Necrosis in Chronic Granulomatous Disease Induces Neutrophil

Home , Burkholderia cenocepacia, Hypochlorite

Burkholderia cenocepacia Induces Neutrophil Necrosis in Chronic Granulomatous Disease Johan Bylund, Paul A. Campsall, Rebecca C. Ma, Barbara-Ann D. Conway and David P. Speert This information is current as of October 10, 2021. J Immunol 2005; 174:3562-3569; ; doi: 10.4049/jimmunol.174.6.3562 http://www.jimmunol.org/content/174/6/3562 Downloaded from References This article cites 35 articles, 13 of which you can access for free at: http://www.jimmunol.org/content/174/6/3562.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 10, 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 © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Burkholderia cenocepacia Induces Neutrophil Necrosis in Chronic Granulomatous Disease1

Johan Bylund, Paul A. Campsall, Rebecca C. Ma, Barbara-Ann D. Conway, and David P. Speert2

Burkholderia cepacia complex is a life-threatening group of pathogens for patients with chronic granulomatous disease (CGD), whose phagocytes are unable to produce reactive oxygen species (ROS). Unlike other CGD pathogens, B. cepacia complex is particularly virulent, characteristically causing septicemia, and is the bacterial species responsible for most fatalities in these patients. We found that a nonmucoid Burkholderia cenocepacia (a predominant species in the B. cepacia complex) isolate was readily ingested by normal human neutrophils under nonopsonic conditions and promoted apoptosis in these cells. The proapoptotic effect was not due to secreted bacterial products, but was dependent on bacterial viability. Phagocytosis was associated with

a robust production of ROS, and the apoptotic neutrophils could be effectively cleared by monocyte-derived macrophages. The Downloaded from proapoptotic effect of B. cenocepacia was independent of ROS production because neutrophils from CGD patients were rendered apoptotic to a similar degree as control cells after challenge. More importantly, neutrophils from CGD patients, but not from normal individuals, were rendered necrotic after phagocytosis of B. cenocepacia. The extreme virulence of B. cepacia complex bacteria in CGD, but not in immunocompetent hosts, could be due to its necrotic potential in the absence of ROS. The Journal of Immunology, 2005, 174: 3562–3569. http://www.jimmunol.org/ hronic granulomatous disease (CGD)3 is a rare primary Their particular virulence in patients with CF has not been immunodeficiency resulting from genetic defects in com- elucidated. C ponents of the phagocyte NADPH-oxidase, rendering the B. cepacia complex infections in CGD patients are often re- patient’s phagocytes unable to produce the reactive oxygen species markably aggressive, and the majority of the fatalities are caused (ROS) needed for proper antimicrobial activity (1). These patients by bacteraemic spread, a condition extremely uncommon with suffer from recurrent infections with a distinct set of fungal and other bacterial species. Furthermore, B. cepacia complex bacteria bacterial species and often display various inflammatory compli- can cause septicemia in CF, a condition very rarely seen with the cations, including granuloma formation (2). The leading bacterial more common CF pathogen, Pseudomonas aeruginosa. This sug- cause of CGD fatalities are members of the Burkholderia cepacia gests that B. cepacia complex possesses virulence factors, making by guest on October 10, 2021 complex, a complex that includes nine distinct species (3). Burk- it unique among bacterial CGD and CF pathogens (4). In both holderia cenocepacia (formerly known as genomovar III) is a par- CGD and CF, B. cepacia complex bacteria most commonly infect ticularly problematic pathogen in patients with cystic fibrosis (CF) the respiratory tract, and the pathology is characterized by exces- (4), but no particular species from the complex appears to pre- sive inflammation with a massive infiltration of neutrophils. dominate in CGD. B. cepacia complex bacteria possess intrinsic Neutrophils are rapidly recruited to a site of infection, where resistance to many antibiotics and can be transmitted from patient they ingest and destroy the invading microorganisms using a com- to patient in CF, making infection with these bacteria a particularly bination of ROS and a variety of cationic peptides and proteolytic difficult clinical challenge. B. cepacia complex bacteria are highly enzymes. Normally, after phagocytosis, the neutrophils undergo resistant to cationic antibacterial peptides (5), crucial components apoptosis (also known as programmed cell death) and are disposed of the oxygen-independent killing machinery of phagocytes, ren- of by other cell types, most commonly macrophages (6). ROS are dering these pathogens especially well suited for infection in CGD. not only crucial bactericidal effectors, but they are also important signaling molecules and are partly responsible for initiating the apoptotic program (7, 8). Antioxidants or inhibitors of the neutro- Department of Pediatrics, University of British Columbia, British Columbia Research phil NADPH-oxidase inhibit apoptosis following phagocytosis of Institute for Children’s and Women’s Health, Vancouver, British Columbia, Canada a variety of bacterial species (7–10). In addition, neutrophils from Received for publication November 4, 2004. Accepted for publication January 4, 2005. CGD patients display delayed spontaneous apoptosis (11) as well as decreased apoptotic rates following ingestion of heat-killed 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 Staphylococcus aureus (9). Yamamoto et al. (9) concluded that with 18 U.S.C. Section 1734 solely to indicate this fact. even though CGD neutrophils exhibited lower apoptotic levels 1 This study was funded by grants from the Canadian Cystic Fibrosis Foundation (to than control cells, the ingestion of S. aureus still promoted apo- D.P.S. and J.B.), the Swedish Society for Medical Research (to J.B.), British Colum- bia Lung Association, and the Canadian Institutes of Health Research (to D.P.S.). ptosis in these cells, indicating that ROS-independent pathways also play a role in phagocytosis-mediated apoptosis. 2 Address correspondence and reprint requests to Dr. David P. Speert, Room 377, Research Centre, 950 West 28th Avenue, Vancouver, British Columbia, V5Z 4H4, Apoptosis and subsequent clearance represent important physi- Canada. E-mail address: [email protected] ological steps for terminating an inflammatory response. However, 3 Abbreviations used in this paper: CGD, chronic granulomatous disease; CF, cystic neutrophils can also undergo a less physiologic form of cell death: fibrosis; DAPI, 4Ј,6Ј-diamidino-2-phenylindole; DPI, diphenylene iodonium; EthD, ethidium homodimer 1; KRG, Krebs-Ringer phosphate buffer; LDH, lactate dehy- necrosis. Necrosis is a more pathological type of cell death in drogenase; MDM, monocyte-derived macrophage; ROS, reactive oxygen species. which the cell membrane loses its integrity, and leakage of cellular

