|Home | About | Journals | Submit | Contact Us | Français|
The mechanism by which bacterial pathogens activate caspase-1 via Nlrp3 remains poorly understood. Here we show that the ability of Staphylococcus aureus, a leading cause of infection in humans, to activate caspase-1 and induce IL-1β secretion resides in culture supernatants of growing bacteria. Caspase-1 activation induced by S. aureus required α-, β- and γ- hemolysins and the host Nlrp3 inflammasome. Mechanistically, α- and β-hemolysins alone did not trigger caspase-1 activation but they did so in the presence of bacterial lipoproteins released by S. aureus. Notably, caspase-1 activation induced by S. aureus supernatant was independent of the P2X7 receptor and the essential TLR adaptors Myd88 and Trif, but was inhibited by extracellular K+. These results indicate that S. aureus hemolysins circumvent the requirement of ATP and the P2X7 receptor to induce caspase-1 activation via Nlrp3. Furthermore, these studies revealed that hemolysins promote in the presence of lipoproteins the activation of the Nlrp3 inflammasome.
Staphylococcus aureus is a Gram-positive bacterium responsible for a wide variety of superficial and life-threatening systemic infections (1). It produces an array of virulence factors that allow the bacteria to colonize epithelial surfaces, invade host tissues and evade the immune response (2). The bacterium also produces a number of disease-relevant exotoxins, such as toxic shock syndrome toxin-1, enterotoxins and membrane-damaging toxins (cytolysins), which include hemolysins (α, β, γ and δ) and Panton-Valentine Leucocidin (3, 4). α-hemolysin is secreted as a 33 KDa monomer and oligomerizes forming heptameric transmembrane pores (5). β-hemolysin is a neutral sphingomyelinase (6). Both α- and β-hemolysins have the ability to damage the membrane of human monocytes and promote the secretion of IL-1β, IL-6 and TNF-α (7-10). Furthermore, α-hemolysin can induce K+ efflux in host cells (11, 12). However, the mechanism by which S. aureus hemolysins promote host immune responses remains poorly understood.
Host recognition of bacterial pathogens including S. aureus is mediated in part by pattern recognition receptors, including membrane-bound TLRs and intracellular nucleotide-binding oligomerization domain receptors (NLRs) (13). TLRs recognize conserved bacterial components such as LPS, flagellin, lipoproteins, lipoteichoic acid and unmethylated CpG DNA (14). There is evidence that TLR2 is critical for S. aureus recognition and host defense (15). More recently, nucleotide-binding oligomerization domain containing 2 (Nod2), a NLR protein which senses muramyl dipeptide produced during the synthesis and/or degradation of peptidoglycan, has been implicated in the host response to S. aureus (16). While TLR2 and Nod2 induce immune responses via the activation of the transcription factor NF-κB and MAP kinases (17), another group of NLRs that include Nlrp3 and Nlrc4 are critical for the activation of caspase-1 and IL-1β secretion in response to bacterial and endogenous stimuli (18). Nlrp3 and Nlrc4 link specific microbial stimuli to caspase-1 activation through the formation of a multi-protein platform called the inflammasome that includes the adaptor apoptosis associated speck-like protein (Asc) (18). Upon microbial stimulation, the IL-1β precursor is induced in macrophages via NF-κB activation (19). Subsequently, pro-IL-1β is processed into mature IL-1β by the enzymatically active heterodimer composed of a 10- and a 20-kDa chain of caspase-1 (20). Nlrp3 is essential for caspase-1 activation in response to a variety of microbial molecules in combination with ATP and particulate matter including urate crystals, silica particles and aluminum salts (21-23). Stimulation of the purinergic P2X7 receptor (P2X7R) with ATP induces K+ efflux via the rapid opening of a non-selective cation channel and the gradual opening of a large pore possibly mediated by pannexin-1, both of which are important for Nlrp3-mediated caspase-1 activation in response to ATP and bacterial products (21, 24). In addition to ATP, activation of the Nlrp3 inflammasome is induced in LPS-stimulated macrophages by pore-forming toxin maitotoxin and the ionophore nigericin, which induce changes in cellular K+ concentration (25). However, the mechanism by which bacterial pathogens activate the Nlrp3 inflammasome is poorly understood.
