Clostridium Scindens Protects Against Vancomycin‐Induced Cholestasis and Liver Fibrosis by Activating Intestinal FXR‐FGF15/19 Signaling

When mice were pretreated with vancomycin for 3 weeks and followed by pBDL operation for 14 days (Figure1A), we observed a decrease in liver/body weight ratio and increases in serum alanine aminotransferase (ALT) and aspartate transferase (AST), while no difference was noted in serum ALP levels (Figure 1B,C). Increased necrosis of liver cells and collagen deposition in vancomycin‐treated mice were as further evidenced by increases in Sirius red positive area and Masson's trichrome staining positive area (Figure 1D). Consistent with this, we also found evaluated mRNA expression of fibrosis‐associated genes (Acta2, Col1a1, Timp1, and Tgf‐β) in vancomycin‐treated mice, as well as an increase in α‐SMA protein levels (Figure 1E,F). Vancomycin‐treated mice also exhibited increased mRNA levels of genes that related to hepatic inflammation, such as Il‐1β, and Il‐6 (Figure 1G). These results established that vancomycin promotes hepatic collagen deposition and hepatic stellate cell (HSC) activation in the pBDL‐induced mouse model.

To further confirm the contribution of vancomycin treatment to cholestasis and fibrosis, we built up another PSC mouse model using a 1% CA diet. Mice were administrated with water containing 0.5 g L⁻¹ vancomycin or regular water for 6 weeks, with free access to a regular diet or a diet supplemented with 1% CA (Figure S1A, Supporting Information). In keeping with those findings we observed 1% CA diet‐fed mice with vancomycin treatment exhibited increased ALT, AST, and ALP levels in serum (Figure S1B, Supporting Information). Moreover, body weight and liver weight were decreased in 1% CA diet‐fed mice with vancomycin treatment (Figure S1C, Supporting Information). We also found increased collagen synthesis and deposition in 1% CA diet‐fed mice with vancomycin treatment, as evidenced by increased Sirius red positive area and protein level of α‐SMA (Figure S1D,E, Supporting Information). 1% CA diet‐fed mice with vancomycin treatment exhibited increased hepatic mRNA expression of Tnf‐α (Figure S1F, Supporting Information).

The characterization of PSC is the abnormal accumulation of bile acids. Therefore, bile acid profiles were detected using HPLC‐MS/MS. Compared to control mice, we observed dramatically increased total and primary bile acids in the liver from vancomycin‐treated mice, along with a decrease in secondary bile acids (Figure2A). Notably, the primary/secondary bile acid ratio was dramatically increased in the liver of vancomycin‐treated mice compared to control mice (Figure 2B). The Vancomycin treatment markedly increased primary bile acids (CA, TCA, β‐MCA, T‐β‐MCA, CDCA, TCDCA, UDCA, TUDCA) in liver tissue (Figure S2A, Supporting Information). Conversely, most of the secondary bile acids (ω‐MCA, T‐ω‐MCA, HDCA, THDCA, DCA, NorDCA, TDCA, LCA, 7_ketoLCA, ACA) were decreased in the liver from vancomycin‐treated mice (Figure S2B, Supporting Information). Additionally, we also observed increased total bile acid, as well as increased T‐α‐MCA, β‐MCA, T‐β‐MCA, CDCA, TCDCA, and TUDCA in serum from vancomycin‐treated mice, while decreases in ω‐MCA, DCA and TDCA (Figure S2C–E, Supporting Information). Consistent with changes observed in mice treated with vancomycin, the bile acid profiles of serum from CLD patients also showed increased total and primary bile acids, along with increased primary/secondary bile acid ratio, compared to healthy controls (Tables S1–S2, Supporting Information).

Vancomycin is a glycopeptide antibiotic, that primarily targets gram‐positive bacteria. Therefore, we also investigated bile acid concentrations in feces. Consistent with the alterations observed in hepatic bile acids, we found a trend toward increased total bile acids and significantly increased primary bile acids in feces from vancomycin‐treated mice (Figure 2C). The primary/secondary bile acid ratio was increased in the feces of vancomycin‐treated mice compared to control mice (Figure 2D). As anticipated, we observed elevated levels of TCA, α‐MCA, and T‐β‐MCA in feces from vancomycin‐treated mice, while levels of ω‐MCA, HDCA, DCA, NorDCA, LCA, 6_ketoLCA and 7_ketoLCA were decreased (Figure S2F,G, Supporting Information).

