ABSTRACT
Staphylococcus aureus is a Gram-positive opportunistic pathogen that causes a variety of diseases. Bloodstream infection is the most severe, with mortality rates reaching 20 to 50%. Exopolysaccharide (EPS) from the probiotic Bacillus subtilis reduces bacterial burden and inflammation during S. aureus bloodstream infection in mice. Protection is due, in part, to hybrid macrophages that restrict S. aureus growth through reactive oxygen species and to limiting superantigen-induced T cell activation and interferon gamma (IFN-γ) production during infection. A decrease in IFN-γ production was observed within 24 h after infection, and here, we investigated how EPS abrogates its production. We discovered that S. aureus uses a rapid, superantigen-independent mechanism to induce host IFN-γ and that this is mediated by interleukin-12 (IL-12) activation of NK cells. Furthermore, we found that EPS limits IFN-γ production by modulating host immunity in a Toll-like receptor 4 (TLR4)-dependent manner, a signaling pathway that is required for EPS-mediated protection from S. aureus infection in vivo. We conclude that EPS protects hosts from acute bloodstream S. aureus infection not only by inducing macrophages that restrict S. aureus growth and inhibit superantigen-activated T cells but also by limiting NK cell production of IFN-γ after S. aureus infection in a TLR4-dependent manner.
INTRODUCTION
Staphylococcus aureus commonly colonizes the skin and anterior nares (1). Upon entering the bloodstream through breaches in skin and mucosal barriers, S. aureus can disseminate and cause multiple organ failure, leading to high mortality rates despite aggressive therapy (2). In addition, multidrug resistance is observed across many clinical isolates of S. aureus, highlighting a need for novel strategies to manage S. aureus infections.
A striking feature of S. aureus is its ability to evade and to manipulate an array of host responses to promote its persistence. One of these immune factors is interferon gamma (IFN-γ), widely accepted as a principal cytokine that activates innate phagocytes and augments their antimicrobial activity (3–7). The host mounts an IFN-γ response during S. aureus infection (8), in part due to S. aureus activation of host T cells through superantigens (SAg) (9). We previously demonstrated that exopolysaccharide (EPS) from a probiotic, Bacillus subtilis, protects mice from S. aureus bloodstream infection, in part by inducing hybrid macrophages that both restrict bacterial growth through reactive oxygen species (ROS) and limit T cell activation by S. aureus SAg (10). In vivo, a striking reduction in serum IFN-γ levels was observed in EPS-treated S. aureus-infected mice (10). While IFN-γ is the classic cytokine that mediates antimicrobial responses, IFN-γ-deficient mice have improved survival during S. aureus bloodstream infection (11), suggesting that S. aureus exploits the host IFN-γ response to promote persistence and pathogenesis. In this study, we sought to understand the host IFN-γ response after S. aureus infection, and we discovered that S. aureus has an alternative, SAg-independent mechanism for inducing IFN-γ. This SAg-independent IFN-γ response is MyD88 dependent and is mediated through interleukin-12 (IL-12)-dependent activation of NK cells. We also show that Toll-like receptor 4 (TLR4) is required for EPS-mediated protection from S. aureus bloodstream infection. These data suggest that S. aureus induces host IFN-γ through both SAg-dependent and -independent mechanisms and that EPS limits this response in a TLR4-dependent manner, leading to improved disease outcomes during systemic S. aureus infection.
RESULTS
SAg-independent IFN-γ response by mouse splenocytes.EPS greatly decreases serum IFN-γ levels after intravenous (i.v.) infection of mice with an epidemic strain of S. aureus (10). Here, we showed that 1 day postinfection (d.p.i.), the levels of IFN-γ in organ homogenates were also decreased in the spleen and liver of EPS-treated mice compared to those of phosphate-buffered saline (PBS)-treated mice; no difference was observed in the kidney (Fig. 1). These data suggest that during systemic S. aureus infection, EPS limits the IFN-γ response in some but not all tissues. This finding was similar to the decrease in S. aureus CFU observed in spleen and liver, but not kidney, of S. aureus-infected EPS-treated mice (10). During murine systemic infection, S. aureus resides in abscesses in the kidney and is not accessible to immune cells, including macrophages that arrest S. aureus growth (10, 12, 13).
