Gene expression profile-based drug screen identifies SAHA as a novel treatment for NAFLD

Abstract

Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease worldwide. Being part of the metabolic syndrome, NAFLD is characterized by the deposition of triglycerides (TGs) as lipid droplets in the cytoplasm of hepatic cells. Recently, the rapid development of high-throughput genome analysis technologies provided opportunities to screen for new drugs for NAFLD. In this study, we screened for potential drugs based on the gene expression profiles of 73 compounds and identified histone deacetylase (HDAC) inhibitors as a novel treatment for the accumulation of lipids in hepatocytes. In the subsequent analysis and experiments, we discovered that SAHA inhibited the fatty acid and lipid metabolism pathways in hepatic cells and induced a significant deficiency of lipid accumulation in HepG2 and SMMC-7721 cells. Furthermore, SAHA inhibited lipid synthesis in hepatic cells by directly suppressing the expression of DGAT2. Hence, our study provides a novel method to screen for effective drugs for liver diseases and identifies SAHA as a potent treatment for NAFLD.

---

1. Introduction

Non-alcoholic fatty liver disease (NAFLD) is progressively becoming one of the leading causes of chronic liver disease worldwide¹ and affects over 20% of the general population.² Owing to the global obesity epidemic, the prevalence of NAFLD has shown a trend of rapid growth, even in younger patients.³ Moreover, NAFLD is closely associated with metabolic syndrome, insulin resistance and cardiovascular disease.² In addition, the first hit in the pathogenesis of NAFLD is the dysfunction of hepatic lipid metabolism, which is characterized by the deposition of triglycerides (TGs) as lipid droplets in the cytoplasm of hepatic cells (hepatic steatosis).⁴ Lifestyle changes and increased insulin sensitivity are the major therapeutic interventions for improving the symptoms of patients with early NAFLD.⁵ In this context, more effective therapeutic options are urgently needed for patients suffering from NAFLD.

Traditionally, drug discovery has been a challenging process that requires more than 10 years and has a high cost.⁶ Recently, with the rapid development of high-throughput analysis technologies, our understanding of diseases has been broadened.⁷ Lamb and his group first proposed Connectivity Map (Cmap) analysis based on the use of gene expression profiles from cultured human cells treated with bioactive small molecules and pattern-matching software to discover the functional connections among diseases, genetic perturbations and drug action.⁸

In the study of cancer therapy, Cmap analysis has made it possible to identify effective drugs for different kinds of cancers.^9–16 Moreover, drugs screened by means of Cmap analysis for inflammatory bowel disease,¹⁷ Alzheimer's disease,¹⁸ pain management,¹⁹ smoking cessation²⁰ and muscle atrophy²¹ also achieved exciting results, and newly identified drugs led to novel recommendations for the treatment of these diseases. However, no published study of NAFLD treatment has used Cmap analysis to identify a potential drug from a library of candidates.

Hence, according to this strategy, we performed an integrated analysis of the gene expression profiles of 73 small molecule drugs in HepG2 cells. We identified candidate drug SAHA as a potent compound for the treatment of NAFLD. In the following experiments, we discovered that lipid accumulation was dramatically decreased in liver cells treated with SAHA, accompanied by the downregulation of genes involved in lipid metabolism. Furthermore, we confirmed that the suppression of HDAC2 by SAHA directly inhibited the expression of DGAT2, which could block TG biosynthesis and reduce the accumulation of lipids in liver cells.

2. Materials and methods

2.1 Bioinformatics analysis

2.1.1 Collection and preprocessing of the gene expression profiles of 73 compounds. The eligible gene expression profiles of 73 compounds in human HepG2 cells were obtained from the National Center for Biotechnology Information Gene Expression Omnibus (GEO; accession numbers GSE28878, GSE40117, GSE4465, GSE51952, GSE5230, GSE6878, GSE12161 and GSE25547; details can be seen in Table S1, ESI†). The interactive web tool GEO2R, which allows users to compare two or more groups of samples in a GEO series, was used to identify the t scores from the eight independent GEO datasets.

2.1.2 Gene set enrichment analysis (GSEA). To determine whether a compound could affect the pathways involved in lipid and fatty acid metabolism, GSEA was conducted, and the normalized enrichment scores (NES) of 73 compounds were displayed in the heat map. The data were uploaded and interpreted using the WEB-based GEne SeT AnaLysis Toolkit (http://www.webgestalt.org/option.php), and the detailed GSEA parameters were as follows: the number of permutations was 1000, the P value was <0.01, the minimum number of genes was 20 and the maximum size of genes was 1000.

