Carboxylesterase 1 Is Regulated by Hepatocyte Nuclear Factor 4α and Protects Against Alcohol- and MCD diet-induced Liver Injury

The liver is a major organ that controls hepatic and systemic homeostasis. Dysregulation of liver metabolism may cause liver injury. Previous studies have demonstrated that carboxylesterase 1 (CES1) regulates hepatic triglyceride metabolism and protects against liver steatosis. In the present study, we investigated whether CES1 played a role in the development of alcoholic liver disease (ALD) and methionine and choline-deficient (MCD) diet-induced liver injury. Both hepatocyte nuclear factor 4α (HNF4α) and CES1 were markedly reduced in patients with alcoholic steatohepatitis. Alcohol repressed both HNF4α and CES1 expression in primary hepatocytes. HNF4α regulated CES1 expression by directly binding to the proximal promoter of CES1. Global inactivation of CES1 aggravated alcohol- or MCD diet-induced liver inflammation and liver injury, likely as a result of increased production of acetaldehyde and reactive oxygen species and mitochondrial dysfunctions. Knockdown of hepatic CES1 exacerbated ethanol-induced steatohepatitis. These data indicate that CES1 plays a crucial role in protection against alcohol- or MCD diet-induced liver injury.

fatty acids (FFA), cholesterol and miR-34a expression. HNF4α controls the basal expression of many genes in the liver, but it is unclear whether HNF4α also regulates CES1 expression.
In the present study, we first investigated hepatic CES1 expression in patients with alcoholic steatohepatitis. We then investigated how alcohol downregulated hepatic CES1 expression and whether CES1 played a role in the development of ALD and non-alcoholic liver injury. Our data show that ethanol reduces hepatic CES1 expression likely through inhibition of HNF4α and that CES1 plays a crucial role in preventing ethanol-or methionine and choline-deficient (MCD) diet-induced liver injury.

Results
Alcohol reduces CES1 and HNF4α expression in patients with alcoholic steatohepatitis and in mice treated with alcohol. To investigate whether CES1 is associated with the development of ALD, we investigated the expression of CES1 in patients with alcoholic steatohepatitis. Hepatic CES1 mRNA level was reduced by 75% (Fig. 1A) and protein level decreased by ~85% (Fig. 1B,C). Interestingly, the nuclear receptor HNF4α was also markedly repressed by > 84% in both mRNA (Fig. 1A) and protein (Fig. 1B,C) levels.
In mice chronically fed a Liber-DeCarli ethanol diet for 10 days plus binge ethanol (5 g/kg) 18 , hepatic Ces1 and Hnf4α mRNA (Fig. 1D) and protein (Fig. 1E) levels were decreased. Consistent with the in vivo data, ethanol treatment repressed Ces1 and Hnf4α mRNA (Fig. 1F) and protein (Fig. 1G) levels in primary hepatocytes. Thus, the data of Fig. 1 indicate that ethanol inhibits hepatic CES1 and HNF4α expression in both mice and humans.
