Staphylococcus aureus is a leading cause of community-acquired and health care-associated pneumonia. The virulence of many clinical isolates of S. aureus correlates with production of α-toxin. This pore-forming toxin contributes to pathogenesis of pneumonia through cytolysis of lung epithelial cells, disruption of lung epithelial barrier, and induction of inflammation. We recently reported that type I IFNs protect lung epithelial cells from the cytolytic action of α-toxin by reducing leakage of cellular ATP into extracellular space. The effect is dependent on induction and protein palmitoylation of phospholipid scramblase 1 (PLSCR1). In IFNα-pretreated cells, PLSCR1 co-localizes with endocytosed α-toxin and associates with cytoskeleton. Depletion of PLSCR1 negates IFN-induced protection from α-toxin and enhances sensitivity to inhaled α-toxin and an α-toxin-producing strain of S. aureus. Here, we discuss the potential implications of these findings.
Induction and Roles of IFNs and IFN-Regulated Genes during Staphylococcal Infections
The complex roles of IFNγ, the only type II IFN, in pathogenesis of staphylococcal infections have been previously well established. IFNγ facilitates bacterial clearance by professional phagocytes and promotes acquired immunity during systemic infections but increases inflammation and tissue injury during localized infections (Zhao et al., Immunology 1998; Sasaki et al., FEMS Immunol Med Microbiol 2006). Recent studies uncovered the diverse and important roles of type I IFNs in bacterial infections. Several distinct mechanisms initiate type I IFN induction during staphylococcal infections. Recognition of lipoteichoic acid from S. aureus by TLR2 results in IRF2-dependent production of IFNα from macrophages (Liljeroos et al., Cell Signal 2008). In the presence of anti-staphylococcal IgGs, plasmacytoid dendritic cells recognize endocytosed bacterial DNA and RNA by TLR9 and TLR7, respectively, and secrete massive amounts of IFNα (Parcina et al., J Immunol 2008). Although the functional significance of IFNα from these sources has not been clearly established, activation of Ifnb gene and IFN-signaling is detrimental in a mouse model of staphylococcal pneumonia caused by prevalent methicillin-resistant S. aureus (MRSA) strain USA300 (Martin et al., J Clin Invest 2009). Apparently, recognition of Xr domain in staphylococcal protein A by lung epithelial cells increases Ifnb gene transcription, activates STAT1 and STAT3 and induces IL-6. This results in increased inflammation and mortality in wild-type, but not in IFNAR1−/− mice, which are deficient in the receptor for type I IFNs. Another report described the detrimental role of type I IFNs in the context of post-viral secondary bacterial pneumonia: induction of type I IFNs by influenza virus inhibits Th17-mediated host defense against S. aureus and impairs bacterial clearance (Kudva et al., J Immunol 2011). These studies showed that the roles of type I IFNs are distinct from the roles of IFNγ despite a substantial overlap in the range of their target genes. They also highlighted the need to dissect the roles of specific subsets of IFN-regulated genes in pathogenesis of staphylococcal infections.
Since we previously observed that pretreatment of host cells with exogenous type I IFNs protects them from α-toxin-induced cell death, we were interested in identifying IFN-regulated pathways and genes with potentially beneficial roles. Type I IFNs exert most of their functions through activation of JAK1 and TYK2 tyrosine kinases and signal transducers and activators of transcription (STAT), which regulate transcription of several hundred genes. In addition, some of the metabolic and antiviral effects of type I IFNs are mediated by p38 MAP-kinase, PI3-kinase, protein kinase Cδ and their substrates. To determine which signaling pathways are involved in protection of lung epithelial cells from α-toxin, we used a screening assay using intracellular ATP as readout for cell viability and a panel of well-characterized pharmacological inhibitors. We found that inhibition of p38 MAP-kinase, PI3-kinase and fatty acid synthase activity significantly affected IFN-induced protection from α-toxin, which is consistent with previously described roles of these pathways in cellular defense from pore-forming toxins. Remarkably, inhibition of protein palmitoylation wiped out the protective effects of IFNα. Taking advantage of the previous gene expression studies, which listed IFN-regulated genes in the cell line that we used in the screening, and proteome-wide analyses of protein palmitoylation, we identified PLSCR1 as the leading candidate gene. Although we cannot exclude the possibility that regulation of other genes may contribute to the protective effects of type I IFNs, PLSCR1 is the first identified IFN-inducible gene associated with cellular defense against α-toxin.
Involvement of PLSCR1 in Host Responses to Staphylococcal α-Toxin
Using shRNA-mediated knockdown and PLSCR1-knockout mice, we validated that PLSCR1 plays a protective role after exposure to purified α-toxin or α-toxin-producing strain of S. aureus. Prior studies implicated PLSCR1 protein in translocation of membrane phospholipids and amplification of transcriptional responses to type I IFNs. However, increased expression of PLSCR1 does not correlate with changes in plasma membrane phospholipids after exposure to α-toxin. In addition, overexpression of nuclear localization mutant PLSCR1, which is excluded from the nucleus, is sufficient to increase cell resistance to α-toxin. Thus, the effects of PLSCR1 on gene transcription are not likely to contribute to protection from α-toxin. Since PLSCR1 co-localizes with endocytosed α-toxin and filamentous actin, it is possible that PLSCR1 participates in removal of α-toxin pores from the cell surface. This would be consistent with the previous studies showing that endocytosis, “detoxification” and exocytosis of α-toxin are critical for cellular defense against α-toxin.
