Journal of Applied Physiology
Vol. 81, No. 4, pp. 1501-1509, October 1996
EXERCISE AND MUSCLE

Wortmannin inhibits both insulin- and contraction-stimulated glucose uptake and transport in rat skeletal muscle

Jørgen F. P. Wojtaszewski, Bo F. Hansen, Birgitte Ursø, and Erik A. Richter

Copenhagen Muscle Research Centre, August Krogh Institute, University of Copenhagen, DK-2100 Copenhagen; Hagedorn Research Center, Gentofte, DK-2820 Copenhagen; and Diabetes Research, Novo Nordisk, DK-2880 Bagsvaerd, Denmark

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCE

ABSTRACT

Wojtaszewski, Jørgen F. P., Bo F. Hansen, Birgitte Ursø, and Erik A. Richter. Wortmannin inhibits both insulin- and contraction-stimulated glucose uptake and transport in rat skeletal muscle. J. Appl. Physiol. 81(4): 1501-1509, 1996.The role of phosphatidylinositol (PI) 3-kinase for insulin- and contraction-stimulated muscle glucose transport was investigated in rat skeletal muscle perfused with a cell-free perfusate. The insulin receptor substrate-1-associated PI 3-kinase activity was increased sixfold upon insulin stimulation but was unaffected by contractions. In addition, the insulin-stimulated PI 3-kinase activity and muscle glucose uptake and transport in individual muscles were dose-dependently inhibited by wortmannin with one-half maximal inhibition values of ~10 nM and total inhibition at 1 µM. This concentration of wortmannin also decreased the contraction-stimulated glucose transport and uptake by ~30-70% without confounding effects on contractility or on muscle ATP and phosphocreatine concentrations. At higher concentrations (3 and 10 µM), wortmannin completely blocked the contraction-stimulated glucose uptake but also decreased the contractility. In conclusion, inhibition of PI 3-kinase with wortmannin in skeletal muscle coincides with inhibition of insulin-stimulated glucose uptake and transport. Furthermore, in contrast to recent findings in incubated muscle, wortmannin also inhibited contraction-stimulated glucose uptake and transport. The inhibitory effect of wortmannin on contraction-stimulated glucose uptake may be independent of PI 3-kinase activity or due to inhibition of a subfraction of PI 3-kinase with low sensitivity to wortmannin.

hindlimb; signaling; phosphatidylinositol 3-kinase

INTRODUCTION

A MAJOR PHYSIOLOGICAL effect of insulin is stimulation of glucose transport into various target tissues. Increasing evidence points to phosphatidylinositol (PI) 3-kinase as having a role in the insulin signaling pathway stimulating glucose transport. Reports have described the inhibitory effect of wortmannin, a selective and irreversible PI 3-kinase inhibitor, on the insulin-stimulated glucose transport in different cell types (3, 6, 22), including recently published data obtained from incubated rat skeletal muscle (16, 19, 32). Furthermore, it has been shown that the impaired insulin-stimulated glucose transport induced by wortmannin was due to inhibition of the insulin-induced translocation of the glucose transporter proteins GLUT-1 and GLUT-4 in adipocytes (6) and GLUT-4 in incubated rat muscle (19). The exact role of PI 3-kinase and/or its products is, however, still unknown. For instance, activation of PI 3-kinase seems in itself to be insufficient for GLUT-4 translocation and stimulation of glucose transport in L6 myoblast (14). In addition, the yeast gene product Vps34, also a PI 3-kinase, is required for protein sorting to the yeast vacuoles and, based on the sequence similarity of the 110-kDa catalytic subunit and Vps34, a similar role for PI 3-kinase in mammalian cells has been proposed (27).

Contraction is the other major physiological event affecting glucose transport in muscle. The effect is mediated through translocation of GLUT-4 transporter proteins from an intracellular storage site to the plasma membrane. Although findings in mammalian skeletal muscles indicate that the increase in intracellular concentration of calcium associated with muscle contractions mediates the increase in glucose transport (7, 11, 33), the cellular signaling mechanism is basically unknown.

The additive effect of maximal insulin stimulation and muscle contractions on glucose transport (23) points to the fact that the signaling pathways for the two stimuli are different but not necessarily independent. The increased insulin sensitivity in skeletal muscle after contractions (25) and the recently reported synergistic effect of contractions and submaximal insulin stimulation on glucose uptake and transport in perfused rat muscle (31) indicate that the intracellular signaling pathways of the two stimuli might have a convergent course. Recent studies on incubated muscles have shown that the PI 3-kinase inhibitor wortmannin inhibits insulin-stimulated but not contraction-stimulated glucose transport (16, 19, 32). This effect is mediated through an inhibition of insulin-induced GLUT-4 translocation (19), and, based on these findings, PI 3-kinase has been attributed to play an essential role in insulin action but not in the signaling for contractions. However, none of these studies measured PI 3-kinase activity and none investigated the sensitivity to wortmannin in different muscle fiber types. Furthermore, the maximal wortmannin concentration used never exceeded the concentration needed for total inhibition of insulin-stimulated glucose transport. Therefore, the existence of a less wortmannin-sensitive PI 3-kinase subfraction involved in contraction-stimulated glucose transport could not be excluded.

