Pancreatic-derived factor promotes lipogenesis in the mouse liver: Role of the Forkhead box 1 signaling pathway

Authors

  • Jing Li,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yujing Chi,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
    • Potential conflict of interest: Nothing to report.

  • Chunjiong Wang,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
  • Jing Wu,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
  • Hang Yang,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
  • Dongjuan Zhang,

    1. Department of Physiology and Pathophysiology, Beijing, China
    Search for more papers by this author
  • Yi Zhu,

    1. Department of Physiology and Pathophysiology, Beijing, China
    2. Peking (Beijing) University Diabetes Center, Beijing, China
    3. Key Laboratory of Molecular Cardiovascular Science, Peking (Beijing) University Health Science Center, Beijing, China
    Search for more papers by this author
  • Nanping Wang,

    1. Peking (Beijing) University Diabetes Center, Beijing, China
    2. Key Laboratory of Molecular Cardiovascular Science, Peking (Beijing) University Health Science Center, Beijing, China
    Search for more papers by this author
  • Jichun Yang,

    Corresponding author
    1. Department of Physiology and Pathophysiology, Beijing, China
    2. Peking (Beijing) University Diabetes Center, Beijing, China
    3. Key Laboratory of Molecular Cardiovascular Science, Peking (Beijing) University Health Science Center, Beijing, China
    • Department of Physiology and Pathophysiology, Peking (Beijing) University Health Science Center, 38 Xueyuan Road, Beijing, China 100191
    Search for more papers by this author
  • Youfei Guan

    Corresponding author
    1. Department of Physiology and Pathophysiology, Beijing, China
    2. Peking (Beijing) University Diabetes Center, Beijing, China
    3. Key Laboratory of Molecular Cardiovascular Science, Peking (Beijing) University Health Science Center, Beijing, China
    • Department of Physiology and Pathophysiology, Peking (Beijing) University Health Science Center, 38 Xueyuan Road, Beijing, China 100191
    Search for more papers by this author

  • Supported by grants from the Natural Science Foundation (30771030, 81030003, 30725033, and 30890041 to Y. G. and J. Y.) and the Ministry of Science and Technology (2010CB912503 to Y. G.). J. Y. is supported by the “New Century Excellent Talents in Universities” Program from the Ministry of Education of China.

Abstract

Pancreatic-derived factor (PANDER) is a pancreatic islet-specific cytokine that cosecretes with insulin and is important for β cell function. Here, we show that PANDER is constitutively expressed in hepatocytes, and its expression is significantly increased in steatotic livers of diabetic insulin-resistant db/db mice and mice fed a high-fat diet. Overexpression of PANDER in the livers of C57Bl/6 mice promoted lipogenesis, with increased Forkhead box 1 (FOXO1) expression, whereas small interfering RNA–mediated knockdown of hepatic PANDER significantly attenuated steatosis, with reduced FOXO1 expression in db/db mice. Hepatic PANDER silencing also attenuated insulin resistance and hyperglycemia in db/db mice. In cultured hepatocytes, PANDER overexpression induced lipid deposition, increased FOXO1 expression, and suppressed insulin-stimulated Akt activation and FOXO1 inactivation. Moreover, FOXO1 overexpression increased PANDER expression in cultured hepatocytes and mouse livers. Conclusion: PANDER promotes lipogenesis and compromises insulin signaling in the liver by increasing FOXO1 activity. PANDER may represent a potential therapeutic target for the treatment of fatty liver and insulin resistance. (HEPATOLOGY 2011;)

Nonalcoholic fatty liver disease (NAFLD) is the most common cause of liver disease in the nonalcoholic, viral hepatitis–negative population in the United States and Europe.1 NAFLD is estimated to be present in up to 30% of the United States population.2 Although the underlying mechanisms remain unclear, NAFLD represents one aspect of metabolic syndrome, and its pathogenesis is associated with hepatic insulin resistance and enhanced liver lipogenesis.3

The cytokine-like protein pancreatic-derived factor (PANDER, FAM3B) was cloned in 2002 and belongs to the superfamily of eukaryotic FAM3 genes.4 It is selectively expressed at a high level in pancreatic islets and at low levels in the small intestine, prostate, and certain neurons.4 The biological role of PANDER remains largely unknown. We and others have shown that recombinant PANDER protein or adenovirus (Ad)-mediated PANDER overexpression induced apoptosis of α and β cells isolated from mouse, rat, and human islets,5, 6 which suggests that PANDER may be involved in regulating the survival of pancreatic islet cells. Recently, increasing evidence has demonstrated that in β cells, PANDER and insulin cosecrete in the same secretary granules, and the expression of PANDER can be induced by glucose treatment through multiple signaling pathways,7, 8 which implies that PANDER may act as a circulating hormone involved in regulating glucose homeostasis. In support of this hypothesis, we have recently found that recombinant PANDER indeed binds to liver cell membranes and inhibits insulin signaling in mouse liver and HepG2 cells.9 Because a large body of evidence points to hepatic insulin resistance playing a critical role in NAFLD,10 this finding also suggests that PANDER may participate in the pathogenesis of hepatic steatosis.

