The candidate probiotic B. dorei BDX-01 was isolated in a stool specimen from a healthy human. The sample was inoculated onto brain heart immersion (BHI, Solarbio Life Sciences, Beijing, China) agar medium supplemented with 5% (v/v) defibrotic sheep blood (Solarbio Life Sciences, Beijing, China) and cultured in a 5% CO₂ anaerobic incubator (No. DG250; Don Whitley Scientific, Bingley, West Yorkshire, United Kingdom) at 37°C for 48 h. The isolated strain had morphophysiological characteristics resembling those of B. dorei, according to Bergey’s Manual of Systematic Bacteriology (Krieg et al., 2001). Genomic DNA was extracted from BDX-01 using a DNA extraction kit (Qiagen, Hilden, Germany). 16S rRNA genes were amplified by PCR using 27F (5′-AGA​GTT​TGA​TCC​TGG​CTC​AG-3′) and 1492R (5′-GGT​TAC​CTT​GTT​ACG​ACT​T-3′) as universal primers. Amplified products were sequenced, and 16S rRNA gene sequences were compared with those of various Bacteroides strains using BLAST software and the NCBI database (https://www.ncbi.nlm.nih.gov).

Total DNA was extracted from strain BDX-01 and the complete genome sequence was determined using an Illumina Hiseq 2,000 sequencing system by Biomarker Technologies. For genome assembly, The filtered reads were assembled by Canu v1.5 software, and then circlator v1.5.5 was taken to cyclizing assembly genome. For genome component prediction, Coding genes prediction was performed by Prodigal v2.6.3. The GenBlastA v1.0.4 program was used to scan the whole genomes after masking predicted functional genes. Putative candidates were then analyzed by searching for non-mature mutations and frame-shift mutations using GeneWise v2.2.0.

Putative virulence factors and antibiotic resistance genes were annotated by BLAST analysis with genes in the Virulence Factors of Pathogenic Bacteria Database (VFDB) (Chen et al., 2005) and Comprehensive Antibiotic Resistance Database (CARD) (Jia et al., 2017), based on a minimum of 50% and 35% amino acid homology, respectively. A BLAST comparison was performed between the genome sequences of strain BDX-01 and B. dorei strain 175 (GenBank accession number NZ_CP046176.1). The average nucleotide identity (ANI) was calculated using CJ Bioscience’s online Average Nucleotide Identity (ANI) calculator (https://www.ezbiocloud.net/tools/ani).

The antibiotic resistance breakpoints of B. dorei BDX-01 and the type strain B. dorei 175 were determined using 14 antibiotics with various antimicrobial mechanisms. The antibiotics included ampicillin, vancomycin, cefoperazone, and meropenem (cell wall synthesis inhibitors); clindamycin, clarithromycin, chloromycetin, tetracycline, streptomycin, and gentamicin (protein synthesis inhibitors); metronidazole, ciprofloxacin, and trimethoprim (nucleic acid synthesis inhibitors); and polymyxin B (cytoplasmic function inhibitor). Overnight 100 μL BDX-01 cultures (10⁷ CFU/mL) were co-cultured anaerobically in sterile 96-well plates (NEST, Wuxi, China) with antibiotics serially diluted between 0.125 and 512 μg/mL. OD₆₀₀ were determined with a microplate reader (Infinite200PRO; Tecan Austria GmbH, Grödig, Austria). The minimal inhibitory concentration (MIC) of different antibiotics was recorded. This metric indicates the lowest doses that inhibits 90% growth of the tested B. dorei strains (D’Aimmo et al., 2007). All experiments were performed in biological triplicate.

The Caco-2 human colon, IEC-6 rat intestinal epithelial and macrophage J774A.1 cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, United States). All cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, United States) supplemented with 10% (v/v) fetal bovine serum (FBS, Nobimpex, South America) at 37°C and under a 5% CO₂ atmosphere.

