The Bacillus strain used in the present study was B. subtilis (CMCC 1.3358), which was commercially available from China General Microbiological Culture Collection Center. For experimental purposes, single colonies of the Bacillus strain were grown in LB broth at 37˚C overnight with shaking (200 rpm). Next, overnight cultures were diluted (1:100) in LB broth and left to grow at 37˚C with shaking (200 rpm) until OD600 was reached 0.5-0.6 (~2-3 h). Bacillus strains were then pelleted by centrifugation (6,000 x g) at room temperature for 2 min, diluted in Fresh PBS solution and mixed to obtain the appropriate bacterial density (B. subtilis: 5x10⁸ CFU/ml).

The animal experiments in the present study were approved by the Ethics and Clinical Research Committee of Nankai University (Project IRM-DWLL-2016121). A total of 40 male C57BL/6J mice (20-22 g; 4-6 weeks old) used in the present study were purchased from the Beijing Hfk Bioscience Co., Ltd. Animals had a sanitary status of SPF and were housed in regulation cages (22±1˚C, relative humidity of 50±10% and 12-h light/dark cycle) with ad libitum access to food and water. All the mice were subdivided into groups (10 mice in each group): Group 1, Control (Ctrl); group 2, ethyl alcohol (EtOH) group; group 3, ethyl alcohol and B. subtilis supplementation (EtOH+BS), group 4, ethyl alcohol and PBS (EtOH+PBS). The alcohol-induced liver injury model was constructed in mice using the Lieber-DeCarli diet as previously described (31). In brief, all mice were allowed to acclimatize to the laboratory conditions for 2 days and to a liquid diet for 5 days, and two mice were housed in each cage. During the acclimatization to liquid diet, prophylactic supplementation with B. subtilis or PBS were performed daily. Following the acclimatization, the alcohol-fed mice were fed with a diet containing 3% (vol/vol) ethanol for 10 days, while the Ctrl mice received an isocaloric amount of maltodextrin as pair-fed control. The mice in the EtOH + BS group were administered bacterial suspension in PBS (200 µl) by gavage every day. Body weights were measured every other day. On day 16, the mice received a single dose of ethanol via oral gavage (5 g/kg body weight), followed by a 9 h-fast, and were sacrificed by anesthesia with isoflurane (4%) prior to excision of tissue samples.

To determine the intestinal permeability, fluorescein isothiocyanate (FITC)-dextran (4 kDa; Sigma-Aldrich; Merck KGaA) was orally administered (600 mg/kg body weight) to mice 4 h before sacrifice. Blood samples were collected quickly and subsequently centrifuged (1,500 x g, 4˚C, 15 min) to prepare serum. Fluorescence was recorded using a spectrophotometer (Tecan Group, Ltd.) at an excitation wavelength of 485 nm and emission wavelength of 528 nm.

Serum aspartate transaminase (AST) and alanine transaminase (ALT) were measured using the Infinity reagent (cat. nos. TR71121 and TR70121; Thermo Fisher Scientific). Hepatic triglyceride (TG) levels were measured using the Triglyceride Liquid Reagents kit (cat. no. BC0625, Beijing Solarbio Science & Technology Co., Ltd.), according to the manufacturer's protocols. Hepatic lipid peroxidation was quantified by measuring malondialdehyde (MDA) using Micro MDA assay kit (cat. no. BC0025, Beijing Solarbio Science & Technology Co., Ltd.), and the oxidative stress of liver tissues was characterized by testing the decreased glutathione content using a reduced glutathione (GSH) assay kit (cat. no. BC1175, Beijing Solarbio Science & Technology Co., Ltd.). Serum LPS was determined using the Mouse Lipopolysaccharides (LPS) ELISA kit (cat. no. CSB-E13066m, Cusabio). The fecal IgA levels were quantified using a Mouse sIgA ELISA kit (cat. no. CSB-E08413m, Cusabio). The expression level of IL-6 in serum was determined using a Mouse IL-6 ELISA kit (cat. no. SEKM-0007, Beijing Solarbio Science & Technology Co., Ltd.). Serum levels of total cholesterol (TCH), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) were detected colorimetrically by mindray 2000 M autoanalyzer according to the manufacturer's protocols.

Total liver and colon RNA were extracted using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.), and the total RNA concentration was quantified using the NanoPhotometer N50 (Implen NanoPhotometers), and cDNA was synthesized at 42˚C for 2 min on an Mastercycler nexus PCR using PrimeScript RT reagent kit with gDNA Eraser (Takara Bio, Inc.). The synthesized cDNA was stored at -20˚C, and RT-qPCR was conducted on a LightCycler^® 96 system (Roche Diagnostics) using a TB Green Premix Ex Taq II (Tli RNaseH Plus; Takara Bio, Inc.). The thermocycling conditions were as follows: Initial denaturation for 10 min at 95˚C, then 40 cycles of 10 sec at 95˚C, 10 sec at 62˚C and 10 sec at 72˚C, followed by 95˚C for 60 sec and dissociation curve analysis. The primer sequences are listed in Table I. Relative gene expression was normalized to the 18S and calculated by the 2^-ΔΔCt method (32).

