Human Neutrophils Kill Streptococcus pneumoniae via Serine Proteases Alistair J. Standish and Jeffrey N. Weiser This information is current as J Immunol 2009; 183:2602-2609; Prepublished online 20 of October 2, 2021. July 2009; doi: 10.4049/jimmunol.0900688 http://www.jimmunol.org/content/183/4/2602 Downloaded from References This article cites 48 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/183/4/2602.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Human Neutrophils Kill Streptococcus pneumoniae via Serine Proteases1

Alistair J. Standish and Jeffrey N. Weiser2

Neutrophils, or polymorphonuclear leukocytes, comprise a crucial component of innate immunity, controlling bacterial and fungal infection through a combination of both oxidative and nonoxidative mechanisms. Indeed, neutrophils are believed to play an important role in controlling infection caused by the major human pathogen Streptococcus pneumoniae. However, the method by which neutrophils kill the pneumococcus as well as other Gram-positive bacteria, is not fully understood. We investigated human neutrophil killing of the pneumococcus in a complement-dependent opsonophagocytic assay. In contrast to other Gram-positive organisms, inhibition of the NADPH oxidase did not affect killing of S. pneumoniae. Supernatant from degranulated neutrophils killed the pneumococcus, suggesting a role for granular products. When neutrophil granule proteases were inhibited with either a protease mixture, or specific serine protease inhibitors 4-(2-Aminoethyl)benzenesulfonylfluoride and diisopropylfluorophos- phate, killing by neutrophils was inhibited in a manner that correlated with increased intracellular survival. All three compounds Downloaded from inhibited intracellular activity of the three major neutrophil serine proteases: elastase, cathepsin G, and proteinase 3. Additionally, purified elastase and cathepsin G were sufficient to kill S. pneumoniae in a serine protease dependent-manner in in vitro assays. Inhibition studies using specific inhibitors of these serine proteases suggested that while each serine protease is sufficient to kill the pneumococcus, none is essential. Our findings show that Gram-positive pathogens are killed by human neutrophils via different mechanisms involving serine proteases. The Journal of Immunology, 2009, 183: 2602–2609. http://www.jimmunol.org/ olymorphonuclear leukocytes (neutrophils) are a major host defense against this Gram-positive diplococcus. Neutrophils component of innate immunity, providing a crucial first are among the first cells recruited during many stages of Spn P barrier against both bacterial and fungal infection. Indeed, pathogenesis, including asymptomatic nasopharyngeal coloniza- neutropenic individuals are highly susceptible to a multitude of life tion, as well as invasive infection of the lungs (2, 3). Recently, we threatening infections as a direct result of the loss of these immune reported that a normally asymptomatic colonizing strain became cells. In healthy individuals, neutrophils circulating in the blood- invasive when mice were depleted of neutrophils (4). Additionally, stream are the first cells recruited to the site of infection, where neutrophil killing early in infection contributed to the subsequent they engulf pathogens into the phagolysosome thus exposing them development of the adaptive immune response. These observa- by guest on October 2, 2021 to a multitude of both oxidative and nonoxidative mechanisms of tions, therefore, confirm that neutrophils play an important role in killing. It has also been suggested that neutrophils produce neu- controlling Spn infection. However, the mechanism by which neu- trophil extracellular traps (NETs),3, which first trap pathogens trophils kill Gram-positive pathogens such as Spn is still largely within chromatin structures, and subsequently kill with various unknown. antimicrobial proteins (1). Neutrophil antimicrobial mechanisms are generally divided into Streptococcus pneumoniae (Spn), commonly known as the two distinct groups, the oxidative and nonoxidative. The oxidative pneumococcus, is a major human pathogen responsible for much burst, which generates reactive oxidative species (ROS) through an morbidity and mortality worldwide. Protection against disease and NADPH oxidase system, was long thought of as the neutrophil’s colonization is mainly mediated by opsonin-dependent phagocy- primary mode of killing pathogens. The essential nature of this tosis. Neutrophils have long been considered key effector cells in defense is seen by the fact that individuals with chronic granulo- tomas disease (CGD), which results in reduced production of ROS, have recurrent bacterial and fungal infections. However, Spn and Department of Microbiology, University of Pennsylvania School of Medicine, Phil- other catalase-negative organisms are not frequently associated adelphia, Pennsylvania 19104 with infection in CGD patients. Indeed, neutrophils isolated from Received for publication March 4, 2009. Accepted for publication June 18, 2009. CGD patients killed Spn just as efficiently as the control in ex vivo The costs of publication of this article were defrayed in part by the payment of page experiments (5). Additionally, mice deficient in enzymes associ- charges. This article must therefore be hereby marked advertisement in accordance ated with phagocyte production of ROS, have shown that ROS do with 18 U.S.C. Section 1734 solely to indicate this fact. not play a role in controlling Spn infection in either the lungs or 1 This work was supported by grants from the U.S. Public Health Service to J.N.W. (AI44231 and AI38446). brain (6, 7). Thus, this suggests that other neutrophil killing mech- 2 Address correspondence and reprint requests to Dr. Jeffrey N. Weiser, 402A John- anisms are necessary to control Spn. son Pavilion, 3610 Hamilton Walk, Philadelphia, PA 19104. E-mail address: Other than ROS, neutrophils also produce a multitude of anti- [email protected] microbial molecules in their granules. The two major granule sub- 3 Abbreviations used in this paper: NET, neutrophil extracellular trap; Spn, Strepto- sets are the azurophilic and specific granules. The azurophilic coccus pneumoniae; ROS, reactive oxygen species; CGD, chronic granulotomas dis- ease; BPI, bactericidal permeability-increasing protein; TS, tryptic soy broth; DPI, granules, are the first to fuse with the phagolysosome and contain