Commensal bacteria contribute to immune homeostasis in the gastrointestinal tract; however, the underlying mechanisms for this are not well understood. A single dose of exopolysaccharide (EPS) from the probiotic spore-forming bacterium Bacillus subtilis protects mice from acute colitis induced by the enteric pathogen Citrobacter rodentium. Adoptive transfer of macrophage-rich peritoneal cells from EPS-treated mice confers protection from disease to recipient mice. In vivo, EPS induces development of anti-inflammatory M2 macrophages in a TLR4-dependent manner, and these cells inhibit T cell activation in vitro and in C. rodentium–infected mice. In vitro, M2 macrophages inhibit CD4+ and CD8+ T cells. The inhibition of CD4+ T cells is dependent on TGF-β, whereas inhibition of CD8+ T cells is dependent on TGF-β and PD-L1. We suggest that administration of B. subtilis EPS can be used to broadly inhibit T cell activation and, thus, control T cell–mediated immune responses in numerous inflammatory diseases.

Trillions of bacteria live in homeostasis within the gastrointestinal tract and provide a variety of benefits to the host, including digestion of food, synthesis of vitamins, and development and maintenance of the immune system. Disruption of the normal microbiota can contribute to a wide range of diseases, including inflammatory bowel disease (IBD), allergy, and diabetes (1). Although the benefits of probiotics and a healthy microbiota in disease prevention are well documented (2, 3), we have limited knowledge of the mechanisms by which bacteria exert these beneficial effects.

Several commensal bacteria were shown to limit colitis through induction and inhibition of immune responses (411). Although polysaccharide A (PSA) from Bacteroides fragilis is the best characterized (12, 13), a few other bacterial molecules, including carbohydrates (14, 15), proteins (16, 17), and sphingolipids (18), were identified as immune modulators. However, for most probiotics, the molecules that mediate protection are not known. We use the Gram-positive spore-forming probiotic, Bacillus subtilis, which protects mice from acute colitis induced by the enteric pathogen Citrobacter rodentium (19). Infection with this pathogen is characterized by diarrhea, colonic hyperplasia, mucosal infiltration of hematopoietic cells, and increases in chemokines and proinflammatory cytokines, similar to the pathology induced by enteropathogenic Escherichia coli in humans (2023). Protection by B. subtilis is not due to decreases in pathogen colonization or to increases in epithelial barrier integrity. Instead, it appears that B. subtilis prevents inflammation by modulating the innate immune response (14).

By screening several B. subtilis mutants, we found that protection from C. rodentium–induced inflammation requires the eps locus, which encodes molecules responsible for the synthesis of exopolysaccharide (EPS) (19, 24). We purified EPS from B. subtilis and showed that i.p. injection of this material protected mice from disease, indicating that EPS is the molecule responsible for protection (14). Although other probiotics and probiotic molecules require repeated treatments (6, 13, 2527), sometimes over the course of weeks, B. subtilis and purified EPS prevent disease after only a single dose (14, 19). Our goal is to elucidate the mechanism by which B. subtilis EPS protects from inflammation caused by C. rodentium infection. In this article, we identify the protective cells as M2 macrophages and show that they inhibit activation of CD4+ and CD8+ T cell responses and that the inhibition is mediated by TGF-β and PD-L1.

All animal experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee at Loyola University Medical Center (Maywood, IL). Specific pathogen–free C57BL/6 and TLR4−/− founders were purchased from The Jackson Laboratory and bred in-house. Mice lacking MyD88 in myeloid cells were generated by crossing Lyz2-Cre–transgenic mice to MyD88-floxed mice, as described (28). Sterile standard chow and tap water were given to mice ad libitum.

All base media and supplements were from Life Technologies (Grand Island, NY). All Abs were from BioLegend (San Diego, CA) unless otherwise indicated. The fluorescent Abs used for flow cytometry include anti-CD16/32 (Fc block), anti-F4/80 (BM8), anti-CD25 (PC61), anti-CD4 (GK1.5), anti-CD8 (53-6.7), anti-CD11b (M1/70), anti-CD44 (IM7), anti–IL-4Rα (I015F8), anti-CD206 (C068C2), and sheep anti-mouse/human arginase-1 (R&D Systems). ELISA reagents for quantitation of IFN-γ, TNF-α, and KC/CXCL1 were from R&D Systems; IL-17A, IL-2, IL-13, and TGF-β matched-pair reagents were from BioLegend. Clodronate and PBS liposomes were from VU Medisch Centrum (Amsterdam, the Netherlands). The anti-CD3ε (145-2C11) for in vivo experiments was from Leinco Technologies (St. Louis, MO). The anti-CD3ε used for in vitro experiments was LEAF-purified anti-CD3ε (145-2C11; BioLegend).

The following neutralizing Abs were used for in vitro assays: TGF-β inhibition, 5 μg/ml anti-TGF-β (1D11; R&D Systems) or equivalent concentration of LEAF-purified mouse IgG1 isotype control; PD-L2 inhibition, 3 μg/ml anti–PD-L2 (TY25) or isotype-control LEAF-purified rat IgG2a; and PD-L1 inhibition, 5 μg/ml anti–PD-L1 (10F.9G2) or isotype-control LEAF-purified rat IgG2b. Experiments with blocking Ab included the addition of 1 μg/ml LEAF-purified anti-CD16/32 (FcγRII/III) to all wells. For small molecule inhibitors, the following concentrations were used: Nor-NOHA (12 μM), exogenous l-arginine (2 mM), and NS-398 (1 μM).

EPS was isolated from B. subtilis DS991 (ΔsinRtasA mutant), a strain that produces and secretes large amounts of EPS (24). The negative control, designated ΔEPS, DS5187 (sinRtasAepsH mutant), does not produce EPS and does not protect from C. rodentium–induced disease (14, 24). EPS was isolated from stationary-phase supernatants of bacteria grown in Luria–Bertani or MSgg medium isolated by 50% EtOH precipitation at −20°C. The precipitate was pelleted (15,000 × g, 4°C, 20 min) and resuspended in 0.1 M Tris (pH 8), and samples were treated with DNase (67 mg/ml) and RNase (330 mg/ml) at 37°C for 2 h, followed by proteinase K (40 mg/ml) digestion at 55°C for 3 h. EPS was further purified by gel filtration on an S1000 column. Dialyzed EPS was quantified by dry weight and phenol sulfuric acid assay.

B. subtilis wild-type (WT) 3610, DS76 (espH mutant) were germinated via exhaustion, as described previously (19). On the day of administration, B. subtilis spores were washed with ice-cold water, resuspended in 100 ml of PBS, and administered to mice via oral gavage. Cells were isolated 5 d postgavage for analysis.

For adoptive transfer, peritoneal cells were isolated by lavage (with RPMI 1640/50% FBS) from mice 3 d posttreatment with EPS (i.p.). Cells in the granulocyte and lymphocyte gates were FACS sorted based on forward scatter and side scatter and injected i.p. into mice. For macrophage-depletion studies, mice were injected i.p. with 200 μl of clodronate-loaded or PBS-loaded liposomes (stock 5 mg/ml). Mice were treated with EPS 4–6 h later, and peritoneal cells were isolated by PBS lavage 3 d later. By flow cytometry, <1% of the transferred cells were macrophages.

C. rodentium ATCC 51459 was cultured for 16 h in Luria–Bertani medium and washed once in PBS. An infectious dose (5 × 108 CFU) was resuspended in 100 μl of sterile PBS and administered to mice by oral gavage. Disease was assessed 11 d postinfection (dpi). Serum cytokine levels were assessed by ELISA, and distal colons were collected and processed for histological analysis as described (19). To assess diarrhea, feces were examined and scored from 1 to 4 (19): 1, no diarrhea (hard, dry pellets); 2, slightly soft stool (mild diarrhea); 3, very soft stool (moderate diarrhea); and 4, unformed stool (severe diarrhea).

For flow cytometry, cells were treated with anti-CD16/32 Fc Block and then stained with surface Abs. Cells were analyzed on a FACSCanto II or LSRFortessa flow cytometer; cell sorting was performed on a FACSAria cell sorter (BD Biosciences). Analyses were performed using FlowJo software (TreeStar, Ashland, OR) by first gating on single cells and then analyzing each population as described.

