Activation of the TGR5/cAMP/PKA/CREB axis in cholangiocytes mediates epithelial-mesenchymal transition and fibrosis in hepatolithiasis

Abstract

Background

The molecular mechanisms linking bile acid signaling to biliary fibrosis in hepatolithiasis are unclear. We investigated the role of the bile acid receptor TGR5 in driving fibrotic progression.

Methods

We employed an integrated strategy. TGR5 expression was assessed in human hepatolithiasis tissues and bioinformatics databases. A hepatolithiasis mouse model was compared with TGR5-knockout (TGR5-/-) mice, analyzing lipid metabolism, cholesterol transport, and fibrosis markers via ELISA, qPCR, and immunohistochemistry. In vitro, human intrahepatic biliary epithelial cells (HIBECs) were treated with bile acids, TGR5 agonists/inhibitors, and cAMP modulators. We assessed the cAMP/PKA/CREB pathway, epithelial-mesenchymal transition (EMT), and fibrosis.

Results

TGR5 was significantly overexpressed in hepatolithiasis tissues. Genetic deletion of TGR5 in mice alleviated metabolic disturbances and markedly reduced fibrosis markers (TGF-β1, α-SMA, Collagen I). In cholangiocytes, TGR5 activation stimulated the cAMP/PKA/CREB signaling cascade, promoting EMT and fibrosis. These pro-fibrotic effects were reversed by inhibiting TGR5 or cAMP. Mechanistically, TGR5-driven cAMP signaling induced CREB phosphorylation, leading to upregulated TGF-β1 expression, which sustains the fibrotic phenotype.

Conclusion

This study defines a critical pathogenic axis in hepatolithiasis, whereby TGR5 activation in cholangiocytes triggers cAMP/PKA/CREB signaling to promote EMT and biliary fibrosis via TGF-β1. Our findings directly implicate TGR5 as a central therapeutic target for mitigating fibrosis in this disease.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00783-7.

Introduction

Hepatolithiasis is a chronic biliary disorder commonly seen in Asian populations, with particularly high prevalence in China and Southeast Asian countries [1, 2]. Clinically, it is characterized by recurrent right upper quadrant pain, chills with fever, and cholestatic jaundice. In severe cases, it can lead to biliary obstruction, ductal dilation, and hepatic dysfunction [2]. More critically, hepatolithiasis is frequently accompanied by biliary stricture and biliary fibrosis, which further contribute to abnormal proliferation of the biliary epithelium and may eventually progress to cholangiocarcinoma, thereby posing a serious threat to long-term patient outcomes [3–5]. Currently, the mainstay of treatment includes surgery, endoscopic drainage, and antimicrobial therapy. However, high recurrence rates and the persistent progression of fibrosis remain significant clinical challenges [6, 7]. Although ongoing research has improved our understanding of the mechanisms underlying biliary stone formation, the molecular basis of biliary fibrosis secondary to hepatolithiasis remains insufficiently explored, thereby limiting the potential for upstream therapeutic interventions. As such, elucidating the pathogenic mechanisms at the cellular and molecular levels is urgently needed to identify novel diagnostic markers and therapeutic targets.

Biliary fibrosis is a common secondary tissue alteration in patients with hepatolithiasis and constitutes a critical pathological basis for biliary dysfunction and disease progression [8]. Evidence suggests that, under the influence of inflammatory stimuli, bile acid accumulation, and oxidative stress, cholangiocytes undergo an epithelial-mesenchymal transition (EMT), a process characterized by the loss of cell polarity, disruption of intercellular adhesion, and an enhanced migratory capacity. This transition is accompanied by excessive secretion of extracellular matrix components such as collagen and α-smooth muscle actin (α-SMA), ultimately leading to fibrotic deposition [9–11]. EMT is recognized as a key mechanism driving fibrosis in various biliary diseases; however, the specific inducers and regulatory pathways of EMT in the context of hepatolithiasis remain poorly defined [8, 12]. In particular, the upstream signals responsible for sustained EMT activation in cholangiocytes have become a focal point of current research. Clarifying the signaling pathways governing this process may facilitate the identification of promising pharmacological targets to halt fibrosis progression and prevent malignant transformation.

TGR5 (G-protein-coupled bile acid receptor 1) is a bile acid-sensing receptor that has recently been recognized for its critical regulatory roles in hepatobiliary diseases [13–15]. It is broadly expressed in cholangiocytes, hepatocytes, and immune cells, where it responds to various bile acid signals and modulates intracellular cAMP levels, thereby activating classical signaling pathways such as PKA/CREB [16, 17]. Previous studies have shown that TGR5 exerts bidirectional effects in cholestatic liver disease, hepatic fibrosis, and inflammatory responses—alleviating certain metabolic imbalances while, under specific pathological conditions, potentially promoting inflammation or fibrotic progression [14, 18]. However, whether TGR5 plays a role in hepatolithiasis-induced biliary fibrosis, particularly through regulation of the cAMP/PKA/CREB pathway and subsequent modulation of TGF-β1 expression and EMT, remains inadequately investigated. Given TGR5’s close association with bile acid metabolism, it may serve as a critical “signaling hub” linking stone formation and biliary epithelial injury. This warrants further exploration of its pathogenic potential and molecular mechanisms.

Building on this background, the present study aims to systematically elucidate the pathogenic role of TGR5 in hepatolithiasis-associated biliary fibrosis and to determine whether it contributes to fibrosis progression by activating the cAMP/PKA/CREB signaling axis, leading to the upregulation of TGF-β1 expression and subsequent induction of EMT in cholangiocytes. Scientifically, the study seeks to elucidate the functional role of bile acid signaling in cholangiocyte fibrogenesis and to address the current gap in understanding TGR5-mediated mechanisms in stone-related biliary disorders. Clinically, this work may establish the potential of TGR5 as a diagnostic biomarker and therapeutic target for biliary fibrosis, offering a theoretical basis for early detection, precise classification, and targeted intervention in chronic hepatobiliary diseases. Moreover, the signaling pathways identified herein may provide a foundation for the development of TGR5-targeted or downstream-directed molecular therapeutics, thereby expanding treatment strategies and improving long-term outcomes for affected patients.

Materials and methods

Clinical sample collection and ethics statement

All human liver tissue samples used in this study were obtained from the Department of Hepatobiliary Surgery at our institution. The samples included liver tissues from patients with hepatolithiasis accompanied by biliary fibrosis (stone group) and normal liver tissues from patients without hepatolithiasis (non-stone group), totaling n = 6 cases, comprising 3 samples in each group. All tissues were freshly collected during surgery, pathologically confirmed, and immediately snap-frozen in liquid nitrogen for subsequent staining and protein extraction. Written informed consent was obtained from all patients before surgery. This study was approved by the Institutional Medical Ethics Committee of Zunyi Medical University, approval number:【2019】1–058.

Source and processing of bioinformatics data

The study utilized publicly available RNA sequencing (RNA-seq) data (GSE202479) from human gallbladder tissue RNA-seq, which were downloaded from the official GEO portal: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE202479. The analysis was based on the authors’ curated normalized expression matrix. A total of seven gallbladder tissue samples were included and grouped by disease status: Control (normal gallbladder): N8_count, N10_count, N20_count (n = 3); Model (chronic cholecystitis caused by gallstones): Y8_count, Y12_count, Y13_count, Y16_count (n = 4). The dataset did not involve drug or genetic interventions; all analyses were case-control comparisons based on untreated tissue samples. Throughout the manuscript, gene nomenclature is standardized: GPBAR1 is consistently referred to as “TGR5”, and statistical comparisons are uniformly reported as Model versus Control. Before analysis, row names were converted to current HGNC-approved symbols (uppercase, whitespace removed). Non-numeric entries were treated as missing values; genes with ≥ 50% missingness were excluded; genes with zero variance were removed to prevent unstable statistical estimates. All data processing preserved the original column order and group labels. To assess overall data quality and between-group separation, the study generated and inspected principal component analysis (PCA) plots (reporting the variance explained by PC1 and PC2), sample-to-sample Pearson correlation heatmaps with hierarchical clustering, and expression density and boxplots. Potential outliers were identified using Mahalanobis distance thresholds (mean ± 3 standard deviations [SD]) and leave-one-out (LOO) PCA diagnostics. As the public dataset lacked explicit batch labels, any future integration of multi-center samples applied ComBat or ComBat-seq for batch correction and adjusted for confounders (e.g., age, gender, inflammation score) using linear model covariates. Sensitivity analyses were conducted to confirm robustness. All analyses were performed directly on the normalized expression values provided by the original authors; to avoid redundant transformations, no additional normalization was applied.

