Research ArticlePhysiology

BTG1 ameliorates liver steatosis by decreasing stearoyl-CoA desaturase 1 (SCD1) abundance and altering hepatic lipid metabolism

See allHide authors and affiliations

Science Signaling  17 May 2016:
Vol. 9, Issue 428, pp. ra50
DOI: 10.1126/scisignal.aad8581

Altering gene expression to prevent fatty liver

Fat accumulation in the liver, a condition called hepatic steatosis, can lead to liver dysfunction and hepatocellular carcinoma. The transcription cofactor BTG1 is decreased in hepatocellular carcinoma. Here, Xiao et al. show that the amount of BTG1 was also decreased in mice that are a genetic model of obesity. BTG1 suppressed the transcription of a gene encoding an enzyme involved in fatty acid synthesis, and overexpression of BTG1 in the livers of obese mice resulted in less hepatic steatosis. In addition, mice that overexpressed BTG1 in the liver were protected against diet-induced fatty liver, suggesting that treatments that increase the activity of BTG1 could be developed to prevent hepatic steatosis.

Abstract

Liver steatosis, a condition in which lipid accumulates in liver cells, is a leading cause of many liver diseases. The livers of patients with hepatocellular carcinoma, a cancer characterized by liver steatosis, have decreased abundance of the transcription cofactor BTG1 (B cell translocation gene 1). We showed that the livers of db/db mice, which are a genetic model of obesity, had decreased BTG1 mRNA and protein abundance. BTG1 overexpression ameliorated liver steatosis in db/db mice, whereas knockdown of BTG1 induced liver steatosis in wild-type mice. Consistent with these changes, we found that BTG1 decreased triglyceride accumulation in cultured hepatocytes. BTG1 overexpression inhibited the expression of the gene encoding stearoyl-CoA desaturase 1 (SCD1), an enzyme involved in the synthesis of fatty acids, by suppressing the activity of activating transcription factor 4 (ATF4). Knockdown of SCD1 prevented liver steatosis in wild-type mice induced by knockdown of BTG1. Conversely, the ability of BTG1 overexpression to ameliorate liver steatosis in db/db mice was negated by ATF4 overexpression. Moreover, BTG1 transgenic mice were resistant to liver steatosis induced by a high-carbohydrate diet. BTG1 abundance was decreased by this diet through a pathway that involved mammalian target of rapamycin (mTOR), ribosomal protein S6 kinase 1 (S6K1), and cAMP response element–binding protein (CREB). Together, our study identifies a role of BTG1 in regulating hepatic lipid metabolism and specifically in preventing ATF4 and SCD1 from inducing liver steatosis.

INTRODUCTION

Nonalcoholic fatty liver disease (NAFLD) is rapidly becoming a common liver disease worldwide and is correlated with metabolic syndrome and hepatocellular carcinoma (1, 2). NAFLD encompasses simple liver steatosis and nonalcoholic steatohepatitis (NASH), which ultimately results in fibrosis, cirrhosis, and liver dysfunction (3). Liver steatosis is caused by increased free fatty acid uptake, augmented de novo lipogenesis, decreased β-oxidation, and/or impaired triglyceride export (48). Some of the important transcription factors and/or cofactors that regulate these processes include sterol regulatory element–binding protein 1C (SREBP1C), which controls the de novo synthesis of fatty acids by affecting the expression of key enzymes involved in lipogenesis, fatty acid synthase (FAS), and stearoyl–coenzyme A (CoA) desaturase 1 (SCD1) (48). Peroxisome proliferator–activated receptor α (PPARα) and PPARγ coactivator 1α (PGC-1α) regulate the expression of the gene encoding carnitine palmitoyltransferase 1α (CPT1α), which is responsible for fatty acid β-oxidation (48).

B cell translocation gene 1 (BTG1) is a member of the BTG antiproliferative protein family [also known as transducer of ErbB2 (Tob)] (9). Initially identified as a translocation gene in a case of B cell chronic lymphocytic leukemia (10), it is conserved in all animal species and ubiquitously distributed in different tissues (11). BTG1 is involved in many important processes including cell growth, differentiation, and survival (12) and interacts with various transcription factors and affects their activities as a cofactor (1316). For example, BTG1 stimulates myoblast differentiation as a coactivator for myogenic factor calmodulin 1 (CMD1) (13) and inhibits the expression of cytokine-encoding genes in macrophages by repressing the activity of nuclear factor κB (NF-κB) (14).

BTG1 is involved in the pathogenesis of several diseases, including multiple sclerosis (17), breast cancer (18), and prostate cancer (19). BTG1 abundance is decreased in the livers of patients with hepatocellular carcinoma (20), a disease that is associated with liver steatosis (2). These results suggest a possible link between BTG1 and liver steatosis. We sought to investigate this possibility and elucidate the underlying mechanisms. We showed that BTG1 regulated hepatic lipid metabolism in vitro and in vivo by affecting the transcriptional activity of activating transcription factor 4 (ATF4), which functions as an upstream regulator for Scd1 expression. In addition, BTG1 abundance was decreased in mice on a high-carbohydrate diet (HCD) through a pathway that involves mammalian target of rapamycin (mTOR), ribosomal protein S6 kinase 1 (S6K1), and cyclic adenosine monophosphate (cAMP) response element–binding protein (CREB).

RESULTS

Overexpression of BTG1 reverses liver steatosis in db/db mice

To investigate the role of BTG1 in the pathogenesis of liver steatosis, we examined BTG1 mRNA and protein abundance in the livers of leptin receptor–mutated (db/db) mice, which develop liver steatosis (21). We found that BTG1 abundance at the mRNA and protein levels was decreased in the livers of db/db mice compared with wild-type mice (Fig. 1A).