For in vitro treatment of peritoneal cells, total peritoneal cells were isolated by lavage with 5 ml of PBS and plated in 24-well plates for 2 h to allow macrophages to adhere. Nonadherent cells were aspirated and washed away. The remaining cells were 80–90% F4/80+CD11b+ macrophages. Cells were treated for 16 h with 1 μg/ml EPS purified from DS991 B. subtilis grown in MSgg medium. For intracellular cytokine analysis, cells were cultured for an additional 2 h with brefeldin A and analyzed by flow cytometry or real-time quantitative PCR. For in vivo treatment, mice were injected i.p. with EPS (100 μg), and peritoneal cells were obtained by lavage 3 d later, as described (14). The level of TGF-β production in serum (1:100) was determined by ELISA after activating TGF-β with 5 μl of 1 N HCl/100 μl for 15 min at room temperature, followed by neutralization with 1 N NaOH.

Splenocytes were labeled with 5 μM CellTrace Violet (Life Technologies), according to the manufacturer’s directions, and were cultured alone (3 × 105 cells) or with total peritoneal cells (104) or purified macrophages (5 × 103) in 96-well flat-bottom tissue culture plates coated with 2 μg/ml anti-CD3. Three days later, nonadherent cells were collected, stained, and analyzed by flow cytometry. For Transwell experiments, 1 × 106 splenocytes were cultured in anti-CD3–coated 24-well plates, with or without Transwell inserts (Corning), containing 5 × 105 peritoneal cells.

Mesenteric lymph nodes (MLNs) were isolated from mice 7 dpi with C. rodentium. Cells (3 × 105) were stimulated with PMA (50 ng/ml) and ionomycin (1 μg/ml) for 4 h. Cytokine production was assessed by ELISA.

RNA was isolated from flow cytometry–sorted F4/80+CD11b+ macrophages using TRIzol Reagent (Invitrogen, Carlsbad, CA), and cDNA was prepared. PCR was performed on a C1000 thermal cycler with a CFX96 Real-Time Detection System (Bio-Rad, Hercules, CA) using the following primers (forward and reverse): Arg1, 5′-AGACCACAGTCTGGCAGTTG-3′ and 5′-CCACCCAAATGACACATAGG-3′; Nos2, 5′-CAGCTGGGCTGTACAAACCTT-3′ and 5′-CATTGGAAGTGAAGCGTTTCG-3′; Ym-1, 5′-CATGAGCAAGACTTGCGTGAC-3′ and 5′-GGTCCAAACTTCCATCCTCCA-3′; FIZZ-1, 5′-TCCCAGTGAATACTGATGAGA-3′ and 5′-CCACTCTGGATCTCCCAAGA-3′; IL-12p40, 5′-GAAGTTCAACATCAAGAGCAGTAG-3′ and 5′-AGGGAGAAGTAGGAATGGGG-3′; and Actb (β-actin), 5′-GGCTGTATTCCCCTCCATCG-3′ and 5′-CCAGTTGGTAACAATGCCATGT-3′. Expression of each target gene was normalized to β-actin expression, and data are presented relative to F4/80+CD11b+ cells isolated from untreated mice.

Statistical significance was determined by an unpaired two-tailed Student t test, unless otherwise indicated, using Prism software (GraphPad, La Jolla, CA). A p value < 0.05 was considered statistically significant.

Previous studies showed that transfer of total peritoneal cells from EPS-treated mice into naive mice protected recipients from development of C. rodentium–induced infectious colitis (14). To identify the cell type that mediates protection, we FACS-purified cells from the lymphocyte and granulocyte gates from the peritoneal cavity of EPS-treated mice. These cells (3 × 104) were adoptively transferred to recipient mice at −1, +1, and +3 dpi with C. rodentium, and disease was assessed 11 dpi. The cells in the granulocyte gate protected recipient mice from disease, whereas lymphocytes did not, as evidenced by increased colonic crypt heights, loose stool, and increased levels of serum proinflammatory chemokine CXCL1 (Fig. 1A, 1B). These data suggest that cells in the granulocyte gate, which are 90% macrophages, mediate protection by EPS. To test whether macrophages traffic from the peritoneal cavity to inhibit disease, we labeled peritoneal cells from EPS-treated mice with CFSE and transferred them into recipient mice by i.p. injection. By flow cytometry, we found CFSE+F4/80+ macrophages in MLNs and a few in the spleen, and some cells were still present in the peritoneum (Fig. 1C). These data suggest that macrophages traffic to the site of inflammation to prevent disease.

FIGURE 1.

Assessment of C. rodentium–induced disease following adoptive transfer of peritoneal granulocytes, lymphocytes, and macrophage-depleted cells from EPS-treated mice. (A) H&E-stained sections of the distal colon on day 11 post–C. rodentium infection. Cells (3 × 104) from EPS-treated mice were transferred on days −1, +1, and +3 relative to C. rodentium infection. Black lines indicate crypt height. Original magnification ×200. (B) Crypt height (left panel), diarrhea score (middle panel), and serum CXCL1 (right panel). Black bars represent PBS-injected mice, no cell transfer (negative control); white bars represent EPS-injected mice, no cell transfer (positive control); checkered bars represent cells transferred from EPS-injected mice, total unsorted (total), granulocytes (Gr), and lymphocytes (Lym). Data are from two independent experiments; n = 5–7 mice total per treatment. (C) Peritoneal cells were isolated from EPS-treated mice at 3 dpi, labeled with CFSE, and transferred into recipient mice. F4/80+CFSE+ cells in the peritoneum, spleen, and MLN were assessed by flow cytometry 24 h posttransfer. Data are representative of three independent experiments. (D and E) Mice were treated with clodronate liposomes (Clod-L) or PBS liposomes (PBS-L) 4–6 h before i.p. injection with EPS (100 μg); peritoneal cells (6 × 104) were transferred on days −1, +1, and +3 relative to C. rodentium infection. (D) H&E-stained sections of the distal colon. Black bar lines indicate crypt height. Original magnification ×200. (E) Crypt height (left panel), diarrhea score (middle panel), and serum CXCL1 (right panel) were measured 11 dpi. Data are from two independent experiments with n = 7–9 mice total per treatment. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05.

FIGURE 1.

Assessment of C. rodentium–induced disease following adoptive transfer of peritoneal granulocytes, lymphocytes, and macrophage-depleted cells from EPS-treated mice. (A) H&E-stained sections of the distal colon on day 11 post–C. rodentium infection. Cells (3 × 104) from EPS-treated mice were transferred on days −1, +1, and +3 relative to C. rodentium infection. Black lines indicate crypt height. Original magnification ×200. (B) Crypt height (left panel), diarrhea score (middle panel), and serum CXCL1 (right panel). Black bars represent PBS-injected mice, no cell transfer (negative control); white bars represent EPS-injected mice, no cell transfer (positive control); checkered bars represent cells transferred from EPS-injected mice, total unsorted (total), granulocytes (Gr), and lymphocytes (Lym). Data are from two independent experiments; n = 5–7 mice total per treatment. (C) Peritoneal cells were isolated from EPS-treated mice at 3 dpi, labeled with CFSE, and transferred into recipient mice. F4/80+CFSE+ cells in the peritoneum, spleen, and MLN were assessed by flow cytometry 24 h posttransfer. Data are representative of three independent experiments. (D and E) Mice were treated with clodronate liposomes (Clod-L) or PBS liposomes (PBS-L) 4–6 h before i.p. injection with EPS (100 μg); peritoneal cells (6 × 104) were transferred on days −1, +1, and +3 relative to C. rodentium infection. (D) H&E-stained sections of the distal colon. Black bar lines indicate crypt height. Original magnification ×200. (E) Crypt height (left panel), diarrhea score (middle panel), and serum CXCL1 (right panel) were measured 11 dpi. Data are from two independent experiments with n = 7–9 mice total per treatment. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05.

Close modal

To test whether macrophages are required for protection, we first injected mice with clodronate liposomes to deplete macrophages. Subsequently, mice were treated with EPS and 3 d later, we adoptively transferred peritoneal cells (6 × 104) to naive recipients at −1, +1, and +3 dpi with C. rodentium; disease was assessed 11 dpi. All mice that received macrophage-depleted peritoneal cells from EPS-injected mice had evidence of disease, whereas only one of seven that received peritoneal cells from EPS-injected mice treated with PBS liposomes developed disease (Fig. 1D, 1E). These data indicate that macrophages are required for EPS-mediated protection from C. rodentium–induced inflammation.

Macrophages polarize into proinflammatory M1 or anti-inflammatory M2 macrophages. We tested whether EPS induces M2 macrophages by culturing peritoneal F4/80+ cells overnight with EPS (1 μg/ml). By flow cytometry, we found upregulation of the M2 markers arginase-1 (Arg-1), CD206, IL-4Rα, and PD-L1 in EPS-treated cells (Fig. 2A). These data indicate that EPS induces M2 macrophages in vitro. These cells also expressed increased levels of IL-13 and IL-4, cytokines required for induction of M2 macrophages (Fig. 2A). We do not find increased expression of IL-13 or IL-4 in CD3+ T cells or IgM+ B cells cultured with EPS (data not shown).