2.1.3 Collection and annotation of ChIP-Seq datasets for HDAC2. The ChIP-Seq dataset obtained with an HDAC2 antibody in HepG2 cells was downloaded as a set of .bed files from the UCSC Genome Browser (http://genome.ucsc.edu). ChIP-seeker, an R Bioconductor package, which was developed by Yu et al.,²² was used to annotate the peaks with their related genes and the distance to the closest transcription start site (TSS).

2.1.4 Gene expression dataset collection and identification of differentially expressed genes (DEGs). The gene expression datasets after SAHA treatment and HDAC2 knockdown were obtained from the GEO datasets (accession number GSE52232). The corresponding gene symbols of the probe sets were obtained through the annotation package, and the expression values were summarized for genes corresponding to multiple probes. t-Tests followed by the Benjamini–Hochberg (BH) correction were applied for the expression matrix, and in the Volcano plot, |fold change| > 2 and adjusted P value <0.01 were used to screen for DEGs.

2.1.5 Functional enrichment analysis. We identified the DEGs from the gene expression datasets of HDAC2 knockdown and control groups. Then, overlapping the DEGs and candidate target genes of HDAC2, which were located from approximately −3000 to +3000 bp around the TSS, screened out the candidate differentially expressed target genes.

2.2 Experiments

2.2.1 Experimental reagents. The mouse monoclonal β-actin antibody (66009-1) was purchased from Proteintech. The ACOX1 rabbit monoclonal antibody (ab184032), anti-DGAT1 rabbit polyclonal antibody (ab54037), anti-DGAT2 rabbit polyclonal antibody (ab102831) and mouse monoclonal antibody specific for CPT1A (ab128568) were purchased from Abcam. The anti-MGAT2 rabbit polyclonal antibody (sc-66963) and PPARα mouse monoclonal antibody (sc-398394) were obtained from Santa Cruz Biotechnology. Suberoylanilide hydroxamic acid (SAHA), oleic acid, Oil Red O (ORO) and other chemical reagents were purchased from Sigma-Aldrich. The dual-luciferase reporter assay kit was purchased from Promega.

2.2.2 Cell culture. Two human hepatocellular carcinoma cell lines, HepG2 and SMMC-7721, were purchased from the American Type Culture Collection and were cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM, Gibco BRL; DMEM, HyClone, Logan, UT) supplemented with 10% (v/v) heated foetal bovine serum (FBS, Gibco, Grand Island, NY), and all the cells were maintained in a 37 °C incubator with 5% CO₂ in air.

2.2.3 Oleic acid-induced hepatic steatosis and ORO staining. To induce steatosis in the in vitro model, 0.3 × 10⁶ HepG2 and SMMC-7721 cells were seeded respectively in 6-well plates and starved in serum-free medium overnight. Then, these cells were treated with a 0.5 mM concentration of an oleic acid–bovine serum albumin (OA–BSA) complex (molar ratio 4 : 1) with or without 10 μM SAHA for 48 hours. The extent of steatosis was quantified by an ORO staining-based colorimetric assay at 510 nm, and the ratio of the absorbance of ORO staining to cell viability was used to evaluate the extent of steatosis.

2.2.4 Total RNA purification and quantitative RT-PCR. Total RNA was extracted using the TRIzol reagent (Invitrogen), and 1 μg of total RNA was used for reverse transcription using the PrimeScript RT reagent kit with gDNA eraser (Takara) according to the manufacturer's instructions. Quantitative RT-PCR was performed with SYBR Green dye (Applied Biosystems) using the ViiA7 Real-Time PCR System (Applied Biosystems). All Q-PCR was performed in triplicate. Relative mRNA expression was calculated by the comparative Ct method using GAPDH as a control.

2.2.5 Western blotting. Treated cells were lysed in RIPA buffer (Solarbio) with protease inhibitors and phosphatase. The concentrations of proteins were detected using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific™), and proteins were separated by 10% SDS-PAGE followed by immunoblotting with antibodies. The signal was detected using a digital gel image analysis system (TANON 5500), and the band intensities were quantified using Tanon GIS software. All experiments were conducted in triplicate.

2.2.6 Plasmid constructs and luciferase activity assay. The possible promoter regions of DGAT2 were amplified by PCR using human genomic DNA as a template, and the fragments were inserted into the pGL3-basic vector (Promega, Madison, WI), which was previously digested with KpnI and HindIII (Takara). The resulting recombinant plasmid was characterized by sequencing. Then 0.5 × 10⁶ HepG2 cells were seeded in 6-well plates, and transient transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the instructions of the manufacturer. After 48 hours of treatment with SAHA or DMSO, the luciferase activity was measured using a luciferase reporter assay system. All experiments were performed in triplicate, and the data are presented as the mean ± SD.

