Hepatic carboxylesterase 1 is essential for both normal and farnesoid X receptor-controlled lipid homeostasis

Authors

  • Jiesi Xu,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yuanyuan Li,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Wei-Dong Chen,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    2. Key Laboratory of Receptor-Mediated Gene Regulation and Drug Discovery, School of Medicine, Henan University, Kaifeng, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yang Xu,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Liya Yin,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
  • Xuemei Ge,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
  • Kavita Jadhav,

    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    Search for more papers by this author
  • Luciano Adorini,

    1. Intercept Pharmaceuticals, New York, NY, USA
    Search for more papers by this author
  • Yanqiao Zhang

    Corresponding author
    1. Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH, USA
    • Address reprint requests to: Dr. Yanqiao Zhang, Department of Integrative Medical Sciences, Northeast Ohio Medical University, Rootstown, OH 44272. E-mail: yzhang@neomed.edu; fax: 330-325-5912.

    Search for more papers by this author

  • See Editorial on Page 1665

    Potential conflict of interest: Dr. Adorini is employed by and owns stock in Intercept.

Abstract

Nonalcoholic fatty liver disease (NAFLD) is one of the major health concerns worldwide. Farnesoid X receptor (FXR) is considered a therapeutic target for treatment of NAFLD. However, the mechanism by which activation of FXR lowers hepatic triglyceride (TG) levels remains unknown. Here we investigated the role of hepatic carboxylesterase 1 (CES1) in regulating both normal and FXR-controlled lipid homeostasis. Overexpression of hepatic CES1 lowered hepatic TG and plasma glucose levels in both wild-type and diabetic mice. In contrast, knockdown of hepatic CES1 increased hepatic TG and plasma cholesterol levels. These effects likely resulted from the TG hydrolase activity of CES1, with subsequent changes in fatty acid oxidation and/or de novo lipogenesis. Activation of FXR induced hepatic CES1, and reduced the levels of hepatic and plasma TG as well as plasma cholesterol in a CES1-dependent manner. Conclusion: Hepatic CES1 plays a critical role in regulating both lipid and carbohydrate metabolism and FXR-controlled lipid homeostasis. (Hepatology 2014;59:1761–1771)

Abbreviations
ATGL

adipose triglyceride lipase

CEH

cholesteryl ester hydrolase

CES1

carboxylesterase 1

FA

fatty acid

FAO

fatty acid oxidation

FXR

Farnesoid X receptor

NAFLD

nonalcoholic fatty liver disease

NASH

nonalcoholic steatohepatitis

SHP

small heterodimer partner

TG

triglyceride

TGH

TG hydrolase

Nonalcoholic fatty liver disease (NAFLD), one of the most common liver diseases worldwide, encompasses a spectrum of liver disorders ranging from hepatic steatosis to nonalcoholic steatohepatitis (NASH), fibrosis, and cirrhosis. NAFLD is often associated with obesity, dyslipidemia, insulin resistance, and type 2 diabetes, and is recognized as the hepatic manifestation of the metabolic syndrome.[1] By far, the molecular mechanisms underlying the development of NAFLD remain to be fully understood.

Hepatic triglyceride (TG) accumulation (steatosis), the hallmark of NAFLD, may result from dysfunctional regulation of hepatic TG synthesis, hydrolysis, and secretion. While TG synthesis and secretion in the liver have been well understood, hepatic TG hydrolysis is largely unknown. Very recently, adipose triglyceride lipase (ATGL), an adipose-enriched TG lipase, has been shown to play an important role in regulating hepatic TG turnover.[2, 3] However, the expression of ATGL in the liver is low,[4] thus raising the possibility that additional TG hydrolase(s) (TGH) may exist in the liver.

Carboxylesterase 1 (CES1) is highly expressed in the liver.[5] Human CES1 is also called cholesteryl ester hydrolase (CEH), as it displays CEH activity.[6] Consistent with its CEH activity, overexpression of human CES1 in macrophages attenuates the development of atherosclerosis in Ldlr−/− mice.[7] Despite high expression in the liver, the role of hepatic CES1 in metabolic control is largely unknown. In an in vitro study, murine Ces1 was shown to prevent TG accumulation in rat McArdle-RH7777 cells.[8] However, the physiological relevance of this finding remains unclear.

