Glucose turnover in response to exercise during high- and low-FIO2 breathing in man.

Glucose turnover in response to exercise during high- and low-F~o, breathing in humans. Am. J. Physiol. 251 (Endocrinol. Metab. 14): E209-E214, 1986.-The purpose of this study was to assess whether breathing high or low concentrations of O2 could affect glucose turnover during exercise in man. Ten healthy subjects performed two constant work-rate exercise tests, one when the fraction of inspired O2 (FIN,) was 0.15 and the other at the same work rate but when the FIN, was 0.80. The work rate for each subject was chosen so that blood lactate would be elevated during hypoxia, but would be lower during hyperoxia. Glucose appearance (R,) and disappearance (Rd) were measured using the primed, constant infusion of [3-3H]glucose. Although the work rate was the same during hypoxia and hyperoxia in each subject, hypoxic exercise was accompanied by a significantly larger rest to exercise increase in Rd ( ARd) compared with hyperoxia by 265%. Simi-larly, AR, was greater during hypoxia than during hyperoxia by 188%. Lactate to pyruvate ratios were significantly higher during hypoxic exercise suggesting a shift in the cell redox to a more reduced state. Insulin and glucagon were not affected by the FIEF, but both epinephrine and norepinephrine were in- creased during hypoxic exercise, which may explain the increase in R,. The regulation of blood glucose during exercise in vivo appears to be dependent on the availability of oxygen to the

Glucose turnover in response to exercise during high-and low=F~o, breathing in man DAN  Physiol. 251 (Endocrinol. Metab. 14): E209-E214, 1986.-The purpose of this study was to assess whether breathing high or low concentrations of O2 could affect glucose turnover during exercise in man. Ten healthy subjects performed two constant work-rate exercise tests, one when the fraction of inspired O2 (FIN,) was 0.15 and the other at the same work rate but when the FIN, was 0.80. The work rate for each subject was chosen so that blood lactate would be elevated during hypoxia, but would be lower during hyperoxia. Glucose appearance (R,) and disappearance (Rd) were measured using the primed, constant infusion of [3-3H]glucose. Although the work rate was the same during hypoxia and hyperoxia in each subject, hypoxic exercise was accompanied by a significantly larger rest to exercise increase in Rd ( ARd) compared with hyperoxia by 265%. Similarly, AR, was greater during hypoxia than during hyperoxia by 188%. Lactate to pyruvate ratios were significantly higher during hypoxic exercise suggesting a shift in the cell redox to a more reduced state. Insulin and glucagon were not affected by the FIEF, but both epinephrine and norepinephrine were increased during hypoxic exercise, which may explain the increase in R,. The regulation of blood glucose during exercise in vivo appears to be dependent on the availability of oxygen to the working muscle cells. hypoxia; hyperoxia; exercise catecholamines; blood glucose regulation; lactate threshold IN MAMMALIAN MUSCLE TISSUE, the level of contractile activity and the PO, of the blood perfusing the cells are known to be determinants of glucose uptake (2,10,17,19,24). It has also been demonstrated that both intense exercise and low Poe are accompanied by increased blood lactate concentration and changes in the cellular redox potential to a more reduced state (2,10). These observations have led to the hypothesis that hypoxia and contractile activity stimulate muscle cell glucose uptake by a single mechanism related to the energy state of the muscle cells (26).
When normal subjects exercise at a constant load, lactate accumulates in the blood only if the work rate is above a certain level (14,30). Breathing a low fraction of inspired O2 (Fro,) reduces the work rate threshold for blood lactate accumulation (12,13). Conversely, the threshold for lactate increase occurs at a higher work rate when the FIN, is high. This suggests that O2 availability can interact with work rate to affect glucose turnover.
The combined effect of hypoxia and contractile activity in exercising man is not known, i.e., the extent to which changing 02 availability to working muscle will affect muscle glucose uptake or hepatic production. The purpose of this investigation was to examine glucose kinetics (uptake and production) and regulation in exercising subjects breathing high and low concentrations of inspired 02, thereby changing the Pop of the blood perfusing the muscle cells, but without altering the work rate. Thus glucose turnover was measured at a single work rate under conditions of high and low Fro,. In each subject this work rate was chosen so that lactate concentrations were elevated during exercise when the FIN, was low, but lactate concentrations were lower when the FIN, was high.

