Strengthening of the intestinal epithelial tight junction by Bifidobacterium bifidum

Abstract

Epithelial barrier dysfunction has been implicated as one of the major contributors to the pathogenesis of inflammatory bowel disease. The increase in intestinal permeability allows the translocation of luminal antigens across the intestinal epithelium, leading to the exacerbation of colitis. Thus, therapies targeted at specifically restoring tight junction barrier function are thought to have great potential as an alternative or supplement to immunology-based therapies. In this study, we screened Bifidobacterium, Enterococcus, and Lactobacillus species for beneficial microbes to strengthen the intestinal epithelial barrier, using the human intestinal epithelial cell line (Caco-2) in an in vitro assay. Some Bifidobacterium and Lactobacillus species prevented epithelial barrier disruption induced by TNF-α, as assessed by measuring the transepithelial electrical resistance (TER). Furthermore, live Bifidobacterium species promoted wound repair in Caco-2 cell monolayers treated with TNF-α for 48 h. Time course (1)H-NMR-based metabonomics of the culture supernatant revealed markedly enhanced production of acetate after 12 hours of coincubation of B. bifidum and Caco-2. An increase in TER was observed by the administration of acetate to TNF-α-treated Caco-2 monolayers. Interestingly, acetate-induced TER-enhancing effect in the coculture of B. bifidum and Caco-2 cells depends on the differentiation stage of the intestinal epithelial cells. These results suggest that Bifidobacterium species enhance intestinal epithelial barrier function via metabolites such as acetate.

Keywords: 1H‐NMR; intestinal epithelial permeability; metabonomics; probiotics; tight junctions.

Figure 1

Screening of bacterial isolates for TJ-barrier-strengthening effects in polarized Caco-2 monolayers at MOI = 1. All the bacteria isolates were incubated with Caco-2 monolayers for 24 h. The average of actual TER values of Caco-2 monolayer used in this experiment was 1057 ± 98 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes in TER ± SD. **P < 0.01, *P < 0.05 compared with the blank (media alone) control group by t-test.

Figure 2

The preventive effect against TNF-α-induced TJ barrier impairment. (A) Schedule of Caco-2 monolayer treatment and TER analysis. The bacterial suspensions were added to the apical side of Caco-2 monolayers an hour before TNF-α stimulation. After 1 h, Caco-2 monolayers were stimulated on the basolateral side with TNF-α, and then incubated for 24 h. (B) The preventive effect of bacterial isolates with barrier-strengthening ability at MOI = 1. The average of actual TER values of Caco-2 monolayer used in this experiment was 976 ± 166 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes of TER ± SD. **P < 0.01, *P < 0.05 compared with the TNF-α control group by t-test.

Figure 3

Repair of TNF-α-induced barrier dysfunction by administration of bacterial isolates post TNF-α stimulation. (A) The schedule of TNF-α and bacterial treatment of Caco-2 monolayers. Caco-2 monolayers were stimulated on the basolateral side with TNF-α, and then incubated for 48 h. After 48 h, the bacterial suspension were added into the apical side, and coincubated for 24 h. (B) Repair effect of four Bifidobacterial species on the TJ function of Caco-2 cell monolayer damaged by basolateral TNF-α stimulation. The average of actual TER values of Caco-2 monolayer used in this experiment was 879 ± 147 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes in TER ± SD. **P < 0.01 compared with the TNF-α treated TER of the same group by t-test.

Figure 4

Characterization of the factor responsible for B. bifidum-induced TJ restoration. (A) Effects of the heat treatment (95°C for 10 min) of bacterial cells on B. bifidum WU12-induced epithelial TJ restoration capacity. The average of actual TER values of Caco-2 monolayer before basolateral TNF-α stimulation used in this experiment was 894 ± 99 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes in TER ± SD before and after administration of B. bifidum WU12. Statistical differences between before and after administration of B. bifidum WU12 in each experimental group were calculated by t-test (**P < 0.01). (B) The cytotoxic effect of MOI on cocultivation of Caco-2 monolayers with alive or heat-killed B. bifidum WU12. The average of actual TER values of Caco-2 monolayer used in this experiment was 753 ± 91 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes of TER ± SD. Statistical differences before and after administration of alive or heat-killed B. bifidum WU12 were calculated by t-test (**P < 0.01).

