Arbutin (more than 98% purity) wad supplied by Aladdin (Shanghai, China). ANIT (more than 98% purity) was purchased from Sigma. UDCA was purchased from the Aladdin (Shanghai, China). Ntcp, Bsep, FXR, and Cyp7a1 as well as GAPDH antibody were from Beyotime Biotech (Beyotime, Shanghai, PR China) and Bioss Biotechnology (Bioss, Beijing, China). Goat anti-mouse IgG-HRP or goat anti-rabbit IgG-HRP were purchased from Santa Cruz Biotechnology (CA, United States). Immunofluorescence antibodies of Alexa Fluor were purchased from Invitrogen Life Science (CA, United States). siFXR ((5′-CAA​GTG​ACC​TCG​ACA​ACA​A-3′) was from Qingke Bioscience (Chengdu, China). Kits for determination of alkaline phosphatase (ALP; cat. no. A059-1), alanine transaminase (C0009-2), aspartate transaminase (C010-2), γ-glutamyltranspeptidase (γ-GT; cat. no. C017-1), total bile acids (TBA; cat. no. E003-1), and direct bilirubin (DBIL; cat. no. C019-2) were obtained from Nanjing Jiancheng. Other chemical reagents were of analytical grade.

Six- to seven-week-old adult C57BL/6 mice (male) were obtained from Dossy Bioscience (Chengdu, China). The animals were supported on a cycle with 12-hour light and 12-hour dark with temperature (25 ± 1°C) and (46 ± 5%) humidity for 1 week prior to experiments. The C57BL/6 mice were distributed into six groups (eight mice in each group) randomly (Table 1). Chemical drug doses and abidance of exposure were based on previous research studies. Arbutin (10, 20, and 40 mg/kg) or 30 mg/kg UDCA or buffer alone was intraperitoneally (i.p.) administered to mice every day for 1 week. On the fifth experimental day, 3 hours after arbutin, UDCA, or vehicle treatment, mice were intragastrically (i.g.) administered α-naphthylisothiocyanate (ANIT) with 75 mg/kg. On the seventh day, all groups were euthanized and fasted to collect blood, bile, and livers for further analysis 4 hours after treatment with arbutin, UDCA, or vehicle buffer. The mice were intragastrically euthanized with 150 mg/kg pentobarbital following the guidelines of the Animal Ethics Committee. The liver, blood, gallbladder, and duodenum were immediately excised after euthanasia and were stored at −80°C. All animal studies were approved by the Animal Ethics Committee of Chengdu University of Traditional Chinese Medicine and performed following the regulations (Figure 2).

After euthanasia, the liver and other tissues were fixed with 4% polyoxymethylene, and the slices were excised from the anterior liver part of the left lobe, embedded in paraffin, and sectioned into 5-µm sections. Each slide was stained with trichrome for hematoxylin and eosin (H&E) and further intuitive liver histological assessment including hepatocellular necrosis, portal inflammation, or hyperplasia. Analysis of the surface areas of the hepatocytes was carried out under a high-resolution microscope (Olympus) with a photographic facility. The samples of blood stored at 4°C for 4 hours were centrifuged for 30 minutes at 12,000 rpm to collect the supernatant. And ALP, γ-GT, ALT, TBIL, AST, and TBA of serum were measured using the Automatic Biochemistry Analyzer according to the manufacturers’ instructions.

Total RNAs were extracted using TRIzol reagent (Life technologies; no. 15596026) according to the manufacturers’ instructions, and the quantity of total RNA was measured. Reverse transcription was performed to synthesis cDNA with BeyoRT™ commercial Kit from Beyotime (Shanghai, China). As a template, cDNA was used for real-time PCR in triplicate with the SYBR Green Master Mix and specific primers on a Real-Time PCR System (Bio-Rad) (Table 2). GAPDH was normalized as an internal control to evaluate the efficiency.

