Two stains, B13 and B14, of C. butyricum were isolated from giant panda feces (Mixed samples of 4 adult giant pandas (two males and two females) were enriched in M17 medium and then screened for two strains using a special medium for Clostridium butyricum (RCM) depending on the amount of butyric acid produced by following a novel pipeline of culturomics (unpublished data). A commercial strain of C. butyricum (denoted CB) was used to compare its results with those of B13 and B14. An inoculum loop of each C. butyricum stored at −80°C was streaked on reinforced Clostridium medium (RCM) agar and incubated in a Bactron Anaerobic Chamber at 37°C under anaerobic conditions (89.9% N₂, 5.1% CO₂, and 5% H₂) for 48 h. A single colony of each strain was then anaerobically grown in RCM in the Bactron Anaerobic Chamber (89.9% N₂, 5.1% CO₂, and 5% H₂) at 37°C for 24 h and subcultured three times. The concentration of each strain was adjusted to 1 × 10⁹ CFU/mL, which was used to gavage the mice.

Specific pathogen-free female C57BL/6 J mice (6–8 weeks old, purchased from Chengdu Dashuo Animal Technology Co., Ltd.) were acclimated for 7 days (23°C ± 2°C, 50% ± 10% relative humidity). All experimental procedures were approved by the Cheng du Research Base of Giant Panda Breeding Institutional Animal Care and Use Committee (SCXK 202015) and performed in accordance with internationally accepted guidelines and ethical principles.

The preventive ability of C. butyricum was evaluated in DSS-induced ulcerative colitis in C57BL/6 J mice for 21 days (Figure 1). Mice were randomly divided into five groups: blank control (BC), disease control (DC), C. butyricum B13 (B13), C. butyricum B14 (B14), and C. butyricum CB (CB). Mice in BC and DC groups were orally administrated with PBS 200 μL/d and in groups B13, B14, and CB were orally administered with C. butyricum B13, B14, and CB (200 μL/d, 1 × 10⁹ cfu/mL) for 14 days, respectively (Figure 1). From day 14 to day 21, DSS (36000–5,000 Da, Yeasen Biotechnology, Shanghai, Co., Ltd) (3% w/v) was dissolved in drinking water to induce inflammation, except for BC. There were 8 mice in each group, and 4 mice were killed in each group before being fed with DSS (Figure 1).

Body weight was monitored throughout the study. Fresh feces were collected daily in sterile centrifuge tubes, snap-frozen, and stored at −80°C for 16S rRNA gene sequencing (Figure 1). DAI (Disease Activity Index) score is calculated based on this formula: DAI score = (body weight loss score + fecal status score + rectal bleeding score) / 3 (Supplementary Table S1). Mice were euthanized by spinal cord dislocation, and colon tissue and blood collected (Figure 1).

A segment of colon tissue located 1 cm from the mouse anus was rinsed with PBS. The colon portion was fixed using 4% paraformaldehyde for 48 h at 4°C. Fixed segments were paraffin-embedded, cut into 5-μm sections, and stained with hematoxylin and eosin (Morin et al., 2016). Sections were scanned using a digital slide scanner for image acquisition and histopathological analysis, following the criteria listed in Supplementary Table S2.

Inflammation was assessed by measuring IL-1β, IL-6, and TNF-α levels in the serum using commercially available ELISA kits (Shanghai Enzyme-linked Biotechnology Co., Ltd).

RNA was extracted from colon tissue using the SPINeasy RNA Kit for Tissue (MP Biomedicals) and was converted to cDNA using Exon Script RT Super Mix with dsDNase kit (Catalog no: A502-02). cDNA was used for transcriptome sequencing and to evaluate the expression of Cldn3, Ocln, Zo1, Zo2.

