Hexokinase II Overexpression Improves Exercise-Stimulated But Not Insulin-Stimulated Muscle Glucose Uptake in High-Fat-Fed C57BL/6J Mice

All procedures performed were approved by the Vanderbilt University Animal Care and Use Committee. Male FVB/NJ mice that selectively overexpress HK II in skeletal muscle were obtained from Dr. David E. Moller (8) and backcrossed onto the C57BL/6J background for at least five generations. The HK II transgene is composed of the human HK II cDNA driven by the rat muscle creatine kinase promoter and yields a three-, five-, and sevenfold increase in HK activity in the soleus, gastrocnemius, and superficial vastus lateralis (SVL) muscles, respectively. This degree of overexpression is within the normal physiological range of HK activity in skeletal muscle (9,10). Male mice carrying the HK II transgene (HK^Tg) were subsequently bred with wild-type (WT) female mice. After a 3-week weaning period, littermates were separated by sex and maintained in microisolator cages. Genotyping for the HK II transgene was performed on genomic DNA obtained from a tail biopsy with the PCR as previously described (4). Mice were fed either a standard diet or a high-fat diet ad libitum and were studied at ∼4 months of age. The high-fat diet contains 35.5% fat (Diet F3282; Bio-Serv, Frenchtown, NJ). This model of dietary-induced insulin resistance creates both fasting hyperglycemia and hyperinsulinemia and thus is a reasonable model for the human condition (5,6).

The surgical procedures were the same as those described previously (4,11,12). Mice were anesthetized with pentobarbital (70 mg/kg body wt). The left common carotid artery was catheterized for sampling with a two-part catheter consisting of PE-10 (inserted into the artery) and silastic (0.025 in outer diameter [OD]). The right jugular vein was catheterized for infusions with a silastic catheter (0.025 in OD). The free ends of catheters were tunneled under the skin to the back of the neck, where they were attached via stainless steel connectors to tubing made of Micro-Renathane (0.033 in OD), which were exteriorized and sealed with stainless steel plugs. Lines were flushed daily with ∼50 μl of saline containing 200 units/ml heparin and 5 mg/ml ampicillin. Animals were individually housed after surgery, and body weight was recorded daily.

Metabolic experiments were performed as previously described (4) and were conducted after a postoperative recovery period of either ∼5 days for hyperinsulinemic-euglycemic clamp experiments or ∼7 days for exercise experiments. The recovery period was a sufficient time for body weight to be restored within 10% of presurgery body weight. For metabolic studies, conscious, unrestrained mice were placed in an ∼1-l plastic container lined with bedding and fasted for 5 h. Approximately 1 h before an experiment, Micro-Renathane (0.033 in OD) tubing was connected to the catheter leads and infusion syringes. After a 1-h acclimation period, a baseline arterial blood sample (150 μl) was drawn for the measurement of arterial blood glucose, hematocrit, and plasma insulin and nonesterified fatty acids (NEFAs). The remaining erythrocytes were washed with 0.9% heparinized saline and reinfused.

For hyperinsulinemic-euglycemic clamp and saline control experiments, an infusion of saline (n = 11 WT standard diet-fed; 11 HK^Tg standard diet-fed; 7 WT high-fat-fed; 7 HK^Tg high-fat-fed) or 4 mU · kg⁻¹ · min⁻¹ insulin (n = 14 WT standard diet-fed; 11 HK^Tg standard diet-fed; 10 WT high-fat-fed; 8 HK^Tg high-fat-fed) was begun at a rate of 1.375 μl/min at t = −90 min. Euglycemia was maintained during insulin experiments by frequently measuring arterial blood glucose (∼5 μl; HemoCue, Mission Viejo, CA) and infusing 50% dextrose as necessary. Mice received saline-washed erythrocytes from a donor mouse as needed to prevent a marked fall in hematocrit (>5%). At t = 0 min, an arterial blood sample (150 μl) was obtained and processed as the baseline blood sample. At t = 5 min, a 12-μCi bolus of 2-deoxy[³H]glucose ([2-³H]DG) was administered to determine R_g, an index of tissue-specific glucose uptake. At t = 7, 10, 15, and 20 min, arterial blood (∼50 μl) was sampled to determine arterial blood glucose and plasma [2-³H]DG. At t = 30 min, a final arterial blood sample (150 μl) was withdrawn and processed as the baseline blood sample with the addition of the determination of plasma [2-³H]DG, and mice were anesthetized with an infusion of sodium pentobarbital. The gastrocnemius (∼6% type IIA, ∼11% type IID, ∼83% type IIB fibers), SVL (∼3% type IIA, ∼10 type IID, ∼87% type IIB fibers), and soleus (∼44% type I, ∼51% type IIA, ∼5% type IID fibers) muscles (13) were excised, immediately frozen in liquid nitrogen, and stored at −70°C until future tissue analysis.

