In the present study, we provide evidence that procaspase-3 is a novel target of proteinase 3 (PR3) but not of human neutrophil elastase (HNE). Human mast cell clone 1 (HMC1) and rat basophilic leukemia (RBL) mast cell lines were transfected with PR3 or the inactive mutated PR3 (PR3S203A) or HNE cDNA. In both RBL/PR3 and HMC1/PR3, a constitutive activity of caspase-3 was measured with DEVD substrate, due to the direct processing of procaspase-3 by PR3. No caspase-3 activation was observed in cells transfected with the inactive PR3 mutant or HNE. Despite the high caspase-3 activity in RBL/PR3, no apoptosis was detected as demonstrated by an absence of 1) phosphatidylserine externalization, 2) mitochondria cytochrome c release, 3) upstream caspase-8 or caspase-9 activation, or 4) DNA fragmentation. In vitro, purified PR3 cleaved procaspase-3 into an active 22-kDa fragment. In neutrophils, the 22-kDa caspase-3 activation fragment was present only in resting neutrophils but was absent after apoptosis. The 22 kDa fragment was specific of myeloid cells because it was absent from resting lymphocytes. This 22-kDa fragment was not present when neutrophils were treated with pefabloc, an inhibitor of serine proteinase. Like in HMC1/PR3, the 22-kDa caspase-3 fragment was restricted to the plasma membrane compartment. Double immunofluorescence labeling after streptolysin-O permeabilization further showed that PR3 and procaspase-3 could colocalize in an extragranular compartment. In conclusion, our results strongly suggest that compartmentalized PR3-induced caspase-3 activation might play specific functions in neutrophil survival.
Neutrophil-derived proteinase 3 (PR3)4 belongs to the family of microbicidal serine proteinase homologues along with human neutrophil elastase (HNE), cathepsin G, and azurocidin (1, 2). However, the role of PR3 is not confined to microbicidal mechanisms or proinflammatory extracellular proteolysis (3). In neutrophils, in contrast to HNE, the subcellular localization of PR3 is not restricted to azurophilic granules (4) because it was possible to detect PR3 in the membrane of secretory vesicles, which subsequently expressed at the plasma membrane (5). Another atypical feature of PR3 is its involvement in myeloid differentiation. This feature explains why PR3 has also been called myeloblastin (6). Moreover, PR3 has a pathophysiologic importance in nephrology because it is the target of antineutrophil cytoplasmic autoantibodies in Wegener’s granulomatosis, an inflammatory disorder characterized by chronic inflammation of the respiratory tract, vasculitis, and glomerulonephritis (7, 8). Wegener’s granulomatosis is strongly associated with antineutrophil cytoplasmic Abs directed against PR3 (80–90% of cases) (9).
To gain insight into the functions that PR3 could mediate in phagocytic cells, we constructed a recombinant cellular model using human mast cell clone 1 (HMC1) and rat basophilic leukemia (RBL) mast cell line that were stably transfected with either PR3 or HNE (10, 11). We recently demonstrated that PR3-transfected cells proliferate at a higher rate compared with HNE-transfected cells (12). Interestingly, the cells transfected with an inactive mutant of PR3 (PR3S203A) did not present this proliferative activity. We showed that p21waf1, a cyclin-dependent kinase inhibitor (13), was absent in the PR3-transfected cells despite the presence of its mRNA. In vitro experiments using purified proteins demonstrated that PR3 could directly cleave p21. It has been proposed that p21 is a part of the cell antiapoptotic machinery. This result is demonstrated through its binding and the subsequent inhibition of different molecular partners, including proteins with known proapoptotic activity, such as caspase-3 and apoptosis-stimulating kinase 1 (14). Because several reports have recently stressed interconnections between proliferation, differentiation, and apoptosis (15), we have investigated whether the increased proliferation observed in PR3-transfected cells could be linked to deregulated apoptosis and especially to caspase activation.
In the present report, we provide evidence that in our model of transfected mast cell lines RBL and HMC1, there was a constitutive activity of caspase-3 in PR3-transfected cells with the generation of a specific 22-kDa fragment. This caspase-3 activation did not result in apoptosis. This 22-kDa fragment of caspase-3 could be detected in resting but disappeared in apoptotic neutrophils. It was present only in the membrane-enriched fraction both in HMC1 transfected with PR3 and in neutrophils. Our data showed that PR3-induced caspase-3 cleavage was compartmentalized in neutrophils, thus preventing the induction of apoptosis. So, it can be suggested that this PR3-induced caspase-3 activation might play a role in neutrophil survival.
Materials and Methods
Cell culture and transfection
The RBL mast cell line (a gift from Dr. M. Dy, Centre National de la Recherche Scientifique, Unité Mixte de Recherche 8603, Paris, France) and the HMC1 line (a gift from Dr. J. Butterfield, Mayo Clinic, Rochester, MN) were cultured in DMEM supplemented with 10% FCS. Cells were transfected with the plasmids pCDNA alone, or pCDNA/PR3, pCDNA/PR3S203A, and pCDNA/HNE, selected in the presence of 1 μg/ml zeocin (Invitrogen Life Technologies) and cloned as previously described (12). To induce apoptosis, HMC1 and RBL cells (10 × 106) were cultured in the presence of HA-14, a pharmacologic inhibitor of Bcl-2 (Calbiochem), or with the chemotherapy drug etoposide (Dakota Pharma) at the indicated time and concentrations. Neutrophils were isolated as previously described (5). Briefly, EDTA anticoagulated blood from voluntary donors was centrifuged for 20 min at 150 × g to eliminate platelets from plasma. Plasma and blood were mixed, put on polymorphoprep (Axis-Shield) and centrifuged 45 min at 700 × g. Neutrophils were washed in PBS before erythrocyte lysis in 0.2% NaCl. After neutrophil numeration, cells were immediately lysed for the control condition or plated in a six-multiwell plate at 2 × 106 cells/ml in RPMI 1640 10% FCS, and incubated at 37°C, 5% CO2. Apoptosis was induced in neutrophils by treatment with gliotoxin (0.1 μg/ml; Sigma-Aldrich) (16) overnight at 37°C or by a 37°C overnight incubation (17). Cells were also treated either with the inhibitor of serine proteinase pefabloc (100 μM, 3 h; Roche) or with the protein synthesis inhibitor cycloheximide (Sigma-Aldrich) at 20 μg/ml, 4 h at 37°C. Lymphocytes were obtained from monocyte/lymphocyte layer. Monocytes were induced to adhere at 2 × 106/ml in RPMI 1640 10% SVF 2 h at 37°C, 5% CO2, and nonadherent lymphocytes were collected from culture medium.