Recent studies showed that the adaptor protein Asc and IL-1β can play a critical role in neutrophil recruitment elicited by S. aureus skin infection and elimination of the bacterium in vivo (26, 27). Furthermore, infection of macrophages with S. aureus induced low level, but significant, caspase-1 activation via Nlrp3 in vitro (25). However, the latter finding remains controversial as some authors did not observe caspase-1 activation and IL-1β secretion in response to S. aureus infection in the absence of exogenous ATP (28). Furthermore, the S. aureus molecules that induce caspase-1 activation and IL-1β secretion remain unknown. The objective of this study was to clarify the mechanism by which S. aureus induces IL-1β and triggers the activation of caspase-1.
Nlrp3-/-, Asc-/-, caspase-1-/- mice have been described (29). P2X7R-/- mice were obtained from Dr. George Dubyak, Case Western Reserve University. MyD88-/- and Myd88/Trif-/- mice were obtained from Dr. Akira, Osaka University. All mice were backcrossed onto the C57BL/6 background at least eight times except for the Nlrp3-/- mice which were in mixed 129/C57BL/6 background. Wild-type (WT) C57BL/6 and 129/C57BL/6 mice were maintained in our Animal Facility. All animal studies were approved by the University of Michigan Committee on Use and Care of Animals.
S. aureus wild-type (wt) strain 8325-4 and its isogenic hemolysin mutants have been previously described (4). S. aureus wt strain SA113 and its isogenic mutant deficient in lipoprotein diacylglyceryl transferase (Δlgt) have also been previously described (30). Clinical isolates of S. aureus were obtained from the laboratory of Clinical Microbiology at the University of Michigan University Hospital with the approval of the Institutional Review Board of the University of Michigan Medical School. Tryptic Soy Broth was used for the growth of the bacteria.
Purified α-hemolysin was purchased from Sigma (H9395) and β-hemolysin from Toxin Technology (BHT1). For caspase-1 activation experiments with purified toxins, macrophages were stimulated for 3 h either with a 1/10 dilution of S. aureus supernatant collected after 20 h of subculture, washed 3 times with PBS and subsequently incubated with 10 μg/ml of α-hemolysin or β-hemolysin in IMDM.
Bacteria in midlogarithmic phase were obtained after subculturing bacteria at a 1:100 dilution of an overnight culture for 3 h. Bone marrow-derived macrophages were infected with S. aureus for 3 h or indicated time at the bacterial-macrophage ratio indicated in Figures and either pulsed for 30 min with 5 mM ATP or left untreated. S. aureus supernatants were prepared by subculturing 1:100 dilutions of overnight S. aureus cultures for different times (0-20 h). The cultures were centrifuged at 4000 g for 10 min and the supernatants were collected and sterilized by filtration. All in vitro stimulation of macrophages with S. aureus supernatants were performed for 3 hours at 1/10 dilution of supernatants collected after 20 h of subculture unless otherwise specified.
The percentage of macrophage death was determined by measurement of the release of LDH after incubation with the CytoTox 96 nonradioactive cytotoxicity assay (Promega). The absorbance at 490 nm was measured, and the percentage of cell death was calculated as follows: [(experimental release - spontaneous release) / (maximum release-spontaneous release)]×100, where ‘spontaneous release’ is that found in untreated macrophages and ‘maximum release’ is the value obtained after lysis of macrophages with a solution of 0.1% Triton-X100.
