Metabolic Responses to Exercise Effects of endurance training and implications for diabetes ERIK A. RICHTER, MD, DSC LORRAINE TURCOTTE, PHD PETER HESPEL, PHD BENTE KIENS, PHD
In this study, some important metabolic responses to exercise will be discussed, and aspects of particular interest for patients with diabetes mellitus will be emphasized. Alterations in the metabolic responses to exercise induced by physical endurance training and consequences of training for metabolism of plasma lipids and lipoproteins will be discussed. Glucoregulation during exercise is not perfect in normal subjects and is less so in patients with diabetes mellitus. For instance, during intense exercise, large increases in the plasma glucose concentration occur and a state of insulin resistance exists for a few hours after intense exercise. Even so, increased sensitivity to insulin is found the day after intense exercise and also shortly after more moderate intensity exercise, both in healthy subjects and in patients with diabetes mellitus. Increased sensitivity to insulin is also found after endurance training, whereas insulin sensitivity is decreased after inactivity. Exercise training increases the ability of muscle to take up and oxidize free fatty acids during exercise and also increases the activity of the enzyme lipoprotein lipase in muscle. The activity of lipoprotein lipase in muscle correlates with muscle insulin sensitivity. This might explain why insulin resistance is often associated with hypercholesterolemia, hypertriglyceridemia, and low high-density lipoprotein cholesterol.
E
xercise is the most powerful perturbant of metabolism in normal life. Yet, mechanisms exist to maintain homeostasis so that the demands of the working muscles for fuel and oxygen are met while the function of the rest of the body is maintained. In an effort to outline some of the key metabolic processes during exercise and recent advances in our understand-
ing of these processes, the topics for this study were selected. Herein, the glycemic response to exercise and the effect of endurance training will be discussed. Next, aspects of fuel selection in working muscle will be reviewed, and finally effects of exercise on insulin sensitivity are considered. The influence of endurance training on these parameters is also discussed.
FROM THE AUGUST KROGH INSTITUTE, UNIVERSITY OF COPENHAGEN, COPENHAGEN, DENMARK; AND THE INSTITUTE OF PHYSICAL EDUCATION, CATHOLIC UNIVERSITY OF LEUVEN, LEUVEN, BELGIUM. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO ERIK A. RICHTER, MD, DSC, AUGUST KROGH INSTITUTE, 13 UNIVERSITETSPARKEN DK-2100, COPENHAGEN, DENMARK.
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER
1992
GLYCEMIC RESPONSE TO EXERCISE— Maintenance of plasma glucose concentration has high priority at rest and during exercise. During moderate intensity exercise, euglycemia is usually maintained, and the prevailing opinion has been that euglycemia is maintained during exercise because of feedback signals from glucosensors that sense a minute fall in glycemia elicited by the increase in peripheral glucose utilization. Such signals then elicit changes in neuroendocrine function so that hepatic glucose production is increased to an extent that matches the exercise-induced increase in glucose utilization and thus maintains euglycemia. Furthermore, a decrease in plasma glucose concentration may by itself directly stimulate hepatic glucose production (1,2). Although evidence for feedback mechanisms regulating hepatic glucose production during exercise is strong (1-4), it apparently is not the only mechanism that operates to increase hepatic glucose production during exercise, especially not during intense exercise. This is clear from observations that indicate that plasma glucose concentrations may increase during intense exercise. From the available evidence a picture emerges: when exercise intensity is low, plasma glucose concentrations are stable or decrease. On the other hand, when intensity of exercise is high, plasma glucose concentrations initially increase but may eventually decrease as exercise continues and hepatic glycogen stores are depleted (Fig. 1). It should be noted that arm exercise elicits a response above that of a leg exercise of the same relative intensity (Fig. 1). This is probably because of the relatively large increase in catecholamine concentrations that arm exercise elicits (5,6). Any deviance from euglycemia is obviously the result of a mismatch between hepatic glucose production and peripheral glucose utilization. Because hepatic glucose production is the primary regulated function, an increase in the blood glucose concentration conse-
1767
Metabolism in exercise
ID (5)
161
Exercise time (min)
18) (7)
Figure 1—Gfycemic response to exercise of different intensities. Numbers refer to the following studies: 1) running at 83% Vo2max (unpublished observations); 2) bicycling at 60% Vo2max (unpublished observations); 3, bicycling at 50% of Vo2max (unpublished observations); 4, onelegged knee extensions at 23% Vo2max (ref. 49); 5) running at 75% Vo2max (unpublished observations); 6) arm cycling at 30% Vo2max (ref. 5); 7) bicycling at 30% Vo2max (ref 5); 8, bicycling at 70% Vo2max (ref. 92); and 9) bicycling at 85% Vo2max (ref. 19).
