All animal experiments were approved by the Laboratory Animal Ethics Committee of Jinan University. Wild-type (WT) mice and whole body FXR knock out (FXR^−/-) mice were purchased from Cyagen Biosciences (China). All mice were on a C57BL/6 background. Animals were maintained under specific pathogen-free conditions and had free access to water and food. Eight-week-old male mice were fed a high-fat diet (HFD; 60% kcal from fat) mixed with different doses of OCA (0.04%, normal-dose; 0.16%, high-dose; 0.40%, over-dose; HFD + OCA group) or without OCA (HFD group). Body weight and food consumption were recorded weekly.

Serum total cholesterol, alanine aminotransferase (ALT), and aspartate aminotransferase (AST), were determined using enzymatic methods kits (Nanjing Jiancheng, China). To measure liver total cholesterol content, 0.1 g of each liver was lysed and measured by a cholesterol kit (Nanjing Jiancheng, China) according to the manufacturer’s protocol.

Liver tissue (70 mg) was hydrolyzed in HCl (6N) as previously described (Trebicka et al., 2007). Hepatic hydroxyproline levels were determined using a hydroxyproline assay kit (Nanjing Jiancheng, China) in accordance with the manufacturer’s protocol.

An equal amount of denatured protein was separated by SDS-PAGE gel and transferred to a PVDF membrane (Millipore, United States). The membranes were blocked and incubated with anti-IL-1β (abcam, ab9722), anti-NLRP3 (ABclonal, A12694) or anti-GAPDH (Proteintech, 60,004-1-lg) overnight. After an incubation with horseradish peroxidase-conjugated antibodies, protein bands were visualized using chemiluminescence kit (Millipore, United States) and LAS Chemiluminescent Imaging System (LAS500, United States). The bands were quantified by ImageJ.

Total RNA from liver was extracted for complementary DNA (cDNA) library construction. Single-end libraries were sequenced by BGISEQ-500, and Bowtie2 software (version 2.2.5) was used to align clean reads to mouse genes and genomes. Gene expression levels (fragments per kilo base of exon per million fragments mapped) were quantified by RSEM (version 1.2.8). Read counts were inputted to DESeq2 to calculate differential gene expression and statistical significance. Differentially expressed genes (DEGs) (HFD + OCA group vs HFD group) were identified as 1) a fold change larger than 2, and 2) Q-value (adjusted p-value) less than 0.001. Pathways overrepresented by DEGs were annotated in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Pathways with a Q-value (corrected p-value) less than 0.05 was considered significantly enriched. Gene Set Enrichment Analysis (GSEA) software (version 4.1) was utilized to identify enriched signaling pathways. Annotated gene sets of gene ontology (GO) biological processes were used as enrichment inputs. The mouse version of MSigDB gene sets were obtained from Walter and Eliza Hall Institute website (https://bioinf.wehi.edu.au/MSigDB/v7.1/). Enriched pathways were identified as a p-value < 0.05 and a false discovery rate (FDR) Q-value < 0.25.

Total RNA was extracted from liver using TRIzol (Thermo Fisher, United States) and reverse transcribed using a reverse transcriptase kit (TAKARA, China). The cDNA and SYBR GREEN Premix EXtaq (TAKARA, China) were used for real-time quantitative PCR in a CFX96 Real-Time PCR Detection System. Primer sequences are listed in Supplementary Table S1.

Data are displayed as mean ± standard error. Comparisons between two groups were evaluated by a two-tailed unpaired Student’s t-test. p < 0.05 was considered statistically significant.

To explore whether FXR is involved in OCA-induced hepatoxicity, wild-type (WT) mice and FXR^−/- mice were administered with an over-dose OCA (0.40% incorporated into an HFD) for 5 days. Mice were humanely euthanatized when they were unable to ambulate. During the 5-days period, over-dose OCA caused all of WT mice died, but none of FXR^−/- mice died (Figure 1A). Liver slices of WT mice showed disordered and necrotic hepatocytes, as well as infiltrated inflammatory cells (Figure 1B). Furthermore, compared to the FXR^−/- mice, WT mice displayed a marked elevation in both serum ALT and serum AST on an over-dose OCA (Figure 1C).

