Surface Proteins and Exotoxins Are Required for the Pathogenesis of Staphylococcus aureus Pneumonia

Abstract

A model of Staphylococcus aureus-induced pneumonia in adult, immunocompetent C57BL/6J mice is described. This model closely mimics the clinical and pathological features of pneumonia in human patients. Using this system, we defined a role for S. aureus strain Newman surface proteins and secreted exotoxins in pneumonia-related mortality.

Staphylococcus aureus is an important bacterial pathogen causing pneumonia in both adult and pediatric populations. In recent reports, workers have described the growing incidence of severe S. aureus pneumonia in otherwise healthy individuals, often caused by multi-drug-resistant strains (8, 9). In addition, S. aureus remains one of the most common causes of ventilator-associated pneumonia, contributing to significant morbidity and mortality (18). At present, little is known about the S. aureus virulence factors that play a role in lower respiratory tract disease. The development of an adult, immunocompetent animal model system recapitulating S. aureus pneumonia would provide a useful tool for investigating such factors.

To date, small-animal models of S. aureus pneumonia have relied on the use of surgical inoculation methods or infection of immunocompromised animals (6, 17). While these models highlight the inflammatory response to intrapulmonary S. aureus, detailed characterization of S. aureus-encoded virulence factors has not been possible as the organisms are rapidly cleared from the lungs. A murine model of pulmonary infection with agar-embedded S. aureus defined a role for coagulase in hematogenous infection (34), while a neonatal mouse model of S. aureus pneumonia revealed the importance of the accessory gene regulator A (agrA), sarA, and staphylococcal protein A (spa) in the development of disease (10, 13). Together, the data suggest that multiple S. aureus virulence factors contribute to the pathogenesis of pneumonia.

We sought to develop a transnasal murine model of S. aureus pneumonia in adult, immunocompetent animals to permit investigation of virulence factors. To define infection parameters leading to evidence of pneumonia in 7-week-old C57BL/6J mice (Jackson Laboratories), groups of 20 animals were inoculated via the intranasal route with either phosphate-buffered saline (PBS) or one of three doses of S. aureus Newman, a human clinical isolate (7). Following 1:100 dilution of an overnight culture into fresh tryptic soy broth, staphylococci were grown with shaking at 37°C to an optical density at 660 nm of 0.5. Culture aliquots (50 ml) were sedimented by centrifugation, and staphylococci were washed and suspended in 750 μl PBS. Animals were anesthetized with ketamine and xylazine as previously described (21). After appropriate anesthesia was documented, 30 μl of bacterial slurry was inoculated into the left nare, and animals were held upright for 1 min postinoculation. All animals were given food and water ad libitum and observed continually for 72 h. Immediately following inoculation, all animals displayed labored breathing marked by a high respiratory rate and exaggerated chest wall excursion. This initial physiologic change resolved within 6 h, and all live animals at this initial time were ambulatory and appeared to be well. A small percentage of animals routinely succumbed within the first 6 h following inoculation, likely from the combined effects of aspiration and anesthesia. These animals were not included in subsequent analyses. Inoculation with 4 × 10⁸ CFU of S. aureus Newman resulted in a mortality rate of approximately 50% at 24 h, and an additional 20% of the animals succumbed to infection within 48 h following inoculation (Fig. 1A). Importantly, all infected animals appeared to be ill, having an increased respiratory rate, hunched posture, and decreased mobility at 24 h. A smaller bacterial inoculum, 8 × 10⁷ CFU, resulted in no mortality, although the infected animals appeared to be ill. The condition of this group of animals improved markedly by 48 h, and the animals resembled uninfected animals. Similar results were obtained with an inoculum of 1.3 × 10⁸ CFU of S. aureus Newman (data not shown). Inoculation with 8 × 10⁸ CFU of S. aureus Newman resulted in nearly 90% mortality by 24 h, which was significantly greater than the mortality observed for an inoculum of 4 × 10⁸ CFU at the same time (P = 0.02); the surviving animals appeared to be ill until 72 h postinfection.

FIG. 1.

