Introduction Recent advances in metagenomic analysis of intestinal bacteria have revealed that inflammatory bowel diseases (IBD) is associated with dysbiosis in the intestinal microflora [1], [2], [3]. In support of these human studies, analysis of mice lacking NLRP6 has revealed that altered composition of intestinal symbiotic bacteria contributes to the pathogenesis of colitis [4]. Probiotics, live microorganisms which confer a health benefit on the host when administered in appropriate amounts, have been used for the treatment of IBD [5]–[8]. Probiotics have been shown to modulate the intestinal symbiotic bacteria leading to the maintenance of intestinal homeostasis [9], [10]. Modulation of microbiota by probiotics has been shown to be elicited by antagonizing pathogenic bacteria through the reduction of luminal pH, inhibition of bacterial adherence, or production of anti-microbial molecules [8]. Probiotics have also been shown to enhance barrier functions of intestinal epithelial cells [11]. Thus, several mechanisms for the cross-talk between probiotics and the host have been postulated. Recent accumulating evidence has indicated that intestinal commensal microbiota has a great influence on the host intestinal immune system [12]–[14]. Commensal microbiota has been shown to induce IgA-mediated responses and development of Th1/Th17 effector T cells as well as regulatory T (Treg) cells [15]–[17]. More recently, a specific microbiota that induces development of Th17 cells or Treg cells has been demonstrated. Segmented filamentous bacteria (SFB), which have been previously shown to induce IgA-producing cells in the small intestine, were shown to induce Th17 cell development in the small intestine of mice [18], [19]. A human symbiotic bacterium, Bacteroides fragilis has been shown to mediate Toll-like receptor 2 (TLR2)-dependent development of Foxp3+ Treg cells in the large intestine [20]–[22]. Clostridium species mediate TLR-independent induction of Foxp3+ Treg cells in the large intestine [23]. Thus, several selective intestinal bacteria promote development of intestinal T cells via distinct mechanisms. Most recently, microbiota-dependent induction of Foxp3+ Treg cells has been shown to be required for the establishment of intestinal CD4+ T cell homeostasis [24]. Additionally, commensal microbiota has been shown to educate Foxp3+ Treg cells to acquire the antigen-specific repertoires of their T cell receptors [25]. Probiotics have also been shown to directly modulate the host immune system, especially the induction of Foxp3+ Treg or TGF-β-bearing Treg cell development [26]–[29]. Thus, several mechanisms for intestinal bacteria-dependent development of Foxp3+ Treg cells have been postulated. Intestinal homeostasis is maintained by regulatory T cell populations consisting of two major CD4+ T cell subsets; Foxp3+ Treg cells and IL-10-producing type 1 regulatory T (Tr1) cells [30]. Tr1 cells modulate immune responses via mechanisms distinct from those used by Foxp3+ Treg cells [31]. Indeed, Tr1 cells do not express the master Treg transcription factor Foxp3, and are induced by distinct cytokines such as IL-10 and IL-27 [32], [33]. Tr1 cells are abundant in the intestinal lamina propria [34], yet it remains unclear how Tr1 cells develop in the intestine. In this study, we analyzed the effect of two probiotic strains, Bifidobacterium breve and Lactobacillus casei, on intestinal T cells responses. Oral administration of B. breve, but not L. casei, resulted in increased IL-10 production from colonic CD4+ T cells, without enhancing Foxp3 expression. B. breve-induced IL-10-producing CD4+ T cells possessed properties of Tr1 cells, as evidenced by expression of cMaf, Il21, and Ahr. B. breve-dependent Tr1 cell induction was mediated by intestinal CD103+ dendritic cells via TLR2/MyD88-dependent production of IL-10 and IL-27. B. breve administration ameliorated intestinal inflammation in immunocompromised mice transferred with naïve CD4+ T cells in an IL-10-dependent manner. These findings establish the mechanisms for Tr1 cell induction by the probiotic B. breve, which modulates the host immune responses. Results B. breve induces IL-10-producing CD4+ T cell in the colon Lactobacillus casei strain Shirota and Bifidobacterium breve Yakult strain have been proven to be beneficial for the treatment of several diseases such as diabetes mellitus, arthritis and inflammatory bowel diseases [35]–[40]. In order to analyze the effect of these probiotic strains on the intestinal homeostasis, we orally treated C57BL/6 mice with L. casei and B. breve (109 bacteria each) for 3 months. We first analyzed fecal microbiota using both quantitative PCR and reverse transcription-quantitative PCR methods targeting rDNA and rRNA, respectively [41]. Administration of L. casei and B. breve did not induce a significant change in the number and composition of microbiota (Text S1, Table S1). Because several microbiota have been shown to induce differentiation of intestinal CD4+ T cells [17], we analyzed production of IL-10, IL-17, and IFN-γfrom CD4+ T cells in the small intestine and large intestine of mice orally treated with L. casei and B. breve. The number of IL-10-, IL-17-, and IFN-γ-producing T cells in both the small intestine and the large intestine was not altered in mice administered with L. casei (Figure 1A, B). In B. breve-treated animals, the number of IL-17- and IFN-γ-producing T cells in the small intestine and the large intestine was not significantly changed. However, the number of IL-10-producing T cells was increased about two-fold in the large intestine, but not altered in the small intestine, spleen, and mesenteric lymph nodes (MLN) (Figure 1C, D and Figure S1). Thus, administration of B. breve in C57BL/6 mice selectively increased the number of IL-10-producing CD4+ T cells in the large intestine without modulating intestinal microbiota. 10.1371/journal.ppat.1002714.g001 Figure 1 Induction of IL-10-producing CD4+ T cells by B. breve in the colonic lamina propria. 6-week-old C57BL/6 mice were fed L. casei or B. breve or placebo daily (each, 1×109) by oral gavage for 3 months (n = 8). Intestinal lamina propria lymphocytes were analyzed for cytokine production by flow cytometry. Percentages of IL-10-, IL-17-, and IFN-γ-producing CD4+ T cells of mice administered with L. casei (A) or B. breve (C) were shown. *P 98%. The cells were used immediately for each of experiment. Isolation of splenic naïve CD4+ cells To prepare single-cell suspensions from spleens, they were ground between glass slides and passed through a 40 µm cell strainer. Splenocytes were treated with RBC lysis buffer (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM EDTA) for 5 min and washed twice with PBS. For FACS sorting, cells were stained with PerCP/Cy5.5-conjugated anti-CD4 (Biolegend), APC-conjugated anti-CD62L, FITC-conjugated anti-CD25 and PE-conjugated anti-CD44 (BD Biosciences). Naïve CD4+ T cells were sorted using a FACSAria for CD4+CD62LhighCD25−CD44low. The purity of the sorted cells was routinely >98%. Intracellular cytokine staining The intracellular expression of IFN-γ, IL-17, and IL-10 in CD4+ T cells was analyzed using the Cytofix/Cytoperm Kit Plus (with Golgistop; BD Biosciences) according to the manufacturer's instructions. In brief, lymphocytes obtained from the intestinal lamina propria were incubated with 50 ng/ml of phorbol myristate acetate (PMA; Sigma) and 5 µM of calcium ionophore A23187 (Sigma) and Golgistop in complete RPMI1640 at 37°C for 4 h. Surface staining was performed with PerCP/Cy5.5-conjufated anti-CD4 for 20 min at 4°C. After Fix/Perm treatment for 20 min, intracellular cytokine staining was performed with PE-conjugated anti-IL-10, FITC-conjugated anti-IFN-γ, and APC-conjugated anti-IL-17 for 20 min. Data were acquired using a FACS Canto II and analyzed using FlowJo software. Alternatively, for intracellular staining for Foxp3 and IL-10, cells were stained