Population
Ten male volunteers ranging in age from 19 to 37 yr in good health and without any prior history of chronic respiratory or other disease, comprised the study population. The group was heterogeneous in terms of fitness as judged by the VO 2max normalized to body weight, but none were in training as athletes at the time of the study. The subjects' age, weight, i70~~Jkg body weight, and specifically selected work rate used in the studies are shown in Table 1 (separated by a 2-wk interval). Since the measurements of glucose uptake and production requires relatively long periods of exercise, only moderate hypoxia (FIN, of 0.15) could be utilized. To maximize the difference in 02 availability to the working muscle, it was decided to compare the hypoxic protocol with exercise done at an FIEF of 0.80. The order of the high-and low-FIN, protocols was selected randomly.
The constant work rate protocols were performed in the morning after an overnight fast. There was a 120min tracer equilibration period before exercise to allow for a steady-state concentration of tritiated glucose. Twenty-five minutes before exercise, the subject began breathing the high-or low-o2 mixture. Blood sampling was performed every 5 min starting 20 min before exercise and continued throughout the 40 min of the exercise period.

Measurement of Gas Exchange Parameters
The subjects breathed through a low-impedance turbine volume transducer for measurement of inspiratory and expiratory volumes. Dead space of the mouthpiece and turbine device was 170 ml. Respired POT and PCO~ were determined by mass spectrometry from a sample drawn continuously from the mouthpiece at 1 ml/s and were displayed digitally. We could therefore verify that no leakage of room air gas occurred during the high-and low-Fro, studies. The electrical signals from these devices underwent analogue to digital conversion for the on-line breath-to-breath computation of O2 uptake (VO,, STPD), COz output (ho2, STPD), and expired ventilation (iTE, BTPS) as previously described (1). Heart rate was measured beat by beat using a modified standard lead I EKG for which three leads were placed on the chest. Saturation of hemoglobin with O2 was measured noninvasively by ear oximetry.
The VOz max was the highest vo2 achieved by the subject. Determination of the lactate threshold from gas exchange measurements is well described (30,32). It is based on the coupling of CO2 production and ventilation. When lactate concentration increases during exercise, the excess hydrogen ion is buffered by bicarbonate and COz is liberated. The increased CO2 production stimulates ventilation but not O2 uptake. Thus the threshold is measured by finding hyperventilation relative to Vo2, i.e., the VoB above which VE/VO~ and end-tidal Po2 increase without an increase in vE/iTC02 or a decrease in end-tidal Pco~. Moreover, in constant work rate protocols done above the threshold, Vo2 continues to increase between the 3rd and 6th min of exercise (30). This allowed an additional means to verify which subjects were above the threshold during the constant work rate protocol.

Tracer Methods
Primed, constant-infusion of [ 3-3H] glucose was used to measure glucose production (Ra), uptake (RJ, and metabolic clearance rate (MCR = Rd/[glucose]) (15). [3-3H]glucose, '5 &i/ml, was infused at a constant rate (0.086 ml/min) via an indwelling catheter in the antecubital vein. A priming dose, equivalent to the amount infused in 140 min, was given at the beginning of each experiment before the constant rate infusion. Ra and Rd were calculated using a "modified" single compartment for glucose as described by Steele et al. (22) for the nonsteady state. These calculations have been validated for the non-steady-state condition (16).
Another indwelling venous catheter was placed in the antecubital vein of the opposite arm. Both arms rested on platforms, thus, no gripping of handle bars occurred during the studies. Blood samples for the measurement of glucose and tracer were obtained every 5 min throughout the 20 min before exercise and the 40-min exercise session (5 base line, 8 exercise). Plasma for tracer glucose determination was deproteinized using BaOH and ZnSOd, evaporated, and redissolved in distilled water. Samples were counted for 100 min by liquid scintillation spectrometry. Plasma glucose concentration was determined by the glucose oxidase method.
Two base-line and five exercise measurements were made of lactate, pyruvate, immunoreactive insulin (IRI), glucagon (IRG), norepinephrine, and epinephrine. Lactate and pyruvate were measured by the method of Hohorst (9). Radioimmunoassays were used to measure IRI (8) and IRG (5). Catecholamines were measured by the enzymatic radioimmunoassay technique (18).

Data Analysis
The test statistics consisted of the means, for each subject, of the base-line (preexercise) and exercise measurements of the variables listed above. Analysis of variance (repeated measurements) was used to test the effect of the high and low FI o2 on the base-line values, and when ANOVA was significant, mean values were compared using a modified t test (27). To compare the effects of hyperoxic exercise with hypoxic exercise on glucose turnover and counterregulation, base-line to exercise differences in each variable (designated, for example, as ARJ were analyzed using paired t tests. Standard techniques of linear regression were used to assess changes in Vop and O2 saturation over the period of the exercise study. Statistical significance was taken at the P c 0.05 level. Unless otherwise stated, values are represented as means t SD.

Effect of FIEF on Lactate Threshold
In all 10 subjects, the lactate threshold occurred at a lower work rate and voz when the FIN, was 0.15 compared with 0.21 (mean decrease in Vo2 at the lactate threshold, 19%; Fig. 1). An additional increase was observed at FIN, of 0.80 compared with the 0.21 studies in 9 of the 10 subjects (mean increase, 15%).

