Probiotic Exopolysaccharide Protects against Systemic Staphylococcus aureus Infection, Inducing Dual-Functioning Macrophages That Restrict Bacterial Growth and Limit Inflammation

Staphylococcus aureus causes severe systemic infection with high mortality rates. We previously identified exopolysaccharide (EPS) from a probiotic, Bacillus subtilis, that induces anti-inflammatory macrophages with an M2 phenotype and protects mice from Citrobacter rodentium-induced colitis.

ABSTRACT

Staphylococcus aureus causes severe systemic infection with high mortality rates. We previously identified exopolysaccharide (EPS) from a probiotic, Bacillus subtilis, that induces anti-inflammatory macrophages with an M2 phenotype and protects mice from Citrobacter rodentium-induced colitis. We tested if EPS could protect from systemic infection induced by S. aureus and found that EPS-treated mice had enhanced survival as well as reduced weight loss, systemic inflammation, and bacterial burden. While macrophages from EPS-treated mice display an M2 phenotype, they also restrict growth of internalized S. aureus through reactive oxygen species (ROS), reminiscent of proinflammatory phagocytes. These EPS-induced macrophages also limit T cell activation by S. aureus superantigens, and EPS abrogates systemic induction of gamma interferon after infection. We conclude that B. subtilis EPS is an immunomodulatory agent that induces hybrid macrophages that bolster antibacterial immunity and simultaneously limit inflammation, reducing disease burden and promoting host survival.

INTRODUCTION

Staphylococcus aureus is a Gram-positive bacterium that causes a wide array of human diseases (1). S. aureus is generally found on skin and in anterior nares with no apparent harm (2), but upon entering the bloodstream, it can disseminate and colonize virtually all organs, resulting in severe systemic disease with up to 50% mortality (1, 3). The host rapidly recognizes the pathogen upon invasion and mounts proinflammatory responses targeted to clear the infection. However, S. aureus evades and resists many aspects of host immunity (4), resulting in devastating outcomes mediated by overt inflammation and tissue damage. Antibiotics that directly target S. aureus are the only option for patients with systemic S. aureus disease, but widespread prevalence of antibiotic-resistant strains limits treatment efficacy and many patients experience persistent bacteremia or succumb to disease despite aggressive therapy (1, 5, 6). Novel approaches for treatment and prevention of systemic S. aureus infections are critically needed.

Within tissues, S. aureus is rapidly recognized by host pattern recognition receptors, and polymorphonuclear cells (PMN) with potent antibacterial activity are recruited (7). However, S. aureus resists antimicrobial activities of PMN, requiring additional immune cells to control infection (4, 7). Macrophages (MΦ) are professional phagocytes that have antibacterial activity (8), facilitate recruitment of immune cells, coordinate adaptive immunity, and promote resolution of inflammation and wound healing (9, 10). MΦ can exhibit several functions due to their extensive heterogeneity (9). Generally, MΦ are classified into two polarized states: classically activated M1 MΦ, associated with antimicrobial activity, and alternatively activated M2 MΦ, associated with immune regulation and wound healing (11). M2 MΦ can be further characterized in four subsets (12, 13), increasing overall heterogeneity of MΦ. This heterogeneity allows the MΦ compartment to coordinate immune responses during infection, limiting bacterial growth and inflammation while promoting tissue repair to restore homeostasis. S. aureus, however, can subvert MΦ functions by each of several mechanisms (8, 14–16), and an attractive strategy to reduce disease burden would be to modulate MΦ in a manner that would bolster antibacterial immunity and promote resolution of inflammation.

Probiotics are microorganisms that can be used to benefit the host, and numerous probiotic products are widely available (17). However, the specific mechanisms by which probiotics function remain mostly unknown. Bacillus subtilis is a Gram-positive spore-forming probiotic bacterium that can limit murine colitis induced by the enteric pathogen Citrobacter rodentium (18, 19). We demonstrated that protection by B. subtilis from disease caused by C. rodentium requires production of exopolysaccharide (EPS) (19), and administration of purified EPS to mice via intraperitoneal (i.p.) injection recapitulates this protection (20). Further, we found that i.p. injection of EPS results in the induction of peritoneal MΦ that display an M2 phenotype and inhibit T cell activation in vitro and in vivo (21), suggesting that EPS mediates its protection by inducing anti-inflammatory M2 MΦ. We hypothesized that this anti-inflammatory property of EPS could be used to limit inflammation during systemic S. aureus infection, preserving tissue integrity and thereby improving outcomes.

