Vanadate treatment of diabetic rats reverses the impaired expression of genes involved in hepatic glucose metabolism: Effects on glycolytic and gluconeogenic enzymes, and on glucose transporter GLUT2

Molecular and Cellular Endocrinology, 0 1993 Elsevier Scientific Publishers

MOLCEL

91

91 (1993) 91-97 Ireland, Ltd. 0303-7207/93/$06.00

02914

Vanadate treatment of diabetic rats reverses the impaired expression of genes involved in hepatic glucose metabolism: effects on glycolytic and gluconeogenic enzymes, and on glucose transporter GLUT2 S.M. Brichard a, B. Desbuquois b and J. Girard a “CNRS, 92190 Meudon, France, and ’ INSERM U30, 75743 Paris, France (Received

Key words: Vanadium;

Diabetes;

Gene

expression;

15 July 1992; accepted

Glucokinase;

Pyruvate

kinase;

16 October

1992)

Phosphoenolpyruvate

carboxykinase;

Glucose

transporter

Summary

The trace element vanadium is a potent insulinomimetic agent in vitro. Oral administration of vanadate to rats made diabetic by streptozotocin (45 mg/kg iv.1 caused a 65% fall in plasma glucose levels without modifying low insulinemia. We studied whether the hypoglycemic effect of vanadate was associated with altered expression of genes involved in key steps of hepatic glucose metabolism. Glucokinase (GK) and L-type pyruvate kinase (L-PK) mRNA levels were decreased respectively by 90% and 70% in fed diabetic rats, in close correlation with changes in enzyme activities. Eighteen days of vanadate treatment partially restored GK mRNA and activity (40% of control levels), and totally restored L-PK parameters. In contrast to the glycolytic enzymes, mRNA levels and activity of the gluconeogenic enzyme, phosphoenolpyruvate carboxykinase (PEPCK) were increased (15- and 2-fold, respectively) in fed diabetic rats. Vanadate treatment normalized both PEPCK mRNA and activity in diabetic rat liver. The 2-fold increase in liver glucose transporter (GLUT21 mRNA and protein, produced by diabetes, was also corrected by this treatment. In conclusion, oral vanadate given to diabetic rats induces a shift of the predominating gluconeogenic flux, with subsequent high hepatic glucose production, into a glycolytic flux by pretranslational regulatory mechanisms.

Introduction

The repressed expression of genes encoding glycolytic enzymes and the stimulated expression of genes encoding gluconeogenic enzymes, as a result of combined insulin deficiency and resistance, lead to a decrease in hepatic glucose consumption and to glucose overproduction (Granner and Pilkis, 19901. This alteration in liver metabolism is a major mechanism that contributes to promote and maintain the diabetic state. The main liver glucose transporter (GLUT21 could also play a role in these perturbations since it appears to be involved in glucose uptake and release’ (Mueckler, 19901.

Correspondence to: J. Girard, Centre de Recherche sur I’Endocrinologie Moleculaire et le Dkeloppement, CNRS, 9, Rue J. Hetzel, 92190 Meudon-Bellevue, France. Tel. 33-I-45.07.58.50; Fax 33-l45.58.90.

Considerable interest has been shown concerning the effects of the trace element vanadium on carbohydrate metabolism. In vitro, vanadate mimics most, but not all, effects of insulin in various cell types (Shechter, 1990; Brichard et al., 1991). In vivo, oral vanadate markedly lowered plasma glucose concentrations in rats made insulin-deficient and diabetic by streptozotocin injection (Heyliger et al., 1985; Meyerovitch et al., 1987; Brichard et al., 1988; Gil et al., 1988; Blonde1 et al., 1989). This glucose lowering effect did not result from a rise in circulating insulin levels which remained low either in the basal state (Heyliger et al., 1985; Brichard et al., 1988) or during glucose tolerance tests (Brichard et al., 1988; Blonde1 et al., 1989) suggesting that insulin target tissues are the site of vanadate action. Experiments using euglycemic-hyperinsulinemic clamps have shown that both the suppressive effect of insulin on hepatic glucose production and the stimulation of peripheral glucose disposal were restored in these animals (Blonde1 et al., 19891. Administration of

vanadate to these rats has also been reported to increase the activity of liver glycolytic enzymes (Gil et al., 1988; Miralpeix et al., 1992) and to stimulate glycogen synthesis (Bollen et al., 1990; Pugazhenti and Khandelwal, 1990). In the present work, we investigated whether vanadate treatment reversed the impaired expression of genes involved in key steps of hepatic glucose metabolism and production in insulin-deficient rats. To this end, we measured mRNA levels of glycolytic (glucokinase (GK) and L-type pyruvate kinase CL-PK)) and gluconeogenic (phosphoenolpyruvate carbo~kinase (PEPCK)) enzymes, and of the glucose transporter isoform (GLUT2) in liver of vanadate-treated animals. Materials and methods Animals and experimental design