using the Foxp3 Staining Buffer set (eBiosciences). In vitro co-culture assays Colonic DC subsets (5×104) were incubated with the same number or the indicated number of L. casei or B. breve in 100 µl of complete RPMI1640 media for 24 h in a round-bottom 96 well plate. DCs were then washed with PBS and naïve CD4+ T cells (5×104) were added into the culture with 2 µg/ml soluble anti-CD3 mAb. After 4 days, T cells were collected, washed and counted. The same numbers of T cells were re-stimulated with plate-bound anti-CD3 mAb (2 µg/ml) and soluble anti-CD28 mAb (2 µg/ml) for 24 h. Re-stimulated T cell cytokine production in the supernatants was analyzed by ELISA (R&D systems). Alternatively, T cells were re-stimulated with 50 ng/ml of PMA and 5 µM of calcium ionophore A23187 for 6 h before intracellular cytokine staining was performed as described above. Golgistop was added for the last 2 h. Quantitative real-time RT–PCR Total RNA was isolated with the RNeasy Mini Kit (Qiagen), and 1–2 µg of total RNA was reverse transcribed using M-MLV reverse transcriptase (Promega) and random primers (Toyobo) after treatment with RQ1 DNase I (Promega). Complementary DNAs were analyzed by qPCR using the GoTaq qPCR Master Mix (Promega) on an ABI 7300 system (Applied Biosystems). All values were normalized to the expression of Gapdh encoding glyceraldhyde-3-phosphate dehydrogenase, and the fold difference in expression relative to that for Gapdh is shown. Amplification conditions were: 50°C (2 min), 95°C (10 min), and 40 cycles of 95°C (15 s) and 60°C (60 s). The following primer sets were used: cMaf, 5′-AATCCTGGCCTGTTTCACAT-3′ and 5′-TGACGCCAACATAGGAGGTG-3′; Il21, 5′-GCCAGATCGCCTCCTGATTA-3′ and 5′-CATGCTCACAGTGCCCCTTT-3′; Il27p28, 5′-TTCCCAATGTTTCCCTGACTTT-3′ and 5′-AAGTGTGGTAGCGAGGAAGCA-3′; Ebi3, 5′-TGAAACAGCTCTCGTGGCTCTA-3′ and 5′-GCCACGGGATACCGAGAA-3′; Il10, 5′-TTTCAAACAAAGGACCAG-3′ and 5′-GGATCATTTCCGATAAGG-3′; and Gapdh, 5′-TGTGTCCGTCGTGGATCTGA-3′ and 5′-CCTGCTTCACCACCTTCTTGA-3′ T-cell-mediated colitis model Naive CD4+CD62LhighCD25−CD44low splenic T cells from BALB/c mice or Il10 −/− mice (BALB/c background) were purified and intraperitoneally transferred into SCID mice (3×105 cells per mouse). B. breve (109 bacteria) were fed by oral gavage from 3 days before the transfer to the end of the experiments. Weight changes were monitored every day. The mice were sacrificed, and the colons were examined histochemically after haematoxylin and eosin staining. Alternatively, the colons were cut into small pieces after wash and cultured for 24 h. Then, culture supernatants were collected and the level of IL-10, IL-17A and IFN-γ was measured by ELISA (R&D systems). Histopathological analysis Paraffin-embedded colon samples were sectioned and stained with hematoxylin and eosin. Severity of colitis was evaluated by the standard scoring system as previously described [68]. Five regions of the colon (cecum; ascending, transverse, and descending of colon; and rectum) were graded semiquantitatively from 0 (no change) to 5 (most severe change). The grading represents an increasing incidence and degree of inflammation, goblet cell loss, ulceration and fibrosis in the lamina propria. The scoring was performed in a blinded manner. Images of hematoxylin and eosin staining and May-Grunwald-Giemsa staining were taken using Biozero (Keyence). Statistical analysis Statistical analysis was performed using PRISM 4 software. Unpaired student's t-test and Mann-Whitney U test were used to determine the significance of experiments. P values of less than 0.05 were considered statistically significant. Supporting Information Figure S1 Percentage of IL-10+ or Foxp3+ CD4+ T cells in MLN or spleens were not changed by oral treatment of B. breve . 