diphenyleneiodonium; PI, protease inhibitor cocktail; AEBSF, 4-(2-Aminoethyl)ben- defensins, bactericidal permeability-increasing protein (BPI) as zenesulfonylfluoride, Hcl; DFP, diisopropylfluorophosphate; wt, wild type; OPH, op- well as serine proteases such as elastase, cathepsin G and protein- sonophagocytic; CytB, cytochalasin B; HNP, human neutrophil peptides. ase 3. Although BPI and defensins function by disrupting anionic Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 bacterial surfaces, the serine proteases generally degrade bacterial www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900688 The Journal of Immunology 2603 proteins, including virulence factors. The invading pathogen is CD18)-deficient mice (Jackson ImmunoResearch Laboratories) have a tar- then exposed to the specific granules that contain additional mi- geted mutation in the gene for integrin ␣ M or CR3 (19). crobicides such as lactoferrin, which has been implicated in de- Opsonophagocytic killing assays fense against Spn (8). The importance of these nonoxidative mechanisms is exempli- Opsonophagocytic (OPH) killing assays were conducted essentially as de- scribed by Davis et al. (2008). Bacterial strains were grown to mid-log fied by Chedaki-Higashi syndrome. This syndrome results in re- phase, PBS-washed, and resuspended in ϩϩϩ solution. In brief, 103 bac- duced levels of microbicidals as well as lower mobilization of its terial cells (in 10 ␮l) were preopsonized with infant rabbit serum shown to granules (9). Patients with this syndrome exhibit recurrent infec- be free of Spn Abs (40 ␮l) (Pel-Freez) for 30 min at 37°C. When Abs were tions and lower life expectancy. Interestingly, neutrophils isolated required, a 1/100 dilution of type 6 specific capsular anitbodies (Statens serum Institut) was added to the preopsonisaton reaction. Neutrophils were from a Chedaki-Higashi syndrome patient had a diminished ca- then added to reactions (105 cells per reaction in 40 ␮l) with ϩϩϩ solution pacity to kill Spn, indicating that nonoxidative mechanisms play an (110 ␮l) and incubated for 45 min at 37°C with rotation. Reactions were important role in neutrophil-mediated killing (10). Furthermore, a stopped by incubation at 4°C, and viable counts of bacteria were deter- recent study showed that granule extract purified from human neu- mined. Percent survival was determined relative to control reactions lack- trophils had activity against Spn (11). In addition, a number of ing neutrophils. To measure intracellular survival, gentimicin sulfate (final concentration 300 ␮g/ml) was added to the OPH reactions for 20 min. neutrophil-derived microbicides possess anti-Spn activity in par- After a 20-min incubation at 37°C, the antibiotic was removed by serial ticular conditions in vitro, while in vivo administration of a de- washing before plating for viable counts. riviative of BPI provided protection against Spn infection (12). Bacterial uptake assay Although the mechanism whereby neutrophils kill Gram-nega- tive bacteria has been extensively studied, little is known concern- Phagocytosis assay was undertaken essentially as described by Lu et al. ing Gram-positives, with most work centered on Staphylococcus (20). In brief, log phase bacteria were pelleted and resuspended in PBS and Downloaded from labeled with FITC (1 mg/ml) for1hat37°C. Bacteria were then washed aureus which requires oxidative mechanisms. As Spn is resistant to four times with PBS to remove unbound FITC. OPH assays were then killing by NETs, we were interested in investigating the antimi- undertaken as described above, and after 20 min reactions were placed on crobial agent(s) responsible for killing Spn following phagocytosis ice and analyzed by flow cytometry. Trypan blue (0.2%) was added to (13). This study used a complement-dependent opsonophagocytic quench signal from extracellular bacteria. Ten thousand events were col- lected per sample, and a single gate was used to exclude cell debris and free assay using human neutrophils to explore this further. bacteria using FlowJo software (Tree Star). http://www.jimmunol.org/ Materials and Methods Determination of serine protease activities Bacterial strains and culture conditions Protease activity assays of elastase, cathepsin G, and proteinase 3 were conducted using specific substrates as previously described; elastase, 0.85 Spn strains were grown in tryptic soy broth (TS) or on tryptic soy agar mM MeOSuc-Ala-Ala-Pro-Val-pNA (Calbiochem) (21); cathepsin G, 0.1 plates supplemented with catalase (5,000 U/plate; Worthington). Agar mM Suc-Ala-Ala-Pro-Phe-pNA (Elastin Products Company) (21, 22); and plates were grown at 37°C in 5% CO . Broth cultures were grown at 37°C 2 proteinase 3, 0.25 mM Boc-Ala-Ala-Nva-SBzl with 0.1 mM 5,5Ј-Dithio- without agitation. Pneumococci used were 1121 (Type 23F capsule) (14), bis(2-nitro-benzoic acid) (Sigma-Aldrich) (Elastin Products Company) T4 (Type 4 capsule) (15), and 6A (Type 6A capsule) (16). Staphylococcus (23). Following incubation of neutrophils with inhibitors, neutrophils were aureus strain 8325–4 (17) was grown on brain heart infusion broth or brain pelleted, washed, and lysed in Triton X-100 (0.01%). Substrates were heart infusion agar. Broth cultures were grown at 37°C with agitation. Agar by guest on October 2, 2021 added to this and the plate was incubated in the dark for 45 min at 37°C plates were grown at 37°C. (the length of the OPH assays). The OD410 of each reaction was read on Reagents Bio-Rad microplate reader model 680. Neutrophils were treated with inhibitors for 30 min at 37°C at the following Neutrophil degranulation concentrations: 20 ␮M cytochalasin D (Sigma-Aldrich), 20 U/ml bovine To degranulate neutrophils, cells were isolated and resuspended in RPMI pancreatic DNase (Sigma-Aldrich), 10 ␮M diphenyleneiodonium (DPI) 1640 with 25 mM HEPES (pH 7.2) at a concentration of 2 ϫ 107/ml. To (Sigma-Aldrich), 1 ϫ Protease Inhibitor (PI) Cocktail Set V, EDTA-Free degranulate primary granules alone, neutrophils were incubated with PMA (Calbiochem), 500 ␮M 4-(2-Aminoethyl)benzenesulfonylfluoride, HCl (10 ng/ml; Sigma-Aldrich) for 30 min at 37°C (24). To degranulate pri- (AEBSF) (Calbiochem), 150 nM aprotinin (Sigma-Aldrich), 1 ␮M trans-Ep- mary and azurophilic granules, neutrophils were incubated first with cy- oxysuccinyl-L-leucylamido(4-guanidino)butane (E-64) (Sigma-Aldrich), 1 tochalasin B (cytB) (5 ␮g/ml; Sigma-Aldrich) for 5 min