Given the increased bile acid accumulation in vancomycin‐treated mice, we examined the most important enzymes for bile acid synthesis. As expected, we observed increased hepatic mRNA levels of Cyp7a1, Cyp8b1, and Cyp7b1 in vancomycin‐treated mice (Figure 2E), as well as an increase in the protein level of CYP7A1 (Figure 2F). It is well‐established that hepatic FXR negatively regulates the transcription of Cyp7a1 through a small heterodimer partner (SHP). Additionally, we observed a trend toward decreased hepatic mRNA level of Fxr and significantly decreased hepatic mRNA level of Shp in vancomycin‐treated mice (Figure S3A, Supporting Information).

The balance of the bile acid pool relies on both synthesis and transport rate, so we assessed the expression level of major bile acid transporters in the liver and ileum. We observed decreases in hepatic mRNA levels of bile acid transporters, including Oatp1a1, Mdr2, Mrp3, and Ost‐α, in vancomycin‐treated mice (Figure S3B, Supporting Information). The mRNA expressions of bile acid transporters in the ileum, such as Asbt, Ibabp, and Ost‐α, were also decreased in vancomycin‐treated mice (Figure S3C, Supporting Information). These findings indicate that vancomycin exacerbates bile acid accumulation by enhancing bile acid synthesis rate and inhibiting bile acid transport rates. Antibiotic treatment has been reported to be involved in intestinal and liver diseases by regulating multiple bile acid‐dependent transcriptional factors, including FXR, VDR, PXR, and so on.^[8^] We only observed decreased mRNA levels of Fxr and downstream gene Fgf15 in the ileum from vancomycin‐treated mice (Figure 2G), while no difference was noted in mRNA expression of Lxr, Pxr, Rorα, Tgr5, and Vdr (Figure S4, Supporting Information). We also observed a decrease in the protein levels of FGF15 in the ileum from vancomycin‐treated mice, compared to control mice (Figure 2H). Consistently, FGF15 levels in serum were decreased in vancomycin‐treated mice (Figure 2I). We also found enhanced bile acid accumulation and inhibition of intestinal FXR‐FGF15/19 signaling by vancomycin administration in the absence of a disease model (sham operation) (Figure S5, Supporting Information). To further verify the effect of vancomycin on the progression of CLD, we conducted a pBDL operation and subsequently treated it with vancomycin or vehicle. Consistent with previous findings, this model demonstrated increased collagen deposition and bile acid accumulation in mice exposed to vancomycin following disease induction (Figure S6, Supporting Information). Although vancomycin treatment resulted in increased fecal and hepatic bile acid concentration, intestinal barrier function was unaffected, as evidenced by no difference in pathological examination and the expression of tight junction proteins (Figure S7, Supporting Information). These findings demonstrate that vancomycin inhibits ileum FXR‐FGF15/19 signaling, thereby promoting bile acid synthesis through increased hepatic CYP7A1 expression.

Consistently, we observed dramatically increased total bile acid levels in the serum of vancomycin‐treated mice fed with a 1% CA diet (Figure S8A, Supporting Information). The livers of 1% CA diet‐fed mice with vancomycin treatment also exhibited decreased mRNA expression of Fxr in liver tissue (Figure S8B, Supporting Information). The protein levels of CYP7A1 were increased in the livers and the protein levels of FGF15 were decreased in the ileum from 1% CA diet‐fed mice with vancomycin treatment (Figure S8C,D, Supporting Information). These results further implicate that vancomycin aggravates cholestasis and fibrosis via inhibiting intestinal FXR‐FGF15/19 signaling and increasing hepatic CYP7A1 expression.

