Bacteroidesacidifaciens Attenuates Concanavalin A–Induced Liver Injury by Remodeling Bile Acids to Suppress CD95-Mediated Hepatocyte Apoptosis via the TGR5–FOSL1 Pathway Authors: Xiang Zhang, Cheng Zhang, Yi Zhou, Yu Han, Yawei Wang, Hesuiyuan Wang, Mengyan Yang, Qinhan Wu, Feng Bai, Fanfan Qu, Jianbin Ji, Hongru Zhang, Zhenglu Wang, Yuebing Wang, Haijin Xu, Zhenghua Zhou, Mingqiang Qiao, Zhenzhou Wu Published: 2026 ABSTRACT AbstractBackground & AimsLiver diseases have profound impacts on human physiological health. Gut microbiota and their metabolites exert significant regulatory roles in various hepatic disorders. Intestinal Bacteroides acidifaciens has been proven to be protective for hepatocyte apoptosis in some liver disease models, such as concanavalin A–induced hepatitis and acute/chronic alcoholic liver diseases. However, the specific metabolites from Bacidifaciens and their regulatory mechanisms need to be further elucidated.MethodsC57BL/6 mice were gavaged with Bacidifaciens, and liver bile acid profiles were analyzed using metabolomic analysis. Glycocholic acid was screened out according to death receptor CD95 expression, and its underlying mechanism was identified using inhibitor and CRISPR technology in the AML-12 cell line. In vivo protection of glycocholic acid was elucidated by drinking 0.2% glycocholic acid for 35 days.ResultsBacidifaciens administration significantly altered hepatic bile acid composition, such as glycocholic acid, β-muricholic acid, and hyocholic acid. Among differential bile acids, glycocholic acid reduced CD95 expression in transcription and membrane-binding protein levels in hepatic cells. Inhibition or deletion of membrane receptor Takeda G protein-coupled receptor 5 abolished the inhibition of glycocholic acid on CD95 expression. FOSL1 was identified as a downstream transcription regulator of Takeda G protein-coupled receptor 5. Mice administered with 0.2% glycocholic acid were less susceptible to concanavalin A–induced hepatitis with reduced CD95 expression and hepatocyte apoptosis.ConclusionsIntestinal Bacidifaciens elevates hepatic glycocholic acid levels. Glycocholic acid binds to Takeda G protein-coupled receptor 5 and upregulates FOSL1. This subsequently downregulates CD95 transcription and inhibits CD95-mediated hepatocyte apoptosis, thereby alleviating concanavalin A–induced liver injury. This study elucidates the molecular mechanism by which Bacidifaciens modulates CD95-dependent hepatocyte apoptosis via glycocholic acid, providing a molecular foundation for exploring the protective role of Bacidifaciens in CD95-associated liver diseases.Keywords: Bacteroides acidifaciens, CD95, Glycocholic Acid, Hepatocyte, TGR5 Summary Our study discovers a novel “Bacteroides acidifaciens–bile acid–Takeda G protein-coupled receptor 5–FOSL1” pathway in the gut–liver axis. It provides a mechanistic foundation for a potential strategy to treat hepatocellular apoptosis in liver disease. What You Need to Know Background Our previous work demonstrated that Bacteroides acidifaciens shows a protective function in a CD95-dependent manner in concanavalin A–induced liver injury. However, the specific metabolites and mechanism mediating this effect remain unclear. Impact Our study delineates a novel “B acidifaciens–bile acid–Takeda G protein-coupled receptor 5–FOSL1” pathway in the gut–liver axis, providing a mechanistic foundation for a potential strategy to treat hepatocellular apoptosis in liver disease. Future Directions Future studies should investigate whether B acidifaciens exerts protective effects in other liver diseases by regulating bile acid metabolism. The liver is one of the most critical metabolic and immune organs, responsible for metabolizing various substances within the body while also harboring a rich amount of immune cells. The occurrence and development of liver disease profoundly impact the body’s physiological health. Gut and liver maintain frequent material exchange through the portal vein, thereby influencing each other. Extensive research on the gut–liver axis has demonstrated that gut microbiota exert significant regulatory functions in the occurrence and development of diverse liver disorders, such as autoimmune liver disease, metabolic dysfunction-associated steatotic liver disease (MASLD), alcoholic liver disease, and hepatocellular carcinoma.1,2 Certain specific genera and species of bacteria have been implicated in liver disease. Research has demonstrated that Bifidobacterium exerts a protective effect in autoimmune liver disease by regulating the intestinal barrier and hepatic immune cells.3 Conversely, other bacteria, such as Veillonella dispar, are positively associated with the development of autoimmune liver disease, with their abundance showing positive correlations with serum alanine aminotransferase (ALT) levels and hepatic inflammation.4 Lacticaseibacillus exerts protective functions in alcoholic liver disease,5 whereas Enterococcus faecalis is closely associated with disease severity.6 Bacteroides uniformis improves MASLD.7 Gut microbiota maintains an exceptionally close relationship with liver health and disease. Bacteroides acidifaciens (BA) was first discovered and isolated from the mouse gut by Japanese researchers in 2000.8 Subsequently, researchers progressively uncovered BA’s significant regulatory role in the liver. In 2021, a Japanese team found that BA colonizing the intestines of germ-free mice promotes acetate production. It modulates hepatic lipid metabolism and insulin sensitivity by conjugating with free fatty acid receptor 2, thereby exerting a protective effect against nonalcoholic steatohepatitis.9 In 2021, a team from China led by Wang et al found that BA colonizing mouse intestines elevated levels of fibroblast growth factor (FGF) 21 in mouse livers, influencing downstream lipid metabolism and thereby improving nonalcoholic steatohepatitis.10 In 2022, another Japanese research team found that increased BA abundance in mouse intestines significantly elevated short-chain fatty acid levels, particularly propionic acid in portal vein blood, suppressed hepatic inflammatory cytokines and promoted antioxidant responses, thereby mitigating ischemia–reperfusion injury in the liver.11 In 2024, Yin’s team from China found that BA could generate unconjugated bile acids via bile salt hydrolase (BSH), activating the intestinal farnesoid X receptor (FXR) and its downstream target FGF15, then promoted glutamate synthesis and ammonia detoxification in hepatocytes, consequently alleviating alcoholic liver injury.12 These outstanding studies represent exciting explorations into the role of BAs in regulating hepatic homeostasis. However, overall, research on the mechanisms of BAs in maintaining liver health remains limited. The etiology of liver disease is highly complex, with hepatocyte apoptosis being a pathologic process present in the progression of multiple liver disorders. Severe liver damage is often accompanied by massive hepatocyte apoptosis.13,14 CD95 (Fas/APO-1) is a death receptor widely expressed on the surface of various cells. Apoptosis triggered by CD95 and CD95L (FasL) combination in hepatocytes is extensively involved in liver disorders such as autoimmune liver disease, viral hepatitis, and alcoholic liver damage.15, 16, 17 Activated natural killer T cells exacerbate alcohol-related liver injury through CD95-mediated hepatocyte apoptosis.18 Hepatitis B virus surface antigens enhance hepatic cell apoptosis by promoting death-inducing signaling complex assembly through increased CD95 polymerization and inhibition of Akt activation, thereby predisposing mice to acute liver failure.19 In human autoimmune liver disease, increased hepatocyte apoptosis represents the primary mode of hepatocyte death, including the extrinsic apoptosis pathway mediated by CD95.20 The concanavalin A (ConA)-induced liver injury murine model is a well-established experimental system for immune-mediated acute hepatitis. Previous studies have shown that the CD95 signaling pathway is pivotal in