Male Wistar rats (S-week-old, 221 + 1 g) were purchased from IFFA Credo (L’Abresle, France). All rats received ad libitum a standard laboratory chow (Extralabo M-25, Pietrement, Provins, France; percent of total energy (calfcal): 60% carbohydrate, 11% fat, 29% protein) and were housed in individual cages at a constant temperature (22°C). A 12-h light/dark cycle was set up (lights on 03.00-15.00 h) to sample liver during the absorptive state of the animals at the time of experiments (between 08.00 and 10.00 h). This was performed in order to observe the largest differences between normal and diabetic rats in mRNA levels and activities of glycolytic and gluconeogenic enzymes. The animals were divided into three experimental groups: group 1, non-diabetic controls (C, n = 6); group II, untreated diabetic rats (D, II = 6); group III, diabetic rats treated with vanadate (V, IZ= 6). Non-ketotic diabetes was induced by an i.v. injection of streptozotocin (45 mg/kg body weight) in a tail vein. Streptozotocin (Sigma, St. Louis, MO, USA) was dissolved in cold 0.1 M citrate buffer (pH 4.5) immediately before use. Control animals received only the vehicle. Five days after injection of streptozotocin, the diabetic rats were assigned to untreated or treated groups. The two groups of diabetic rats were matched for pretreatment body weight and fed plasma glucose 1eveIs. They were also matched, as seen a posteriori, for fed plasma insulin levels (Table 1). The treated group received increasing amounts of sodium metavanadate (NaVO,; Merck, Darmstadt, Germany) in drinking solutions (up to 0.4 mg/ml>, as previously described (Brichard et al., 1989). This latter concentration was reached after 12 days of treatment and was maintained during the last week of the study. This progressive increase permitted us to partially overcome the aversion of the rats for vanadate. Every 2 or 3 days during the study, tail vein blood was collected from fed animals for determination of

plasma glucose levels. Plasma insulin levels were also measured in the last samples. After 18 days of treatment, animals were killed between 08.00 and 10.00 h (i.e., in the absorptive state). Liver was immediately removed, frozen in liquid nitrogen and stored at - 80°C for subsequent mRNA extraction, crude membrane preparation and enzyme activity measurement. RNA extraction and ~ori~ern blot analytic

Total RNA was isolated by a single extraction with an acid guanidinium thiocyanate-phenol-chloroform mixture (Chomczynski and Sacchi, 1987). The concentration of RNA was determined by absorbance at 260 nm. All samples had a 260/280 absorbance ratio of about 2. For Northern blot analysis, RNA (30 pg) was denaturated in a solution containing 2.2 mM formaldehyde and 50% formamide (v/v) by heating at 95°C for 2 min. RNA was then size-fractionated by 1% agarose gel electrophoresis, transferred to a Hybond-N membrane (Amersham, Bucks, UK) and crosslinked by UV irradiation. The integrity and relative amounts of RNA were assessed by methylene blue staining of the blot. The glucokinase, L-pyruvate kinase, phosphoenolpyruvate carboxykinase and GLUT2 cDNA probes were kindly supplied by Drs. P. Iynedjian (Iynedjian et al., 19871, A. Kahn (Simon et al., 19831, R.W. Hanson (Yoo-Warren et al., 1983) and B. Thorens (Thorens et al., 19881, respectively. Probes were labeled with j2P using the Megaprime labeling system kit (Amersham, Bucks, UK). Hybridizations were carried out in medium containing denatured salmon sperm DNA (0.3 mg/ml), 42% deionized formamide, 7.5% dextran sulfate, 8 X Denhardt’s solution, 40 mM-Tris/H~l (pH 7.51, 0.1% sodium pyrophosphate and 1% SDS at 42°C overnight. The membrane was washed twice for 30 min in 2 X sodium saline citrate (SSCJ/O.l% SDS at 42°C and once for 15-30 min in 0.1-0.5 x SSC/O.l% SDS at 55°C (GK, PEPCK, GLUT21 or 65°C (L-PK). The filter was then exposed to Hyperfilm MP (Amersham) for 16-96 h at -80°C with intensifying screens. After autoradiography, the membrane was washed in 0.1 X SSC/O.l% SDS at 95°C before rehybridization to a new probe. Intensity of the mRNA bands on the blot was quantified by scanning densitomet~ (Hoefer, San Francisco, CA, USA). Membrane preparation of Western blot analysis

