Wistar rats weighing 300-350 g and aged 12 weeks were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. All procedures were approved by the Ethics Committee of Shaoxing People's Hospital and conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (revised 2011). The rats had ad libitum access to food and water and were maintained at 20˚C, with 50% humidity under 12-h light/dark cycles.

Rats were randomly grouped into sham, bile duct ligation (BDL) and BDL + lipopolysaccharide (LPS) groups. The animals were anesthetized with an intraperitoneal injection of 50 mg/kg pentobarbital. In the BDL and BDL + LPS groups, the distal common bile ducts were dissociated and ligated with 6-0 silk, and PE-10 polyethylene catheters long enough to reach the skin were inserted into the proximal bile ducts as described in our previous study (18,19). Intra-bile duct infusion was performed by injecting 0.2 ml saline or LPS (2 mg/ml, purified from Escherichia coli O111:B4; Sigma-Aldrich; Merck KGaA) into the proximal bile ducts through the catheter. The catheter was sealed with a sealing cap and the abdominal cavity was covered using silk sutures following the introduction of 0.1 ml air. The rats in the sham group underwent a sham procedure, which involved dissecting bile duct carefully without injury and closing the abdominal cavity. The rats in the sham, BDL and BDL + LPS groups were further divided into three groups according to the intraperitoneal injections they received after surgery. Briefly, in each group, one-third of the rats were intraperitoneally injected with normal saline (NS), another one-third were intraperitoneally injected with 2 mg/kg 17-DMAG (MedChemExpress) and the remaining one-third were intraperitoneally injected with 5 mg/kg 17-DMAG. The intraperitoneal injections were performed daily. A total of 108 rats were randomly classified into the aforementioned three groups according to the surgical procedure (n=36/group). Each of these groups comprised rats treated with NS, 2 mg/kg 17-DMAG and 5 mg/kg 17-DMAG (n=12/treatment), respectively. At 24 and 72 h after the surgery, a fraction (n=6/group/time point) of the rats in each group were euthanized. Caspase-3 activity detection and pathology experiments, including tissue staining and immunohistochemistry, were performed on samples taken at 72 h. Other experiments were performed on samples taken at 24 h. The rats were sacrificed via cervical dislocation under 2.5% sevoflurane inhalation anesthesia. The death of the animals was confirmed by examining the cessation of vital signs. Blood and liver tissue samples were harvested from the euthanized animals. Prior to analysis, blood samples were centrifuged (1,000 x g for 5 min at 4˚C) and the serum was stored at -80˚C. Liver tissues were snap-frozen in liquid nitrogen or fixed with 10% formalin.

Clinical tissue samples were obtained from patients who underwent hepatectomy at Shaoxing People's Hospital (Shaoxing, China) from January 2014 to December 2017. Normal liver tissues (n=5; age range, 42-63 years; inclusion criteria: Definite pathological diagnosis of hemangioma; exclusion criteria: Liver cirrhosis, fatty liver, hepatitis virus positive, liver tumor and hepatolithiasis) were obtained from patients who underwent hepatectomy to treat liver hemangioma; non-hemangioma tissues were selected for pathological examination. Intrahepatic cholangitic liver tissues (n=5; age range: 45-69 years; inclusion criteria: Definite pathological diagnosis of hepatolithiasis; exclusion criteria: Liver cirrhosis, fatty liver, hepatitis virus positive and liver tumor, such as intrahepatic cholangiocarcinoma) were obtained from patients who underwent hepatectomy to treat intrahepatic stones and cholangitis. There were two male patients and three female patients in the two groups, respectively. All participants provided written informed consent, and Shaoxing People's Hospital Institutional Review Board approved the tissue acquisition protocol.

An Automated Chemical Analyzer (Dimension RxL Max HM; Siemens AG) was used to quantify the levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT) and total bilirubin (TBIL) in rat serum to determine the degree of liver injury.