at 37°C, followed ␮M leupeptin hemisulfate (Sigma-Aldrich), 1 mM diisopropylfluorophosphate Ϫ by incubation with fMLP (10 7 M; Sigma-Aldrich) for a further 30 min (DFP) (Sigma-Aldrich), 200 ␮M chymostatin (Sigma-Aldrich), and Elastase (25). In both cases, neutrophils were then pelleted and then the supernatant Inhibitor IV (HNE IV) 100 ␮M (Calbiochem). was taken and used in killing assays. Bacteria were grown to mid-log phase 5 Isolation of human neutrophils and ϳ10 6A were diluted in RPMI 1640 with 25 mM HEPES (pH 7.2). Fifty microliters 6A was then added to 50 ␮l of degranulated neutrophil Human neutrophils were isolated from heparinized whole blood from supernatant for1hat37°C, following which viable counts of Spn were healthy donors by gradient centrifugation using Polymorphprep according determined. to manufacturer’s instructions (Axis Shield). The neutrophil-enriched frac- Degranulated neutrophil supernatants were separated by SDS-PAGE on tion was collected, RBC lysed, and washed with Hank’s buffer without a 10% Tris-HCl gel (Bio-Rad), and proteins were transferred to polyvi- Ca2ϩ or Mg2ϩ (Invitrogen) with 0.1% gelatin (–ϩ solution). Cells were nylidene difluoride transfer membrane (Thermo Scientific). After transfer, counted using trypan blue staining and adjusted to a density of 7 ϫ 106 the membrane was probed with polyclonal rabbit antisera against elastase cells/ml in Hank’s buffer with Ca2ϩ and Mg2ϩ (Life Technologies), and (Calbiochem) and a polyclonal mouse antisera against lactoferrin (Sig- 0.1% gelatin (ϩϩϩ solution). ma-Aldrich), and then reacted with IgG alkaline phosphatase conjugate (BioRad). Membrane was subsequently developed using NBT/BCIP Isolation of murine neutrophils (Sigma-Aldrich). Murine neutrophils were isolated as previously described (18). In brief, In vitro microbicidal assays phagocytes were obtained by lavage of the peritoneal cavity (8 ml/animal with PBS containing 20 mM EDTA) of mice treated 24 h and again 2 h The in vitro bactericidal activities of elastase and cathepsin G was deter- before cell harvest by intraperitoneal administration of 10% casein in PBS mined essentially as described by Miyasaki and Bodeau (1991) (26). (1 ml/dose). Cells collected from the peritoneal cavity lavage were en- In brief, Spn were grown to mid-log phase, washed twice in PBS and then riched for neutrophils using separation by a Ficoll density gradient cen- ϳ107/ml Spn were incubated in the presence or absence of purified human trifugation according to the manufacturer’s protocol (MP Biomedicals, neutrophil elastase or cathepsin G at indicated concentrations (Elastin http://www.mpbio.com/). Neutrophils were collected and washed with 5 ml Products Company). Viable counts were determined after1hat37°C in a of –ϩ solution. Mice (C57BL6 wild type (wt) and Mac1Ϫ/Ϫ) were pur- total volume of 100 ␮l of 10 mmol/L sodium phosphate containing 1% TS. chased from Jackson ImmunoResearch Laboratories. Mac-1 (CD11b/ To inhibit their serine protease activity, proteases were preincubated with 2604 NEUTROPHIL SERINE PROTEASES KILL Streptococcus pneumoniae

FIGURE 2. Neutrophils do not kill Spn via ROS during complement- mediated opsonophagocytosis. Human neutrophils were incubated with rabbit serum opsonized Spn and Staphylcoccus aureus (Sa), and survival was assessed following a 45 min incubation. Percent bacterial survival was calculated based on viable counts (CFU/ml) relative to no neutrophil con- trols. Spn 6A was also opsonized with rabbit serum containing a 1/100 dilution of type 6 capsular Ab (6A ϩ Ab). Neutrophils were pretreated with NADPH and oxidative burst inhibitor, diphenyleneiodonium (DPI), before Downloaded from .(p Ͻ 0.001 ,ءءء) OPH assay. DPI only had an effect on the killing of Sa Values represent Ն three independent determinations in duplicate Ϯ SE.

ically dependent on MAC1 complement receptor (Fig. 1B). Thus, killing in this ex vivo assay was shown to require neutrophils, complement, and opsonophagocytosis. http://www.jimmunol.org/ As we were interested specifically in OPH killing, we wanted to ensure that NETs did not play a role in our assay. To confirm this, FIGURE 1. Neutrophil killing is dependent on phagocytosis, comple- neutrophils were pretreated with bovine pancreatic DNase before ment and MAC1. A, Human neutrophils were incubated with rabbit serum the OPH assay (28). As expected, this had no effect on killing (data opsonized Spn, and survival was assessed following a 45 min incubation. not shown). We were also interested in whether Spn was able to Percent survival of Spn was calculated based on viable counts (CFU/ml) survive intracellularly. For this, OPH reactions were treated with relative to no neutrophil controls. Preincubating neutrophils with actin- gentimicin for 20 min to kill any extracellular bacteria (29). Few, cytoskeleton inhibitor cytochalasin D (cytD) or heat inactivation (HI) of if any, Spn were recovered, suggesting that Spn cannot survive p Ͻ 0.001; by guest on October 2, 2021 ,ءءء) opsonizing serum abrogated killing of all three Spn strains p Ͻ 0.05 relative to corresponding wt strain). Values represent Ն two intracellularly in human neutrophils, and that the killing of Spn by ,ء independent determinations in duplicate Ϯ SE. B, Murine neutrophils neutrophils after uptake is an efficient process (data not shown). isolated from the peritoneal cavity of congenic wt or MAC 1 (CD11b/ When the respiratory oxidative burst was inhibited by preincu- CD18; CR3) deficient mice were incubated with mouse serum opso- bation of neutrophils with DPI, a known inhibitor of neutrophil nised Spn (1121), and survival assessed following a 45 min incubation NADPH oxidase, there was no significant effect on the ability of -p Ͻ 0.001).Values represent three independent determinations in the neutrophils to kill three different isolates of Spn (Fig. 2). How ,ء) duplicate Ϯ SE. ever, DPI pretreatment significantly abrogated killing of the ROS- sensitive bacterium Staphylcoccus aureus ( p Ͻ 0.001) (Fig. 2). Similar results were seen when killing was made more efficient inhibitor AEBSF for2hat37°C. Investigation of activity of each protease with the addition of specific capsular Abs to preopsonisation mix- was undertaken as described above. ture (Fig. 2), suggesting that neutrophil killing of Spn via both Fc and MAC 1 receptors were independent of the production of ROS. Statistical analysis Statistical differences between groups were tested