Research ArticleInfectious Disease

Staphylococcus aureus α toxin potentiates opportunistic bacterial lung infections

See allHide authors and affiliations

Science Translational Medicine  09 Mar 2016:
Vol. 8, Issue 329, pp. 329ra31
DOI: 10.1126/scitranslmed.aad9922

Toxic eviction

There are many benefits to a good roommate, but the wrong choice can be toxic. Now, Cohen et al. examine the effects of co-habitation on lung infection. They found that α toxin produced by Staphylococcus aureus can worsen lung co-infection by Gram-negative bacteria by preventing acidification of bacteria-containing phagosomes, increasing proliferation, spread, and lethality. However, early treatment or prophylaxis with a neutralizing antibody to α toxin prevented this effect and promoted S. aureus clearance in a humanized mouse model. If this eviction occurs in humans, this approach may reduce co-infection risk in patients colonized with S. aureus.

Abstract

Broad-spectrum antibiotic use may adversely affect a patient’s beneficial microbiome and fuel cross-species spread of drug resistance. Although alternative pathogen-specific approaches are rationally justified, a major concern for this precision medicine strategy is that co-colonizing or co-infecting opportunistic bacteria may still cause serious disease. In a mixed-pathogen lung infection model, we find that the Staphylococcus aureus virulence factor α toxin potentiates Gram-negative bacterial proliferation, systemic spread, and lethality by preventing acidification of bacteria-containing macrophage phagosomes, thereby reducing effective killing of both S. aureus and Gram-negative bacteria. Prophylaxis or early treatment with a single α toxin neutralizing monoclonal antibody prevented proliferation of co-infecting Gram-negative pathogens and lethality while also promoting S. aureus clearance. These studies suggest that some pathogen-specific, antibody-based approaches may also work to reduce infection risk in patients colonized or co-infected with S. aureus and disparate drug-resistant Gram-negative bacterial opportunists.

INTRODUCTION

From the advent of the antibiotic era, the empiric use of broad-spectrum antibiotics has resulted in the emergence and cross-species spread of antibiotic resistance in many opportunistic bacteria (1). Multidrug resistance, coupled with the lack of innovative antibiotic development, threatens to send us back into a preantibiotic era (2, 3). An increasing awareness that broad-spectrum antibiotic use disrupts the beneficial microbiome and increases the risk for multiple metabolic, inflammatory, and infectious diseases has strengthened the rationale for pathogen-specific antibacterial strategies. However, one substantial concern with any pathogen-specific antibacterial approach is that targeting a single species may not provide sufficient coverage in a polymicrobial infection.

Bacterial pneumonia is an important cause of morbidity and mortality, particularly in the hospital setting, in immunocompromised individuals and among the very young and the elderly (4, 5). Although one infecting pathogen may apparently predominate, infection often results from initial colonization or subclinical infection with multiple pathogens (68). For example, Staphylococcus aureus and Pseudomonas aeruginosa are often observed in sequence or together in the lung, and respiratory co-infections with these opportunists correlate with worse clinical outcome than mono-infection with either pathogen (915). Coinfection with S. aureus and Klebsiella pneumoniae resulted in increased lung pathology relative to mono-infections (14). Mortality rate was higher among co-infected individuals; however, this might not be solely attributable to the mixed infection because these patients also had increased rates of underlying illness such as cardiac diseases. Also, >30% of patients with ventilator-associated pneumonia (VAP) suffer from polymicrobial infections, and S. aureus is one of the most commonly cultured pathogens in polymicrobial VAP (16). The severity of the co-infections relative to mono-infections was not clear from this study. Future clinical studies are required to determine if polymicrobial infections are associated with worse outcomes in VAP patients. Here, we investigated whether S. aureus could potentiate infection with Gram-negative bacteria and which S. aureus virulence factors are responsible for this potentiation.

Monoclonal antibody (mAb)–based approaches to prevent pneumonia by precise targeting of individual virulence factors have shown promise in the prevention and treatment of S. aureus or P. aeruginosa pneumonia in preclinical models (1719). We demonstrate that multiple Gram-negative organisms take advantage of the effects of S. aureus α toxin (AT) on mucosal host defense, specifically macrophage function, resulting in proliferation and dissemination of the co-infecting Gram-negative pathogens. It was noteworthy that passive immunization with anti-AT mAb, MEDI4893*, promoted clearance of both S. aureus and the co-infecting Gram-negative pathogen. These data suggest that co-infection with S. aureus expressing AT may be a risk factor for Gram-negative disease and demonstrate that pathogen-specific, antibody-based approaches for prophylaxis or therapy could be effective against mixed opportunistic infections without resorting to empiric broad-spectrum antibiotics.