cDNA prepared from DC, B14, and CB colon tissues were selected for transcriptome sequencing. Paired-end libraries were prepared using an ABclonal mRNA-seq Lib Prep Kit (ABclonal, China), and library quality was assessed using an Agilent Bioanalyzer 4,150 system. Finally, the library preparations were sequenced on an Illumina MGISEQ-T7 and 150-bp paired-end reads were generated. Low-quality reads and any adapter sequences were removed as previously described (Zhao et al., 2013). High-quality reads were mapped to the reference genome¹ in orientation mode using HISAT2 software (Kim et al., 2015). The final efficiency of RNA-seq read alignments varied from 96.91 to 98.40% (Supplementary Table S3). Differentially expressed genes (DEGs; | log₂FC | > 1 and p_adj < 0.05) were analyzed using DESeq2 (Love et al., 2014). Gene ontology (GO) and pathway enrichment analysis of DEGs was performed using the “enricher” function in the “clusterProfler” package in R (Yu, 2018).

Fast SYBR Green qPCR Master Mix UDG (Catalog no: A402-01) was used for RT-PCR. Negative controls containing water as a template and the housekeeping gene Gapdh as an internal control were included in all RT-qPCR runs. The 2^−ΔΔCt method was used to calculate the relative gene expression based on Gapdh expression. The gene primer sequences of Cldn3, Ocln, Zo1, and Zo2 were found in PrimerBank and are shown in Supplementary Table S4. According to the instructions, the cycling conditions included: 1 cycle of 50°C for 120 s; 1 cycle of 94°C for 600 s; 1 cycle of 95°C for 180 s; 40 cycles of 95°C for 5 s and 60°C for 30 s.

Total DNA from the feces samples was extracted using a Magnetic Soil and Stool DNA Kit (TIANGEN). DNA concentration and purity were monitored on 1% agarose gels. V3 + V4 regions of 16S rRNA genes were PCR amplified using a pair of universal primers (Forward: CCTAYGGGRBGCASCAG, Reverse: GGACTACNNGGGTATCTAAT) tagged with 6-bp barcodes (Zhai et al., 2019). PCR conditions were 94°C for 4 min, followed by 30 cycles of 94°C for 30 s, 54°C for 30 s and 72°C for 30 s and then 72°C for 5 min. Single amplification was performed in 25-μL reactions with 50 ng of template DNA with the Phusion^®High-Fidelity PCR Master Mix (New England Biolabs). Sequencing libraries were generated using the TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina). Paired-end reads were obtained using the Illumina MiSeq PE-250 platform according to the standard protocols of Shanghai Applied Protein Technology Co., Ltd. Microbial raw sequences were analyzed using the QIIME2 pipeline (version 2021.2) (Bolyen et al., 2019). Paired-end reads were merged using FLASH (version 1.2.7) (Magoč and Salzberg, 2011). The DADA2 plugin was used to denoise and quality filter reads, and a feature table of amplicon sequence variants (ASVs) was prepared for subsequent analysis. After selecting the representative sequences of each ASV using QIIME 2 software, all representative sequences were aligned to the database for annotation against the SILVA reference database (version 138) (Pruesse et al., 2007) in QIIME2. Based on the ASV cluster analysis results and annotation information, the α diversity index, β diversity index, and species abundance at each taxonomic level were analyzed using Microbiome Analyst (Dhariwal et al., 2017).²

Differences between groups were analyzed via one-way analysis of variance using the IBM SPSS Statistics 27 software with least-squares deconvolution. Data are expressed as mean ± SEM. The LEfSe of the Huttenhower Galaxy Server³ was also used to compare the differences in gut microbiota among groups (p < 0.05; LDA score > 2.0).

At day 14, i.e., when no DSS were administered into mice, supplementation with C. butyricum had no significant effect on hematoxylin and eosin staining of colonic tissues (Supplementary Figure S1).

