INTRODUCTION

Glucose transport is a rate-limiting step in the metabolism of many cell types and therefore in energy production (1). This process is mediated by a family of transmembrane glycoproteins differing in their kinetics and tissue distribution (2). Among the six known glucose transporters, Glut1 is the ubiquitous glucose transporter. Its gene expression is regulated in different cell types by various stimuli. These include hypoglycemia (3), hypoxia (4), mitochondria inhibitors (5), prolonged insulin exposure (6), tumor necrosis factor  (7), and iron chelators (8). It appears that under such stress conditions, energy requirements are increased, and induction of the Glut1 isoform is therefore instrumental in cellular adaptation. However, the cellular mechanisms involved in this process are still unclear.

Increased oxidative stress has been suggested to play a role in many pathophysiological conditions. Altered transcriptional regulation of various genes, postulated to be mediated by transcriptional activation factors such as AP-1 and NF-B, is a well described cellular reaction to oxidative stress (9-11). Striated muscle appears to be specifically exposed to oxidative stress in ischemia-reperfusion injury (12, 13) as well as in conditions resulting in systemic increase in oxidative stress parameters, including diabetes mellitus (14, 15). In this disease, oxidative stress has been suggested to be related to impaired insulin action, supporting a physiologically relevant consequence of skeletal muscle exposure to increased oxidative stress (16-18). We have recently assessed the effect of oxidative stress on glucose transport and metabolism in L6 myotubes and 3T3-L1 adipocytes (19, 20). We observed that exposure of both cell types to prolonged low grade oxidative stress caused a time- and dose-dependent increase in Glut1 expression at both the protein and mRNA levels, resulting in increased basal glucose transport activity.

Recently, the transcriptional regulation of the Glut1 gene has been shown to involve regulatory elements that include a promotor and two enhancers (21, 22). These elements include potential binding sites for various transcriptional activation factors, including the serum-responsive element (SRE),¹ the cAMP-responsive element (CRE), and AP-1-binding sites (TRE). In this study, we aimed to elucidate the cellular mechanisms responsible for Glut1 activation in response to oxidative stress. We demonstrate that following exposure to oxidative stress, the Glut1 gene transcription rate is enhanced, concomitant with increased stability of mRNA transcripts. This transcriptional activation is at least partly mediated by enhancer 1 of the Glut1 gene and requires de novo protein synthesis. Increased binding activity of AP-1 and SRE to DNA is observed in response to oxidation and may mediate Glut1 transcriptional activation in response to oxidative stress.

EXPERIMENTAL PROCEDURES

Tissue culture medium, serum, and reagents were obtained from Biological Industries (Beit-Haemek, Israel). Cytochalasin B, 2-deoxyglucose, cycloheximide, xanthine, xanthine oxidase, and glucose oxidase were obtained from Sigma. Recombinant human insulin was from Novo Nordic (Bagsvaerda, Denmark). Rapamycin was from Calbiochem. 2-Deoxy[³H]glucose and 5-[-³²P]dCTP were obtained from Rotem Industries (Dimona, Israel). [-³²P]ATP and [³H]acetyl coenzyme A were obtained from NEN Life Science Products. EcoRI and BglI were from Pharmacia LKB (Uppsala, Sweden). The polyclonal antibodies raised against the C-terminal sequences of Glut1 were obtained from East Acres Biologicals. The plasmids pGT1-1.3CAT, pGT1-1.3CAT/SBb, pGT1-1.3CAT/Bxb, and pGT1-1.3CAT/Bxb/SBb were constructed as described previously (22).

L6 muscle cells were grown in monolayers to the stage of myotubes as described previously (4, 20). The cells were grown in 24-well plates for transport determinations or in 10-cm diameter dishes for RNA, transcriptionally active nucleus, and nuclear protein preparations.

Reactive oxygen species were generated continuously by two different sources of oxidant-generating systems as described previously (19, 20) using the enzyme/substrate mixture of xanthine/xanthine oxidase (Xan/XO) or glucose/glucose oxidase (Glc/GO). The first catalyzes the conversion of xanthine to uric acid with reduction of O₂ to O₂, H₂O₂, and OH^·, whereas the second catalyzes the reaction of glucose to glucuronic acid and H₂O₂.

Hexose uptake was measured for 10 min using 10 µM 2-deoxy[³H]glucose (1 µCi/ml) as described previously (4, 20).

Total cellular RNA was extracted from L6 myotubes after incubation in the indicated conditions using a single-step extraction procedure with acid guanidium thiocyanate/phenol/chloroform as described by Chomczynski and Sacchi (23). Northern blot analysis was performed according to Sambrook et al. (24) as described previously (19, 20). Autoradiographs were performed with a BioImaging BAS1000 analyzer (Fuji Film Co., Tokyo, Japan).

