Cystic fibrosis (CF) airways are characterized by bacterial infections, excess mucus production, and robust neutrophil recruitment. The main CF airway pathogen is Pseudomonas aeruginosa. Neutrophils are not capable of clearing the infection. Neutrophil primary granule components, myeloperoxidase (MPO) and human neutrophil elastase (HNE), are inflammatory markers in CF airways, and their increased levels are associated with poor lung function. Identifying the mechanism of MPO and HNE release from neutrophils is of high clinical relevance for CF. In this article, we show that human neutrophils release large amounts of neutrophil extracellular traps (NETs) in the presence of P. aeruginosa. Bacteria are entangled in NETs and colocalize with extracellular DNA. MPO, HNE, and citrullinated histone H4 are all associated with DNA in Pseudomonas-triggered NETs. Both laboratory standard strains and CF isolates of P. aeruginosa induce DNA, MPO, and HNE release from human neutrophils. The increase in peroxidase activity of neutrophil supernatants after Pseudomonas exposure indicates that enzymatically active MPO is released. P. aeruginosa induces a robust respiratory burst in neutrophils that is required for extracellular DNA release. Inhibition of the cytoskeleton prevents Pseudomonas-initiated superoxide production and DNA release. NADPH oxidase inhibition suppresses Pseudomonas-induced release of active MPO and HNE. Blocking MEK/ERK signaling results in only minimal inhibition of DNA release induced by Pseudomonas. Our data describe in vitro details of DNA, MPO, and HNE release from neutrophils activated by P. aeruginosa. We propose that Pseudomonas-induced NET formation is an important mechanism contributing to inflammatory conditions characteristic of CF airways.
Cystic fibrosis (CF) is an inherited disorder resulting from mutations in the CF transmembrane conductance regulator (CFTR) anion channel (1). Several organs are affected in CF, but lung complications remain the main cause of disease morbidity and mortality (2). Lack of apical plasma membrane expression of a fully functional CFTR in respiratory epithelial cells leads to altered fluid absorption, ion secretion, and mucous retention within the airways. As a result, CF airways become obstructed and infected with bacteria, leading to robust neutrophil infiltration and chronic inflammation (3). A wide variety of bacterial species have been detected in CF airways, but Pseudomonas aeruginosa represents the predominant pathogen infecting most CF patients (4–6).
The main pharmacological approaches used in the clinical management of CF airway disease focus on treating bacterial infections with antibiotics, promoting airway clearance by lysing DNA within mucous plugs, and rehydrating the airway surface (7–10). Survival and quality of life have improved significantly for people with CF over the last two decades; however, for most patients, the disease still cannot be adequately managed, let alone cured, and more effective treatments are urgently needed (8). Recently, promising new drug candidates have emerged that address the core problem in CF: the dysfunctional CFTR membrane protein (11–13). Although these CFTR “protein-assist” therapies are very promising, currently there is only one Food and Drug Administration–approved drug that potentiates CFTR function, in the case of defective channel gating, and no drugs have been approved to date to correct the defective cellular trafficking of the common CFTR F508del mutation (13).
Airway inflammation in CF is the product of a complex set of innate immune interactions. The neutrophil is a pivotal cellular player influencing the outcome of these interactions. Neutrophil density in CF sputum has been shown to correlate with CF disease severity (measured as forced expiratory volume in 1 s) (14, 15). Sputum and blood concentrations of human neutrophil elastase (HNE) and myeloperoxidase (MPO) in CF patients are associated with declines in lung function (14–18). IL-8, a major neutrophil chemoattractant that both airway epithelial cells and neutrophils produce, has also been associated with CF lung function decline (15, 16). IL-1β, a proinflammatory cytokine mainly produced by macrophages but also secreted by activated neutrophils, has also been linked to CF lung damage (14). These data obtained from clinical samples of CF patients clearly show that neutrophil recruitment and “activation” are major contributors to lung damage.
It is therefore important to understand why recruited neutrophils release their granule contents in CF. Formation of neutrophil extracellular traps (NETs) offers a possible mechanism (19). In NETs, extracellular DNA is associated with neutrophil granule components (HNE, MPO) and histones (19, 20). Neutrophils release NETs in response to bacterial and inflammatory stimuli (19, 21). Negative correlations were found between CF sputum extracellular DNA concentrations and lung function measures (16, 22). Neutrophils undergoing NET formation were detected in CF sputum samples (23–25). Histone citrullination, a histone modification characteristic for NET formation, was also detected in CF sputum samples (23).
