Hepatic SIRT1 Attenuates Hepatic Steatosis and Controls Energy Balance in Mice by Inducing Fibroblast Growth Factor 21

Gastroenterology 2014;146:539–549

BASIC AND TRANSLATIONAL—LIVER Hepatic SIRT1 Attenuates Hepatic Steatosis and Controls Energy Balance in Mice by Inducing Fibroblast Growth Factor 21 Yu Li,1,§ Kimberly Wong,1 Amber Giles,1 Jianwei Jiang,1 Jong Woo Lee,1 Andrew C. Adams,2 Alexei Kharitonenkov,2 Qin Yang,3 Bin Gao,4 Leonard Guarente,5 and Mengwei Zang1,§ 1

Department of Medicine, Boston University School of Medicine, Boston, Massachusetts; 2Lilly Research Laboratories, Lilly Corporate Center, Indianapolis, Indiana; 3Division of Endocrinology, Diabetes and Metabolism, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; 4Laboratory of Liver Diseases, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, Maryland; 5Department of Biology, Paul F. Glenn Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts

Keywords: Liver-Specific Disruption of Sirt1; HepatocyteDerived Hormone; Metabolic Homeostasis; Obesity.

D

eregulation of lipid metabolic homeostasis is a common characteristic of metabolic disorders such as fatty liver disease, obesity, and diabetes. The liver functions as a major metabolic buffering system for metabolic homeostasis, allowing extrahepatic tissues such as the brain and heart to function normally under nutrient stress and deprivation.1 An important element in the transcriptional response to nutrient deprivation is the nicotinamide adenine dinucleotide (NAD)-dependent deacetylase SIRT1, which tightly regulates fatty acid metabolism through multiple nutrient sensors such as AMP-activated protein kinase (AMPK), sterol regulatory element binding protein-1 (SREBP-1), and peroxisome proliferator-activated receptor-gamma coactivator 1 a ( PGC-1a).2–4 Polyphenolic SIRT1 activators, including resveratrol and the synthetic polyphenol S17834, prevent hepatic steatosis and hyperlipidemia in mice with type 1 and type 2 diabetes.5,6 Hepatic overexpression of SIRT1 ameliorates hepatic steatosis and glucose intolerance in obese mice.7 However, the underlying mechanism of hepatic SIRT1 actions remains incompletely understood. In our effort to identify a novel downstream regulator of SIRT1, gene expression chip assays show fibroblast growth factor 21 (FGF21), the hepatocyte-derived

§

Author’s share co-corresponding authorship.

Abbreviations used in this paper: Ad, adenovirus; BAT, brown adipose tissue; CPT1a, carnitine palmitoyltransferase 1a; FAS, fatty acid synthase; FGF21, fibroblast growth factor 21; MCAD, medium-chain acyl-CoA dehydrogenase; MEF, mouse embryonic fibroblast; mRNA, messenger RNA; NAD, nicotinamide adenine dinucleotide; PPARa, peroxisome proliferator activated receptora; SIRT1, NAD-dependent protein deacetylase sirtuin-1; SIRT1 LKO mice, liver-specific SIRT1 knockout mice; SREBP-1, sterol regulatory element binding protein-1; VCO2, carbon dioxide production; VO2, oxygen consumption; WAT, white adipose tissue; WT, wild-type. © 2014 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2013.10.059

BASIC AND TRANSLATIONAL LIVER

BACKGROUND & AIMS: The hepatocyte-derived hormone fibroblast growth factor 21 (FGF21) is a hormone-like regulator of metabolism. The nicotinamide adenine dinucleotide–dependent deacetylase SIRT1 regulates fatty acid metabolism through multiple nutrient sensors. Hepatic overexpression of SIRT1 reduces steatosis and glucose intolerance in obese mice. We investigated mechanisms by which SIRT1 controls hepatic steatosis in mice. METHODS: Liver-specific SIRT1 knockout (SIRT1 LKO) mice and their wild-type littermates (controls) were divided into groups that were placed on a normal chow diet, fasted for 24 hours, or fasted for 24 hours and then fed for 6 hours. Liver tissues were collected and analyzed by histologic examination, gene expression profiling, and real-time polymerase chain reaction assays. Human HepG2 cells were incubated with pharmacologic activators of SIRT1 (resveratrol or SRT1720) and mitochondrion oxidation consumption rate and immunoblot analyses were performed. FGF21 was overexpressed in SIRT1 LKO mice using an adenoviral vector. Energy expenditure was assessed by indirect calorimetry. RESULTS: Prolonged fasting induced lipid deposition in livers of control mice, but severe hepatic steatosis in SIRT1 LKO mice. Gene expression analysis showed that fasting upregulated FGF21 in livers of control mice but not in SIRT1 LKO mice. Decreased hepatic and circulating levels of FGF21 in fasted SIRT1 LKO mice were associated with reduced hepatic expression of genes involved in fatty acid oxidation and ketogenesis, and increased expression of genes that control lipogenesis, compared with fasted control mice. Resveratrol or SRT1720 each increased the transcriptional activity of the FGF21 promoter (–2070/þ117) and levels of FGF21 messenger RNA and protein in HepG2 cells. Surprisingly, SIRT1 LKO mice developed late-onset obesity with impaired whole-body energy expenditure. Hepatic overexpression of FGF21 in SIRT1 LKO mice increased the expression of genes that regulate fatty acid oxidation, decreased fastinginduced steatosis, reduced obesity, increased energy expenditure, and promoted browning of white adipose tissue. CONCLUSIONS: SIRT1-mediated activation of FGF21 prevents liver steatosis caused by fasting. This hepatocyte-derived endocrine signaling appears to regulate expression of genes that control a brown fat-like program in white adipose tissue, energy expenditure, and adiposity. Strategies to activate SIRT1 or FGF21 could be used to treat fatty liver disease and obesity.

540

Li et al

hormone, as the most markedly down-regulated gene in liver-specific SIRT1 knockout (SIRT1 LKO) mice. FGF21, a fasting-induced hepatokine, rapidly is gaining interest as a metabolic regulator,8,9 although the mechanism of the nutrient regulation of FGF21 is unclear. FGF21 is expressed predominantly in the liver, adipose tissue, and pancreas, with most circulating FGF21 originating from the liver.10 The autocrine/paracrine and endocrine actions of FGF21 hormone are mediated through FGF receptors complexed with b-klotho.11 Both pharmacologic administration of FGF21 and transgenic overexpression of FGF21 in mice protect against body weight gain and metabolic dysfunction in diabetic rodents, monkeys, and human beings.8,12–16 Previous studies have reported that SIRT1 regulates FGF21 expression in hepatocytes in vitro and in vivo.17,18 However, the relative contribution of FGF21 to SIRT1’s effects on overall energy metabolism has not been investigated. The present study characterizes FGF21 as a critical downstream regulator of SIRT1 that protects against hepatic steatosis, enhances expression of brown fat-like genes in white adipose tissue, and increases whole-body energy expenditure. Our in vivo and in vitro studies illustrate the following: (1) hepatic SIRT1 is required for fasting-induced production and secretion of FGF21 in the liver; (2) defective FGF21 caused by hepatic SIRT1 ablation exacerbates fasting-induced hepatic steatosis by impairing fatty acid oxidation and increasing lipogenesis; (3) FGF21 is essential for SIRT1 to stimulate hepatocyte fatty acid oxidation; and (4) hepatic overexpression of FGF21 enhances systemic energy expenditure and ameliorates obesity.

Materials and Methods BASIC AND TRANSLATIONAL LIVER

Animals Hepatocyte-specific deletion of the SIRT1 gene in mice (SIRT1 LKO) was achieved by crossing albumin-Cre recombinase transgenic mice with floxed SIRT1Dex4 mice containing the deleted SIRT1 exon 4, which encodes 51 amino acids of the conserved SIRT1 catalytic domain, as described previously.19 The protocol for this study was approved by the Boston University Medical Center Institutional Animal Care and Use Committee.

Animal Fasting and Refeeding Experiments SIRT1 LKO mice and their WT littermates were divided into 3 groups: fed, fasted, and re-fed. The fed group was placed on a normal chow diet, the fasted group was fasted for 24 hours, and the re-fed group was fasted for 24 hours and then fed for 6 hours.

In Vivo Adenoviral Gene Transfer The adenovirus producing full-length FGF21 was generated and purified as described previously.6,20 Adenovirus-mediated overexpression of FGF21 in vivo was accomplished via tail vein injection as described previously.2,7,20

Statistical Analysis Values are expressed as the mean  SEM. Statistical significance was evaluated using an unpaired 2-tailed t test or a 1-way analysis of variance (ANOVA) for more than 2 groups.

Gastroenterology Vol. 146, No. 2 Differences were considered significant at a P level of less than .05.

Results Hepatocyte-Specific Deletion of SIRT1 in Mice Increases the Susceptibility to Fasting-Induced Fatty Liver To explore whether liver-tissue–specific SIRT1 is a driving force in maintaining lipid homeostasis, metabolic phenotypes of hepatic SIRT1 LKO mice were characterized. As shown in Figure 1, blood glucose and plasma insulin levels appeared comparable between wild-type (WT) and SIRT1 LKO mice in both fed and fasted states. In the fed state, plasma cholesterol and triglyceride levels were slightly higher in SIRT1 LKO mice compared with those in the WT littermates. In the fasted state, plasma lipids were similar between WT and SIRT1 LKO mice. In liver, prolonged fasting induced lipid deposition in WT mice, as shown by Oil Red O staining and direct triglyceride measurements. Strikingly, pronounced hepatic steatosis was observed in fasted SIRT1 LKO mice, as evidenced by increased stained areas and increased hepatic triglyceride content. Hepatic cholesterol content was comparable among groups. These data indicate that hepatic SIRT1 may play a critical role in regulating lipid metabolism under fasting conditions.