2.2.7 Statistical analysis. All results were expressed as the mean ± SD obtained from three independent experiments. Differences between means were tested by t-tests, with a P value ≤0.05 as the significance threshold.

3. Results

3.1 Identification of a histone deacetylase inhibitor (HDACi) as a candidate drug for NAFLD from 73 compounds

As lipid metabolism disorders play an important role in the pathogenesis of NAFLD, we hypothesized that we might discover a pharmacological inhibitor of NAFLD that could downregulate the mRNA expression of genes related to lipid metabolism. In the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, we collected six pathways associated with lipid and fatty acid metabolism (Fig. 1).

Fig. 1 The 6 pathways involved in fatty acid and lipid metabolism in humans from KEGG.

To select microarray datasets for the Cmap analysis, we performed a comprehensive search of GEO databases from the NCBI, and used “HepG2 cells” as the keywords in our search. The datasets we choose should meet these criteria: (a) the studies had to be performed in HepG2 cells treated with compounds; (b) the gene expression data had to be profiled on the Affymetrix HU-133 plus 2.0 array platform; (c) the number of test samples and control samples had to be three or greater than three. Then, according to these inclusion criteria, eight GEO datasets were obtained: GSE28878, GSE40117, GSE4465, GSE51952, GSE5230, GSE6878, GSE12161 and GSE25547. To find effective drugs for NAFLD treatment, we downloaded these gene expression profiles from NCBI and a total of 73 compounds in HepG2 cells (Table S1, ESI†) were analyzed and screened to determine whether some of the compounds showed potency in affecting the pathways involved in lipid and fatty acid metabolism.

Fig. 2A showed that the 73 compounds were classified into three clusters based on the NES from GSEA analysis. Forty-one compounds, which are presented in the red area, induced upregulation of most of the six pathways; in the blue area, there were 19 compounds that promoted fatty acid biosynthesis (hsa00061) while significantly inhibiting the other 5 pathways; and the green area included 13 compounds that upregulated the glycerolipid metabolism (hsa00561) pathways and downregulated the other pathways. According to this comprehensive appraisal, the compounds in the blue area exhibited more inhibitory effects on lipid and fatty acid metabolism than the compounds in the green area. Therefore, we selected the top three compounds that significantly inhibited these pathways, which were named PCB153, IQ and TSA.

Fig. 2 Screening of the gene expression microarray data for 73 compounds in HepG2 cells to identify potent therapeutic drugs for NAFLD. (A) A heatmap revealed the regulation of 6 pathways associated with fatty acid and lipid metabolism after treatment with 73 compounds in the HepG2 cell line after normalization based on the NES from GSEA. To detect the effect of PCB153, IQ and TSA on the 6 fatty acid and lipid metabolism pathways, we used the differentially expressed genes (DEGs) obtained from treatment with the 3 compounds compared with control. Gene-set enrichment analysis (GSEA) was performed on the DEGs and all the genes in the 6 pathways. The enrichment plots of (B) PCB153, (C) IQ and (D) TSA for DEGs involved in fatty acid and lipid metabolism are shown.

GSEA analysis was conducted to elucidate the effects of PCB153, IQ and TSA on lipid and fatty acid metabolism in HepG2 cells, and the results showed that most of the genes in the six pathways were downregulated after treatment with these three compounds (Fig. 2B–D). PCB153 is a member of the polychlorinated biphenyls (PCBs), which are widespread environmentally persistent organic pollutants that have various adverse effects on the reproductive,^23,24 endocrine²⁵ and nervous systems.²⁶ IQ (2-amino-3-methylimidazo[4,5-f]quinoline) is a known carcinogen that causes multiple cancers in rodents, including hepatocellular carcinomas and intestinal tumours.²⁷ Hence, PCB153 and IQ were not viable candidate drugs due to toxicity.

TSA (trichostatin A), a hydroxamic acid, is a widely used HDACi with broad spectrum. HDACs catalyse the deacetylation of histones, maintaining the equilibrium state between histone acetylation and deacetylation, and are closely related to gene transcription and expression, affecting cell proliferation, differentiation, and apoptosis.²⁸ Nunes et al. revealed that TSA could decrease cholesterol levels in neuronal cells by modulating genes involved in cholesterol synthesis, uptake and efflux.²⁹ Therefore, we speculated that HDACis represent a potential therapeutic approach for NAFLD.