contents occurs (12). If neutrophils die by necrosis, their toxic LPS purification and generation of bacterial supernatants contents may leak out and cause additional tissue damage and LPS was purified by suspending a freeze-dried bacterial pellet in an ex- perpetuation of the inflammatory reaction. Necrotic cell death (in- traction mixture consisting of phenol (90%):chloroform:petroleum spirit in stead of apoptosis) could result in chronic inflammation rather than the proportions 2:5:8. The mixture was stirred on ice for 10 min and cen- resolution of the inflammatory response (13). trifuged (10,000 ϫ g, for 15 min), after which the supernatant was filtered. We studied the interaction of human neutrophils with B. ceno- The filtrate was air dried and repurified (17), washed in ice-cold ethanol, and resuspended in endotoxin-free water plus 0.2% triethylamine after dry- cepacia and investigated how this interaction affected neutrophil ing. When cell-free bacterial supernatants were used, the bacteria were cell death. Furthermore, we evaluated whether ROS affected the grown in RPMI 1640 medium (without antibiotics) overnight and the su- outcome of neutrophil interaction with B. cenocepacia, thus pro- pernatants were harvested by sedimenting out the bacteria in a microcen- viding clues as to why these bacteria display such extraordinary trifuge (9000 ϫ g, for 2 min) and then filtered (0.2 ␮m) before use. virulence in CGD. Our data demonstrate that a clinical B. cenocepacia isolate was readily ingested nonopsonically by neutrophils Neutrophil NADPH-oxidase activity from normal individuals. Ingestion was accompanied by ROS pro- The NADPH-oxidase activity was determined using a luminol-ECL system duction, and the interaction resulted in the induction of neutrophil (18) with a Victor3 (PerkinElmer Life and Analytical Sciences) plate reader ␮ apoptosis. Challenge of neutrophils from CGD patients with this and disposable 96-well plates containing 220- l reaction mixtures. Each well contained 106 neutrophils, luminol (a cell-permeable chemilumines- isolate induced slightly higher levels of apoptosis than in normal cence substrate; 2 ϫ 10Ϫ5 M), and cell-impermeable scavengers superox- neutrophils, suggesting that ROS-independent pathways are oper- ide dismutase (20 U) and catalase (2000 U). The plates were equilibrated ative. More importantly, however, B. cenocepacia also induced in the Victor3 for 10 min at 37°C, after which the stimulus (30 ␮l) was significant levels of necrosis in CGD neutrophils, suggesting that added. The light emission was recorded continuously, and data are ex- Downloaded from ROS are crucial for controlling ingested B. cenocepacia and ab- pressed as counts per second. rogating cytotoxic/necrotic effects. In summary, these results could Bacterial challenge explain why certain B. cepacia complex bacteria are capable of causing such severe, necrotizing infections in CGD patients while Freshly isolated neutrophils were diluted to 107 cells/ml in KRG supple- posing little or no threat to immunocompetent hosts. mented with 0.3% BSA (to minimize cellular aggregation and adhesion to the tubes) and incubated at 37°C for 10 min in the presence or absence of Ϫ5

