ATR (SA9830, 97% pure) was purchased from Solarbio Life Science (Beijing, China). ATR was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 2 mg/mL as a stock solution; then, the working solution was diluted at concentrations of 0.1 mg/ml and 0.2 mg/ml, respectively. The dose of ATR (2 and 4 mg/kg of body weight) used in this study was chosen according to the previous study (Zhang et al., 2020).

All experimental procedures performed in this study were approved by the Animal Care and Use Committee (Approved number: GDY2002225, date: August 19, 2020) of Guangdong Medical University (Zhanjiang, China).

32 male ICR wide-type (WT) mice (6–8-week-old, 40–45 g) were obtained from Experimental Animal Center of Guangdong Medical University (Guangdong, China, production license: SCXK (Yue) 2018-0008) and allowed to acclimatize to the animal facility environment for a week before the experiment. Four mice were housed per cage in a specific pathogen-free laboratory (temperature 24 ± 2°C, humidity 50 ± 5%, and a standard 12 h light/dark cycle) with access to food and water. A total of 32 mice were distributed according to a single factorial arrangement. All mice were randomly divided into 4 treatment groups (8 mice per group and 4 mice per cage), including the normal control group, HFD control group, HFD with low ATR group (2 mg/kg ATR), and HFD with high ATR group (4 mg/kg ATR). Control mice were fed with a maize- and soybean-based basal diet (16.7% calories from fat, 3.77 Kcal/g diet), and mice in the other three groups were feed with a high-fat diet (GD60 rodent diet, Table 1, 60% calories from fat, 5.24 Kcal/g diet). All the food was purchased from Guangdong Animal Laboratory Animal center. The total feeding duration was 56 days (8 weeks), and ATR solution was given by intraperitoneal injection at a volume of 0.4 ml for 2 weeks before the feeding duration ends. Feed and water were provided ad libitum throughout the experiment.

The body weights of mice were recorded every 3 days from the 7th week to the 8th week, and the food consumption of each cage was recorded weekly. The final liver weight and epididymal fat weight were also recorded and the results were expressed as a relative weight to body weight (g/kg of body weight).

At the end of the experiment, about 1 ml blood samples were collected from the angular vein of mice under anesthesia by pentobarbital. Serum samples were obtained by centrifugation at 1790 × g for 10 min at room temperature and later stored in microtubes at −80°C. The levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglyceride (TGs) were estimated using commercially available kits purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, Jiangsu, China) in accordance with the manufacturer’s instruction.

Each mouse’s liver was collected and placed in a cryogenic vial after being washed with ice-cold sterile saline solution to remove blood contamination. For the determination of hepatic lipid contents, about 0.5 g of liver was used to prepare the homogenate. The minced tissue was homogenized in an ice-cold phosphate buffer solution (w/v, 1:9) using a glass homogenate tube and was centrifuged at 1,346 × g for 10 min at room temperature, and then the supernatant was collected and stored at −80°C until further analysis. The TG level in each sample was estimated using a commercially available kit purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, Jiangsu, China) in accordance with the manufacturer’s instructions. The protein concentrations of the tissue homogenates were determined via bicinchoninic acid assay (Walker, 2009).

For the histological analysis, liver tissues were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. The fixed liver samples were embedded with optimal cutting temperature compound (OCT); then, each sample was used to prepare 5 slides and each slide had 3 sections (10 μm thickness), which were stained with hematoxylin and eosin (H&E) and then sealed by a neutral resin size. For oil red O staining, the liver slides were stained with freshly prepared oil red O solution for 10 min. Afterward, the slides were washed thrice with distilled water and imaged under an inverted light microscope (Axio Vert. A1, Carl Zeiss, Germany). The ratio of oil red O was subjected to densitometric analysis based on the average percent of LDs accumulated hepatocytes per field at 200 magnification in 20 random fields by ImageJ 1.45 software (NIH, United States). For the NAFLD activity score (NAS), the liver slides were each read twice by 3 pathologists from the pathology department of our hospital. Weighted kappa scores were used to measure the degree of inter-rater and intra-rater agreement between and within multiple pathologists. The NAS is defined as the unweighted sum of the score of steatosis (0–3), lobular inflammation (0–3), and ballooning (0–2). Thus, the score is ranging from 0 to 8 (Kleiner et al., 2005). In brief, the severity of steatosis was graded from 0 to 3 (0 for <5%, 1 for 5–33%, 2 for 33–66%, 3 for >66%). Ballooning was determined by the number of prominent ballooning cells and scored from 0 to 3 (0 for none, 1 for few balloon cells, 2 for many cells/prominent ballooning). The lobular inflammation was assessed by counting the number of inflammatory foci per 200× magnification field in 20 random fields and scored 0–3 (0 for no foci, 1 for <2 per field, 2 for 2–4 per field, and 3 for >4 per field. The fibrosis was scored 0–4 (0 for no fibrosis, 1 for perisinusoidal or periportal fibrosis, 2 for perisinusoidal and portal/periportal fibrosis, 3 for bridging fibrosis, and 4 for cirrhosis). Notably, fibrosis is less reversible and generally thought to be a result of disease activity, which is not included as a component of the activity score (Kleiner et al., 2005). Cases with NAS of 0–3 were largely considered not diagnostic with steatohepatitis; on the other hand, cases with a score of >4 were diagnosed as steatohepatitis, and the scores between 3 and 4 were diagnosed with the possibility of steatohepatitis. Additionally, cases without ballooning, lobular inflammation, and fibrosis, but with the ratio of steatosis more than 33%, were diagnosed as NAFLD, and cases with the ratio of steatosis between 5 and 33% were diagnosed as hepatocytes steatosis.