3.2 SAHA shows inhibitory effects on lipid accumulation in OA-induced hepatic steatosis

SAHA (vorinostat), which is an optimized structure with TSA as the lead compound, acquired US Food and Drug Administration (FDA) approval for the treatment of cutaneous T cell lymphoma in 2006. To explore whether SAHA has inhibitory effects on lipid metabolism, similar to TSA, the gene expression profiles of liver cells with and without SAHA treatment were acquired from the GEO database and analysed utilizing GSEA analysis. Notably, all six pathways involved in lipid and fatty acid metabolism were downregulated significantly by SAHA treatment (Fig. 3A).

Fig. 3 SAHA shows inhibitory effects on lipid accumulation in HepG2 and SMMC-7721 cells. (A) GSEA measuring the correlations between the genes and the 6 pathways involved in fatty acid and lipid metabolism following treatment with SAHA and DMSO. (B) After HepG2 and SMMC-7721 cells were incubated with 10 μM SAHA and DMSO in the presence of 0.5 mM oleic acid for 48 hours, the cells were stained with the Oil Red O stain for 30 min, and the representative images (left) and statistical analysis (right) are presented. In the images, the scale bar below each picture is 20 μm. (C) The mRNA expression of genes related to fatty acid and lipid metabolism in HepG2 and SMMC-7721 cells treated with increasing concentrations of SAHA or DMSO for 24 hours was detected by quantitative RT-PCR. (D) The expression of genes involved in fatty acid and lipid metabolism was detected by immunoblotting using different antibodies in HepG2 and SMMC-7721 cells treated with 10 μM SAHA or DMSO for 24 hours. The intensities were calculated by normalizing against β-actin. All data are represented as the mean ± SD (n = 3; *P < 0.05 vs. DMSO; **P < 0.01 vs. DMSO; ***P < 0.001 vs. DMSO).

To confirm the impact of SAHA on lipid synthesis in vitro, we set up an OA-induced hepatic steatosis model in HepG2 and SMMC-7721 cells. Lipid accumulation in the cells was examined by ORO staining. As expected, lipid accumulation decreased dramatically in HepG2 cells treated with SAHA compared to that of the control in the presence of oleic acid (Fig. 3B). Similarly, SAHA significantly reduced the lipid content in SMMC-7721 cells.

Then, we analysed the mRNA expression levels of some enzymes that are involved in the metabolism of lipids by real-time RT-PCR in HepG2 and SMMC-7721 cells. As shown in Fig. 3C, compared to the control treatment, SAHA dramatically suppressed the mRNA expression of the key rate-limiting enzymes of triacylglycerol synthesis (MGAT2 and DGAT1) and the key lipogenic enzyme FASN in a dose-dependent manner. Furthermore, PPARα, which is closely associated with the biological processes of gluconeogenesis and fatty acid beta-oxidation, was remarkably upregulated after SAHA treatment. Additionally, the mRNA expression of the genes ACOX1 and CPT1A, the rate-limiting enzymes of fatty acid beta-oxidation, also increased progressively with increasing concentrations of SAHA.

To further confirm that treatment with SAHA can alleviate the accumulation of lipids in hepatocellular carcinoma (HCC) cells, we detected the expression of proteins involved in lipid metabolism in HepG2 and SMMC-7721 cells by western blotting. As expected, in comparison with that of the control group, the expression of MGAT2, DGAT1 and FASN was significantly decreased by SAHA. Similarly, the expression levels of the ACOX1, CPT1A and PPARα proteins were increased significantly (Fig. 3D). Analysis of these data by repeated measures demonstrated that SAHA could ameliorate hepatic steatosis by inhibiting lipid accumulation in HCC cells.

3.3 Inhibition of HDAC2 by SAHA suppresses expression of the DGAT2 transcript

SAHA is a well-accepted HDAC2 inhibitor, and HDACs modulate histones and non-histone proteins to regulate chromatin structure and gene expression.³⁰ To further elucidate the function of SAHA in ameliorating the accumulation of lipids in NAFLD, we analysed the ChIP-Seq data for HDAC2 from the Encyclopedia of DNA Elements (ENCODE) and obtained a total of 18 647 enriched peaks. Among these peaks, 6206 were located from approximately −3000 to +3000 bp around the TSS of the corresponding genes (Fig. 4A). We also screened the gene expression data (GSE52232) obtained after the depletion of HDAC2 in HepG2 cells by RNA interference. There were 943 candidate genes that exhibited considerable changes in expression (P value <0.01); these genes consisted of 420 upregulated genes and 523 downregulated genes. In addition, to further explore the biological significance of HDAC2 in the transcriptional regulation of genes related to lipid metabolism in liver cells, we integrated the predicted HDAC2 target genes from the ChIP-Seq datasets with these DEGs and observed that the expression of DGAT2, which was one of the predicted target genes of HDAC2, was dramatically decreased (Fig. 4B). The mRNA level and protein levels of DGAT2 were verified, and we discovered the inhibitory effect of SAHA on the expression of DGAT2 (Fig. 4C and D).