Here, we demonstrate that mice treated with EPS exhibit reduced weight loss, systemic inflammation, and bacterial load during S. aureus bloodstream infection, leading to increased survival. Mechanistically, we found that EPS abrogates early induction of gamma interferon (IFN-γ) in vivo during infection and that MΦ from EPS-treated mice restrict growth of internalized S. aureus ex vivo through enhanced production of reactive oxygen species (ROS). Further, these MΦ retained their capacity to suppress T cell activation induced by S. aureus superantigen (SAg)-containing culture medium. These data suggest that EPS induces cells possessing both antibacterial and anti-inflammatory properties, thereby providing the means to bolster S. aureus clearance and also maintain tissue integrity, processes that would improve the outcomes of lethal S. aureus infection.

RESULTS

Attenuation of S. aureus bloodstream infection by B. subtilis-derived EPS.

EPS prevents systemic inflammation by inducing M2 MΦ that suppress T cell activation (21), and we hypothesized that EPS could also improve the outcome of systemic infection by S. aureus, where inflammation plays a role in pathogenesis (4, 22). We administered EPS 1 day prior to and 1 day after systemic infection with an epidemic strain of S. aureus USA300 (LAC) and found that EPS-treated mice lost less weight than phosphate-buffered saline (PBS)-treated infected controls (Fig. 1A). Even a single dose of EPS benefitted the host, since we observed reduction in weight loss by 1 day postinfection (dpi), prior to the second administration of EPS (Fig. 1A). EPS-treated mice also had improved survival compared to that of PBS-treated mice (Fig. 1B). Given that EPS limits inflammation in disease caused by infection with C. rodentium (20, 21) and reduces weight loss early in systemic S. aureus infection, we assessed if the levels of serum proinflammatory chemokines and cytokines were decreased in EPS-treated mice. Indeed, these mice had lower levels of MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), tumor necrosis factor (TNF), and gamma interferon (IFN-γ) than did PBS-treated mice at 1 dpi (Fig. 1C), indicating that EPS limits systemic inflammation early during S. aureus infection. While EPS does not appear to affect pathogen colonization in C. rodentium-induced colitis and anti-inflammatory M2 MΦ are not generally associated with enhanced antibacterial immunity (19, 23, 24), we nevertheless assessed bacterial load within organs of S. aureus-infected mice. We did not observe any difference in S. aureus CFU in the kidneys from EPS-treated mice at either 1 or 3 dpi (Fig. 2A). However, the bacterial burden in these mice was reduced 3-fold in the blood 6 h postinfection (hpi) (Fig. 2B). In addition, the number of S. aureus CFU was reduced in the spleen and the liver 8-fold and 3-fold, respectively, 1 dpi (Fig. 2C and D) and by 4-fold in the liver 3 dpi (Fig. 2D). These data indicate that EPS limits bacterial burden during systemic S. aureus infection.

FIG 1.

Effect of B. subtilis EPS on disease after S. aureus bloodstream infection. Mice were treated with EPS (1 to 3 mg/kg) or PBS 1 day prior to and 1 day after systemic infection with S. aureus. (A) Body weight loss. Differences were analyzed using two-way analysis of variance (ANOVA) with Bonferroni’s multiple-comparison test. Error bars represent standard deviations (SD). (B) Percent survival as determined by >20% body weight loss. Differences in curves were analyzed using log-rank (Mantel-Cox) test. n = 17 (PBS) or n = 22 (EPS). Data were pooled from 7 independent experiments. (C) Serum level (1 dpi) of cytokines and chemokines measured by CBA. n = 12 to 16. Data were pooled from 3 to 4 independent experiments. Each dot represents data from individual mice. Bar represents means. Data were analyzed using unpaired, two-tailed Student's t test. IFN-γ data were analyzed using one-sample t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 2.

