Restoration of Hypoxia-stimulated Glucose Uptake in GLUT4-deficient Muscles by Muscle-specific GLUT4 Transgenic Complementation*

To investigate whether GLUT4 is required for exercise/hypoxia-induced glucose uptake, we assessed glucose uptake under hypoxia and normoxia in extensor digitorum longus (EDL) and soleus muscles from GLUT4-deficient mice. In EDL and soleus from wild type control mice, hypoxia increased 2-deoxyglucose uptake 2–3-fold. Conversely, hypoxia did not alter 2-deoxyglucose uptake in either EDL or soleus from either male or female GLUT4-null mice. Next we introduced the fast-twitch skeletal muscle-specific MLC-GLUT4 transgene into GLUT4-null mice to determine whether changes in the metabolic milieu accounted for the lack of hypoxia-mediated glucose transport. Transgenic complementation of GLUT4 in EDL was sufficient to restore hypoxia-mediated glucose uptake. Soleus muscles from MLC-GLUT4-null mice were transgene-negative, and hypoxia-stimulated 2-deoxyglucose uptake was not restored. Although ablation of GLUT4 in EDL did not affect normoxic glycogen levels, restoration of GLUT4 to EDL led to an increase in glycogen under hypoxic conditions. Male GLUT4-null soleus displayed reduced normoxic glycogen stores, but female null soleus contained significantly more glycogen under normoxia and hypoxia. Reduced normoxic levels of ATP and phosphocreatine were measured in male GLUT4-null soleus but not in EDL. However, transgenic complementation of GLUT4 prevented the decrease in hypoxic ATP and phosphocreatine levels noted in male GLUT4-null and control EDL. In conclusion, we have demonstrated that GLUT4 plays an essential role in the regulation of muscle glucose uptake in response to hypoxia. Because hypoxia is a useful model for exercise, our results suggest that stimulation of glucose transport in response to exercise in skeletal muscle is totally dependent upon GLUT4. Furthermore, the compensatory glucose transport system that exists in GLUT4-null soleus muscle is not sensitive to hypoxia/muscle contraction.

GLUT4 is the predominant mammalian facilitative glucose transporter isoform expressed in insulin-sensitive tissues including skeletal muscle, adipose tissue, and heart (1)(2)(3)(4)(5). In skeletal muscle, glucose transport is mediated by at least two different pathways: one stimulated by insulin (6 -10), insulin mimicking agents (11), and insulin-like growth factors (12) and one activated by exercise/muscle contraction (6,7,13) or hypoxia (8,14,15). When these two pathways are stimulated congruently, glucose transport activity is increased in an additive manner (6), providing evidence for separate and distinct signaling pathways to glucose transport (14,16,17). Insulinstimulated glucose transport is mediated by phosphatidylinositol 3-kinase (14,16,17), whereas muscle contraction and hypoxia are believed to activate glucose transport by a calciummediated mechanism (8,18,19). The molecular mechanism by which insulin and/or muscle contractions/hypoxia increase glucose transport involves translocation of GLUT4 from an intracellular compartment to plasma membrane and transverse tubules (20 -27). Hypoxia is thought to activate glucose transport by mimicking the effects of muscle contraction on calcium release from sarcoplasmic reticulum (8). However, important differences exist between hypoxia and contraction in stimulating glucose uptake (28). Nevertheless, evidence exists for two distinct intracellular vesicular pools of GLUT4 that are sensitive to either insulin and IGF-1 or exercise and hypoxia (22,23,26).
Several groups have shown that GLUT4 overexpression in skeletal muscle leads to enhanced basal and insulin-stimulated glucose uptake (29 -31), and this is attributed to increased glucose transport in skeletal muscle (32,33). However, recent studies with GLUT4-overexpressing transgenic mice were not able to provide a clear link between increased GLUT4 content and enhanced hypoxia-stimulated glucose uptake (33). Thus, overexpressed GLUT4 may be targeted to an insulin-sensitive, rather than a muscle contraction/hypoxia-sensitive pool. This raises the question of whether exercise/hypoxia-induced glucose transport is totally dependent upon GLUT4 or whether other members of the glucose transporter family play a role.