quently occurs when hepatic glucose production is increased more than necessary to keep up with peripheral glucose utilization, whereas a decrease in glycemia occurs when hepatic glucose production fails to keep pace with peripheral glucose utilization. It follows that, during intense exercise, the stimulation of hepatic glucose production is larger than the increase in peripheral glucose utilization. It is beyond the scope of this study to go into detail with mechanisms regulating hepatic glucose production during exercise, but changes in concentrations of insulin (7,8), glucagon (8,9), glucose itself (1,2), and sympathoadrenal activity (10,11) probably all play a role. Nevertheless, these mechanisms may be redundant, and the absence of one mechanism can be compensated for by the others. The concept has been put
1768
forward that the glycemic response to exercise in addition to feedback regulation is subject to feed-forward regulation (i.e., that, at the onset of exercise, impulses from the working muscles and motor centers increase neuroendocrine activity directly dependent on work intensity, and this increased neuroendocrine activity then enhances hepatic glucose production) (12). This primary response to exercise is subsequently modulated by feedback mechanisms. Thus, when exercise intensity is high, central command elicits such changes in neuroendocrine function that an initial overshoot in hepatic glucose production occurs. The glycemic response to exercise also depends on the nutritional status. When subjects are fed a carbohydrate-rich diet for 4 days, plasma glucose concentration is higher during exercise than when they are fed 4 days of a fatrich diet (13). This probably in part relates to a higher hepatic glycogen store when fed carbohydrates compared with fats, because it has been shown in rats that the higher the preexercise glycogen concentration in the liver, the more likely is plasma glucose to increase during exercise (14,15). The conclusion from the available studies then is that, in healthy subjects, maintainence of euglycemia during exercise is often not achieved. This is especially true at the onset of high intensity exercise when plasma glucose is likely to increase. It is not surprising that, in patients with diabetes, the glycemic response to exercise is even more variable. In insulin-dependent diabetes (type I), it is generally believed that the glycemic response to moderate intensity exercise is dependent on the preexercise metabolic status (16). Thus, when the patients are in good metabolic control and plasma insulin concentrations are rather high, they are likely to decrease plasma glucose concentrations during exercise, whereas when exercise is conducted in the ketotic, insulin-deficient state, a worsening of ketosis and a fur-
ther increase in glycemia can be expected. Given the facts that, even in nondiabetic subjects glucoregulation during exercise is not perfect and that patients with insulin-dependent diabetes cannot regulate their plasma insulin concentration during exercise, it is to be expected that glucoregulation during various types of exercise in diabetics might be quite variable and to some extent unpredictable. Part of the unpredictability stems from the rather variable effect of exercise on absorption of insulin from a subcutanous depot (17,18). From the studies in healthy individuals, it would appear that maintenance of stable blood glucose concentrations is most likely at exercise intensities of - 5 0 - 6 0 % of Vo 2max) although prolonged exercise at this intensity may lead to hypoglycemia. 50-60% of Vo 2max is probably also the exercise intensity most likely to maintain blood glucose concentrations stable in patients with diabetes in good metabolic control. In fact, in agreement with predictions from exercise studies in healthy subjects (19,20), intense exercise has been shown to increase plasma glucose concentrations in both type I (21) and non-insulin-dependent (type II) diabetes (22; Fig. 2). Furthermore, like in healthy individuals, this increase may last for several hours (21,22) and is larger the higher the preexercise plasma glucose concentration is in patients with type I diabetes (21). Effects of endurance training The glycemic response to exercise is influenced by the training status of the individual. It has been repeatedly shown that well-trained subjects at a given moderate to high exercise intensity increase their plasma glucose concentration more than untrained subjects do (20,23). Trained individuals are better able to maintain their plasma glucose concentrations during prolonged exercise than are untrained subjects (24,25). These observations agree well with the well-known effect of endurance training to increase the combustion of lipid fuels at the ex-
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER
1992
Richter and Associates
GLUCOSE 210 •
170
mg-dl-'
endurance training in diabetic patients (33), endurance training is expected to result in basically similar metabolic changes in patients with diabetes as in healthy individuals.