RESULTS

S. aureus enhances Gram-negative bacterial proliferation and lethality

To examine the impact and growth of S. aureus and P. aeruginosa in a mixed infection, mice were coinfected intranasally with sublethal doses of S. aureus (SF8300, CC8) and P. aeruginosa (6077, ExoU+). Although all animals survived sublethal challenge with either P. aeruginosa or S. aureus, simultaneous intranasal infection with S. aureus + P. aeruginosa resulted in 90 to 100% mortality by 7 days after infection (Fig. 1A). Numbers of S. aureus recovered from both mono- and mixed-infected lungs were similar up to 24 hours after challenge (Fig. 1B and table S1). However, at 48 hours, the number of S. aureus colony-forming units (CFU) in the S. aureus + P. aeruginosa–infected mice trended higher (P = 0.286) than in the S. aureus mono-infected mice. P. aeruginosa burden was similar in the mono- and mixed-infected lungs at up to 8 hours, and by 24 hours after infection, P. aeruginosa numbers were reduced in the lungs of mono-infected mice. In contrast, striking proliferation of P. aeruginosa in the lungs of S. aureus + P. aeruginosa–co-infected mice was evident from 8 to 24 hours (P = 0.0005) and 48 hours (P < 0.0001) after infection with significantly more P. aeruginosa recovered (P < 0.0001) compared to mono-infection. Pseudomonas proliferation in the lung was accompanied by dissemination as evidenced by increased numbers of P. aeruginosa in the spleen (P = 0.018) 48 hours after infection relative to animals that had been challenged with P. aeruginosa alone (Fig. 1C and table S1). To investigate whether S. aureus could potentiate infection with other Gram-negative bacteria, mice were challenged intranasally with either K. pneumoniae or Acinetobacter baumannii alone or in combination with S. aureus. Although most animals survived the mono-infections, 90 to 100% lethality was observed upon co-infection coupled with increased K. pneumoniae and A. baumannii bacterial burden in the lungs and spleens of co-infected animals (Fig. 1, D to I, and table S1). These data indicate that S. aureus can potentiate proliferation, dissemination, and disease with a range of clinically relevant Gram-negative opportunistic bacterial pathogens.

Fig. 1. S. aureus promotes Gram-negative infection.

(A) Survival of mice infected with S. aureus (5e7 CFU) (Sa), P. aeruginosa (1e5 CFU) (Pa), or S. aureus + P. aeruginosa. (B) S. aureus or P. aeruginosa CFU recovered from the lungs of mice infected as described above at various time points. (C) P. aeruginosa CFU recovered from the spleens of mice after infection with P. aeruginosa or S. aureus + P. aeruginosa at various time points. (D) Survival of mice infected with S. aureus (5e7 CFU), K. pneumoniae (5e1 CFU) (Kp), or S. aureus + K. pneumoniae. (E) S. aureus or K. pneumoniae CFU recovered from the lungs of mice infected as described above at various time points. (F) K. pneumoniae CFU recovered from the spleens of mice after infection with K. pneumoniae or S. aureus + K. pneumoniae at various time points. (G) Survival of mice infected with S. aureus (5e7 CFU), A. baumannii (6e5 CFU) (Ab), or S. aureus + A. baumannii. (H) S. aureus or A. baumannii CFU recovered from the lungs of mice infected as described above at various time points. (I) A. baumannii CFU recovered from the spleens of mice after infection with A. baumannii or S. aureus + A. baumannii at various time points. Significance was determined by log-rank test (A, D, and G) or analysis of variance (ANOVA) followed by Dunnett’s test (B, C, E, F, H, and I). ΦP < 0.02, compared to mixed infection; #P < 0.0001, P. aeruginosa versus S. aureus + P. aeruginosa (P. aeruginosa) and P < 0.03, S. aureus + P. aeruginosa (P. aeruginosa) versus S. aureus + P. aeruginosa (t = 4 hours) (P. aeruginosa); and *P < 0.03 mixed infection (Gram-negative) versus Gram-negative. Data are representative of at least three independent experiments. Exact P values in table S1.

Pathogen-specific mAbs reduce co-infection disease severity

We previously demonstrated that treatment with MEDI3902, a bispecific mAb targeting P. aeruginosa, protected animals from intranasal challenge with P. aeruginosa (17). Here, immunoprophylaxis with MEDI3902 significantly increased survival (P < 0.0001) of mice co-infected with S. aureus + P. aeruginosa and reduced both S. aureus and P. aeruginosa CFU in the lungs compared to an isotype control immunoglobulin G (c-IgG) (P = 0.0022 and P = 0.0023, respectively) (Fig. 2, A and B). Likewise, immunoprophylaxis with MEDI4893*, a mAb targeting S. aureus AT also shown to protect mice against lethal S. aureus challenge (18, 20), increased survival of S. aureus + P. aeruginosa–co-infected animals and reduced the lung burden of both bacteria by greater than 1 log (Fig. 2, C and D). MEDI4893* had a similar effect when animals were infected with P. aeruginosa along with other strains of S. aureus, NRS387 (CC5) and NRS261 (CC30) (fig. S1), as well as S. aureus + K. pneumoniae and S. aureus + A. baumannii mixed infections (fig. S2, A to D).

Fig. 2. Pathogen-specific mAbs reduce disease severity.