Effect of EPS on S. aureus burden during bloodstream infection. Mice were treated with EPS or PBS 1 day prior to systemic infection with S. aureus, and bacterial CFU in different organs were determined. (A) Numbers of S. aureus CFU per organ in kidney 1 and 3 dpi. (B) Numbers of S. aureus CFU/ml in blood 6 hpi. (C) Numbers of S. aureus CFU per organ in spleen 1 dpi. (D) Numbers of S. aureus CFU per organ in liver 1 and 3 dpi. n = 8 to 10 (6 hpi or 1 dpi) or n = 17 to 18 (3 dpi). Data were pooled from 2 to 5 independent experiments. Each dot represents an individual mouse. Bars represent means. Data were analyzed using unpaired, two-tailed Student's t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

ROS-mediated inhibition of S. aureus growth by EPS-induced MΦ.

With acute S. aureus infection, monocytes are recruited to sites of infection, where they differentiate into classically activated, proinflammatory M1 MΦ that mediate antibacterial immunity and control bacterial load (8, 9). On the contrary, M2 MΦ generally have reduced antimicrobial functions, such as nitric oxide production and secretion of proinflammatory cytokines, and are thought to enhance bacteremia and sepsis (23, 24). How, then, could EPS treatment, which induces M2 MΦ, lead to reduced bacterial load? We hypothesized that EPS could increase the uptake of S. aureus by immune cells and/or enhance their bactericidal activity, leading to reduced numbers of bacteria during infection. To test this, we isolated MΦ (F4/80⁺) from the peritoneal cavity of EPS-treated mice and incubated them with serum-opsonized S. aureus. After washing the cells free of bacteria, we determined S. aureus CFU in cell lysates. No significant difference in the number of internalized S. aureus CFU between cells from PBS- or EPS-treated mice was found (Fig. 3A, 0.5 h, left). We continued to culture the infected cells and found that by 5 h, the numbers of S. aureus CFU were significantly lower in cultures of cells from EPS-treated mice (Fig. 3A, left), indicating that cells from these mice restrict growth of internalized S. aureus more than cells from PBS-treated mice. EPS did not directly suppress S. aureus growth, since its growth in vitro was unaffected by the presence of EPS (up to 0.5 mg/ml) in culture medium (see Fig. S1 in the supplemental material). Further, F4/80⁻ cells from EPS-treated mice did not limit growth of S. aureus (Fig. 3A, right), indicating that EPS exerts its effect specifically through MΦ.

FIG 3.

Effect of EPS on MΦ capacity to inhibit growth of internalized S. aureus. Mice were treated with EPS or PBS, and 3 days later peritoneal cells were isolated. (A) Peritoneal cells were sorted into F4/80⁺ and F4/80⁻ cells using magnetic selection and infected with S. aureus for 30 min; CFU were monitored over time. Differences were analyzed using two-way ANOVA with Bonferroni’s multiple-comparison test. n = 5 to 8. Data were pooled from 3 to 4 independent experiments. (B) Peritoneal cells were infected with S. aureus and stained with CellROX green. Cells were analyzed by flow cytometry after gating on CD11b^high F4/80^high cells. Data were normalized to uninfected cells from PBS-treated mice. Differences were analyzed using one-way ANOVA with Bonferroni’s multiple-comparison test. n = 9. Data were pooled from 3 independent experiments. (C) Same setup as that for panel A but in the presence or absence of 2 μM DPI. Differences were analyzed using 2-way ANOVA with Bonferroni’s multiple-comparison test. n = 6. Data were pooled from 3 independent experiments. Error bars represent SD in panels A and C, and bars represent means in panel B. Each dot represents data from a single mouse. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Phagocytes utilize multiple mechanisms to respond to internalized pathogens (8), but one common mechanism is through production of ROS that induce oxidative damage and kill bacteria (8, 25). After uptake, MΦ generate ROS through an NADPH oxidase complex that catalyzes the reaction between O₂ and NADPH to generate O₂⁻, which is used to generate a variety of reactive oxidants, a process known as respiratory burst (25). Since EPS-induced MΦ inhibit growth of internalized S. aureus, we thought that EPS may increase ROS levels in MΦ during infection with S. aureus, thereby enhancing their antibacterial capacities. We isolated total peritoneal cells from EPS-treated mice and infected them with serum-opsonized S. aureus. After washing infected cells, we measured the level of ROS in CD11b^high F4/80^high MΦ by flow cytometry using an ROS indicator, CellROX green. Although we did not find differences in levels of CellROX staining between uninfected MΦ isolated from PBS- or EPS-treated mice (Fig. 3B), upon infection, the MΦ from EPS-treated mice had increased CellROX green staining while those from PBS-treated mice did not (Fig. 3B), indicating that MΦ from EPS-treated mice increase cellular ROS levels in response to S. aureus infection. We hypothesized that this increase in ROS levels is responsible for inhibition of S. aureus growth by MΦ from EPS-treated mice and tested this by repeating the assay shown in Fig. 3A in the presence of diphenyleneiodonium (DPI), a noncompetitive NADPH oxidase inhibitor. We found that the inhibition of growth of internalized S. aureus by MΦ from EPS-treated mice was abrogated in the presence of DPI (Fig. 3C), suggesting that these MΦ inhibit growth of S. aureus through ROS.