Loss of SIRT1 in the Liver Suppresses Fasting-Induced Gene Expression, Production, and Secretion of Hepatic FGF21 As part of a directed screen to identify a novel secreted protein that could be modulated by genetic manipulation of hepatic SIRT1, gene expression chip analysis showed that FGF21 was reduced by 83% in livers of SIRT1 LKO mice compared with those of WT mice under feeding conditions. Prolonged fasting resulted in an 11.7-fold induction of FGF21 in livers of WT mice. Hepatic SIRT1 deletion ablated approximately 60% of the FGF21 induction under fasting conditions (Supplementary Table 2). Peroxisome proliferator activated receptor (PPAR)a gene expression was increased approximately 2-fold in WT mice in response to fasting and decreased 40% in SIRT1 LKO mice. However, gene expression profiles of proinflammatory mediators, cytokines, growth factors, oxidative stress, and redox modulators appeared comparable between both genotypes of mice under nutrient stress conditions (Supplementary Table 2). We next determined whether hepatic SIRT1 deficiency might interfere with fasting-induced production and release of hepatic FGF21. As shown in Figure 2, protein expression of hepatic SIRT1 was induced by nutrient deprivation and repressed by nutrient availability. Similar to the gene chip data (Supplementary Table 2), messenger RNA (mRNA) levels of FGF21 were increased robustly 15-fold upon fasting and decreased nearly 50% upon refeeding as observed previously.9 Consistently, FGF21 protein production was increased 5-fold in WT mice after a 24-hour fast, and this induction was decreased approximately 40% in SIRT1 LKO

Regulation of FGF21 by SIRT1 in the Liver

541

Figure 1. Hepatocyte-specific deletion of SIRT1 in mice increases susceptibility to developing fastinginduced fatty liver. (A) Representative immunoblots for SIRT1 in 2 mouse livers from each group are shown. The expected deletion mutant protein in SIRT1 LKO mice, which migrates slightly faster than WT SIRT1, was visualized by immunoblotting. (B) Blood glucose and plasma insulin levels in WT and SIRT1 LKO mice at 6 and 7 months of age that were fed a normal diet (Fed) or fasted for 24 hours (Fasted) (n ¼ 10–18). (C) Plasma triglyceride and cholesterol levels in mice (n ¼ 4–6). (D and E) Hepatic steatosis was assessed by Oil Red O staining and quantified by measuring Oil Red O–stained areas. (F) Hepatic triglyceride and cholesterol levels in WT and SIRT1 LKO mice (n ¼ 4–6). Results are presented as the mean  SEM. *P < .05 vs the fed WT mice; #P < .05 vs the fasted WT mice.

mice. These results suggest that hepatic SIRT1 and FGF21 both are regulated physiologically by a feeding-fastingrefeeding cycle, and that SIRT1 is a major regulator of fasting-inducible FGF21. Importantly, plasma FGF21 levels similarly were increased 8-fold in WT mice in response to fasting, and this induction was suppressed approximately 40% in SIRT1 LKO mice. In addition, moderate but consistent reduction of FGF21 production and secretion also was seen in the fed SIRT1 LKO mice. The effect of refeeding on FGF21 was comparable between both genotypes. Therefore, hepatic SIRT1 is necessary for prolonged fasting-induced gene transcription, protein production, and secretion of FGF21.

Loss of SIRT1 in the Liver Alters Expression of Key Genes Involved in Lipid Metabolism It has been shown that FGF21 regulates hepatic fatty acid oxidation and ketogenesis.9,21 To determine the functional consequence of defective FGF21 in SIRT1 LKO mice, we measured the expression of key genes involved in lipid metabolism. mRNA amounts of rate-limiting enzymes of

fatty acid oxidation, carnitine palmitoyl transferase 1a (CPT1a), and medium-chain acyl-CoA dehydrogenase (MCAD) were increased approximately 9- and 5-fold in fasted WT livers, respectively, but up-regulated only 4- and 2.5-fold in fasted SIRT1 LKO livers, respectively (Figure 3A); these changes were shown previously in hepatic FGF21 knockdown mice fed a ketogenic diet.9 The results suggest that induction of hepatic FGF21 and b-oxidation during fasting largely is diminished in SIRT1 LKO mice, a metabolic defect that could explain the pronounced hepatic steatosis found in SIRT1 LKO mice. Consistent with the concomitant induction of hepatic SIRT1 and FGF21 by nutrient deprivation (Figure 2), mRNA levels of several enzymes required for ketone body synthesis, including acetyl-CoA acetyltransferase 1 (ACAT1), 3-hydroxy3-methylglutaryl-CoA synthase 2 (HMGCS2), 3-hydroxy3-methylglutaryl–CoA lyase (HMGCL), and b-hydroxybutyrate dehydrogenase 1 (BDH1), were increased by 4- to 8-fold in WT mice under fasting conditions. Consequently, plasma b-hydroxybutyrate levels were increased 8-fold by fasting and fully suppressed by refeeding. Strikingly, fasting-induced adaptive ketogenesis was blunted markedly in SIRT1 LKO

BASIC AND TRANSLATIONAL LIVER

February 2014

542

Li et al

Gastroenterology Vol. 146, No. 2

encoding SREBP-1c, FAS, and stearoyl-CoA desaturase 1 (SCD1) were up-regulated more than 2-fold in fasted SIRT1 LKO mice (Figure 3E). Taken together, hepatic steatosis in SIRT1 LKO mice with triglyceride accumulation likely is attributable to the reduction of fatty acid oxidation and the stimulation of fatty acid synthesis.

FGF21 Is Essential for SIRT1 to Regulate Hepatocyte Fatty Acid Oxidation

BASIC AND TRANSLATIONAL LIVER

Figure 2. Production and secretion of hepatic FGF21 are reduced in SIRT1 LKO mice in response to nutrient deprivation. SIRT1 LKO mice and their WT littermates at 6 and 7 months of age (n ¼ 4–7) were subjected to feeding (Fed), fasting for 24 hours (Fasted), or refeeding for 6 hours after a 24-hour fast (Refed). (A) Hepatic SIRT1 expression was increased by fasting and decreased by refeeding in WT mice, and the induction was eliminated in SIRT1 LKO mice (n ¼ 4–7). (B and C) Fastinginduced gene expression and protein production of hepatic FGF21 were diminished in SIRT1 LKO mice (n ¼ 4–7), as determined by real-time polymerase chain reaction and enzyme-linked immunosorbent assays (ELISA). (D) The circulating levels of FGF21 were increased robustly by fasting in WT mice, and the induction was impaired in SIRT1 LKO mice. *P < .05 vs the fed WT mice; #P < .05 vs the fasted WT mice.

mice (Figure 3B–D). Collectively, hepatic SIRT1 deficiency inhibits hepatic fatty acid use and energy supply to extrahepatic tissues, likely via suppression of the FGF21-regulated processes of b-oxidation and ketogenesis. Our previous studies showed that SIRT1 suppresses high glucose-induced fatty acid synthase (FAS) expression and lipid accumulation in HepG2 cells.2 To further delineate the mechanisms responsible for the steatotic phenotype in SIRT1 LKO mice, gene expression of a key lipogenic transcription factor, SREBP-1c, and its target genes, were assessed by real-time polymerase chain reaction. mRNAs

To investigate a causal relationship between SIRT1 and FGF21 in a cell-autonomous manner, human HepG2 cells were used to investigate the transcriptional regulation of FGF21 by SIRT1 because this cell line expresses both genes2,20 and because they respond well to SIRT1 activators to decrease hepatocellular lipids.2 As shown in Figure 4A–D, luciferase-based reporter assays with a human FGF21 promoter (2090 to þ117)20 showed that FGF21 promoterdriven reporter activity was stimulated progressively by increasing the expression of SIRT1. FGF21 mRNA levels were increased dose-dependently by SIRT1 overexpression in HepG2 cells, consistent with the previous observation in primary hepatocytes.17 To determine the functional consequence of SIRT1-induced FGF21 on fatty acid oxidation, immunoblotting analysis showed that protein production of FGF21 was induced more than 2-fold by SIRT1 activators including resveratrol (10 mmol/L) and SRT1720 (5 mmol/L), an effect that was well correlated with their ability to increase the gene expression of CPT1a. In parallel, cellular oxygen consumption rates, an index of mitochondrial oxidation function and a readout of energy expenditure in vitro,22 were measured using an extracellular flux analyzer of live cells in real time. Compared with those of SIRT1þ/þ mouse embryonic fibroblasts (MEFs), the bioenergetic profiles of SIRT1-/- MEFs showed a decrease in basal and maximal mitochondrial respiratory capacity, consistent with the repressed CPT1a and MCAD expression found in SIRT1 LKO livers (Figure 3A) and decreased rates of fatty acid oxidation seen in SIRT1-/- MEFs.3 Conversely, basal respiration was increased moderately but significantly by SRT1720, and maximal respiration caused by the uncoupling agent, carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) was stimulated further more than 2-fold by SRT1720 (5 mmol/L) in HepG2 cells, an effect also seen with resveratrol (10 mmol/L). Notably, induction of CPT1a by SIRT1 activators was correlated closely with their ability to increase oxygen consumption rates. Taken together, our data show that SIRT1 potently induces FGF21 and fatty acid oxidation in vitro. To investigate whether the autocrine/paracrine effect of FGF21 is responsible for the up-regulation of fatty acid oxidation by SIRT1, we found that expression of CPT1a was increased approximately 2-fold by resveratrol, but such an induction largely was diminished by small interfering RNAmediated knockdown of FGF21 (Figure 4E). By using a competitive antagonist of FGF21, an N-terminally truncated FGF21 protein (FGF21DN17) that previously was described to block in vivo FGF21 signaling,14 we observed that the stimulatory effect of resveratrol on CPT1a was blocked significantly by the FGF21DN17 mutant (Figure 4F). These

Regulation of FGF21 by SIRT1 in the Liver

543

Figure 3. Expression of metabolic genes involving fatty acid oxidation, ketogenesis, and lipogenesis is altered in livers of SIRT1 LKO mice. (A) Hepatic expression of key genes of fatty acid oxidation such as CPT1a and MCAD was induced by fasting in WT mice and inhibited in SIRT1 LKO mice (n ¼ 4–7). (B and C) Fasting-induced expression of ketogenic enzymes including acetyl-CoA acetyltransferase 1 (ACAT1), 3-hydroxy-3-methylglutaryl–CoA synthase 2 (HMGCS2), 3-hydroxy-3methylglutaryl-CoA lyase (HMGCL), and b-hydroxybutyrate dehydrogenase 1 (BDH1) was suppressed in livers of SIRT1 LKO mice. (D) Effects of fasting and refeeding on circulating ketone bodies in mice (n ¼ 5–10). (E) mRNA amounts of SREBP-1c and key lipogenic enzymes including FAS and stearoyl-CoA desaturase (SCD)1 were up-regulated in livers of SIRT1 LKO mice (n ¼ 4–7). *P < .05 vs the fed WT mice; #P < .05 vs the fasted WT mice.

studies of the genetic and pharmacologic manipulation of SIRT1 or FGF21 define FGF21 as a novel downstream mediator of SIRT1 to stimulate hepatocyte b-oxidation.