It has been reported recently that PANDER gene-deficient mice exhibit glucose intolerance due to impaired insulin secretion from islet β cells, but hepatic glucose production during hyperinsulinemic–euglycemic clamp was significantly repressed in PANDER gene-deficient mice, suggesting that liver-produced PANDER may also be involved in the regulation of hepatic glucose metabolism.11 In addition, Wilson et al.12 further demonstrated that overexpression of PANDER in normal mice induces hyperglycemia and glucose intolerance. These observations suggest that liver-produced PANDER may be a negative regulator of hepatic insulin sensitivity. However, to date, the role of PANDER in the regulation of lipid metabolism and in the pathogenesis of NAFLD remains uncharacterized. The purpose of the present study was to determine the role of liver-produced PANDER in the progression of NAFLD.

Abbreviations

Ad, adenovirus; AMPK, adenosine monophosphate–activated protein kinase; cDNA, complementary DNA; CHO, cholesterol; FAS, fatty acid synthase; FOXO1, Forkhead box O1; G6Pase, glucose-6-phosphatase; GFP, green fluorescent protein; HFD, high-fat diet; IgG, immunoglobulin G; mRNA, messenger RNA; NAFLD, nonalcoholic fatty liver disease; pAMPK, phosphorylated adenosine monophosphate–activated protein kinase; PANDER, pancreatic-derived factor; PPARγ, peroxisome proliferator-activated receptor γ; RT-PCR, reverse-transcription polymerase chain reaction; siRNA, small interfering RNA; SREBP, sterol regulatory element binding protein; TG, triglyceride; VLDL, very low-density lipoprotein.

Materials and Methods

Antibodies and Other Reagents.

The anti-PAN1 is an affinity-purified rabbit polyclonal antibody against the peptide PVLKAPAPKRQ-C mapping amino acids 51-63 of mouse PANDER (Biosen Biotechnology Co., Beijing, China). All other antibodies used in the study were purchased from Santa Cruz Biotechnology or Cell Signaling. An adenovirus expressing murine PANDER (Ad-PANDER) was constructed by SinoGenoMax Co. (Beijing, China) with a full-length PANDER complementary DNA (cDNA) coding sequence kindly provided by Bryan A. Wolf5; adenoviruses expressing the full-length Forkhead box 1 (Ad-FOXO1) and green fluorescent protein (Ad-GFP) were a gift from H. Dong (University of Pittsburgh).13 Phosphorylated proteins examined in this study include p-Akt(Ser 473), pAMPK(Thr 172), and pFOXO1(Thr24).

Cell Culture.

HepG2 cells (American Type Culture Collection, Rockville, MD) within the 8th passage were cultured at 37°C in 5% CO2–95% air and were infected with the indicated doses of Ad-GFP or Ad-PANDER for 48 hours.

Animals.

Eight-week-old male C57Bl/6 mice (Jackson Laboratory, Bar Harbor, ME) were used to determine the tissue distribution of PANDER and hepatic overexpression of PANDER. Eight- to 12-week-old male insulin-resistant db/db mice on a C57BKS background and age- and sex-matched db/m mice were used to determine PANDER expression in fatty liver and the effect of physical training (swimming 40 minutes/day for 7 weeks) and small interfering RNA (siRNA) on hepatic PANDER expression. All procedures were approved by the Institutional Animal Care and Use Committee of Peking University Health Science Center.

Tissue Distribution of PANDER as Assessed by Way of Reverse-Transcription Polymerase Chain Reaction Analysis.

Tissue distribution of PANDER messenger RNA (mRNA) in mouse was examined by way of reverse-transcription polymerase chain reaction (RT-PCR) with two sets of primers, which amplified a 200-bp portion of PANDER cDNA and a 708-bp full-length PANDER cDNA. RT-PCR involved livers from three 10-week-old male Sprague-Dawley rats and three male human volunteers aged 40-45 years. The use of human samples was approved by the Human Research Ethics Committee at Peking University Health Science Center. All primers for RT-PCR or real-time PCR assays are given in Supporting Table 1.

Immunohistochemistry.