diphenylene iodonium (DPI; final concentration 10 M) before bacteria http://www.jimmunol.org/ Materials and Methods were added, giving a final volume of 1 ml and a multiplicity of infection Patients of 30 bacteria/neutrophil. Where indicated, cytochalasin D was included at a concentration of 10 ␮g/ml. All bacterial challenges took place in polypro- Peripheral blood was obtained from healthy volunteers and three CGD pylene tubes slowly tumbling at 37°C for 1 h, after which the cells were patients, two male gp91phox-deficient patients (X-linked CGD) and one fe- subjected to repeated washes in PBS (110 ϫ g, for 10 min, to remove male gp47phox-deficient patient (autosomal recessive CGD). None of the nonattached bacteria) and resuspended in 0.5 ml of RPMI 1640 medium patients was infected at the time of the experiments. All donors (healthy supplemented with 10% heat-inactivated FCS, 250 ␮g/ml meropenem Ϫ5 controls and CGD patients) provided informed consent, and the study was (RPMI 1640 medium), and DPI (10 M) when required. Samples were approved by the Clinical Research Ethics Board of the University of British withdrawn for cytospinning, and the tubes were incubated standing at 37°C Columbia. in a humidified CO2 incubator (5% CO2, 95% air). After 18–20 h of incubation, the supernatants were collected and the cells were washed in PBS by guest on October 10, 2021 Isolation of cells before FACS analysis. Human neutrophils were purified using dextran sedimentation and Ficoll- Lactate dehydrogenase (LDH) assay Paque gradient centrifugation (14). The cells were washed and resuspended (107/ml) in Krebs-Ringer phosphate buffer (KRG, pH 7.3) containing glu- Supernatants (300 ␮l) were spun in a microcentrifuge, and 250 ␮l of cell- cose (10 mM), Ca2ϩ (1 mM), and Mg2ϩ (1.5 mM), and stored on melting free supernatant was analyzed by a cytotoxicity detection kit (Roche Ap- ice until use. This protocol routinely produced a neutrophil population of plied Science) to quantify released levels of LDH, according to the man- ϳ95% purity, as judged by visual inspection of Giemsa-stained slides. ufacturer’s instructions. In parallel with the bacterially challenged samples, About 95% of the freshly isolated neutrophils were viable and negative for control samples including medium only (blank) and 2.5 ϫ 106 neutrophils both annexin V and ethidium homodimer 1 (EthD). Human monocyte- in 1% Triton X-100 (total) were also analyzed, and data are presented as derived macrophages (MDMs) were obtained as previously described (15), LDH released as percentage of total LDH after subtraction of blank sample except that cells were allowed to differentiate in Teflon pots with 100 ng/ml readings. M-CSF instead of autologous serum, and used for apoptotic neutrophil uptake assays on day 6. Flow cytometric measurements of neutrophil cell death ␮ Bacterial strains and growth Washed neutrophil cell pellets were resuspended in 300 l of annexin V-binding buffer containing 5 ␮l of Annexin VFITC (BD Biosciences) and B. cenocepacia, strain C8963, was a nonmucoid, genomovar III-A B. ce- 3 ␮M EthD (Molecular Probes). EthD is a cell-impermeable nucleic acid- pacia complex isolate from a CF patient. The mucoid variant, strain C9343, binding dye that only permeates leaky (necrotic) cell membranes and de- was a sequential isolate from the same CF patient, and both strains have tects necrotic cell death (19). The cells were incubated in the dark for 15 been previously described (16). Throughout the text, the term B. cenoce- min, and analysis was performed on Ͼ10,000 cells using a FACSCalibur pacia refers to the nonmucoid isolate C8963, and the mucoid variant is system and CellQuest software (BD Biosciences). Neutrophils were gated always referred to as mucoid B. cenocepacia. The isolates are deposited in to exclude cellular debris and unbound bacteria, the amount of annexin V the Canadian B. cepacia Complex Research and Referral Repository. The (FL1) and EthD (FL3) binding to the cells was measured, and the data were strains were stored at Ϫ80°C or maintained on Columbia agar containing analyzed using WinMDI 2.8 software, defining cells as viable (annexin 5% sheep blood (PML Microbiologicals) for a maximum of two passages. Vlow/EthDlow), apoptotic (annexin Vhigh/EthDlow), or necrotic (annexin Unless otherwise specified, bacteria were grown in Luria broth at 37°C, Vhigh/EthDhigh). 250 rpm overnight, washed, and cultured in KRG plus 0.3% BSA for 1 h 8 until reaching the desired OD (corresponding to ϳ10 CFU/ml). For some Neutrophil phagocytosis experiments, an aliquot (2 ml) was UV killed by incubation for 50 min in 24-well tissue culture plates placed on a UV table (UV Transilluminator; After cytospinning, slides were air dried overnight, fixed in methanol (5 Ultraviolet Products), equipped with four 15-W lamps. Viable counts con- min), Giemsa stained, and mounted. Slides were studied using oil immer- firmed that treatment for 50 min reduced viability of strain C8963 by sion and ϫ1000 magnification; Ͼ200 cells were scored as either positive Ͼ99.99%. The NADPH-oxidase assay was performed on bacteria that or negative for bacterial association, and the number of bacteria/cell was were concentrated using a microcentrifuge (9000 ϫ g, for 2 min) and counted for the first 50 positive cells. Phagocytosis is expressed as mean resuspended to a density of 109 CFU/ml in KRG (without BSA). number of bacteria (bound or ingested)/cell. 3564 NEUTROPHIL NECROSIS IN CGD

MDM uptake of apoptotic neutrophils Neutrophil apoptosis following phagocytosis Freshly prepared neutrophils were labeled using Vybrant CFDA SE cell After bacterial challenge, cells were washed and incubated in tracker kit (Molecular Probes), according to the manufacturer’s instruc- RPMI 1640 medium supplemented with bacteriostatic concentra- ␮ tions, using a dye concentration of 1 M. Fluorescent cells were washed in tions of meropenem. After 18 h of incubation, cells were washed PBS and either subjected to challenge with B. cenocepacia (as above) or resuspended in RPMI 1640 medium supplemented with 1 ␮g/ml FAS li- and stained with annexin V and the cell-impermeable nucleic acid- gand (mouse anti-human CD95; eBiosciences) and 1 ␮g/ml cross-linking binding dye EthD, to assess apoptosis and necrosis. Approximately Ab (anti-mouse IgM; eBiosciences) to induce apoptosis. After overnight 30–50% of untreated cells were annexin V-positive/EthD-nega- incubation, the apoptotic cells (levels of apoptosis were analyzed morpho- tive, representing spontaneous levels of apoptosis. A signiﬁcantly logically and were ϳ70% for both conditions in the experiment shown) were washed and resuspended in HBSS plus 0.1% gelatin. MDMs were washed and resuspended in HBSS plus 0.1% gelatin before mixing with apoptotic neutrophils in polypropylene tubes (105 MDMs and 106 neutrophils in 400 ␮l) in the presence or absence of cytochalasin D (5 ␮g/ml incubated with the MDMs for 15 min at 37°C before addition of

neutrophils) and left in a CO2 incubator at 37°C for 90 min. After washing, the MDMs were stained using an anti-CD14 allophycocyanin Ab (Caltag Laboratories)1honiceandanalyzed by flow cytometry. Samples were first gated based on forward scatter vs side scatter, and the gate was set around MDMs using an MDM-only sample (cells within this gate were Ͼ90% positive for CD14). This gate also included occasional neutrophils and was used to plot CFSE vs CD14. During analysis, cells were defined as MDMs (CFSElow/CD14high), neutrophils (CFSEhigh/CD14low), or Downloaded from MDMs associated with neutrophils (CFSEhigh/CD14high). Data are expressed as percentage of CD14-positive cells also expressing CFSE fluorescence in the presence or absence of cytochalasin D. After FACS analyses, the remaining cells were cytospun, stained with 4Ј,6Ј-diamidino-2- phenylindole (DAPI), mounted, and inspected visually using a Zeiss Axioplan 2 microscope with FITC and DAPI fluorescence filters. http://www.jimmunol.org/

Results Neutrophil uptake of B. cenocepacia To study the initial interaction between neutrophils and B. cenocepacia, neutrophils puriﬁed from peripheral blood of healthy volunteers were coincubated with the nonmucoid B. cenocepacia isolate, C8963, for 1 h. After challenge, the cells were washed at low

speed to remove the majority of noningested bacteria, and the up- by guest on October 10, 2021 take of bacteria was determined microscopically. The bacteria were readily taken up by the neutrophils under nonopsonic conditions, and most cells had ingested multiple bacteria. Fig. 1A shows neutrophils containing multiple ingested bacteria, many of which resided within phagosome-like organelles (arrows). When the bacteria were UV killed before challenge, the level of phagocytosis was not signiﬁcantly different from that obtained using viable bacteria (Fig. 1B). Phagocytosis was not affected by the presence of DPI, an inhibitor of the neutrophil NADPH-oxidase (Fig. 1B).