Cells were disrupted using a Dounce homogenizer and ice-cold buffer containing 10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl₂, and 0.5 (v/v) Nonidet P-40. Nuclei were isolated using the procedure described by Mahajan and Thompson (25). The final pellet was resuspended in 0.05 ml of 75 mM NaCl, 20 mM Tris-HCl (pH 8.0), 1 mM dithiothreitol, and 0.5 mM EDTA in 50% glycerol and frozen at 70 °C.

Transcription assays were carried out in a 0.2-ml reaction volume in buffer containing 32% glycerol, 100 mM KCl, 50 mM HEPES (pH 7.4), 5 mM magnesium acetate, 4 mM dithiothreitol, 0.1 mM EDTA, 4 units of RNasin, 1 mM CTP, 1 mM GTP, and 1 mM ATP plus 0.5 µCi of [-³²P]UTP. The reaction was allowed to proceed for 30 min at 30 °C and terminated by the addition of 1 mM MgCl₂ and 7000 units/µl RNase-free DNase I for 15 min at 37 °C. The reaction mixture was then digested with 300 µg/ml proteinase K in the presence of 0.1% SDS. The labeled RNA transcripts were isolated using the acid phenol method (23) and resuspended in 100 µl of water.

Nylon filters (Zeta Probe, Bio-Rad) containing fixed and denatured plasmids with either Glut1 or -actin (10 µg) cDNA probes were prehybridized for 1 h at 50 °C in 6 × SSPE (1 × SSPE = 0.18 M NaCl, 10 mM NaPO₄ (pH 7.7), and 1 mM EDTA), 10 × Denhardt's reagent, 1.0% SDS, and 100 µg/ml denatured salmon sperm DNA. Hybridization was carried out at 65 °C for 36 h in 50% formamide, 6 × SSPE, 1.0% SDS, and 100 µg/ml denatured salmon sperm DNA plus 7 × 10⁶ cpm/ml radiolabeled RNA transcripts. Subsequent to hybridization, the filters were washed three times under low stringency conditions (2 × SSPE and 1.0% SDS) for 60 min at 37 °C. The washed filters were subjected to autoradiography and quantitated by laser scanning densitometry.

293 cells were cultured in Dulbecco's modified Eagle's medium for 24 h and then transfected with a reporter gene construct by the calcium phosphate method coupled with hyperosmotic shock as described previously (26). Cells were treated 24 h after transfection for 3 h with Xan/XO, and CAT and -galactosidase activities were determined. The -galactosidase promotor is the long terminal repeat of the Moloney murine leukemia virus. Assay of CAT activity was performed as described previously (24): cells were washed twice with phosphate-buffered saline, lysed in 250 mM Tris (pH 7.8) by freeze-thawing, and then scraped and collected in 1.5-ml microcentrifuge tubes. The tubes were heated to 65 °C for 15 min to inactivate endogenous acetylases and then spun at 12,000 × g for 10 min at 4 °C to remove cell debris. An aliquot of cell extract (50 µl) was mixed with 200 µl of reaction mixture containing 25 µmol of Tris (pH 8.0), 0.25 µmol of chloramphenicol, and 0.1 µCi of [³H]acetyl coenzyme A (1.4 Ci/mmol). 5 ml of nonaqueous scintillation fluid was added to the reaction mixture, and then the mixture was incubated at 37 °C. The amount of radioactivity in the scintillation fluid (the acetylated chloramphenicol diffuses into the organic scintillation fluid) was determined as a function of time up to 24 h. -Galactosidase activity was determined as described previously (24).

Gel shift assays were performed with nuclear extracts of L6 myotubes as described by Dignam et al. (27). The probes used in gel mobility shift assays were as follows: CRE, 5-GATCAGCGCTGTGGCGTCATGACCTCGCTGACAG-3; SRE, 5-GGATGTCCATATTAGGACACATCTG-3; and TRE, 5-GATCCCTCGGGGTGACTCATGGGCTAG-3. The oligonucleotides were labeled with [-³²P]ATP using polynucleotide kinase, annealed, and subsequently purified on a Sephadex G-50 column. Equivalent protein amounts (10 µg) were incubated in a final reaction volume of 20 µl containing 0.5 ng of ³²P-5-end-labeled double-stranded probe, 2 µg of DNA duplex poly(dI-dC), 20 µg of bovine serum albumin, 20 mM HEPES (pH 8.0), 4 mM Tris (pH 7.9), 50 mM KCl, 2 mM EDTA, and 500 µM dithiothreitol. After incubation at room temperature for 20 min, samples were electrophoretically separated on a nondenaturing 5% polyacrylamide gel. The gel was dried and autoradiographed. For the competition experiments, a 100-fold molar excess of unlabeled oligonucleotide or nonrelated oligonucleotide was added to the binding reaction.