P. aeruginosa is likely to activate neutrophils in CF airways. This organism is associated with diminished pulmonary function in CF and, similar to many other pathogens, triggers robust NET release in vivo (15, 26) and in vitro (27, 28). Pyocyanin, a major toxin of P. aeruginosa, which can be relatively abundant in the CF airway, enhances NET formation and has been linked to lung function decline in this disorder (26, 28–30). Pseudomonas and neutrophils are found in close proximity, and neutrophils have been shown to phagocytose bacteria in airway samples of CF patients (31). These data suggest that Pseudomonas–neutrophil interactions are highly relevant to CF airway pathophysiology. The major objective of this study was to provide insight into the mechanism of Pseudomonas-induced NET formation, a process highly relevant to neutrophil dysfunction in the CF airway. Our hypothesis was that P. aeruginosa–stimulated NET formation contributes to release of MPO, HNE, and extracellular DNA. Our approach to test this hypothesis was to quantify NET formation, MPO, and HNE release and activities after exposure of human neutrophils in vitro to laboratory standard strains and CF isolates of P. aeruginosa.
Materials and Methods
The Institutional Review Board of the University of Georgia (UGA) approved the human subject study to collect peripheral blood from healthy volunteers (UGA #2012-10769-06). Enrolled healthy volunteers were >18 y of age, were not pregnant, were >110 pounds, had no infectious disease complications, and provided written informed consent (32).
Human subjects also included CF patients enrolled in an observational study of CF lung disease severity, “Genetics of CF Lung Disease” (Seattle Children’s Hospital Institutional Review Board approved protocol 10855 and Partners Healthcare Systems/Massachusetts General Hospital Institutional Review Board approved protocol 2011P000544). The protocols and informed consent procedures were approved by the Institutional Review Boards of the Seattle Children’s Hospital and Partners Healthcare Systems/Massachusetts General Hospital, and the studies were performed in accordance with the ethical guidelines of the Declaration of Helsinki. All 10 CF patients in this study were homozygous for the F508del allele of the CFTR gene and were categorized as having “severe” lung disease: they were in the lowest quintile for age of airway obstruction, as assessed by their median forced expiratory volume during the initial second of exhalation (for more details, see Ref. 28).
Patients with X-linked (gp91phox-deficient) and autosomal (p47phox-deficient) chronic granulomatous disease (CGD) and healthy volunteers were recruited under the National Institute of Allergy and Infectious Diseases Institutional Review Board–approved protocol (National Institutes of Health #05-I-0213, “Evaluation of Patients with Immune Function Abnormalities”). Human subjects provided written informed consent for participation. Ten milliliters blood was drawn from CGD patients and was processed in parallel with healthy volunteers’ blood. Two X-CGD patients (Patients 2 and 3) and one p47phox-deficient CGD patient participated in the study; all subjects were characterized in previous studies with regard to their genetic defects and residual NADPH oxidase activities: Patient 2 (1.7 nmol/106 cells/h, gp-146, 0.75%) and patient 3 (2.38 nmol/106 cells/h [1.05%]) (33). The p47phox−/− patient (p47-16) had NADPH oxidase activity of 2.34 nmol/106 cells/h (1.04%), and superoxide production in healthy neutrophils was 226 nmol/106 cells/h (100%) (28, 33).
Human neutrophil isolation
Neutrophils were isolated as described previously (32). In brief, whole blood was drawn at the Health Center of UGA (Athens, GA) or at the Department of Transfusion Medicine of the National Institutes of Health (Bethesda, MD), and coagulation was prevented with heparin. RBCs were removed by dextran sedimentation (GE Healthcare, Amersham, U.K.). Leukocytes were layered on top of a 5-step Percoll gradient (65, 70, 75, 80, and 85%; Sigma, St. Louis, MO), and the 70-75-80% Percoll layers containing neutrophils were harvested. Cells were stored in a mix of autologous serum and RPMI 1640 medium (Life Technologies, Grand Island, NY) until use at room temperature (RT). Trypan blue dye extrusion was used to determine cell viability (>99%). Purity of the neutrophil preparations was confirmed by cytospin. Serum was prepared from coagulated blood by centrifugation and sterile filtration. Calcium- and magnesium-containing HBSS (Mediatech, Manassas, VA) containing 1% autologous serum, 5 mM glucose, and 10 mM HEPES was used as assay buffer.