In vitro BDX-01 safety was determined in Caco-2 and IEC-6 cells using a lactate dehydrogenase (LDH) activity assay (Solarbio Life Sciences, Beijing, China). The Caco-2 and IEC-6 cells were inoculated in 96-well plates at a density of 5 × 10⁴/well and incubated at 37°C overnight. Each well contained 100 μL bacterial suspension at 5 × 10⁷ CFU/mL (MOI = 100) or 1 × 10⁸ CFU/mL (MOI = 200). After co-incubation at 37°C under a 5% (v/v) CO₂ atmosphere for 4 h, the LDH content was determined by microplate analyzer at 572 nm and the cytotoxicity was calculated as follows: c y t o t o x i c i t y % = s a m p l e – b a c k g r o u n d / t o t a l   L D H   r e l e a s e × 100

Total LDH was released by lysing the cells in 1% (v/v) TritonX-100.

Before in vitro anti-inflammatory assays, Caco-2 cells were seeded in six-well culture plates (NEST, Wuxi, China) and the medium was changed every other day for 14 days to allow complete differentiation. BDX-01 culture (1 × 10⁷ CFU/mL), GW4064 (10 μM), and TβMCA (100 μM) were added to the culture plates, and the suspensions were incubated at 37°C for 12 h. Culture suspensions were removed from each well and cells were then treated with 100 ng/mL TNF-α for 4 h. J774A.1 cells were seeded in six-well culture plates and incubated at 37°C overnight. GW4064 (10 μM) and TβMCA (100 μM) were added to the culture plates, and the suspensions were incubated at 37°C for 12 h. The culture suspensions were removed from each well and a new medium was added according to different groups. For BDX-01 treatment, the cells were treated with 1 μg/mL lipopolysaccharide (LPS) and BDX-01 (1 × 10⁷ CFU/mL) for 4 h. For LPS treatment, 1 μg/mL LPS was subjected to cells for 4 h. For the separation of different components of BDX-01, the intact bacterial suspension was directly preserved, and then part of the BDX-01 culture suspension was centrifuged at 9,000 rpm for 10 min to separate into bacterial precipitate and culture supernatant. The bacterial precipitate was heat inactivated at 100°C for 10 min, and then the BDX-01 culture supernatant (BDS) was filtered through a cellulose acetate filter with a pore size of 0.2 μm. The Caco-2 cells were pretreated with live BDX-01 (MOI = 1:100), heat-killed BDX-01 (MOI = 1:100), and BDS (50%) for 4 h and TNF-α (100 ng/mL) for another 4 h. After incubation, cells were collected for qRT-PCR.

The experimental protocol was approved by the Institutional Animal Care and Use Committee of Southern Medical University (approval number L2018225), and all in vivo experiments were performed under the guidelines of our institution for the use of laboratory animals. All SPF male C57BL/6J mice aged 6–8 weeks and 20–25 g in body weight were purchased from the Animal Center of the Southern Medical University, Guangzhou, China. To assess in vivo BDX-01 safety, mice were assigned randomly into two groups (n = 5–8/group). Each group was given either 5 × 10¹¹ CFU/d BDX-01 in phosphate-buffered saline (PBS, Solarbio Life Sciences, China) or 0.5 mL/d PBS by oral gavage for 5 days. General condition and body weight were monitored daily for 14 days after intragastric administration. At the end of the experiment, mice were sacrificed and blood samples were collected for hematological and serum biochemical analysis. The mice were anesthetized with 100 mg/kg pentobarbital sodium (Nanfang Hospital, Southern Medical University) and sacrificed by cervical dislocation. The stomach, colon, liver and spleen were weighed and subjected to histopathological examination.

For the acute colitis experiment, colitis induction was performed as previously described with slight modifications (Chassaing et al., 2014). Briefly, 2.5% (w/v) dextran sulfate sodium (DSS; MW = 36,000–50,000 Da; Macklin Biochemical Co., Ltd., Shanghai, China) was added to the drinking water of mice and administered continuously for 8 days. Changes in disease activity index (DAI) were assessed as previously described (Jang et al., 2019) according to the following criteria: (1) weight loss (%), (2) stool consistency, and (3) fecal blood (Supplementary Table S1).

