Endocrinology and Metabolism

Abstract

We showed that the rat Na+/Pi cotransporter-1 (RNaPi-1) gene was regulated by insulin and glucose in rat hepatocytes. The aim of this work was to elucidate signaling pathways of insulin-mediated metabolic regulation of the RNaPi-1 gene in H4IIE cells. Insulin increased RNaPi-1 mRNA abundance in the presence of glucose and decreased RNaPi-1 mRNA in the absence of glucose, clearly establishing an involvement of metabolic signals for insulin-induced upregulation of the RNaPi-1 gene. Pyruvate and insulin increased RNaPi-1 expression but downregulated L-pyruvate kinase, indicating the existence of gene-specific metabolic signals. Although fructose, glycerol, and lactate could support insulin-induced upregulation of the RNaPi-1 gene, compounds entering metabolism beyond pyruvate oxidation, such as acetate and citrate, could not, suggesting that RNaPi-1-specific metabolic signals are generated at or above pyruvate oxidation. Wortmannin, LY-294002, and rapamycin abolished the insulin effect on the RNaPi-1 gene, whereas expression of dominant negative Asn17 Ras and mitogen-activating protein kinase (MAPK) kinase (MEK) inhibitor PD-98059 exhibited no effect. Thus we herein propose that metabolic regulation of RNaPi-1 expression by insulin is mediated through the phosphatidylinositol 3-kinase/p70 ribosomal S6 kinase pathways, but not the Ras/MAPK pathway.

  • rat Na+/Pi cotransporter-1
  • insulin
  • Ras
  • phosphatidylinositol 3-kinase
  • p70 ribosomal S6 kinase

a major physiological function of insulin is to regulate the key enzymes and membrane transporters that are involved in glucose and lipid metabolism (7, 27). For insulin to exert these functions, it binds to its receptor at the plasma membrane of target cells and activates the tyrosine kinase that is associated with the cytoplasmic tail of the receptor (7, 27, 40). This causes phosphorylation of the receptor and many endogenous substrates, such as insulin receptor substrate-1 and -2, and Shc (7, 27). Phosphorylated substrates in turn engage in the formation of the signaling complexes via phosphotyrosine-containing binding motifs with Srchomology-2 domains found in molecules like phosphatidylinositol 3-kinase (PI 3-kinase) (11). These divergent intermolecular interactions underlie the basic mechanism for the multiple effects of insulin on the cell. For instance, two distinct signal pathways are involved in gene regulation by insulin: the Ras/mitogen-activating protein kinase (MAPK) pathway, which activates many transcription factors and leads to c-fos induction, and the PI 3-kinase pathway, which activates many genes involved in glucose and lipid metabolism (7, 29).

Cytosolic inorganic phosphate (Pi) plays a central role in cellular energy metabolism and in hormone-regulated glucose homeostasis (1, 2, 9). Administration of carbohydrates increases cellular Pi uptake, and extracellular Pi regulates hepatic glucose output (1). The hypoglycemic effect of insulin is markedly diminished when serum Pi levels are low (9). The direct effects of Pi on several key metabolic enzymes have also been well documented (2). For example, Pi activates 6-phosphofructo-1-kinase, which catalyzes the conversion of fructose 6-phosphate to fructose 1,6-diphosphate and increases glycolysis (2). It is now clear that intracellular Pi concentrations are tightly regulated by a variety of mechanisms including insulin (6, 12). After the original discovery that Pi transport across the luminal brush-border membrane of mammalian kidney proximal tubule cells was carrier mediated and Na+ dependent (16, 26), Na+/Pi cotransporters have been found in many different cells, including rat hepatocytes (21, 22). On the basis of amino acid sequence similarity, three types of Na+/Pi cotransporters have been cloned recently (26). Rat Na+/Pi cotransporter-1-related type I cotransporters are primarily expressed in the liver and kidney, whereas NaPi-2-related type II cotransporters are kidney specific and play a key role in kidney Pi reabsorption (22, 26). Recently, expression study in Xenopus oocytes indicates that RNaPi-1 type cotransporters may also act as Cl channels (3). The surface receptor for amphotropic murine retrovirus (Ram-1)-related type III cotransporters expresses in a variety of cells and may represent a housekeeping type of transporter (18). Although Pidepletion regulates both type II and type III cotransporters (32, 39), we have shown that expression of RNaPi-1 in rat hepatocytes, like many other genes of glycolytic and lipogenic enzymes (23-25), is regulated by insulin and glucose and that metabolic signals are involved in this regulation (21). In this study, we aimed to identify the metabolic components and the signaling pathways that mediate the effects of insulin and glucose on RNaPi-1 gene expression in H4IIE rat hepatoma cells that are devoid of endogenous gluconeogenic activity (28, 34, 37).

