Angiopoietin-Related Growth Factor Suppresses Gluconeogenesis through the Akt/Forkhead Box Class O1-Dependent Pathway in Hepatocytes

Type 2 diabetes mellitus (T2DM) is one of the most common endocrine diseases in developed countries (Fagot-Campagna et al., 2005). Excessive hepatic glucose production is thought to be a major contributor to the T2DM status (Spiegelman and Flier, 2001; Basu et al., 2004; Radziuk and Pye, 2004). Prolonged elevation of blood glucose levels can lead to many diabetic complications, such as cardiovascular disease, stroke, diabetic neuropathy, and retinopathy, etc. (Fagot-Campagna et al., 2005). Thus, the discovery of antidiabetic agents that can inhibit hepatic glucose production is a focus of research for the pharmaceutical community. Although some hypoglycemic drugs have been developed for the treatment of diabetes (Curtis et al., 2005; Das and Chakrabarti, 2005), more efficacious agents are needed.

AGF is a liver-derived circulating factor (Oike et al., 2003) that counteracts high-fat-induced obesity and related insulin resistance through increased energy expenditure (Oike et al., 2005). Both visceral and s.c. fat depots were significantly increased in AGF-deficient mice, and sections of white adipose tissue from these mice showed increased adipocyte size relative to those from wild-type mice. Furthermore, a large amount of lipid accumulation in liver, skeletal muscle, and brown adipose tissue was observed in AGF-deficient mice. These mice also showed significant decreases in whole-body oxygen consumption rates compared with wild-type mice, whereas no increase was observed in daily food intake in AGF-deficient mice compared with wild-type mice, indicating that inactivation of AGF in vivo leads to decreased energy expenditure and obesity (Oike et al., 2005). Thus, AGF is a potential target for pharmacological interventions counter-acting obesity and related metabolic diseases. However, the target tissues and the molecular mechanisms by which AGF counteracts obesity and related insulin resistance are not well defined.

The liver plays a central role in maintaining glucose homeostasis by regulating enzymes that are involved in the processes of glycogenolysis and gluconeogenesis (Nordlie et al., 1999). In this study, we provide the first evidence that AGF inhibits glucose production through decreased expression of glucose-6-phosphatase (G6Pase) in rat hepatocytes. We also demonstrate a detailed mechanism by which AGF inhibits gluconeogenesis in H4IIEc3 cells.

Materials and Methods

Cell Culture. The rat hepatoma cell line H4IIE was cultured in minimal essential medium containing 10% fetal bovine serum, 0.1 mM nonessential amino acid solution, sodium pyruvate solution, and 100 U/ml penicillin/100 μg/ml streptomycin and maintained at 37°C in 5% CO₂. Primary rat hepatocytes were isolated as described previously (Shih and Towle, 1999). In brief, primary hepatocytes were isolated from male Wistar rats (6W-9W) using the collagenase perfusion method. Hepatocytes were plated on collagen-coated 3.5-cm dishes at 1 × 10⁶ cells/dish in HepatoZYME-SFM (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum and 100 U/ml penicillin/100 μg/ml streptomycin and maintained at 37°C in 5% CO₂.

Purification of Recombinant AGF Protein. Purification of recombinant AGF protein was described by Oike et al. (2004). In brief, the conditioned medium from transfected human embryonic kidney 293 cells was collected and passed through a 0.22-μm-pore size filter (Millipore, Bedford, MA). To purify AGF protein, the filtered conditioned medium was transferred to an anti-FLAG antibody (M2) affinity gel (Sigma-Aldrich, St. Louis, MO), and only AGF protein was trapped in the gel. After the gel was washed with phosphate-buffered saline (PBS), protein was eluted by adding Gly-HCl, pH 3.0, and immediately neutralized with Tris, pH 8.0. AGF protein was dialyzed in PBS overnight at 4°C. Protein concentration was measured with the BCA protein assay kit (Pierce Chemical, Rockford, IL).

Glucose Production Assay. Cells were cultured on 48-well plates at a density of 2 × 10⁵ (H4IIEc3 cells) or 1 × 10⁵ (primary hepatocytes) cells/well. After a 6-h attachment period, cells were treated with or without AGF in serum-free medium for 21 h. These cells were washed twice with PBS to remove glucose and then incubated for 3 h in 100 μl of glucose production medium [glucose- and phenol red-free Dulbecco's modified Eagle's medium containing sodium pyruvate (2 mM)]. Fifty microliters of medium was sampled for measurement of glucose concentration using an Amplex Red Glucose/Glucose Oxidase Assay Kit (Invitrogen). Glucose concentrations were normalized to cellular protein concentrations measured by the BCA protein assay kit. Student's t test was used for statistical analysis.