(A) Survival of mice treated with MEDI3902 (15 mg/kg) or c-IgG at the indicated times before or after infection with S. aureus (5e7 CFU) + P. aeruginosa (1e5 CFU). (B) S. aureus or P. aeruginosa CFU recovered from the lungs of infected (24 hours, S. aureus + P. aeruginosa) mice treated with MEDI3902 (15 mg/kg) or c-IgG 24 hours before infection. (C) Survival of mice treated with MEDI4893* (15 mg/kg) at the indicated times before or after infection with S. aureus + P. aeruginosa. (D) S. aureus or P. aeruginosa CFU recovered from the lungs of infected (24 hours, S. aureus + P. aeruginosa) mice treated with MEDI4893* (15 mg/kg) or c-IgG 24 hours before infection. Significance was determined by log-rank test (A and C) or Mann-Whitney test (B and D). *P < 0.0001, #P = 0.0105. Data are representative of at least three independent experiments (A to D).

To assess whether pathogen-specific mAbs could treat a co-infection, MEDI4893* and MEDI3902 were administered at various times after S. aureus + P. aeruginosa infection. Whereas MEDI3902 conferred significant (P < 0.0001) protection when administered as late as 8 hours after co-infection (Fig. 2A), MEDI4893* afforded significant (P = 0.0105) protection only when administered up to 1 hour after infection (Fig. 2C). These data suggest that at the onset of infection, S. aureus suppresses innate defense mechanisms, enabling the proliferation of P. aeruginosa, which ultimately results in death of the co-infected animals.

AT is sufficient for promotion of enhanced Gram-negative pneumonia

The above studies suggested that AT plays a key role in the S. aureus–mediated potentiation of Gram-negative growth and dissemination during a mixed infection; however, S. aureus produces an array of virulence factors that contribute to disease pathogenesis (21). To confirm a role for AT and identify other potential virulence factors that may promote lethality in the mixed infection, S. aureus isogenic in-frame deletion mutants, ΔagrA, Δpvl, ΔspA, ΔclfA, ΔisdH, Δpsm α-type 1,2,3,4, Δhla, and Δhla comp:hla, were tested in the P. aeruginosa mixed infection model (Fig. 3, A and B) (2226). P. aeruginosa co-infection with all of the S. aureus mutants except for those lacking the gene for AT (Δhla), or the virulence regulator Δagr, resulted in mortality similar to that observed with P. aeruginosa + wild-type S. aureus. Significantly fewer S. aureus (P = 0.0118) and P. aeruginosa (P < 0.0001) were recovered from the lungs of mice infected with P. aeruginosa + Δhla compared to those infected with P. aeruginosa + wild-type S. aureus (Fig. 3, C and D). Similar results were obtained during K. pneumoniae + Δhla and A. baumannii + Δhla mixed infections (fig. S3, A to D).

Fig. 3. AT is sufficient for potentiation of P. aeruginosa.

(A) Survival of mice infected with S. aureus mutants and P. aeruginosa. (B) Survival of mice infected with S. aureus (5e7 CFU), Δhla S. aureus, Δhla S. aureus comp:hla, AT (0.05 μg), or ATH35L (0.05 μg) and P. aeruginosa (1e5 CFU). (C) P. aeruginosa CFU recovered from lungs of mice 24 hours after infection with S. aureus (5e7 CFU), Δhla S. aureus, AT (0.05 μg), or ATH35L (0.05 μg) and P. aeruginosa (1e5 CFU). (D) S. aureus CFU recovered from lungs of mice 24 hours after infection with S. aureus (5e7 CFU) and Δhla S. aureus and P. aeruginosa (1e5 CFU). (E) S. aureus or P. aeruginosa CFU recovered from the lungs or spleen of infected (24 hours, S. aureus + P. aeruginosa) humanized mice treated with MEDI4893* (15 mg/kg) or c-IgG 24 hours before infection. Significance was determined by Mann-Whitney test (C to E). Data are representative of at least three (A to D) or two (E) independent experiments.

To further confirm a pivotal role for AT in co-infection potentiation, mice were challenged with either a mixture of P. aeruginosa + 0.05-μg purified AT protein or P. aeruginosa + 0.05-μg ATH35L, an AT mutant defective for pore formation. A high rate of mortality was observed in the sublethal infections only with the functional wild-type AT protein (Fig. 3, B and C). Likewise, sublethal doses of K. pneumoniae or A. baumannii + AT resulted in high rates of mortality, whereas mixtures of K. pneumoniae and A. baumannii with ATH35L did not (fig. S3, E to H). These data indicate that active AT is necessary and sufficient for enabling P. aeruginosa, K. pneumoniae, and A. baumannii proliferation and associated mortality in this murine lung infection model.