Pancreases and livers were fixed, dehydrated, and embedded in paraffin wax. Sections (5 μm) were incubated with the anti-PAN1 antibody (1:200) and a nonimmune immunoglobulin G (IgG) overnight at 4°C and then incubated with a polyperoxidase-conjugated goat anti-rabbit IgG (Zhongshan Golden Bridge, Beijing, China) for 30 minutes at 37°C. The slides were counterstained with 3,3′-diaminobenzidine or hematoxylin.

Overexpression and Knockdown of PANDER in Mouse Livers.

Overexpression of PANDER was achieved by way of tail vein injection of Ad-PANDER in normal C57B/6 mice at 1.5 × 109 plaque-forming units as described.14 To knockdown hepatic PANDER expression in db/db mice, a mixture of three sets of stealth siRNA against mouse PANDER cDNA coding sequence was synthesized by Invitrogen (sense1, 5′ UUUGCUGCAAC AAACACCCAGCUUG-3′, antisense1, 5′-CAAGCUGG GUGUUUGUUGCAGCAAA-3′; sense2, 5′-UUGACG ACAGCAAUGUUUAUCCC-3′, antisense2, 5′-AGAAA CAUUGCUGUCGUCAACU-3′; sense3, 5′-CGGAU GUUGUAGAGAGUGCUGGA-3′, antisense3, 5′-UCC AGCACUCUCUAC

AACAUCCGAA-3′). The siRNA mixture was administrated to db/db mice by way of tail vein injection at 2.5 mg/kg body weight in 100 μL sterile saline as described.15, 16 The same amount of scramble sequences from Invitrogen was used as a control. Three days later, fasting blood glucose, glucose tolerance, and minimal model analysis were monitored, and tissues were collected for Oil Red O staining, real-time PCR, and immunohistochemical assays.

Measurement of Hepatic Triglyceride and Cholesterol Content and Serum Lipid Profile Analysis.

Total triglyceride (TG) and cholesterol (CHO) were extracted from mouse liver and quantitated by use of TG and CHO assay kits.17 Serum samples were obtained from mice for TG, CHO, and insulin measurements. The serum lipid profile was determined by high-performance liquid chromatography as described.18

Glucose Tolerance Test.

Mice were fasted for 6 hours (8 AM to 2 PM) before oral glucose tolerance test. Blood glucose levels were measured immediately before glucose injection and then at 30, 60, 90, and 120 minutes after oral administration of glucose (2 g/kg body weight). Blood glucose levels were determined by use of a Freestyle Brand Glucometer (Roche) with blood collected from the tail vein.

Minimal Model Analysis of Insulin Sensitivity.

The protocol for minimal model analysis of insulin sensitivity has been described.19 In brief, the mice were fasted for 6 hours and then anesthetized by way of intraperitoneal injection of sodium pentobarbital. After 30 minutes, a blood sample (30 μL) was collected. Then, D-glucose was injected through the jugular vein at a dose of 1 g/kg. Additional blood samples (30 μL each) were taken 0, 2, 4, 8, 19, 22, 30, 40, 50, 70, 90, and 180 minutes after glucose injection. Insulin (0.03 U/kg) was given at the time point of 19 minutes by way of tail vein injection. At the time points indicated above, tail vein blood samples were collected for glucose measurement, and the serum was immediately separated for insulin measurement using an enzyme-linked immunosorbent assay kit.

Immunoblot Assays.

To determine the expression levels of selected proteins, 200 μg of liver protein was separated by way of 10% or 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Western blot analysis was performed as described.8 Protein expression levels were quantitated with use of Image J (version 1.42). All data were normalized to the control value.

Quantitative Real-Time PCR Assays of mRNA Levels.

Total RNA was isolated with use of TRIzol reagent (Invitrogen). Quantitative PCR was performed with CHROMO4 (MJ Research) and SYBR Green 1 (Molecular Probes, Eugene, OR) used as a fluorescent probe. Target gene mRNA level was normalized to that of β-actin in the same sample as described.5, 20 All the primers for real-time PCR or RT-PCR assays are given in Supporting Table 1.

Statistical Analysis.

Data are presented as the mean ± SEM. Statistical significance of differences between two groups was analyzed by way of paired Student t test. When more than two groups were compared, the Kruskal-Wallis test and Dunn's test were used. P < 0.05 was considered statistically significant.

Results

PANDER Is Constitutively Expressed in Mouse, Rat, and Human Liver.

RT-PCR analysis revealed PANDER mRNA expressed in most mouse tissues examined and in rat and human livers (Fig. 1A, Supporting Fig. 1) and in human hepatocarcinoma cell line HepG2 cells (data not shown). The level of PANDER protein in mouse liver was approximately three-fold lower than that in the pancreas (Fig. 1B). Immunohistochemical analysis revealed PANDER protein diffusely expressed in the liver (Fig. 1C). Western blot further confirmed that full-length PANDER protein was expressed in mouse liver (Fig. 1D).

