Detrimental role of SIX1 in hepatic lipogenesis and fibrosis of non‐alcoholic fatty liver disease

Nonalcoholic fatty liver disease (NAFLD) is the most common liver disease worldwide. Aberrant lipid metabolism and accumulation of extracellular matrix proteins are hallmarks of the disease, but the underlying mechanisms are largely unknown. This study aims to elucidate the key role of sine oculis homeobox homologue 1 (SIX1) in the development of NAFLD.


| INTRODUC TI ON
Nonalcoholic fatty liver disease (NAFLD) is becoming a major global health threat and is recognized as the most common risk factor for hepatocellular carcinoma (HCC) and liver transplantation. 1,2 NAFLD encompasses a disease spectrum ranging from simple steatosis to nonalcoholic steatohepatitis (NASH), which has manifestations of lipid accumulation, inflammation, and varying degrees of fibrosis. 3 A variety of metabolic conditions such as obesity, insulin resistance, and diabetes have been linked to NAFLD. Furthermore, Eslam et al. suggested a name change from NAFLD to metabolic associated fatty liver disease (MAFLD) to emphasize the association of NAFLD with metabolic risk factors. 4,5 Currently, diet-and lifestyle-associated obesity is a major risk factor for NAFLD, which is characterized by the excessive accumulation of fat in hepatocytes. 6 Imbalanced fat intake, synthesis, and oxidation increase the accumulation of lipids in the liver [7][8][9] 10 In addition, LXR can also regulate lipid synthesis through sterol regulatory element-binding protein 1c (SREBP1c) and carbohydrate response element binding protein (ChREBP). [11][12][13] The abnormal regulation of LXR increased DNL and was significantly associated with NAFLD. 14,15 To date, a number of protein factors impinging on the transcriptional activity of LXR have been elucidated.
As the disease progresses, NAFLD can progress to advanced fibrosis and cirrhosis, a condition with excess accumulation of extracellular matrix in the liver. Hepatic stellate cell (HSC) activation is the major reason for fibrogenesis. Several different profibrotic cytokines can activate HSCs, and transforming growth factor β (TGFβ) is one of the most potent inducers. TGFβ binds to the type Ι and II TGFβ receptor (TGFβRI/TGFβRII) and facilitates the translation of mothers against decapentaplegic homologue (Smad) family members 2 and 3 (Smad2/3) to the nucleus, which can further increase the transcription of fibrogenic genes. Given the heavy involvement of lipid synthesis and the TGFβ pathway in the progression of NAFLD, there is great potential in discovering therapies targeting these two processes.
Sine oculis homeobox homologue 1 (SIX1) is a transcription factor that is conserved from Drosophila to humans and is reported to be involved in a wide range of biological processes. 16 SIX1 is a key regulator of organogenesis but remains at low or undetectable levels in most normal adult tissues. 17 However, the hepatic function of the SIX1 system is still unknown. A recent study of SIX1 has attracted attention to its ability to modulate energy metabolism. It has been reported that SIX1 can increase the transcription of glycolytic genes, thereby promoting glycolysis. 18 In addition, overexpression of SIX1 can also activate TGFβ signalling in breast cancer. These results prompted us to explore the potential effect of SIX1 in NAFLD progression.
In this work, we found that SIX1 was upregulated in humans and experimentally induced mouse models of NAFLD.
Hepatocyte-specific overexpression of SIX1 exacerbated steatosis, inflammation, and fibrosis in mice fed a methionine-and choline-deficient diet (MCD) and high-fat high-cholesterol (HFHC) diet. Furthermore, we defined the impact of SIX1 on lipid synthesis-related genes by directly increasing the expression of LXRα and LXRβ. In addition, SIX1 also played a pivotal role in fibrosis by mediating the cascade of the TGFβ pathway in HSCs.
Collectively, we demonstrated that SIX1 emerges as a novel regulator in hepatic steatosis and fibrosis, providing new insight into the potential treatment of NAFLD.

| Human liver samples
Human liver tissues of NAFLD acquired from bariatric surgery patients in Xijing Hospital have been reported previously. 19 Patients were stratified based on the histological assessments with or without hepatic steatosis (NAFL) and lobular inflammation (NASH) as described previously. 19 All individuals provided written informed consent, and all the procedures were approved by the ethics committee of Xijing Hospital.