On day 21, after DSS had been administered for 7 days, significant damage to mouse colons was observed. The BC group (without DSS) had the lowest DAI (p < 0.05), the longest colon length (p < 0.001), and the lowest histological scores (p < 0.01) among the five groups (Figures 2, 3). In the absence of C. butyricum interventions, the DC group had the highest DAI score, lowest weight gain, shortest colon length, and highest histological scores among the five groups (Figures 2, 3). Among the three groups B13, B14, and CB with C. butyricum interventions, the CB group had a significantly lower DAI score and higher weight gain than that of the DC group (p < 0.05) but with no significant difference between B13 and B14 (Figures 2A,B). The B14 and CB groups had a significantly longer colon length than the DC group but did not significantly differ between each other (p > 0.05) (Figure 2C); the CB group had the lowest histological scores but was not significantly different from other groups (Figure 2D). Thus, C. butyricum could enhance the integrity of colonic mucosa and decrease tissue damage in DSS-induced mice; furthermore, administration of C. butyricum B14 and CB produced more protective effects than that of C. butyricum B13.

The serum levels of IL-6 and TNF-α was significantly higher following DSS treatment in the DC group than that in the BC group (p < 0.05); no significant difference was present between the DC and BC groups for IL-1β without the protection of C. butyricum (Figure 4). The C. butyricum intervention significantly decreased the serum levels of IL-6 in B14 and CB and TNF-α in the CB group compared with those in the DC group (p < 0.01 and p < 0.05, respectively), but serum IL-6, IL-1β, and TNF-α levels in the B13 group were not significantly decreased (Figure 4). The differences in IL-6, IL-1β, and TNF-α lacked significant difference among B13, B14, and BC groups, although the B13 group had a relative higher value than the other two groups (Figure 4). The levels of serum IL-6 and TNF-α in B13 were significantly higher than those in the BC group (p < 0.01), but the differences in the levels of IL-6 and TNF-α were not significantly different between BC and B14 or CB groups (Figure 4). The difference in the level IL-1β did not significantly differ among the five groups (Figure 4B). Thus, C. butyricum B14 and CB could significantly decrease the inflammation in DSS-induced colitis in mice.

The above results show that C. butyricum B14 and CB had favorable protective effects against DSS-induced colitis in mice. Therefore, we used transcriptomic profiling to analyze the impact of C. butyricum B14 and CB on colitis development. RNA extracted from colon tissue from adult male mouse was sequenced using DNBSEQ-T7. Sequencing generated 66.12G clean data for mouse samples. The sequences of each sample were the aligned to the genome sequence of the respective species. RNA-seq read alignment efficiency varied from 96.91 to 98.4% (Supplementary Table S5).

A total of 573 DEGs were identified in B14 vs. DC, of which 292 were upregulated and 281 were downregulated (Figure 5B). A total of 983 DEGs were identified in CB vs. DC, of which 352 were upregulated and 631 were downregulated (Figure 5C). Among these DEGs, 235 DEGs overlapped between the two groups (Figure 5A). The top 30 GO terms of these DEGs were related to the immune system, inflammatory, and metabolic activity (Supplementary Figure S2).

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed, and the top 20 KEGG pathways of DC vs. B14 showed that the downregulated pathways were related to immune and inflammatory processes, such as cell adhesion molecules, inflammatory bowel disease, and B cell receptor signaling pathway and the upregulated pathways were related to metabolic activity pathways, such as retinol metabolism, steroid hormone biosynthesis, metabolic pathways, and PPAR signaling pathway (Figures 5D,E). Similar to DC vs. B14, downregulated pathways in DC vs. CB were enriched in immune and inflammatory pathways, including the IL-17 signaling pathway, TNF signaling pathway, and chemokine signaling pathway, and upregulated pathways were enriched in metabolic activity pathways, including metabolic pathways, Drug metabolism-other enzymes, and PPAR signaling pathway (Figures 5F,G). In summary, C. butyricum B14 and CB could downregulate inflammation-related pathways and upregulate metabolism-related pathways to protect the host and decrease the inflammation level induced by DSS in mice models.

Among these DEGs, Tifa and Igkv12-89, which are involved in inflammatory and immune-related signaling pathways, and Snai2, which is related to the intestinal barrier, had a significantly lower expression in the B14 and CB groups (p < 0.05) (Figures 6A–C), and upregulated DEGs, including Muc4, Lama3, and Cldn4, are related to Nr1d1, which is associated with the intestinal barrier and is involved in immune-related signaling pathways (p < 0.05) (Figures 6D–G). These seven genes should be the key genes contributing to the decrease in inflammation in DSS-induced mice for C. butyricum B14 and CB.