Differential gene expression analysis

After importing the normalized expression matrix into the Xiantao Academic platform, group comparisons were set as Model vs. Control. The platform’s default differential analysis workflow was used, which automatically applies commonly adopted statistical frameworks (e.g., limma or DESeq2) to construct models and perform multiple testing corrections. Statistical significance was determined by the Benjamini-Hochberg (BH)-adjusted false discovery rate (FDR, p.adjust [padj]), with the primary threshold set at padj < 0.05. For visualizations such as volcano plots, an additional threshold of |log₂FC| ≥ 0.58 (approximately a 1.5-fold change) was optionally applied to enhance interpretability. All statistical comparisons were consistently conducted as Model relative to Control. Output fields included gene symbols, log₂fold change (log₂FC), p-values, and padj, and were directly used for downstream visualizations (e.g., volcano plots, heatmaps) and enrichment analyses.

Correlation analysis between TGR5 and fibrosis-related genes

Expression values of TGR5 and fibrosis-related genes (COL1A1, COL3A1, FN1, VIM, TGFB1, MMP2, and MMP9) were extracted from the GSE202479 dataset. Spearman’s rank correlation analysis was performed separately in the control group and the model group to assess the correlation coefficients between TGR5 and each fibrosis-related gene. Because the sample size in each group was relatively small, the correlation coefficient (R) was used to describe the monotonic trend of co-variation, rather than for formal statistical significance inference. Scatter plots with linear regression lines were used to visualize the co-variation patterns under disease conditions.

Gene ontology (GO)/Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis

Over-representation analysis (ORA) was performed on the Xiantao Academic platform using the list of differentially expressed genes (DEGs). Upregulated and downregulated genes were derived from the Model vs. Control comparison using the primary significance threshold of padj < 0.05, optionally combined with |log₂FC| ≥ 0.58. The background set consisted of all genes retained in the expression matrix after quality control. KEGG and GO Biological Process (BP) categories were analyzed separately. Statistical significance was assessed using BH-padj (FDR). For each pathway or term, -log₁₀(padj), GeneRatio (k/n), and gene count were reported. The top 15 enriched terms from each category were presented in the main text. Results were exported as GOKEGG.xlsx from the platform. For reproducibility, the same analyses can be replicated using the clusterProfiler package with enrichKEGG(organism = “hsa”, pAdjustMethod = “BH”, minGSSize = 10, universe = QC genes) and enrichGO(ont = “BP”, OrgDb = org.Hs.eg.db). Gene ID-to-symbol conversion followed HGNC annotations and the org.Hs.eg.db database.

Key term analysis and gene-level differential expression

In this study, we located the KEGG: Bile secretion term within GOKEGG.xlsx and extracted its associated gene list (automatically recognizing the geneID or gene column, with support for delimiters such as /, ;, or, ). The intersection of this gene list with the expression matrix was computed, after which the ΔMean for each gene was calculated as mean(Model)-mean(Control). Genes were ranked by this difference for visualization and interpretation within the pathway. Where appropriate, permutation tests (two-sided, 10,000 iterations) were performed for gene-level comparisons within the term, followed by BH-FDR correction within the same gene family. The number of matched genes (Count) and the GeneRatio were retained to ensure analytical reproducibility.

Module/axis scoring construction

To evaluate sample-level pathway or axis activity, we independently scored predefined gene sets using the same expression matrix. The gene sets included:

CREB axis = {PRKACA, PRKACB, CREB1}

Fibrosis = {COL1A1, COL3A1, FN1, VIM, TGFB1, MMP2, MMP9}

Bile secretion = intersection of the KEGG Bile secretion term with the genes present in the expression matrix (modules with < 3 intersecting genes were designated as NA and excluded from statistical analysis).

The scoring procedure consisted of the following steps:

Step 1 (Standardization): For each gene g, a row-wise z-score was computed across all 7 samples:

Where x_{g, i} represents the normalized expression value of gene g in sample i, with n = 7; when σ_g = 0, it is set to 1 (resulting in a z-score of 0). In cases where individual samples have missing values, the missing entry is first imputed using the mean expression of that gene across the remaining samples, followed by a z-score calculation.

Step 2 (Sample Score): The module score for sample i was calculated as the arithmetic mean of z-scores for all genes in the module:

A higher score indicates global upregulation of the corresponding module in that sample. No sign flipping was applied to putative negative regulators; however, if such genes are explicitly defined in future analyses, their z-scores can be multiplied by -1 before averaging to preserve directional consistency.

Sample-wise scores for the CREB axis, fibrosis module, and bile secretion module were exported for visualization and statistical analysis. Group comparisons were performed using the two-sided Mann-Whitney U test, with Cliff’s δ (defined as Model-Control) and its 95% confidence interval (CI) reported (BCa bootstrap, 10,000 iterations). Correlation analyses (e.g., TGR5 expression vs. CREB axis score, CREB axis score vs. fibrosis/bile secretion score) were assessed using Spearman’s correlation coefficient (ρ). When evaluating multiple correlations within the same family, the BH-FDR correction was applied.

Receptor profiling and side-by-side comparison

The target receptors included TGR5 (GPBAR1), S1PR2, FXR (NR1H4), PXR (NR1I2), CAR (NR1I3), and VDR. For each gene, ΔMean (Model-Control) was calculated, and receptor-level “waterfall” or “forest” plots were generated. BH-FDR correction was applied within the receptor family. In parallel, a side-by-side comparison of TGR5 and S1PR2 was performed using boxplots overlaid with individual data points. The Mann-Whitney U test and Cliff’s δ (with 95% CIs) were used to evaluate both effect direction and magnitude. No cross-gene numerical comparisons were conducted, and GPBAR1 was consistently referred to as “TGR5” throughout the analysis.

Correlation and integrative analysis

Spearman’s rank correlation (two-sided) was used to assess three predefined relationships: TGR5 expression vs. CREB axis score, CREB axis score vs. fibrosis score, and CREB axis score vs. bile secretion score. To enhance robustness in this small-sample setting, permutation testing (10,000 iterations) was used to estimate empirical p-values, and BCa bootstrapping (10,000 iterations) was applied to calculate 95% CIs for the ρ. These three correlations were treated as a single statistical family, and p-values were adjusted using the BH-FDR method (reported as q-values). Regression lines were included for visualization only and were not used for inferential purposes.

Robustness and sensitivity analysis

LOO analysis was performed by iteratively removing each sample and recalculating TGR5 ΔMean, CREB axis score Δ, fibrosis score Δ, and all three correlation coefficients. The absolute deviation and coefficient of variation (CV) from the full-sample estimates were reported. A CV < 0.3 with consistent directionality was considered indicative of insensitivity to individual samples. In cases of rank ties or extreme values, exact rank-based tests or mid-p adjustments were additionally employed as evidence of statistical robustness.

Cell culture and treatment

Human intrahepatic biliary epithelial cells (HIBECs; ScienCell, Cat# 4510) were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin, and maintained in a humidified incubator at 37 °C with 5% CO₂. When cell confluence reached approximately 70–80%, cells were treated with 500 µM taurodeoxycholic acid (TDCA; Sigma, Cat# T4009) for 24 h to establish a bile acid-induced biliary fibrosis model. Each condition was tested in triplicate, and the entire experiment was independently repeated twice to ensure reproducibility and data reliability.

H-89 and forskolin treatment

Based on the TDCA-induced cholangiocyte fibrosis cell model, additional pharmacological intervention groups were established as follows (3 technical replicates per group; each experiment was independently repeated twice): Control group (Control): no TDCA induction and no drug treatment; Model group (Model): treated with 500 µM TDCA for 24 h; Forskolin treatment group (Model + Forskolin): cells were co-treated with 500 µM TDCA and 50 µM forskolin (HY-15371, MCE) for 24 h [19]; H-89 treatment group (Model + H-89): cells were co-treated with 500 µM TDCA and 100 µM H-89 (HY-15979, MCE) for 24 h [20]; Combined treatment group (Model + Forskolin + H-89): cells were co-treated with 500 µM TDCA, 50 µM forskolin, and 100 µM H-89 for 24 h; TGR5 inhibitor combination group (Model + SBI-115 + Forskolin): cells were co-treated with 500 µM TDCA, 10 µM SBI-115 (TGR5 inhibitor), and 50 µM forskolin for 24 h.

