Skeletal Muscle of Stroke-Prone Spontaneously Hypertensive Rats Exhibits Reduced Insulin-Stimulated Glucose Transport and Elevated Levels of Caveolin and Flotillin

5-Aminoimidazole-4-carboxamide ribonucleoside (AICAR), cytochalasin B, Pederson Fetuin, collagenase type 1A, and bovine serum albumin fraction V were supplied by Sigma (Poole, U.K.). Hams-F10, α-minimal essential medium (MEM), and fetal bovine serum (FBS) were obtained from Gibco/Life Technologies (Paisley, U.K.). Dr. G. Danielson (Novo Nordisk, Bagsvaerd, Denmark) provided porcine insulin. AMARA peptide (AMARAASAAALARRR) was a gift from Professor D. G. Hardie, University of Dundee, U.K. PKB substrate peptide (RPRAATF) was from Dr. R. Plevin, University of Strathclyde, U.K. All other reagents were as previously described (20,24,25).

Professor K. Siddle (University of Cambridge) and Dr. E.M. Gibbs (Pfizer, Groton, CT) supplied anti–IRS-1 antibodies. Anti–IRS-2, anti-PKB (PH domain), and anti-p85α antibodies were from Upstate Biotechnologies (TCS Biologicals, Bucks, U.K.). Anti-syntaxin 4 was from Professor D.E. James (University of Queensland) (26), and anti-AMPKα1 and α2 were provided by Prof. D.G. Hardie (University of Dundee) and are described by Woods et al. (27). Anti–endothelial nitric oxide synthase (eNOS) antibodies were from Sigma. Anti–caveolin-1 antibodies were purchased from Becton Dickinson (Oxford, UK). Anti–VAMP-2 antibody was purchased from Synaptic Systems (Gottingen, Germany). Anti-GLUT4 antibody was used as previously described (24). Anti-flotillin antibodies were from Transduction Laboratories (San Diego, CA).

Male Wistar-Kyoto (WKY) rats and SHRSP rats were reared in house as previously described (4,28). Rats were housed under controlled conditions of temperature (21°C) and light (12-h light-dark cycle) and fed on normal rat diet (rat and mouse no. 1 maintenance diet; Special Diet Services, Edinburgh, U.K.) and water ad libitum. Blood pressure was measured as outlined previously (4,28). Details of the animals are presented in Table 1. Note that differences in body weight in age-matched animals are consistent with previous studies (4,28). Although the SHRSP rats had a slightly smaller body weight than the WKY rats, their muscle weight was slightly increased (Table 1).

Primary skeletal muscle cultures were grown from satellite cells isolated from dissociated muscle tissue by methods adapted from Sarabia et al. (29). Briefly, male rats were killed by CO₂ overdose, and the extensor digitorum longus (EDL) muscle was dissected and cleaned. The muscles from one rat were washed twice in sterile phosphate-buffered saline (room temperature), minced, and enzymatically dissociated by three rounds of mixing at 37°C with 10 ml of 300 units/ml collagenase (type 1A), made up in 0.25% wt/vol trypsin. The supernatant was removed, neutralized with 10% FBS, and centrifuged at 1,000g for 5 min. The pellet was resuspended in 4 ml of growth medium (Hams-F10, 20% FBS, 1% penicillin/streptomycin, 1% l-glutamine, 500 mg/l Pederson Fetuin, 500 mg/l BSA, 1 μmol/l dexamethasone, and 1 nmol/l insulin). Resuspended cells were preplated on a noncoated tissue culture flask at 37°C in a 5% CO₂ humidified atmosphere for 2 h, and then the supernatant was transferred to collagen-coated flasks. Cells were fed every third day with growth medium and were passaged once they were 50% confluent. Cells were plated onto collagen-coated plasticware and differentiated when 70–80% confluent by replacing growth medium with differentiation medium (α-MEM, 2% FBS, 1% l-glutamine, 1% penicillin/streptomycin, and 1 nmol/l insulin). Cells were used between 8 and 10 days postdifferentiation.

