The kimchi sample used in this study (Daesang Corporation, Seoul, Republic of Korea) was serially diluted in de Man-Rogosa-Sharpe broth medium (MRS; Becton, Dickinson and Company, Sparks, MD, United States). Thereafter, aliquots of each serial dilution were spread on MRS agar plates, which were then incubated at 30^°C for 2 days. The colonies thus obtained were analyzed using 16S rRNA gene sequencing as previously described (Lim et al., 2017). Subsequently, Lactiplantibacillus (Lp.) plantarum DSR J266 and Levilactobacillus (Lv.) brevis DSR J301 isolated from the kimchi were cultured in MRS broth at 30^°C for 2 days and harvested via centrifugation at 8,000 g and 4^°C for 15 min. Thereafter, the harvested cells were washed three times with phosphate-buffered saline (PBS), followed by the resuspension of the two strains in PBS to a final concentration of 5 × 10⁸ colony forming units (CFU)/100 μL and equal mixing for administration via gavage. Lacticaseibacillus (Lb.) rhamnosus GG was isolated and prepared using the same process, and the final concentration was adjusted to 1 × 10⁹ CFU/200 μL for administration via gavage.

Six-week-old wild-type male C57BL/6N mice were purchased from Orient Bio Inc. (Seongnam, Republic of Korea). The animals were housed (2–3 mice per individually ventilated cage) in a pathogen-free animal facility at the World Institute of Kimchi (Gwangju, Republic of Korea, Republic of Korea), and maintained at 22 ± 2^°C and 55 ± 5% relative humidity under a 12-h/12-h light-dark cycle. After the animals were acclimatized for 1 week before the experiments, they were randomly divided into the following six diet groups (n = 5 per group, Figure 1): normal diet and vehicle (NC); normal diet and Lp. plantarum DSR J266 + Lv. brevis DSR J301 (NL); normal diet and Lb. rhamnosus GG (NG); alcohol diet and vehicle (AC); alcohol diet and Lp. plantarum DSR J266 + Lv. brevis DSR J301 (AL); and alcohol diet and Lb. rhamnosus GG (AG). Mice were fed a nutritionally adequate Lieber-DeCarli liquid diet (LDC; Doo Yeol Biotech, Seoul, Republic of Korea) throughout the study period (Lieber and DeCarli, 1982). Alcohol liver injury was induced in mice using the modified NIAAA model (Bertola et al., 2013; Guo et al., 2018). Further, mice in the NC, NL, and NG groups were fed the LDC liquid diet for 15 days, whereas those in the AC, AL, and AG groups were fed the LDC liquid diet containing alcohol. During the first 5 days, the alcohol concentration in the diet was gradually increased to 4.5%. This was to allow the animals adapt to the alcohol diet. Subsequently, the LDC liquid diet containing 4.5% alcohol was fed to the mice in the AC, AL, and AG groups for the remaining 10 days. During the same period, mice in the NL and AL groups were gavaged daily with Lp. plantarum DSR J266 and Lv. brevis DSR J301, while those in the NG and AG groups were gavaged daily with Lb. rhamnosus GG. Furthermore, mice in the AC and NC groups were gavaged daily with PBS (vehicle). The gavage was always performed in the morning (9–10 am), and in the afternoon, the feeding tubes and LDC liquid diet were replaced with new ones (4–5 pm). The calories in the normal and alcohol liquid diets were adjusted using maltose dextrin (Guo et al., 2018). On day 16, mice in the alcohol diet groups (AC, AL, and AG groups) were administered a single alcohol binge (5 g kg^–1 body weight, 20% alcohol), whereas mice in the normal diet groups (NC, NL, and NG) were gavaged with isocaloric maltose dextrin. After 9 h, when serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and inflammatory cytokines peaked (Bertola et al., 2013), the all mice were euthanized and their blood, tissue, and fecal samples were collected.

Serum was separated from the collected blood samples via centrifugation at 3,000 g for 15 min at 4°C. Thereafter, serum AST and ALT levels were determined using a FUJI DRI-CHEM NX500iVC biochemical analyzer with their respective FUJI DRI-CHEM slide kits (FujiFilm Corporation, Tokyo, Japan). In addition, serum levels of inflammatory cytokines, including interleukin (IL)-6, IL-10, monocyte chemotactic protein (MCP)-1, interferon (IFN)-γ, tumor necrosis factor (TNF)-α, and IL-12p70, were detected using the BD™ Cytometric Bead Array kit and the BD FACSCanto II flow cytometry system (BD Biosciences, San Jose, CA, United States).

The collected liver tissue samples were fixed in 10% formalin at 25^°C for 24 h followed by embedding in paraffin, sectioning, mounting on glass slides, and staining with hematoxylin and eosin (H&E). Thereafter, representative photomicrographs were captured using an Olympus BX41 microscope (Olympus, Tokyo, Japan) at 100 × magnification. Further, to analyze the captured images, Olympus CellSens Standard software version 1.6 was used.