FIGURE 2. Viable, but not UV-killed, B. cenocepacia induce neutrophil apoptosis. A, Effect of challenge with viable B. cenocepacia on neu- FIGURE 1. Nonopsonic interaction of B. cenocepacia and human neutrophil apoptosis assessed by annexin V binding alone or in combination trophils. A, Giemsa-stained slides of neutrophils after 1-h nonopsonic in- with EthD, or by surface expression of CD16. Representative experiments teraction with viable B. cenocepacia. Arrows indicate bacteria residing in are shown. B, The effect on neutrophil apoptosis by challenge with viable phagosome-like organelles. B, Cellular association with bacteria (viable or or UV-killed bacteria, as determined by flow cytometry. Apoptotic cells UV killed) in the presence or absence of DPI was determined by micro- were defined as annexinV high/EthD low. Data are mean ϩ SEM of three scopic evaluation and is expressed as mean number of bacteria (bound or to seven independent experiments, and asterisks indicate statistically sig- ingested)/cell ϩ SEM of three to six independent experiments. nificant differences (p Ͻ 0.05). The Journal of Immunology 3565 higher proportion of bacterially challenged neutrophils was apoptotic compared with unchallenged cells (Fig. 2). Annexin V binding is but one characteristic feature of apoptotic neutrophils; other characteristics include down-regulation of certain cell surface re- ceptors and DNA fragmentation, followed by nuclear condensation (20). The proapoptotic effect of this B. cenocepacia strain was confirmed also by microscopic evaluation of cellular morphology (data not shown) and loss of surface-located CD16 (Fig. 2A). When the bacteria were UV-killed before challenge, no induction of apoptosis was seen; instead, UV-killed bacteria had slight antiapoptotic effects (not statistically significant), indicating that the proapoptotic effect was dependent on bacterial viability (Fig. 2B). The proportion of necrotic neutrophils (i.e., positive for both annexin V and EthD) was low (ϳ5%) irrespective of bacterial challenge.

Antiapoptotic effects of bacterial-derived molecules Previous studies have shown that B. cepacia complex can secrete

a hemolysin that induces apoptosis in human neutrophils (21). To Downloaded from investigate whether the proapoptotic effect was due to secreted bacterial products, neutrophils were incubated in the presence of cell-free bacterial supernatants. Instead of inducing apoptosis, however, bacterial supernatants had a marked antiapoptotic effect and inhibited spontaneous apoptosis in a dose-dependent manner (Fig. 3). Puriﬁed LPS from the B. cenocepacia isolate also was http://www.jimmunol.org/ antiapoptotic in a dose-dependent manner in the nanogram per milliliter range (Fig. 3). Thus, we concluded that no soluble factors released from the bacteria were responsible for the induction of apoptosis, and that released factors, such as LPS, instead constituted potent survival signals. That B. cenocepacia induced neutrophil apoptosis suggested that the uptake of viable bacteria was a proapoptotic event capable of overriding the effects of antiapoptotic molecules released from the bacteria.

FIGURE 4. A mucoid B. cenocepacia variant that is resistant to phago- by guest on October 10, 2021 Neutrophil apoptosis is dependent on bacterial uptake cytosis inhibits spontaneous neutrophil apoptosis. A, Bacterial association A mucoid B. cenocepacia variant, previously shown to be genet- of the nonmucoid vs mucoid B. cenocepacia isolates in the presence or absence of cytochalasin D (10 ␮g/ml) was determined by microscopic ically very closely related to the nonmucoid strain C8963, inter- evaluation and is expressed as mean number of bacteria (bound or ingest- acted signiﬁcantly less with human neutrophils than the nonmu- ed)/cell. B, Effect of challenge with the mucoid B. cenocepacia variant on coid isolate (16). When this mucoid strain was used to challenge neutrophil apoptosis assessed by annexin V binding in combination with neutrophils, very low levels of association were measured, and in EthD. Representative experiments are shown. contrast to the nonmucoid isolate, association was not inhibitable by cytochalasin D (Fig. 4A), indicating that the mucoid strain was not phagocytosed. Challenge with the mucoid isolate did not in- spontaneous apoptosis (Fig. 4B). LPS and supernatant from the duce neutrophil apoptosis; on the contrary, it protected cells from mucoid strain were also potently antiapoptotic (data not shown), indicating that in the absence of phagocytosis, antiapoptotic molecules secreted from the bacteria constituted powerful survival signals.