Nuclear receptors are ligand-activated transcription factors with important regulatory roles in both development and adult physiology. The farnesoid X receptor (FXR; NR1H4), a member of the nuclear receptor superfamily, is important for maintaining bile acid, lipid, and glucose homeostasis. Activation of FXR displays beneficial effects on lipid and carbohydrate metabolism.[9] First, it lowers plasma TG and cholesterol levels[10-12] and improves insulin sensitivity.[12-15] Second, it increases reverse cholesterol transport[16] and protects against atherosclerosis.[17, 18] Lastly, it lowers hepatic TG levels and protects against NAFLD.[12, 19-22] The beneficial effects of FXR activation on hepatic TG homeostasis indicate that FXR is a therapeutic target for treatment of NAFLD.[23, 24]

Mechanistically, Watanabe et al.[19] attributed the FXR-mediated reduction in hepatic TG level to an FXR-SHP-SREBP-1c pathway, in which FXR induces small heterodimer partner (SHP) and SHP in turn represses sterol regulatory element binding protein 1c (SREBP-1c), a master regulator of genes involved in fatty acid (FA) biosynthesis. SHP is an atypical nuclear receptor that functions as a repressor by interacting with other nuclear receptors or transcription factors, and is a known FXR target gene.[25, 26] In contrast with the data published by Watanabe et al.,[19] subsequent studies from other laboratories failed to support the role of the FXR-SHP-SREBP-1c pathway in FXR-mediated reduction in hepatic TG levels. First, transgenic expression of human SHP in mouse livers increases both SREBP-1c expression and TG level in the liver.[27] Second, Shp−/− mice have reduced hepatic TG level when fed a high-fat diet.[28] Third, knockout of Shp in ob/ob mice prevents hepatic TG accumulation.[29] Lastly, although activation of FXR represses SREBP-1c expression, activation of FXR does not suppress SREBP-1c target genes, such as fatty acid synthase.[30] Collectively, these data strongly suggest that the FXR-SHP-SREBP-1c pathway is unlikely to be the underlying mechanism by which activation of FXR reduces hepatic TG levels.

In this study we investigated the role of hepatic CES1 in both normal and FXR-controlled lipid metabolism. Overexpression of CES1 in the liver lowers hepatic TG levels and improves glucose tolerance in diabetic mice. In contrast, knockdown of CES1 in the liver increases the levels of hepatic TG and plasma non-HDL-C (nonhigh-density lipoprotein cholesterol) levels. The effect of hepatic CES1 on lipid metabolism likely results from its TGH activity, and subsequent changes in fatty acid oxidation (FAO) and/or lipogenesis. Hepatic CES1 is inducible by FXR and is a direct FXR target gene. Activation of FXR lowers the levels of hepatic TG, plasma TG, and plasma cholesterol in a CES1-dependent manner. Thus, our study reveals a critical role for hepatic CES1 in regulating both lipid and carbohydrate metabolism and in FXR-controlled lipid homeostasis.

Materials and Methods

Mice, Diets, and Ligands

C57BL/6 mice, ob/ob mice, db/db mice, and Fxr−/− mice were purchased from the Jackson Laboratories (Bar Harbor, ME). All mice were fed a standard chow diet. Specific FXR agonists GW4064[31] (30 mg/kg, twice a day) and OCA (INT-747)[32] (30 mg/kg/d) have been described previously and were administered by gavage. Unless otherwise stated, male mice were used and all mice were fasted for 5-6 hours prior to euthanization. All the animal studies have been approved by the Institutional Animal Care and Use Committee at Northeast Ohio Medical University.

Real-Time PCR

RNA was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA). Messenger RNA (mRNA) levels were determined by quantitative reverse-transcription polymerase chain reaction (qRT-PCR) on a 7500 real-time PCR machine from Applied Biosystems (Foster City, CA) by using SYBR Green Supermix (Roche, Indianapolis, IN). Results were calculated using Ct values and normalized to 36B4 mRNA level.

Lipid and Lipoprotein Analysis

Approximately 100 mg liver was homogenized in methanol and lipids were extracted in chloroform/methanol (2:1 v/v) as described.[33] Hepatic triglyceride and cholesterol levels were then quantified using Infinity reagents from Thermo Scientific (Waltham, MA). Hepatic fatty acid profile was quantified using gas chromatography (GC)-mass spectrometry at the Mouse Metabolic Phenotyping Center (MMPC) of Case Western Reserve University (Cleveland, OH). Hepatic total free fatty acids and free cholesterol were quantified using kits from BioVision (Milpitas, CA). Plasma lipid and glucose levels were also determined using Infinity reagents. Briefly, after 100 mL plasma was injected, lipoproteins were run at 0.5 mL/min in a buffer containing 0.15 M NaCl, 0.01 M Na2HPO4, 0.1 mM EDTA, pH 7.5, and separated on a Superose 6 10/300 GL column (GE Healthcare) using the BioLogic DuoFlow QuadTec 10 System (Bio-Rad, Hercules, CA). A 500-mL sample per fraction was collected.

