|Home | About | Journals | Submit | Contact Us | Français|
Staphylococcus aureus is a dangerous pathogen that can cause necrotizing infections characterized by massive inflammatory responses and tissue destruction. Staphylococcal α-hemolysin is an essential virulence factor in severe S. aureus pneumonia. It activates the nucleotide-binding domain and leucine-rich repeat containing gene family, pyrin domain containing 3 (NLRP3) inflammasome to induce production of interleukin-1β and programmed necrotic cell death. We sought to determine the role of α-hemolysin–mediated activation of NLRP3 in the pathogenesis of S. aureus pneumonia. We show that α-hemolysin activates the NLRP3 inflammasome during S. aureus pneumonia, inducing necrotic pulmonary injury. Moreover, Nlrp3−/− mice have less-severe pneumonia. Pulmonary injury induced by isolated α-hemolysin or live S. aureus is independent of interleukin-1β signaling, implicating NLRP3-induced necrosis in the pathogenesis of severe infection. This work demonstrates the exploitation of host inflammatory signaling by S. aureus and suggests the NLRP3 inflammasome as a potential target for pharmacologic interventions in severe S. aureus infections.
Community-acquired methicillin-resistant Staphylococcus aureus (CA-MRSA) is now the primary cause of skin infections requiring emergency medical attention in the United States . In addition, CA-MRSA can also cause severe, life-threatening infections including necrotizing pneumonias and fasciitis, which are associated with high mortality rates, even in previously healthy patients [2, 3]. These necrotizing infections are among the most severe complications of S. aureus infection and are characterized by localized necrosis and inflammation.
All S. aureus produce secreted exotoxin virulence factors, including several cytolysins: α-hemolysin, β-hemolysin, and bicomponent leukocidins . α-Hemolysin is one of several critical virulence factors in a murine model of S. aureus necrotizing pneumonia, including those caused by CA-MRSA [5, 6]. Purified α-hemolysin induces pulmonary inflammation in rats and rabbits [7–9]. Immunization with inactive α-hemolysin or pharmacologic inhibition of α-hemolysin can prevent or reduce the severity of S. aureus pneumonia in mice [10–12]. The mechanisms by which α-hemolysin increases the virulence of S. aureus in necrotizing pneumonia are not fully understood.
The nucleotide-binding domain and leucine-rich repeat containing gene family, pyrin domain containing 3 protein (NLRP3) inflammasome is a signaling complex that activates procaspase-1, the processing and secretion of the cytokines interleukin (IL) 1β and IL-18, and the initiation of programmed cellular necrosis [13–16]. The NLRP3 inflammasome is activated in response to many pathogen-derived molecules, sterile inducers of inflammation, and microbial pore-forming toxins [17–20]. We recently demonstrated that S. aureus α-hemolysin induces NLRP3-mediated signaling in cultured cells . S. aureus β-hemolysin and γ-hemolysin also activate the NLRP3 inflammasome, suggesting that numerous S. aureus virulence factors converge on this common host-signaling pathway . We sought to investigate whether the activation of the NLRP3 inflammasome was important in the pathogenesis of these infections.
All protocols were conducted in accordance with National Institutes of Health guidelines for the care and use of laboratory animals and approved by the Institutional Animal Care and Use Committee of the University of North Carolina at Chapel Hill.
S. aureus strain Newman and the α-hemolysin–deficient isogenic strain (Newman hla::erm) were provided by Dr Juliane Bubeck Wardenburg (University of Chicago) [5, 6]. Wild-type mice were from Jackson Laboratory (Bar Harbor, Maine). Nlrp3−/− and Il1r1−/− mice, back-crossed onto the C57BL/6J genetic background for 9 and 6 generations, respectively, were from Dr John Bertin (Millennium Pharmaceuticals) and Dr Jacques Peschon (Immunex; Amgen). S. aureus α-hemolysin (Sigma-Aldrich) was previously determined to be ~50% pure and to activate the NLRP3 inflammasome without triggering Toll-like receptor–dependent pro–IL-1β production . The endotoxin content of the α-hemolysin was determined to be <0.003 International Endotoxin Units/μg of toxin using a chromogenic limulus amebocyte lysate (ToxinSensor; Genscript).