mediating hepatocyte apoptosis in this model,21,22 and its inhibition robustly attenuates hepatic damage in mice.23,24 These studies indicate that CD95/CD95L–mediated hepatocyte apoptosis has a vital role in maintaining liver homeostasis. CD95 expression is regulated by multiple factors, including N6-methyladenosine demethylase,25 transcription factor NF-κB,26 and post-translational modifications.27 However, current research on the transcriptional regulation of CD95 remains unclear and warrants further investigation. Metabolites derived from gut microbiota, including bile acids, lipopolysaccharides, short-chain fatty acids, and tryptophan derivatives, could enter the liver via the portal vein and influence hepatic physiological functions. These metabolites play key roles in the occurrence and development of diverse liver disorders, such as autoimmune liver disease, alcoholic liver disease, and hepatocellular carcinoma.28,29 Bile acids are synthesized in the liver, stored in the gallbladder, released into the intestines, partially modified by gut microbiota, and ultimately excreted into feces or reabsorbed back into the liver. Acting as signaling molecules, bile acids bind to specific receptors to exert diverse biological effects, including modulation of immune function, glucose, lipid and energy metabolism.30 Many research works indicate that regulation of bile acid composition by intestinal flora is significantly related to the development of some liver diseases, such as alcoholic liver disease, MASLD, and liver injury. B uniformis colonizing in mouse intestines synthesizes the secondary bile acid 3-succinoylcholic acid via β-lactamase, and reshapes the gut microbiota to improve intestinal barrier function and anti-inflammatory effects, thereby alleviating MASH.31 Parabacteroides distasonis reduces intestinal reabsorption by enhancing BSH activity and suppressing ileal FXR and decreases taurochenodeoxycholic acid (TCDCA) levels, thus inhibiting MPT-Caspase-11 pyroptosis in mouse hepatocytes and liver fibrosis.32 Research on the bile acid regulation function by BA is limited. Some researchers reported that BA could increase unconjugated bile acids abundance in both mouse intestines and coculture systems in vitro.12,33 However, regulation of bile acid composition in mouse livers by BA and its meaning remains unknown. The relationship between bile acids and CD95-mediated apoptosis is rarely studied. It was reported that glycochenodeoxycholic acid (GCDCA) could induce CD95 membrane localization and activate hepatocyte apoptosis.34 Taurolithocholic acid (TLCA) could bind to Takeda G protein-coupled receptor 5 (TGR5) and promote Ser/Thr phosphorylation of CD95, thus inhibiting its membrane transport and localization in cholangiocytes and following apoptosis without affecting its transcription.35 Nevertheless, the regulatory mechanisms of bile acid components on the expression and function of CD95, especially at the transcriptional level, remain largely unknown and require further elucidation. In the publication of our group in 2022, we reported that BA shows a protective function in a CD95-dependent manner in ConA-induced mouse liver injury.36 In this study, we aim to identify which BA-derived metabolite mediates this protection and to elucidate its mechanism of action. Transcriptomic and metabolomic analyses of liver tissues from BA-treated mice revealed that BA modulates bile acid composition in the liver with significant increases in some bile acids such as glycocholic acid (GCA) and β-muricholic acid (β-MCA). Screening of differential bile acid components using cell coculture shows only GCA could inhibit CD95 expression in a hepatocyte cell line AML-12. Free drinking of 0.2% sodium GCA solution decreased mice’s susceptibility to ConA-induced liver injury, inhibiting CD95 expression and apoptosis in liver tissues. Mechanistically, the results indicate GCA binds to the bile acid receptor TGR5 to upregulate the expression of the transcriptional regulator FOSL1. Increased expression of FOSL1 negatively regulates CD95 transcription, consequently exerting a protective effect against CD95-mediated hepatocyte apoptosis. This study elucidates a clear molecular mechanism by which BA regulates CD95-mediated hepatocyte apoptosis, establishing a theoretical foundation for investigating the protective effects of BA against CD95-related liver diseases. Results BA Protects Against Liver Injury and Affects Metabolite Compositions in the Mouse Liver We found that BA-treated mice exhibited lower serum ALT levels in the ConA-induced liver injury model (Figure 1A). Hematoxylin and eosin (H&E) staining showed a significant reduction in the necrotic area in the livers of BA-treated mice (Figure 1B). CD95 immunohistochemistry (IHC) revealed that hepatic CD95 expression was significantly lower in BA-treated mice compared with the control group (Figure 1C). Subsequently, hepatic injury was induced using the CD95 agonistic monoclonal antibody Jo2, which directly activates the CD95 pathway and triggers acute hepatocyte apoptosis. Results also showed that serum ALT levels were significantly reduced in BA-treated mice (Figure 1D). These results indicate that oral administration of BA protects mice from ConA-induced liver injury, and this effect is CD95-dependent. To investigate the mechanism of BA’s protection, transcriptomic analysis was performed using the liver tissues of BA-treated and control mice. Kyoto Encyclopedia of Genes and Genomes pathway analysis showed significant enrichment in bile acid secretion and primary bile acid synthesis pathways in the mouse liver following BA administration, suggesting that BA modulates these hepatic processes (Figure 1E). Untargeted metabolomic analysis of the livers from BA-treated and control mice was conducted. Principal component analysis revealed that the metabolic composition in the liver changed significantly after BA administration (Figure 1F). Using the Kyoto Encyclopedia of Genes and Genomes database to annotate all identified metabolites, we found significant enrichment in bile acid–related pathways the same as transcriptomic analysis results (Figure 1G). Preliminary analysis of differential metabolites revealed significant differences in various bile acid components, including GCA, tauroursodeoxycholic acid (TUDCA), taurodehydrocholic acid (TDHCA), and taurodeoxycholic acid (TDCA) (Figure 1H and I). These data suggest that BA administration alters bile acid compositions in the mouse liver. BA influences metabolite compositions in the mouse liver and exerts a protective effect against ConA-induced liver injury. (A) Serum ALT levels 12 hours after ConA tail vein injection in mice (n = 6 mice per group). (B) Representative H&E liver sections and necrotic area statistics (n = 6 mice per group). Scale bar, 200 μm. (C) CD95 IHC staining and analysis (n = 6 mice per group). Scale bar, 100 μm. (D) Serum ALT levels 12 hours after Jo2 intraperitoneal injection in mice (n = 8 mice per group). (E) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment results for differentially expressed genes in liver transcriptome (n = 6 mice per group). (F) Principal component analysis (PCA) of liver nontarget metabolites (n = 8 mice per group). (G) KEGG metabolic pathway enrichment analysis of liver nontarget metabolites. (H) Volcano plot of differentially expressed metabolites in liver nontarget metabolites. (I) Differentially expressed bile acids in liver nontarget metabolites. Data from 3 independent experiments combined. BA, oral administration of BA; Control/CT, normal mouse fecal microbiota transplantation. Statistical analysis: mean ± standard deviation (SD). ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. BA Alters Enterohepatic Bile Acid Composition To further clarify the changes in hepatic bile acid profiles following BA administration, we performed targeted bile acid metabolomic analysis on the livers of BA-treated and control mice (another independent experiment distinguished from untargeted metabolomic analysis). Principal component analysis, hierarchical clustering analysis, and volcano plots revealed significant alterations in the bile acid compositions of mouse livers following BA administration compared with the control group (Figure 2A–C). Total bile acid levels were notably increased (Figure 2D), with elevated levels of primary bile acids and decreased levels of secondary bile acids (Figure 2E). Elevated levels of both conjugated and unconjugated bile acids were observed (Figure 2F). Within the conjugated bile acids, both taurine-conjugated and glycine-conjugated bile acids exhibited increased concentrations (Figure 2G). Compared with controls, the levels of GCA, β-MCA, hyocholic acid (HCA), and TUDCA were notably elevated in the livers of BA-administered mice, whereas levels of hyodeoxycholic acid (HDCA), taurohyocholic acid (THCA), taurohyodeoxycholic acid (THDCA), and deoxycholic acid (DCA) were significantly decreased (Figure 2H). To delineate the impact of BA colonization on bile acid metabolism, we conducted targeted metabolomic profiling of murine fecal bile acids. The results showed that BA administration increased total and primary bile acids, reduced secondary bile acids, significantly increased conjugated bile acids, and left unconjugated bile acids unchanged (Figure 2I–L). This pattern was generally consistent with hepatic alterations, except for unconjugated bile acids and glycine-conjugated bile acids. These results indicate that BA colonization remodels the enterohepatic bile acid pool in mice. BA altersenterohepatic bile acid composition. (A) Principal component analysis of liver bile acid–targeted metabolomics (n = 6 mice per group). (B) Heatmap clustering analysis of liver bile acid–targeted metabolomics. (C) Volcano plot of liver bile acid–targeted metabolome. (D) Comparison of total bile acid content in the liver. (E) Comparison of primary and secondary bile acid content in the liver. (F) Comparison of conjugated and unconjugated bile acid content in the liver. (G) Comparison of taurine- and glycine-conjugated bile acid content in the liver. (H) Differentially altered bile acids in the liver. (I) Comparison of total bile acid content in feces (another independent experiment) (n = 6–7 mice per group). (J) Comparison of primary and secondary bile acid content in feces. (K) Comparison of conjugated and unconjugated bile acid content in feces. (L) Comparison of taurine- and glycine-conjugated bile acid content in feces. BA, oral administration of BA; Control/CT, normal mouse fecal microbiota transplant. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. GCA Reduces Hepatocellular CD95 Expression To verify the regulatory effects on CD95 expression in hepatocytes of the differential bile acid components gained in the targeted bile acid metabolomic analysis, we screened these bile acids cocultured with hepatocyte cell line AML-12 based on parameters such as concentration, fold change, and significance. We selected the highest treatment dose of each bile acid component to be no more than 1/2 IC50. Flow cytometry results showed GCA significantly decreased the proportion of CD95+ AML-12 cells in a dose-dependent effect, whereas other bile acids, such as TUDCA, β-MCA, and HCA, did not show this effect (Figure 3A and B; Figure 4A–H). Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) was applied to further examine the impact of these potential bile acids on CD95 transcription in AML-12 cells. The results showed that GCA significantly reduced CD95 transcription levels in a dose-dependent manner (Figure 3D), and the other bile acids (eg, HDCA, THDCA, and THCA) did not cause substantial changes in CD95 transcription (Figure 3F). To validate the generalizability of GCA’s effect across different liver cell lines, we further incubated Hepa1-6 cells with GCA. Flow cytometry and qRT-PCR results indicated that GCA reduced both the membrane CD95 and CD95 transcription levels in a dose-dependent manner, similarly to AML-12 cells (Figure 3C and E). These results collectively demonstrate that GCA significantly downregulates both transcriptional and membrane protein expression of CD95 in hepatocytes. GCA reduces hepatocyte CD95 expression. (A) Flow cytometry analysis of CD95-positive cell proportion after coculture of AML-12 cells with various differential bile acids. (B) CD95-positive cell proportion after coculture of AML-12 cells with different concentrations of GCA. (C) CD95-positive cell proportion after coculture of Hepa 1-6 cells with different concentrations of GCA. (D) qRT-PCR analysis of CD95 transcription levels in AML-12 cells cocultured with different concentrations of GCA. (E) qRT-PCR analysis of CD95 transcription levels in Hepa 1-6 cells cocultured with different concentrations of GCA. (F) qRT-PCR analysis of CD95 transcription levels in AML-12 cells cocultured with differential bile acids. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. Coculture of AML-12 cells with different bile acids in vitro. (A–H) Changes in the proportion of CD95+ AML-12 cells detected by flow cytometry. (I–L) Changes in CD95 transcription levels detected by qRT-PCR. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. GCA Regulates CD95 Expression via the Bile Acid Receptor TGR5 To identify the bile acid receptor to which GCA binds to regulate CD95 expression in AML-12 cells, we conducted coculture experiments using various bile acid receptor inhibitors, including FXR inhibitor DY-268, sphingosine-1-phosphate receptor 2 (S1PR2) inhibitor JTE-013, and TGR5 inhibitor SBI-115. Flow cytometry results showed that only SBI-115 could reverse the inhibitory effect of GCA on CD95 expression (Figure 5A; Figure 6), suggesting that GCA may inhibit CD95 expression via TGR5. Further qRT-PCR analysis revealed that SBI-115 effectively reversed GCA’s suppression of CD95 transcription (Figure 5B), consistent with the flow cytometry results. To further validate this result, we designed CRISPR/Cas9 single guide RNA to knock out TGR5 in AML-12 cells using the https://crispor.gi.ucsc.edu tool (Figure 7; Table 1). After knocking out TGR5 (TGR5 knockout [KO]), the inhibitory effects of GCA on CD95 transcription levels and membrane protein levels (Figure 5C–E) were abolished. These results indicate that GCA reduces hepatocellular CD95 transcription and membrane protein expression through binding to TGR5. GCA combined with TGR5 reduces hepatocyte CD95 expression. (A) Flow cytometry analysis of CD95-positive cell proportion after coculture of AML-12 cells with various bile acid receptor inhibitors. (B) qRT-PCR detection of CD95 transcript levels after SBI-115 coculture. (C) Flow cytometry analysis of CD95-positive cell proportion after TGR5 knockout. (D) Changes in fluorescence intensity and bar chart statistics. (E) qRT-PCR analysis of CD95 transcription levels after TGR5 knockout. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. Coculture of AML-12 cells with different bile acid receptor inhibitors in vitro. (A–C) Flow cytometry analysis of changes in the proportion of CD95+ AML-12 cells. (D) qRT-PCR analysis of changes in CD95 transcription levels. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. CRISPR/Cas9 knockout sites. Table 1Single Guide RNA DesignSingle guide RNA nameSingle guide RNA sequence (5'-3')KO TGR5 1-FCACCGTAGTGGTGGGCGACGCTCATKO TGR5 1-RAAACATGAGCGTCGCCCACCACTACKO TGR5 2-FCACCGCTGTGTGAGATCCGCCGACKO TGR5 2-RAAACGTCGGCGGATCTCACACAGCKO CREB5 1-FCACCGTTCCCGTTCGACATGGCGGCKO CREB5 1-RAAACGCCGCCATGTCGAACGGGAACKO CREB5 2-FCACCGCTTCGCCGGCGTCCCCCTGTKO CREB5 2-RAAACACAGGGGGACGCCGGCGAAGCKD FOSL1 1-FCACCGCTCGTATGACTCCTGGCCGKD FOSL1 1-RAAACCGGCCAGGAGTCATACGAGCKD FOSL1 2-FCACCGTCCGCTCGCGTCTCACCCTGKD FOSL1 2-RAAACCAGGGTGAGACGCGAGCGGAC Single Guide RNA Design GCA Regulates CD95 Expression via the Transcription Factor FOSL1 To identify transcription factors downstream of GCA binding to TGR5 that regulate CD95 transcription, transcriptomic analysis on GCA-cocultured AML-12 cells was performed, followed by transcription factor database predictions (Figure 8). Several potential transcription factors, such as CREB5, STAT5B, and FOSL1, were selected (Figure 9A and B) and subsequently validated by qRT-PCR. As shown in Figure 9C and Figure 10, CREB5 and FOSL1 transcription levels were reversed when AML-12 cells were treated with GCA plus SBI-115 compared with GCA treatment only, whereas no consistent reversion was observed for the other transcription factors. CREB5 and FOSL1 transcriptions were also reversed in TGR5 KO AML-12 cells compared with wild-type cells (Figure 9D). To distinguish which one of CREB5 and FOSL1 is the key transcription regulator of CD95, CREB5 KO or FOSL1 knockdown AML-12 cells (KO may be lethal) were created respectively using CRISPR/Cas 9 (Figure 7; Table 1). Flow cytometry and qRT-PCR analysis showed that only FOSL1 knockdown significantly restored GCA-induced downregulation of CD95 transcription and membrane protein levels (Figure 9E–G). KO of CREB5 upregulated CD95 transcription but failed to restore CD95 membrane protein levels (Figure 11). To further verify whether FOSL1 and CREB5 directly bind to the CD95 promoter region, we performed computational prediction followed by experimental validation of transcription factor binding sites. The results revealed 2 putative FOSL1-binding sites and 1 putative CREB5-binding site within the CD95 promoter, with the latter overlapping the first FOSL1-binding site (Figure 9H). Chromatin immunoprecipitation–qPCR analysis showed that FOSL1 was significantly enriched at both predicted binding sites in the CD95 promoter. Notably, GCA treatment enhanced FOSL1 enrichment, particularly at site 2 (Figure 9I). CREB5 also exhibited increased enrichment at its predicted site 1 following GCA exposure, although to a lower degree than observed with FOSL1 (Figure 9I). Agarose gel electrophoresis further showed that the expected bands were detected in samples immunoprecipitated with the corresponding antibodies, whereas no obvious specific bands were observed in the IgG negative control, further supporting the enrichment of CREB5 and FOSL1 at the predicted binding sites within the CD95 promoter (Figure 12). These data indicate that both CREB5 and FOSL1 bind directly to the CD95 promoter region, with FOSL1 exhibiting significantly stronger binding affinity than CREB5. Collectively, these results indicate that GCA binds to TGR5, upregulates downstream transcription regulator FOSL1, and subsequently downregulates CD95 transcription and membrane expression in hepatocytes, whereas CREB5 may exert an auxiliary regulatory role in this process. Prediction of regulatory transcription factors for CD95, integrated with GCA-treated AML-12 cell transcriptomics analysis. GCA suppresses CD95 expression in hepatocytes by regulating the transcription factor FOSL1. (A) Volcano plot of differentially expressed genes between the control (CT) and GCA-treated groups. (B) Potential transcription factor expression changes predicted by transcription factor databases and screened by transcriptomics. (C) qRT-PCR detection of transcription level changes in potential downstream transcription factors of TGR5 activated by GCA. (D) qRT-PCR detection of potential transcription factor expression changes after TGR5 knockout. (E) Flow cytometry detection of CD95-positive cell proportion after FOSL1 knockdown. (F) Fluorescence intensity changes and bar statistics. (G) qRT-PCR detection of CD95 expression changes after FOSL1 knockdown. (H) Predicted binding sites. (I) Validation by chromatin immunoprecipitation–qPCR. n = 3 per group. Statistical analysis: mean ± standard deviation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. Screening for transcription factors mediating the effects of GCA-bound TGR5 on CD95 in AML-12 cells. (A–M) qRT-PCR analysis of the mRNA expression of the indicated transcription factors under the indicated treatments. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. Examining whether GCA regulates CD95 in AML-12 cells by modulating CREB5. (A) Flow cytometry analysis of changes in the proportion of CD95+ AML-12 cells after CREB5 KO. (B) Changes in mean fluorescence intensity (MFI) and bar chart statistics. (C) qRT-PCR detection of changes in CD95 transcription levels after CREB5 KO. n = 3 per group. Statistical analysis: mean ± standard deviation. ∗∗∗P < .001; ∗∗∗∗P < .0001. Agarose gel electrophoresis validation of the binding of transcription factors FOSL1 and CREB5 to the CD95 promoter. GCA Alleviates ConA-Induced Liver Injury by Downregulating CD95 Expression in Mouse Liver To evaluate whether GCA exerts protective effects against liver injury by inhibiting CD95-mediated hepatocyte apoptosis, mice were provided with drinking water containing 0.2% GCA for 5 weeks, followed by induction of liver injury via tail vein injection of ConA (Figure 13A). Liver tissue samples were collected 12 hours later for analysis. Results showed that serum ALT levels were significantly reduced in GCA-treated mice (Figure 13B), and the area of hepatic necrosis was markedly decreased (Figure 13C), compared with the control group. IHC and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining further revealed that GCA treatment significantly downregulated CD95 expression (Figure 13D) and substantially reduced hepatocyte apoptosis in liver tissues (Figure 13E). These results demonstrate that GCA protects against ConA-induced liver injury by suppressing CD95 expression and inhibiting hepatocyte apoptosis in the mouse liver. GCA reduces CD95 expression in mouse liver and alleviates ConA-induced liver injury. (A) Experimental strategy. (B) Serum ALT levels in mice (n = 7–8 per group) after 5 weeks of free water/0.2% GCA drinking water and tail vein injection of ConA 12 hours prior. (C) Representative H&E liver sections and necrotic area statistics (n = 9–10 mice per group). Scale bar, 200 μm. (D) CD95 IHC staining and analysis (n = 3 mice per group). Scale bar, 100 μm. (E) TUNEL assay and analysis (n = 3 mice per group). Scale bar, 100 μm. Statistical analysis: mean ± standard deviation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. Discussion The regulatory role of gut commensal BA on hepatic bile acid composition and its meaning in liver diseases are rarely studied. We demonstrated that gut-colonized BA protects against ConA-induced hepatitis in a CD95-dependent manner, which is consistent with our previous work.36 In the present study, we hypothesize that BA suppresses CD95 expression in hepatocytes by modulating the composition of specific metabolites in the liver. Our results demonstrate that increasing BA abundance in the mouse gut modulates hepatic bile acid composition, elevating the proportion of bile acids such as GCA and β-MCA in liver tissue. Among these bile acids, GCA binds to the membrane receptor TGR5 on hepatocyte surfaces, thereby upregulating the expression of the transcription regulator FOSL1. This subsequently downregulates CD95 transcription, inhibiting CD95-mediated apoptosis of hepatocytes. Mice drinking sodium GCA solution exhibited significantly reduced susceptibility to ConA-induced liver injury, along with decreased hepatic CD95 expression and apoptosis. This study first reports the molecular mechanism by which intestinal BA protects hepatocytes by regulating hepatic GCA levels. Given that CD95-mediated hepatocyte injury is widely implicated in various liver diseases, including autoimmune, alcoholic, and viral hepatitis,15, 16, 17 our work demonstrates that the regulatory function of intestinal BA on bile acid metabolism holds significant importance for hepatocyte protection across multiple hepatic pathologies. In this study, BA administration significantly elevated total hepatic bile acid levels in mice. Notably, the expression of bile acid transporters, including ileal apical sodium-dependent bile acid transporter and organic solute transporter α/β, as well as hepatic sodium taurocholate cotransporting polypeptide and organic anion transporting polypeptides, remained unchanged (Figure 14), suggesting the preserved enterohepatic reabsorption capacity that likely enhances