Distinct B cell subsets in Peyer's patches convey probiotic effects by Limosilactobacillus reuteri

Abstract

Background: Intestinal Peyer's patches (PPs) form unique niches for bacteria-immune cell interactions that direct host immunity and shape the microbiome. Here we investigate how peroral administration of probiotic bacterium Limosilactobacillus reuteri R2LC affects B lymphocytes and IgA induction in the PPs, as well as the downstream consequences on intestinal microbiota and susceptibility to inflammation.

Results: The B cells of PPs were separated by size to circumvent activation-dependent cell identification biases due to dynamic expression of markers, which resulted in two phenotypically, transcriptionally, and spatially distinct subsets: small IgD⁺/GL7^-/S1PR1⁺/Bcl6, CCR6-expressing pre-germinal center (GC)-like B cells with innate-like functions located subepithelially, and large GL7⁺/S1PR1^-/Ki67⁺/Bcl6, CD69-expressing B cells with strong metabolic activity found in the GC. Peroral L. reuteri administration expanded both B cell subsets and enhanced the innate-like properties of pre-GC-like B cells while retaining them in the sub-epithelial compartment by increased sphingosine-1-phosphate/S1PR1 signaling. Furthermore, L. reuteri promoted GC-like B cell differentiation, which involved expansion of the GC area and autocrine TGFβ-1 activation. Consequently, PD-1-T follicular helper cell-dependent IgA induction and production was increased by L. reuteri, which shifted the intestinal microbiome and protected against dextran-sulfate-sodium induced colitis and dysbiosis.

Conclusions: The Peyer's patches sense, enhance and transmit probiotic signals by increasing the numbers and effector functions of distinct B cell subsets, resulting in increased IgA production, altered intestinal microbiota, and protection against inflammation. Video abstract.

Keywords: Gut microbiome; Inflammatory bowel disease; Innate-like B lymphocytes; PD-1 dependent; Probiotics; R2LC.

Fig. 1

Phenotypic and transcriptional distinction between B cell subsets in Peyer’s patches. a Flow cytometry of PPs live CD3^-CD19⁺B220⁺ B lymphocytes (gated as shown in Fig. S1A) and separated into the small and large B cell subsets by forward scatter area and expressed as mean FSC-A (size), and the number of small-B and large-B cells per mg tissue (n = 6 mice left panel; n = 24 mice right panel). b Imaging flow cytometry of the areas (μm²) of small-B and large-B cell from three independent experiments. c, d Immunohistochemistry of PPs stained with anti-B220 (magenta), anti-CD138 (yellow) and Hoechst (blue) in c or anti-B220 (white), anti-GL7 (green), anti-Ki67 (cyan), anti-IgA (yellow), and anti-IgD (blue) in d. Scale Bars equal 100 μm or 10 μm in closed-up reviews. e Enrichment of gene ontology categories (Biological Process, BP) for genes differentially expressed in large B (upper panel) versus small B cells (lower panel) determined by microarray of lin^-CD19⁺B220⁺ cells sorted on FSC-A (n = 4 samples per group), the number of genes in each functional category is shown. Data were adjusted by false discovery rate control (FDR). f, g Displays of heat maps of genes expression by q-RT-PCR in sorted small and large B cells. The expression was normalized to the mean value of large B and each column represents one sample. h, i Heat maps and histograms depicting expression of surface markers on small and large B cells (n = 6 mice per group). The frequency of indicated markers (normalized to the mean value of large B cells). i MFI of B cells positive for GL7 and S1PR1, and percentage in each subset. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 using two-tailed Student’s t test

Fig. 2

L. reuteri increases the population of Peyer’s patches B cells and enhances their effector functions. a Flow cytometry of small-pre-GC-like and large-GC-like B cell subsets from control mice or mice treated with 10⁸ cfu L. reuteri (R2LC, 4659 or 6475) perorally for 7 consecutive days (n = 5–12 mice per group). b Numbers of S1PR1⁺ small pre-GC-like B cells (CD3^-CD19⁺B220⁺) and their S1PR1 expression (MFI, n = 5–12 mice per group). c Illustration of enzymes regulating sphingosine-1-phosphate (S1P) homeostasis (top). q-RT-PCR analysis of gene expression (fold change, n = 6 mice per group) of the S1P pathway (bottom). d Experimental design evaluating the effect of the sphingosine-1-phosphate (S1P) receptor modulator FTY720 (top), and flow cytometry of B cell subsets (lower panels, n = 10–11 mice per group). e Principal component analysis of the microarray data, where each dot represents one sample (n = 4 samples per group). f Enrichment of gene ontology categories (BP) for genes upregulated by L. reuteri in small B cells. g Flow cytometry of DEFa3⁺ B cells (CD11C^-CD19⁺B220⁺) in PPs and DEFa3 expression (MFI, n = 5–6 mice per group). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01 using two-tailed Student’s t test