Constant Work Rate Protocols
Gas exchange. Vo2 increased significantly with the duration of exercise between the 3rd and 6th min of hypoxic exercise (mean correlation coefficient r of 0.52) confirming, by gas exchange criteria, that the exercise done during hypoxia was above the threshold. The mean O2 saturation during hypoxic exercise was 91 t 1%. By contrast, during hyperoxic exercise, O2 saturation did not change and remained between 98 and 100% in all subjects.
Glucose kinetics. Base-line glucose concentration during hypoxia (mean, 96.4 t 8.1 mg/dl) did not differ from that measured during hyperoxia (92.4 t 7.9 mg/dl). Glucose concentrations did not change appreciably during hypoxic or hyperoxic exercise. Glucose concentration during hypoxic exercise was 97.6 t 5.1 mg/dl compared with hyperoxia, 97.4 t 8.6 mg/dl. Under hypoxic conditions, mean base-line values of Rd increased from 3.16 t 1.39 mg* min-' l kg-' to a mean of 5.43 & 2.13 mgemin-' l kg-' during exercise. In contrast, under hyperoxic conditions, mean base-line values of Rd increased from 3.52 t 1.15 mg. min-' l kg-' to a mean of 4.77 t 1.73 mg*min-l l kg-' during exercise. The increase in glucose uptake during hypoxic exercise was greater than during hyperoxic exercise in 8 of the 10 subjects. The average ARd during hypoxia was greater than during 0 1. Effect of FIEF on lactate (anaerobic) threshold in 10 normal subjects. FI o2 is on the x-axis and Voz, in ml OJmin, is on the y-axis. Vo2 corresponding to LT was greater in all '10 subjects at an FIN, of 0.21 compared with 0.15. A further but generally smaller increase occurred at an FIN, of 0.80 in 9 of 10 subjects. hyperoxia by 265%, P C 0.05 (Fig. 2). In addition, AMCR was significantly greater during hypoxic exercise (Fig. 2). The mean resting MCR during hypoxia was 2.94 t 1.17 ml. min-' l kg-' and increased to 5.66 t 2.23 mlmin-' l kg-' during hypoxic exercise. The mean resting MCR during hyperoxia was 4.13 t 1.17 ml l min-' l kg-' and was 4.84 t 1.63 ml. min-' l kg-' during hyperoxic exercise.
Mean base-line R, was higher during hyperoxia compared with hypoxia by 23% (3.90 t 0.95 compared with 3.00 t 1.28 mgemin-' . kg-'), but this difference was not statistically significant. AR, during hypoxic exercise was greater than during hyperoxic exercise in 9 of the 10 subjects. The average AR, during hypoxia was significantly greater than during hyperoxia by 188% (Fig. 2). The mean exercise R, was 5.56 t 2.13 mg=min-' l kg-' during hypoxia and 5.27 t 1.63 mgemin-' l kg-' during hyperoxia.
Lactate, pyruvate, and lactate-to-pyruvate ratios. There were no differences in the mean base-line lactate, pyruvate, or lactate-to-pyruvate ratios between the hypoxic and hyperoxic studies (Fig. 3). As expected the increases in lactate (Alactate) and pyruvate (Apyruvate) were significantly higher during hypoxic exercise. In addition,  2. Mean base-line to exercise increase in glucose uptake (A&), production (AR,), and metabolic clearance rate (AMCR) in response to exercise during hypoxic (hatched bars) and hyperoxic (dear bars) conditions in 10 healthy males. Hypoxia led to significantly greater increases in Rd, R,, and MCR during exercise than did hyperoxia even though work rate increment was same for both conditions. Shown here are means t SE. *Significant difference.  3. Lactate, pyruvate, and lactate-to-pyruvate ratio (L/P) in response to exercise during hypoxic and hyperoxic conditions. Clear bars represent mean t SE of base-line conditions. Hatched bars represent mean t SE of increase in each of these variables (Alactate, Apyruvate, AL/P). Alactate, Apyruvate, and AL/P were all significantly greater during exercise under hypoxic compared with hyperoxic conditions as indicated by asterisk.

E212
the increase in the ratio of lactate to pyruvate during exercise was significantly higher during hypoxia.
Insulin and glucagon. Mean IRI was lower during exercise than during the base-line period in both hypoxia and hyperoxia, but these differences were not significant ( Fig. 4). There were no significant differences between base-line glucagon in the hypoxic and hyperoxic studies. Similarly, during exercise the hypoxic values did not differ from the hyperoxic values. Epinephrine and norepinephrine.
There were no significant differences between base-line epinephrine in the hypoxic and hyperoxic studies (Fig. 5). However, the increase in epinephrine during hypoxic exercise was significantly greater than during hyperoxic exercise. Similarly, base-line values for norepinehrine did not differ between hypoxia and hyperoxia, but the norepinephrine increase over base line during hypoxic conditions was significantly greater than during hyperoxic conditions.