EXPERIMENTAL PROCEDURES

Materials.

Chemicals of the highest purity available were from Sigma (St. Louis, MO) and Boehringer Mannheim (Indianapolis, IN). TRI reagent for RNA isolation was from Molecular Research Center (Cincinnati, OH), and32P-labeled radionucleotides (∼3,000 Ci/mmol) were from Du Pont NEN (Boston, MA). Rabbit polyclonal anti-ACTIVE MAPK pAb and anti-p42/44 antibodies were obtained from Promega (Madison, WI) and New England Biolabs (Beverly, MA), respectively. An enhanced chemiluminescence kit was from Amersham (Cleveland, OH). All protein kinase inhibitors were purchased from Calbiochem (San Diego, CA). Optitran and Nytran membranes were obtained from Schleicher and Schuell (Keene, NH).

Cell culture.

H4IIE cells were cultured in DMEM supplemented with 10% fetal bovine serum, 0.1 μM dexamethasone, and penicillin (100 U/ml) and streptomycin (0.1 mg/ml) antibiotics. Cells were incubated in 95% air-5% CO2 in humidified incubators at 37°C. Approximately 14 h before treatment with effectors, the cultures were refed glucose- and serum-free DMEM supplemented with 0.1 μM dexamethasone and antibiotics as before.

Northern blot analysis.

Total RNA was isolated from cultured H4IIE cells using TRI reagent, as we previously described (21). RNA was electrophoresed through 1% agarose gels containing 1.9% formaldehyde, transferred to nylon membranes, and fixed by ultraviolet (UV) cross-linking. RNaPi-1 cDNA probe was made as previously described (21). Human β-actin and rat Ram-1 clones were kindly provided by Dr. Craig Thompson (University of Michigan, Ann Arbor, MI) and Dr. A. Dusty Miller (Fred Hutchinson Cancer Research Center, Seattle, WA), respectively. Rat L-type pyruvate kinase (L-PK) cDNA was PCR-cloned from rat liver mRNA and confirmed by DNA sequence. The cDNA probes were random labeled with α-[32P]dATP. The intensities of autoradiograms were scanned using a Bio-Rad imaging densitometer (model GS-670), and areas were quantified. For quantitation of the relative amount of mRNAs, the same blots were used for hybridizations with β-actin. The amounts of the RNaPi-1 and other mRNAs were normalized to the corresponding β-actin mRNA, as previously described (21).

Nuclear run-on transcription assay.

Nuclei were isolated from H4IIE cells as previously described (17). Briefly, cells were collected with ice-cold saline solution and pelleted at 700 g for 5 min. After two washes with the same saline solution, cells were suspended in 40 ml of lysis buffer [0.5 M sucrose, 50 mM NaCl, 0.5 mM spermidine, 0.15 mM phenylmethylsulfonyl fluoride (PMSF), 1% aprotinin, 7 mM β-mercaptoethanol, and 0.25% Nonidet P-40] and incubated on ice for 3 min. Nuclei were collected by centrifugation at 1,000g for 5 min and washed twice with the same lysis buffer without Nonidet P-40. The nuclei pellet was resuspended in glycerol buffer (50% glycerol, 5 mM MgCl2, 0.1 mM EDTA, and 50 mM Tris ⋅ HCl, pH 7.5). After counting, nuclei were frozen at −80°C at a concentration of 2–4 × 108/ml. Nuclei (10 × 106) were labeled with α-[32P]UTP as previously described (17).32P-labeled RNA was purified using TRI reagent according to the manufacturer's instructions, and total radioactivity was determined using scintillation counting. An equal amount of radioactivity was used in hybridization. cDNA probes were loaded onto nylon membrane with slit-blot apparatus and immobilized by UV cross-linking.