EPS abrogation of local IFN-γ levels during S. aureus bloodstream infection. Mice were treated with PBS or EPS by i.p. injection 1 day prior to i.v. infection with S. aureus. Levels of IFN-γ in organ homogenates were determined 1 d.p.i. using CBA. Bars represent means. Each triangle represents one mouse. Data pooled from 2 independent experiments. n = 8. Analyzed using Student's t test. **, P < 0.01.
We hypothesized that the EPS-mediated decrease in the IFN-γ response during S. aureus infection was due to inhibition of SAg-mediated T cell activation. However, when we infected mice with a staphylococcal enterotoxin-like (SEI)-Q-deficient S. aureus strain (JE2 seq::erm), the only SAg encoded in strain JE2 that stimulated murine T cells in our prior studies (10), we found that the levels of serum proinflammatory cytokines 1 d.p.i., including IFN-γ, were not significantly different than those in mice infected with the wild type (WT) JE2 S. aureus strain (Fig. 2A). Similarly, the levels of IFN-γ in spleen and liver 1 d.p.i. were similar in mice that received WT JE2 or the JE2 seq::erm mutant S. aureus strain (Fig. 2B), indicating that some of the IFN-γ produced in response to S. aureus is SAg independent. We also found no decrease 1 d.p.i. in bacterial burden in spleen or liver of mice infected with an SEl-Q-deficient mutant (seq::erm) (Fig. 2C), suggesting that early in infection, SAg is not the principal mechanism promoting disease or production of IFN-γ. We conclude that S. aureus harbors a SAg-independent mechanism to induce an IFN-γ response.
Cytokine and S. aureus CFU in tissues after bloodstream infection with an seq::erm mutant strain of S. aureus. Mice were infected i.v. with WT JE2 and the SEl-Q-deficient (seq::erm) mutant, and disease was assessed. (A) Serum levels of cytokines and chemokines determined 1 d.p.i. using CBA. (B) IFN-γ levels in organ homogenates determined 1 d.p.i. using CBA. (C) CFU of S. aureus in organ homogenates 1 d.p.i. Bars represent means. Each triangle represents one mouse. Data pooled from 2 to 3 independent experiments. Analyzed using Student's t test. n = 8 to 11.
S. aureus secretes a variety of virulence factors other than SAg, including hemolysins, leukocidins, proteases, and other immune modulators that promote disease (14). We tested if the early SAg-independent IFN-γ response was due to one of these other secreted factors or to a nonsecreted S. aureus molecule. We first stimulated splenocytes with supernatant from the SEl-Q-deficient mutant, generated in both the JE2 and AH1263 strain backgrounds (seq::erm), in vitro and measured cytokine production by cytometric bead array (CBA) to determine if a secreted molecule other than SAg stimulated IFN-γ production. We found that almost no IFN-γ was produced in response to supernatants from all S. aureus strains (Fig. 3A), indicating that INF-γ production was not stimulated by secreted molecules. The supernatant, however, harbors activating molecules driving the production of tumor necrosis factor (TNF) and SAg-mediated interleukin-2 (IL-2) production. We conclude that the SAg-independent production of IFN-γ is not mediated by supernatant molecules.
SAg-dependent and SAg-independent induction of splenic IFN-γ response. Mouse splenocytes were stimulated with WT (AH1263 and JE2, both derived from USA300 LAC) and SEl-Q-deficient (seq::erm) mutant S. aureus culture supernatants (A) or heat-killed S. aureus (B), and levels of IFN-γ and TNF were assessed using CBA. Error bars represent standard deviation (SD). Representative data from at least 3 independent experiments. n = 3 to 6. Data analyzed using one-way analysis of variance (ANOVA). (C) IFN-γ levels in splenocytes from WT JE2 or MyD88-deficient mice stimulated with heat-killed S. aureus for 24 h. Bars represent means. Dashed line represents detectable level. Each triangle represents one mouse. Data pooled from 3 independent experiments. Analyzed using 1-sample t test. n = 9 (WT), 6 (MyD88 KO). *, P < 0.05; ***, P < 0.001.