Lung sections were incubated in 2.5 mg/mL collagenase/0.5 mg/mL DNase-1 (Sigma-Aldrich) at 37°C for 60 minutes. Red blood cells were lysed with ammonium chloride potassium-containing lysing buffer (Gibco). Single-cell suspensions were adjusted to 2 × 106/mL in complete medium and then stimulated with heat-killed S. aureus (HKSA; 1 particle/cell; Invivogen), 10 μg/mL S. aureus α-hemolysin (Sigma-Aldrich), or both for 4 hours. CD11b+ cells were purified using anti-CD11b magnetic microbeads according to the protocol provided by the manufacturer (Miltenyi Biotec).
Lactate dehydrogenase release, enzyme-linked immunosorbent assay (ELISA) for IL-1β and tumor necrosis factor (TNF) α, and high-mobility group box 1 (HMGB1) immunoblot analyses were carried out as described by Craven et al . HMGB1 levels were analyzed using the HMGB1 ELISA kit (IBL International). IL-6 and macrophage inflammatory protein 1α levels were determined using BioPlex multiplex cytokine analysis (BioRad).
Inoculi were prepared from S. aureus grown for 5 hours at 37°C in tryptic soy broth after 1:100 dilution of an overnight culture. The bacteria were washed and suspended in phosphate-buffered saline (PBS) at a concentration of 2 × 109 colony-forming units/mL, confirmed by colony counting of plated dilutions of the suspension. Age-matched (10–13-week-old) and sex-matched mice were anesthetized, hung in an upright position by their incisors, and inoculated with 50 μL of PBS containing S. aureus, HKSA, or α-hemolysin into the buccal cavity; the nares were blocked to induce aspiration. Intrarectal temperature of infected mice was monitored with a Thermalert TH-5 (Physitemp). Arterial oxygen saturation and pulse distention were determined using the MouseOx pulsoximeter (STARR Life Sciences) in nonanesthetized mice. For studies of lung mechanics, infected mice were given vancomycin (3 mg/d) intraperitoneally starting 4 hours after inoculation.
Mice were euthanized by CO2 asphyxiation, the trachea was canulated, and the lungs were flushed 3 times with 1 mL of 0.5 mmol/L ethylenediaminetetraacetic acid (EDTA-Dulbecco's PBS; Invitrogen). Total cell counts were determined using a Neubauer hemacytometer. Polymorphonuclear cell counts were determined by microscopic analysis of cells prepared using the CytoSpin 4 cytocentrifuge (Thermoscientific) and visualized using Quik-Dip Stain (Mercedes Medical). Bronchoalveolar lavage fluid (BALF) and aseptically excised lungs, spleens, and kidneys were collected from euthanized animals. Organ homogenates were prepared in sterile Whirl-Pack bags (Nasco), and serial dilutions were plated in duplicate on mannitol salt agar plates (BD). Bacterial burdens were determined by colony counting after overnight incubation.
Mice were anesthetized and then paralyzed with sequential intraperitoneal injection of 70–90 mg/kg pentobarbital sodium (American Pharmaceutical Partners) and 0.8 mg/kg pancuronium bromide (Baxter Healthcare). Invasive measurements of lung mechanics, including dynamic resistance and dynamic and static compliance, were performed using a computer-controlled small animal ventilator (flexiVent system; Scireq), as described elsewhere .
Mouse lungs were inflated with 10% buffered formalin (20-cm pressure), processed, stained with hematoxylin-eosin stain, and analyzed microscopically. The extent of lung pathology was scored as described in Supplemental Table 1. The total pathologic score for each mouse was calculated as the sum of scores from each category for that individual. All hematoxylin-eosin–stained sections were scored by a veterinary pathologist (Jackson Laboratory), who was blinded to treatment conditions.
Statistical analysis was performed using GraphPad software by Prism. All data are expressed as means ± standard error of the mean. In experiments in which only 2 groups were studied, an unpaired Student t test was used to determine significance. Significance in differences between multiple groups was analyzed by analysis of variance with Bonferroni as a posttest. Survival curves were created using the Kaplan–Meier method and compared using the log-rank (Mantel-Cox) test. In all cases, differences were considered statistically significant at P < .05.