Hepatic Lipogenesis

Mice were fasted for 4 hours and then injected intraperitoneally with 2H2O (20-30 mL/g). After 4 hours, liver and plasma were snap-frozen in liquid nitrogen. The newly synthesized palmitate, triglycerides, and cholesterol were measured by mass spectrometry at MMPC of Case Western Reserve University.

Statistical Analysis

Statistical significance was analyzed using an unpaired Student t test or analysis of variance (ANOVA) (GraphPad Prism, San Diego, CA). All values are expressed as mean ± SEM. Differences were considered statistically significant at P < 0.05.

Other detailed Materials and Methods are provided in the Supporting Information.

Results

Overexpression of Hepatic CES1 Lowers Hepatic Triglyceride Levels and Improves Glucose Homeostasis

Our data have shown that hepatic CES1 is down-regulated by fasting (Supporting Fig. S1). To investigate whether hepatic CES1 plays a role in lipid and/or carbohydrate metabolism, we generated adenovirus expressing Ces1 (Ad-Ces1-GFP), which was subsequently injected intravenously into C57BL/6 mice. Hepatic expression of CES1 had no effect on plasma cholesterol or TG levels (Fig. S2A), but significantly lowered plasma glucose levels by ∼30% (Fig. 1A, left panel) and hepatic TG levels by 60% (Fig. 1A, right panel). Hepatic cholesterol levels also tended to decrease (P = 0.14) (data not shown).

Figure 1.

Hepatic expression of CES1 lowers hepatic triglyceride levels and improves glucose homeostasis. (A) C57BL/6 mice were intravenously injected with either Ad-GFP or Ad-Ces1 (n = 7-8 mice per group). After 7 days, mice were fasted for 5 hours. Plasma glucose (left panel) and hepatic TG (right panel) levels were determined. (B-F) ob/ob mice were intravenously injected with Ad-GFP or Ad-Ces1 (n = 5 mice per group). After 7 days and a 5-hour fast, mice were euthanized. Plasma glucose (B, left panel) and hepatic TG levels (B, right panel) were determined. Representative liver images are shown in (C, top panel) and representative hematoxylin and eosin (H&E) staining (C, middle panel) or oil red O staining (C, bottom panel) of the liver sections are shown in (C). Glucose tolerance test (GTT) was performed after a 16-hour fast (D). Western blot assays were performed using liver lysates (E) and protein levels quantified using ImageJ software (F).

To determine whether hepatic CES1 has similar effects on lipid and glucose homeostasis in obese and diabetic mice, we overexpressed Ces1 in the liver of ob/ob mice. Hepatic expression of Ces1 in diabetic ob/ob mice lowered plasma glucose levels by 32% (Fig. 1B, left panel) and hepatic TG levels by 60% (Fig. 1B, right panel), which was also evidenced by the appearance of more reddish liver (Fig. 1C, top panel) and smaller lipid droplets in hepatocytes (Fig. 1C, lower panels). In addition, hepatic expression of Ces1 significantly improved glucose tolerance in a glucose tolerance test (Fig. 1D). Consistent with the latter finding, hepatic expression of Ces1 increased hepatic p-AKT level by 4-fold (Fig. 1E,F), indicating that hepatic CES1 expression enhances hepatic insulin signaling.

Overexpression of Hepatic CES1 Does Not Affect Fatty Acid Lipogenesis or Very-Low-Density Lipoprotein (VLDL) Secretion

To determine the underlying mechanism by which overexpression of hepatic CES1 has beneficial effects on lipid and carbohydrate metabolism, we first determined hepatic gene expression. Hepatic expression of Ces1 in C57BL/6 mice selectively regulated the expression of only certain genes (Fig. S2B,C). Consistent with hepatic cholesterol levels that tended to decrease (data not shown), genes involved in cholesterol biosynthetic pathways, including Srebp-2, HMG-CoA reductase (Hmgcr), and HMG-CoA synthase (Hmgcs), were reduced by 20%-25% (Fig. 2A). In contrast, hepatic genes involved in TG synthesis, including Srebp-1c, acetyl-coA carboxylesterase 1 (Acc1), Fas, acyl-CoA:diacylglycerol acyltransferase 1 (Dgat1), and Dgat2 were not significantly different between the two groups (Fig. 2A). Interestingly, Acc2, which represses fatty acid oxidation (FAO), was inhibited, whereas a few peroxisome proliferator-activated receptor α (PPARα) target genes, including fatty acid translocase Cd36, pyruvate dehydrogenase kinase 4 (Pdk4), angiopoietin-like protein 4 (Angptl4),[34-37] were significantly induced (Fig. 2A). Carnitine palmitoyltransferase 1b (Cpt1b), a rate-limiting enzyme in mitochondrial FAO, was also induced (Fig. 2A). In addition, expression of hepatic Ces1 also induced glucokinase (Gck) expression and tended to reduce hepatic Pepck expression (P = 0.06) (Fig. 2A). Hepatic Gck, which facilitates glucose uptake by phosphorylating glucose to glucose-6-phosphate, is transcriptionally regulated by insulin.[38] Finally, in ob/ob mice hepatic expression of Ces1 reduced the expression of gluconeogenic genes Pepck and G6pase (Fig. 2A, right panel), consistent with enhanced insulin signaling in these mice (Fig. 1D-F). Thus, overexpression of hepatic Ces1 may result in increased PPARα activity and insulin sensitivity in the liver.