The subcutaneous abscess model has been previously described (26). Briefly, the mice were inoculated subcutaneously with 106 CFUs of midlogarithmic growth phase S. aureus strain 8325-4 in 50 μl of PBS. Lesional skin biopsies were collected 24 h later and homogenized in the presence of proteinase inhibitors (Roche).
Membranes were probed with antibodies against caspase-1 (a gift of Peter Vandenabeele (Ghent University, Belgium), mature IL-1β (p17, Asp116, Cell Signaling), pro-IL-1β (pIL-1β) (R&D Systems) (27) and α-hemolysin antiserum (Sigma). As a positive control for caspase-1 activation, WT macrophages were incubated with 1 μg/ml of LPS for 3 h and stimulated for 30 min with 5 mM ATP. Levels of IL-1β, mature IL-18 and MIP-2α were determined by ELISA (MBL International).
Statistical significance between groups was determined by two tailed Student’s t test. Animals studies were analyzed by one-way ANOVA. Differences were considered significant when p < 0.05.
The ability of S. aureus to active caspase-1 in macrophages remains controversial (25, 28). Consistent with our previous results (28), infection of macrophages with S. aureus did not induce detectable activation of caspase-1 at different bacterial-macrophage ratio for up to 24 hrs of culture (Fig. 1A). However, incubation of macrophages with the highest MOI of S. aureus tested and for extended periods of time (12-24 hours) led to significant release of IL-1β (Fig. 1B) which was associated with elevated cytotoxicity (Fig. 1C). To further assess the ability of S. aureus to activate caspase-1, we incubated macrophages with S. aureus supernatants collected after bacterial growth in culture for different times. Notably, stimulation of macrophages for 3 hrs with culture supernatants from S. aureus collected after 2 h of growth elicited significant caspase-1 activation, which increased by 6 h and peaked by 12 h of bacterial culture (Fig. 1D). Consistently, a parallel increase in IL-1β secretion was elicited by culture supernatants of growing S. aureus (Fig. 1E). Caspase-1 activation and IL-1β secretion triggered by culture supernatant from S. aureus were comparable to that observed with LPS, but unlike that induced with the TLR4 agonist, that triggered by the bacterium supernatant did not require exogenous ATP stimulation (Fig. 1D and E). These results indicate that S. aureus releases factor(s) during bacterial growth that activate caspase-1 in the absence of exogenous ATP.
S. aureus secretes several toxins including hemolysins that function as pores and/or membrane-damaging factors (4, 7, 8). To begin to assess a role for hemolysis in caspase-1 activation, we determined the production of α-hemolysin by S. aureus during bacterial growth culture. Immunoblotting analysis revealed that α-hemolysin was first detectable in the culture supernatants by 6 h and increased by 12 h of bacterial culture (Figure 2A). To assess whether α-hemolysin was induced during infection, we determined the production of α-hemolysin during culture of S. aureus with macrophages under conditions similar to those shown in Fig. 1A in which caspase-1 was not activated. Under these conditions, we did not observe detectable production of α-hemolysin (Fig. 2B). To assesses the role of hemolysins in caspase-1 activation, culture supernatants from wt S. aureus or its isogenic mutants lacking α-, β- or γ-hemolysins and all possible combinations were tested for their ability to induce caspase-1 activation in macrophages. The culture supernatants from single hemolysin mutant strains, particularly α-hemolysin, induced significant, although reduced, levels of caspase-1 activation when compared to wt S. aureus supernatant (Fig. 2C). Notably, combined deficiency of α-, β- and γ-hemolysins abolished the ability of the S. aureus culture supernatant to activate caspase-1 (Fig. 2C) and to induce IL-1β secretion (Fig. 2D). Thus, α-, β- and γ-hemolysins released by S. aureus play an essential, although redundant, role in the activation of caspase-1 and the secretion of IL-1β in macrophages.