using Student’s unpaired Neutrophil granule components kill Spn t test (GraphPad Prism 4; GraphPad Software). Other than ROS, neutrophils contain an array of antimicrobial products in cytoplasmic granules. These granules fuse with the Results phagolysosome, releasing a barrage of antimicrobial products to Phagocytic killing of S. pneumoniae does not require the kill the engulfed pathogen. Neutrophils can be artificially induced oxidative burst to degranulate, spilling out their granular products into the super- To investigate interactions between neutrophils and Spn, a rolling natant. Incubation of neutrophils with PMA results in degranula- tube ex vivo OPH killing assay was used (27). Killing of Spn was tion of specific granules, while cytB/fMLP degranulates both spe- dependent on phagocytosis as killing of three isolates of different cific and azurophilic granules (24, 25). Degranulated supernatants serotypes was blocked by incubation with cytochalasin D (an actin were run on a Western blot and probed with Abs against either cytoskeleton inhibitor which blocks phagocytosis) (Fig. 1A). Ac- elastase (azurophilic granule marker) or lactoferrin (specific gran- tive complement was also required, as heat inactivation at 56°C for ule marker). As expected, PMA supernatants showed increased 30 min completely abolished killing (Fig. 1A). Furthermore, while levels of lactoferrin and no detectable elastase compared with su- murine neutrophils isolated from the peritoneal cavity readily pernatants from undegranulated neutrophils (Fig. 3A). Degranula- killed Spn, neutrophils from MAC1Ϫ/Ϫ (CR3; CD11b/CD18) de- tion induced by cytB/fMLP resulted in similar levels of lactoferrin ficient mice did not, providing further evidence that this assay as PMA, but also contained elastase, indicating that both types of required active phagocytosis, and that bacterial uptake was specif- granules had been degranulated (Fig. 3A). Furthermore, while The Journal of Immunology 2605

FIGURE 3. Azurophilic granular components kill Spn. A, Supernatants of 2 ϫ 106 un- (1), PMA- (2), and cytB/fMLP- (3) degranulated neutrophils were separated by SDS-PAGE, transferred to polyvinylidene difluoride Downloaded from transfer membrane and then probed with Abs against elastase or lactoferrin as markers for azurophilic and specific granules, respectively. Sizes are shown in kDa. B, Supernatants from the 107 neutrophils degranulated by the method indicated were incubated with 105 Spn for 1 h. The effects of FIGURE 4. Proteases contribute to neutrophil-mediated killing of Spn. supernatant on survival of strain 6A is shown relative to incubation of Spn Human neutrophils were incubated with rabbit serum opsonized Spn and in supernatant from undegranulated neutrophils. Only cytB/fMLP- survival was assessed following a 45 min incubation. Percent Spn survival http://www.jimmunol.org/ p Ͻ 0.001). Values represent Ն ,ءءء) degranulated supernatants killed Spn was calculated based on viable counts (CFU/ml) relative to no neutrophil three independent determinations in duplicate Ϯ SE. controls. A, Before OPH assays, neutrophils were incubated with PI Cock- tail V for 30 min. Each Spn strain tested showed a significant increase in Ͻ ءء Ͻ ء there was little elastase and cathepsin G activity in supernatants survival when incubated with PI-inhibited neutrophils ( , p 0.05; , p -p Ͻ 0.001). Values represent Ն three independent determina ,ءءء ;0.01 from undegranulated and PMA-degranulated supernatants, activity tions in duplicate Ϯ SE. B, Survival of intracellular 6A was based on of these proteases increased 50- and 5-fold in cytB/fMLP-degranu- bacterial counts following treatment of OPH reactions with gentimicin sul- lated neutrophils, respectively (data not shown). fate to kill extracellular bacteria. Before OPH assays, neutrophils were Following degranulation, neutrophils were pelleted and super- ϫ 6

incubated with PI for 30 min. Results are presented as CFU per 1.4 10 by guest on October 2, 2021 p Ͻ 0.001). Values represent three independent ,ءءء) natants from the equivalent of 107 neutrophils were added to 105 neutrophil Ϯ SE CFU of Spn 6A. Although supernatant from PMA-degranulated determinations in duplicate Ϯ SE. neutrophils did not affect Spn viability compared with an unde- granulated control, supernatants from cytB/fMLP treated neutro- phils significantly decreased Spn survival (Fig. 3B). This suggested not shown). Additionally, killing was abolished when supernatants that neutrophil granular products, in particular from the azurophilic from cytB/fMLP degranulated neutrophils were inhibited with granules, kill Spn. AEBSF, suggesting that the inhibitor was having a direct effect on essential antimicrobial products from neutrophil granules (Fig. Proteases contribute to killing of Spn 5B). To ensure that the observed differences seen were the result of One subset of antimicrobial products from the azurophilic granules the specific inhibition of serine proteases, another serine protease are proteases, which degrade bacterial proteins. Thus, we at- inhibitor was used. DFP directly inhibits neutrophil serine pro- tempted to inhibit neutrophil proteases in ex vivo OPH assays teases without having adverse effects on other neutrophil functions using PI Cocktail Set V (Calbiochem). When neutrophils were such as the oxidative burst and phagocytosis (30–32). Neutrophils preincubated with this mixture for 30 min, survival of Spn was preincubated with DFP also showed reduced ability to kill Spn significantly enhanced (Fig. 4A). This was evident in three isolates (Fig. 5A). The inhibition of neutrophil killing with DFP and of different capsular types. The PI mixture had no effect on either AEBSF was similar whether neutrophils were washed to remove neutrophil viability, as assessed by trypan blue staining, or on inhibitor or the inhibitor remained during the OPH assay. Addi- growth of Spn (data not shown). When gentimicin was added fol- tionally, when Spn was pretreated with AEBSF and then washed to lowing the OPH reaction, intracellular survival of Spn strain 6A remove the inhibitor, increased survival of Spn was not seen (data was significantly increased in neutrophils treated with the mixture not shown), suggestive that inhibition of killing was the result of (Fig. 4B). Thus, the protease inhibitor enhanced the ability of Spn inhibition of neutrophil, and not Spn, serine proteases. to survive inside the neutrophil following