intrahepatic bile acid accumulation. Conversely, after BA administration, fecal bile acid excretion was correspondingly enhanced as the total enterohepatic bile acid pool increased (Figure 2I). Hepatic transcriptomics revealed significant downregulation of bile acid synthetic enzymes, including Cyp7a1, Cyp27a1, Cyp7b1, and Cyp8b1 (Figure 15), indicating either reduced de novo synthesis or negative feedback in response to elevated intrahepatic bile acids. Critically, our end point analyses (7 days post BA administration) reflect steady-state adaptation outcomes rather than dynamic regulatory events. Stage-dependent profiling of bile acid synthesis/transport/metabolism following BA administration is warranted to identify early mechanistic driving factors. Changes in the expression of bile acid transporters in the liver and ileum of mice following oral administration of BA. n = 6 per group. Statistical analysis: mean ± standard deviation. ns, not significant. Changes in the expression of genes related to bile acid synthesis in the livers of mice following oral administration of BA. n = 6 per group. Statistical analysis: mean ± standard deviation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. We further assessed whether the marked increase in hepatic total bile acids after BA administration potentiates cholestatic liver injury. We employed an α-naphthyl isothiocyanate (ANIT)-induced cholestatic liver injury mouse model following BA administration. Notably, BA-treated mice exhibited comparable levels of serum ALT, total bilirubin (TBil), and direct bilirubin (DBil) vs controls, with no increase in hepatic necrosis by H&E staining (Figure 16). Liver bile acid targeted metabolomic analysis (Figure 2D–H) indicated that BA administration shifted the hepatic bile acid profile toward hydrophilic and low-toxicity pattern, with a marked decrease in hydrophobic secondary bile acids, likely attenuating cytotoxicity. Therefore, the hepatic bile acid composition remodeled by BA likely mitigates hepatotoxic risk despite elevated total bile acids. Effects of BA gavage on ANIT-induced cholestasis in mice. (A) Serum ALT levels. (B) Serum DBil levels. (C) Serum TBil levels. (D) Representative H&E-stained liver sections, scale bar: 100 μm. n = 7 per group. Statistical analysis: mean ± standard deviation. ns, not significant. There has been relatively little research on the regulatory function of BA on bile acids, but some researchers have made certain explorations in this area. Shen et al reported that BA can alter the intestinal bile acid composition in mice by expressing BSH, reducing the levels of conjugated bile acids such as THCA and TUDCA, while increasing the contents of unconjugated bile acids such as cholic acid (CA) and DCA.12 Similarly, Jiao et al demonstrated that coculturing BA with human- or mouse-derived bile acids in vitro increases the proportion of unconjugated bile acids.33 These findings collectively indicate that BA possesses a notable capacity to modulate bile acid composition. In the present study, targeted metabolomic analysis of mouse liver showed that both conjugated and unconjugated bile acids increased following BA treatment, which is not entirely the same as those reported in the aforementioned studies. However, both lines of research consistently observed that BA increased the level of unconjugated bile acid CA in the intestine and in vitro. Given that most of the bile acids in the intestines can be reabsorbed into the liver and reconjugated with glycine or taurine to form conjugated bile acids, the increase in the content of intestinal CA resulting from BA is expected to lead to a subsequent increase in the content of GCA in the liver. The observation in this study that BA significantly increases GCA levels in the liver is consistent with the conclusions of the aforementioned research. BSH-mediated deconjugation is the initial step in secondary bile acid formation. However, the complete biotransformation requires the activity of key downstream enzymes, such as 7α-dehydroxylases and hydroxysteroid dehydrogenases. We found that, following BA administration, secondary bile acids decreased in both the liver and feces of mice, whereas primary and conjugated bile acids concomitantly increased. In terms of alterations in specific bile acid compositions, the levels of CA and GCA were increased in the feces of mice following BA administration, whereas the contents of DCA, TLCA and TDCA were decreased (Figure 17). This metabolic shift indicates neither enhanced deconjugation nor secondary bile acid enrichment. Therefore, we propose that although BA expresses BSH activity, its capacity for terminal conversion is constrained by low activity of key enzymes (eg, 7α-dehydroxylase). Consequently, BA-dominated microbiota establishes a biochemical bottleneck limiting secondary bile acid production. However, 1 limitation of this study is that it did not conduct an in-depth analysis of the enzymatic activity of BA in the metabolism of secondary bile acids. Differential changes in some bile acids in mouse feces following oral administration of BA. n = 6–7 per group. Statistical analysis: mean ± standard deviation. ∗∗P < .01; ∗∗∗P < .001. CD95-/CD95L-mediated hepatocyte apoptosis plays a crucial role in liver homeostasis. CD95 expression could be regulated by various metabolites, including bile acids. Katona et al reported that LCA could induce increased localization of CD95 on the membrane of human colon cancer cells.37 Graf et al demonstrated that taurolithocholic acid-3 sulfate induces CD95 membrane transport in rat hepatocytes, promoting CD95-mediated apoptosis, whereas TUDCA prevents CD95 membrane transport and attenuates the apoptotic response.38 Lian et al indicated that CDCA downregulates CD95 expression in hepatocytes and ameliorates ConA-induced liver injury.39 In this study, liver bile acid–targeted metabolomics analysis revealed that there was no correlation between CD95 expression or liver injury severity and the bile acids mentioned above, LCA, TLCA, CDCA, and TUDCA (Figure 2H; Figure 4). GCA showed inhibition on CD95 expression and CD95-mediated apoptosis in hepatocytes. However, THDCA also exhibited inhibitory effect on CD95 membrane protein levels when cocultured with AML-12 cells, although this bile acid was not associated with liver injury severity according to the bile acid–targeted metabolomics analysis. Bile acids, as a class of signaling molecules, exert biological functions by binding to various kinds of bile acid receptors, such as FXR, TGR5, S1PR2, and constitutive androstane receptor (CAR). Several studies have shown that a few bile acids interact with the corresponding receptors to play roles in expression or post-translation regulation. Chenodeoxycholic acid (CDCA) acts as an agonist of FXR to reduce hepatic CD95 expression.39 Besides CDCA, HDCA and DCA are also FXR agonists.40 TLCA was reported to bind to TGR5 and promote serine/threonine phosphorylation of CD95, thus inhibiting its membrane transport and localization in cholangiocytes.35 CDCA, HDCA, DCA, and TLCA all decrease in the liver tissue of BA-treated mice according to the bile acid–targeted metabolomics analysis in this research, suggesting these bile acids are not the key components that mediate the protection of BA. Our results show GCA could suppress CD95 transcription, not post-translational modification, via TGR5, rather than FXR, S1PR2, or CAR (Figure 6). Furthermore, Shen et al demonstrated that BA elevates unconjugated bile acid levels in the intestine, activates intestinal FXR to enhance FGF15 expression. FGF15 enters into liver through portal vein and increases glutamate-related metabolites such as glutathione to exert protection to liver.12 To validate whether BA regulates hepatic CD95 expression via FGF15, recombinant FGF15 protein and its receptor FGFR4 inhibitor were applied to handle AML-12 cells in vitro. The results indicated that FGF15 did not exhibit inhibitory effect on CD95 expression (Figure 18C). Exploration of different pathways to reduce CD95 