Fig. 1 Overexpression of BTG1 reverses liver steatosis in db/db mice.

(A) Btg1 expression (left) and BTG1 abundance (right) were analyzed in the livers of male C57BL/6J wild-type (WT) and db/db mice. (B to E) Male C57BL/6J WT and db/db mice were injected with Ad-BTG1 or Ad-GFP and analyzed for Btg1 expression and BTG1 abundance (B), hepatic steatosis through staining with Oil Red O or H&E (C) (scale bars, 100 μm), liver (D) or serum (E) triglycerides (TG), total cholesterol (TC), or free fatty acids (FFAs). Means ± SEM for (A), (B), (D), and (E) are representative of at least two independent experiments, with five to seven mice per group in each experiment. Images in (C) are representative of six mice in each group. Legend in (B) applies to the graphs in (D) and (E). Statistical significance was calculated using two-tailed Student’s t test for the effects of db/db mice compared to WT mice (*P < 0.05) in (A), or one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls test for the effects of any group compared to the Ad-GFP WT mice (*P < 0.05), compared to Ad-GFP in db/db mice (#P < 0.05), and Ad-BTG1 db/db mice compared to Ad-BTG1 WT mice (&P < 0.05) in (B), (D), and (E). A.U., arbitrary units.

To test the role of BTG1 in liver steatosis, we injected wild-type and db/db mice with adenovirus expressing BTG1 (Ad-BTG1), resulting in increased BTG1 abundance in the livers of these mice compared with control mice, which were injected with adenovirus expressing green fluorescent protein (Ad-GFP) (Fig. 1B and fig. S1A). Body weight and food intake were not altered by BTG1 overexpression in wild-type mice (table S1). BTG1 overexpression decreased liver steatosis [as demonstrated by hematoxylin and eosin (H&E) and Oil Red O staining], hepatic triglycerides, and cholesterol concentrations in db/db mice, but not in wild-type mice (Fig. 1, C and D). Concentrations of serum triglycerides and cholesterol were increased in BTG1-overexpressing wild-type and db/db mice injected with Ad-BTG1 (Fig. 1E), although concentrations of hepatic and serum free fatty acids were not changed (Fig. 1, D and E). Consistently, liver weight was decreased in wild-type mice by BTG1 overexpression, with no change in muscle and adipose tissue weight (table S1). In HepG2 cells, overexpression of BTG1 significantly decreased triglyceride content compared with control cells (fig. S1, B and C).

Knockdown of BTG1 induces liver steatosis in wild-type mice

To further explore the impact of BTG1 on liver steatosis in vivo, we injected wild-type mice with an adenovirus encoding a short hairpin RNA (shRNA) directed against BTG1 (Ad-shBTG1), which decreased hepatic BTG1 abundance at the mRNA and protein levels (Fig. 2A). Body weight and food intake in wild-type mice were not altered by BTG1 knockdown (table S2). These mice had moderate liver steatosis, as indicated by H&E and Oil Red O staining (Fig. 2B), increased concentrations of hepatic triglycerides, and decreased concentrations of serum triglycerides compared with control mice injected with an adenovirus encoding a scrambled version of the BTG1-targeted shRNA (Ad-scrambled) (Fig. 2, C and D). Concentrations of serum and hepatic cholesterol and free fatty acids were not affected by BTG1 knockdown (Fig. 2, C and D). BTG1 knockdown consistently increased liver weight without affecting muscle or adipose tissue weight in wild-type mice (table S2). As observed in vivo, knockdown of BTG1 by Ad-shBTG1 or small interfering RNA (siRNA) increased triglyceride content in HepG2 cells compared with control cells (fig. S2).

Fig. 2 Knockdown of BTG1 induces liver steatosis in WT mice.

(A to D) Male C57BL/6J WT mice injected with Ad-shBTG1 or Ad-scrambled (Ad-scr) were analyzed for Btg1 expression (left) and BTG1 abundance (right) (A), hepatic steatosis through staining with Oil Red O or H&E (B) (scale bars, 100 μm), liver (C) or serum (D) triglycerides, total cholesterol, or free fatty acids. Means ± SEM for (A), (C), and (D) are representative of at least two independent experiments, with five to seven mice per group in each experiment. Images in (B) are representative of six mice in each group. Legend in (A) applies to the graphs in (C) and (D). Statistical significance was calculated using two-tailed Student’s t test for the effects of Ad-shBTG1 compared to the control group (*P < 0.05).

BTG1 ameliorates liver steatosis by decreasing SCD1 abundance

A lipid metabolism disorder is one of the major causes for liver steatosis (5). We therefore examined the effects of BTG1 on genes and proteins related to lipid metabolism in the livers of mice overexpressing BTG1 or with BTG1 knockdown. Genes encoding enzymes involved in lipogenesis include Fas, Scd1, acetyl-CoA carboxylase (Acc), Srebp1c, malic enzyme (Me), glycerol-3-phosphate acyltransferase (Gpat), carbohydrate-responsive element–binding protein (Chrebp), Pparγ, Srebp2, and 3-hydroxy-3-methylglutaryl-CoA synthase (Hmgcs2) (5). Except for Fas and Scd1, other genes encoding lipogenic enzymes were not affected by overexpression or knockdown of BTG1 (Fig. 3, A and B). Genes related to fatty acid uptake [cluster of differentiation 36 (Cd36) and fatty acidbinding protein (Fabp)], fatty acid oxidation (Pparα and Cpt1α), and triglyceride secretion [apolipoprotein B (ApoB) and ApoE] (5) were also examined. These genes showed inconsistent changes in the livers of mice overexpressing BTG1 or with BTG1 knockdown (Fig. 3, A and B). The expression of genes related to lipid metabolism, however, was not altered in muscle or adipose tissue of either group of mice (fig. S3).