Total liver membranes were prepared according to Thorens et al. (1988). Briefly, tissue c&200 mg) was homogenized with 10 strokes of a motor-driven Teflon pestle in 20 ~01s. 0.25 M sucrose/25 mM Hepes buffer (pH 7.4) containing 3 mM dithiothreitol, 0.26 U/ml aprotinin, 0.1 mM phenylmethylsulfonyl fluoride. This homogenate was centrifuged for 10 min at 8000 X g

93

and the supernatant was centrifuged again for 20 min at the same speed. The cytosol was then centrifuged for 40 min at 150,000 x g and the membrane pellet was resuspended in phosphate-buffered saline. Protein content was measured by the method of Bradford (1976) fBio-Rad, Munich, Germany), using BSA as standard. Equal amounts of membrane proteins (100 pg) were solubilized in Laemmli buffer (Laemmli, 1970), submitted to a 12% SDS-PAGE and electrophoretically transferred to nitrocellulose membranes (Schleicher and Schuell, Dassel, Germany). Rainbow protein markers were used as molecular weight standards and also to assess the efficiency of the transfer. Filters were then processed as previously described (Thorens et al., 1988). After blocking, the membranes were incubated with a 10e2 dilution of anti-peptide antibody directed against GLUT2 (a gift from Dr. B. Thorens). The detection was performed with 0.1 $X/ml ‘251-protein A, followed by autoradiography and subsequent quantification by scanning densitometry.

Enzyme activities

All enzyme activities were carried out at 37°C. Glucokinase (EC 2.7.1.1) activity was determined using a spectrophotometric assay (Newgard et al., 1983) by subtracting the hexokinase activity measured at 0.5 mM glucose from the activity measured at 100 mM glucose. Results were expressed as ,umol NADH produced per min/g liver. L-Pyruvate kinase (EC 2.7.1.40) activity was measured according to Blair et al. (1976) at a subsaturating concentration of phosphoenolpyruvate (1.3 mMI (active form of the enzyme) and at a saturating concentration of the substrate (6.6 mM) (maximal activity of the enzyme unaffected by its phosphorylation state). Results were expressed as pmol NADH oxidized per min/g liver. Phosphoenolpyruvate carboxykinase (EC 4.1.1.32) activity was determined in

liver cytosolic fractions using the NaH[‘“C]O, fixation assay of Chang and Lane (1966). Results were expressed as pmol NaH[‘4C]0, fixed per min/g liver. Other anai~~~~al procedures

Plasma glucose was measured by a glucose oxidase method (Peridochrom, Boehringer, Mannheim, Germany). Plasma insulin was determined by radioimmunoassay using a commercial kit (Oris Industrie, Gif sur Yvette, France). Statistical

analysis

Results are given as the mean -t_SEM for the indicated number of rats. Comparisons between C, D, and V rats were carried out by analysis of variance followed by the Newman-Keuls test for multiple comparisons (Sokal and Rohlf, 1969). Differences were considered statistically significant at p < 0.05. Results Diabetic rats gained weight at a much lower rate than controls (Table 1). As previously observed (Brichard et al., 19881, this growth retardation was not improved by vanadate treatment. Average plasma glucose levels in fed diabetic rats were above 25 mmol/l and their plasma insulin levels were significantly reduced. Vanadate administration resulted in a 65% fall in glucose concentrations compared with untreated diabetic rats (Table 1). This decrease was already observed after 3 days of treatment (39%, p < 0.01 vs. D rats). This effect was not due to a rise in plasma insulin levels, which remained low in fed treated and untreated diabetic rats compared with controls (Table 1). The attenuation of hyperglycemia in V-treated rats was accompanied by a 86% decrease in fluid consumption (31 k 3 ml/day vs. 216 & 14 ml/day in D rats;

TABLE 1 BODY WEIGHT, PLASMA GLUCOSE AND INSULIN LEVELS IN FED RATS OF THE THREE EXPERIMENTAL

GROUPS

Values are means & SEM for six rats in each group. Diabetes was induced by an i.v. injection of streptozotocin 5 days before the treatment. A group of diabetic rats was untreated; another group received NaVO, in drinking solutions from day 0 to day 18. Control rats were injected with citrate buffer and did not receive vanadate. Three samples were taken for plasma glucose measurements during the last week of the study; the values obtained were averaged for each rat. Plasma glucose (mmol/l)