Fig. 4 Screening of candidate HDAC2 target genes associated with fatty acid and lipid metabolism by bioinformatics analysis in the HepG2 cell line. (A) The distribution of HDAC2 binding areas across the genome after the analysis of ChIP-Seq datasets in the HepG2 cell line from the Encyclopedia of DNA Elements (ENCODE) using ChIP-seeker. (B) Based on the combined analysis of the ChIP-Seq and gene expression datasets, the Volcano plot was used to summarize DEGs between the HDAC2 knockdown and the control group and show target genes of HDAC2. (C) The mRNA expression of DGAT2 in HepG2 and SMMC-7721 cells treated with increasing concentrations of SAHA or DMSO for 24 hours was detected by quantitative RT-PCR. (D) The protein expression of DGAT2 was detected by immunoblotting in HepG2 and SMMC-7721 cells treated with 10 μM SAHA or DMSO for 24 hours. The DGAT2 intensity was calculated by normalizing against β-actin. All data are represented as the mean ± SD (n = 3; *P < 0.05 vs. DMSO; **P < 0.01 vs. DMSO; ***P < 0.001 vs. DMSO).

DGAT2 is the enzyme that catalyses the final step in the biosynthesis of TGs and has been recognized as a potential therapeutic target for obesity and diabetes in recent years.^31,32 Subsequently, we investigated whether HDAC2 could directly regulate the transcription of DGAT2 through a luciferase assay. We constructed a series of luciferase reporter plasmids driven by the HDAC2 binding site-containing DGAT2 promoter (Fig. 5A) and performed luciferase reporter assays using HepG2 cells. As shown in Fig. 4D, the downregulation of HDAC2 by SAHA markedly decreased luciferase activity driven by the full-length DGAT2 promoter or short fragments of the DGAT2 promoter but not the region from approximately −217 to −62 in HepG2 cells; in addition, a DGAT2 promoter fragment (approximately −83 to +128) was sufficient for HDAC2 binding and regulation (Fig. 5B). Consistent with the initial analysis, these results further confirmed that directly suppressing the binding of HDAC2 to the DGAT2 promoter by SAHA treatment led to a decrease in DGAT2 expression, which could inhibit lipid accumulation and attenuate hepatic steatosis development and progression.

Fig. 5 SAHA inhibits lipid synthesis in hepatic cells by directly suppressing the expression of DGAT2 by HDAC2. (A) The predicted binding sites of HDAC2 in the transcription start site of DGAT2 in the HepG2 cell line. (B) A series of DGAT2-luciferase reporter plasmids were transfected into HepG2 cells. Relative DGAT2 promoter activities were measured 48 hours after transfection by the dual-luciferase assay. All data are represented as the mean ± SD (n = 3; ***P < 0.001 vs. DMSO).

4. Discussion

HDACs are a family of enzymes that lead to the deacetylation of lysine residues in histone and non-histone proteins, and they control histone acetylation with histone acetyltransferases.³³ Increasing evidence indicates that HDACs can regulate the expression of multiple genes associated with the pathogenesis of many diseases, such as gastric cancer,^34–36 liver cancer,³⁷ nasopharyngeal carcinoma,² cervical cancer³⁸ and lymphocytic leukaemia.³⁹ Recently, more investigators have focused on the effect of HDACis on diseases involved in metabolic disorders. In HepG2 and F9 cells, TSA treatment resulted in clear repression of genes involved in the cholesterol biosynthesis, fatty acid biosynthesis and glycolysis pathways and showed potency in the control of cholesterol levels in humans.⁴⁰ SAHA, which is known as vorinostat, is structurally related to the natural product TSA. It has also been revealed that the HDACis TSA and SAHA could increase both mRNA and protein levels of CLA-1/SR-BI in macrophages, which played an anti-atherosclerotic role in atherosclerosis.⁴¹