130
FUEL METABOLISM DURING EXERCISE— During exercise, the metabolic rate of muscle increases many 0< times above the rate at rest. The selection • • N1DDM of fuels for this increased metabolic rate • • Control is a process that has been studied for years, but nevertheless many questions remain unanswered. In an effort to understand the mechanism that regulates selection of glucose versus muscle glycogen during exercise, we recently performed studies using the isolated perfused rat hindlimb 60 120 0 20 180 preparation. Rats were preconditioned 0-12 Recovery, min Exercise* by different dietary and exercise regimes to have preperfusion muscle glycogen Figure 2—Effect of graded short-term (12- concentrations between 60 and 10 min) bicycle exercise on plasma glucose and in-fimol/g wet weight. They were perfused sulin concentrations in non-insulin-dependent with a standard medium containing 6 diabetic patients and in controls. NIDDM, non- mM glucose and no insulin at rest and insulin-dependent diabetes mellitus. From Kjcerfor 15 min of intermittent subtetanic et al. (22). © by the American Physiological contractions. Glucose uptake increased Society. with muscle contractions in all groups, but the increase was larger the lower the muscle glycogen concentration (34). pense of carbohydrates during exercise This is in accordance with findings in performed at the same absolute intensity man (35). In fact, a linear negative cor(24,26). The sparing of carbohydrate fu- relation between initial muscle glycogen els include a decreased rate of muscle concentration and glucose uptake during glycogen breakdown (27,28). A slower contractions was obtained (Fig. 3). The rate of glycogen depletion in the liver has mechanism behind this phenomenon apbeen found in trained compared with parently is multifactorial. In this hindlimb preparation, muscle glycogen untrained rats during exercise (27). It was shown in humans that en- breakdown during contractions is lindurance training also decreases plasma early related to preexercise muscle glyglucose turnover and oxidation during cogen concentrations (34,36). The high moderate intensity exercise (29). The ex- rate of glycogen breakdown leads to high act mechanism behind this decrease in intramuscular concentrations of glucoseplasma glucose utilization is at present 6-phosphate (G-6-P) (34), which inunknown, but may include changes in hibit hexokinase (37). This in turn reglucose transport and changes in glucose sults in increased intracellular muscle utilization, secondary to the training- glucose concentrations (34) that decrease induced increased enzymatic capacity for the gradient of glucose from the interstioxidative metabolism of lipid (27,30- tial space to the cytoplasm, in turn de32). Because similar adaptations in mus- creasing net glucose uptake. However, cle enzymatic capacity are induced by this was apparently not the only mecha-
a 3 10
90
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER 1992
Initial glycogen concentration (pmol /g w.w.l
Figure 3—Relationship between muscle glycogen concentrations and glucose uptake during contractions in perfused rat hindquarter. From Hespel and Richter (34). © by the Physiological Society, London.
nism by which high preexercise muscle glycogen concentrations decreased muscle glucose uptake. When muscle glucose transport was estimated using the uptake of [14C]-3-O-methylglucose during contractions, it was found that glucose transport was —25% lower when glycogen concentrations were high than when they were low (34), suggesting that muscle membrane permeability was affected by the intracellular muscle milieu. Although the exact mechanism behind this phenomenon is not known it could be related to changes in intracellular pH. When glycogen concentrations are high, lactate production is higher than when glycogen stores are low (36). Thus, intracellular pH is probably lower in contracting muscle with high glycogen concentrations than in muscles with low glycogen. Preliminary data in isolated muscle membrane vesicles show that glucose transport is decreased by a fall in pH from the optimal value of pH 7.2 (unpublished observations). This is in accordance with findings in erythrocytes (38). Thus, a larger decrease in pH during contractions in muscles with high glycogen concentrations compared with muscles with low glycogen may in part be responsible for the lower rate of glucose transport observed in the former group.