Serum-starved myoblasts were washed two times in 3 ml of Krebs-Ringer HEPES buffer (118 mmol/l NaCl, 5 mmol/l NaHCO₃, 4.7 mmol/l KCl, 1.2 mmol/l KH₂PO₄, 1.2 mmol/l MgSO₄.7H₂O, 25 mmol/l HEPES, 2.5 mmol/l CaCl₂, pH 7.4) and incubated for 30 min in the presence or absence of 1 nmol/l to 1 μmol/l insulin or 500 μmol/l AICAR (in the presence or absence of 100 nmol/l wortmannin, as indicated in the figure legends). Uptake was initiated by addition of 10 μmol/l 2-deoxy-d-glucose (deGlc) (1 μCi/ml) for 10 min. Specific carrier-mediated deGlc uptake was determined by calculating the cell-associated radioactivity in the presence of 10 μmol/l cytochalasin B. Uptake was stopped by washing in ice-cold phosphate-buffered saline. Cells were lysed in 1 ml of 1% Triton-X100. DeGlc uptake was linear for a period of >30 min under these conditions (data not shown).

Flexor digitorum brevis (FDB) muscles were dissected and cleaned of connective tissue without stretching and with tendons intact. Muscles were placed in Krebs-Hensliet buffer (118 mmol/l NaCl, 4.7 mmol/l KCl, 1.2 mmol/l K₂HPO₄, 1.2 mmol/l MgSO₄.7H₂O, 25 mmol/l NaHCO₃, 2.5 mmol/l CaCl₂, pH 7.4, supplemented with 25 mmol/l d-glucose and 0.1% [wt/vol] BSA) and were allowed to recover for 30 min at 37°C with continuous gassing with 95% O₂/5% CO₂. Muscles were pinned at resting length and incubated for 30 min in the presence or absence of 1 μmol/l insulin or 500 μmol/l AICAR; subsequently, the muscles were rapidly washed three times with glucose-free Krebs-Hensliet buffer containing 1% BSA and were incubated for 10 min at 37°C in uptake buffer (Krebs-Hensliet buffer) containing 1 μmol/l insulin, 10 μmol/l deGlc (1 μCi/ml), and [¹⁴C]sucrose (0.2 μCi/ml), which was used as an extracellular marker (in the presence or absence of 100 nmol/l wortmannin, as indicated in the figure legends). Muscles were washed three times with ice-cold saline blotted on filter paper and then weighed and solubilized in 1 ml of 0.5N NaOH at 60°C for 45 min, and tissue-associated radioactivity was determined. In control studies, we have found that the uptake of deGlc is linear for at least 20 min under these conditions in both SHRSP and WKY FDB muscle (data not shown).

Freshly dissected muscles were homogenized using a Polytron in ice-cold lysis buffer (50 mmol/l HEPES pH7.4, 1% NP40, 150 mmol/l NaCl, 5 mmol/l Na₄P₂O₇, 5 mmol/l Na₃VO₄, 1 mmol/l EDTA, 1 mmol/l EGTA, 1 mmol/l diothiothreitol (DTT), 10% glycerol, and protease inhibitors), followed by 20 strokes in a Dounce homogenizer. The homogenate was centrifuged at 14,000g for 5 min at 4°C to remove unsolubilized material. Aliquots of the solubilized protein were analyzed by immunoblotting as described below. Note that this procedure is sufficient to solubilize >95% of the cellular caveolin and flotillin (data not shown).

Muscles were isolated and snap frozen. Frozen tissue was powdered on dry ice and resuspended in HES buffer (20 mmol/l HEPES, pH 7.4, 1 mmol/l EDTA, 250 mmol/l sucrose, 2 mmol/l sodium orthovanadate, 10 mmol/l NaF, and 1 mmol/l sodium pyrophosphate), supplemented with proteinase inhibitor tablets (Boehringer Mannheim), before homogenization. The homogenate was centrifuged at 12,000g for 20 min at 4°C and the pellet was collected (heavy membrane fraction). The supernatant was further centrifuged at 140,000g for 1 h at 4°C, and the pellet (light membrane fraction) and supernatant (soluble protein fraction) were collected. These fractions were assayed for protein and resuspended in SDS-PAGE sample buffer before immunoblot analysis.