Measurement of phosphorylation of p42/44 MAPKs.

Activation of p42/44 MAPKs in cultured H4IIE cells was determined by Western blot with a rabbit polyclonal antibody raised against dually phosphorylated p42/44 MAPKs (19). Briefly, after cells were exposed to insulin, reaction was terminated by the replacement of medium with 200 μl of ice-cold lysis buffer (10 mM Tris ⋅ HCl, pH 7.4, 150 mM NaCl, 1 mM NaF, 1 mM Na3VO4, 1 mM EGTA, 1 mM PMSF, 50 mM tetrasodium pyrophosphate, 10 nM okadaic acid, 1% Triton X-100, 0.25% sodium deoxycholate, 10 μg/ml aprotinin, and 10 μg/ml leupeptin). For Western blot analysis, cell lysates (60 μg/lane) were electrophoresed on 10% SDS-polyacrylamide gels and transferred to an Optitran membrane. The membranes were probed with an anti-ACTIVE MAPK pAb that detects p42/44 MAPKs only when they are activated by phosphorylation at T202 and Y204. To ensure equal loading and protein transfer, the same blots were stripped and probed with a polyclonal antibody recognizing both phosphorylated and nonphosphorylated p42/44 MAPKs.

Preparation of replication-defective adenovirus Asn17Ras and adenovirus infection of H4IIE cells.

A replication-defective adenovirus expressing dominant negative Asn17 Ras was generated as previously described (15, 19). Virus was amplified in human kidney 293 cells, and the viral particles were purified from 293 cell lysates by cesium chloride gradient ultracentrifugation and then desalted by dialysis (15, 19). The concentration of recombinant adenovirus was determined on the basis of absorbance at 260 nm, where 1 optical density unit corresponds to 1012 particles/ml. An identical adenovirus containing the β-galactosidase gene instead of the Asn17 Ras gene was used as a virus control.

Acetate incorporation assay.

To determine whether H4IIE cells utilize acetate, cells were incubated with [3H]acetate for various times. Cells were then washed with cold PBS and precipitated with 10% TCA. TCA-insoluble materials were dissolved in 0.2 N NaOH and 0.1% SDS solution. Aliquots were counted in a scintillation counter, and the total cellular protein was determined by the Lowry method with BSA as standard.

Statistics.

Data are given as means ± SE. Statistical analysis was performed using Student's t-test, and significance was accepted atP < 0.05.

RESULTS

Insulin transcriptionally regulates the RNaPi-1 gene through signals derived from glucose metabolism.

We have demonstrated that the RNaPi-1 gene is regulated by insulin-activated metabolic signaling pathways in rat hepatocytes in primary culture. To further elucidate RNaPi-1-specific metabolic signaling pathways, H4IIE cells were used in the following studies. These cells, derived from a rat hepatoma cell line H35, retain the glycolytic but lose the gluconeogenic activity of hepatocytes (28, 34). Moreover, these cells exhibit high sensitivity to various hormones, such as insulin and glucocorticoids (34, 35, 37). When the effects of insulin on RNaPi-1 mRNA were measured, insulin increased steady-state levels of RNaPi-1 mRNA in the presence of glucose but decreased RNaPi-1 mRNA in the absence of glucose (Fig.1 A), clearly establishing a requirement of glucose for insulin-induced upregulation of RNaPi-1 mRNA. The effects of insulin on RNaPi-1 mRNA were dose dependent. For example, a significant induction of RNaPi-1 mRNA in the presence of glucose was observed at 0.1 nM insulin and reached maximum at 10 nM (data not shown). The effects of insulin appeared to be specific for the RNaPi-1 cotransporter, insofar as it did not significantly alter the mRNA levels of Ram-1, another Na+/Picotransporter (Fig. 1 B). This supports our previous observations that Na+/Pi cotransporters are differentially regulated (21). When time-dependent changes were measured under these experimental conditions, significant effects of insulin on RNaPi-1 mRNA were observed after 6 h of exposure, reached maximum after 24 h of exposure, and lasted for ≥36 h (Fig.2, A and B). When RNaPi-1 mRNA levels were determined as a function of glucose concentration, a dose-dependent induction was observed in the presence of insulin, and maximal induction was reached at 5 mM glucose (data not shown). However, glucose exhibited no significant effect on RNaPi-1 mRNA in the absence of insulin. Because glucagon represses insulin-induced gene expression by increasing intracellular cAMP levels (21, 37), it was of interest to determine whether cAMP regulates basal and insulin-induced RNaPi-1 expression in H4IIE cells. As depicted in Fig.3, insulin induction of RNaPi-1 expression in the presence of glucose was prevented by an increase in intracellular cAMP in forskolin-treated H4IIE cells. However, forskolin had no effect on basal levels of RNaPi-1 mRNA when cells were cultured in the absence of insulin.