transcription levels in AML-12 cells. (A) Coculture of AML-12 cells with the cAMP analog bucladesine sodium in vitro; qRT-PCR detection of changes in CD95 and CREB5 transcription levels. (B) Coculture of AML-12 cells with the PI3K inhibitor LY294002 in vitro; qRT-PCR detection of changes in CD95 and CREB5 transcription levels. (C) FGF15/FGFR4 inhibitor was cocultured with AML-12 cells in vitro, and qRT-PCR was used to detect changes in CD95 transcription levels. n = 3 per group. Statistical analysis: mean ± standard deviation. ns, not significant; ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. TGR5, as a bile acid–activated G protein-coupled receptor, is widely distributed across various tissues and organs, and plays crucial roles in multiple physiological pathways. Some bile acids, such as TLCA, DCA, and TDCA, are recognized as potent TGR5 ligands with higher affinity than GCA. Although GCA suppresses CD95 expression via TGR5 activation in this study, we investigated whether high-affinity TGR5 agonists exert similar effects. Treatment of AML-12 cells with TLCA or DCA within maximally tolerated concentrations revealed that DCA could not significantly inhibit CD95 expression, and TLCA marginally reduced CD95 transcription (P < .05) (Figure 4D, L). These results suggest ligand-specific divergences may exist in TGR5 downstream signaling, which is consistent with established pathway heterogeneity for this receptor.41 Hepatic bile acid metabolomics data in BA-treated mice demonstrated significant reductions in TLCA, DCA, and TDCA, yet these high-affinity agonists, particularly TDCA, exhibited relatively high levels (Figure 2H). Their minimal contribution to CD95 suppression implies that in vivo physiological effects extend beyond receptor affinity, involving multifactorial determinants, such as divergent downstream signaling, receptor competition dynamics, bile acid hydrophobicity, and cell type, rather than the effect of a single receptor or pathway alone. In addition, the data processing methods for bile acid metabolomics may also serve as an important contributing factor. Methodologically, TLCA, DCA, and TDCA fell below detection thresholds in some samples. Imputation of missing values as half-minimums according to the data processing method provided by the metabolomics service company, likely overestimated their abundance, which may be one of the reasons why these high-affinity TGR5 ligands do not show the inhibitory effect on CD95 expression through TGR5 activation in vivo. Studies have shown that TLCA can bind to TGR5, activate cAMP and downstream protein kinase A, thereby inhibiting the membrane localization of CD95 in cholangiocytes.35 In this study, we found that the cAMP analogue bucladesine could reduce CD95 transcription in AML-12 cells (Figure 18A), and the cAMP-responsive protein CREB5 was negatively relevant to CD95 transcription (Figure 9B; Figure 10). However, CREB5 knockout could not reverse the membrane CD95 levels according to flow cytometry analysis in AML-12 cells, despite upregulation of CD95 transcription (Figure 11). This indicates that CREB5 is weakly involved in the negative regulation of GCA on CD95-mediated hepatocyte apoptosis. Given that phosphoinositide 3-kinase (PI3K)/Akt signaling is also a potential downstream pathway of TGR5,42 we performed a preliminary validation using a PI3K pathway inhibitor. The results showed that PI3K inhibition reversed the suppressive effect of GCA on CD95 expression (Figure 18B), suggesting that the PI3K signaling pathway may be involved in the regulation of CD95 expression by GCA. However, because this observation has not yet been further validated through genetic interventions, pathway activation assays, or other complementary approaches, the precise role of PI3K signaling in this process remains to be elucidated. Further integration of transcription factor prediction analysis and in vitro screening revealed that FOSL1, a downstream transcription regulator of TGR5, suppresses both CD95 transcription and membrane protein levels. However, we failed to set up a correlation between FOSL1 and cAMP. FOSL1 is a crucial transcription factor belonging to the AP-1 family of transcription complexes. FOSL1 was reported to play essential roles in biological processes such as proliferation, differentiation, and survival, and participate in the initial stage and progression of various diseases.43 Previous studies have shown FOSL1 could inhibit apoptosis in osteosarcoma cells and promote drug resistance by regulating the expression of apoptosis-related proteins BCL2 and BAX.44 No report has shown any relevance between FOSL1 and CD95 so far. The function in liver diseases has been rarely discussed. In this study, our findings report for the first time that FOSL1 is regulated by TGR5 and participates in CD95-mediated hepatocyte apoptosis. There were several limitations in this study. We verified GCA’s function on CD95 in mouse and human hepatocellular carcinoma cell lines and found that CD95 downregulation was weaker in Hepa1-6 and HepG2 cells than in AML-12 cells (Figure 3C and E; Figure 19). Due to the limited availabilities of liver cell lines and the challenges in isolating and culturing primary hepatocytes in vitro (isolated primary hepatocytes highly expressed CD95 because of the oxidative environment), we could not identify our findings in more hepatocyte cells or cell lines. Future work will involve in-depth validation using organoids derived from primary liver cells. In addition, we also observed that GCA promoted CD95 expression in the mouse lymphoma cell line EL4 (Figure 20), which is opposite to the findings in the hepatocyte line AML-12. We hypothesize that bile acids may exert distinct biological functions by binding different receptors in various tissues and cell types. We will further explore the function patterns of bile acids across different tissues and cells. Furthermore, questions about how BA influences bile acid synthesis, transport, and metabolism to alter hepatic and intestinal bile acid composition, as well as its regulatory mechanisms in other liver disease models, remain unresolved. These questions warrant further investigation to comprehensively evaluate the role of BA in regulating bile acids in the hepatobiliary system within hepatobiliary-related diseases. GCA regulates CD95 expression in HepG2 cells via TGR5. (A) Changes in CD95 membrane protein levels following coculture with different concentrations of GCA. (B) Changes in CD95 transcriptional levels following coculture with different concentrations of GCA. (C) Changes in CD95 transcriptional levels following coculture with the TGR5 inhibitor SBI-115. n = 3 per group. Statistical analysis: mean ± standard deviation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. Detection of GCA effects on different cell types. (A) Flow cytometry analysis of CD95+ EL4 cells proportion after coculture with different GCA concentrations. (B) qRT-PCR analysis of CD95 transcription levels in EL4 cells co-cultured with GCA. n = 3 per group. Statistical analysis: mean ± standard deviation. ∗P < .05; ∗∗P < .01. Materials and Methods Animal Experiments All animal experiments were conducted in accordance with institutional guidelines and approved by the Laboratory Animal Ethics Committee of Nankai University. Healthy C57BL/6 mice used to establish breeding colonies were obtained from Vital River Laboratory Animal Technology Co, Ltd. Experiments were performed with age-matched (6–10 weeks) and sex-matched mice (male or female mice were used). All animals were maintained under specific pathogen-free conditions at the Nankai University Laboratory Animal Center. Mice were maintained in groups of 4 per cage within temperature-controlled rooms (24 ± 2 °C) under a 12:12-hour light/dark cycle and had free access to sterile food and water. Antibiotic mixture depleted the gut microbiota of 6-week-old mice. Soluble amphotericin B (0.1 mg/mL; Solarbio), vancomycin hydrochloride (5 mg/mL; Aladdin), metronidazole (10 mg/mL; Sangon), and neomycin sulfate (10 mg/mL; Solarbio) were administered by oral gavage at 200 μL per mouse twice daily. Concurrently, ampicillin (1 mg/mL; Solarbio) was supplied to the drinking water for 2 weeks. Gut microbiota clearance rates were validated using 16S ribosomal RNA sequencing analysis. After 2 weeks, gavage was discontinued for 2 days, and normal drinking water was restored to prepare for bacterial inoculation. Fecal microbiota transplantation treatment Approximately 150 mg of fresh feces from donor mice was collected, homogenized, and resuspended in 3 mL of sterile, prechilled phosphate-buffered saline (PBS). After thorough mixing, the suspension was allowed to stand for 5 minutes to allow impurities to settle freely. The supernatant was then collected and administered by oral gavage (200 μL per mouse) to mice with depleted gut microbiota. This procedure was performed twice daily for 1 week. Bacteroides acidifaciens treatment BA was isolated from the feces of C57BL/6J mice in previous studies.36 BA was enriched for 36 hours under anaerobic conditions at 37 °C in sterile BLHK medium (brain–heart infusion 38.5 g/L, Na2S·9H2O 0.25 g/L, L-cysteine 0.25 g/L, hemin 5 μg/mL, vitamin K1 10 μg/mL). The culture was centrifuged (9200 rpm; 5 minutes), washed twice with sterile PBS, and resuspended to 3 × 109 colony-forming units in 2 mL sterile PBS. The live bacteria were then inoculated into mice with depleted gut microbiota (200 μL/mouse, twice daily, for a week). Free-water group Mice were provided a standard diet and allowed free access to sterile water, which was changed every 2 days for a period of 5 weeks. Free-drinking glycocholic acid solution group Mice were provided a normal diet and allowed free access to water containing 0.2% GCA (0.1 g sodium glycocholic acid powder dissolved in 50 mL sterile water) for free drinking. Water was changed every 2 days for 5 weeks. Establishment and Evaluation of Liver Disease Models Concanavalin A–induced hepatitis ConA (10 mg/kg; Sigma-Aldrich) was dissolved in sterile, prechilled PBS and administered via tail vein injection to C57BL/6 mice. Twelve hours postinjection, liver injury was analyzed by measuring serum ALT levels (Rongsheng) and performing histologic staining and statistical analysis of liver tissue. Liver tissue was obtained from mice (from the same anatomic location), fixed in 4% paraformaldehyde (Biosharp), and prepared into paraffin-embedded tissue sections. H&E staining, IHC staining, and TUNEL assays were performed to quantify areas of hepatic necrosis, CD95 expression, and hepatocyte apoptosis. α-Naphthyl isothiocyanate–induced cholestatic liver injury ANIT (Aladdin) was dissolved in prewarmed corn oil to prepare an 8 mg/mL solution. Mice received a single oral gavage of ANIT at a dose of 80 mg/kg body weight, corresponding to an administration volume of 0.1 mL/10 g body weight. Forty-eight hours after ANIT administration, the mice were euthanized, and blood and liver tissues were collected. Serum ALT, DBil (Biosino Bio-Technology & Science, Inc), and TBil (Biosino Bio-Technology & Science, Inc) levels were measured. Liver tissues were collected from the same anatomic region, fixed in 4% paraformaldehyde, embedded in paraffin, and subjected to H&E staining to assess histopathologic changes associated with cholestatic liver injury. Jo2-induced CD95-mediated liver injury CD95-mediated liver injury was induced by intraperitoneal injection of the agonistic anti-mouse CD95 monoclonal antibody Jo2 (0.2 μg/g body weight; BD Biosciences). Twelve hours after injection, blood samples were collected, and serum ALT levels were measured to evaluate liver injury. Cell Culture and Processing AML-12 cells Dulbecco’s Modified Eagle Medium/F12 medium (10% fetal bovine serum, 1% insulin–transferrin–selenium, 40 ng/mL dexamethasone, 1% penicillin-streptomycin). Hepa1-6, EL4, and HepG2 cells: Dulbecco’s Modified Eagle Medium basic medium (10% fetal bovine serum, 1% penicillin-streptomycin). Bile acids or drugs at various concentrations were dissolved in the medium for cell treatment. After 48 hours of incubation, cells were harvested for flow cytometry, qRT-PCR, and other experiments. Flow cytometry Cells were digested and stained with anti-mouse/human CD95 antibody (BioLegend) via flow cytometry. Analysis was performed using the BD FACS Calibur and Beckman Cytoflex systems. RNA extraction and quantitative reverse-transcription polymerase chain reaction Triquick Reagent (Solarbio) was used to extract total RNA from cells and tissues. Messenger RNA expression was analyzed using reverse transcription reagents and qPCR reagents (TaKaRa) on a Roche LightCycler 96 instrument. The internal control for normalization was glyceraldehyde 3-phosphate dehydrogenase. Primer sequences are shown in Table 2. Table 2Primers for qRT-PCR in This StudyGene namePrimer sequence (5′-3′)mGapdh ForwardTGGCCTTCCGTGTTCCTAC ReverseGAGTTGCTGTTGAAGTCGCAmCd95 ForwardGCGGGTTCGTGAAACTGATAA ReverseGCAAAATGGGCCTCCTTGATAmCreb5 ForwardAGGATCTTCTGCCGTCTTGAT ReverseGCGCAGCCTTCAGTCTCATmPou3f1 ForwardTCGAGGTGGGTGTCAAAGG ReverseGGCGCATAAACGTCGTCCAmDlx3 ForwardCACTGACCTGGGCTATTACAGC ReverseGAGATTGAACTGGTGGTGGTAGmEtv4 ForwardCGCACAGACTTCGCCTACG ReverseCAGACATCATCTGGGAATGGTCmJun ForwardCCTTCTACGACGATGCCCTC ReverseGGTTCAAGGTCATGCTCTGTTTNF-κB1 ForwardGGAGGCATGTTCGGTAGTGG ReverseCCCTGCGTTGGATTTCGTGmNr4a1 ForwardTTGAGTTCGGCAAGCCTACC ReverseGTGTACCCGTCCATGAAGGTGmRunx1 ForwardGCAGGCAACGATGAAAACTACT ReverseGCAACTTGTGGCGGATTTGTAmFosl1 ForwardATGTACCGAGACTACGGGGAA ReverseCTGCTGCTGTCGATGCTTGmStat5a ForwardCGCCAGATGCAAGTGTTGTAT ReverseTCCTGGGGATTATCCAAGTCAATmStat5b ForwardCGATGCCCTTCACCAGATG ReverseAGCTGGGTGGCCTTAATGTTCmSox9 ForwardCGTGCAGCACAAGAAAGACC ReverseGGACCCTGAGATTGCCCAGAmFos ForwardCGGGTTTCAACGCCGACTA ReverseTTGGCACTAGAGACGGACAGAmASBT ForwardGTCTGTCCCCCAAATGCAACT ReverseCACCCCATAGAAAACATCACCAmOSTα ForwardAGGCAGGACTCATATCAAACTTG ReverseTGAGGGCTATGTCCACTGGGmOSTβ ForwardGCATCCTGGCAAACAGAAATCG ReverseTCTTTCAACTCAGGTTCCCCChGAPDH ForwardACAACTTTGGTATCGTGGAAGG ReverseGCCATCACGCCACAGTTTChCD95 ForwardAGATTGTGTGATGAAGGACATGG ReverseTGTTGCTGGTGAGTGTGCATTqRT-PCR, quantitative reverse-transcription polymerase chain reaction. Primers for qRT-PCR in This Study qRT-PCR, quantitative reverse-transcription polymerase chain reaction. Reagents Reagents used were ConA (Sigma-Aldrich), Jo2 (BD Biosciences), ANIT (Aladdin), brain–heart infusion broth (Hopebio), L-cysteine (BBI Life Sciences), Na2S·9H2O (Yien), hemin (Hopebio), vitamin K1 (Hopebio), insulin–transferrin–selenium–X media supplement (100×) (Beyotime), dexamethasone (MedChemExpress), anti-mouse CD95 (BioLegend), anti-human CD95 (BioLegend), GCA (Aladdin), sodium CA (Solarbio), TUDCA (Aladdin), β-MCA (MedChemExpress), HCA (Aladdin), DCA (Aladdin), HDCA (Aladdin), THDCA (Aladdin), THCA (MedChemExpress), CDCA (Aladdin), TLCA (MedChemExpress), FXR receptor inhibitor DY-268 (MedChemExpress), S1PR2 receptor inhibitor JTE-013 (MedChemExpress), TGR5 receptor inhibitor SBI-115 (MedChemExpress), CAR receptor inhibitor CINPA1 (MedChemExpress), cAMP analog bucladesine sodium (MedChemExpress), PI3K inhibitor LY294002 (MedChemExpress), recombinant FGF15 protein (MedChemExpress), FGFR4 inhibitor BLU9931 (MedChemExpress), CD95 antibody (BOSTER), TUNEL Detection Kit (Servicebio), FRA1 rabbit antibody (Cell Signaling Technology), and CREB5 rabbit antibody (ABclonal). CRISPR/Cas 9 Single guide RNA was designed using https://crispor.gi.ucsc.edu. Primers were annealed, the CRISPRv2 vector was digested, and the digestion products were run on a gel. The linearized vector was recovered using the Gel Extraction Kit (Omega). After ligating the annealed products with the linearized vector, the ligation products were transformed into DH5α-competent cells (TIANGEN). Plates were spread, and single colonies were picked for sequencing verification. Plasmid was extracted using a plasmid extraction kit (TIANGEN). Viruses were packaged in 293T cells, and viral lysates were collected. Target cells were infected, and single-clone cells were picked for expansion and subsequent