Gut microbiota is essential to maintain bile acid homeostasis, mainly relying on the conversion of primary bile acids to secondary bile acids. Furthermore, altered gut microbiota has been implicated in the pathogenesis of PSC. Feces from vancomycin‐treated and control mice were analyzed by 16S rDNA. We observed lower alpha diversity in vancomycin‐treated mice than in control mice (Figure3A). The results of 16S rRNA sequencing revealed decreased abundances of Clostridium XIVα, Barnesiella, and Bacteroides in vancomycin‐treated mice, while Klebsiella, Escherichia/Shigella, Mucispirillum, and Morganella were increased (Figure 3B). Our study aims to identify potential probiotics involved in vancomycin‐induced cholestatic liver injury, with Clostridium XIVα, Barnesiella, and Bacteroides considered as candidate species. The gut microbiota participates in bile acid homeostasis by converting primary bile acids into secondary bile acids, with Clostridium XIVα recognized as the primary bacterial cluster involved in this metabolic pathway.^[9^] However, both Barnesiella and Bacteroides, as fiber‐degrading bacteria, primarily play roles in the intestinal barrier, immunity, and brain health.^[10^] Therefore, we next investigated whether changes induced by vancomycin in Clostridium XIVα correlate with bile acids, Spearman correlation analysis was conducted. We found that the hepatic concentration of T‐α‐MCA, CDCA, MuroCA, THDCA, TDCA, 7_ketoLCA, α‐MCA, NorDCA and DCA has significantly positive correlations with the abundance of Clostridium XIVα, while hepatic concentration of LCA, HDCA, T‐ω‐MCA, TCDCA, CA, TUDCA, TCA and ACA has significantly negative correlations with it. In addition, the fecal concentration of 7_ketoLCA, UDCA, LCA, DCA, NorDCA,6_ketoLCA, HDCA, ω‐MCA, and α‐MCA had significantly positive correlations with the abundance of Clostridium XIVα (Figure 3C). This correlation may be attributed to the 7‐alpha‐dehydroxylation enzyme activity in Clostridium XIVα. Clostridium scindens (C. scindens) is a predominant bacterium in Clostridium XIVα, known for its high levels of bile acid 7‐dehydroxylating activity.^[11^] qPCR results showed decreased levels of C. scindens in vancomycin‐treated mice and CLD patients, compared to controls (Figure 3D,E). We also analyzed the correlation between C. scindens and serum bile acid concentrations. The abundance of C. scindens was negatively correlated with serum bile acid levels in CLD patients (Figure 3F). These findings suggest that vancomycin treatment alters gut microbiota, especially decreasing the abundance of C. scindens, which may contribute to bile acid accumulation and liver fibrosis.

To confirm whether vancomycin aggravates bile acid accumulation and liver fibrosis by decreasing the abundance of C. scindens. C57BL/6J mice were exposed to water containing 0.5 g L⁻¹ vancomycin for 3 weeks, followed by daily gavage of C. scindens or vehicle, and then administrated to pBDL operation (Figure4A). Consistent with previous results, vancomycin treatment led to elevated serum ALT and AST levels, and C. scindens gavage reduced the vancomycin‐induced ALT and AST levels in serum (Figure 4B). Histologic analysis revealed decreased collagen synthesis and deposition in the C. scindens group, with decreases in Sirius red positive area, and Masson's trichrome staining positive area (Figure 4C). Consistently, we also found decreased mRNA levels of Acta2, Col1a1 and Tgf‐β after gavage of C. scindens, along with decreased α‐SMA protein level (Figure 4D,E). Taken together, these results indicate that restoration of C. scindens alleviates vancomycin‐induced hepatic fibrosis.