Flow cytometry analysis of PR3 and HNE intracellular expression
Intracellular expression of serine proteinase (PR3 or HNE) in stably transfected HMC1 and RBL cells was performed as previously described (18). Briefly, 5 × 105 cells were incubated in cell permeabilizing solution (BD Biosciences), washed in PBS (0.5% BSA, 0.1% NaN3) and then saturated with 1 mg/ml heat-aggregated goat IgG, 5% FCS diluted in PBS (1% BSA, 0.1% NaN3). Cells were first incubated for 30 min with either mouse monoclonal anti-PR3 CLB 12.8 (CLB) or anti-HNE (clone 39A; Biogenesis) Ab followed by FITC-conjugated F(ab′)2 goat anti-mouse IgG (Immunotech). A cell fix solution was added and cells were analyzed for fluorescence (FL1) on a FACScan flow cytometer (BD Immunocytometry Systems).
Colorimetric or fluorometric measurement of caspase-3, caspase-8, and caspase-9 activities
Cells (10 × 106) treated or not with apoptosis inducers at indicative concentration and time were centrifuged for 10 min at 250 × g. Pellets were lysed at 100 × 106/ml in lysis buffer (50 mM HEPES, pH 7.4, 0.1% CHAPS, 5 mM DTT, 0.1 mM EDTA) for 15 min at 4°C, according to the manufacturer’s instructions (Biomol). After centrifugation at 10,000 × g for 15 min, the bicinchoninic acid assay (Pierce) was used to determine protein concentrations in supernatants. Samples (10 μl) adjusted at 3 mg/ml protein and 80 μl of assay buffer (50 mM HEPES, pH 7.4, 100 mM NaCl, 0.1% CHAPS, 10 mM DTT, 1 mM EDTA, 10% glycerol) were distributed into a 96-well plate and 10 μl of specific chromogenic caspase substrates Ac-DEVD-pNA (Biomol), or Ac-IETD-pNA or Ac-LEDH-pNA (Chemicon International) were added at a final concentration of 200 μM, to measure caspase-3, caspase-8, and caspase-9 activity, respectively. The OD value was measured at 405 nm following a 4-h incubation period at 37°C with a microplate reader (Dynatech Laboratories). To ascertain specificity, the selective inhibitor Ac-DEVD-CHO or Ac-IETD-CHO or Ac-LEDH-CHO was added to inhibit caspase-3, caspase-8, and caspase-9 activity, respectively. For the fluorometric measurement of caspase-3 activity on whole cell lysate or on cell fractions, the same protocol was used with the Ac-DEVD-AMC caspase-3 specific fluorometric substrate. Its emission was detectable at 460 nm after a 360 nm excitation on a fluorometer microplate reader (Berthold).
Cell cycle analysis by flow cytometry
HMC1 cells were plated into a six-well plate (5 × 105 cells/well) in DMEM 10% FCS in the presence or the absence of 100 μM caspase-3 inhibitor DEVD-Fmk (Bachem) and allowed to seed for 24 h in culture. 5-Bromo-2′-deoxyuridine (BrdU, 10 μM; BD Pharmingen) was incorporated during 1 h at 37°C. After successive resuspension in PBS 3% FCS and PBS 1% BSA, cells were fixed in frozen 70% ethanol for 15 h at 4°C. Denaturation was performed with 2 N HCl 0.5% Triton X-100 during 24 h at 4°C and neutralization with 0.1 M Na2B4O7 during 10 min at 4°C. Cells were treated with DNase (300 μg/ml), 1 h at 37°C, labeled with monoclonal FITC-conjugated anti-BrdU (BD Pharmingen) and propidium iodide (2.5 μg/ml) for 30 min at room temperature in the dark. BrdU incorporation was analyzed by flow cytometry (FL1) and the percentages of cells in the G1, S, and G2/M phases of the cell cycle were determined by CellQuest (FL2).
Analysis of phosphatidylserine (PS) externalization by annexin V labeling
During the early phase of apoptosis, the membrane PS is translocated from the inner to the outer leaflet of the plasma membrane, thereby exposing PS to the external cellular environment where it can bind to annexin V in the presence of calcium. 7-Aminoactinomycin D (7-AAD) staining was used to label necrotic cells. RBL (5 × 105/ml) were cultured with or without etoposide (10 μM, 15 h) and neutrophils (2 × 106/ml) were labeled immediately or incubated overnight at 37°C with or without gliotoxin (0.1 μg/ml) to induce apoptosis or with pefabloc (100 μM). Then cells were labeled with PE-conjugated annexin V and 7-AAD as previously described (18). Annexin V-PE (FL2) and 7-AAD (FL3) labeling were analyzed immediately by flow cytometry.