P. aeruginosa strains
The following P. aeruginosa strains were used in this study: PA14 (kind gift from Dr. Frederick Ausubel, Massachusetts General Hospital, Boston, MA), PAO1 (American Type Culture Collection), PA2192 and GFP-expressing PAO1 (kindly provided by Dr. Joanna Goldberg, Emory University, Atlanta, GA), and PA10145 (ATCC 10145). CF isolates of P. aeruginosa were obtained as described previously (28). Bacteria were cultured in Luria–Bertani broth overnight. Late exponential phase cultures were washed twice and suspended in HBSS. Bacterial density was set to 109/ml using OD = 0.6 at 600 nm.
Quantification of NETs
NETs were quantified essentially as described with some modifications (32). Neutrophils were attached to poly d-lysine–coated 96-well black transparent-bottom plates (Thermo Scientific, Rochester, NY) in assay medium containing 10 μM Sytox Orange (Life Technologies, Grand Island, NY). For inhibitor tests, cells were pretreated with specific inhibitors for 15 min before bacterial infection. Neutrophils were infected with 1–50 multiplicity of infection (MOI) P. aeruginosa PA14 for dose-dependent extracellular DNA release assay or 10 MOI for inhibitor experiments. Fluorescence (excitation: 530 nm, emission: 590 nm) was measured in a fluorescence plate reader (Varioskan Ascent, Thermo Scientific) for up to 5 h at 37°C. Fluorescence of samples containing 500 μg/ml saponin (Sigma) with neutrophils was used as the maximal signal for DNA release (100%). Rise in fluorescence was referred to as “extracellular DNA release” and expressed as relative fluorescence units or normalized on saponin-treated as maximal signal (referred to as “% of max”). The following inhibitor concentrations were used: diphenylene iodonium [DPI], NADPH oxidase inhibitor, 10 μM; Sigma), U0126 (MEK1/2 inhibitor, 25 μM; Sigma), MEK162 (MEK1/2 inhibitor, 50 μM; Sigma), PD98059 (MEK1 inhibitor, 20 μM; Sigma), and cytochalasin D (cytoskeleton inhibitor, 10 μM; Sigma).
Immunostaining and fluorescent microscopy
Neutrophils were seeded on 12-mm glass coverslips (VWR International, Radnor, PA) in 24-well plates (Thermo Scientific, Rochester, NY). PA14 (opsonized with 10% autologous serum, 37°C, 30 min) was added to neutrophils and incubated for 3 h at 37°C. After fixation with 4% paraformaldehyde (Affymetrix, Cleveland, OH), samples were blocked with 5% normal donkey serum (Jackson Immunoresearch Laboratories, West Grove, PA) and 0.1% saponin (Sigma) in PBS for 30 min at RT. The following Abs were used: monoclonal mouse anti-human MPO/FITC Ab (1:500, clone MPO-7; Dako), polyclonal rabbit anti-histone H4 (citrulline 3; 1:1000; Millipore, Billerica, MA), and rabbit anti-HNE (1:1000) overnight at 4°C. For HNE and citrullinated histone H4 (citH4) staining, Alexa Fluor 594–labeled donkey anti-rabbit secondary Ab was used for 1 h (1:2000; Molecular Probes, Grand Island, NY). Samples were stained with DAPI (2 min, RT, 1:20,000; Molecular Probes) and washed in PBS twice. Specimens mounted with ProLong Antifade Kit (Molecular Probes) were analyzed with Zeiss AxioCam HRM fluorescence microscope (Axioplan2 imaging software).
Measurement of superoxide production
Superoxide release was assessed by Diogenes chemiluminescence superoxide-detection kit (National Diagnostics, Atlanta, GA). Adherent neutrophils were stimulated by P. aeruginosa PA14 (50 MOI), PMA (100 nM), or were left unstimulated. Chemiluminescence was measured by Varioskan Flash luminescence microplate reader (Thermo Scientific, Logan, UT) for 90 min. Data are shown as kinetics of representative curves (relative luminescence units) or integral superoxide production values (integrated relative luminescence units) obtained by multiplying accumulated luminescence for the entire duration of the measurement with the ratio of the measurement cycle length and integration time.
Supernatant collection for ELISA and enzyme activity measurements
Supernatants of neutrophils were used to measure secretion and activities of MPO and HNE. A total of 100,000 neutrophils/well were seeded into 96-well poly-d-lysine–coated transparent plates (Thermo Scientific, Rochester, NY), stimulated with 100 nM PMA or P. aeruginosa PA14 (50 MOI), respectively. After 3 h of incubation at 37°C in HBSS including 1% serum, cell supernatants were collected, centrifuged for 10 min at 10,000 × g to remove debris or bacteria, and stored at −80°C for further analysis or used immediately. Each supernatant was diluted 1:30 for HNE ELISA, 1:100 for MPO ELISA, or left undiluted for MPO activity test.