Total protein was extracted from livers using a RIPA buffer containing 1% protease inhibitor and 1% phenylmethylsulfonyl fluoride (cat. no. R0010; Beijing Solarbio Science & Technology Co., Ltd.) and was measured by the BCA assay. Total protein (50 µg per lane) was separated by 10% SDS-polyacrylamide gel, transferred onto a polyvinylidene difluoride membrane (Whatman plc), and blocked with 5% skimmed milk at room temperature for 1 h. Membranes were immunostained with primary antibodies against TLR4 (cat. no. 14358; dilution, 1:1,000; CST) at 4˚C overnight, as well as antibody against reference protein β-actin (cat. no. K200058M; dilution, 1:1,000; Beijing Solarbio Science & Technology Co., Ltd.). Following incubation with a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (cat. no. SE134; dilution, 1:5,000; Beijing Solarbio Science & Technology Co., Ltd.) was carried out at room temperature for 1 h. Immunoreactive proteins were stained with ECL Western Blotting Substrate (Beijing Solarbio Science & Technology Co., Ltd.). Images were digitalized using the ChemiDoc™ XRS system (Bio-Rad Laboratories, Inc.). β-actin (Beijing Solarbio Science & Technology Co., Ltd.) was used as an internal standard.

For the histological analysis, the liver and colonic tissues were stained with hematoxylin and eosin (H&E). In brief, tissues were fixed in 10% formalin at room temperature for 24 h, and paraffin-embedded sections (5 µm) were stained with hematoxylin (10 min) and eosin (30 sec) at room temperature.

Total genomic DNA was extracted from 150-200 mg of stool samples using the QIAamp PowerFecal DNA kit (Qiagen GmbH). The hypervariable V3-V4 region (341F and 805R) of the the prokaryotic 16S rRNA gene was amplified. Following purification, these amplicons were equally combined and then subjected to a sequencing library preparation according to the manufacturer's protocols. DNA concentration and quality were determined by the NanoPhotometer N50 (Implen NanoPhotometers) and Qubit Fluorometer (Thermo Fisher Scientific, Inc.). Deep DNA pyrosequencing procedures were performed on a paired-end Illumina MiSeq PE300 (2x300 bp) platform at the Novogene Corporation, according to the manufacturer's protocols. These raw sequences were processed following the QIIME (v1.9.1) pipeline (33), and the analysis of gut microbiota diversity and composition of fecal samples was determined, of which Shannon indexes and Simpson indexes were calculated with the operational taxonomic unit (OTU) table (34). The β diversity of fecal microbiota was assessed by principal coordinate analysis based on Bray-Curtis distance.

All experimental results were obtained from at least three independent times, and representative data are presented as the mean ± standard deviation. One-way analysis of variance was used to determine whether the groups were statistically different (P<0.05). GraphPad Prism 7.0 (GraphPad Software, Inc.) was applied to all statistical analysis.

To investigate the effect of the prophylactic dietary B. subtilis supplementation on alcohol-induced injury, a murine model was constructed as shown in Fig. 1A. Notably, alcohol feeding caused continuous body weight loss in the EtOH group compared with the Ctrl group; however, dietary B. subtilis supplementation notably prevented alcohol-induced body weight loss (Fig. 1B). Following these mice being sacrificed, it was observed that ethanol feeding significantly increased the murine liver weight and the liver-body weight ratio compared with Ctrl mice, and dietary B. subtilis supplementation significantly decreased the liver weight (Fig. 1B and C). Notably, it was found that supplementation with PBS-bacterial vehicle did not cause notable effects in EtOH-fed mice (Fig. 1B-D); therefore, the effects of PBS supplementation in EtOH-fed mice were excluded from further experiments. In conclusion, dietary B. subtilis supplementation likely alleviated alcohol-induced injury.

The development of alcohol-induced liver injury (EtOH group, Fig. 2A) was confirmed, as indicated by notably increased levels of ALT, AST and hepatic triglycerides compared with the Ctrl group (Fig. 2B-E). In addition, it was investigated whether dietary B. subtilis supplementation may relieve alcohol-induced liver injury in mice. Microscopic morphological examination confirmed the protective effect of dietary B. subtilis supplementation with a decrease in clear vacuoles and neutrophil infiltration in H&E-stained images (Fig. 2A). Dietary B. subtilis supplementation significantly decreased hepatic triglyceride levels in treated mice compared with EtOH-fed mice (Fig. 2C). Additionally, the hepatic dysfunction induced by alcohol was improved as demonstrated by decreased serum ALT and AST levels (Fig. 2B and E), as well as the partial decrease in serum TCH and LDL-C level (Fig. S1). Furthermore, hepatic oxidative stress was induced by alcohol consumption with increased levels of hepatic thiobarbituric acid-reactive substances (TBARS) in the EtOH mice (1.53±0.07 µmol/g liver) compared with the Ctrl mice (1.03±0.09 µmol/g liver), and dietary B. subtilis supplementation significantly decreased hepatic MDA levels to 1.17±0.06 µmol/g liver (Fig. 2D). GSH depletion was another trait of ethanol-induced liver injury, and the administration of B. subtilis notably increased the content of reduced GSH (Fig. S2). Consistent with these protective effects, dietary B. subtilis supplementation significantly decreased the expression of peroxisome proliferator-activated receptor-γ (PPAR-γ) and fatty acid translocase (CD36; Fig. 2F and G), which likely ameliorated the hepatic lipid dysfunction induced by alcohol.