Preparation of Total RNA and Quantitative PCR. Cells were plated on 35-mm dishes at a density of 2 × 10⁶ (H4IIEc3 cells) or 1 × 10⁶ (primary hepatocytes) cells/dish. After a 6-h attachment period, cells were treated with or without AGF in serum-free medium for 24 h, and then total RNA was isolated using the QIAGEN RNeasy mini kit (QIAGEN, Valencia, CA). Quantitative PCR was performed with ABI Prism 7700 and SYBR Green Reagent (Applied Biosystems, Foster City, CA). cDNA was synthesized from 0.5 μg of total RNA using SuperScript II reverse transcriptase (Invitrogen). Absolute cDNA abundance was calculated using the standard curve obtained from rat genomic DNAs. PCR conditions were as follows: 10 min at 95° C, then 50 cycles of 15 s at 94°C, 30 s at 60°C, and 1 min at 72°C. Levels of mRNA were normalized using β-actin as the housekeeping gene. The following oligonucleotide DNA primers were synthesized: for rat G6pc, the forward primer was 5′-TGAAACTTTCAGCCACATCCG-3′, and the reverse primer was 5′-GCAGGTAAAATCCAAGTGCGAA-3′; for rat Pck1, the forward primer was 5′-TCCCATTGGCTACGTCCCTAA-3′, and the reverse primer was 5′-CCACCTCCTTCTCCCAGAATTC-3′; and for rat β-actin, the forward primer was 5′-GGTCGTACCACTGGCATTGTG-3′, and the reverse primer was 5′-GCTCGGTCAGGATCTTCATGAG-3′.

Isolation of Whole-Cell, Nuclear, and Cytoplasmic Extracts and Western Blot Analyses. Cells were plated on 35-mm dishes at a density of 2 × 10⁶ (H4IIEc3 cells) or 1 × 10⁶ (primary hepatocytes) cells/dish. After a 6-h attachment period, cells were treated with or without AGF in serum-free medium for the indicated time. These cells were washed twice with ice-cold PBS and treated with lysis buffer [1 mM EDTA, 1% SDS, Complete phosphatase inhibitors cocktail (Sigma-Aldrich), 10 mM HEPES-HCl, pH 7.5] The cell lysates were boiled for 10 min and centrifuged to obtain the supernatants as cell extracts. Cellular protein concentrations were measured with the BCA protein assay kit (Pierce Chemical). Cytoplasmic or nuclear proteins from H4IIEc3 cells were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents based on the manufacturer's instruction (Pierce Chemical). SDS sample buffer (2×) was added to the cell lysates. After boiling for 5 min, the samples were separated by 10% SDS-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane using a semidry apparatus. After soaking in 5% nonfat milk for 1 h at room temperature, the membrane was incubated with primary antibodies for 16 to 20 h at 4°C. The specific signals were amplified by addition of horseradish peroxidase-conjugated secondary antibodies and visualized using an enhanced chemiluminescence system.

Immunohistochemistry. Cells were plated at 5 × 10⁴ cells/dish. After a 6-h attachment period, cells were treated with or without AGF in serum-free medium for the indicated time and fixed with 3.7% formaldehyde in PBS and permeabilized with 0.5% Triton X-100. FoxO1 was visualized using anti-FoxO1 polyclonal antibody (1 μg/ml, final concentration) and fluorescein isothiocyanate-conjugated Affinity Purified donkey anti-rabbit IgG (20 μg/ml, final concentration). Slides were examined using an Eclipse TE 300 fluorescence microscope (Nikon, Tokyo, Japan).

Results

AGF Suppressed Gluconeogenesis via Decreased Expression of G6Pase in Hepatocytes. In both type 1 and type 2 diabetes, excessive hepatic glucose production is a major contributor of both fasting and postprandial hyperglycemia. To examine the influence of AGF on hepatic glucose metabolism, we measured glucose production in H4IIEc3 cells. Treatment of the cells with AGF induced a concentration-dependent reduction in the amount of glucose produced (Fig. 1A). The rate of gluconeogenesis is controlled principally by the activities of unidirectional enzymes such as phosphoenolpyruvate carboxykinase (PEPCK) and G6Pase. PEPCK catalyzes one of the rate limiting steps of gluconeogenesis, the conversion of oxaloacetate to phosphoenolpyruvate, whereas G6Pase catalyzes the final step of gluconeogenesis, the production of free glucose from glucose 6-phosphate. The genes of these gluconeogenic enzymes are controlled at the transcriptional level by hormones, mainly insulin, glucagon, and glucocorticoid. Insulin inhibits gluconeogenesis by suppressing the expression of PEPCK and G6Pase, whereas glucagon and glucocorticoids stimulate hepatic glucose production by inducing these genes (O'Brien and Granner, 1996). To examine the influence of AGF on the expression of these enzymes, we measured the levels of Pck and G6pc mRNAs using quantitative PCR in AGF-treated H4IIEc3 cells. Decreased expression of G6pc mRNA was observed after 6 h of treatment with 1 μg/ml AGF and remained depressed for at least 24 h (Fig. 1C). Likewise, expression of G6pase protein was also decreased by 1 μg/ml AGF after 24-h treatment (Fig. 1). In contrast, PEPCK was not affected by AGF at either the mRNA or protein levels (Fig. 1B). We also assessed the effects of AGF on glucose production and expression of G6Pase in primary rat hepatocytes. AGF reduced glucose output by 30% in a concentration-dependent manner (Fig. 2A). At 1 μg/ml, AGF decreased expression of G6Pase mRNA (Fig. 2B) and protein (Fig. 2C) in cultured rat hepatocytes.

