Staphylococcus aureus (S. aureus) is a serious human health threat because of its ability to cause a wide array of diseases, ranging from skin and soft tissue infections to more life-threatening conditions such as sepsis, endocarditis, and brain abscess (Lowy 1998). Methicillin-resistant Staphylococcus aureus (MRSA) has emerged as a frequent cause of hospital- and community-associated infections, which complicates treatment (Drago et al. 2007) and MRSA isolates have been recovered from brain abscesses (Sifri et al. 2007; Naesens et al. 2009; Arora et al. 2012; David et al. 2012). Brain abscesses can originate from many sources, including: (i) subacute or chronic otitis media, mastoiditis, frontal or ethmoid sinusitis, or dental infection; (ii) hematogenous dissemination from peripheral sites of infection, including bacterial endocarditis or lung abscesses (Madsen 1983; Kassis et al. 2010; Wadhwa et al. 2012); (iii) immunocompromised patients with HIV infection, organ transplantation, chemotherapy, or steroid usage (Young and McGwire 2005; Muzumdar et al. 2011); (iv) penetrating trauma to the calvarium; or (v) infectious complications following neurosurgical procedures. Without timely treatment intervention, a brain abscess can rupture, which is associated with a high mortality rate. Therefore, elucidating mechanisms of protection as well as pathophysiology elicited by S. aureus after CNS colonization may unveil novel therapeutic targets.
Numerous proinflammatory cytokines are elicited during acute CNS S. aureus infection, including interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-alpha (TNF-α) (Kielian et al. 2004, 2007b; Stenzel et al. 2005; Garg et al. 2009; Xiong et al. 2012). Our prior studies utilizing IL-1β and IL-1RI knockout (KO) mice demonstrated a critical role for IL-1 signaling during early brain abscess pathogenesis, as both mouse strains displayed marked susceptibility to infection and elevated bacterial burdens compared to wild-type (WT) animals (Kielian et al. 2004; Xiong et al. 2012). Our recent report showed that microglial IL-1β production in response to live S. aureus was mediated by the Nod-like receptor protein 3 (NLRP3) inflammasome (Hanamsagar et al. 2011), which has been studied extensively for its role in IL-1β secretion in response to diverse pathogenic signals, including S. aureus α-hemolysin (Craven et al. 2009). However, little is known about the functional importance of inflammasomes during CNS bacterial infection, with only one report documenting that IL-1β release during S. pneumoniae meningitis is controlled via the NLRP3 inflammasome (Hoegen et al. 2011). To further dissect the underlying mechanisms mediating IL-1β release during early CNS infection, we focused on the NLRP3 inflammasome utilizing a well-established mouse model of S. aureus-induced brain abscess that recapitulates various aspects of human disease. We found that ASC (apoptosis-associated speck-like protein containing a caspase-1 recruitment domain) and caspase 1/11 were critical for survival during the first 24 h after infection, whereas NLRP3 was dispensable, revealing the involvement of a NLRP3-independent, ASC-dependent inflammasome during acute CNS S. aureus infection. In addition to NLRP3, other ASC-dependent inflammasomes have been described including absent in melanoma 2 (AIM2), NLRC4, and NLRP6 (Fernandes-Alnemri et al. 2009; Hornung et al. 2009; Broz et al. 2010; Rathinam et al. 2010; Elinav et al. 2011). When considering alternative inflammasomes that could play a role during CNS S. aureus infection, AIM2 seemed the most plausible since intracellular survival of S. aureus has been reported in phagocytes enabling the detection of cytoplasmic dsDNA, whereas NLRC4 and NLRP6 recognize motifs that are unrelated to S. aureus (Gresham et al. 2000; Watanabe et al. 2007). Subsequent experiments identified this upstream adaptor as AIM2, which was unexpected, since prior to our study, AIM2 had been implicated in sensing dsDNA in macrophages following viral or L. monocytogenes infection (Fernandes-Alnemri et al. 2010; Rathinam et al. 2010; Warren et al. 2010). To our knowledge, this is the first study to examine the functional role of inflammasome components during CNS S. aureus infection and reveals a unique role for AIM2 in S. aureus detection in the brain.