phagocytosis. To confirm that these compounds inhibited neutrophil serine To elucidate which component of the mixture was responsible proteases, activity of three major neutrophil serine proteases, elas- for this effect, OPH assays were undertaken with each of the in- tase, cathepsin G, and proteinase 3 was assayed. Neutrophils were dividual inhibitors. Of the four components of the PI Cocktail only treated with inhibitors, washed and then lysed in Triton X-100 one, AEBSF, resulted in increased survival of Spn (Fig. 5A). such that intracellular activity was investigated. To this a substrate AEBSF is a general serine protease inhibitor previously shown to specific for each protease was added as indicated in Materials and inhibit oxidative burst and the shedding of neutrophil activation Methods. Although other inhibitors such as DPI, and broad spec- marker CD43 (30). AEBSF and the PI mixture did not affect bac- trum cysteine protease E64, had no affect on activity of the pro- terial uptake during phagocytosis as seen by flow cytometry (data teases as expected, addition of the PI mixture, AEBSF, or DFP, 2606 NEUTROPHIL SERINE PROTEASES KILL Streptococcus pneumoniae Downloaded from http://www.jimmunol.org/

FIGURE 5. Inhibition of neutrophil serine proteases inhibits neutrophil FIGURE 6. Purified human neutrophil elastase and cathepsin G kill Spn 7 by guest on October 2, 2021 killing. A, Human neutrophils were incubated with serum opsonized Spn in vitro. In brief, 10 of Spn strain 6A was incubated in PBS with 1% TS and survival was assessed following a 45-min incubation. Percent survival with either elastase (A) or cathepsin G (B) at indicated concentration for 1 h was calculated based on viable counts (CFU/ml) relative to no neutrophil and then plated to determine CFU. Data is presented as survival relative to ,p Ͻ 0.01). C ,ءء ;p Ͻ 0.001 ,ءءء) .controls. Neutrophils were preincubated with protease inhibitors aprotinin, incubation in PBS with 1% TS alone ␮ ␮ leupeptin, trans-Epoxysuccinyl-L-leucylamido(4-guanidino)butane (E64), Before the killing assay 3.39 M of elastase and 1 M of cathepsin G was Aminoethyl)benzenesulfonylfluoride, HCl (AEBSF), and diisopropylflu- incubated with AEBSF for2hat37°C to inhibit serine protease activity. p Ͻ 0.001). Values AEBSF significantly abrogated killing by both elastase and cathepsin G ,ءءء) orophosphate (DFP) for 30 min before OPH assay p Ͻ 0.001). Values represent Ն two independent determinations in ,ءءء) /represent Ն three independent determinations in duplicate Ϯ SE. B, cytB fMLP degranulated neutrophil supernatants were incubated with AEBSF duplicate Ϯ SE. before incubation with 105 6A. Survival was relative to incubation with p Ͻ 0.001). Values represent Ն ,ءءء) undegranulated control supernatant three independent determinations in duplicate Ϯ SE. C, Proteolytic activity says were undertaken. Addition of either elastase or cathepsin G ϳ ␮ of elastase, cathepsin G and proteinase 3 were determined following in- killed Spn strain 6A with an IC50 of 3.39 and 0.75 M, respec- cubation of neutrophils with indicated inhibitors. Neutrophils were then tively (Fig. 6, A and B). Furthermore, these proteases were able to pelleted, washed, and lysed in Triton X-100 to investigate intracellular kill another Spn strain, T4, at even lower IC50s (data not shown). proteolytic activity. Specific substrates for each serine protease was then Killing of bacteria via these serine proteases can be both depen- added, and activity was assayed after 45 min. Percent protease activity was dent and independent of proteolysis. To investigate their mecha- determined relative to the untreated control. Values represent Ն two in- dependent determinations in duplicate Ϯ SE. nism of action against Spn, elastase and cathepsin G were prein- cubated with the serine protease inhibitor AEBSF for 2 h which reduced levels of proteolysis to ϳ15 and 5% for elastase and ca- abrogated activity levels of all three proteases (Fig. 5C). Thus, thepsin G, respectively. This treatment significantly inhibited their these inhibitors affected neutrophil serine protease activity as well antimicrobial activities (Fig. 6C). Thus, elastase and cathepsin G as abrogating neutrophil killing of Spn. kill Spn via their serine protease activity. Furthermore, this would indicate that the inhibitory effect of serine protease inhibitors seen Elastase and cathepsin G kill Spn in the OPH assays was the result of the direct inhibition of neu- Numerous functions have been attributed to elastase, cathepsin G trophil serine protease antimicrobial activity. and proteinase 3, including both cell signaling and antimicrobial affects. In particular, much work has focused on the antimicrobial Inhibition of individual neutrophil serine proteases does not properties of elastase and cathepsin G. To understand whether in- inhibit killing of Spn creased Spn survival was the result of direct inhibition of antimi- To investigate whether one or multiple neutrophil serine proteases crobial activity, in vitro elastase and cathepsin G microbicidal as- are required for killing, we made use of a commercial intracellular The Journal of Immunology 2607

as many other Gram-positive pathogens, has to date remained un- clear. In this study, we identified neutrophil serine proteases as the mediators responsible for the complement-dependent OPH killing of Spn by human neutrophils. OPH is crucial for the eradication of Spn by the host. Indeed, the polysaccharide capsule is the patho- gen’s major virulence factor due to its ability to act as an antiph- agocytic factor (34). Unencapsulated Spn are almost completely avirulent as they are taken up and readily killed by phagocytes. Once taken up into the phagolysosome, killing is an efficient pro- cess, as only limited numbers of Spn were detected intracellularly in our assays. Although numerous antimicrobial molecules present within neutrophil granules can kill Spn in in vitro assays, our study identifies a factor that can specifically kill Spn inside the phagoly- sosome (8, 11, 35). The neutrophil oxidative burst was long thought of as the pri- mary antimicrobial mechanism of these immune cells. However, in recent times this has been brought into question (36, 37). When we inhibited the neutrophil oxidative burst with a specific NADPH inhibitor, neutrophil killing of Spn was unaffected. However, kill- ing of another Gram-positive pathogen, Staphylococcus aureus, Downloaded