Endocrinology and Metabolism

Effect of epinephrine on glucose disposal during exercise in humans: role of muscle glycogen

Matthew J. Watt, Mark Hargreaves


This study examined the effect of epinephrine on glucose disposal during moderate exercise when glycogenolytic flux was limited by low preexercise skeletal muscle glycogen availability. Six male subjects cycled for 40 min at 59 ± 1% peak pulmonary O2 uptake on two occasions, either without (CON) or with (EPI) epinephrine infusion starting after 20 min of exercise. On the day before each experimental trial, subjects completed fatiguing exercise and then maintained a low carbohydrate diet to lower muscle glycogen. Muscle samples were obtained after 20 and 40 min of exercise, and glucose kinetics were measured using [6,6-2H]glucose. Exercise increased plasma epinephrine above resting concentrations in both trials, and plasma epinephrine was higher (P < 0.05) during the final 20 min in EPI compared with CON. Muscle glycogen levels were low after 20 min of exercise (CON, 117 ± 25; EPI, 122 ± 20 mmol/kg dry matter), and net muscle glycogen breakdown and muscle glucose 6-phosphate levels during the subsequent 20 min of exercise were unaffected by epinephrine infusion. Plasma glucose increased with epinephrine infusion (i.e., 20–40 min), and this was due to a decrease in glucose disposal (Rd) (40 min: CON, 33.8 ± 3; EPI, 20.9 ± 4.9 μmol · kg−1 · min−1,P < 0.05), because the exercise-induced rise in glucose rate of appearance was similar in the trials. These results show that glucose Rd during exercise is reduced by elevated plasma epinephrine, even when muscle glycogen availability and utilization are low. This suggests that the effect of epinephrine does not appear to be mediated by increased glucose 6-phosphate, secondary to enhanced muscle glycogenolysis, but may be linked to a direct effect of epinephrine on sarcolemmal glucose transport.

  • glucose transport
  • exercise

although previous studies have demonstrated a reduction in glucose disposal during exercise when plasma epinephrine levels are elevated (12, 14,31), the mechanism or mechanisms underlying the decreased glucose disposal have yet to be fully elucidated. The most widely held view is that glucose disposal is reduced because of inhibition of glucose phosphorylation by elevated glucose 6-phosphate (G-6-P), secondary to greater flux through glycogenolysis. Indeed, we recently demonstrated that decreased glucose uptake during moderate-intensity cycle exercise with epinephrine infusion was associated with increased skeletal muscle glycogenolysis and G-6-P accumulation (31).

Another possibility is that epinephrine directly affects GLUT4-mediated glucose transport across the sarcolemma, although this is equivocal. GLUT4 is phosphorylated via β-adrenergic pathways (13), and this may inhibit GLUT4 transporter activity, as demonstrated in rat adipocytes (16) and adipocyte plasma membrane vesicles (24). Bonen et al. (2) reported GLUT4 translocation to the sarcolemma but reduced glucose transport, implying decreased GLUT4 intrinsic activity, in rat skeletal muscle with epinephrine administration. Although β-adrenergic receptor stimulation results in phosphorylation of GLUT4 in rat skeletal muscle (19), no inhibitory effect of epinephrine on insulin (1, 19)- and contraction (1)-stimulated 3-O-methyl-d-glucose transport has been reported.

In the present study, we attempted to examine the effect of elevated plasma epinephrine on glucose disposal during exercise with reduced muscle glycogen availability. Under such conditions, exercise results in less muscle glycogenolysis (9) and G-6-Paccumulation (29). Thus we hypothesized that if the epinephrine-induced reduction in glucose disposal was primarily mediated via inhibition of glucose phosphorylation by G-6-P, secondary to increased glycogenolysis, elevated epinephrine would have no, or a reduced, effect on glucose disposal in the glycogen-depleted state.



Six recreationally active males (23 ± 1 yr; 74 ± 3 kg) volunteered to participate in this study, which was approved by the Deakin University Human Research Ethics Committee. All experimental procedures and possible risks were explained to subjects, verbally and in writing, and subjects provided their written consent. Peak pulmonary oxygen uptake (V˙o 2 peak) was determined during an incremental cycling test to volitional exhaustion on an electromagnetically braked cycle ergometer (Lode Instruments, Groningen, The Netherlands) and averaged 60.5 ± 2.2 ml · kg−1 · min−1.