HNF4α regulates CES1 expression via binding to a DR-1 binding. The finding that alcohol inhibits both HNF4α and CES1 expression suggests that HNF4α may regulate CES1 expression. Over-expression of HNF4α induced CES1 mRNA level by 2.3 fold in HepG2 cells ( Fig. 2A). HNF4α over-expression also induced Next, we examined how HNF4α regulates CES1 expression. Promoter-luciferase assays showed that HNF4α induced the luciferase activity by 10 fold, 4 fold and 6 fold in the − 1.9 kb, − 0.98 kb and − 0.3 kb Ces1 promoters, respectively, but not in the − 0.25 kb Ces1 promoter (Fig. 3A), suggesting that HNF4α may bind to the response element(s) located between − 0.3 kb and − 0.25 kb of the Ces1 promoter. Indeed, there was a potential DR-1 element (direct repeat separated by one base pair) between 300 bp and 287 bp upstream of the transcription start site (Fig. 3B). Mutation of the DR-1 element abolished the induction of Ces1 promoter activity by HNF4α (Fig. 3A). Electrophoretic mobility shift assays showed that HNF4α protein bound to the DR-1 element and this binding was inhibited by wild-type but not mutant DR-1 oligonucleotides (Fig. 3C, left panel). In addition, an HNF4α antibody was able to supershift the DNA/protein complex (Fig. 3C, right panel), indicating that HNF4α binds to the DR-1 element in vitro. Finally, chromatin immunoprecipitation assays showed that HNF4α protein bound . mRNA levels were determined by qRT-PCR (A). Protein levels were determined by Western blotting (B) and then quantified (C). (D,E) C57BL/6 mice were fed a Lieber-DeCarli diet for 5 days, followed by fed a Lieber-DeCarli diet containing 5% ethanol or pair-fed a control diet for 10 days. On the 16 th day, mice were gavaged with a single dose of ethanol (5 g/kg) (n = 8). mRNA levels were then determined by qRT-PCR (D) and protein levels determined by Western blots (E). (F,G) Mouse primary hepatocytes were treated with 0, 50 or 100 mM ethanol for 24 h (n = 3). mRNA (F) and protein (G) levels were determined. *p < 0.05, **p < 0.01. Unpaired Student t-test was used for statistical analysis. to the Ces1 promoter containing the DR-1 element in the liver (Fig. 3D). Collectively, the data of Figs 2 and 3 demonstrate that HNF4α regulates Ces1 expression by binding to a DR-1 element located between 300 bp and 287 bp upstream of the transcription start site.  Global inactivation of CES1 aggravates MCD diet-induced liver inflammation. Hepatic CES1 is known to regulate lipid and glucose metabolism 11,12,14 . However, the role of CES1 in liver inflammation or injury is unknown. We have therefore generated global Ces1 −/− mice, which did not express any Ces1 mRNA or protein in the liver or other tissues (see Supplementary Fig. S1). When Ces1 +/+ mice and Ces1 −/− mice were fed an MCD diet for 4 weeks, there was a 21% increase in hepatic TG level but hepatic cholesterol level was unchanged (Fig. 4A). Ces1 −/− mice had more fat accumulation than Ces1 +/+ mice (5.03% vs. 3.37%) but reduced lean mass (Fig. 4B). Ces1 −/− mice also had increased plasma alanine aminotransferase (ALT) level (Fig. 4C) as well as hepatic mRNA levels of tumor necrosis factor α (Tnfα ) (Fig. 4D), interleukin 1β (Il1β ) (Fig. 4E), and interleukin 6 (Il6) (Fig. 4F). These data demonstrate that loss of CES1 aggravates MCD diet-induced liver inflammation and liver injury.

Global inactivation of CES1 exacerbates alcohol-induced hepatitis.
To test whether global inactivation of CES1 affects the development of ALD, Ces1 +/+ mice and Ces1 −/− mice were fed a Lieber-DeCarbli diet containing 5% ethanol or pair-fed a control diet for 10 days followed by a single gavage of ethanol (3 g/kg) as described 18 . The chronic binge drinking did not affect their body weight (see Supplementary Fig. S2). Ethanol feeding significantly increased plasma levels of ALT ( Global inactivation of CES1 aggravates ethanol-induced acetaldehyde and oxidative stress production and mitochondrial dysfunction. To understand how CES1 inactivation aggravates ethanol-induced liver injury, we first measured acetaldehyde level in the liver. Acetaldehyde is a reactive compound that is highly toxic to hepatocytes because it can sensitize cells to oxidative stress or other damaging signals, ultimately leading to mitochondrial damage and cell death 2,19 . At the basal level, hepatic acetaldehyde level was increased by 2.2 fold in Ces1 −/− mice in comparison to Ces1 +/+ mice (Fig. 6A). Upon ethanol feeding, hepatic acetaldehyde level was increased by 1.7 fold in Ces1 +/+ mice and 2 fold in Ces1 −/− mice (Fig. 6A). Hepatic acetaldehyde level in ethanol-fed Ces1 −/− mice was 2.4 fold higher than in ethanol-fed Ces1 +/+ mice (Fig. 6A). Consistent with these observations, the mRNA level of acetaldehyde dehydrogenase 2 (Aldh2) was reduced by 50% in Ces1 −/− mice (p < 0.05; data not shown).