Previously we examined the consequences of genetic ablation of GLUT4 on basal and insulin-stimulated glucose transport in skeletal muscle (34). In EDL 1 muscles from male or female GLUT4-null mice, basal 2-deoxyglucose uptake was somewhat lower than wild type control muscles, and insulin had no significant effect. Interestingly, basal 2-deoxyglucose uptake was increased 2-fold in soleus from male GLUT4-null mice compared with controls, with no further increase noted under insulin-stimulated conditions. More surprisingly, in soleus from female GLUT4-null mice, insulin induced a 2-fold increase in 2-deoxyglucose uptake, with no difference under basal condi-tions. These results provide evidence for the existence of a compensatory glucose transport system in oxidative muscles of GLUT4-null mice. However, because the response to exercise/ hypoxia has not been investigated in these animals, it is not known whether this compensatory mechanism is sensitive to stimuli such as exercise or hypoxia.
To investigate whether GLUT4 is necessary for exercise/ hypoxia-induced glucose uptake, we determined glucose uptake under hypoxic and normoxic conditions in EDL and soleus skeletal muscles from GLUT4-deficient mice. Hypoxia exposure led to a 2-3-fold increase in 2-deoxyglucose uptake in EDL and soleus muscle from male and female control mice. Conversely, in muscle from male and female GLUT4-null mice, hypoxia exposure did not stimulate 2-deoxyglucose uptake. Because the magnitude of increase in hypoxia-stimulated glucose uptake can be influenced by the duration of exposure to a hypoxic environment, intramuscular glycogen, ATP, phosphocreatine, and lactate levels were measured to assess muscle viability and to compare metabolic effects of hypoxia between GLUT4-null and control muscles. With the exception that male soleus muscle lacking GLUT4 exhibited lower ATP content, no significant differences were observed in the levels of these metabolites between GLUT4-null and wild type control groups in both muscle types from both sexes after hypoxia treatment. These results suggest that the lack of hypoxia-stimulated 2-deoxyglucose uptake in GLUT4-null muscles cannot be attributed to differences in metabolite levels following hypoxia. GLUT4null mice display altered whole body glucose and lipid metabolism (35,36). Thus, we have introduced a GLUT4 transgene driven by the myosin light chain 1 promoter (MLC) that targets expression to fast-twitch skeletal into GLUT4-null mice to provide direct evidence that GLUT4 ablation, rather than changes in the metabolic milieu, accounts for the lack of hypoxia-mediated glucose transport. The resulting MLC-GLUT4-null mice display normal glucose homeostasis and insulin action (36). In EDL muscles of MLC-GLUT4-null mice, GLUT4 complementation fully restored hypoxia-stimulated 2-deoxyglucose uptake. Soleus muscles from MLC-GLUT4-null mice were transgene-negative (36) and hypoxia-stimulated 2-deoxyglucose uptake was not restored. These results strongly suggest that GLUT4 is essential for hypoxia-mediated glucose transport in skeletal muscle.

EXPERIMENTAL PROCEDURES
Animals-MLC-GLUT4-null mice were obtained by crossing GLUT4null mice to MLC-GLUT4 mice (36). The GLUT4 locus segregates independently from the MLC-GLUT4 transgene locus. MLC-GLUT4null mice were genotyped for the MLC-GLUT4 transgene and GLUT4 gene disruption as described previously (31,35). Mice between 10 and 14 weeks old were used in this study. Animals were fed ad libitum and maintained in a murine hepatitis virus-free barrier facility on a 12-h light and dark cycle. All protocols have been approved by the Animal Care and Use Committee of the Albert Einstein College of Medicine in accordance with the Public Health Service Animal Welfare Policy.
Muscle Incubation Procedure-Animals were killed by cervical dislocation, and soleus and EDL muscles were rapidly isolated. Muscles from an individual mouse were randomly divided such that one was used for normoxic and the other for hypoxic incubations. Muscles were incubated in 1 ml of Krebs-Ringer bicarbonate (KRB) buffer (118.6 mM NaCl, 4.76 mM KCl, 1.19 mM KH 2 PO 4 , 1.19 mM MgSO 4 , 2.54 mM CaCl 2 , 5 mM Hepes, pH 7.4) supplemented with 0.1% bovine serum albumin (fraction V, RIA grade, Sigma). Normoxic medium was pre-gassed with 95% O 2 /5% CO 2 , and hypoxic medium was pre-gassed with 95% N 2 /5% CO 2 . Muscles were pre-incubated under either normoxic or hypoxic conditions for 45 min (30°C) in KRB medium supplemented with 5 mM D-glucose. The specific gas mixture was maintained throughout the pre-exposure period.