The significance of each result was compared only with the control using Student's t test.

RESULTS

A 24-h exposure of L6 myotubes to 50 milliunits/ml glucose oxidase in medium containing 5 mM glucose (Glc/GO) or to 20 milliunits/ml xanthine oxidase in the presence of 50 µM xanthine (Xan/XO) caused an ~2-fold increase in glucose transport activity attributed to elevated levels of Glut1 protein (Table I). Cycloheximide, a protein synthesis inhibitor, prevented elevation of both glucose transport activity and Glut1 protein expression (20), indicating that reactive oxygen species stimulated the de novo synthesis of this glucose transporter. The activation of glucose transport activity was reversible when, following treatment, cells were further exposed to fresh medium. Full recovery to pretreatment 2-deoxyglucose uptake activity was observed after 40 h, with a t1/2 of 17.5 h. Steady-state mRNA levels were increased to a similar extent by oxidation (Table I), suggesting an increased transcription rate of the Glut1 gene and/or increased mRNA stability. To address this question, nuclear run-on and mRNA stability assays were performed, respectively. Following treatment with oxidant-generating systems for 24 h, transcriptionally active nuclei were isolated from cells, and in vitro transcription was performed for 30 min as described under "Experimental Procedures." This was followed by Northern dot-blot analysis using an empty plasmid or a plasmid containing the full-length cDNA probes of either Glut1 or -actin. The empty plasmid did not react with nascent mRNA (data not shown). However, nuclei isolated from cells exposed to either Glc/GO or Xan/XO had a higher Glut1 transcription rate compared with the control, with no similar effect observed on the transcription rate of -actin (Fig. 1). Densitometric analysis of three independent experiments indicated 2.0- and 2.4-fold increases in Glut1 transcription rates in Glc/GO- and Xan/XO-pretreated cells, respectively, compared with control cells. This suggests a specific effect of oxidative stress on the transcriptional regulation of the Glut1 gene. Under the same conditions, Glut1 mRNA stability was evaluated. Following a 24-h exposure to Glc/GO or Xan/XO, cells were further incubated with the mRNA synthesis inhibitor actinomycin D. The half-life of pre-existing Glut1 mRNA was increased in both Glc/GO- and Xan/XO-treated cells compared with control cells (t1/2 = 7.8 ± 1.3, 6.0 ± 2.0, and 2.4 ± 0.5 h, respectively; p < 0.05 for each treatment versus the control) (Fig. 2). These results suggest that both increased transcription rate and mRNA stability may account for the increased steady-state Glut1 mRNA content following oxidation.

Table I. Effect of a 24-h exposure to Glc/GO or Xan/XO on 2-deoxyglucose uptake and Glut1 protein and mRNA content in L6 myotubes L6 myotubes were incubated for 24 h with 50 milliunits/ml glucose oxidase in medium containing 5 mM glucose (Glc/GO) or with 50 µM xanthine and 20 milliunits/ml xantine oxidase (Xan/XO), after which 2-deoxyglucose uptake and Glut1 protein and mRNA levels were determined as described under "Experimental Procedures." Values are means ± S.E. The number of independent experiments is indicated in parentheses. Glc/GO Xan/XO 2-DGa uptake (-fold of control) 2.65  ± 0.43 (10) 2.2  ± 0.7 (8) Glut1 protein (arbitrary units) 2.1  ± 0.4 (5) 1.95  ± 0.2 (5) Glut1 mRNA (arbitrary units) 2.3  ± 0.7 (4) 2.0  ± 0.3 (4) a 2-DG, 2-deoxyglucose.

[View Larger Version of this Image (44K GIF file)]

[View Larger Version of this Image (45K GIF file)]