Fig. 3 BTG1 prevents liver steatosis by suppressing SCD1.

(A to D) Male C57BL/6J WT mice were injected with Ad-BTG1 or Ad-GFP (A and C), or with Ad-shBTG1 or Ad-scrambled (B and D). Mice were analyzed for mRNA expression of genes encoding lipogenic enzymes (A and B) and SCD1 abundance (C and D). (E and F) Primary hepatocytes were infected with Ad-BTG1 or Ad-GFP (E), or with Ad-shBTG1 or Ad-scrambled (F). The abundance of BTG1 and SCD1 proteins was determined. (G to J) Male C57BL/6J WT mice were injected with Ad-shBTG1, Ad-shSCD1, and/or Ad-scrambled as indicated. Mice were analyzed for Btg1 and Scd1 expression (left) and protein abundance (right) (G), hepatic steatosis through staining with Oil Red O or H&E (H) (scale bars, 100 μm), liver (I) or serum (J) triglycerides, total cholesterol, or free fatty acids. Means ± SEM are representative of at least three independent experiments (E and F) or at least two independent experiments (A to D, G, I, and J), with five to seven mice per group in each experiment. Images in (H) are representative of six mice in each group. Statistical significance was calculated using two-tailed Student’s t test for the effects of Ad-BTG1 or Ad-shBTG1 compared to the control group in (A) to (F), or one-way ANOVA followed by the Student-Newman-Keuls test for the effects of any group compared to the control group (*P < 0.05) in (G), (I), and (J), and to compare mice injected with Ad-SCD1 and Ad-shBTG1 to mice injected with Ad-shBTG1 (#P < 0.05) in (I) and (J).

Furthermore, we found that liver lipogenic index (22) was significantly increased in mice with BTG1 knockdown (fig. S4A), although serum concentrations of 3-β-hydroxybutyrate (3-HB), a measure of fatty acid oxidation (23), were not affected by overexpression or knockdown of BTG1 (fig. S4B). Fatty acid uptake in HepG2 cells was not affected by BTG1 overexpression (fig. S4C). A triglyceride secretion assay (24) showed that very low density lipoproteins (VLDLs) were largely enhanced in BTG1-overexpressing mice (fig. S4D).

We next investigated the possible involvement of SCD1 in mediating BTG1’s effect on liver steatosis. We found that hepatic SCD1 protein abundance was significantly decreased by BTG1 overexpression or increased by BTG1 knockdown (Fig. 3, C and D). Overexpression of BTG1 in db/db mice also decreased SCD1 abundance (fig. S4, E and F). Liver and serum fatty acid composition measurement revealed that the concentration of C16:0 remained the same in liver, but increased in serum, whereas the ratios of C18:1n-9/C18:0 and C16:1n-7/C16:0, which are an index of SCD1 activity (22, 25), were higher in mice with BTG1 knockdown compared with control mice (fig. S4, G and H, and table S3). Additionally, SCD1 abundance was significantly decreased by BTG1 overexpression or increased by BTG1 knockdown by Ad-shBTG1 or BTG1 siRNA in primary cultured mouse hepatocytes compared with control cells (Fig. 3, E and F, and fig. S5A).

We then assessed a role for SCD1 in BTG1-regulated liver steatosis in wild-type mice in which BTG1 and/or SCD1 abundance had been decreased. Injection of Ad-shSCD1 reduced SCD1 abundance in liver (Fig. 3G and fig. S5B). As expected, simultaneous knockdown of BTG1 and SCD1 triggered liver steatosis, as demonstrated by H&E and Oil Red O staining and measurement of hepatic triglycerides (Fig. 3, H and I). Knockdown of SCD1 alone also decreased hepatic triglycerides (Fig. 3I). Knockdown of SCD1 did not affect the decrease in serum triglycerides caused by BTG1 knockdown (Fig. 3J). Hepatic and serum cholesterol and free fatty acid contents were similar among these different groups of mice (Fig. 3, I and J).

BTG1 decreases SCD1 abundance and ameliorates liver steatosis by suppressing the transcriptional activity of ATF4

BTG1 regulates downstream target genes by affecting the activity of the upstream transcription factors (1316). PPARγ and SREBP1C (5, 26), transcription factors that target Scd1 and Fas, are unlikely to mediate the action of BTG1 on lipid metabolism because PPARγ activity is not controlled by BTG1 (13). In addition, we did not detect an association between BTG1 and SREBP1C coexpressed in human embryonic kidney (HEK) 293T cells and BTG1 did not affect the abundance of SREBP1C (fig. S6).

Hepatic SCD1 abundance is decreased in ATF4-null mice and increased by adenovirus-mediated overexpression of ATF4 (27, 28), suggesting that BTG1 may decrease SCD1 abundance by suppressing ATF4 abundance and/or activity. Consistent with this possibility, we found that BTG1 interacted with ATF4 when coexpressed in HEK293T cells (Fig. 4, A and B, and fig. S7A). In addition, BTG1 most likely influenced the transcriptional activity of ATF4, because BTG1 did not directly affect hepatic ATF4 mRNA or protein abundance, but changed the mRNA and protein abundance of two ATF4 targets: C/EBP (CCAAT/enhancer binding protein) homologous protein (CHOP) and tribbles homolog 3 (TRB3) (Fig. 4, C and D) (29). Furthermore, BTG1 overexpression partially or fully attenuated the increase in SCD1 mRNA and protein induced by ATF4 overexpression in primary cultured mouse hepatocytes (Fig. 4E and fig. S7B). ATF4-stimulated Scd1 promoter activity was also prevented by coexpression of BTG1 in HEK293T cells (Fig. 4F). Moreover, chromatin immunoprecipitation (ChIP) assay in primary hepatocytes overexpressing BTG1 and/or ATF4 showed that BTG1 reduced the binding of ATF4 to the Scd1 promoter (fig. S7C).