We have reported that human acute monocytic leukemia cell line, THP1 cells, and murine peritoneal macrophages process and secrete IL-1β in response to treatment with HKSA and purified recombinant α-hemolysin. In these studies, HKSA induced pro–IL-1β production, whereas α-hemolysin triggered NLRP3-dependent caspase-1 activation and pro–IL-1β processing . Because α-hemolysin is a critical virulence factor in murine pneumonia models, we sought to determine whether it could induce NLRP3 signaling in cells from mouse lungs. Mixed lung cell preparations treated with α-hemolysin secreted modest levels of IL-1β, regardless of whether they were pretreated with HKSA (Figure 1A). To examine the response of pulmonary macrophages to α-hemolysin treatment, CD11b+ cells were isolated from these mixed preparations. These cells secreted higher levels of IL-1β after treatment with either α-hemolysin or HKSA followed by α-hemolysin than preparations depleted of CD11b+ cells (Figure 1B). α-Hemolysin did not induce IL-1β secretion in mixed or CD11b+ cells from Nlrp3−/− mice (Figure 1A and and1B).1B). α-hemolysin–induced IL-1β secretion did not require prestimulation (Figure 1B). Supernatants from lung cell preparations induced pro–IL-1β expression when applied toTHP1 cells (Figure 1C), suggesting that pro–IL-1β is up-regulated in the lung cells as a result of the preparation method rather than constitutive high-level expression.
We also sought to determine whether NLRP3-mediated signaling was activated by α-hemolysin in vivo. HKSA, α-hemolysin, and HKSA+α-hemolysin were administered to mice by intratracheal instillation. Six hours after instillation of HKSA (5 × 108 particles, a lethal dose of living S. aureus), BALF contained minimal IL-1β and robust levels of TNF-α (Figure 1D and and1E).1E). Treatment with α-hemolysin alone did not induce TNF-α or IL-1β secretion. Instillation of HKSA+α-hemolysin induced detectable IL-1β, which was markedly diminished in the Nlrp3−/− mice (Figure 1D). Because instillation of both HKSA and α-hemolysin was required for pulmonary IL-1β production, it is likely that a priming step that induces pro–IL1β production is required for α-hemolysin–induced IL-1β secretion in vivo, as has been documented in cell culture systems . The addition of HKSA+α-hemolysin induced less TNF-α in BALF from wild-type mice compared with HKSA treatment alone, and BALF TNF-α followed a similar trend in Nlrp3−/− mice (Figure 1E). Thus, HKSA-induced TNF-α secretion may be carried out by cells that do not express NLRP3, or α-hemolysin may interfere with signaling pathways for TNF-α production independent of NLRP3.
We also characterized the pulmonary pathology associated with HKSA+α-hemolysin treatment. Despite the difference in IL-1β secretion between wild-type and Nlrp3−/− mice treated with HKSA+α-hemolysin at 6 hours, no animals survived >12 hours (data not shown). To further investigate α-hemolysin–induced pulmonary pathology, mice were given a sublethal dose of HKSA+α-hemolysin (HKSA, 5 × 107 particles; α-hemolysin, 0.5 μg). After 24 hours, the BALF of HKSA+α-hemolysin–treated animals demonstrated robust neutrophilia, which was markedly diminished in the Nlrp3−/− animals (Figure 2A). Nlrp3−/− mice also exhibited less lung pathology in a composite histopathologic index (Figure 2B and and2C)2C) and individual indices of alveolar and vasculitic inflammation (data not shown). Thus, NLRP3 activation plays an important role in α-hemolysin–induced pulmonary injury and inflammation.
Intratracheal HKSA+α-hemolysin induced severe hypothermia and labored respiration, a syndrome similar to pneumonia induced by living bacteria. The clinical consequences of intratracheal HKSA (1 × 108 particles), α-hemolysin (1 μg), or HKSA+α-hemolysin were studied in wild-type mice. Although mice completely recovered from treatment with HKSA alone, treatment with α-hemolysin resulted in ~40% mortality and treatment with HKSA+α-hemolysin induced 80%–100% mortality (Figure 2D and 2E). Nlrp3−/− mice were protected from death when compared with their wild-type counterparts (Figure 2E). Surviving Nlrp3−/− mice were less hypothermic and recovered more rapidly than wild-type mice (Figure 2F). These data suggest that NLRP3-induced inflammation plays a significant role in the clinical phenotype of α-hemolysin–induced pneumonitis.