To test if nonsecreted factors could drive the SAg-independent IFN-γ response, we stimulated splenocytes with washed, heat-killed S. aureus and tested for cytokines by CBA. We found readily detectable levels of IFN-γ in cultures stimulated with both AH1263 and JE2, with high levels after stimulation with seq::erm isogenic mutants (Fig. 3B). By comparison, IL-2 levels, which would presumably be produced by activated T cells, were not detectable in any of the cultures (data not shown), indicating that heat-killed S. aureus drives SAg-independent IFN-γ production without stimulating IL-2. As in Fig. 3A, levels of TNF stimulated by the seq::erm mutant were not reduced (Fig. 3B), indicating that the seq::erm mutant did not have an overall defect in stimulating splenocytes. These data indicate that S. aureus can stimulate production of IFN-γ not only by secreted SAg but also by nonsecreted factors, independent of SAg.
S. aureus activates innate immune cells primarily through TLRs (15, 16), leading us to use MyD88-deficient mice to test if the SAg-independent induction of IFN-γ is mediated through TLRs. We found no evidence of IFN-γ production in MyD88−/− splenocytes stimulated with heat-killed S. aureus (Fig. 3C), showing that the SAg-independent mechanism for production of IFN-γ requires signaling through TLRs.
NK cells as source of SAg-independent IFN-γ.While T cells are the major source of IFN-γ (17), other cells, including NK, NKT cells, and myeloid cells, can also produce IFN-γ (18). To identify the cell type responsible for the SAg-independent production of IFN-γ in response to heat-killed S. aureus, we depleted CD3+ T and NKT cells, NK1.1+ NK and NKT cells, or CD11b+ myeloid cells from spleen and stimulated them with heat-killed S. aureus. We found that depletion of CD3+ splenocytes only partially reduced the IFN-γ response (Fig. 4A) but that depletion of NK1.1+ (a marker present on most peripheral NK cells) or CD11b+ splenocytes resulted in complete abrogation of IFN-γ (Fig. 4A). These data suggest that NK/NKT cells and myeloid cells, rather than T cells, are responsible for the SAg-independent IFN-γ response to heat-killed S. aureus. As discussed below, we think it likely that the NK/NKT cells and myeloid cells function together to produce this IFN-γ.
Mechanism of SAg-independent induction of splenic IFN-γ response. (A) Total splenocytes or splenocytes depleted of CD3+, NK1.1+, or CD11b+ cells from mice were stimulated with heat-killed S. aureus for 24 h, and IFN-γ levels were determined using CBA. Data pooled from 3 independent experiments. n = 4 to 7. Data analyzed using one-way ANOVA. (B) Mouse splenocytes were stimulated with heat-killed S. aureus in the presence of neutralizing anti-IL-12p40 (1 μg/ml) or anti-IL-18 (10 μg/ml) antibodies or irrelevant antibody (isotype, 10 μg/ml) for 24 h, and IFN-γ levels were determined using CBA. Error bars represent SD. Representative from 3 independent experiments. n = 3. Analyzed using 1-way ANOVA. **, P < 0.01.
NK cell activation is often mediated by IL-12 (19), and we tested if stimulation of splenocytes with heat-killed S. aureus in the presence of neutralizing anti-IL-12 antibodies would abrogate the IFN-γ response to heat-killed S. aureus. Whereas the addition of irrelevant or neutralizing anti-IL-18 antibodies did not significantly affect the IFN-γ response, anti-IL-12 almost completely inhibited the response (Fig. 4B). These data, together with the above data, lead us to conclude that the SAg-independent production of IFN-γ by S. aureus is mediated in large part by NK and/or NKT cells, rather than T cells, and requires TLR signaling and IL-12.
EPS abrogation of SAg-independent IFN-γ response in a TLR4-dependent manner.EPS was previously demonstrated to inhibit activation of T cells, including activation by S. aureus SAg (10, 20), and here, we tested if EPS can inhibit the SAg-independent induction of IFN-γ from NK cells. We isolated splenocytes 1 day after mice were treated with EPS and stimulated them with heat-killed S. aureus and found almost no IFN-γ produced compared to that of cells from PBS-treated mice (Fig. 5), suggesting that EPS can inhibit the SAg-independent IFN-γ response. We also observed decreased levels of IL-6 and increased levels of IL-10 in splenocytes from EPS-treated mice (Fig. 5); no differences in IL-2 levels were found, which is consistent with this IFN-γ production being T cell and SAg independent (Fig. 5).