Both RBL and HMC1 cells were resuspended in ice cold fractionation buffer, pH 7.4 (25 mM sucrose, 20 mM HEPES, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 400 μM leupeptine, 4 mM PMSF, 400 μM pepstatine, 1 mM orthovanadate) at a concentration of 100 × 106/ml. Cells were disrupted using a potter (20 strokes at 4°C). Intact cells and nuclei were pelleted by centrifugation (750 × g, 10 min, 4°C) and the supernatants were centrifuged at 10,000 × g for 25 min at 4°C. The resulting pellet containing granules was washed in the fractionation buffer. The supernatant was centrifuged at 100,000 × g for 1 h at 4°C to obtain a pellet containing plasma membranes, which was subsequently washed in the fractionation buffer, and the supernatant, which represented the cytosol. Neutrophils were fractioned using nitrogen cavitation as previously described (5) to obtain the cytosol, the granule, and the membrane fractions.
Western blot analysis
For the analysis of caspase-3 cleavage fragments, HMC1, neutrophils, lymphocytes, and cell fractions were lysed (100 × 106 cells/ml, 15 min, at 4°C) in RIPA or in 1% Nonidet P-40 buffer as indicated, supplemented with antiproteases (0.1% SDS, 0.5% desoxycholate, 1% Nonidet P-40, 150 mM NaCl, 50 mM Tris, pH 8, 1 mM DTT, 1 mM orthovanadate, 2 mM EDTA, 1 mM EGTA, 400 μM pepstatine, 4 mM PMSF, 400 μM leupeptine). Protein concentration was determined using the bicinchoninic acid assay and Western blot was performed as previously described (5). The primary Ab was a rabbit polyclonal anti-caspase-3 (Neomarkers). For cytochrome c detection in RBL subcellular fractions, the primary Ab was a mouse monoclonal anti-cytochrome c (BD Pharmingen). The secondary Abs were conjugated with HRP, and a chemiluminescence detection kit (Pierce) was used.
Analysis of DNA fragmentation
DNA ladder analysis: RBL cells (1 × 106/ml) were washed in PBS, placed 30 min at −20°C, and resuspended in 10 mM Tris-HCl, pH 8, 400 mM NaCl, 2 mM EDTA, 20 mg/ml proteinase K, and 10% SDS before being incubated overnight at 37°C. Pellets were mixed with saturated NaCl and 1 M MgCl2, precipitated in ethanol, and incubated 1 h at −80°C. Dry pellets were incubated overnight at 4°C in 10 mM Tris-HCl, pH 8, 1 mM EDTA. RNase-DNase-free (1.5 μg/ml) was added for 3 h at 37°C and 5 min at 70°C. DNA fragmentation was thus analyzed on 2% agarose gel containing ethidium bromide and revealed by UV. To determine cytoplasmic histone-associated DNA fragments (mono- and oligonucleosomes) by ELISA, RBL cells (1 × 106/ml) were solubilized in lysis buffer for 30 min at room temperature, according to the manufacturer’s instructions (Roche). Supernatants, which contained cytoplasmic fractions, were added on the streptavidin-precoated microplate. Pellets, which represent cell nuclei containing high m.w. and unfragmented DNA, were removed. Then, a mixture of biotin-conjugated antihistone Abs that bind to histones (H1, H2A, H2B, H3, and H4), and peroxidase-conjugated anti-DNA Ab, which reacts with the DNA component of the nucleosomes, were added for 2 h at room temperature. After washing, the substrate solution was added to quantitatively determine the amount of cytoplasmic histone-associated DNA fragments using the OD at 405 nm after a 10-min incubation period.
In vitro procaspase-3 cleavage by PR3 or HNE
PR3 and HNE (generous gifts from Dr. D. Pidard, Institut Pasteur, Paris, France) were purified from human neutrophils as previously described (19, 20). Commercially available purified procaspase-3 (10 μg/ml; Biomol) was incubated with purified PR3 (12 nM) during 30 min at 37°C. Caspase-9 (0.1 U/μl; Chemicon International) was used as a positive control of procaspase-3 activation. Caspase-3 activity of the PR3-cleaved fragment was assessed using the colorimetric DEVD-pNA substrate as described. The cleavage products were analyzed by Western blot using the rabbit polyclonal anti-caspase-3 (Neomarkers). To test the in vitro cleavage of PR3 and HNE on procaspase-3 from control HMC1, cells were lysed by sonication at 4°C (100 mM HEPES, 250 mM NaCl, pH 7.4) and incubated for 1 h at 37°C in the presence of PR3 or HNE at concentrations ranging from 50 to 300 nM.
Measurement of serine protease activity
The PR3 or HNE serine protease activity was measured using the hydrolysis of the tripeptide thiobenzyl ester Boc-Ala-Pro-Nva-SBzl (Sigma-Aldrich) as previously described (19). Briefly, 20 μl of sample was diluted in reaction buffer (100 mM HEPES, 500 mM NaCl, pH 7.5) containing Boc-Ala-Pro-Nva-SBzl (325 μM) in the presence of 800 μM 5′,5′-dithio-bis(2-nitrobenzoic acid) (DTNB; Sigma-Aldrich). Absorbance was measured at 405 nm using a microplate reader (Dynatech Laboratories).