The DEGs excluded the intestinal barrier-associated genes Cldn3, Ocln, Zo1, or Zo2; therefore, qPCR was subsequently used to detect the expression levels of these four genes to evaluate the mucosal barrier function in the intestine (Figures 6H–K). Expression levels of these genes in the DC group treated with DSS but not with C. butyricum intervention was significantly decreased compared with those in the BC group (p < 0.01) (Figures 6H–K). Compared with that in the DC group, intervention with C. butyricum B14 did not significantly increase the expression of these genes, whereas they were significantly increased with C. butyricum B14 treatment (p < 0.05); C. butyricum B13 only significantly increased the expression of Zo2 (p < 0.05) (Figures 6H–K). Thus, C. butyricum B14 had the greatest potential to decrease the damage induced by DSS in the colon.

16S rRNA amplicon sequencing was used to analyze the microbiota in the 40 fecal samples on day 14 and 21^t in the five groups (Supplementary Table S6). A total of 4,875,721 high-quality 16S rRNA gene sequences were obtained by quality filtering and paired-end sequence assembly, averaging 121,893 per sample, and ranging from 97,622 to 139,137 (Supplementary Table S6). These 4,875,721 reads were classified into 7,989 ASVs, belonging to 35 phyla (Firmicutes, Bacteroidetes, etc.) and 384 genera.

At the phylum level, the predominant bacterial communities were Firmicutes and Bacteroidetes in mice (Supplementary Figures S3A,C). At the genus level, before DSS administration, treatment with C. butyricum decreased the relative abundance of Alistips, especially in the B14 and CB groups, where the relative abundance of Alistips in B14_14 and CB_14 was lower than that in BC_14 (p < 0.05) (Figure 7A).

Following DSS administration, the relative abundance of Bifidobacterium and Faecalibaculum was significantly higher in B14_21 and CB_21 than in DC_21 (p < 0.05) (Figure 7F). In addition, the relative abundance of Dubosiella in the three experimental groups with C. butyricum intervention was significantly lower than that in DC_21 (p < 0.05) (Figure 7F). Moreover, the relative abundance of Actinobacteriota was significantly increased in B14_21 and CB_21 compared with that in DC_21 (p < 0.05) (Supplementary Figure S3).

The Venn diagram (Figure 7D) shows that 315 ASVs were shared between mice on day 14 before exposure with DSS and only 103 common ASVs were shared between mice after DSS administration on day 21 (Figure 7E). The α-diversity results showed no significant difference in the ACE, Chao1, Shannon, or Simpson indices among the five groups on day 14 (p > 0.05) (Supplementary Figures S3A,C). In contrast to BC_14, the gut microbiota richness did not significantly differ after the administration of C. butyricum (Supplementary Figure S3B). Compared with BC_21, there was a significantly decreased gut microbiota richness in the other four groups after DSS administration (p < 0.05) (Supplementary Figure S3D). Compared with DC, only the Simpson index was significantly higher in B13 (p < 0.05), and no differences were observed in the other three indexes (p > 0.05) (Supplementary Figure S3D). In addition, no differences were observed in ACE, Chao1, Shannon, and Simpson indices between B14 and CB (p > 0.05) (Supplementary Figure S3D).

The principal coordinate analysis (PCoA) based on weighted UniFrac metrics showed that the B14 and CB groups formed two distinct clusters and that the BC and B13 samples had a larger overlap than samples from the B14 and CB groups on day 14 (Figure 7B); the gut microbiota in BC_21 was significantly different from the other four groups while CB_21 and B14_21 clustered together but were divided with DC_21 (Figure 7C).