Inhibition of S. aureus superantigens by EPS-induced MΦ.

EPS induces anti-inflammatory M2 MΦ that limit T cell activation (21), but M2 MΦ are not generally associated with bactericidal activity (24). Paradoxically, as shown above, we found that EPS-induced MΦ inhibit S. aureus growth through ROS, like that of proinflammatory M1 MΦ (24). MΦ have a broad spectrum of functions (11) and are plastic in nature, constantly modulating their functions in response to environmental factors (26). A major pathway by which innate immune cells are activated by S. aureus is the Toll-like receptor 2 (TLR2)-myeloid differentiation primary response 88 (MyD88) pathway (27), which recognizes S. aureus-derived lipoproteins (28) and induces proinflammatory activation of MΦ (24). We hypothesized that upon exposure to S. aureus and its activating ligands, EPS-induced MΦ could change to a proinflammatory phenotype and gain antibacterial function. In this case, we expected that M2 MΦ polarization would be lost and the MΦ would no longer inhibit T cell activation. To test this possibility, we devised a T cell activation assay using S. aureus culture supernatants that contain both S. aureus lipoproteins that drive proinflammatory activation of MΦ (8, 28) and superantigens (SAg) that drive T cell activation and proliferation (29). The results show that whereas S. aureus culture supernatants of LAC and JE2 strains (both USA300 lineages), which produce staphylococcal enterotoxin-like (SEl)-K, -Q, and -X SAgs (30), stimulated splenic T cell proliferation, the supernatant from an SEl-Q-deficient mutant (selQ::Tn) did not (Fig. 4 and Fig. S2), indicating that T cell activation was dependent on SEl-Q SAg. Deficiency in another gene encoding a putative type A enterotoxin (USA300_1559::Tn) did not affect T cell activation (Fig. 4 and Fig. S2). To test if EPS-induced MΦ inhibit SEl-Q-mediated T cell proliferation in the presence of inflammatory factors of S. aureus, we cultured peritoneal F4/80⁺ MΦ from PBS- or EPS-treated mice with naive splenocytes that were stimulated with culture supernatant of S. aureus USA300 (LAC). Flow cytometric analysis showed a decrease in CD25⁺ CD44⁺-activated CD4⁺ and CD8⁺ T cells (Fig. 5A and Fig. S3) as well as reduced T cell proliferation (Fig. 5B and Fig. S3), showing that EPS-induced MΦ still inhibit T cell activation by S. aureus SEl-Q SAg, even in the presence of other proinflammatory S. aureus factors.

FIG 4.

Effect of S. aureus SAg on CD4⁺ and CD8⁺ T cell proliferation. CTV-labeled splenocytes from naive mice were cultured in the presence of 33% culture supernatant of S. aureus or SAg-deficient mutants for 4 days and analyzed for T cell proliferation by flow cytometry. Proliferative index represents the percentage of events within the proliferated gate, as shown in Fig. S2. Error bars represent SD. n = 4 to 6. Data were pooled from 2 independent experiments. Data were analyzed using one-way ANOVA with Bonferroni’s multiple-comparison test. ***, P < 0.001.

FIG 5.

Effect of EPS-induced MΦ on CD4⁺ and CD8⁺ T cell activation by S. aureus SAg. CTV-labeled naive splenocytes were cocultured with peritoneal F4/80⁺ cells from PBS- or EPS-treated mice for 4 days in the presence of 33% S. aureus culture supernatant and analyzed for T cell activation (A) and proliferation (B) by flow cytometry. Data were pooled from 3 independent experiments. Bars represent means. n = 8. Data were analyzed using Student's t test. *, P < 0.05; **, P < 0.01.