FIGURE 2.

EPS-induced changes in phenotype of peritoneal macrophages in vitro and in vivo. (A) Peritoneal macrophages were incubated for 16 h with EPS in vitro and analyzed by flow cytometry. M2 marker expression (Arg-1, CD206, IL-4Rα, and PD-L1) and M2 cytokine expression (IL-13 and IL-4) by F4/80+CD11b+ peritoneal macrophages. Data are representative of three independent experiments. (BD) WT and TLR4−/− mice were injected with EPS, and peritoneal cells were examined by flow cytometry and real-time quantitative PCR 3 d later. (B) Representative flow cytometric profiles of M2 macrophage marker expression (Arg-1, CD206, IL-4Rα, and PD-L1) on F4/80+CD11b+ cells from WT mice (upper panels) or TLR4−/− mice (lower panels). (C) Fold change in mean fluorescence intensity (MFI) of Arg-1 and IL-4Rα on F4/80+CD11b+ cells from WT and TLR4−/− mice (compared with mice injected with ΔEPS). Average of four independent experiments; n = 10 mice total per treatment. (D) Fold change by real-time quantitative PCR of Ym-1, FIZZ-1, IL-12p40, and Nos2 expression in F4/80+CD11b+ cells in WT and TLR4−/− mice injected with EPS (compared with mice injected with ΔEPS) (average of four independent experiments, n = 7 mice total per group).*p ≤ 0.05. ND, not detectable.

FIGURE 2.

EPS-induced changes in phenotype of peritoneal macrophages in vitro and in vivo. (A) Peritoneal macrophages were incubated for 16 h with EPS in vitro and analyzed by flow cytometry. M2 marker expression (Arg-1, CD206, IL-4Rα, and PD-L1) and M2 cytokine expression (IL-13 and IL-4) by F4/80+CD11b+ peritoneal macrophages. Data are representative of three independent experiments. (BD) WT and TLR4−/− mice were injected with EPS, and peritoneal cells were examined by flow cytometry and real-time quantitative PCR 3 d later. (B) Representative flow cytometric profiles of M2 macrophage marker expression (Arg-1, CD206, IL-4Rα, and PD-L1) on F4/80+CD11b+ cells from WT mice (upper panels) or TLR4−/− mice (lower panels). (C) Fold change in mean fluorescence intensity (MFI) of Arg-1 and IL-4Rα on F4/80+CD11b+ cells from WT and TLR4−/− mice (compared with mice injected with ΔEPS). Average of four independent experiments; n = 10 mice total per treatment. (D) Fold change by real-time quantitative PCR of Ym-1, FIZZ-1, IL-12p40, and Nos2 expression in F4/80+CD11b+ cells in WT and TLR4−/− mice injected with EPS (compared with mice injected with ΔEPS) (average of four independent experiments, n = 7 mice total per group).*p ≤ 0.05. ND, not detectable.

Close modal

To determine whether EPS also induces an M2 macrophage phenotype in vivo, we examined peritoneal cells from mice 3 d after i.p. injection with EPS. We found increased M2 macrophage marker expression on F4/80+CD11b+ cells but not with treatment with the negative control, ΔEPS (Fig. 2B, 2C). Additionally, we FACS-purified F4/80+CD11b+ macrophages from EPS-injected mice and found that the M2 macrophage transcripts Ym-1 and FIZZ-1 were upregulated 5- and 9-fold, respectively, compared with macrophages from ΔEPS-treated mice. No upregulation of the M1 macrophage markers IL-12p40 or Nos2 (inducible NO synthase) was found (Fig. 2D). We conclude that in vivo administration of EPS induces cells with an M2 macrophage phenotype.

EPS protection from C. rodentium–induced disease requires TLR4, and peritoneal cells from TLR4−/− mice were unable to prevent disease following transfer to naive WT mice (14). These findings suggest that EPS induction of M2 macrophages requires TLR4 signaling. After injecting TLR4−/− mice with EPS, we found no change in the expression of the M2 markers Arg-1, CD206, IL-4Rα, or PD-L1 (Fig. 2B, 2C). Similarly, little to no upregulation of Ym-1 and FIZZ-1 transcripts occurs in macrophages from EPS-injected TLR4−/− mice (Fig. 2D). As expected, EPS also does not alter expression of the M1 macrophage markers IL-12p40 and Nos2 (inducible NO synthase) (Fig. 2D). These data show that the induction of M2 macrophages by EPS requires TLR4 signaling. Consistent with this finding, no M2 macrophages were found in the peritoneal cavity of EPS-injected myeloid-specific MyD88−/− mice (data not shown), which are not protected by EPS from C. rodentium–induced disease (14). These data suggest that polarization of M2 macrophages by EPS requires TLR4 and MyD88 in myeloid cells.

Much of the pathology associated with C. rodentium is driven by hyperinflammatory CD4+ T cell responses (20, 21). We hypothesized that EPS-induced M2 macrophages prevent disease by inhibiting T cell activation. To test this possibility, we FACS-purified peritoneal macrophages (F4/80+CD11b+) from EPS-treated mice or ΔEPS-treated mice and tested whether they inhibited T cell proliferation. Peritoneal macrophages from EPS-treated (or ΔEPS-treated) mice were cocultured with anti-CD3–stimulated splenocytes labeled with CellTrace Violet. As determined by flow cytometry, macrophages from EPS-treated mice inhibited the proliferation of CD4+ T cells (Fig. 3A), as well as CD8+ T cells (Fig. 3B), indicating that EPS-induced M2 macrophages broadly inhibit T cell responses. In addition, these macrophages decreased T cell activation, as measured by expression of CD25 and CD44 (Fig. 3A, 3B, right panels). As a control, macrophages from ΔEPS-treated mice did not inhibit T cell activation or proliferation (Fig. 3A, 3B). The inhibition of T cell responses by F4/80+CD11b+ cells was specific, because F4/80 peritoneal cells (CD3+ T cells and IgM+ B cells) did not inhibit T cell responses (Fig. 3C). As expected, peritoneal cells from EPS-treated TLR4−/− or myeloid MyD88−/− mice, which do not generate M2 macrophages, also did not inhibit T cell proliferation (Fig. 3D, 3E). We conclude that EPS-induced anti-inflammatory M2 macrophages inhibit T cell activation and proliferation in vitro. Similar to EPS, administration of B. subtilis spores (by oral gavage) resulted in the generation of peritoneal cells that inhibited T cell responses, whereas peritoneal cells from mice gavaged with EPS-deficient epsH B. subtilis spores did not inhibit T cell proliferation (Fig. 3F). We conclude that B. subtilis and EPS induce systemic anti-inflammatory responses.

FIGURE 3.

Inhibition of anti-CD3–stimulated T cell proliferation by peritoneal cells from EPS and B. subtilis–treated mice. CTV-labeled spleen cells were stimulated with anti-CD3 and cultured with F4/80+CD11b+ cells from EPS- or ΔEPS-treated mice. (A) Proliferation of CD4+ T cells in three independent experiments (left panel). Horizontal line represents the percentage of proliferating cells in three independent experiments. Percentage of proliferating CD4+ T cells in three independent experiments (middle panel). Percentage of activated CD4+ T cells (CD44+CD25+) in three independent experiments. n = 7 mice total per group. (B) Proliferation of CD8+ T cells in three independent experiments (left panel). Horizontal line represents the percentage of proliferating cells. Percentage of proliferating CD8+ T cells in three independent experiments (middle panel). Percentage of activated CD8+ T cells in three independent experiments (CD44+CD25+) (right panel). n = 7 mice total per group. (C) Proliferation of CD4+ T cells cocultured with peritoneal CD3+ T cells (left panel) or IgM+ B cells (right panel) from EPS- or ΔEPS-treated mice. Horizontal lines represent the percentage proliferation. (D) Proliferation of CD4+ T cells cultured with peritoneal cells from EPS-treated TLR4−/− or EPS-treated WT mice. Horizontal line represents the percentage of proliferating cells. (E) Percentage of proliferating CD4+ T cells in cocultures with peritoneal cells from TLR4−/− mice (left panel) and myeloid (mye) MyD88−/− mice (right panel) in four independent experiments. None represents splenocytes alone. n = 9 mice total per group. (F) Proliferation of CD4+ T cells cultured with peritoneal cells from mice gavaged with WT or epsH B. subtilis spores (left panel). Horizontal line represents the percentage of proliferating cells. Percentage of proliferating CD4+ T cells in three independent experiments (right panel). None represents splenocytes alone. n = 5 mice total per group. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. ns, not significant (p > 0.05).

FIGURE 3.