MPO and HNE ELISA
Concentrations of human MPO in neutrophil supernatants were measured by commercial ELISA kit (R&D Systems, Minneapolis, MN). Serial dilutions prepared from MPO standard provided in the kit (stock: 125 ng/ml) were used to quantify MPO concentrations of unknown samples. HNE release was assessed by ELISA. Supernatant samples diluted with coating buffer (25 mM carbonate, 25 mM bicarbonate, pH 9.6) were incubated overnight at 4°C in 96-well high binding microlon ELISA plates (Greiner bio-one, Fricken-hausen, Germany). After blocking with 1% BSA for 1 h, anti-HNE rabbit polyclonal Ab (1:500 in PBS [Calbiochem], 481001 [EMD Millipore]) was added for 2 h at RT. After repeated washes, samples were incubated with horseradish peroxidase–linked donkey anti-rabbit Ab (1:2000 in PBS, NA934V; GE Healthcare) for 1 h. Blue coloration developed in the presence of 3,3′,5,5′-tetramethylbenzidine (Thermo Scientific, Rockford, IL) peroxidase substrate solution. Reaction was stopped by adding 1N hydrochloric acid (Sigma), and absorption was read at 450-nm wavelength with Eon microplate photometer (BioTek, Winooski, VT). Purified HNE standard (stock: 1 mg/ml; Cell Sciences, Canton, MA) was used to determine HNE concentrations in unknown samples.
Measurement of peroxidase activity
Peroxidase activity was measured by hydrogen peroxide–dependent oxidation of Amplex Red. Fifty microliters undiluted neutrophil supernatants was added to 96-well nontransparent black microplates (Costar, Corning, NY) and mixed with assay solution containing 100 μM Amplex Red (Molecular Probes, Eugene, OR) and 100 μM hydrogen peroxide (Sigma). Production of the fluorescent product was measured with fluorescence plate reader (Varioskan Ascent, Thermo Scientific) after 30 min at 560-nm excitation and 590-nm emission wavelengths. Standards with known MPO concentrations (stock: 125 ng/ml MPO; R&D Systems) were used to determine peroxidase activities of unknown samples. Results are expressed as “equivalent ng/ml MPO activity.” This way of data presentation was chosen (instead of showing unit enzymatic activities per time and volume) to enable comparison between MPO concentration and activity in neutrophil supernatants.
Measurement of “NET-free” and “NET-linked” bacterial numbers
PA14 was added to human neutrophils (100,000/well) attached to poly-d-lysine–coated transparent 96-well plates (Thermo Scientific, Rochester, NY) in HBSS. Bacteria were added at two different doses: PA14/PMN MOI = 10:1 or 50:1. After 3 h, supernatants were carefully collected and referred to as “NET-free” fraction. Absence of DNA in this fraction was confirmed by DNA gel electrophoresis. Equal volume of assay medium containing 30 μg/ml DNAse I (Roche Applied Science, Indianapolis, IN) was added back to neutrophils and incubated for 15 min; this volume was referred to as “NET-linked” fraction. After centrifugation, concentration of living PA14 bacteria in each sample was determined by a microplate-based bacterial growth assay as described previously (34).
Results were analyzed by Student t test or one-way ANOVA with Tukey posttest for multiple comparisons. Each experiment was independently performed at least three times with neutrophils isolated from different donors. Correlation was determined by computing the Pearson correlation coefficient (r). The significance of correlation was (two-tailed probability) determined using the correlation coefficient and the sample size (20) using online statistical software (Statistics Calculators, version 3.0 Beta). Statistically significant differences were considered as follows: *p < 0.05, **p < 0.01, ***p < 0.001.