Our study provides initial evidence that type I IFN-induced PLSCR1-dependent protection of lung epithelial cells is downstream of ADAM10 (a disintegrin and metalloprotease 10). This zinc-dependent metalloprotease acts as a protein receptor for α-toxin and mediates α-toxin-induced cytotoxicity, cleavage of E-cadherin, and disruption of the lung barrier functions (Inoshima et al., Nat Med 2011). Pretreatment with IFNα does not have significant effects on α-toxin binding or oligomerization, loss of intracellular potassium, activation of p38 MAP-kinase or cleavage of E-cadherin. Furthermore, IFNα pretreatment or deletion of PLSCR1 does not have any significant effect on inflammatory responses to α-toxin or α-toxin-producing S. aureus. However, pretreatment of cells with IFNα or overexpression of PLSCR1 significantly reduces leakage of intracellular ATP into extracellular space. This finding raises the following questions: What is the mechanism of ATP release into extracellular space after exposure to α-toxin? Does extracellular ATP (eATP) contribute to α-toxin-induced cytotoxicity? Are type I IFNs or IFN-regulated genes involved in regulation of ATP release or responses to eATP? Below, we will contemplate involvement of eATP in responses to α-toxin.
ATP Release and Purinergic Receptor Signaling in Responses to α-Toxin and S. aureus
Intracellular ATP is rapidly depleted from host cells after exposure to α-toxin due to reduction of ATP biosynthesis, increased consumption of ATP and release of intracellular ATP into extracellular space, albeit the relative contribution of each of these processes has not been precisely determined. Moreover, it is not known whether ATP escapes the cells through the channels in α-toxin pores or by other mechanisms. Lung epithelial cells use connexin and pannexin hemichannels, maxi-anion channels and exocytosis to release ATP into extracellular space in response to mechanical stress, osmotic shock, fluxes of calcium or proteolysis. In addition, early apoptotic cells release ATP into extracellular space through pannexin hemichannels as “find-me” signals for recruitment of phagocytes. Since α-toxin riggers changes in intracellular sodium and potassium, calcium fluxes, proteolysis and apoptosis, it is likely that ATP is released into extracellular space via multiple mechanisms. A recent study showed that pharmacological inhibition of pannexin channels inhibits α-toxin-induced hemolysis (Skals et al., Pflugers Arch 2011). Our preliminary unpublished data suggest that vesicular exocytosis and connexin hemichannels contribute to ATP release after exposure to α-toxin and that IFNα pretreatment reduces ATP release through connexin hemichannels. We are currently working to determine whether inhibition of ATP release through exocytosis or connexin hemichannels may preserve intracellular ATP and reduce α-toxin cytotoxicity against lung epithelial cells.
Extracellular ATP is rapidly hydrolyzed to ADP, AMP and adenosine by phosphatases and host ectonucleotidases such as CD39 and CD73. Even transient increases of ATP and ADP are readily sensed by P2 purinergic receptors, which are usually associated with proinflammatory and prothrombotic effects of extracellular nucleotides. Seven P2X receptors are ATP-gated membrane ion channels and eight P2Y receptors are G-protein coupled receptors with diverse ligand sensitivity (ATP, ADP, UTP, UDP and UDP-glucose). Signaling through P2X1 and P2X7 receptors apparently contributes to hemolysis by various pore-forming toxins, including staphylococcal α-toxin. In addition, stimulation of P2X7 receptor with extracellular ATP results in activation of NLRP3 inflammasome, which initiates innate immune responses and promotes inflammation. Our own data suggest that oxidized ATP (a nonselective antagonist of P2 receptors) significantly inhibits α-toxin-induced cell death in vitro and reduces the toxic effects of inhaled α-toxin in vivo. Thus, extracellular ATP and its metabolites may contribute to pathogenesis of staphylococcal infections by increasing inflammation and tissue injury.
Sequential hydrolysis of ATP, ADP and AMP by host phosphatases and ectonucleotidases may increase extracellular adenosine. One of the key enzymes in this pathway is CD73 ecto-5′-nucleotidase, which produces adenosine from AMP. Although type I IFNs induce the expression and activity of CD73 expression, it remains to be determined whether upregulation of CD73 occurs during staphylococcal infections. Importantly, pathogenic strains of S. aureus express adenosine synthase, which may utilize extracellular ATP, ADP and AMP as substrates for adenosine synthesis (Thammavongsa et al., J Exp Med 2009). Since adenosine exerts potent anti-inflammatory effects through P1 adenosine receptors (mainly A2A and A2B), accumulation of adenosine may allow S. aureus to escape from phagocytic clearance and killing. Thus, S. aureus may utilize pore-forming toxins to increase the availability of substrates for host ectonucleotidases and staphylococcal adenosine synthase, which may result in excessive accumulation of extracellular adenosine and subsequent impairment of bacterial clearance.
Concluding Remarks
We identified PLSCR1 as a mediator of type I IFN-induced protection from staphylococcal α-toxin. Our study uncovered new functions for PLSCR1 and a mechanism for potentially beneficial effects of type I IFNs during staphylococcal infections. Evidently, PLSCR1 acts downstream of ADAM10 and reduces leakage of cellular ATP into extracellular space, which may have broad implications on pathogenesis of staphylococcal pneumonia.
Acknowledgments
This work has been supported by National Institute of Health grant 5R21AI79322.