Fig. 4 BTG1 regulates liver steatosis by suppressing the transcriptional activity of ATF4.

(A and B) HEK293T cells were cotransfected with (+MYC-ATF4) or without (−MYC-ATF4) MYC-tagged ATF4 plasmid and with (+FLAG-BTG1) or without (−FLAG-BTG1) FLAG-tagged BTG1 plasmid. Immunoprecipitation (IP) and immunoblotting were performed using the antibodies indicated. (C and D) Male C57BL/6J WT mice were injected with Ad-BTG1 or Ad-GFP (C), or with Ad-shBTG1 or Ad-scrambled (D). The abundance of the indicated mRNAs (left) and the corresponding proteins were determined (right). (E) Primary hepatocytes were infected with Ad-GFP, Ad-BTG1, and/or Ad-ATF4 as indicated. Scd1 expression and SCD1 protein abundance were assessed. (F) Luciferase activity was assessed in HEK293T cells expressing the Scd1 promoter vector, BTG1 or ATF4. (G to J) Male db/db mice were injected with Ad-GFP, Ad-BTG1, and/or Ad-ATF4 as indicated. Mice were analyzed for Scd1 expression and SCD1 abundance (G), hepatic steatosis through staining with Oil Red O or H&E (H) (scale bars, 100 μm), and liver (I) and serum (J) triglyceride, total cholesterol, and free fatty acid concentrations. Means ± SEM are representative of at least three independent experiments (A, B, E, and F) or at least two independent experiments (C, D, G, I, and J), with five to seven mice per group in each experiment. Images in (H) are representative of six mice in each group. Statistical significance was calculated using two-tailed Student’s t test for the effects of Ad-BTG1 or Ad-shBTG1 compared to the control group in (C) and (D), or one-way ANOVA followed by the Student-Newman-Keuls test for the effects of any group compared to the control group (*P < 0.05) in (E) to (G), (I), and (J), to compare Ad-BTG1– and Ad-ATF4–injected mice to Ad-ATF4–injected mice (#P < 0.05) in (E) and (F), and to compare Ad-BTG1– and Ad-ATF4–injected mice to Ad-BTG1–injected mice (#P < 0.05) in (G), (I), and (J).

In primary cultured mouse hepatocytes, overexpression of ATF4 blocked the decrease in SCD1 mRNA and protein abundance caused by overexpression of BTG1 (fig. S7, D and E). ATF4 overexpression enhanced hepatic lipid accumulation in db/db mice (fig. S8). In db/db mice, the suppressive effects of BTG1 on SCD1 abundance and liver steatosis in db/db mice were also reversed by ATF4 (Fig. 4, G and H, and fig. S9). Moreover, ATF4 also prevented BTG1 from decreasing hepatic triglycerides and cholesterol and increasing serum triglycerides and cholesterol (Fig. 4, I and J). Hepatic and serum free fatty acids did not change among the three groups of mice (Fig. 4, I and J).

ATF4 is involved in the regulation of endoplasmic reticulum (ER) stress (30) and inflammation (31), and BTG1 inhibits proliferation (9), suggesting that BTG1 may regulate liver steatosis through signals in these processes. The protein abundance and gene expression of several ER stress markers (30), however, were not affected by overexpression or knockdown of BTG1 (fig. S10, A to D). The gene expression of some liver regeneration markers was altered, whereas some were not affected by overexpression or knockdown of BTG1 (fig. S10, E and F). NF-κB activity, as assessed by a luciferase reporter, was decreased by overexpression of BTG1 in HEK293T cells (fig. S10G).

Transgenic overexpression of BTG1 protects mice from liver steatosis induced by an HCD

To further determine the effects of BTG1 on liver steatosis, we generated BTG1 transgenic mice, which exhibited significantly increased hepatic BTG1 mRNA and protein abundance compared with control mice (Fig. 5A). In addition, the mRNA and protein abundance of SCD1 and TRB3, but not those of ATF4, were also decreased in the livers of BTG1 transgenic mice (Fig. 5A). Hepatic Fas and Chop mRNA was decreased in BTG1 transgenic mice (fig. S11A). Because BTG1 mRNA and protein abundance decreased in the livers of mice fed an HCD for 4 weeks (Fig. 5B), we determined whether BTG1 transgenic mice were resistant to HCD-induced liver steatosis (32). HCD caused lipid accumulation in the livers of wild-type mice, but not BTG1 transgenic mice, compared with mice under a control diet (Fig. 5C). Consistently, hepatic triglyceride and free fatty acid concentrations were lower in BTG1 transgenic mice than those in wild-type mice under HCD, although cholesterol concentrations were similar (Fig. 5D). Serum cholesterol, triglycerides, and free fatty acids were similar between wild-type and BTG1 transgenic mice under HCD (Fig. 5E). Again, hepatic Fas and Scd1 mRNA was lower in transgenic mice compared with wild-type mice maintained on an HCD (fig. S11B).

Fig. 5 Transgenic overexpression of BTG1 protects mice from liver steatosis induced by an HCD.