CCK-8 assay

HIBECs were digested and seeded into 96-well plates at a density of 1 × 10⁴ cells per well, with three technical replicates per group. After incubation at 37 °C with 5% CO₂ for 4–6 h to allow adherence, cells were treated with various concentrations of the following compounds: TDCA (Sigma, Cat# T4009; 250, 500, 1000 µM), TGR5 agonist hyodeoxycholic acid (HDCA; MedChemExpress, Cat# HY-113554; 50, 100, 200 µM), and TGR5 inhibitor SBI-115 (Sigma, Cat# SML1790; 5, 10, 20 µM). A drug-free control group and blank wells were included for comparison.

After 24 h of treatment, the medium was removed and replaced with 100 µL of serum-free medium containing 10% CCK-8 reagent (Dojindo, Cat# CK04), followed by a 2-hour incubation. Optical density (OD) was measured at 450 nm using a microplate reader (BioTek ELx800), and the background absorbance from blank wells was subtracted before analysis. All experiments were independently repeated twice to ensure consistency of the results.

Cell co-culture

Cholangiocytes: Human intrahepatic biliary epithelial cells (HiBECs; ScienCell, Cat#4510) were cultured under the same conditions as described above in RPMI-1640 supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37 °C in 5% CO₂.

Hepatic stellate cells: Human LX-2 cells (CL-0560, Wuhan Procell Life Science & Technology Co., Ltd.) were cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37 °C in 5% CO₂. Cells in the logarithmic growth phase were used for experiments.

A Transwell co-culture system with 0.4 μm pore size inserts (Corning, Cat#3412) was used. HiBECs (2 × 10⁴ cells/well) were seeded in the upper chamber, and LX-2 cells (5 × 10⁴ cells/well) were seeded in the lower chamber, ensuring that the two cell types had no direct contact and interacted only through secreted factors, including cytokines and small signaling molecules. After HiBECs were pretreated according to the experimental grouping for 24 h, the upper inserts were transferred into the lower wells containing LX-2 cells and co-cultured for an additional 48 h.

The specific groups were as follows: Control co-culture group (HiBEC + LX-2): untreated HiBECs + LX-2 cells. Model co-culture group (Model + LX-2): TDCA (500 µM)-treated HiBECs + LX-2 cells. TGR5 agonist co-culture group (HDCA + LX-2): TDCA (500 µM) + HDCA (100 µM)-treated HiBECs + LX-2 cells. TGR5 inhibitor co-culture group (SBI-115 + LX-2): TDCA (500 µM) + SBI-115 (10 µM)-treated HiBECs + LX-2 cells [21].

Immunofluorescence staining

HIBECs were seeded onto coverslips placed in 24-well plates. After cell adherence, the cells were washed three times with PBS (3 min each), fixed with 4% paraformaldehyde (Solarbio, Cat# P0099) for 15 min, and washed again three times with PBS. Permeabilization was performed using 0.5% Triton X-100 (Solarbio, Cat# T8200) in PBS for 20 min at room temperature. Following three additional PBS washes, the excess liquid around the coverslips was gently blotted using absorbent paper, and 5% BSA (Solarbio, Cat# A8020) was applied for blocking in a 37 °C incubator for 30 min.

For patient samples and mouse liver tissues, specimens were fixed with 4% paraformaldehyde, embedded in OCT, and cut into 4 μm serial sections using a cryostat. Sections were washed three times with PBS (3 min each), fixed again with 4% paraformaldehyde for 15 min, and washed three times with PBS. They were then permeabilized with 0.5% Triton X-100 at room temperature for 20 min. After three PBS washes, the area around the sections was dried with absorbent paper, and the sections were blocked with 5% BSA for 30 min at 37 °C [22].

After blocking, the BSA solution was removed without additional washing, and primary antibody working solutions were added to cover the cells. The primary antibodies included: TGR5 (Abcam, Cat# ab72608, 1:200), α-SMA (Proteintech, Cat# 14395-1-AP, 1:200), CK19 (Abcam, Cat# ab133496, 1:200) and ki67(Abcam, Cat#ab16667, 1:200). Cells were incubated overnight at 4 °C. On the following day, cells were washed three times with PBS (5 min each), then incubated at 37 °C for 30 min with fluorescent secondary antibodies (e.g., Alexa Fluor 488/594 conjugates; Invitrogen, 1:500). After additional PBS washes, DAPI staining solution (Beyotime, Cat# C1002, 1 µg/mL) was added and incubated in the dark for 3 min. Finally, the coverslips were rinsed and mounted using 50% glycerol.

Images were captured using a fluorescence microscope (Olympus CX43), with excitation wavelengths adjusted according to the respective fluorophores of the antibodies. Fluorescence intensity analysis was performed using ImageJ software. Each group included three technical replicates, and all experiments were independently repeated at least twice to ensure reproducibility.

Animal model construction

This study was approved by the institutional ethics committee of Zunyi Medical University(Approval No.: KLLY-2019-165). Male C57BL/6J mice (8 weeks old) and TGR5 knockout (TGR5^-/-) mice (provided by GemPharmatech, Jiangsu, License No.: SCXK (Su) 2018-0008) were used. All animals were housed in a specific pathogen-free (SPF) facility under controlled conditions (temperature: 20–26 °C; humidity: 40%-70%; 12 h light/dark cycle) with ad libitum access to food and water. After one week of acclimatization, model induction was initiated.

The mice were randomly divided into four groups (n = 6 per group): (1) Control group (C57BL/6J, fed a standard diet); (2) Model group (C57BL/6J, fed a lithogenic diet for 8 consecutive weeks); (3) TGR5^-/- control group (fed a standard diet); (4) TGR5^-/- model group (fed a lithogenic diet for 8 consecutive weeks). The lithogenic diet was custom-formulated by the institutional animal center and consisted of 2.5% cholesterol, 0.8% bile acids, 16% fat, and 80.7% basal feed.

At week 8, intrahepatic hepatolithiasis modeling surgery was performed. Mice were anesthetized with 1% pentobarbital sodium (intraperitoneal injection, 40 mg/kg). Following shaving and disinfection, a ~ 2 cm midline abdominal incision was made to expose the hepatic hilum. The bile duct was traced to the duodenal junction, and the portal vein and left hepatic duct were bluntly dissected. A 0.08 mm metal wire was used to lift the left hepatic duct, which was then ligated below the wire using 7 − 0 surgical sutures; the wire was removed to complete the ligation. The abdominal wall was closed in layers, and gentamicin (4 mg/kg, intraperitoneal) was administered postoperatively for analgesia and infection prevention.

Following surgery, mice were maintained until postoperative week 8. At the endpoint, under terminal anesthesia, serum, bile, liver tissue, and intrahepatic bile ducts were collected for subsequent analyses.

Hematoxylin and eosin (H&E) staining

Mouse liver tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned into consecutive slices of 4 μm thickness using a microtome. Sections were baked at 60 °C for 1 h, followed by deparaffinization in xylene and rehydration through a graded ethanol series. The slides were stained with hematoxylin solution (ZSGB-BIO, Cat# ZLI-9610) for 5 min, rinsed with tap water, and differentiated in 1% hydrochloric acid ethanol for 15 s. After another rinse, bluing was performed using Scott’s solution (Solarbio, Cat# G1865) for 3 min. Sections were then counterstained with eosin (Solarbio, Cat# G1100) for 3 min, dehydrated through 70%, 80%, 90%, and 100% ethanol, cleared in xylene, and mounted with neutral resin (CWBIO, Cat# CW0136).

Stained sections were examined under a fluorescence microscope (Olympus CX43) to evaluate hepatocyte morphology, portal area architecture, bile duct dilation, inflammatory cell infiltration, and necrosis. For each mouse, liver samples were collected from the left lobe, and three consecutive sections were stained to ensure consistency of results.

Masson’s trichrome staining

Mouse liver tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 4 μm. Following baking at 60 °C for 1 h, the sections were deparaffinized with xylene and rehydrated using a graded ethanol series. Masson’s Trichrome Staining Kit (Servicebio, Cat# G1006) was used in accordance with the manufacturer’s protocol. The main steps included hematoxylin staining, acidic ethanol differentiation, ponceau acid fuchsin staining, phosphomolybdic acid differentiation, and aniline blue counterstaining. After staining, sections were dehydrated with absolute ethanol, cleared in xylene, and mounted using neutral resin (CWBIO, Cat# CW0136).