In light of these observations, we decided to reinvestigate the role of wortmannin on insulin- and contraction-stimulated muscle glucose transport in a physiological system allowing simultaneous characterization of the metabolic response in the three different muscle fiber types. In the present study, we demonstrate that wortmannin is able to inhibit both insulin- and contraction-stimulated glucose transport in perfused rat skeletal muscle.

MATERIALS AND METHODS

Male Wistar rats (n = 92) weighing 249 ± 1 g were maintained on a constant 12:12-h light-dark cycle and received normal rat chow and water ad libitum.

The rats were anesthetized by intraperitoneal injection of pentobarbital sodium (5 mg/100 g body wt) and prepared surgically for hindquarter perfusion as described by Ruderman et al. (26). We used a cell-free perfusate with a low concentration of bovine serum albumin in an attempt to control the free concentration of wortmannin. Thus the perfusate consisted of Krebs-Henseleit buffer solution, 0.1% bovine serum albumin (fraction V, Sigma Chemical) dialyzed for 24 h against 11 vol of Krebs-Henseleit buffer solution (pore size 10-15 kDa), 8 mM glucose, 0.15 mM pyruvate, 4.2 IU/ml heparin, and 1 mM mannitol. During the hindlimb perfusion experiments, the arterial perfusate was continuously gased with a mixture of 95% O₂-5% CO₂, which, on average, yielded an arterial pH of 7.42 ± 0.01 (n = 92), CO₂ pressure of 36.0 ± 0.2 Torr (n = 92), and O₂ pressure of 541 ± 5 Torr (n = 92). The temperature of the perfusate was 35°C, which resulted in a muscle temperature in the calf muscles of ~32°C. The first 25 ml of perfusate that passed though the hindquarter were discarded, whereupon the perfusate was recirculated at a flow of 15 ml/min. Only one leg was perfused in experiments that included electrical stimulation. The leg was, from the beginning of the perfusion, immobilized by a pin under the patella tendon that was fixed onto the perfusion platform; a hook pin was also placed around the Achilles' tendon. The latter was connected to an isometric muscle tension transducer. A hook electrode was placed around the sciatic nerve and connected to an impulse generator (DISA Impulse Generator). During the first 0.5 min of contractions, the resting length of the calf muscles and the stimulation voltage were adjusted to obtain maximum active tension upon stimulation. Muscles were made to contract isometrically by stimulation of the sciatic nerve with either moderate or intense stimulation protocols. 1) Moderate stimulation: supramaximal (25 ± 2 V) trains of 100 ms were delivered every 2.5 s for 10 min (n = 20). The impulse frequency and duration within the train were 67 Hz and 1 ms, respectively. 2) Intense stimulation: supramaximal (28 ± 1 V) trains of 200 ms were delivered every 1 s for 10 min (n = 42). The impulse frequency and duration within the train were 100 Hz and 0.1 ms, respectively. During stimulation, flow was 20 ml/min and tension developed by the calf muscle was measured by the isometric muscle tension transducer and recorded by a pen writer (model 220, Clevite Brush Mark). The effects of wortmannin on different parameters were investigated during four different perfusion conditions: basal, maximal insulin stimulation, moderate electrical stimulation, and intense electrical stimulation. When wortmannin was included in the perfusate, it was added at the beginning of the perfusion. Because wortmannin was dissolved in dimethyl sulfoxide (DMSO), the perfusate of the corresponding control experiment included DMSO in the same concentration (1:5,000). In experiments where insulin (human insulin, Actrapid) was included in the perfusate, it was added so that it reached the hindlimb 10 min after wortmannin or DMSO, and the concentration was 10,000 µU/ml in all experiments. Electrical stimulation was always started 20 min after wortmannin or DMSO had reached the hindlimb. Independent of intervention, perfusate arterial and venous samples were taken every 5 min starting 10 min after the beginning of the perfusion.