Inoculum-based mortality and proliferation of S. aureus Newman in murine lung tissue. (A) C57BL/6J animals were inoculated with either PBS or various doses of live S. aureus Newman via the intranasal route. The levels of survival were recorded at 24, 48, and 72 h postinfection. Animals that appeared to be moribund were killed and counted as dead animals at the appropriate time. The results were analyzed to determine statistical significance using Fisher's exact test (P ≤ 0.02) (indicated by asterisks). (B) Animals were inoculated with 3 × 10⁸ to 4 × 10⁸ CFU of S. aureus Newman, and the bacterial CFU in both lungs (5 min) or the right lung (6, 24, 48, and 72 h) were enumerated at different times postinfection. Data were analyzed to determine significance using Student's t test.

To assess the kinetics of bacterial growth and clearance in the lung, animals were infected with 3 × 10⁸ to 4 × 10⁸ CFU of wild-type S. aureus Newman. At different times postinfection, animals were killed by forced CO₂ inhalation, in compliance with the University of Chicago Institute of Animal Care and Use Committee guidelines. The right lung of each animal was excised using aseptic techniques and suspended in 1 ml of PBS, and the tissue was homogenized. Serial dilution and plating were performed to determine the staphylococcal burden in the lung tissue. Immediately following infection, approximately one-third of the inoculum could be recovered from the lungs (Fig. 1B); this level of recovery was not significantly different from that at 6 h postinfection. Interestingly, by 24 h, in most animals there were significant increases in the number of staphylococci in lung tissues (P = 0.05), indicating that S. aureus Newman proliferated following infection. The level of recovery of S. aureus decreased at 48 to 72 h, corresponding to clinical improvement in the animals.

To discern whether pulmonary infection with S. aureus in this murine model was capable of causing pathological lesions observed in human patients, we examined the lungs of infected animals for gross pathological changes, as well as histopathologic evidence of infection. The lung tissue of infected animals was red and had a firm texture (Fig. 2A). In contrast, the lungs of uninfected animals were light pink and spongy. Inspection of the dissected left lung from a representative infected animal further revealed a heterogeneous red color, consistent with marked congestion (Fig. 2B, right panel).

FIG. 2.

Gross pathology of animals infected with S. aureus via the intranasal route. Representative infected animals were compared to uninfected animals in order to obtain gross pathological findings for lungs in situ (A) or following dissection of the left lung (B).

For histopathologic analysis, the left lung was dissected and placed in 1% formalin. Formalin-fixed tissues were processed, stained with hematoxylin and eosin, and visualized by light microscopy. Histopathologic examination revealed the consequences of S. aureus infection for lung parenchyma. As a control, we observed normal alveolar architecture in uninfected animals, in which thin-walled air spaces were defined by a single layer of pneumocytes (Fig. 3A). As early as 6 h following inoculation with S. aureus, aggregates of dark purple-stained immune cells were observed in the lungs of infected animals (Fig. 3B). The overall lung architecture was preserved at this time, and no bacteria were evident in tissues. In contrast, by 24 h, significant alveolar destruction had occurred along with infiltration of large numbers of immune cells (Fig. 3C). Interestingly, large foci of staphylococci were found in lung tissues at this time, consistent with bacterial proliferation. Dense, eosinophilic staining consistent with proteinaceous edema was observed to fill the alveolar space in infected animals (Fig. 3D). By 48 h, the size of these bacterial foci was reduced or foci were absent, and the reemergence of air-filled spaces was evident (Fig. 3E). At 72 h, significant air space had been restored; however, the alveolar walls remained thickened (Fig. 3F). Together, these data established a murine model of S. aureus pneumonia that closely mimics the clinical and histopathologic findings for human patients. It is likely that both the size of the inoculum and the mouse strain utilized contribute to the development of pneumonia in this animal model. This combination was not examined in previous studies. The large inoculum required to cause pneumonia in these animals speaks to the remarkable ability of the murine immune system to eliminate this pathogen from the lung, raising the possibility that an extension of this model system to other strains of immunocompetent mice may enhance our understanding of pulmonary immunity against S. aureus.