(A) The abundance of the indicated mRNAs and the corresponding proteins were analyzed in the livers of C57BL/6J WT and BTG1 transgenic (Tg) mice. (B to E) Male C57BL/6J WT and BTG1 transgenic mice were fed a normal diet (ND) or HCD. Mice were analyzed for Btg1 expression and BTG1 abundance (B), hepatic steatosis through staining with Oil Red O or H&E (C) (scale bars, 100 μm), and liver (D) and serum (E) triglycerides, total cholesterol, and free fatty acid concentrations. Means ± SEM are representative of at least two independent experiments for (A), (B), (D), and (E), with five to seven mice per group in each experiment. Images in (C) are representative of six mice in each group. Statistical significance was calculated using two-tailed Student’s t test for the effects of BTG1 transgenic compared to control mice in (A), the mice fed with HCD compared to normal diet in (B) (*P < 0.05), or one-way ANOVA followed by the Student-Newman-Keuls test for the effects of any group compared to control mice fed with normal diet (*P < 0.05), to compare HCD to normal diet in BTG transgenic mice (#P < 0.05), and to compare BTG transgenic to control mice under HCD treatment (&P < 0.05) in (D) and (E).

HCD regulates BTG1 abundance through a pathway involving mTOR, S6K1, and CREB

The mTOR-S6K1 signaling pathway couples cellular nutrient sensing to organism metabolic homeostasis (33) and is activated by HCD (34), suggesting a possible role of this pathway in regulating BTG1 abundance by HCD. Consistent with previous reports (34), phosphorylation of mTOR, S6K1, and the downstream target S6 was increased in the livers of mice on an HCD compared with mice fed a control diet (Fig. 6A). However, the mTOR-S6K1 pathway was not affected by BTG1 overexpression or knockdown (fig. S12, A and B). In contrast, BTG1 abundance in HepG2 cells was decreased or increased upon expression of a constitutively active S6K1 mutant or knockdown of S6K1, respectively (fig. S12, C and D). Accordingly, knockdown of S6K1 significantly reversed the decrease in hepatic BTG1 mRNA and protein abundance caused by HCD feeding (Fig. 6B).

Fig. 6 HCD feeding regulates BTG1 abundance through a pathway involving mTOR, S6K1, and CREB.

(A) The phosphorylation (p) and total abundance of the indicated proteins were analyzed in the livers of C57BL/6J WT mice fed a normal diet or HCD. (B) C57BL/6J WT mice were fed an HCD before being injected with Ad-shS6K1 or Ad-scrambled. Mice were analyzed for S6k1 and Btg1 expression and the abundance of the indicated proteins. (C) Luciferase activity was assessed in HEK293T cells expressing a Btg1 promoter vector with or without CREB overexpression. (D) Primary hepatocytes overexpressing CREB or with CREB knockdown were analyzed for Btg1 expression. (E) HEK293T cells were cotransfected with HA-tagged constitutively active S6K1 (CA-S6K1) and/or MYC-tagged CREB plasmids. Immunoprecipitation and immunoblotting were performed using the antibodies indicated. (F) BTG1 abundance was determined in primary hepatocytes transfected with CREB siRNA and/or constitutively active S6K1. Means ± SEM are representative of at least three independent experiments for (C) to (F) or at least two independent experiments for (A) and (B), with five to seven mice per group in each experiment. Statistical significance was calculated using two-tailed Student’s t test for the effects of the mice fed with HCD compared to normal diet in (A), Ad-shS6K1 compared to the control group in (B), CREB overexpression or CREB siRNA compared to the control group in (C) and (D) (*P < 0.05), or one-way ANOVA followed by the Student-Newman-Keuls test for the effects of any group compared to control (*P < 0.05) and to compare to hepatocytes coexpressing constitutively active S6K1 and CREB siRNA to hepatocytes expressing constitutively active S6K1 only (#P < 0.05) in (F). (G) Working model. BTG1, a co-regulator for ATF4, regulates hepatic lipid metabolism by decreasing SCD1 abundance.

Because S6K1 can directly phosphorylate CREB (35), and we found three potential CRE sites in the Btg1 promoter (36) using Genomatix, we speculated that S6K1 may regulate Btg1 expression through CREB. Consistent with this possibility, phosphorylation of hepatic CREB was increased by HCD, which was reversed by Ad-shS6K1 (Fig. 6, A and B). Overexpression of CREB decreased Btg1 promoter activity and Btg1 mRNA abundance, whereas two different siRNAs directed against CREB increased Btg1 mRNA abundance (Fig. 6, C and D, and fig. S13A). Consistent with a previous study (35), exogenously expressed S6K1 and CREB interacted (Fig. 6E). Finally, CREB knockdown significantly reversed the decrease in BTG1 abundance induced by the constitutively active S6K1 mutant (Fig. 6F and fig. S13, B and C).

DISCUSSION

The BTG family is unique to metazoans, and BTG1 is widely distributed throughout the body, such as brain, spleen, lung, liver, and kidney (11). Most progress in the search of BTG1 function comes from the analysis of various cell types (12). Results obtained from BTG1-null mice indicate that BTG1 is required for generating adult new neurons (37). Here, we showed that BTG1 mRNA and protein abundance was significantly decreased in the livers of db/db mice and overexpression of BTG1 reverses liver steatosis in these mice. Conversely, knockdown of BTG1 caused liver steatosis in wild-type mice. In addition, BTG1 transgenic mice were resistant to HCD-induced liver steatosis. Not all of the phenotypes, however, were the same between BTG1 transgenic mice and wild-type mice that overexpressed BTG1, which could be due to the different BTG1 distribution pattern. The transgenic mice used in our current study overexpressed BTG1 throughout the whole body, whereas the adenovirus-mediated overexpression of BTG1 occurs preferentially in livers. Nonetheless, our results demonstrate a function of BTG1 in regulating liver steatosis in mice. We speculate that BTG1 has a direct effect on liver steatosis based on results obtained from our in vitro study, showing that triglyceride content in HepG2 cells was decreased or increased by overexpression or knockdown of BTG1, respectively.