In the stained sections, collagen fibers appeared blue, nuclei black, and muscle and cytoplasmic components red. Observations were performed using an Olympus CX43 microscope. For each group, a minimum of three consecutive liver sections were stained to assess the extent of peribiliary fibrosis and collagen deposition in intrahepatic bile ducts.

ELISA and biochemical assays

Bile samples were first clarified by low-speed centrifugation and then diluted at a ratio of 1:10 for the measurement of triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C). Corresponding assay kits were purchased from Nanjing Jiancheng Bioengineering Institute (TG: Cat# A110-1-1; HDL-C: Cat# A112-1-1; LDL-C: Cat# A113-1-1). All measurements were based on enzymatic colorimetric reactions and conducted according to the manufacturer’s protocols. Absorbance was recorded at 450 nm using a fully automated microplate reader (SuPerMax 3100, Shanghai Suntrip), and concentrations were calculated based on standard curves.

Approximately 50 mg of liver tissue was homogenized (1:10 weight-to-volume ratio in PBS), and the supernatant was collected after centrifugation for the determination of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels. The corresponding assay kits were obtained from Shandong Box Biological Technology Co. (AST: Cat# 70110; ALT: Cat# 70111). The assays were performed using a fully automated biochemical analyzer (BK-600, Shandong Box) based on the colorimetric method, and results were expressed in U/L.

All samples were tested in triplicate, and the experiments were independently repeated twice to ensure data consistency. The manufacturers’ instructions were strictly followed, and all procedures and sample handling were conducted in a blinded manner throughout the entire process.

Total bile acid measurement

Total bile acids (TBA) were measured in undiluted mouse plasma samples using a bile acid assay kit (Abcam, Cambridge, UK; Cat#ab239702) according to the manufacturer’s instructions. Absorbance was measured using a Tecan Infinite^® M1000 Pro microplate reader (Männedorf, Switzerland) following the standard protocol [23].

siRNA transfection

Mouse TGR5-specific siRNA oligonucleotides were obtained from Invitrogen (Carlsbad, CA, USA). The sequences of TGR5 siRNA were as follows: sense, 5′-CAC UUC UUC UUC CUC UTT-3′; antisense, 5′-AGA AGG AGU AGG GTT-3′. A double-stranded RNA oligonucleotide, NNGCGCGCUUGUGUGAUGAUCA (5′-3′), was used as the negative control siRNA.

HIBECs were seeded into 6-well plates and allowed to adhere overnight. At the time of transfection, cell confluence reached approximately 80–85%. According to the manufacturer’s instructions, siRNA was transfected into HIBECs using Lipofectamine 3000 (L3000001, Invitrogen Corp., Carlsbad, CA, USA) at a final concentration of 20 µM. Cells were cultured for 48 h after transfection before being used for subsequent biochemical analyses [24].

Quantitative polymerase chain reaction (qPCR) detection

Total RNA was extracted using Trizol reagent (Invitrogen, Cat# 15596026). RNA concentration and purity were assessed using a microvolume spectrophotometer (NanoPhotometer NP80, IMPLEN), with the OD260/280 ratio serving as an indicator of purity. A total of 1 µg RNA was reverse-transcribed into cDNA using the HiScript II Q RT SuperMix (with gDNA wiper, Vazyme, Cat# R223-01). qPCR was performed using ChamQ SYBR qPCR Master Mix (Vazyme, Cat# Q711-02) in a 20 µL reaction system containing 10 µL SYBR Mix, 0.4 µL forward primer (10 µM), 0.4 µL reverse primer (10 µM), 1 µL cDNA template, and 8.2 µL nuclease-free water. Amplification was carried out on a CFX Connect™ Real-Time PCR System (Bio-Rad) under the following conditions: initial denaturation at 95 °C for 10 min; followed by 40 cycles of 95 °C for 10 s, 58 °C for 30 s (annealing), and 72 °C for 30 s (extension). A melting curve analysis was subsequently performed to confirm the specificity of amplification. Each sample was tested in triplicate, and the experiment was independently repeated twice to ensure reproducibility. Relative gene expression levels were calculated using the 2^-ΔΔCt method, with β-actin used as the internal control. Primer sequences are listed in Supplementary Tables 1–2. All primers were synthesized by TSINGKE, targeting fibrosis-related genes including TGR5, α-SMA, CK19, COL1A1, COL3A1, and TGF-β1.

Western blot (WB)

Tissues or cells were lysed on ice for 30 min using RIPA buffer containing protease inhibitors (Beijing Pulilai, Cat# C1053), followed by centrifugation at 12,000 × g for 15 min at 4 °C to collect the supernatant. Protein concentrations were quantified using the BCA assay (Elabscience, Cat# E-BC-K318-M), and all samples were standardized to 5 mg/mL. A total of 30 µg of protein from each sample was mixed with 5× SDS loading buffer and denatured at 95 °C for 5 min. Proteins were separated by 10% SDS-PAGE and transferred onto methanol-activated PVDF membranes (Millipore, Cat# IPVH00010) using wet transfer at 300 mA for 90 min on ice in pre-chilled Towbin buffer containing 20% methanol. Membranes were blocked at room temperature for 1 h in 5% non-fat milk in TBST. For phospho-proteins (p-PKA, p-CREB), blocking was performed in 5% BSA-TBST instead. Primary antibodies were diluted according to Supplementary Tables 3 and incubated with membranes overnight at 4 °C. The next day, membranes were washed three times (10 min each) with TBST and incubated for 1 h at room temperature with HRP-conjugated secondary antibodies (goat anti-rabbit/goat anti-mouse, ZSGB-BIO, Beijing, 1:5,000). After additional washes with TBST, chemiluminescent signals were developed using ECL substrate (Thermo, Cat# RJ239676) and captured using a Bio-Rad ChemiDoc™ XRS + or Tanon-5200 imaging system, ensuring exposure within the linear dynamic range. Band intensities were quantified using ImageJ software. Target proteins were normalized to GAPDH, and phospho-proteins were normalized to their respective total proteins (p-PKA/PKA, p-CREB/CREB).

Each group included three biological replicates, with at least two technical replicates per sample. Independent experiments were repeated across no fewer than two batches.

Statistical analysis

Wet-lab data were presented as mean ± SD. Comparisons between groups were performed using one-way analysis of variance (ANOVA) in SPSS 20.0. If the data met assumptions of normality and homogeneity of variance, Tukey’s HSD test was used for post hoc comparisons; otherwise, Welch’s ANOVA followed by Dunnett’s T3 adjustment was applied. A two-tailed p-value < 0.05 was considered statistically significant. All figures were generated using GraphPad Prism 9.0. Densitometric analysis of WB bands was performed with ImageJ (NIH, Version 6.0) and normalized to internal controls (e.g., GAPDH). Each sample was subjected to at least two technical replicates, and all experiments were independently repeated across a minimum of two batches.

For bioinformatics analyses, unless otherwise specified, the BH-FDR correction was applied to control for multiple comparisons, with a significance threshold set at q < 0.05. In small-sample scenarios (n < 10), effect size metrics such as Cliff’s δ and Spearman’s ρ, along with their 95% CIs, were reported in parallel to avoid overreliance on p-values. All differences and correlations were consistently expressed in the direction of Model-Control. For predefined gene sets or pathways (e.g., receptor families, genes within the KEGG “Bile secretion” pathway), BH-FDR correction was applied independently. Bioinformatics analyses were conducted in a Python environment (version ≥ 3.10), primarily utilizing the following packages: pandas (v1.5+), numpy (v1.23+), scipy (v1.10+), matplotlib, and PyPDF2.

Results

TGR5 expression was significantly upregulated in liver tissues from patients with hepatolithiasis and hepatobiliary fibrosis

To identify key bile acid receptors potentially involved in the pathogenesis of hepatolithiasis-associated fibrosis, we first performed differential expression analysis on publicly available human gallbladder RNA-seq data (GSE202479). PCA revealed a clear separation between the Control group (non-hepatolithiasis) and the Model group (hepatolithiasis with chronic inflammation) along both PC1 and PC2 axes (Fig. S1A), indicating distinct global transcriptional profiles between the two groups. Further volcano plot analysis demonstrated that multiple members of the bile acid receptor family exhibited significant expression changes (Fig. S1B). Notably, TGR5 was markedly upregulated in the Model group, with S1PR2 also showing an upward trend, whereas nuclear receptors such as FXR and PXR were substantially downregulated (Fig. S1C-D). In the group-wise mean difference waterfall plot, TGR5 ranked among the top upregulated genes, suggesting a potential role in the pathological processes associated with hepatolithiasis (Fig. S1E). To assess the robustness of the differential-expression findings and to delineate the broader relationship between TGR5 and fibrosis-related signaling, we performed LOO sensitivity testing coupled with an integrated receptor–axis–phenotype visualization. Sequential exclusion of individual samples did not alter the magnitude of TGR5 dysregulation or its correlations with the CREB axis and fibrosis scores (Fig. S1F). In parallel, a compact heat-map revealed a coordinated up-regulation of TGR5 with CREB-axis and fibrosis-associated genes (Fig. S1G), reinforcing the notion that TGR5 plays a pivotal role in the pathological processes associated with hepatolithiasis.