Experimental design.

At least 2 days after the determination ofV˙o 2 peak, subjects performed fatiguing cycle exercise late in the afternoon. Exercise consisted of constant load exercise at power outputs that varied between 60 and 80%V˙o 2 peak, interspersed with a series of three 1-min sprints at 100% V˙o 2 peakwith 2-min recovery periods. Subjects ceased cycle exercise after 2 h and began arm-cranking exercise, which consisted of four 5-min intervals at a power output of 35–50 W. The arm exercise was employed to provide an alternate site for glucose disposal after the completion of the fatiguing exercise to minimize glycogen synthesis in leg skeletal muscle. After arm exercise, subjects recommenced cycle exercise, which consisted of 1-min efforts at 100%V˙o 2 peak with 2-min recovery periods. Fatigue was defined as the point when subjects were unable to maintain their cadence >60 rpm. Upon completion of exercise, subjects were provided with a standardized meal (4.5 MJ, 73% fat, 9% carbohydrate, 18% protein) and were instructed to abstain from caffeine, alcohol, tobacco, and exercise.

On the next morning (8 AM) after an overnight fast, subjects arrived at the laboratory, voided, and rested on a couch. An indwelling Teflon catheter was inserted into an antecubital vein of one arm to permit the infusion of [6,6-2H]glucose and epinephrine or saline solutions and in a contralateral arm vein for venous blood sampling. After a blood sample was obtained for subsequent determination of background isotopic enrichment, a primed (3.3 mmol) continuous (47.4 ± 2.7 mmol/min) infusion of [6,6-2H]glucose (Cambridge Isotope Laboratories, Cambridge, MA) was begun using a peristaltic pump (Minipuls 3, Gilson, Villiers Le Bel, France). With the subject under local anesthesia with lidocaine, two incisions were made in the skin and fascia overlying the vastus lateralis muscle for subsequent muscle sampling.

Subjects commenced cycle exercise at 59 ± 1%V˙o 2 peak and continued for 40 min. After 20 min, exercise was interrupted and, while subjects remained on the cycle ergometer, a muscle sample was obtained from the vastus lateralis, by use of the percutaneous needle biopsy technique with suction, and quickly frozen in liquid nitrogen for later analysis. The time delay between cessation and recommencement of exercise was 1 min. During the rest period, an infusion of saline (CON) or epinephrine (EPI, 7.43 ng · kg−1 · min−1) was started by syringe pump (IVAC P3000, IVAC Medical Systems, Hampshire, UK). This infusion rate was chosen on the basis of our previous study (31) and was designed to produce plasma epinephrine levels in the high physiological range. Immediately upon cessation of exercise, a second muscle sample was obtained. The order of trials was counterbalanced, and they were conducted ≥1 wk apart.

Venous blood samples were obtained at 5-min intervals 10 min before and throughout exercise for determination of glucose, [2H]glucose enrichment, and lactate. Additional blood was drawn at rest and during 10-min intervals throughout exercise for determination of free fatty acids (FFA), catecholamines, and insulin. Blood for glucose, [2H]glucose enrichment, lactate, and insulin was placed in lithium-heparin tubes and rolled. For epinephrine, norepinephrine, and FFA, 1.5 ml of whole blood were added to 30 μl of EGTA and reduced glutathione. All blood was spun, and the plasma was stored at −20°C. Plasma for catecholamine analysis was stored at −80°C. Expired gases were collected and analyzed on-line (Gould 2900 Metabolic System) for 5 min at 10-min intervals throughout exercise, with the average of the values during the last minute being recorded. Heart rate was measured continuously via telemetry (Polar sports tester, Polar Electro, Finland) and recorded every 5 min. Whole body carbohydrate oxidation was estimated from the equation 4.585 V˙co 2 − 3.226V˙o 2 (23), whereV˙co 2 andV˙o 2 are carbon dioxide production and oxygen uptake, respectively.