Immunofluorescence labeling of PR3 and caspase-3 by confocal microscopy
After isolation, neutrophils were induced to adhere on coverslips precoated with poly-l-lysine (Sigma-Aldrich) for 45 min at room temperature. Cells were fixed in PBS (3% paraformaldehyde) for 15 min. Streptolysin-O (SLO, 0.05 U/ml; Sigma-Aldrich) diluted in permeabilization buffer (50 mM HEPES, 100 mM KCl, 20 mM NaCl, 1 mM EGTA, 0.1% dextrose) was added 5 min at 4°C and after several washing in PBS, cells were incubated in permeabilization buffer 15 min at 37°C to allow plasma membrane pore formation (21). For Triton X-100 (Sigma-Aldrich) and methanol permeabilization, cells were incubated in 0.5% Triton X-100 or ice-cold methanol for 15 min and 20 min, respectively. After saturation in PBS, 1% BSA for 20 min. Abs were successively added 30 min at 37°C and for caspase-3 labeling: polyclonal anti-caspase-3 (40 μg/ml; Neomarkers), followed by biotinylated anti-rabbit IgG (10 μg/ml; Dako), and Alexa Fluor 555-conjugated streptavidin (10 μg/ml; Molecular Probes); for PR3: monoclonal anti-PR3 CLB12--8 (20 μg/ml; Sanquin) followed by FITC-conjugated anti-mouse IgG (7 μg/ml; Jackson ImmunoResearch Laboratories). For colocalization experiments, labeling was performed by adding successively anti-caspase-3, anti-PR3, biotinylated anti-rabbit IgG, FITC-conjugated anti-mouse IgG, and Alexa Fluor 555-conjugated streptavidin. Slides were mounted using Vectashield mounting medium (Vector Laboratories) and analyzed by fluorescent microscopy or by confocal microscopy, with a Zeiss LSM-510 confocal scanning laser microscope equipped with a 25 mW argon laser and a 1 mW helium-neon laser, using a Plan-Apochromat ×63 objective (numerical aperture 1.40, oil immersion). Green fluorescence was observed with a 505–550 nm band-pass emission filter under 488 nm laser illumination. Red fluorescence was observed with a 560 long-pass emission filter under 543 nm laser illumination. Pinhole diameters were set to get 0.6 μm thick optical slices. Stacks of images were collected every 0.3 μm along the z-axis. Images corresponding to FITC and Alexa fluor 555 were obtained using the multitrack mode. Percentages of colocalization were determined using LSM software on 20 distinct cells double labeled in five independent experiments.
Statistical analysis was performed using the Statview software package. Comparisons were made by ANOVA. Data were expressed as mean ± SEM. Differences were considered significant for values of p < 0.05 (*) or p < 0.01 (**).
Increased caspase-3 activity in stably transfected HMC1/PR3 and RBL/PR3
HMC1 and RBL were transfected with either PR3 or PR3S203A, an inactive mutant of PR3 or HNE, a PR3 homologue, as previously described (12, 18). The intracellular expression of PR3, PR3S203A, and HNE in HMC1-transfected cells was analyzed by flow cytometry after cell permeabilization (Fig. 1,A). This analysis showed that HMC1 transfected with pcDNA plasmid alone (CT) was not labeled by mouse monoclonal anti-PR3 or mouse monoclonal anti-HNE, whereas >93% of HMC1/PR3 and the PR3 inactive mutant HMC1/PR3S203A expressed the recombinant protein, respectively, PR3 and PR3S203A. Analysis of HNE expression using anti-HNE allowed us to detect ∼96% of HMC1 cells labeled with anti-HNE. Similar results were obtained in RBL transfectants with a high expression of PR3 measured in RBL/PR3 and RBL/PR3S203A, and a high expression of HNE in RBL/HNE, as compared with control RBL (CT) transfected with pcDNA plasmid alone (Fig. 1 B).
Caspase-3 activity was measured by spectrophotometry using the synthetic substrate Ac-DEVD-pNA. As shown in Fig. 1,C, a strong caspase-3 activity was detected in HMC1/PR3. Caspase-3 activity was totally dependent of PR3 serine protease activity because inactive mutated PR3-transfected cells (HMC1/PR3S203A) had the same low caspase-3 activity as control cells. Interestingly, in HMC1/PR3, caspase-3 activity is about six times higher than in control cells. The specificity of this assay was determined using a specific caspase-3 DEVD-CHO inhibitor, which induced total inhibition of caspase-3 activity. Similar results were obtained in RBL transfectants with a very high caspase-3 activity (about a 7-fold increase) as compared with control RBL, RBL/PR3S203A, and RBL/HNE (Fig. 1 D). We therefore concluded that constitutive PR3-induced caspase-3 activity was dependent on PR3 enzymatic serine protease activity.
Increased proliferation in PR3-transfected cells is not related to constitutive caspase-3 activity
We have previously described (12) that, among these transfectants, only HMC1/PR3 and RBL/PR3 showed increased cell proliferation as compared with controls (CT: pcDNA alone; PR3S203A: inactive PR3 and HNE). The proliferative effect of PR3 is illustrated in Fig. 2,A, which depicts the proportion of cells in different phases of the cell cycle determined by flow cytometry analysis. This cell cycle analysis showed a significant increase in the percentage of cells in S phase in HMC1/PR3 (70%) as compared with control (46%). No increase in cells in the subG1 phase, which represents apoptotic cells, could be detected in any of our transfectants (data not shown). Incubation of HMC1/PR3 in the presence of DEVD-Fmk, a specific inhibitor of caspase-3, did not reverse the proliferative effect of PR3 thereby demonstrating that caspase-3 is not directly involved in PR3-induced proliferation (Fig. 2 B). BrdU incorporation was significantly increased in PR3-transfected cells as compared with control (71 and 50%, respectively) and remained unchanged after incubation with DEVD-Fmk (71 and 52%, respectively).
No apoptosis could be detected in RBL/PR3 despite significant caspase-3 activity
To determine whether caspase-3 activation could result in apoptosis in PR3-transfected cells and to compare the intensity of endogenous caspase-3 activity in RBL/PR3 to caspase-3 activity triggered by apoptosis, control RBL (CT) and RBL/PR3 were treated with etoposide to induce apoptosis, which was then evaluated by different methods.