A common criticism of murine models is that murine cells are not sensitive to all S. aureus toxins; therefore, we tested the ability of AT neutralization to protect against a mixed infection in human interleukin-3/granulocyte-macrophage colony-stimulating factor (IL-3/GM-CSF) knock-in humanized mice (27). Immunoprophylaxis with MEDI4893* resulted in greater than 1-log reduction in the numbers of S. aureus and P. aeruginosa recovered from either the lung or spleen (Fig. 3E). These data suggest that prophylaxis with pathogen-specific mAbs can protect against a co-infection and that neutralization of a single S. aureus virulence factor, AT, can reduce proliferation and accompanying mortality associated with multiple Gram-negative pathogens.

AT affects alveolar macrophage phagocytosis of S. aureus in vivo

We first hypothesized that the mortality and Gram-negative proliferation associated with the mixed infection was due to excessive inflammation resulting from the presence of multiple pathogens. However, cytokine levels in bronchoalveolar lavage fluid (BALF) of mice 8 hours after S. aureus + P. aeruginosa infection were comparable to those observed in mice infected with S. aureus alone (fig. S4A). Prophylaxis with MEDI4893* did not alter lung cytokine levels with the exception of IL-1β, which was reduced. In addition, animals challenged with mixtures of P. aeruginosa + AT had elevated IL-1β lung levels compared to mice challenged with sublethal doses of P. aeruginosa alone. However, inhibition of IL-1 receptor signaling with Anakinra did not protect mice from a lethal mixed infection, suggesting that AT-mediated proliferation of Gram-negative bacteria was not caused by changes in IL-1β levels (fig. S4B) (28). Histological differences were not observed between mono- and mixed-infected lungs 24 hours after infection, and epithelial barrier dysfunction could be attributed entirely to the presence of P. aeruginosa (fig. S4, C and D). Overall, our results did not indicate that AT drives excessive lung inflammation in this co-infection model.

To better elucidate how AT alters innate immunity, we first studied its impact on S. aureus clearance in mono-infections. Both alveolar macrophages (AMs; CD11c+CD11b) and neutrophils (CD11cCD11b+Ly6G+) contribute to host clearance of S. aureus in mono-infection; depletion of either cell type resulted in greater quantities (P < 0.0001) of bacteria in the lungs (Fig. 4, A and B, and fig. S5, A and B) (29). Similar to previous reports, MEDI4893* prophylaxis significantly (P < 0.004) improved S. aureus clearance from the lungs of infected mice (Fig. 4C) (18). AT neutralization did not significantly alter AM, neutrophil, or inflammatory monocyte numbers in the lungs 4 and 24 hours after infection relative to c-IgG–treated mice (Fig. 4, D to F). Expression of the regulatory markers PD-L1 or CD200R on AMs and neutrophils was unaltered by MEDI4893* prophylaxis (fig. S5, C to F). These data indicated that although macrophages and neutrophils are required for optimal host defense against S. aureus, AT does not influence the numbers of these cells in the lung nor their immunologic phenotype.

Fig. 4. Macrophages and neutrophils contribute to clearance of S. aureus.

(A) S. aureus CFU recovered from the lungs of control [phosphate-buffered saline (PBS)]– or AM (clodronate)–depleted mice 24 hours after infection with S. aureus (5e7 CFU). (B) S. aureus CFU recovered from the lungs of control (c-IgG)– or neutrophil (anti-Ly6G)–depleted mice 24 hours after infection with S. aureus (5e7 CFU). (C) S. aureus CFU recovered from the lungs of infected (S. aureus 5e7 CFU, 24 hours) mice treated with MEDI4893* or c-IgG 24 hours before infection. (D) Numbers of AMs in the lungs of naïve mice or mice treated with MEDI4893* or c-IgG 24 hours before infection with S. aureus (5e7 CFU) for 4 or 24 hours. (E) Numbers of neutrophils in the lungs of naïve mice or mice treated with MEDI4893* or c-IgG 24 hours before infection with S. aureus (5e7 CFU) for 4 or 24 hours. (F) Numbers of inflammatory monocytes in the lungs of naïve mice or mice treated with MEDI4893* or c-IgG 24 hours before infection with S. aureus (5e7 CFU) for 4 or 24 hours. (G) Fluorescence-activated cell sorting (FACS) analysis of S. aureus association with macrophages in the lungs of MEDI4893* (15 mg/kg)– or c-IgG–treated (24 hours prior) mice infected with S. aureus (5e7 CFU, 4 hours). (H) FACS analysis of S. aureus association with neutrophils in the lungs of MEDI4893* (15 mg/kg)– or c-IgG–treated (24 hours prior) mice infected with S. aureus (5e7 CFU, 4 hours). (I) Confocal images of CD11c+ (green) cells containing S. aureus (yellow, identified by red arrows) recovered from BALF of MEDI4893* (15 mg/kg)– or c-IgG–treated (24 hours prior) mice infected with S. aureus (5e7 CFU, 4 hours). n = 3 images per animal, 3 animals per experiment, 3 experiments. Scale bars, 4 μm. Significance was determined by Mann-Whitney test. Data are representative of at least three independent experiments (A to H).