To further validate the relationship between TGR5 and the fibrotic molecular program, we analyzed the correlations between TGR5 and several classical fibrosis-related genes (COL1A1, COL3A1, FN1, VIM, TGFB1, MMP2, and MMP9) in the GSE202479 dataset. As shown in Figure S2, in the model group, TGR5 exhibited a consistent positive correlation trend with multiple extracellular matrix and pro-fibrotic genes, including COL1A1, COL3A1, FN1, TGFB1, and MMP2, whereas this coupling pattern was much weaker in the control group. Notably, FN1 showed a perfectly monotonic correlation with TGR5 in the model group (R = 1.00), suggesting that TGR5 is highly coordinated with extracellular matrix remodeling. Although statistical significance could not be established because of the limited sample size, this disease-specific co-variation pattern further supports a key role for TGR5 in hepatolithiasis-associated fibrosis.

To validate the above transcriptomic findings and further characterize the tissue expression pattern, we collected liver tissue samples from patients with hepatolithiasis-associated fibrosis and from non-stone controls for histopathological and molecular analyses. H&E and Masson’s trichrome staining showed marked collagen deposition, disruption of hepatic lobular architecture, and inflammatory cell infiltration in liver tissues from hepatolithiasis patients, indicating obvious cholestatic injury and fibrosis (Fig. 1A–C). Immunofluorescence further confirmed that, compared with non-stone samples, the expression of TGR5 and collagen I (TGR5+, collagen I+, and TGR5+/collagen I+ signals) was significantly increased in the liver tissues of hepatolithiasis patients. The upregulated TGR5 signal was mainly localized to cholangiocytes in the portal area, where the red fluorescence was markedly intensified, whereas CK19 expression was decreased, suggesting a transition of epithelial cells toward a profibrotic phenotype (Fig. 1D). Consistently, Western blot analysis also showed that TGR5 protein levels were significantly elevated in patients with hepatolithiasis compared with the control group (Fig. 1E).

Fig. 1.

Pathological and Molecular Differences Between Patients with Non-Stone and Hepatolithiasis. Note: (A) H&E staining of liver tissue to observe histological alterations; magnification ×100. (B) Masson’s trichrome staining of liver sections to evaluate collagen fiber deposition; magnification ×100. (C) Quantification of collagen fiber area from Masson-stained images using image analysis software. (D) Immunofluorescence staining to detect TGR5, CK19 and collagen I expression in liver tissue. DAPI labeled the nuclei (blue); the secondary antibody for TGR5 was labeled with red fluorescence; the secondary antibody for CK19 was labeled with green fluorescence; and the secondary antibody for collagen I was labeled with green fluorescence; magnification ×100. The right panel shows a statistical analysis of integrated optical density (IOD) from fluorescence images. (E) WB analysis of TGR5 and associated signaling proteins (PKA, p-PKA, CREB, p-CREB, and TGF-β1) in liver tissue. β-Actin was used as a loading control for normalization. Sample size for each group: n = 3. Data are presented as mean ± SD. Asterisks (*) indicate statistically significant differences compared with the non-stone group (p < 0.05)

Collectively, both RNA-seq and histological data indicate sustained overexpression of TGR5 in liver tissues associated with hepatolithiasis, suggesting its potential regulatory role in the disease process and laying the groundwork for subsequent mechanistic investigations.

TGR5^-/- alleviates hepatobiliary fibrosis in a mouse model of hepatolithiasis

To investigate the in vivo role of TGR5 in the development of hepatolithiasis-associated hepatobiliary fibrosis, a mouse model of hepatolithiasis was established, and TGR5^-/- mice were included for comparative analysis (Fig. 2A). Upon model establishment, liver morphology and gallstone formation were assessed (Fig. 2B). In wild-type (WT) model mice, pronounced gallbladder distension and yellow granular gallstones were observed. In contrast, gallstone formation was markedly reduced in the TGR5^-/- model group. Compared with the control group, WT model mice exhibited hepatomegaly, increased tissue firmness, and pale discoloration of the liver, suggestive of cholestasis and parenchymal alteration.

Fig. 2.

Impact of TGR5^-/- on Hepatobiliary Structure and Tissue Fibrosis in a Hepatolithiasis Mouse Model. Note: (A) Schematic diagram of the hepatolithiasis mouse model construction. (B) Following model induction, the in situ liver and ex vivo gallbladder and liver morphology were examined. Representative gross images of the hepatobiliary region were captured. Blue arrows indicate the gallbladder; white arrows, the bile ducts; green arrows, the liver margins. (C) H&E staining of liver tissue (×100) was used to assess histological structure and inflammatory cell infiltration. (D) Masson’s trichrome staining (×100) was used to evaluate collagen fiber deposition; the bar graph on the right shows the quantitative analysis of collagen area ratio. (E) Immunofluorescence staining was used to detect the expression of TGR5, CK19, and collagen I in liver tissues. DAPI labeled the nuclei (blue); the secondary antibody for TGR5 was labeled with red fluorescence; the secondary antibody for CK19 was labeled with green fluorescence; and the secondary antibody for collagen I was labeled with green fluorescence. Magnification, ×100. The panels on the right show the statistical analysis of the mean integrated optical density (IOD) of the fluorescence images. (F) Schematic diagram illustrating the proposed mechanism by which TGR5 regulates gallstone formation and fibrosis. Sample size for each group: n = 6. Data are presented as mean ± SD. Asterisks (*) denote statistically significant differences compared with the WT control group (p < 0.05)

H&E staining revealed prominent inflammatory cell infiltration and periductal erythrocyte dispersion in the WT model group (Fig. 2C), whereas the inflammatory response was notably attenuated in TGR5^-/- mice. Masson’s trichrome staining further showed substantial collagen deposition around the bile ducts in the WT model group, while the extent of fibrosis was significantly reduced in TGR5^-/- mice (Fig. 2D). Furthermore, these findings were corroborated by immunofluorescence analysis. The results showed that the expression of TGR5 and Collagen I was significantly upregulated in the liver tissues of the model group, with TGR5 upregulation predominantly localized to cholangiocytes in the portal area. In contrast, the expression of TGR5 and Collagen I was significantly attenuated in the TGR5-/- model group. Conversely, while CK19 expression was decreased in the model group, it was significantly enhanced in the TGR5-/- model group (Fig. 2E). The bar graph on the right indicates that TGR5^-/- significantly lowered the proportion of fibrotic area (p < 0.05).

Together, these findings demonstrate that TGR5 promotes fibrosis progression in hepatolithiasis, and TGR5^-/- effectively mitigates structural damage and collagen accumulation in the hepatobiliary system (Fig. 2F).

TGR5^-/- ameliorated lipid metabolic dysregulation and cholesterol transport impairment in a mouse model of hepatolithiasis

To evaluate the role of TGR5 in lipid metabolism under gallstone-associated conditions, ELISA was used to measure three major lipid parameters—TG, LDL-C, HDL-C and TBA—in mouse bile (Fig. 3A-D). The results revealed significant increases in TG, TBA and LDL-C levels (p < 0.05), along with a marked decrease in HDL-C levels in the model group. In contrast, the TGR5^-/- model group exhibited a further increase in TG, a pronounced reduction in LDL-C, TBA, and a restoration of HDL-C levels to near-control values (p < 0.05).

Fig. 3.