FIG. 3.

Histopathologic findings following intranasal inoculation of S. aureus. Lung tissue harvested from animals infected with S. aureus Newman was prepared and visualized by hematoxylin and eosin staining. Representative histology for an uninfected control having normal lung parenchyma is shown along with a series of images obtained for animals examined at the times indicated. Aggregates of purple-stained immune cells were seen as early as 6 h postinfection (arrowhead), and dense accumulation of bacteria was evident in tissues at 24 h postinfection (arrowhead).

To define S. aureus virulence factors critical for infection of the lower respiratory tract, mortality following pulmonary infection of mice with wild-type S. aureus Newman or isogenic mutants of this strain was assessed. S. aureus Newman strains carrying a deletion in srtA and srtB have been described previously (24, 27). agrA, spa, hla, and icaA mutants harboring bursa aurealis insertions were transduced into wild-type S. aureus Newman using isolates of the Phoenix transposon library (2). All mutant strains were cultured in tryptic soy broth supplemented with erythromycin (10 μg/ml). When mice were inoculated with wild-type strain Newman, slightly more than 70% of the infected animals succumbed over a 72-h period (Fig. 4A). Sortase A mutants (srtA) of S. aureus strain Newman are unable to anchor surface proteins with LPXTG sorting signals to the staphylococcal cell wall envelope; srtA mutations effectively disrupt the surface display of 17 polypeptides (Spa, FnBPA, FnBPB, ClfA, ClfB, SdrC, SdrD, SdrE, IsdA [SasE], IsdB [SasJ], IsdH [SasI], SasA, SasB, SasC, SasD, SasF, and SasH) involved in staphylococcal adherence to host tissues or immune evasive strategies (24, 25, 27). Compared to the mortality of animals challenged with the same dose of wild-type staphylococci, there was a significant reduction in the mortality of animals infected with sortase A mutants (P = 0.001). Protein A, a surface protein with five immunoglobulin-binding modules, captures antibodies via their Fc portion (5, 14). S. aureus Newman insertion mutants with mutations in spa, with defects in protein A synthesis and in staphylococcal binding to immunoglobulin, also displayed a significant defect for S. aureus-induced mortality. These data corroborate previous observations concerning the requirement of protein A for the pathogenesis of staphylococcal pneumonia in newborn mice (10). Sortase B (SrtB) anchors IsdC, a heme-binding protein, to the cell wall envelope, and mutants with a deletion in srtB have defects in staphylococcal heme iron scavenging (22, 26). Deletion of srtB in S. aureus strain Newman resulted in only a small reduction in mortality, suggesting that heme iron scavenging may not be essential for the pathogenesis of staphylococcal pneumonia. The exopolysaccharide poly-N-acetylglucosamine (PNAG) is synthesized by icaABC products (12, 29). PNAG conjugates may function as a vaccine as immunization of mice with this compound can protect the animals against invasive staphylococcal disease (20, 29). Furthermore, icaABC mutations cause a reduction in virulence in a mouse model of abscess formation in kidney tissues (19). However, icaA mutants displayed no defect in virulence, suggesting that the PNAG exopolysaccharide is not required for the pathogenesis of staphylococcal pneumonia in mice.

FIG. 4.

S. aureus mutants lacking protein A (spa) or all surface proteins (srtA) or exoproteins (agrA and hla) are defective in the ability to cause pneumonia-related mortality. Animals infected with 3 × 10⁸ to 4 × 10⁸ CFU of wild-type S. aureus Newman or isogenic mutant strains were scored for acute lethal disease, which demonstrated that there was a significant reduction in mortality for animals infected with both the srtA and spa strains (A). Analysis of mutants with bursa aurealis insertions in agrA and hla (alpha-toxin) likewise demonstrated that there was a marked reduction in the ability to cause acute lethal disease (B). Statistical significance was evaluated by Fisher's exact test (one asterisk, P = 0.001; two asterisks, P < 0.002).