Figure 1.

PANDER expression in normal and steatotic mouse livers. (A) Ubiquitous expression of PANDER mRNA in mouse tissues, including the liver and pancreas, as assessed by way of RT-PCR. An RNA sample without reverse-transcription was used as a negative control. (B) Western blot analysis of PANDER protein expression in the liver and pancreas. *P < 0.05 versus liver (n = 5). (C) Intrahepatic distribution of PANDER protein in normal C57B/6 mouse livers as analyzed by way of immunohistochemistry. Hepatic (left and middle) and pancreatic tissue (right) sections were stained with specific PANDER antibody (left and right) and nonspecific IgG (middle). Diffused PANDER immunoreactivity was observed in the liver (left). Original magnification ×400. (D) Full-length (27 kD) but not secreted (23 kD) PANDER protein detected in mouse liver by way of western blot analysis. H, high dose (multiplicity of infection, 100) of Ad-PANDER; L, low dose (multiplicity of infection, 20) of Ad-PANDER. (E) Western blot analysis demonstrating an increased hepatic PANDER protein expression in db/db mice. *P < 0.05 versus db/m mice, n = 5. (F) Physical exercise (swimming 40 minutes/day for 7 weeks) attenuated PANDER protein levels in the livers of db/db mice. *P < 0.05 versus db/m, #P < 0.05 versus db/db (n = 5). The results are expressed as the mean ± SEM.

Increased Hepatic PANDER Expression in db/db Mice and High-Fat Diet–Fed Mice.

To investigate the role of PANDER in the pathogenesis of fatty liver, the PANDER expression in the liver of db/db mice and mice fed a high-fat diet (HFD) was evaluated. Hepatic PANDER mRNA (data not shown) and protein levels (Fig. 1E) were increased approximately two- and four-fold, respectively, in the livers of db/db mice as compared with db/m mice. Similarly, PANDER protein expression was increased in the livers of mice fed an HFD for 12 weeks (Supporting Fig. 2). Physical exercise for 7 weeks significantly improved insulin resistance, ameliorated fatty liver, and attenuated hepatic PANDER mRNA (data not shown) and protein expression in db/db mice (Fig. 1F, Supporting Fig. 3). Similarly, increased hepatic PANDER expression in the livers of HFD-fed rats was also reversed by physical exercise, concomitant with an amelioration of fatty liver (data not shown).

Induction of Hepatic Steatosis and Insulin Resistance by Hepatic Overexpression of PANDER in C57Bl/6 Mice.

Hepatic overexpression of PANDER was achieved by way of tail vein injection of Ad-PANDER and confirmed by a 100-fold increase in mRNA level (data not shown) and a four-fold increase in protein level (Fig. 2A). Hepatic overexpression of PANDER significantly increased liver weight (P < 0.05) (data not shown) and TG content (P < 0.01) (Fig. 2B), but had little effect on hepatic CHO level (Supporting Fig. 4). PANDER-overexpressed livers exhibited typical morphological changes of steatosis with excessive neutral lipid deposition as assessed by way of Oil Red O staining (Fig. 2C) and resulted in a marked increase in serum TG level (Fig. 2D), very low-density lipoprotein (VLDL)-TG levels (Fig. 2E) and VLDL-CHO levels (Supporting Fig. 4), with little effect on fasting blood glucose levels (data not shown). In addition, serum insulin levels were found to be significantly elevated by hepatic overexpression of PANDER (data not shown). Hepatic PANDER overexpression also induced insulin resistance as assessed by a minimal model analysis (Supporting Fig. 5A,B). Liver overexpression of PANDER resulted in a significant increase in the number of balloon cells, with little change in glycogen, collagen, and tumor necrosis factor α levels. Hepatic monocyte chemoattractant protein 1 expression appeared to be induced by PANDER overexpression (Supporting Fig. 6). No change of hepatic and adipose adiponectin expression and serum adiponectin concentrations were observed after PANDER overexpression (Supporting Fig. 7).

Figure 2.

Hepatic and plasma lipid levels in C57Bl/6 mouse liver with PANDER overexpression by Ad-PANDER. (A) Protein expression of hepatic PANDER. *P < 0.05 versus Ad-GFP (n = 10). (B) Effect of hepatic overexpression of PANDER on liver TG content. **P < 0.01 (n = 10). (C) Macroscopic appearance (upper) and Oil Red O staining (lower) of livers infected with Ad-GFP (left) and Ad-PANDER (right). (D) Serum TG level at 0 and 4 days after treatment. ***P < 0.001 (n = 10). (E) Quantitative determination of serum lipid profile by way of high-performance liquid chromatography. The results are expressed as the mean ± SEM.