Human neutrophils release extracellular DNA in response to P. aeruginosa
We show that human neutrophils release extracellular DNA in response to GFP-expressing P. aeruginosa PAO1 (Fig. 1A). Bacteria are entrapped in NETs (Fig. 1A and insets). Neutrophils also release DNA in response to another P. aeruginosa strain, PA14 (Fig. 1B, 1C). PA14-induced DNA release starts at the same time when induced by the positive control, PMA (Fig. 1B). Bacteria (PA14) alone do not release any DNA (Fig. 1B). DNA release from adherent human neutrophils is induced in a dose-dependent manner by P. aeruginosa strain PA14 (Fig. 1C). Quantitation reveals an average release of 3.01 ng/μl DNA from 100,000 neutrophils on PA14 exposure (Fig. 1D). PAO1 and two other P. aeruginosa strains (PA10145, PA2192) also stimulate DNA release in a dose-dependent manner (Fig. 1E). We next tried to estimate the proportion of live bacteria attached to NETs after a 3-h incubation time. The supernatant of Pseudomonas–neutrophil suspensions was very carefully removed after incubation, did not contain NETs, and was therefore termed “NET-free” (Fig. 1F). Absence of DNA in the NET-free fraction was confirmed by running the samples on DNA agarose gels (data not shown). An equal volume of new assay medium containing DNAse I was added back and was referred to as “NET-linked.” As shown in Fig. 1F, we found that 41.3 ± 3.4% of live bacteria (mean ± SEM, n = 3) were associated with NETs when a MOI of 10:1 was used and 29.7 ± 13.6% (mean ± SD, n = 3) with a MOI of 50:1. These measurements reveal that in vitro NETs are capable of trapping large amounts of P. aeruginosa.
Pseudomonas-induced NETs contain citH4
The signaling steps leading from stimulation to NET formation in human neutrophils have not been well defined, but citrullination of histones (H3 and H4) has been implicated and confirmed by several groups (35–37). Citrullinated histone H3 was detected in CF sputum samples by immunofluorescence (23). In this article, we show that histone H4 is citrullinated on amino acid residue 3 and colocalizes with extracellular DNA after Pseudomonas exposure (Fig. 2A).
MPO is associated with Pseudomonas-triggered extracellular DNA traps
MPO is stored in azurophilic granules in neutrophils, becomes associated with extracellular DNA in NETs, and is also implicated in the progress of NET formation (19, 20). Our indirect immunofluorescence analysis detects colocalization of MPO with extracellular DNA and citH4 in Pseudomonas-stimulated NETs (Fig. 2A). These data prove that MPO released from neutrophils in the presence of P. aeruginosa is associated with NETs.
Neutrophil elastase is bound to NETs after Pseudomonas challenge
HNE is a major inflammatory marker of CF airways (14–16, 38). HNE is associated with extracellular DNA in NETs stimulated in vitro by PMA and in CF sputum samples (19, 24). Although HNE is stored in primary granules in resting neutrophils, it becomes associated with extracellular DNA after stimulation with P. aeruginosa PA14 (Fig. 2B, 2C). Thus, the main CF airway pathogen P. aeruginosa activates human neutrophils to release NETs that contain extracellular DNA, HNE, MPO, and cit3H4.
MPO and HNE are released after neutrophil exposure to P. aeruginosa
To confirm our immunofluorescence data, we measured MPO and HNE release in supernatants of Pseudomonas-stimulated human neutrophils by ELISA. Throughout the entire article we measured total MPO or HNE concentrations and MPO activity in our samples. We did not distinguish between NET-associated and NET-free proteins. Quantitation of MPO release by ELISA detected 176.0 ± 14.5 ng/ml MPO in neutrophil supernatants after Pseudomonas (PA-14) stimulation and only 29.2 ± 1.7 ng/ml with basal release (mean ± SEM, n = 8; Fig. 3A). We also observed an increased peroxidase activity in supernatants of Pseudomonas-activated neutrophils (basal: 28.9 ± 6.0, bacterial stimulus: 77.7 ± 11.9 ng/ml MPO activity, mean ± SEM, n = 8; Fig. 3B). A similarly large increase was observed in HNE release on exposure to PA14 (Fig. 3C). Background release of HNE (165.1 ± 19.8 ng/ml) was enhanced 2.7-fold by bacterial challenge (447.2 ± 40.4 ng/ml; mean ± SEM, n = 8). We did not detect any significant peroxidase activity with bacteria alone (data not shown).
Human neutrophils release DNA, MPO, and HNE in the presence of CF isolates of P. aeruginosa
To confirm that our previous findings are not limited to Pseudomonas laboratory strains, we exposed human neutrophils to CF clinical isolates of P. aeruginosa and measured DNA, MPO, and HNE releases. All clinical isolates induced release of extracellular DNA (Fig. 4A), MPO (Fig. 4B), and HNE (Fig. 4C) at comparable levels with PAO1. This confirms that clinical isolates behave similarly to the laboratory strains and that PAO1 and PA14 are good models to study Pseudomonas-initiated neutrophil activation. DNA, MPO, and HNE release data show similar patterns among isolates. To determine the extent of correlation among the parameters we measured, we plotted extracellular DNA, MPO, and HNE release of clinical isolates against each other (Fig. 4D–F). We calculated the Pearson correlation coefficients (r) to measure the strength of a linear association and the significance between any of these two variables: MPO versus HNE: r = 0.89157226975503, p = 0.00000013; MPO versus DNA: r = 0.80978489595945, p = 0.00001517; and HNE versus DNA: r = 0.62058387267863, p = 0.00350493. Correlations in each combination were determined to be significant.