Fig. 1.

Effects of insulin on steady-state levels of rat Na+/Pi cotransporter-1 (RNaPi-1) mRNA.A: representative autoradiogram of insulin effects. Cells were exposed for 24 h either to buffer alone (lanes 1 and 3) or to 10 nM insulin (lanes 2 and 4) in the presence (lanes 3 and 4) or absence (lanes 1 and2) of 10 mM glucose. Total RNA was isolated and assayed for RNaPi-1, amphotropic murine retrovirus (Ram-1), and β-actin mRNA levels. B: combined data from 5 independent experiments. Values of RNaPi-1 and Ram-1 mRNAs were normalized to those of corresponding β-actin measured on the same blots and expressed relative to a control value of one. Values are means ± SE. * P < 0.05; ** P < 0.01 vs. control.

Fig. 2.

Time course of effects of insulin and glucose on RNaPi-1 mRNA. Cells were treated with 10 mM of glucose in the presence or absence of 10 nM insulin for various times. Total RNA was isolated and assayed for RNaPi-1 and β-actin mRNAs as in Fig. 1. A: representative autoradiogram of insulin and glucose effects. B: combined data from 3 independent experiments. Values are means ± SE. * P <0.05; ** P < 0.01 vs. time 0.

Fig. 3.

Effects of forskolin on insulin-induced RNaPi-1 gene expression. Cells were pretreated with 0.1 mM forskolin for 15 min before exposure for 24 h to 10 nM insulin in the presence of 10 mM glucose. Total RNA was isolated and assayed for RNaPi-1 mRNA as in Fig. 1. Values are means ± SE of 4 experiments. ** P < 0.01 vs. insulin-treated cells.

To determine whether insulin increases transcription rate of the RNaPi-1 gene in the presence of glucose, nuclear run-on experiments were performed. As shown in Fig. 4, after 24-h exposure, insulin increased newly synthesized RNaPi-1, but not Ram-1, transcripts in the presence of glucose. These data indicate that insulin transcriptionally regulates the RNaPi-1 gene in H4IIE cells.

Fig. 4.

Nuclear run-on experiments showing effects of insulin and glucose on RNaPi-1 transcription. Cells were treated with 10 nM insulin in the presence of 10 mM glucose for 24 h before isolation and labeling of nuclei with α-[32P]UTP. An equal amount of radioactive run-on RNA was used in the hybridization. Insert: a representative autoradiogram. Graph, combined data from 3 experiments. Intensities of signals of RNaPi-1 were corrected by subtracting signals of a vector pBluescript (pBS), and they are represented in graph as means ± SE. Insulin significantly stimulated transcription of the RNaPi-1 gene in the presence of glucose (P < 0.01, Student's t-test).

Metabolic signals leading to activation of the RNaPi-1 gene are generated at or above pyruvate oxidation.