Body weight(g)

Control Diabetic Diabetic + vanadate

Plasma insulin (I*U/ml)

Day 0

Day 1X

Day 0

Days 12-15-18

Day 0

Day 18

253i3 236&4 ** 23654 **

323+7 259i9 ** 256*7 **

7.0+0.1 27.2+1.5 ** 25.6+1.3 **

6.9+0.1 2%3+10 ** 10.1;1:1 *,+

37*4 20+4 * 25+3 *

49+3 20+3 ** 27+2 **

Statistical significance of differences (analysis of variance followed by Newman-Keuls ‘p <: 0.01 compared with untreated diabetic rats.

test): *p < 0.05; **p < 0.01 compared with controls;

GK mRNA

L-PK

4--

mRNA

PEPCK

mRNA

GLUT2

mRNA

2.4 kb

3.2 kb

2.6 kb

2.8 kb

t C

D

V

Fig. 1. Northern biot analysis of gfucokinase (GK), L-type pyruvate kinase (L-PK), pbospho~noipyruvate carboxykinase (PEPCK) and GLUT2 transporter mRNA in liver from control iC)* untreated diabetie (I)) and vanadate-treated diabetic fV) rats. All lanes were loaded with 30 pg of total RNA and the filter was hybridized successively with the different radiolabeled cDNA probes. This figure is representative af six separate experiments.

p < 0.01). Their average vanadate intake was 49 rt: 5 mg/kg per day (mean value of measurements made during the last week). To determine whether the blood glucose lowering effect of vanadate was associated with altered expression of genes involved in key steps of hepatic glucose metabolism, we measured mRNA Ievels of glycolytic and glu~oneogeni~ en~mes, and of GLUT2 transporter isoform. Yields of total RNA in liver did not differ between test groups (in mgfg: 5.0 rir0.1 (cf, 4.4 i 0.2 CD), 4.8 + 0.2 001. The abundance of mRNA teveis coding for the glycolytic enzymes, GK and L-P& (2.4 and 3.2 kb transcripts, respectively) was assessed

oC

DV

Ol-

C

D

V

Fig. 2. Effects of vanadate treatment on glucokinase (GK), L-type pyruvate kinase (L-PK), phosphoenolpyru~~te carboxykinas~ (PEPCK) and GLUT2 transporter mRNA levels in liver of diabetic rats. Values are means rt SEM for six control CC), untreated diabetic iDI and vanadate-treated diabetic rats. mRNA Ieve& were quantified by scanning densitometry of autoradjographj~ signals obtained from a Northern blot hybridized successively with the different radioiabeied cDN.4 probes. Results were expressed as percentages of values in control rats. Statistical significance of differences (analysis of variance followed by Newman-Keuls test): * p < 0.05, * * y < 0.01 compared with controls; ‘p < 0.05, ++p < 0.01 compared with untreated diabetic rats.

by Northern blot analysis (Fig. 11. Q~tantification of GK and L-PK mRNA autoradiographi~ signals revealed a 90% and 70% decrease in diabetic rats compared with controls (Fig. 21. The activities of GK and L-PK were lowered accordingly (Table 2). Eighteen days of vanadate treatment partially restored GK mRNA and activity (38% and 45% of control levels,

TABLE 2 EFFECTS OF VANADATE TREATMENT RATS

ON GK, L-P& PEPCK ACITVITIES AND ON GLUT2 PROTEIN IN LIVER OF DfABETXC

Values are means & SEM for six rats in each group. The animals were killed in the prandiai state after 18 days of treatment. GLUT2 protein was measured by Western blot analysis, quantized by scanning densitomet~ and expressed as a percentage of control rats. Abbreviations: GK, glucokinase; L-PK, L-type pyruvate kinase; PEPCK, phosphoenolpyruvate carboxykinase; PEP, phosphoenoip~~ruvate. CiKa

Control Diabetic Diabetic+vanadate Statistical +p < 0.01 a pmol of h Percent

1.61+0.12 0.15+0.03 ** 0.73F0.10 **.+

LPK a PEP 1.3 mM

PEP 6.6 mM

45.2 * 4.5 7.6rt 1.0 ** 44.7f5.6 +

73.3 * 6.5 17.5 f 1.4 * * 71.2&.7.1+

PE.PCK a

GLUT;! protein h

1.15 f0.08 2.65rtO.14 ** 1.06 Ifr0.06 +

lOO* 16 227+56 * 119*20+

significance of differences (analysis of variance followed by Newman-Keuls test): * p < 0.05: * * p < 0.01 compared with controls; compared with untreated diabetic rats. substrate transfurrned/min per g of liver. of controis.