K E Y W O R D S
hepatic lipogenesis, liver fibrosis, non-alcoholic fatty liver disease, sine oculis homeobox homologue 1, TGFβ signalling pathway

Key points
In this study, we found that SIX1 is involved in hepatic lipogenesis and fibrosis modulation, and the overexpression of SIX1 increased liver lipid accumulation and hepatic fibrosis by inducing the expression of LXRα/β and activating of TGFβ signalling pathway, thus contributing to the progression of NAFLD.
The mice fed with HFHC diet were simultaneously supplied with sterile water containing 23 g/L fructose. The ingredients of diets were described in the Supporting Information.
Adeno-associated virus serotype 9 (AAV9) was used to obtain liver-specific SIX1 overexpression or knockdown mice. Albumin-Cre (Alb-Cre) mice (male, 8 weeks old) were injected with the AAV9-SIX1, AAV9-SIX1-KO, or AAV9-control virus at a dose of 10 12 vg per 200 μL per mouse. 20 After injection, one group of mice was fed MCD diet for 4 weeks to trigger steatohepatitis or for 8 weeks to obtain fibrosing steatohepatitis. Another group of mice was fed HFHC diet for 16 weeks to induce steatohepatitis. The body weights of the HFHC-diet mice were recorded every week. All animals were fasted, and the serum and tissues were harvested at the end of the experiment. Adeno-associated viruses stably overexpressing (pAAV2-hEF1a-DIO-SIX1-EGFP-WPRE-pA) or silencing (pAAV2-hTBG-DIO-miRNAi-EGFP-3Flag) SIX1 were described in the Supporting Information.

| Cell culture and treatments
Mouse immortalized hepatocytes (AML12) and human hepatocytes (HepG2) were cultured in Dulbecco's modified Eagle's medium (DMEM)/Ham's F12 medium and DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin, respectively. The human HSC line LX-2 was cultured in DMEM supplemented with 2% FBS and 1% penicillin/streptomycin. AML12, HepG2, and LX-2 cells were transfected with the lentivirus carrying the SIX1 expression cassette or control. Cells were cultured in serum-free medium overnight before experiments. Then, the cells were cultured in medium supplied with MCD or palmitic acid (PA), as indicated. In addition, SR9243 (1 μM, Glpbio, USA) was added to AML12 and HepG2 cells for 6 h. LX-2 cells were pre-treated for 16 h with 100 nM LY2157299 (Glpbio, USA) and then stimulated with 5 ng TGFβ for 3 h.

| Western blot analysis
Liver tissues or cells were extracted using radioimmunoprecipitation assay (RIPA) buffer combined with protease inhibition. Lysates were then centrifuged (15 000× g for 15 min), and the concentration of the protein in the supernatant was determined by the Bradford method.
A total of 30 μg of protein was separated by 10% SDS/PAGE and transferred to a polyvinylidene fluoride membrane. Then, the primary antibodies were incubated at 4°C overnight, and the secondary antibodies were incubated for 1 h at room temperature. The antibodies used are outlined in the Supporting Information.

| RNA isolation and real-time reverse transcription (RT)-PCR
Total RNA was isolated with an mRNA isolation kit (Qiagen, Germany), and 1 μg of total RNA was reverse transcribed to cDNA with a cDNA Synthesis Kit (TaKaRa, Japan). To calculate the relative expression of the cDNA, 18S was used as an internal control. Real-time PCR was performed in a 20 μL reaction by using qPCR Master Mix (Qiagen, Germany) and the Bio-Rad CFX-96 system. The gene-specific primer sequences used are outlined in Supporting Information.