We next investigated whether AT alters phagocytosis of S. aureus in vivo. AT neutralization significantly increased S. aureus association with AM (P = 0.0039), not neutrophils, 4 hours after infection as measured by flow cytometry (Fig. 4, G and H, and fig. S6). A similar increase in bacterial uptake by AMs is observed in mice infected with Δhla S. aureus as compared with wild-type S. aureus (fig. S7). Confocal microscopy confirmed that these bacteria were not surface-associated but internalized by AM (Fig. 4I). An increased number of S. aureus were observed within CD11c+ cells recovered from MEDI4893*-treated animals as compared to c-IgG–treated animals, confirming bacterial uptake into these phagocytic cells.

Macrophage phagocytosis of S. aureus is thought to be influenced by natural killer (NK) cells; thus, we tested the effect of AT on NK cell function in vitro. Human NK cells were exquisitely sensitive to AT-mediated lysis due to their high level of ADAM10 expression, the receptor for AT (Fig. 5, A and B). At sublytic concentrations, AT prevented interferon-γ (IFN-γ) production by human NK cells in response to IL-12 stimulation (Fig. 5C). Consistent with the in vitro results, neutralization of AT in vivo significantly increased the number of IFN-γ+ NK+CD3 cells (not CD3+ cells) in the lung (P = 0.0015) 24 hours after infection as compared with c-IgG (Fig. 5, D and E), suggesting that AT may influence macrophages indirectly by altering NK function. To investigate the impact of NK cells on bacterial clearance from lungs, anti-NK1.1 was used to deplete NK cells before S. aureus infection (Fig. 5F and fig. S8) (30). Mice treated with anti-NK1.1 cleared S. aureus from their lungs as effectively as mice treated with c-IgG despite reduced IFN-γ levels (P = 0.03) in the airways (Fig. 5, G and H), suggesting that in vivo AT reduces S. aureus phagocytosis by AMs directly and not through modulation of NK activation.

Fig. 5. NK cell function is altered by AT.

(A) AT induced cytotoxicity of human neutrophils, NK cells, and PBMCs over 4 hours of incubation, measured by live/dead staining. (B) FACS analysis of ADAM10 expression on human PMNs (polymorphonuclear leukocyte), PBMCs and NK cells. MFI, mean fluorescence intensity. (C) In vitro stimulation of IFN-γ release from human NK cells by IL-12 in the presence of increasing amounts of AT. (D) Percent of NK cells expressing IFN-γ in the lungs of MEDI4893* (15 mg/kg)– or c-IgG–treated (24 hours prior) mice infected with S. aureus (5e7 CFU, 24 hours). (E) Percent of CD3+ cells expressing IFN-γ in the lungs of MEDI4893* (15 mg/kg)– or c-IgG–treated (24 hours prior) mice infected with S. aureus (5e7 CFU, 4 hours). (F) Numbers of NK cells in the lungs of mice 24 hours after treatment with c-IgG or anti-NK1.1 antibody. (G) S. aureus CFU recovered from the lungs of control (c-IgG)– or NK cell (anti-NK1.1)–depleted mice 24 hours after infection with S. aureus (5e7 CFU). (H) IFN-γ levels in BALF 24 hours after infection with S. aureus + P. aeruginosa in control (c-IgG)– or NK cell (anti-NK1.1)– depleted mice. Significance was determined by Mann-Whitney test (D to H). Data are representative of at least two independent experiments.

AT prevents lysosome acidification and bacterial killing via calpain activation

Although neutralization of AT improves S. aureus phagocytosis by AMs in vivo, it was unclear if neutralization of AT increases bacterial killing within macrophages. Processing of S. aureus was visualized in RAW 264.7 cells or human peripheral blood monocytes (hPBMCs) (31) that were pretreated with LysoTracker to visualize acidic cellular compartments. In both murine macrophages and hPBMCs, S. aureus colocalized with LysoTracker-positive compartments in the presence of MEDI4893* but not in c-IgG–treated cells (Fig. 6A and figs. S9 and S10). Macrophages were incubated with fluorescein isothiocyanate (FITC)–labeled S. aureus, and loss of FITC signal, which correlates with acidification of the bacteria’s microenvironment, was measured by flow cytometry. FITC signal intensity was significantly reduced in macrophages incubated with MEDI4893* as compared with c-IgG (P < 0.0001) or S. aureus Δhla as compared with wild-type S. aureus (P = 0.0075) (Fig. 6B). S. aureus Δhla colocalized with LysoTracker-positive compartments in hPBMCs co-incubated with ATH35L but not AT, confirming the role of AT in altering cellular processing of S. aureus (Fig. 6C).

Fig. 6. AT prevents lysosomal acidification.