Analytic techniques.

Plasma glucose (G) and lactate were measured using an automated analyzer (EML105, Radiometer, Copenhagen, Denmark). Plasma FFA were analyzed using an enzymatic, colorimetric method (Wako NEFA-C test kit, Wako Chemicals, Richmond, VA). Plasma catecholamines were determined using a single-isotope radioenzymatic method (TRK 995, Amersham, Buckinghamshire, UK), and plasma insulin (Phadeseph, Pharmacia & Upjohn, Uppsala, Sweden) was measured by radioimmunoassay. Plasma [2H]glucose enrichment (E) was measured as previously described (31). The rates of glucose appearance (glucose Ra) and disappearance (glucose Rd) were calculated using a modified one-pool non-steady-state model (30), with the assumption of a pool fraction of 0.65 and an estimate of the glucose space as 25% of body mass, to estimate the volume of distribution (V)Ra={FV[(G1+G2)/2][(E2E1)/(t2t1)]×[(E2+E1)/2]1} Rd=Ra{V×[(G2G1)/(t2t1)]} Glucose Ra was assumed to equal hepatic glucose production in the fasted state. After freeze-drying, muscle samples were dissected free of visible fat, blood, and connective tissue and were powdered and divided into two aliquots. One aliquot was extracted according to the methods of Harris et al. (10), and the homogenate was analyzed for lactate, creatine phosphate (CP), creatine, and glucose 6-phosphate (G-6-P) by use of enzymatic fluorometric methods (20). A second aliquot was extracted in 2 N HCl, boiled for 2 h, and neutralized with 0.67 N NaOH. The homogenate was analyzed for glycogen according to the method of Passonneau and Lauderdale (22).

Statistical analysis.

Results from the two trials were compared using a two-way ANOVA for repeated measures, with statistical significance defined asP < 0.05. Where appropriate, specific differences were located by the Newman-Keuls post hoc test. All data are presented as means ± SE (n = 6).


Plasma epinephrine levels were similar between trials before and throughout the initial 20 min of exercise. The infusion of epinephrine at 20 min increased (P < 0.05) the plasma epinephrine concentration at 30 and 40 min compared with CON (Fig.1). Plasma norepinephrine increased (P < 0.05) during exercise in both trials, and epinephrine infusion further enhanced (P < 0.05) this response (see Table 2).

Fig. 1.

Plasma epinephrine concentrations at rest and during 40 min of exercise at 59 ± 1% peak pulmonary O2 uptake (V˙o 2 peak) with (EPI, ●) or without (CON, ○) epinephrine infusion starting at 20 min. Values are means ± SE;n = 6. *P < 0.05, significantly different from corresponding time point.

During exercise, V˙o 2 increased (P < 0.05) with time, and there was no difference between trials (Table 1). Respiratory exchange ratio was not different between trials, and estimated carbohydrate oxidation was not different between trials during the first 20 min of exercise. Epinephrine infusion increased (P < 0.05) carbohydrate oxidation in EPI at 30 min; however, no difference between trials was observed at 40 min (Table 1). Heart rate was elevated (P < 0.05) in EPI at 30 and 40 min of exercise compared with CON (Table 1).

View this table:
Table 1.

Physiological responses during exercise at 59 ± 1%V˙o 2 peak with or without epinephrine infusion

Plasma lactate increased (P < 0.05) at the onset of exercise in both trials, and the infusion of epinephrine resulted in higher lactate levels during exercise (Table2). Plasma FFA were elevated (P < 0.05) late in exercise but were not different between trials (Table 2). Plasma insulin decreased (P < 0.05) during exercise and was unaffected by epinephrine infusion (Table 2).

View this table:
Table 2.

Plasma metabolite and hormone concentrations during exercise at 59 ± 1% V˙o 2 peak with or without epinephrine infusion

The combination of fatiguing exercise and a low carbohydrate diet on the preceding day resulted in low muscle glycogen levels after the initial 20 min of exercise that were not different between trials. Although exercise from 20 to 40 min decreased (P < 0.05, time effect) muscle glycogen further, very little net glycogenolysis occurred in each trial, and no difference between trials was evident (Table 3). G-6-Plevels during exercise were not different between trials. Muscle lactate, CP, creatine, and ATP were not different with either treatment or time (Table 3).