95

GLUT2

protein

t

53 kDa

CDVCDVCDV Fig. 3. Western blot analysis of GLUT2 protein in liver crude membranes prepared from control (C), untreated diabetic (D) and vanadate-treated (V) rats. All lanes were loaded with 100 pg of membrane protein. This figure is representative of two separate experiments.

respectively), whereas PK parameters were totally restored (Figs. 1 and 2; Table 2). In contrast to the glycolytic enzymes, mRNA levels of the gluconeogenic enzyme, PEPCK were increased 15-fold in diabetic rats where hepatic glucose production is unrestrained. Autoradiographic signals of PEPCK transcripts were only barely detectable in fed controls (Figs. 1 and 2). Diabetes induced a 2-fold rise in PEPCK activity (Table 2). Vanadate administration normalized both PEPCK mRNA and activity in liver of treated rats (Figs. 1 and 2; Table 2). Glucose transport in liver is mostly mediated by the predominant GLUT2 glucose transporter isoform, which appears to have bidirectional transport capabilities, thus moving glucose into or out the hepatocytes. Examination of GLUT2 mRNA transcript (2.8 kb) abundance on Northern blots indicated that levels of these liver transcripts were increased 2.5-fold by diabetes (Figs. 1 and 2). Interestingly, a good correlation was found between GLUT2 and PEPCK gene expression in rat liver (coefficient of correlation from linear regression analysis: r = 0.88; p < 0.001; n = 18), whereas no correlation existed between the former parameter and glucokinase mRNA (r = 0.38). The amount of GLUT2 protein was measured by Western blot analysis in crude liver membranes. Yields of liver membrane proteins were not different in the three groups of rats (in mg/g liver: 17 k 1 (C), 16 + 1 (D), 19 + 1 (V>I. Quantification of autoradiographic signals showed a 2-fold rise of GLUT2 protein in diabetic rats. Altered abundance of GLUT2 mRNA and protein was corrected upon treatment of diabetic rats with vanadate (Fig. 3). Discussion Vanadate exerts a marked lowering effect on plasma glucose levels in insulin-deficient diabetic rats and this was associated with a restoration of hepatic L-PK and GK activities, in agreement with previous reports (Gil et al., 1988; Miralpeix et al., 19921, and with a normalization of both PEPCK activity and GLUT2 protein. This contrasts with the lack of effect of vanadate on

blood glucose levels and on hepatic enzyme activities observed in normal rats (Heyliger et al., 1985; Gil et al., 1988; Blonde1 et al., 1989, 1990; Bollen et al., 1990; Pugazhenti and Khandelwal, 1990). All alterations in liver enzyme activities and in GLUT2 protein amounts found in vanadate-treated diabetic rats were explained by changes in mRNA levels. The low expression of GK and L-PK genes in diabetic rats was increased and partially or totally reversed by vanadate; conversely, the enhanced expression of PEPCK and GLUT2 genes was suppressed by vanadate treatment. Vanadate appears therefore able to exert in vivo opposite effects on genes involved in liver glycolytic and gluconeogenic pathways. Moreover, vanadate was recently reported to restore, in addition to L-PK, mRNA levels of 6-phosphofructo-2-kinase (PFK-2) in diabetic rat liver (Miralpeix et al., 1992). Both L-PK and PFK-2 genes require the presence of glucose for the effects of insulin on their expression (Decaux et al., 1989; Cifuentes et al., 1991). A glucose metabolite is thought to be responsible for the effect of insulin on gene transcription (reviewed in Granner and Pilkis, 1990). In the present work, vanadate treatment has also been shown to restore expression of GK and PEPCK genes, whose insulin-regulated transcription appears to be glucose-independent (Sasaki et al., 1984; Iynedjian et al., 1989). The in vivo demonstration of these effects of vanadate are of particular interest due to the potential therapeutic use of vanadium compounds as antidiabetogenic agents (Shechter, 1990; Brichard et al., 1991). At least two mechanisms may contribute to the effects of vanadate treatment on gene expression in diabetic rat liver. First, an insulin-like action similar to that reported in vitro is likely to play a major role. Addition of vanadate has been found to induce L-PK mRNA in primary cultures of hepatocytes (Miralpeix et al., 1991) and to inhibit PEPCK expression in hepatoma cells (Bosch et al., 1990). However, this latter effect requires vanadate concentrations (0.5-2 mM) (Bosch et al., 1990) one or two orders of magnitude higher than those reached in the plasma (N 20 FM) (Meyerovitch et al., 1987; Brichard et al., 1989; Pugazhenti and Khandelwal, 1990) or in the liver (- 20 pmol/kgI (Gil et al., 1988; Pugazhenti and Khandelwal, 1990) of treated rats. Moreover, the in vitro effects of vanadate cannot be extrapolated a priori in vivo and vice versa. For example, vanadate has been shown to inactivate in vitro (Bosch et al., 1987), but to stimulate in vivo liver glycogen synthase (Bollen et al., 1990; Pugazhenti and Khandelwal, 1990). No change or a decrease in plasma insulin levels occurred in diabetic rats given vanadate (Heyliger et al., 1985; Brichard et al., 1988, 1989; Pugazhenti and Kandelwal, 19901, while an increase in insulin release was observed after addition of vanadate to isolated pancreatic islets (Fagin et