In H4IIE cells that are devoid of gluconeogenesis, glucokinase (GK) is replaced by other hexokinases capable of phosphorylating glucose independently of insulin (34). Because glucose had no effect on RNaPi-1 mRNA in the absence of insulin (Fig. 2), it appears that glucose 6-phosphate is not a proximate regulator of the RNaPi-1 gene. To evaluate this hypothesis, cells were treated with 2-deoxyglucose and fructose in the presence and absence of insulin. Like glucose, fructose supported insulin-induced RNaPi-1 expression, whereas 2-deoxyglucose decreased RNaPi-1 mRNA (Fig. 5). Because 2-deoxyglucose can only be phosphorylated, but not further metabolized, and because fructose is metabolized to trioses, but not to glucose 6-phosphate in these cells, these data confirm that glucose 6-phosphate was not a proximate regulator of the RNaPi-1 gene. Because fructose had no effect on RNaPi-1 expression in the absence of insulin, the signal must be generated through triose metabolism. This was supported by the fact that glycerol, lactate, and pyruvate induced RNaPi-1 expression in the presence of insulin (Table 1).

Fig. 5.

Effects of fructose and 2-deoxyglucose on RNaPi-1 mRNA levels. Cells were cultured in glucose-free medium and treated with either 10 mM fructose (Fruc) or 10 mM 2-deoxyglucose in the presence or absence of 10 nM insulin (Ins) for 24 h. RNaPi-1 mRNA levels were measured as in Fig. 1. Values are means ± SE of 4 different experiments. * P< 0.05; ** P < 0.01 vs. control.

View this table:
Table 1.

Effect of carbohydrates on RNaPi-1 induction

To determine whether compounds that enter metabolism beyond pyruvate oxidation can support insulin-induced RNaPi-1 expression, cells were cultured in the glucose-free medium and treated with insulin in the presence of either acetate or citrate. As depicted in Table 1, neither acetate nor citrate could support insulin-induced RNaPi-1 expression. In the absence of insulin, none of the compounds induced RNaPi-1 expression (Table 1). To ensure that these cells could use acetate, effects of insulin on [3H]acetate incorporation were determined, as shown in Fig. 6. The data clearly demonstrate that H4IIE cells can use acetate in an insulin-dependent manner.

Fig. 6.

Time-dependent [3H]acetate incorporation. Cells were cultured in glucose-free medium and exposed to 5 mM acetate and 1 μCi [3H]acetate for various times. Incorporation of [3H]acetate was measured as described in experimental procedures. Activity was expressed as counts ⋅ min−1 ⋅ 100 μg protein−1. Insert: effects of insulin on [3H]acetate incorporation. Cells were treated with 10 nM insulin for various times, as indicated, and then exposed to [3H]acetate for 2 h. Values are means ± SE of 4 measurements.

RNaPi-1 and L-PK use different metabolic signals.

Because the L-PK gene is regulated by metabolic signals generated above pyruvate metabolism, to corroborate the above findings it was of interest for us to determine whether pyruvate regulates L-PK and RNaPi-1 differently in H4IIE cells. In the experiments of Fig.7, cells were exposed to pyruvate in the presence of insulin and assayed for RNaPi-1 and L-PK mRNAs. The data showed that, whereas pyruvate and insulin upregulated the RNaPi-1 gene, they downregulated L-PK. Furthermore, whereas addition of pyruvate to a glucose-free medium showed no effect on L-PK mRNA in the absence of insulin (data not shown), addition of glucose increased the steady-state mRNA levels of L-PK, but not RNaPi-1, under the same experimental conditions (Fig. 7). These data support the proposition that L-PK and RNaPi-1 are regulated by different metabolic signals.

Fig. 7.

Effects of glucose and pyruvate on RNaPi-1 and L-PK mRNA levels. Cells were incubated with either 10 mM glucose or 20 mM pyruvate in the presence or absence of 10 nM insulin for 24 h. Total RNA was isolated and assayed for RNaPi-1 and rat L-type pyruvate kinase (L-PK) mRNAs. Autoradiogram is representative of 4 independent experiments.

Insulin regulates the metabolic signals through wortmannin and rapamycin-sensitive pathways.