The neutrophil respiratory burst is activated by P. aeruginosa
NADPH oxidase activity is required for NET formation triggered by bacteria and PMA (39, 40). The requirement of neutrophil respiratory burst activity for Pseudomonas-induced NETs has not been studied. First, we wanted to see whether adherent human neutrophils respond to P. aeruginosa with superoxide production. Neutrophils stimulated by PA14 released large amounts of superoxide that was dependent on the dose of bacterial challenge (Fig. 5A). As expected, the flavoenzyme inhibitor DPI diminished the Pseudomonas-initiated respiratory burst (Fig. 5B). We also exposed neutrophils of X-linked CGD patients to PA14 and measured the chemiluminescence signal. CGD neutrophils were entirely unresponsive to the bacterium, whereas oxidase-competent control neutrophils responded with large respiratory burst activity (Fig. 5C). Thus, human neutrophils respond to P. aeruginosa with oxidative activity.
Reactive oxygen species are required for neutrophil activation by P. aeruginosa
We then examined whether treating neutrophils with the NADPH oxidase inhibitor DPI before Pseudomonas challenge would affect extracellular DNA release and observed a significant reduction by prior DPI treatment (Fig. 5D). Neutrophils from both a gp91phox- and a p47phox-deficient patient remained unresponsive to PAO1 or PA14, further confirming the requirement for a functional NADPH oxidase in Pseudomonas-induced NET formation (Fig. 5E). DPI pretreatment also greatly reduced Pseudomonas-induced total MPO release (Fig. 5F) and extracellular peroxidase activity (Fig. 5G). PA14-triggered MPO release (146.8 ± 14.8 ng/ml) was reduced by up to 77.5% by DPI (33.2 ± 7.0 ng/ml; mean ± SEM, n = 8; Fig. 5F). PA14-stimulated extracellular peroxidase activity (51.6 ± 7.3) was inhibited by up to 63.6% by NADPH oxidase inhibition (18.8 ± 4.6; relative fluorescence units, mean ± SEM, n = 8; Fig. 5G). PA14-triggered secretion of total HNE (282.1 ± 29.4 ng/ml) was also reduced by 61.5% after DPI pretreatment (108.7 ± 23.2 ng/ml; mean ± SEM, n = 8; Fig. 5H). Our results demonstrate that NADPH oxidase mediates Pseudomonas-induced NET formation, and this process is responsible for the majority of MPO and HNE release from neutrophils on Pseudomonas challenge.
Pseudomonas-stimulated respiratory burst and DNA release requires an intact cytoskeleton
Human neutrophils phagocytose bacteria; however, the extent to which this process is required for Pseudomonas-activated NET formation is unknown. We stimulated human neutrophils with P. aeruginosa PA14 in the presence of the cytoskeleton inhibitor cytochalasin D and measured superoxide production and NET formation. Cytochalasin D treatment blocked both the Pseudomonas-triggered respiratory burst (Fig. 6A) and DNA release (Fig. 6B).
Role of MEK/ERK signaling in Pseudomonas-triggered NET formation
Activation of the MEK/ERK signaling pathway has been implicated in PMA-stimulated NET formation (41, 42). We tested the effect of U0126, a compound with known inhibitory effect on MEK1/2, on Pseudomonas-triggered neutrophil activation. U0126 reduced PA14-induced extracellular DNA release and caused a 67.0% decline in PA14-stimulated superoxide production (Fig. 7A). U0126 had an inhibitory action of 42.6% on Pseudomonas-triggered MPO release, but entirely blocked peroxidase activity in the supernatant (Fig. 7A). The fact that U0126 blocked MPO activity but had only moderate-to-medium inhibitory action on the other parameters raised concerns that U0126 can interfere with MPO activity (Fig. 7A). We therefore tested U0126 and two other, well-characterized MEK inhibitors (MEK162 and PD98059) in a cell-free system measuring MPO activity. Surprisingly, we found that U0126 strongly inhibited in vitro MPO function, whereas MEK162 and PD98059 were without effect (Fig. 7B).
