Male C57 mice (body weight, 23±1 g; 6 weeks old) were obtained from the Animal Center, Nanjing Drum Tower Hospital (Nanjing, China) and the in vivo experiment was performed in the same facility. Mice were maintained under controlled conditions (25°C, 55% humidity, 12 h light/dark cycle) and fed standard laboratory food. Mice were administered 3% (wt/vol) dextrose sulfate sodium (DSS) (molecular weight, 35,000–44,000; MP Biomedicals, Inc., Aurora, OH, USA) via drinking water for seven days. Additionally, mice were treated daily with different reagents via gavage (catheter diameter, 1.2 mm), including normal saline (NS; n=8) or B. subtilis (R179; Beijing Hanmi Pharm Co., Ltd., Beijing, China) at a high (1×10⁹ CFU/mouse/day; n=8) or low (1×10⁸ CFU/mouse/day; n=8) dosage until the end of the study. On day eight, mice were weighed and then sacrificed via ether exposure (200 mg/l; Shanghai National Medicine Group, Shanghai, China) in an airtight container in a biosafety cabinet. Colons were harvested, measured and fixed in 4% formalin for subsequent histological examination. Animal experiments were approved by the Ethics Committee of Medical Research, Huashan Hospital of Fudan University (Shanghai, China).

Following the initiation of DSS treatment, daily changes in body weight and clinical signs of colitis, such as rectal bleeding, diarrhea and piloerection, were examined. The disease activity index consisted of scoring for rectal bleeding (0–4), as previously reported (13). Hemoccult SENSA (Beckman Coulter, Inc., Brea, CA, USA) was used to examine rectal bleeding.

Alcian blue staining was performed according to a previous report (14). Tissue sections (6 µm thick) were immersed in 100% ethanol for 10 min, rinsed in water for 10 min, immersed in 3% acetic acid for 2 min and subsequently stained in 1% alcian blue 8GX in 3% acetic acid (pH 2.5) for 2.5 h. To remove non-specific staining, 3% acetic acid and water was used to rinse the sections for 10 min. Slides were subsequently oxidized in 1% periodic acid in water at room temperature for 10 min, washed in water for 5 min, immersed in Schiff's reagent for 10 min, rinsed in water for 5 min and three times in 0.5% sodium metabisulphite prior to a final wash in water. To reveal O-acetylated oligosaccharides, sections were treated with 0.1 M KOH for 30 min and 1 mM periodic acid prior to the Schiff reagent.

Frozen tissue sections (6 µm thick) were immunostained with 1:100 primary antibodies against ZO-1 (clonality, H-300; cat. no. sc-10804) and claudin (clonality, D-4; cat. no. sc-137121; species: mouse, rat, human, equine, canine, bovine, porcine) (both Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Images were analyzed using a BIOREVO immunofluorescence microscope (Keyence Corp., Osaka, Japan). Each result was obtained from at least three separate experiments. Six mice per group were prepared for each experiment.

Intestinal permeability was determined according to a previously described method (15). DSS-treated mice with high/low-dose B. subtilis or control saline (n=5; 4 days) were fasted for 12 h prior to oral gavage of disaccharide permeability probes [100 mg/ml lactulose and 50 mg/ml mannitol (both Sigma-Aldrich; Merck Millipore, Darmstradt, Germany) dissolved in 2 ml water] and urine was collected 12 h later. Urine volume was measured and the concentrations of lactulose and mannitol were determined by high-performance liquid chromatography with an NH2 column (Bischoff Chromatohraphy, Leonberg, Germany) and acetonitrile (70%) based elution. The ratio of the amount of probe in urine to the amount administered as lactulose or mannitol recovery rate was calculated accurately. Intestinal permeability was evaluated as a ratio of lactulose recovery rate to mannitol recovery rate.