The treatment groups were normal (Control), DSS model (Model), mesalazine (Mesalazine), and B. dorei (BDX-01), with eight mice randomly assigned per group. During the 7-day pre-intervention, mice in the BDX-01 group were administered by gavage a viable overnight culture of 1 × 10⁹ CFU/kg BW/d BDX-01. Simultaneously, the Mesalazine and Model groups were administered equivalent volumes of the culture medium by gavage. Except for those in the Control group, mice in each group received 2.5% (w/v) DSS in drinking water for 8 days. Starting 3 days before DSS intervention, the Mesalazine group was administered 200 mg/kg BW/d mesalazine (Macklin Biochemical Co., Ltd.) dissolved in PBS. Body weight and mental conditions were monitored throughout the experiment. The stool samples were collected on the day the experiment ended. On day 8 after DSS treatment, the mice were euthanized, and their colons and feces were collected.

Mouse gut microbiota were depleted as previously described with minor modifications (Ke et al., 2021). Briefly, for 4 days before the BDX-01/culture medium treatment and the colitis induction, the mice were administered by gavage an antibiotic cocktail (Abx) comprising 100 mg/kg/d vancomycin, 200 mg/kg/d metronidazole, 40 mg/kg/d gentamicin, and 200 mg/kg/d ampicillin. Feces were collected on day 5 and subjected to gel electrophoresis to assess the effects of antibiotics on depleting gut microbiota. Then, the Abx, Abx + DSS, and Abx + DSS + BDX-01 groups were designated, and eight mice were randomly assigned to each. The interventions used were consistent with those administered to the corresponding colitis groups described above.

The tissues were fixed with 4% paraformaldehyde and embedded in paraffin, cut into 4-μm sections, and stained with hematoxylin-eosin (HE) for pathological analysis. The histological scores were classified into one of the following groups: (1) loss of epithelium, (2) crypt damage, (3) depletion of goblet cells, or (4) infiltration of inflammatory cells (Supplementary Table S2) (Jang et al., 2019).

Total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, United States) according to the operating instructions. The mRNA was transformed into cDNA with reverse transcriptases (TaKaRa Bio Inc., Kusatsu, Shiga, Japan). Quantitative real-time PCR was performed using SYBR Premix Ex Taq (Takara, Japan) in a Roche Light Cycler 96 Real-time System (Roche Diagnostics, Basel, Switzerland). Each sample was run in triplicate. Each sample was detected in triplicate 10-μL reaction volumes and assessed using the comparative cycle (2^−ΔΔCT) method (Livak and Schmittgen, 2001). The results were normalized to the expression of GADPH and calculated relative to the control group. The primers used are listed in Supplementary Table S3.

Colon samples were collected and homogenized with a tissue breaker in 1× radioimmunoprecipitation assay (RIPA, Solarbio Life Sciences, Beijing, China) buffer with phenylmethylsulfonyl fluoride (PMSF, Solarbio Life Sciences, Beijing, China; 1:1,000). The mixtures were centrifuged at 15,000 × g and 4°C for 10 min. The supernatant was collected and the protein concentration was determined using a bicinchoninic acid (BCA) protein assay kit (Solarbio Life Sciences, Beijing, China). Proteins were separated by 12% (w/v) SDS-polyacrylamide gel and transferred to polyvinylidene difluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA, United States). The membranes were blocked for 1 h with 5% (v/v) skimmed milk and incubated at 4°C overnight with rabbit anti-caspase-1 monoclonal antibody (1:1,000 dilution; Cell Signaling Technology (CST), Danvers, MA, United States), rabbit anti-NLRP3 monoclonal antibody (1:1,000 dilution; Proteintech Group, Rosemont, IL, United States), rabbit anti-IL-1β monoclonal antibody (1: 1,000 dilution; Proteintech Group), rabbit anti-GSDMD monoclonal antibody (1:1,000 dilution; CST), rabbit anti-FXR monoclonal antibody (1: 1,000 dilution; Proteintech Group), rabbit anti-G-protein bile acid receptor-1 (TGR5) monoclonal antibody (1: 1,000 dilution; Proteintech Group), rabbit anti-vitamin D receptor (VDR, NR1H1) monoclonal antibody (1: 1,000 dilution; Proteintech Group), rabbit anti-pregnane-X-receptors (PXR, NR1H2) monoclonal antibody (1: 1,000 dilution; Proteintech Group), rabbit anti-FGF15 monoclonal antibody (1: 1,000 dilution; Proteintech Group) and rabbit anti-β-actin monoclonal antibody (1:10,000 dilution; ABclonal, Woburn, MA, United States). The membranes were then incubated with horseradish peroxidase (HRP)-labeled secondary antibodies at 25°C for 1 h. Protein bands were visualized with an enhanced chemiluminescence (ECL) kit (Fdbio Science, Hangzhou, Zhejiang, China). Band intensities were normalized relative to β-actin.