LEfSe showed that the abundance of class Clostridia and genus Akkermansia in B14_14 dramatically increased compared with that in BC_14 (Supplementary Figures S4A,G). The CB_14 sample had an increase in the relative abundance of Bacilli at the class level and Bifidobacterium and Faecalibaculum at the genus level compared with that in BC_14 (Supplementary Figures S4B,H). Compared with that in the BC_21, sample the relative abundance of Erysipelotrichaceae increased at the family level in DC_21 (Supplementary Figures S4C,I). The abundance of Dubosiella significantly increased at the genus level in DC_21 (Supplementary Figures S4C,J), and the abundance of Faecalibaculum and Bifidobacterium at the genus level and Bifidobacteriaceae at the phylum level significantly increased in B14_21 and CB_21 (Supplementary Figures S4D,E,I,F).

Inflammation can activate the immune response and prompt sustained activity of the adaptive immune system (Hespel and Moser, 2012). In this study, we found that the levels of inflammatory factors were downregulated following treatment, indicating that the immune response may be suppressed. Therefore, we analyzed the DEGs in the DC, B14, and CB groups and identified several genes related to immune response, including Tifa, Igkv12-89, and Nr1d1. TIFA (T2BP) is an inflammatory signaling adaptor that includes an FHA domain (Shen et al., 2015). TIFAs are important proteins that link the TNF signaling and NF-κB pathways that are a part of the proinflammatory stress response (Kanamori et al., 2002); moreover, they are involved in innate immunity induced by pathogen-associated molecular patterns (Li Y. et al., 2023). TNF-α–mediated signaling is attenuated when endogenous Tifa is knocked out. However, TIFA dimers bind with each other through intermolecular FHA-pT9 upon stimulation with inflammatory cytokines, such as TNF-α, resulting in TRAF6 oligomerization and subsequent NF-κB activation (Huang et al., 2023) (Figure 8). The NF-κB pathway is a common immune response pathway (Aggarwal, 2006). We found that Tifa expression was significantly downregulated after the administration of C. butyricum, demonstrating that inflammatory pathways were indeed inhibited. The KEGG enrichment results also supported this conclusion. Another related gene, Igkv12-89, is downstream of immunomodulation (Brüggemann et al., 2019). Igkv12-89 can be transcribed into IGKV12-89, which belongs to the immunoglobulin Kappa variable cluster (IGKV). During humoral immunity (induced by DSS or invasion of pathogenic bacteria), lymphocytes (e.g., macrophages) recognize and bind antigens, contact B cells to activate helper T lymphocytes, secrete interleukins and other cytokines, and together with T lymphocytes activate B cells to produce effector B cells that secrete IGKV (Catera et al., 2017) (Figure 8). Our results suggest that the downregulation of the signaling pathway upstream of TIFA may affect the downregulation of downstream Igkv12-89 expression. We found that the gene expression of a negatively regulated immune system is upregulated-Nr1d1 in our results. NR1D1 can negatively regulate the proinflammatory cytokine IL-6 in macrophages (Goswamy, 2022) (Figure 8), which is consistent with the reduced levels of IL-6 in our results. Zhao et al. (2019) show that C. butyricum can also reduce the level of the inflammatory factor IL-6. In addition, NR1D1 is a circadian clock component that integrates circadian rhythm and metabolism and plays a role in promoting metabolism (Everett and Lazar, 2014). According to KEGG analysis, most upregulated pathways are enriched in metabolism-related pathways, and we suspect that this may be related to the upregulation of Nr1d1. The effect of the above changes in these DEGs is consistent with the downregulation of the immune pathways shown in KEGG results. Studies have found that C. butyricum can inhibit the activation of pathways such as NF-κB and PI3K-Akt, thereby reducing the inflammatory response and protecting the intestine (Wang et al., 2022). Xie et al. (2020) research shows that after giving C. butyricum to mice with enteritis, its inflammation-related pathways are also significantly down-regulated, but mainly about IL-17 inflammatory factor-related pathways. Therefore, Tifa and Nr1d1 can be considered as therapeutic targets for inflammatory/immune diseases. Thus, C. butyricum B14 and CB treatment reduced the expression of proinflammatory factors. When the expression of Tifa is downregulated, the TNF-α–mediated signaling is attenuated, inhibiting the expression of inflammatory cytokines (Huang et al., 2023). The downregulation of IGKV expression also indicates that C. butyricum downregulates humoral immunity. Therefore, the downregulation of Tifa and lgkv12-89 and the upregulation of Nr1d1 all reflected a reduction in the immune response.