Clearance of apoptotic neutrophils Apoptotic cells are normally cleared through uptake by other, viable phagocytes, in particular macrophages (13). To determine whether neutrophils rendered apoptotic by phagocytosis of nonmucoid B. cenocepacia could be cleared by other phagocytes, we assessed the uptake of apoptotic neutrophils by MDMs using a flow cytometric assay similar to one previously described (22). As a control, we induced apoptosis by cross-linking of neutrophil CD95 (FAS), using a mouse anti-human CD95 Ab with an anti- mouse Ab. Cross-linking of FAS was a potent proapoptotic stimulus and yielded similar levels of apoptosis as challenge with B. FIGURE 3. Antiapoptotic effects of bacterial supernatants and purified cenocepacia (data not shown). LPS. Normal neutrophils were incubated with cell-free supernatants or When MDMs (identified by their high CD14 expression) were purified LPS from B. cenocepacia, and apoptosis was quantified on basis of mixed with apoptotic neutrophils labeled with the cell tracker dye annexin V binding. The figure shows mean ϩ SEM of three to four inde- CFSE, a substantial proportion of CD14-positive MDMs acquired pendent experiments. CFSE fluorescence, indicating that the MDMs had associated with 3566 NEUTROPHIL NECROSIS IN CGD apoptotic neutrophils (Fig. 5A). The actin filament-disrupting agent cytochalasin D largely inhibited the association, demonstrat- ing that most of these interactions represented ingestion of the apoptotic neutrophils by MDMs (Fig. 5A). This was confirmed by fluorescent microscopy of DAPI-stained slides, showing MDMs engulfing one or more CFSE-positive neutrophils (Fig. 5B). The proportion of CD14-positive cells displaying CFSE fluorescence in the presence or absence of cytochalasin D was determined for neutrophils rendered apoptotic either by ingestion of B. cenocepa- FIGURE 6. B. cenocepacia induces intracellular ROS production in cia or by cross-linking of FAS. Both types of apoptotic neutrophils neutrophils. Normal neutrophils were stimulated with viable or UV-killed were ingested to a similar extent by the MDMs (Fig. 5C). Thus, we bacteria (multiplicity of infection 1:30), and the intracellular ROS produc- concluded that neutrophils having phagocytosed viable B. cenoce- tion was measured continuously by luminol-ECL in the presence of extra- pacia became apoptotic and could be cleared by MDMs. cellular scavengers superoxide dismutase and catalase. Also shown are control stimulations using PMA (100 ng/ml) or viable bacteria in the pres- Neutrophil ROS production ence of DPI (10Ϫ5 M). A representative experiment is shown. Oxidative stress has long been known as an important proapoptotic factor (23), and neutrophils from patients with CGD (which are ϳ unable to form ROS) display delayed spontaneous apoptosis in with B. cenocepacia, peaking after 60 min and then declining slowly (Fig. 6). A very similar oxidative burst was noted when

vitro (11). Furthermore, intracellular production of ROS can in- Downloaded from duce neutrophil apoptosis (7), and ROS scavengers and/or neutrophils were stimulated with UV-killed bacteria, indicating NADPH-oxidase inhibitors prevent apoptosis following phagocy- that bacterial viability did not influence the oxidative response of tosis of a variety of microbes (8–10, 24, 25). The neutrophil the phagocytes. That UV-killed B. cenocepacia failed to induce NADPH-oxidase is localized in two cellular pools: in the mem- apoptosis, even though the bacterial cells were both phagocytosed brane of intracellular organelles (specific granules), and in the and induced ROS production, indicated that ROS generation was plasma membrane, and neutrophils can produce ROS intracellu- not sufficient to induce neutrophil apoptosis. The oxidative burst http://www.jimmunol.org/ larly as well as extracellularly (18). Phagocytosis of bacteria is induced by phagocytosis of bacteria (viable or UV killed) was often accompanied by activation of the NADPH-oxidase, resulting comparable in magnitude and duration, but differed kinetically, in ROS generation inside the phagosome. Robust intracellular from that induced by PMA and could be inhibited by DPI (Fig. 6). ROS production was measured when neutrophils were stimulated Apoptosis or necrosis and the implications of ROS The fact that UV-killed B. cenocepacia did not induce neutrophil apoptosis even though it was phagocytosed (Fig. 1B) and induced an oxidative burst (Fig. 6) indicated that the proapoptotic effect of viable bacteria was not exclusively due to the formation of ROS. by guest on October 10, 2021 To further investigate the role of ROS, we performed challenge experiments on neutrophils from CGD patients. Consistent with previous publications (9, 11), levels of spontaneous apoptosis were consistently lower for CGD neutrophils than for control neutrophils (Fig. 7), although not statistically significant. Challenge with B. cenocepacia induced apoptosis also in CGD neutrophils, and the levels of apoptosis, defined as percentage of cells positive for annexin V and negative for EthD, were not suppressed in the absence of ROS (Fig. 7). In fact, when neutrophils were unable to mount an oxidative response (CGD neutrophils or normal neutrophils in the presence of DPI), challenge resulted in slightly higher proportions of apoptotic cells than did challenge of normal neutrophils with the ability to produce ROS unaltered. These differences, however, failed to

FIGURE 5. Apoptotic neutrophils are cleared by MDMs. A, CD14- stained MDMs acquired CFSE fluorescence of apoptotic neutrophils when coculturing took place in the absence of cytochalasin D. B, Cells were cytospun and DAPI stained to visualize phagocytosis of apoptotic neutrophils (green with condensed blue nucleus) by MDMs (blue nucleus). C, The percentage of MDMs (CD14ϩ) acquiring CFSE fluorescence in the FIGURE 7. B. cenocepacia induces apoptosis in a ROS-independent presence or absence of cytochalasin D (5 ␮g/ml) was determined using manner. Normal neutrophils (in the presence or absence of DPI) or CGD flow cytometry. Neutrophils were rendered apoptotic due to challenge with neutrophils were challenged with B. cenocepacia, and apoptosis was as- B. cenocepacia or incubation with ␣-FAS ϩ cross linking Ab (as described sessed by flow cytometry. Cells positive for annexin V and negative for in the text) and the uptake quantified using flow cytometry. The figure EthD were defined as apoptotic. Data are mean Ϯ SEM from three to seven shows mean ϩ SEM of three different experiments. independent experiments. The Journal of Immunology 3567

Table 1. The absence of ROS promotes necrotic cell death after B. cenocepacia challenge

Necrotic Neutrophils (% of total Ϯ SD)a

Control Control ϩ DPI CGD Treatment (n ϭ 7) (n ϭ 6) (n ϭ 3)