experiments. Chromatin Immunoprecipitation Assay Experiments were performed in accordance with the manufacturer’s instructions for the SimpleChIP Plus Enzymatic Chromatin IP Kit (Magnetic Beads) (Cell Signaling Technology). Cells were cross-linked with 1% formaldehyde for 10 minutes, and the reaction was quenched with glycine. Cells were then harvested and subjected to nuclear isolation. Chromatin was digested with micrococcal nuclease and further processed by sonication to generate chromatin fragments. An aliquot of each sample was reserved as the input fraction, whereas the remainder was incubated overnight at 4 °C with antibodies against CREB5 (ABclonal) or FOSL1/FRA1 (Cell Signaling Technology), with normal IgG serving as a negative control. Protein G magnetic beads were then added to enrich immunocomplexes. After sequential washing, elution, reverse cross-linking, and proteinase K digestion, DNA was purified. The enrichment of the CD95 promoter region was finally determined by qPCR. Primer sequences are shown in Table 3. Table 3Primers for Chromatin Immunoprecipitation qPCRPrimer namePrimer sequence (5’-3’)mCd95-site 1-FCCCCAGTCTCTGTGAGGTCAmCd95-site 1-RAGAGGGGATGGACAAGACCAmCd95-site 2-FCTGTGCAGTGGTGAGTCAGTmCd95-site 2-RCCCAAGGGTTGCGAGAAAGTqPCR, quantitative polymerase chain reaction. Primers for Chromatin Immunoprecipitation qPCR qPCR, quantitative polymerase chain reaction. Transcriptomics Analysis Extraction of liver tissue or cellular RNA RNA integrity and RNA content were precisely assessed using Agilent 2100 Bioanalyzer (Agilent Technologies), using Oligo dT magnetic beads to enrich messenger RNA from total RNA. After a series of processing steps, the library was established. Quantification was performed using a Qubit instrument and real-time qPCR, with fragment size distribution analyzed via bioanalyzer. After quality control, sequencing libraries were pooled in proportion to their molar concentrations and subjected to Illumina sequencing-by-synthesis. Fluorescent incorporations were base-called to generate nucleotide sequences. Raw reads were processed with fastp to trim adapters and remove low-quality bases, yielding clean reads. The reference genome was indexed with HISAT2 (v2.0.5), and paired-end reads were aligned to the index. Gene-level quantification was performed with featureCounts (v1.5.0-p3) to obtain read counts and fragments per kilobase of transcript per million mapped reads. Differential expression between groups was assessed using DESeq2 (v1.20.0). Functional annotation, pathway enrichment analysis, and bioinformatics studies such as protein–protein interaction networks for differentially expressed genes are conducted. Metabolomics Analysis Untargeted metabolomics We accurately weighed 50 mg of sample and added the prechilled extraction solvent (water:acetonitrile:methanol = 1:2:2, v/v/v, concentration of internal standard 2 mg/L, 1000 μL). We then added ceramic homogenization beads and homogenized using a homogenizer (45 Hz, 10 minutes) and sonicated for 10 minutes in an ice-water bath. We incubated the compound at −20 °C for 1 hour to precipitate the protein. After centrifugation (4 °C, 12,000 rpm, 15 minutes), we aspirated 500 μL of supernatant and dried it in a vacuum concentrator. We resuspended the metabolite residue in resuspension buffer (water:acetonitrile = 1:1, v/v), vortexed, and sonicated again. After centrifugation under identical conditions, we transferred the supernatant to an injection vessel for instrument analysis. Simultaneously, samples were pooled from each batch to prepare quality control samples. The liquid chromatography–mass spectroscopy (LC-MS) systems used were the Acquity I-Class PLUS and Xevo G2-XS QTof (Waters). The column used was the Acquity UPLC HSS T3 column (Waters). MassLynx V4.2 raw data were subjected to peak extraction and alignment using Progenesis QI software. Compound identification was performed by querying the METLIN database and BMG’s proprietary database, with theoretical fragment matching allowing for a mass deviation of up to 100 ppm. Targeted metabolomics Liver: After grinding liver tissue samples with liquid nitrogen, we dissolved them in mass spectrometry water. One hundred microliters of the dilution solution was taken and mixed with 300 μL of the precipitating agent containing the mixed internal standard (methanol:acetonitrile = 2:8) to precipitate the protein. The reaction mixture was incubated on ice for 30 minutes, then centrifuged at 4 °C at 12,000 rpm. The entire supernatant was obtained for LC-MS analysis. Feces: A total of 100 mg of fresh or frozen fecal samples was collected and mixed with a prechilled extraction solution at a fixed mass-to-volume ratio. Samples were thoroughly homogenized and vortexed to ensure consistent extraction of bile acids. After incubation at low temperature to precipitate proteins and particulate impurities, the mixture was centrifuged at 12,000 rpm at 4 °C, and the entire resulting supernatant was subjected to LC-MS analysis. Bile acid quantification: Ultra–high-performance liquid chromatography coupled with tandem mass spectrometry (ExionLC AD UHPLC-QTRAP 6500+ system, AB SCIEX CorpA). Separation was performed using ACQUITY UPLC BEH C18 column (Waters; 50 °C). LC-MS analysis was performed using bile acid standards of known concentrations to determine bile acid concentrations in the samples. Statistics Statistical analysis was performed using ImageJ, FlowJo v10.9.0, SPSS, R, and GraphPad Prism v10.1.2. Data are presented as mean ± standard deviation. All experiments were performed in at least 3 independent replicates. Differences between groups were analyzed using unpaired t tests, Mann–Whitney U tests or 1-way analysis of variance (∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001). Acknowledgments CRediT Authorship Contributions Xiang Zhang (Conceptualization: Lead; Data curation: Lead; Formal analysis: Lead; Investigation: Lead; Methodology: Lead; Software: Lead; Validation: Lead; Visualization: Lead; Writing – original draft: Lead) Cheng Zhang (Formal analysis: Supporting; Investigation: Supporting; Methodology: Equal; Software: Equal; Validation: Equal) Yi Zhou (Investigation: Supporting; Methodology: Equal; Software: Equal; Validation: Equal) Yu Han (Investigation: Supporting; Methodology: Equal; Software: Equal; Validation: Equal) Yawei Wang (Methodology: Supporting; Software: Equal; Validation: Equal) Hesuiyuan Wang (Methodology: Supporting; Software: Equal; Validation: Equal) Mengyan Yang (Methodology: Supporting; Software: Supporting) Qinhan Wu (Methodology: Supporting; Software: Supporting) Feng Bai (Methodology: Supporting; Validation: Supporting) Fanfan Qu (Funding acquisition: Lead; Project administration: Lead; Supervision: Supporting) Jianbin Ji (Methodology: Supporting; Supervision: Supporting) Hongru Zhang (Methodology: Supporting; Supervision: Supporting) Zhenglu Wang (Funding acquisition: Lead; Project administration: Lead; Supervision: Lead) Yuebing Wang (Supervision: Supporting; Writing – review & editing: Supporting) Haijin Xu (Supervision: Supporting; Writing – review & editing: Supporting) Zhenghua Zhou (Supervision: Supporting; Writing – review & editing: Supporting) Mingqiang Qiao (Supervision: Supporting; Writing – review & editing: Supporting) Zhenzhou Wu (Conceptualization: Lead; Data curation: Lead; Formal analysis: Lead; Funding acquisition: Lead; Investigation: Lead; Methodology: Lead; Project administration: Lead; Supervision: Lead; Writing – review & editing: Lead) Footnotes Conflicts of interest The authors disclose no conflicts. Funding This work was supported by the National Natural Science Foundation of China (Nos. 32470952, 32170715 and 82374351, 82505442). This work was also supported by the Natural Science Foundation of Tianjin City (22JCYBJC01230). Data Availability All data are deposited in Mendeley Data (https://data.mendeley.com/preview/r776dnx8jc?a=5484d1f0-26fb-4439-9ca2-2c00395b51e8). All data are available from the corresponding author upon reasonable request. References