- Top of page
- Materials and methods
- Acknowledgments and conflict of interest disclosure
We previously established a pivotal role for the NLRP3 inflammasome in mediating IL-1β release from microglia in response to live S. aureus (Hanamsagar et al. 2011). In the current report, we have extended these findings to identify the molecular composition of the inflammasome that is critical for protection during acute CNS bacterial infection in vivo. Prior work from our laboratory demonstrated that IL-1RI KO mice succumb to CNS S. aureus infection within the first 24 h following bacterial exposure (Xiong et al. 2012); therefore, we predicted that similar results would be observed for both NLRP3 and ASC KO animals, since this inflammasome platform has the largest range of molecular triggers leading to IL-1β processing and release. Unexpectedly, ASC, but not NLRP3 KO mice were exquisitely sensitive to CNS S. aureus infection. The observed disconnect between ASC and NLRP3 phenotypes differs from most of the current inflammasome literature (Gris et al. 2010; Jha et al. 2010; Zaki et al. 2010; Hoegen et al. 2011), although divergent involvement of NLRP3 versus ASC has been reported in models of experimental autoimmune encephalomyelitis and bacterial pneumonia (Shaw et al. 2010; van Lieshout et al. 2014). Differential requirements for NLRP3 versus ASC in eliciting IL-1β production following S. aureus infection could be one explanation, since cytokine levels were significantly reduced in ASC KO mice but not NLRP3 KO animals (data not shown). The inability to elicit wild-type IL-1β production in ASC KO mice functionally equates with the inability of the cytokine to signal via the IL-1RI, which agrees with the finding that both ASC KO and IL-1RI KO animals in our previous study (Xiong et al. 2012) displayed a similar heightened sensitivity to CNS S. aureus infection, with approximately 33–50% of mice succumbing within the first 24 h. Collectively, these findings highlight the essential role of IL-1 signaling for protection after acute CNS bacterial exposure. IL-1β exerts numerous functions within the CNS, including modulation of blood–brain barrier permeability, activation of resident glia, and indirectly affecting leukocyte recruitment by augmenting chemokine and adhesion molecule expression at the blood–brain barrier (Gosselin and Rivest 2007). In the context of CNS S. aureus infection, IL-1β is critical for survival and bacterial containment, and plays an indirect role in dictating leukocyte influx by augmenting chemokine expression (Kielian et al. 2004; Xiong et al. 2012). However, we did not detect any significant differences in peripheral leukocyte infiltrates or bacterial burdens between WT and ASC or AIM2 KO mice in this study, which may result from the combined actions of IL-1β and IL-1α loss in IL-1RI KO animals in our previous report (Xiong et al. 2012), as both cytokines have overlapping biological activities. In addition, it is likely that IL-1β exerts beneficial effects during acute CNS infection that remain to be defined. A similar protective role for IL-1β has been reported in West Nile encephalitis (Ramos et al. 2012). Despite the marked reduction in IL-1β in ASC KO mice, residual cytokine remained, suggesting that pro-IL-1β may also be processed by currently unknown alternative mechanisms, albeit relatively minor. For example, caspase 1 may be activated via other inflammasomes responsive to S. aureus-related triggers or danger-associated molecular patterns released from necrotic cells (Willingham et al. 2009; Lamkanfi et al. 2010; Davis et al. 2011). Alternatively, other caspases or enzymes (i.e., cathepsins) may be involved in pro-IL-1β processing (and potentially IL-18 cleavage) in this model (Terada et al. 2010).
One unexpected finding during our in vivo studies was the distinct outcomes in NLRP3 and ASC KO mice following CNS S. aureus infection, which differs from other inflammatory models that generally reported similar phenotypes in both strains (Gris et al. 2010; Jha et al. 2010; Zaki et al. 2010; Hoegen et al. 2011). This raised the possibility that alternative ASC-dependent inflammasome sensor(s) were likely upstream effectors for triggering IL-1β release and host survival during CNS S. aureus infection. Other inflammasomes known to recruit ASC for caspase 1 activation include AIM2, NLRC4, and nod-like receptor protein 6 (Hornung et al. 2009; Broz et al. 2010; Elinav et al. 2011). The AIM2 inflammasome senses cytosolic dsDNA following bacterial or viral infections (Hornung et al. 2009). Phagocytosis and intracellular killing of S. aureus by host cells may lead to the release of bacterial DNA into the cytoplasm and subsequent AIM2 inflammasome activation. Alternatively, numerous reports have demonstrated intracellular survival of S. aureus (Gresham et al. 2000; Watanabe et al. 2007), where the organism may be capable of triggering cytosolic AIM2. Indeed, here we show that AIM2 KO mice displayed a phenotype that closely paralleled that of ASC KO animals in terms of survival after infection and impaired proinflammatory mediator production. Collectively, these findings reveal a novel role for AIM2, and not the NLRP3 inflammasome, for host survival following S. aureus infection in the CNS. However, another layer of complexity is that AIM2 can also act independent of caspase 1, by recruiting caspase 8 and inducing caspase 8-dependent apoptosis, as shown in a study of Francisella infection (Pierini et al. 2012). In addition, several inflammasomes, including NLRP3, AIM2, and NLRC4 are known to act cooperatively, likely as an evolutionary mechanism to ensure maximal pathogen eradiation or as a means to add multiple layers of regulatory control over cytokine secretion (Kim et al. 2010; McCoy et al. 2010; Wu et al. 2010; Coll and O'Neill 2011). Therefore, it is possible that the concerted action of several ASC-dependent inflammasomes is elicited following CNS S. aureus infection, although this remains to be determined (Fig. 4). Assessing the cellular localization and spatial distribution patterns of these inflammasome molecules may provide insights into their role in disease pathogenesis.