Hepatic Overexpression of PANDER Increases FOXO1 Expression and Suppresses Akt and Adenosine Monophosphate–Activated Protein Kinase Activity in Mouse Liver.

Western blot analysis revealed that PANDER overexpression resulted in a significant increase in FOXO1 levels, with decreased level of phosphorylated-FOXO1 (Fig. 3A). PANDER overexpression also resulted in a marked decrease in phosphorylated Akt (pAkt) and phosphorylated adenosine monophosphate–activated protein kinase (pAMPK) levels without affecting total Akt and AMPK expression (Fig. 3A).

Figure 3.

Hepatic PANDER overexpression increased FOXO1 level and lipogenic gene expression in mouse liver. (A) Western blot analysis of expression of FOXO1, Akt, and AMPK. *P < 0.05 versus Ad-GFP (n = 6). (B) Immunoblotting assay of expression of PPARγ, SREBP-1, and FAS. *P < 0.05, **P < 0.01 (n = 5). (C) Real-time PCR analysis of lipogenic gene expression of PPAR-γ, SREBP-1, and FAS in the liver. *P < 0.05 (n = 5). The results are expressed as the mean ± SEM.

Effect of PANDER Overexpression on Expression of Genes Involved in Lipid Metabolism in the Liver.

Western blot analysis demonstrated that hepatic peroxisome proliferator-activated receptor γ (PPARγ), sterol regulatory element binding protein (SREBP)-1C and fatty acid synthase (FAS) expression was markedly up-regulated by PANDER overexpression (Fig. 3B). Consistently, real-time PCR analysis further revealed that the mRNA levels of PPARγ, SREBP-1, and FAS were up-regulated by two- to 10-fold with PANDER overexpression (Fig. 3C). In addition, the mRNA levels of GPAT, medium chain acyl coenzyme A dehydrogenase, and stearoyl coenzyme A desaturase-1 were also significantly up-regulated (data not shown).

Knockdown of Hepatic PANDER Ameliorates Fatty Liver and Insulin Resistance in db/db Mice.

Hepatic PANDER expression was decreased approximately 70% at both mRNA and protein levels in db/db mice with PANDER knockdown by siRNA (Fig. 4A,B). Hepatic PANDER gene silencing significantly reduced mouse liver weight (data not shown) and hepatic TG content (Fig. 4C). PANDER knockdown also significantly ameliorated hepatic steatosis of db/db mice as evaluated by morphological changes and Oil Red O staining (Fig. 4D). Although serum TG concentration was not significantly reduced (Fig. 4E), VLDL-TG level was decreased after PANDER knockdown (Fig. 4F). Hepatic PANDER inhibition had no effect on liver CHO content, serum CHO levels, or plasma CHO profile (Supporting Fig. 8). Notably, hepatic PANDER gene silencing significantly reduced fasting blood glucose levels and improved insulin resistance in db/db mice (Supporting Figs. 5C,D and 9). There was no change in serum alanine aminotransferase and aspartate aminotransferase activity, indicating that knockdown of PANDER had little effect on liver enzymes (Supporting Fig. 10).

Figure 4.

Knockdown of hepatic PANDER gene by siRNA ameliorated fatty liver in db/db mice. (A) Hepatic PANDER mRNA expression. *P < 0.05 versus scramble siRNA (n = 8). (B) Hepatic PANDER protein level. **P < 0.01 versus scramble siRNA (n = 8). (C) Hepatic TG content. ***P < 0.001 versus scramble siRNA (n = 8). (D) Macroscopic appearance (upper) and Oil Red O staining (lower) of the livers of a db/m mouse (left) and db/db mouse receiving scramble siRNA (middle) or siRNA PANDER (right). (E) Serum TG levels. (F) Serum TG profile. The results are expressed as the mean ± SEM. db+NS, db/db mice treated with saline; db+scra, db/db mice treated with scramble siRNA; db+siPAN, db/db mice treated with PANDER siRNA.

Hepatic Knockdown of PANDER Reduces FOXO1 Expression and Increases Akt and AMPK Activity in db/db Mice.

PANDER knockdown in the liver resulted in a marked reduction in FOXO1 expression (Fig. 5A). Notably, specific knockdown of PANDER resulted in a marked increase in pAkt and pAMPK levels and the ratios of pAkt/Akt and pAMPK/AMPK (Fig. 5A).

Figure 5.