1769
Metabolism in exercise
Another possibility of fuel competition is between carbohydrate and lipids. Since the description of the "glucosefatty acid cycle" almost 30 yr ago (39,40), it has been assumed that an increased supply of fatty acids would decrease carbohydrate oxidation during exercise. The original observations were done in perfused rat hearts and in hemidiaphragms (39,40), and evidence for the operation of the glucose-fatty acid cycle during exercise in humans has been scarce. Support for its operation during exercise comes from a study (41) in which muscle glycogen breakdown during running was decreased when plasma free fatty acids (FFAs) were increased by a fatty meal and heparin injection, and from a study in which muscle glycogen breakdown was increased when FFA were decreased by administration of nicotinic acid (42). In contrast, there was no effect of increased FFA on carbohydrate oxidation during prolonged, moderate intensity exercise (43). In rats, slower muscle glycogen utilization and higher blood glucose concentrations were found when FFAs were elevated (44,45), whereas in perfused rat muscle controversy exists as to whether FFAs decrease muscle glucose uptake and glycogen utilization (46-48). In view of the controversy regarding effects of FFA on muscle metabolism during exercise in humans, we recently conducted a study in which muscle metabolism was carefully studied during kneeextensor exercise, first with one leg at a normal (—500 jxM) and then with the other leg at an elevated (-~1100 |xM) plasma concentration of FFA (49). The increase in FFA was obtained by infusion of intralipid and heparin, and substrate metabolism was studied by arteriovenous catherization and by taking muscle biopsies. Glucose uptake was significantly lower during infusion of intralipid both at rest, during exercise, and after 10 min of recovery (Fig. 4). On the other hand, glycogen breakdown was identical in the two legs (Fig. 5) and so was release of
1770
•3
50
j
4.8
O
4.4 4.2
Figure 4—Blood glucose concentration and thigh glucose uptake at rest, during knee extensions, and after 10-min recovery. Subjects were studied either in control situation or during an infusion of intralipid and heparin. Values are means ± SE of 11 observations. X denotes P < 0.05 compared with control. From Hargreaves et al. (49). © by the American Physiological Society.
lactate, pyruvate, and citrate (49). At the end of the 1-h knee extension, muscle G-6-P and glucose were identical whether FFAs were raised or not (49). This makes it difficult to attribute the decrease in glucose uptake to the operation of the glucose fatty acid cycle, because this would require an increase in both G-6-P and glucose in muscle, and such an increase was not found. However, glucose and G-6-P were measured in muscle homogenates, and it cannot be excluded that physiologically important changes in the concentration of these compounds might occur in small compartments in the cytoplasm. Yet, if the compartments are small, their contribution to the concentration in a homogenate may be undetectable. Alternatively, if an increase in FFA does not increase G-6-P and glucose in muscle, then it is tempting to speculate that the decrease in glucose uptake during intralipid infusion was caused by a direct effect of increased FFA on muscle glucose trans-
port. There is some experimental evidence for a direct effect of FFA on cardiac muscle glucose transport (40), but the question clearly needs more investigation. It is difficult to reconcile why an effect of FFA on glucose uptake is so inconsistently found. Even in our own hands, we find an effect of elevated FFA on glucose uptake (Fig. 4), whereas we repeatedly have been unable to detect any effect of palmitate on glucose uptake (48,50) in perfused muscle. Part of the explanation may lie in the type of fatty acid used. Intralipid consists mainly of soy bean oil that on hydrolysis liberates mainly the unsaturated fatty acids linolic acid (18:2) and oleic acid (18:1). In the only study of perfused muscle, which has shown an effect of increasing FFA on glucose uptake, oleate was used (47), whereas the studies that have shown no effect of FFA on glucose uptake used palmitate, a saturated fatty acid (46,48, 50), octanoate (46), or simply ketone bodies (46). It is not inconceivable that the different fatty acids may have different effects on muscle glucose transport and metabolism, but direct experimental evidence for this notion is lacking.
1
600
£00
200
Pre Post Control
Pre Post Intralipid
Figure 5—Muscle glycogen concentrations before and after 60 min of knee extensions during control conditions and infusion of intralipid and heparin. Values are means ± SE of 11 observations. From Hargreaves et al. (49). © by the American Physiological Society.