Oxidative stress is an important contributor to the pathogenesis of ALD. Mitochondrial reactive oxygen species (ROS) triggers pro-inflammatory cytokine production, damages mitochondrial DNA and promotes lipid peroxidation [20][21][22] . We therefore examined whether the aggravated liver inflammation in Ces1 −/− mice was associated with increased ROS. Indeed, upon alcohol feeding hepatic levels of hydrogen peroxide (H 2 O 2 ) (Fig. 6B) and malondialdehyde (MDA) (Fig. 6C) in Ces1 −/− mice were increased by 5.5 fold and 3 fold, respectively, whereas these increases were less pronounced in Ces1 +/+ mice (Fig. 6B,C). In addition, hepatic H 2 O 2 and MDA levels were increased by 3 fold and 2 fold, respectively, in ethanol-fed Ces1 −/− mice than in ethanol-fed Ces1 +/+ mice (Fig. 6B,C), indicating that CES1 is critical for preventing ethanol-induced oxidative stress.  Ethanol feeding significantly increases acetate level and therefore enhances the synthesis of acetyl-CoA and β -hydroxybutyrate (β -HB). On the other hand, elevated mitochondrial FAO is also accompanied by increased β -HB synthesis. As such, plasma β -HB level may reflect mitochondrial functions under normal conditions. As shown in Fig. 6D, ethanol significantly increased plasma β -HB level in Ces1 +/+ mice but not in Ces1 −/− mice. In contrast, ethanol induced a 2.5-fold increase in hepatic FFA level in Ces1 −/− mice but not in Ces1 +/+ mice (Fig. 6E). The increase in hepatic FFA level was not accompanied by any change in genes involved in fatty acid synthesis (Srebp1c, Fas, Acc1, Acc2), transport (CD36, Fabp1) or oxidation (Pparα , Cpt1α , Mcad, Acox1, and Acox2) (see Supplementary Fig. S2). These data suggest that Ces1 −/− mice may have an impaired mitochondrial function.
Hepatic CES1 deficiency exacerbates alcohol-induced liver steatosis. In addition to the liver, CES1 is also expressed in other tissues, such as intestine, adipose tissue and macrophages. We therefore investigated whether the aggravated liver injury observed in global Ces1 −/− mice was due to inactivation of hepatic Ces1. Consistent with a role in TG hydrolysis, over-expression of Ces1 prevented alcohol-induced TG accumulation in AML12 cells (see Supplementary Fig. S3), a murine hepatocyte cell line. We then fed C57BL/6 mice a Lieber-DeCarli control diet for 5 days. On day 6, the mice were fed either a Lieber-DeCarli control diet or 5% ethanol diet and also injected i.v. with adenoviruses expressing Ces1 shRNA (Ad-shCes1) or LacZ shRNA (Ad-shLacZ; control) as described 12 . On day 16, the mice were gavaged with a single dosage of ethanol (5 g/kg). Mice infected with Ad-shCes1 had a ~97% reduction in hepatic Ces1 mRNA level (See Supplementary Fig. S4A). Body weight and food intake were comparable between the two genotypes (see Supplementary Fig. S4B and S4C). Alcohol treatment led to increased plasma and hepatic levels of TG in both the Ces1-deficient mice and control mice (Fig. 7A,B). In addition, alcohol-fed Ces1-deficient mice had ~2 fold increase in plasma or hepatic TG levels than alcohol-fed control mice (Fig. 7A,B). Similar results were observed for hepatic FFAs (Fig. 7C). In contrast, there was no change in hepatic cholesterol levels among the different groups (see Supplementary Fig. S4D). Oil red O staining of liver sections also confirmed hepatic TG accumulation (Fig. 7D). Consistent with the change in hepatic TG levels, knockdown of Ces1 in ethanol-fed mice markedly increased mRNA levels of lipogenic genes, including Srebp-1, Fas, Acc1, Acc2, diacylglycerol O-acyltransferase 2 (Dgat2) and peroxisome proliferator-activated receptor γ (Pparγ ) (Fig. 7E). Thus, the data of Fig. 7 indicate that hepatic CES1 deficiency aggravates ethanol-induced liver steatosis via inducing lipogenic genes.