SAgs are a classical virulence factor of S. aureus that cross-link T cell receptors (TCR) and major histocompatibility complex (MHC) molecules on accessory cells to drive polyclonal T cell activation and proinflammatory cytokine production, resulting in toxic shock syndrome (31). Importantly, SAg induces IFN-γ production early in S. aureus systemic infection, thereby promoting pathogen survival in vivo (32). Because EPS abrogated levels of early IFN-γ production during S. aureus infection (Fig. 1C), we hypothesized that EPS inhibits the stimulatory effects of S. aureus SAgs. We tested this by isolating splenocytes from PBS- or EPS-treated mice and stimulating them ex vivo with S. aureus culture supernatants. Indeed, splenocytes from EPS-treated mice had reduced T cell proliferation compared to that of cells from PBS-treated mice (Fig. 6 and Fig. S4). We conclude that EPS-induced MΦ retain their anti-inflammatory function to limit T cell activation by S. aureus in vitro and that splenic T cells from EPS-treated mice have a diminished response to S. aureus SAg.

FIG 6.

Effect of EPS on splenic CD4⁺ and CD8⁺ T cell responses to S. aureus SAg. CTV-labeled splenocytes from PBS- or EPS-treated mice were cultured for 4 days in the presence of 33% S. aureus culture supernatant and analyzed by flow cytometry. Data were pooled from 3 independent experiments. NT, not tested. Bars represent means. n = 6 to 7. Data were analyzed using one-way ANOVA with Bonferroni’s multiple-comparison test. *, P < 0.05; ***, P < 0.001.

Induction of hybrid M1-M2 MΦ by EPS.

MΦ are generally described as proinflammatory M1 or anti-inflammatory M2. MΦ have been studied primarily using in vitro polarization with lipopolysaccharide (LPS) and IFN-γ, which induce classical M1 MΦ that express proinflammatory cytokines and inducible nitric oxide synthase (iNOS), and with interluekin-4 (IL-4) and IL-13, which induce alternative M2 MΦ, as defined by arginase-1 (Arg-1) expression (11, 12). Both iNOS and Arg-1 utilize arginine as their substrate, and it is thought that these two enzymes negatively regulate each other, contributing to polarization of MΦ (33). However, MΦ are plastic in nature and hybrid M1/M2 MΦ have been reported (34). Because EPS-induced MΦ inhibit S. aureus growth through ROS, as do M1 MΦ, and also inhibit T cell activation by S. aureus SAg, as do M2 MΦ, we hypothesized that EPS-induced MΦ represent hybrid M1- and M2-like MΦ. To test this possibility, we assessed expression of iNOS and Arg-1 in peritoneal CD11b^high F4/80^high MΦ in EPS-treated mice and found a significant increase in both iNOS and Arg-1 expression, indicating that EPS-induced MΦ coexpress M1 and M2 MΦ markers (Fig. 7). These data further indicate that EPS induces a hybrid M1- and M2-like MΦ in vivo.

FIG 7.

Effect of EPS on M1 and M2 MΦ marker expression levels. Peritoneal CD11b^high F4/80^high MΦ from PBS (solid line)-, ΔEPS (dotted line)-, or EPS (dashed line)-treated mice were analyzed for M2 (Arg-1) and M1 (iNOS) markers by flow cytometry. (A) Representative flow cytometry histogram plot. (B) Median fluorescence intensity compared to that of PBS-treated MΦ. Data were pooled from three independent experiments. n = 5 to 10. Data were analyzed using one-way ANOVA with Bonferroni’s multiple-comparison test. **, P < 0.01; ***, P < 0.001.