from was significantly inhibited. Spn is killed normally by neutrophils from CGD patients, whereas killing of Staphylococcus aureus is abrogated (5). Additionally, like other catalase-negative patho- gens, Spn is not found in increased rates in CGD individuals. FIGURE 7. Inhibition of individual serine proteases does not inhibit Taken together, these results suggest that the oxidative burst is not A neutrophil killing. , Proteolytic activity of elastase, cathepsin G and pro- http://www.jimmunol.org/ teinase 3 were determined following incubation of neutrophils with indi- important for neutrophil killing of this pathogen, and also show cated inhibitors. Neutrophils were then pelleted, washed, and lysed in Tri- that Gram-positive pathogens are not killed by a common mech- ton X-100 to investigate intracellular proteolytic activity. Specific anism. Recently, Selva et al. (38) suggested that S. aureus was substrates for each serine protease were then added, and activity was as- susceptible to hydrogen peroxide (one antimicrobial product of sayed after 45 min. Percent protease activity was determined relative to the ROS) due to the induction of the SOS response in lysogenic untreated control. Values represent Ն two independent determinations Ϯ strains. Furthermore, they showed that no such response occurs in SE. B, Human neutrophils were incubated with rabbit serum opsonized Spn lysogenic Spn, which may be intrinsically resistant to the effects of and survival was assessed following a 45 min incubation. Percent survival hydrogen peroxide. Thus, this may provide one explanation why S. was calculated based on viable counts (CFU/ml) relative to no neutrophil aureus is susceptible to ROS while Spn is not. controls. Neutrophils were preincubated with HNE IV, chymostatin by guest on October 2, 2021 Our results using supernatants from degranulated neutrophils (CHYM), both together or serine protease inhibtor AEBSF for 30 min p Ͻ 0.001 relative to control). Values represent Ն indicated that the particular microbicide(s) which kill Spn are ,ءءء) before OPH assay two independent determinations in duplicate Ϯ SE. present in the azurophilic granule. The azurophil granules contain myeloperoxidase, bactericidal permeability-increasing protein, ␣ inhibitor of human neutrophil elastase, HNE IV, also known as defensin human neutrophil proteins (HNP), and a family of serine ONO-5046 or sivelestat. This inhibitor suppressed elastase activity proteases. Recently, Beiter et al. (13) suggested that HNPs, found to similar levels as seen when neutrophils were preincubated with in azurophilic granules, were the bactericidal agents from neutro- AEBSF, while both proteinase 3 and cathepsin G levels remained phils responsible for killing Spn. However, mouse neutrophils, largely unaffected (Fig. 7A). When HNE IV-treated neutrophils which lack HNPs, still efficiently kill Spn (see Fig. 1B), suggesting were used in ex vivo OPH assays, no significant effect on killing other agents are primarily responsible. Additionally, this paper fo- of Spn was seen (Fig. 7B). Chymostatin is an inhibitor of many cused on killing by HNPs in vitro, and not within the specific proteases including chymostrypsin-like serine proteases and lyso- conditions of the phagolysosome, where the majority of killing of somal cysteine proteases. Chymostatin inhibited cathepsin G ac- Spn is likely to occur. tivity while elastase and proteinase 3 remained at control levels Other than HNPs, the three serine proteases elastase, cathepsin (Fig. 7A). However, this inhibition also had no effect on the ability G and proteinase 3 are perhaps the major antimicrobial compo- of neutrophils to kill Spn (Fig. 7B). Thus, inhibition of either elas- nents, present in neutrophils in concentrations in the mM range tase or cathepsin G alone does not alter a neutrophils capacity to (33, 39). All have direct bactericidal properties when tested in vitro kill Spn. (40–43). Additionally, studies using mice deficient in elastase and When HNE IV and chymostatin were used in combination, pro- cathepsin G have shown that they are important for the in vivo teinase 3 activity was left at control levels (Fig. 7B). This inhibi- clearance of a number of bacteria including Klebisella pneu- tion of elastase and cathepsin G together resulted in a small but moniae, and S. aureus (36, 44). When enzymatic activity of all significant increase in survival of Spn. This suggests that these three proteases were inhibited via incubation of neutrophils with serine proteases contribute to neutrophil killing of Spn. The re- protease inhibitors, neutrophils were unable to kill Spn, suggestive maining activity of proteinase 3, the most abundant of the neutro- that nonenzymatic killing mechanisms against Spn are of lesser phil serine proteases in the neutrophil, is likely responsible for the importance. Additionally, inhibition of serine proteases activity in remaining capacity of the neutrophils to kill Spn (33). degranulated neutrophil supernatants abolished killing, suggesting that serine proteases were directly responsible for the antimicrobial Discussion activity. Neutrophil serine proteases also play a role in that acti- Neutrophils are critical immune cells in the clearance of S. pneu- vation of other neutrophil antimicrobials such as LL37 (45), how- moniae. However, the mechanism by which they kill Spn, as well ever, purified elastase and cathepsin G were able to kill Spn at 2608 NEUTROPHIL SERINE PROTEASES KILL Streptococcus pneumoniae physiological levels in vitro, similar to those seen for other bacteria 4. Matthias, K. A., A. M. Roche, A. J. Standish, M. Shchepetov, and J. N. Weiser. (42, 44). Additionally, this killing was dependent on serine pro- 2008. Neutrophil-toxin interactions promote antigen delivery and mucosal clear- ance of Streptococcus pneumoniae. J. Immunol. 180: 6246–6254. tease activity. Furthermore, the kinetics of neutrophil killing was 5. Kaplan, E. L., T. Laxdal, and P. G. Quie. 1968. Studies of polymorphonuclear similar to that for purified serine proteases (data not shown). Thus, leukocytes from patients with chronic granulomatous disease of childhood: bac- these results provide strong evidence that serine proteases are re- tericidal capacity for streptococci. Pediatrics 41: 591–599. 