⁻Ly6C^lo), neutrophils (CD11b⁺/Ly6G⁺), and monocytes (CD11b⁺F4/80⁻Ly6G⁻Ly6C⁺) were identified as previously described (29). Fluorescent cells were detected using a BD LSR Fortessa with Diva software (BD Biosciences) and analyzed using FlowJo software (TreeStar).

Hematopoietic stem cells (Lin⁻Sca1⁺cKit⁺ [LSK] cells) were purified from the BM of mice. For each replicate, the BM from two WT or hPR3Tg was combined, and a single-cell suspension was generated. Cells were stained with a mixture of anti-mouse lineage (Lin) Abs, including anti-CD3, -B220, –Gr-1, -Ter119, and -CD11b, and Lin⁻ cells were isolated using magnetic beads, according to the manufacturer’s instructions (Dynal). For sorting LSK cells, Lin⁻ cells were stained with PE-conjugated anti-Sca1 and FITC-conjugated anti-cKit Abs, and the cells were sorted using a FACSAria III (BD). The sorted LSK cells were cultured in IMDM supplemented with 10% FCS, murine stem cell factor (50 ng/ml), human FMS-like tyrosine kinase 3 ligand (100 ng/ml), human IL-6 (50 ng/ml), and human IL-11 (10 ng/ml). Cell counts were performed every second day for a period of 4 d, and cell expansion was calculated. Apoptosis was measured using Annexin V–FITC and 7-aminoactinomycin D (7AAD) after 4 d in culture.

WT and hPR3Tg received a single i.v. injection of EdU (0.2 mg). Whole blood and BM were collected 24, 48, and 72 h after EdU injection. Neutrophils were identified using an Ab against mouse Ly6G (1A8) and then cells were stained for intranuclear detection of EdU integrated into genomic DNA, using a Click-iT Plus EdU assay kit (Thermo Fisher), according to the manufacturer’s instructions. Flow cytometry analysis was performed using a BD Accuri C6 flow cytometer.

Neutrophils were collected from the peritoneal lavage of WT and hPR3Tg 4 h after i.p. injection of 1 mg of zymosan. Production of NADPH oxidase–dependent oxidants was measured by chemiluminescence in a single-photon luminometer (AutoLumat LB 953; Berthold) with the cyclic hydrazide luminol (5-amino-2, 3-dihydro-1,4-phthalazinedione). For each test, 7.5 × 10⁵ neutrophils were resuspended in 260 μl of HBSS in polystyrene tubes containing 200 μl of luminol (0.15 mM) and 40 μl of HBSS, opsonized zymosan (100 μg/ml final concentration), or PMA (1 μg/ml final concentration; Sigma-Aldrich). Luminescence activity was measured in duplicate over 40 min and expressed as integrated total counts, as described previously (30).

Neutrophils were isolated using a Ficoll gradient from the BM or peritoneal lavage fluid of WT and hPR3Tg 4 h after the i.p. injection of 1 mg of zymosan. BM cells were isolated by flushing femurs and tibias of 8–22-wk-old mice with HBSS. Spontaneous apoptosis and l-leucyl-l-leucine methyl ester (LLME)-induced cell death were investigated using whole BM after erythrocyte lysis by NH₄Cl, as previously described (28). Cells were cultured in DMEM (4 mM l-glutamine, 25 mM d-glucose, 1 mM sodium pyruvate) (Life Technologies) containing 1% FCS at 2 × 10⁶ cells per milliliter, with or without 100 μM LLME. At the indicated time points, cells were harvested, and viability was assessed using Annexin V–FITC and 7AAD. After gating on Ly6G⁺ events, viable cells were identified as FSC^hi7AAD⁻AnnexinV⁻, and apoptotic cells were identified as FSC^hi7AAD⁻AnnexinV⁺ or FSC^hi7AAD⁺AnnexinV⁺ to account for early and late apoptotic cells, respectively. To assess neutrophil survival in zymosan-elicited neutrophils, 1 × 10⁶ cells were cultured in 500 μl of RPMI 1640 supplemented with 10% FCS and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin). Neutrophils were cultured for 16 h at 37°C in 5% CO₂ in the presence or absence of TNF-α (25 μM) or G-CSF (50 μM), and apoptosis was measured using Annexin V–FITC and 7AAD (BD Biosciences), according to the manufacturer’s instructions. Flow cytometry analysis was performed using a BD Accuri C6 flow cytometer.