| Immunofluorescence
Cells were grown on Millicell EZ SLIDE 4-well glass (Millipore, USA) and treated with MCD medium or PA stimulation, as indicated.
Cells were fixed with 4% paraformaldehyde for 15 min and washed twice with phosphate-buffered saline (PBS) solution. Then, the cells were permeabilized with .5% Triton X-100 and blocked with 10% goat serum. Cells were incubated overnight with the primary antibody against SIX1 (D4A8K, Cell Signalling Technology, USA); cells were then incubated with secondary antibody and stained with 4′,6-diamidino-2-phenylindole (DAPI).

| Glucose uptake assays
The Glucose Uptake Assay Kit (Promega, USA) was used to determine glucose uptake as previously described. 21 Detailed experimental steps were described in the Supporting Information.

| Immunohistochemistry
Mice liver tissues were fixed in the 4% paraformaldehyde solution overnight, then were subjected to haematoxylin and eosin (H&E),

Oil red O, Sirius Red and α-SMA staining by Wuhan Servicebio
Technology Ltd (Wuhan, China).

| Cleavage under targets and tagmentation assay
Cleavage under targets and tagmentation (CUT&Tag) assays, which performed by Romics Biotechnology Ltd (Shanghai, China), were used to map the chromatin fragments to the genome. We performed this assay using HepG2 cells with anti-SIX1 antibody as previously reported. 22 The detailed experimental steps were described in the Supporting Information.

| Dual-luciferase reporter assay
Full length or deletion of the LXRα and LXRβ promoters was amplified and cloned into the pGL3-Basic vector to generate reporter plasmids. HEK293T cells were seeded into 96-well plates and transfected with 100 ng of reporter plasmid, 100 ng of full-length human SIX1 plasmid, and 10 ng of pRL-TK plasmid, along with negative control plasmids using Lipofectamine 2000 (Invitrogen USA, #11668030). After 24 h transfection, dual-luciferase assays were performed using the dual-luciferase reporter assay system (Promega, USA) with a Victor X machine (PerkinElmer, USA). Firefly luciferase values were normalized to Renilla luciferase values. All the plasmids were confirmed by DNA sequencing.

| Statistical analysis
Values were obtained from three independent experiments and are presented as the mean ± standard deviation (SD). Pearson's correlation coefficient and Spearman's correlation coefficient were used to estimate the association of SIX1 levels and several factors of interest in NAFLD patient samples. Data were compared between two groups by Student's unpaired t-test. Comparisons between the control group and several experimental groups were performed by one-way analysis of variance (ANOVA). Two-way ANOVA with multiple comparisons was used to compare the differences within two groups for time-dependent measurements. p < .05 was considered statistically significant.

| SIX1 is overexpressed in livers of nonalcoholic steatohepatitis patients and three different male mice models with diet-induced nonalcoholic fatty liver disease
To investigate whether SIX1 is dysregulated in NAFLD, we first evaluated the protein expression of SIX1 in liver biopsies from eight patients with NASH, eight patients with nonalcoholic fatty liver (NAFL), and eight control individuals by western blotting.
The clinical data of these patients have been reported previously. 19 We found that the mRNA level of SIX1 was upregulated in patients with NAFL and NASH compared that in control individuals ( Figure 1A). Furthermore, SIX1 was positively correlated with alanine aminotransferase (ALT), aspartate aminotransferase (AST) and triglyceride (TG) levels ( Figure 1B-D). We further examined the protein level of SIX1 in human liver tissues. The results showed that SIX1 expression was significantly elevated in patients with NAFL and NASH compared with the control individuals ( Figure 1E). In addition, we also tested SIX1 expression in C57BL6/J male mice with steatosis induced by different diets, including MCD, HFHC, and HFD after STZ injection. The results showed that the expression level of SIX1 was increased in the animal models fed a special diet compared to that in male mice fed the control diet ( Figure 1F-H). Furthermore, the results of immunochemistry staining showed that the SIX1 almost located in the nucleus of hepatocytes and its expression exhibited much higher levels in three different male mice models with dietinduced NAFLD compared to that in mice fed with the control diet ( Figure 1I). These results indicate that SIX1 expression is increased in human and male murine livers with steatosis and steatohepatitis.