6. Schaper, M., S. L. Leib, D. N. Meli, R. P. Brandes, M. G. Ta¨uber, and S. Christen. sponsible for the neutrophil-mediated OPH killing of Spn. 2003. Differential effect of p47phox and gp91phox deficiency on the course of When either elastase or cathepsin G were inhibited alone, no pneumococcal meningitis. Infect. Immun. 71: 4087–4092. 7. Marriott, H. M., L. E. Jackson, T. S. Wilkinson, A. J. Simpson, T. J. Mitchell, loss of killing was seen. Two recent studies investigating the in D. J. Buttle, S. S. Cross, P. G. Ince, P. G. Hellewell, M. K. B. Whyte, and vivo effects of the elastase inhibitor used in this study found that D. H. Dockrell. 2008. Reactive oxygen species regulate neutrophil recruitment such inhibition did not render mice or hamsters more susceptible to and survival in pneumococcal pneumonia. Am. J. Respir. Crit. Care Med. 177: 887–895. Spn infection (46, 47). Thus, this suggests that loss of a single 8. Shaper, M., S. K. Hollingshead, W. H. J. Benjamin, and D. E. Briles. 2004. PspA protease both in vivo and in our ex vivo OPH assays may not be protects Streptococcus pneumoniae from killing by apolactoferrin, and antibody sufficient to inhibit neutrophil killing of Spn. When both elastase to PspA enhances killing of pneumococci by apolactoferrin [corrected]. Infect. Immun. 72: 5031–5040. and cathepsin G were inhibited concurrently, only a small but sig- 9. Ganz, T., J. A. Metcalf, J. I. Gallin, L. A. Boxer, and R. I. Lehrer. 1988. Micro- nificant abrogation of killing was evident. Proteinase 3 is the most bicidal/cytotoxic proteins of neutrophils are deficient in two disorders: Chediak- abundant neutrophil serine protease, with mean concentrations in Higashi syndrome and “specific” granule deficiency. J. Clin. Invest. 82: 552–556. 10. Root, R. K., A. S. Rosenthal, and D. J. Balestra. 1972. Abnormal bactericidal, azurophillic granules of 13.4 mM (33). Activity assays showed metabolic, and lysosomal functions of Chediak-Higashi Syndrome leukocytes. that preincubation with chymostatin and HNE IV did not inhibit J. Clin. Invest. 51: 649–665. proteinase 3 activity. Thus, we hypothesize that this activity may 11. Beiter, K., F. Wartha, R. Hurwitz, S. Normark, A. Zychlinsky, and

B. Henriques-Normark. 2008. The capsule sensitizes Streptococcus pneumoniae Downloaded from be responsible for the remaining bactericidal capacity when elas- to ␣-defensins human neutrophil proteins 1 to 3. Infect. Immun. 76: 3710–3716. tase and cathepsin G are inhibited. This suggests a redundancy in 12. Srivastava, A., H. Casey, N. Johnson, O. Levy, and R. Malley. 2007. Recombi- neutrophil killing mechanisms, with all three serine proteases be- nant bactericidal/permeability-increasing protein rBPI21 protects against pneu- mococcal disease. Infect. Immun. 75: 342–349. ing sufficient, but none essential to kill Spn. 13. Beiter, K., F. Wartha, B. Albiger, S. Normark, A. Zychlinsky, and The differences seen between Spn and Staphylococcus aureus B. Henriques-Normark. 2006. An endonuclease allows Streptococcus pneu- are intriguing for a number of reasons. Reeves et al. (36) suggested moniae to escape from neutrophil extracellular traps. Curr. Biol. 16: 401–407. 14. McCool, T. L., T. R. Cate, G. Moy, and J. N. Weiser. 2002. The immune response http://www.jimmunol.org/ that the major contribution of ROS was not directly microbicidal, to pneumococcal proteins during experimental human carriage. J. Exp. Med. 195: but rather was its ability to activate proteases in the phagolyso- 359–365. 15. Tettelin, H., K. E. Nelson, I. T. Paulsen, J. A. Eisen, T. D. Read, S. Peterson, some. The oxidative burst contributes to an increase in phagoly- J. Heidelberg, R. T. DeBoy, D. H. Haft, R. J. Dodson, et al. 2001. Complete sosomal pH, which in turn provides an optimal environment for genome sequence of a virulent isolate of Streptococcus pneumoniae. Science 293: maximal activity of proteases, and also allows their release from 498–506. 16. King, S. J., K. R. Hippe, and J. N. Weiser. 2006. Deglycosylation of human the anionic proteoglycan matrix. Treatment of neutrophils with glycoconjugates by the sequential activities of exoglycosidases expressed by DPI not only inhibits the production of ROS, but also results in the Streptococcus pneumoniae. Mol. Microbiol. 59: 961–974. loss of this temporary increase in pH. In our study, while Spn is 17. Novick, R. 1967. Properties of a cryptic high-frequency transducing phage in Staphylococcus aureus. Virology 33: 155–166. susceptible to serine proteases, the pathogen is still killed in the 18. Luo, Y., and M. Dorf. 2001. Isolation of mouse neutrophils. Curr. Prot. Immunol. by guest on October 2, 2021 absence of ROS. Thus, either serine proteases are able to kill Spn 3.20.1–3.20.6. at lower pHs as is the case for Neisseria gonorrhoeae (48), or other 19. Coxon, A., P. Rieu, F. J. Barkalow, S. Askari, A. H. Sharpe, U. H. von Andrian, M. A. Arnaout, and T. N. Mayadas. 1996. A novel role for the ␤ 2 integrin mechanisms of serine protease activation are present in the neu- CD11b/CD18 in neutrophil apoptosis: a homeostatic mechanism in inflammation. trophil during killing of Spn. S. aureus is thought to be killed by Immunity 5: 653–666. cathepsin G via a mechanism independent of its protease activity, 20. Lu, L., Z. Ma, T. S. Jokiranta, A. R. Whitney, F. R. DeLeo, and J. Zhang. 2008. Species-specific interaction of Streptococcus pneumoniae with human comple- whereas we have shown the proteolysis is critical for both cathep- ment factor H. J. Immunol. 181: 7138–7146. sin G and elastase to kill Spn (49). Additionally, elastase does not 21. Nakajima, K., J. C. Powers, B. M. Ashe, and M. Zimmerman. 1979. Mapping the kill S. aureus (44). This may suggest that Spn is simply more extended substrate binding site of cathepsin G and human leukocyte elastase. Studies with peptide substrates related to the ␣ 1-protease inhibitor reactive site. sensitive to the presence of all three neutrophil serine proteases in J. Biol. Chem. 254: 4027–4032. the phagolysosome. 22. Blocher, M., P. Walde, and I. J. Dunn. 1999. Modeling of enzymatic reactions in ␣ This study has identified serine proteases as the primary killing vesicles: the case of -chymotrypsin. Biotechnol. Bioeng. 62: 36–43. 23. Kam, C. M., J. E. Kerrigan, K. M. Dolman, R. Goldschmeding, mechanism during OPH killing of Spn by human neutrophils. Al- A. E. Von dem Borne, and J. C. Powers. 1992. Substrate and inhibitor studies on though other studies have shown Spn to be susceptible to neutro- proteinase 3. FEBS Lett. 297: 119–123. phil antimicrobial molecules in vitro, this is the first study to show 24. Zarember, K. A., J. A. Sugui, Y. C. Chang, K. J. Kwon-Chung, and J. I. Gallin. 