The regulatory elements of the Glut1 gene include a promotor (1.3kb to +137 base pairs) and two enhancers (enhancer 1, 3.3kb to 2.7kb; and enhancer 2, +16.7kb to +18kb), which have been shown to include consensus sequences for the binding of various transcription factors, including SRE, CRE, and the AP-1-binding site (TRE) (21, 22). To assess the involvement of these regulatory elements in oxidative stress-induced Glut1 gene activation, plasmids containing the Glut1 promotor and the CAT reporter gene either alone or with enhancer 1, enhancer 2, or both (Fig. 3) were transfected into 293 cells. These cells exhibited an ~50% increase in 2-deoxyglucose uptake activity following exposure to Xan/XO (data not shown). To control for CAT activity, cells were simultaneously transfected with plasmids containing -galactosidase as described above. Cells were then exposed for 3 h to Xan/XO, after which both CAT and -galactosidase activities were measured. Cells containing a plasmid with the promotor alone (pGT1-1.3CAT) did not exhibit an increased CAT/-galactosidase activity ratio following oxidation compared with control cells. A maximal 4.2 ± 0.25-fold increase in the CAT/-galactosidase activity ratio was observed following oxidation of cells transfected with a plasmid containing both enhancers (pGT1-1.3CAT/Bxb/SBb). This effect could be attributed mainly to enhancer 1, as a 2.4 ± 0.21-fold increase in the CAT/-galactosidase activity ratio was observed in cells containing the pGT1-1.3CAT/SBb plasmid, without a similar effect in cells transfected with a plasmid containing enhancer 2. 

Effect of Xan/XO on the CAT/-galactosidase activity ratio in 293 cells transfected with plasmids containing various regulatory regions of the Glut1 gene.

[View Larger Version of this Image (19K GIF file)]

Enhancer 1 of the Glut1 gene has been shown to contain consensus sequences for the binding of various transcriptional activation factors. These include two AP-1-binding sites (TRE), one SRE, and one CRE (21). To evaluate the potential role of the respective transcription factors in inducing Glut1 gene transcription, we assessed whether oxidation caused increased binding activity of nuclear protein extracts to prelabeled oligonucleotides utilizing the gel electromobility shift assay. L6 myotubes were incubated for 8 h in the absence or presence of Glc/GO or Xan/XO, after which nuclear protein extracts were prepared as described. Fig. 4 demonstrates increased AP-1 and SRE (but not CRE) binding activity (data not shown) following exposure of the cells to oxidative stress.

[View Larger Version of this Image (114K GIF file)]

AP-1 activation may involve activation of its two precursors, c-fos and c-jun. To further evaluate the respective role of the AP-1 transcription factor in Glut1 gene transcriptional activation, steady-state mRNA levels of c-fos and c-jun were measured in L6 myotubes exposed to Glc/GO or Xan/XO. As shown in Fig. 5A, c-fos and c-jun mRNA levels were increased following oxidation, without a similar effect on -actin mRNA levels (data not shown). The ability of cycloheximide to inhibit both AP-1 activation and Glut1 mRNA elevation following oxidation was assessed. Fig. 5B demonstrates that activation of AP-1 DNA-binding activity was inhibited when L6 myotubes were exposed to oxidative stress in the presence of 5 µg/ml cycloheximide. Under similar conditions, the elevation of Glut1 mRNA levels was prevented by cycloheximide (Fig. 5C). Taken together, these results confirm the role of de novo protein synthesis in both activation of AP-1 DNA-binding activity and induction of the Glut1 gene in response to oxidative stress.

[View Larger Version of this Image (58K GIF file)]

Fig. 4 demonstrates the increased activation of SRE binding to DNA following oxidative stress. This transcription factor has been shown to mediate Glut1 transcriptional activation in response to insulin exposure (22). In L6 myotubes, chronic insulin-induced elevation of Glut1 protein is inhibited by rapamycin, a pp70^S6K inhibitor, but not by the expression of a Ras dominant-negative mutant (28). In accordance, the Ras/mitogen-activated protein kinase cascade was not involved in increasing the expression of Glut1 in L6 myotubes in response to oxidative stress, as no increase in immunoreactive active mitogen-activated protein kinase was observed (data not shown). To compare the insulin-induced activation of the Glut1 gene with the effect of oxidative stress, Glut1 mRNA levels and glucose transport activity were evaluated following exposure to Glc/GO or insulin for 24 h in the absence and presence of rapamycin. L6 myotubes were incubated with either 100 nM insulin or Glc/GO for 24 h in the absence and presence of 30 or 100 ng/ml rapamycin. Under these conditions, cell viability as detected by protein recovery and the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide test was not affected. Fig. 6B demonstrates that rapamycin did not exert an effect on 2-deoxyglucose uptake activity of control cells. However, 100 ng/ml rapamycin totally prevented insulin-induced elevation of Glut1 mRNA levels (Fig. 6A) and glucose transport activity (Fig. 6B). In contrast, no similar effect of rapamycin was observed on oxidative stress-induced elevation of Glut1 mRNA levels and glucose transport activity. Taken together, these results provide evidence that oxidative stress-induced Glut1 transcriptional activation is distinct from Glut1 activation by insulin.