Because IL-1β is important in host clearance of S. aureus from skin infection models, we used a murine model to further delineate the role of NLRP3 inflammasome activation in S. aureus pneumonia. Wild-type mice experienced severe hypothermia and markedly diminished pulse distention consistent with severe systemic inflammatory response within 24 hours of S. aureus administration. Nlrp3−/− mice were less hypothermic and had increased pulse distention compared with their wild-type counterparts (Figure 3A and and3B).3B). In mice surviving to 48 hours, Nlrp3−/− animals were less hypoxemic than wild-type mice (Figure 3C). The overall mortality in this model of S. aureus pneumonia was lower in Nlrp3−/− mice (14% mortality) than in wild-type mice (37% mortality), but this difference was not statistically significant (data not shown). Compared with wild-type mice, Nlrp3−/− mice exhibited significantly better survival free of severe pneumonia, as defined by death, oxygen saturation <95%, or a temperature drop >10°C (Figure 3D).
The consequences of severe S. aureus pneumonia on lung function were determined in surviving mice. Wild-type and Nlrp3−/− mice were treated with vancomycin, an antibiotic used to treat severe S. aureus infections, at the first signs of illness after inoculation with S. aureus. All mice survived and normalized their temperature within 3 days (data not shown), and pulmonary function was assessed on day 4. Postpneumonia mice had reduced dynamic lung compliance and increased airway resistance compared with uninfected mice (Figure 3E and and3F).3F). Along with milder clinical disease, lung compliance was not reduced in Nlrp3−/− mice after pneumonia. Thus, the NLRP3 inflammasome plays a key role in mediating pulmonary injury during S. aureus pneumonia.
Given the unexpected findings that the NLRP3 inflammasome was not protective in murine S. aureus pneumonia and that the disease process was worse in Nlrp3+/+ mice, further studies of the effects of NLRP3 on the host and bacteria during S. aureus pneumonia were carried out. Quantitative culture of BALF and lung homogenates revealed that wild-type and Nlrp3−/− mice had equivalent S. aureus burdens after 24 hours and had almost completely cleared the bacteria by 48 hours (Figure 4A and and4B).4B). S. aureus burdens in the kidneys were low, with no significant difference between wild-type and Nlrp3−/− mice with S. aureus pneumonia (Figure 4C). No dissemination to the spleen was observed in either mouse strain (data not shown). Thus, NLRP3 inflammasome activity does not influence clearance of S. aureus during pulmonary infection.
We also examined the role of NLRP3 in the inflammatory response to live S. aureus pneumonia. As expected, Nlrp3−/− mice had reduced levels of IL-1β in BALF and lung homogenates (Figure 4D and and4E).4E). TNF-α (Figure 4F), IL-6 (not shown), and macrophage 300 inflammatory protein 1α (not shown) levels in BALF were similar between wild-type and Nlrp32/2 mice. Thus, Nlrp3−/− mice had reduced IL-1β levels during S. aureus pneumonia, while production of other inflammatory cytokines and chemokines was largely intact.
Wild-type mice with S. aureus pneumonia had hemorrhagic lungs on gross examination, and the Nlrp3−/− mice had pink, healthy-appearing lung tissue (Figure 4G). The BALF samples from Nlrp3−/− mice with S. aureus pneumonia contained fewer neutrophils than their wild-type counterparts (Figure 4H). Histologic examination also revealed significantly less pulmonary pathology in the Nlrp3−/− mice using a composite index (Figure 4I). In total, our data suggest that activation of the NLRP3 inflammasome during pneumonia leads to excessive inflammation that is deleterious, rather than protective, to the host lung tissue.