Bacteria genomic DNA was extracted from mice’ feces with a TIANGEN stool DNA kit (No. DP328; TIANGEN Biotech Co., Ltd., Beijing, China) according to operating instructions and diluted to 1 ng/μL. Fecal DNA was amplified with V4 primers 514F (5′-GTGCCAGCMGCCGCGGTAA- 3′) and 805R (5′-GGACTACHVGGGTWTCTAAT-3′) and a TIANGEN 2 × Taq PCR MasterMix (No. KT201; TIANGEN Biotech Co., Ltd.) according to operating instructions (Yin et al., 2015).

Total genomic DNA was extracted from the mice’s stool using the TIANamp Stool DNA Kit (Tiangen, Beijing, China). The V3–V4 region of the 16S rRNA gene was amplified using 338F and 806R primers and sequenced using an Illumina HiSeq 2,500 platform (Illumina, San Diego, CA, United States) by Majorbio BioPharm Technology Co., Ltd. (Shanghai, China). Paired end reads were merged via the rapid length adjustment of short reads, followed by sequence analysis using UPARSE (Edgar, 2013). Subsequently, sequences with ≥97% similarity were subsequently assigned to the same operational taxonomic units (OTUs) (Magoč and Salzberg, 2011). Then, representative sequences were selected using the QIIME software from each OTU, which were aligned and annotated based on the SILVA database (v138) (Caporaso et al., 2010).

BSH activity was determined by measuring the amount of amino acids released from conjugated bile salts, as previously described with minor modifications (Kumar et al., 2012). BSH activity of BDX-01 was measured in vitro, and Lactobacillus plantarum strain lp91 was used as a positive control. Fecal bacterial BSH activity was assayed according to CA production from fecal TCA. Briefly, protein extracts were prepared by sonicating 0.5 g of stool samples in 1 mL of PBS. The samples were centrifuged at 12,000 × g for 10 min and 1 mL of supernatant was collected and incubated with 1.8 mL of PBS and 0.1 mL of 0.1 M trichloroacetic acid (Shanghai Macklin Biochemical Co., Ltd., Shanghai, China) at 37°C for 30 min. Reactions were stopped by adding with 0.1 mL of trichloroacetic acid for 1 min. The mixtures were centrifuged and 1 mL of supernatant collected was mixed with 1 mL of 2 M trichloroacetic acid buffer and 1 mL of ninhydrin reagent [0.5 mL of 1% (w/v) ninhydrin in 0.5 M citrate buffer (pH 5.5), 1.2 mL of 30% (v/v) glycerol, and 0.2 mL of 0.5 M citrate buffer (pH 5.5)]. The mixture was vortexed to mix well, boiled for 15 min, and then cooled on ice. The taurine standard was then read for absorbance at 570 nm. Unit activity of BSH was expressed as the amount of enzyme that releases 1 mmol amino acid (AA)/min from the substrate.

Stock solutions were prepared by diluting the standard to 10 mM. Calibration standard solutions were prepared by stepwise dilution of the standard containing isotopically labeled internal standard mixture identical in concentration to the samples. Each stool sample (25 mg) was weighed and transferred to the Eppendorf tube (Eppendorf GmbH, Hamburg, Germany). Then 1,000 μL extract solution (2:2:1 acetonitrile-methanol-water with 0.1% (v/v) formic acid and isotopically labeled internal standard mixture) was added, and the samples were vortexed, sonicated in an ice bath, incubated, and centrifuged. The supernatants were transferred to liquid chromatography-mass spectrometry (LC-MS) vials for ultrahigh-performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) analysis by BIOTREE (Shanghai, China).