Figure 2.

Hepatic expression of CES1 selectively regulates gene expression and has no effect on lipogenesis or VLDL secretion. (A) Hepatic mRNA levels in wild-type (left panel) or ob/ob mice (right panel) were determined by qRT-PCR. (B-D) De novo lipogenesis was determined in mice after injection of 2H2O (n = 5 mice per group). The levels of newly synthesized [2H]palmitate (B), [2H]TG (C), or [2H]cholesterol (D) in the liver were quantified. (E) Hepatic protein level was assessed by western blot assays. (F) VLDL secretion rate was determined (n = 6 mice per group).

The data of Fig. 2A also suggest that overexpression of CES1 may have little effect on lipogenic pathways. We then investigated de novo lipogenesis by injection of mice with 2H2O. Consistent with the gene expression data (Fig. 2A), overexpression of hepatic Ces1 had no effect on hepatic de novo biosynthesis of palmitate (Fig. 2B) or TG (Fig. 2C), but reduced hepatic cholesterol biosynthesis (Fig. 2D). Finally, we also determined the effect of hepatic Ces1 expression on VLDL secretion. Overexpression of hepatic Ces1 had no effect on the protein levels of microsomal triglyceride transfer protein (MTP) or ApoB (Fig. 2E), or VLDL secretion rate (Fig. 2F). Thus, lipogenesis or VLDL secretion does not contribute to the reduced hepatic TG levels following hepatic CES1 expression.

Overexpression of Hepatic CES1 Increases Hepatic TG Hydrolysis and Stimulates Fatty Acid Oxidation

The data of Fig. 2A suggest that CES1 overexpression may induce PPARα activity. To test this hypothesis, we cotransfected a Ces1-expressing plasmid and a luciferase-reporter plasmid containing three copies of PPARα responsive element (PPRE). Overexpression of CES1 highly induced PPARα activity (Fig. 3A, top panel). Free fatty acids (FFAs), which may derive from TG hydrolysis, have been well documented as endogenous ligands for PPARα.[39] Thus, we hypothesized that CES1 had TGH activity (Fig. 3A, bottom panel). Overexpression of Ces1 significantly increased the release of [3H]FFAs from [3H]triolein in both COS-7 cells (Fig. 3B) and the liver (Fig. 3C), indicating that CES1 has TGH activity.

Figure 3.

Hepatic expression of CES1 increases triglyceride hydrolase activity and activates PPARα. (A) HepG2 cells were transfected with a control plasmid or a Ces1-expressing plasmid together with a 3xPPRE-Luc reporter plasmid. Luciferase activity was determined (top panel). In the bottom panel, the diagram shows that CES1 hydrolyzes TG and releases FFAs, which bind to PPARα/RXR complex and then induce PPARα activity. (B,C) TGH activity was assessed using lysates from COS-7 cells (B) or the liver (C). (D,E) Hepatic fatty acid profile was determined by GC-mass spectrometry (D) and hepatic FFA levels were quantified (n = 8 mice per group) (E). (F) FAO was performed in the liver cell line AML12 cells that were infected with Ad-GFP or Ad-CES1 for 48 hours, or treated with either vehicle or carnitine (1 mM) (n = 3-5 per group). Carnitine treatment serves as a positive control. Veh, vehicle.

Although the expression of CES1 increased hepatic TGH activity (Fig. 3C), hepatic total fatty acids and a number of long-chain fatty acids, such as C14:0, C16:0, C16:1, and C18:2, were significantly reduced (Fig. 3D). Hepatic FFA levels were also significantly reduced (Fig. 3E). The repression of Acc2 and induction of Cd36, Pdk4, Angptl4, and Cpt1b (Fig. 2A), all of which are involved in FAO, suggest that CES1 may regulate FAO. As predicted, overexpression of CES1 increased FAO (Fig. 2F). In addition, hepatic expression of CES1 also tended to increase plasma levels of β-hydroxybutyrate (P = 0.07) (Fig. S3). Thus, the reduced hepatic FFA levels in Ces1-overexpressing mice may be a net effect of increased TGH activity and increased FAO.