(A) Confocal images of RAW cells infected with S. aureus [multiplicity of infection (MOI) of 10, 1 hour] in the presence of MEDI4893* or c-IgG. (B) FACS analysis of acidification of the bacterial microenvironment in macrophages infected with S. aureus in the presence of c-IgG, MEDI4893*, or Δhla S. aureus. (C) Confocal images of lysosomal acidification (red) in hPBMCs infected with Δhla S. aureus (MOI 10, 1 hour) in the presence of AT (1 μg/ml) or ATH35L (1 μg/ml); DNA (bacterial and eukaryotic) is labeled blue. (D) Calpain activation in RAW cells treated with AT or ATH35L for 1 hour at the indicated concentrations. (E) Confocal images of lysosomal acidification (red) in hPBMCs infected with S. aureus (MOI 10, 1 hour) in the presence of calpain inhibitor (100 μM) or dimethyl sulfoxide (DMSO); DNAs (both bacterial and eukaryotic) are labeled blue. (F) Percent of S. aureus killed after a 1-hour incubation with RAW cells (MOI 1) in the presence of MEDI4893* (10 μg/ml) or c-IgG. (G) Percent of S. aureus killed after a 1-hour incubation with RAW cells (MOI 1) in the presence of calpain inhibitor (100 μM) or DMSO. (H) Percent of S. aureus killed after a 1-hour incubation with hPBMCs (MOI 1) in the presence of MEDI4893* (10 μg/ml) or c-IgG. (I) Percent of S. aureus killed after a 1-hour incubation with human PMNs (MOI 1) in the presence of MEDI4893* (10 μg/ml) or c-IgG. Percent killed calculated as a ratio of (innocula − recovered CFU)/innocula. Significance was determined by t test (B and E to H). Data are representative of at least three independent experiments. Scale bars, 4 μm (A, C, and E).

Clostridium septicum AT, a structurally related pore-forming toxin, causes membrane Ca2+ fluxes and activates the Ca2+-dependent protease calpain that cleaves the lysosomal protein Lamp2 (31). S. aureus AT, but not ATH35L, resulted in a dose-dependent increase in calpain activation in macrophages (Fig. 6D). Incubation of hPBMCs with a calpain inhibitor restored LysoTracker colocalization with internalized S. aureus, suggesting that AT-dependent calpain activation prevents lysosomal acidification (Fig. 6E). Bacterial killing assays were performed to determine if AT neutralization or calpain inhibition promotes S. aureus killing by macrophages. Significantly fewer bacteria (P = 0.0177) were recovered from macrophages in the presence of MEDI4893* as compared with c-IgG, or the calpain inhibitor as compared with a DMSO control (P = 0.0026), demonstrating increased bacterial killing when either AT or calpain activation was neutralized (Fig. 6, F and G). Killing of S. aureus by hPBMCs, not human neutrophils, was increased (P = 0.0249) in the presence of MEDI4893* (Fig. 6, H and I). These results indicate that AT alters S. aureus uptake and killing within macrophages via calpain activation.

Altered macrophage function inhibits killing of Gram-negative pathogens

The results above indicate that AT alters S. aureus killing within macrophages, but not neutrophils, and that this effect is independent of NK cell function in vivo. To gain further insight into the role for these cells during a Gram-negative lung infection, mice lacking macrophages, neutrophils, or NK cells were infected intranasally with P. aeruginosa or K. pneumoniae. Although impaired clearance of both Gram-negative bacterial species was observed in mice lacking macrophages, P. aeruginosa but not K. pneumoniae clearance was impaired by depleting neutrophils, and K. pneumoniae but not P. aeruginosa lung burden was affected by a reduction in NK cells (Fig. 7, A to C, and fig. S11, A to C). We used a green fluorescent protein (GFP)–P. aeruginosa strain to monitor phagocytosis of P. aeruginosa during an S. aureus + P. aeruginosa mixed infection after MEDI4893* or c-IgG prophylaxis. Similar numbers of GFP–P. aeruginosa were associated with macrophages or neutrophils in mice treated with MEDI4893* or c-IgG 4 hours after infection (fig. S11, D and E). However, although AT did not have an impact on P. aeruginosa phagocytosis, we found that the ability of hPBMCs to kill P. aeruginosa (P = 0.0145) or K. pneumoniae (P = 0.0004) in vitro was impaired in the presence of recombinant AT, whereas human neutrophil bactericidal function was not affected (Fig. 7, D and E, and fig. S11, F and G). hPBMCs were exposed to GFP–P. aeruginosa alone or with S. aureus at a 10:1 (GFP–P. aeruginosa/S. aureus) ratio, and bacterial processing was observed by confocal microscopy (Fig. 7F and fig. S12). In cells infected with GFP–P. aeruginosa alone, bacteria colocalized with LysoTracker-positive compartments, whereas GFP–P. aeruginosa were dispersed throughout the cytoplasm in cells infected with GFP–P. aeruginosa + S. aureus. Neutralization of AT restored GFP–P. aeruginosa trafficking to LysoTracker-positive compartments in cells co-infected with GFP–P. aeruginosa + S. aureus (Fig. 7G and fig. S12). We next tested the effect of AT neutralization on a mixed infection in mice lacking AMs (Fig. 7H and fig. 11H). Mortality in mice lacking AM and infected with either S. aureus + P. aeruginosa or S. aureus + K. pneumoniae was not reduced by MEDI4893*. Collectively, these data indicate that AT adversely modulates macrophage killing of Gram-positive and Gram-negative bacteria in the lung, promoting Gram-negative proliferation and enhancing lethality during infection with multiple pathogens.