In previous work, researchers reported that tracheal instillation of S. aureus strain 8325-4 into the lungs of anesthetized Sprague-Dawley rats causes damage to alveolar epithelia and erythrocytes in a manner requiring hla, which encodes staphylococcal alpha-toxin, the secreted hemolysin expressed by virtually all S. aureus strains (16, 28). After binding to receptor sites on cell surfaces, alpha-toxin forms a heptameric assembly and funnel-shaped pore that perforates host cell membranes (3, 35). S. aureus mutants lacking hla have reduced virulence in invasive disease models as larger numbers of staphylococci are required to kill mice following either intraperitoneal or intramammary infection (4, 33). These observations prompted us to examine the virulence of S. aureus Newman hla mutants in murine pneumonia. Interestingly, animals infected with the hla mutant strain appeared to be moderately ill within 24 h postinoculation; however, only a small number of these animals succumbed to the infection (Fig. 4B). The death of these animals was delayed, occurring more than 48 h postinoculation. Expression of many staphylococcal genes is regulated by agr, the accessory gene regulatory locus. This locus provides both quorum sensing and regulatory control of virulence (31). Briefly, AgrA and AgrC, a response regulator and a sensory kinase, perceive the environmental abundance of autoinducer peptide to activate expression of an array of genes, including hla and other exotoxin genes, at the threshold level (15). The autoinducer peptide, synthesized from an AgrD proinducer, is processed and secreted by AgrB (23). Mutations in agrA are known to abrogate quorum sensing (32). S. aureus Newman variants carrying a bursa aurealis insertion in agrA are avirulent in the murine pneumonia model, as none of the experimental animals succumbed to infection (Fig. 4). These findings can be explained by the regulatory defect of agrA mutations, which abrogate expression of many virulence genes, including genes encoding α-hemolysin, β-hemolysin, γ-hemolysin, and δ-hemolysin, as well as leukocidins (31). S. aureus Newman cannot express β-hemolysin, as this strain has a phage insertion in the hlb gene (1). However, three secreted γ-hemolysins (HlgA, HlgB, and HlgC) assemble into heterooligomeric toxins with a structure and function similar to the structure and function reported for α-hemolysin (11). Thus, the observed virulence defect of agrA mutants in the murine pneumonia model is likely due to the aggregate loss of all secreted hemolysins and toxins (31).

The inability of agrA and hla mutant strains to contribute to lethality in experimental animals raises the interesting possibility that S. aureus exotoxins may play a pivotal role in lung parenchymal injury. It is readily appreciated that insults to the alveolar epithelium contribute to impaired gas exchange. Furthermore, there are detrimental systemic effects of pulmonary inflammation, as patients with acute lung injury are susceptible to multiple-organ dysfunction and increased mortality. These systemic effects are likely mediated by the combined effects of inflammatory cytokines, such as interleukin-1 and interleukin-8, along with the products of arachadonic acid metabolism, including thromboxane A2 and prostaglandins. Our observation that agrA and α-hemolysin mutants do not induce mortality may provide insight into the specific mechanism by which S. aureus-induced lung injury contributes to the significant morbidity and mortality associated with severe S. aureus pneumonia. Together with the observation that protein A is required for inflammatory responses in the lung following S. aureus infection, our data suggest that one of the principal functions of S. aureus virulence factors may be to cause lung parenchymal insult, facilitating bacterial survival and evasion of host defenses.

Multiple recent studies have highlighted the association of the Panton-Valentine leukocidin (PVL) with S. aureus strains isolated from patients with severe necrotizing pneumonia (8, 30). Like alpha-toxin and other hemolysins, PVL is a pore-forming toxin whose expression is regulated by agr. The precise role of PVL in pulmonary infection has not been elucidated yet. Considering the data presented here, it is plausible to speculate that S. aureus alpha-toxin and PVL may both have the ability to induce pulmonary inflammation, resulting in systemic manifestations of disease and concomitant mortality. The murine model system described here should allow more rigorous assessment of the role of these cytotoxins and other staphylococcal virulence factors in the pathogenesis of pulmonary infection.