For measurements of muscle membrane glucose transport, 3-O-[glucose-^l4C(U)]methyl-D-glucose (specific activity 315 mCi/mmol) and D-mannitol-[1-³H(N)] (specific activity 22.5 Ci/mmol) (New England Nuclear) were added simultaneously to the perfusate, yielding an activity of 0.050 and 0.075 µCi/ml, respectively. To avoid efflux of 3-O-[glucose-^l4C (U)]methyl-D-glucose from the intracellular space, the duration of isotope exposure was varied depending on the expected glucose transport during the different interventions. In basal conditions, the hindlimb was exposed to the isotopes for 20 min starting 20 min after the perfusion commenced. During electrical stimulation, independent of stimulation protocol, the exposure time was 7 min starting 3 min after muscle stimulation. During maximal insulin stimulation, the exposure time was 5 min starting 10 min after addition of insulin. In experiments with maximal insulin stimulation and increasing concentrations of wortmannin (1-2,000 nM), the exposure time was increased in parallel from 10 to 20 min but always beginning 10 min after addition of insulin. At the time when the isotopes reached the muscle, recirculation was stopped and one-way perfusion was started to secure a constant specific activity for glucose in the arterial perfusate during the period of glucose transport measurement. At the end of perfusion, muscle biopsies were taken from three different portions of the calf muscle: the superficial medial portion of the gastrocnemius (consisting mainly of fast-twitch white fibers), the soleus muscle (consisting mainly of slow-twitch red fibers), and the deep medial portion of the gastrocnemius muscle (consisting mainly of fast-twitch red fibers) (1). The biopsies were cut out, trimmed of connective tissue, blotted and freeze clamped with aluminum clamps cooled in liquid N₂, and stored at 80°C until analyzed. Uptake of 3-O-[glucose-^l4C(U)]methyl-D-glucose in muscles was detected in perchloric acid extracts and corrected for label in the extracellular space as determined by the ³H counts for mannitol. Radioactivity was measured in a liquid scintillation counter (model 2000 Tri-Carb, Packard Instruments). From the intracellular accumulation of 3-O-[glucose-^l4C (U)]methyl-D-glucose, the rate of glucose transport was calculated by using a "specific activity of glucose" determined by the glucose concentration and 3-O-[glucose-^l4C(U)]methyl-D-glucose counts in the perfusate.

Perfusate samples were held on ice and analyzed for glucose concentration in duplicate within 1-2 h by using a dual-channel glucose-lactate analyzer (model YSI-2700 Select, Yellow Springs Instruments). At rest and after 5 min of electrical stimulation, O₂ pressure, CO₂ pressure, and pH were determined within 20 min on perfusate samples obtained anaerobically (model ABL 510 acid-base analyzer, Radiometer). Rates of glucose and O₂ uptake of the hindquarter were calculated by multiplying the measured arteriovenous concentration differences by the flow rate. Perfusate flow rate was measured by using timed collection of the venous effluent. In the basal and insulin-stimulated state, uptake is expressed relative to perfused muscle mass. This muscle mass was considered to be 8.3 and 16.6% of the rat body weight in one- and two-leg perfusions, respectively (26). In the electrical-stimulated state, uptake was expressed relative to the stimulated muscle mass, whereas unstimulated muscle was assumed to have the same uptake as before electrical stimulation. The stimulated muscle mass was considered to be 2.73% of the rat body mass (29). Muscle temperature was measured in preliminary experiments by insertion of a thermistor (Ellab-Instrument) in the calf region. Muscle concentrations of ATP and phosphocreatine (PCr) were determined by standard enzymatic methods (18). Protein concentration was measured with "Bicinchoninic Acid Protein Assay Reagent" (Pierce Chemical).

PI 3-kinase was immunoprecipitated as previously described, with minor modifications (10). Muscle biopsies, weighing 30-40 mg, were homogenized (model OMNI 2000, OMNI International) in ice-cold solubilization buffer A (1:15 wt/vol) with an OMNI 1005 generator operating at maximum speed for two times 20 s and left to sit on ice for 1 h. The solubilization buffer A was freshly made and composed of the following mixture: 50 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES; pH 7.4), 137 mM NaCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM NaF, 10 mM Na₂P₂O₇, 2 mM EDTA, 1% Nonidet P-40, 10% glycerol, 2 mM Na₃VO₄, 34 µg/ml phenylmethylsulfonyl fluoride (PMSF), 2 µg/ml aprotinin, 5 µg/ml leupeptin, 0.5 µg/ml pepstatin, 10 µg/ml antipain, and 1.5 mg/ml benzamidine. Insoluble material was removed by centrifugation (4°C, 16.900 g) in a Heraus 1379 rotor (model 17RS Biofuge, Heraus Sepatech) for 1 h. Insulin-receptor substrate (IRS)-1 was immunoprecipitated overnight at 4°C from aliquots (150 µl) of the supernatant with anti-IRS-1 (8 µg/ml) (UBI) or preimmune rabbit immunoglobulin (8 µg/ml) (Dakopat) followed by protein A Sepharose 6 MB (4°C, 2 h) (Pharmacia). Alternatively, anti-p85 antibodies (whole serum, 6 µl/ml) (UBI) or preimmune rabbit serum (6 µl/ml) (Dakopat) was used. The immunoprecipitates were washed successively in fresh ice-cold phosphate-buffered saline (pH 7.4) containing 1% Nonidet P-40 and 100 µM Na₃VO₄ (twice), 100 mM tris(hydroxymethyl)aminomethane (pH 7.4) containing 500 mM LiCl₂ and 100 µM Na₃VO₄ (twice), and 10 mM tris(hydroxymethyl)aminomethane (pH 7.4) containing 100 mM NaCl, 1 mM EDTA, and 100 µM Na₃VO₄ (3 times). The pellet was resuspended in 30 µl of 20 mM HEPES (pH 7.4), 0.4 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N,N-tetraacetic acid, and 0.4 mM Na₂HPO₄. The aliquot was placed at 20°C until the PI 3-kinase activity was measured.