Lipid accumulation in liver is likely to reflect an imbalance in hepatic triglyceride synthesis, β-oxidation, uptake, and/or secretion of fatty acids (5). Our results suggest that β-oxidation and fatty acid uptake are unlikely to contribute to the effects of BTG1 on liver steatosis. Consistently, mRNA and protein abundance for SCD1, the rate-limiting enzyme in the biosynthesis of monounsaturated fatty acids (38, 39), was greatly reduced or increased by overexpression or knockdown of BTG1 in vivo and in vitro, respectively. SCD1 knockdown reversed the effects of BTG1 knockdown on liver steatosis, thus confirming the importance of SCD1 in BTG1’s effects. In addition to the decreased lipogenesis, however, the increased triglyceride export may also contribute to the amelioration of liver steatosis by BTG1 overexpression in mice. The increased serum triglyceride and cholesterol concentrations were possibly due to the increased triglyceride export in BTG1-overexpressing mice.

ATF4 is a transcription factor that belongs to the family of basic zipper–containing proteins (40) and involved in various physiological processes, including formation of eye lens fibers (41) and glucose metabolism (42, 43). We have previously shown that SCD1 abundance is significantly decreased in white adipose tissue and liver of ATF4 knockout mice compared with control mice (28), suggesting that ATF4 might be involved in the ability of BTG1 to prevent an increase in hepatic SCD1 abundance and liver steatosis. We found that BTG1 interacted with ATF4 and attenuated the increase in SCD1 abundance and promoter activity induced by ATF4 in vitro. Conversely, overexpression of ATF4 reversed the ability of BTG1 to prevent an increase in hepatic SCD1 abundance and lipid accumulation in db/db mice. Consistent with our results, several studies have demonstrated that BTG1 regulates gene expression by acting as a cofactor for various transcription factors (1316). These results demonstrate a relationship between BTG1 and ATF4, which may also provide insights into the other functions of BTG1 and ATF4. We have previously shown that global deletion of ATF4 significantly decreases Fas expression in white adipose tissue compared with control mice (28), and that overexpression of ATF4 increases Fas mRNA in primary hepatocytes, suggesting that ATF4 might also be involved in the regulation of Fas expression by BTG1. Extensive experiments will be required to test this possibility in the future.

Btg1 expression is increased in rat liver during protein-calorie malnutrition (44), although nutritional regulation of BTG1 mRNA and protein abundance is poorly understood. mTOR and S6K1 are serine-threonine protein kinases that are essential for protein synthesis, growth, development, and proliferation (4547), and the activity of these kinases is sensitive to nutritional status (34, 48). As shown previously (34), we also found that mTOR/S6K1 activity was increased in liver of mice fed an HCD, and S6K1 was required for HCD to decrease BTG1 abundance in vivo. Furthermore, we provided evidence showing that S6K1 regulates BTG1 abundance through CREB, a transcription factor (49), that may directly bind to CRE sites at Btg1 promoter. Detailed mechanisms underlying CREB control of Btg1 expression remain to be further explored.

Phosphorylation of S6K1 and S6 is increased in livers of db/db mice compared with control mice (50, 51). Furthermore, we have previously shown that BTG1 mRNA and protein abundance is induced by decreased S6K1 activity under leucine deprivation (52). Here, we provided in vitro evidence showing that S6K1 regulates BTG1 abundance through CREB. Thus, we speculate that BTG1 abundance in the livers of db/db mice was decreased in an S6K1- and CREB-dependent pathway.

Autophagy, ER stress, and apoptosis, three important factors closely linked to liver steatosis (5355), might be involved in BTG1/ATF4 regulation of liver steatosis. ATF4 is a critical regulator for autophagy (56), ER stress (30), inflammation (31), and apoptosis (57). These results raise the possibility that BTG1 regulates hepatic lipid metabolism through signal pathways related to these processes. Our results showed that the abundance of some of the proteins or the expression of genes related to ER stress was altered by manipulation of BTG1, and future studies will be necessary to examine the role of autophagy and apoptosis in BTG1-mediated regulation of liver steatosis.

Here, we also found that serum cholesterol concentrations were significantly increased by BTG1 overexpression but not significantly decreased by BTG1 knockdown. We speculated that these findings could be due to the low cholesterol amount in the serum of wild-type mice. Because there is normally not much cholesterol in the serum of wild-type mice, the reduction of hepatic cholesterol by BTG1 knockdown may not be that obvious, making it difficult to see any changes. However, when secretion was increased by overexpression of BTG1, an increase in serum cholesterol amount might be easier to detect because the basal cholesterol content was low in the serum of wild-type mice. Similar reasons might explain the discrepancy between the increase in liver triglycerides upon BTG1 knockdown and the lack of significant change in liver triglycerides upon BTG1 overexpression in wild-type mice. Similarly, BTG1 overexpression reduced liver triglycerides in db/db mice but failed to reduce triglycerides in wild-type mice, even though BTG1 overexpression reduced Fas and Scd1 gene expression.