Hepatic Overexpression of FGF21 Ameliorates Hepatic Steatosis in SIRT1 LKO Mice Given that fasting-induced fatty liver is associated with FGF21 insufficiency in SIRT1 LKO mice (Figure 2), rescue experiments with adenovirus encoding FGF21 (Ad-FGF21) were performed to determine the effect of FGF21 gain-offunction on the steatotic phenotype of SIRT1 LKO mice. As shown in Figure 5, in vivo adenoviral gene transfer of FGF21 into SIRT1 LKO mice was accomplished successfully and shown by robustly increased hepatic production and circulating levels of FGF21 in mice at 2 weeks postinjection. Fasting-induced fatty liver in SIRT1 LKO mice was attenuated by FGF21 overexpression, as reflected by reductions in Oil Red O–stained areas and hepatic triglyceride content. Plasma triglyceride and cholesterol levels were not altered

(Figure 5D). Remarkably, reduced fatty acid oxidation genes (CPT1a and MCAD) in the fasted SIRT1 LKO mice were normalized by reconstitution of FGF21, and the impairment of the transcription of ketogenic genes (HMGCS2 and HMGCL) in SIRT1 LKO mice also was restored. Consequently, the reduction of plasma b–hydroxybutyrate concentrations in fasted SIRT1 LKO mice was counteracted by FGF21 overexpression, reaching nearly normal levels in control green fluorescent protein (GFP)–expressing mice. Ad-FGF21 treatment did not alter SREBP1c and FAS expression in livers of SIRT1 LKO mice (Supplementary Figure 1). Therefore, it appears to suggest that hepatic SIRT1 deficiency exacerbates fasting-induced hepatic steatosis largely through hepatic impairment of FGF21 and fatty acid oxidation.

SIRT1 LKO Mice Develop Late-Onset Obesity With Decreased Whole-Body Energy Expenditure FGF21 has emerged as an important regulator of energy expenditure and body weight.12,16 Circulating FGF21 was

BASIC AND TRANSLATIONAL LIVER

February 2014

544

Li et al

Gastroenterology Vol. 146, No. 2

BASIC AND TRANSLATIONAL LIVER

Figure 4. SIRT1 stimulates mitochondrial fatty acid oxidation in an FGF21-dependent manner in HepG2 cells. (A) Increased expression of SIRT1 progressively stimulated transcriptional activity of the FGF21 (-2090/þ117) promoter-driven reporter and increased FGF21 gene expression. Overexpression of FLAG-tagged mouse SIRT1 (FLAG-mSIRT1) in HEK293T cells was confirmed by immunoblots. (B) Protein production of FGF21 and mRNA amounts of CPT1a were stimulated by SIRT1 activators resveratrol (10 mmol/L) and SRT1720 (5 mmol/L). (C) Oxygen consumption rates were determined by analyzing a bioenergetics profile of SIRT1þ/þ and SIRT1-/- MEFs using the Seahorse Bioscience (North Billerica, MA). The oxygen consumption rates of basal respiration, adenosine triphosphate (ATP) turnover, and respiratory capacity were measured and calculated as averages for each phase. (D) The mitochondrial oxygen consumption rate was increased by SIRT1 activators. (E) The ability of resveratrol (10 mmol/L) to induce CPT1a was abrogated by small interfering RNA (siRNA)-mediated FGF21 knockdown in HepG2 cells. (F) Resveratrolinduced CPT1a expression was diminished by a competitive inhibitor of FGF21 (FGF21DN17, 100 nmol/L) in HepG2 cells (n ¼ 3–4). *P < .05 vs control cells; #P < .05 vs treatment cells. pcDNA, an empty pcDNA plasmid as a negative control.

reduced in SIRT1 LKO mice (Figure 2D), therefore we sought to determine whether hepatocyte-specific deletion of SIRT1 would affect energy expenditure and adiposity. As shown in Figure 6A–C, compared with WT littermates fed a standard chow diet, SIRT1 LKO mice were more susceptible to late-onset body weight gain, as evidenced by a moderate but statistically significant increase in body weight. SIRT1 LKO mice appeared to have increased adiposity with more

visible fat, possibly owing to increased adipocyte cell size. Body composition analysis showed that SIRT1 LKO mice had increased fat mass without significantly affecting lean mass. To understand the physiological mechanisms by which SIRT1 LKO mice are more susceptible to developing obesity, the major components of energy metabolism such as food intake, body temperature, energy expenditure, and physical activity were examined. Daily food intake was similar

Regulation of FGF21 by SIRT1 in the Liver

545

Figure 5. Hepatic overexpression of FGF21 rescues hepatic steatosis and metabolic defects in fasted SIRT1 LKO mice. (A) An adenovirus vector encoding FGF21 (Ad-FGF21) was overexpressed in HEK293A cells. (B) Adenoviral overexpression of FGF21 in SIRT1 LKO mice (n ¼ 4–7) was shown by increased hepatic gene expression, protein production, and circulating levels of FGF21. (C and D) Fasting-induced hepatic steatosis in SIRT1 LKO mice was ameliorated by hepatic overexpression of FGF21, as indicated by decreased Oil Red O–stained areas and reduced hepatic triglyceride levels. (E and F) Suppression of hepatic fatty acid oxidation and ketogenesis in SIRT1 LKO mice was prevented by hepatic overexpression of FGF21. *P < .05 vs Ad-GFP–injected WT mice; #P < .05 vs Ad-GFP–injected SIRT1 LKO mice. HMGCS2, 3hydroxy-3-methylglutaryl– CoA synthase 2; HMGCL, 3-hydroxy-3-methylglutar yl–CoA lyase; BDH1, bhydroxybutyrate dehydr ogenase 1.

between WT and SIRT1 LKO mice on a regular chow diet. No significant difference in body temperature was noted at 10 AM and 5 PM (Figure 6D). Comprehensive metabolic cage studies indicated that the rates of oxygen consumption (VO2) and carbon dioxide production (VCO2) were reduced in SIRT1 LKO mice under feeding conditions. The calculated energy expenditure was reduced by 20% and 23% in SIRT1 LKO mice during the light and dark phases, respectively. Similarly, the rates of VO2, VCO2, and energy expenditure were much lower in SIRT1 LKO mice than those in WT mice during a 24-hour fast through light and dark cycles, although metabolic rates were decreased markedly in both genotypes of mice in the fasting state. Moreover, the respiratory quotients (respiratory quotient ¼ VCO2/VO2) appeared indistinguishable in WT and SIRT1 LKO mice in fed and fasted states, likely owing to the combined decrease in both VO2 and VCO2. This may reflect the relatively equal use of carbohydrates vs lipids as an energy source in whole body. Circadian locomotor activity, as measured by ambulatory movement over a 12hour block of light and dark phases, was similar between the 2 genotypes of mice (Figure 6E and F). Collectively, the

obesity-prone phenotypes of SIRT1 LKO mice were characterized by an increase in fat mass, a reduction in energy expenditure, but no significant alterations in lean mass, food intake, body temperature, or physical activity.

Adenoviral Delivery of FGF21 Into SIRT1 LKO Mice Increases Energy Expenditure and Promotes the Browning of White Adipose Tissue To directly assess whether suppression of FGF21 is responsible for energy disturbances in SIRT1 LKO mice, indirect calorimetry was examined in a cohort of SIRT1 LKO mice at 2 weeks after injection of either adenovirus encoding green fluorescent protein (Ad-GFP) or Ad-FGF21. As shown in Figure 7A–C, hepatic overexpression of FGF21 caused a 15% and 10% increase in Vo2 rates in SIRT1 LKO mice under feeding conditions throughout the light and dark cycles. Hepatic overexpression of FGF21 normalized the decreased energy expenditure in SIRT1 LKO mice without significantly affecting respiratory quotient or locomotor activity. Reactivation of FGF21 decreased body weight in SIRT1 LKO mice

BASIC AND TRANSLATIONAL LIVER

February 2014

546

Li et al

Gastroenterology Vol. 146, No. 2

BASIC AND TRANSLATIONAL LIVER

Figure 6. SIRT1 LKO mice are susceptible to developing obesity with increased adiposity and decreased total oxygen consumption and energy expenditure. (A) Chow-fed SIRT1 LKO mice at 6 and 7 months of age (n ¼ 8) gained more body weight. Body weight curves of WT and SIRT1 LKO mice over a 15-month period (n ¼ 8–16). (B) A representative photograph of increased adiposity and adipocyte size in WAT of SIRT1 LKO mice. (C) Body composition analysis of WT and SIRT1 LKO mice. (D) Food intake (n ¼ 10–12) and body temperature (n ¼ 6) were similar between WT and SIRT1 LKO mice. (E) The rates of VO2, VCO2, and energy expenditure were measured by comprehensive metabolic monitoring in mice (n ¼ 4) over a 24-hour period with food and over a 24-hour fast and normalized to lean body mass. Upper panels: circadian changes in metabolic parameters in mice in the fed state during light and dark cycles, as indicated by the white and gray areas. Lower panels: metabolic rates in mice in fed and fasted states. (F) Respiratory quotient and locomotor activity were determined in metabolic cages. The physical activity is expressed as total counts measured by summing X and Y beam breaks, and the bar graph represents average locomotor activity over a 12-hour block of light and dark phases. *P < .05 vs WT mice.

with an approximately 10% reduction in fat mass and no significant change in lean mass. These data indicate that FGF21 may play an important role in mediating the effect of hepatic SIRT1 on energy expenditure and adiposity. Because FGF21 has emerged as a regulator of brown-fatlike features in white adipose tissue (WAT),23 we further investigated whether enhanced energy expenditure by FGF21 treatment in SIRT1 LKO mice could be attributed to altered transcription of key genes related to the browning program in WAT or thermogenic genes in brown adipose tissue (BAT). Strikingly, mRNA amounts of genes that govern the browning of WAT, such as uncoupling protein 1 (UCP1) and the BAT program coactivator PRD1-BF1-RIZ1 homologous domain containing 16 (PRDM16), as well as expression

of other BAT markers, type II iodothyronine deiodinase (DIO2), cell death activator CIDE-A (Cidea), and cytochrome oxidase subunit VIIa polypeptide 1 (Cox7a1), were increased 2- to 6-fold in WAT of SIRT1 LKO mice injected with AdFGF21, suggesting that increased circulating FGF21 in vivo promotes the development of brown fat-like characteristics in WAT (Figure 7D). Notably, mRNA amounts of other white adipose genes such as adiponectin, but not leptin, also were increased (Figure 7E). Ad-FGF21 treatment appeared to enhance BAT activity because expression of PRDM16 and Cidea in BAT was stimulated. Collectively, activation of SIRT1-FGF21 signaling represents an important liveradipose tissue axis that stimulates the browning of WAT and acts to maintain systemic energy balance.