al., 1987; Zhang et al., 1991). Although the insulin-like effects have usually been demonstrated when the element was added to tissues or cells isolated from normal rats, vanadate treatment exerts little or no effect in the whole animal whose glucose homeostasis is not impaired (Heyliger et al., 1985; Gil et al., 1988; Blonde1 et al., 1989, 1990; Bollen et al., 1990; Pugazhenti and Khandelwal, 1990). Second, the plasma glucagon levels, which are elevated in insulin-deficient diabetic rats, have been reported to be lowered by vanadate treatment (Pugazhenti and Khandelwal, 1990). This could contribute to the in vivo reversal of impaired GK, L-PK and PEPCK gene expression (reviewed in Granner and Pilkis, 1990). A puzzling finding in this unlike insulin treatment work is that vanadate, (Iynedjian et al., 19881, did not fully restore GK mRNA in diabetic liver. Admittedly, the dose of vanadate may have been suboptimal. However, gene expression of other key regulatory enzymes was normalized under the present experimental conditions. Alternatively, vanadate might exert a less potent insulin-like effect on gene expression of GK than of the other enzymes studied. This is in line with the observations that vanadate did not always duplicate all the biological actions of the hormone (reviewed in Brichard et al., 1991). Eventually, the potent glucagon-mediated repression of GK transcription, which is an essential feature in the regulation of this enzyme (Iynedjian et al., 1989) may be involved in the present finding. In primary cultured hepatocytes, glucagon has been shown to inhibit GK transcription at a concentration IO-fold lower than that required to induce PEPCK mRNA (P. Bossart, J.-F. Decaux and J. Girard, unpublished data). Although decreased, the persistence of slightly elevated glucagon levels in vanadate-treated rats could be sufficient to maintain a partial inhibition of GK gene expression. There remains much controversy in the cellular and molecular mechanisms involved in the insulin-like action of vanadate (reviewed in Shechter, 1990; Brichard et al., 1991). Its insulin-like properties have usually been ascribed to enhanced phosphorylation of the insulin receptor (Tamura et al., 1984). This may result either from the activation of the tyrosine kinase of the receptor (Tamura et al., 1984) or from the inhibition of a phosphotyrosyl phosphatase (Swarup et al., 1982). Strong evidence has, however, been presented by in vitro and in vivo studies suggesting that vanadate action actually occurs at a step distal to the receptor. The insulin-like effects of vanadate were preserved in adipocytes depleted of insulin receptors by trypsin or insulin and Tris treatment (Green, 1986). In addition, vanadate administration to mildly (Blonde1 et al., 1990) or severely (Venkatesan et al., 1991) diabetic rats did not induce any detectable activation of insulin receptor under conditions where glucose tyrosine kinase,