For exercise and sedentary control experiments, mice were acclimated to treadmill running with a single 10-min bout of exercise (0.5–0.6 mph, 0% grade) ∼5 days postsurgery. Experiments were performed 2 days after the treadmill acclimation trial. After the baseline arterial blood sampling, mice either remained sedentary (n = 9 WT standard diet-fed; 12 HK^Tg standard diet-fed; 7 WT high-fat-fed; 8 HK^Tg high-fat-fed) or ran (n = 9 WT standard diet-fed; 10 HK^Tg standard diet-fed; 8 WT high-fat-fed; 8 HK^Tg high-fat-fed) on a motorized treadmill (0.6 mph, 0% grade) for 30 min. At t = 5 min, a 12-μCi bolus of [2-³H]DG was administered to determine R_g. At t = 10, 15, and 20 min, arterial blood (∼50 μl) was sampled to determine blood glucose and plasma [2-³H]DG. At t = 30 min, a final arterial blood sample (150 μl) was taken and processed as the baseline blood sample with the addition of the determination of plasma [2-³H]DG, mice were anesthetized, and muscles were excised as described for the clamp studies.

Immunoreactive insulin was assayed with a double-antibody method (14). NEFAs were measured spectrophotometrically by an enzymatic colorimetric assay (Wako NEFA C kit; Wako Chemicals, Richmond, VA). After deproteinization with barium hydroxide [Ba(OH)₂, 0.3 N] and zinc sulfate [ZnSO₄, 0.3 N], [2-³H]DG radioactivity of plasma was determined by liquid scintillation counting (Packard TRI-CARB 2900TR; Packard, Meriden, CT) with Ultima Gold (Packard) as scintillant. Muscle samples were weighed and homogenized in 0.5% perchloric acid. Homogenates were centrifuged and neutralized with KOH. One aliquot was counted directly to determine [2-³H]DG and [2-³H]DG-6-phosphate ([2-³H]DGP) radioactivity. A second aliquot was treated with Ba(OH)₂ and ZnSO₄ to remove [2-³H]DGP and any tracer incorporated into glycogen (15) and then counted to determine [2-³H]DG radioactivity. [2-³H]DGP is the difference between the two aliquots. In all experiments, the accumulation of [2-³H]DGP was normalized to tissue weight.

Muscle glycogen was determined by the method of Chan and Exton (16) on the contralateral gastrocnemius and SVL muscles. Soleus glycogen could not be determined because both of the small muscles were used to assay for [2-³H]DG and [2-³H]DGP. After deproteinization with 0.5% perchloric acid, tissue glucose and G6P were measured enzymatically (17) and expressed as millimoles per liter tissue water.

R_g was calculated by the equation (18)

where [2-³H]DGP_tissue is the [2-³H]DGP radioactivity in the muscle in dpm/g, AUC [2-³H]DG_plasma is the area under the plasma [2-³H]DG disappearance curve in dpm · ml⁻¹ · min⁻¹, and glucose_plasma is the average blood glucose in mmol/l during the experimental period.

Data are presented as mean ± SE. Differences between groups were determined by ANOVA. The significance level was set at P < 0.05.

Baseline descriptive characteristics for 5-h-fasted mice are reported in Table 1. HK II overexpression had no effect on body weight. High-fat feeding increased body weight by ∼10 g in both WT and HK^Tg mice. Baseline arterial blood glucose, insulin, and NEFA concentrations were not influenced by HK II overexpression. High-fat feeding increased arterial blood glucose in WT mice and caused hyperinsulinemia. HK II overexpression protected mice against diet-induced fasting hyperglycemia but not hyperinsulinemia. NEFAs were not altered by high-fat feeding regardless of genotype.

All measurements were made after 90 min of saline or insulin infusion to ensure steady-state conditions. Arterial blood glucose was not affected by high-fat feeding or HK II overexpression in saline-infused mice (Fig. 1A). Arterial blood glucose was similar in all groups during the hyperinsulinemic-euglycemic clamp experiments (Fig. 1B). In addition, there were no detectable differences in arterial blood glucose between saline- and insulin-infused mice.

The glucose infusion rates (GIRs) required to maintain euglycemia averaged for the final 30 min of the experiment are reported in Fig. 2. GIR was increased by HK II overexpression in standard diet-fed mice, but this increase was not statistically significant (67 ± 8 vs. 58 ± 3 mg · kg⁻¹ · min⁻¹; P = 0.11). High-fat feeding decreased GIR, and HK II overexpression was unable to correct the reduced GIR associated with high-fat feeding.