Effect of hepatic PANDER knockdown on expression of genes involved in lipid and glucose metabolism in db/db mouse liver. (A) Immunoblot assay of FOXO1, Akt, and AMPK expression. *P < 0.05 versus scramble siRNA for all of Fig. 5 (n = 5). (B) Western blot analysis of PANDER gene silencing on PPARγ, SREBP-1, and FAS protein expression. *P < 0.05 (n = 5). (C) Real-time PCR analysis of mRNA expression of PPARγ, SREBP-1, and FAS in the liver after PANDER knockdown. (D) Real-time PCR analysis of mRNA expression of glucokinase and G6Pase after PANDER knockdown. *P < 0.05 (n = 5). The results are expressed as the mean ± SEM.

Effect of Hepatic PANDER Knockdown on Expression of Genes Involved in Lipid and Glucose Metabolism in db/db Mice.

Immunoblot analysis demonstrated that knockdown of PANDER significantly lowered the protein expression of PPARγ, SREBP-1, and FAS in the livers of db/db mice (Fig. 5B). Real-time PCR assay further showed that the mRNA expression of PPARγ, SREBP-1, and FAS was markedly reduced in the livers of db/db mice with hepatic PANDER gene knockdown (Fig. 5C). The mRNA levels of other lipogenic genes including acetyl coenzyme A carboxylase-1, medium chain acyl coenzyme A dehydrogenase, and CD36 were also reduced by hepatic PANDER gene knockdown (data not shown). In addition, hepatic PANDER knockdown significantly reduced glucokinase and G6P expression in the livers of db/db mice (Fig. 5D).

PANDER Induces Lipid Accumulation and Blocks Insulin-Induced Akt Activation in Hepatocytes.

Infecting HepG2 cells with Ad-PANDER for 48 hours to overexpress PANDER markedly induced lipid accumulation (Fig. 6A), with a significant up-regulation of FOXO1 (Fig. 6B). PANDER overexpression significantly attenuated Akt activity at basal levels and after treatment with 100 nM insulin for 5, 10, and 15 minutes (Fig. 6C,D). In addition, insulin-stimulated phosphorylation of FOXO1 was also significantly inhibited by PANDER overexpression by approximately 60% (Fig. 6E,F).

Figure 6.

PANDER overexpression with Ad-PANDER promoted lipid accumulation and blocked insulin-stimulated Akt and FOXO1 phosphorylation in HepG2 cells. (A) Morphologic change (upper) and Oil Red O staining (lower) of HepG2 cells. Overexpression of PANDER by Ad-PANDER for 48 hours markedly increased lipid droplets and neutral lipid content in HepG2 cells. (B) Western blot analysis of induction of FOXO1 protein expression by PANDER overexpression in HepG2 cells. *P < 0.05 (n = 4). (C) Suppression of insulin (INS)-stimulated Akt phosphorylation by PANDER overexpression. HepG2 cells were infected with Ad-PANDER or Ad-GFP (multiplicity of infection, 50) for 48 hours and then treated with 100 nM insulin for 0, 5, 10, and 15 minutes. (D) Quantitative measurement of the levels of pAkt. *P < 0.05 versus basal levels in Ad-GFP infected cells. &P < 0.05 versus pAkt levels in Ad-GFP-treated cells at each time point. (E) Western blot analysis of FOXO1 and Akt phosphorylation by PANDER overexpression. Ad-PANDER–treated HepG2 cells were stimulated with insulin (INS) for 5 minutes, and the levels of pAkt and pFOXO1 in the cell lysate were assayed. (F) Quantitative measurement of the levels of pFOXO1 with PANDER overexpression. The results are expressed as the mean ± SEM of at least five independent experiments.

FOXO1 Induces PANDER Expression In Vitro and In Vivo.

As shown in Fig. 7A, FOXO1 overexpression by way of tail vein delivery of Ad-FOXO1 resulted in a marked accumulation of neutral lipids in the livers of C57Bl/6mice as described.21 Hepatic overexpression of FOXO1 also significantly increased PANDER protein expression (Fig. 7B) and resulted in a significant increase in gene expression of FAS, acetyl coenzyme A carboxylase-1, and GPAT in the liver (data not shown). Consistently, overexpression of FOXO1 also markedly increased PANDER protein level (Fig. 7C) in cultured HepG2 cells.

Figure 7.