metabolism was stimulated. The suggestion that vanadate and insulin did not always share a common pathway to elicit their effects was recently fostered by the observation that vanadate inhibited PEPCK gene transcription in hepatoma cells by acting on a target region in the promoter which was totally different from that involved in insulin action (Bosch et al., 1990). Since the discovery that facilitated glucose transport is mediated by a family of tissue-specific membrane proteins, much attention has been paid to the potential regulatory role of vanadate on these proteins. Vanadate has been reported to increase the expression of the ubiquitous glucose transporter (GLUTl) in cultured fibroblasts and this was associated with an increase in glucose uptake (Mountjoy and Flier, 1990). Moreover, vanadate treatment corrected the diminished expression of the insulin-responsive glucose transporter (GLUT41 in skeletal muscle of streptozotocin-diabetic rats (Strout et al., 1990). However, vanadate had no effect on GLUT4 in obese insulin-resistant fu/fa rats, despite a marked improvement of glucose uptake by muscle (Brichard et al., 1992). The influence of vanadate on the main liver glucose transporter (GLUT2) has not yet been studied. Although glucose transport is not the rate limiting step of liver glucose metabolism, GLUT2 is involved in either net glucose uptake or release, depending on the prevailing metabolic conditions (Mueckler, 1990). In agreement with some (Oka et al., 1990; Yamamoto et al., 1991), but not all (Thorens et al., 1990) previous reports, a 2-fold increase in both GLUT2 mRNA and protein was observed in liver of streptozotocin-diabetic rats. This could contribute to the increased glucose output from diabetic liver (Oka et al., 1990). This suggestion is strengthened by the finding of a good linear correlation between the expression of GLUT2 and PEPCK. Vanadate, like insulin treatment (Oka et al., 1990), decreased the amount of GLUT2 towards normal levels. Oral vanadate appears, therefore, to be able to correct the impaired expression of both GLUT2 and GLUT4 glucose transporter isoforms, respectively in liver and skeletal muscle of insulin-deficient animals. This is in keeping with the results of euglycemic-hyperinsulinemic clamp experiments which have shown a reversal of impaired insulin action both at the level of liver and peripheral tissues in vanadate-treated diabetic rats (Blonde1 et al., 1990). In conclusion, oral vanadate given to diabetic rats induces a shift of the predominating gluconeogenic flux, with subsequent hepatic glucose overproduction, into a glycolytic flux by pretranslational regulatory mechanisms. This largely contributes to reverse the impaired glucose homeostasis. These observations may help in elucidating both the mechanisms of insulin action on gene transcription and the causes of their perturbations. This may also raise new therapeutic

97

possibilities resistance.

in the yet unsolved

problem

of insulin

Acknowledgments

We thank Drs. P. Iynedjian (Geneva, Switzerland), A. Kahn (Paris, France), R.W. Hanson (Cleveland, OH, USA) and B. Thorens (Lausanne, Switzerland) for kindly providing us with the cDNA probes and the antibody used in this study. This work was supported in part by Fondation de la Recherche Mkdicale. S.M.B. acknowledges receipt of a fellowship from Institut National de la Sante et de la Recherche Mtdicale. References Blair, J.B., Cimbala, M.A., Foster, J.L. and Morgan, R.A. (1976) J. Biol. Chem. 251, 3756-3762. Blondel, O., Bailbe, D. and Portha, B. (1989) Diabetologia 32, 185-190. Blondel, O., Simon, J., Chevalier, B. and Portha, B. (1990) Am. J. Physiol. 258, E459-E467. Bollen, M., Miralpeix, M., Ventura, F., Toth, B., Bartrons, R. and Stalmans, W. (1990) Biochem. J. 267, 269-271. Bosch, F., Arino, J., Gomez-Foix, A.M. and Guinovart J.J. (1987) J. Biol. Chem. 262, 218-222. Bosch, F., Hatzoglou, M., Park, E.A. and Hanson, R.W. (1990) J. Biol. Chem. 265, 13677-13682. Bradford, M. (1976) Anal. Biochem. 72, 248-254. Brichard, S.M., Okitolonda, W. and Henquin, J.C. (1988) Endocrinology 123, 2048-2053. Brichard, SM., Pottier, A.M. and Henquin, J.C. (1989) Endocrinology 125, 2510-2516. Brichard, S.M., Lederer, J. and Henquin, J.C. (1991) Diab&te MCtabol. 17, 435-440. Brichard, S.M., Assimacopoulos-Jeannet, F. and Jeanrenaud, J. (1992) Endocrinology 131, 311-317. Chang, H.C. and Lane, M.D. (1966) J. Biol. Chem. 241, 2413-2420. Chomczynski, P. and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. Cifuentes, M.E., Espinet, C., Lange, A.J., Pilkis, S.J. and Hod, Y. (1991) J. Biol. Chem. 266, 1557-1563. Decaux, J.F., Antoine, B. and Kahn, A. (1989) J. Biol. Chem. 264, 11584-11590. Fagin, J.A., Ikejiri, K. and Levin, S.R. (1987) Diabetes 36, 1448-1452. Gil, J., Miralpeix, M., Carreras, J. and Bartrons, R. (1988) J. Biol. Chem. 263, 1868-1871. Granner, D. and Pilkis, S. (1990) J. Biol. Chem. 265, 10173-10176. Green, A. (1986) Biochem. J. 238, 663-669. Heyliger, C.E., Tahiliani, A.G. and McNeill, J.H. (1985) Science 227, 1474-1477.