[View Larger Version of this Image (41K GIF file)]

DISCUSSION

With accumulating evidence suggesting a significant role for oxidative stress in diverse pathophysiological conditions, there is increasing interest in understanding cellular response to oxidative stress. The ability of oxidative stress to alter gene expression has been shown in many cell types and in various experimental models (10, 11). We have previously demonstrated that the expression of the glucose transporter isoform Glut1 is increased in both L6 myotubes and 3T3-L1 adipocytes in response to oxidative stress, resulting in increased glucose transport and metabolism (19, 20). These are believed to provide the elevated energy required for the recruitment of repair and antioxidant defense mechanisms. In this study, we investigated further the cellular mechanisms by which Glut1 expression is increased following exposure to oxidative stress.

The results of this study indicate that in L6 myotubes, the increase in steady-state Glut1 mRNA levels following oxidative stress can be attributed to both an enhanced Glut1 transcription rate, as determined by nuclear run-on assay (Fig. 1), and increased mRNA stability (Fig. 2). The fact that steady-state Glut1 mRNA levels are increased by not more than 2-2.5-fold (Table I) suggests that additional mechanisms may be involved in regulating Glut1 mRNA levels following oxidative stress. For example, increased mRNA stability may develop as a consequence of an earlier occurring enhancement of the transcription rate if mRNA degradation mechanisms become rate-limiting.

A combination of increased mRNA stability and elevated transcription rate in the regulation of this glucose transporter gene expression has been observed in clone 9 cells exposed to the oxidative phosphorylation inhibitor sodium azide (29). Increased stability of Glut1 mRNA was reported in cells exposed to tumor necrosis factor  (7) and phorbol ester and glucose deprivation (30), whereas an increased transcription rate without an alteration in mRNA stability was reported in ARL 15 cells treated with thyroid hormone (31, 32). The direct cellular mechanism responsible for elevated steady-state Glut1 mRNA levels in response to anoxia, iron chelators, malignant transformation, heat shock, and viral infection has yet to be determined. However, it appears that increasing Glut1 gene expression represents a common cellular response to various stress conditions, including oxidative stress, by mechanisms that result in transcriptional activation and/or stabilization of pre-existing mRNA transcripts. Alternative mechanisms to increase Glut1 protein content to enhance glucose flux may occur at the protein stability level.

The transcriptional regulation of Glut1 in response to oxidative stress is mediated at least partly by enhancer 1 of the regulatory region of this gene (Fig. 3). Among the three transcription factors known to have a potential binding site on enhancer 1, AP-1 and SRE are shown to be activated in response to oxidative stress (Fig. 4), whereas CRE is not. While this excludes the direct or indirect (through the activation of another gene) involvement of CRE in Glut1 transcriptional activation, the contribution of AP-1 or SRE activation to this process cannot be conclusively sorted out by this study. The dependence of this process on de novo protein synthesis is demonstrated by the ability of cycloheximide to inhibit elevation of Glut1 mRNA levels following oxidation. Under the same conditions, oxidative stress-induced AP-1 activation is also inhibited, which may provide circumstantial evidence for a direct or indirect role for AP-1 in Glut1 transcriptional activation following oxidative stress.

H₂O₂ has been shown to increase cellular protein tyrosine phosphorylation by altering the tyrosine kinase-to-phosphatase balance and was further demonstrated to exhibit insulinomimetic effects (33, 34). The effect of chronic exposure to insulin on Glut1 induction was studied using various experimental models. In L6 myotubes, the activation of Glut1 following an 18-h exposure to insulin was inhibited by rapamycin, a pp70^S6K inhibitor (28). In NIH/3T3HIR3.5 cells transfected with hybrid genes containing the enhancer and promotor of the Glut1 gene, a 1-h exposure to insulin resulted in activation of SRE and enhancer 1 (22). To compare the effect of chronic insulin with the effect of prolonged oxidative stress on Glut1 gene expression, we assessed the ability of rapamycin to prevent Glut1 activation. While rapamycin completely prevented the activation of glucose transport and Glut1 following chronic exposure to insulin, no similar effect on the induction of Glut1 in response to oxidative stress was observed (Fig. 6). Thus, the activation of Glut1 by prolonged low grade oxidative stress may not be identical to that exerted by chronic exposure to insulin.

In conclusion, the cellular response of L6 myotubes to oxidative stress involves the transcriptional activation of the Glut1 gene, mediated at least partly by enhancer 1 of the Glut1 gene, and requires de novo protein synthesis. This effect may not be attributed to an insulinomimetic effect of H₂O₂. The respective role of various transcription factors including AP-1 and SRE in mediating this response requires further study.