DIABETES CARE, VOLUME 15,
SUPPLEMENT 4,
NOVEMBER
1992
Richter and Associates
An important point in fatty acid metabolism is the concentration dependency of FFA uptake. It is generally believed that the uptake and oxidation of FFA in muscle is linearly related to the plasma concentration of FFA (51,52). This has been interpreted as evidence supporting that the transport of FFA across the plasma membrane occurs by simple diffusion. However, until recently, it has been ignored that, by far, the largest part of FFA in vivo is bound to albumin and that only the small unbound fraction of FFA (usually
In this study, some important metabolic responses to exercise will be discussed, and aspects of particular interest for patients with diabetes mellitus will be emphasized. Alterations in the metabolic responses to exercise induced by physical endurance training and consequences of training for metabolism of plasma lipids and lipoproteins will be discussed. Glucoregulation during exercise is not perfect in normal subjects and is less so in patients with diabetes mellitus. For instance, during intense exercise, large increases in the plasma glucose concentration occur and a state of insulin resistance exists for a few hours after intense exercise. Even so, increased sensitivity to insulin is found the day after intense exercise and also shortly after more moderate intensity exercise, both in healthy subjects and in patients with diabetes mellitus. Increased sensitivity to insulin is also found after endurance training, whereas insulin sensitivity is decreased after inactivity. Exercise training increases the ability of muscle to take up and oxidize free fatty acids during exercise and also increases the activity of the enzyme lipoprotein lipase in muscle. The activity of lipoprotein lipase in muscle correlates with muscle insulin sensitivity. This might explain why insulin resistance is often associated with hypercholesterolemia, hypertriglyceridemia, and low high-density lipoprotein cholesterol.
E
xercise is the most powerful perturbant of metabolism in normal life. Yet, mechanisms exist to maintain homeostasis so that the demands of the working muscles for fuel and oxygen are met while the function of the rest of the body is maintained. In an effort to outline some of the key metabolic processes during exercise and recent advances in our understand-
ing of these processes, the topics for this study were selected. Herein, the glycemic response to exercise and the effect of endurance training will be discussed. Next, aspects of fuel selection in working muscle will be reviewed, and finally effects of exercise on insulin sensitivity are considered. The influence of endurance training on these parameters is also discussed.
FROM THE AUGUST KROGH INSTITUTE, UNIVERSITY OF COPENHAGEN, COPENHAGEN, DENMARK; AND THE INSTITUTE OF PHYSICAL EDUCATION, CATHOLIC UNIVERSITY OF LEUVEN, LEUVEN, BELGIUM. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO ERIK A. RICHTER, MD, DSC, AUGUST KROGH INSTITUTE, 13 UNIVERSITETSPARKEN DK-2100, COPENHAGEN, DENMARK.
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER
1992
GLYCEMIC RESPONSE TO EXERCISE— Maintenance of plasma glucose concentration has high priority at rest and during exercise. During moderate intensity exercise, euglycemia is usually maintained, and the prevailing opinion has been that euglycemia is maintained during exercise because of feedback signals from glucosensors that sense a minute fall in glycemia elicited by the increase in peripheral glucose utilization. Such signals then elicit changes in neuroendocrine function so that hepatic glucose production is increased to an extent that matches the exercise-induced increase in glucose utilization and thus maintains euglycemia. Furthermore, a decrease in plasma glucose concentration may by itself directly stimulate hepatic glucose production (1,2). Although evidence for feedback mechanisms regulating hepatic glucose production during exercise is strong (1-4), it apparently is not the only mechanism that operates to increase hepatic glucose production during exercise, especially not during intense exercise. This is clear from observations that indicate that plasma glucose concentrations may increase during intense exercise. From the available evidence a picture emerges: when exercise intensity is low, plasma glucose concentrations are stable or decrease. On the other hand, when intensity of exercise is high, plasma glucose concentrations initially increase but may eventually decrease as exercise continues and hepatic glycogen stores are depleted (Fig. 1). It should be noted that arm exercise elicits a response above that of a leg exercise of the same relative intensity (Fig. 1). This is probably because of the relatively large increase in catecholamine concentrations that arm exercise elicits (5,6). Any deviance from euglycemia is obviously the result of a mismatch between hepatic glucose production and peripheral glucose utilization. Because hepatic glucose production is the primary regulated function, an increase in the blood glucose concentration conse-
1767
Metabolism in exercise
ID (5)
161
Exercise time (min)
18) (7)
Figure 1—Gfycemic response to exercise of different intensities. Numbers refer to the following studies: 1) running at 83% Vo2max (unpublished observations); 2) bicycling at 60% Vo2max (unpublished observations); 3, bicycling at 50% of Vo2max (unpublished observations); 4, onelegged knee extensions at 23% Vo2max (ref. 49); 5) running at 75% Vo2max (unpublished observations); 6) arm cycling at 30% Vo2max (ref. 5); 7) bicycling at 30% Vo2max (ref 5); 8, bicycling at 70% Vo2max (ref. 92); and 9) bicycling at 85% Vo2max (ref. 19).