Rat basophilic leukemia (RBL) cells stably transfected to express recombinant hPR3 or the control RBL cells transfected with an empty plasmid were cultured as previously described (31). Apoptosis of RBL cells was induced by UV-B (5 min, l = 312 nm, 0.68 mW/cm²); the cells were cultured for an additional 20 h before being evaluated by flow cytometry with Annexin V–PE/7AAD (BD Biosciences) and labeled with the fluorescent dye TAMRA, as previously described (14). Apoptotic RBL cell clearance by resident peritoneal macrophages was studied in 12-wk-old WT and hPR3Tg following a single i.p. injection of 1 × 10⁷ apoptotic RBL cells. After 30 min or 2 h, peritoneal cells were recovered by lavage with PBS, and phagocytosis was assessed after macrophage immune labeling with an FITC-coupled rat anti–mouse-F4-80 Ab (AbD Serotec) and analyzed by flow cytometry. The percentage of double-positive macrophages was considered to represent the percentage of macrophages that had phagocytosed TAMRA-labeled apoptotic RBL cells.

Data are shown as mean ± SEM. Statistical analyses were performed using Prism GraphPad 6 software. Comparisons were made using a Student t test or ANOVA analysis, where appropriate. Differences were considered significant at p < 0.05.

hPR3 cDNA encoded by PRTN3 was cloned under control of the rEF1 promoter in a pWHERE-rEF1 plasmid containing mH19 insulators. The purified construct was microinjected into the male pronucleus of fertilized mouse oocytes. Two founder lines were identified by PCR analysis and transmitted the transgene to their progeny, and hPR3Tg were maintained as hemizygous. Initial experiments were performed on both founder lines, including assessing the expression levels of PR3, as well as peritonitis induced by CLP; results were identical for both lines. Subsequently, the results for only one founder line are presented throughout the article. hPR3 mRNA was detected in neutrophils isolated from BM of hPR3Tg but not their WT littermates (Fig. 1A). Western blot analysis of the peritoneal lavage cells using an Ab capable of detecting human and mouse PR3 demonstrated significantly higher levels of protein in hPR3Tg compared with WT (Fig. 1B). The enzymatic activity of PR3 within these cells was measured using ABZ substrate, a fluorescent substrate specifically cleaved by PR3 but not NE (Fig. 1C, 1D) or other serine proteases (25). Recruited peritoneal cells of hPR3Tg exhibited a significant increase in PR3 activity compared with WT controls, suggesting that the human transgene was fully processed and enzymatically active (Fig. 1C, 1D). It is important to note that there was no significant difference in the proportion or number of neutrophils within the peritoneal lavage fluid of WT and hPR3Tg (80.38 ± 2.14 versus 75.58 ± 2.35%, respectively) at the time point examined, and the increase in PR3 protein and activity cannot simply be attributed to an increase in neutrophils, which are known to contain very high levels of serine proteases.