Effects of TGR5^-/- on Biliary Lipid Levels and Expression of Related Metabolic Genes in a Gallstone Mouse Model. Note: (A-D) ELISA was used to quantify the concentrations of TG, LDL-C, HDL-C and TBA in bile, expressed in mmol/L. (E-J) qPCR was performed to assess mRNA expression levels of hepatic lipid metabolism-related genes, including PPAR-α, FXR, LXR, and cholesterol transport-associated factors ABCG5, ABCG8, and ABCG11. Gene expression was normalized to β-Actin and presented as relative expression levels. (K) Schematic diagram illustrating the proposed mechanism by which TGR5 regulates lipid metabolism and cholesterol transport in the hepatobiliary system. Sample size for each group: n = 6. Data are presented as mean ± SD. *p < 0.05 vs. control group; #p < 0.05 vs. model group; $p < 0.05 vs. TGR5^-/- control group

To further investigate the regulatory role of TGR5 in bile acid and cholesterol transport, qPCR was performed to examine the hepatic expression of key genes, including PPAR-α, LXR, FXR, as well as ABC transporters ABCG5, ABCG8, and ABCG11 (Fig. 3E-J). In the model group, PPAR-α, LXR, and FXR were significantly downregulated, whereas ABCG5, ABCG8, and ABCG11 were markedly upregulated. These trends were partially reversed in the TGR5^-/- model group, suggesting that TGR5^-/- may alleviate metabolic imbalance by restoring the expression of transport-related genes. Furthermore, we validated the protein levels of these key molecules using Western blot analysis. The results demonstrated that the expression of PPAR-α and FXR was significantly downregulated, while ABCG8 expression was markedly upregulated in the model group. In the TGR5-/- model group, however, these trends were partially reversed, suggesting that TGR5 deletion may alleviate lipid metabolic dysregulation by restoring the expression of transport-related factors (Fig. S3).

Collectively, these findings indicate that TGR5 plays a regulatory role in lipid metabolism and cholesterol homeostasis in the context of hepatolithiasis. TGR5^-/- appears to mitigate metabolic disturbances and normalize the expression of associated pathways (Fig. 3K).

TGR5^-/- suppresses CREB pathway activation and attenuates the biliary fibrosis phenotype

To elucidate the role of TGR5 in the development of biliary fibrosis, DEGs between the model and control groups were subjected to GO and KEGG enrichment analyses. At the KEGG level, the DEGs were significantly enriched in pathways related to digestion and secretion, with consistent enrichment of the “Bile secretion” pathway. At the GO-BP level, enrichment was observed in immune activation, extracellular matrix remodeling, and membrane compartmentalization processes (Fig. S4A-B), suggesting that the TGR5 signaling pathway participates in multiple aberrant BP under hepatolithiasis conditions.

Module scoring analysis revealed statistically significant differences between the model and control groups in both the CREB signaling axis and bile secretion modules. Correlation analysis further demonstrated a positive association between TGR5 expression and CREB module scores (Spearman ρ = 0.64, p = 0.014). In contrast, the CREB score was negatively correlated with the bile secretion module and positively correlated with fibrosis scores (Fig. S4C-F), implying a potential regulatory cascade linking TGR5, CREB activation, and fibrosis progression. Gene-level analysis confirmed that most transport-related genes within the “Bile secretion” pathway—such as ABCB11 and SLC family members—were consistently downregulated in the model group compared to controls, thereby validating the robustness of the pathway-level enrichment results (Fig. S4G).

We further assessed the expression of key proteins downstream of the TGR5 signaling axis by WB analysis (Fig. 4A). Compared with the normal control group, the model group exhibited significantly elevated levels of TGR5, p-PKA/PKA, p-CREB/CREB, TGF-β1, α-SMA, and collagen I (p < 0.05). In the TGR5^-/- control group, the expression levels of these proteins were markedly lower than those in WT controls, indicating that TGR5^-/- alone reduced their baseline expression. No significant differences in these protein levels were observed between the TGR5^-/- control and TGR5^-/- model groups. However, when compared with the model group, the TGR5^-/- model group showed significantly decreased expression of all target proteins, suggesting that TGR5^-/- reversed the activation status induced by the disease model.

Fig. 4.

TGR5^−/− Reduces the Expression of TGR5-Associated Proteins and Fibrosis Markers. Note: (A) WB analysis of protein expression levels of TGR5, PKA, p-PKA, CREB, p-CREB, TGF-β1, α-SMA, and collagen I across different groups. Quantification is presented as relative grayscale intensity (n = 6). (B) Immunohistochemical staining of intrahepatic bile duct tissues showing expression of α-SMA and collagen I (magnification ×100); IOD quantification is shown below (n = 6). (C) qPCR analysis of TGF-β1 and α-SMA mRNA relative expression levels, reflecting transcriptional changes of key fibrotic markers (n = 6). (D) Schematic diagram illustrating the mechanism by which TGR5^−/− attenuates cholangiocyte activation and fibrosis via inhibition of the PKA-CREB-TGF-β1 signaling axis. Data are presented as mean ± SD. Statistical significance: *p < 0.05 vs. WT control group; #p < 0.05 vs. model group; $p < 0.05 vs. TGR5^−/− control group

Immunohistochemistry further corroborated these findings: the model group exhibited robust positive staining for α-SMA and collagen I, whereas the staining intensities were substantially reduced in the TGR5^-/- model group (Fig. 4B). In addition, qPCR analysis revealed that mRNA levels of TGF-β1 and α-SMA were significantly upregulated in the model group (p < 0.05) and significantly downregulated in the TGR5^-/- control group. The TGR5^-/- model group also demonstrated markedly lower expression of these genes compared with the model group (p < 0.05), in alignment with the observed protein expression trends (Fig. 4C). Corroborating these results with human clinical samples (Fig. 1E), we found that TGF-β1 protein levels were significantly higher in hepatolithiasis tissue than in non-hepatolithiasis controls, suggesting that TGF-β1 upregulation is not limited to the animal model but is also evident in human pathology. This cross-species consistency supports the clinical relevance of the findings.

Collectively, these results indicate that TGR5^-/- mitigates biliary fibrosis by suppressing PKA/CREB pathway activation and downregulating TGF-β1 signaling (Fig. 4D).

TGR5 regulates cholangiocyte proliferation via bidirectional modulation

To investigate the regulatory role of TGR5 in cholangiocyte proliferation, an in vitro hepatolithiasis model was established by treating HIBECs with TDCA. As shown by the CCK-8 assay (Fig. 5A), increasing TDCA concentrations initially promoted cell proliferation, which then declined at higher concentrations. qPCR analysis revealed that TGR5 expression was significantly upregulated in a dose-dependent manner with TDCA treatment (Fig. 5B). Based on these results, a concentration of 500 µM TDCA was selected for subsequent experiments.

Fig. 5.

Dose-Dependent Effects of TGR5 Agonists and Antagonists on Cholangiocyte Proliferation. Note: (A-B) HIBECs were treated with TDCA to induce a cholangiocyte model. Cell proliferation was assessed by CCK-8 assay (A), and TGR5 mRNA expression was measured by qPCR (B), confirming the modeling concentration as 500 µM. (C-D) CCK-8 assays were performed under different concentrations of HDCA (C) and SBI-115 (D) to determine optimal experimental doses (100 µM and 10 µM, respectively). (E) CCK-8 analysis of cell proliferation in modeled cells treated with either a TGR5 agonist or antagonist showed that TGR5 activation partially restored proliferative capacity, whereas inhibition further suppressed it. Data are presented as mean ± SD (n = 3). Statistical significance: *p < 0.05 vs. control group; #p < 0.05 vs. model group. (F) Schematic diagram illustrating the bidirectional regulatory role of TGR5 in modulating cholangiocyte proliferation. Statistical analysis was conducted using one-way ANOVA followed by Tukey’s post hoc test

Next, optimal treatment concentrations for the TGR5 agonist HDCA and the antagonist SBI-115 were determined (Fig. 5C-D), with 100 µM HDCA and 10 µM SBI-115 chosen for further intervention studies. As shown in the CCK-8 assay (Fig. 5E), TDCA treatment significantly reduced cell proliferation. HDCA partially reversed this inhibitory effect, while SBI-115 further suppressed proliferation. These findings indicate that TGR5 exerts a bidirectional regulatory effect on cholangiocyte proliferation. In addition, we constructed TGR5 siRNA to further verify the regulatory role of TGR5 in cholangiocyte proliferation. The results showed that si-TGR5 markedly reduced TGR5 expression (Fig. S5A). CCK-8 assay showed that cell viability/proliferative activity was significantly decreased in the TDCA-induced model group compared with the control group, and was further reduced in the Model + si-TGR5 group compared with the Model + siControl group (Fig. S5B), indicating that TGR5 knockdown aggravated the proliferation inhibition of cholangiocytes under model conditions.