View this table:
Table 3.

Muscle metabolite concentrations during exercise at 59 ± 1%V˙o 2 peak with or without epinephrine infusion

Plasma glucose was higher (P < 0.05) in EPI compared with CON at 35 and 40 min of exercise (Fig.2). Glucose Ra increased (P < 0.05, time effect) in both trials, but no differences between trials were observed (Fig. 2). Thus the increased plasma glucose was due to a lower glucose Rd in EPI. Glucose Rd increased (P < 0.05) from rest during both trials, and although glucose Rd continued to rise in CON, the infusion of epinephrine decreased (P< 0.05) Rd in EPI after 30 and 40 min of exercise (Fig.2).

Fig. 2.

Plasma glucose concentration (A), glucose rate of appearance (Ra, B), and glucose rate of disappearance (Rd, C) at rest and during 40 min of exercise at 59 ± 1% V˙o 2 peakwith (EPI, ●) or without (CON, ○) epinephrine infusion starting at 20 min. Values are means ± SE;n = 6. *P < 0.05, significantly different from corresponding time point.


The present study has demonstrated that the decrease in glucose disposal observed with elevated epinephrine during moderate exercise is unlikely to be due to inhibition of glucose phosphorylation, secondary to enhanced muscle glycogenolysis and increased muscle G-6-Plevels. Rather, our results suggest, albeit indirectly, that the effect of epinephrine may be mediated via effects on sarcolemmal glucose transport.

Previous studies have demonstrated an inhibitory effect of epinephrine on glucose disposal during exercise. The infusion of epinephrine to physiological levels decreased glucose uptake during exercise in adrenalectomized, epinephrine-deficient humans (12) and in men who underwent anesthesia of the celiac ganglion to block endogenous epinephrine production (17). Furthermore, the present investigation and a recent study from our laboratory (31) demonstrated a decrease in glucose Rd during moderate cycle exercise when plasma epinephrine was elevated. In addition, in studies employing β-adrenergic blockade, an increase in glucose uptake across a range of exercise intensities has been observed (27,28). Taken together, these findings support the hypothesis that epinephrine decreases glucose disposal during exercise.

The mechanism or mechanims underlying the decreased glucose Rd during exercise are possibly related to effects on sarcolemmal glucose transport, glucose phosphorylation, and/or glucose delivery. It has been suggested previously that the increase in glycogenolysis associated with elevated plasma epinephrine (7,14, 25, 31) results in an accumulation of cytosolic G-6-P, which inhibits hexokinase (21) and glucose phosphorylation. Indeed, we have previously demonstrated a twofold increase in glycogen utilization, greater accumulation of G-6-P (6.5–8 mmol/kg dry matter), and reduced whole body glucose Rd with epinephrine infusion during moderate exercise (31). In the present study, the combination of fatiguing exercise and a low carbohydrate diet was successful in reducing muscle glycogen levels and flux through glycogenolysis and resulted in G-6-P levels during exercise that were only 25–30% of those we observed in our previous study (31), even with epinephrine infusion. Despite the marked differences in muscle G-6-P levels, elevation of plasma epinephrine to similar levels reduced glucose Rd to the same extent in both studies. Thus the inhibitory effect of epinephrine on glucose Rd appears to be independent of the muscle G-6-P level; this lends support to the possibility of a direct effect of epinephrine on sarcolemmal glucose transport.

It has been suggested previously that β-adrenergic stimulation inhibits glucose transport via a cAMP-dependent pathway (18), and this was supported by the findings that isoproterenol caused phosphorylation of GLUT4 (13) and markedly decreased insulin-stimulated glucose uptake in rat adipocytes (13, 16). β-Adrenergic stimulation was also shown to phosphorylate GLUT4 in rat skeletal muscle (19); however, no change in insulin (1, 3, 19)- or contraction (1)-stimulated glucose transport was observed with epinephrine infusion in skeletal muscle. In contrast, Bonen et al. (2) reported decreased GLUT4 intrinsic activity in a rat skeletal muscle with epinephrine administration. Moreover, epinephrine reduced insulin-stimulated glucose transport in a dose-dependent manner in rat hindlimb preparations, despite unaltered GLUT4 translocation (8).