HSPCs located within the BM are self-renewing and multipotent cells that are capable of generating all cells of the blood and immune system (32). During myeloid cell differentiation, HSPCs undergo progressive commitment to produce common myeloid progenitor cells (CMPs), which, in turn, commit to megakaryocyte-erythrocyte progenitor cells (MEPs) or granulocyte-monocyte progenitor cells (GRMPs) (33). Finally, GRMPs can produce terminally differentiated granulocytes and monocytes. PR3, also called myeloblastin, has been shown to enhance the proliferation of hematopoietic cells responding to G-CSF during granulopoiesis (34–36). The consequences of hPR3 expression on the composition of BM progenitor compartments was examined. Although there was no difference in the percentage of LSK cells, including HSPCs, there was a significant increase in the population of cells enriched for myeloid progenitors, defined as LK (Lin⁻Sca1⁻cKit⁺) cells (Fig. 2A). Because this LK population contains CMPs, GRMPs, and MEPs, differential Ab staining and flow cytometry analysis were used to determine whether a particular subset was preferentially altered in hPR3Tg. Despite the increase in LK cells, no difference in the proportion of CMPs, GRMPs, and MEPs was observed in hPR3Tg compared with WT controls (Fig. 2B). Lastly, the proportion and total number of differentiated neutrophils, monocytes, and B cells were comparable between WT and hPR3Tg with regard to percentiles and numbers (Fig. 2C). To assess progenitor proliferation, the LSK population was isolated and cultivated in the presence of murine stem cell factor, hFMS-like tyrosine kinase 3 ligand, human IL-6, and human IL-11 for 4 d. There was no difference in the rate of proliferation between LSK cells isolated from WT and hPR3Tg (Fig. 2D), and they exhibited a similar level of apoptosis after 4 d in culture (Fig. 2E). Under certain conditions, PR3 has been shown to be a proapoptotic factor in mature neutrophils (8). To determine whether hPR3 affected the survival of differentiated BM PMNs in the steady-state, BM cells were placed in culture, and the spontaneous apoptosis of PMNs was evaluated by flow cytometry. No difference was observed in the percentage of neutrophils undergoing spontaneous apoptosis in hPR3Tg compared with WT controls (Fig. 2F). Similarly, no difference was observed in the death of neutrophils of both genotypes following granule membrane permeabilization with LLME at any time points examined (Fig. 2G).

The composition of the blood was also examined, and expression of the hPR3 transgene did not affect the total number of WBCs, RBCs, or platelets per cubic millimeter (Fig. 3A–C). Of note, although there was no difference in the lymphocyte, PMN, eosinophil, or basophil populations, there was a small, but significant, increase in the proportion and total number of monocytes per cubic millimeter within the blood (Fig. 3D). To test whether the hPR3 transgene influenced neutrophil homeostasis and turnover, mice were pulsed with a single injection of EdU, and the frequency of Ly6G⁺EdU⁺ cells was assessed in the BM and peripheral blood. Flow cytometry analysis found that there was no difference in the frequency of EdU⁺ neutrophils within the BM or circulation at any time point examined between WT and hPR3Tg (Fig. 3E, 3F, respectively).

A model of zymosan-induced peritonitis was used to study the effects of hPR3 on immune cell recruitment during a sterile inflammatory response, as well as the resolution of this inflammation. Injection of zymosan resulted in a rapid influx of cells into the peritoneal cavity of mice after 4 h, and this cellular content gradually increased over the next 72 h (Fig. 4A). Overall, there was a small, but statistically insignificant, increase in total cells within the peritoneal cavity of hPR3Tg compared with WT controls from 24 h onward. Differential Ab staining and flow cytometry analysis showed that injection of zymosan elicited a rapid recruitment of neutrophils (Gr1⁺F4/80⁻), which peaked at 4 h in WT and steadily declined over the next 72 h. In contrast, although hPR3Tg displayed the same degree of neutrophil infiltration at 4 h, neutrophil numbers in the peritoneal cavity were maintained up to 48 h, until a decrease was observed at 72 h (Fig. 4B). These data suggest that, although hPR3 did not alter the initial recruitment of neutrophils, it may hamper the resolution of inflammation and facilitate prolonged neutrophil accumulation. The recruitment of macrophages and B and T lymphocytes gradually increased over 72 h after the induction of peritonitis, and there was no significant difference in the number of any of these cells types in hPR3Tg compared with WT controls at any time point (Fig. 4B).