2-Deoxyglucose Uptake in Isolated Muscle-Following pre-incubation, muscles were transferred to oxygenated KRB containing 2 mM pyruvic acid and 3 mM mannitol (Sigma). Muscles were incubated for 15 min at 30°C under a gas phase of 95%O 2 /5%CO 2 . This incubation step was included to remove glucose from the extracellular space before measuring glucose transport and, in the case of hypoxia-stimulated muscle, to allow for resynthesis of high energy phosphates (8). 2-Deoxyglucose uptake was measured as described previously (37). Rate of intracellular 2-deoxyglucose accumulation is expressed as nmol ϫ mg muscle protein Ϫ1 ϫ 20 min Ϫ1 . Protein was determined by the BCA assay (Pierce).
Biochemical Assays-Following incubation as described above, muscles were immediately freeze-clamped on a block of dry ice and stored at Ϫ80°C for subsequent metabolic analysis as described earlier (38). ATP, phosphocreatine, lactate, and glycogen were assessed fluorometrically on perchloric acid extracts of freeze-dried muscle according to Lowry and Passonneau (39) with modifications indicated by Wallberg-Henriksson et al. (38). Glycogen was assessed in the extract and pellet, and results were combined for total muscle glycogen concentration. Results are expressed as mmol glucose units ϫ kg Ϫ1 dry weight.
Statistics-To asses effects between genotypes (control, GLUT4-null, and MLC-GLUT4-null), statistical differences were determined by oneway ANOVA. When a significant F ratio was obtained, Fisher's PLSD post hoc test was employed to identify statistical differences (p Ͻ 0.05) between the means. To evaluate differences between treatment (normoxic versus hypoxic) within a genotype, Student's unpaired t test was employed.

Glucose Uptake Following Normoxic or Hypoxic Exposure-
Isolated EDL and soleus muscles were incubated for 45 min under normoxic (95% O 2 /5% CO 2 ) or hypoxic (95% N 2 /5% CO 2 ) atmosphere. Thereafter, muscles were exposed to oxygenated glucose-free medium for 15 min, followed by assessment of 2-deoxyglucose uptake under aerobic conditions. Hypoxia induced a significant 2.0 -2.5-fold increase (p Ͻ 0.005) in 2-deoxyglucose uptake in EDL ( Fig. 1) and soleus ( Fig. 2) muscle from male and female wild type control mice. No effect of hypoxia was noted on 2-deoxyglucose uptake in either EDL or soleus muscle from male or female GLUT4-null mice. Restoration of GLUT4 in EDL muscle from male or female MLC-GLUT4-null mice led to a significantly marked increase in hypoxia-mediated glucose transport by 3.7-and 3.1-fold, respectively (p Ͻ 0.0001) when compared with GLUT4-nulls. Soleus muscles from MLC-GLUT4-null mice were transgenenegative (36), and hypoxia-stimulated 2-deoxyglucose uptake was not restored to control levels.