Measurement of caspase-3 activity was performed with or without an apoptotic inducer, either etoposide or HA-14. Etoposide caused a significant increase in caspase-3 activity in control RBL and in RBL/PR3 (Fig. 3,A). Interestingly, the endogenous caspase-3 basal activity detected in untreated RBL/PR3 was of similar magnitude as caspase-3 activity obtained in etoposide-treated control RBL. After apoptosis induction, caspase-3 activity increased about four times in control cells and one time and half in RBL/PR3. Similar results were obtained with HA-14, a pharmacologic compound that binds to Bcl-2 to inhibit its antiapoptotic effects. As shown in Fig. 3 B, 10 μM etoposide for 15 h triggered apoptosis in control RBL as well as in RBL/PR3 with 37 and 31% of apoptotic cells represented by annexin V labeling, respectively. Apoptotic cells are labeled only with PE-annexin V but not with 7-AAD. In the absence of etoposide, no binding of annexin V was observed in either untreated RBL/PR3 (6%) or control RBL (9%). Same results were obtained in HMC1 cells (data not shown).
To understand the molecular mechanisms resulting in caspase-3 activation, the mitochondria pathway has been investigated with the subsequent cytochrome c release. The release of cytochrome c from the mitochondrial intermembrane space to the cytosol, which is indicative of a variation in the mitochondrial permeability such as its transmembrane potential variation, has been studied either in control RBL or in RBL/PR3 after cell fractionation. Under basal conditions, cytochrome c was localized within the mitochondrial fraction and was absent from cytosol both in control RBL and in RBL/PR3 (Fig. 4,A). Of note, etoposide treatment triggered the cytochrome c release within the cytosolic fraction both in control RBL and in RBL/PR3. So constitutive caspase-3 activity measured in RBL/PR3 was not linked to cytochrome c release from mitochondria to cytosol. To test whether caspase-3 activation in PR3 transfectants could result from upstream caspase activation, caspase-8 and caspase-9 activities were measured using specific colorimetric substrates Ac-IETD-pNA and Ac-LEDH-pNA for caspase-8 and caspase-9, respectively. The reaction specificity was ascertained using the specific IETD-CHO and LEDH-CHO caspase-8 and caspase-9 inhibitor, respectively (data not shown). As shown in Fig. 4,B, no significant difference was detected in caspase-8 and caspase-9 basal activity in RBL/PR3 compared with control RBL (CT), RBL/PR3S203A, and RBL/HNE. In the presence of etoposide, no difference in the activity of caspase-8 and caspase-9 could be detected in all transfectants. So, increased caspase-3 activity in PR3-transfected cells was not related to the activation of upstream caspases such as caspase-8 or caspase-9. After analyzing the DNA fragmentation on agarose gel, it was confirmed that etoposide induced apoptosis in all transfectants (Fig. 4,C). No DNA fragmentation could be detected in untreated RBL/PR3 like in control RBL consistent with an absence of apoptosis. Measuring levels of cytoplasmic histone-associated DNA fragments by ELISA, indicative of apoptosis, also showed significant increase in etoposide-treated control RBL (CT) or RBL/PR3 or RBL/S203A, or RBL/HNE compared with untreated control RBL, RBL/PR3, RBL/S203A, and RBL/HNE, respectively (Fig. 4 D). No histone-associated proteins were detected in RBL/PR3 in basal condition.
This set of data demonstrated that the intensity of endogenous caspase-3 activation observed specifically in RBL/PR3 appeared to be physiologically relevant because it represents the same magnitude as caspase-3 activity obtained in control RBL during apoptosis. However, this caspase-3 activation is not associated with apoptosis evaluated by PS externalization and DNA fragmentation. Moreover, caspase-3 activation in RBL/PR3 is not a consequence of mitochondrial depolarization or upstream caspase-8 and caspase-9 activation. Likewise, an analysis of apoptosis was performed in control HMC1, compared with HMC1/PR3, in the presence or absence of apoptosis induction and the results corroborated those obtained in RBL transfectants with a strong caspase-3 activity in HMC1/PR3 in the absence of apoptotic phenotype. This is in keeping with the results of the cell cycle analysis showing an absence of a subG1 peak (data not shown).
Characterization of caspase-3 cleavage fragments in HMC1 cells and in vitro
We performed Western blot analysis to characterize caspase-3 activation fragments generated in HMC1/PR3 under basal condition and after apoptosis induction by HA-14 (40 μM, 2 h), using a polyclonal anti-caspase-3 Ab, which recognized both the procaspase-3 and its active forms. Under basal conditions, procaspase-3 appeared as a doublet at 32 and 26 kDa, which was present in all HMC1 transfectants (Fig. 5,A). In contrast, in HMC1/PR3, two additional fragments were observed under basal conditions: a major fragment at 22 kDa and a minor fragment at 19 kDa. The absence of this 22-kDa caspase-3 fragment in HMC1/PR3S203A, transfected with inactive mutated PR3, demonstrates that this cleavage resulted from PR3 enzymatic activity. To compare the cleavage fragments in HMC1/PR3 with the caspase-3 fragments obtained during apoptosis, all HMC1 transfectants were treated with HA-14. Apoptosis then generated others fragments of caspase-3 at 19, 17, and 15 kDa in control HMC1, HMC1/PR3S203A, and HMC1/HNE. The predominant fragment at 17 kDa has been described as the caspase-3 active fragment during apoptosis (22). Interestingly, these fragments were also present in HMC1/PR3 treated with HA-14 in addition to the 22-kDa caspase-3 fragment. The same caspase-3 activation pattern was obtained in RBL and when apoptosis was induced with etoposide (data not shown). To test the hypothesis of a direct cleavage of procaspase-3 by PR3, neutrophil-purified PR3 was incubated with recombinant procaspase-3. Western blot