TLR4-dependent reduction of SAg-independent IFN-γ response by EPS. Splenocytes from PBS- or EPS-treated WT or TLR4-deficienct mice were isolated 1 day after treatment and stimulated ex vivo with heat-killed S. aureus. After 24 h, levels of cytokines were determined using CBA. Bars represent means. Each triangle represents data from one mouse. Analyzed using Student's t test. n = 12 to 15 (WT), 6 (TLR4 KO). *, P < 0.05; ***, P < 0.001; ****, P < 0.0001.
EPS protection from Citrobacter rodentium colitis requires TLR4 as does the induction of anti-inflammatory macrophages (20, 21). We tested if EPS requires TLR4 to limit the NK cell IFN-γ response to S. aureus. WT and TLR4-deficient mice were treated with EPS, and splenocytes were treated with heat-killed S. aureus ex vivo, and after testing the IFN-γ response, we found that EPS did not reduce IFN-γ levels in TLR4-deficient mice (Fig. 5). Because the IFN-γ response of splenocytes to heat-killed S. aureus is NK-derived, we conclude that TLR4 is required for EPS inhibition of IFN-γ production by NK cells. No differences in IL-6 and IL-10 levels were observed in heat-killed S. aureus-stimulated cells from EPS-treated TLR4-deficient mice (Fig. 5), suggesting that EPS modulation of the host response to heat-killed S. aureus requires TLR4. Together, these data indicate that EPS modulates host immunity through TLR4, resulting in an altered response to S. aureus that reduces proinflammatory IFN-γ and IL-6 responses while enhancing anti-inflammatory IL-10 production.
Requirement for TLR4 for EPS protection from S. aureus bloodstream infection.We previously showed that TLR4 is required for the generation of anti-inflammatory macrophages by EPS (20), and above, we showed that TLR4 is required for EPS inhibition of SAg-independent production of IFN-γ in response to heat-killed S. aureus. To test if TLR4 is required for EPS protection from S. aureus bloodstream infection, we treated WT and TLR4-deficient mice with PBS or EPS and infected with S. aureus. We measured cytokine levels 1 d.p.i. and found that while EPS reduced the levels of IFN-γ, keratinocyte chemoattractant (KC), and monocyte chemoattractant protein 1 (MCP-1) in WT mice, the reduction was not significant in TLR4-deficient mice (Fig. 6A), indicating that EPS reduces inflammation during S. aureus infection through TLR4. We also examined the weights of infected mice and found that while EPS treatment reduced weight loss in WT mice at 1 d.p.i., it did not prevent weight loss in TLR4-deficient mice (Fig. 6B), indicating that EPS reduces disease burden mediated by TLR4. Lastly, we assessed bacterial load within the organs of these mice and found that EPS-treated WT mice had reduced S. aureus CFU in spleen and liver as reported previously (10), while in TLR4-deficient mice, this difference in CFU was not observed in the liver and was much reduced in the spleen (Fig. 6C), suggesting that EPS reduction of bacterial burden is also mediated by TLR4. Together, these data indicate that EPS reduces bacterial burden and cytokine production and prevents production of IFN-γ during S. aureus bloodstream infection through modulation of TLR4, leading to reduced disease burden.
EPS attenuation of S. aureus bloodstream infection through TLR4. WT or TLR4-deficient mice were treated with PBS or EPS by i.p. injection 1 day prior to i.v. infection with S. aureus. (A) Serum cytokine and chemokine levels at 1 d.p.i. determined using CBA. (B) Body weight at 1 d.p.i. relative to weight prior to infection. (C) S. aureus CFU within organ homogenates at 1 d.p.i. Bars represent means. Each triangle represents data from one mouse. Data pooled from 3 independent experiments. Analyzed using Student's t test. n = 7 to 12. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.