Hepatic CES1 deficiency exacerbates alcohol-induced hepatitis. Next, we determined whether hepatic CES1 deficiency affected liver injury. Ethanol induced plasma ALT (Fig. 8A) and AST (Fig. 8B) levels in both the control mice and Ces1-deficient mice. Hepatic Ces1 deficiency further increased plasma ALT and AST levels (Fig. 8A,B). In addition, ethanol significantly induced hepatic mRNA levels of Tnfα (Fig. 8C), Il1β (Fig. 8D), Il6 (Fig. 8E) and Mcp1 (Fig. 8F) in Ces1-deficient mice but not in the control mice except for Il6. Furthermore, hepatic Tnfα , Il1β , Il6 and Mcp1 mRNA levels were increased by > 2 fold in ethanol-fed Ces1-deficient mice than in ethanol-fed control mice (Fig. 8C-F). Finally, ethanol increased plasma β -HB levels in the control mice but reduced plasma β -HB levels in the Ces1-deficient mice (Fig. 8G), suggesting that hepatic CES1 deficiency may cause mitochondrial dysfunction. Collectively, the data of Figs 7 and 8 demonstrate that hepatic CES1 deficiency is sufficient to aggravate alcohol-induced steatohepatitis.

Discussion
CES1 is a drug-metabolizing enzyme. Its role in liver inflammation or liver injury has not been investigated previously. In the current study, we demonstrate that CES1 protects liver from alcohol-or MCD diet-induced liver inflammation or liver injury. Importantly, we also show that hepatic CES1 is markedly reduced in patients with alcoholic steatohepatitis. Mechanistically, our data suggest that ethanol inhibits CES1 expression likely through reducing HNF4α expression. Thus, the present study suggests that the HNF4α -CES1 pathway may play an important role in the protection against the development of ALD.
CES1 has the ability to hydrolyze amide or ester bonds 24 . Our previous studies show that TG is a substrate for CES1 12 . Defective TG hydrolase activity is responsible for elevated TG accumulation in the liver of Ces1-deficient mice. Ethanol is unlikely to be a direct substrate of CES1. However, the metabolites of CES1 may affect alcohol metabolism and mitochondrial functions. When CES1 is deficient, the change in the levels of metabolite(s) and/ or substrate(s) of CES1 may cause accumulation of toxic substances, such as acetaldehyde and ROS. These toxic substances are sufficient to cause damages on mitochondria and cell membranes, eventually resulting in liver inflammation and damage. One of our future directions will be to determine which metabolite(s) and substrate(s) cause an increase in these toxic substances. In agreement with our hypothesis, ALDH2 is inhibited in Ces1 −/− mice. A previous report shows that HNF4α also controls ALDH2 transcription 25 , supporting our current finding that the HNF4α -CES1 pathway regulates the development of ALD.
HNF4α is known to control the basal expression of a variety of genes associated with lipid, glucose, bile acid and drug metabolism, differentiation and morphogenesis 26 . The current study identifies CES1, a TG hydrolase in the liver 12 , as a direct target gene of HNF4α , suggesting that HNF4α may also regulate lipolysis. Under the conditions of common metabolic stress (diabetes, obesity and NAFLD, HFD feeding), elevated FFAs, cholesterol, p53 and miR-34a orchestrate to inhibit HNF4α expression 27 . In addition, alcoholism leads to hypozincemia 28 and zinc deprivation suppresses the DNA-binding activity of HNF4α in HepG2 cells 29 . Thus, ethanol treatment inhibits HNF4α expression and DNA binding activity, likely through inducing lipogenesis and FFA levels and reduced zinc level. It is also possible that other unknown mechanism(s) are involved in inhibiting HNF4α expression. An unexpected finding in this study is that liver-specific knockdown, but not global inactivation, of CES1 aggravates ethanol-induced liver steatosis. Loss of hepatic CES1 is known to induce lipogenesis and liver steatosis 12 . In addition to the liver, CES1 is also expressed in the intestine, in which CES1 is reported to regulate chylomicron production 30 . We speculate that loss of intestinal CES1 may affect alcohol metabolism and fat absorption, thus resulting in unchanged hepatic TG levels in Ces1 −/− mice.