As a 7α‐dehydroxylation bacteria, C. scindens could convert primary bile acids to secondary bile acids, we next determined bile acid profiles in the liver and feces from C. scindens group mice and control mice. Compared with mice in the vancomycin group, we observed decreased total and primary in the liver from mice in the C. scindens group, with a decrease in the primary/secondary bile acid ratio (Figure5A–C). 7α‐hydroxy‐4‐cholesten‐3‐one (C4) is used as a biomarker for the rate of bile acid synthesis.^[12^] As expected, we observed a trend to increase in hepatic C4 levels in the vancomycin group, and gavage of C. scindens had a trend to reduce the C4 level (Figure 5D). Concentrations of five tauro‐conjugated bile acids were decreased in the liver of mice treated with C. scindens, including T‐α‐MCA, T‐β‐MCA TCA, TCDCA, and TUDCA (Figure 5E). We observed elevated levels of total taurine‐conjugated bile acids, accompanied by an increased ratio of taurine to glycine‐conjugated bile acids (Figure 5F). Consistently, hepatic mRNA levels of bile acid CoA: amino acid N‐acyltransferase (BAAT) and bile acid‐CoA: amino acid N‐acyltransferase (BACS), in the livers of mice treated with vancomycin compared to control mice. However, these increases were reversed in mice treated with C. scindens (Figure 5G). Our in vitro experiments showed that Caco‐2 cells treated with varying concentrations of TCA, TCDCA, and T‐β‐MCA for 48 h exhibited decreased mRNA levels of Fxr, Shp, and Fgf15 in a dose‐dependent manner (Figure S9A—C, Supporting Information). Additionally, LX2 cells treated with TCA showed increased mRNA expression of Acta2, Col1a1, and Tgf‐β (Figure S9D, Supporting Information). These distinct bile acids, identified in mice treated with C. scindens, inhibited intestinal FXR‐FGF15/19 signaling and promoted HSC activation in vitro. Considering that C. scindens limited bile acids accumulation, we next determined the expression of enzymes for bile acids synthesis. Gavage of C. scindens reduced the protein level of CYP7A1 compared to the vancomycin group (Figure 5H). As we precious showed, vancomycin treatment inhibited intestinal FXR‐FGF15/19 signaling. We next tested whether the gavage of C. scindens could affect the intestinal FXR‐FGF15 axis. Our results showed that the protein level of FGF15 was increased in mice with C. scindens gavage, compared to mice in the vancomycin group (Figure 5I). We also observed that vancomycin treatment resulted in decreased serum FGF15 levels, whereas C. scindens gavage increased serum FGF15 levels (Figure 5J). These results implicate that restoration of C. scindens alleviates vancomycin‐induced bile acid synthesis and accumulation by activating intestinal FXR‐FGF15/19 signaling and inhibiting the expression of hepatic CYP7A1.

It is well‐established that the 7‐α‐dehydroxylation strain shares the BaiCD, BaiE, BaiF, BaiG, BaiH, and BaiI genes, which encode enzymes responsible for converting primary bile acids to secondary bile acids.^[13^] Site‐directed mutagenesis of BaiE from C. scindens confirmed that baiE is the rate‐determining enzyme in this pathway.^[14^] We constructed engineered bacteria EcN‐BaiE and investigated its function in PSC model mice treated with vancomycin (Figure6A,B). Mice administrated with EcN‐BaiE stains exhibited decreased serum ALT, AST, and total bile acid (TBA) levels, compared to the EcN group (Figure 6C,D). Similar observations were made with C. scindens transplantation, gavage of EcN‐BaiE stains also resulted in reduced protein expression of CYP7A1 in the liver (Figure 6E). Histologic analysis revealed reduced collagen synthesis and deposition in the EcN‐BaiE group, as evidenced by decreases in Sirius red positive area and Masson's trichrome staining positive area, compared to the EcN group (Figure 6F). Consistently, we also found decreased protein levels of α‐SMA after administration of EcN‐BaiE stains (Figure 6G). All these findings demonstrate that bile acid 7α‐dehydratase, which is expressed in C. scindens, leads to the inhibition of hepatic CYP7A1 expression, thereby attenuation of liver fibrosis.

To further confirm the contributions of intestinal FXR‐FGF15/19 to vancomycin‐induced liver fibrosis, we utilized Fexaramine (Fex), an intestinal‐restricted FXR agonist, to activate intestinal FXR in vivo (Figure7A). We observed decreased ALT and AST levels in serum from Fex‐treated mice (Figure 7B). The mRNA level of downstream genes for FXR activation, such as Shp and Fgf15, were dramatically increased in the ileum from Fex‐treated mice, compared to control mice treated with vehicle (Figure 7C). We observed decreased TBA level in serum from Fex‐treated mice (Figure 7D). Intestinal FXR‐FGF15/19 signaling represses CYP7A1 in the liver.^[15^] Mice treated with Fex exhibited decreased protein levels of CYP7A1 in the liver (Figure 7E), which may be correlated with the decreased Sirius red positive area, and Masson's trichrome staining positive area (Figure 7F). The protective function of Fex against vancomycin‐induced liver fibrosis was further evidenced by decreased protein levels of α‐SMA (Figure 7G). We also found that the administration of recombinant protein FGF19 also reversed vancomycin‐induced cholestasis and fibrosis by suppressing the expression of CYP7A1 (Figure S10, Supporting Information). These results further implicate that vancomycin‐aggravated liver fibrosis depends on the suppression of intestinal FXR‐FGF15/19 signaling.