Under normoxic conditions, 2-deoxyglucose uptake in EDL muscles were similar between MLC-GLUT4-null and wild type control male and female mice (Fig. 1). However, 2-deoxyglucose uptake was 40 -45% lower in EDL muscle from GLUT4-null versus control and MLC-GLUT4-null animals. In soleus muscle (Fig. 2), similar rates of 2-deoxyglucose uptake were noted between male and female control, GLUT4-null, and MLC-GLUT4-null mice under normoxic conditions. Glycogen Content Following Normoxic or Hypoxic Exposure-We next examined glycogen content in EDL and soleus muscles from male and female wild type control, GLUT4-null, and MLC-GLUT4-null mice. Muscles were incubated under normoxic or hypoxic conditions, as described above for glucose transport measurements. Although glycogen content tended to be lower in EDL (Fig. 3) and soleus (Fig. 4) muscles following hypoxia exposure, significant decreases were only noted in muscles from control female mice (p Ͻ 0.05). In EDL muscle, glycogen content was similar between the different genotypes under normoxic conditions and similar between control and GLUT4-null mice under hypoxic conditions. However, glycogen content in EDL muscle following hypoxia was approximately 2-fold lower in male GLUT4-null than MLC-GLUT4-nulls (Fig.  3A). Similarly, female MLC-GLUT4-null EDL displayed an approximately 30 -40% increase in glycogen content when compared with control or GLUT4-null EDL after hypoxic treatment (Fig. 3B). In contrast to EDL, soleus muscle from male GLUT4 null and MLC-GLUT4-null mice contained approximately 40% less glycogen than controls under normoxic conditions (Fig.  4A). In female GLUT4-null mice, soleus muscle glycogen levels were increased approximately 1.8-fold under normoxic conditions compared with control and MLC-GLUT4-null mice (Fig.  4B). Under hypoxic conditions, soleus muscle glycogen content was similar among control, GLUT4-null, and MLC-GLUT4null soleus, except in female GLUT4-null mice where a 2.0-fold increase compared with control mice was noted following hypoxia exposure (Fig. 4).
ATP, Phosphocreatine, and Lactate Content Following Normoxic or Hypoxic Exposure-We next assessed ATP, phosphocreatine, and lactate in perchloric acid extracts of freeze-dried muscle exposed to normoxic or hypoxic conditions as described above for glycogen determinations. Generally, ATP and phosphocreatine levels were slightly lower in hypoxia-stimulated muscles. Levels of ATP (Table I) and phosphocreatine (Table II) in EDL or soleus were not significantly different between female wild type control and GLUT4-null mice under either normoxic or hypoxic conditions. In males, levels of ATP and phosphocreatine did not differ between control, GLUT4-null, and MLC-GLUT4-null EDL muscle under normoxic conditions or between control and GLUT4-null EDL muscle under hypoxic conditions. In contrast, ATP and phosphocreatine levels in normoxic GLUT4-null soleus were decreased 37 and 51%, re-spectively. Similar reductions in ATP and phosphocreatine levels were measured in the transgene-negative soleus of male MLC-GLUT4-null mice. Upon hypoxia treatment, the phosphocreatine levels of GLUT4-null and MLC-GLUT4-null soleus muscles were not significantly different from normoxic levels. However, the ATP levels of GLUT4-null and MLC-GLUT4-null soleus muscles under hypoxic conditions were further reduced when compared with normoxic conditions. Under hypoxic conditions, ATP levels in GLUT4-null and MLC-GLUT4-null soleus muscles were decreased by 57 and 38%, respectively, compared with controls.
Transgenic complementation of GLUT4 in EDL muscle of MLC-GLUT4-null mice maintained ATP and phosphocreatine levels to wild type control levels under normoxic conditions. Following hypoxia treatment, ATP and phosphocreatine levels in male MLC-GLUT4-null EDL muscles were increased 30 and 34%, respectively, when compared with control EDL. Lactate levels were not different in either EDL or soleus muscle between control, GLUT4-null, and MLC-GLUT4-null mice of both sexes (Table III). Lactate levels were generally higher in hypoxia-exposed muscles, with statistical increases noted in muscle from male MLC-GLUT4-null mice and soleus muscles from male controls and female MLC-GLUT4-null mice. DISCUSSION We provide the first direct evidence that GLUT4 is required for hypoxia-mediated glucose transport in skeletal muscle. This observation is based on our finding that GLUT4 ablation in isolated soleus and EDL muscles prevented increased 2-deoxyglucose uptake following exposure to hypoxia. Furthermore, transgenic complementation of GLUT4 was sufficient to fully restore hypoxia-induced glucose transport in EDL muscle from MLC-GLUT4-null. In vitro exposure of isolated skeletal muscle to hypoxia, followed by re-oxygenation, is known to increase glucose transport (8,14,15,33,40,41) and GLUT4 recruitment to the cell surface (8,14). Because hypoxia appears to be a relevant model for exercise-stimulated glucose transport (8), our results suggest that GLUT4 is the only downstream effector of the muscle contraction/hypoxia pathway to glucose transport. In contrast, proteins other than GLUT4 may mediate the insulin-sensitive glucose transport process. Although we have reported that EDL muscles from GLUT4-null mice fail to demonstrate a significant increase in insulin-stimulated 2-deoxyglucose uptake (34,36), soleus muscles from female GLUT4null mice elicit a significant increase in insulin-stimulated glucose transport, which cannot be accounted for by changes in GLUT1 protein expression (34). These results provide evidence for a novel compensatory glucose transport system that is sensitive to insulin but not hypoxia.