We further validated this finding at the tissue level by Ki67 immunofluorescence staining. Compared with the control group, the number of Ki67-positive cells among CK19-positive cholangiocytes was significantly decreased in the model group. Moreover, compared with the model group, the TGR5^-/- model group showed a further reduction in Ki67-positive cholangiocytes (Fig. S5C), suggesting that TGR5 deficiency further suppresses cholangiocyte proliferation.

To illustrate the underlying regulatory mechanism of TGR5, a schematic diagram was constructed (Fig. 5F). TDCA stimulation increases TGR5 expression, which subsequently modulates epithelial cell proliferation through downstream signaling pathways. Activation of TGR5 by HDCA promotes proliferation, whereas inhibition by SBI-115 suppresses it, demonstrating the bidirectional nature of TGR5-mediated regulation.

TGR5 activation enhances the PKA/CREB pathway and upregulates fibrosis markers, while inhibition has the opposite effect

To further elucidate the regulatory mechanism of TGR5 in hepatolithiasis-associated fibrosis, this study integrated findings from multiple levels of evidence. First, analysis of clinical samples (Fig. 1E) revealed that the expression levels of TGR5, p-PKA, p-CREB, and TGF-β1 were significantly elevated in liver tissues from patients with hepatolithiasis, indicating that this pathway is activated in the context of biliary injury. Animal experiments further confirmed that TGR5^-/- markedly suppressed CREB pathway activation and reduced the expression of fibrosis-associated markers such as TGF-β1, α-SMA, and collagen I (Fig. 4). In parallel, in vitro experiments demonstrated that TGR5 activation enhanced epithelial cell proliferation, whereas its inhibition exerted a suppressive effect (Fig. 5).

However, whether TGR5 directly regulates the fibrotic process in an in vitro biliary fibrosis model, and whether its effect is causally linked to the PKA/CREB signaling pathway, remains to be determined. To address this, an in vitro hepatolithiasis model was established by stimulating HIBECs with TDCA, followed by treatment with either the TGR5 agonist HDCA or the antagonist SBI-115, to evaluate their respective effects on the PKA/CREB pathway and fibrosis-related markers.

Immunofluorescence results (Fig. 6A) show that TGR5, α-SMA, and collagen I are significantly upregulated in the model group, while CK19 is downregulated, indicating a shift from epithelial to fibroblastic phenotype. Following TGR5 activation, the expression of TGR5 and collagen I was further elevated, α-SMA exhibited a mild increase, and CK19 was further suppressed. Conversely, TGR5 inhibition led to the opposite pattern of expression changes. These results suggest a positive correlation between TGR5 activity and the degree of fibrosis. WB analysis (Fig. 6B) further confirmed that TGR5, TGF-β1, α-SMA, collagen I, p-PKA/PKA and p-CREB/CREB were markedly upregulated in the model group, while CK19 expression was significantly reduced. Treatment with the TGR5 agonist enhanced the expression of TGR5, TGF-β1, collagen I, and p-PKA, whereas the TGR5 inhibitor broadly suppressed fibrosis markers and attenuated pathway activation.

Fig. 6.

Bidirectional Regulation of PKA/CREB Activity and Fibrosis by Pharmacological Modulation of TGR5 in an In Vitro Hepatolithiasis Model. Note: (A) Immunofluorescence staining was performed to detect TGR5, CK19, α-SMA, and collagen I (magnification ×100; channel labels are shown in the figure legend). Quantification of signal intensity was presented as IOD/area (n = 3). (B) WB analysis was used to evaluate the protein expression levels of TGR5, TGF-β1, CK19, α-SMA, collagen I, and PKA/CREB pathway proteins. Bar graphs display relative protein expression normalized to β-actin; phosphorylated proteins were expressed as p/total ratios (n = 3). (C) Immunofluorescence staining of TGR5, CK19, α-SMA, and Collagen I in the co-culture system (magnification ×100; channel labels are shown in the figure legend). The right panel shows the quantification of signal intensity (IOD/area) (n = 3). Data are presented as mean ± SD. Statistical analysis: One-way ANOVA followed by Tukey’s post hoc test. Significance indicators: *p < 0.05 vs. control group; #p < 0.05 vs. model group

To better simulate the biliary fibrosis microenvironment, we established a co-culture system using human intrahepatic biliary epithelial cells (HiBECs) and hepatic stellate cells (LX-2). Immunofluorescence staining was performed to assess the expression of TGR5, α-SMA, Collagen I, and CK19 in LX-2 cells. The results showed that in the model co-culture group, the expression of TGR5, α-SMA, and Collagen I was significantly enhanced, while CK19 expression was decreased, indicating a transition toward a fibrotic phenotype. Following TGR5 activation, the levels of TGR5 and Collagen I were further elevated, α-SMA was slightly upregulated, and CK19 was further suppressed. Conversely, TGR5 inhibition reversed these trends, suggesting a positive correlation between TGR5 activity and the degree of fibrosis (Fig. 6C).

Taken together, evidence from transcriptomic enrichment analysis, clinical samples, animal models, and in vitro experiments demonstrates that TGR5 regulates the fibrotic transformation of cholangiocytes via the PKA/CREB signaling pathway. Pharmacological modulation of TGR5 directly alters the extent of the fibrotic phenotype, identifying this axis as a potential therapeutic target for hepatolithiasis-associated fibrosis.

TGR5 promotes fibrotic transformation of cholangiocytes via the cAMP/PKA/CREB pathway

To further elucidate the molecular mechanism by which TGR5 regulates cholangiocyte fibrosis, particularly its downstream cAMP-mediated effects, we modulated intracellular cAMP levels in an in vitro model and examined subsequent alterations in cellular behavior and signaling activity. Prior results indicated that TGR5 activation enhances CREB phosphorylation and upregulates TGF-β1 expression (Fig. 6B), leading us to hypothesize that the cAMP/PKA/CREB axis serves as a key mediator in this process.

First, HIBECs were treated with various concentrations of the PKA inhibitor H-89. The CCK-8 assay results indicated that a concentration of 20 µM induced significant cytotoxicity (Fig. 7A); therefore, lower concentrations were selected for subsequent experiments. ELISA results showed that cAMP levels in the model group were higher than those in the control group (Fig. 7B). Treatment with the cAMP agonist forskolin further elevated cAMP levels, whereas both the PKA inhibitor and the TGR5 inhibitor significantly reduced them. In the combined treatment group, cAMP levels remained higher than with TGR5 inhibition alone but were lower than with the agonist alone, suggesting that TGR5 positively regulates cAMP production.

Fig. 7.

TGR5 Regulates Cholangiocyte Fibrosis via the cAMP/PKA/CREB Signaling Pathway. Note: (A) CCK-8 assessed cell viability after 24 h treatment with various concentrations of H-89 (n = 3). (B) ELISA was used to measure intracellular cAMP levels across different experimental groups. (C) WB analysis of TGR5, p-CREB/CREB, and TGF-β1 protein expression. (D) qPCR quantification of EMT- and fibrosis-related gene expression, including E-cadherin, CK19, α-SMA, Vimentin, N-cadherin, and collagen I. (E) Schematic diagram depicting the cAMP/PKA/CREB axis mediating TGR5-induced fibrotic transformation. Data are presented as mean ± SD (n = 3). Statistical significance: *p < 0.05 vs. control group; #p < 0.05 vs. model group; $p < 0.05 vs. agonist group

Western blot analysis further corroborated these trends (Fig. 7C). In the model group, the expression of p-CREB/CREB was upregulated. The PKA agonist forskolin further enhanced this expression, while treatment with either the PKA inhibitor or the TGR5 inhibitor resulted in its downregulation; notably, TGR5 expression remained largely unchanged under these conditions. The results from the combined treatment group also supported the role of TGR5 as an upstream regulator of cAMP signaling.

Furthermore, qPCR analysis of fibrosis- and EMT-related gene expression (Fig. 7D) revealed that in the model group, the expression of E-cadherin and CK19 was decreased, while the expression of α-SMA, Vimentin, N-cadherin, and Collagen I was increased. The PKA agonist further exacerbated these trends, whereas inhibitor treatment reversed this phenomenon. In the combined treatment group, most markers showed a tendency toward attenuation of the agonist’s effects.

In summary, TGR5 promotes the transformation of cholangiocytes toward a fibrotic phenotype by activating the cAMP/PKA/CREB signaling axis, suggesting that this pathway plays a critical pathogenic role in the development of hepatolithiasis-associated fibrosis (Fig. 7E).