DISCUSSION
S. aureus bloodstream infections are a significant challenge for health care professionals despite aggressive therapeutic measures. The success of S. aureus as a pathogen is attributed to the vast array of virulence factors that manipulate and subvert the host response, promoting immune evasion and pathogenesis. IFN-γ is a prime example in that it is known as a classic antimicrobial mediator but is used by S. aureus to enhance its pathogenesis. IFN-γ-deficient mice have improved survival rates after i.v. S. aureus challenge (11), and EPS limits IFN-γ production after S. aureus bloodstream infection (10). EPS inhibits S. aureus SAg-induced T cell activation (10), and it is well known that this stimulation induces production of IFN-γ (22). We found that IFN-γ is also induced by S. aureus molecules other than SAg, presumably by a surface molecule present on the heat-killed bacterium. After S. aureus infection, IFN-γ levels have biphasic kinetics in which low levels are detectable early after infection (day 1), undetectable levels are found during the next couple of days, and high levels appear by 5 d.p.i. (8). We suggest that the SAg-independent induction of IFN-γ is part of the early host response, whereas the SAg induction of IFN-γ occurs later during infection. We found SAg-independent reduction of IFN-γ production by EPS in both serum and in tissues, especially the spleen and liver. The kidney showed no reduction in IFN-γ levels or in bacterial load after administration of EPS, suggesting that the kidney has an alternative immunological response to infection.
SAg-independent IFN-γ induction from NK cells.We searched for the cells responsible for the SAg-independent induction of IFN-γ by S. aureus and found that spleen cells depleted of NK1.1+ cells lost the capacity to produce IFN-γ in response to heat killed S. aureus, absent SAg. While we conclude that NK cells are the cells responsible for producing the IFN-γ, some of these may be NKT cells because depletion of CD3+ T cells from splenocytes stimulated with heat-killed S. aureus also reduced the amount of IFN-γ produced (Fig. 4A). This SAg-independent IFN-γ response was eliminated if anti-IL-12 antibody was added to the cultures. We think that the IL-12 is likely produced by myeloid cells after TLR2 stimulation by S. aureus lipopeptides and that this in turn activates NK (and potentially NKT) cells to produce IFN-γ. This view is consistent with our finding that NK cells, myeloid cells, and MyD88 signaling are required for SAg-independent IFN-γ induction. We suggest that EPS inhibits this IFN-γ production either by reducing myeloid cell production of IL-12 or by reducing NK cell sensitivity to IL-12 activation. EPS functions in a TLR4-dependent manner and, accordingly, may directly inhibit production of IFN-γ in the NK cell response to heat-killed S. aureus or decrease expression of IL-12R. Levels of IL-12, however, were below the limit of our detection, and we cannot distinguish between these two possibilities. We also found reduced IL-6 and increased IL-10 production in EPS-treated mice in response to heat-killed S. aureus, suggesting that EPS may have induced an anti-inflammatory phenotype in myeloid cells. The finding that S. aureus induces NK (and potentially NKT) cells to produce IFN-γ in a SAg-independent manner demonstrates yet another means by which S. aureus hijacks the innate immune system for its benefit.
TLR4-mediated induction of an anti-inflammatory state by EPS.EPS induces anti-inflammatory macrophages that limit T cell activation through TLR4 (20), and it also limits NK cell production of IFN-γ in response to S. aureus in a TLR4-dependent manner. Similarly, EPS protection from S. aureus bloodstream infection requires TLR4. Activation of TLR4 is not generally associated with limiting host responses, as lipopolysaccharide (LPS), the classical TLR4 agonist, is known to induce IFN-γ production (23). However, activation of TLR4 by a ligand is a complex process involving multiple coreceptors that can alter cellular responses (24). In the case of bacterial ligands, polysaccharide A (PSA) from Bacteroides fragilis is known to limit inflammation through TLR2-dependent activation of dendritic cells, and large polysaccharides from Helicobacter hepaticus modulate macrophages to have an anti-inflammatory signature mediated by TLR2 (25–29). We found that splenocytes from EPS-treated mice produced lower levels of IL-6 and increased levels of IL-10 upon stimulation with heat-killed S. aureus, indicating that EPS induced an anti-inflammatory state. The induction of such an anti-inflammatory state in a TLR4-dependent manner is consistent with the view that some molecules can induce TLR4-mediated anti-inflammatory signatures, including production of IL-10 (30–33). For example, LPS induces higher levels of IL-10 expression upon restimulation of macrophages, a signature associated with endotoxin tolerance (34). Many commensal bacteria likely also produce TLR ligands but seemingly coexist with the host without uncontrolled inflammatory processes. We think the anti-inflammatory state induced by EPS through TLR4 reflects the complexity of TLR signaling, presumably influenced by the nature of the bacterial ligand.