All data are displayed as mean ± standard deviation (SD). Data were analyzed using an unpaired t-test, Kruskal–Wallis test, Mann–Whitney U test, or one-way analysis of variance (ANOVA). A p-value of < 0.05 was considered statistically significant. Statistical analyses were performed using Prism v. 9.0.0 (GraphPad Software, Inc., La Jolla, CA, United States) and SPSS v. 19.0 for Windows (SPSS Inc., Chicago, IL, United States).

The bacterial strain B. dorei BDX-01 (BDX-01) was isolated from the stool of a healthy human and morphologically resembled B. dorei. Gram staining and direct microscopic observation revealed that this strain was Gram-negative and had a short rod shape with rounded ends. The 16S rRNA gene sequence analysis demonstrated that the strain shared 99% nucleotide sequence similarity with the B. dorei 175 strain. The complete genome was 5,521,745 bp in length and contained on a single chromosome, with an average GC content of 42.01%. BDX-01 and B. dorei 175 strain had an ANI of 99.97%. Hence, the isolate was a novel B. dorei strain (Figures 1A–C).

Before the colitis study, we assessed the virulence factor, antibiotic resistance, cytotoxicity, and animal toxicity of BDX-01 to confirm its in vitro and in vivo safety. According to the blast analysis of the VFDB database, 14 virulence factor homologs were identified in BDX-01 genome (Table 1), most of which correlate with cellular structure or physiological activities, and none have been reported as the pathogenesis of B. dorei. The antibiotic resistance genes identified in the genome of BDX-01 are summarized in Table 2. All drug-resistance genes were located on the chromosome rather than on plasmids. Based on the clinical resistance classification, its MIC was >32 μg/mL (D’Aimmo et al., 2007). It was sensitive to cefoperazone, meropenem, tetracycline, and ciprofloxacin and particularly sensitive to clindamycin and clarithromycin. The latter two antibiotics inhibited protein synthesis in BDX-01 (MIC = 0.125 μg/mL and 0.5 μg/mL, respectively) (Table 3).

Cytotoxicity assays using IEC-6 and Caco-2 cells showed no significant difference between the treatment and control groups (p > 0.05). Therefore, intact viable BDX-01 cells were not cytotoxic in vitro (Figures 2A, B). In the acute toxicity experiments, neither death nor treatment-related toxicity occurred in mice. Body weight did not differ significantly between the groups (Figure 2C). No histopathological injury to the stomach, colon, liver, and spleen was found in both groups (Figure 2D). In addition, hematologic analysis revealed no significant differences in blood routine indices and hepatic and renal function between the groups (Supplementary Table S4).

We investigated the effects of BDX-01 on a mouse DSS-induced colitis model (Figure 3A). DSS administration significantly reduced body weight, shortened colon length, and increased the DAI score in mice. Weight loss (Figures 3B, C) and colon length reduction (Figures 3D, E) were significantly reduced in the Mesalazine and BDX-01 groups compared with the Model group. Furthermore, BDX-01 treatment markedly reduced the DAI score of mice (Figures 3D, E). Hence, BDX-01 treatment improved the symptoms of DSS-induced acute colitis.

HE staining revealed inflammatory cell infiltration, crypt loss, and mucosal injury in the colon tissue of the Model group (Figure 3F). In contrast, the Mesalazine and BDX-01 groups displayed nearly complete colon tissue integrity. The histopathological scores of the Model group (8.00 ± 0.71) were significantly increased and significantly decreased in the Mesalazine (5.40 ± 0.89) and BDX-01 (4.80 ± 0.84) groups (Figure 3G). Therefore, BDX-01 treatment alleviated DSS-induced colon histopathological damage.

Dysbiosis of the intestinal microbiota is characteristic of UC. Therefore, we investigated whether BDX-01 modulates the composition of the gut microbiota and confers protection against UC. Alpha-diversity was not significantly different among the four groups based on their Shannon indices (Figure 4A). The beta-diversity was evaluated based on a principal coordinate analysis (PCoA) of the weighted UniFrac distance. The gut microbiota of DSS-treated mice differed significantly from that of control mice. However, the administration of BDX-01 ameliorated the shift in the gut microbiota induced by DSS challenge. In contrast, mesalazine had no apparent significant effect on gut dysbiosis (Figure 4B).