Insulin regulates genes through two major signaling pathways, namely, the Ras/p42/44 MAPK pathway and the PI 3-kinase pathway in hepatocytes (7, 27). To determine whether the Ras/MAPK pathway is involved in insulin-induced upregulation of the RNaPi-1 gene, cells were infected with a recombinant adenovirus expressing an Asn17 dominant negative mutant of Ras for 12 h, as previously described (15, 19). The virus-transduced cells were then treated with insulin in the presence of glucose and measured for RNaPi-1 mRNA. Adenovirus β-galactosidase-infected cells served as control. As depicted in Fig.8, expression of dominant negative Ras showed no effects on insulin-stimulated RNaPi-1 expression in H4IIE cells. To further elucidate whether p42/44 MAPKs are involved in regulation of RNaPi-1, the effects of insulin on RNaPi-1 mRNA were determined in the presence of MEK inhibitor PD-98059. The data showed that 30 μM PD-98059 exhibited no effect on insulin-induced RNaPi-1 expression (Fig. 9). As previously reported (15), control experiments showed that expression of dominant negative Ras and MEK inhibitor PD-98059 blocked insulin-induced activation of p42/44 MAPK in H4IIE cells under the same experimental conditions. Insulin caused a 2.2 ± 0.4-fold increase in p42 MAPK phosphorylation and preincubation with 30 μM PD-98059 for 15 min decreased insulin-induced phosphorylation to 1.1 ± 0.2-fold over control. Like PD-98059, expression of Asn17 Ras reduced the stimulation to 1.2 ± 0.3-fold.

Fig. 8.

Effects of expression of Asn17 dominant negative mutant of Ras on insulin-induced RNaPi-1 expression. Cells were infected with Asn17 Ras adenovirus at a concentration of 2,000 particles/cell for 12 h, washed, and then exposed to 10 nM insulin in the presence of 10 mM glucose for 24 h. The β-galactosidase virus-infected cells served as control. Total RNA was isolated and measured for RNaPi-1 mRNA as in Fig. 1. Values are means ± SE of 3 independent experiments. ** P < 0.01 vs. control.

Fig. 9.

Effects of PD-98059, wortmannin, and LY-294002 on insulin-induced RNaPi-1 expression. Cells were pretreated with 30 μM PD-98059 (PD) or 100 nM wortmannin (Wort.) or 50 μM LY-294002 (LY) for 15 min before exposure to 10 nM insulin in the presence of 10 mM glucose for 16 h. During treatment, cells were refed with wortmannin every 4 h as previously described (29). Total RNA was isolated and assayed for RNaPi-1 mRNA as in Fig. 1. Values are means ± SE of 3 experiments. * P < 0.05; ** P < 0.01 vs. control.

To determine whether insulin-stimulated metabolic regulation of RNaPi-1 is mediated through the PI 3-kinase pathway, the effects of insulin on RNaPi-1 mRNA were determined in the presence of PI 3-kinase inhibitors (100 nM wortmannin and 50 μM LY-294002). As shown in Fig. 9, insulin-induced RNaPi-1 expression was significantly blocked when the cells were treated with either wortmannin or LY-294002. One of the signaling pathways downstream from PI 3-kinase requires activation of p70s6k (15, 29, 35). To determine whether this pathway is involved in insulin regulation of the RNaPi-1 gene, rapamycin, a p70s6k inhibitor, was used in experiments shown in Fig.10. The data showed that rapamycin inhibited insulin-induced RNaPi-1 expression in H4IIE cells in a dose-dependent manner. At high concentration it also significantly decreased basal levels of RNaPi-1 mRNA.

Fig. 10.

Effects of rapamycin (Rap) on RNaPi-1 expression. Cells were pretreated with different concentrations of rapamycin for 15 min before exposure to 10 nM insulin in the presence of 10 mM glucose for 16 h. RNaPi-1 mRNA was assayed as in Fig. 1. Values are means ± SE of 3–5 experiments. ** P < 0.01 vs. control.

DISCUSSION

The PI 3-kinase signal pathway mediates insulin-activated metabolic regulation of the RNaPi-1 gene.