Previous studies have characterized gut barrier dysfunction in alcohol-induced endotoxemia and liver damage (35-37), and disordered intestinal integrity may be due to the direct injury induced by alcohol and indirect inflammatory reactions (38,39). In the present study, alcohol consumption resulted in a significant increase in intestinal permeability. Specifically, serum FITC concentrations increased to 1,183±76.18 ng/ml from 439.2±23.53 ng/ml in Ctrl mice (Fig. 3A), and evidence of colon damage was noted through microscopic morphological examination of H&E-stained samples (Fig. 3B). However, dietary B. subtilis supplementation significantly restored intestinal permeability to a certain degree, decreasing FITC concentrations to 786.8±45.8 ng/ml (Fig. 3A). In addition, the damaged gut barrier resulted in the translocation of LPS from lumen to blood, and dietary B. subtilis supplementation significantly decreased LPS levels to 57.43±3.35 ng/ml from 88.04±7.6 ng/ml in EtOH mice (Fig. 3C). Consistent with amelioration of intestinal barrier dysfunction, dietary B. subtilis supplementation significantly restored Muc2, Zo-1 and occludin expression levels in the colon (Fig. 3D). Additionally, alcohol consumption notably damaged immunoglobulin A (IgA) production and secretion with fecal IgA levels decreased by 24% in the EtOH group compared with the Ctrl group, and dietary B. subtilis supplementation significantly restored the level of IgA secretion (Fig. 3E). Furthermore, dietary B. subtilis supplementation notably reversed the alcohol-induced decrease in Reg3g and Reg3b mRNA expression to levels equivalent to those noted in the Ctrl group (Fig. 3D), thereby improving intestinal barrier function.

Alcohol consumption increases gut bacteria-derived PAMPs (e.g., LPS) in the bloodstream, and PAMPs trigger hepatic inflammation, which contributes toward the development of alcohol-induced liver diseases (21). In the present study, alcohol consumption increased LPS levels in the serum, which stimulated TLR4 to induce the release of critical proinflammatory cytokines that are required to activate potent immune responses. Alcohol-induced increases in TLR4 protein levels were accompanied by significant upregulation of the mRNA levels of TLR4 target genes (Fig. 4A-C), including NF-κB, tumor necrosis factor alpha (TNF-α), IL-6 and monocyte chemoattractant protein-1 (MCP-1), indicating activation of the alcohol-induced hepatic TLR4 pathway. However, dietary B. subtilis supplementation significantly decreased hepatic TLR4 protein levels by 53% (P<0.01) with concomitant significant decreases in TLR4-regulated mRNA levels of NF-κB, TNF-α, IL-6 and MCP-1 (P<0.01) in the treated group, compared with the EtOH group (Fig. 4C).

It has been reported that dysbiosis of the gut microbiome and metabolome is associated with the progression of alcohol-induced liver injury (5,40,41). The present study characterized the protective effects of dietary B. subtilis supplementation on alcohol-induced perturbation of intestinal microbiota composition. OTU-based principal coordinate analysis revealed that the gut microbiota structures of each tested group exhibited individual variations and that the experimental groups significantly differed from the Ctrl group (Fig. 5A). The diversity of gut microbiota in EtOH-fed mice compared with Ctrl mice was significantly decreased as evidenced by decreased Shannon indexes and increased Simpson indexes; however, dietary B. subtilis supplementation reversibly increased microbiota diversity (Fig. 5B). Analysis of the gut microbiota at phylum levels revealed a significantly altered composition of gut microbiota in Ctrl, EtOH and EtOH + BS groups (Fig. 5C). The phyla Firmicutes, Bacteroidetes and Proteobacteria dominated the fecal microbial communities (Fig. 5C), and alcohol consumption significantly increased the abundance of Proteobacteria and decreased that of Firmicutes (Fig. 5D-F); however, dietary B. subtilis supplementation markedly decreased the level of Proteobacteria to that noted in the Ctrl group (Fig. 5F). Notably, further analysis at the family level revealed a marked decrease in Lachnospiraceae in EtOH-fed mice (Fig. 5E), while dietary B. subtilis supplementation decreased this trend. Members of the Lachnospiraceae family are important butyrate producers that reside in the intestinal microbiota (42). Furthermore, in the EtOH groups, the relative increase in Bacteroide abundance and decrease in Ruminococcaceae_UCG-013 abundance at the species level were reversed by dietary B. subtilis supplementation (Fig. 5G). In summary, dietary B. subtilis supplementation partly alleviated alcohol-induced intestinal microbiota dysbiosis.