Total RNA was isolated from liver tissue samples or cells using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc.) and reverse transcribed into cDNA using a PrimeScript™ RT reagent Kit (cat. no. RR047A; Takara Bio, Inc.), according to the manufacturer's protocols. Briefly, RT was conducted at 37˚C for 15 min and 85˚C for 5 sec, and the cDNA was stored at 4˚C until further use. qPCR was performed using SYBR^® Premix Ex Taq™ II (cat. no. RR820A; Takara Bio, Inc.) and an ABI 7500 Real-time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. The primer sequences used are shown in Table I. β-actin was used as an endogenous control. The qPCR conditions comprised 95˚C for 30 sec, and a total of 40 cycles at 95˚C for 5 sec and extension at 60˚C for 34 sec, followed by a dissociation curve analysis. All assays were performed three times. Relative expression levels were then determined using the 2^-ΔΔCq method (20).

A Caspase-3 Activity Assay Kit (cat. no. C1116; Beyotime Institute of Biotechnology) was used according to the manufacturer's protocol to detect the activity of caspase-3 in the rat liver tissues. The absorbance was measured at 405 nm and the relative caspase-3 activity to the activity of the sham group was determined. Six samples from each group were analyzed.

The liver tissues were weighed and homogenized immediately in 10 volumes of saline after washing in saline. Supernatants were collected to perform ELISAs following centrifugation (12,000 x g for 10 min at 4˚C). ELISA kits (cat. nos. BMS630 and KRC2341, eBioscience; Thermo Fisher Scientific, Inc.) were used according to the manufacturer's instructions to measure the content of IL-1β and IL-18 in the liver samples. Each sample was tested in triplicate.

Liver tissue samples obtained from the rats were fixed with 10% formalin. The formalin-fixed specimens (3 µm) were stained using automatic H&E slide stainer (Leica ST5020 Multistainer; Leica Microsystems, Inc.) according to the manufacturer's instructions for histological evaluation. The stained liver tissues were examined for histopathological evidence of pathological damage under light microscope (magnification, x100).

HSP90 expression in the rat liver tissues was detected by IHC. Deparaffinization and rehydration of the sections (3 µm thickness) were performed with gradient xylene and a descending ethanol gradient. Antigen retrieval was performed with Citrate Antigen Retrieval solution (1:50; cat. no. P0083; Beyotime Institute of Biotechnology) at 95˚C for 20 min. Endogenous peroxidase activity was quenched using 0.3% hydrogen peroxide (25˚C, 10 min). The sections were blocked with 1% BSA (Beyotime Institute of Biotechnology) for 2 h at room temperature. The sections were then incubated with primary anti-HSP90 antibody (1:100; cat. no. 13171-1-AP; ProteinTech Group, Inc.) overnight at 4˚C, followed by incubation with horseradish peroxidase-conjugated secondary antibody (1:50; cat. no. A0208; Beyotime Institute of Biotechnology) for 1.5 h at room temperature. After thoroughly washing with washing buffer (cat. no. P0106L; Beyotime Institute of Biotechnology), the sections were developed using a 3,3'-diaminobenzidine kit (cat. no. P0202; Beyotime Institute of Biotechnology) and counterstained with hematoxylin staining solution (cat. no. C0107; Beyotime Institute of Biotechnology) for 10 min at room temperature. Each stained sample was observed under high power magnification using a light microscope (magnification, x200; Leica Microsystems, Inc.).

BDL and sham rat models were prepared using the aforementioned procedure. After 1 week, cholestatic blood samples were harvested from the rats and cholestatic serum was obtained after centrifugation (2,000 x g for 10 min at 4˚C). Cholestatic serum samples from all BDL model rats were mixed in one tube and non-cholestatic serum samples from all sham model rats were mixed in another tube. The sera were filtered using a 0.22-µm filter (Millipore; Merck KGaA) for sterilization. The automated chemical analyzer was used to examine the cholestatic and non-cholestatic serum. The cholestatic serum contained 188.02 µmol/l total bilirubin (TBIL) and 146.6 µmol/l direct bilirubin (DBIL), while the non-cholestatic serum contained 2.3 µmol/l TBIL and 1.72 µmol/l DBIL. The two types of serum were used in the subsequent experiments.