FOXO1 overexpression with Ad-FOXO1 induced PANDER expression in mouse liver and cultured hepatocytes. (A) FOXO1 promotes hepatic lipogenesis in C57Bl/6 mice. Ad-FOXO1 and Ad-GFP (1.5 × 109 plaque-forming units) were injected by way of the tail vein. Oil Red O staining of livers was performed 1 week after viral infection to assay lipid accumulation in the liver with FOXO1 overexpression. (B) FOXO1 increased PANDER expression in the liver. Western blot analysis of hepatic PANDER expression in the livers with FOXO1 overexpression is shown. (C) FOXO1 up-regulated PANDER expression in HepG2 cells. Western blot analysis of PANDER protein level in HepG2 cells infected with Ad-FOXO1 and Ad-GFP (multiplicity of infection, 20) for 48 hours is shown. (D) Schematic model showing the proposed major mechanisms by which PANDER regulates lipid and glucose metabolism in the liver.

Discussion

Because hepatic insulin resistance is involved in the development of NAFLD, the purpose of the current study was to determine whether liver-produced PANDER participates in the pathogenesis of hepatic steatosis. We found PANDER constitutively expressed in the liver, where its expression was increased in db/db and HFD-fed mice. Selective overexpression of PANDER in the liver resulted in a massive lipid accumulation with increased FOXO1 expression and suppression of Akt and AMPK activity in normal mice. Hepatic knockdown of PANDER significantly attenuated liver steatosis with decreased FOXO1 expression and increased Akt and AMPK activity in db/db mice. PANDER overexpression induced lipogenesis, increased FOXO1 expression, and blocked basal and insulin-elicited Akt activation in cultured hepatocytes. Moreover, FOXO1 increased PANDER expression both in cultured hepatocytes and in normal mouse livers. Thus, hepatic PANDER promotes lipogenesis through a FOXO1-dependent pathway and may represent a potential therapeutic target for treating fatty liver.

In islet β cells, PANDER cosecretes with insulin in response to glucose8 and antagonizes insulin action in both an autocrine and a paracrine manner.8, 9 PANDER may also act as a hormone, because it can bind to liver cell membranes and inhibit insulin signaling in HepG2 cells.9 We found PANDER expression significantly up-regulated in the livers of db/db mice and mice fed an HFD and attenuated by swimming exercise, which was concomitant with an amelioration of fatty liver, thereby supporting the idea that PANDER promotes lipogenesis in the liver. Indeed, overexpression of PANDER induced massive hepatic lipid accumulation, with elevated plasma VLDL-TG levels in normal mice, whereas silencing PANDER expression resulted in a marked attenuation of fatty liver, with a significant reduction in plasma VLDL-TG concentration in db/db mice. Therefore, hepatic PANDER is an important endogenous factor in regulating lipid metabolism and contributes to the pathogenesis of fatty liver by enhancing lipogenesis in the liver.

Multiple mechanisms are involved in the regulation of liver lipid homeostasis.1 Among them, FOXO1 plays a key role in hepatic glucose and lipid metabolism, and increased activity of FOXO1 is associated with the progression of nonalcoholic fatty liver13, 21 and increased hepatic production of VLDL-TG.13, 22 We found that PANDER overexpression significantly increased FOXO1 expression in normal mice. In contrast, PANDER knockdown markedly attenuated FOXO1 expression in db/db mice. Given that overexpression of FOXO1 in the liver induced a phenotype similar to PANDER overexpression (up-regulation of lipogenic genes and increased VLDL secretion),13, 23 these findings suggest that PANDER may accelerate lipogenesis by way of increased FOXO1 activity. Furthermore, we found that FOXO1 is an important regulator of PANDER expression. FOXO1 can induce PANDER expression both in cultured hepatocytes and in mouse livers, which implies that a positive feedback loop between PANDER and FOXO1 may play an important role in sustaining excessive lipogenesis seen in the steatotic liver (Fig. 7D).

Although it is clear FOXO1 plays an important role in lipid metabolism, whether FOXO1 promotes lipogenic processes in the liver remains under debate. Gain-of-function studies using mice with liver overexpression of FOXO1 and loss-of-function experiments using mice with liver-specific deletion of FOXO1 have yielded conflicting and controversial results, with claims ranging from strongly positive associations1, 21, 22 to strongly negative associations.24–26 Differences are likely to be due to a number of factors, including the genetic backgrounds of the mice, the level of overexpression or insufficiency of FOXO1, insulin sensitivity, as well as feeding protocol (fed versus fasting). However, consistent with a previous report,21 our present study supports the possibility that FOXO1 overexpression enhances lipogenesis in the liver.