The total genomic DNA corresponding to the collected fecal samples was extracted via mechanical and chemical lysis, and purified using a Mo Bio PowerSoil DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, CA, United States) according to the manufacturer’s instructions. Thereafter, the quality of the DNA was measured via PicoGreen assay, performed using a NanoDrop UV spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States). Next, genomic DNA was amplified using primers targeting the V3 to V4 hypervariable regions of the 16S rRNA gene (S-D-Bact-0341-b-S-17, 5′-CCTACGGGNGGCWGCAG-3′; S-D-Bact-0785-a-A-21, 5′-GACTACHVGGGTATCTAATCC-3′) (Herlemann et al., 2011; Klindworth et al., 2013). Paired-end sequencing was further performed by Macrogen Inc. (Seoul, Republic of Korea), using an Illumina MiSeq platform (Illumina Inc., San Diego, CA, United States). The generated sequencing reads were analyzed using QIIME2 software version 2019.10 pipeline (Bolyen et al., 2019), which utilizes 16S rDNA sequences in the Silva v132 database (Quast et al., 2013; Yilmaz et al., 2014), and to predict the functional profiles of the mouse gut microbial community based on the 16S rRNA gene dataset, PICRUSt software (Langille et al., 2013) was used. The 16S rRNA gene sequencing data were deposited in the National Center for Biotechnology Information (NCBI) GenBank under accession no. SRX12860986.

All the results for qualitative data were presented as the mean ± standard error of mean (SEM). The serological characteristics that differed significantly between the diet groups were determined using one-way ANOVA and Tukey’s HSD post-hoc test (p ≤ 0.001). The level of significance was set at *p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001.

All animal procedures were performed according to the National Institute of Health Guidelines for the Humane Treatment of Animals, with approval from the Institutional Animal Care and Use Committee of the World Institute of Kimchi (WIKIM IACUC 201936). The animals were sacrificed via forced CO₂ inhalation and all efforts were made to minimize suffering.

To investigate the effect of LAB strains (Figures 2A,B) on the degree of alcohol-induced liver damage, the serum ALT and AST levels of mice in the different treatment groups were measured (Bertola et al., 2013; Torruellas et al., 2014). Thus, we observed that in the alcohol-fed mice, AST levels were approximately 2-fold ALT levels. Previous studies have reported that an AST/ALT ratio of 1.5 or 2.0 indicates alcoholic hepatitis (Cohen and Kaplan, 1979; Alves et al., 1981; Correia et al., 1981). Further, compared with the normal diet groups, AST and ALT levels were significantly higher in the alcohol diet groups (Figures 2A,B; ANOVA, p < 0.05); this is consistent with the results of a previous study (Bertola et al., 2013). We also observed that the normal diet groups showed no significant differences with respect to AST and ALT levels. The differences in AST and ALT levels among the alcohol diet groups could be attributed to the presence, absence, or type of LAB ingested. Further, compared with the AC group, AST and ALT levels in the AL group were significantly decreased (ANOVA, p < 0.05). These results indicated that LAB treatment in the normal diet groups did not affect AST and ALT levels, whereas LAB treatment in the alcohol diet groups decreased AST and ALT levels.

To further investigate the effects of LAB treatment on ALD mice, the levels of inflammatory cytokines (IL-6, IL-10, MCP-1, IFN-γ, TNF-α, and IL-12) in sera samples from mice in the different treatment groups were analyzed. Thus, IL-6, IL-12, and MCP-1 were identified as the main inflammatory cytokines in the samples (Figures 2C–E); IL-10, IFN-γ, and TNF-α were not detected in any of the samples. Further, IL-6 was not detected in samples from the normal diet groups, but was observed in samples from the alcohol diet-fed mice (Figure 2C). The average IL-6 concentration corresponding to the AC group was higher than that corresponding to the AL and AG groups, indicating a decrease in inflammatory cytokine levels following LAB treatment. Reportedly, IL-6 is a representative inflammatory cytokine, and its level is elevated in animal models of alcohol-related liver injury (Latvala et al., 2005). Furthermore, the AL and AG groups displayed relatively low IL-6 levels, suggesting that LAB treatment protected these mice from alcohol-induced hepatocyte damage. Moreover, the levels of proinflammatory cytokines, IL-12 and MCP-1, were relatively higher in the alcohol diet groups than in the normal diet groups (Figures 2D,E), indicating that IL-12 and MCP-1 may be closely involved in the progression of alcoholic liver injury.

Liver tissue samples from mice in the different treatment groups were subjected to H&E staining to confirm ALD induction and the accumulation of neutral lipid (triglyceride) droplets (Figure 3). Mice in the normal diet groups (the NC, NL, and NG groups) showed healthy liver tissues, whereas those in the alcohol diet groups showed fat accumulation in their livers. However, the AL and AG groups exhibited relatively less liver fat accumulation than the AC group. These results suggested that although LAB intake did not affect fat accumulation in the livers of mice that consumed a normal diet, it prevented fat accumulation in the hepatic tissues of alcohol diet-fed mice (Cai et al., 2019; Hsieh et al., 2021).