Human hepatocyte (L-02) cells were cultured in DMEM containing 10% FBS, 100 U/ml antibiotics with streptomycin, penicillin, and streptomycin. The viability of L-02 cells was analyzed by MTT experiments. The cells were cultured with indicated concentrations of arbutin (25, 50, 100, and 200 μM) for 24 hours; vehicle groups were cultured with the same final concentration of DMSO (the final concentration must be <0.001). Then, MTT was added to each well to incubate for 3 hours at 37°C. After discarding the supernatant, DMSO was used to dissolve the mixture, and absorbance was measured at a wavelength of 570 nm by thhe Microplate Reader system (BIO-RAD). For transient transfection experiment, L-02 cells were seeded with a proper density for 24-well plates at 7 × 10⁵ cells each well. A control siRNA was used as a negative control. Cells were transfected using Lipofectamine ™ 3000 (Thermo Fisher Scientific). These transfected cells were treated with drugs mentioned before for another 24 hours after transfection. Cell proliferation was detected by the EdU assay kit (RiboBio, C10310) according to the instruction completely. And images were obtained with a fluorescence microscope. For lactate dehydrogenase release (LDH) assay, cell cytotoxicity was assessed by a commercial kit from Beyotime Biotech (Beyotime, Shanghai, PR China). The experiment was carried out in accordance with the supplier’s instructions.

Total proteins of liver were lysate by RIPA buffer, and the concentration of protein samples were evaluated by a BCA Protein Assay Kit (Thermo Fisher) according to the instructions. Equivalent samples were subjected to SDS-PAGE and then transferred to the polyvinylidene difluoride (PVDF) membranes (Millipore Corporation, MA, United States). After sealing with 5% non-fat milk, the PVDF membranes were washed and co-incubated with specific primary antibodies at 4°C overnight. Subsequently, the second antibody was incubated with the membrane for 1 hour at room temperature. Specific protein–antibody complexes were measured using enhanced chemiluminescence (ECL) reagents and collected with the gel imaging system (ChmiScope, Clinx, China).

Cells were seeded on the glass cover slides (WHB scientific). After indicated treatment, cells were fixed in 4% formaldehyde, permeabilized by 0.3% Triton X-100, and blocked with 5% BSA. The slides were then incubated with indicated antibodies above at 4°C overnight and Alexa Fluor 488/594 antibodies (1:200 dilution) for 1 hour at room temperature. Nuclei were stained with 1:2000 diluted DAPI solution (Solarbio, C0060) at room temperature for 15 minutes, and the images were captured by using a DM2500 fluorescence microscope (Leica).

For molecular docking studies, the AutoDock Tools were used to identify the potential FXR agonists. The 3D structure file of NR1H4 (FXR, code: 4QE6) was downloaded from Protein Data Bank (Kudlinzki et al., 2019). The original structure of NR1H4 was then obtained by AutoDock Tools v1.5.6 to preserve charge and prepare pdbqt file for docking. Then, the hydrogenated SDF structures of arbutin were obtained from PubChem database. And, the structure with the lowest docking energy was measured by minimization. Finally, the resultant protein poses were ranked by scoring, and the image of top-ranked protein pose was further displayed by Discovery Studio Visualizer 2019 and PyMOL v1.8 software.

All data were presented as mean ± SD. One-way analysis of variance (ANOVA) was used to analyze differences between multiple groups. And non-parametric Mann–Whitney U test and unpaired Student’s t-test were used for two groups with GraphPad Prism software v8. As a threshold, p < 0.05 was shown as statistically significant.

To investigate the cytotoxic of arbutin on hepatocyte cells, MTT assay of L-02 cells treated with arbutin was performed. The results suggested that 0–200 μM of arbutin for 24 hours showed no obvious proliferative inhibition (Figures 1B,C). Moreover, the IC50 value of arbutin was 618.2 μM (Supplementary Figure S1A); Consistently, EdU incorporation assays and colony formation analysis confirmed that treatment with arbutin has no effects on the proliferation of hepatocyte cells (Figures 1E,F). The results also did not show enhanced cytotoxicity monitored via lactate dehydrogenase (LDH) assay (Figure 1D) Additionally, the serum levels of AST and ALT did not change significantly in mice treated with arbutin 40 mg/kg (data not shown), therefore implicating that arbutin did not exert significant cytotoxicity in vitro and in vivo. Accordingly, arbutin may have no significant cytotoxic effects on hepatocytes.