The BRL rat hepatic cell line was obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences. Liver sinusoidal endothelial cells (LSECs) were isolated from the rat liver tissues with modifications to the previously described protocol (21). Briefly, the rat liver tissue was perfused with collagenase (Invitrogen; Thermo Fisher Scientific, Inc.). The single-cell suspension was subjected to velocity and density centrifugations (50 x g for 2 min at 25˚C) in Percoll gradients to separate hepatocyte and nonparenchymal cell suspensions. The nonparenchymal cell suspension was incubated with CD146 (LSEC)-biotin antibodies (1:50; cat. no. 130-111-844; Miltenyi Biotec, Inc.) for cell surface staining (30 min at room temperature). The LSEC cells were positively selected by magnetic separation using Anti-Biotin MicroBeads (Miltenyi Biotec, Inc.) according to the manufacturer's instructions.

Prior to stimulation with serum, BRL cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.), and LSECs were cultured with a medium comprising 40% MCDB 131 (Gibco; Thermo Fisher Scientific, Inc.), 40% RPMI-1640 (Gibco; Thermo Fisher Scientific, Inc.) and 20% FBS. The ‘cholestatic culture’ was initiated when the FBS was replaced with the cholestatic serum or non-cholestatic serum obtained from the rats.

BRL cells were seeded into a 96-well plate at a density of 2x10³ cells/well and allowed to attach overnight at 37˚C. The cells were then cultured in a medium containing non-cholestatic serum or cholestatic serum, with or without 17-DMAG (0.5 µM) for 48 h. Cell proliferation was determined using the CCK-8 assay (MedChemExpress). Briefly, a 1/10 volume of CCK-8 solution was added to each well and the cells were cultured for 2 h. Cell viability was then determined by measuring the absorbance at 450 nm using a plate reader (BioTek Instruments, Inc.). Six repeated wells for each cell condition were analyzed.

BRL cells were cultured in a cholestatic medium with or without 17-DMAG (0.5 µM) at 37˚C. After 24 h, the cells were collected and stained using Annexin V-FITC and propidium iodide (Becton, Dickinson and Company) according to the manufacturer's instructions. Apoptotic cells were quantified by flow cytometry using a Flow Cytometer (Cytomics FC500; Beckman Coulter, Inc.). The data was analyzed with CXP software version 1.0 (Beckman Coulter). The experiments were performed in triplicates.

Data are presented as the mean ± SD. The statistical significance of a difference between two groups was determined using the Student's t-test. Two-way ANOVA followed by Bonferroni's adjustment were used to evaluate differences among multiple groups. P<0.05 was considered to indicate a statistically significant result. All statistical analyses were conducted using SPSS 13.0 software (SPSS, Inc.).

Serum levels of ALT, AST and TBIL were examined to confirm the successful establishment of the animal model. The significantly increased ALT, AST and TBIL levels in the BDL group compared with sham group were indicative of substantial cholestasis and liver injury in the BDL group (Fig. 1A). In addition to simple biliary obstruction, the gram-negative pathogen LPS is another factor that aggravates biliary obstruction-induced liver injury (22). Therefore, a BDL + LPS rat model was established to simulate frequently occurring clinical cases. As shown in Fig. 1A, LPS administration further aggravated the BDL-induced liver injury.

The mRNA expression levels of three HSP90 family members in the liver tissues of the BDL and BDL + LPS rat models were compared with those in the sham group to clarify the association between HSP90 expression and cholestatic liver injury. In the BDL and BDL + LPS groups, the results showed that Hsp90aa1 and Hsp90b1 mRNA levels were significantly upregulated compared with those in the sham group (Fig. 1B). IHC showed that the HSP90 protein expression was also upregulated in the liver tissues of rats in the BDL and BDL + LPS groups (Fig. 1C). Whether a similar phenomenon exists in human liver tissues was also examined. Liver specimens obtained from patients with hepatolithiasis or intrahepatic cholangitis who had undergone hepatectomy were analyzed using IHC. The results demonstrated that HSP90 expression was upregulated in the patients with intrahepatic cholangitis (Fig. 1D). These results imply that aberrant HSP90 expression is associated with cholestasis and cholangitis.

Previous studies have supported the potential of 17-DMAG, a water-soluble HSP90 inhibitor, as a therapeutic agent (23,24). The effect of 17-DMAG on cholestasis-induced liver injury was therefore investigated in vivo to identify whether the inhibition of HSP90 enhances the transcription of HSP70, as indicated by previous studies (25-27). The elevated mRNA levels of HSP70 members 1b and 4 (Hspa1b and Hspa4, respectively) in the rats following 17-DMAG administration indicated that the activity of HSP90 was effectively inhibited (Fig. 1E).