The total protein concentration in the peritoneal exudates was measured as a surrogate indicator of inflammation. There was no significant difference in total proteins in the lavage fluid at 4 h in WT compared with hPR3Tg. At 24 h, total proteins in both groups were reduced relative to 4 h, yet hPR3Tg displayed significantly higher protein concentrations within the lavage fluid compared with the WT controls (Fig. 4C). The levels of the proinflammatory cytokine IL-6 and the chemokine MCP-1 (also known as CCL2) were assessed at 4 and 24 h after the induction of peritonitis. Although the amounts of IL-6 and MCP-1 were higher 4 h after injection of zymosan compared with 24 h, there was no significant difference between WT and hPR3Tg (Fig. 4D). AnxA1 is an anti-inflammatory protein that is known to decrease PMN recruitment, is cleaved by PR3, and is normally secreted during inflammation and apoptosis. AnxA1 in the peritoneal exudate was detected by Western blot analysis at all time points examined following zymosan injection in WT and hPR3Tg (Fig. 4E). Notably, there was a significant reduction in full-length AnxA1 in hPR3Tg compared with the WT controls 24 h after the injection of zymosan. No significant difference was observed at any other time point investigated. All results were normalized to total protein content within the lavage fluid. These results suggest that AnxA1 cleavage by PR3 may contribute to the PMN accumulation in the peritoneal cavity at 24 and 48 h.

During inflammation initiated by zymosan, hPR3Tg display increased neutrophil accumulation. To investigate the effects of hPR3 on neutrophil function and survival, in vitro experiments were performed using neutrophils isolated from the peritoneal lavage 4 h after the administration of zymosan. The ability of these neutrophils to produce reactive oxygen species (ROS) was assessed by luminol-amplified chemiluminescence. The results showed that neutrophils from hPR3Tg were able to produce comparable levels of ROS basally and following treatment with PMA or opsonized zymosan (Fig. 5A). To assess intrinsic survival of zymosan-recruited peritoneal neutrophils, neutrophils were purified using a Ficoll gradient to exclude macrophages, and isolated neutrophils were cultured at 37°C, with or without TNF-α or G-CSF. As expected, TNF-α was proapoptotic and produced more cell death compared with untreated controls, whereas G-CSF was prosurvival and prevented apoptosis (Fig. 5B). Importantly, compared with WT controls, hPR3Tg neutrophils were more viable after 16 h in culture, regardless of whether they underwent spontaneous apoptosis or were treated with TNF-α or G-CSF. Accordingly, under all conditions, there was a significant decrease in the percentages of AnnexinV⁺ hPR3Tg neutrophils compared with control neutrophils (Fig. 5B). Taken together, these data suggest that hPR3 enhances the survival of neutrophils during inflammation; this may explain, in part, the increased accumulation of neutrophils observed during peritonitis.

To determine whether the accumulation of neutrophils during inflammation in hPR3Tg was not only the result of enhanced survival, but could also be due to impaired clearance by macrophages, in vivo experiments were performed involving the injection of apoptotic RBL cells into the peritoneal cavity of mice. In these experiments, RBL cells stably transfected with hPR3 were used because they possess neutrophil-like characteristics but do not express other serine proteases, meaning that PR3-mediated processes can be studied exclusively (10, 14). TAMRA-labeled apoptotic RBL or RBL/PR3 cells were injected into the peritoneal cavity of WT and hPR3Tg, and macrophage phagocytosis was quantified by flow cytometry (Fig. 5C). There was no significant difference in the ability of hPR3Tg and WT resident macrophages to phagocytose apoptotic cells at 30 min or 2 h.