Knockdown of Hepatic CES1 Causes Hepatic Steatosis and Elevated Plasma Cholesterol Levels

To determine whether hepatic CES1 is required for maintaining normal lipid and glucose homeostasis, we generated three adenoviruses expressing short hairpin RNA (shRNA) specifically against murine Ces1 (Ad-shCes1). One of the three Ad-shCes1 adenenoviruses was particularly efficient in knocking down endogenous Ces1 expression in cultured cells (Fig. S4), and was used for subsequent studies. Ad-shLacZ (control) or Ad-shCes1 was intravenously injected into C57BL/6 mice. Remarkably, hepatic expression of Ces1 shRNA reduced hepatic levels of Ces1 mRNA (Fig. 4A) and protein (Fig. 4B) by ∼95%, increased plasma total cholesterol (TC) levels by ∼2-fold (Fig. 3C), but had no effect on plasma TG (Fig. 4C) or glucose levels (data not shown). Analysis of plasma by fast protein liquid chromatography (FPLC) indicated that hepatic Ces1 deficiency markedly increased plasma non-HDL-C levels (Fig. 4D), but had no much effect on plasma TG lipoprotein profile (data not shown). Hepatic Ces1 deficiency also caused development of fatty liver (Fig. 4E), which was associated with more than a 2-fold increase in hepatic TG levels but unchanged hepatic TC levels (Fig. 4F). Thus, hepatic CES1 is indispensable for maintaining normal hepatic TG homeostasis and plasma cholesterol levels.

Figure 4.

Loss of hepatic CES1 causes fatty liver and increased plasma cholesterol level. (A-F) C57BL/6 mice were intravenously injected with either Ad-shLacZ or Ad-shCes1 (n = 8 mice per group). After a 6-hour fast, mice were euthanized. Hepatic mRNA (A) and protein (B) levels were determined. Plasma TG and total cholesterol (TC) levels were quantified (C). Plasma cholesterol lipoprotein profile was determined by FPLC (D). Representative liver images (top panel) and oil red O staining of liver sections (lower panel) are shown (E). Hepatic TG and TC levels were quantified (F).

Hepatic CES1 Deficiency Results in Increased Lipogenesis

The data of Fig. 4 show that the effect of hepatic CES1 deficiency on lipid metabolism is quite striking. Analysis of hepatic gene expression indicated that many SREBP- or cholesterol/LXR (liver X receptor)-regulated genes, including Hmgcs, Acc-1, Dgat1, Dgat2, ATP citrate lyase (Acl),[40, 41] Abca1, and Cd36,[42] were significantly induced (Fig. 5A). The induction of these genes was selective, as many other genes were unchanged (Fig. S5). Consistent with the induction of numerous lipogenic genes, hepatic de novo lipogenesis of palmitate (Fig. 5B), TG (Fig. 5C), and cholesterol (Fig. 5D), were induced by 2.1-, 2.9-, and 2.1-fold, respectively, indicating that loss of hepatic CES1 has profound effects on lipogenesis. Investigations of hepatic mature/nuclear SREBPs (nSREBPs) showed that loss of hepatic CES1 resulted in an increase in the levels of hepatic nSREBP-1 and nSREBP-2 by 2-fold and 2.4-fold, respectively (Fig. 5E). The latter data suggest that elevated mature SREBP levels are responsible for increased lipogenesis.

Figure 5.

Loss of hepatic CES1 induces de novo lipogenesis. (A) Hepatic mRNA levels were determined by qRT-PCR (n = 8 per group). (B-D) De novo lipogenesis was determined in mice after injection of 2H2O (n = 5 per group). The levels of newly synthesized [2H]palmitate (B), [2H]TG (C), or [2H]cholesterol (D) in the liver were quantified. (E) Hepatic protein levels were determined by western blot assays (E, left panel) and then quantified by using ImageJ software (E, right panel). nBP-1, nuclear form SREBP-1. nBP-2, nuclear form SREBP-2. (F) Hepatic ratio of CE to FC was determined (n = 8 mice per group). The CE/FC ratio in shLacZ-treated mice was set at 1.