quently occurs when hepatic glucose production is increased more than necessary to keep up with peripheral glucose utilization, whereas a decrease in glycemia occurs when hepatic glucose production fails to keep pace with peripheral glucose utilization. It follows that, during intense exercise, the stimulation of hepatic glucose production is larger than the increase in peripheral glucose utilization. It is beyond the scope of this study to go into detail with mechanisms regulating hepatic glucose production during exercise, but changes in concentrations of insulin (7,8), glucagon (8,9), glucose itself (1,2), and sympathoadrenal activity (10,11) probably all play a role. Nevertheless, these mechanisms may be redundant, and the absence of one mechanism can be compensated for by the others. The concept has been put
1768
forward that the glycemic response to exercise in addition to feedback regulation is subject to feed-forward regulation (i.e., that, at the onset of exercise, impulses from the working muscles and motor centers increase neuroendocrine activity directly dependent on work intensity, and this increased neuroendocrine activity then enhances hepatic glucose production) (12). This primary response to exercise is subsequently modulated by feedback mechanisms. Thus, when exercise intensity is high, central command elicits such changes in neuroendocrine function that an initial overshoot in hepatic glucose production occurs. The glycemic response to exercise also depends on the nutritional status. When subjects are fed a carbohydrate-rich diet for 4 days, plasma glucose concentration is higher during exercise than when they are fed 4 days of a fatrich diet (13). This probably in part relates to a higher hepatic glycogen store when fed carbohydrates compared with fats, because it has been shown in rats that the higher the preexercise glycogen concentration in the liver, the more likely is plasma glucose to increase during exercise (14,15). The conclusion from the available studies then is that, in healthy subjects, maintainence of euglycemia during exercise is often not achieved. This is especially true at the onset of high intensity exercise when plasma glucose is likely to increase. It is not surprising that, in patients with diabetes, the glycemic response to exercise is even more variable. In insulin-dependent diabetes (type I), it is generally believed that the glycemic response to moderate intensity exercise is dependent on the preexercise metabolic status (16). Thus, when the patients are in good metabolic control and plasma insulin concentrations are rather high, they are likely to decrease plasma glucose concentrations during exercise, whereas when exercise is conducted in the ketotic, insulin-deficient state, a worsening of ketosis and a fur-
ther increase in glycemia can be expected. Given the facts that, even in nondiabetic subjects glucoregulation during exercise is not perfect and that patients with insulin-dependent diabetes cannot regulate their plasma insulin concentration during exercise, it is to be expected that glucoregulation during various types of exercise in diabetics might be quite variable and to some extent unpredictable. Part of the unpredictability stems from the rather variable effect of exercise on absorption of insulin from a subcutanous depot (17,18). From the studies in healthy individuals, it would appear that maintenance of stable blood glucose concentrations is most likely at exercise intensities of - 5 0 - 6 0 % of Vo 2max) although prolonged exercise at this intensity may lead to hypoglycemia. 50-60% of Vo 2max is probably also the exercise intensity most likely to maintain blood glucose concentrations stable in patients with diabetes in good metabolic control. In fact, in agreement with predictions from exercise studies in healthy subjects (19,20), intense exercise has been shown to increase plasma glucose concentrations in both type I (21) and non-insulin-dependent (type II) diabetes (22; Fig. 2). Furthermore, like in healthy individuals, this increase may last for several hours (21,22) and is larger the higher the preexercise plasma glucose concentration is in patients with type I diabetes (21). Effects of endurance training The glycemic response to exercise is influenced by the training status of the individual. It has been repeatedly shown that well-trained subjects at a given moderate to high exercise intensity increase their plasma glucose concentration more than untrained subjects do (20,23). Trained individuals are better able to maintain their plasma glucose concentrations during prolonged exercise than are untrained subjects (24,25). These observations agree well with the well-known effect of endurance training to increase the combustion of lipid fuels at the ex-
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER
1992
Richter and Associates
GLUCOSE 210 •
170
mg-dl-'
endurance training in diabetic patients (33), endurance training is expected to result in basically similar metabolic changes in patients with diabetes as in healthy individuals.