Although we predict that AIM2 and ASC are primarily active in microglia during acute CNS infection, it remains possible that astrocytes and/or infiltrating monocytes/macrophages may also play a role. However, unpublished data from our laboratory show that highly purified astrocytes do not secrete IL-1β in response to live S. aureus challenge in vitro. It is possible that in an in vivo setting, astrocytes may interact with microglia to secrete IL-1β, which is supported by a recent report (Holm et al. 2012); however, any such release might be small and insignificant in comparison with microglia. Macrophages and monocytes express both AIM2 and ASC (Mariathasan et al. 2005; Fernandes-Alnemri et al. 2009; Rathinam et al. 2010) and are activated during bacterial infection. However, we previously showed that CNS intrinsic cells were critical for the initial sensing of CNS S. aureus infection (Garg et al. 2009) and the rapid kinetics of susceptibility of AIM2 and ASC KO mice to intracerebral S. aureus challenge suggests the involvement of CNS intrinsic sensors of infection. Therefore, although it is possible that the lack of AIM2 or ASC in infiltrating monocytes/macrophages is involved in regulating IL-1β release and infection outcome at later stages in the disease process, the sum of currently available evidence points to a key role for CNS intrinsic cells during the early phase of infection. However, this prediction remains to be tested and would require the use of Cre-specific mice to ablate AIM2 or ASC expression in astrocytes or microglia/macrophages combined with the use of radiation bone marrow chimeras to discriminate between the involvement of the latter two populations.
Reactive oxygen species (ROS) have been shown to be an important upstream activator of the NLRP3 inflammasome in response to numerous stimuli (Martinon 2010; Zhou et al. 2011). However, we have previously demonstrated that ROS does not induce IL-1β secretion in the context of live S. aureus challenge in microglia (Hanamsagar et al. 2011). Similarly, a recent study has shown that mycobacterium-induced AIM2 inflammasome activation is not ROS dependent (Yang et al. 2013). Therefore, we do not anticipate a role of ROS in the observed pathology of AIM2 or ASC KO mice reported here; however, this remains to be examined.
In conclusion, our studies reveal the existence of a complex regulatory network for inflammasome activation and cytokine secretion during CNS S. aureus infection that is primarily AIM2/ASC-dependent and NLRP3-independent. To our knowledge, a role for the AIM2 inflammasome during S. aureus infection either in the CNS or the periphery has not yet been demonstrated. Although sustained inflammation following CNS infection can be detrimental, it is clear that the early release of proinflammatory cytokines, such as IL-1β, is essential for host survival. Therefore, it will be critical to design future studies to understand the temporal aspects of inflammasome function to develop therapeutic interventions that are safe and effective based on the nature of the insult.
Acknowledgments and conflict of interest disclosure
- Top of page
- Materials and methods
- Acknowledgments and conflict of interest disclosure
This work was supported by NIH National Institute of Neurological Disorders and Stroke (NINDS) 2R01 NS040730 to T.K. The authors thank Dr Vishva Dixit at Genetech for providing NLRP3 and ASC KO mice, Debbie Vidlak for performing the MILLIPLEX analysis, Amanda Angle for excellent technical assistance, and Dr Charles Kuszynski and Victoria Smith in the UNMC Cell Analysis Facility for assistance with FACS analysis. The authors thank Ms. Kari Nelson, supported by NIH/NIAID P01 AI083211, for editorial assistance and Dr Paul Drew for critical review of the manuscript. All experiments were conducted in compliance with the ARRIVE guidelines. The authors have no conflict of interest to declare.