IFN-γ is the classic cytokine that enhances antimicrobial functions of phagocytes (6, 7), and it is paradoxical that IFN-γ is hijacked by S. aureus to promote its persistence and pathogenesis. While S. aureus is known to induce synthesis of IFN-γ by SAg stimulation of T cells, here, we discovered that S. aureus also induces NK/NKT cells to produce IFN-γ independent of SAg. While we have not identified the molecule that stimulates NK cells, we think it is likely lipoprotein or another TLR2 agonist that resides on heat-killed S. aureus. We showed that this SAg-independent production of IFN-γ is inhibited by EPS derived from the probiotic B. subtilis. We think that by inhibiting IFN-γ production, EPS interferes with some aspect of S. aureus pathogenesis, thereby reducing disease burden and improving outcomes. The inhibition of SAg-independent production of IFN-γ by EPS is TLR4 dependent. Further, in a TLR4-dependent manner, EPS inhibits disease symptoms, including weight loss, cytokine production, and reduction in CFU in spleen and liver after bloodstream infection with S. aureus. We conclude that EPS modulates host immunity to S. aureus through TLR4 to induce an anti-inflammatory state, resulting in reduced disease burden and improved host outcomes. We suggest that TLRs may be a key therapeutic target not only to stimulate host immunity but also to limit inflammation and promote host survival during deadly diseases, such as S. aureus bloodstream infection.
MATERIALS AND METHODS
Mice and reagents.C57BL/6J and TLR4−/− mice were purchased from Jackson Laboratories (Bar Harbor, ME) and bred in-house at Loyola University Chicago. Four- to eight-week-old mice were used for all studies. We did not observe any noticeable differences between male and female mice. All experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee at Loyola University Chicago. Cell culture base medium and supplements were from Life Technologies (Grand Island, NY), bacterial growth media were from BD (Franklin Lakes, NJ), and antibodies were from BioLegend (San Diego, CA). Starter cultures of S. aureus USA300 strain AH1263 (erythromycin-sensitive derivative of strain LAC [35]), JE2 (USA300, parental strain for the University of Nebraska transposon mutant library, also derived from strain LAC [36]), or seq::erm (SEl-Q-deficient transposon mutant [36]) were cultured in tryptic soy broth (TSB). For bloodstream infection experiments, overnight cultures were diluted 1:100 into fresh growth medium, incubated at 37°C until reaching exponential phase (∼3 h), and normalized to an inoculum of 108 CFU/ml in PBS on the day of infection.
Generation of AH1263 seq::erm.A marked seq::erm insertional disruption in parental strain AH1263 was generated by bacteriophage-mediated transduction using phage ϕ11 packaged with DNA from transposon mutant NE1605 of the Nebraska transposon mutant library (36). Transduction was carried out as previously described (37), and the insertional disruption in seq was confirmed by PCR using oligonucleotides Buster (TGTTTTTTAAGTAAATCAAGTAC) and seq (TTAACTAGAGAGTTCCCATT).
Preparation of B. subtilis-derived EPS.EPS was prepared as previously described (10, 20, 21) with few modifications. B. subtilis DS991 (sinR::kan tasA::spec, overproduces and secretes EPS) or DK4623 (ΔEPS, sinR::kan tasA::spec sdpABC::mls skf::tet lytC::cat ΔPBSX ΔSPB ΔpBS32, DS991 with lytic genes [38] and prophages deleted), generously provided by Daniel B. Kearns (Indiana University, Bloomington, IN), was cultured in 1% tryptone-phosphate broth (1% tryptone, 25 mM phosphate, 0.1 M NaCl) or MSgg (minimal salts glutamate glycerol) medium. EPS was obtained from stationary-phase culture supernatants (10, 20) or from bacterial lawns on Luria-Bertani agar plates by 75% ethanol precipitation at −20°C. The precipitate was pelleted by centrifugation (13,700 × g, 4°C, 30 min), resuspended in 0.1 M Tris (pH 8), and treated with DNase (67 μg/ml) and RNase (330 μg/ml) at 37°C for 2 h, followed by proteinase K (40 μg/ml) digestion at 55°C for 2 h. EPS was then purified on DEAE-cellulose (Whatman, Maidstone, UK) ion exchange and/or gel filtration on Sephacryl S-500. Carbohydrate-positive fractions were identified by a modified phenol sulfuric acid assay (39, 40) and desalted by gel filtration (Pharmacia Fine Chemicals, Piscataway, NJ). Total carbohydrate content was measured using the modified phenol sulfuric acid assay (39, 40). All EPS preparations contained undetectable levels of protein or nucleic acid by absorption at 280 nm and 260 nm, respectively, and were shown to induce peritoneal hybrid macrophages using flow cytometry (10, 20).