Over-dose OCA (0.4% OCA) is lethal, therefore, a non-lethal but toxic dose was used for exploring the mechanisms leading to hepatoxicity. WT mice and FXR^−/- mice were fed with 0.04% (normal dose) OCA for 14 weeks, after which the dose of OCA was increased to 0.16% (high dose) for 8 weeks. During the 22-week period, none of the mice died. However, the body weight of WT mice dropped sharply after the dosage increase (Figure 2A, left panel), while the decrease in body weight of FXR^−/- mice was mild (Figure 2A, right panel). Furthermore, high-dose OCA led to increased serum AST levels (Figure 2B, right panel) in WT mice, which are indicative of liver injury. The high-dose OCA induced hepatic fibrosis in WT mice but not in FXR^−/- mice, as shown by Masson’s staining liver slices (Figure 2C) and increased hepatic hydroxyproline content (Figure 2D). The mRNA expression of tissue inhibitor of matrix metalloproteinase 1 (Timp1) and connective tissue growth factor (Ctgf), markers of hepatic stellate cell activation, were markedly upregulated in the livers of WT mice (Figure 2E, left panel). FXR deficiency abolished the effect of high-dose OCA on hepatic stellate cell activation (Supplementary Figure S1A). Hence, FXR was essential for high-dose OCA-induced hepatic stellate cell activation and liver fibrosis. Furthermore, high-dose OCA decreased the hepatic TG content in WT mice and FXR^−/- mice (Supplementary Figures S2A,B).

To explore the mechanism underlying high-dose OCA-induced hepatic fibrosis, RNA-seq was performed on the livers of WT mice (WT group) and FXR^−/- mice (FXR^−/- group) fed high-dose OCA, and WT mice fed normal-dose OCA (WT-N group). A total of 5,509 DEGs were found in WT group, while 468 DEGs in FXR^−/- group and 918 DEGs in WT-N group (Figure 3A). KEGG enrichment analysis of DEGs highlighted six enriched pathways in WT group (compared to FXR^−/- group) that were related to 1) inflammasome activation and inflammatory responses (cytokine-cytokine receptor interaction, NOD-like receptor signaling pathway, c-type lectin receptor signaling pathway), and 2) fibrogenic responses (osteoclast differentiation, MAPK signaling pathway, and Ras signaling pathway) (Figure 3B). The enrichment of these pathways was almost consistent in the WT-N group and the FXR^−/- group (Figure 3B). Furthermore, high-dose OCA upregulated proinflammatory mediators and fibrosis gene expression in livers of WT mice, but not in FXR^−/- mice (Figure 3C). However, normal-dose OCA did not induce the expression of these genes in WT mice (Figure 3C). The evidence demonstrated that OCA caused an FXR-mediated inflammatory and fibrogenic response in a dose-dependent manner. GSEA analysis of WT group indicated that the most significantly enriched pathway was positive regulation of IL-1β production (Figure 3D, left panel). Notably, the majority (6 out of 10, 60%) of the most significant pathways in WT group were involved in the IL-1β pathway (Figure 3D, right panel). The mRNA and protein expression of IL-1β were upregulated in livers of WT group (Figures 4A,E,G), but not in FXR^−/- group (Figures 4B,F,G). In addition, the expression of Nlrp3 and Tnfa also upregulated in the presence of FXR (Figures 4A–D).

FXR plays a crucial role in cholesterol homeostasis (Chavez-Talavera et al., 2017). We found that the high-dose OCA decreased serum cholesterol and caused cholesterol accumulation in the livers of WT mice (Figure 5A). The mRNA expression of 3-hydroxy-3-methylglutaryl-CoA reductase (Hmgcr), the rate-limiting enzyme for cholesterol synthesis, was upregulated by high-dose OCA in the livers of WT mice (Figures 5B,C). High-dose OCA also upregulated the expression of Srb1, which is responsible for hepatic uptake high-density lipoprotein (HDL) cholesterol from circulation. Moreover, high-dose OCA markedly suppressed the expression of enzymes involved in cholesterol conversion to bile acids, such as Cyp7a1, Cyp8b1, and Cyp27a1 (Figures 5B,C) in the livers of WT mice. Conversely, the expression of these genes in the livers of FXR^−/- mice had no such differences with high-dose OCA treatment (Figures 5B,D). The results revealed that in the presence of FXR, high-dose OCA upregulated genes involved in cholesterol acquisition and downregulated those involved in secretion of cholesterol in the livers. Together, these changes contributed to the cholesterol accumulation in livers. Cholesterol trigger NLRP3 activation, leading to the production of IL-1β, which triggers an inflammatory response, and promotes hepatic stellate cell proliferation (Ioannou, 2016). Therefore, high-dose OCA may have resulted in hepatic fibrosis via cholesterol accumulation and increased the production of IL-1β in an FXR-mediated manner (Figure 5E).