| Liver SIX1 overexpression aggravates high-fat high-cholesterol-induced nonalcoholic fatty liver disease progression in mice
To determine whether SIX1 affects NAFLD progression, Alb-Cre mice were administered the AAV9 vector for SIX1 overexpression ( Figure 2A). Western blot and qPCR analyses confirmed that SIX1 was overexpressed in the livers of AAV9-SIX1 mice compared to the AAV9-control group ( Figure 2B). Then, the SIX1-overexpressing mice and control mice were fed with HFHC diet for 16 weeks. The haematoxylin and eosin (H&E) and Oil red O staining indicated that the livers from HFHC-treated SIX1overexpressiong mice showed significantly increased lipid accumulation and more severe liver injury, which was consistent with the significant increase in serum ALT, AST levels, and TG contents of the serum ( Figure 2C,F), while SIX1 overexpression did not cause steatohepatitis or liver fibrosis in mice fed with the normal diet ( Figure S1). In addition, ectopic expression of SIX1 significantly aggravated diet-induced metabolic disorders, SIX1overexpressing mice fed with HFHC showed significantly more body weight gains, higher body weight and liver weight, and liver/body weight ratio than the AAV9-control mice fed with the same diet ( Figure 2D,E). In addition, HFHC-fed liver-specific SIX1 overexpression mice exhibited more glucose tolerance, as determined by intraperitoneal glucose tolerance test (IPGTT) assay ( Figure 2G). Additionally, SIX1-overexpressing mice presented more hepatic collagen deposition and liver triglyceride content quantified by Sirius Red staining and hepatic triglyceride assay ( Figure 2C,H). These date indicated the detrimental role of SIX1 in metabolic disorder and hepatic steatosis.
To further confirm the crucial role of SIX1 in liver fibrosis and hepatic inflammation, we detected the levels of α-smooth muscle actin (α-SMA) and myeloperoxidase positive (MPO + ), F4/80 + cells in hepatic compartment. IHC and IF results revealed significantly increased α-SMA levels, more infiltrated neutrophils, and macrophages in SIX1-overexpressing mice liver sections than in the control mice ( Figure 2I). Furthermore, more severe liver fibrosis and HSCs activation in SIX1-overexpressing mice were indicated by increased α-SMA and TGFB1 mRNA levels ( Figure 2J) and collagen protein levels ( Figure 2K). Finally, we examined NASH-related proinflammatory cytokines, including Mcp1, IL-10, IL-16, IL-1β, Tnfα, Mip1α, and inos by RT-PCR and increased pro-inflammatory genes expression was displayed in SIX1-overexpressing mice with HFHC diet ( Figure 2L). Therefore, these data collectively indicate that SIX1 overexpression aggravates the severity of HFHCinduced NAFLD.

| SIX1 promotes lipid accumulation and cell injury in cultured cell line
To elucidate the molecular mechanisms of SIX1 in NAFLD, we performed in vitro cell culture experiments. We used the mouse hepatic cell line AML12 and the human hepatoma cell line HepG2 and treated these hepatocytes with PA stimulation. As illustrated in Figure 3A, SIX1 expression levels were increased when AML12 and HepG2 cells were treated with PA stimulation. In addition, we found that PA stimulation could also induce the increased expression of SIX1 in the nucleus ( Figure 3B). To further address the functional significance of SIX1 in steatohepatitis, we overexpressed SIX1 in AML12 and HepG2 cells. When AML12 and HepG2 cells were exposed to PA stimulation, increased lipid accumulation was found with Oil Red O staining in the LV-SIX1 group compared to that in the control group ( Figure 3C). In addition, there was an increase in the levels of ALT and AST after PA stimulation in the SIX1 overexpression group ( Figure 3D). These results indicate that SIX1 could increase the lipid content and cell injury observed in both human and mouse hepatocytes.