Fig. 7. Macrophage processing of P. aeruginosa is altered by AT.

(A) P. aeruginosa CFU recovered from the lungs of control (PBS)– or AM (clodronate)–depleted mice 24 hours after infection with P. aeruginosa (1e5 CFU). (B) P. aeruginosa CFU recovered from the lungs of control (c-IgG)– or neutrophil (anti-Ly6G)–depleted mice 24 hours after infection with P. aeruginosa (1e5 CFU). (C) P. aeruginosa CFU recovered from the lungs of control (IgG)– or NK cell (anti-NK1.1)–depleted mice 24 hours after infection with P. aeruginosa (1e5 CFU). (D) Percent of P. aeruginosa killed after 1 hour incubation with hPBMCs (MOI 1) in the presence of AT (0.1 μg/ml) or ATH35L (0.1 μg/ml). (E) Percent of K. pneumoniae killed after 1-hour incubation with hPBMCs (MOI 1) in the presence of AT (0.1 μg/ml) or ATH35L (0.1 μg/ml). (F) Confocal images of human macrophages infected with P. aeruginosa or S. aureus + P. aeruginosa (1:10 ratio) for 1 hour. Blue, DNA; magenta, actin; green, P. aeruginosa; yellow, LysoTracker. (G) Confocal images of human macrophages infected with S. aureus + P. aeruginosa (1:10 ratio) for 1 hour in the presence of MEDI4893* (10 μg/ml) or c-IgG. Scale bars, 4 μm. (H) Survival of macrophage-depleted (clodronate) mice treated with MEDI4893* (15 mg/kg) or c-IgG 24 hours before infection with S. aureus (5e7 CFU) + P. aeruginosa (1e5 CFU). Significance was determined by Mann-Whitney test (A to C), t test (D and E), or log-rank test (G). Data are representative of at least three independent experiments.

DISCUSSION

Pathogen-specific approaches for preventing or treating serious bacterial infections are currently progressing through the antibacterial pipeline. Whether these approaches can work against mixed infections is a pressing question. Aspiration of bacteria from the diverse flora of the upper airway or gastrointestinal tract into the smaller airways of the lower lung can result in life-threatening bacterial pneumonias, often involving two or more bacterial species (3336). Epidemiologic studies identified a strong correlation between S. aureus colonization in the upper airway and onset of VAP, and a transition from S. aureus to P. aeruginosa as the primary pathogens in early- to late-onset VAP (37, 38). A similar transition is observed in cystic fibrosis patients (3941). This sequence of infections suggests that S. aureus colonization and co-infection may promote infection by other bacterial species.

We describe a murine lung infection model in which S. aureus in combination with different Gram-negative bacteria results in significantly increased lethality compared to mono-infection (Fig. 8). These studies suggest that AT can potentiate growth and dissemination of multiple Gram-negative bacterial species including those that are the most problematic with respect to antibiotic resistance. Neutralization of AT with MEDI4893* reduced S. aureus CFU, prevented outgrowth of Gram-negative pathogens, and reduced mortality in this model. The potential of a pathogen-specific approach was further highlighted by targeting proliferating P. aeruginosa with MEDI3902, which also resulted in protection without resulting in S. aureus proliferation and lethality.

Fig. 8. S. aureus enhances Gram-negative proliferation through the activity of AT.

(A) Gram-negative mono-infection. (B) Mixed infection.

Previous studies have described mechanisms through which P. aeruginosa outcompetes or interacts with the host immune system to eradicate S. aureus (10, 42, 43). Our data describe a mechanism in which P. aeruginosa takes advantage of S. aureus AT to evade clearance by macrophages in the lung. In contrast to a lethal S. aureus mono-infection, we did not observe significant S. aureus–dependent epithelial damage during lethal mixed infection. Furthermore, AT neutralization during a mixed infection did not dampen the inflammatory cytokine response, nor did it influence phagocytosis of S. aureus or P. aeruginosa by neutrophils (18, 44). At sublytic levels, AT inhibited endosomal/lysosomal processing of phagocytosed bacteria in macrophages, not in neutrophils. The effect of AT correlated with their respective levels of ADAM10 expression, as we, and others, observe low ADAM10 levels on neutrophils in comparison with macrophages (44).

AMs are the primary phagocytic cells present in the uninfected lung and are therefore the first immune cells encountered by infecting pathogens, coordinating the innate immune response to bacterial pathogens (4548). Our data suggest that they play a substantial role in the acute response to bacterial pathogens in the airway. Others have described macrophages as protective cells in models of respiratory bacterial infection, although some studies have found that AMs can contribute to disease pathology (46, 47, 4953). NK cells have been described to contribute to host defense by regulating macrophage killing function (52, 54, 55). Whereas AMs contributed to bacterial clearance, NK cells were only required for clearance of K. pneumoniae in our model, despite increased sensitivity (and ADAM10 expression) of NK cells to AT as compared to either neutrophils or PBMCs. Our results demonstrate that AT impairs AM clearance of both S. aureus and the co-infecting Gram-negative pathogen during a mixed lung infection.