We next studied the effects of MEK162 and PD98059 on neutrophils. Both inhibitors had minor but significant inhibition of Pseudomonas-triggered extracellular DNA release: MEK162 inhibited 33% of DNA release (n = 4) and PD98059 inhibited 38% of DNA release (n = 4; Fig. 7C, first panel). MEK162 and PD98059 did not influence release or activity of MPO (Fig. 7C). MEK162 (but not PD98059) mildly but significantly reduced HNE release (Fig. 7C). These data confirm that whereas MEK/ERK signaling has minor contribution to extracellular DNA secretion, it does not generally influence release of MPO and HNE.
Neutrophil airway recruitment and activation in CF is poorly understood and has not been targeted therapeutically. In this study, we explored details of neutrophil responses to the main CF airway pathogen, P. aeruginosa. This mechanism is of high importance to understand CF lung inflammation because both presence of P. aeruginosa and neutrophil activation have been strongly linked to diminished lung function in CF (14–18, 26, 38, 43–45).
We found that the main effector response of neutrophils in response to PA14 in vitro is NET formation (Figs. 1–3). The backbone of NETs is extracellular DNA (19). DNA has been detected in CF airways, and its presence correlates with accelerated decline of lung function in CF patients (16, 22). Moreover, DNAse treatment is routinely used to ease CF respiratory symptoms (16, 22, 46, 47). Our data show that neutrophils stimulated by different P. aeruginosa strains (PA14, PAO1, PA10145, and PA2192) release large amounts of DNA (Fig. 1). This is consistent with data obtained by several other groups (25, 27, 28, 48–50). Our data allow comparison of different bacterial strains used in Pseudomonas research and found that PA14 and PA10145 are the strongest NET inducers, whereas PAO1 and PA2192 release NETs to a lesser extent (Fig. 1). Similarly, CF isolates show a wide variation in NET induction: most strains stimulate DNA release at levels comparable with that of PAO1, whereas a few (#28, #64, #300) provoke marked responses from human neutrophils (Fig. 4A). Although our data used human neutrophils isolated from peripheral blood of healthy volunteers, these data confirm that P. aeruginosa is a strong inducer of DNA release from neutrophils and suggest that P. aeruginosa is likely to stimulate NET formation in CF airways.
Neutrophil azurophilic granule components MPO and HNE are found in large amounts in CF airways and contribute to tissue destruction through oxidative stress and proteolytic damage (14, 16, 18). Recently, HNE was found in the airways of CF infants, where its presence was associated with persistent bronchiectasis (38). Colocalization of neutrophil granule proteins (HNE, MPO) and histones with DNA is required to define the mechanism of DNA release as NET formation (19, 40). Citrullination of histones (H3, H4) by peptidyl arginine deiminase 4 is required for in vitro and in vivo NET release (35–37, 51). Citrullinated histone H3 has previously been detected in CF sputum samples (23). Pseudomonas-stimulated NETs contain citH4, MPO, and HNE that colocalize with extracellular DNA, as detected by immunofluorescence (Fig. 2). Our data show that when neutrophils are in the presence of Pseudomonas in vitro, DNA release occurs mainly through NET formation. Physical association of DNA, MPO, and HNE was found previously in CF sputum samples (23, 24). Absolute quantitation of total MPO release detected by ELISA versus enzymatic activity data suggests that ∼45–50% of released MPO is enzymatically active after Pseudomonas challenge (Figs. 3, 5F, 5G, 7A, and 7C). MPO has been shown to remain active in its NET-bound state (52).
Our survey of CF isolates of P. aeruginosa shows for the first time, to our knowledge, that these clinical strains stimulate release of extracellular DNA, MPO, and HNE from human neutrophils (Fig. 4). Previously, only NET-mediated killing of CF isolates of P. aeruginosa was measured (27). Linear correlation between extracellular DNA, MPO, and HNE indicate that their release is not independent but coordinated, through mechanisms such as NET formation (Fig. 4D–F). All clinical isolates tested trigger DNA/MPO/HNE release, but to a different extent. Future studies are required to determine whether the extent of NET induction by the clinical isolates correlates with clinical measures of the CF patients from whom they were isolated.
Oxidative stress has been implicated in the pathogenesis of CF (53). In the inflammatory CF lung, neutrophils are the main source of released reactive oxygen species. P. aeruginosa stimulates a robust neutrophil respiratory burst that is required for DNA release and contributes to MPO and HNE release (Fig. 5). Our data confirm previous findings that DPI inhibits extracellular DNA release from neutrophils stimulated by a clinical isolate of P. aeruginosa of unknown origin (49). Beyond producing tissue-damaging reactive oxygen species, neutrophil NADPH oxidase is required for maximal release of DNA, MPO, and HNE after Pseudomonas challenge. This suggests that NADPH oxidase–dependent NET formation stimulated by Pseudomonas occurs in CF airways. The need for NADPH oxidase to mediate NET formation could depend on the nature of the stimuli (54). The observation that DPI blocks DNA/NET release in Pseudomonas-treated neutrophils suggests that DPI treatment can distinguish between NET-dependent and -independent release mechanisms (Fig. 5). Incubation with DPI inhibited a major part of both MPO and HNE release (Fig. 5F–H), suggesting that NETs are the main mechanism of neutrophil inflammatory marker release.