On the 4th day following DSS treatment, mice were anesthetized with ether (200 mg/l; Shanghai National Medicine Group) in an airtight container within a biosafety cabinet and blood was collected from the retrobulbar venous plexus using pyrogen-free heparinized syringes. Cytokine [interleukin (IL)-10, IL-12 p70, IL-17A, and IL-23] levels were analyzed by ELISA according to the manufacturer's protocol (R&D Systems, Inc., Minneapolis, MN, USA). Plasma endotoxin was measured using a Limulus amebocyte lysate pyrogen test kit (Xiamen Houshiji, Ltd., Xiamen, China; cat. no. KC48).

Fresh mice fecal samples were collected from the cages, weighed and stored at −80°C. Fecal samples were mixed with distilled water and centrifuged (2,500 × g). The supernatant was removed, filtered and mixed with ether and sulfuric acid. Following centrifugation (2,500 × g), the ether layer was collected and measured in an Agilent 6890N gas chromatograph machine (Agilent Technologies, Inc., Santa Clara, CA, USA) to determine the total SCFA concentrations.

Microbiota composition was assessed by 454 pyrosequencing (GS FLX TI technology, Genoscreen, Lille, France) targeting the V3-V4 region of the bacterial 16S rRNA gene (V3, forward 5′-TACGGRAGGCAGCAG-3′ and V4 reverse 5′-GGACTACCAGGGTATCTAAT-3′). Sequences were binned for a minimal sequence length of 300 pb, a minimal base quality threshold of 30 cycles and a maximum homopolymer length of 6 cycles. Resulting sequences were assigned to different taxonomic levels, from phylum to genus using the Ribosomal Database Project (16). Sequences were further clustered into operational taxonomic units (OTUs) or phylotypes at 97% of identity using the Quantitative Insights into Microbial Ecology pipeline and CD-HIT (17,18). OTUs were assigned to their closest taxonomic neighbors and relative bacterial species using Seqmatch (Michigan State University, East Lansing, MI, USA) and Blastall (National Centre for Biotechnology Information, Bethesda, MD, USA).

Data were expressed as the mean ± standard error of the mean. Differences were analyzed using Student's t-test, Chi-square test, or one-way analysis of variance with Tukey's post-hic test for multiple group comparison. P<0.05 was considered to indicate a statistically significant difference.

To examine the role of B. subtilis in the amelioration of UC in vivo, mice were initially exposed to a lethal dose of 4% DSS (20 ml/d), a pharmacological agent used to induce UC that also causes severe secondary symptoms. Mice were subsequently treated orally with either a high- or low-dose of B. subtilis preparation, or phosphate-buffered saline for control at eight days post-administration. The results indicated that the high dose of B. subtilis solution protected mice from the lethal effect of DSS-induced UC (Fig. 1A). At the end of the study, ~50% survival of probiotic-treated mice was observed. In contrast, only ~1/3 of mice who received normal saline survived for the same period as the high-dose probiotic-treated group and this difference was significant (P=0.0280; Fig. 1A). In addition, a high-dose of B. subtilis administration significantly ameliorated weight loss compared with the normal saline group (P=0.0120; Fig. 1B). As DSS can promote intestinal damage, such as hematochezia and intestinal bleeding, the anuses and intestinal tracts of the three groups were examined. For mice in the high-dose B. subtilis-treated group, the anus and intestinal tract exhibited reduced bleeding and anabrosis than the control and low dose groups (Fig. 1C and D). Mice in the control and low-dose groups suffered more severe colon necrosis and shorter colons compared with the high-dose B. subtilis-treated group (Fig. 1E and F). These results indicate that high-dose B. subtilis administration alleviates DDS-induced colon damage and that B. subtilis induces a dose-dependent effect.

DSS induces UC by causing serious intestinal mucosal damage (19). In accordance with the findings mentioned, low or high doses of B. subtilis may attenuate the symptoms of DSS-induced UC. In light of this, it was hypothesized that B. subtilis may produce its effect by protecting the intestinal mucosa from damage and by reinforcing its repair. To test this hypothesis, alcian blue staining was conducted to determine how the intestinal mucosa reacted to the administration of B. subtilis following DSS treatment. As shown in Fig. 2A, mice treated with high-dose B. subtilis exhibited increased expression levels of mucins compared with the control group, which indicated repair of the colon mucosa. These results suggest that the high dose of the B. subtilis probiotic promoted the restoration of intestinal mucosa.