| Identification of SIX1 as a key regulator of lipogenetic gene expression
SIX1 is reported to play an important role in glucose metabolism by inducing the expression of glycolysis-related genes. 18 We also found that overexpression of SIX1 could increase glucose uptake in AML12 and HepG2 cells, and glucose could be the main carbon precursor of DNL ( Figure 4A). Since SIX1 has the powerful ability to metabolize energy, we hypothesized that SIX1 may also augment DNL progression. To obtain a comprehensive understanding of the role of  (Table S1). The gene function analyses of CUT&Tag results indicated a significant enrichment of metabolic pathways ( Figure 4D), and we found that nuclear receptor subfamily 1 group h member 2 (NR1H2), which encodes the central regulator of the lipogenic progression LXRβ, was the potential target of SIX1. Furthermore, using a promoter luciferase assay, we found that SIX1 could activate LXRβ promoter activity ( Figure 4E, lower lane). Interestingly, a luciferase assay also showed that the promoter of LXRα could also be activated by SIX1, which highlighted the role of SIX1 in lipogenic progression ( Figure 4E, upper lane). Next, we demonstrated that hepatic expression of LXRα and LXRβ was upregulated in AAV9-SIX1 mice compared to that in control mice ( Figure 4F). In addition, we tested whether LXRα and LXRβ levels were altered in response to SIX1 overexpression in cultured cell lines and discovered an increase in the levels of LXRα and LXRβ ( Figure 4G). Moreover, we also found that PA-induced LXRα and LXRβ could be reversed by SIX1 knockdown ( Figure 4H). In addition, while SIX1 increased the levels of lipogenesis genes, it failed to do so after LXRα/LXRβ were inhibited, implying that SIX1 could regulate the downregulation of lipogenic gene expression via LXRα/ LXRβ in hepatocytes ( Figure 4I, Figure S2B). These results suggested that SIX1 augmented lipid content by inducing DNL by directly targeting LXRα/LXRβ.

| SIX1 overexpression promotes hepatic fibrogenesis though the TGFβ pathway in MCD diet-induced steatohepatitis
Our results have shown that SIX1 overexpression promotes lipid accumulation and induces more severe liver injury. In addition, we also observed more collagen deposition in HFHC-induced it did better mimick the pathogenesis of severe human NASH than other dietary models. 23 The mice fed with MCD diet developed more quickly liver fibrosis than fed with an HFHC diet. 24,25 First of all, 8-week-old mice were administered the AAV9 vector for SIX1 overexpression. Western blot and qPCR results confirmed that SIX1 was successfully overexpressed in the livers of AAV9-SIX1 mice compared to the AAV9-control group ( Figure 5A). SIX1-overexpressing mice with MCD diet developed more severe steatohepatitis and liver fibrosis than the control group, as indicated by Oil Red O, Sirius Red, and α-SMA staining ( Figure 5B). However, SIX1 overexpression did not have a phenotype on mice fed with the normal diet ( Figure S1). In addition, overexpression of hepatic SIX1 indeed increased the serum levels of ALT and AST ( Figure 5C). The hepatic triglyceride content and collagen deposition were significantly increased in SIX1-overexpressing mice with MCD diet, as measured by liver triglyceride assay and Sirius Red staining ( Figure 5D). Furthermore, the development of severe fibrosis was indicated by higher mRNA levels of α-SMA and TGFB1 ( Figure 5E  To further explore the mechanisms of SIX1 in fibrosis, we performed qPCR to verify fibrosis-related genes in SIX1-overexpressing mice. We found that a few representative genes, including TGFB, which is one of the most potent inducers of fibrogenesis, in hepatic fibrosis were increased in the livers of SIX1-overexpressing mice treated with an MCD diet ( Figure 5E,G). To confirm that TGFβ signalling was altered, we analysed the levels of phosphorylation of Smad3 (phosphorylated Smad3, p-Smad3), a downstream effector of the TGFβ pathway. Western blot analysis revealed increased levels of p-Smad3 ( Figure 5H) in the liver tissues of SIX1-overexpressing mice. Previous work shows that SIX1 increases TGFβ signaling, 26,27 and we also found that TGFβreceptor 1 (TGFBR1) and TGFβ receptor 2 (TGFBR2) were in the CUT&Tag results (Table S1). Since HSCs are major contributors to fibrogenesis through TGFβ, we next investigated the role of SIX1 in fibrosis using the human HSC cell line LX-2 and found an increase in the levels of TGFβRI and TGFβRII in response to SIX1 overexpression ( Figure 5I). In addition, the results also showed that stable SIX1-overexpressing LX-2 cells displayed increased p-Smad3 levels ( Figure 5J). To test whether SIX1-induced TGFβRI and TGFβRII could influence the activation of fibrogenicgenes, we treated LX-2 SIX1-overexpressing cells with a TGFβ receptor inhibitor (LY2157299) and examined the expression of the related genes in hepatic fibrosis. The results showed that inhibition of the TGFβ receptor could reverse the phosphorylation of Smad3 and the upregulation of fibrosis-related genes induced by SIX1 ( Figure 5K,L). Together, these results are consistent with the hypothesis that TGFβRI and TGFβRII are the targets of SIX1 in LX-2 cells and suggest that the regulation of TGFβRI and TGFβRII may be an underlying mechanism of SIX1-dependent activation of TGFβ signalling and the induction of liver fibrosis.