Spontaneous 3.8 Ϯ 1.7 5.3 Ϯ 1.6 1.9 Ϯ 0.6 (annexin Vϩ/EthDϩ) B. cenocepacia 7.7 Ϯ 2.0 18.9 Ϯ 2.5b 20.2 Ϯ 3.3b (annexin Vϩ/EthDϩ) Spontaneous 10.7 Ϯ 2.2 11.3 Ϯ 1.7 8.5 Ϯ 2.4 (LDH release) B. cenocepacia 24.9 Ϯ 5.9 40.4 Ϯ 5.2 56.4 Ϯ 16.2b (LDH release)

a Neutrophils were cultured without treatment (Spontaneous) or after challenge with viable B. cenocepacia, and necrosis was quantified either by using flow cytometry (necrotic cells were defined as annexin Vϩ/EthDϩ) or by measuring release of cellular LDH to the supernatants. b Significantly different ( p Ͻ 0.05) from control cells receiving the same treatment and analyzed using the same method. Downloaded from

sequently being ingested by macrophages, processes that are critical for resolution of inflammation (13). The uptake of apoptotic neutrophils facilitates noninflammatory removal of the phlogistic mediators and proteolytic enzymes residing within these cells, and at the same time actively suppresses secretion of inflammatory mediators from the ingesting macrophages (26). Neutrophils that http://www.jimmunol.org/ die by necrosis, instead of apoptosis, release their toxic contents in an uncontrolled fashion, and macrophages ingesting necrotic cells may become activated to secrete high levels of proinflammatory cytokines (27). Thus, apoptosis and clearance are doubly beneficial FIGURE 8. B. cenocepacia causes necrotic cell death in the absence of in the termination of an inflammatory response, but necrosis may ROS. A, Normal or CGD neutrophils were cultured in the presence of have profound toxic implications on the outcome of an infection. viable B. cenocepacia, after which the cells were stained with annexin V The molecular mechanisms underlying neutrophil apoptosis