To determine whether α-hemolysin from live S. aureus was driving the NLRP3-induced lung pathology, we used an isogenic strain of S. aureus with a deletion of the hla gene. As reported elsewhere, the hla-deficient S. aureus was less virulent than α-hemolysin-expressing S. aureus (data not shown) . To observe a clinical phenotype, we induced S. aureus pneumonia with hla::erm S. aureus with a larger dose than was used with wild-type S. aureus. There was no difference between wild-type and Nlrp3−/− mice in the recovered bacteria in BALF or lung homogenates after challenge with hla-deficient S. aureus (Figure 5A and and5B).5B). In contrast to findings with wild-type S. aureus, the mouse strains challenged with hla::erm S. aureus were equally hypothermic, with no difference in BALF neutrophil counts (Figure 5C and and5D).5D). These data support the hypothesis that activation of NLRP3 is a major mechanism by which α-hemolysin mediates induction of severe S. aureus pneumonia.
Anti-IL-1 therapy is highly efficacious in the treatment of periodic fever syndromes in patients with mutations in the Nlrp3 gene, suggesting that processing and secretion of IL-1β is the primary downstream signaling event of NLRP3 inflammasome activation [24–26]. We sought to determine whether excessive IL-1β signaling induced by α-hemolysin was involved in S. aureus virulence. First, we studied the effects of intratracheal HKSA+α-hemolysin intoxication in Il1r1−/− mice. Unlike Nlrp3−/− mice, HKSA+α-hemolysin–treated Il1r1−/− mice had mortality and hypothermia equivalent to that in wild-type mice (Figure 6A and and6B).6B). We also studied pneumonia with live S. aureus in wild-type and Il1r1−/− mice. Survival free from severe pneumonia (Figure 6C) and absolute mortality (not shown) was also indistinguishable between strains. Il1r1−/− mice with S. aureus pneumonia also had worse blood oxygen saturation than wild-type mice (Figure 6D). Combined, these data suggest that the pathogenic mechanism of α-hemolysin–mediated NLRP3 activation during S. aureus pneumonia does not require production of IL-1β.
Because α-hemolysin enhanced S. aureus virulence through a mechanism that was NLRP3 dependent and IL-1 signaling independent, we sought to examine additional NLRP3-dependent signaling pathways in CD11b+ cells and in mice. Similar to our findings with IL-1β secretion, α-hemolysin induced NLRP3-dependent death that was accompanied by release of HMGB1, a marker of programmed necrosis in CD11b+ cells (Figure 6E—G). Cell death was not significantly induced in the CD11b-negative cells by this concentration of α-hemolysin (data not shown). HMGB1 levels were also increased in BALF from HKSA+α-hemolysin–treated wild-type mice but not Nlrp3−/− mice, confirming that NLRP3 mediates α-hemolysin–induced necrotic cell death in vivo (Figure 6H). α-Hemolysin also induced IL-18 secretion from wild-type but not Nlrp3−/− CD11b+ cells (Figure 6I). IL-18 levels were increased in BALF from wild-type mice with S. aureus pneumonia but not Nlrp3−/− mice (Figure 6J).
S. aureus produces many virulence factors, including both surface proteins and exotoxins that have been shown to be important in murine pneumonia models. Recent studies have focused on the role of α-hemolysin in this model . Numerous host cell types are susceptible to this cytolytic toxin. Although lysis of red blood cells to provide heme as an iron source may play a role in α-hemolysin–mediated virulence in humans, this is not the case in mice, because murine hemoglobin is not recognized by the S. aureus hemoglobin receptor . Cytolysis of host epithelial cells and leukocytes must be the primary mechanism by which this toxin mediates virulence in the murine pneumonia. Several studies indicated that the NLRP3 inflammasome may be an important target pathway for α-hemolysin; however, in vivo evidence was lacking [21, 22]. In this study we show that pulmonary macrophages are a target of α-hemolysin activity among cells isolated from murine lungs. S. aureus α-hemolysin activates the NLRP3 inflammasome both in cultured cells in vitro and in the lungs in vivo. Interestingly, α-hemolysin’s activity toward other cell types in the respiratory system may be mediated through signaling systems independent of NLRP3, such as a disintegrin and metalloprotease 10 (ADAM10), which is a cellular receptor for α-hemolysin and was recently shown to play an important role in murine S. aureus pneumonia [28, 29]. Mice lacking NLRP3 are not completely protected from α-hemolysin-expressing S. aureus, suggesting that α-hemolysin promotes virulence through NLRP3-dependent and NLRP3-independent mechanisms.