In our study, Cmap analysis based on eight independent datasets screened out TSA as a compound that could inhibit most of the pathways associated with lipid and fatty acid metabolism, and we chose the optimized HDACi SAHA to conduct our experiments. To investigate the correlation of lipid metabolism and the expression of related genes following treatment with SAHA and DMSO in HepG2 cells, GSEA analysis was carried out, and the results demonstrated that SAHA significantly inhibited all the pathways of lipid metabolism. Furthermore, ORO staining also verified that SAHA significantly decreased lipid accumulation in both HepG2 and SMMC-7721 cells. In addition, the marked reduction in the expression levels of lipogenesis genes (FASN, DGAT1 and MGAT2) demonstrated that SAHA reduced hepatic cell lipid accumulation by decreasing lipid synthesis. We also observed significant increases in the mRNA and protein expression of important regulators of lipid beta-oxidation (PPARA, CPT1A and ACOX1) in HepG2 and SMMC-7721 cells treated with SAHA, implying that SAHA also induced lipid beta-oxidation in hepatic cells. In a model of obesity and diabetes (db/db mice), SAHA treatment induced a significant reduction of circulating TGs, non-esterified fatty acids and lipids that accumulate in the liver of db/db mice and dramatically decreased both fasting glycaemia and circulating insulin.⁴⁰ Based on this evidence, SAHA, as a safe and effective drug that is already available commercially, exhibits potency for the treatment of metabolic diseases and activity against multiple solid tumours and haematological malignancies.

Recently, a study reported that SAHA significantly improved insulin resistance, learning and memory performance and ameliorated biochemical abnormalities induced by a high-fat diet in mice.⁵ To further elucidate the function of SAHA in ameliorating the accumulation of lipids in NAFLD, the ChIP-Seq data for HDAC2 from ENCODE and the gene expression data obtained after HDAC2 knockdown from the GEO database (GSE52232) identified 187 downregulated and 122 upregulated genes among all candidate target genes of HDAC2. We found that DGAT2 was among the dramatically downregulated candidate target genes of HDAC2. DGAT2, which is in the glycerolipid metabolism pathway, is the critical rate-limiting enzyme catalysing the de novo synthesis of triacylglycerols from newly synthesized fatty acids and nascent diacylglycerols. The results of real-time quantitative PCR and western blotting assays verified that the expression of DGAT2 was significantly decreased in HepG2 and SMMC-7721 cells treated with SAHA. Although the DGAT1 and DGAT2 enzymes catalyse the final step in the synthesis of TGs, they have different catalytic activities. It has been reported that DGAT2 overexpression causes more TG accumulation in rat hepatoma cells than DGAT1 overexpression, and in mice, DGAT2 deficiency results in a severe decrease in TG, while DGAT1 deficiency leads to a moderate reduction.^42,43 Moreover, the result from our luciferase reporter assay further confirmed that the inhibition of HDAC2 by SAHA directly suppressed the transcription of DGAT2 in HepG2 cells through the predicted binding sites in the DGAT2 promoter area. Furthermore, Victor et al. reported that DGAT2 is the link between glycaemia and triglyceridaemia, which are important markers of the metabolic syndrome.⁴⁴ Based on the results of this study and the previous reports, the inhibition of HDAC2 by SAHA might be a potential therapeutic strategy for NAFLD.

5. Conclusion

Our results highlight the pivotal role of SAHA, an HDACi, in the epigenetic regulation of transcription of genes involved in lipid metabolism based on transcriptome data and bioinformatics analytical techniques. Subsequent experiments further confirmed the anti-NAFLD effects of SAHA and suggested that SAHA is a potential effective drug for the treatment of NAFLD.

Author contributions

Libin Deng designed the experiments and interpreted the data. Zhujun Cheng, Yusong Wen, Bowen Liang, Siyang Chen, Yujun Liu, Zang Wang and Jiayu Cheng performed the experiments. Xiaoli Tang, Hongbo Xin and Zhujun Cheng analysed the data. Zhujun Cheng and Libin Deng wrote the paper. All authors read and approved the manuscript.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was supported in part by the National Natural Science Foundation of China (81660041 to Libin Deng and 81200631 to Xiaoli Tang) and the Natural Science Foundation of Jiangxi Province (20181BAB205008 and 20133BCB23007 to Libin Deng and 20171BAB205109 to Xiaoli Tang).