Muscles were isolated by dissection, placed in Krebs-Hensliet buffer supplemented with 25 mmol/l d-glucose and 0.1% (wt/vol) BSA, and allowed to recover for 30 min at 37°C with continuous gassing with 95% O₂/5% CO₂. Muscles were incubated for 30 min in the presence or absence of 1 μmol/l insulin, after which the muscles were rapidly washed three times with glucose-free Krebs-Hensliet buffer and snap frozen. Frozen tissue was powdered using a mortar and pestle on dry ice and resuspended in lysis buffer (50 mmol/l Tris-HCl pH 7.4, 50 mmol/l NaF, 1 mmol/l sodium pyrophosphate, 1 mmol/l EDTA, 1 mmol/l EGTA, 50 mmol/l mannitol, 1 mmol/l DTT, 0.1 mmol/l benzamidine, 0.1 mmol/l phenylmethylsulfonyl fluoride, 1 mmol/l Na₃VO₄, 5 μg/ml soybean trypsin inhibitor, and 1% Triton X-100). Muscle lysates were spun in a microfuge at 14,000 rpm for 10 min to remove insoluble material. PKB immunoprecipitation and assay were carried out using triplicate samples (200 μg) of soluble muscle lysate, exactly as described by Salt et al. (25).

Protein samples were separated on SDS-PAGE and transferred to nitrocellulose, as outlined by Maier et al. (20). Immuno-labeled proteins were visualized using horseradish-peroxidase–conjugated secondary antibody and the enhanced chemiluminescence system (Amersham, Amersham, U.K.). Bands were quantified by densitometry using a BioRad GS700 system. To quantify the relative levels of expression, increasing loads of protein were loaded into adjacent lanes, and the linearity of the immunoblot signal was determined by densitometric analysis of blots developed using enhanced chemiluminescence. Multiple exposures of X-ray film were performed to insure linearity of response to film to signal. All immunoblot signals were quantified from linear regions of the protein titration curve, as outlined by Maier et al. (20).

Statistical analysis was performed using Student’s t test with StatView software (Abacus, Stanford, CA). P ≤ 0.05 was taken as a significant difference between groups.

The characteristics of the animals used in this study are shown in Table 1. Blood pressure was higher in the SHRSP compared with the WKY animals. We have previously shown that adipocytes from these animals exhibit reduced insulin-stimulated glucose transport in isolated adipocytes and a blunted ability of insulin to suppress isoproternol-stimulated lipolysis compared with the isogenic WKY strain (4). We therefore examined the ability of insulin to stimulate deGlc uptake in isolated FDB muscle from these animals. Figure 1 compares the response to insulin and AICAR, an activator of AMPK, on deGlc transport. As shown, muscles from SHRSP exhibit decreased insulin-stimulated deGlc uptake compared with WKY muscles, but the response to AICAR is essentially identical between the two groups. Consistent with other studies, wortmannin (100 nmol/l) completely blocked insulin-stimulated deGlc transport but was without effect on AICAR-stimulated glucose transport (data not shown).

The stimulation of glucose transport in skeletal muscle is mediated by the GLUT4 glucose transporter (10). We therefore determined the total cellular levels of GLUT4 in skeletal muscle membranes from SHRSP and WKY animals by quantitative immunoblotting. Figure 2 demonstrates that GLUT4 levels are similar in different muscles from SHRSP and WKY animals. These data indicate that reduced GLUT4 expression cannot explain the reduced ability of insulin to stimulate deGlc uptake in SHRSP muscle.

In an attempt to identify differences between the SHRSP and WKY muscle that could account for the observed insulin resistance, we have characterized the cellular content and distributions of proteins involved in insulin signaling and GLUT4 trafficking in muscles of these animals. Figure 3 shows an immunoblot analysis of the levels and distribution of IRS-1, IRS-2, and the p85α subunit of PI3K in muscle fractions. We have compared the levels of these proteins in three crude subcellular fractions of muscle, specifically a heavy membrane fraction, a light membrane fraction (enriched in intracellular membranes, as evidenced by the majority of cellular GLUT4 being present in this fraction), and a soluble protein fraction (Fig. 3). As has been reported in adipocytes (30,31,32), these key insulin signaling molecules are localized predominantly in the light membrane and soluble protein fractions of muscle. Neither the absolute level nor the distribution among these fractions was significantly different between SHRSP and WKY animals (Figs. 3 and 5). Hence, it would seem unlikely that defective expression of IRS-1, IRS-2, or PI3K can explain the insulin resistance observed in SHRSP muscles.