In a previous study using transgenic mice overexpressing GLUT4 using its endogenous promoter, contraction of EDL muscles by in situ electrical stimulation resulted in a greater increase in 2-deoxyglucose uptake compared with wild type control muscle (32), suggesting that contraction-induced glucose uptake is mediated by GLUT4. However, a similar GLUT4 transgenic model was shown to exhibit increased insulin-mediated, but not hypoxia-mediated, glucose transport in skeletal muscle (37). The discrepancy between the reported findings with in situ contraction and in vitro exposure of muscle to hypoxia are unclear because hypoxia is believed to activate glucose transport via the same mechanism as muscle contraction (8). Our results clearly demonstrate that GLUT4 is essential for hypoxia-mediated glucose uptake. Furthermore, we show that GLUT4 overexpressed from a transgene can be targeted to a hypoxia-sensitive compartment. Consistent with Brozinick and co-workers (33), we did not observe a greater increase in hypoxia-stimulated glucose uptake in EDL muscle of MLC-GLUT4-null compared with control mice. GLUT4 content in EDL muscle from female MLC-GLUT4 mice is similar to controls (36); thus, a greater level of hypoxia-stimulated glucose uptake between transgenic complemented and control mice might not be expected. However, in male MLC-GLUT4null mice, GLUT4 content in EDL muscle is increased by 360% (36), consequently, a greater increase in hypoxia-mediated glucose uptake can be predicted. Indeed, we reported that insulinstimulated glucose uptake is significantly greater in EDL muscle from male and female MLC-GLUT4-null mice compared with wild type controls (36). Thus, differences may exist be- FIG. 3. Skeletal muscle glycogen levels in EDL muscle from wild type control, GLUT4-null, and MLC-GLUT4-null mice. EDL muscles were isolated from male (A) or female (B) mice and incubated as described in the legend to Fig. 1 under oxygenated (open bars) or hypoxic (closed bars) conditions. Glycogen was determined in extracts prepared from freeze-dried muscle and in the pellet (see "Experimental Procedures"). Results were combined to give total muscle glycogen content. Results are expressed as mmol glucose units ϫ kg Ϫ1 dry weight. Values are the means Ϯ S.E. for 7-9 muscles/group. Statistical difference between genotypes was assessed by one-way ANOVA. *, p Ͻ 0.05, significantly different from control mice; #, p Ͻ 0.05, significantly different from GLUT4-null mice.
FIG. 4. Skeletal muscle glycogen levels in soleus muscle from wild type control, GLUT4-null, and MLC-GLUT4-null mice following exposure to oxygenated or hypoxic conditions. Soleus muscles were isolated from male (A) or female (B) mice and incubated as described in the legend to Fig. 1 under oxygenated (open bars) or hypoxic (closed bars) conditions. Total muscle glycogen concentration was determined (see "Experimental Procedures"). Results are expressed as mmol glucose units ϫ kg Ϫ1 dry weight. Values are the means Ϯ S.E. for 7-9 muscles/group. Statistical difference between genotypes was assessed by one-way ANOVA. *, p Ͻ 0.05, significantly different from control mice; **, p Ͻ 0.01, significantly different from control and MLC-GLUT4-null mice.
tween the recruitment machinery for insulin-and hypoxia/ exercise-induced GLUT4 translocation. Our results indicate that there may be a limit in the hypoxia-induced glucose transport system, which cannot be overcome by overexpression of GLUT4. To achieve a further increase in hypoxia-mediated glucose uptake, overexpression of additional components of the translocation machinery for GLUT4 may be required.