PI 3-kinase activity in immunoprecipitates was measured by in vitro phosphorylation of PI. In brief, sonicated PI (10 µl) (1 mg/ml in 5 mM HEPES) was added to each sample. The PI 3-kinase reaction (30°C) was started by addition of 10 µl of the following mixture: 50 mM MgCl₂, 250 µM ATP, 0.5 µCi/µl [-³²P]ATP (specific activity 5,000 Ci/mmol; Amersham International) in a buffer consisting of 20 mM HEPES (pH 7.4), 0.4 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N,N-tetraacetic acid, and 0.4 mM Na₂HPO₄. The reaction was stopped by addition of 15 µl of 4 N HCl; and lipids were harvested by methanol-chloroform (1:1 by vol) extraction, dried during pelleting, resuspended in 10 µl of chloroform, and applied to a silica gel thin-layer chromatography plate coated with 1% potassium oxalate (Wattman). The thin-layer chromatography plates were developed in chloroform-methanol-ammonia(25%)-water (45:35:3:7 by vol), air dried, scanned, and quantified by a phosphorimager (Molecular Dynamics).

Insulin-receptor tyrosine kinase (IRTK) activity was measured by the method described by Klein et al. (15). Supernatents from the muscle biopsies were obtained in the following manner. Muscle biopsies were homogenized with a Potter homogenizer in ice-cold solubilization buffer [10 mg muscle per 200 µl buffer: 20 mM HEPES (pH 7.4), 8 mM EDTA, 0.2 mM Na₃VO₄, 10 mM Na₄P₂O₇, 2.5 mM PMSF, 1 mg/ml aprotinin, 2.5 mg/ml benzamidine, 2.5 µg/ml pepstatin, 2.5 µg/ml leupeptin, 160 mm NaF, 2 mM dichloroacetic acid, and 1% Triton X-100]. After 20 min at 4°C, the samples were spun at 20,000 g for 60 min at 4°C (model TI 70.1, Beckman Instrument) to remove cell debris. Microtiter wells were coated with anti-insulin-receptor antibody (10 µg/ml; Ab-3, Oncogene Science) as described by Klein et al. (15), and 30 µl of supernatant were added to each well. In contrast to the methods of Klein et al., the quantification of incorporated phosphate into Poly-Glu-Tyr (4:1) was performed by scanning of filters using a phosphorimager (Molecular Dynamics).

Statistical evaluation of data from two groups was done by unpaired Student's t-test, and when more than two groups were compared, one-way analysis of variance was used. Differences between groups were considered statistically significant at P < 0.05. When one-way analysis of variance was used, the P values in the post hoc tests were corrected for multiple comparisons (Bonferonni correction). Data are presented as means ± SE.

RESULTS

In the basal non-insulin-stimulated state, wortmannin (1 µM) did not influence glucose uptake across the hindlimb (Fig. 1). Glucose transport in individual muscles did not differ significantly among the three investigated muscle groups in the basal state and were not affected by wortmannin (1 µM) (Table 1).

Table 1. Effect of wortmannin on glucose transport in basal state and during moderate electrical stimulation Glucose Transport, µmol · g1 · h1 WG RG SOL Basal Control 1.0 ± 0.2  0.4 ± 0.1  0.4 ± 0.1  Wortmannin (1 µM) 0.7 ± 0.3  0.3 ± 0.3  0.2 ± 0.1  Moderate ES Control 14.0 ± 1.6  11.1 ± 0.9  4.0 ± 0.7  Wortmannin (1 µM) 11.6 ± 1.7  8.1 ± 0.9* 1.4 ± 0.5* Values are means ± SE of 6 basal and 10 moderate electrical stimulation (ES) experiments. WG, white gastrocnemius; RG, red gastrocnemius; SOL, soleus. * Significantly different compared with control values, P < 0.05.