Another possibility that cannot be discounted is a vasoconstrictor effect of increased epinephrine and norepinephrine (Table 2), which could potentially reduce glucose delivery to contracting skeletal muscle and limit glucose Rd. We have no data in the present study to support or refute such a mechanism; however, it should be noted that any reduction in skeletal muscle blood flow on glucose Rd may have been partly offset by the higher plasma glucose concentration after epinephrine infusion (Fig. 2).

The observation that glycogen utilization was not different between trials was somewhat unexpected, as epinephrine has been shown to increase muscle glycogen utilization during exercise (7, 14, 25,26, 31). Furthermore, the magnitude of epinephrine-stimulated glycogen breakdown is thought to be independent of muscle glycogen concentration in rat epitrochlearis (15), which suggests that glycogenolysis should have been greater in EPI. Although we detected no significant change in net glycogen utilization between trials (19 ± 9 and 25 ± 8 mmol/kg dry matter for CON and EPI, respectively), we cannot rule out the possibility that greater muscle glycogenolysis may have occurred at the onset of epinephrine infusion, but this small change was not detected over the 20-min sampling period. The absence of an effect of epinephrine on glycogenolysis during exercise in the glycogen-depleted state may be due to the release of phosphorylase from the glycogen-protein-sarcoplasmic reticulum complex (6). The release of phosphorylase would cause an uncoupling of phosphorylase, phosphorylase kinase, and Ca2+, all of which are essential for the activation of the kinase (6). Although epinephrine stimulation results in phosphorylation and activation of the kinase, any effect of epinephrine would be inhibited by the uncoupling of phosphorylase from the kinase and Ca2+. Partial support for this hypothesis was reported in glycogen-depleted rats, where the ability of epinephrine to activate phosphorylase was blunted by exhaustive exercise (4). Although the underlying mechanism is unknown, our results suggest that, in contrast to situations in which preexercise glycogen concentration is normal (7,31), epinephrine may not enhance glycogen breakdown when preexercise glycogen concentrations are low.

Finally, it is worth noting the effect of muscle glycogen on glucose Rd during exercise. In the present study, in which preexercise muscle glycogen levels were ∼150 mmol/kg dry matter, glucose clearance over the first 20 min of exercise in CON averaged 113 ± 16 ml/kg. This was higher (P < 0.05) than the value of 71 ± 10 ml/kg for the corresponding trial in our previous study (31), when preexercise muscle glycogen was ∼500 mmol/kg dry matter. Subject characteristics and absolute and relative exercise intensities were similar, whereas plasma glucose and insulin levels were lower (P < 0.05) in the present study. The higher glucose clearance with low muscle glycogen is consistent with previous results obtained in contracting, perfused rat muscle (5, 11) and is thought to be due to effects of muscle glycogen availability on both sarcolemmal glucose transport and intracellular glucose metabolism.

In summary, the results of the present study have shown that glucose uptake is reduced during exercise with elevated plasma epinephrine and that this was not associated with elevated G-6-P, secondary to enhanced muscle glycogenolysis. Rather, the inhibitory effect of epinephrine on glucose disposal during moderate exercise may be due to reduced sarcolemmal glucose transport; this warrants further investigation.


We acknowledge the excellent medical assistance of Dr. Andrew Garnham, School of Health Sciences, Deakin University, and we thank Dr. Mark Febbraio, Department of Physiology, The University of Melbourne, for use of laboratory facilities in the plasma catecholamine analyses, and Dom Caridi, Department of Chemistry and Biology, Victoria University, for technical assistance in measuring plasma [2H]glucose enrichment.


  • This study was supported by the Australian Research Council.

  • Address for correspondence: M. Hargreaves, School of Health Sciences, Deakin Univ., Burwood, Victoria 3125, Australia (E-mail: mharg{at}deakin.edu.au).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • May 21, 2002;10.1152/ajpendo.00098.2002


View Abstract