To study the host recognition system involved in hemolysin-mediated caspase-1 activation, we analyzed macrophages deficient in several inflammasome components as well as critical TLR adaptors and the P2X7R. Caspase-1 activation and IL-1β secretion induced by S. aureus supernatant was abolished in macrophages deficient in Nlrp3 or Asc (Fig. 3A and 3B). In contrast, Nlrc4, a NLR family member required for caspase-1 activation in response to Salmonella, was dispensable for caspase-1 activation and IL-1β secretion induced by S. aureus supernatant (data not shown). Nlrp3-mediated caspase-1 activation induced by bacterial molecules including LPS requires ATP-mediated stimulation of the P2X7R (25). However, processing of pro-caspase-1 and IL-1β secretion induced by S. aureus supernatant was unimpaired in macrophages deficient in the P2X7R (Fig. 3A and 3B). Furthermore, caspase-1 activation proceeded normally in macrophages deficient in MyD88 or both MyD88 and Trif, the adaptors that are essential for TLR signaling (Fig. 3A). In contrast, IL-1β secretion induced by S. aureus supernatant was abrogated in macrophages deficient in MyD88 or MyD88 and Trif (Fig. 3B). Thus, TLR signaling is required for IL-1β secretion but not caspase-1 activation induced by S. aureus supernatant.
Activation of the Nlrp3 inflammasome in response to ATP or particulate matter in combination with TLR agonists is inhibited by high concentration of extracellular K+ (31, 32). Furthermore, the K+ efflux elicited by α- and β- hemolysin is necessary for IL-1β secretion (8, 33). Consistently, incubation of macrophages in medium rich in K+, but not Na+, impaired caspase-1 activation and IL-1β secretion induced by S. aureus supernatant (Fig. 3C and 3D). These results suggest that the K+ efflux triggered by hemolysins released during S. aureus growth regulates Nlrp3 inflammasome activation independently of the P2X7R.
To study the role of hemolysins in caspase-1 activation in vivo we used a previously described model of S. aureus subcutaneous infection (26, 27). WT mice were injected subcutaneously with 1 × 106 CFUs of wt S. aureus or its isogenic hemolysin mutants and tissue extracts of the lesions were prepared 24 h post-infection. Immunoblotting of the skin tissue revealed that infection with S. aureus induced the production of proIL-1β (Fig. 4A). Importantly, S. aureus lacking α-, β- and γ-hemolysins was impaired in the production of the processed IL-1β (p17) and IL-18 in the tissue, whereas that induced by single or double mutants were slightly reduced or unimpaired (Fig. 4B and 4C). In contrast, MIP-2α induction by the triple hemolysin S. aureus mutant was unimpaired (Fig. 4D), providing specificity for the response. Counting the number of viable bacteria in the lesions revealed similar bacterial burdens in mice infected with wild-type and the isogenic S. aureus mutants (Fig. S1).
We next determined the role of the Nlrp3 inflammasome in the processing of pro-IL-1β and secretion of IL-18 in vivo. Immunoblotting of skin tissue extracts prepared after infection with S. aureus showed that production of the mature IL-1β (p17) was reduced in tissue from mice deficient in Nlrp3, Asc and caspase-1 (Fig. 4E). Furthermore, the production of IL-18 was abrogated in mice deficient in Nlrp3, Asc and caspase-1 (Fig. 4F). Taken together, these results indicate that S. aureus α-, β- and γ-hemolysins play a critical role in the production of mature IL-1β and IL-18 through the Nlrp3 inflammasome in vivo.
To validate the results obtained with the S. aureus laboratory strain 8325-4, we next determined whether bacterial strains isolated from patients suffering from S. aureus infection induce caspase-1 activation via Nlrp3. In line with our previous results, α-hemolysin was released to the culture media during in vitro growth of clinical S. aureus clinical isolates (Fig. 5A). Importantly, caspase-1 activation induced by the culture supernatant of the clinical S. aureus strains required Nlrp3 but was independent of MyD88 and Trif (Fig. 5B). These results indicate that the culture supernatant from S. aureus strains isolated from infected patients can induce caspase-1 activation via Nlrp3 independently of TLR signaling.