After maximal (10,000 µU/ml) insulin stimulation, hindlimb glucose uptake increased by 12.6 ± 0.7 (n = 3) to 15.6 ± 0.7 (n = 3) µmol ^· g^{1 ·} h¹ (Fig. 1) and glucose transport increased significantly in all three muscle groups to 15 ± 2 (n = 3), 22 ± 3 (n = 3), and 27 ± 6 (n = 3) µmol ^· g^{1 ·} h¹ in the white and red gastrocnemius and soleus muscles, respectively (Fig. 2). By adding wortmannin in increasing concentrations from 1 nM to 2 µM, we obtained a dose-dependent decrease in the insulin-stimulated hindlimb glucose uptake (Fig. 1). In addition, upon insulin stimulation, glucose transport in individual muscles also showed a wortmannin dose-dependent decrease with maximal inhibition at 1 µM and one-half maximal inhibition (IC₅₀) values at ~10 nM for all three muscle groups (Fig. 2).

On moderate electrical stimulation, glucose uptake in contracting muscles increased significantly by 15 ± 3 µmol ^· g^{1 ·} h¹ from 3.6 ± 0.4 µmol ^· g^{1 ·} h¹ (n = 10) (Fig. 3). In the presence of 1 µM wortmannin, the increase was significantly reduced to only 7.7 ± 1.9 (n = 10) µmol ^· g^{1 ·} h¹. During moderate electrical stimulation, glucose transport in individual muscles increased significantly in all three muscle groups compared with the basal condition (Table 1). Adding 1 µM of wortmannin to the perfusate, however, resulted in significantly lower transport values in red gastrocnemius and soleus but not in white gastrocnemius muscle.

Perfusion with maximal insulin stimulation (10,000 µU/ml) resulted in a significant increase in IRTK activity in all muscle groups compared with the basal state, but neither basal nor insulin-stimulated IRTK activity was inhibited by 1 µM wortmannin (Table 2).

Table 2. Effect of wortmannin on IRTK activity in basal state and during maximal insulin stimulation IRTK Activity, fmol Pi incorporated · h1 · mg muscle1 WG RG SOL Basal Control 0.49 ± 0.09  0.52 ± 0.09  0.40 ± 0.09  Wortmannin (1 µM) 0.43 ± 0.05  0.48 ± 0.09  0.36 ± 0.09  Insulin Control 4.5 ± 1.4* 3.3 ± 0.7* 3.4 ± 0.7* Wortmannin (1 µM) 3.1 ± 0.4* 4.8 ± 0.1* 3.0 ± 0.7* Values are means ± SE of 6 basal and 3 maximal insulin- stimulation (10,000 µU/ml) experiments. IRTK, insulin-receptor tyrosine kinase; WG, white gastrocnemius; RG, red gastrocnemius; SOL, soleus. * Significantly different compared with basal values, P < 0.05.

PI 3-kinase activities were estimated in anti-p85 and anti-IRS-1 immunoprecipitates from red gastrocnemius muscle only. PI 3-kinase activity in anti-IRS-1 immunoprecipitates increased significantly (6.5-fold above basal) upon insulin stimulation (Fig. 4). This increase could be inhibited in a dose-dependent manner by wortmannin (IC₅₀ of ~15 nM, total inhibition at 1 µM). There was no significant change in IRS-1-associated PI 3-kinase activity upon moderate electrical stimulation. Wortmannin (1 µM) induced a significant reduction in IRS-1-associated PI 3-kinase activity in the basal condition and during electrical stimulation. The pattern of results obtained in immunoprecipitates with anti-p85 were similar to the results mentioned above, with two important differences: 1) it was not possible to detect an effect of insulin on the p85-associated PI 3-kinase activity compared with the basal state, and 2) the absolute PI 3-kinase activity level was ~30- and 230-fold greater during interventions with and without insulin in anti-p85 compared with anti-IRS-1 immunoprecipitates.

During moderate electrical stimulation, the O₂ uptake in contracting muscle increased by approximately sevenfold compared with basal value. The O₂ uptake was not affected by 1 µM wortmannin either in the basal state or during electrical stimulation (Table 3).

Table 3. Effect of wortmannin on O2 uptake Wortmannin, µM O2 Uptake, µmol · g1 · h1 Basal Moderate ES Intense ES 0 (Control) 8.1 ± 0.6 (6) 48 ± 3* (10) 55 ± 2* (18) 1 7.2 ± 0.5 (6) 44 ± 2* (10) 49 ± 2* (15) 3 39 ± 2 (6) 10 34 ± 3 (3) Values are expressed per gram perfused muscle (basal) or per gram contracting muscle (moderate ES and intense ES). Values are means ± SE with no. of experiments in parentheses. Significantly different (P < 0.05) compared with: * basal values; control values.