Global or liver-specific loss of CES1 causes liver inflammation in the presence or absence of alcohol feeding, underscoring the importance of CES1 in controlling inflammatory response. CES1 deficiency causes elevated levels of acetaldehyde, H 2 O 2 , MDA and FFAs as well as mitochondrial dysfunctions, which are known to contribute to liver inflammation and liver damage. Hepatic FAEE, which is known to cause cytotoxicity, is increased upon alcohol feeding 31 . CES1 is shown to play a role in FAEE metabolism 32,33 . Thus, we cannot rule out that FAEE also plays a role in liver damage in Ces1 −/− mice.
Our recent data show that CES2 is also regulated by HNF4α 34 . CES2 prevents liver steatosis by regulating lipolysis, ER stress and lipogenesis 34 . Similarly, CES1 prevents lipolysis by modulating lipolysis and lipogenesis 12 . These functional similarities between CES1 and CES2 likely result from their triglyceride hydrolase activity. In addition to HNF4α , CES1 is also induced by glucose 14 and farnesoid X receptor 12 , but repressed during fasting 14 and alcohol. In contrast, CES2 is repressed in db/db mice, high fat diet-fed mice and patients with non-alcoholic steatohepatitis. Therefore, the regulation of CES1 and CES2 is largely different. It is unclear whether CES2 also plays a role in ALD.
In summary, the current study demonstrates that CES1 protects liver from ethanol-or MCD diet-induced liver inflammation and damage. The impaired HNF4α -CES1 pathway may contribute to the pathogenesis of ALD. Targeting hepatic CES1 may represent a novel strategy for prevention and treatment of ALD and non-alcoholic liver injury.

Methods
Mice, diets and human ALD samples. Liver-specific Hnf4α −/− mice were generated by crossing Hnf4α fl/fl mice with albumin-Cre mice (all from the Jackson Laboratory. Bar Harbor, ME). Ces1 −/− mice were generated by replacing exon 1 with an ACN-LacZ cassette as described 33 . Detailed characterization of the Ces1 −/− mice will be reported elsewhere. Ces1 −/− mice were backcrossed with C57BL/6 mice for at least 6 generations prior to experimentation. MCD diets were purchased from Harlan Laboratories (Cat # TD.90262; Madison, WI). Lieber-DeCarli diets were purchased from Bio-Serv (Flemington, NJ). Human liver samples with alcoholic steatohepatitis (with fibrosis) were obtained from the Liver Tissue Cell Distribution System at University of Minnesota. All the animal experiments were approved by the Institutional Animal Care and Use Committee at Northeast Ohio Medical University (NEOMED). The use of human tissue samples was approved by the Institutional Review Board at NEOMED. All the experiments were performed in accordance with the relevant guidelines and regulations.
Chronic plus binge alcohol drinking. C57BL/6J mice (12 weeks old) were administered a Lieber-DeCarli control liquid diet (Cat # F1259SP, BioServ) for 5 days. On the 6 th day, mice were fed a Lieber-DeCarli liquid diet (Cat # F1258SP, BioServ) containing 5% (vol/vol) ethanol or pair-fed a Lieber-DeCarli control liquid diet for 10 days 18 . On the 16 th day, mice were gavaged with a single dose of ethanol (3 or 5 g/kg body weight) or isocaloric maltose dextrin. For some studies, on the 6 th day mice were also injected i.v. with either Ad-shLacZ or Ad-shCes1. These adenoviruses have been described previously 12 . RNA isolation and quantitative real-time PCR. Total RNA was isolated using TRIzol Reagent (Life Technologies, NY). 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). Relative mRNA levels were calculated using the comparative cycle threshold (Ct) method and were normalized to the values of 36B4 mRNA levels.

Primary hepatocyte isolation.
Mouse primary hepatocytes were isolated as described 36,37 . Mice were anaesthetized by intraperitoneal injection of 50 mg/kg pentobarbital. The portal vein was cannulated with a 23-gauge plastic cannula. Mouse livers were perfused with Hank's Balanced Salt Solution (HBSS, Cat # 14170-112, Thermo Fisher Scientific) containing 0.19 g/L EDTA. Simultaneously, the inferior vena cava was cut open. Subsequently, livers were perfused with HBSS, calcium, magnesium buffer (Cat # 14025092, Thermo Fisher Scientific) with 0.8 mg/mL Collagenase from Clostridium histolyticum type IV (Sigma, St. Louis, MO). Primary hepatocytes were released and collected in a 50 mL centrifuge tube. After centrifugation at 50 × g for 5 minutes and washing with DMEM, cells were cultured in DMEM/10%FBS (Atlanta Biologicals, Georgia, USA) in 6-well plates pre-coated with 0.1% gelatin (Sigma-Aldrich).