Our previous data indicate that NET formation triggered by calcium pyrophosphate dihydrate crystals requires an intact cytoskeleton (32). Similarly, bacteria represent microscopic particles that neutrophils routinely phagocytose. The observation that cytochalasin D inhibits Pseudomonas-stimulated superoxide production and NET formation further suggests that NET release requires an intact cytoskeleton (Fig. 6). Cytochalasin B treatment has a similar effect on Staphylococcus aureus–triggered NET formation (55). It is not clear whether phagocytosis of microbes is required or not for microbe-stimulated NET formation. One school of thoughts argues that neutrophils kill microbes either by phagocytosis (and consequent intracellular killing) or by extracellular NET formation, but not using both mechanisms (56).
By studying the effect of the MEK/ERK pathway in Pseudomonas-triggered NET formation, to our surprise we found that U0126 was very efficient at blocking neutrophil effector functions, whereas MEK162 and PD98059 were not (Fig. 7). The fact that U0126 is an efficient blocker of MPO activity could explain this difference. This raises concerns about using U0126 as an MEK inhibitor in neutrophil studies because it clearly blocks MPO activation as well. U0126 has been previously used to implicate the MEK-ERK signaling pathway in PMA- or Helicobacter pylori–stimulated NET formation (41). However, this supports that MPO activity is required for NET formation stimulated by PMA but also by P. aeruginosa (55, 57). The requirement of MPO for Pseudomonas-induced NETs is, however, controversial because other studies found that MPO-deficient human neutrophils or murine neutrophils do not have impaired NET release (48, 49). The mechanism of inhibitory action of U0126 on MPO function is not known. Its chemical structure (four primary amines and two disulfide bonds) suggests that U0126 could be an excellent scavenger of hypochlorous ions (55, 58). MEK162- and PD98059-mediated blockade of the MEK-ERK pathway decreased Pseudomonas-stimulated DNA release but had no significant effect on MPO measures or HNE release (except for MEK162 on HNE ELISA; Fig. 7). This suggests that the MEK/ERK pathway contributes to Pseudomonas-induced NET release to a similar extent as in PMA induction (41). However, this minor effect on NET inhibition does not appear to contribute significantly to MPO or HNE release, as NET formation is not their only release mechanism.
P. aeruginosa acquires resistance to NET-mediated killing in CF airways (27). Clinical data clearly indicate that neutrophils are unable to remove Pseudomonas effectively from the CF lung, as both NET-mediated and classical intracellular killing mechanisms fail. Neutrophils instead release their dangerous granule contents into the airway lumen. Although the CF airway environment is very complex and multiple mechanisms can be responsible for neutrophil dysfunction, our detailed in vitro characterization of Pseudomonas-induced NET formation (DNA associated with active MPO and HNE) presents a likely mechanism for excess inflammatory mediator release from neutrophils in CF. Our study helps to understand neutrophil dysfunction, a clinically relevant mechanism of CF airway disease.
We thank the staff of the UGA Health Center: Angela Standridge and Vickie Cromer for phlebotomy; and Adam Davis, Houston Taylor, and Dr. Ronald L. Forehand for their continuous support. We thank Sandra Anaya-O’Brien for CGD patient recruitment and Dr. Douglas Kuhns for archived patient data (National Institutes of Health, National Institute of Allergy and Infectious Diseases). We are grateful to Drs. Brian Condie and Kristina Buac (Department of Genetics, UGA) for providing access to and help with the AxioZeiss fluorescence microscope. We also thank Dr. Joanna B. Goldberg (Emory University) for providing the GFP-expressing PAO1 and the PA2192 strains.
This work was supported by startup funds from the Office of the Vice President for Research, University of Georgia (to B.R.) and by the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Diseases.
Abbreviations used in this article:
CF transmembrane conductance regulator
chronic granulomatous disease
citrullinated histone H4
human neutrophil elastase
multiplicity of infection
neutrophil extracellular trap
University of Georgia.
The authors have no financial conflicts of interest.