Discussion

TGR5, a bile acid-sensing receptor, plays a pivotal role in the physiological regulation of various hepatobiliary diseases [25]. Previous studies have primarily focused on its anti-inflammatory and hepatoprotective effects in conditions such as cholestasis, hepatic fibrosis, and metabolic syndrome [14, 26]. For instance, it has been reported that TGR5 activation improves hepatocellular lipid metabolism and modulates bile acid homeostasis, thereby mitigating bile acid toxicity and liver injury [15, 27]. However, no prior study has systematically investigated the role of TGR5 in hepatolithiasis-associated biliary fibrosis. The present study is the first to demonstrate, across clinical specimens, animal models, and in vitro cellular systems, that TGR5 exerts a pro-fibrotic effect in this pathological context, primarily through activation of the cAMP/PKA/CREB signaling axis. These findings provide novel evidence supporting a bidirectional role of bile acid receptors in the pathogenesis of chronic biliary diseases.

In both clinical tissue samples and RNA-seq database analyses, we observed elevated TGR5 expression in hepatolithiasis-associated liver tissues, along with enrichment in pathways related to bile secretion, inflammation, and fibrosis. This pattern aligns with findings in certain hepatobiliary diseases that also report high TGR5 expression [28, 29], yet the functional implications differ. Whereas previous studies linked TGR5 upregulation with tissue protection and anti-inflammatory responses, our results reveal that TGR5 activation in this setting promotes fibrotic progression. This divergence may reflect context-specific factors such as cell type, tissue microenvironment, and downstream signaling bias, suggesting that TGR5 may exhibit functional duality under different pathological conditions. Our findings underscore the importance of distinguishing the biological role of TGR5 in cholangiocytes from its functions in hepatocytes or immune cells.

Through the TGR5^-/- mouse model, this study further confirmed that TGR5^-/- significantly alleviated the biliary fibrosis phenotype, as evidenced by improved lipid metabolism and downregulation of key fibrogenic markers such as TGF-β1 and α-SMA. This finding stands in sharp contrast to previous studies in hepatic fibrosis, where TGR5 was implicated in antifibrotic responses [30]. The discrepancy may be attributed to the cellular context: the present study focused on cholangiocytes rather than hepatocytes. Moreover, under the pathological condition of hepatolithiasis, bile acid metabolism is severely disrupted, potentially leading to aberrant TGR5 activation and a reversal of its expected effect. These results suggest that the role of TGR5 in biliary fibrosis exhibits marked tissue specificity.

As a key initiating event in biliary fibrosis, EMT has recently gained attention for its upstream regulatory signaling networks [31, 32]. While the classical TGF-β1/Smad axis is widely recognized as the major driving force, our findings reveal that TGR5 promotes CREB phosphorylation via the cAMP/PKA/CREB pathway, which in turn induces TGF-β1 expression and facilitates the EMT process. This hierarchical cascade—from a noncanonical to a canonical signaling axis—offers a novel perspective on the mechanisms underlying cholangiocyte phenotypic transformation. Moreover, it positions TGR5 as a potential upstream modulator of TGF-β1 activation, thereby expanding the current understanding of EMT regulatory pathways.

In vitro experiments, this study simulated bile acid accumulation and found that treatment with a TGR5 agonist enhanced fibrotic features in cholangiocytes, whereas inhibition of TGR5 reversed this trend. Further validation using cAMP agonists and inhibitors confirmed the involvement of this signaling axis. This chain-based validation model—encompassing receptor, signaling cascade, and effector molecules—offers a deeper mechanistic insight compared with previous studies that were limited to correlative expression analyses, thereby strengthening the credibility of causality. Notably, although TGR5 represents a potential pharmacological target, its functional heterogeneity across various disease contexts underscores the need for cautious evaluation of agonist or antagonist applications, particularly with respect to disease specificity and tissue selectivity.

This study provides systematic evidence of TGR5’s functional role in biliary fibrosis; however, several limitations remain. First, advanced translational models such as human-derived organoids have not yet been employed. Second, the role of TGR5 in other biliary cell populations, including immune cells and biliary smooth muscle cells, has not been investigated, which may lead to an underestimation of its multifaceted regulatory roles within the complex tissue microenvironment. Moreover, whether bypass branches beyond the cAMP/PKA/CREB axis exist remains an open question for future exploration.

In conclusion, this study is the first to elucidate the pathogenic role of TGR5 in hepatolithiasis-associated biliary fibrosis and demonstrates that it promotes fibrotic progression via the cAMP/PKA/CREB axis by upregulating TGF-β1 and inducing EMT. These findings not only enrich the mechanistic landscape of biliary fibrosis but also provide a theoretical basis for targeting TGR5 as a potential therapeutic strategy. Future investigations incorporating multicellular organoid models, large-scale clinical validation, and targeted drug design may accelerate the translational application of TGR5 as a precision therapeutic target for patients with hepatobiliary diseases.

Conclusion

This study systematically elucidated the key pathogenic mechanisms of TGR5 in hepatobiliary fibrosis associated with hepatolithiasis. By integrating clinical tissue samples, bioinformatic analyses, animal models, and in vitro cellular experiments, we demonstrated that TGR5 expression was markedly upregulated in response to bile acid stimulation. This upregulation activated the cAMP/PKA/CREB signaling pathway, promoting the transdifferentiation of cholangiocytes into a mesenchymal phenotype and enhancing the expression of fibrosis-associated factors such as TGF-β1, α-SMA, and collagen I, thereby contributing to the progression of biliary fibrosis. The use of TGR5^-/- mice further confirmed its pro-fibrotic role in stone formation, fibrosis exacerbation, and lipid metabolism disorders. Moreover, pharmacological modulation of its downstream signaling axis effectively reversed the fibrotic phenotype in vitro, highlighting the TGR5/cAMP/PKA/CREB pathway as a central driver in this pathological process.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ALT

Alanine Transaminase

ANOVA

Analysis of Variance

AST

Aspartate Aminotransferase

BH

Benjamini-Hochberg

BP

Biological Process

CI

Confidence Interval

CV

Coefficient of Variation

DEGs

Differentially Expressed Genes

EMT

Epithelial-Mesenchymal Transition

FDR

False Discovery Rate

GO

Gene Ontology

H&E

Hematoxylin and Eosin

HDCA

Hyodeoxycholic Acid

HDL-C

High-Density Lipoprotein Cholesterol

HIBECs

Human Intrahepatic Biliary Epithelial Cells

KEGG

Kyoto Encyclopedia of Genes and Genomes

LDL-C

Low-Density Lipoprotein Cholesterol

LOO

Leave-One-Out

log₂FC

Log₂Fold Change

ORA

Over-Representation Analysis

OD

Optical Density

padj

p.Adjust

PCA

Principal Component Analysis

qPCR

Quantitative Polymerase Chain Reaction

RNA-seq

RNA Sequencing

SD

Standard Deviation

TDCA

Taurodeoxycholic Acid

TG

Triglycerides

TGR5

G-Protein-Coupled Bile Acid Receptor 1

TGR5^−/−

TGR5 Knockout

WB

Western Blot

WT

Wild-Type

α-SMA

α-Smooth Muscle Actin

ρ

Correlation Coefficient

Author contributions

Dan Tang and Lijin Zhao conceived and designed the study. Jiali Yang, Xinhao Shi, and Xuanyu Gu performed the experiments and collected the data. Jiali Yang and Xinhao Shi conducted data analysis and interpretation. Dan Tang drafted the manuscript, and Lijin Zhao critically revised it for important intellectual content. Lijin Zhao supervised the project and provided overall guidance. All authors reviewed and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (No. 82460133, No. 81960125), the Science and Technology Foundation Project of the Health Commission of Guizhou Province (No. gzwkj2021-172), and the Guizhou Provincial Basic Research Program (Natural Science) (No. [2020]1Y302).

Data availability

All data generated or analyzed during this study are included in this article and/or its supplementary material files. Further enquiries can be directed to the corresponding author.

Declarations

Ethical approval

All human experiments were conducted in accordance with the ethical principles of the Declaration of Helsinki and were approved by the Institutional Ethics Committee of Zunyi Medical University (Approval No.: [2019] 1–058). All animal experiments complied with the ARRIVE guidelines and the National Institutes of Health (NIH) guide for the care and use of laboratory animals, and were approved by the Animal Ethics Committee of Zunyi Medical University (Approval No.: KLLY-2019-165).

Competing interests

The authors declare no competing interests.