Translation of these preclinical results to the clinical setting is limited by the use of murine models. Humans, unlike mice, are sensitive to a wider range of S. aureus toxins, which could contribute to the pathogenesis of human infection (56, 57). Members of the leukotoxin family (HlgAB, HlgCB, LukAB, LukSF, and LukDE) and the S. aureus superantigens likely contribute to the pathogenesis of S. aureus disease in humans (58). Phenol-soluble modulins can also enable S. aureus escape from the lysosome of nonprofessional and professional phagocytes (22). To bridge the gap from murine models to the clinical setting, we used IL-3/GM-CSF knock-in humanized mice. These animals have a humanized immune system and murine nonhematopoietic cells and are sensitive to a wider range of S. aureus toxins as suggested by our data (Fig. 3E) and others (59). Neutralization of AT in the humanized mice promoted clearance of both S. aureus and P. aeruginosa. In vitro, we demonstrated that lysosomal acidification properly occurs in human monocytes during infection with bacteria lacking only AT, and neutralization of this toxin in the context of wild-type S. aureus improves bacterial killing. We are also limited by the acute nature of in vivo murine models. Modeling chronic infection in a murine model has proven difficult, and therefore, we are unable to determine how S. aureus might alter a preexisting infection. Our results suggest that S. aureus might exacerbate a preexisting infection, but clinical studies are required to demonstrate this empirically. Therefore, although targeting alternate toxins may provide additional benefit in the clinical context of mixed infections, the wide strain coverage of an anti-AT antibody such as MEDI4893* and the established importance of AT as a pivotal virulence factor of S. aureus make AT an attractive therapeutic target against S. aureus.

In conclusion, we demonstrate in a murine mixed infection model that S. aureus AT greatly potentiates expansion and rapid dissemination of different opportunistic Gram-negative bacteria. Further, a mAb targeting a key virulence factor such as AT, or a mAb targeting the proliferating Gram-negative opportunist, protected against a lethal mixed bacterial infection in this murine model. These studies indicate that pathogen-specific mAb therapies have the potential to reduce the risk of serious respiratory infections with different bacterial pathogens and hold potential to treat them in the context of a mixed infection. Clinical studies are necessary to further test this hypothesis and to better define patient populations who could benefit from alternative or adjunctive precision medicine strategies to the current broad-spectrum antibiotic treatment paradigm.

MATERIALS AND METHODS

Additional materials and methods, as well as source data for all figures (table S3), can be found in the Supplementary Materials.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/8/329/329ra31/DC1

Materials and Methods

Fig. S1. MEDI4893* protects against mixed infection with diverse S. aureus strains.

Fig. S2. MEDI4893* prevents mortality associated with mixed infection of S. aureus and either K. pneumoniae or A. baumannii.

Fig. S3. AT potentiates infection with either K. pneumoniae or A. baumannii.

Fig. S4. Mixed infection does not result in excessive tissue damage or inflammation.

Fig. S5. Immune cell populations in the lung.

Fig. S6. AT neutralization increases phagocytosis of S. aureus by AMs.

Fig. S7. AT reduces phagocytosis of S. aureus by AMs.

Fig. S8. Anti-NK1.1 depletion of NK cells.

Fig. S9. AT prevents colocalization of S. aureus with acidic lysosomes.

Fig. S10. AT reduces lysosomal acidification in human cells.

Fig. S11. Contribution of immune cells to the clearance of P. aeruginosa or K. pneumoniae from the lung.

Fig. S12. Confocal imaging analysis of P. aeruginosa internalization by hPBMCs.

Fig. S13. Flow cytometry gating strategy and isotype controls.

Table S1. Exact P values for CFU data in Fig. 1.

Table S2. Bacterial strains.

Table S3. Source data for all figures (Excel).

REFERENCES AND NOTES

Acknowledgments: We thank M. Mazaitis for his work in designing Fig. 8. Funding: Research was supported by MedImmune, a member of the AstraZeneca group, and by the U.S. Public Health Service Grant NIH R01 AI087674 to B.A.D. Author contributions: T.S.C., J.J.H., C.K.S., and B.R.S. designed the studies; T.S.C., J.J.H., O.J.-N., A.E.K., C.T., A.D., M.H., M.P., and Q.W. performed the experiments; L.C. analyzed lung histology; T.O. conducted statistical analysis; B.A.D. and V.T.M.L. provided bacterial strains; and T.S.C., J.J.H., B.A.D., L.C., J.S., C.K.S., and B.R.S. wrote the manuscript. Competing interests: This work was funded by MedImmune, a member of the AstraZeneca group. All authors except B.A.D. and V.T.M.L. were employed by MedImmune when work was executed and may currently hold AstraZeneca stock or stock options. Data and materials availability: Patents describing the work in this paper have been filed by MedImmune: WO/2012/109285.
View Abstract

Navigate This Article