analysis showed that PR3 cleaved procaspase-3 into a 22-kDa fragment (Fig. 5,B). This fragment was the same size as the fragment generated in PR3-transfected cells (data not shown). Procaspase-3 activated with 12 nM PR3 showed a significant caspase-3 activity of the same magnitude as that obtained with caspase-9 used as a positive control for caspase-3 activation (Fig. 5,C). We also verified that PR3 could not directly cleave the DEVD substrate in the absence of caspase (data not shown). Therefore, it can be concluded that the direct cleavage of procaspase-3 by PR3 generated a 22-kDa active fragment. To evaluate whether HNE could directly cleave procaspase-3, purified HNE or PR3 was incubated in the presence of procaspase-3 from a lysate of HMC1 cells. As shown in Fig. 5,D, procaspase-3 was cleaved by purified PR3 in a dose-dependent manner beginning at 50 nM. No fragment at 22 kDa could be detected after incubation of procaspase-3 with HNE. Similar data were obtained using purified procaspase-3 (data not shown). The in vitro cleavage of procaspase-3 resulted from the disappearance of the 26-kDa fragment and the appearance of the 22-kDa fragment. In HMC1/PR3, we have also detected this 22-kDa caspase-3 fragment, but at the same time, the intensity of the procaspase-3 fragments were not modified. It could be hypothesized that, in HMC1/PR3, the majority of caspase-3 exist under its proform and the proportion of procaspase-3, which is cleaved by PR3, represented only a small part. On the contrary, in vitro, the ratio of procaspase-3 to PR3 was different, thus allowing a total cleavage of the 26-kDa fragment. The serine proteinase activity of either PR3 or HNE were determined using the chromogenic substrate Boc-Ala-Pro-Nval-Sbzl, thus providing evidence that the HNE used for in vitro assays was as active as PR3, at the same concentrations (Fig. 5 E). It thus can be concluded that, in contrast with PR3, HNE at nanomolar concentrations did not cleave procaspase-3 in vitro. This data demonstrated that the absence of caspase-3 activity in HNE-transfected cells could be due to both a different subcellular localization between HNE and PR3 and also to an increased ability of PR3 to process procaspase-3, as compared with HNE.
Subcellular localization of the 22-kDa caspase-3 fragment in HMC1/PR3
To determine the intracellular localization of the 22-kDa fragment resulting from PR3-activated caspase-3, subcellular fractionation into granules, membranes and cytosol was performed in HMC1/PR3. Western blot analysis of caspase-3 showed that the 22-kDa fragment, which was detected in the whole cell lysate, was present within the plasma membrane-enriched fraction but was absent from both the granule and the cytosolic fractions (Fig. 6,A). Moreover, using the specific fluorometric caspase-3 substrate Ac-DEVD-AMC, a higher caspase-3 activity was measured in the plasma membrane fraction as compared with the caspase-3 activity measured within the granule or the cytosolic fractions. No caspase-3 activity could be detected in control HMC1 using the fluorometric substrate (data not shown) thus confirming the results showed in Fig. 1,C (Fig. 6 B). Interestingly, within the membrane fraction, the 22-kDa caspase-3 fragment resulted from the cleavage of the 32 kDa procaspase-3 proform because a strong decrease in the intensity of the upper caspase-3 proform at 32 kDa was observed. Altogether this set of data suggest that in HMC1/PR3, PR3 can activate and cleave procaspase-3 into a specific fragment of 22 kDa in the plasma membrane compartment, without triggering apoptosis.
Evidence of the 22-kDa fragment of caspase-3 in resting neutrophils
To investigate whether PR3-activated caspase-3 could have a physiological relevance, Western blot analysis of caspase-3 was performed in neutrophils under resting conditions or following apoptosis (Fig. 7,A). Like in HMC1, in neutrophils the procaspase-3 appeared as a doublet at 32 and 26 kDa. Interestingly, a band at 22 kDa that lines up with the one observed in PR3-transfected HMC1 cells (Fig. 7,C) could be observed in resting neutrophils. Almost no apoptosis was detected in resting neutrophils because <5% of neutrophils were labeled with annexin V (Fig. 7,B). In accordance with our results obtained in PR3-transfected HMC1 and RBL, our data showed that caspase-3 cleavage by PR3 in neutrophils did not result in apoptosis. In neutrophils, apoptosis could be triggered by the treatment with gliotoxin (0.1 μg/ml, overnight at 37°C) or by a 37°C overnight incubation, as shown by annexin V labeling in Fig. 7,B (84 and 55%, respectively). Under apoptotic conditions using gliotoxin treatment, only the 17-kDa fragment of caspase-3 could be detected as previously described (22, 23, 24). After the 37°C overnight incubation, both the 22-kDa and the 17-kDa caspase-3 cleavage fragments were present because of the presence of both apoptotic (55%) and nonapoptotic neutrophils as determined by annexin V labeling. Moreover, when apoptosis was delayed by G-CSF treatment (23), the 22 kDa was present (data not shown). It thus becomes apparent that the presence of the 22-kDa caspase-3 fragment is inversely correlated with apoptosis. Interestingly, this fragment was not observed in neutrophils preincubated with pefabloc, a serine proteinase inhibitor thus strongly suggesting that a serine proteinase was involved in the generation of this 22-kDa caspase-3 fragment in the absence of apoptosis (9% of annexin V labeling). Treatment of neutrophils with cycloheximide (20 μg/ml, 4 h, 37°C), a protein synthesis inhibitor, did not affect the 22-kDa caspase-3 fragment, thus demonstrating that this caspase-3 processing by PR3 did not require de novo protein synthesis (data not shown). Moreover, the 22-kDa fragment of caspase-3 was observed in monocytes (data not shown) but was not observed in resting lymphocytes (Fig. 7 A) thus suggesting that it was myeloid-specific.