Although the expression of NaPi-2-type transporters and Ram-1 is regulated by Pi depletion through Ca2+signaling (32, 39), that of RNaPi-1 is regulated by insulin and glucose (21). In this study, we further demonstrated that insulin regulated RNaPi-1 mRNA through a transcriptional mechanism in H4IIE cells. However, it remains to be determined whether insulin increases RNaPi-1 protein levels as well as Na+-dependent Picotransport activity in these cells. It is well established that binding of insulin to its receptor results in phosphorylation of endogenous substrates that allow for intermolecular interactions between phosphoproteins and SH-2 domain-containing proteins (7, 27). In H4IIE cells, insulin activated both Ras/p42/44 MAPK and PI 3-kinase pathways (15, 35). Insulin also increases cell volume, which in turn can further activate PI 3-kinase (20). Although activation of the Ras/MAPK pathway plays a key role in the effects of insulin on mitogenesis and c-fos expression, activation of the PI 3-kinase pathway leads to insulin regulation of many metabolic enzymes, such as phosphoenolpyruvate carboxykinase (PEPCK) and hexokinase II (HKII) (15, 29, 35). To determine how insulin generates RNaPi-1-specific signals, a recombinant adenoviral vector and several well characterized pathway-specific protein-kinase inhibitors were used in this study (15, 19). Although expression of dominant negative Ras and MEK inhibitor PD-98059 blocked insulin-induced activation of p42/44 MAPK, they failed to repress insulin-induced RNaPi-1 expression (Figs.8 and 9), suggesting that activation of a Ras/MAPK pathway is not required for insulin regulation of the RNaPi-1 gene. In contrast, when cells were treated with PI 3-kinase inhibitors wortmannin and LY-294002, insulin-induced RNaPi-1 expression was abolished (Fig. 9), indicating that the PI 3-kinase pathway is involved in insulin regulation of the RNaPi-1 gene. Interestingly, a recent study has demonstrated that synthesis of a NaPi-2-type transporter was also mediated by activation of PI 3-kinase in OK cells (30). Two downstream signaling pathways have been described for PI 3-kinase, Akt-GSK3 and p70s6k, which can be distinguished by their sensitivity to rapamycin (7, 8, 31). It has been demonstrated that rapamycin specifically blocks insulin-induced activation of p70s6k in H4IIE cells (35). p70s6k controls the ribosomal protein S6 phosphorylation in response to mitogens and plays an essential role in controlling the translation machinery (4, 31). Because significant increases in RNaPi-1 expression occurred after the cells were exposed to insulin for 6 h (Fig. 2), it is quite possible that the induction of RNaPi-1 requires increased protein synthesis. Accordingly, specific inhibition of p70s6k by rapamycin was found to be sufficient to block insulin-induced upregulation of RNaPi-1 expression (Fig. 10). These data suggest that activation of p70s6k by insulin may increase translation of the proteins that are involved in generation of RNaPi-1-specific metabolic signals. However, it remains to be determined as to which proteins are involved in regulation of RNaPi-1 and how they work. Interestingly, like RNaPi-1, insulin regulation of HIIK is also sensitive to rapamycin (29). On the other hand, insulin-induced downregulation of PEPCK and translocation of glucose transporter 4 are sensitive only to wortmannin, but not rapamycin (13, 35). These data point out that, downstream from PI 3-kinase, both p70s6k-dependent and p70s6k-independent pathways are used by insulin in regulation of metabolic enzymes and transporters.

RNaPi-1-specific metabolic signals may be generated at or above pyruvate oxidation.

Many metabolic enzymes are regulated by insulin in a glucose-dependent manner (36, 38). As shown in Figs. 1 and 2, the stimulatory effects of insulin on the RNaPi-1 gene in H4IIE cells were clearly glucose dependent. Because insulin most likely augmented glucose metabolism in H4IIE cells, we postulated that, like many other proteins, insulin-mediated transcriptional upregulation of RNaPi-1 was mediated by metabolic signals generated from glucose and other carbohydrates (10, 14, 33). Interestingly, in addition to this glucose-dependent pathway, an apparent glucose-independent pathway was also involved in regulation of RNaPi-1 mRNA in rat hepatocytes in primary culture, because insulin increased RNaPi-1 mRNA in the absence of glucose (21). However, it is quite possible that the glucose-independent regulation of RNaPi-1 is also mediated by metabolic signals. As discussed below, insulin-induced metabolic signals are generated at or above pyruvate oxidation. Because rat hepatocytes in primary culture exhibit high gluconeogenic activity, glucose metabolites, especially those not heavily controlled by insulin (e.g., the production of pyruvate), may accumulate through the gluconeogenic process. These metabolites may then suffice to mediate insulin induction of RNaPi-1 expression in rat hepatocytes. In contrast, because H4IIE cells lost gluconeogenic capacity (28, 34, 37), they cannot accumulate glucose and glucose metabolites at levels sufficient to mediate an insulin effect. Therefore, the effects of insulin on RNaPi-1 are completely dependent on glucose or other carbohydrates in H4IIE cells. Alternatively, insulin may activate different signaling pathways in hepatocytes from those in H4IIE cells.