Disappearance of [2-³H]DG from arterial plasma is shown in Fig. 3. Neither genotype nor diet significantly altered the disappearance of [2-³H]DG in saline-infused mice. The disappearance of [2-³H]DG was increased in insulin-compared with saline-infused mice irrespective of genotype or diet. Standard diet-fed HK^Tg mice had an accelerated [2-³H]DG disappearance compared with standard diet-fed WT and high-fat-fed HK^Tg mice during insulin infusion.

R_g’s for the gastrocnemius, SVL, and soleus muscles during saline and insulin infusions are shown in Fig. 4. There were no differences in R_g when saline alone was infused. R_g was significantly increased during insulin infusion in all muscles regardless of genotype or diet. HK II overexpression augmented the insulin-stimulated R_g in all muscles of standard diet-fed mice. High-fat feeding blunted insulin-stimulated R_g, and as with GIR, HK II overexpression was unable to correct the impaired response.

Muscle glycogen content was unaltered by HK II overexpression or high-fat feeding independently (Table 2). However, glycogen content was reduced in these muscles when HK II overexpression and high-fat feeding were combined. Muscle glucose was increased by high-fat feeding in the gastrocnemius of WT saline-infused mice, and this increase was normalized by HK II overexpression (Table 3). Stimulation by insulin led to a fall in muscle glucose in the gastrocnemius and SVL of high-fat-fed WT mice, raising the possibility of impaired glucose delivery in muscles that consist of type II fibers. No significant differences in muscle G6P were detected between groups (Table 4).

Arterial blood glucose concentrations during sedentary and exercise experiments are shown in Fig. 5. There were no differences in arterial blood glucose between standard diet-fed WT and HK^Tg and high-fat-fed HK^Tg mice under sedentary conditions (Fig. 5A). The elevated arterial blood glucose observed in sedentary, high-fat-fed WT mice waned such that by the end of the study, concentrations were no longer greater than standard diet-fed mice. The relative hyperglycemia observed in high-fat-fed WT mice, however, persisted throughout exercise (Fig. 5B). There was a general trend for arterial blood glucose to rise to a plateau by 20 min of exercise in all groups.

Disappearance of [2-³H]DG from arterial plasma is shown in Fig. 6. Genotype and diet did not alter [2-³H]DG disappearance in sedentary mice (Fig. 6A). HK II overexpression augmented [2-³H]DG disappearance during exercise in both standard diet- and high-fat-fed mice (Fig. 6B).

R_g’s for soleus, gastrocnemius, and SVL during sedentary and exercise conditions are shown in Fig. 7. There were no differences in sedentary R_g between any of the groups. Exercise significantly increased R_g in the gastrocnemius and SVL regardless of genotype and diet. Exercise was able to increase R_g only in the soleus muscle of fed HK^Tg mice. The lack of an increase in soleus R_g of WT mice during exercise is due to the high tonic activity of this muscle. High-fat feeding blunted the exercise-stimulated R_g in the gastrocnemius and the SVL of WT mice. In addition, the exercising soleus R_g was reduced in high-fat-fed WT mice compared with standard diet-fed controls. The impairment of exercise-stimulated R_g as a result of high-fat feeding was markedly improved in the gastrocnemius and SVL muscles of HK^Tg mice. In addition, exercise-stimulated R_g in the gastrocnemius and SVL muscles was significantly increased in high-fat-fed HK^Tg compared with high-fat-fed WT mice. However, the attenuated exercise-stimulated R_g in the soleus as a result of high-fat feeding was not corrected by HK II overexpression.

Exercise led to a significant fall in glycogen content in the gastrocnemius and SVL of standard diet-fed WT mice and in the gastrocnemius of standard diet-fed HK^Tg mice (Table 2). Glycogen content in the gastrocnemius was greater in high-fat-fed compared with standard diet-fed WT mice after exercise. Muscle glucose was lowered by HK II overexpression in the gastrocnemius of sedentary standard diet-fed mice, the soleus of sedentary high-fat-fed mice, and the SVL of exercising high-fat-fed mice (Table 3). Exercise increased muscle G6P in the gastrocnemius of high-fat-fed WT mice and SVL of high-fat-fed HK^Tg mice (Table 4). HK II overexpression increased G6P in the gastrocnemius of sedentary, high-fat-fed mice and in the soleus of both sedentary and exercising, high-fat-fed mice.