and EthD to differentiate between apoptotic (lower right quadrant) and have been extensively studied (reviewed by Simon (23) and Akgul by guest on October 10, 2021 necrotic (upper right quadrant) cells. A representative experiment is et al. (12)), and the process can be either accelerated or delayed by shown. B, Cellular LDH in cell-free supernatants was quantified after chal- death and survival signals, respectively. Upon phagocytosis of a lenge of normal or CGD neutrophils. Data are from one representative microbial prey, in most cases the neutrophils activate its apoptotic experiment and expressed as means of duplicate samples compared with program (28). However, some pathogens, such as Streptococcus the total LDH content of cells from each donor. pyogenes (28) and Chlamydia pneumoniae (29), modulate the neutrophil response to inhibit apoptosis. In both cases, the neutrophils fail to kill the ingested bacteria, and the delay of apoptosis may be reach statistical significance. These results indicated that the in- a bacterial strategy for establishing intracellular parasitism. In gen- duction of apoptosis occurred in a ROS-independent manner. eral, ROS appear to be important proapoptotic mediators, but the More dramatically, when neutrophils from CGD patients were exact nature of their interaction with other death molecules, e.g., challenged with B. cenocepacia, a significant proportion of cells caspases, has not been determined (12, 23). Regardless of how exhibited EthD fluorescence, indicative of a leaky plasma mem- ROS affect apoptotic processes, studies on neutrophils from CGD brane (Fig. 8 and Table I). This suggested that CGD neutrophils patients have shown that these cells display a slower rate of spon- became necrotic after B. cenocepacia challenge, and similar results taneous apoptosis compared with neutrophils from normal donors were obtained using normal neutrophils in the presence of DPI (9, 11). CGD neutrophils also fail to induce the proapoptotic reg- (Table I). Necrotic cell death was also evaluated by measuring the ulator BAX upon phagocytosis (30), and this could be the reason leakage of cellular LDH after challenge. These analyses clearly for the apoptotic defects displayed after the binding of death re- showed that substantial necrosis occurred only when neutrophils ceptors (11) and phagocytosis of IgG- and C3bi-coated latex incapable of producing ROS were challenged with viable bacteria beads (30). (Fig. 8 and Table I). Taken together, these data suggest that ROS Our data demonstrated that normal neutrophils produced sub- are not involved in promoting neutrophil apoptosis following stantial amounts of ROS during phagocytosis of B. cenocepacia, phagocytosis of B. cenocepacia, but are critical for detoxifying and subsequently became apoptotic. Using both neutrophils from ingested bacteria to abrogate necrosis of the ingesting neutrophil. CGD patients and normal neutrophils in the presence of DPI, we next investigated whether the ROS production was directly linked Discussion to the induction of apoptosis. We found that apoptosis after bac- Although the inflammatory reaction is crucial for host defense terial challenge was not reduced in the absence of ROS. Moreover, against infection, this process is a double-edged sword that when only viable B. cenocepacia promoted apoptosis, despite the fact persistent may cause a number of complications and even facilitate that UV-killed and live bacteria were ingested and induced ROS bacterial spread. Neutrophils are short-lived, potent inflammatory production at comparable levels. Taken together, these data dem- cells that are normally cleared by undergoing apoptosis and sub- onstrate that ROS production was neither sufficient, nor necessary 3568 NEUTROPHIL NECROSIS IN CGD for inducing apoptosis. This is in contrast to most earlier studies In summary, we have demonstrated that B. cepacia complex that, using DPI or different antioxidants, report that in the absence bacteria have the innate capacity to cause devastating necrosis un- of ROS the proapoptotic effects induced by the phagocytosis of der conditions in which ROS production is compromised. Patients microbes are inhibited (8–10, 25). Although chemicals such as with two dissimilar diseases, CGD and CF, may be particularly DPI have been used to mimic the conditions in CGD (31), they susceptible to this opportunistic pathogen because either ROS are may affect cells in other ways than by simply preventing ROS not produced (CGD) or their capacity to kill bacteria is abrogated formation (32). Neutrophils from CGD patients have been used to due to redox imbalance at the site of infection (CF). In both dis- demonstrate that ROS-dependent pathways are, at least partially, eases, novel interventions might be designed by taking advantage responsible for induction of neutrophil apoptosis following inges- of these observations. tion of heat-killed S. aureus (9). Thus, there appear to exist both ROS-dependent and ROS-independent mechanisms for inducing Acknowledgments neutrophil apoptosis. We thank Renee Wang and Mary Kinloch for excellent technical assis- tance, and Astra-Zeneca for supplying meropenem. A striking and clinically relevant difference between normal and CGD neutrophils challenged with B. cenocepacia was the induc- Disclosures tion of necrotic cell death in CGD neutrophils. After bacterial chal- The authors have no financial conflict of interest. lenge, apoptotic cell death was induced in both normal and CGD neutrophils; only in CGD cells, however, could necrosis be de- References tected. We measured necrosis using two different techniques, cel- 1. Heyworth, P. G., A. R. Cross, and J. T. Curnutte. 2003. Chronic granulomatous disease. Curr. Opin. Immunol. 15:578. lular uptake of the nuclear binding dye EthD and leakage of cel- 2. Goldblatt, D. 2002. Current treatment options for chronic granulomatous disease. Downloaded from lular LDH; both techniques showed significantly more necrosis Exp. Opin. Pharmacother. 3:857. when the challenged cells were unable to produce ROS (Table I). 3. Winkelstein, J. A., M. C. Marino, R. B. Johnston, Jr., J. Boyle, J. Curnutte, J. I. Gallin, H. L. Malech, S. M. Holland, H. Ochs, P. Quie, et al. 2000. Chronic The exact molecular mechanisms by which the different types of granulomatous disease: report on a national registry of 368 patients. Medicine cell death are mediated and where ROS fit into these processes are 79:155. 4. Speert, D. P. 2002. Advances in Burkholderia cepacia complex. Paediatr. Respir. largely unknown. It is possible that the different fate of CGD neu- Rev. 3:230. trophils after bacterial challenge is due to an altered degranulation 5. Speert, D. P., M. Bond, R. C. Woodman, and J. T. Curnutte. 1994. Infection with http://www.jimmunol.org/ pattern and impaired phagolysosomal fusion compared with nor- Pseudomonas cepacia in chronic granulomatous disease: role of nonoxidative killing by neutrophils in host defense. J. Infect. Dis. 170:1524. mal neutrophils. Another intriguing possibility is that ROS some- 6. Haslett, C. 1999. Granulocyte apoptosis and its role in the resolution and control how influence the activity of different types of caspases that may of lung inflammation. Am. J. Respir. Crit. Care Med. 160:S5. 7. Lundqvist-Gustafsson, H., and T. Bengtsson. 1999. Activation of the granule pool dictate the outcome. Regardless of the underlying molecular de- of the NADPH oxidase accelerates apoptosis in human neutrophils. J. Leukocyte terminants, our data clearly show that neutrophils unable to pro- Biol. 65:196. duce ROS become necrotic after challenge with B. cenocepacia. 8. Watson, R. W., H. P. Redmond, J. H. Wang, C. Condron, and D. Bouchier-Hayes. 1996. Neutrophils undergo apoptosis following ingestion of Escherichia coli. Taken together with the remarkable resistance to nonoxidative J. Immunol. 156:3986. means of neutrophil killing displayed by this group of bacteria (5), 9. Yamamoto, A., S. Taniuchi, S. Tsuji, M. Hasui, and Y. Kobayashi. 2002. Role of it may be that neutrophils from CGD patients, challenged with a reactive oxygen species in neutrophil apoptosis following ingestion of heat-killed by guest on October 10, 2021 Staphylococcus aureus. Clin. Exp. Immunol. 129:479. bacterium resistant to nonoxidative killing, are more prone to un- 10. Blomgran, R., L. Zheng, and O. Stendahl. 2004. Uropathogenic Escherichia coli dergo necrotic cell death simply because the engulfed prey is not triggers oxygen-dependent apoptosis in human neutrophils through the coopera- tive effect of type 1 fimbriae and lipopolysaccharide. Infect. Immun. 72:4570. properly killed. 11. Kasahara, Y., K. Iwai, A. Yachie, K. Ohta, A. Konno, H. Seki, T. Miyawaki, and Our data could explain why B. cenocepacia is often such a prob- N. Taniguchi. 1997. Involvement of reactive oxygen intermediates in spontane- lematic pathogen in CGD patients, while only rarely, if ever, caus- ous and CD95 (Fas/APO-1)-mediated apoptosis of neutrophils. Blood 89:1748. 12. Akgul, C., D. A. Moulding, and S. W. Edwards. 2001. Molecular control of ing any problems in normal individuals. Healthy individuals are neutrophil apoptosis. FEBS Lett. 487:318. able to effectively cope with a B. cenocepacia infection, and their 13. Savill, J. S., P. M. Henson, and C. Haslett. 1989. Phagocytosis of aged human neutrophils by macrophages is mediated by a novel “charge-sensitive” recogni- neutrophils become apoptotic after ingestion of the bacteria. These tion mechanism. J. Clin. Invest. 84:1518. apoptotic neutrophils can then be cleared by macrophages in an 14. Boyum, A., D. Lovhaug, L. Tresland, and E. M. Nordlie. 1991. Separation of anti-inflammatory manner, ensuring a proper termination of the leukocytes: improved cell purity by fine adjustments of gradient medium density and osmolality. Scand. J. Immunol. 34:697. inflammatory process. In CGD patients, however, infection with B. 15. Davidson, D. J., A. J. Currie, G. S. Reid, D. M. Bowdish, K. L. MacDonald, cenocepacia could be highly destructive if the ingesting neutro- R. C. Ma, R. E. Hancock, and D. P. Speert. 2004. The cationic antimicrobial peptide LL-37 modulates dendritic cell differentiation and dendritic cell-induced phils failed to kill the bacteria and then became necrotic. Instead of T cell polarization. J. Immunol. 172:1146. resolving the inflammation, such a neutrophil-bacterial interaction 16. Conway, B. A., K. K. Chu, J. Bylund, E. Altman, and D. P. Speert. 2004. Pro- would promote tissue destruction and a persistent inflammatory duction of exopolysaccharide by Burkholderia cenocepacia results in altered cell- surface interactions and altered bacterial clearance in mice. J. Infect. Dis. process. The resulting tissue damage could also provide a means 190:957. for bacteraemic spread of this organism, a pathological feature that 17. Hirschfeld, M., Y. Ma, J. H. Weis, S. N. Vogel, and J. J. Weis. 2000. Cutting edge: repurification of lipopolysaccharide eliminates signaling through both hu- is rare with other bacterial species in CGD patients (4). man and murine Toll-like receptor 2. J. Immunol. 165:618. B. cenocepacia is also a very problematic pathogen in patients 18. Dahlgren, C., and A. Karlsson. 1999. Respiratory burst in human neutrophils. with CF. Unlike the more common CF pathogen, P. aeruginosa, B. J. Immunol. Methods 232:3. 19. Levesque, A., A. Paquet, and M. Page. 1995. Improved fluorescent bioassay for cenocepacia can cause necrotizing pneumonia and life-threatening the detection of tumor necrosis factor activity. J. Immunol. Methods 178:71. septicemia. Numerous reports suggest that an oxidant/antioxidant 20. Dransfield, I., A. M. Buckle, J. S. Savill, A. McDowall, C. Haslett, and N. Hogg. 1994. Neutrophil apoptosis is associated with a reduction in CD16 (Fc␥RIII) imbalance exists in the lungs of CF patients, even though CF neu- expression. J. Immunol. 153:1254. trophils are not impaired in ROS production in vitro; molecules 21. Hutchison, M. L., I. R. Poxton, and J. R. Govan. 1998. Burkholderia cepacia found in CF secretions (e.g., bacterial mucoid exopolysaccharide produces a hemolysin that is capable of inducing apoptosis and degranulation of mammalian phagocytes. Infect. Immun. 66:2033. and host mucin) can quench ROS (33–35). It is possible that such 22. Schrijvers, D. M., W. Martinet, G. R. De Meyer, L. Andries, A. G. Herman, and quenching would protect B. cepacia complex bacteria from ROS- M. M. Kockx. 2004. Flow cytometric evaluation of a model for phagocytosis of cells undergoing apoptosis. J. Immunol. Methods 287:101. mediated killing, creating a situation akin to CGD with necrotic 23. Simon, H. U. 2003. Neutrophil apoptosis pathways and their modifications in cell death and persistent destructive inflammation. inflammation. Immunol. Rev. 193:101. The Journal of Immunology 3569