Inhibition of anti-CD3–stimulated T cell proliferation by peritoneal cells from EPS and B. subtilis–treated mice. CTV-labeled spleen cells were stimulated with anti-CD3 and cultured with F4/80+CD11b+ cells from EPS- or ΔEPS-treated mice. (A) Proliferation of CD4+ T cells in three independent experiments (left panel). Horizontal line represents the percentage of proliferating cells in three independent experiments. Percentage of proliferating CD4+ T cells in three independent experiments (middle panel). Percentage of activated CD4+ T cells (CD44+CD25+) in three independent experiments. n = 7 mice total per group. (B) Proliferation of CD8+ T cells in three independent experiments (left panel). Horizontal line represents the percentage of proliferating cells. Percentage of proliferating CD8+ T cells in three independent experiments (middle panel). Percentage of activated CD8+ T cells in three independent experiments (CD44+CD25+) (right panel). n = 7 mice total per group. (C) Proliferation of CD4+ T cells cocultured with peritoneal CD3+ T cells (left panel) or IgM+ B cells (right panel) from EPS- or ΔEPS-treated mice. Horizontal lines represent the percentage proliferation. (D) Proliferation of CD4+ T cells cultured with peritoneal cells from EPS-treated TLR4−/− or EPS-treated WT mice. Horizontal line represents the percentage of proliferating cells. (E) Percentage of proliferating CD4+ T cells in cocultures with peritoneal cells from TLR4−/− mice (left panel) and myeloid (mye) MyD88−/− mice (right panel) in four independent experiments. None represents splenocytes alone. n = 9 mice total per group. (F) Proliferation of CD4+ T cells cultured with peritoneal cells from mice gavaged with WT or epsH B. subtilis spores (left panel). Horizontal line represents the percentage of proliferating cells. Percentage of proliferating CD4+ T cells in three independent experiments (right panel). None represents splenocytes alone. n = 5 mice total per group. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. ns, not significant (p > 0.05).

Close modal

To determine whether EPS reduces T cell responses during C. rodentium infection, we assessed IL-2 production in the MLNs of mice pretreated with EPS. Seven dpi, we found that MLN cells from infected mice treated with EPS had reduced IL-2 production compared with mice infected with C. rodentium alone (Fig. 4A, left panel). To directly test whether EPS-induced M2 macrophages alter T cell responses during C. rodentium infection, we purified F4/80+ peritoneal macrophages from EPS-treated mice and transferred them into recipient mice infected with C. rodentium. Seven dpi, we assessed T cell cytokine production by MLN cells. Infected mice that received F4/80+ macrophages had reduced IL-2 production compared with mice that received F4/80 peritoneal cells from EPS-treated mice (Fig. 4A, right panel). These data suggest that EPS-induced M2 macrophages prevent C. rodentium–induced inflammation by inhibiting T cells.

FIGURE 4.

Inhibition of T cell responses in EPS-treated mice. (A) Quantification of IL-2 by ELISA of PMA- and ionomycin-stimulated cells from the MLN of C. rodentium–infected mice that were left untreated or were treated with EPS (left panel). Quantification by of IL-2 by ELISA of PMA- and ionomycin-stimulated MLN cells from C. rodentium–infected mice that received adoptive transfer of F4/80+ or F4/80 peritoneal cells from EPS-treated mice (right panel). Data are representative of two independent experiments with n = 7–10 mice total per group. (B) Quantification by ELISA of serum TNF-α, IFN-γ, and IL-2 in EPS-treated or untreated (NT) mice 2 h after i.p. injection of anti-CD3. Average of three independent experiments with n = 8 mice total per group. (C) Quantification by ELISA of IFN-γ, IL-17A, and IL-13 in culture supernatants of anti-CD3–stimulated splenocytes from EPS- or ΔEPS-treated mice. Average of four independent experiments with n = 8 mice total per group. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05.

FIGURE 4.

Inhibition of T cell responses in EPS-treated mice. (A) Quantification of IL-2 by ELISA of PMA- and ionomycin-stimulated cells from the MLN of C. rodentium–infected mice that were left untreated or were treated with EPS (left panel). Quantification by of IL-2 by ELISA of PMA- and ionomycin-stimulated MLN cells from C. rodentium–infected mice that received adoptive transfer of F4/80+ or F4/80 peritoneal cells from EPS-treated mice (right panel). Data are representative of two independent experiments with n = 7–10 mice total per group. (B) Quantification by ELISA of serum TNF-α, IFN-γ, and IL-2 in EPS-treated or untreated (NT) mice 2 h after i.p. injection of anti-CD3. Average of three independent experiments with n = 8 mice total per group. (C) Quantification by ELISA of IFN-γ, IL-17A, and IL-13 in culture supernatants of anti-CD3–stimulated splenocytes from EPS- or ΔEPS-treated mice. Average of four independent experiments with n = 8 mice total per group. ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05.

Close modal

To further test whether EPS inhibits T cell responses in vivo, we induced M2 macrophages by administering EPS and injected mice 3 d later with anti-CD3 (0.25 mg/kg body weight) to induce potent T cell activation. We expected decreased production of TNF-α, IFN-γ, and IL-2 in EPS-treated mice compared with untreated mice. Indeed, we found a 2-fold decrease in the production of these cytokines in serum of EPS-injected mice compared with untreated mice (Fig. 4A). We tested whether EPS inhibited a specific T cell subset by measuring the levels of Th1, Th17, and Th2 inflammatory cytokines, IFN-γ, IL-17A, and IL-13, respectively, in cultures of splenocytes from EPS- or ΔEPS-treated mice stimulated with anti-CD3. The secretion of each of the cytokines was decreased in EPS-treated mice compared with ΔEPS-treated mice (Fig. 4B), indicating that EPS induces a systemic anti-inflammatory response that broadly suppresses Th1, Th17, and Th2 cytokine production.

M2 macrophages inhibit T cell responses by multiple factors, including Arg-1, PGE2, IL-10, PD-L1/2, and TGF-β (2931). When separated from T cells in a Transwell, the M2 macrophages did not inhibit T cell proliferation, indicating that cell-to-cell contact is required for inhibition (Fig. 5A, 5B). To identify the molecule(s) produced by EPS-induced M2 macrophages that prevents T cell activation, we added inhibitors to the T cell–inhibition assay. Although only 14% of CD4+ T cells proliferated in cultures with peritoneal cells from EPS-treated mice, proliferation was greatly increased by the addition of anti–TGF-β (69% CD4+ T cells) (Fig. 5C), indicating that EPS-induced M2 macrophages produce TGF-β, which inhibits CD4+ T cell proliferation in vitro. In contrast, inhibition of Arg-1 activity by Nor-NOHA, or addition of exogenous l-arginine, did not restore T cell proliferation (Fig. 5D), even though we find increased expression of Arg-1 in EPS-induced M2 macrophages (Fig. 2A, 2B). Further, the addition of NS-398, a COX2 inhibitor that prevents PGE2 production, or addition of neutralizing Abs to PD-L2, PD-L1, and IL-10, did not restore CD4+ T cell proliferation (Fig. 5D), suggesting that TGF-β is responsible for the inhibitory effect of M2 macrophages on CD4+ T cells. Consistent with this conclusion, we found a small, but significant, increase in levels of total TGF-β in serum from EPS-treated mice 3 d after treatment (Fig. 5E).

FIGURE 5.

Restoration of T cell proliferation by inhibitors of M2 macrophage function. Proliferation of CD4+ (A) or CD8+ (B) T cells cocultured with peritoneal cells from untreated (NT)- or EPS-treated mice in direct contact (No Transwell, left panel) or cocultured with peritoneal cells in a Transwell insert (right panel). Horizontal bar represents the percentage of proliferation. Data are representative of three independent experiments. (C) Proliferation of CD4+ T cells of anti-CD3–stimulated splenocytes cultured with peritoneal cells from EPS-treated mice in the presence of neutralizing anti-TGF-β Ab or mouse (Ms) IgG1 isotype control. Horizontal line represents the percentage of proliferation. Data are representative of three independent experiments. (D) Percentage of proliferation of CD4+ T cells cultured with peritoneal cells from EPS-treated mice and containing inhibitors of M2 macrophage function. Data are from three independent experiments each (n = 6 mice total per group). (E) Quantification by ELISA of total TGF-β in serum 3 d after EPS treatment or no treatment (NT). n = 4–5 mice total per group. Data were analyzed by Student t test. *p ≤ 0.05. (F) Percentage proliferation of CD8+ T cells cultured with peritoneal cells from EPS-treated mice containing inhibitors of M2 macrophage function. (D and F) Black bars represent NT peritoneal cells alone (negative control); white bars represent peritoneal cells from EPS-treated mice (positive control); and checkered bars represent cultures with inhibitors of M2 macrophages, anti-TGF-β, Nor-NOHA, l-arginine, NS-398, anti–PD-L2, anti–PD-L1, and anti–IL-10. Data were analyzed by ANOVA in combination with a Bonferroni test for multiple comparisons. ***p ≤ 0.001, *p ≤ 0.05.