| Liver-specific SIX1 deficiency ameliorates diet-induced nonalcoholic fatty liver disease pathogenesis
To further confirm the effects of SIX1 in the progression of NAFLD, AAV-mediated SIX1 knockdown was generated in mouse model fed with HFHC or MCD diet. The western blotting and qPCR results confirmed the efficacy of AAV-KO system ( Figure 6A). First of all, to rule out the influence of SIX1 itself on hepatic metabolism, we detected hepatic lipid content and liver fibrosis using H&E, Oil red O, and Sirius Red staining in mice fed with the normal diet. Results indicated that SIX1 knockdown did not cause a phenotype on the control diet ( Figure S1). SIX1 knockdown in mice fed with HFHC for 16 weeks showed significantly alleviated hepatic steatosis and liver fibrosis, as indicated by H&E, Oil red O, Sirius Red, and α-SMA staining ( Figure 6B). After SIX1 ablation, the mice exhibited less body weight gains, lower body weight, liver weight, and body/liver weight ratio than the control mice fed with the same diet ( Figure 6C,D).
Serum ALT, AST and TG levels was markedly downregulated in livers of mice with hepatic SIX1 knockdown ( Figure 6E). The liver triglyceride and collagen content was also markedly decreased in SIX1-KO mice compared to the control ( Figure 6F). In keeping with this, HFHC-fed liver-specific SIX1-KO mice exhibited improved glucose tolerance, as determined by IPGTT assay ( Figure 6G). To verify the effect of SIX1 on hepatic inflammation observed in Figure 2I