As described in our working model (Fig. 6G), we demonstrated that overexpression of BTG1 ameliorates liver steatosis in db/db mice and knockdown of BTG1 caused hepatic lipid accumulation in wild-type mice. Furthermore, the effects of BTG1 on relieving liver steatosis in db/db mice are mediated by preventing ATF4 from increasing SCD1 abundance. Moreover, BTG1 also plays an important role in HCD-induced liver steatosis. Together, our study identifies a role of BTG1 in hepatic lipid metabolism by inhibition of ATF4 and SCD1, which will provide important insights in targeting BTG1 for treating liver steatosis and NAFLD.

MATERIALS AND METHODS

Animals and treatments

Eight- to 10-week-old male C57BL/6J wild-type mice and db/db mice were obtained from Model Animal Research Center of Nanjing University (Nanjing, China). Here, we used BTG1 transgenic mice, which overexpress BTG1 in the whole body, under a C57BL/6J background that was produced by crossing our originally generated FVB/N background BTG1 transgenic mice (Cyagen Biosciences Inc.) with C57BL/6J wild-type mice for at least four generations. Tail biopsies were analyzed by genomic polymerase chain reaction (PCR) (forward primer, 5′-TGGCTAACTAGAGAACCCACT-3′; reverse primer, 5′-CTGTCTACCATTTGCACGTT-3′). All the mice were maintained on a 12-hour light/dark cycle at 25°C and provided free access to commercial rodent chow and tap water before initiation of the experiments. Male C57BL/6J wild-type and BTG1 transgenic mice were continuously fed ad libitum for 4 weeks on either an HCD (Research Diets Inc.) or control diet, as described previously (27). These experiments were conducted in accordance with guidelines of the Institutional Animal Care and Use Committee of the Institute for Nutritional Science, Shanghai Institutes for Biological Sciences (SIBS), Chinese Academy of Sciences (CAS).

Plasmids and cell treatments

The DNA fragments encoding BTG1, ATF4, or CREB were amplified from mouse liver genomic complementary DNA (cDNA) and inserted into the expressing vector pcDNA3.1/myc–His A or p3XFLAG-CMV-10. Hemagglutinin (HA)–tagged rat constitutively active S6K1 expression plasmid was the Addgene plasmid 8991 (58). Plasmids were transfected into cells with Effectene Transfection Reagent (Qiagen). The double-stranded siRNA targeting CREB and BTG1 was from GenePharma. The sequence is #1 5′-GCCCAGCAACCAAGTTGT-3′ and #2 5′-GCAGCTCATGCAACATCA-3′ for CREB and 5′-GCTGTAAGGAGGAACTTC-3′ for BTG1. Cells were transfected with siRNA using X-tremeGENE siRNA Transfection Reagent (Roche Diagnostics). HepG2 cells were incubated with 200 μM sodium oleate (Sigma-Aldrich) for 24 hours to induce lipid accumulation (59). Primary hepatocytes were prepared by collagenase perfusion as described previously (59). Cells were maintained in Dulbecco’s modified Eagle’s medium with 25 mM glucose (Gibco), 10% fetal bovine serum, and penicillin and streptomycin (50 mg/ml) at 37°C and 5% CO2–95% air.

Generation and administration of recombinant adenoviruses

Recombinant adenoviruses expressing mouse ATF4 (Ad-ATF4) were generated as previously described (27). Recombinant adenoviruses expressing MYC-tagged mouse BTG1 (Ad-BTG1) were generated using the AdMax Adenoviral Vector Creation System (Microbix Biosystems Inc.) according to the manufacturer’s instruction. Adenoviruses expressing scrambled (Ad-scrambled) or shRNA against mouse BTG1 (Ad-shBTG1), S6K1 (Ad-shS6K1), and SCD1 (Ad-shSCD1) were generated using the BLOCK-iT Adenoviral RNAi Expression System (Invitrogen) according to the manufacturer’s instructions. The scrambled sequence is 5′-TTCTCCGAACGTGTCACGT-3′. The shRNA sequence for BTG1 is 5′-GGATCAGGTTACCGTTGTATT-3′, for S6K1 is 5′-GGGAGTTGGACCATATGAACT-3′, and for SCD1 is 5′-GAGATCTCCAGTTCTTACA-3′, which targets mouse and/or human BTG1, S6K1, or SCD1, respectively. High-titer stocks of amplified recombinant adenoviruses were purified as previously described (58). Adenoviruses were diluted in phosphate-buffered saline and administered at a dose of 107 plaque-forming units (PFU) per well in 12-well plates or through tail vein injection using 109 PFU per mice.

Measurement of triglycerides, total cholesterol, and free fatty acids

Hepatic and cellular lipids were extracted with chloroform/methanol (2:1) as previously described (59). Triglycerides, total cholesterol, and free fatty acids were determined using a triglyceride kit (BHKT Clinical Reagents), a cholesterol kit (SSUF-C), and a free fatty acid kit (Wako Pure Chemical Industries), respectively. All of these assays were performed according to the manufacturer’s instructions.

Functional assays for lipid metabolism

The lipogenic index was determined from the C16:0/C18:2n-6 ratio in the liver (22). Serum 3-HB was determined using a 3-HB assay kit (Nanjing Jiancheng Bioengineering Institute). Fatty acid uptake was determined using QBT Fatty Acid Uptake Assay Kit (Molecular Devices). For the VLDL secretion assay, wild-type mice were injected with Ad-BTG1 or Ad-GFP for 14 days and then fasted for 4 hours, followed by intraperitoneal injection of poloxamer 407 at a dose of 1 mg/g. Triglycerides were measured in serum of tail vein blood taken at different time points (24).