The concentrations of ATP and PCr were not significantly different during basal conditions compared with in vivo. During moderate electrical stimulation, ATP concentrations remained unchanged, whereas PCr concentrations decreased in all three muscle fibers compared with basal value. Wortmannin (1 µM) did not affect ATP and PCr concentrations during basal state or electrical stimulation (Table 4).

Table 4. Effect of wortmannin and ES on muscle ATP and phosphocreatine concentrations Wortmannin, µM No. of Experiments ATP, mmol/kg dry wt Phosphocreatine, mmol/kg dry wt WG RG SOL WG RG SOL In vivo (rest) 6 27 ± 1  25 ± 1  18 ± 1  67 ± 5  56 ± 5  43 ± 4  Basal 0 6 29 ± 2  28 ± 2  25 ± 1  58 ± 5  48 ± 3  52 ± 4  1 6 30 ± 1  27 ± 2  25 ± 3  75 ± 7  63 ± 4  48 ± 5  Moderate ES 0 6 21 ± 2  28 ± 3  21 ± 1  26 ± 4* 18 ± 5* 13 ± 3* 1 6 29 ± 5  31 ± 4  23 ± 2  24 ± 6* 24 ± 7* 16 ± 5* Intense ES 0 8 16 ± 2* 18 ± 3* 15 ± 3* 23 ± 4* 22 ± 6* 18 ± 4* 1 8 15 ± 1* 22 ± 1  14 ± 1* 21 ± 2* 21 ± 4* 14 ± 3* 3 6 25 ± 3 27 ± 2  18 ± 1  32 ± 9* 29 ± 9* 11 ± 1* 10 3 29 ± 1 34 ± 4 24 ± 1  44 ± 2  65 ± 4 44 ± 7 Values are means ± SE. Significantly different (P < 0.05) compared with: * basal values without wortmannin; values of intense ES without wortmannin.

Wortmannin (1 µM) did not cause any differences in maximal or mean force development during moderate electrical stimulation (Table 5).

Table 5. Effect of wortmannin on maximal and mean force development during ES Wortmannin, µM No. of Experiments Maximal Force, N/g Mean Force, N/g Moderate ES 0 10 0.95 ± 0.02  0.71 ± 0.02  1 10 0.89 ± 0.07  0.66 ± 0.05  Intense ES 0 18 2.14 ± 0.05* 0.48 ± 0.02* 1 15 2.22 ± 0.06* 0.49 ± 0.03* 3 6 1.54 ± 0.2 0.38 ± 0.02  10 3 1.48 ± 0.08 0.14 ± 0.02 Values are expressed as force per gram contracting muscle during 10-min period of stimulation. Values are means ± SE. Significantly different (P < 0.05) compared with values from: * moderate ES; intense ES without wortmannin.

Taken together these data indicate that in perfused skeletal muscle wortmannin, in a concentration (1 µM) that fully blocks insulin-stimulated glucose transport and uptake, causes an ~30-65% reduction in contraction-stimulated glucose transport and uptake without confounding effects on contractile performance or O₂ uptake. Nevertheless, it might be argued that the applied electrical stimulation was possibly not eliciting maximal glucose transport and uptake. Therefore, it was uncertain whether wortmannin was able to impair maximal contraction-stimulated glucose transport and uptake. Furthermore, it was of interest to study whether higher concentrations of wortmannin possibly could inhibit contraction-stimulated glucose uptake to a greater extent. Therefore, we extended our experiments to include higher concentrations of wortmannin (3 and 10 µM) and, at the same time, we intensified the electrical stimulation to secure maximal stimulation by using a stimulation protocol previously shown to elicit a maximum response in glucose transport (23). The increase in glucose uptake in contracting muscles during intense electrical stimulation was significantly higher compared with the previous more moderate stimulation [24 ± 2 (n = 18) vs. 14 ± 3 (n = 10) µmol ^· g^{1 ·} h¹]. A wortmannin concentration of 1 µM significantly reduced the effect by 68% to 7.7 ± 1.3 µmol ^· g^{1 ·} h¹ (n = 15), and the effect became even more marked at 3 and 10 µM wortmannin [88% and ~100% inhibition of contraction-stimulated glucose uptake, respectively; 3 ± 2 (n = 6) and 0.1 ± 0.2 (n = 3) µmol ^· g^{1 ·} h¹]. By increasing the wortmannin concentration during intense electrical stimulation, we obtained a dose-dependent decrease in the stimulated glucose transport in all three muscle groups. The inhibitory effect at 10 µM was 68, 88, and 80% for white gastrocnemius, red gastrocnemius, and soleus, respectively (Fig. 5). During intense electrical stimulation, maximal force was higher and mean force was lower compared with moderate electrical stimulation. A wortmannin concentration of 1 µM did not affect either of these parameters during intense electrical stimulation, whereas at 3 µM the maximal force was reduced by 28% (mean force unaffected) and at 10 µM both maximal and mean force were decreased by 31 and 71%, respectively (Table 5). O₂ uptake in contracting muscle during intense electrical stimulation was unaffected by 1 µM but was decreased by 3 and 10 µM wortmannin (Table 3).