The effect of hPR3 on survival during CLP, an acute inflammatory model of sepsis, was investigated. This model was chosen because it represents an infectious model of inflammation and, therefore, involves different physiopathological mechanisms compared with peritonitis induced by zymosan. Severe CLP was induced according to a previously established protocol (26), and the survival rate of WT was 50% after 7 d. In contrast, no hPR3Tg survived this model of sepsis, with all animals dying within 5 d of surgery (Fig. 6A). The health and well-being of the mice were assessed using a scoring system at 24 and 48 h postsurgery. hPR3Tg scored significantly higher, indicating a more severe physiological response to CLP, associated with increased pain and distress (Fig. 6B). In addition to the increased mortality observed in hPR3Tg, these animals displayed a significant increase in the total number of cells recruited to the peritoneal cavity 6 h after the induction of sepsis; this was due to a larger infiltrate of neutrophils (Fig. 6C, 6D). No difference in the recruitment of monocytes, macrophages, or lymphocytes was observed between hPR3Tg and WT controls. To determine whether the increased mortality in hPR3Tg resulted from an increased bacterial load, the number of bacteria within the peritoneum was assessed 18 h after the induction of CLP. There was no difference in the total number of CFU in the peritoneal cavity of hPR3Tg compared with WT controls (Fig. 6E). To further examine bacterial clearance, mice were inoculated intranasally with 8 × 10⁶ CFU P. aeruginosa; no difference in the number of bacteria in the lungs was observed between WT and hPR3Tg after 24 h, confirming that the expression of hPR3 did not result in any alteration of bacterial killing during an infection (Supplemental Fig. 1). The level of AnxA1 protein at 6 and 20 h after the induction of CLP was examined by Western blot, and results were normalized to the total protein content within the lavage fluid. AnxA1 was detected within the lavage fluid at both time points examined; although no difference was observed at 6 h, there was a significant decrease in AnxA1 protein levels at 20 h in hPR3Tg compared with WT controls (Fig. 6F, 6G). Total protein within the lavage fluid was decreased in hPR3Tg at 6 h; however, at 20 h after the induction of CLP, hPR3Tg had significantly more protein within the peritoneal fluid (Fig. 6H). Although PR3 can cleave various inflammatory cytokines and chemokines, we did not observe any difference in the absolute levels of IL-1β, TNF-α, and MCP-1 proteins in the lavage fluid 20 h after CLP (Supplemental Fig. 2).

This study centers on a novel transgenic mouse model expressing the human ortholog of PR3 that possesses unique structural and biochemical features that may amplify its pathogenic activity compared with murine PR3, including a hydrophobic patch that facilitates membrane anchorage (15, 22, 37). Furthermore, although many of the substrates cleaved by murine PR3 and hPR3 are well conserved, some substrates are unique to either human or mouse (21, 38). These fundamental differences between human and mouse PR3 highlight the need for a mouse model expressing hPR3 to fully understand the molecular pathways that underlie the proinflammatory activities of this protein in human pathologies (23). To generate hPR3Tg, the transgene was placed under control of the rEF1 promoter; although this results in expression in almost all cell types, PR3 is initially expressed as a proform (39–41), and only myeloid cells are able to process the proform into the mature and enzymatically active serine protease (39). In line with this, mature and enzymatically active PR3 was detected within neutrophils elicited by zymosan.