FIGURE 5.

Restoration of T cell proliferation by inhibitors of M2 macrophage function. Proliferation of CD4+ (A) or CD8+ (B) T cells cocultured with peritoneal cells from untreated (NT)- or EPS-treated mice in direct contact (No Transwell, left panel) or cocultured with peritoneal cells in a Transwell insert (right panel). Horizontal bar represents the percentage of proliferation. Data are representative of three independent experiments. (C) Proliferation of CD4+ T cells of anti-CD3–stimulated splenocytes cultured with peritoneal cells from EPS-treated mice in the presence of neutralizing anti-TGF-β Ab or mouse (Ms) IgG1 isotype control. Horizontal line represents the percentage of proliferation. Data are representative of three independent experiments. (D) Percentage of proliferation of CD4+ T cells cultured with peritoneal cells from EPS-treated mice and containing inhibitors of M2 macrophage function. Data are from three independent experiments each (n = 6 mice total per group). (E) Quantification by ELISA of total TGF-β in serum 3 d after EPS treatment or no treatment (NT). n = 4–5 mice total per group. Data were analyzed by Student t test. *p ≤ 0.05. (F) Percentage proliferation of CD8+ T cells cultured with peritoneal cells from EPS-treated mice containing inhibitors of M2 macrophage function. (D and F) Black bars represent NT peritoneal cells alone (negative control); white bars represent peritoneal cells from EPS-treated mice (positive control); and checkered bars represent cultures with inhibitors of M2 macrophages, anti-TGF-β, Nor-NOHA, l-arginine, NS-398, anti–PD-L2, anti–PD-L1, and anti–IL-10. Data were analyzed by ANOVA in combination with a Bonferroni test for multiple comparisons. ***p ≤ 0.001, *p ≤ 0.05.

Close modal

Anti–TGF-β also partially restored CD8+ T cell proliferation, as did neutralizing anti-PD-L1 Ab (Fig. 5F). The addition of both Abs resulted in complete restoration of CD8+ T cell proliferation (Fig. 5F), indicating that TGF-β and PD-L1 contribute to the inhibition of CD8+ T cells. Neutralizing anti–IL-10 did not restore proliferation of CD8+ T cells. We conclude that EPS-induced M2 macrophages inhibit T cells through TGF-β and PD-L1 in a cell contact–dependent manner (Table I).

Table I.
Commensal bacteria with beneficial immunomodulatory effects on inflammatory diseases
BacteriaElicited Immune ResponseBacterial Molecule Responsible for ProtectionInflammatory DiseaseRefs.
B. subtilis TLR4-dependent induction of M2 macrophages EPS C. rodentium–induced colitis (14, 19
B. fragilis TLR2-dependent induction of iTregs by IL-10+ DCs PSA Helicobacter hepaticus–induced colitis, TNBS-induced colitis, EAE (5, 12, 13, 47
Bifidobacterium spp. Induction of Tr1 cells and Tregs; reduction in B. breve–specific Ab production Possibly EPS C. rodentium–induced colitis (6, 15
B. fragilis Negative regulation of invariant NKT cells in neonatal mice Sphingolipid GSL-Bf717 Oxazolone-induced colitis (18
F. prausnitzii Inhibition of NF-κB in intestinal epithelial cells Anti-inflammatory protein MAM DNBS-induced colitis (16, 17
C. butyricum TLR2-dependent induction of IL-10+ F4/80+CD11b+CD11cint cells ND DSS-induced colitis (9
Clostridia spp. Induction of IL-10+ Tregs; upregulates IL-22 by ILCs ND DSS-induced colitis, food allergy (7, 8, 46
Lactobacillus spp. Reduction of proinflammatory mucosal cytokines ND H. hepaticus–induced IBD (25
Segmented filamentous bacteria Induction of Th17 cells; IgA production ND C. rodentium–induced colitis, T1D (11, 49
VSL#3 Induction of TNF-α production by epithelial cells; IL-10+ TGF-β+ T cells ND Chronic CD-like ileitis (SAMP mice), TNBS-induced colitis (4, 44
BacteriaElicited Immune ResponseBacterial Molecule Responsible for ProtectionInflammatory DiseaseRefs.
B. subtilis TLR4-dependent induction of M2 macrophages EPS C. rodentium–induced colitis (14, 19
B. fragilis TLR2-dependent induction of iTregs by IL-10+ DCs PSA Helicobacter hepaticus–induced colitis, TNBS-induced colitis, EAE (5, 12, 13, 47
Bifidobacterium spp. Induction of Tr1 cells and Tregs; reduction in B. breve–specific Ab production Possibly EPS C. rodentium–induced colitis (6, 15
B. fragilis Negative regulation of invariant NKT cells in neonatal mice Sphingolipid GSL-Bf717 Oxazolone-induced colitis (18
F. prausnitzii Inhibition of NF-κB in intestinal epithelial cells Anti-inflammatory protein MAM DNBS-induced colitis (16, 17
C. butyricum TLR2-dependent induction of IL-10+ F4/80+CD11b+CD11cint cells ND DSS-induced colitis (9
Clostridia spp. Induction of IL-10+ Tregs; upregulates IL-22 by ILCs ND DSS-induced colitis, food allergy (7, 8, 46
Lactobacillus spp. Reduction of proinflammatory mucosal cytokines ND H. hepaticus–induced IBD (25
Segmented filamentous bacteria Induction of Th17 cells; IgA production ND C. rodentium–induced colitis, T1D (11, 49
VSL#3 Induction of TNF-α production by epithelial cells; IL-10+ TGF-β+ T cells ND Chronic CD-like ileitis (SAMP mice), TNBS-induced colitis (4, 44

CD, Crohn’s disease; DCs, dendritic cells; DNBS, dinitrobenzene sulfonic acid; ILC, innate lymphoid cell; MAM, microbial anti-inflammatory molecule; T1D, type 1 diabetes; TNBS, 2,4,5-trinitrobenzene sulfonic acid.

Naive CD4+ T cells polarize into effector T cell subsets according to environmental conditions. In the presence of TGF-β, CD4+ T cells upregulate expression of Foxp3 and become inducible regulatory T cells (iTregs) (32). Because M2 macrophages produce TGF-β, we hypothesized that these could also induce Tregs. We examined Foxp3 expression in peritoneal CD4+CD25+ T cells 3 d post-EPS treatment and found an increase in Foxp3+ cells in EPS-treated mice compared with untreated mice (Fig. 6A, 6B). Further, peritoneal cells from EPS-treated mice induced more CD4+CD25+Foxp3+ T cells in IL-2–containing cocultures of peritoneal cells with anti-CD3–stimulated splenocytes compared with cocultures from untreated mice (Fig. 6C, 6D). Together, these data suggest that EPS has the capacity to induce Tregs, presumably through TGF-β produced by M2 macrophages (Fig. 7).

FIGURE 6.

Induction of Tregs by EPS. (A) Representative example of Foxp3 expression in CD4+CD25+ T cells in the peritoneal cavity of untreated (NT; upper panels) or EPS-treated (EPS; lower panels) mice. (B) Percentage of CD4+CD25+Foxp3+ cells in three independent experiments with n = 6 mice total per group. (C) CD4+CD25+Foxp3+ T cells in cocultures of anti-CD3–stimulated splenocytes, IL-2 (50 ng/ml), and peritoneal cells from EPS-treated or untreated (NT) mice. (D) Percentage of CD4+CD25+Foxp3+ cells in three independent experiments with n = 4 mice total per group. **p ≤ 0.01.

FIGURE 6.

Induction of Tregs by EPS. (A) Representative example of Foxp3 expression in CD4+CD25+ T cells in the peritoneal cavity of untreated (NT; upper panels) or EPS-treated (EPS; lower panels) mice. (B) Percentage of CD4+CD25+Foxp3+ cells in three independent experiments with n = 6 mice total per group. (C) CD4+CD25+Foxp3+ T cells in cocultures of anti-CD3–stimulated splenocytes, IL-2 (50 ng/ml), and peritoneal cells from EPS-treated or untreated (NT) mice. (D) Percentage of CD4+CD25+Foxp3+ cells in three independent experiments with n = 4 mice total per group. **p ≤ 0.01.

Close modal
FIGURE 7.

Model for EPS modulation of immune responses. B. subtilis and purified EPS induce M2 macrophages that inhibit CD4+ and CD8+ T cells through production of TGF-β and PD-L1 and potentially through induction of Tregs.