Regulation of FGF21 by SIRT1 in the Liver

547

BASIC AND TRANSLATIONAL LIVER

February 2014

Figure 7. Hepatic overexpression of FGF21 rescues the deregulation of whole-body energy metabolism and promotes browning of WAT in SIRT1 LKO mice. (A) Circadian changes in VO2 were measured by indirect calorimetry over 48 hours in mice in fed and fasted states. (B) The rates of VO2 and energy expenditure were increased by hepatic overexpression of FGF21 in SIRT1 LKO mice in fed and fasted states through light and dark cycles (n ¼ 4). Respiratory quotient (RQ) and locomotor activity were similar between the 2 genotypes of mice. (C) Body weight and body composition in mice (n ¼ 4–7). (D and E) Adenoviral delivery of FGF21 increased the transcription of brown-fat-like genes and adiponectin in (D) WAT and (E) BAT of SIRT1 LKO mice. *P < .05 vs Ad-GFP–injected SIRT1 LKO mice. (F) The proposed model for nutrient regulation of FGF21 through SIRT1. Prolonged fasting increases the NAD/ form of NADH ratio and stimulates hepatic SIRT1, which in turn promotes expression, production, and release of hepatic FGF21. Induction of FGF21 by hepatic SIRT1 stimulates fatty acid oxidation and ketogenesis, enhances the metabolic adaptation to fasting, and improves fatty liver disease. Activation of the SIRT1-FGF21 cascade may represent a liver-adipose tissue axis that contributes to reduced adiposity and body weight by promoting white fat browning and increasing energy expenditure.

548

Li et al

Discussion The characterization of phenotypic changes of SIRT1 LKO mice, such as hepatic steatosis, increased adiposity, and reduced whole-body oxygen consumption, establishes SIRT1 as a novel upstream regulator of hormone FGF21 for the autocrine/paracrine modulation of hepatic fatty acid metabolism as well as the endocrine regulation of WAT browning and whole-body energy homeostasis. Hepatic SIRT1 deficiency results in diminished hepatic production and release of FGF21, impaired b-oxidation, stimulated lipogenesis, and reduced energy expenditure. The discovery of SIRT1-dependent induction of FGF21 also is supported by the fact that hepatic steatotic and obesity-prone phenotypes in SIRT1 LKO mice can be rescued by reactivation of FGF21. Figure 7F shows the novel function of FGF21 as a downstream mediator of local and systemic effects of hepatic SIRT1 on hepatic steatosis, white fat browning, energy homeostasis, and body weight control.

Regulation of FGF21 by SIRT1 in the Liver

BASIC AND TRANSLATIONAL LIVER

The most important implication of the present study is the identification of novel cross-talk between nutrient sensors SIRT1 and FGF21. Several lines of evidence have suggested similar properties of SIRT1 and FGF21. First, hepatic SIRT1 activity is induced by prolonged fasting via an increased NAD/ NADH ratio. Hepatic expression of FGF21 also is stimulated by fasting via PPARa.9 Second, SIRT1 gain-of-function in mice is protective against obesity-induced hepatic steatosis and glucose intolerance.18 Likewise, transgenic mice overexpressing FGF21 in the liver are resistant to diet-induced obesity and insulin resistance.8 Pharmacologic administration of FGF21 to ob/ob or db/db mice attenuates hepatic steatosis and enhances insulin sensitivity.8,12 Third, the metabolic defects seen in SIRT1 LKO mice are similar to those observed in FGF21 KO mice.21 Fasting-induced hepatic gene expression and protein secretion of FGF21 are dramatically repressed in SIRT1 LKO mice. Metabolic defects of the adaptive response to fasting seen in SIRT1 LKO mice are reconstituted by hepatic overexpression of FGF21. In further support of the notion that FGF21 is necessary for up-regulation of b-oxidation by SIRT1, our in vitro studies indicate that SIRT1 induces gene and protein expression of FGF21 in hepatocytes. Importantly, the ability of SIRT1 to stimulate CPT1a expression is abrogated by FGF21 knockdown or treatment with the FGF21DN17 mutant. Because FGF21 is identified as a target gene of nuclear receptors retinoic acid receptor b (RARb) and PPARa,9,20 it is possible that SIRT1 integrates with nuclear receptors for transcriptional regulation of FGF21 at multiple levels.

Ablation of SIRT1 in the Liver Increases the Propensity for Hepatic Steatosis and Late-Onset Obesity by Suppressing FGF21 Production One major finding of the present study is that pronounced hepatic steatosis is developed in SIRT1 LKO mice and can be attributed at least in part to suppression of FGF21, fatty acid oxidation, and ketogenesis, along with stimulation of lipogenic genes in liver. These results are consistent with the

Gastroenterology Vol. 146, No. 2

hepatic lipid accumulation seen in SIRT1 LKO mice challenged with high-fat diets.17 Even though FGF21, in the setting of fasting, is not required for the inhibition of SREBP-1 by SIRT1, our data could not rule out the possibility that FGF21, in the setting of hyperinsulinemia and obesity, may suppress SREBP-1. The importance of FGF21 in SIRT1 actions is supported further by the fact that gain-of-function of FGF21 protects against hepatic steatosis and obesityprone phenotypes in SIRT1 LKO mice by increasing hepatic fatty acid oxidation and whole-body energy expenditure. Because the alleles of 2 common SIRT1 variants are associated with a higher risk of being overweight or obese in human beings,24 our studies may provide mechanistic insight into a role of defective hepatic SIRT1 in triggering FGF21 impairment and energy metabolic perturbation in overweight individuals. Our studies support the rationale for targeting FGF21 as a novel therapy to treat human beings in obesity associated with SIRT1 dysfunction.

Regulation of the Browning of White Adipose Tissue by SIRT1-Dependent Induction of FGF21 The data presented here strongly suggest that induction of FGF21 by SIRT1 contributes to enhanced energy expenditure and weight loss, likely through white fat browning, because gain-of-function of FGF21 in SIRT1 LKO mice stimulates expression of browning-related genes such as UCP1 and PRDM16. In support of our findings, recent studies by the Shulman group also showed that FGF21 administration increases WAT browning and energy expenditure in highfat–fed mice.15 This underlying mechanism also may explain the increased energy expenditure seen in SIRT1 transgenic mice.18 Interestingly, increased expression of adiponectin in WAT is evident in mice injected with Ad-FGF21, which is consistent with recent studies that FGF21 stimulates adiponectin production in WAT and has a therapeutic impact on metabolic disorders.25 Promotion of white fat browning by targeting the SIRT1-FGF21 axis represents an attractive therapeutic approach to combat the obesity epidemic. In conclusion, the present study shows that the loss of hepatic SIRT1 reduces hepatic and circulating FGF21 levels and subsequently inhibits hepatic b-oxidation and energy expenditure. The convergence of these defects exacerbates hepatic steatosis and ultimately leads to increased weight gain. SIRT1 is an endogenous activator of FGF21 in hepatocytes that transduces fasting signals to the stimulation of hepatic b-oxidation and ketogenesis as well as systemically controls WAT browning and energy homeostasis. This SIRT1-FGF21 axis represents a novel hepatocyte-derived endocrine signaling pathway to potentially combat hepatic steatosis and obesity in human beings.

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http://dx.doi.org/10.1053/ j.gastro.2013.10.059.

References 1. Cahill GF. Fuel metabolism in starvation. Ann Rev Nutr 2006;26:1–22. 2. Hou X, Xu S, Maitland-Toolan KA, et al. SIRT1 regulates hepatocyte lipid metabolism through activating AMPactivated protein kinase. J Biol Chem 2008;283: 20015–20026. 3. Gerhart-Hines Z, Rodgers JT, Bare O, et al. Metabolic control of muscle mitochondrial function and fatty acid oxidation through SIRT1/PGC-1 alpha. EMBO J 2007; 26:1913–1923. 4. Ponugoti B, Kim DH, Xiao Z, et al. SIRT1 deacetylates and inhibits SREBP-1C activity in regulation of hepatic lipid metabolism. J Biol Chem 2010;285:33959–33970. 5. Zang MW, Xu SQ, Maitland-Toolan KA, et al. Polyphenols stimulate AMP-activated protein kinase, lower lipids, and inhibit accelerated atherosclerosis in diabetic LDL receptor-deficient mice. Diabetes 2006;55: 2180–2191. 6. Li Y, Xu SQ, Mihaylova MM, et al. AMPK phosphorylates and inhibits SREBP activity to attenuate hepatic steatosis and atherosclerosis in diet-induced insulin-resistant mice. Cell Metab 2011;13:376–388. 7. Li Y, Xu SQ, Giles A, et al. Hepatic overexpression of SIRT1 in mice attenuates endoplasmic reticulum stress and insulin resistance in the liver. FASEB J 2011; 25:1664–1679. 8. Kharitonenkov A, Shiyanova TL, Koester A, et al. FGF-21 as a novel metabolic regulator. J Clin Invest 2005; 115:1627–1635. 9. Badman MK, Pissios P, Kennedy AR, et al. Hepatic fibroblast growth factor 21 is regulated by PPAR alpha and is a key mediator of hepatic lipid metabolism in ketotic states. Cell Metab 2007;5:426–437. 10. Kharitonenkov A, Larsen P. FGF21 reloaded: challenges of a rapidly growing field. Trends Endocrinol Metab 2011; 22:81–86. 11. Kharitonenkov A, Dunbar JD, Bina HA, et al. FGF-21/ FGF-21 receptor interaction and activation is determined by beta Klotho. J Cell Physiol 2008;215:1–7. 12. Xu J, Lloyd DJ, Hale C, et al. Fibroblast growth factor 21 reverses hepatic steatosis, increases energy expenditure, and improves insulin sensitivity in diet-induced obese mice. Diabetes 2009;58:250–259. 13. Kharitonenkov A, Wroblewski VJ, Koester A, et al. The metabolic state of diabetic monkeys is regulated by fibroblast growth factor-21. Endocrinology 2007;148:774–781. 14. Adams AC, Coskun T, Irizarry Rovira AR, et al. Fundamentals of FGF19 & FGF21 action in vitro and in vivo. PLoS One 2012;7:e38438. 15. Camporez JP, Jornayvaz FR, Petersen MC, et al. Cellular mechanisms by which FGF21 improves insulin sensitivity in male mice. Endocrinology 2013;154:3099–3109.

