Japanese white rabbits were provided by the Laboratory Animal Center of Xi'an Jiaotong University. A total of 20 male rabbits (16 weeks old, about 2.0 kg body weight) were randomly divided into three groups: The control group (n=4) was fed with a chow diet, HCD group (n=8) was fed with a chow diet supplemented with 1% cholesterol and high fat and high cholesterol diet (HFCD) group (n=8) was fed a chow diet containing 6.7% lard and 1% cholesterol for 12 weeks. Rabbits were fed with a restricted diet (100 g/day/animal) with free access to water.

Cholesterol was purchased from Wako Pure Chemical Industries, Ltd., (Osaka, Japan). The experimental protocols were approved by the Animal Administration Committee of Xi'an Medical University (Shaanxi, China) and performed according to the Xi'an Medical University Guidelines for Animal Experimentation and the Guide for the Care and Use of Laboratory Animals Published by the US National Institutes of Health (NIH publication no. 85-23, revised 1996).

Rabbits were fasted for 16 h prior to blood collection. Blood samples were collected via the ear artery and put into tubes containing EDTA and the plasma separated by centrifugation at 2,000 rpm/min (20 min, 4°C). Plasma total cholesterol (TC) and triglyceride (TG) levels were measured biweekly using commercial assay kits (BioSino Bio-technology and Science, Inc., Beijing, China). At 12 weeks, blood samples were collected to measure plasma free fatty acid (FFA) levels using commercial assay kits (BioVision, Inc., Milpitas, CA, USA). At the same time, plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured using commercial assay kits (Pars Azmun, Tehran, Iran).

To evaluate the effect of high cholesterol diet (HCD) on glucose metabolism, rabbits were fasted overnight and an intravenous glucose tolerance test (IVGTT) was performed using the method as described previously (22). A bolus of glucose (0.6 g/kg body weight) was injected through the ear vein and blood samples were collected through the ear artery at 5, 10, 15, 20, 30, 45, 60, 75 and 120 min. Plasma glucose levels were measured using commercial assay kits (Wako Pure Chemical Industries, Ltd., Osaka, Japan). The incremental area under the curve (AUC) was calculated according to the trapezium rule as described previously (23).

At the end of the experiment, the rabbits were euthanized by overdose of injection of sodium barbital solution. A piece of liver was quickly put into liquid nitrogen for homogenate and RNA extraction. About 100 mg liver tissue was homogenized in 0.75 ml of methanol and chloroform (2:1), and lipids were extracted from the chloroform fraction and N2-dried to remove the chloroform. Then, the hepatic TC and TG contents were determined using Wako kits (Wako Pure Chemical Industries, Ltd.). For the hepatic FFA assay, liver tissue was homogenized in 0.2 ml of chloroform with 1% Triton X-100. Fatty acids were extracted in the chloroform fraction and N2-dried to remove the chloroform. Then, the hepatic tissue FFA contents were determined using a FFA quantification kit (BioVision, Mountain View, CA, USA) (24). For reverse transcription-quantitative polymerase chain reaction (RT-qPCR) reactions, total RNA was extracted from liver samples using TRIzol Plus (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to manufacturer's instructions. A total of 10 µg total RNA was reverse transcribed into cDNA using Takara SuperScript™ RT reagent kit with gDNA eraser (cat. no. RR047A; Takara Biotechnology Co., Ltd., Dalian, China). A mixture of dNTPs were included in the PrimeScript RT enzyme Mix, and OligodT primers and random 6 mers were included in the RT primer mix of the SuperScript™ RT reagent kit. ERS-associated gene expression was measured by RT-qPCR. Forward and reverse primer sequences of ERS-associated gene are presented in Table II. GAPDH was used as the reference gene and the primer sequences are included in Table II. The thermocycling conditions used for qPCR were as follows: 95°C for 30 sec, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec, and then, 95°C for 15 sec, 60°C for 30 sec and 95°C for 15 sec. RT-qPCR analysis was performed using Takara SYBR Green kit (cat. no. DRR820A; Takara Biotechnology Co., Ltd.) by TaKaRa TP800 (Takara Bio, Inc., Otsu, Japan). Values of the quantification cycle (Cq) were used for quantification with the 2^−ΔΔCq method, as described previously (25).