C57BL/6J mice were housed in a barrier facility on a 12 h light/dark cycle(22 ± 2 °C). Seven‐week‐old mice were divided randomly into 5 mice/cages and housed for 1 week to adjust to the environment. They were then assigned to the H₂O or vancomycin‐treatment group. Mice in the vancomycin group were provided with drinking water containing vancomycin (0.5 g L⁻¹, Sigma‐Aldrich, St. Louis, MO, USA) for 3 weeks, with fresh antibiotic water replaced every other day. After 3 weeks, mice were administrated to either a sham or pBDL operation and received continuous vancomycin treatment for another 2 weeks. In rescue experiments, mice were exposed to water containing vancomycin (0.5 g L⁻¹) for 3 weeks, followed by daily gavage of C. scindens (10¹⁰ CFU per mouse) (American Type Culture Collection (ATCC), Manassas, VA, USA) (10¹⁰ CFU per mouse) or vehicle for 4 days. Subsequently, all mice were administrated to pBDL, repeated with 4 days of vancomycin treatment and 4 days of C. scindens/EcN‐BaiE or vehicle. The mice were euthanized 16 days after pBDL and liver tissues were fixed in formalin for histology.

Following 2 weeks of vancomycin treatment, all mice were subjected to pBDL and then half the mice were gavaged with an intestine‐restricted FXR agonist fexaramine (FEX)/ human fibroblast growth factor‐19 recombinant protein (FGF19) /EcN‐BaiE or vehicle (MedChemExpress, Shanghai, China). In selected experiments, after 2 weeks of oral vancomycin treatment, mice then had free access to a diet supplemented with 1% CA (Feed‐Science, Nanjing, China) and continuously received vancomycin for 4 weeks. Animal use and euthanasia protocols were approved by the Institutional Animal Care and Use Committee of the Xiangya Hospital of Central South University (Ethical Approval No. CSU‐2022‐0061).

Escherichia coli Nissle 1917 (EcN) was purchased from ATCC (Manassas, VA, USA). The information and coding sequences of the BaiE (bile acid 7alpha‐dehydratase) gene of C.scindens were extracted from the NCBI database and optimized these genes for expression in EcN. The region encoding of the BaiE enzyme was cloned into sgRNA/Cas9 cloning plasmid and then introduced into EcN through heat stimulation to create the engineered bacteria (EcN‐BaiE). Colonies were selected on LB agar containing kanamycin (50 µg mL⁻¹). Finally, pCP20 was utilized to eliminate the kanamycin resistance cassettes, resulting in the generation of the final antibiotic‐free wild‐type and EcN‐BaiE strains.

Wild‐type and EcN‐BaiE strains were cultured with LB and incubated on a shaker at 37 °C overnight. The Wild‐type and EcN‐BaiE strains were then harvested by centrifugation at 3000–5000 × g, washed three times with sterile PBS, and resuspended in sterile ice‐cold PBS (the final concentration of ≈10¹¹ CFU mL⁻¹).

Fecal and serum samples were obtained from 17 healthy controls and 24 CLD patients, who were recruited from the Xiangya Hospital of Central South University between 2020 and 2023. The CLD patients and healthy controls were matched based on age and sex. All the participants confirmed that they had not taken any antibiotics within the preceding 3 months before sample collection. This study was approved by the ethics committee of Xiangya Hospital of Central South University (Ethical Approval No. 202008140).

Serum ALT, AST, and ALP levels were assessed using a colorimetric enzymatic assay kit according to the manufacturer's instructions (Sigma‐Aldrich, St. Louis, MO, USA). In brief, 100 µL of the Master Reaction Mix was prepared for each well. Subsequently, 100 µL of the Master Reaction Mix was added to each of the standard, positive control, and test wells (5 µL well⁻¹). The wells were mixed and then incubated at room temperature for 3 min. The initial measurement was taken at 570 nm wavelength, followed by further incubation at 37 °C with measurements taken every 5 min. ALT/AST/ALP activity was analyzed according to the manufacturer's instructions.