Overall, hPR3Tg had a defect in the resolution of inflammation, and in two models of peritonitis, neutrophils, which contain significantly more PR3 protein compared with other myeloid cells (42), were specifically affected. Neutrophils are typically the first immune cell mobilized to the site of injury or infection and are essential for mounting an appropriate immune response (43), because they possess various mechanisms that are capable of killing pathogens, including multiple proteases and an ability to generate ROS. However, these same mechanisms can also cause extensive tissue damage if the influx of neutrophils or release of these mediators is not controlled. Therefore, the key to mounting an effective immune response while avoiding unnecessary tissue damage is by controlling neutrophil activation and survival (44). During zymosan-induced peritonitis, hPR3Tg exhibited sustained neutrophil accumulation and a clear delay in the resolution phase of inflammation. Interestingly, during severe sepsis induced by CLP, a huge increase in neutrophils was observed as early as 6 h, suggesting that, under different inflammatory conditions, hPR3 may also modulate neutrophil recruitment. The different pathophysiological mechanisms involved in zymosan-induced peritonitis and CLP likely account for this, including the presence of live bacteria, and the fact that CLP represents a significantly more severe inflammatory response, which might affect the early steps of inflammation, including the interactions between neutrophils and the endothelium. In fact, during CLP-induced sepsis, all hPR3Tg died within 5 d compared with only 50% of WT controls. Given that there was no increase in bacterial load, we suggest that the massive neutrophil burden in the hPR3Tg contributed to the increased mortality. When uncontrolled, neutrophils and their intrinsic factors, including proteases found within granules, can cause extensive tissue damage (5). Furthermore, because PR3 is known to cleave and activate various proteins within the neutrophil and surrounding microenvironments, it seems likely that increased PR3 in the neutrophils of the transgenic mice could amplify their toxic potential.

One mechanism that is likely to contribute to the proinflammatory phenotype of hPR3Tg is the cleavage of the powerful anti-inflammatory and proresolving protein AnxA1. AnxA1 is a glucocorticoid-regulated protein released into the inflammatory microenvironment that can act as an endogenous downregulator of the inflammatory process by limiting neutrophil recruitment and activation or inhibiting the production of proinflammatory mediators (45). This protein can also induce neutrophil apoptosis, promote the clearance of apoptotic neutrophils by macrophages, and may contribute to the reprogramming of macrophages toward a proresolving phenotype (46–49). Mice lacking AnxA1 or its receptor FPR2/3 exhibit exacerbated inflammation during immune challenge that is characterized by increased neutrophil accumulation (50, 51). Furthermore, following release of AnxA1 into the inflammatory microenvironment, numerous studies demonstrated that this protein is cleaved (52–54); in stark contrast to the full-length form, the 33-kDa cleavage product may possess proinflammatory properties (55). Administration of serine protease inhibitors during LPS challenge in mice increased the levels of intact AnxA1, which, in turn, accelerated the resolution of inflammation (56). Notably, PR3 has been shown to specifically cleave AnxA1, and an AnxiA1 mutant resistant to PR3 cleavage controlled inflammation and accelerated its resolution in vitro, in vivo, and during models of chronic inflammation, such as arthritis (13, 57). Our present findings showing that hPR3 expression facilitated the cleavage of AnxA1 during inflammation in vivo is in keeping with these previous studies. Given the powerful anti-inflammatory and proresolving properties of this protein, we believe that loss of AnxA1 in hPR3Tg contributes to the increased neutrophil accumulation and inflammation observed during sterile and infectious peritonitis. Although AnxA1 is cleaved during inflammation in our transgenic mice, we do not exclude the possibility that hPR3 can also cleave another unidentified protein to perpetuate the neutrophil-driven inflammation. We also observed that inflammatory neutrophils expressing hPR3 had increased survival in vitro during spontaneous apoptosis and when treated with TNF-α or G-CSF. The increased survival of hPR3Tg neutrophils likely explains why neutrophil numbers remained constant for the first 48 h after the induction of inflammation. Because AnxA1 can promote neutrophil apoptosis (49), cleavage of this protein may also be responsible for the increased neutrophil survival observed in vitro.

In contrast to inflammatory neutrophils collected from peritoneal lavage fluid, neutrophils contained in the BM did not show a delayed spontaneous apoptosis in vitro. Likewise, cell death induction by permeabilization of neutrophil granules with LLME, which releases granule contents into the cytoplasm (58), was unaltered in hPR3Tg. These results suggest that the increased level of PR3 in neutrophil granules was not sufficient to offset the cytoplasmic antiprotease shield, such as that provided by Serpinb1. Indeed, Serpinb1 inhibits PR3, CG, and NE and protects neutrophils from spontaneous and LLME-induced cell death in vivo and in vitro (28, 59), although it should be noted that evidence suggests that, in both death pathways, CG is actually the major protease required for inducing neutrophil death in the absence of Serpinb1 (28).