To further evaluate possible signaling defects associated with insulin action in these muscles, we assayed the ability of insulin to stimulate PKB activity in these muscles because PKB has been suggested to be involved in the insulin-stimulated translocation of GLUT4 to the plasma membrane of both adipocytes and muscle (33,34,35,36,37). The result of this analysis is shown in Fig. 4. No differences in the basal or insulin-stimulated PKB activity were observed in soleus or EDL muscles isolated from SHRSP and WKY animals.

Figure 5 shows an immunoblot analysis in which we compared the total cellular content of VAMP2, VAMP3/cellubrevin, and syntaxin 4, which are the SNARE proteins that mediate GLUT4 translocation (38) in total muscle extracts. Quantification of these data revealed that VAMP2 expression was elevated 1.9 ± 0.2–fold in EDL and 1.7 ± 0.3–fold in tibialis muscles from the insulin-resistant SHRSP strain compared with WKY animals (P = 0.02 for both). VAMP3/cellubrevin levels were also raised in both muscles (53% in EDL and 1.5 ± 0.13–fold in tibialis), but this increase only reached statistical significance in the tibialis muscle (P < 0.05). Syntaxin 4 levels were also increased in EDL (1.08 ± 0.2–fold, P = 0.03) and tibialis (1.8 ± 0.05–fold, P = 0.02) muscles from SHRSP animals. By contrast, cellular levels of AMPKα2, PKB/cAKT, and eNOS did not differ significantly between strains in either muscle (Fig. 4). AMPKα1 was barely detectable in these tissues, consistent with published results (data not shown) (39).

Recent studies have implicated caveolae and/or lipid rafts as possible components of the insulin signaling apparatus. We therefore examined the level of expression of caveolin isoforms in muscle tissue from SHRSP and WKY animals, as well as the expression of flotillin, a protein known to be enriched in membrane rafts and one that has been implicated as a component of the insulin signaling cascade. The results are shown in Fig. 6. Levels of expression of caveolin-1 and flotillin were found to be markedly elevated in SHRSP compared with controls. Caveolin expression was increased in EDL muscle by 2.4 ± 0.4–fold (P = 0.005) and in tibialis by 3.34 ± 0.4–fold (P = 0.002). Similarly, flotillin levels were also increased by 1.9 ± 0.05–fold and 1.7 ± 0.12–fold in EDL and tibialis muscles, respectively (P < 0.05 for both).

It is possible that the insulin resistance observed in SHRSP may be secondary to changes in physiology, such as altered blood flow to the muscle beds by defective vasodilatation in response to insulin or by changes in circulating factors that regulate peripheral insulin sensitivity. To address this, we isolated myoblast cultures from muscle satellite cells isolated from SHRSP or WKY animals. These multinucleated cells exhibited high levels of expression of the muscle-specific markers MyoD, myogenin desmin, and sarcomeric-α-actin (data not shown). We therefore examined the ability of insulin and AICAR to stimulate deGlc uptake in these cells. Figure 7A shows a typical dose response of deGlc uptake to insulin. We routinely observed both a decreased maximal rate of insulin-stimulated deGlc uptake and a rightward shift of the dose-response curve in myoblasts from SHRSP compared with WKY animals. Figure 7B demonstrates that myoblast cell cultures from SHRSP muscle exhibit blunted insulin-stimulated deGlc uptake compared with WKY, but the response to AICAR was similar between the two cultures. Consistent with data on intact muscle, pretreatment with 100 nmol/l wortmannin completely blocked insulin-stimulated deGlc uptake but was without effect on AICAR-stimulated deGlc uptake.

Glucose transport in skeletal muscle is stimulated by both insulin and exercise. Insulin-stimulated glucose transport is reduced in type 2 diabetes and insulin resistance, but several studies have suggested that exercise-stimulated glucose transport is unaffected in such individuals (17,18,19,40). Understanding the molecular basis for defective insulin action is of fundamental importance, given the greatly increased incidence of cardiovascular disease in individuals with type 2 diabetes (41,42). In this study, we have examined the ability of insulin and AICAR (an activator of AMPK) to modulate glucose transport in skeletal muscle from SHRSP compared with the normotensive, insulin-sensitive WKY strain. We have shown that muscle from SHRSP exhibits blunted insulin-stimulated glucose transport (Fig. 1), and these data are in agreement with our previous observation that adipocytes from these animals also exhibit defective insulin action (4).