Serum TBA concentrations were detected using a fluorometric detection kit according to the manufacturer's instructions (Sigma‐Aldrich, St. Louis, MO, USA). In brief, prepare 250 µL of 80 µm sodium cholate (20 µL Standard + 230 µL ultrapure water) for the Internal Standard. Transfer 20 µL of serum to each of the three wells (Sample, Internal Standard, Sample Blank). Add 5 µL ultrapure water to the Sample and Sample Blank wells, and 5 µL Internal Standard to the Internal Standard well. For the Internal Standard and Sample wells, prepare the Working Reagent by mixing 75 µL Assay Buffer, 8 µL NAD, 4 µL Probe, 1 µL Enzyme A, and 1 µL Enzyme B. For the Sample Blank wells, prepare the Blank Reagent by mixing 75 µL Assay Buffer, 8 µL NAD, 4 µL Probe, and 1 µL Enzyme B (no Enzyme A). Add 80 µL of Working Reagent to the Internal Standard and Sample wells, and 80 µL of Blank Reagent to the Sample Blank wells. Mix by tapping, incubate for 20 min in the dark and read fluorescence intensity at 530 nm excitation and 585 nm emission.

Serum FGF15 levels were detected using an Elisa kit according to the manufacturer's instructions (Cloud‐Clone Crop, Wuhan, China). Add 100 µL of each dilution of standard, blank, and serum into the appropriate wells. Incubate covered with a plate sealer for 1 h at 37 °C. Remove the liquid from each well without washing. Add 100 µL of Detection Reagent A working solution to each well, cover with the plate sealer, and incubate for 1 h at 37 °C. Aspirate the solution and wash each well with 350 µL of 1× Wash Solution, repeating 3 times. Add 100 µL of Detection Reagent B working solution to each well, cover, and incubate for 30 min at 37 °C. Repeat the aspiration/wash process 5 times. Add 90 µL of Substrate Solution to each well and incubate for 15 min at 37 °C, while protecting from light. Add 50µL of Stop Solution to each well, causing the liquid to turn yellow. Ensure no water drops or fingerprints on the plate bottom and no bubbles on the liquid surface. Immediately conduct measurement at 450nm using a microplate reader.

The liver and ileum tissues were fixed in 4% neutralized formalin for 2 d, followed by processing and paraffin embedding. Sections were then subjected to deparaffinization and stained with Hematoxylin and Eosin (H&E), picro‐sirius or biebrich scarlet‐acid fuchsin solution. Slides were visualized using a Zeiss Axio Imager M1 microscope (Carl Zeiss AG, Oberkochen, Germany).

Tissue (20–40 mg) was homogenized in RIPA buffer containing Protease Inhibitor Cocktail and Phosphate Inhibitor Cocktail Tablets (Roche, Indianapolis, IN, USA), using a homogenizer. Protein concentrations were measured using the BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA). 20–30 µg of protein was separated by SDS‐PAGE gel and then transferred onto a polyvinylidene difluoride membrane. The membranes were blocked for 1 h using 5% BSA w/v dissolved in TBST v/v, and then incubated with primary antibodies overnight at 4 °C. Blots were then probed with secondary antibodies (Sigma‐Aldrich, St. Louis, MO, USA) for 1 h, and detected the signal according to the manufacturer's recommendations (Thermo Fisher Scientific, Waltham, MA, USA). All antibodies are listed in the Table(Supporting Information). S3

According to the manufacturer's instructions, total RNA was extracted from liver tissue or cells using Trizol (Thermo Fisher Scientific, Waltham, MA, USA). 1 µg of RNA was treated with DNase I (Invitrogen, Waltham, MA, USA) and incubated at 37 °C for 30 min, then used as a template for cDNA synthesis using a High‐Capacity cDNA Reverse transcription Kit (Applied Biosystems, Foster City, CA, USA). Power SYBR Green PCR Master Mix (Fisher Scientific, Springfield Township, NJ) was used to quantify relative mRNA expression levels (Fisher Scientific, Springfield Township, NJ) on a QuantStudio 7 Flex Real‐Time PCR System (Applied Biosystems, Foster City, CA, USA). The expression levels were calculated using the ΔΔCt method according to MIQE guidelines.^[31^] β‐actin or 16S was used for normalization. Primer sequences are listed in Table S4 (Supporting Information).