130
FUEL METABOLISM DURING EXERCISE— During exercise, the metabolic rate of muscle increases many 0< times above the rate at rest. The selection • • N1DDM of fuels for this increased metabolic rate • • Control is a process that has been studied for years, but nevertheless many questions remain unanswered. In an effort to understand the mechanism that regulates selection of glucose versus muscle glycogen during exercise, we recently performed studies using the isolated perfused rat hindlimb 60 120 0 20 180 preparation. Rats were preconditioned 0-12 Recovery, min Exercise* by different dietary and exercise regimes to have preperfusion muscle glycogen Figure 2—Effect of graded short-term (12- concentrations between 60 and 10 min) bicycle exercise on plasma glucose and in-fimol/g wet weight. They were perfused sulin concentrations in non-insulin-dependent with a standard medium containing 6 diabetic patients and in controls. NIDDM, non- mM glucose and no insulin at rest and insulin-dependent diabetes mellitus. From Kjcerfor 15 min of intermittent subtetanic et al. (22). © by the American Physiological contractions. Glucose uptake increased Society. with muscle contractions in all groups, but the increase was larger the lower the muscle glycogen concentration (34). pense of carbohydrates during exercise This is in accordance with findings in performed at the same absolute intensity man (35). In fact, a linear negative cor(24,26). The sparing of carbohydrate fu- relation between initial muscle glycogen els include a decreased rate of muscle concentration and glucose uptake during glycogen breakdown (27,28). A slower contractions was obtained (Fig. 3). The rate of glycogen depletion in the liver has mechanism behind this phenomenon apbeen found in trained compared with parently is multifactorial. In this hindlimb preparation, muscle glycogen untrained rats during exercise (27). It was shown in humans that en- breakdown during contractions is lindurance training also decreases plasma early related to preexercise muscle glyglucose turnover and oxidation during cogen concentrations (34,36). The high moderate intensity exercise (29). The ex- rate of glycogen breakdown leads to high act mechanism behind this decrease in intramuscular concentrations of glucoseplasma glucose utilization is at present 6-phosphate (G-6-P) (34), which inunknown, but may include changes in hibit hexokinase (37). This in turn reglucose transport and changes in glucose sults in increased intracellular muscle utilization, secondary to the training- glucose concentrations (34) that decrease induced increased enzymatic capacity for the gradient of glucose from the interstioxidative metabolism of lipid (27,30- tial space to the cytoplasm, in turn de32). Because similar adaptations in mus- creasing net glucose uptake. However, cle enzymatic capacity are induced by this was apparently not the only mecha-
a 3 10
90
DIABETES CARE, VOLUME 15, SUPPLEMENT 4, NOVEMBER 1992
Initial glycogen concentration (pmol /g w.w.l
Figure 3—Relationship between muscle glycogen concentrations and glucose uptake during contractions in perfused rat hindquarter. From Hespel and Richter (34). © by the Physiological Society, London.
nism by which high preexercise muscle glycogen concentrations decreased muscle glucose uptake. When muscle glucose transport was estimated using the uptake of [14C]-3-O-methylglucose during contractions, it was found that glucose transport was —25% lower when glycogen concentrations were high than when they were low (34), suggesting that muscle membrane permeability was affected by the intracellular muscle milieu. Although the exact mechanism behind this phenomenon is not known it could be related to changes in intracellular pH. When glycogen concentrations are high, lactate production is higher than when glycogen stores are low (36). Thus, intracellular pH is probably lower in contracting muscle with high glycogen concentrations than in muscles with low glycogen. Preliminary data in isolated muscle membrane vesicles show that glucose transport is decreased by a fall in pH from the optimal value of pH 7.2 (unpublished observations). This is in accordance with findings in erythrocytes (38). Thus, a larger decrease in pH during contractions in muscles with high glycogen concentrations compared with muscles with low glycogen may in part be responsible for the lower rate of glucose transport observed in the former group.