Murine model of S. aureus bloodstream infection.C57BL/6J or TLR4-deficient (TLR4−/−) mice were treated with 3 mg/kg of body weight EPS by intraperitoneal (i.p.) injection in 200 μl PBS at −1 and 1 day postinfection (d.p.i.). Control mice were injected with equal volume PBS. On day 0, anesthetized mice were infected with 107 CFU of S. aureus in 100 μl PBS by retro-orbital injection (10, 41, 42). Mice with >20% body weight loss that showed signs of lethargy were euthanized during the experiment. On 1 d.p.i., S. aureus CFU in spleen, liver, and kidney homogenates of euthanized mice were quantified by plating on tryptic soy agar plates and enumerating CFU after incubation for 12 h at 37°C. Tissue homogenates were cleared by centrifugation at 2,000 × g for 5 min, and supernatants were collected along with serum samples for cytokine and chemokine analysis.
Cytokine and chemokine measurements.Levels of proinflammatory cytokines and chemokines in serum, tissues homogenates, and culture supernatants were measured in triplicate using cytometric bead array (CBA) (BD Biosciences) according to the manufacturer’s specifications. Samples were analyzed on an LSRFortessa (BD Biosciences), and data were analyzed using FCAP Array software v3.0 (BD Biosciences) or FlowJo software (Ashland, OR).
Stimulation of splenocytes.Splenocytes (3.5 × 105) from C57BL/6J or MyD88−/− mice were seeded in a 96-well flat-bottom plate and stimulated with either S. aureus cell-free culture supernatant at 33% final concentration to evaluate superantigen-dependent induction of IFN-γ (10) or heat-killed S. aureus at a multiplicity of infection (MOI) of 10 to evaluate superantigen-independent induction of IFN-γ. To collect S. aureus cell-free culture supernatant, starter cultures (wild-type [WT] and seq::erm) were subcultured 1:100 in RPMI medium without antibiotics for 9 h. LAC and JE2 S. aureus strains encode three known SAg (SEl-K, SEl-Q, and SEl-X), but only SEl-Q has SAg activity in C57BL/6J mice (10). Therefore, a seq::erm mutant was used in all studies testing superantigen dependency. After normalizing cultures to optical density at 600 nm, supernatants were collected by centrifuging cultures at 2,000 × g and passing supernatant through a 0.2-μm filter to remove bacterial cells. To prepare heat-killed S. aureus, bacterial cell pellets from overnight cultures were washed 3 times with PBS and normalized to 3.5 × 108 CFU/ml in PBS followed by incubation at 65°C for 1 h to kill all bacteria. At various time points after stimulation of splenocytes with bacterial supernatant or heat-killed bacteria, supernatants were collected, and the levels of cytokines and chemokines were quantified using CBA. For experiments with EPS, C57BL/6J or TLR4−/− mice were treated with 3 mg/kg EPS by i.p. injection 1 day prior to tissue collection. For depletion experiments, CD3−, NK1.1−, or CD11b− cells were purified using the BD IMag cell separation system (BD Biosciences) after incubation with anti-CD16/32 (Mouse BD Fc Block, clone 93) and biotinylated anti-CD3 (145-C211), anti-NK1.1 (PKC146), or anti-CD11b (M1/70) antibodies. For neutralization experiments, 10 μg/ml rat IgG2a isotype control (RTK2758), 1 μg/ml anti-IL-12p40 (C17.8), or 10 μg/ml anti-IL-18 (YIGIF74-1G7, Bio X Cell) antibodies were included in the medium.
ACKNOWLEDGMENTS
We acknowledge Daniel B. Kearns (Indiana University) for generously providing the B. subtilis strains used in this study and Mae Kingzette, Pi-Chen Yam, and Jesús Zamora-Pineda for purification of EPS.
FOOTNOTES
- Received 23 August 2019.
- Returned for modification 27 September 2019.
- Accepted 3 January 2020.
- Accepted manuscript posted online 13 January 2020.
- Copyright © 2020 American Society for Microbiology.