Several glucose metabolites have now been identified to play important roles in regulation of a number of enzymes that are involved in glucose and lipid metabolism (10, 23-25). An increase in glucose-6-phosphate concentration, for example, is found to stimulate genes such as fatty acid synthase and acetyl-CoA carboxylase (5, 14). However, glucose-6-phosphate does not seem to act as a proximate regulator of RNaPi-1 gene because glucose had no effect on RNaPi-1 mRNA in the absence of insulin although insulin-dependent GK is replaced by other hexokinases capable of phosphorylating glucose independently of insulin in H4IIE cells (34). This notion is supported by the observations that in the presence of insulin, fructose induces RNaPi-1 expression while it is metabolized to trioses, but not to glucose-6-phosphate in these cells in the presence of insulin. Because 2-deoxyglucose can be phosphorylated to 2-deoxyglucose-6-phosphate, but not further metabolized, that 2-deoxyglucose decreased RNaPi-1 mRNA further supports that RNaPi-1 is regulated by signals generated downstream of glucose-6-phosphate.

Recently, downstream from glucose 6-phophate, a pentose pathway metabolite xylulose 5-phosphate has been identified as a proximate regulator of several enzymes including L-PK, PEPCK, and glucose-6-phosphatase (10, 24). In the presence of glucose, insulin induces GK and increases xylulose 5-phosphate concentration in rat hepatocytes by enhancing the conversion of glucose to glucose 6-phosphate, resulting in an insulin- and glucose-dependent upregulation of L-PK and downregulation of PEPCK (10, 24). However, the RNaPi-1 gene appeared to be regulated by metabolic signals other than xylulose 5-phosphate, because although the RNaPi-1 gene was induced, the L-PK gene was downregulated by pyruvate and insulin (Fig. 7). The proximate metabolic signals for RNaPi-1 appear to be generated at or above pyruvate oxidation because glycerol, lactate, and pyruvate induced RNaPi-1 expression in the presence of insulin (Table 1). It is clear that pyruvate itself is not a direct regulator of the RNaPi-1 gene, because in the absence of insulin pyruvate failed to induce RNaPi-1 gene expression. Interestingly, synthesis of malic enzyme in hepatocytes is also regulated by signals generated from pyruvate oxidation (23). Like malic enzyme (23), exposure of H4IIE cells to insulin in the presence of acetate and citrate failed to increase RNaPi-1 mRNA abundance (Table 1). Because H4IIE cells use acetate in an insulin-dependent manner (Fig. 6), these data indicate that, once glucose is irreversibly converted to cytosolic acetyl-CoA, its stimulatory effects on RNaPi-1 are lost.

In summary, we have demonstrated that the RNaPi-1 gene is regulated by insulin through a unique signaling pathway that includes insulin activation of PI 3-kinase/p70s6k and generation of metabolic signals (Fig. 11). It remains to be determined which metabolic signals generated by insulin and glucose are responsible for RNaPi-1 induction and how these metabolic signals regulate RNaPi-1 expression at gene levels.

Fig. 11.

A putative model for metabolic regulation of the RNaPi-1 gene by insulin and glucose. PI 3-kinase, phosphatidylinositol 3-kinase.

Acknowledgments

This work was supported by National Institutes of Health Grants HL-36573 and DK-43051 and by a grant-in-aid from the American Heart Association with funds contributed in part by the American Heart Association, Ohio-West Virginia Affiliate.

Footnotes

  • Address for reprint requests and other correspondence: Z. Xie, Department of Pharmacology, Medical College of Ohio, 3035 Arlington Ave., Toledo, OH 43614–5804 (E-mail: zxie{at}mco.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

REFERENCES

View Abstract