Fig. 3

L. reuteri expands the GC-like B cell population and reprograms its functional gene signature. a, c Immunohistochemistry of the GL7⁺ area stained with anti-B220 (magenta), anti-GL7 (green) and anti-Ki67 (cyan). Scale bars equal 100 μm. Representative images of PPs from the control or L. reuteri-treatment are shown in a and c, and the corresponding quantification shown in Fig. S3G, H. b, d Flow cytometry quantification of GL7⁺ B cells (in CD3⁻CD19⁺B220⁺, n = 5–6 mice per group) and Ki67⁺ cells (n = 5–6 mice per group). e In vivo proliferation assay was performed by flow cytometry following i.p. injection of EdU. A histogram depicting EdU expression in GL7⁻, GL7⁺ or EdU⁺CD19⁺B220⁺B cells and the number of EdU⁺GL7⁺ B cells (n = 6 mice per group). f Expression of Mki67, Gsk-3α, and Hif-1α in large GC-like B cells from control or L. reuteri-treated mice (GOI, gene of interest; A.U., arbitrary unit). g q-RT-PCR analysis of gene expression levels (n = 6 mice per group) and expression of TGFβR1 (microarray analysis). h Flow cytometry analyzing the numbers of TGFβ1⁺ B cells as well as the TGFβ1 levels (MFI, n = 6–18 mice per group). i The PPs tissue was analyzed for cytokine/chemokine production by Multi-Plex Mesoscale and ELISA, normalized to tissue protein content (n = 6 mice per group). j Expression of α germline transcripts (αGT) in PPs (n = 6 mice per group). k The numbers of B cells of different subsets in PPs in untreated mice and in response to L. reuteri R2LC and L. reuteri R2LC_ΔADO (n = 6 mice per group). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01 using Student’s t test or ANOVA with Tukey’s post hoc test

Fig. 4

L. reuteri promotes B cell IgA-responses in a Tfh-PD-1-dependent manner. a The area of IgA⁺ plasma cells (anti-CD138) quantified in ileal and colonic tissues, n = 5 mice per group (duplicate slides per mouse, scale bars equal 100 μm). b Free IgA (μg/ml, ELISA) in the lumen of ileum and colon (n = 10–11 mice per group) and c Serum IgA and IgG concentrations (μg/ml, ELISA, n = 5-6 mice per group). d Flow cytometry of IgA⁺ bacteria in ileum (n = 16-17 mice per group). e, f Ileal microbiome assessed by 16S rRNA gene amplicon sequencing. Microbial community diversity was calculated as α-diversity in mice treated with L. reuteri and/or DSS compared to controls, n = 4–9 mice per group. Data are presented as median values. g Localization of PD-1⁺ T cells in PPs stained with anti-PD-1 (red), anti-CD3 (green), and Hoechst (blue, scale bar equals 100 μm) from three independent experiments. h Flow cytometry of Tfh cells and PD-1 expression (n = 13-14 mice per group). A histogram of PPs and spleen indicating a uniquely high expression of PD-1 in PPs T cells. i Design of PD-1 monoclonal antibodies blocking (blue arrows) and CD4⁺ T cell depletion (black arrows) experiments and effects on the number of B cell subsets (n = 6–13 mice per group). j, k Free IgA production and quantification of IgA⁺ bacteria in the ileum from PD-1 (mAb) blocking experiments (n = 6 mice per group). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 using two-tailed Student’s t test or ANOVA with Tukey’s post hoc test

Fig. 5

L. reuteri provides protection against DSS-induced colitis, disruption of Peyer’s patches and intestinal dysbiosis. a Macroscopic analysis of the PPs surface area (mean area of PPs per mouse, mm², n = 5 mice per group) in WT mice (control), WT mice receiving DSS (DSS, 3% in drinking water for 7 consecutive days) as well as mice receiving L. reuteri R2LC was given daily to WT mice for 14 days starting 7 days prior to DSS-treatment (L. reuteri-DSS). Histological analysis of the PPs height (μm, serially sectioned with all slides per PP analyzed, n = 5 mice per group, scale bars equal 200 μm). b Flow cytometry of live CD3^-CD19⁺B220⁺ B cells, small-pre-GC-like and large-GC-like B cells in PPs (n = 6–17 mice per group). c Body weight change and disease activity index of DSS-treated mice and mice co-treated with L. reuteri (n = 9 mice per group) or FTY720 (n = 6 mice per group). d Pearson correlation of PPs B cell number and body weight loss (%) (n = 24 mice). e Bar graph depicts bacterial community composition of individual mouse, n = 4 mice per group. f–h Relative abundance of Erysipelotrichaceae, S24-7 and Lactobacillaceae in the ileal microbiota (n = 4–9 mice per group). Data are presented as median values. Contrast analysis demonstrated that bacterial taxa changed significantly with DSS-treatment (diseased P < 0.05) and was preserved by L. reuteri-treatment (protected P < 0.05). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 using ANOVA with Tukey’s post hoc test