Figure 5

The kinetics of TER and TJ-related mRNA expression during B. bifidum WU12-induced Caco-2 monolayer restitution. (A) B. bifidum WU12-induced Caco-2 monolayer restitution over time as estimated by TER. Caco-2 monolayers incubated without B. bifidum WU12 were used as the negative control. The average of actual TER values of Caco-2 monolayer before basolateral TNF-α stimulation used in this experiment was 822 ± 86 Ω cm². Experiments were carried out in triplicate, and data represent the means of relative changes of TER ± SD. (B–D) Temporal changes of mRNA expression changes in TJ protein genes (Claudin-1, Occludin, ZO-1) during Caco-2 monolayer restitution induced by B. bifidum WU12. The time course of mRNA expression of nontreated group (w/o B. bifidum) was used as the negative control. Experiments were carried out in triplicate, and data represent the means of relative changes in ± SD. Statistical differences were calculated by t-test, comparing conditions with the negative control group at the same time point (**P < 0.01).

Figure 6

Principle component analysis (PCA) score plots and loading plots of the metabolic profile affected by B. bifidum WU12. (A) PCA score plot was computed from blank (square), Caco-2 mono-culture 48 h (with TNF-α; diamond), time-dependent change of the coculture group (TNF-α treated; circle), time-dependent change in B. bifidum WU 12 mono-culture group (triangle), B. bifidum WU12 coculture control (dash). (B) PCA loading plot derived from the information on the PCA score plots. Experiments were carried out in duplicate.

Figure 7

Increased production of acetate and formate by the cocultures of Caco-2 monolayer and B. bifidum WU. Acetate, formate, and lactate concentrations in the culture supernatant after 24 h incubation were determined by ion chromatography. Experiments were carried out in triplicate, and data represent the means ± SD. Statistical differences were calculated by t-test (**P < 0.01).

Figure 8

Production of acetate and formate in the coculture supernatant varied in a strain-dependent manner. (A–C) Acetate, formate, and lactate concentrations in the supernatant after 24 h cocultures of Caco-2 monolayer and seven different Bifidobacterial strains were determined by ion chromatography. Experiments were carried out in triplicate, and data represent the means ± SD. Statistical differences were calculated by t-test (**P < 0.01, *P < 0.05).

Figure 9

Acetate induced the TJ barrier restorative capacity. (A) Effects of acetate and formate on the TJ barrier restoration. Caco-2 monolayers were stimulated on the basolateral side with TNF-α, and then incubated for 48 h. After 48 h, acetate or formate were added into the apical side, and incubated for 24 h. The average of actual TER values of Caco-2 monolayer used in this experiment was 846 ± 96 Ω cm². Data represent the average of two or three independent experiments carried out in triplicate. (B) Effects of acetate or B. bifidum WU12 on mRNA expression of TJ protein genes. TJ protein gene expressions in Caco-2 cells treated with acetate or B. bifidum WU12 for 24 h was were determined by quantitative RT-PCR. Data represent the average of three independent experiments carried out in duplicate or triplicate. Statistical differences were calculated by t-test (*P < 0.05).

Figure 10

Characterization of B. bifidum-induced strengthening epithelial barrier function at different time points during Caco-2 monolayer polarization. (A) B. bifidum-induced TER enhancement: Day7, nonpolarization (actual TER values, 307 ± 77 Ω cm²); Day14, early-polarization (actual TER values, 508 ± 19 Ω cm²); Day20, stable-polarization (actual TER values, 806 ± 49 Ω cm²). TER changes were calculated from the TER value measured before and 24 h after the coincubation with B. bifidum WU12. Data represent the average of three independent experiments carried out in triplicate. (B) Short chain fatty acids production in nonpolarized (Day 7) or stable-polarized Caco-2 cells (Day 20) treated with B. bifidum WU12. Experiments were carried out in triplicate, and data represent the means ± SD. Statistical differences were calculated by t-test (**P < 0.01, *P < 0.05).