Regulation of FGF21 by SIRT1 in the Liver

549

16. Gaich G, Chien JY, Fu H, et al. The effects of LY2405319, an FGF21 analog, in obese human subjects with type 2 diabetes. Cell Metab 2013;18:333–340. 17. Purushotham A, Schug TT, Xu Q, et al. Hepatocytespecific deletion of SIRT1 alters fatty acid metabolism and results in hepatic steatosis and inflammation. Cell Metab 2009;9:327–338. 18. Banks AS, Kon N, Knight C, et al. SirT1 gain of function increases energy efficiency and prevents diabetes in mice. Cell Metab 2008;8:333–341. 19. Chen D, Bruno J, Easlon E, et al. Tissue-specific regulation of SIRT1 by calorie restriction. Genes Dev 2008; 22:1753–1757. 20. Li Y, Wong K, Walsh K, et al. Retinoic acid receptor beta stimulates hepatic induction of fibroblast growth factor 21 to promote fatty acid oxidation and control wholebody energy homeostasis in mice. J Biol Chem 2013; 288:10490–10504. 21. Badman MK, Koester A, Flier JS, et al. Fibroblast growth factor 21-deficient mice demonstrate impaired adaptation to ketosis. Endocrinology 2009;150:4931–4940. 22. Gesta S, Bezy O, Mori MA, et al. Mesodermal developmental gene Tbx15 impairs adipocyte differentiation and mitochondrial respiration. Proc Natl Acad Sci U S A 2011;108:2771–2776. 23. Fisher FM, Kleiner S, Douris N, et al. FGF21 regulates PGC-1 alpha and browning of white adipose tissues in adaptive thermogenesis. Genes Dev 2012;26:271–281. 24. Zillikens MC, van Meurs JBJ, Rivadeneira F, et al. SIRT1 genetic variation is related to BMI and risk of obesity. Diabetes 2009;58:2828–2834. 25. Holland WL, Adams AC, Brozinick JT, et al. An FGF21adiponectin-ceramide axis controls energy expenditure and insulin action in mice. Cell Metab 2013;17:790–797. Author names in bold designate shared co-first authors. Received February 12, 2013. Accepted October 24, 2013. Reprint requests Address requests for reprints to: Mengwei Zang, MD, PhD, Department of Medicine, Boston University School of Medicine, Boston, 650 Albany Street, X725, Boston, Massachusetts 02118. e-mail: [email protected]; fax: (617) 638-7113, or Yu Li, PhD, Department of Medicine, Boston University School of Medicine, 650 Albany Street, Boston, Massachusetts 02118. e-mail: [email protected]. Acknowledgments The authors greatly appreciate Dr Richard A. Cohen for helpful discussion. The authors are grateful to Dr Francesca Seta for technical assistance in gene chip assays and Dr Markus Michael Bachschmid for the gift of SIRT1 vector. Conflicts of interest The authors disclose no conflicts. Funding This work was supported by National Institutes of Health grants RO1 DK076942 and R21 AA021181, the Robert Dawson Evans Faculty Merit Award, the Wing Tat Lee Award (M.Z.), and by the Boston University Clinical and Translational Science Institute (CTSI) subsidy fund UL1RR025771 (Y.L.).

BASIC AND TRANSLATIONAL LIVER

February 2014

549.e1

Li et al

Gastroenterology Vol. 146, No. 2

Supplementary Materials and Methods Reagents and Antibodies The Quantikine Mouse FGF21 enzyme-linked immunosorbent assay kit (cat. MF2100) was from R&D Systems (Minneapolis, MN). The b-Hydroxybutyrate LiquiColor Test kit (cat. 2440-058) was from Stanbio Laboratory (Boerne, TX). The Dual-Luciferase Reporter Assay kit was purchased from Promega (Madison, WI). The anti-FGF21 antibody (monoclonal, clone 84-E12; raised against human FGF-21 protein) was described previously.1 Rabbit polyclonal SIRT1 antibody (cat. 07-131) was from Millipore (Billerica, MA). SIRT1 antibody (H-300; cat. sc-15404) as well as horseradish-peroxidase–conjugated anti-mouse and antirabbit secondary antibodies were obtained from Santa Cruz Biotechnology, Inc (Santa Cruz, CA).

Liver-Specific SIRT1 Knockout (SIRT1 LKO) Mice SIRT1 LKO mice were generated by crossing albumin-Cre recombinase transgenic mice with floxed SIRT1Dex4 mice containing the deleted SIRT1 exon 4, which encodes 51 amino acids of the conserved SIRT1 catalytic domain.2,3 SIRT1 LKO mice of the C57BL/6 genetic background were described previously.3 These mice were genotyped with the following primers: 50 CCCCATTAAAGCAGTATGTG 30 (F) and 50 CATG TAATCTCAACCTTGAG 30 (R) for the floxed SIRT1 gene; and 50 GTTAATGATCTACAGTTATTGG 30 (F), 50 CGCA TAACCAGTGAAACAGCATTGC 30 (R) for the Alb-Cre gene. SIRT1 LKO mice (Creþ/-/SIRT1flox/flox) and their littermates (Cre-/-/SIRT1flox/flox) were used for experiments. Mice were maintained on a 12-hour light-dark cycle in a temperaturecontrolled barrier facility with free access to water and food. These procedures were approved by the Boston University Medical Center Institutional Animal Care and Use Committee.

Animal Fasting and Refeeding Experiments SIRT1 LKO mice and WT littermates were divided into 3 groups: fed, fasted, and re-fed, as described previously.4,5 The fed group was placed on a normal chow diet, the fasted group was fasted for 24 hours, and the re-fed group was fasted for 24 hours and additionally re-fed for 6 hours before the end of experiments. When mice were killed under isoflurane anesthesia, tissues were taken rapidly, freshly frozen in liquid nitrogen, and stored at 80 C until needed for biochemical analysis.

Gene Expression Array Gene expression chip array was performed using the Fluidigm BioMark system (South San Francisco, CA). Predesigned TaqMan probes and primers were obtained from Applied Biosystems (ABI, Foster City, CA).

Quantitative Real-Time Polymerase Chain Reaction by

First-strand complementary reverse-transcription using

DNA total

was generated RNA. Real-time

polymerase chain reaction was performed using the Applied Biosystems StepOnePlus Real-Time Polymerase Chain Reaction System and the SYBR green kit (Applied Biosystems, Foster City, CA), as described previously.5–7 Primers were designed using Primer3 (Supplementary Table 1). The mRNA abundance of tested genes was determined by quantitative reverse-transcription polymerase chain reaction, normalized to that of b-actin, and expressed as relative mRNA levels.

Immunoblotting Analysis

Mouse liver tissues or cultured cells were lysed at 4 C in lysis buffer containing 20 mmol/L Tris-HCl, pH 8.0, 1% (vol/vol) Nonidet P-40, 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L ethylene glycol-bis(b-aminoethyl ether)N,N,N0 ,N0 -tetraacetic acid, 1 mmol/L sodium orthovanadate, 25 mmol/L b-glycerolphosphate, 1 mmol/L dithiothreitol, 1 mmol/L phenylmethylsulfonyl fluoride, 2 mg/mL aprotinin, 2 mg/mL leupeptin, and 1 mg/mL pepstatin, as described previously.5,6,8–10 The protein levels of SIRT1 or FGF21 were visualized by immunoblots and normalized to b-actin levels.

Measurement of Plasma FGF21 Levels Plasma FGF21 levels were measured using the Quantikine Mouse FGF21 ELISA kit (R&D Systems) with a minimum detectable dose of 3.81 pg/mL, according to the manufacturer’s instructions. Mouse plasma samples (50 mL) were mixed with 50 mL of assay diluent in a 96-well plate, and the reactions were incubated on a plate shaker at room temperature for 2 hours. After washing the plate 5 times, 100 mL of mouse FGF-21 conjugate was added to each well, and the reactions were incubated at room temperature for 2 hours. After washing the plate 5 times, 100 mL of substrate solution was added into each well, followed by plate incubation at room temperature for 30 minutes in the dark. Once 100 mL of stop reaction solution was added to each well, the optical density at 450 nmol/L was measured in a microplate reader (Infinite M1000; Tecan, Männedorf, Switzerland). Plasma FGF21 concentrations were calculated according to the standard curve.

Measurement of Hepatic Protein Concentrations of FGF21 Liver tissue was homogenized in Lysing Matrix D bead tubes (cat. 6913-500; MP Biomedicals, Santa Ana, CA) in 0.4 mL ice-cold RIPA buffer (cat. 89901; Thermo Scientific, Waltham, MA) with 1 Halt Protease and Phosphatase Inhibitor Cocktail (cat. 1861281; Thermo Scientific) at 4 C. Homogenates were solubilized through incubation in cold room for 30 minutes and subjected to centrifugation for 10 minutes at 12,000  g at 4 C, after which the supernatant was collected and the protein concentration was determined (cat. 1856210, Coomassie Plus Protein Assay Reagent; Pierce). FGF21 concentrations then were determined using an enzyme-linked immunosorbent assay (cat. MF2100; R&D Systems) and normalized to protein concentration.