To further assess the ability of hemolysins to activate the Nlrp3 inflammasome, we tested whether purified α- and β-hemolysins could restore the ability to activate caspase-1 of the supernatant from S. aureus deficient in α-, β- and γ-hemolysins. Notably, incubation of macrophages with purified hemolysins did not activate caspase-1 (Fig. 6A). However, addition of purified α- or β-hemolysin to supernatant from α-, β- and γ-deficient S. aureus restored its ability to induce caspase-1 activation and IL-1β secretion in macrophages (Fig. 6A and 6B). Consistent with our previous results, this activity was abrogated in macrophages deficient in Nlrp3 or Asc (Fig. 6A and 6B), suggesting that that α- and β-hemolysins cooperate with other factor(s) released by the bacteria to activate the Nlrp3 inflammasome.
The bacterial cell wall of S. aureus contains several immunostimulatory molecules which are released during bacterial growth (30, 34). To identify the factor(s) present in the S. aureus supernatant that is required with hemolysins to activate caspase-1, we next investigated the role of the lipoproteins released by S. aureus in the secretion of IL-1β. The enzyme lipoprotein diacyglyceryl transferase, encoded in the lgt locus, catalizes the first step of lipoprotein biosynthesis (30). To determine if lipoproteins are required for IL-1β secretion, we incubated WT and TLR2-/- macrophages with culture supernatants from the wt S. aureus strain SA113 or its isogenic lgt deletion mutant (Δlgt) and assessed the induction of proIL-1β. Immunoblotting analysis revealed that the mutant Δlgt strain was impaired in its ability to induce pro-IL-1β when compared to the wt SA113 strain (Fig. 7A). Consistently, TLR2 deficient macrophages failed to induce pro-IL-1β in response to wt supernatant (Fig. 7A). In line with previous studies (30, 34), we found that IL-6 secretion was abrogated and TNF-α secretion greatly reduced in macrophages stimulated with lipoprotein deficient bacterial supernatants (Fig. 7B and 7C).
We next tested whether bacterial lipoproteins present in the S. aureus supernatant contribute to caspase-1 activation. Unlike the 8325-4 strain, the culture supernatant from the wt SA113 S. aureus did not induce detectable caspase-1 activation (Fig. 7D) and this was associated with undetectable levels of α-hemolysin (Fig. 7E). Therefore, we tested whether purified α-hemolysin could restore the ability of strain SA113 supernatant to activate caspase-1. We found that the addition of α-hemolysin to macrophages previously stimulated with culture supernatant from the wt strain SA113 induced caspase-1 activation and IL-1β secretion, but each factor alone did not (Fig. 7F and 7G). Importantly, culture supernatant from Δlgt mutant strain exhibited impaired caspase-1 activation and IL-1β secretion in the presence of purified α-hemolysin, when compared to that induced by supernatant from the wt parental strain (Fig. 7F and 7G). These results indicate that lipoproteins released by S. aureus are important in inducing caspase-1 activation and IL-1β secretion.