Fresh liver samples were homogenized in lysis buffer at 4°C followed by centrifugation at 10,000 rpm at 4°C for 10 min. The resultant supernatants were collected and subjected to western blotting. Briefly, 15 µg lysates were fractionated on 10% SDS-polyacrylamide gels and then transferred to Sequi-Blot polyvinylidene fluoride membranes (Bio-Rad, Hercules, CA, USA). The membranes were incubated with each primary antibody (Ab) (1:1,000) at 4°C overnight, as recommended in the manufacturer's instructions. After washing 3 times, they were incubated with horseradish peroxidase conjugated secondary Ab for 2 h. Signals were detected using the Immobilon reagent (Millipore, Billerica, MA, USA) and visualized using an LAS-400 Lumino Image Analyzer (Fujifilm, Co., Tokyo, Japan) (26). Primary Abs of CHOP (L63F7), JNK (9251L) and Caspase-12 (2202P) were purchased from Cell Signaling Technology, Inc., (Danvers, MA, USA). Anti-GRP78 (ab89789) was purchased from Abcam (Cambridge, UK).

At the end of the experiment, a piece of liver was stocked in liquid nitrogen and embedded in optimal cutting temperature (OCT) compound and cut in 8-µm sections followed by oil red O staining. The areas stained by oil red O were quantified to assess the areas of connective tissue and lipid deposition, respectively, using an image analysis system (WinROOF v6.5) (24,27). At the same time, formalin-fixed liver was embedded in paraffin and cut in 4-µm thick sections and stained with hematoxylin and eosin (HE).

Data are expressed as the mean ± standard error of the mean. Statistical analysis was performed using analysis of variance and Dunnett's post hoc test, and the data with an equal F value, using GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference.

Plasma TC levels were remarkably increased in both HCD and HFCD groups compared with control group (P<0.05 or P<0.01). Plasma TG significantly increased in HFCD group compared with control group (P<0.05). At the end of the experiment, plasma FFA was measured and the results showed that FFA was significantly increased in HCD and HFCD groups compared with control group (P<0.05). We also performed a glucose tolerance test and found that glucose clearance was delayed in HCD and HFCD groups compared with control group (P<0.05 or P<0.01; Fig. 1). Plasma ALT and AST levels were not remarkably different among the three groups (Table I).

Compared with the liver of the control group, the surface of the liver of HCD and HFCD groups appeared obviously yellow and greasy (Fig. 2, left). Body weight and the ratio of the liver weight to body weight were not statistically significant in the three groups (Fig. 2A and B).

The analysis of liver histology revealed that there were a large number of vacuoles in the hepatocytes of the HCD and HFCD groups compared with control group as demonstrated by both HE and oil red O staining (Fig. 3). The areas stained by oil red O were measured to evaluate the severity of connective tissue and lipid deposition as previously described (24,26) and the results showed that NAFLD model was successfully established in both HCD and HFCD groups even though liver histological features were not significantly different (Fig. 3). To evaluate lipid compositions of liver, TC, TG and FFA contents were measured using liver homogenate. The results showed that hepatic TC and FFA contents were significantly increased in HCD and HFCD groups compared with control group, and hepatic TG content was significantly increased in HFCD group compared with control group. However significant differences were not observed between the HCD and HFCD groups (P<0.05 or P<0.01; Fig. 4).

Expression levels of ERS-associated liver genes were compared by RT-qPCR to investigate whether ERS plays any role in the pathogenesis of NAFLD. The genes and primer sequence were showed in Table II. Glucose regulation protein 78 (GRP78) mRNA expression was significantly decreased in the HCD and HFCD groups compared with the control group. JNK mRNA expression was increased in HCD group compared with control and HFCD group, while there was no significant difference between HFCD and control groups. Caspase-12 mRNA expression was also increased in the HCD and HFCD groups compared with control, and there was a significant difference between the HCD and HFCD groups (P<0.05 or P<0.01). CCAAT/enhancer-binding protein homologous protein (CHOP) mRNA expression were not statistically significant in the three groups. The protein expression levels of the ERS indicators were measured and the results exhibited a similar trend to the mRNA expression levels (Fig. 5).