We then used a Wilcoxon rank test to compare the four groups in terms of their bacterial distributions at various taxonomic levels. The relative abundances of Firmicutes and Actinobacteriota were significantly lower, and that of Desulfobacterota was significantly higher in the Model and Mesalazine groups than in the Control group. The abundances of Verrucomicrobiota was significantly higher in the Mesalazine group. The abundances of Firmicutes and Actinobacteriota were significantly higher in the BDX-01 group than in the other groups (Supplementary Figure S2).

In the family-level (Supplementary Figure S3), the abundances of Erysipelotrichaceae, Lactobacillaceae, and Akkermansiaceae were lower, and those of Bacteroidaceae, Rikenellaceae, Oscillospiraceae, Prevotellaceae, Ruminococcaceae, and Desulfovibrionaceae were higher in the Model and Mesalazine groups than in the Control group. Erysipelotrichaceae in the BDX-01 group and Akkermansiaceae in the Mesalazine group were markedly higher than those in the Model group.

In the genus-level, the relative abundances of Dubosiella and Lactobacillus were significantly lower in the Model group than in the Control group. However, mesalazine treatment significantly increased the relative abundance of Akkermansia and Lachnoclostridium. Supplementation with BDX-01 notably increased the relative abundances of Dubosiella and Faecalibaculum (Supplementary Figure S4). Dubosiella, Faecalibaculum_rodentium, and Bacteroides_dorei were higher in the BDX-01 group than in the Model group (Figures 4C, D). Thus, BDX-01 alleviates colitis and regulates gut microbiota composition.

Changes in BA composition and metabolism have been linked to UC. A fecal bile acid analysis revealed that BDX-01 treatment significantly altered the fecal BA profile (Figure 5A). Enzyme activity assays in patients with IBD showed impaired BA deconjugation due to decreased BSH activity (Duboc et al., 2013). BDX-01 exhibited high BSH activity in vitro (Supplementary Figure S5). Gut microbiota analyses revealed that Firmicutes, Actinobacteria, and Erysipelotrichaceae have high BSH activity (Jones et al., 2008; Labbé et al., 2014). These taxa were significantly increased in response to the BDX-01 treatment (Figure 5B). An assay of fecal samples confirmed that BDX-01 significantly increased intestinal BSH activity (Figure 5C). Compared to the Control group, the Model group presented significantly decreased total fecal BA level and unconjugated primary bile acid (UPBA):conjugated primary bile acid (CPBA) and SBA:PBA ratios. Therefore, intestinal BA excretion, deconjugation, and 7α-dehydroxylation were altered in the Model group (Figures 5D–F). The CA/TCA and β-MCA/Tβ-MCA ratios and DCA level were significantly reduced in the Model group but could be reversed by BDX-01 treatment (Figures 5G–J). Spearman rank correlation analysis (Figure 5K) showed significant positive correlations between fecal CA/TCA and the relative abundances of Faecalibaculum_rodentium, Dubosiella, and Bacteroides_dorei. βMCA/TβMCA, DCA, and LCA were positively correlated with the aforementioned bacteria. In contrast, BAs were negatively correlated with the abundances of Bacteroides_acidifaciens, Bacteroides_sartorii, and Alistipes in the Model group.

Elevated fecal CA/TCA and β-MCA/Tβ-MCA ratios and DCA in response to BDX-01 treatment may regulate host metabolism and immunity by acting on different bile acid receptors. The various bile acids activate several nuclear receptors in the gut: the TGR5 that is activated by LCA and DCA; the PXR that are activated by CDCA, LCA, and DCA; and the VDR that is activated by LCA and DCA, in addition to vitamin D (Fiorucci et al., 2022). Western blotting analysis confirmed that BDX-01 only upregulated FXR among the receptors (Figure 5L; Supplementary Figure S6). These results suggest that BDX-01 may alleviate colitis by regulating the bile acid-FXR pathway. Detailed relative BA alterations in feces are shown in Supplementary Table S5.