24. Rotstein, D., J. Parodo, R. Taneja, and J. C. Marshall. 2000. Phagocytosis of multiply in neutrophil granulocytes and delay their spontaneous apoptosis. J. Im- Candida albicans induces apoptosis of human neutrophils. Shock 14:278. munol. 172:1768. 25. Perskvist, N., M. Long, O. Stendahl, and L. Zheng. 2002. Mycobacterium tuber- 30. Kobayashi, S. D., J. M. Voyich, K. R. Braughton, A. R. Whitney, W. M. Nauseef, culosis promotes apoptosis in human neutrophils by activating caspase-3 and H. L. Malech, and F. R. DeLeo. 2004. Gene expression proﬁling provides insight altering expression of Bax/Bcl-xL via an oxygen-dependent pathway. J. Immunol. into the pathophysiology of chronic granulomatous disease. J. Immunol. 172:636. 168:6358. 31. Ellis, J. A., S. J. Mayer, and O. T. Jones. 1988. The effect of the NADPH oxidase 26. Savill, J., I. Dransﬁeld, C. Gregory, and C. Haslett. 2002. A blast from the past: inhibitor diphenyleneiodonium on aerobic and anaerobic microbicidal activities clearance of apoptotic cells regulates immune responses. Nat. Rev. Immunol. of human neutrophils. Biochem. J. 251:887. 2:965. 32. Bowie, A., and L. A. O’Neill. 2000. Oxidative stress and nuclear factor-␬B ac- 27. Stern, M., J. Savill, and C. Haslett. 1996. Human monocyte-derived macrophage tivation: a reassessment of the evidence in the light of recent discoveries. Bio- ␣ ␤ phagocytosis of senescent eosinophils undergoing apoptosis: mediation by v 3/ chem. Pharmacol. 59:13. CD36/thrombospondin recognition mechanism and lack of phlogistic response. 33. Learn, D. B., E. P. Brestel, and S. Seetharama. 1987. Hypochlorite scavenging by Am. J. Pathol. 149:911. Pseudomonas aeruginosa alginate. Infect. Immun. 55:1813. 28. Kobayashi, S. D., K. R. Braughton, A. R. Whitney, J. M. Voyich, T. G. Schwan, 34. Konig, B., P. Friedl, S. S. Pedersen, and W. Konig. 1992. Alginate: its role in J. M. Musser, and F. R. DeLeo. 2003. Bacterial pathogens modulate an apoptosis neutrophil responses and signal transduction towards mucoid Pseudomonas differentiation program in human neutrophils. Proc. Natl. Acad. Sci. USA aeruginosa bacteria. Int. Arch. Allergy Immunol. 99:98. 100:10948. 35. Cross, C. E., A. van der Vliet, S. Louie, J. J. Thiele, and B. Halliwell. 1998. 29. Van Zandbergen, G., J. Gieffers, H. Kothe, J. Rupp, A. Bollinger, E. Aga, Oxidative stress and antioxidants at biosurfaces: plants, skin, and respiratory tract M. Klinger, H. Brade, K. Dalhoff, M. Maass, et al. 2004. Chlamydia pneumoniae surfaces. Environ. Health Perspect. 106 (Suppl. 5):1241. Downloaded from http://www.jimmunol.org/ by guest on October 10, 2021