DISCUSSION

Systemic infection by S. aureus is a severe challenge to the host, requiring strong antibacterial immunity while limiting overt inflammation. We demonstrated that B. subtilis-derived EPS protects mice from systemic S. aureus infection, and that this protection is likely due to reduction in bacterial load and limitation of inflammation by MΦ that both suppress S. aureus growth through ROS and limit T cell activation by SAg. EPS-treated mice display reduced serum levels of proinflammatory cytokines and chemokines within 24 h after S. aureus infection. A surprising finding was the reduced level of IFN-γ in EPS-treated mice infected with S. aureus. Classically, IFN-γ is the defining agent of type 1 immunity (35) and mediates protection against bacterial infections by enhancing antimicrobial functions of phagocytes (36–38). Although administration of recombinant IFN-γ led to improved survival in murine systemic S. aureus infection in one study (39), other studies showed that neutralization of IFN-γ or IFN-γ deficiency improved survival of S. aureus-infected mice with reduced bacterial burden in the kidney (40, 41), suggesting that IFN-γ accentuates disease. Our data are consistent with the idea that IFN-γ contributes to S. aureus pathogenesis, because disease was attenuated in EPS-treated mice that had decreased levels of IFN-γ.

IFN-γ is produced primarily by CD4⁺ T cells (29, 37, 42), and SAgs are major S. aureus virulence factors that induce T cells as well as NKT cells to produce IFN-γ (29, 43). Whereas infection with wild-type S. aureus increases serum IFN-γ levels within 8 h, infection with a Δsea mutant (deficient in staphylococcal enterotoxin A SAg) does not, and bacterial load is reduced in the liver by 4 dpi (32). These findings suggest that SAg drives the production of IFN-γ, which promotes bacterial survival, although the pathogenic process for this is not known. Our finding that EPS-induced MΦ inhibit S. aureus SAg activation of splenic T cells suggests that EPS functions in vivo by limiting SAg activation of T cells and subsequent IFN-γ production, thereby neutralizing the pathogenic effects of IFN-γ. The precise role of IFN-γ, a cytokine classically associated with enhanced antibacterial immunity, plays in S. aureus pathogenesis provides an interesting avenue of study to better understand how pathogens exploit host immunity to promote its survival and disease.

The mechanism by which EPS-induced MΦ inhibit S. aureus-induced T cell activation is not known. Although it could result from decreased expression of MHC class II molecules, which are recognized by SAgs, EPS does not alter MHC class II expression on MΦ (unpublished data). Instead, because we previously showed that EPS-induced MΦ inhibit anti-CD3ε activation of T cells through PD-L1 and transforming growth factor beta (TGF-β) and that EPS promotes regulatory T (T_reg) cells, we suggest that the inhibition of T cell activation by S. aureus is due to the upregulated expression of inhibitory molecules such as PD-L1, PD-L2, TGF-β, or IL-10 or to the generation of T_reg cells (21). Alternatively, the inhibition may be due to depletion of arginine through iNOS and Arg-1 (44–46), both of which are expressed in EPS-induced MΦ.

EPS reduces bacterial burden during S. aureus infection in blood, spleen, and liver but not in the kidney. We suggest that the difference in S. aureus CFU is due to differences in the response to SAg stimulation. Infection with a SAg-defective mutant resulted in reduced numbers of S. aureus CFU in the liver but not in kidney (32), indicating that SAg promotes colonization in the liver but not kidney. Kidney, compared to liver and spleen, contains relatively few T and NKT cells, targets of SAg. Because EPS inhibits SAg stimulation of T cells, we think the effect of EPS is primarily in the liver and spleen, where large numbers of T and NKT cells reside (29, 43). Another explanation for the difference in the number of bacterial CFU between organs may come from the myeloid compartment. In the kidney, dendritic cells (DC) serve as sentinels for infection and recruit neutrophils (47), which leads to abscess formation. In contrast, liver harbors Kupffer cells, unique tissue resident MΦ, that clear activated neutrophils and cellular debris (48, 49). Clearance of neutrophils would reduce abscess formation within hepatic tissues and not only help clear bacteria but also maintain MΦ access to S. aureus. In addition, peritoneal MΦ are known to traffic directly to the liver during liver injury to clear activated neutrophils and promote wound repair (50), providing another means of reducing pathogenic neutrophil activation and supplementing MΦ-mediated immunity against S. aureus. The spleen also contains specialized resident MΦ subsets known to promote immunity against bloodborne pathogens (51), and they could contribute to clearing activated neutrophils and provide additional immunity against S. aureus.