To evaluate the effects of arbutin on liver injury induced by cholestasis, C57BL/6 mice were treated with ANIT (75 mg/kg) to induce cholestasis. In brief, the vehicle and ANIT groups were treated with vacant solvent. The UDCA treatment with 30 mg/kg and several arbutin treatment groups were, respectively, pretreated with indicated drugs for 5 days. Then, all groups, except the control, were treated with ANIT that resolved in olive oil. Two days later, the serum and liver tissue samples were collected. Our results showed that the levels of ALT and AST in serum and the sensitive blood biochemical index of liver injury were significantly upregulated after ANIT treatment. The levels of TBIL, γ-GT, TBA, and ALP in serum, the key biochemical indicators relevant to liver damage, were also elevated by administration with ANIT (Figure 2). Accordingly, these results indicated that intrahepatic cholestatic liver damage had been successfully induced by ANIT. Notably, treatment with UDCA or arbutin visibly reduced the serum ALT and AST levels compared with ANIT treatment. Meanwhile, the levels of serum indicators in the UDCA- and arbutin-treated groups decreased significantly. Moreover, arbutin also reversed the liver injury induced by ANIT (10, 20, and 40 mg/kg arbutin) in an adaptive response way (p < 0.01 or <0.05) (Figure 2). Together, these data suggest that arbutin can alleviate the hepatotoxicity in ANIT-induced cholestasis in vivo.

Morphological and histopathological data of liver were assessed to analyze the hepatoprotective effect of arbutin in vivo. The tissues acquired from the ANIT-treated mice displayed visible hepatic necrosis, inflammatory factor infiltration, and edema (Figure 3; Table 3). We also found that ANIT led to a gallbladder filling and a blackened color (data not shown). Congruously, treatment with UDCA and arbutin could significantly attenuate these pathological changes. Arbutin showed a hepatoprotective effect in an adaptive response way (Figure 3).

In order to investigate the mechanism of arbutin in alleviating liver injury, the genes related to BA metabolism were measured by Western blot or real-time PCR. In our results, ANIT treatment attenuated the transcriptional levels of basolateral uptake transporters, Oatp1b2 as well as Ntcp, which was restored by UDCA or arbutin. The mRNA levels of BSEP and Mrp2, two canalicular efflux transporters, were decreased by ANIT, but arbutin could rescue Bsep, rather than Mrp2 (Figure 4B). For the BA transporters, the mRNA levels of Mrp3 and Mrp4 were reduced in ANIT-treated mice, which could not be affected by arbutin. Furthermore, arbutin reversed the reduced transcriptional levels of Cyp7a1 or Cyp8b1, bile acid synthesis enzymes, caused by ANIT. The levels of Cyp7a1 and Cyp8b1 also decreased when treated with UDCA. Meanwhile, UDCA and arbutin both could restore the expressions of bile acid detoxification enzymes, Sult2a1 as well as Ugt1a1, which are attenuated by ANIT (Figure 4). Altogether, arbutin can reverse the altered several genes expression related to BAs homeostasis.

As the biosynthesis or transport of bile acid is controlled by the FXR signaling pathway, we analyzed the expression levels of FXR and downstream proteins, Cyp8b1, Bsep, and Ugt1a1 by Western blot. As expected, the suppressed expression of FXR in ANIT-treated mice was restored by UDCA and arbutin.

We also observed that the downstream biosynthetic enzymes (Cyp8b1, Cyp7a1, and Ugt1a1) and transporters (Bsep, Ntcp, and Oatp1b2) of bile acid could be reversed by arbutin treatment in a dose-dependent way (Figures 5A–C). To further validate whether arbutin could be a potential agonist for FXR, we transfected with siRNA to knockdown the FXR in L-02 cells; we then analyzed the levels of the FXR by Western blot and fluorescence absorbance (Figures 5D,E). Together, these results indicated that the FXR might have participated in arbutin regulated genes involved in BAs metabolism, suggesting arbutin may be a potential and efficient agonist for FXR to exert the hepatoprotection against cholestasis caused by ANIT.

To confirm that arbutin has potential to bind to the FXR, molecular docking analysis with arbutin and the residues from the NR1H4 3D structure (code: 4QE6) was performed. Obeticholic acid (OCA) is a well-known potent agonist of the FXR and docked with a score of −11.1 kJ/mol (Supplementary Figure S1B) (Zhang et al., 2020). Similar to the interaction between the FXR and OCA, the energy for binding between arbutin and FXR was −7.4 kJ/mol, implicating that the interaction was stable (Figure 6A). As shown in results, arbutin binds to a hydrophobic pocket consisting of amino acid residues (Met328, Ser332, Ile335, Arg331, His294, Met290, Ala291, Leu287, Phe336, Phe366, Met365, Ile362, Tyr369, Tyr361, and Ile352). Moreover, arbutin formed two essential hydrogen bonds with Met328 and Ser260, and π–π interaction with Tyr369 residues of FXR, which further enhanced the bonding affinity (Figure 6B). Together, the docking model demonstrated that arbutin could bind to the FXR, and arbutin might be a potential FXR agonist (Figure 7).