Iynedjian,

P.B., Ucla,

C. and Mach,

6032-6038. Iynedjian, P.B., Gjinovci,

B. (1987) J. Biol. Chem.

A. and Renold,

262,

A.E. (1988) J. Biol. Chem.

263, 740-744. Iynedjian, P.B., Jotterand, D., Nouspikel, T., Asfari, M. and Pilot, P.-R. (1989) J. Biol. Chem. 264, 21824-21829. Laemmli, U.K. (1970) Nature 227, 680-685. Meyerovitch, J., Farfel, Z., Sack, J. and Schechter, Y. (1987) J. Biol. Chem. 262, 6658-6662. Miralpeix, M., Decaux, J.F., Kahn, A. and Bartrons, R. (1991) Diabetes 40, 462-464. Miralpeix, M., Carballo, E., Bartrons, R., Crepin, K., Hue, L. and Rousseau, G.G. (1992) Diabetologia 35, 243-248. Mountjoy, K.G. and Flier, J.S. (1990) Endocrinology 26, 2778-2787. Mueckler, M. (1990) Diabetes 39, 6-11. Newgard, C.B., Hirsch, L.J., Foster, D.W. and McGarry, J.D. (1983) J. Biol. Chem. 258, 8046-8052. Oka, Y., Asano, T., Shibasaki, Y., Lin, J.-L., Tsukuda, K., Akanuma, Y. and Takaku, F. (1990) Diabetes 39,441-446. Pugazhenti, S. and Khandelwal, R.L. (1990) Diabetes 39, 821-827. Sasaki, K., Cripe, T.P., Koch, S.R., Andreone, T.L., Petersen, D.D., Beale, E.G. and Granner. D.K. (1984) J. Biol. Chem. 259, 1524215251. Schulz, L.O. (1988) Ann. Nutr. Metab. 32, 289-296. Shechter. Y. (1990) Diabetes 39, l-5. Simon, M.P., Besmond, C., Cottreau, D., Weber, A., Chaumet-Riffaud, P., Dreyfus, J.C., Sala-Trepat, J., Marie, J. and Kahn, A. (1983) J. Biol. Chem. 258, 14576-14584. Sokal, R.R. and Rohlf, F.J. (1969) in Biometry. The Principles and Practice of Statistics in Biological Research, pp. l-776, Freeman, San Francisco, CA. Strout, H.V., Vicario, P.P., Biswas, C., Saperstein, R., Brady, E.J., Pilch, P.F. and Berger, J. (1990) Endocrinology (Baltimore) 126, 2728-2732. Swarup, G., Speeg, K.V., Cohen, S. and Garbers, D.L. (1982) J. Biol. Chem. 257, 7298-7301. Tamura, S., Brown, T.A., Whipple, J.H., Fujita-Yamaguchi, Y., Dubler, R.E., Cheng, K. and Larner, J. (1984) J. Biol. Chem. 259, 6650-6658. Thorens, B., Sarkar, H.K., Kaback, H.R. and Lodisch, H.F. (1988) Cell 55, 281-290. Thorens, B., Flier, J.S., Lodish, H.F. and Kahn, B.B. (1990) Diabetes 39, 712-719. Venkatesan, N., Avidan, A. and Davidson, M.B. (1991) Diabetes 40, 492-498. Yamamoto, T., Fukumoto, H., Koh, G., Yano, H., Yasuda, K., Masuda, K., Ikeda, H., Imura, H. and Seino, Y. (1991) Biochem. Biophys. Res. Commun. 175, 995-1002. Yoo-Warren, H., Monahan, J.E., Short, J., Short, H., Bruzel, A., Wynshaw-Boris, A., Meismer, H.M., Samols, D. and Hanson, R.W. (1983) Proc. Natl. Acad. Sci. USA 80, 3656-3660. Zhang, A., Gao, Z.Y., Gilon, P., Nemquin, M., Drews, G. and Henquin, J.C. (1991) J. Biol. Chem. 266, 21649-21656.