Compared with basal values, ATP and PCr concentrations decreased with intense electrical stimulation (Table 4). Neither ATP nor PCr concentrations were affected by 1 µM wortmannin. Interestingly, the decrease compared with basal values was diminished when the wortmannin concentration was increased to 3 and 10 µM. Thus the decrease in tension development and O₂ uptake during electrical stimulation with 3 and 10 µM wortmannin is not due to interference with ATP or PCr synthesis.

Independent of protocol (1- and 2-leg perfusions) and subsequent interventions, perfusion pressure was unaffected by wortmannin at all concentrations used (data not shown).

DISCUSSION

In this study, we have provided evidence that the fungal toxin wortmannin has an inhibitory effect on both the insulin- and the contraction-stimulated glucose uptake and transport in perfused rat muscles. The IC₅₀ for wortmannin inhibition of insulin-stimulated glucose transport was ~10 nM in all muscle fiber types and is comparable to IC₅₀ values reported in other systems, including incubated rat muscle (3, 16, 19, 22, 32). At 1 µM wortmannin, the insulin effect on glucose uptake and transport was totally blocked, in agreement with previous studies (16, 19, 32). Interestingly, in contrast to these previous studies in incubated muscle, the present study shows that wortmannin at 1 µM inhibits contraction-stimulated glucose transport and uptake by ~30-70% without any confounding effects on muscle contractility or O₂ uptake.

The effect of wortmannin on insulin-stimulated glucose transport was not due to inhibition of IRTK (Table 2), interferences with insulin binding to the insulin receptor (22), or tyrosine phosphorylation of the -subunit of the insulin receptor and IRS-1 (22) but is most likely due to inhibition of PI 3-kinase (Fig. 4).

The sensitivity of both insulin- and contraction-stimulated glucose transport to inhibition by wortmannin does not necessarily imply that the effect on both processes is due to inhibition of PI 3-kinase. In fact, our data could be interpreted to suggest that PI 3-kinase may not be involved in the contraction-stimulated glucose transport. This is suggested by the fact that PI 3-kinase activity precipitated either with anti-IRS-1 antibody or anti-p85 antibody was not increased by muscle contractions (Fig. 4). Furthermore, PI 3-kinase activity precipitated with either antibody was maximally inhibited by 1 µM of wortmannin, which caused only a partial reduction in glucose uptake and transport when electrical stimulation was applied. On the other hand, it cannot be excluded that a small subfraction (isoform) of PI 3-kinase might be activated by contractions and might be less sensitive to wortmannin. Such a subfraction might escape detection because we cannot immunoprecipitate it with a specific antibody and because it might only constitute a few percent of the total PI 3-kinase pool (in analogy with IRS-1-precipitated PI 3-kinase activity; see below). The unchanged anti-p85 immunoprecipitated PI-3 kinase activity upon maximal insulin stimulation suggests that most of the total (anti--p85 and anti--p85 immunoprecipitated) PI 3-kinase activity is in fact not stimulated by insulin or that stimulation such as direct binding to the insulin receptor (17, 30) might not be detectable with this assay. On the other hand, the anti-IRS-1 immunoprecipitated PI 3-kinase showed a 6.5-fold increase with maximal insulin (Fig. 4). In absolute terms, the increase in PI 3-kinase activity associated with IRS-1 would only cause a ~3% increase in p85-associated PI 3-kinase activity that is not detectable with the assay.

There is no obvious explanation for the discrepancy between the former studies demonstrating no effect of 1 µM wortmannin on contraction-stimulated glucose transport (16, 19, 32) and the present study in which 1 µM wortmannin caused a ~30-70% reduction in contraction-stimulated glucose transport and uptake, except for the model used. It could be speculated that the supply of wortmannin was superior in our model due to the use of an intact circulation. However, the similarity in the wortmannin dose-dependency of insulin-stimulated glucose transport between the present study and previous studies indicates that this is not a likely explanation. Furthermore, because the effect of wortmannin on both insulin- and contraction-stimulated glucose transport was nearly similar in the three different muscle groups (Fig. 2 and Table 1), fiber type differences do not offer an explanation either.