AMPK is activated in response to exercise in both human and rodent muscle (8,9). AMPK has therefore been suggested to play a crucial role in exercise-stimulated glucose transport, based on studies showing that treatment of muscle with AICAR can activate glucose transport by a mechanism similar to that induced by exercise (8,9). Strikingly, AICAR is able to activate glucose transport to a similar extent in muscle from both SHRSP and WKY strains (Fig. 1), suggesting that the insulin resistance observed in the SHRSP is confined to an insulin-specific step in the activation of glucose transport. This step is not due to aberrant expression of the major insulin-responsive glucose transporter, GLUT4, because levels of expression of this protein were found to be similar in all muscles examined (EDL, tibialis, and FDB) (Fig. 2).

To further address the mechanism of the observed insulin resistance, we isolated myoblast cultures from SHRSP and WKY animals. These cultures exhibited high levels of expression of a range of skeletal muscle markers and expressed GLUT4 (data not shown). Like intact muscle, these cultures exhibit insulin- and AICAR-stimulated glucose transport (Fig. 7). Remarkably, the defect observed in insulin-stimulated glucose transport in intact muscles from the SHRSP was also observed in myoblast cultures (Fig. 7B). We have found that this diminution of insulin sensitivity was evident throughout all passages of the cells (we have examined up to passage 7). These data strongly argue that the defect observed in isolated muscle strips is an inherent genetic defect and does not reflect an adaptive response of the muscle to altered blood flow or circulating factors.

In an attempt to identify possible sites of lesion in the insulin signaling cascade, we have examined the levels of expression and subcellular localization and activity of several key proteins implicated in insulin-stimulated glucose transport, specifically IRS-1, IRS-2, and PI3K (Figs. 3,4,5) (43,44). The subcellular distribution of IRS-1 and IRS-2 is modulated by insulin (44). The association of IRS-1 with a cytoskeletal ‘scaffold’ adjacent to the plasma membrane has been suggested to be a fundamental aspect of insulin signaling in adipocytes (30). In response to insulin treatment, IRS-1 is released from this scaffold and appears to behave as a soluble protein, moving from the particulate to the cytosolic fraction upon subcellular analysis (30). In an effort to address whether IRS proteins were cytosolic or particulate in muscle, we subjected muscle homogenates to subcellular fractionation. Analysis of the subcellular distribution of IRS-1 and IRS-2 revealed that these molecules were associated with both heavy and light membrane fractions, with some immunoreactivity evident in the cytosolic fraction (Fig. 3). However, comparison of these fractions from SHRSP and WKY strains revealed no differences in either total level of expression or subcellular distribution (Fig. 3). Hence, it would appear unlikely that the primary defect in insulin action in the SHRSP lies at the level of expression or localization of these proteins. Similarly, levels of expression of AMPKα2 (the major isoform in skeletal muscle) (39) and both α- and β-isoforms of PKB/cAKT and eNOS were similar between the two groups (Fig. 5).

In an effort to address whether the extent of activation of downstream signaling molecules differed between the two strains, we assayed PKB/cAKT activity in both soleus and EDL muscles from SHRSP and WKY animals. As shown in Fig. 4, no difference in either basal or insulin-stimulated PKB activity was observed between these strains. These data suggest that the defect in insulin action observed in the insulin-resistant SHRSP is not a consequence of defective activation of this important kinase.