FIGURE 7.

Model for EPS modulation of immune responses. B. subtilis and purified EPS induce M2 macrophages that inhibit CD4+ and CD8+ T cells through production of TGF-β and PD-L1 and potentially through induction of Tregs.

Close modal

Until recently, studies of bacterial products that modulate the immune system have focused primarily on pathogenic molecules that elicit proinflammatory responses or contribute to evasion of the immune system. These studies identified host pattern recognition receptors, such as TLRs and NODs, and their cognate ligands (e.g., LPS, lipoteichoic acid, and flagellin), along with the downstream signaling pathways. Less understood are the mechanisms by which commensal bacteria and commensal-derived products circumvent proinflammatory responses, allowing select bacteria to live in homeostasis with the host. We showed previously that a single oral dose of the probiotic B. subtilis or a single i.p. injection of B. subtilis EPS prevents inflammatory responses against the enteric pathogen C. rodentium. In this study, we showed that B. subtilis EPS induces M2 macrophages that broadly inhibit in vitro and in vivo T cell responses through TGF-β and PD-L1.

During C. rodentium infection, mice display increased expression of the Th17 cytokine, IL-17A, in Peyer’s patches (23) and Th1 cytokines, IL-12, IFN-γ, and TNF-α, in the colon (20, 21). These cytokines lead to increased production of chemokines that recruit innate immune cells to the site of infection, exacerbating disease. EPS-induced M2 macrophages efficiently block this complex immune response to C. rodentium infection and also reduce CD4+ and CD8+ T cell responses after mice are injected with anti-CD3. Increases in M2 macrophages are found in the peritoneum and spleen after i.p. or i.v. injection of EPS, as well as after oral administration of B. subtilis spores (data not shown). We also find donor M2 macrophages in the MLN in mice receiving adoptive transfer of peritoneal cells from EPS-treated mice, further indicating that injection of EPS induces M2 macrophages that can migrate throughout the body. Because EPS and B. subtilis can elicit an anti-inflammatory response when administered through several routes, we think that EPS and/or B. subtilis are good candidates as therapeutic agents for treating inflammatory diseases.

EPS does not protect TLR4−/− mice from disease caused by C. rodentium, and we showed that M2 macrophages did not develop in these mice or in mice defective in MyD88 signaling in myeloid cells. The dependency on TLR4 is not likely due to LPS, because B. subtilis is Gram-positive and produces little to no LPS. In contrast to other bacterial polysaccharides, notably PSA from B. fragilis that signals through TLR2, EPS is unlikely to function through TLR2 because EPS induces M2 macrophages in TLR2−/− mice (data not shown). We hypothesize that EPS uses a coreceptor of the TLR4 signaling complex, similar to LPS, which binds to the coreceptors CD14 and MD2 (33). In addition, EPS could bind C-type lectins or scavenger receptors, which serve as carbohydrate-binding pattern recognition receptors and can associate with TLR4 (3437). EPS binding to these receptors could upregulate M2 macrophage–inducing transcription factors, such as STAT6, IRF4, and peroxisome proliferator-activated receptor γ, which regulate transcription of Arg-1 and CD206 (38, 39). We hypothesize that EPS binding to TLR4 coreceptors directly modifies macrophages and induces M2 macrophage–specific transcription factors. The upregulation of M2 macrophage markers on EPS-treated primary macrophages (Fig. 2A) and on RAW264.7 macrophages (data not shown) will allow us to directly test this hypothesis.

M2 macrophages mediate an anti-inflammatory response by TGF-β and PD-L1, as well as by Arg-1, IL-10, and PD-L2 (30, 31, 40, 41). Such molecules protect from colitis, promote tissue repair and metabolic homeostasis, and provide protective immunity to helminth infections (42). Although EPS-induced M2 macrophages produce multiple anti-inflammatory molecules, in our in vitro cocultures of EPS-induced M2 macrophages and proliferating T cells, we find restoration of CD4+ T cell proliferation only by interfering with TGF-β. In contrast, CD8+ T cell restoration occurs through PD-L1, as well as TGF-β. CD8+ T cells are more sensitive to PD-1 ligation (43), which may explain why PD-L1 blockade increases proliferation of CD8+ T cells but not CD4+ T cells. Inhibition of other molecules, including Arg-1, PGE2, PD-L2, and IL-10, had no effect on T cell inhibition. We conclude that TGF-β and PD-L1 are the main inhibitory factors of the EPS-induced M2 macrophages. Although PD-L1 predominantly inhibits T cells, TGF-β is a pleiotropic cytokine that acts on virtually all cell types. In this study, we focus mainly on inhibition of T cells by EPS, but this does not exclude the potential to regulate a variety of cells and immune responses, given the multifaceted nature of TGF-β signaling. In certain circumstances, TGF-β promotes wound healing, which is another mechanism through which EPS may help to resolve C. rodentium–induced inflammation.

Several commensal bacteria other than B. subtilis prevent inflammatory diseases by modulating the immune response (Table I). The beneficial effects of commensals are described mostly for models of IBD, where, with a few exceptions (4, 11, 15), protection appears to be mediated, in large part, by TLR2 signaling, iTregs, and IL-10. For example, B. fragilis, Bifidobacterium infantis, and Clostridia spp. induce Tregs that ameliorate disease in chemically induced colitis models (8, 10, 12, 13), and VSL#3 and Bifidobacterium breve induce IL-10–producing Tr1 cells (6, 44). Although we find a slight increase in Tregs after administration of B. subtilis EPS, we do not know whether they contribute to protection. The probiotic Clostridium butyricum mediates protection from dextran sulfate sodium (DSS)-induced colitis in a Treg-independent manner through TLR2-dependent, IL-10–producing F4/80+CD11b+CD11cint macrophages (9). In contrast, B. subtilis EPS-induced M2 macrophages mediate protection through TLR4 signaling and do not produce detectable levels of IL-10 (data not shown). We conclude that B. subtilis EPS induces an anti-inflammatory response distinct from that of all previously described probiotics. With the exception of B. subtilis EPS, B. fragilis PSA and sphingolipids, and Faecalibacterium prausnitzii microbial anti-inflammatory molecule protein (5, 14, 1618), most of the bacterial molecule(s) responsible for protection have not been identified or purified. Undoubtedly, additional molecules from a large number of commensal bacteria will be found to regulate immune responses.

In addition to the probiotics indicated in Table I, several others are known to limit diseases, although the mechanisms of immune modulation are unknown (2527, 45). Although most known probiotics target gastrointestinal disease, many probiotics will likely be useful for treating or preventing other inflammatory diseases, including diabetes, allergy, and experimental autoimmune encephalomyelitis (EAE) (7, 4648). In fact, segmented filamentous bacteria–induced Th17 cells prevent the spontaneous development of type-1 diabetes in NOD mice (49). Clostridia spp. alter innate lymphoid cells and prevent the development of food allergy (7), and B. fragilis PSA prevents the development of EAE (47). As our understanding of the mechanisms by which specific commensal bacteria modulate the immune system increases, we will likely identify more diseases for which these probiotics will be beneficial.

This study highlights a unique mechanism by which a Gram-positive commensal EPS induces an anti-inflammatory environment, as modeled in Fig. 7. EPS from B. subtilis, and presumably other organisms, induces the generation of anti-inflammatory M2 macrophages in a TLR4-dependent manner. These M2 macrophages produce multiple immune-inhibitory molecules, including TGF-β, PD-L1, and Arg-1, all of which inhibit T cell responses and prevent inflammatory diseases. Understanding the mechanisms by which EPS and other molecules from commensal organisms regulate the immune system will lead to new rationally designed therapeutics for inflammatory diseases.

This work was supported by National Institutes of Health Grants R21AI098187, R01AI110586, R01AI050260, and F31DK104541.

Abbreviations used in this article:

Arg-1

arginase-1

dpi

day postinfection

DSS

dextran sulfate sodium

EAE

experimental autoimmune encephalomyelitis

EPS

exopolysaccharide

IBD

inflammatory bowel disease

iTreg

inducible regulatory T cell

MLN

mesenteric lymph node

PSA

polysaccharide A

WT

wild-type.