At the present, the mechanisms involved in the reciprocal activation between PANDER and FOXO1 remain unclear. FOXO1 level can be regulated by multiple posttranslational modifications that affect its subcellular localization, DNA binding, and transcriptional activity. A major form of regulation is Akt-mediated phosphorylation of FOXO1 in response to insulin or growth factors. Akt phosphorylates FOXO1 and deactivates its transcriptional activity.27, 28 We found that PANDER overexpression markedly reduced Akt activity, whereas PANDER knockdown significantly increased it in mouse livers. PANDER also ameliorated insulin-elicited Akt phosphorylation in cultured hepatocytes. These findings suggest that PANDER induces FOXO1 expression, at least in part, through an Akt-dependent mechanism. In addition to Akt, AMPK can phosphorylate FOXO1 protein and deactivate its activity.29 Suppression of AMPK activation by PANDER may therefore contribute to increased FOXO1 activity in the liver as well (Fig. 7D).

In addition to the Akt-FOXO1 pathway, other mechanisms may also contribute to the lipogenic effect of PANDER in the liver. Notably, in PANDER-overexpressed mouse livers, PPARγ level was significantly up-regulated, whereas hepatic knockdown of PANDER gene expression resulted in a marked reduction in PPARγ expression in db/db mouse livers. Because of the important role of PPARγ in adiponenesis30, 31 and hepatic lipid metabolism,32, 33 our findings strongly suggest that induction of PPARγ expression is involved in the steatotic effect of PANDER. Consistent with these findings, PPARγ target genes responsible for lipogenesis, including SREBP-1, FAS, and CD36, were significantly up-regulated after PANDER overexpression and markedly attenuated with PANDER gene silencing in mouse livers. The induction of the PPARγ gene by PANDER may not be associated with increased FOXO1 expression, because several lines of evidence demonstrate that FOXO1 binds to the PPARγ promoter to repress its transcriptional activity,34 and FOXO1 gene semideficiency is associated with an increase in adipose PPARγ expression.35 Taken together, although the mechanism by which PANDER induces PPARγ expression is currently unknown, increased levels of PPARγ could at least in part account for the excessive lipogenic process in the liver induced by PANDER, thus contributing to the pathogenesis of hepatic steatosis.

Akt, AMPK, and FOXO1 are crucial regulators in hepatic glucose metabolism.29, 35, 36 We failed to observe glucose intolerance and hyperglycemia, a finding that to some extent was inconsistent with that of Wilson et al. that overexpression of PANDER induced glucose intolerance and hyperglycemia in wild-type mice. However, we did observe that overexpression of PANDER in the liver resulted in a significant increase in circulating insulin levels and hepatic glucose-6-phosphatase (G6Pase) expression, which were in line with the observations of this study.12 The difference may be due to the time length of PANDER overexpression. In our study, we performed glucose tolerance test 3 days after virus injection, whereas Wilson et al. conducted their study 7 days after virus administration. Further supporting the role of PANDER in hepatic glucose metabolism, hepatic PANDER gene silencing significantly lowered fasting blood glucose levels, with a significant reduction in G6Pase expression in livers, in insulin-resistant db/db mice. Furthermore, according to a minimal model analysis, PANDER overexpression decreased peripheral insulin sensitivity, whereas knockdown of hepatic PANDER increased insulin sensitivity. Collectively, these results indicate that PANDER could enhance liver gluconeogenesis, possibly by increasing G6Pase activity, which may be a result of suppressed Akt/FOXO1 and AMPK/FOXO1 signaling pathways.

In the present study, we provide evidence that PANDER is expressed in human liver samples. In the human hepatic cell line HepG2, PANDER represses Akt activation, up-regulates FOXO1 activity, and induces lipid deposition. In addition, free fatty acids significantly activate PANDER expression in HepG2 cells (data not shown). Because FOXO1 expression and activity have been reported to be increased in human steatohepatitic livers and are correlated with the severity of nonalcoholic steatohepatitis,21 we speculate that PANDER expression may be increased in human steatotic liver, which contributes to the progression of human fatty liver and other related metabolic disorders through a FOXO1-mediated pathway (Fig. 7D).

It should be noted that the experimental animal models used in the current study may not resemble NAFLD. The effect of long-term hepatic overexpression of PANDER on liver lipid metabolism may provide more insight into the role of PANDER in the pathogenesis of NAFLD. Generation of a transgenic mouse with selective PANDER expression in the liver will be useful in addressing this important issue.

In conclusion, we demonstrate that PANDER is a critical metabolic regulator in the liver. Hepatic PANDER promotes lipogenesis and compromises insulin signaling in the liver, mainly through an increase in FOXO1 activity. The reciprocal activation between PANDER and FOXO1 and attenuated Akt/FOXO1 and AMPK/FOXO1 signaling may represent major mechanisms involved in PANDER-mediated hepatic lipogenesis and insulin resistance. PANDER may be a potential therapeutic target for the treatment of hepatic steatosis and related metabolic disorders.

Acknowledgements

We thank Laura Heraty and Tiao Guan for assistance in editing the manuscript.

Ancillary

Advertisement