The animal experiments were approved by the Institutional Animal Care and Research Advisory Committee at Guizhou Medicine University, Guiyang, China. A total of 18 8-week-old (weight, 18–20 g) male C57BL/6 mice (Experimental Animal Center of Hubei) were randomly divided and fed with a normal chow diet containing 4% fat (NCD group, n=6) or a HFD containing 60% fat and 2% cholesterol (HFD group, n=6) (23), or a HFD plus daily infused TMZ (40 mg/kg) by gavage (HFD + TMZ group, n=6) for 8 weeks before sacrifice. All mice were housed in a temperature-controlled environment in 12/12 h light/dark cycles with free access to food and water.

HepG2 cells were maintained in Dulbecco's modified Eagle's medium (DMEM, Sigma-Aldrich; Merck KGaA) with 10% fetal bovine serum (FBS, Beijing Transgen Biotech Co., Ltd.), in an incubator kept at 37°C with 5% CO₂. At 70–80% confluence, the cells were incubated in serum-free medium or serum-free medium containing 0.25 mM palmitic acid (PA; Sigma-Aldrich; Merck KGaA), or 10 µM TMZ (Laboratoires Servier) or 1 µM Compound C (MedChemExpress) for another 24 h.

The livers were fixed in 4% formaldehyde at room temperature and embedded in paraffin, sections were at a thickness of 5 µm. All staining was performed at room temperature. The hepatic morphology was detected by haematoxylin and eosin (H&E) staining and Oil Red O (ORO) staining. Staining with Sirius red was performed to visualize liver tissue fibrosis.

TG and total cholesterol (TC) levels in plasma or HepG2 cells were determined by GRO-PAP analysis kits (Nanjing Jiancheng Bioengineering Institute). TG and TC in liver were measured on an AEROSET Clinical Chemistry System (Abbott Laboratories).

siRNA targeting FOXO1 (NM_002015) and the negative control (NC) were purchased from Guangzhou RiboBio Co., Ltd. siRNA transfections were performed with Lipofectamine^® 2000 (Invitrogen, Thermo Fisher Scientific, Inc.) at a final concentration of 100 nM. After 48 h cells were used for subsequent experiments.

Total RNA was isolated using a TRIzol reagent kit (Invitrogen, Thermo Fisher Scientific, Inc.) and quantified by qPCR using a SYBR Premix Ex Taq II mix (Takara Bio, Inc.), according to the manufacturer's instructions. The quantitative PCR was performed in an ABI 7900 Fast Real-Time PCR System (Applied Biosystems Inc.), following the manufacturer's protocols. mRNA levels were quantified using the 2^−ΔΔCq method (24) and normalized to the internal gene β-actin, primers are listed in Table I.

Protein was extracted using RIPA lysis buffer (Wuhan Boster Biological Technology, Ltd.). Protein (20 µg) was separated by 8% SDS-PAGE and transferred on to PVDF membrane. The membranes were incubated with antibodies of ChREBP (1:1,000, Abcam, cat. no. ab92809), fatty acid synthase (FASN; 1:500, Wuhan Boster Biological Technology, Ltd., cat. no. PB0909), acetyl-CoA carboxylase (ACC; 1:500, Wuhan Boster Biological Technology, Ltd., cat. no. BM4414), AMPK (1:1,000, Cell Signaling Technology, Inc., cat. no. 2532), phosphorylated (p)-AMPK (Thr172; 1:1,000, Cell Signaling Technology, Inc., cat. no. 2535), AKT (1:1,000, Cell Signaling Technology, Inc., cat. no. 9272), p-AKT (Ser473; 1:1,000, Cell Signaling Technology, Inc., cat. no. 9271), GAPDH (1:2,000, Wuhan Boster Biological Technology, Ltd., cat. no. BM1623) and FOXO1 (1:500, Wuhan Boster Biological Technology, Ltd., BM4249) overnight at 4°C. Then, membranes were incubated with rabbit anti-mouse lgG (cat. no. BA1048) and goat anti-rabbit lgG antibodies (cat. no. BA1039; both from Wuhan Boster Biological Technology, Ltd.). Finally, the results were detected by ECL reagents (Beyotime Institute of Biotechnology) and semi-quantified by densitometry using the Tanon 5200 automatic chemiluminescent imaging system (Tanon Science and Technology Co., Ltd).

The HepG2 cells were subjected to DMEM with 5 nM 22-NBD cholesterol (Invitrogen; Thermo Fisher Scientific, Inc.) for 6 h. The cholesterol efflux was measured following 2-h incubation with apolipoprotein (apo)A-I (25 mg/ml) or high-density lipoprotein (HDL; Guangzhou Yiyuan Biotech Co., Ltd.) particles isolated from human plasma (50 mg/ml).