1
Hooper
L. V.
,
Littman
D. R.
,
Macpherson
A. J.
.
2012
.
Interactions between the microbiota and the immune system.
Science
336
:
1268
1273
.
2
Belkaid
Y.
,
Hand
T. W.
.
2014
.
Role of the microbiota in immunity and inflammation.
Cell
157
:
121
141
.
3
Palm
N. W.
,
de Zoete
M. R.
,
Flavell
R. A.
.
2015
.
Immune-microbiota interactions in health and disease.
Clin. Immunol.
159
:
122
127
.
4
Pagnini
C.
,
Saeed
R.
,
Bamias
G.
,
Arseneau
K. O.
,
Pizarro
T. T.
,
Cominelli
F.
.
2010
.
Probiotics promote gut health through stimulation of epithelial innate immunity.
Proc. Natl. Acad. Sci. USA
107
:
454
459
.
5
Mazmanian
S. K.
,
Round
J. L.
,
Kasper
D. L.
.
2008
.
A microbial symbiosis factor prevents intestinal inflammatory disease.
Nature
453
:
620
625
.
6
Jeon
S. G.
,
Kayama
H.
,
Ueda
Y.
,
Takahashi
T.
,
Asahara
T.
,
Tsuji
H.
,
Tsuji
N. M.
,
Kiyono
H.
,
Ma
J. S.
,
Kusu
T.
, et al
.
2012
.
Probiotic Bifidobacterium breve induces IL-10-producing Tr1 cells in the colon.
PLoS Pathog.
8
:
e1002714
.
7
Stefka
A. T.
,
Feehley
T.
,
Tripathi
P.
,
Qiu
J.
,
McCoy
K.
,
Mazmanian
S. K.
,
Tjota
M. Y.
,
Seo
G.-Y.
,
Cao
S.
,
Theriault
B. R.
, et al
.
2014
.
Commensal bacteria protect against food allergen sensitization.
Proc. Natl. Acad. Sci. USA
111
:
13145
13150
.
8
Atarashi
K.
,
Tanoue
T.
,
Shima
T.
,
Imaoka
A.
,
Kuwahara
T.
,
Momose
Y.
,
Cheng
G.
,
Yamasaki
S.
,
Saito
T.
,
Ohba
Y.
, et al
.
2011
.
Induction of colonic regulatory T cells by indigenous Clostridium species.
Science
331
:
337
341
.
9
Hayashi
A.
,
Sato
T.
,
Kamada
N.
,
Mikami
Y.
,
Matsuoka
K.
,
Hisamatsu
T.
,
Hibi
T.
,
Roers
A.
,
Yagita
H.
,
Ohteki
T.
, et al
.
2013
.
A single strain of Clostridium butyricum induces intestinal IL-10-producing macrophages to suppress acute experimental colitis in mice.
Cell Host Microbe
13
:
711
722
.
10
O’Mahony
C.
,
Scully
P.
,
O’Mahony
D.
,
Murphy
S.
,
O’Brien
F.
,
Lyons
A.
,
Sherlock
G.
,
MacSharry
J.
,
Kiely
B.
,
Shanahan
F.
,
O’Mahony
L.
.
2008
.
Commensal-induced regulatory T cells mediate protection against pathogen-stimulated NF-kappaB activation.
PLoS Pathog.
4
:
e1000112
.
11
Ivanov
I. I.
,
Atarashi
K.
,
Manel
N.
,
Brodie
E. L.
,
Shima
T.
,
Karaoz
U.
,
Wei
D.
,
Goldfarb
K. C.
,
Santee
C. A.
,
Lynch
S. V
, et al
.
2009
.
Induction of intestinal Th17 cells by segmented filamentous bacteria.
Cell
139
:
485
498
.
12
Round
J. L.
,
Mazmanian
S. K.
.
2010
.
Inducible Foxp3+ regulatory T-cell development by a commensal bacterium of the intestinal microbiota.
Proc. Natl. Acad. Sci. USA
107
:
12204
12209
.
13
Shen
Y.
,
Giardino Torchia
M. L.
,
Lawson
G. W.
,
Karp
C. L.
,
Ashwell
J. D.
,
Mazmanian
S. K.
.
2012
.
Outer membrane vesicles of a human commensal mediate immune regulation and disease protection.
Cell Host Microbe
12
:
509
520
.
14
Jones
S. E.
,
Paynich
M. L.
,
Kearns
D. B.
,
Knight
K. L.
.
2014
.
Protection from intestinal inflammation by bacterial exopolysaccharides.
J. Immunol.
192
:
4813
4820
.
15
Fanning
S.
,
Hall
L. J.
,
Cronin
M.
,
Zomer
A.
,
MacSharry
J.
,
Goulding
D.
,
Motherway
M. O.
,
Shanahan
F.
,
Nally
K.
,
Dougan
G.
,
van Sinderen
D.
.
2012
.
Bifidobacterial surface-exopolysaccharide facilitates commensal-host interaction through immune modulation and pathogen protection.
Proc. Natl. Acad. Sci. USA
109
:
2108
2113
.
16
Sokol
H.
,
Pigneur
B.
,
Watterlot
L.
,
Lakhdari
O.
,
Bermúdez-Humarán
L. G.
,
Gratadoux
J.-J.
,
Blugeon
S.
,
Bridonneau
C.
,
Furet
J.-P.
,
Corthier
G.
, et al
.
2008
.
Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients.
Proc. Natl. Acad. Sci. USA
105
:
16731
16736
.
17
Quévrain
E.
,
Maubert
M. A.
,
Michon
C.
,
Chain
F.
,
Marquant
R.
,
Tailhades
J.
,
Miquel
S.
,
Carlier
L.
,
Bermúdez-Humarán
L. G.
,
Pigneur
B.
, et al
.
2016
.
Identification of an anti-inflammatory protein from Faecalibacterium prausnitzii, a commensal bacterium deficient in Crohn’s disease.
Gut
65
:
415
425
.
18
An
D.
,
Oh
S. F.
,
Olszak
T.
,
Neves
J. F.
,
Avci
F. Y.
,
Erturk-Hasdemir
D.
,
Lu
X.
,
Zeissig
S.
,
Blumberg
R. S.
,
Kasper
D. L.
.
2014
.
Sphingolipids from a symbiotic microbe regulate homeostasis of host intestinal natural killer T cells.
Cell
156
:
123
133
.
19
Jones
S. E.
,
Knight
K. L.
.
2012
.
Bacillus subtilis–mediated protection from Citrobacter rodentium–associated enteric disease requires espH and functional flagella.
Infect. Immun.
80
:
710
719
.
20
Higgins
L. M.
,
Frankel
G.
,
Douce
G.
,
Dougan
G.
,
MacDonald
T. T.
.
1999
.
Citrobacter rodentium infection in mice elicits a mucosal Th1 cytokine response and lesions similar to those in murine inflammatory bowel disease.
Infect. Immun.
67
:
3031
3039
.
21
Chan
J. M.
,
Bhinder
G.
,
Sham
H. P.
,
Ryz
N.
,
Huang
T.
,
Bergstrom
K. S.
,
Vallance
B. A.
.
2013
.
CD4+ T cells drive goblet cell depletion during Citrobacter rodentium infection.
Infect. Immun.
81
:
4649
4658
.
22
Higgins
L. M.
,
Frankel
G.
,
Connerton
I.
,
Gonçalves
N. S.
,
Dougan
G.
,
MacDonald
T. T.
.
1999
.
Role of bacterial intimin in colonic hyperplasia and inflammation.
Science
285
:
588
591
.
23
Li
L.
,
Shi
Q. G.
,
Lin
F.
,
Liang
Y. G.
,
Sun
L. J.
,
Mu
J. S.
,
Wang
Y. G.
,
Su
H. B.
,
Xu
B.
,
Ji
C C.
, et al
.
2014
.
Cytokine IL-6 is required in Citrobacter rodentium infection-induced intestinal Th17 responses and promotes IL-22 expression in inflammatory bowel disease.
Mol. Med. Rep.
9
:
831
836
.
24
Guttenplan
S. B.
,
Blair
K. M.
,
Kearns
D. B.
.
2010
.
The EpsE flagellar clutch is bifunctional and synergizes with EPS biosynthesis to promote Bacillus subtilis biofilm formation.
PLoS Genet.
6
:
e1001243
.
25
Peña
J. A.
,
Rogers
A. B.
,
Ge
Z.
,
Ng
V.
,
Li
S. Y.
,
Fox
J. G.
,
Versalovic
J.
.
2005
.
Probiotic Lactobacillus spp. diminish Helicobacter hepaticus–induced inflammatory bowel disease in interleukin-10-deficient mice.
Infect. Immun.
73
:
912
920
.
26
Johnson-Henry
K. C.
,
Nadjafi
M.
,
Avitzur
Y.
,
Mitchell
D. J.
,
Ngan
B.
,
Galindo-Mata
E
,
Jones
N. L.
,
Sherman
P. M.
.
2005
.