TNF-α and IL-6 can exacerbate intestinal inflammation by increasing colonic mucosal permeability, disrupting the intestinal tight junction protein barrier, and increasing intestinal penetration of tubular antigens, in addition to being involved in regulating inflammatory responses (Al-Sadi et al., 2009; Farré et al., 2020). Basement membrane laminin, macrophage transmigration, and associated loss of the intestinal tight junction barrier are key factors in the induction of enteritis (Nighot et al., 2021). Our transcriptome results show that Muc4, Lama3, and Cldn4, which are related to the intestinal barrier, were significantly upregulated and that the B14 strain had a stronger effect than the B13 strain. The normal intestinal surface is covered with a layer of mucus that comprises secretory mucin, which forms an external loose layer and an internal dense layer, and membrane-bound mucin, which forms a glycocalyx covering and protects the surface epithelial cells (McGuckin et al., 2011). These glycoproteins play an important role in protecting epithelial cells and are associated with epithelial renewal and differentiation. When intestinal inflammation occurs, the intestinal mucin barrier is disrupted, which aggravates intestinal damage. Mucin-4 (Muc4) is a large transmembrane glycoprotein composed of a heavily O-glycosylated extracellular α subunit and membrane-anchored β subunit, which usually acts to lubricate and protect the apical epithelial surface (Das et al., 2015; Rowson-Hodel et al., 2017). Mucin molecules may also promote mucosal repair and healing (Longman et al., 2000). Disruption of the basement membrane (BM) barrier function leads to intestinal inflammatory diseases. Laminin332 is an important constituent protein of the BM (Tayem et al., 2021). The protein encoded by lama3 is an α3 chain of laminin332 and plays a vital role in nucleation during hemidesmosome (HD) assembly and in maintaining the structural integrity of the HD (Schneider et al., 2007; Li Y. et al., 2020). When lama3 is mutated or missing, complete loss of laminin332 can co-occur in the BM and affect HD formation, which may eventually cause borderline epidermolysis bullosa and an inflammatory phenotype (Herrmann et al., 2021). The processed peptide laminin, which is a component of the extracellular matrix, protects epithelial tissue from pathogen attack and is part of the innate defense mechanism (Senyürek et al., 2010). Laminin α3 expression and processing are enhanced after infection and in chronic wounds, resulting in faster wound healing (Senyürek et al., 2014). Laminin can impact the microenvironmental response to inflammation in the intestine and likely participate in the regeneration process. In transgenic mice overexpressing laminin, the inflammatory response to DSS was attenuated, further suggesting that it had a protective effect against enteritis (Ljubimov et al., 2014). In our study, the expression of Lama3 was visibly elevated after the addition of C. butyricum, and C. butyricum B14 promoted the healing of damaged parts of enteritis. Lama3 has also been shown to be expressed in Crohn’s disease (Bouatrouss et al., 2000). Moreover, the overexpression of laminins can be involved in the occurrence of inflammatory bowel disease-related cancers (Ljubimov et al., 2014).

Transcriptome results showed that the expression of the intestinal tight junction membrane protein CLDN4 is increased under the action of C. butyricum B14. QPCR results showed increased expression of mRNA of four other intestinal tight junction proteins, including claudin-3, occludin, ZO-1, and ZO-2. The barrier function of the intestinal epithelium is closely related to the expression of these functional proteins (Liévin-Le Moal and Servin, 2013), indicating that C. butyricum B14 treatment promoted the repair of the damaged intestinal barrier. Moreover, Ma et al. (2022) showed that C. butyricum and its derived extracellular vesicles upregulate tight junction-related protein genes. Tight junctions act as physical barriers that prevent solutes and water from freely passing through the paracellular space between the epithelial or endothelial cell sheets. They also play a crucial role in maintaining cell polarity and signal transduction (Ma et al., 2018). This suggests that the administration of C. butyricum B14 can enhance the function of the colonic barrier and reduce intestinal permeability more prominently than C. butyricum CB. These results were consistent with those of several studies that revealed the enhancement in the expression of intestinal barrier genes upon the addition of C. butyricum (Ma et al., 2012; Mao Hagihara et al., 2019; Zhao et al., 2020).