Bioinformatics analyses (http:///jaspar.genereg.net/) demonstrated that there were a number of FOXO1 binding sites in the ChREBP promoter region. The 2,256 bp ChREBP promoter sequence was amplified from human DNA by PCR, with Mlu1 or HindIII restriction enzyme cutting site at 5′ and 3′ end, respectively. The primers used were 5′-CTCTGAACCCATCTTTGAGGC-3′ and 5′-AGTCTGTGTCCGAGTCCGAgT-3′. PCR products were digested and inserted into a pGL-3 luciferase reporter vector (Beijing Augct DNA-Syn Biotechnology Co., Ltd.). 293T cells (1×10⁶ cells per well) in 24-well plates were co-transfected with 400 ng of plasmid, 50 ng of Renilla luciferase plasmid and 100 nM si-FOXO1 or 100 nM si-NC, using Lipofectamine^® 2000 (Invitrogen, Thermo Fisher Scientific, Inc.). Cells were incubated at 37°C and 5% CO₂. After 48 h, the cells were washed using cold PBS and lysed with passive lysis buffer (Promega Corporation). Luciferase activities were then measured by a luminometer (Sirius; Berthold Detection Systems GmbH) according to the manufacturer's instructions. Luciferase expression levels were adjusted with reference to Renilla luciferase activity. Six independent experiments were performed for each reporter vector.

Statistical analyses were performed with SPSS 24.0 (IBM Corp.). Quantitative variables were expressed as mean ± standard deviation of n experiments. Comparisons of quantitative variables between two groups were performed by the paired or unpaired Student's t-test. One-way ANOVA was used to compare multiple variables followed by a Tukey's post hoc test. The differences among three or more groups were analyzed by a multiple comparisons test. All probability values were two-sided and P<0.05 was considered to indicate a statistically significant difference.

Metabolic disorder was induced in mice with continuous HFD exhibiting body weight gain, hyperlipidemia, hyperglycemia and glucose intolerance compared with NCD mice (Fig. 1). To clarify the potential effect of TMZ on HFD-mediated metabolic disorder, the administration of TMZ by gavage was used. The TMZ significantly decreased the liver weight of HFD + TMZ mice compared with the HFD mice, but it did not change the whole body weight as shown in Fig. 1A and B. HFD + TMZ mice exhibited 25–30% lower plasma TG/TC levels than those of HFD mice (Fig. 1C and D) and improved glucose tolerance compared with the HFD mice (Fig. 1E and F). All these data led to the hypothesis that TMZ might ameliorate hepatic lipid metabolism in HFD mice.

To investigate the hypothesis above, the liver was investigated. As shown in Fig. 1G, H&E and ORO staining demonstrated excessive lipid accumulation in HFD mice liver compared with the NCD mice liver. TMZ markedly decreased hepatocyte bullous steatosis and hepatic fibrosis in HFD + TMZ mice liver compared with the HFD mice (Fig. 1G and H). In addition, TMZ decreased the expression of transforming growth factor-β and Collagen IV in livers of HFD + TMZ mice compared with those of the HFD + TMZ mice, which also suggested that TMZ protected hepatic fibrosis in the livers of HFD + TMZ mice compared with those of the HFD + TMZ mice (Fig. 1I).

The hepatic TG/TC level in HFD + TMZ mice was markedly decreased compared with HFD mice (Fig. 2A and B). Notably, TMZ treatment demonstrated a lower mRNA and protein expression level of ChREBP, which was a critical transcription factor in hepatic lipogenesis in the liver of HFD + TMZ mice compared with those of the HFD mice (Fig. 2C and D). As expected, TMZ had the same effects on mRNA expression of ACC and FASN in livers of HFD + TMZ mice compared with those of the HFD mice. Furthermore, HFD mice demonstrated decreased phosphorylation of AMPK (Thr172) compared with NCD mice (Fig. 2D). TMZ treatment significantly reversed the reduced phosphorylation of AMPK.

To verify the protective action of TMZ on NAFLD in vitro, the palmitate-treated liver cancer cell line HepG2 was used to confirm the animal experiment results. As demonstrated in Fig. 3A and B, ORO staining demonstrated that TMZ reduced excessive lipid accumulation, as before. There was no evidence to show that TMZ effected cholesterol efflux to apoA-I and HDL in liver cancer cell lines (Fig. 3C). A number of fat synthesis-related genes and cholesterol efflux-related genes were detected in liver cancer cells treated with TMZ and palmitate. TMZ only decreased the ChREBP mRNA expression. Liver X receptor (LXR)α, sterol regulatory element-binding 1 (SREBP1), CD36, SCARB1, trafficking kinesin protein 2, ATP-binding cassette subfamily A member 1 and ATP-binding cassette subfamily G member 1 demonstrated no significant difference in expression (Fig. 3D). In addition, TMZ also significantly decreased the mRNA and protein expression of key lipogenic genes, such as pyruvate kinase L/R, ACC and FASN (Fig. 3E and G).

It was further investigated how TMZ affected fatty acid and fat synthesis in the liver cancer cell line. The level of TG and cholesterol in HepG2 cells are shown in Fig. 4A and B. The reduced TG and cholesterol level in TMZ-treated liver cancer cells were both increased following intervention with the AMPK inhibitor Compound C. It was then established that TMZ increased the PA-related activation of AMPK and suppressed the PA-related activation of AKT in vitro. All these effects were reversed when the Compound C was added (Fig. 4C-E). The expression of FOXO1 was associated with the AMPK/AKT pathway (Fig. 4F). Nevertheless, the relationship between the AMPK/AKT pathway and ChREBP-mediated lipid accumulation remains to be determined. As shown in Fig. 4G and H, suppression of FOXO1 decreased the expression of ChREBP in the HepG2 and TMZ-PA-treated HepG2 cells. Furthermore, the suppression of FOXO1 decreased the luciferase reporter activity of the vector in 293T cells (Fig. 4I). These results suggested that TMZ suppressed hepatic lipogenesis by regulating ChREBP expression via AMPK-FOXO1 pathway.