Previous studies have shown that IL-1β signaling and the adaptor protein Asc play a critical role in the clearance of S. aureus infection in the skin (26, 27). However, the mechanism by which the bacterium induces such host responses remained poorly understood. Here we provide evidence both in vitro and in vivo for a critical but redundant function of S. aureus α-, β- and γ-hemolysins in eliciting caspase-1 activation via the Nlrp3 inflammasome. This redundancy may explain, at least in part, why a previous study analyzing single hemolysin S. aureus mutants (25) did not reveal any significant role for these exotoxins in caspase-1 activation. Unlike the present work, these authors observed weak but significant caspase-1 activation after S. aureus infection in LPS-stimulated macrophages (25). These differences in results could be due to the use of different culture conditions or the priming step with LPS prior to the stimulation with the bacteria. We also show that the main caspase-1 activating activity resides in the supernatants of growing S. aureus which is consistent with the fact that hemolysins are exotoxins secreted by the bacteria (2). Experiments with purified α- and β-hemolysins showed that these toxins are necessary but not sufficient to activate the Nlrp3 inflammasome, as they required the presence of other bacterial products to activate caspase-1. Specifically, we provide evidence that lipoproteins released by the bacterium are important for the induction of proIL-1β and the activation of the Nlrp3 inflammasome. Although S. aureus supernatant did not require P2X7R signaling to activate the Nlrp3 inflammasome, caspase-1 activation was blocked by incubation of macrophages in medium containing high concentration of K+. These results suggest that S. aureus hemolysins released by S. aureus elicit K+ efflux independently of P2X7R which is important for the activation of caspase-1. In line with these results, purified α-hemolysin can induce K+ efflux (11) and this activity is thought to be necessary for IL-1β maturation (8, 35).
TLR2 plays an important role in the innate immune response to S. aureus and mediates the production of proinflammatory cytokines in macrophages and monocytes (36, 37). In this study, we examined the role of TLRs and bacterial lipoproteins in caspase-1 activation and IL-1β secretion induced by S. aureus in macrophages. Using macrophages that are deficient in the adaptor proteins MyD88 and Trif, and that therefore cannot signal via TLRs, we found that caspase-1 activation induced by S. aureus supernatant is independent of TLR signaling. The results suggest that products released by S. aureus induce TLR-independent caspase-1 activation in combination with pore forming toxins. In contrast, lipoproteins in the bacterial supernatant are important for pro-IL-1β induction through TLR2 signaling. Collectively, these results reveal distinct but critical roles for lipoproteins and hemolysins as well as the host TLRs and the Nlrp3 inflammasome in IL-1β secretion in response to S. aureus in macrophages.
Several bacteria activate caspase-1 through the formation of membrane pores. Salmonella, for instance, requires the presence of a functional type III secretion system to activate the Nlrc4 inflammasome (29). Similarly, L. monocytogenes requires the expression of the pore-forming protein listerolysin O to activate caspase-1 (38). In the current work, we show that α-, β- and γ-hemolysins produced by S. aureus are critical for activating the Nlrp3 inflammasome both in vitro and in vivo. A common feature of these observations is the formation of pores which alter the permeability of host membranes and may allow the cytosolic delivery of microbial molecules. Consistent with this hypothesis, we have previously shown that ATP triggers through pannexin-1 the intracellular delivery of MDP, which induces caspase-1 activation via Nlrp3 (39). Alternatively, lipoproteins and other microbial molecules released by S. aureus may provide a priming signal, at least in part, independently of TLRs for inflammasome activation and hemolysins may mimic ATP and P2X7R signaling to activate caspase-1. Because a direct association between microbial molecules and Nlrp3 proteins has not yet been identified, a likely model is that the activation of the inflammasome by microbial ligands is indirectly induced via an endogenous mediator. Consistently, there is evidence that reactive oxygen species, calcium-independent phospholipase A2, and cathepsin B may contribute to the activation of the Nlrp3 inflammasome at least in response to some stimuli (23, 32, 40, 41). Further work is needed to understand the mechanism whereby bacterial pathogens and host factors interact to induce inflammasome activation.
We thank Millenium Pharmaceuticals, Shizuo Akira and George Dubyak for generously supplying mutant mice, Timothy Foster and Friedrich Götz for kindly providing bacterial strains, Joel Whitfield for technical support, Peter Vandenabeele for providing anti caspase-1 antibody and Sherry Koonse for excellent animal husbandry.
1This work was supported by grants AI063331 and AI064748 (to G. N.). R. M.-P. was supported in part by a Fulbright-Ramón Areces Fellowship and L. F. was supported by a Fellowship grant from the Arthritis Foundation.
The authors declare that no competing interests exist.