SREBP processing is known to be regulated by the change in intracellular sterol levels; when cellular sterol levels are low, mature SREBPs are increased by inducing SREBP processing.[43] Consistent with previous data showing that CES1 has CEH activity,[6, 44] loss of hepatic Ces1 significantly reduced hepatic free cholesterol (FC) level (Fig. S6A), whereas the hepatic FFA level remained unchanged (Fig. S6B). Ces1-deficient mice also had significantly higher levels of cholesterol esters (CE) to FC (Fig. 5F). Thus, the reduced hepatic FC levels may account, at least in part, for the increased levels of hepatic nSREBP-1 and nSREBP-2.

Finally, we investigated the effect of hepatic Ces1 deficiency on VLDL secretion. Loss of hepatic Ces1 had no effect on the protein levels of MTP, ApoB100, or ApoB48 (Fig. S6C), but tended to increase the hepatic VLDL secretion rate (Fig. S6D). Thus, loss of hepatic CES1 causes hepatic steatosis and hyperlipidemia (Fig. 4), likely resulting from increased lipogenesis.

Activation of FXR Induces Hepatic CES1 Expression

FXR is highly expressed in the liver and plays an important role in regulating lipid and carbohydrate metabolism. Activation of FXR by the specific agonist GW4064 reduced hepatic TG levels by ∼40% but did not alter hepatic cholesterol levels (Fig. 6A). Analysis of hepatic gene expression showed that GW4064 treatment induced hepatic mRNA levels of Shp and Ces1, but had no effect on Ces2 or Ces3 mRNA levels (Fig. 6B). GW4064 treatment also highly induced hepatic CES1 protein levels, but did not alter hepatic ApoB100 or MTP protein levels (Fig. 6C), suggesting that GW4064 does not affect VLDL secretion. In addition to GW4064, hepatic Ces1 mRNA was also induced by treatment with obeticholic acid (OCA, INT-747) (Fig. 6D), another potent and selective FXR agonist,[32] or cholic acid, the endogenous FXR ligand (Fig. 6E). In contrast, the hepatic Ces1 mRNA level was reduced in Fxr−/− mice (Fig. 6F). Thus, hepatic CES1 expression is regulated by FXR.

Figure 6.

Hepatic CES1 is regulated by FXR. (A-C) C57BL/6 mice were gavaged with vehicle (0.5% CMC [carboxymethyl cellulose]) or GW4064 (30 mg/kg, twice a day) for 7 days (n = 8 mice per group). Hepatic TG and TC levels were determined (A). Hepatic mRNA levels were quantified by qRT-PCR (B) and hepatic protein levels were determined by western blot assays (C). (D) C57BL/6 mice were gavaged with either 0.5% CMC (vehicle) or INT-747 (OCA, 30 mg/kg/d) for 7 days (n = 5 mice per group). Hepatic mRNA levels were determined. Cyp7a1 and Cyp8b1 served as positive controls. (E) C57BL/6 mice were fed a chow diet or 0.5% cholic acid (CA) for 7 days. Hepatic mRNA levels were quantified. (F) Hepatic mRNA levels were quantified by qRT-PCR in wild-type or Fxr−/− mice (n = 8 mice per group).

CES1 Is a Direct FXR Target Gene

To investigate how FXR regulates CES1 expression, we first determined the FXR response element(s) (FXRE) in the Ces1 gene promoter. Analysis of the promoter activity by using a serial of 5′-deletion constructs (Fig. 7A) and mutagenesis (Fig. 7B) uncovered an FXRE located between 300 bp and 150 bp upstream of the transcription start site. Electrophoretic mobility shift assays showed that the FXR/RXR complex bound to this FXRE in vitro and this binding was competed away by cold wild-type oligonucleotides but not by mutant oligonucleotides (Fig. 7C). Finally, chromatin immunoprecipitation assay showed that FXR bound to the promoter of Akr1b7, a known FXR target gene,[45, 46] as well as the Ces1 promoter in the liver (Fig. 7D). Collectively, the data of Fig. 7 demonstrate that Ces1 is a direct FXR target gene, and that CES1 is induced by FXR through an FXRE located between 150 bp and 300 bp upstream of the transcription start site.

Figure 7.

CES1 is a direct FXR target gene. (A,B) Transient transfection assays were performed using promoter-luciferase constructs containing a serial of 5′-deletions (A) or mutations (B). (C) EMSA assays were performed using in vitro transcribed/translated proteins. Wild-type (WT) and mutant (MUT) oligos were used in the competition assays. (D) Chromatin immunoprecipitation (ChIP) assays were performed using liver lysates (n = 3 per group). Akr1b7 served as a positive control.