References

1. A. C. Tuyama and C. Y. Chang, J. Diabetes, 2012, 4, 266–280.
2. X. L. Yang, C. D. Zhang, H. Y. Wu, Y. H. Wu, Y. N. Zhang, M. B. Qin, H. Wu, X. C. Liu, X. Lina and S. M. Lu, Asian Pac. J. Cancer Prev., 2014, 15, 4663–4670.
3. J. M. Pappachan, S. Babu, B. Krishnan and N. C. Ravindran, J. Clin. Transl. Hepatol., 2017, 5, 384–393.
4. X. Zhang, J. H. Zhang, X. Y. Chen, Q. H. Hu, M. X. Wang, R. Jin, Q. Y. Zhang, W. Wang, R. Wang, L. L. Kang, J. S. Li, M. Li, Y. Pan, J. J. Huang and L. D. Kong, Antioxid. Redox Signaling, 2015, 22, 848–870.
5. S. Sharma and R. Taliyan, Neuropharmacology, 2016, 105, 285–297.
6. M. P. Licciardello, A. Ringler, P. Markt, F. Klepsch, C. H. Lardeau, S. Sdelci, E. Schirghuber, A. C. Muller, M. Caldera, A. Wagner, R. Herzog, T. Penz, M. Schuster, B. Boidol, G. Durnberger, Y. Folkvaljon, P. Stattin, V. Ivanov, J. Colinge, C. Bock, K. Kratochwill, J. Menche, K. L. Bennett and S. Kubicek, Nat. Chem. Biol., 2017, 13, 771–778.
7. A. Julia, J. Invest. Dermatol., 2017, 137, 1000–1002.
8. J. Lamb, E. D. Crawford, D. Peck, J. W. Modell, I. C. Blat, M. J. Wrobel, J. Lerner, J. P. Brunet, A. Subramanian, K. N. Ross, M. Reich, H. Hieronymus, G. Wei, S. A. Armstrong, S. J. Haggarty, P. A. Clemons, R. Wei, S. A. Carr, E. S. Lander and T. R. Golub, Science, 2006, 313, 1929–1935.
9. D. C. Hassane, M. L. Guzman, C. Corbett, X. Li, R. Abboud, F. Young, J. L. Liesveld, M. Carroll and C. T. Jordan, Blood, 2008, 111, 5654–5662.
10. F. M. Siu, D. L. Ma, Y. W. Cheung, C. N. Lok, K. Yan, Z. Yang, M. Yang, S. Xu, B. C. Ko, Q. Y. He and C. M. Che, Proteomics, 2008, 8, 3105–3117.
11. M. H. Chen, W. L. Yang, K. T. Lin, C. H. Liu, Y. W. Liu, K. W. Huang, P. M. Chang, J. M. Lai, C. N. Hsu, K. M. Chao, C. Y. Kao and C. Y. Huang, PLoS One, 2011, 6, e27186.
12. S. Claerhout, J. Y. Lim, W. Choi, Y. Y. Park, K. Kim, S. B. Kim, J. S. Lee, G. B. Mills and J. Y. Cho, PLoS One, 2011, 6, e24662.
13. A. K. Reka, R. Kuick, H. Kurapati, T. J. Standiford, G. S. Omenn and V. G. Keshamouni, J. Thorac. Oncol., 2011, 6, 1784–1792.
14. M. Sirota, J. T. Dudley, J. Kim, A. P. Chiang, A. A. Morgan, A. Sweet-Cordero, J. Sage and A. J. Butte, Sci. Transl. Med., 2011, 3, 96ra77.
15. G. Wang, Y. Ye, X. Yang, H. Liao, C. Zhao and S. Liang, PLoS One, 2011, 6, e14573.
16. L. Yang and P. Agarwal, PLoS One, 2011, 6, e28025.
17. J. T. Dudley, M. Sirota, M. Shenoy, R. K. Pai, S. Roedder, A. P. Chiang, A. A. Morgan, M. M. Sarwal, P. J. Pasricha and A. J. Butte, Sci. Transl. Med., 2011, 3, 96ra76.
18. F. Chen, Q. Guan, Z. Y. Nie and L. J. Jin, Am. J. Alzheimer's Dis., 2013, 28, 693–701.
19. M. Chang, S. Smith, A. Thorpe, M. J. Barratt and F. Karim, Mol. Pain, 2010, 6, 56.
20. J. O. Boyle, Z. H. Gumus, A. Kacker, V. L. Choksi, J. M. Bocker, X. K. Zhou, R. K. Yantiss, D. B. Hughes, B. Du, B. L. Judson, K. Subbaramaiah and A. J. Dannenberg, Cancer Prev. Res., 2010, 3, 266–278.
21. S. D. Kunkel, M. Suneja, S. M. Ebert, K. S. Bongers, D. K. Fox, S. E. Malmberg, F. Alipour, R. K. Shields and C. M. Adams, Cell Metab., 2011, 13, 627–638.
22. G. Yu, L. G. Wang and Q. Y. He, Bioinformatics, 2015, 31, 2382–2383.