Given previous reports showing that α-hemolysin activates NLRP3, it was not unanticipated that IL-1β secretion is diminished in Nlrp3−/− mice after intratracheal α-hemolysin administration or S. aureus pneumonia. However, because IL-1β and other inflammasome components are important in the clearance of S. aureus in skin infection models, it is surprising that Nlrp3−/− mice are not more susceptible to S. aureus pneumonia . S. aureus lacking O-acetylated peptidoglycan induces increased inflammasome activation and has decreased virulence in a skin infection model compared with S. aureus with O-acetylated peptidoglycan . However, O-acetylation of peptidoglycan also protects the bacteria from lysozyme, making it unclear whether the mechanism leading to the diminished virulence is related to increased inflammasome activation. We report that S. aureus pneumonia is slightly more severe in Il1r1−/− mice, paralleling previous reports that IL-1β signaling is involved in clearance of S. aureus in other models of infection. Mice lacking NOD2, an NLR protein involved in recognition of muramyl dipeptide, have decreased sensitivity to S. aureus pneumonia and increased sensitivity to intraperitoneal and cutaneous S. aureus infections [32–34]. NOD2 signaling up-regulates production of pro–IL-1β, antimicrobial peptides, and other host defense mechanisms, leaving the mechanism underlying the site-specific phenotypes of S. aureus infection in Nod2−/− mice still to be determined . It also remains to be determined whether NLRP3 deficiency will carry a consistent phenotype with varying sites of infection.
Although S. aureus pneumonia in mice may differ from human disease, α-hemolysin does activate the NLRP3 inflammasome in both human and murine cells. Thus, inhibition of the NLRP3 inflammasome may someday prove to be useful adjunctive therapy to antibiotics in severe S. aureus pneumonia. At this time, specific NLRP3 inflammasome inhibitors are not available; however, glyburide, an antihyperglycemic medication, has been shown to have NLRP3-inhibitory effects . Interestingly, a recent analysis of patient survival during infection with Burkholderia pseudomallei demonstrated that diabetics taking glyburide had better survival rates than nondiabetics, hinting that inhibition of the NLRP3 inflammasome may indeed be beneficial in some acute human infections . Because the protection from S. aureus pneumonia observed in Nlrp3−/− mice was independent of IL-1β signaling, IL-1 antagonists, which are efficacious in ameliorating fever syndromes associated with NLRP3 signaling hyperactivity, are unlikely to be beneficial and may be harmful in the setting of S. aureus infection .
S. aureus produces multiple pore-forming toxins that are implicated in the pathogenesis of severe infections and are capable of activating the NLRP3 inflammasome [38, 39]. The presence of NLRP3-activating, pore-forming toxins in other respiratory pathogens, like pneumolysin in Streptococcus pneumoniae, suggests that the NLRP3-induced inflammation may be an important factor in many bacterial pneumonias [40–42]. The ability to activate programmed necrosis at a distance may protect these bacteria from host phagocytes, and inflammatory damage to the pulmonary architecture may facilitate bacterial penetration of the epithelial barrier [7, 43]. Further studies are needed to clarify the role of toxin-mediated activation of NLRP3 in other bacterial pneumonias.
Supplementary materials are available at the Journal of Infectious Diseases online (http://www.oxfordjournals.org/our_journals/jid/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.
We thank Jenny P. Y. Ting and P. Fredrick Sparling for their insights into these studies. The Imaging and Biomarker Analysis Core of the Southeastern Regional Center of Excellence for Emerging Infections and Biodefense (SERCEB) carried out multiplex cytokine analysis.
Studies were supported by the Burroughs Wellcome Fund: Career Award for Medical Scientists (to J. A. D.) and the National Institutes of Health (grants AI057157 [SERCEB Career Development Award] and AI088255 to J. A. D., AI077437 [Lung Disease Model Core] and HL071802 to S. L. T.; and HL048160 to C.M.D.). P. M. B. is a trainee of the University of North Carolina STD and HIV Training Program (grant T32AI007001).
All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.