PR3, also known as myeloblastin, is expressed in human CD34⁺ hematopoietic progenitor cells and promotes granulocyte proliferation and differentiation (34). This protein is downregulated during the differentiation of the leukemic cell line HL-60 and human peripheral blood monocytes, and inhibition of PR3 in HL-60 cells resulted in the inhibition of proliferation and, as a consequence, cellular differentiation (34). In the current study, we demonstrated that the proportion of LSK cells within the BM of hPR3Tg is similar to WT controls, and their levels of proliferation and apoptosis were identical, at least in vitro. However, there was a significant increase in committed myeloid progenitor populations, and the potential pathophysiological relevance is unknown. This increase did not lead to any change in mature leukocyte numbers in the BM, and although most of the peripheral blood leukocyte populations were unaffected, there was a small, but significant, increase in the proportion and absolute number of monocytes. The homeostasis and turnover of neutrophils within the BM and circulation were also examined and found to be unaffected by the expression of hPR3.

One limitation of our mouse model is that we are unable to determine whether the proinflammatory phenotype observed is simply due to increased protein and enzymatic activity or is a result of a function specific to hPR3 due to its hydrophobic patch, for instance, and not the mouse homolog (37). In the future, generation of an hPR3-knockin mouse model would allow us to explore this question further. However, our novel transgenic mouse may represent a useful tool in developing an animal model for granulomatosis with polyangiitis, the systemic autoimmune disease involving ANCA directed against PR3 (60, 61). To date, attempts to produce an animal model of anti-PR3 vasculitis have been unsuccessful. For example, although immunization of mice with chimeric human/mouse PR3 led to the generation of autoantibodies, no evidence of vascular lesions in the kidney or lung was observed (62). More recently, two methods designed to induce vasculitis in mice expressing both hPR3 and CD177, including passive transfer of anti-PR3 Abs and BM transplantation from WT immunized with hPR3, failed to induce glomerulonephritis (63). Failure to induce vasculitis in these mice was attributed to the fact that mouse dipeptidyl peptidase I was unable to process human pro-PR3 into the mature form; therefore, the signaling complex among PR3, CD177, and CD11b, which promotes neutrophil activation by anti-PR3 Abs, was unable to form. We do not encounter the same issue in our model, because mature hPR3 protein and increased enzymatic activity in neutrophils of these mice were detected. Furthermore, although some studies suggest that CD177 is required for the membrane expression of hPR3 (64), we demonstrated previously that it was independent of CD177 in RBL cells (14, 22, 23), and although basal expression of membrane PR3 is related to CD177, apoptosis-induced PR3 expression is independent of CD177 (14).

In conclusion, expression of hPR3 perpetuates inflammation, which is characterized by neutrophil accumulation and cleavage of the powerful anti-inflammatory protein AnxA1. Our newly described transgenic mouse model could be used in the future to decipher the molecular mechanisms that contribute to PR3-dependent diseases and inflammation, including vasculitis associated with ANCAs, and may ultimately be used to develop new therapeutic strategies.

We acknowledge Muriel Andrieu and Karine Labroquère (Cochin Cytometry and Immunobiology Facility) and the staff of the Animal Care Facilities (Cochin Institute), Dr. Thomas Hellmark for providing anti-PR3 Ab, Dr. Adam Lesner (University of Gdansk) for the generous gift of the PR3 ABZ substrate, Dr. Anne Lombès for assistance in establishing the PR3-activity assay, Roberto Adelfio for technical help, and Aurélie Durand and Stephanie Bessoles for aid in quantifying leukocytes within the blood.

The authors have no financial conflicts of interest.