In addition, we observed a significant decrease in Snai2 expression in the B14 and BC groups. SNAI2 (SNAIL2/Slug), encoded by Snai2, is an important regulator of epithelial-to-mesenchymal transition (EMT) and can directly act on e-cadherin (Serrano-Gomez et al., 2016). E-cadherin regulates cell-to-cell adhesion and maintains the structural and functional integrity of epithelial tissue (van Roy and Berx, 2008). Snai2 also plays an important role in intestinal barrier protection. SNAI2 plays a crucial role in cancer progression (Emadi Baygi et al., 2010; Guo et al., 2022). Several studies have shown that Snai2 expression is upregulated in enteritis (Zidar et al., 2016; Guo et al., 2023), which is consistent with our findings. When C. butyricum is added, the expression level of Snai2 is significantly inhibited, whereas that of e-cadherin tends to be normal, intercellular adhesion is enhanced, the EMT process is blocked, and colon fibrosis is reduced (Guo et al., 2023). Fibrosis causes the colon to shorten and is an important sign of enteritis (Zidar et al., 2016). Therefore, our results show that the increase in colon length and the decrease in histological index after the addition of C. butyricum may reflect the decrease in fibrosis.

The gut microbiota plays an important role in the host immune system, and an imbalance in the gut microbiota can disrupt the intestinal epithelial barrier and lead to the infiltration of inflammatory cytokines, thereby causing gut-related diseases (Neurath, 2019). Our results showed that C. butyricum significantly improved the gut microbiota, especially with strains B14 and CB (Figure 8).

At the genus level, after administration of C. butyricum, the relative abundance of Alistips decreased. Alistipes is considered a potential pathogen in the intestine, and may cause inflammation (Li et al., 2022). Similarly, Ma et al. (2022) showed that C. butyricum MIYAIRI II 588 could treat DSS-induced colitis by reducing levels of pathogenic bacteria in intestines. The relative abundance of Akkermansia increased in B14_14. Akkermansia stimulates mucus production and thickens the mucus layer to alleviate gut inflammation (Mithieux, 2018) and improves the intestinal barrier (Wang et al., 2020). This is consistent with the increased expression of Muc4 shown in the transcriptome results. In addition, the main increase in CB_14 was the relative abundance of Bifidobacterium and Faecalibaculum.

After DSS administration, the relative abundance of the phylum Erysipelotrichaceae and genus Dubosiella was significantly increased in the DC group. Erysipelotrichaceae is harmful to intestinal diseases (He et al., 2021). Studies have shown that Dubosiella can act as a probiotic that could prevent enteritis (Zhai et al., 2019; Wan et al., 2022) and that Dubosiella can play a role in improving obesity and antiaging (Liu et al., 2023).

Both the B14 and CB groups showed increased abundances of Bifidobacterium and Faecalibaculum. Bifidobacterium is an important component of the human and animal gut microbiota (Biavati and Mattarelli, 2015) and has biological barrier and nutritional effects, antitumor properties, and can enhance the immune system, improve gut function, combat aging (Bahmani et al., 2019; Din et al., 2020; Pizzo et al., 2020; Yao et al., 2021), and treat depression (Tian et al., 2020, 2022). Faecalibaculum also plays a role in regulating gut microbiota by promoting metabolism, preventing obesity, and acting as an antitumor agent (Ivanov et al., 2009; Zagato et al., 2020). Faecalibaculum can produce short-chain fatty acids (e.g., butyric acid) that inhibit the production of inflammatory factors, promote the production of mucins as well as the synthesis of tight junction proteins and antimicrobial peptides, and strengthen the intestinal barrier (Hu, 2023; Li L. et al., 2023). The downregulation of Tifa and Igkv12-89 and the upregulation of Muc4, Lama3, Cldn4, Cldn3, Zo1, Zo2, and Snai2 may be caused by these factors.