Hepatic CES1 Is Essential for Activated FXR to Improve Lipid Homeostasis

Activation of FXR has beneficial effects on lipid homeostasis. The findings that hepatic CES1 is important for maintaining lipid homeostasis and is regulated by FXR prompted us to study the role of hepatic CES1 in FXR-controlled lipid homeostasis. C57BL/6 mice were injected intravenously with either Ad-shLacZ or Ad-shCes1, followed by treatment with either vehicle or OCA for 7 days. As expected, knockdown of hepatic Ces1 increased plasma cholesterol levels by ∼4 fold (Fig. 8A). Interestingly, OCA treatment reduced plasma cholesterol levels in the control mice but not in Ces1-deficient mice (Fig. 8A). Similarly, OCA treatment also reduced plasma TG levels in the control mice but not in the Ces1-deficient mice (Fig. 8B). In contrast, OCA lowered plasma glucose levels in both the control mice and Ces1-deficient mice (Fig. S7A). Analysis of plasma by FPLC showed that loss of hepatic Ces1 markedly increased plasma non-HDL-C levels (Fig. S7B) and also slightly increased LDL triglyceride levels (Fig. S7C), and these changes were exacerbated following OCA treatment (Fig. S7B,C). Thus, hepatic CES1 is critical for an FXR agonist to lower plasma TG and cholesterol levels.

Figure 8.

Essential roles of hepatic CES1 in FXR-regulated lipid homeostasis. (A-E) C57BL/6 mice were intravenously injected with Ad-shLacZ or Ad-shCes1. The next day these mice were gavaged with either vehicle (0.5% CMC) or INT-747 (OCA, 30 mg/kg/d) for 7 days (n = 8-10 mice per group). After a 5-hour fast, mice were euthanized. Plasma total cholesterol (TC) (A) and TG (B) levels as well as hepatic TG levels (C) were determined. Hepatic mRNA levels of Shp (D) and Ces1 (E, left panel) were determined by qRT-PCR. Western blot assays were also performed (E, right panel). NS, not significant. In (C), treatment of the control group versus the Ces1-deficient group by OCA was significantly different (P < 0.01). (F) Central roles of hepatic CES1 in lipid and carbohydrate metabolism and in FXR signaling. Expression of hepatic CES1 lowers hepatic TG level by increasing TG hydrolysis and FAO. As a result, the hepatic FFA level is reduced, which in turn causes an increase in hepatic insulin sensitivity and a reduction in plasma glucose level. When hepatic CES1 is deficient, hepatic free cholesterol (FC) levels are reduced, resulting in an increase in hepatic SREBP processing. The increased nSREBP1 and nSERBP2 levels promote lipogenesis, resulting in increased levels of hepatic TG and plasma cholesterol. Hepatic CES1 can be induced FXR and bile acids. The induction of hepatic CES1 is critical for activated FXR to improve lipid homeostasis.

In the liver, OCA treatment had no effect on hepatic cholesterol levels in control mice but they were increased in Ces1-deficient mice (Fig. S7D). As expected, OCA treatment reduced hepatic TG levels in control mice (Fig. 8C). Interestingly, OCA treatment had no significant effect on hepatic TG levels in Ces1-deficient mice (Fig. 8C). Thus, hepatic CES1 is important for activated FXR to lower hepatic TG levels.

Analysis of hepatic gene expression showed that OCA induced the FXR target gene Shp in both control mice and Ces1-deficient mice (Fig. 8D). OCA also induced hepatic Ces1 expression in control mice, but this induction was blunted in Ces1-deficient mice (Fig. 8E).

We also treated ob/ob mice with OCA. Similar to what we observed in C57BL/6 mice, we found that OCA treatment lowered plasma cholesterol and TG levels as well as hepatic TG levels in a CES1-dependent manner (Fig. S8A,B,D), but had no effect on hepatic cholesterol levels (Fig. S8C). Interestingly, loss of hepatic CES1 did not further increase hepatic TG levels in ob/ob mice that were deficient in hepatic Ces1 (Fig. S8D). One possibility is that hepatic TG accumulation in ob/ob mice had reached the maximum level. Analysis of hepatic gene expression showed that OCA treatment induced hepatic mRNA levels of Ces1 in the control mice (Fig. S8E) and Shp in both the control and Ces1-deficient mice (Fig. S8F). Together, the data of Fig. 8 and Fig. S8 demonstrate that hepatic CES1 is indispensable for activated FXR to regulate both plasma lipid and hepatic TG levels.