We next turned our attention to other molecules involved in the translocation of GLUT4, particularly the SNARE proteins, in which some notable differences were observed between the two strains (Fig. 5). Firstly, we observed elevated expression of VAMP2 and syntaxin 4 in the insulin-resistant SHRSP muscle compared with control animals. Insulin-stimulated GLUT4 translocation represents a highly regulated vectorial delivery of defined intracellular cargo to the plasma membrane. The fidelity of this response is mediated at least in part by SNARE proteins (38,45). Studies have identified homologs of SNAREs in adipocytes and muscle, in particular syntaxin 4 and VAMP2. GST-fusion proteins for each of these SNARE proteins inhibit insulin-stimulated GLUT4 translocation in muscle and fat cells, arguing cogently for a crucial functional role of these proteins in insulin action (24,26,46,47,48,49,50,51). We have shown that VAMP2 and syntaxin 4 SNARE protein expression is significantly upregulated in the insulin-resistant Zucker diabetic fatty rat (20). Similar data were observed in the present study; VAMP2 expression was doubled in both EDL and tibialis muscle from SHRSP compared with WKY (P < 0.01), and syntaxin 4 levels were also increased. Such data suggest that upregulation of SNARE proteins may be evident in a range of insulin-resistant models. However, whether these changes are causal or adaptive remains to be determined. Interestingly, elevated levels of both VAMP2 and syntaxin 4 were also observed in isolated myoblast cultures from SHRSP compared with WKY animals (data not shown). Such observations could suggest that the overexpression of SNARE proteins may be more than simply an adaptive change to whole-animal insulin sensitivity. However, further work will be required to address this issue.

One other major difference between these two strains was identified in this study. Specifically, the expression of caveolin and flotillin were markedly elevated in muscle from the insulin-resistant SHRSP compared with WKY animals (Fig. 6). Caveolin is the major structural protein of caveolae, which are flask-like invaginations of the cell surface (52). Caveolae have been implicated as components of the spatial organization of intracellular signaling cascades, and several studies have reported localization of insulin receptors to these structures (21,22,23,53). Caveolin expression is induced upon differentiation of adipocytes (54), and insulin stimulates tyrosine phosphorylation of caveolin in fat cells (23). More recently, the localization of insulin-regulated CAP proteins to caveolae has argued strongly for an important (but as yet undefined) role for caveolae in insulin signal transduction (21,55). The overexpression of caveolin observed in skeletal muscle of the SHRSP strain may therefore be an important contributory factor to the insulin resistance observed in these animals. Whether increased expression of caveolin could result in the sequestration of important insulin signaling components in an inactive environment or act as an inhibitor of downstream effects remains to be determined. Strikingly, the lipid raft–associated protein, flotillin, is also markedly overexpressed in SHRSP muscles compared with WKY. Flotillin has been proposed to localize a protein complex of CAP-Cbl proteins to lipid rafts in adipocytes that may play a crucial role in propagation of the insulin signal, at least in adipocytes (21). The fact that both caveolin and flotillin are overexpressed in muscle from the SHRSP suggests that overexpression of caveolae-associated proteins may act to blunt the ability of insulin receptors to signal to proximal molecules in the insulin signaling cascade. Further studies will be required to test this hypothesis in a systematic fashion. It is interesting to note that caveolin and flotillin were overexpressed in myoblast cultures from SHRSP animals compared with WKY cultures by similar magnitudes (data not shown).

In summary, we have shown that skeletal muscle of SHRSP exhibits reduced insulin-stimulated glucose transport compared with WKY isogenic controls. However, the response to activation of AICAR was indistinguishable between the two strains, suggesting that the defect in insulin action lies on a specific arm of the machinery or signaling cascade activated by insulin. We further show that SHRSP, like the insulin-resistant Zucker diabetic fatty rat, exhibits elevated VAMP2 and syntaxin 4 expression in skeletal muscle, suggesting that this defect may be common to a range of rodent models of insulin resistance. Finally, we report for the first time that caveolin and flotillin are overexpressed in the insulin-resistant muscles. This result may be of considerable importance, given the proposed central role of caveolae in the organization of the insulin signaling system.

This work was supported by the Medical Research Council (MRC) (Cooperative Group Grant G9800025 and Component Grant G9901492 to J.M.C.C., A.F.D., and G.W.G.), the Wellcome Trust (equipment grant to G.W.G.), the British Heart Foundation (project grant to G.W.G. and A.F.D., and a fellowship to I.P.S.), and the University of Glasgow (J.M.C.C. and G.W.G.). D.J.J. is supported by a Collaborative Awards in Science and Engineering PhD studentship with Glaxo SmithKline. The authors are grateful to S.J. Yeaman and D. Turnbull (University of Newcastle) for help in establishing myoblast cultures and to all colleagues who generously supplied reagents and advice during this study.