Bifidobacterium breve reduces apoptotic epithelial cell shedding in an exopolysaccharide and MyD88-dependent manner

Abstract

Certain members of the microbiota genus Bifidobacterium are known to positively influence host well-being. Importantly, reduced bifidobacterial levels are associated with inflammatory bowel disease (IBD) patients, who also have impaired epithelial barrier function, including elevated rates of apoptotic extrusion of small intestinal epithelial cells (IECs) from villi-a process termed 'cell shedding'. Using a mouse model of pathological cell shedding, we show that mice receiving Bifidobacterium breve UCC2003 exhibit significantly reduced rates of small IEC shedding. Bifidobacterial-induced protection appears to be mediated by a specific bifidobacterial surface exopolysaccharide and interactions with host MyD88 resulting in downregulation of intrinsic and extrinsic apoptotic responses to protect epithelial cells under highly inflammatory conditions. Our results reveal an important and previously undescribed role for B. breve, in positively modulating epithelial cell shedding outcomes via bacterial- and host-dependent factors, supporting the notion that manipulation of the microbiota affects intestinal disease outcomes.

Keywords: Bifidobacterium; epithelial cell shedding; exopolysaccharide; inflammatory bowel disease.

© 2017 The Authors.

Figures

Figure 1.

LPS challenge induces cell shedding from small intestinal villi. C57 BL/6 mice were administered either (a) PBS (control) or (b) LPS by IP injection and proximal small intestines removed after 1.5 h. Processed tissue was sectioned and stained by immunohistochemistry for CC3 (i.e. brown cells indicate shedding event), also highlighted by arrows. A representative picture for each group is shown (12 mice per group, two independent experiments).

Figure 2.

Bifidobacterium breve UCC2003 protects against LPS-induced cell shedding. C57 BL/6 mice received three daily oral gavage doses of (a) PBS or (b) approximately 1 × 109B. breve UCC2003 followed by IP challenge with LPS 24 h later. Representative images are shown. Formalin-fixed, paraffin-embedded intestinal sections were sectioned and stained with anti-CC3 and (c) quantified using the WinCrypts and Score programs, 20 well-orientated hemi-villi were counted per mouse. Data are mean ± s.d., n = 12 (two independent experiments) analysed with a Mann–Whitney U-test.

Figure 3.

The cytoprotective effect of B. breve is not mediated by the TNF-α signalling pathway. C57 BL/6 mice were gavaged with PBS or B. breve and challenged with PBS or LPS for 1.5 h. Columns show TNF-α levels (via ELISA) in (a) whole small intestine intestinal homogenates or (b) plasma ± s.d. (c) Representative immunohistochemical staining for F4/80+ macrophages (brown cells) in control or B. breve-colonized mice. (d) Western blot analysis (F4/80 or housekeeping β-actin) of whole small intestinal homogenates, with (e) columns show relative density of F4/80 from (from d) whole intestinal homogenates. (f) Columns show F4/80 expression via RT-PCR ± SD. (g) Representative histology image of epithelial cell stripping protocol (modified Weiser method) leaving lamina propria intact (as indicated by arrows) and (h) FACS analysis for purity (anti-CD45). (i) Columns shown TNF-R1 expression via RT-PCR ± SD and (j) western blotting for protein expression in isolated intestinal epithelial cells, with (k) columns showing relative density of TNF-R1 (from (j)). n = 9 mice per group, representative of three experiments analysed with ANOVA Kruskal–Wallis test with Dunn's multiple comparison test (a), and with Mann–Whitney U-test (b,e,f,i,k).

Figure 4.

The cytoprotective effect of B. breve is MyD88 dependent. (a,b) IEC MyD88+/+ mice and (c,d) IEC MyD88−/− mice were gavaged with PBS (control) or B. breve UCC2003 and challenged with LPS. Paraffin-embedded intestinal sections were stained with anti-CC3 and quantified using the WinCrypts and Score programs. (e) Columns shown TLR2 expression via RT-PCR. Data are mean ± s.d., n = 12 (two independent experiments) analysed with Mann–Whitney U-test.

Figure 5.

Bifidobacterium breve EPS plays a role in modulating the cytoprotective effect. C57 BL/6 mice were gavaged with either B. breve UCC2003 or (a,b) B. breve UCC2003del (i.e. EPS-negative) or (c,d) B. breve UCC2003inv (i.e. EPS2). Formalin-fixed, paraffin-embedded intestinal sections were stained with anti-CC3 and quantified using the WinCrypts and Score programs. Data are mean ± s.d., n = 12 (two independent experiments) analysed with Mann–Whitney U-test.

Figure 6.

Bifidobacterium breve EPS attenuates inflammatory and apoptosis signalling. Whole small intestinal homogenates from LPS-challenged (a) B. breve UCC2003-EPSdel and (b) B. breve-colonized mice compared with control (i.e. PBS) were analysed using a custom RT-PCR array. Data are mean ± SD, n = 6 (two independent experiments), *p < 0.05 and **p < 0.01 and analysed with Mann–Whitney U test.