1769
Metabolism in exercise
Another possibility of fuel competition is between carbohydrate and lipids. Since the description of the "glucosefatty acid cycle" almost 30 yr ago (39,40), it has been assumed that an increased supply of fatty acids would decrease carbohydrate oxidation during exercise. The original observations were done in perfused rat hearts and in hemidiaphragms (39,40), and evidence for the operation of the glucose-fatty acid cycle during exercise in humans has been scarce. Support for its operation during exercise comes from a study (41) in which muscle glycogen breakdown during running was decreased when plasma free fatty acids (FFAs) were increased by a fatty meal and heparin injection, and from a study in which muscle glycogen breakdown was increased when FFA were decreased by administration of nicotinic acid (42). In contrast, there was no effect of increased FFA on carbohydrate oxidation during prolonged, moderate intensity exercise (43). In rats, slower muscle glycogen utilization and higher blood glucose concentrations were found when FFAs were elevated (44,45), whereas in perfused rat muscle controversy exists as to whether FFAs decrease muscle glucose uptake and glycogen utilization (46-48). In view of the controversy regarding effects of FFA on muscle metabolism during exercise in humans, we recently conducted a study in which muscle metabolism was carefully studied during kneeextensor exercise, first with one leg at a normal (—500 jxM) and then with the other leg at an elevated (-~1100 |xM) plasma concentration of FFA (49). The increase in FFA was obtained by infusion of intralipid and heparin, and substrate metabolism was studied by arteriovenous catherization and by taking muscle biopsies. Glucose uptake was significantly lower during infusion of intralipid both at rest, during exercise, and after 10 min of recovery (Fig. 4). On the other hand, glycogen breakdown was identical in the two legs (Fig. 5) and so was release of
1770
•3
50
j
4.8
O
4.4 4.2
Figure 4—Blood glucose concentration and thigh glucose uptake at rest, during knee extensions, and after 10-min recovery. Subjects were studied either in control situation or during an infusion of intralipid and heparin. Values are means ± SE of 11 observations. X denotes P < 0.05 compared with control. From Hargreaves et al. (49). © by the American Physiological Society.
lactate, pyruvate, and citrate (49). At the end of the 1-h knee extension, muscle G-6-P and glucose were identical whether FFAs were raised or not (49). This makes it difficult to attribute the decrease in glucose uptake to the operation of the glucose fatty acid cycle, because this would require an increase in both G-6-P and glucose in muscle, and such an increase was not found. However, glucose and G-6-P were measured in muscle homogenates, and it cannot be excluded that physiologically important changes in the concentration of these compounds might occur in small compartments in the cytoplasm. Yet, if the compartments are small, their contribution to the concentration in a homogenate may be undetectable. Alternatively, if an increase in FFA does not increase G-6-P and glucose in muscle, then it is tempting to speculate that the decrease in glucose uptake during intralipid infusion was caused by a direct effect of increased FFA on muscle glucose trans-
port. There is some experimental evidence for a direct effect of FFA on cardiac muscle glucose transport (40), but the question clearly needs more investigation. It is difficult to reconcile why an effect of FFA on glucose uptake is so inconsistently found. Even in our own hands, we find an effect of elevated FFA on glucose uptake (Fig. 4), whereas we repeatedly have been unable to detect any effect of palmitate on glucose uptake (48,50) in perfused muscle. Part of the explanation may lie in the type of fatty acid used. Intralipid consists mainly of soy bean oil that on hydrolysis liberates mainly the unsaturated fatty acids linolic acid (18:2) and oleic acid (18:1). In the only study of perfused muscle, which has shown an effect of increasing FFA on glucose uptake, oleate was used (47), whereas the studies that have shown no effect of FFA on glucose uptake used palmitate, a saturated fatty acid (46,48, 50), octanoate (46), or simply ketone bodies (46). It is not inconceivable that the different fatty acids may have different effects on muscle glucose transport and metabolism, but direct experimental evidence for this notion is lacking.
1
600
£00
200
Pre Post Control
Pre Post Intralipid
Figure 5—Muscle glycogen concentrations before and after 60 min of knee extensions during control conditions and infusion of intralipid and heparin. Values are means ± SE of 11 observations. From Hargreaves et al. (49). © by the American Physiological Society.
DIABETES CARE, VOLUME 15,
SUPPLEMENT 4,
NOVEMBER
1992
Richter and Associates
An important point in fatty acid metabolism is the concentration dependency of FFA uptake. It is generally believed that the uptake and oxidation of FFA in muscle is linearly related to the plasma concentration of FFA (51,52). This has been interpreted as evidence supporting that the transport of FFA across the plasma membrane occurs by simple diffusion. However, until recently, it has been ignored that, by far, the largest part of FFA in vivo is bound to albumin and that only the small unbound fraction of FFA (usually