The total protein samples were prepared by extracting liver in RIPA lysis buffer (Beyotime Biotechnology, Jiangsu, China), followed by centrifugation at 12,000 rpm for 15 min at 4°C. Supernatants containing total proteins were quantified with a BCA protein estimation kit (Meilun Biotechnology, Dalian, China) according to the previous method (Walker, 2009), and 50 μg of the total protein samples was resolved with 8, 10, or 15% polyacrylamide gel (depending on the molecular size of the proteins to be analyzed). Thereafter, immunoblotting was performed by transferring the resolved protein onto a PVDF membrane in a trans-buffer at 100 V for 1 or 2 h depending on the molecular size of the protein. Furthermore, the PVDF membranes were blocked with 5% skimmed milk or 5% BSA (for phosphorylated protein) in TBST buffer (20 mM Tris-HCl, pH 7.5, 150 mM sodium chloride, and 0.05% Tween 20) for 2 h, washed thrice with TBST buffer for 5 min each, and incubated overnight with primary antibodies including ANT2 (A15639, ABclonal), AMPKα (5831, CST), phospho-AMPKα (2535, CST), Raptor (2280, CST), mLST8 (3274, CST), mTOR (2983, CST), phospho-mTOR (2974, CST), SQSTM1/p62 (23214, CST), LC3A/B (3868, CST), and ATG7 (8558, CST) at 4°C. Afterward, the membranes were washed thrice with TBST buffer for 5 min each time and incubated with anti-rabbit or anti-mouse (depending on the species of primary antibody) secondary antibodies conjugated to horseradish peroxidase (1:2,000) for 2 h at room temperature. The blots were developed in the dark by using an ECL detection kit. The developed blots were subjected to densitometric analysis using ImageJ 1.45 software (NIH, United States) and normalized with GAPDH (5174, CST) as the internal control.

The liver slides were prepared as before; the fixed liver slides were permeabilized with 0.1% Triton X-100 for 10 min and then washed and blocked with 2% BSA for 1 h at room temperature. After blocking, the slides were incubated with LC3B (83506, CST) mouse antibody (1:300) overnight at 4°C, followed by incubation with the Cy3 goat anti-mouse IgG (AS008, ABclonal) secondary antibody (1:500) for 1 h at room temperature. Furthermore, the slides were incubated with perilipin 2 (PLIN2; A6276, ABclonal) rabbit antibody (1:200) overnight at 4°C, followed by incubation with the FITC goat anti-rabbit IgG (ab6717, Abcam) secondary antibody (1:500) for 1 h at room temperature. Finally, the slides were counterstained with DAPI for 5 min to stain the nucleus. The coverslips were mounted on slides and visualized under a confocal microscope equipped with 493, 550, and 455 nm lasers (LSM 800, Carl Zeiss, Germany).

Quantitative data were presented as the mean ± SD and statistically evaluated via one-way ANOVA followed by Tukey’s multiple comparison test using the Graph Pad Prism 6 software (GraphPad software, San Diego, CA, United States). In the data analysis, p < 0.05 was considered statistically significant.