MGU consists of three serial steps: the delivery of glucose from the blood to the muscle membrane, the transport across the muscle membrane, and the phosphorylation to G6P by an HK. As with other metabolic pathways (19,20), the regulation of MGU is distributed and may be altered by physiological conditions at more than one step. Likewise, dysregulation, resulting in impaired MGU, may occur at multiple steps of the glucose uptake pathway. The aim of the present study was to determine, in conscious, unrestrained C57BL/6J mice, the functional sites of impairment created by high-fat feeding and, more specifically, to determine whether a diet-induced form of insulin resistance creates a functional impairment to glucose phosphorylation. To this end, glucose fluxes (i.e., R_g’s) were measured during two independent physiological conditions characterized by high fluxes (i.e., stimulation by insulin or exercise) in both standard diet- and high-fat-fed mice with and without HK II selectively overexpressed in skeletal muscle. Here we show that high-fat feeding results in a functional impairment in glucose phosphorylation capacity in conscious, unrestrained C57BL/6J mice. It is also apparent from these studies that diminished glucose phosphorylation capacity is not the only deficit contributing to high-fat feeding-induced impairments in MGU in vivo.

High-fat feeding resulted in insulin resistance, as evidenced by the ∼30-mg/dl increase in fasting arterial blood glucose, the more than doubling of fasting plasma insulin concentration, and the marked reduction in exogenous glucose required to maintain euglycemia during insulin infusion. Whereas it has been demonstrated that high-fat feeding elicits deficits in glucose transport (21–25), less is known regarding high-fat feeding-induced deficits to steps both proximal and distal to glucose transport. Functional impairments to steps distal to glucose transport (e.g., glucose phosphorylation) can occur as a result of high-fat feeding (26) or accumulation of intracellular lipids (27). A recent report demonstrated that overexpression of glucokinase, an HK isozyme that is not inhibited by G6P, in skeletal muscle prevented glucose intolerance and insulin resistance in high-fat-fed mice (28). Our laboratory has provided evidence from the high-fat-fed conscious rat using a modeling approach (7) that diet-induced insulin resistance can be manifested by impairments to both delivery of glucose to the muscle and intracellular phosphorylation to G6P. More recently it was shown that the ability of insulin to increase bulk blood flow and capillary recruitment and thus glucose delivery is impaired in the obese Zucker rat, another animal model of insulin resistance (29). A recent study by Kim et al. (30) reported impaired glucose disposal in high-fat-fed rats despite unaltered GLUT4 content or HK activity, again emphasizing the possibility of impaired glucose delivery. The current study provides further direct and indirect evidence that multiple steps of MGU are impaired in a diet-induced model of insulin resistance as HK II overexpression rescued impaired MGU during exercise but neither in the basal state nor during insulin stimulation.

In non-insulin-resistant states, different physiological stimuli that increase glucose fluxes can be used in an attempt to differentially expose the barriers to MGU. Under basal conditions, glucose transport is the primary rate-controlling step of glucose uptake (31–33). However, upon stimulation by insulin (34,35) or exercise (36–38), GLUT4 translocates to the sarcolemma, making it highly permeable to glucose. GLUT4 translocation functionally lowers the barrier to glucose transport and, in effect, shifts the onus of control to glucose delivery and phosphorylation. In agreement with previous work in standard diet-fed FVB/NJ mice (4), glucose phosphorylation is a significant barrier to insulin- and exercise-stimulated MGU as transgenic overexpression of HK II selectively in skeletal muscle was able to augment both insulin- and exercise-stimulated R_g in the muscles of conscious, unrestrained, standard diet-fed C57BL/6J mice (see Figs. 4 and 7). It is important to note that under conditions of high resistance to glucose transport (e.g., saline infusion and sedentary protocols), HK II overexpression had no effect on R_g, providing further support for glucose transport’s important role in controlling MGU under basal, sedentary conditions.

It is interesting that the same transgene resulted in subtle phenotypic differences between the C57BL/6J and FVB/NJ mouse strains. For example, in the previous report, HK II overexpression augmented GIR and thus whole-body glucose disposal by ∼25%. Here, GIR was only increased by ∼15 in HK^Tg mice (P = 0.11). Whereas HK II overexpression augmented indexes of insulin-stimulated MGU in both mouse strains, the effectiveness of the transgene was increased in the C57BL/6J mice. Differences in physiological parameters between mouse strains are common; thus, one needs to exercise caution when comparing the phenotypes created by genetic manipulations.