Following the administration of 17-DMAG or NS to the rats, the serum levels of ALT and AST were assessed. ALT and AST levels in the BDL and BDL + LPS groups exhibited significant reductions upon treatment with 17-DMAG compared with those in the NS-treated rats (Fig. 1A). However, 17-DMAG administration did not significantly mitigate the increase of TBIL levels after the onset of cholestasis (Fig. 1A).

A morphological assessment was performed to evaluate the degree of injury exhibited by the rat liver tissue. Pathological examination revealed dilation of the bile capillaries and increased local necrosis in the livers of rats with BDL-induced cholestasis (Fig. 1F). In addition, following LPS administration, large areas of necrosis and increased inflammatory cell infiltration were observed (Fig. 1F). However, compared with NS treatment, 17-DMAG treatment reduced local necrosis and inflammatory cell infiltration during cholestasis with or without LPS (Fig. 1F).

Caspase-3 activity and the ratio of Bax to Bcl2 are often used to evaluate the apoptotic response (28). Notably, caspase-3 activity increased significantly in the liver tissues of the BDL and BDL + LPS groups compared with the sham group. However, treatment with 17-DMAG effectively attenuated this change (Fig. 1G). Furthermore, the Bax to Bcl2 ratio exhibited similar changes (Fig. 1H). These results indicate that 17-DMAG effectively alleviated cholestasis-induced liver injury.

A ‘cholestatic culture’ experiment was conducted to further clarify the function of 17-DMAG during cholestasis-induced cell damage in vitro. Cholestatic serum from BDL model rats was added to the culture medium of BRL cells to simulate the cholestatic microenvironment in vitro. CCK-8 assays were then performed to evaluate cell viability. Moderate cytotoxicity was observed when BRL cells were treated with cholestatic serum (Fig. 2A). However, 17-DMAG administration significantly alleviated this cytotoxicity, whether or not LPS was present in vitro (Fig. 2A). In addition, the flow cytometric analysis of cell apoptosis indicated that 17-DMAG administration promoted the survival of BRL cells and reduced the number of apoptotic cells in the cholestatic serum-containing medium (Fig. 2B).

Pro-inflammatory cytokines induced by bile acid play an important role in cholestasis-induced liver injury (29,30). IL-1β and IL-18 are important cytokines, as they have a role in the induction of apoptosis during liver injury (15,31). Therefore, they were analyzed in the present study. The results showed that the mRNA expression levels of IL-1β and IL-18 increased significantly during cholestasis-induced liver injury, and treatment with 17-DMAG substantially attenuated these changes (Fig. 3A). ELISAs of the liver tissues also demonstrated that the cholestasis-induced increases in IL-1β and IL-18 levels were significantly attenuated following treatment with 17-DMAG (Fig. 3B).

LSECs serve an important role in the immune response and liver injury (32,33). LSECs were isolated from rats for in vitro analyses to study the underlying mechanism. The results showed that the addition of cholestatic medium significantly increased the expression of IL-1β and IL-18 by LSECs, compared with non-cholestatic medium (Fig. 3C and D). Notably, the administration of 17-DMAG significantly attenuated the increase in the expression of IL-1β and IL-18 induced by cholestatic medium (Fig. 3C and D).

Mammalian TLRs detect microbial infection and regulate the cell death-associated cellular response during stress. The HSP90 family, particularly HSP90B1, have been identified as TLR chaperones that inhibit TLR signaling and the biogenesis of downstream pro-inflammatory cytokines (16,17). TLR9 has been reported to induce liver injury via the regulation of IL-1β and IL-18 (15,34). Therefore, the TLR9 inhibitor E6446 dihydrochloride was used to treat the LSECs. The administration of E6446 also reversed the cholestatic medium-induced expression of IL-1β and IL-18, in a comparable manner to 17-DMAG (Fig. 3C and D). These results suggest a potential association between the protective effect of HSP90 inhibitors on the liver and the regulation of pro-inflammatory cytokine biogenesis by TLR9.