An antidote for Staphylococcus aureus pneumonia?

S. aureus encodes a remarkable repertoire of virulence factors. These molecules promote host colonization, facilitate evasion of the human innate immune system, and alter immune responses (for review see reference 8). For the purposes of this commentary, we will limit our discussion to a few S. aureus surface molecules, some of which have been used previously as vaccine targets.

Human neutrophils are a primary cellular defense against bacterial infections. Previous studies have shown that host opsonins, such as serum complement and antibody, play a major role in the phagocytosis of S. aureus by neutrophils (9–11). S. aureus makes several molecules, including protein A, serotype 5 or 8 capsular polysaccharide (CP5 or CP8), and clumping factor A (ClfA), which inhibit phagocytosis (12–14). But despite the bacteria's efforts to evade neutrophils, normal human serum contains a sufficient number of opsonins to promote their rapid uptake by these cells (15). The majority of clinical isolates, including USA300, encode ClfA and CP5 or CP8 (14). Because antibodies specific for CP5 or CP8 enhance phagocytosis, CP5 and CP8 have been evaluated extensively as vaccine antigens (16–18). In the end, however, S. aureus vaccines designed to enhance bacterial uptake have had limited success.

One possible reason for this outcome is the lack of correlation between uptake of the bacteria by neutrophils and their subsequent destruction. For instance, the most prominent CA-MRSA isolates survive relatively well inside neutrophils, probably in part because of their ability to resist the effects of neutrophil-derived reactive oxygen species and antimicrobial peptides (19, 20). The neutrophils, on the other hand, undergo rapid lysis after uptake of these strains (Fig. 1) (15, 21). The ability of S. aureus to survive after phagocytosis has lead some to suggest that neutrophils could be a vector for disseminating bacteria (22, 23). S. aureus can also persist inside macrophages for several days, ultimately causing the death of the cells in a process that depends on Hla (24).

Because uptake does not necessarily correlate with the killing of S. aureus, high titers of anticapsule antibodies, which facilitate uptake, may not protect against disease. This notion is not new, as it has long been known that virtually all humans have circulating antistaphylococcal antibody, and yet some still become infected (25). The idea that antibodies against the bacterial capsule may not provide protection was borne out in two unsuccessful phase III clinical trials designed to test the efficacy of active immunization against the S. aureus antigens CP5, CP8, and ClfA (26).

Hla, a pore-forming cytolytic toxin that assembles as a heptameric β-barrel structure in the plasma membrane of susceptible cells, is arguably the most widely studied S. aureus toxin (for review see reference 27). The toxin is known to cause destruction of a wide-range of host cells, including erythrocytes, epithelial cells, fibroblasts, and monocytes. Hla gained notoriety in 1928 when it was implicated in the deaths of 12 Australian children who had received a diphtheria toxoid vaccine that was later found to be contaminated with an Hla-producing S. aureus strain (27). Although anti-Hla antibody therapy was studied intensively, interest in this approach waned during the antibiotic era. S. aureus encodes numerous other extracellular cytolytic toxins, including δ-hemolysin, γ-hemolysin, Panton-Valentine leukocidin (PVL), leukocidin D/E, a leukocidin homologue (LukM/F′-PV), and the newly described phenol-soluble modulin-like peptides (28). The relative contribution of Hla to human disease as compared with these other virulence factors is not known, in part because susceptibility to Hla varies among different animal species.

Until now, vaccination against Hla has not been tested in an S. aureus pneumonia model. In this issue, Wardenburg and Schneewind show that immunization against Hla prevents S. aureus pneumonia (7). The authors first show that the severity of lung disease in mice correlates with the levels of Hla produced by a particular S. aureus isolate (7). These findings are consistent with a recent study from the same group demonstrating that Hla is important for the pathogenesis of CA-MRSA pneumonia (29).

In the new study, mice were immunized with a nonpore-forming Hla variant, Hla_H135L, and challenged intranasally 3 wk later, a protocol that typically induces lethal pneumonia (29). Immunization with Hla_H135L protected 90–100% mice against all S. aureus strains tested (7). Vaccine-induced protection correlated directly with reduced inflammation and less severe destruction of lung tissue. Passive immunization with Hla antibody 24 h before intranasal challenge with S. aureus also protected animals against an otherwise lethal intranasal challenge with CA-MRSA or an antibiotic-sensitive S. aureus strain (7).

Antibodies against Hla also protected human lung epithelial cells from S. aureus–induced lysis (7). Although these results indicate that Hla contributes to lung tissue destruction, it is not yet clear whether the animals' death resulted from direct destruction of lung cells by the toxin, from an excessive inflammatory response, or from both. Passive transfer of Hla antibodies significantly reduced circulating levels of interleukin 1β, a cytokine known to accompany acute lung injury. Therefore, it is reasonable to conclude that the inflammatory response may contribute to Hla-mediated lung damage (Fig. 2).

There has been considerable debate about whether another S. aureus toxin, PVL, is essential for CA-MRSA virulence. In their report, Wardenburg and Schneewind found that, unlike antibodies specific for Hla, antibodies specific for PVL did not protect mice against S. aureus pneumonia (7). This finding was consistent with an earlier study by this group, which also suggested that PVL is not required for CA-MRSA–induced pneumonia in mice (29).

This result appears to conflict with earlier data suggesting an essential role for PVL (30). But that study relied on laboratory strains of S. aureus that overexpressed PVL, thus clouding the physiological significance of these findings. Indeed, studies conducted using animal models of CA-MRSA disease have unanimously suggested that PVL is dispensable for bacterial virulence (21, 29). Recent work has, however, highlighted the potential importance of other virulence factors in CA-MRSA disease (28, 29).

There are numerous examples of successful vaccination against bacterial toxins, including botulinum, diphtheria, and tetanus toxins. However, these toxins are known to be the primary causative agents of disease induced by their respective organisms. In contrast, S. aureus produces many toxins, and it has been generally accepted that no single S. aureus extracellular molecule can trigger disease on its own. This idea is called into question by the finding that Hla alone is required for S. aureus pneumonia (7, 29). The high level of infections caused by the S. aureus isolate USA300 and the abundance of Hla produced by this strain in vitro suggest that targeting Hla during invasive CA-MRSA infections may be a promising therapeutic approach.

This article was supported in part by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.