Subcellular localization of PR3-activated caspase-3 in membrane compartment of neutrophils
To investigate the intracellular localization of the 22-kDa caspase-3 fragment, subcellular fractionation was performed on neutrophils. Western blot analysis of caspase-3 showed that the 22-kDa fragment strictly lined up with the fragment obtained in HMC1/PR3. This fragment was absent from both the granule and the cytosolic fraction in neutrophil (Fig. 7,C). However, this fragment was detected only within the neutrophil plasma membrane-enriched fraction with a parallel decrease in the 32 kDa procaspase-3 band, in accordance with what we have described in HMC1/PR3. Interestingly, in the membrane fraction, the decrease of the 32-kDa intensity band was associated with the presence of the 22-kDa caspase-3 fragment. Moreover, the fluorometric measurement of caspase-3 activity performed on neutrophil fractions showed an increased caspase-3 activity in the membrane fraction as compared with either the cytosol or the granules. In addition, the specific membrane-associated caspase-3 activity was not measured in the membranes of resting RBL or lymphocytes where no procaspase-3 cleavage was observed (Fig. 7 D). As a positive control, caspase-3 activity increased about six times as compared to control neutrophil after apoptosis induction by gliotoxin (data not shown).
To investigate whether caspase-3 could colocalize with PR3 in neutrophils, double immunofluorescence labeling for caspase-3 and PR3 was analyzed. Permeabilization was performed using three distinct protocols. SLO, a streptococcus toxin that induces pore formation into membrane, interacting with cholesterol, permeabilized only the plasma membrane and does not affect the membranes of the granules, thus precluding the detection of granular proteins. By contrast, Triton X-100 or methanol permeabilized all the membranes of the cells, including plasma membrane and granule membranes, thus allowing detection of intragranular proteins. Myeloperoxidase (MPO) labeling using a polyclonal Ab was taken as a control to validate the different permeabilization protocols. MPO, which is exclusively localized within azurophilic granules, presents a granular pattern of fluorescence after Triton X-100 or methanol permeabilization. By contrast, MPO cannot be detected without granular permeabilization by SLO (Fig. 8,A). Next, immunofluorescence labeling was performed with the rabbit polyclonal anti-caspase-3 Ab, on resting neutrophils permeabilized with SLO (Fig. 8 B). This immunofluorescence of caspase-3 labeling (red fluorescence) clearly showed that caspase-3 had a cytosolic localization with a homogenous pattern and was partly associated with the membrane compartment. As a control, we have verified that the biotinylated Ab did not show an aspecific labeling (data not shown). Likewise, PR3 (green fluorescence) was detected in the same compartment, after SLO permeabilization. Thus confirming that the presence of PR3 was not strictly restricted to granules. Confocal microscopy analysis provided evidence of a possible colocalization of PR3 and caspase-3 in resting neutrophils (merge). From this set of experiments, it could be concluded that procaspase-3 and PR3 could colocalize in resting neutrophils in a subcellular compartment, which is distinct from granules. It can be suggested that this colocalization, which allows the cleavage of membrane-associated procaspase-3 into the 22-kDa fragment, could take place at the inner face of the plasma membrane.
To discover new molecular targets of PR3 to further understand its function within neutrophils, and to discriminate between PR3 and its homologue HNE, we have first used a cellular model of mast cell stably transfected with PR3 and then transpose our findings into neutrophils. In the present study, the new findings are 1) the activation of procaspase-3 by PR3 into a specific 22-kDa fragment, in the absence of apoptosis, 2) the presence of the 22-kDa fragment in resting neutrophils and the absence of this fragment in apoptotic neutrophils, and 3) the presence of this 22 kDa in the membrane compartment of both HMC1/PR3 and neutrophils, where PR3 and procaspase-3 can colocalize.
We have demonstrated that PR3 can cleave procaspase-3, thus resulting in a strong increase in caspase-3 activity. However, PR3-induced procaspase-3 activation did not result in apoptosis. In addition to this high level of caspase-3 activity, we could not detect a single feature of apoptosis in PR3-transfected cells. On the contrary, no caspase-3 activity could be detected in HNE-transfected cells. This absence of apoptosis in the PR3-transfected cells could be surprising because activation of caspases by serine proteinase has already been shown to result into apoptosis (25). Cathepsin G has been shown to cleave procaspase-7 in vitro, leading to an active form (26). However, the importance of this phenomenon in a cellular model with subcellular compartmentalization has not been investigated. PR3 has also been described to have proapoptotic properties using cellular models such as endothelial cells treated with extracellular purified PR3. Both active and inactivated PR3 could mediate this apoptotic effect (27). When active PR3 was used to induce apoptosis, the cleavage of several cell cycle regulatory proteins was observed including p21, thereby confirming our study and the cleavage of NF-KB, which has been described as necessary in some cell types during apoptosis. No proapoptotic effect of exogenous PR3 could be demonstrated on epithelial or myeloid cells. In the promyelocytic HL-60, it was reported that an HL-60 variant, which expressed less PR3 than the original HL-60, was resistant to apoptosis induced by doxorubicin, thus suggesting that PR3 fulfilled a proapoptotic role (28). In this report, those cells with high levels of PR3 showed increased caspase-3 activity and increased DNA fragmentation after doxorubicin treatment but did not demonstrate mitochondrial depolarization or caspase-8 or caspase-9 activation. In all previous studies, no clear identification was made of a molecular, PR3-specific target, and no comparisons exist using neutrophil elastase.
PR3 very original role in activating caspase-3 without triggering apoptosis, which we describe in this study, must be explained by the special PR3 subcellular localization. PR3 can interact with procaspase-3 and activate it in a subcompartment of the cell, thus precluding activated caspase-3 to have access to its usual substrates, which mediate apoptosis.