Involvement of Akt in the AGF-Induced Suppression of Gluconeogenesis. Next, we tried to explore the upstream mechanism of AGF-mediated depression of G6pase expression. It has been suggested that Akt signaling is involved in epidermal proliferation induced by AGF (Oike et al., 2003). Moreover, Akt modulates G6Pase expression by suppressing the activity of FoxO1, a member of the forkhead family of transcription factors (Accili and Arden, 2004; Barthel et al., 2005) and regulates glycogen synthesis by inhibition of glycogen synthase kinase-3β (GSK-3β) (Summers et al., 1999). These observations suggested that Akt might be involved in the AGF-modulated reduction in glucose production.

To determine whether Akt actually plays a role in the AGF-mediated reduction of glucose production, we used a variety of PI3K-Akt signaling inhibitors and monitored glucose production. The reduction of glucose production mediated by 1 μg/ml AGF was blocked by pretreatment of the cells with 1l-6-hydroxymethyl-chiro-inositol 2-[(R)-2-O-methyl]-3-O-octadecylcarbonate (Akt inhibitor I) (Hu et al., 2000), Triciribine (Akt inhibitor V) (Kim et al., 2005), and LY294002, a specific PI3K inhibitor, in a concentration-dependent manner (Fig. 3).

AGF-Induced Phosphorylation of Akt, GSK-3β, and FoxO1. Treatment of H4IIEc3 hepatoma cells with AGF increased the phosphorylation of Akt (Fig. 4A) in a concentration-dependent manner. AGF-mediated enhancement of Akt phosphorylation reached a maximal level 10 min after AGF addition (Fig. 4B), followed by subsequent phosphorylation of GSK-3β and FoxO1 (Guo et al., 1999) (Fig. 4, A and B).

AGF-Accelerated Nuclear Export of FoxO1. FoxO1 is expressed in the liver, where it can stimulate the expression of several genes promoting hepatic glucose production, including G6Pase, both in liver cells and in vivo (Puigserver et al., 2003). Signaling by insulin and growth factors results in the phosphorylation of FoxO proteins at three highly conserved Akt kinase sites, corresponding to Thr24, Ser256, and Ser319 in human FoxO1 (Brunet et al., 1999; Kops et al., 1999; Rena et al., 1999). Phosphorylation at these sites suppresses transcriptional activity and promotes the redistribution of FOXO proteins outside of the nucleus and localization in the cytoplasmic compartment (Guo et al., 1999; Kops et al., 1999; Rena et al., 2001). The regulation of the nuclear/cytoplasmic distribution of FoxO proteins is thought to be important in determining their biological activity. Thus, we used subcellular fractionation and Western blotting to ask whether treatment with AGF influenced the localization of FoxO1. In H4IIEc3 cells, translocation of FoxO1 from nucleus to cytoplasm by treatment with 1 μg/ml AGF appeared 1 h after stimulation and lasted for at least 6 h (Fig. 5A). In the absence of AGF, fluorescent signals derived from FoxO1 were present in both nucleus and cytoplasm. After AGF treatment at 1 μg/ml for 3 h, FoxO1 fluorescence translocated to the cytoplasm (Fig. 5B). These findings demonstrate for the first time that nuclear export of FoxO1 was stimulated by AGF in H4IIEc3 hepatoma cells.