| DISCUSS ION
Our findings here show for the first time that SIX1 is overexpressed in the liver of patients and mice with NAFLD and that the upregulation of SIX1 promotes lipid synthesis and liver fibrosis, thereby participating in the progression of liver steatosis and steatohepatitis. SIX1 remains at a low level in adult tissues, but its reexpression is known to contribute to tumour initiation and progression. 26, 28-31 SIX1 has been reported to be overexpressed in breast, 32,33 cervical, 34 and ovarian cancer. 29 In addition, SIX1 was upregulated and promoted liver cancer progression. However, SIX1 has not been investigated in noncancerous adult tissues such as NAFLD. Herein, we sought to investigate the potential role of LXRβ are the sensors that mediate glucose and FA metabolism. 35 In this sense, in the liver, SIX1 activates LXR and converts excess glucose into FAs. These results indicate that SIX1 can directly F I G U R E 6 Liver-specific SIX1 deficiency ameliorates diet-induced NAFLD pathogenesis. (A) The mRNA and protein levels of SIX1 in HFHC-induced liver-specific SIX1 knockdown by AAV9 system. (B) Representative images of liver sections stained by H&E, Oil-red O, Sirius red and α-SMA to evaluate the severity of steatohepatitis in indicated groups. (C) Body weight gains of control and SIX1-deficient mice fed with HFHC for 16 weeks. Data represent mean ± SEM. **p < .01, two-way ANOVA with multiple comparisons. (D) Body weight, liver weight, and liver/body weight ratio of HFHC mice were measured in indicated groups. (E) Serum ALT, AST, and TG levels were measured in indicated groups. (F) Hepatic triglyceride content and collagen deposition analysed by liver TG assay and Sirius Red staining (G) Glucose tolerance test was examined in AAV-control and AAV-SIX1-KO mice fed with HFHC. **p < .01, ***p < .001, and two-way ANOVA with multiple comparisons. (H) Representative images of H&E, Oil-red O, Sirius red, and α-SMA staining in SIX1 knockdown and control mice fed with MCD diet. (I) The levels of SIX1 in MCD-induced NASH were evaluated by western blotting and real-time PCR. (J) Hepatic triglyceride and collagen content quantified by liver TG assay and Sirius Red staining. (K) Serum ALT and AST levels were measured in indicated groups. (L) α-SMA and TGFB1 levels were measured by real-time PCR in SIX1 ablation and control mice fed with MCD diet. Statistics: n = 6, for each group. ns: not significant, *p < .05, **p < .01, and ***p < .001. regulate lipid metabolism in the liver. To our knowledge, this is the first report that shows the potential role of SIX1 in lipid metabolism. However, in addition to LXRα and LXRβ, we also found that SIX1 can bind to the promoters of ACC1 and SCD1. Therefore, the potential role of SIX1 in lipogenic progression may be more complex and should be further assessed.
A link between SIX1 and liver fibrosis was also observed in our in vivo study. Our results show that SIX1 can increase the expression of various fibrogenic genes such as collagen type I alpha 1 (COL1A1), collagen type III alpha1 (COL3A1), smooth muscle alpha 2 actin (ACTA2), TGF-β1, and tissue inhibitor of metalloproteinase 1 (Timp1) in the MCD-induced NAFLD model. Previous studies have demonstrated that SIX1 can modulate TGFβ signalling by increasing TGFβRI expression in mammary carcinoma cells 27 and TGFβRII in oesophageal cancer cells. 36 Micalizzi et al. also reported that the overexpression of SIX1 increased the levels of p-Smad3 and thus increased TGFβ signalling in breast cancer. 26 In our study, we also find that SIX1 can promote HSC activation by increasing the expression of TGFβRI, TGFβRII, and p-Smad3 in HSCs. In addition, the results show that SIX1-induced fibrogenic genes can be reversed by a TGFBR inhibitor. These data suggest that SIX1 increases the expression of fibrogenic genes through the induction of the TGFβ signalling pathway and promotes the progression of liver fibrosis. TGFβ release is required for the role of SIX1 in HSCs, and we found that TGFβ levels are increased in SIX1-overexpression mice. Further studies are warranted to elucidate the relationship between increased TGFβ and SIX1 in the pathogenesis of steatosis and NASH. Besides the above important findings, we admitted that there were some limitations in our study. All mice used in our study are male, while accumulated evidences have proven that sexual category, sex hormones, and gender habits could interact with numerous NAFLD factors including cytokines, stress, and environmental factors, thus altering the risk profiles and phenotypes of NAFLD. 37 Thus, it is essential to explore whether the gender would effect the phenotype induced by SIX1 overexpression or knockdown under the condition of NAFLD. Besides, as we known, transcription factors are notoriously difficult to target. A previous study has indicated that the small molecule inhibitor of SIX1 and its transcriptional complex eyes absent 2 (EYA2) were sufficient to suppress breast cancer metastasis without severe toxicity at a dose of 25 mg/kg for 21 days. 38 In addition, in our study, SIX1 knockdown with AAV was well tolerated in mice and did not reveal major toxicities. It still needs more robust evidences and pre-clinical validations when considering it as a therapeutic target.
Our findings suggest a detrimental function of SIX1 in the progression of NAFLD. The direct regulation of LXRα/β and TGFβ signalling by SIX1 provides a new regulatory mechanism in hepatic steatosis and fibrosis.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have declared that no competing interest exists.

PATI ENT CO N S ENT S TATEM ENT
Fully informed consent was obtained from all patients.

PE R M I SS I O N TO R E PRO D U CE M ATE R I A L FRO M OTH E R S O U RCE S
No materials reproduced from other sources in this manuscript.