Discussion

CES1 is highly expressed in the liver, but the role of hepatic CES1 in lipid and carbohydrate metabolism has not been investigated before. Herein, we used both gain- and loss-of-function approaches to demonstrate that hepatic CES1 plays a critical role in regulating both lipid and carbohydrate metabolism. In addition, we also demonstrate that FXR regulates hepatic CES1 expression and that the induction of hepatic CES1 is indispensable for activated FXR to improve lipid homeostasis (Fig. 8F). Numerous studies have demonstrated that FXR is a therapeutic target for treatment of the metabolic disease,[9, 24] particularly for treatment of NAFLD.[47] Our data suggest that hepatic CES1 is a promising therapeutic target for treatment of NAFLD. Consequently, the findings from the current study may have significant therapeutic implications.

One key finding in our study is the demonstration that hepatic CES1 has TGH activity and is essential for maintaining hepatic TG homeostasis. Hepatic TG mobilization has been poorly understood. Previous data suggest that ATGL may play an important role in mobilizing hepatic TG.[2, 3] However, ATGL expression is low in hepatocytes.[4, 48] Since CES1 is highly expressed in the liver and plays an essential role is controlling hepatic TG levels, CES1 may represent one of the major TG hydrolases in the liver.

Previous data have shown that CES1 also has CEH activity.[6, 44] Thus, the effect of hepatic CES1 on lipid and carbohydrate metabolism may depend on its TGH activity and/or CEH activity (Fig. 8F). Expression of hepatic CES1 lowers hepatic TG levels and improves glucose homeostasis likely through increased TG hydrolysis and subsequent increase in PPARα activity and FAO, as neither lipogenesis nor VLDL secretion is changed (Figs. 2, 3). In contrast, loss of hepatic CES1 increases hepatic TG levels and plasma cholesterol levels likely through increased lipogenesis, which may result from loss of CEH activity and subsequent increase in SREBP processing (Figs. 6, 8F).

Very recently and after this study was completed, Quiroga et al.[49] reported that global Ces1−/− mice develop hepatic steatosis, increased obesity, and hyperlipidemia. Given that CES1 is expressed in multiple tissues, including liver, intestine, and other metabolic tissues[50] (data not shown), one cannot conclude which tissue(s) play a role in the observed metabolic changes. Moreover, intestinal CES1 is also reported to regulate chylomicron secretion,[50] which can contribute significantly to both hepatic TG and plasma lipid metabolism. Thus, it is critical to study the relative roles of hepatic versus intestinal CES1 in lipid and carbohydrate metabolism. Our study provides the first evidence demonstrating that hepatic CES1 plays a critical role in regulating both lipid and glucose homeostasis.

The finding that expression of hepatic CES1 improves glucose homeostasis is intriguing (Fig. 1). Our data show that expression of hepatic CES1 enhances hepatic insulin signaling, as hepatic glucokinase and p-AKT levels are increased, whereas hepatic gluconeogenic genes are repressed (Figs. 1, 2). The increased hepatic insulin signaling may result from reduced hepatic FFA level (Fig. 3E).

FXR has been recognized as an interesting target for treatment of fatty liver disease.[24, 47] However, the mechanism underlying FXR-mediated alleviation of hepatic steatosis remains undetermined. Our study provides compelling evidence demonstrating that CES1 is a direct downstream target gene of FXR (Figs. 6, 7) and is necessary for activated FXR to lower plasma TG and cholesterol levels (Fig. 8). We also demonstrate that activation of FXR lowers hepatic TG level, at least in part through induction of hepatic CES1 (Fig. 8). Thus, hepatic CES1 plays a crucial role in FXR-controlled lipid homeostasis. Interestingly, CES1 overexpression is not sufficient to lower plasma TG or cholesterol levels (Fig. S2). One possibility is that CES1 is just one of the downstream components of FXR signaling that are responsible for FXR-controlled lipid homeostasis.

Both human CES1 and mouse CES1 are highly expressed in the liver and share 74% amino acid homology. However, nothing has been known about the role of human CES1 in triglyceride or glucose metabolism. One of our future directions will be to investigate whether human CES1 and mouse CES1 have similar effects on lipid and glucose homeostasis.

In summary, we have shown that hepatic expression of CES1 has beneficial effects on lipid and carbohydrate metabolism, whereas loss of hepatic CES1 causes fatty liver and pro-atherogenic lipid profile. Such effects of CES1 may result from its TGH and/or CEH activity and subsequent change in FAO and/or lipogenesis. In addition, activation of FXR induces hepatic CES1, which is critical for FXR-mediated improvement of lipid homeostasis. Together, our data indicate that hepatic CES1 is an important component of FXR signaling in metabolic control, and also suggest that hepatic CES1 may represent a therapeutic target for treatment of NAFLD.

Ancillary