To further test this hypothesis, intestinal permeability was measured, as DSS-induced intestinal mucosa damage may lead to an increase in intestinal permeability. The results of the present study demonstrated that intestinal permeability was recovered in the B. subtilis-treated group (Fig. 2B). This further supports the hypothesis that treatment with B. subtilis may protect intestinal mucosa from DSS-induced damage and attenuate the inflammatory reaction.

The tight junction complex of the intestinal mucosa is considered to be the first ‘firewall’ for gut immunity (20). ZO-1 and claudins are two large families of the tight junction complex. To further explore whether B. subtilis was able to repair DSS-induced damage to the tight junctions, the intestines of mice treated with or without B. subtilis were harvested and probed with ZO-1 and claudins. The samples were then observed under a confocal laser scanning microscope, and the results demonstrated that the two tight junction-associated markers increased following high-dose B. subtilis treatment (Fig. 3A and B). This indicated that B. subtilis was involved in the repair process of DSS-induced mucosal damage and restored the mucosal tight junction complex.

Restoration of intestinal permeability may be related to the relief of the inflammation reaction (21). Therefore, the effect of B. subtilis administration on gut inflammation was explored using ELISA. An increase in plasma cytokines has previously been reported, including IL-12, IL-17 and IL-23, whereas IL-10 decreased in IBD (22). In the present study, the mean plasma levels of IL-12, IL-17 and IL-23 in the high-dose B. subtilis group were significantly reduced (P=0.0411, 0.0087 and 0.0152, respectively) and the mean IL-10 plasma levels were significantly increased (P=0.0450) (Fig. 4) compared with the NS group. However, low-dose B. subtilis treatment produced a less-marked effect.

Gut microbiota is a primary source of LPS endotoxin, which is a damage-associated pathogen that promotes gut inflammatory reactions and systemic inflammation (23,24). Plasma LPS of mice was tested and a difference in LPS was detected between the high-and low-dose B. subtilis-treated groups and the control group. These results indicated a significant decrease of LPS concentration in the probiotic group with the administration of high-dose B. subtilis compared with the control group (P<0.001; Fig. 5A).

SCFAs have anti-inflammatory functions via interaction with G protein-coupled receptor 43 and have been demonstrated to induce pro-inflammatory cytokines in various models of colitis (25,26). The concentration of total SCFAs was significantly higher in the high-dose B. subtilis group compared with the control group (P=0.0055; Fig. 5B), suggesting that high-dose B. subtilis was beneficial in maintaining SCFA content, which in turn reduced gut inflammation.

There are 100 trillion microorganisms housed in the human body. These microorganisms are maintained as commensals on the gut mucosa, and are associated with metabolism, including maintaining the internal environment and regulating the immune system (27). Consequently, 16s-rDNA sequencing analysis was performed in the present study to examine changes in the microbiota. As demonstrated in Fig. 6, it was observed that the gut microbiota was markedly altered in the high-dose B. subtilis-treated group compared with the control group. Specifically, a reduction of Acinetobacter sp., Ruminococcus sp., Clostridium spp. and Veillonella sp. was detected upon high-dose B. subtilis treatment, whereas levels of Bifidobacterium sp., Lactobacillus sp., and Butyricicoccus sp. were increased. Acinetobacter spp., Ruminococcus spp. Clostridium sp. and Veillonella sp. have been shown to be overrepresented in IBD patients, whereas Bifidobacterium spp., Lactobacillus spp., and Butyricicoccus sp. are decreased (28–32). These results demonstrated the role of B. subtilis in the amelioration of DSS-induced dysbiosis in the gut of model mice.