6-week-old C57BL/6 mice were fed with B. breve or placebo daily by oral gavage for 3 months (n = 8). MLNs and spleens were taken, and analyzed for expression of cytokines and Foxp3 by flow cytometry. Representative FACS dot plots were shown gated on CD4+ T cells. A: MLN, B: Spleen. (PDF) Click here for additional data file. Figure S2 B. breve induces IL-10-producing Tr1 cells in a dose-dependent manner. CD11chigh CD11b−CD103+ DCs (CD103+ DCs) (5×104) were isolated from the colonic lamina propria of C57BL/6J mice, and treated with the increasing numbers of B. breve (5×101 to 5×105) for 24 h in round-bottom 96-well plate. After washing, splenic naïve CD4+ T cells (5×104) were co-cultured with B.breve-treated CD103+ DC in the presence of anti-CD3 mAb for 4 days. Then, T cells were harvested and re-stimulated. IL-10 production in the culture supernatants was analyzed by ELISA. Data are representative of two independent experiments. Error bars, S.D. *P<0.05, **P<0.01. (PDF) Click here for additional data file. Figure S3 Retinoic acid-independent induction of Tr1 cells by B. breve . B. breve-treated CD103+ DCs were co-cultured with splenic naïve CD4+ T cells in the presence of an inhibitor of retinoic acid receptor (2 µM of LE540, WAKO chemicals, JAPAN) for 4 days. IL-10 production by re-stimulated T cells was quantified by ELISA. Data are representative of two independent experiments. Error bars, S.D. *P<0.01, N.S, not significant. (PDF) Click here for additional data file. Figure S4 TLR4/TLR9-independent induction of Tr1 cells by B. breve . Intestinal CD103+ DCs from wild-type, Tlr4 −/− and Tlr9 −/− mice were treated with B. breve for 24 h, and then co-cultured with splenic naïve CD4+ T cells for 4 days. IL-10 production by re-stimulated T cells was quantified by ELISA. Data are representative of two independent experiments. Error bars, S.D. *P<0.01. (PDF) Click here for additional data file. Figure S5 TLR2-dependent induction of Tr1 cells. Intestinal CD103+ DCs were stimulated with B. breve or TLR ligands such as LPS (TLR4 ligand), Pam3 (TLR2 ligand) or flagellin (TLR5 ligand) for 24 h, and then co-cultured with splenic naïve CD4+ T cells for 4 days. IL-10 production by re-stimulated T cells was quantified by ELISA. Data are representative of two independent experiments. Error bars, S.D. *P<0.01. (PDF) Click here for additional data file. Figure S6 B. breve directly acts on CD103+ DCs to induce Tr1 cells. CD103+ DCs were treated by B.breve or culture supernatant (10-fold concentrated) of B. breve for 24 h. After washing, naïve CD4+ T cells were co-cultured with treated CD103+ DCs for 4 days. Then, T cells were harvested and re-stimulated by anti-CD3 and anti-CD28 mAbs. IL-10 concentration in the supernatants was quantified by ELISA. Representative data were shown from two independent experiments. Error bars, S.D. N.D, not detected. (PDF) Click here for additional data file. Figure S7 Induction of Tr1 cell development by killed B. breve . CD103+ DCs were treated by live, UV killed or sonicated B. breve for 24 h, and then, co-cultured with naïve CD4+ T cells for 4 days. T cells were harvested and re-stimulated by anti-CD3 and anti-CD28 mAbs. IL-10 concentration in the supernatants was quantified by ELISA. Data were representative of three independent experiments. Error bars, S.D. *P<0.01, N.S, not significant. (PDF) Click here for additional data file. Table S1 Composition of fecal commensal microflora in probiotics-fed mice. 6-week-old C57BL/6 mice were fed with L. casei, B. breve or placebo daily (1×109) by oral gavage for 3 months (n = 5, respectively). Fecal samples were collected, weighed and suspended in 9 volumes of sterilized anaerobic transfer medium. Total RNA and DNA fractions extracted from each sample were assessed by RT-qPCR or qPCR with the specific primers. “Number” indicates CFU of each bacteria calculated using control cultured bacteria. (x/5) indicated the right side of “number” show detection rate of mice analyzed. (PDF) Click here for additional data file. Text S1 Supplemental methods. Methods for “Analysis of Fecal Microbiota” and “Culture and Killing of B. breve” are described with references. (PDF) Click here for additional data file.