Measurement of hepatic acetaldehyde level using high performance liquid chromatography (HPLC). Hepatic acetaldehyde level was determined using derivatization with DNPH, followed by HPLC separation as described previously 38 . 100 mg liver tissue was homogenized in 3 M perchloric acid (Fisher Scientific). The pH of the solution was adjusted immediately to 4.0 using 2 volumes 3 M sodium acetate buffer (pH 9.0) (Fisher Scientific). After centrifugation at 12000 rpm at 4 o C for 20 minutes, the supernatant was transferred into an ice-cold tube, followed by addition of 80-fold molar excess 2,4-dinitrophenylhydrazine (DNPH, Cat # 119266, Sigma-Aldrich) in 6 N HCl. The mixture was then placed on a shaker for 1 h at room temperature. Derivatization was stopped with 3 volumes of 3M sodium acetate buffer (pH 9.0). Two volumes of acetonitrile (Sigma-Aldrich) were added to extract AcH-DNP. After centrifugation at 10000 × g for 5 minutes at 4 o C, the organic phase was condensed to 50 μL. The ultra high performance liquid chromatograph (UHPLC) machine was purchased from Shimadzu Corp. (Columbia, MD). A Restek C18 HPLC column (25 cm × 4.6 mm i.d. 5 μm) coupled with an Ultra C18 guard column (10 mm × 4 mm i.d.) were purchased from Fisher Scientific (Pittsburgh, PA). The elution program was described previously 38 . An AcH-DNP standard (Sigma-Aldrich, St.Louis. MO) was used to create a standard curve. The values of the area under the curve of the AcH-DNP peaks were determined to calculate the concentration of acetaldehyde in each sample.
Malondialdehyde (MDA) assay. MDA assay was performed as described 35 . Buffer I was made by dissolving 2-Thiobarbituric acid (Sigma-Aldrich) in 10% perchloric acid to a final concentration 0.67%. Assay buffer was made by adding 20% trichloroacetic acid to buffer I (2:3,v/v). 1,1,3,3,-tetraethoxypropane (Sigma-Aldrich) was used as a standard. 30 mg liver tissue was homogenized in 500 μL saline. After centrifugation, 100 μL supernatant was collected and added to 1 ml assay buffer. The reaction mixture was incubated at 95 °C for 30 min, cooled down and centrifuged at 3000 rpm for 10 min. The absorbance was measured at OD532 nm.
Mitochondrial H 2 O 2 was detected using Amplex Red assay as described 39 . The working solution was prepared using 100 μL of 10 mM Amplex Red reagent (Thermo Fisher Scientific), 2 μL of 1000 U/ml horseradish peroxidase (HRP) and 10 ml of 50 mM potassium phosphate (pH7.7) with 0.5 mM diethylenetriaminepentaacetic acid (Sigma-Aldrich). 50 μL of sample or standard was pipetted into 96 plates, followed by addition of 50 μl of the Amplex Red reagent/HPR working solution. The reaction mixture was protected from light and incubated at room temperature for 30 min. The fluorescence was measured with excitation in the range of 530-560 nm and emission at 590 nm. Mitochondrial H 2 O 2 levels were normalized to protein levels.
Lipid and body fat analysis. Plasma triglycerides and cholesterol were measured using Infinity reagents from Thermo Scientific (Waltham, MA). Plasma and hepatic free fatty acids were measured using a kit from Wako Diagnostics (Richmond, VA). To measure lipids in liver, approximately 100 mg liver tissue was homogenized in methanol and extracted in chloroform/methanol (2:1 v/v). Hepatic triglyceride and cholesterol levels were then quantified. Body fat content was determined by EchoMRI (Houston, TX).

Plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) analysis.
Plasma ALT and AST levels were determined using Infinity reagents (Middletown, VA) following the manufacture's instructions.
Statistical Analysis. The data were analyzed using unpaired Student t test and ANOVA (GraphPad Prisim, CA). All values were expressed as mean ± SEM. Differences were considered statistically significant at P < 0.05.