Discussion

T2DM results from a subnormal response of tissues to insulin (insulin resistance) and a failure of the insulin-secreting β cells to compensate. Hepatic insulin resistance is characterized by the inefficiency of insulin to suppress hepatic glucose production. The disruption of hepatic glucose metabolism is involved in the development of insulin resistance (Wu et al., 2007). Overexpression of G6Pase in liver is sufficient to perturb whole-animal glucose and lipid homeostasis, possibly contributing to the development of metabolic abnormalities associated with diabetes (Trinh et al., 1998). Impairments in the insulin signal transduction pathway and hepatic glucose metabolism appear to be critical pathologic factors contributing to insulin resistance and T2DM. In the present study, we demonstrated for the first time that AGF reduced glucose production through decreased expression of G6Pase in rat cultured H4IIEc3 cells (Fig. 1) and primary hepatocytes (Fig. 2). It was also demonstrated that the Akt was involved in AGF-induced suppression of gluconeogenesis resulting from phosphorylation of glycogen synthase kinase 3β and Foxo1, substrates of Akt. These results may explain how AGF counteracts obesity and insulin resistance. AGF is, therefore, implicated in the progression toward diabetes, especially T2DM.

G6Pase and PEPCK are such proteins whose expression is regulated by insulin, whereas expression of PEPCK was not affected by AGF. The insulin receptor is a tyrosine kinase that, when activated by insulin, phosphorylates members of the family of insulin receptor substrates. This leads to the recruitment of PI 3-kinase to the plasma membrane, where it is involved in the phosphorylation of phosphatidylinositol 4,5-bisphosphate to phosphatidylinositol 3,4,5-trisphosphate, the second messenger of the insulin receptor, as well as a variety of growth factor receptors. Phosphatidylinositol 3,4,5-trisphosphate indirectly activates Akt. Activation of Akt in hepatoma cells is sufficient to repress the glucocorticoid- and cAMP-promoted induction of PEPCK gene transcription (Agati et al., 1998; Liao et al., 1998), and the overexpression of the gene for Akt in cultured hepatocytes decreases the levels of mRNA for PEPCK and G6Pase (Schmoll et al., 2000). This suggests that the downstream targets of Akt, transcription factors such as FoxO1, hepatocyte nuclear factor-3β, and cAMP response element binding, are involved in the effect of insulin on the transcription of the genes for these two gluconeogenic enzymes. However, Kotani et al. (1999) have shown that a dominant negative form of Akt does not prevent insulin's inhibition of transcription for the PEPCK gene promoter in hepatoma cells. Furthermore, overexpression of Foxo1 markedly increased the expression of the catalytic subunit of G6Pase. In contrast, both basal and dexamethasone/cAMP-induced levels of PEPCK mRNA were unaffected by Foxo1-overexpression (Barthel et al., 2001). These observations suggested that suppression of PEPCK expression by insulin was not only regulated by Akt-FoxO1 signaling cascades, but the other cascades coregulate this function. Moreover, AGF mediated the adhesion and migration cellular events through interaction with Arg-Gly-Asp (RGD)-binding integrins (Zhang et al., 2006). The integrin signaling via RGD peptides induced an increase of phospho-Akt (Pinkse et al., 2005). Taking into consideration these reports, we assumed that AGF activates Akt using the different signaling cascades from insulin. Although activation of Akt occurs, AGF does not affect PEPCK expression.

The most important finding of the present study is that AGF stimulates nuclear export of FoxO1. K14-AGF transgenic mice (in which expression of AGF is forced in epidermal cells) show keratinocyte proliferation and increased cutaneous wound healing, suggesting that AGF promotes keratinocyte growth (Oike et al., 2003). In addition, significant immunoreactivity is detected in the epidermis of K14-AGF mice by using an antibody directed against phospho-Ser473 of Akt compared with that of controls, suggesting that Akt signaling is involved in epidermal proliferation induced by AGF (Oike et al., 2003). Furthermore, K14-AGF mice exhibit increased vascular permeability and increased numbers of blood vessels, suggesting that AGF has angiogenic activity like other angiopoietins and angiopoietin-related proteins, such as Ang-1, angiopoietin-related protein 2, and ANGPTL3 (Oike et al., 2004). In addition, inhibition of FoxO1 transcriptional activities by nuclear export or gene silencing promotes endothelial cell proliferation (Fosbrink et al., 2006) and angiogenesis and postnatal neovascularization (Potente et al., 2005). Taken together with these reports, FoxO1 may be a key molecule in the function of AGF in glucose metabolism, cell proliferation, and angiogenesis.

In summary, we found that AGF induced PI3K/Akt/FoxO1 signaling cascades and thereby reduced G6Pase protein expression in hepatocytes. These results suggest that AGF plays a crucial role in regulating glucose metabolism and in the pathogenesis of T2DM. The elucidation of the molecular and cellular mechanisms underlying AGF-induced modulation of metabolic gene expression via FoxO1-dependent transcription may enable us to develop new therapies for diabetes.