The body weights of HFD mice were significantly higher than WT mice. Starting from the 4th day, injection of ATR compromised the increased body weight of HFD mice, and the body weight was significantly decreased by a high concentration of ATR at the end of the experiment (14th day; Figure 1A). In contrast to the WT group, HFD induced the relative weight (RW) of liver and epididymal fat, but ATR inclusion reversed the increased RW of those internal organs (p < 0.05; Figures 1C, D). The serum AST level was markedly increased in the HFD group, but the lower level of AST was observed in the ATR group as compared with the HFD group (p < 0.05; Figure 1F). No significant impact of ATR on food intake (Figure 1B) and serum ALT level was found (p > 0.05; Figure 1E). These findings suggested that ATR can reverse the changes of body weight, RW of liver and epididymal fat, and the serum AST level induced by HFD.

The livers of HFD feed mice were paler in colors than those of WT mice, but the colors of the livers in the ATR group were redder (Figure 2A, up channel) than those in the HFD mice. On histological analysis, the score of lobular inflammation and hepatocellular ballooning both were 0, as no obvious changes were observed in the H&E staining (Figure 2A, middle channel). Additionally, the ratios of steatosis were all between 5 and 33% in the liver of HFD mice, and the steatosis scores were all 1. Thus, the NAS were the weighted value of steatosis score (the value is 0.375 for all, no significant impact of ATR was observed). The intracellular lipid was analyzed via oil red O staining-lipid droplets (LDs; Figure 2A, down channel), and the results showed that LDs in HFD mice liver (area ratio of oil red O is 13%; Figure 2B) were significantly higher than that in WT mice (area ratio of oil red O is 0.3%; Figure 2B); however, the LDs in HFD mice liver were reduced in the presence of ATR, the area ratio of oil red O is 8.25% for low-ATR, and 6.37% for high-ATR (Figure 2B). Furthermore, the serum TG level in HFD mice was significantly reduced by a low concentration of ATR (p < 0.05; Figure 2C), and the hepatic TG level was decreased by ATR, both in the low and high concentration of ATR (p < 0.05; Figure 2D). Additionally, fibrosis of the liver was analyzed by Masson trichrome staining, the results showed that the hepatic fibrosis of HFD mice was observed in perisinusoidal or periportal space (Supplementary Figure S1B), and about 0.79% fibrosis was measured (Supplementary Figure S1B). ATR treatment decreased hepatic fibrosis, in both low (0.61% fibrosis) and high concentration (0.35% fibrosis) (Supplementary Figure S1B). Altogether, these results implied that ATR can improve the NAFLD steatosis and decrease HFD-induced lipid accumulation in the liver.

As shown in Figure 3, HFD treatment increased the protein expression of ANT2 (p < 0.05) and decreased the p-AMPKα protein expression, the value of p-AMPKα/AMPKα, and the Raptor expression (p < 0.05 for all); furthermore, HFD treatment increased the protein expression of p-mTOR and the value of p-mTOR/mTOR (p < 0.05 for all). In contrast to the HFD group, ATR treatment decreased the ANT2 protein expression (p < 0.05) and increased the p-AMPKα protein expression, the value of p-AMPKα/AMPKα, and the Raptor expression (p < 0.05 for all); moreover, ATR decreased the protein expression of p-mTOR and the value of p-mTOR/mTOR (p < 0.05 for all). No significant change was observed in the protein expression of mLST8.

The protein expressions related to autophagy were determined to analyze the role of autophagy in liver steatosis induced by HFD. As shown in Figure 4, HFD treatment decreased the protein expression of ATG7 and LC3A/B-Ⅱ (p < 0.05 for all). In contrast to the HFD group, ATR treatment increased the protein expression of LC3A/B-Ⅱ both in the low and high concentration of ATR (p < 0.05), and only a high concentration of ATR increased the ATG7 protein expression (p < 0.05). No significant change was observed in the protein expression of P62.

Colocalization of LC3B and Perilipin 2 (PLIN2) was performed to determine whether the reduced LDs were associated with autophagy or not. As shown in Figure 5A, the liver slice of HFD mice exhibited a limited amount of colocalization of LC3B and PLIN2 (Figure 5A, row 2), as evidenced by the coefficient of Rr (−0.621) in the HFD group (Figures 5B,C). Meanwhile, the liver co-treated with HFD and ATR showed an increased amount of colocalization of LC3B and PLIN2 (Figure 5A, row 3 and 4), as indicated by the coefficients of Rr (0.519 for low-ATR and 0.967 for high-ATR) in the ATR group (Figures 5B,C). Besides, the optical density analysis showed that ATR treatment reduced the green fluorescence of FITC-stained PLIN2 and increased the red fluorescence of Cy3-stained LC3B (Figure 5D; p < 0.05 for all). These data revealed the direct association between LC3B and PLIN2 in ATR-treated HFD mice livers.