EPS induces macrophages that coexpress M1 and M2 markers and display functions of both M1 and M2 macrophages. Subdivision of macrophages into M1 or M2 is based on in vitro polarization conditions, and in vivo, the functions of macrophages span a wide spectrum (11, 12, 26). For example, Arg-1⁺ myeloid cells (M2-like) are associated with persistent S. aureus biofilms (23), but they also inhibit growth of planktonic S. aureus during infection (M1-like) (52). Because of this, we think that EPS-induced MΦ display both M1- and M2-like properties due to as-yet-unidentified environmental cues generated during EPS treatment in vivo. While we cannot rule out the possibility that EPS-induced MΦ contain two distinct groups, one M1-like and one M2-like, hybrid states of MΦ have been described (34), and by flow cytometry, it appears that essentially all MΦ upregulate expression of M1 and M2 markers in response to EPS.

Probiotics are widely marketed for their health benefits (17, 53), but the specific mechanisms by which these products benefit the host are mostly unknown (54). Further, the molecules and their mechanism of action are poorly understood (20, 55–59), with a few exceptions, including polysaccharide A (PSA) and α-galactosylceramide from Bacteroides fragilis (60–62). PSA is a zwitterionic polysaccharide that modulates T cell responses through DCs in a TLR2-dependent manner (61) and protects mice from experimental colitis through an IL-10-dependent mechanism (63). α-Galactosylceramide functions through CD1d-mediated recognition by NKT cells (62). While the structure of B. subtilis EPS is not yet known, it harbors immunomodulatory properties different from those of other known probiotics, because the induction of anti-inflammatory responses is TLR4 dependent and the inhibition of T cell responses is not through IL-10 but through TGF-β and PD-L1 (21). Further, these MΦ not only limit inflammation but also promote antibacterial immunity against an invasive pathogen. Additional knowledge of the structural and functional properties of EPS is needed to understand and appreciate its full potential as an immunomodulatory molecule.

In conclusion, B. subtilis EPS represents a recently discovered probiotic-derived agent that protects hosts from systemic S. aureus infection by bolstering antibacterial immunity of phagocytes through enhanced ROS production and by limiting inflammatory activation of immune cells. We suggest that this multifunctionality allows EPS to counteract several aspects of S. aureus pathogenesis to improve outcomes, an attractive strategy to combat complex systemic infections.

MATERIALS AND METHODS

Mice and reagents.

C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME) and bred in-house at Loyola University Chicago. Four- to 8-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 media were from BD (Franklin Lakes, NJ), and antibodies were from BioLegend (San Diego, CA). Starter cultures of S. aureus USA300 strain LAC or its plasmid-cured derivative, AH1263 (64), were prepared in tryptic soy broth (TSB). These two strains were used interchangeably, and we did not observe notable differences in outcomes. Transposon insertion mutants (sel-k::Tn, sel-q::Tn, sel-x::Tn, and USA300_1559::Tn) from the University of Nebraska transposon mutant library were used for SAg studies (65). For bloodstream infection experiments, overnight cultures were diluted 1:100 into fresh growth medium, incubated at 37°C to exponential phase (∼3 h), and subsequently normalized to an inoculum of 10⁸ CFU/ml in PBS on the day of infection.

Preparation of B. subtilis-derived EPS.

EPS was prepared as previously described (17, 18), with a 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 [66] 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 (21) 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 by DEAE-cellulose (Whatman, Maidstone, UK) ion exchange chromatography and/or gel filtration on Sephacryl S-500. Carbohydrate-positive fractions were identified by a modified phenol sulfuric acid assay (67, 68) and desalted by gel filtration (Pharmacia Fine Chemicals, Piscataway, NJ). Total carbohydrate content was measured using the modified phenol sulfuric acid assay. All EPS preparations were assessed for protein and nucleic acid content by spectrometry and also for their ability to induce peritoneal M2 MΦ, as previously described (21), prior to use.

Murine model of S. aureus bloodstream infection.

C57BL/6J mice were treated with 3 mg/kg of body weight EPS by intraperitoneal (i.p.) injection in a 200-μl final volume in PBS 1 day before and 1 day after infection. Control mice were injected with equal volumes of PBS. On day 0, anesthetized mice were infected with 10⁷ CFU S. aureus in 100 μl PBS by retro-orbital injection (69, 70). Mice with >20% body weight loss that showed signs of lethargy were euthanized during the experiment. While most mice were assessed until 3 dpi, any mice that lost >20% body weight were considered dead for assessing survival in order to avoid utilizing subjective criteria for moribund status. For disease assessment, numbers of S. aureus CFU in spleen, liver, and kidney homogenates of euthanized mice were determined by plating on tryptic soy agar plates for 12 h at 37°C. Some of the surviving EPS-treated mice were randomly selected (4 mice from 2 independent experiments) and were monitored for 13 days for long-term survival analysis.