February 2014

Measurement of Blood Glucose, Plasma Insulin, Plasma, and Hepatic Lipids Blood glucose concentrations were determined using the AlphaTRAK Glucose Meter from Abbott Laboratories (North Chicago, IL). Mouse plasma insulin was measured with an ultrasensitive mouse insulin enzyme-linked immunosorbent assay kit (Mercodia AB, Uppsala, Sweden) as described previously.7 Plasma and liver triglyceride and cholesterol levels were analyzed as described previously.6,7,9,11

Liver Oil Red O Staining When animals were killed, livers were embedded in ornithine carbamyl transferase (OCT) (Tissue-Tek, Laborimpex) and rapidly frozen on the dry ice. Hepatic steatosis was assessed by Oil Red O staining in ornithine carbamyl transferase–embedded samples using the Oil Red O Stain Kit from American MasterTech (Lodi, CA) according to the manufacturer’s protocol. Cryosections were fixed in 60% isopropanol for 10 minutes, stained with 0.3% Oil Red O in 60% isopropanol for 30 minutes, and subsequently washed with 60% isopropanol. Sections were counterstained with Gill’s hematoxylin, washed with acetic acid solution (4%), and mounted with aqueous solution. Staining images were captured and digitalized using an Olympus UC30 digital camera attached to an Olympus BX40 microscope (Shinjuku, Tokyo, Japan). Relative areas of steatosis were quantified by histomorphometry using a computerized image analysis system (ImageJ, National Institutes of Health, Bethesda, MD) and expressed as the percentage of area stained with Oil Red O. For each image, the individual staining pattern was generated through the use of ImageJ’s color deconvolution according to previously described methods.12

Body Composition Analysis Body composition was determined with the nuclear magnetic resonance system using a Body Composition Analyzer Echo 900 (Echo Medical Systems, Houston, TX) as described previously.5 Body fat, lean mass, body fluids, and total body water were measured in live conscious mice with ad libitum access to chow.

Comprehensive Metabolic Monitoring Energy expenditure was assessed by indirect calorimetry measurements using Comprehensive Laboratory Animal Monitoring Systems (Columbus Instruments, Columbus, OH) in the Metabolic Phenotyping Core at Boston University School of Medicine, as described previously.5,13,14 Mice of both genotypes were housed individually, acclimatized to respiratory chambers for 24 hours, and allowed free access to food and water. The data for VO2, VCO2, and locomotor activity were recorded simultaneously over a 48-hour period, which was divided into a 24-hour feed and a 24-hour fast. The VO2 and VCO2 rates were expressed as average values measured every 18 minutes over a 12-hour block in the light and dark cycles. The energy expenditure (kcal/kg/h) was calculated with the following formula: (3.815 þ 1.232  VCO2/VO2)  VO2  0.001, and normalized

Regulation of FGF21 by SIRT1 in the Liver

549.e2

to lean body mass as described previously.15,16 Physical activity was measured on X- and Y- axes by using infrared beams to count the beam breaks in the same Comprehensive Laboratory Animal Monitoring Systems cage. The values then were summed (Xamb þ Yamb) over 12-hour intervals of the light and dark cycles, respectively.

Food Intake and Body Temperature Measurements After 1 week of adaptation, food intake of each individual mouse, as measured in grams of food, was recorded daily for 1 week. During each 24-hour period, each mouse was housed with a thin layer of bedding, and a known quantity of diet was provided. After 24 hours, the remaining food was weighed. Body temperature was assessed with implanted programmable subcutaneous biocompatible microchip transponders (IPTT-300 Extended Accuracy Calibration; Bio Medic Data Systems, Seaford, DE) over the mouse’s shoulder blades as instructed by the manufacturer. This procedure involved light anesthesia using isoflurane, disinfection of the back area with surgical scrub (Betadine, Stamford, CT), tenting of the skin over the shoulder blade area, quick insertion of a needle delivery device containing the microchip, and depression of the plunger on the device that expelled the microchip from the delivery device. MB-10 was used to maintain sterility in the surgical environment. Body temperature was measured using a DAS-6007 Probe (Bio Medic Data Systems, Seaford, DE).

In Vivo Adenoviral Gene Transfer The adenovirus-producing mouse FGF21 was generated and purified by using the Adenovirus Purification Kit (Puresyn, Malvern, PA), as described previously.5,6,8,10,11 For the amplification and purification of high-titer recombinant adenoviruses, a large number of HEK293A cells were infected with viral supernatant at a multiplicity of infection of approximately 10 pfu per cell. Adenovirus-mediated overexpression of FGF21 in the liver of mice was accomplished through intravenous injection. One hundred microliters of adenovirus (5  109  1  1010 pfu) per mouse were injected into the tail vein using a 0.1-mL syringe with a 29.5-gauge needle. Two weeks after injection, each group of mice was killed under isoflurane anesthesia.

Dual Luciferase Activity Assays A human FGF21 promoter-driven luciferase plasmid encoding a 50 -flanking fragment of the human FGF21 (-2090/þ117) promoter was generated and described previously.5 HepG2 cells in a 12-well plate were cotransfected with 0.5 mg of plasmids encoding FGF21-Luc promoter, along with 20 ng of Renilla luciferase plasmid pRL-SV40 (Promega, Madison, WI) as an internal control, as described previously.5,6 Dual-luciferase assays for Firefly luciferase and Renilla luciferase activities were performed in duplicate according to the manufacturer’s protocols (Promega). Luciferase activity was measured using a Turner ED-20e Luminometer (Turner Designs Inc, Sunnyvale, CA). The Firefly luciferase activity was normalized to the Renilla

549.e3

Li et al

luciferase activity (Firefly luciferase/Renilla luciferase) and presented as relative luciferase activity.

Cell Treatments and RNA Interference Human HepG2 cells, MEFs derived from wild-type and SIRT1-/- mice, human embryonic kidney 293T cells (HEK293T), and human embryonic kidney 293A cells (HEK293A) were cultured as described previously.6,9–11,17 HepG2 cells were treated without or with either resveratrol (10 mmol/L) or SRT1720 (5 mmol/L) in serum-free Dulbecco’s modified Eagle medium for 24 hours. For RNA interference, FGF21 knockdown was performed using a small interfering RNA oligonucleotide targeting human FGF21 (GCCUUGAAGCCGGGAGUUA, cat. J-013305-06-0005) from Dharmacon RNAi Technology (Thermo Fisher Scientific, Lafayette, CO). On-Target plus Non-targeting Pool (cat. D001810-10-05) from Dharmacon RNAi Technology was used as the control small interfering RNA oligonucleotide.

XF24 Oxygen Consumption Assays and Bioenergetics Profiles Oxygen consumption rates in live cells were measured using the XF24 extracellular flux analyzer from Seahorse Bioscience (North Billerica, MA), as described previously.18 HepG2 cells or MEFs were seeded in a 24-well XF24 V28 cell culture microplate (Seahorse Bioscience) at a density of 20,000 cells per well in 200 mL of Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum for 16 hours in a 5% CO2 incubator at 37 C before the experiments. Cells were washed and incubated in 630 mL of nonbuffered Dulbecco’s modified Eagle medium (without sodium carbonate) at 37 C in a non-CO2 incubator. Triplicates of each treatment were included in the experiments, and 4 wells evenly distributed in the microplate were used for the correction of temperature variation. During the experiments, oxygen consumption rates were measured for 3 minutes within a 7-minute interval, which also included a 2minute mixing period and a 2-minute waiting period. A bioenergetics profile composed of basal mitochondrial respiration, adenosine triphosphate turnover, Hþ leak, mitochondrial respiratory capacity, and nonmitochondrial respiration was determined by calculating the area under the curve. The oxygen consumption rate in the basal condition was first determined and used to calculate the basal mitochondrial respiration. After this, the adenosine triphosphate (ATP) synthase was inhibited by oligomycin (2.5 mmol/L), creating an accumulation of protons in the mitochondrial intermembrane space and a decrease in the electron transport chain activity, reducing the oxygen consumption rate. This reduction in oxygen consumption rate from the basal state reflected the respiration driving ATP synthesis in the cells (ATP turnover). The remaining oxygen consumption rates were composed of the proton leak, which maintained minimal electron transport chain activity and nonmitochondrial respiration. In this situation of high proton gradient and low electron transport chain activity, addition of a mitochondrial uncoupler, carbonyl cyanide-p-

Gastroenterology Vol. 146, No. 2

trifluoromethoxyphenylhydrazone (FCCP, 1.5 mmol/L), allowed the electron transport chain to resume its pumping activity by dissipating the proton gradient across the inner mitochondrial membrane. This induced an increase in the oxygen consumption rate, and FCCP continually dissipated the protons across the membrane and the electron transport chain activity was challenged to its maximal capacity. The oxygen consumption rates during this phase reflected the maximal mitochondrial respiratory capacity. Furthermore, treatment with antimycin (4 mmol/L), an inhibitor of electron transfer, completely suppressed electron transport chain activity. As a result, the oxygen consumption rate decreased dramatically, and the oxygen consumed in this situation by the cells came from a nonmitochondrial respiration. Finally, the basal mitochondrial respiration was calculated by subtracting the nonmitochondrial oxygen consumption rate in antimycin A–treated cells from the oxygen consumption rate in pretreated cells.

Statistical Analysis Values are expressed as the mean  SEM. Statistical significance was evaluated using an unpaired 2-tailed t test or a 1-way ANOVA if there were more than 2 groups. Differences were considered significant at a P value of less than .05.