An alternative explanation for the lack of a further increase in hypoxia-stimulated glucose uptake in GLUT4-overexpressing muscles is that the hypoxic conditions used in the present study were unable to generate a stimulatory signal sufficient for the translocation of overexpressed GLUT4. Unlike male control and male and female GLUT4-null EDL muscles, ATP levels in transgene expressing MLC-GLUT4-null EDL muscles were not lower with hypoxia treatment. Male MLC-GLUT4null EDL muscles also exhibited higher phosphocreatine and glycogen levels than control and GLUT4-null EDL muscles following hypoxia. These results indicate that MLC-GLUT4null EDL muscle was not affected by hypoxia to the same extent as GLUT4-null or control EDL. It is possible that increased basal glucose uptake seen in male MLC-GLUT4-null EDL muscle (36) can confer a superior ability to maintain glycogen and high energy phosphate levels and minimize cellular stress caused by hypoxia. It is unknown whether hypoxia stimulates glucose transport solely by mobilizing calcium stores or also by causing cellular stress. Although no correlation exists between levels of in vitro muscular glucose transport and ATP concentration, the degree of increase in glucose transport is affected by the amount of time exposed to hypoxic environment (8). In addition, glucose transport rates are negatively correlated with intramuscular concentrations of phosphocreatine following hypoxia (43). These previous studies suggest that hypoxia-induced cellular stress is a potentiating factor in determining the magnitude of hypoxia-mediated increase in glucose uptake.
Although overexpression of GLUT4 in MLC-GLUT4 EDL muscle prevented the decrease in ATP and phosphocreatine levels with hypoxia treatment, lack of GLUT4 did not appear to influence the level of ATP or phosphocreatine in EDL muscle from either male or female mice under hypoxic and normoxic conditions. In soleus muscle, GLUT4 ablation resulted in dramatically reduced levels of ATP and phosphocreatine in male but not female animals. Decreased ATP levels in male GLUT4null and MLC-GLUT4-null soleus muscle following hypoxia exposure can be attributed to decreased ATP and phosphocreatine levels under normoxic conditions. In addition, one striking difference between soleus muscle from male and female GLUT4-null mice is the level of muscle glycogen, with approximately 50% lower levels noted in male versus female GLUT4null mice. Although glycogen content in male GLUT4-null soleus was reduced 43%, glycogen content in female GLUT4-null soleus was increased 75% when compared with sex-matched controls. Reduced muscle glycogen levels may be one factor   leading to the compromised energy state observed in the soleus muscle from male GLUT4-null mice. Similarly, elevated glycogen content in female GLUT4-null soleus muscle appears to have a sparing effect on depletion of ATP and phosphocreatine following hypoxia exposure. We have previously reported sexually dimorphic responses to GLUT4 ablation (34 -36). Female GLUT4 null soleus muscle displays significant, although blunted, insulin-stimulated glucose uptake (34). Upon careful examination of the glycogen, ATP, and phosphocreatine levels of female soleus muscles under normoxic and hypoxic conditions, a strong similarity can be drawn to the control male soleus under similar conditions ( Fig. 4 and Tables I and II). The underlying factor(s) responsible for this observation is presently unknown. Several important points can be made regarding the effects of variable GLUT4 expression on skeletal muscle glycogen content. First, transgenic complementation of GLUT4 in EDL muscle from male and female GLUT4-null mice led to a significant increase in muscle glycogen levels. This finding is in agreement with several reports whereby increased muscle glycogen levels have been noted in GLUT4 transgenic mice (29,32,33,42) and suggests that enhanced glucose flux, presumably because of increased expression of GLUT4, leads to greater glycogen storage in skeletal muscle. Our finding that glycogen content in soleus muscle of male GLUT4-null mice was 45% lower than control levels supports this concept. However, the relationship between GLUT4 expression and glycogen content may not be as strong as previously suggested. For example, glycogen levels were significantly elevated in soleus muscle from female GLUT4-null mice compared with controls under both hypoxic and normoxic conditions. This finding is consistent with our previous report of increased glycogen levels and enhanced glucose incorporation to glycogen in soleus muscle from female GLUT4-null mice (34), and provides evidence that GLUT4 expression may not be directly linked to muscle glycogen content.
In conclusion, we have demonstrated that GLUT4 plays an essential role in the regulation of skeletal muscle glucose uptake in response to hypoxia. As hypoxia is a useful model for exercise, our results suggest that exercise, a potent stimulator of glucose uptake in skeletal muscle, is totally dependent upon GLUT4. Furthermore, the compensatory glucose transport system expressed in soleus muscle of GLUT4-null mice is not sensitive to hypoxia.