Liver samples (10 mg) were homogenized in 20 µL ultrapure water and 200 µL acetonitrile/methanol (v/v = 8:2) solution containing 10 µL internal standard using 25 mg precooled grinding beads. Centrifuge at 13 500 rpm for 20 min at 4 °C and lyophilize the supernatant on a freeze dryer. Redissolve samples in 100 µL 1:1 mixture solution of acetonitrile/methanol (v/v = 8:2) and ultrapure water. Centrifuge at 13 500 rpm for 20 min at 4 °C and supernatant was transferred into a new 96‐well plate. Frozen fecal samples were homogenized in 20 µL ultrapure water and 200 µL acetonitrile/methanol (v/v = 8:2) solution containing 10 µL internal standard using 25 mg precooled grinding beads. Centrifuge at 13 500 rpm for 20 min at 4 °C. Collect 20 µL of supernatant and add 80 1:1 mixture solution of acetonitrile /methanol (v/v = 8:2) and ultrapure water. Inject 5 µL of standards or prepared liver or fecal sample or serum to ultra‐performance liquid chromatography coupled to tandem mass spectrometry (UPLCMS/MS) system (ACQUITY UPLC‐Xevo TQ‐S, Waters Corp., Milford, MA, USA). The bile acid profile (Table S5 Supporting Information) of the liver and feces was measured and analyzed using the QuanMET v1.0 software of Metabo‐Profile Co. (Shanghai, China).^[32^]

Total DNA was extracted from feces using EasyPure Stool Genomic DNA Kit according to the manufacturer's instructions (TransGen Biotech Co, Bejing, China). Briefly, 200 mg of feces were homogenized with glass beads in a 1.5 mL tube. Add 1 mL of LB21 and mix well by vortex. Homogenate was then incubated at 70 °C for 10 min and centrifuged at 15 000 g for 2 min. The supernatant was transferred to a new tube and 250 µL of PB21 was added. Mixed them by vortex and incubated on ice for 5 min. Centrifuge at 15 000 g for 2 min and transfer the supernatant to a new tube. Add the same volume of ethanol and BB21 to the supernatant and mix well by vortex. Transfer the mixtures to a spin column and centrifuge at 15 000 g for 30 s. Add 500 µL of CB21 and WB21 to the column in turn and centrifuge at 15 000 g for 30 s. Discard the flow‐through and centrifuge at 15 000 g for 2 min. Place the spin column into a new collect tube and add 100 µL elution buffer to the center of the column. Incubate at room temperature for 1 min and centrifuge at 15 000 g for 2 min to elute the DN.

DNA was extracted from 1–2 g feces using a QIAamp DNA Microbiome Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. The extracted DNA samples were sent to Novogene Bio‐Technology Co., Ltd. (Nanjing, China) for sequencing and analysis. All the primers were designed based on the V4 hypervariable regions and 16S rRNA genes were amplified with PCR reaction mix. Then sequencing libraries were prepared using the 16S Library Preparation Protocol from the Illumina Miseq platform. Run the sequencing machine of the Miseq system and obtain FASTQ files. Process the overlapping paired‐end FASTQ files in a data curation pipeline implemented in QIIME 2 version 2017.7.0.

Human hepatic stellate cell line (LX2) and Caco‐2 (an immortalized cell line of human colorectal adenocarcinoma cells) were purchased from ATCC (Manassas, VA, USA). LX2 was cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin‐streptomycin (Invitrogen, Waltham, MA, USA). Caco‐2 was cultured in high glucose Dulbecco's modified Eagle's medium (Invitrogen, Waltham, MA) with 10% FBS and 1% penicillin‐streptomycin (Invitrogen, Waltham, MA, USA).

Statistical analysis was determined using GraphPad Prism 9 (GraphPad Software, Inc., La Jolla, CA, USA). Data are presented as mean values with error bars representing S.E.M. Two‐tailed unpaired Student's t‐test or Mann‐Whitney U test or one‐way ANOVA were used to determine differences between two or among three groups, respectively. Spearman correlation analysis was performed between the microbiota taxa and individual bile acids. Statistical outliers were determined using Grubb's test. Statistical difference was defined as p < 0.05.

The data that support the findings of this study are available from the corresponding author upon reasonable request.