Indeed, we have already shown in PR3-transfected RBL, that there is a pool of PR3 associated with the plasma membrane, which allowed the PR3 to be expressed at the plasma membrane upon apoptosis, in the absence of degranulation. In contrast, in HNE-transfected RBL, no HNE could be expressed at the plasma membrane using similar experimental conditions thereby demonstrating that HNE has a different subcellular compartmentalization (18). These data demonstrated the presence of an extragranular pool of PR3 thus allowing PR3 to have specific target proteins, which are not localized within granules, but are located in cytosol such as p21 (12), or at the inner face of the plasma membrane, such as procaspase-3. Likewise, our data provide evidence that, in neutrophils, PR3-activated caspase-3 is restricted to the plasma membrane-enriched compartment. We previously demonstrated that PR3 was present within the membrane-enriched fraction in resting neutrophils (5). Now, we also showed that PR3 could be detected by immunofluorescence after SLO permeabilization, which does not permeabilize the granule membrane, thus precluding to detect granular proteins such as MPO. Therefore, the data obtained in our model of PR3-transfected mast cells combined with those obtained in mature neutrophils are in agreement with the presence of PR3 in a subcellular compartment distinct from the granules. Interestingly, in lymphocytes, granzyme B, a granular serine proteinase homologue, was also able to process procaspase-3 in activated nonapoptotic CD8+ T cells. Interestingly, it was described that small amounts of granzyme B were exported from cytosolic granules to the cytosol (29). The presence of the 22-kDa fragment of caspase-3 is found only in resting neutrophils or in G-CSF-treated neutrophils, in which apoptosis has been delayed, thus strongly suggesting that the 22-kDa caspase-3 fragment is associated with neutrophil survival. After apoptosis, the 22-kDa fragment is not detected but the 17-kDa fragment of caspase-3, typical of apoptosis could be detected thus providing evidence that the 22 kDa is not related to apoptosis but should play another role in resting neutrophils. Further studies are needed to investigate whether the membrane-associated 22-kDa fragment has specific protein substrates and whether it is associated with neutrophil survival. It should be noted that the detection of the doublet of procaspase-3 and the 22-kDa caspase-3 fragment depends on the anti-caspase-3 Ab used. In several reports, Western blot analysis of caspase-3 in neutrophils showed a single band of procaspase-3 at 32 kDa and the active fragment at 17 kDa in apoptotic neutrophils (23, 24). Interestingly, in several recent studies the 22-kDa caspase-3 fragment in resting neutrophils has already been shown (30, 31). In a particular study, focused on the effect of c-reactive protein on neutrophil apoptosis, the authors did not comment or explain the presence of this 22-kDa atypical caspase-3 fragment. The Western blot analysis of caspase-3 in resting neutrophils showed exactly the same bands as those we have observed: the doublet of procaspase-3 and the 22-kDa caspase-3 fragment thus strongly suggesting that the rabbit polyclonal anti-caspase-3 used in the study was the same as our anti-caspase-3 (30).
Finally, our present work strongly suggests that caspase-3 activation might have some biological functions in the absence of apoptosis. Although this hypothesis is uncommon, it already has been put forward that caspases might mediate other physiologic processes than apoptosis (32) and are associated with differentiation processes in various cell types, including lens fiber (33), keratinocyte (34), skeletal muscle (35), platelet (36), erythrocyte (37), osteoblast (38), and myeloid cells (39), and with proliferation in T cells (40). Interestingly, in megakaryocytes induced apoptosis, a diffuse caspase activation was described whereas a restricted caspase activation may induce differentiation, thus suggesting that caspase activation switch from a localized to a diffuse form to have access to its apoptotic substrates.
We described a novel role for caspase-3 in the absence of apoptosis. Our work therefore revealed a novel and unexplored track of research for caspase-3 functions. We described a potential novel role for PR3, which can mediate compartmentalized caspase-3 activation in resting neutrophils, related to neutrophil survival mechanisms or to terminal differentiation. Because we have previously reported that an increased expression of membrane-associated PR3 could be a risk factor for chronic inflammatory disorders such as vasculitis or rheumatoid arthritis (41), it could be speculated that membrane-associated PR3 could interfere with normal neutrophil apoptosis. Indeed, the delay in neutrophil apoptosis has been already associated with several, acute chronic inflammation diseases (42, 43).
We are indebted to Dr. Yaël Zermati and Dr. Stéphanie Vicca for extremely pertinent and helpful advice. We thank Patrick Nusbaum for expertise in flow cytometry, and Sandrine Canteloup and Xavier Tirvengadum for technical assistance. We thank Valérie Nicolas (Plateau Technique, Imagerie Cellulaire, Institut Fédératif de Recherche 75, Faculté de Pharmacie, Chatenay-Malabry, France) for excellent assistance in elaborating the confocal microscopy data, Yves Lepelletier for help in colocalization analysis, and Gérard Delrue of Institut Fédératif de Recherche Necker Enfants-Malades for Western blot scanning.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by fellowships from the Association de Recherche sur la Polyarthrite (to M.P.) and from Vaincre la Mucoviscidose (to C.K.), and by the Leg Poix. We thank Baxter Extramural Grant Program (to V.W.S.) and Amgen for their generous financial support.
This study was presented in part at the 38th Annual Scientific Meeting of the European Society for Clinical Investigation at the Phagocyte Workshop in Utrecht, the Netherlands, April 14–17, 2004, by M. Pederzoli, and V. Witko-Sarsat.
Abbreviations used in this paper: 7-AAD, 7-aminoactinomycin D; BrdU, 5-bromo-2′-deoxyuridine; HNE, human neutrophil elastase; HMC, human mast cell; MPO, myeloperoxidase; PR3, proteinase 3; PS, phosphatidylserine; RBL, rat basophilic leukemia; SLO, streptolysin-O.