Histological analysis of tissues

Frozen sections of liver were stained with Oil Red O. Paraformaldehyde-fixed and paraffin-embedded sections of liver were stained with H&E for histology.

RNA isolation and relative quantitative real-time PCR

RNA isolation and relative quantitative real-time PCR (RT-PCR) were performed as described previously (27). The sequences of primers used for RT-PCR are listed in table S4.

Luciferase assay

Scd1 promoter (−849 to +100) was generated in pGL3-Basic vector. Btg1 promoter was generated in pGL2-Basic vector as previously described (36). HEK293T cells were cotransfected with the internal control vector pRL Renilla (Promega) and plasmids indicated using Lipofectamine 2000. The firefly and Renilla luciferase activities were assayed using Dual-Glo Luciferase Assay System (Promega).

Antibodies, immunoblotting, and coimmunoprecipitation

The antibodies against FLAG, MYC, HA, mTOR phosphorylated at Thr2448, mTOR, S6K1 phosphorylated at Ser389, S6K1, S6 phosphorylated at Ser235/236, S6, CREB phosphorylated at Ser133, CREB, IRE1α (inositol-requiring enzyme 1α), phosphorylated eIF2α (eukaryotic initiation factor 2α), eIF2α, phosphorylated PERK, and PERK were from Cell Signaling Technology; antibodies against BTG1, SCD1, ATF4, TRB3, CHOP, and SREBP1C were from Santa Cruz Biotechnology Inc.; antibodies against ACTIN and phosphorylated IRE1α were from Sigma-Aldrich and Novus Biologicals, respectively. Immunoblotting and coimmunoprecipitation assays were previously described (60).

ChIP assay

ChIP assays were performed according to the manufacturer’s protocol (Millipore) with anti-ATF4 antibody (1:50; Santa Cruz Biotechnology Inc.) or normal rabbit immunoglobulin G (1:50; Santa Cruz Biotechnology Inc.) for negative control. Immunoprecipitated Scd1 promoter was quantified using PCR with primers designed to amplify the 150–base pair (bp) region encompassing the CRE site (forward, 5′-CGGGCTTCACAGGAGGCA-3′; reverse, 5′-AGAGAGAGGGCGGGAC-3′) or a 150-bp upstream region not involved in ATF4 response (forward, 5′-TTCTGGTTCCACTGGTGA-3′; reverse, 5′-GTACCTCTTTAAAGATTTAT-3′).

Statistics

All data are expressed as means ± SEM. Significant differences were assessed either by two-tailed Student’s t test or by one-way ANOVA, followed by the Student-Newman-Keuls test for most of the results.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/9/428/ra50/DC1

Fig. S1. Overexpression of BTG1 decreases triglyceride content in vivo and in vitro.

Fig. S2. Knockdown of BTG1 by Ad-shBTG1 or BTG1 siRNA increases triglyceride content in vitro.

Fig. S3. The effects of BTG1 on the expression of genes encoding lipid-metabolizing enzymes in muscle and white adipose tissue.

Fig. S4. The effects of BTG1 on lipid metabolism in the liver.

Fig. S5. BTG1 suppresses SCD1 in vitro and knockdown of BTG1 and SCD1 in vivo.

Fig. S6. Lack of an interaction between BTG1 and SREBP1C.

Fig. S7. BTG1 overexpression attenuates ATF4-stimulated increase in SCD1 mRNA and protein abundance in vitro.

Fig. S8. ATF4 overexpression exacerbates liver steatosis in db/db mice.

Fig. S9. BTG1 regulates SCD1 abundance through ATF4.

Fig. S10. The effects of BTG1 on the expression of genes and abundance of proteins related to ER stress and regeneration and on NF-κB activity.

Fig. S11. Gene expression in the livers of BTG1 transgenic mice on an HCD.

Fig. S12. The possible link between BTG1 and the mTOR-S6K1 pathway.

Fig. S13. S6K1 regulates BTG1 abundance through CREB.

Table S1. Metabolic parameters in wild-type mice injected or not with Ad-BTG1.

Table S2. Metabolic parameters in wild-type mice injected or not with Ad-shBTG1.

Table S3. Liver and serum fatty acid composition in wild-type mice injected or not with Ad-shBTG1.

Table S4. List of oligonucleotide primer pairs used in RT-PCR analysis.

REFERENCES AND NOTES

Acknowledgments: We thank J. Fang (Institute for Nutritional Science, Shanghai, China) for providing the FOXO3aTM and NF-κB–driven reporter plasmids and Y. Chang (Chinese Academy of Medical Sciences and Peking Union Medical College) for giving the human mature SREBP1C plasmid. Funding: This work was supported by grants from the National Natural Science Foundation (81130076, 81325005, 31271269, 81100615, 81390350, 81300659, and 81570777), Basic Research Project of Shanghai Science and Technology Commission (13JC1409000), International Science and Technology Cooperation Program of China (Singapore 2014DFG32470), and research supported by the CAS/State Administration of Foreign Experts Affairs international partnership program for creative research teams. F.G. was also supported by the One Hundred Talents Program of the CAS. F.X. was supported by China Postdoctoral Science Foundation–funded project (2012M520950 and 2013T60473), CAS-funded project (2013KIP310), Youth Innovation Promotion Association of the CAS, and Sanofi-Aventis–SIBS scholarship program. Author contributions: F.X. and J.D. researched data and wrote, reviewed, and edited the article. Y.G., Y.N., F.Y., and J.Y. researched data and contributed to discussion. S.C. researched data and provided research material. F.G. directed the project, contributed to discussion, and wrote, reviewed, and edited the article. F.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article