High-fat feeding yielded a dramatic reduction in insulin-stimulated MGU that was not improved by HK II overexpression. This observation can be accounted for by one of three possibilities. The first possibility is that glucose phosphorylation during insulin stimulation is not impaired in skeletal muscles of mice, and thus HK II overexpression would not improve insulin-stimulated R_g. The second possibility is that insulin-stimulated translocation of GLUT4 is impaired by high-fat feeding, and, therefore, HK II overexpression alone would not improve insulin-stimulated R_g. The third possibility is that high-fat feeding impairs the function of all steps proximal to glucose phosphorylation during insulin stimulation, masking a functional impairment in glucose phosphorylation. Work from our laboratory using an independent method to assess functional deficits to the steps that compose MGU and another rodent model of insulin resistance, the high-fat-fed rat, suggest the latter (7). The data presented here cannot directly distinguish between potential deficits in glucose delivery and membrane transport.

Monogenic forms of insulin resistance have been created in mice by disrupting nitric oxide synthase (39) and GLUT4 (40,41). These genetic manipulations perturb skeletal muscle glucose delivery or transport and independently impair MGU. In addition, a reduction in HK II activity can impair MGU in an alternative stimulated physiological condition such as exercise (42). Because independent manipulations to each step of the MGU pathway can impair MGU, it is possible that insulin resistance created by high-fat feeding results from an impairment in any of the steps individually or in combination.

In contrast to stimulation by insulin, the marked impairment in exercise-stimulated MGU created by high-fat feeding in WT mice was improved by HK II overexpression in the gastrocnemius and SVL. A primary difference between stimulation by exercise versus insulin is that exercise results in a massive hyperemia. The result of the hyperemia is that the barrier to glucose delivery is lowered. Therefore, these results in conjunction with those from the hyperinsulinemic-euglycemic clamp experiments suggest that high-fat feeding increases the resistance to MGU at sites upstream of glucose phosphorylation (i.e., glucose delivery and membrane transport). However, when the barriers to both glucose transport and delivery are minimized during exercise as a result of hyperemia and contraction-stimulated GLUT4 translocation, an impairment in glucose phosphorylation that can be improved by HK II overexpression is exposed. In agreement with the glucose flux data, after high-fat feeding, G6P concentrations were lower in two of the muscles studied in WT compared with HK^Tg mice (see Table 4), further suggesting a functional impairment in glucose phosphorylation.

Control of metabolic pathways is more sensitively assessed using flux measurements as opposed to concentrations of metabolites alone (43). Therefore, it can be difficult to draw conclusions solely from the concentrations of glucose and G6P in tissue homogenates. Several factors complicate the interpretation of tissue glucose and G6P measurements from crude homogenates, including cellular compartmentation, bidirectionality of metabolic substrate flux, and breakdown during the tissue excision process. In addition, flux through a metabolic pathway may be increased without commensurate increases in metabolic intermediates. These factors underscore the value of in vivo flux measurements. Although the [2-³H]DG method is free from concerns associated with measuring metabolites, there are still issues that one must consider (e.g., analog discrimination, tracer recycling). Glucose and [2-³H]DG have slightly different affinities for GLUTs and HKs. A correction factor, the lumped constant, can be applied to adjust for analog discrimination. The lumped constant for 2-deoxyglucose has been determined to be ∼1.2 in human muscle in a variety of conditions (44–46) and ∼0.9 in rat muscle (47,48). Thus, in rodent muscle, R_g determined from [2-³H]DG accumulation may slightly underestimate MGU. Nevertheless, the measures of R_g provide the most sensitive means for studying the regulation of tissue-specific glucose uptake in vivo and the impact that dietary manipulations have on it.

In conclusion, high-fat feeding impairs both insulin- and exercise-stimulated MGU, but only exercise-stimulated MGU is increased by HK II overexpression. These findings indicate that 1) the impairment in MGU during insulin stimulation in high-fat-fed mice is due to functional deficits in glucose delivery and/or membrane transport and 2) deficits in glucose phosphorylation can best be exposed when the barriers to glucose delivery and transport are lowered such as they are during stimulation by exercise. Just as the regulation of MGU can be distributed among glucose delivery, transport, and phosphorylation, dysregulation can occur at multiple steps. Therefore, the most successful treatment strategies for correcting impairments in MGU will involve targeting multiple if not all of the steps of MGU.

We acknowledge the Department of Veterans Affairs and the National Institute of Diabetes and Digestive and Kidney Diseases for Grants R01 DK-54902 (to D.H.W.), U24-DK-59637 (to D.H.W.), and DK-20593 (to D.K.G. for the Vanderbilt Diabetes Research and Training Center).

We thank Wanda Snead and Greg Poffenberger of the Vanderbilt Hormone Assay Core for performing the insulin assays.