We examined the proinflammatory cytokines of the colon tissue to evaluate the effects of BDX-01 on the inflammatory response. DSS administration significantly upregulated the proinflammatory cytokines TNF-α, IL-1β, and IL-6 compared with the control. In contrast, they were markedly downregulated in the Mesalazine and BDX-01 groups (Figure 6A). The expression of the anti-inflammatory cytokine IL-10 was downregulated and relatively upregulated in the DSS and BDX-01 groups, respectively (Figure 6A).

The FXR agonism exert anti-inflammatory effects by inhibiting inflammatory signaling pathways, including the NF-κB signaling pathway and the NLRP3 inflammasome signaling pathway (Gadaleta et al., 2011; Hao et al., 2017). The mRNA and/or the protein expression levels of NF-κB pp65, NLRP3, ASC, caspase-1, gsdmd, and IL-1β were significantly increased in the Model group colon tissues. Administration of BDX-01 and mesalazine markedly reversed the overexpression of the NLRP3 inflammasome signaling pathway but did not inhibit NF-κB activation (Figures 6A, B; Supplementary Figure S7). Thus, BDX-01 could modulate the NLRP3 inflammasome signaling pathway and alleviate inflammation.

Mice were given an oral antibiotic (Abx) to deplete the gut microbiota and determine their roles in alleviating colitis with BDX-01 (Figure 7A). All microbiota were effectively depleted after oral antibiotic cocktail administration for 4 days (Supplementary Figure S1). As expected, mice in the Abx + DSS group showed symptoms of colitis (weight loss and elevated DAI score) and colonic histopathological damage compared with those in the Abx group. Body weight loss (Figure 7B), DAI score (Figure 7C), colon length (Figures 7D, E), and histopathology score (Figures 7F, 7G) were markedly decreased in the Abx + DSS + BDX-01 group compared with the Abx + DSS group. Fecal BSH activity did not differ significantly among the three groups (Figure 7H). However, treatment with BDX-01 upregulated FXR and its downstream factor FGF15 and downregulated NLRP3 and IL-1β compared with the Abx + DSS group (Figure 7I; Supplementary Figure S8). These data indicate that, in addition to regulating the gut microbiota, BDX-01 mitigates DSS-induced colitis by directly activating the FXR-NLRP3 pathway.

Enterocytes secrete cytokines and chemokines that promote macrophage and immunocyte infiltration into intestinal inflammatory sites (Kagnoff and Eckmann, 1997). In active UC, NLRP3, and IL-1β are expressed mainly in lamina propria immune cells (Ranson et al., 2018). We determined the anti-inflammatory effects of BDX-01 on intestinal cells and macrophages and function of FXR. We subjected human intestinal epithelial Caco-2 cells and mouse J774A.1 macrophages to TNF-α and LPS, respectively, with or without BDX-01, the FXR agonist GW4064, and the FXR antagonist T-βMCA. In response to the TNF-α/LPS treatment, Nlrp3 and Il-1β mRNAs were dramatically upregulated compared with the control. Neither BDX-01, GW4064, nor T-βMCA alone affected Il-1β expression in intestinal cells (Figures 8A, B) or macrophages (Figures 8E, F). We confirmed the transcription of Fxr in the intestinal cells in response to BDX-01 treatment compared to GW4064 and TβMCA (Figures 8C, D). BDX-01 significantly downregulated Nlrp3 and Il-1β mRNA induced by TNF-α and LPS and blocked by TβMCA (Figures 8A–F). Therefore, FXR activation counteracts the expression of the NLRP3 inflammasome and proinflammatory factor IL-1β expression in intestinal cells and macrophages.

We co-treated Caco-2 cells with TNF-α, BDX-01 culture supernatant (BDS), and live and heat-killed bacterial cells to determine the anti-inflammatory effects of the BDX-01 components. Whereas BDS treatment upregulated Fxr and downregulated Nlrp3 and Il-1β, the bacterial cells had no such effects (Figures 8G–I).