23. Y. Chen, M. S. McLachlan, S. Kaserzon, X. Wang, L. Weijs, M. Gallen, L. M. Toms, Y. Li, L. L. Aylward, P. D. Sly and J. F. Mueller, Environ. Res., 2016, 147, 259–268.
24. I. Ali, B. Julin, A. Glynn, J. Hogberg, M. Berglund, J. E. Johansson, S. O. Andersson, O. Andren, E. Giovannucci, A. Wolk, U. Stenius and A. Akesson, Carcinogenesis, 2016, 37, 1144–1151.
25. A. Mesnier, S. Champion, L. Louis, C. Sauzet, P. May, H. Portugal, K. Benbrahim, J. Abraldes, M. C. Alessi, M. J. Amiot-Carlin, F. Peiretti, P. Piccerelle, G. Nalbone and P. H. Villard, PLoS One, 2015, 10, e0128847.
26. A. Szczepkowska, C. Lagaraine, V. Robert, J. Mlynarczuk, L. Dufourny, J. C. Thiery and J. Skipor, Ecotoxicol. Environ. Saf., 2016, 124, 303–308.
27. S. A. Bol, J. Horlbeck, J. Markovic, J. G. de Boer, R. J. Turesky and A. Constable, Carcinogenesis, 2000, 21, 1–6.
28. V. Moresi, M. Carrer, C. E. Grueter, O. F. Rifki, J. M. Shelton, J. A. Richardson, R. Bassel-Duby and E. N. Olson, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 1649–1654.
29. M. J. Nunes, M. Moutinho, M. J. Gama, C. M. Rodrigues and E. Rodrigues, PLoS One, 2013, 8, e53394.
30. Y. H. Lee, D. Seo, K. J. Choi, J. B. Andersen, M. A. Won, M. Kitade, L. E. Gomez-Quiroz, A. D. Judge, J. U. Marquardt, C. Raggi, E. A. Conner, I. MacLachlan, V. M. Factor and S. S. Thorgeirsson, Cancer Res., 2014, 74, 4752–4761.
31. S. Friedel, K. Reichwald, A. Scherag, H. Brumm, A. K. Wermter, H. R. Fries, K. Koberwitz, M. Wabitsch, T. Meitinger, M. Platzer, H. Biebermann, A. Hinney and J. Hebebrand, BMC Genet., 2007, 8, 17.
32. C. S. Choi, D. B. Savage, A. Kulkarni, X. X. Yu, Z. X. Liu, K. Morino, S. Kim, A. Distefano, V. T. Samuel, S. Neschen, D. Zhang, A. Wang, X. M. Zhang, M. Kahn, G. W. Cline, S. K. Pandey, J. G. Geisler, S. Bhanot, B. P. Monia and G. I. Shulman, J. Biol. Chem., 2007, 282, 22678–22688.
33. X. X. Zheng, T. Zhou, X. A. Wang, X. H. Tong and J. W. Ding, Atherosclerosis, 2015, 240, 355–366.
34. Q. Shen, W. Tang, J. Sun, L. Feng, H. Jin and X. Wang, Am. J. Transl. Res., 2014, 6, 538–547.
35. X. Zhu, J. Liu, X. Xu, C. Zhang and D. Dai, Int. J. Oncol., 2015, 46, 195–204.
36. Y. L. Li, T. S. Yang, W. M. Ruan, W. Cui, Y. Jin and X. M. Zou, Int. J. Clin. Exp. Med., 2014, 7, 1958–1966.
37. Q. Q. Shi, G. W. Zuo, Z. Q. Feng, L. C. Zhao, L. Luo, Z. M. You, D. Y. Li, J. Xia, J. Li and D. L. Chen, Asian Pac. J. Cancer Prev., 2014, 15, 7849–7855.
38. D. H. Yang, J. W. Lee, J. Lee and E. Y. Moon, PLoS One, 2014, 9, e97352.
39. L. Peiffer, S. J. Poll-Wolbeck, H. Flamme, I. Gehrke, M. Hallek and K. A. Kreuzer, J. Cancer Res. Clin. Oncol., 2014, 140, 1283–1293.
40. S. V. Chittur, N. Sangster-Guity and P. J. McCormick, BMC Genomics, 2008, 9, 507.
41. Y. Bao, Y. Yang, L. Wang, L. Gao, W. Jiang, L. Wang, S. Si and B. Hong, Atherosclerosis, 2009, 204, 127–135.
42. S. J. Stone, H. M. Myers, S. M. Watkins, B. E. Brown, K. R. Feingold, P. M. Elias and R. V. Farese, Jr., J. Biol. Chem., 2004, 279, 11767–11776.
43. S. J. Smith, S. Cases, D. R. Jensen, H. C. Chen, E. Sande, B. Tow, D. A. Sanan, J. Raber, R. H. Eckel and R. V. Farese, Jr., Nat. Genet., 2000, 25, 87–90.
44. V. A. Zammit, Biochem. J., 2013, 451, 1–12.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8mo00214b

---