To describe the dose dependency of wortmannin on the maximal contraction-stimulated glucose uptake and transport, we investigated the effects of even higher wortmannin concentrations (3 and 10 µM). The effects on both parameters became more marked. Thus at 10 µM the inhibitory effect was nearly total. At these concentrations, wortmannin reduced the contractility and O₂ uptake significantly but not totally (Tables 3 and 5). However, evidence from amphibian muscles points to the fact that the contraction-stimulated glucose transport is independent of the amount of work performed but rather is dependent on the frequency of stimulation (12). Furthermore, because in incubated epitrochlearis muscles it has been reported that wortmannin decreases muscle performance but not glucose transport (32), the decrease in muscle performance in the present study may not be an important factor in decreasing glucose uptake and transport. We cannot, however, exclude the possibility that part of the effect on glucose uptake and transport of wortmannin at the higher concentrations is secondary to the decrease in muscle performance. Interestingly, as performance declined, muscle ATP and PCr concentrations in contracting muscle increased toward resting levels, indicating that decreased availability of ATP and PCr is not the reason for the decreased contractile performance at 3 and 10 µM wortmannin.

The selectivity of wortmannin has been questioned even at concentrations <1 µM because myosin light chain kinase (21), PI 4-kinase (20), and phospholipase A₂ (8) have been reported to be inhibited by wortmannin. To our knowledge, none of these are thought to be involved in insulin-stimulated glucose transport in muscle. However, Nakanishi et al. (21) did also find an inhibitory effect of wortmannin on the calcium-dependent protein kinase (PKC) activity in smooth muscles (<10% and ~50% at 1 and 10 µM, respectively), and there is some evidence supporting a role for PKC activity in contraction-induced stimulation of muscle glucose transport. Thus, when PKC was downregulated by prolonged preincubation with phorbol esters or inhibited with polymyxin B, contraction-stimulated glucose transport was impaired (7, 11, 34). Inhibition of PKC might be a possible explanation for the observed effect on the contraction-stimulated glucose transport at higher concentrations of wortmannin.

In the present study, we used cell-free perfusate with a low albumin concentration to avoid binding of wortmannin to erythrocytes and albumin (9). The use of cell-free perfusate has been reported previously and the viability of the preparation has been evaluated and was found to be satisfactory during resting conditions (4), and, even during electrical stimulation, the preparation has been found to function well when an appropriate increase in flow is applied (28). In our model, the O₂ uptake (32°C) during resting conditions was similar to previously reported values in perfused muscle without erythrocytes and in the lower range of reported values with erythrocytes (4) (temperature coefficient of 2.5 was applied to compare estimates at 37°C). Furthermore, in resting muscle, ATP and PCr concentrations remained unchanged during 45 min of perfusion compared with values obtained in muscles from anesthetized rats (Table 4). The approximately six- to sevenfold increase in O₂ uptake in contracting muscle is comparable with the findings of Shiota and Sugano (28). Nevertheless, in the present study the increase in O₂ uptake during contractions is two to four times less than that found during contractions when perfused with erythrocytes (13, 31). Thus oxygenation was presumably not optimal in the present study. However, the decrease in PCr concentrations during moderate stimulation and the additional decrease in ATP concentrations during intense electrical stimulation are in agreement with findings during moderate and strenuous exercise in vivo (2, 5). Furthermore, the force development during both moderate and intense electrical stimulation is comparable with previous findings in perfusions with erythrocytes (24, 31). Thus our findings support the notion that for selected studies cell-free perfusate is a suitable alternative even during contractions.

In conclusion, three main findings were obtained in the present study. 1) In contrast to recent studies in incubated muscles, we find that wortmannin at 1 µM inhibits contraction-stimulated glucose uptake and transport in perfused skeletal muscle without confounding effects on muscle performance. At higher concentrations (3 and 10 µM), inhibition is almost total, but then contractility is decreased, albeit not abolished. The effect of wortmannin could be through a PI 3-kinase-independent mechanism, but it cannot be excluded that the effect is mediated through a subfraction of PI 3-kinase that is less sensitive to inhibition by wortmannin. 2) Neither IRS-1- nor p85-associated PI 3-kinase is activated by contractions. 3) PI 3-kinase seems to have an essential role in insulin stimulation of glucose transport in all three skeletal muscle fiber types.

ACKNOWLEDGEMENTS

We thank H. H. Klein for the gift of anti-insulin-receptor antibody-coated microwells.

FOOTNOTES

Address for reprint requests: J. Wojtaszewski, Copenhagen Muscle Research Centre, August Krogh Institute, Univ. of Copenhagen, 13 Universitetsparken, DK-2100 Copenhagen, Denmark.

Received 22 February 1996; accepted in final form 13 May 1996.

REFERENCE