The bacterial 16S rRNA gene amplicon obtained from mice fecal samples was sequenced to confirm differences in gut bacterial community composition owing to alcoholic liver injury and LAB intake. A total of 3,776,969 sequencing reads for bacterial 16S rRNA genes were generated from the 30 collected fecal samples. After removing low-quality sequences, noise, unmerged sequences, chimeras, and singletons, 2,240,092 high-quality bacterial reads were obtained, with an average of 74,670 reads per sample (Supplementary Table 1). Further, we observed that the bacterial alpha-diversity indices of the alcohol diet groups was lower than those of the normal diet groups, based on the ACE/Chao1 indices (Figures 4A,B) and Shannon-Simpson indices (Figures 4C,D) obtained (Fairfield and Schnabl, 2020). Nevertheless, the decrease in bacterial diversity in the AC group was alleviated by LAB treatment based on the diversity indices obtained for the AL and AG groups. These results suggested that alcohol consumption reduced the microbial diversity in the intestine, which however, was restored by LAB treatment. The beta-diversity of the bacterial community was analyzed using weighted and unweighted UniFrac distances (Figures 4E,F), which revealed a significant difference between the normal and alcohol diet groups (PERMANOVA, p < 0.05). In addition, the distances between normal diet group plots were similar, whereas the alcohol diet group plots were widely separated, confirming that the alcohol diet induced greater differences in microbial diversity between the treatment groups than the normal diet.

To further investigate the effect of LAB intake on alcohol-induced liver injury in mice, gut microbiome composition was analyzed. Thus, we observed that the microbiomes of the alcohol and normal diet groups differed substantially (Figure 5 and Supplementary Figure 1). The abundances of the genera Akkermansia and Monoglobus were relatively high in the normal diet group, but low in the alcohol diet groups. However, the abundances of the genera Escherichia-Shigella and Enterococcus were relatively higher in the alcohol diet groups compared with their abundances in the normal diet groups (Fairfield and Schnabl, 2020). We also observed that the abundances of the genus Escherichia-Shigella and Enterococcus differed depending on the LAB treatment. Specifically, Escherichia-Shigella and Enterococcus were the most dominant genera in the AC group, but showed decreased abundances in the AL and AG groups. Reportedly, species of the family Enterobacteriaceae, to which the genus Escherichia-Shigella belongs, are prevalent in the intestine of patients with alcoholic cirrhosis (Chen et al., 2011; Bajaj et al., 2012; Tuomisto et al., 2014). Most of the sequences classified as the genus Enterococcus were confirmed to be from Enterococcus faecalis ASV1 (Figure 6). Further, Enterococcus faecalis ASV1 was present at a very low frequency in the samples from the normal diet groups; however, it was significantly increased in the AC group (Figure 6), and in the AG group Enterococcus faecalis ASV1 was significantly decreased to the level observed for the normal diet groups.

To investigate differences in microbial metabolism following the LAB treatment of normal and alcohol diet fed mice, metabolic functions were predicted using PICRUSt based on the obtained 16S rRNA gene sequencing data (Figures 7A,B) and KEGG BRITE metabolic categories. Thus, we observed substantial differences in gene abundance for each metabolic category according to the treatment diets. Further, the gene abundance according to the type of LAB administered was comparable among the normal diet groups, but considerably differed among the alcohol diet groups. Specifically, genes associated with protein families (i.e., signaling and cellular processes, membrane transport, and signal transduction) were present in relatively high proportions in the alcohol diet groups than in the normal diet groups. In contrast, genes associated with amino acid metabolism, translation, nucleotide metabolism, and replication and repair categories were present in relatively low proportions in the alcohol diet groups compared with the normal diet groups. Additionally, genes associated with the metabolism and biosynthesis of alanine, aspartate, glutamate, arginine, cysteine, methionine, glycine, serine, threonine, histidine, lysine, phenylalanine, tyrosine, tryptophan, valine, leucine, and isoleucine, were present in relatively lower proportions in the alcohol diet groups than in the normal diet groups (Figure 7B). The alteration of intestinal microflora by alcohol administration also affected the abundance of aminoacyl-tRNA biosynthesis-related genes (Figure 8). Specifically, the abundance of genes related to aminoacyl-tRNA biosynthesis was similar among the normal diet groups, but decreased in the alcohol diet groups; however, their abundance recovered slightly after LAB treatment. Decreased expression of aminoacyl-tRNA biosynthesis-related genes may affect amino acid synthesis by intestinal bacteria, but can be recovered via LAB treatment. Previous studies have confirmed that acute ALD can be alleviated by regulating the pathways related to aminoacyl-tRNA biosynthesis (Yi et al., 2021). Positive and negative correlation analyses between gut bacterial community composition and the predicted metabolic categories were performed to identify the bacteria involved in microbial metabolism (Figure 9). Thus, we observed that the frequency of amino acid metabolism genes showed negative correlation with Escherichia-Shigella and Enterococcus, which was relatively predominant in the alcohol group. However, Akkermansia, which was dominant in the normal diet group, showed a significant positive correlation.