Serum cytokine measurements.

Serum was collected from mice 1 dpi, and levels of proinflammatory cytokines and chemokines were measured using a 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 FlowJo software (Ashland, OR) by gating on individual beads and examining the geometric mean of detection reagent fluorescence intensity.

S. aureus uptake and growth with peritoneal phagocytes.

Three days after EPS injection, peritoneal cells were harvested by lavage with RPMI medium. F4/80⁺ macrophages 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-F4/80 (BM8) monoclonal antibodies. F4/80⁺ and F4/80⁻ cells (5 × 10⁵) or total peritoneal cells (10⁶) were infected with S. aureus and opsonized by incubation with 10% mouse serum for 30 min at 37°C, at multiplicity of infection of 1, for 30 min in antibiotic-free RPMI. S. aureus uptake was assessed by washing cells three times, lysing cells with 0.1% saponin, and plating on solid medium to quantify S. aureus CFU. S. aureus growth was assessed after 2 to 6 h in culture, followed by quantifying S. aureus CFU after lysis with 0.1% saponin and plating on solid medium. For ROS experiments, cells were washed of free bacterium and then further incubated in medium containing 1.25 μM CellROX green (Thermo Fischer Scientific, Waltham, MA) for 1 h in rotating tubes. Cells were then incubated with anti-CD16/32 (93), stained with anti-CD11b-allophycocyanin (APC) (M1/70) and anti-F4/80-APC-Cy7 (BM8) antibodies, and analyzed using an LSRFortessa (BD Biosciences). Flow cytometric data were analyzed using FlowJo software by gating first for nonlymphocytes and then for CD11b^high F4/80^high MΦ, followed by assessing median fluorescence intensity of CellROX green staining. Data were normalized to those for uninfected cells from PBS-treated mice for each experiment. For assessing the effect of ROS on bacterial growth, 2 μM diphenyleneiodonium (DPI; Sigma-Aldrich, St. Louis, MO) was included in the medium.

S. aureus SAg-mediated activation of splenocytes.

Splenocytes from PBS- or EPS-treated C57BL/6J mice were labeled with CellTrace violet (CTV; Life Technologies) according to the manufacturer’s descriptions, and 3.5 × 10⁵ cells were cultured in 96-well flat-bottom plates (Corning, Corning, NY). Cells were stimulated with cell-free supernatants of stationary-phase cultures of S. aureus AH1263 or JE2 transposon library mutants in RPMI medium at a final dilution of 1:3. After 4 days, nonadherent cells were stained with anti-CD4-APC (GK1.5), anti-CD8a-PE-Cy7 (53-6.7), anti-CD25-APC-Cy7 (PC61), and anti-CD44-PE (IM7) antibodies. Cells were analyzed using an LSRFortessa or FACSCanto II, and data were analyzed using FlowJo software by gating on CD4⁺ and CD8a⁺ lymphocytes. For coculture studies, peritoneal F4/80⁺ MΦ (2.5 × 10⁴) were obtained from PBS- or EPS-treated mice using the BD IMag cell separation system (BD Biosciences).

Flow cytometric analysis of M1 and M2 markers.

Peritoneal cells from PBS- or EPS-treated C57BL/6J mice were incubated with anti-CD16/32 (93) and stained with anti-CD11b-APC (M1/70) and anti-F4/80-APC-Cy7 (BM8) antibodies. Cells were then fixed and permeabilized (BD Cytofix/Cytoperm) for intracellular staining with anti-Arg-1-fluorescein isothiocyanate (R&D Systems) and anti-iNOS-phycoerythrin-eFluor610 (BD Pharmingen) antibodies. Cells were analyzed using LSRFortessa, and data were analyzed using FlowJo software by gating on CD11b^high F4/80^high MΦ.

Statistical analyses.

Statistical significance was determined by an unpaired, two-tailed Student's t test, unless otherwise noted, using GraphPad Prism software (La Jolla, CA).