2007. Human polymorphonuclear leukocytes inhibit Aspergillus fumigatus conid- an effect within the conditions of the human neutrophil phagoly- ial growth by lactoferrin-mediated iron depletion. J. Immunol. 178: 6367–6373. sosome during a step critical to the human defense against this 25. Bentwood, B. J., and P. M. Henson. 1980. The sequential release of granule major pathogen. constitutents from human neutrophils. J. Immunol. 124: 855–862. 26. Miyasaki, K. T., and A. L. Bodeau. 1991. In vitro killing of oral Capnocytophaga by granule fractions of human neutrophils is associated with cathepsin G activity. J. Clin. Invest. 87: 1585–1593. Disclosures 27. Davis, K. M., H. T. Akinbi, A. J. Standish, and J. N. Weiser. 2008. Resistance to The authors have no financial conflict of interest. mucosal lysozyme compensates for the fitness deficit of peptidoglycan modifi- cations by Streptococcus pneumoniae. PLoS Pathog. 4: e1000241. 28. Wartha, F., K. Beiter, B. Albiger, J. Fernebro, A. Zychlinsky, S. Normark, and B. Henriques-Normark. 2007. Capsule and D-alanylated lipoteichoic acids pro- References tect Streptococcus pneumoniae against neutrophil extracellular traps. Cell Micro- 1. Brinkmann, V., U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D. S. Weiss, biol. 9: 1162–1171. Y. Weinrauch, and A. Zychlinsky. 2004. Neutrophil extracellular traps kill bac- 29. Lysenko, E. S., T. B. Clarke, M. Shchepetov, A. J. Ratner, D. I. Roper, teria. Science 303: 1532–1535. C. G. Dowson, and J. N. Weiser. 2007. Nod1 signaling overcomes resistance of 2. Dallaire, F., N. Ouellet, Y. Bergeron, V. Turmel, M. C. Gauthier, M. Simard, and S. pneumoniae to opsonophagocytic killing. PLoS Pathog. 3: e118. M. G. Bergeron. 2001. Microbiological and inflammatory factors associated with 30. Remold-O’Donnell, E., and D. Parent. 1995. Downregulation of neutrophil CD43 the development of pneumococcal pneumonia. J. Infect. Dis. 184: 292–300. by opsonized zymosan. Blood 85: 337–342. 3. van Rossum, A. M. C., E. S. Lysenko, and J. N. Weiser. 2005. Host and bacterial 31. Amrein, P. C., and T. P. Stossel. 1980. Prevention of degradation of human factors contributing to the clearance of colonization by Streptococcus pneu- polymorphonuclear leukocyte proteins by diisopropylfluorophosphate. Blood 56: moniae in a murine model. Infect. Immun. 73: 7718–7726. 442–447. The Journal of Immunology 2609

32. Mark, D. E., K. G. Lazzari, and E. R. Simons. 1988. Are serine proteases in- 41. Mak, P., J. Pohl, A. Dubin, M. S. Reed, S. E. Bowers, M. T. Fallon, and volved in immune complex activation of neutrophils? J. Leukocyte Biol. 44: W. M. Shafer. 2003. The increased bactericidal activity of a fatty acid-modified 441–447. synthetic antimicrobial peptide of human cathepsin G correlates with its enhanced 33. Campbell, E. J., M. A. Campbell, and C. A. Owen. 2000. Bioactive proteinase 3 capacity to interact with model membranes. Int. J. Antimicrob. Agents 21: 13–19. on the cell surface of human neutrophils: quantification, catalytic activity, and 42. Hirche, T. O., R. Benabid, G. Deslee, S. Gangloff, S. Achilefu, M. Guenounou, susceptibility to inhibition. J. Immunol. 165: 3366–3374. F. Lebargy, R. E. Hancock, and A. Belaaouaj. 2008. Neutrophil elastase mediates 34. Melin, M., H. Jarva, L. Siira, S. Meri, H. Ka¨yhty, and M. Va¨keva¨inen. 2009. innate host protection against Pseudomonas aeruginosa. J. Immunol. 181: Streptococcus pneumoniae capsular serotype 19F is more resistant to C3 depo- 4945–4954. sition and less sensitive to opsonophagocytosis than serotype 6B. Infect. Immun. 43. Belaaouaj, A., K. S. Kim, and S. D. Shapiro. 2000. Degradation of outer mem- 77: 676–684. brane protein A in Escherichia coli killing by neutrophil elastase. Science 289: 35. Braconier, J. H., and H. Odeberg. 1982. Granulocyte phagocytosis and killing 1185–1188. virulent and avirulent serotypes of Streptococcus pneumoniae. J. Lab. Clin. Med. 44. Belaaouaj, A., R. McCarthy, M. Baumann, Z. Gao, T. J. Ley, S. N. Abraham, and 100: 279–287. S. D. Shapiro. 1998. Mice lacking neutrophil elastase reveal impaired host de- 36. Reeves, E. P., H. Lu, H. L. Jacobs, C. G. M. Messina, S. Bolsover, G. Gabella, fense against gram negative bacterial sepsis. Nat. Med. 4: 615–618. E. O. Potma, A. Warley, J. Roes, and A. W. Segal. 2002. Killing activity of 45. Sørensen, O. E., P. Follin, A. H. Johnsen, J. Calafat, G. S. Tjabringa, neutrophils is mediated through activation of proteases by Kϩ flux. Nature 416: P. S. Hiemstra, and N. Borregaard. 2001. Human cathelicidin, hCAP-18, is pro- 291–297. cessed to the antimicrobial peptide LL-37 by extracellular cleavage with protein- 37. Mayer-Scholl, A., R. Hurwitz, V. Brinkmann, M. Schmid, P. Jungblut, ase 3. Blood 97: 3951–3959. Y. Weinrauch, and A. Zychlinsky. 2005. Human neutrophils kill Bacillus an- 46. Hagio, T., K. Kishikawa, K. Kawabata, S. Tasaka, S. Hashimoto, N. Hasegawa, thracis. PLoS Pathog. 1: e23. and A. Ishizaka. 2008. Inhibition of neutrophil elastase reduces lung injury and 38. Selva, L., D. Viana, G. Regev-Yochay, K. Trzcinski, J. M. Corpa, I. Lasa, bacterial count in hamsters. Pulm. Pharmacol. Ther. 21: 884–891. R. P. Novick, and J. R. Penade´s. 2009. Killing niche competitors by remote- 47. Yanagihara, K., Y. Fukuda, M. Seki, K. Izumikawa, Y. Miyazaki, Y. Hirakata, control bacteriophage induction. Proc. Natl. Acad. Sci. USA 106: 1234–1238. K. Tsukamoto, Y. Yamada, S. Kamhira, and S. Kohno. 2007. Effects of specific 39. Liou, T. G., and E. J. Campbell. 1995. Nonisotropic enzyme–inhibitor interac- neutrophil elastase inhibitor, sivelestat sodium hydrate, in murine model of severe tions: a novel nonoxidative mechanism for quantum proteolysis by human neu- pneumococcal pneumonia. Exp. Lung Res. 33: 71–80. trophils. Biochemistry 34: 16171–16177. 48. Shafer, W. M., V. C. Onunka, and L. E. Martin. 1986. Antigonococcal activity of 40. Gabay, J. E., R. W. Scott, D. Campanelli, J. Griffith, C. Wilde, M. N. Marra, human neutrophil cathepsin G. Infect. Immun. 54: 184–188. Downloaded from M. Seeger, and C. F. Nathan. 1989. Antibiotic proteins of human polymorpho- 49. Odeberg, H., and I. Olsson. 1975. Antibacterial activity of cationic proteins from nuclear leukocytes. Proc. Natl. Acad. Sci. USA 86: 5610–5614. human granulocytes. J. Clin. Invest. 56: 1118–1124. http://www.jimmunol.org/ by guest on October 2, 2021