References 1. Izumiya Y, Bina HA, Ouchi N, et al. FGF21 is an Aktregulated myokine. FEBS Lett 2008;582:3805–3810. 2. Cheng HL, Mostoslavsky R, Saito S, et al. Developmental defects and p53 hyperacetylation in Sir2 homolog (SIRT1)-deficient mice. Proc Natl Acad Sci U S A 2003; 100:10794–10799. 3. Chen D, Bruno J, Easlon E, et al. Tissue-specific regulation of SIRT1 by calorie restriction. Genes Dev 2008; 22:1753–1757. 4. Engelking LJ, Kuriyama H, Hammer RE, et al. Overexpression of Insig-1 in the livers of transgenic mice inhibits SREBP processing and reduces insulin-stimulated lipogenesis. J Clin Invest 2004;113:1168–1175. 5. Li Y, Wong K, Walsh K, et al. Retinoic acid receptor beta stimulates hepatic induction of fibroblast growth factor 21 to promote fatty acid oxidation and control wholebody energy homeostasis in mice. J Biol Chem 2013; 288:10490–10504. 6. Li Y, Xu SQ, Mihaylova MM, et al. AMPK phosphorylates and inhibits SREBP activity to attenuate hepatic steatosis and atherosclerosis in diet-induced insulin-resistant mice. Cell Metab 2011;13:376–388. 7. Li Y, Xu SQ, Giles A, et al. Hepatic overexpression of SIRT1 in mice attenuates endoplasmic reticulum stress and insulin resistance in the liver. FASEB J 2011; 25:1664–1679. 8. Zang MW, Zuccollo A, Hou XY, et al. AMP-activated protein kinase is required for the lipid-lowering effect of metformin in insulin-resistant human HepG2 cells. J Biol Chem 2004;279:47898–47905. 9. Zang MW, Xu SQ, Maitland-Toolan KA, et al. Polyphenols stimulate AMP-activated protein kinase, lower lipids, and inhibit accelerated atherosclerosis in diabetic LDL receptor-deficient mice. Diabetes 2006;55:2180–2191.

February 2014 10. Zang M, Gong J, Luo L, et al. Characterization of Ser338 phosphorylation for Raf-1 activation. J Biol Chem 2008; 283:31429–31437. 11. Hou X, Xu S, Maitland-Toolan KA, et al. SIRT1 regulates hepatocyte lipid metabolism through activating AMP-activated protein kinase. J Biol Chem 2008;283:20015–20026. 12. Ruifrok AC. Quantification of immunohistochemical staining by color translation and automated thresholding. Anal Quant Cytol Histol 1997;19:107–113. 13. Maeda K, Cao HM, Kono K, et al. Adipocyte/macrophage fatty acid binding proteins control integrated metabolic responses in obesity and diabetes. Cell Metab 2005;1:107–119. 14. Bence KK, Delibegovic M, Xue B, et al. Neuronal PTP1B regulates body weight, adiposity and leptin action. Nat Med 2006;12:917–924.

Regulation of FGF21 by SIRT1 in the Liver

549.e4

15. Hofmann SM, Zhou L, Perez-Tilve D, et al. Adipocyte LDL receptor-related protein-1 expression modulates postprandial lipid transport and glucose homeostasis in mice. J Clin Invest 2007;117:3271–3282. 16. Tang T, Zhang J, Yin J, et al. Uncoupling of inflammation and insulin resistance by NF-kappa B in transgenic mice through elevated energy expenditure. J Biol Chem 2010; 285:4637–4644. 17. Zang MW, Dong MQ, Pinon DI, et al. Spatial approximation between a photolabile residue in position 13 of secretin and the amino terminus of the secretin receptor. Mol Pharmacol 2003;63:993–1001. 18. Gesta S, Bezy O, Mori MA, et al. Mesodermal developmental gene Tbx15 impairs adipocyte differentiation and mitochondrial respiration. Proc Natl Acad Sci U S A 2011;108:2771–2776.

Supplementary Figure 1. Hepatic expression of SREBP-1c and FAS in fasted SIRT1 LKO mice is not affected by hepatic overexpression of FGF21. Adenoviruses (5  109 pfu) encoding either FGF21 or control green fluorescent protein (GFP) were delivered into 8- to 10-month-old mice (n ¼ 4–7) via tail vein injection. The animals were killed in a 24-hour fasted state after 2 weeks of injection.

549.e5

Li et al

Gastroenterology Vol. 146, No. 2

Supplementary Table 1.Quantitative Reverse-Transcription Polymerase Chain Reaction Primers Gene FGF21 CPT1a MCAD ACAT1 HMGCS2 HMGCL BDH1 b-actin UCP1 DIO2 Cidea Cox7a1 PRDM16 Adiponectin Leptin FGF21 CPT1a b-actin

Species Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Human Human Human

Forward primers

Reverse primers

CTGGGGGTCTACCAAGCATA CCAGGCTACAGTGGGACATT GATCGCAATGGGTGCTTTTGATAGAA CCCCATTGATTTTCCACTTG ATACCACCAACGCCTGTTATGG ACTACCCAGTCCTGACTCCAA GAAGATGCTGTCCGGCTAAG CCACAGCTGAGAGGGAAATC CACCTTCCCGCTGGACACT AGAGTGGAGGCGCATGCT ATCACAACTGGCCTGGTTACG CAGCGTCATGGTCAGTCTGT CAGCACGGTGAAGCCATTC GTTGCAAGCTCTCCTGTTCC TGAAGGATCCGGAAGTGTTC ACCTGGAGATCAGGGAGGAT GCTTTGGAGACCTGCTTTTG GATGAGATTGGCATGGCTTT

CACCCAGGATTTGAATGACC GAACTTGCCCATGTCCTTGT AGCTGATTGGCAATGTCTCCAGCAAA AGCACAACCACACTGAATGC CAATGTCACCACAGACCACCAG TAGAGCAGTTCGCGTTCTTCC CACCACTGGTCTGTTTGCAG AAGGAAGGCTGGAAAAGAGC CCCTAGGACACCTTTATACCTAATGG GGCATCTAGGAGGAAGCTGTTC TACTACCCGGTGTCCATTTCT AGAAAACCGTGTGGCAGAGA GCGTGCATCCGCTTGTG ATCCAACCTGCACAAGTTCC TCCAAATGTTCCGGAAAGAG GCACAGGAACCTGGATGTCT AGGCACTGGGTTGTGAAGAC GTCACCTTCACCGTTCCAGT

ACAT1, acetyl-CoA acetyltransferase 1; BDH1, b-hydroxybutyrate dehydrogenase 1; HMGCS2, 3-hydroxy-3-methylglutarylCoA synthase 2; HMGCL, HMG-CoA lyase.

February 2014

Regulation of FGF21 by SIRT1 in the Liver

549.e6

Supplementary Table 2.Gene Expression Profile in the Liver of WT and SIRT1 LKO Mice in the Fed or Fasted States Fed WT (n ¼ 4) Genes Acacb Angpt1 Angpt2 Arg1 Arg2 Atp2a2

Atp2a3

Cat CD3e Cdh5 Cox4i1 Cpt1a Crmp1 ctgf cybb Edn1 Fgf1 Fgf2 Fgf21 Flt1 Glrx Glrx2 Gpx1 Gsr Hif1a Hmox1 Hras1 Icam1 Igf1 Il18 Il1b IL33 Il4r Il6 Kdr Keap1 LIgl1/Mgl1 Mcp1 Mfn1 Mfn2 Mip1a MMP19 mmp2 Mmp9

Fasted

LKO (n ¼ 5)

WT (n ¼ 4)

LKO (n ¼ 7)

Definition

AB assay ID

Mean

SEM

Mean

SEM

Mean

SEM

Mean

SEM

ACC2 (acetyl-CoA carboxylase b) Angiopoietin-1 Angiopoietin-2 Arginase-1 Arginase-2 SERCA 2 (sarcoplasmic/ endoplasmic reticulum calcium adenosine triphosphatase 2) SERCA 3 (sarcoplasmic/ endoplasmic reticulum calcium adenosine triphosphatase 3) Catalase CD3 VE-cad/cadherin 5 Cox 4 (cytochrome c oxidase subunit IV isoform 1) Cpt1a (carnitine palmitoyltransferase 1a, liver) Drp-1 (collapsin response mediator protein 1) Connective tissue growth factor NOX2 (gp91phox) Endothelin 1 FGF-1 (a-FGF) FGF-2 (b-FGF) FGF 21 Flt (vascular endothelial growth factor receptor 2, VEGFR2) Glutaredoxin-1 Glutaredoxin-2 Glutathione peroxidase 1 Glutathione reductase Hypoxia-inducible factor 1 a (HIF1a) Heme oxygenase-1 H-Ras Intercellular adhesion molecule 1 (ICAM1) Insulin-like growth factor 1 IL18 IL1b IL33 IL4 receptor IL6 Flk-1 Keap1 (kelch-like ECH-associated protein 1) Macrophage galactose lectin Monocyte chemotactic protein 1 Mitofusin 1 Mitofusin 2 Macrophage inflammatory protein 1a Matrix metallopeptidase 19 Matrix metallopeptidase 2 Matrix metallopeptidase 9

Mm01204683_m1 Mm01129232_m1 Mm00545822_m1 Mm00475988_m1 Mm00477592_m1 Mm01201431_m1

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

0.5117 0.2290 0.2621 0.2670 0.2433 0.2704

1.3948 1.4178 1.0307 0.9790 0.6524 1.0778

0.3598 0.4986 0.1019 0.1597 0.2324 0.1100

0.7854 2.3787 0.7156 1.4233 0.5805 1.9577

0.1605 0.3903 0.1379 0.2435 0.1753 0.2442

0.2748 1.3176 1.6696 1.1904 0.5691 1.3220

0.0420 0.3568 0.8270 0.2566 0.1356 0.18769

Mm00443898_m1 1.0000 0.4660 0.6998 0.1829

0.6585 0.1605 0.7141 0.1315

Mm00437992_m1 Mm01179194_m1 Mm00486938_m1 Mm00438289_g1

0.4778 0.8108 0.5950 0.4665

1.0000 1.0000 1.0000 1.0000

0.3910 0.8818 0.2946 0.2825

0.5127 0.4703 0.8344 0.5266

0.0309 0.2537 0.1268 0.0688

0.0606 0.2463 0.0691 0.0576

0.4242 0.1905 0.4508 0.5237

0.0546 0.0085 0.0292 0.0387

Mm00550438_m1 1.0000 0.2640 1.2399 0.1592

2.8582 0.1042 2.2033 0.1471

Mm00514390_m1 0.0000 0.0000 0.0000 0.0000

0.0000 0.0000 0.0000 0.0000

Mm01192933_g1 Mm01287743_m1 Mm00438656_m1 Mm0043806_m1 Mm00433287_m1 Mm00840165_g1 Mm00438980_m1