Amelioration of the effects of Citrobacter rodentium infection in mice by pretreatment with probiotics.
J. Infect. Dis.
191
:
2106
2117
.
27
Peran
L.
,
Camuesco
D.
,
Comalada
M.
,
Bailon
E.
,
Henriksson
A.
,
Xaus
J.
,
Zarzuelo
A.
,
Galvez
J.
.
2007
.
A comparative study of the preventative effects exerted by three probiotics, Bifidobacterium lactis, Lactobacillus casei and Lactobacillus acidophilus, in the TNBS model of rat colitis.
J. Appl. Microbiol.
103
:
836
844
.
28
Gais
P.
,
Reim
D.
,
Jusek
G.
,
Tanja
R.-B.
,
Weighardt
H.
,
Pfeffer
K.
,
Altmayr
F.
,
Janssen
K.-P.
,
Holzmann
B.
.
2012
.
Cutting edge: divergent cell-specific functions of MyD88 for inflammatory responses and organ injury in septic peritonitis.
J. Immunol.
188
:
5833
5837
.
29
Taylor
M. D.
,
Harris
A.
,
Nair
M. G.
,
Maizels
R. M.
,
Allen
J. E.
.
2006
.
F4/80+ alternatively activated macrophages control CD4+ T cell hyporesponsiveness at sites peripheral to filarial infection.
J. Immunol.
176
:
6918
6927
.
30
Pesce
J. T.
,
Ramalingam
T. R.
,
Mentink-Kane
M. M.
,
Wilson
M. S.
,
Kasmi
K. C.
,
Smith
A. M.
,
Thompson
R. W.
,
Cheever
A. W.
,
Murray
P. J.
,
Wynn
T. A.
.
2009
.
Arginase-1-expressing macrophages suppress Th2 cytokine-driven inflammation and fibrosis.
PLoS Pathog.
5
:
e1000371
.
31
Huber
S.
,
Hoffmann
R.
,
Muskens
F.
,
Voehringer
D.
.
2010
.
Alternatively activated macrophages inhibit T-cell proliferation by Stat6-dependent expression of PD-L2.
Blood
116
:
3311
3320
.
32
Coombes
J. L.
,
Siddiqui
K. R.
,
Arancibia-Cárcamo
C. V.
,
Hall
J.
,
Sun
C. M.
,
Belkaid
Y.
,
Powrie
F.
.
2007
.
A functionally specialized population of mucosal CD103+ DCs induces Foxp3+ regulatory T cells via a TGF-beta and retinoic acid-dependent mechanism.
J. Exp. Med.
204
:
1757
1764
.
33
da Silva Correia
J.
,
Soldau
K.
,
Christen
U.
,
Tobias
P. S.
,
Ulevitch
R. J.
.
2001
.
Lipopolysaccharide is in close proximity to each of the proteins in its membrane receptor complex. transfer from CD14 to TLR4 and MD-2.
J. Biol. Chem.
276
:
21129
21135
.
34
Stewart
C. R.
,
Stuart
L. M.
,
Wilkinson
K.
,
van Gils
J. M.
,
Deng
J.
,
Halle
A.
,
Rayner
K. J.
,
Boyer
L.
,
Zhong
R.
,
Frazier
W. A.
, et al
.
2009
.
CD36 ligands promote sterile inflammation through assembly of a Toll-like receptor 4 and 6 heterodimer.
Nat. Immunol.
11
:
155
161
.
35
Mukhopadhyay
S.
,
Varin
A.
,
Chen
Y.
,
Liu
B.
,
Tryggvason
K.
,
Gordon
S.
.
2011
.
SR-A/MARCO-mediated ligand delivery enhances intracellular TLR and NLR function, but ligand scavenging from cell surface limits TLR4 response to pathogens.
Blood
117
:
1319
1328
.
36
Amiel
E.
,
Alonso
A.
,
Uematsu
S.
,
Akira
S.
,
Poynter
M. E.
,
Berwin
B.
.
2009
.
Pivotal advance: Toll-like receptor regulation of scavenger receptor-A-mediated phagocytosis.
J. Leukoc. Biol.
85
:
595
605
.
37
Yu
H.
,
Ha
T.
,
Liu
L.
,
Wang
X.
,
Gao
M.
,
Kelley
J.
,
Kao
R.
,
Williams
D.
,
Li
C.
.
2012
.
Scavenger receptor A (SR-A) is required for LPS-induced TLR4 mediated NF-κB activation in macrophages.
Biochim. Biophys. Acta
1823
:
1192
1198
.
38
Lawrence
T.
,
Natoli
G.
.
2011
.
Transcriptional regulation of macrophage polarization: enabling diversity with identity.
Nat. Rev. Immunol.
11
:
750
761
.
39
Shirey
K. A.
,
Pletneva
L. M.
,
Puche
A. C.
,
Keegan
A. D.
,
Prince
G. A.
,
Blanco
J. C.
,
Vogel
S. N.
.
2010
.
Control of RSV-induced lung injury by alternatively activated macrophages is IL-4R alpha-, TLR4-, and IFN-beta-dependent.
Mucosal Immunol.
3
:
291
300
.
40
Gobert
A. P.
,
Cheng
Y.
,
Akhtar
M.
,
Mersey
B. D.
,
Blumberg
D. R.
,
Cross
R. K.
,
Chaturvedi
R.
,
Drachenberg
C. B.
,
Boucher
J.-L.
,
Hacker
A.
, et al
.
2004
.
Protective role of arginase in a mouse model of colitis.
J. Immunol.
173
:
2109
2117
.
41
Yue
Y.
,
Huang
W.
,
Liang
J.
,
Guo
J.
,
Ji
J.
,
Yao
Y.
,
Zheng
M.
,
Cai
Z.
,
Lu
L.
,
Wang
J.
.
2015
.
IL4I1 is a novel regulator of M2 macrophage polarization that can inhibit t cell activation via L-tryptophan and arginine depletion and IL-10 production.
PLoS One
10
:
e0142979
.
42
Little
M. C.
,
Hurst
R. J.
,
Else
K. J.
.
2014
.
Dynamic changes in macrophage activation and proliferation during the development and resolution of intestinal inflammation.
J. Immunol.
193
:
4684
4695
.
43
Carter
L.
,
Fouser
L. A.
,
Jussif
J.
,
Fitz
L.
,
Deng
B.
,
Wood
C. R.
,
Collins
M.
,
Honjo
T.
,
Freeman
G. J.
,
Carreno
B. M.
.
2002
.
PD-1:PD-L inhibitory pathway affects both CD4(+) and CD8(+) T cells and is overcome by IL-2.
Eur. J. Immunol.
32
:
634
643
.
44
Di Giacinto
C.
,
Marinaro
M.
,
Sanchez
M.
,
Strober
W.
,
Boirivant
M.
.
2005
.
Probiotics ameliorate recurrent Th1-mediated murine colitis by inducing IL-10 and IL-10-dependent TGF-beta-bearing regulatory cells.
J. Immunol.
174
:
3237
3246
.
45
Grabig
A.
,
Paclik
D.
,
Guzy
C.
,
Dankof
A.
,
Baumgart
D. C.
,
Erckenbrecht
J.
,
Raupach
B.
,
Sonnenborn
U.
,
Eckert
J.
,
Schumann
R. R.
, et al
.
2006
.
Escherichia coli strain Nissle 1917 ameliorates experimental colitis via toll-like receptor 2- and toll-like receptor 4-dependent pathways.
Infect. Immun.
74
:
4075
4082
.
46
Cao
S.
,
Feehley
T. J.
,
Nagler
C. R.
.
2014
.
The role of commensal bacteria in the regulation of sensitization to food allergens.
FEBS Lett.
588
:
4258
4266
.
47
Ochoa-Repáraz
J.
,
Mielcarz
D. W.
,
Wang
Y.
,
Begum-Haque
S.
,
Dasgupta
S.
,
Kasper
D. L.
,
Kasper
L. H.
.
2010
.
A polysaccharide from the human commensal Bacteroides fragilis protects against CNS demyelinating disease.
Mucosal Immunol.
3
:
487
495
.
48
Hu
Y.
,
Peng
J.
,
Tai
N.
,
Hu
C.
,
Zhang
X.
,
Wong
F. S.
,
Wen
L.
.
2015
.
Maternal antibiotic treatment protects offspring from diabetes development in nonobese diabetic mice by generation of tolerogenic APCs.
J. Immunol.
195
:
4176
4184
.
49
Kriegel
M. A.
,
Sefik
E.
,
Hill
J. A.
,
Wu
H.-J.
,
Benoist
C.
,
Mathis
D.
.
2011
.
Naturally transmitted segmented filamentous bacteria segregate with diabetes protection in nonobese diabetic mice.
Proc. Natl. Acad. Sci. USA
108
:
11548
11553
.

The authors have no financial conflicts of interest.