DISCUSSION
The results of this study demonstrate that exerciseinduced increases in glucose uptake and production in human subjects can be affected by the O2 concentration of the inspired gas. During hypoxia, the metabolic demand imposed by the moderate work rates led to significantly larger increases in R, and Rd than during hyperoxia (Fig. 2). The increased glucose uptake was accompanied by increased concentrations of lactate and pyruvate and larger lactate to pyruvate ratios (Fig. 3), suggesting a mechanism related to changes in the cellular redox state. In addition, our data indicate that the increased glucose production during hypoxia was mediated by elevated levels of catecholamines (Fig. 5). Glucose uptake can be directly affected by insulin, by blood flow to the tissues, and by the redox state of the cells. It is unlikely that insulin levels were directly responsible for the increased glucose uptake during hypoxic exercise, since the levels of insulin were the same for both hypoxic and hyperoxic exercise (Fig. 3). Muscle blood flow isknown to increase with exercise and can be further affected by FIN,. Welch et al. (31) showed that in exercising subjects breathing 100% 02 the increase in muscle blood flow was -10% less than in normoxia. Some investigators have hypothesized that glucose delivery to the muscle (the product of muscle blood flow and glucose concentration) may be an important modulator of muscle glucose uptake (21). Recently,however,Rennie et al. (19) demonstrated in vitro that the glucose transport rate is largely independent of the flow rate during muscular contraction. Thus small differences in blood flow that may have occurred in the present protocol appear unlikely to have been a major determinant of the increase in glucose uptake during hypoxic exercise.
The increase in both lactate and pyruvate was greater under hypoxic conditions as would be expected to result from more rapid glycolysis (mass action effect). But when glucose uptake in muscle tissue is induced by hypoxia, it is accompanied by an increase in the intracellular lactateto-pyruvate ratio (2,26). The lactate-to-pyruvate ratio is thought to reflect the cell redox state,' i.e., the ratio of NADH to NAD (33). During the 40 min of constant load exercise in the present studies, it is likely that serum values of lactate and pyruvate reflect general trends at the cellular level. Therefore, the elevated lactate-to-pyruvate ratios consistently observed during hypoxic compared with hyperoxic exercise probably indicate a more reduced muscle cell redox state.
The data presented here are consistent with other in vitro and in vivo studies demonstrating that tissue hypoxia can be a critical determinant of glucose uptake. Experiments done in electrically stimulated, perfused rat hindlimb preparations have shown that hypoxic perfusates increase glucose uptake (10, 26). When O2 availability to exercising muscle is limited by anemia, glucose uptake and production are significantly increased, as recently demonstrated in exercising dogs (29). Finally, hypoxic exercise has been demonstrated to increase splanchnic glucose production in humans (20).
Base-line levels of norepinephrine and epinephrine were unaffected by the gas mixture (Fig. 5). Consistent with previous studies (4), both substances increased during exercise and the increases were greater in the hypoxic state indicating additional adrenergic drive. Although catecholamines have been shown to inhibit glucose uptake by muscle tissue (29, in the present study, glucose uptake was elevated during hypoxic exercise at the same time that catecholamines were increased. The stimulus for glucose uptake in hypoxic exercise appears to be strong enough to mitigate possible inhibition by catecholamines. Hepatic glucose production was also elevated during hypoxic compared with hyperoxic exercise (Fig. 2). While glucagon has been shown to be an important regulator of glucose production during exercise (6,11,23,28), the levels of IRG during exercise in the present study were unchanged by the FI oz. Catecholamines have been shown to stimulate R, even in the absence of changes in the pancreatic hormones (3,7); thus, the exaggerated catecholamine response observed in hypoxic exercise may have been responsible for the generally greater rate of hepatic glucose production observed during hypoxic exercise. Alternatively, hypoxia may have a direct effect on hepatic glucose production.
The results of this investigation emphasize the robust nature of the homeostatic mechanisms for glucose regulation. The stimulus of hypoxia significantly increased glucose uptake during moderate exercise, but also resulted in an increased glucose production so that blood glucose concentration was maintained within a narrow, presumably "optimal" range. In diseases of the heart, lungs, and blood vessels, reduced tissue O2 availability is commonly encountered with increased lactate concentration. The finding of a normal blood glucose in these conditions may conceal significant alterations in substrate utilization. D. M. Cooper is a Clinician-Scientist of the American Heart Association, Greater Los Angeles Affiliate. D. Wasserman was an Edward Christie Postdoctoral Fellow at the University of Toronto in the Department of Physiology.
This work was supported in part by a Research Feasibility grant of the American Diabetes Association.