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

0.2215 0.4454 0.2456 0.2998 0.4996 0.1696 0.2697

0.8117 1.0373 1.2701 1.5380 0.4960 0.1692 0.8372

0.1994 1.8046 0.4443 0.2099 0.7960 0.1667 0.1947 2.0998 0.3121 0.3327 2.8674 0.1196 0.2211 0.7858 0.2887 0.0753 11.7127 2.3898 0.0873 0.7286 0.0711

2.4001 0.5178 1.4642 2.1000 0.3407 4.8701 0.7318

0.52257 0.0421 0.4324 0.2372 0.0566 1.2228 0.0495

Mm00728386_s1 Mm00469836_m1 Mm00656767_g1 Mm00439151_m1 Mm00468878_m1

1.0000 1.0000 1.0000 1.0000 1.0000

0.1787 0.2556 0.2409 0.3248 0.2588

0.8447 1.0936 1.2193 1.0281 0.8141

0.1279 0.1320 0.1213 0.3312 0.1414

1.5824 1.2093 1.1752 1.2390 0.7386

0.1758 0.0931 0.1589 0.2689 0.0549

2.3928 1.2259 1.9959 1.9215 0.9154

0.2937 0.1109 0.2282 0.3767 0.0810

Mm00516005_m1 1.0000 0.2603 0.8348 0.1711 Mm00476174_m1 1.0000 0.2789 0.7701 0.1367 Mm00516023_m1 1.0000 0.3320 2.0389 0.5216

0.8178 0.1612 1.0266 0.1879 0.4860 0.0238 0.5071 0.0530 2.0826 0.3839 1.0721 0.0510

Mm00439560_m1 Mm00434225_m1 Mm01336189_m1 Mm00505403_m1 Mm00439634_m1 Mm00446191_m1 Mm00440111_m1 Mm00497268_m1

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

0.2460 0.3622 0.4104 0.3890 0.2218 0.4654 0.2575 0.2742

1.5780 1.2349 0.4996 0.8764 1.1877 0.3319 0.7166 0.6221

0.2497 0.1763 0.0897 0.2169 0.1755 0.1489 0.1041 0.0916

1.8687 0.5289 0.3316 0.6842 1.5164 0.2192 0.6135 0.4285

0.1330 0.0848 0.1089 0.0879 0.1485 0.0963 0.0815 0.0319

0.9627 0.3755 0.3578 0.3317 0.9488 0.4521 0.5912 0.6287

0.0522 0.0368 0.0585 0.0267 0.0969 0.1148 0.0366 0.0650

Mm00839416_m1 Mm00441242_m1 Mm00612599_m1 Mm00500120_m1 Mm00441259_g1

1.0000 1.0000 1.0000 1.0000 1.0000

0.3641 0.6410 0.2512 0.2749 0.4831

1.0118 0.8619 1.3021 0.8954 0.4924

0.4457 0.2021 0.1529 0.1207 0.1535

0.7495 0.2270 1.5877 1.0762 0.6378

0.1142 0.0899 0.0603 0.0588 0.1442

1.5269 0.2177 1.0406 0.8386 0.5285

0.7431 0.0985 0.0830 0.0830 0.1438

Mm00491300_m1 1.0000 0.2950 1.7328 0.2234 Mm00439498_m1 1.0000 0.2790 0.7600 0.2063 Mm00442991_m1 1.0000 0.3522 0.4628 0.1160

1.6701 0.2319 1.3121 0.1588 0.4622 0.0672 0.7735 0.1539 0.2910 0.0993 0.3804 0.0812

549.e7

Li et al

Gastroenterology Vol. 146, No. 2

Supplementary Table 2. Continued Fed WT (n ¼ 4) Genes Ncf1 Nos3

Definition

P47phox NOS3 (endothelial nitric oxide [eNOS]) Nox4 NOX 4 nrf2 NRF-2 Pecam1 Platelet endothelial cell adhesion molecule 1 Pparg PPARg Ppargc1a Peroxisome proliferative activated receptor g, coactivator 1 a Prdx1 Peroxiredoxin 1 Prdx2 Peroxiredoxin 2 Prkaa2 AMPKa2 (protein kinase, AMPactivated, a2 catalytic subunit) Pten Phosphatidylinositol-3,4,5trisphosphate 3-phosphatase Saa1 Saa1 Sesn2 Sestrin2 Sirt1 Sirt 1 Sirt3 Sirt 3 Slc2a2 GLUT2 (solute carrier family 2 [facilitated glucose transporter], member 2) Sod1 CuZnSOD (Superoxide dismutase 1) Sod2 MnSOD (Superoxide dismutase 2) Sod3 ecSOD (Superoxide dismutase 3) SPHK1 Sphingosine kinase 1 stat3 Signal transducer and activator of transcription 3 Tfam Mitochondrial transcription factor A Timp2 Tissue inhibitors of metalloproteinases Tlr2 Toll-like receptor 2 Tlr4 Toll-like receptor 4 tnf Tumor necrosis factor-a traf6 Traf6 TSP-1/Thbs1 Thrombospondin-1 Txn1 Thioredoxin-1 Ucp2 Mitochondrial uncoupling protein 2 PPARa PPARa Vcam1 Vascular cell adhesion protein 1 Vegfa VEGF A vegfc VEGF C Vegfa VEGF A vegfc VEGF C

AB assay ID

Mean

SEM

Fasted

LKO (n ¼ 5) Mean

SEM

WT (n ¼ 4) Mean

SEM

LKO (n ¼ 7) Mean

SEM

Mm00447921_m1 1.0000 0.5729 2.2559 0.2873 Mm00435217_m1 1.0000 0.5343 0.6085 0.1247

2.0047 0.2231 1.0705 0.2517 1.1143 0.1564 1.3341 0.2486

Mm00479246_m1 1.0000 0.4836 0.7746 0.2557 Mm00477784_m1 1.0000 0.2599 0.9895 0.1457 Mm01242584_m1 1.0000 0.3152 1.1876 0.1714

0.1860 0.1121 0.6163 0.1978 0.6496 0.0847 0.5163 0.0462 1.3326 0.1450 1.1502 0.1091

Mm01184322_m1 1.0000 0.3099 1.1401 0.1728 Mm01208835_m1 1.0000 0.3709 0.7149 0.1216

1.9204 0.2300 0.7253 0.1921 1.7404 0.3575 1.0149 0.2214

Mm01621996_s1 1.0000 0.1192 0.7690 0.1615 Mm00448996_m1 1.0000 0.3051 1.3354 0.1868 Mm01264789_m1 1.0000 0.2720 1.1832 0.2194

0.6356 0.0897 0.8471 0.0816 1.3023 0.1204 1.0232 0.1556 0.6857 0.0283 0.6936 0.0716

Mm00477210_m1 1.0000 0.2844 1.0477 0.0770

1.7122 0.1402 1.0341 0.0934

Mm00656927_g1 Mm00460679_m1 Mm00490758_m1 Mm00452129_m1 Mm00446229_m1

1.0000 1.0000 1.0000 1.0000 1.0000

0.8935 0.8680 1.4393 1.4374 0.7685

Mm01344233_g1

1.0000 0.2560 0.6467 0.0740

0.5695 0.0907 0.5345 0.0425

Mm00449726_m1 Mm01213380_s1 Mm00448841_g1 Mm01219775_m1

1.0000 1.0000 1.0000 1.0000

1.1511 0.5923 0.1973 0.6417

0.5075 0.7513 0.4092 0.2944 0.3220

0.2787 0.5314 0.9171 0.2727

1.3438 0.2567 0.2159 1.6288 0.9282

0.9394 0.5825 0.2371 0.8284

0.5165 0.0650 0.0471 0.2655 0.1579

0.1206 0.1995 0.1526 0.1520

0.6454 0.1318 0.0605 0.1119 0.0891

0.0807 0.1841 0.1064 0.0636

1.2437 0.4010 0.2240 1.3337 0.4633

0.8085 1.1623 0.5016 0.5640

0.8670 0.0716 0.0275 0.1502 0.0674

0.0549 0.2100 0.1450 0.0375

Mm00447485_m1 1.0000 0.2309 1.1577 0.1679

1.2896 0.1122 0.7700 0.0633

Mm00441825_m1 1.0000 0.2987 0.4112 0.0909

0.5110 0.0772 0.5483 0.0473

Mm00442346_m1 Mm00445273_m1 Mm00443258_m1 Mm00493836_m1 Mm01335418_m1 Mm00726847_s1 Mm00627598_m1

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

0.5849 0.3537 0.6672 0.2679 0.7880 0.3675 0.3502

0.7702 1.3637 0.6038 0.9540 0.2388 0.4339 1.0747

0.1966 0.3313 0.1246 0.2180 0.0872 0.1156 0.2100

0.6207 1.3749 0.3754 1.0494 0.1471 0.4704 2.2272

0.1367 0.3503 0.1505 0.0656 0.0150 0.1704 0.2312

0.3315 1.1109 0.1895 0.7193 0.2598 1.1259 1.3722

0.0768 0.1783 0.0371 0.0915 0.1877 0.1979 0.1322

Mm00440939_m1 Mm01320970_m1 Mm01281449_m1 Mm00437310_m1 Mm01281449_m1 Mm00437310_m1

1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

0.3614 0.3666 0.2601 0.3998 0.2601 0.3998

1.3762 1.1314 0.8977 1.4047 0.8977 1.4047

0.2961 0.1918 0.0763 0.2579 0.0763 0.2579

1.9633 1.9640 0.6418 1.2690 0.6418 1.2690

0.1783 0.3449 0.0560 0.3027 0.0560 0.3027

2.2963 1.6713 0.5183 0.6408 0.5183 0.6408

0.3649 0.2136 0.0278 0.1590 0.0278 0.1590