INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

SOME STUDIES ON MUSCULAR FATIGUE have shown that glucose supplementation can delay fatigue induced by physical exercise or pointed out that supplementing carbohydrates during exercise significantly lengthened its duration (10, 13, 19).

Early research by Levine et al. (26) in the 1920's established the link between hypoglycemia and extreme fatigue. Later, work performed in Scandinavia (6, 34) revealed the existence of a link between fatigue and peripheral glycogen utilization. However, all fatigue-related phenomena cannot be explained by this sole observation. In 1983, Coyle et al. (11) showed that, in many cases, ingested glucose cannot diminish glycogen utilization during exercise. Experiments were carried out by Koslowski et al. (25) in 1981 in dogs in which glucose was directly infused in the carotid artery. Results showed that supplying glucose to the brain could delay fatigue, pointing on an action of glucose supplementation on central fatigue.

Mechanisms and kinetics of glucose transport across the blood-brain barrier (BBB) and inside the brain are well known (12, 18, 27). They involve glucose transporters of the GLUT family; GLUT-1 is exclusively involved along the BBB, whereas, inside the brain, transporters are mainly GLUT-1 and GLUT-3, both on neural cell membranes (9, 14, 20, 38). However, an interesting question remains concerning glucose regulation and transport between the peripheral compartment and brain extra- and intracellular compartments in some physiological situations, such as physical exercise and fasting.

In this study, we first focused on the relationship between glycemic variations and cerebral glucose variations after exercise and fasting, where glucose deficits could influence several central regulations. Then we studied the influence of a subsequent glucose supplementation. We aimed, first, at determining whether or not exercise could specifically modify cerebral glucose regulation, compared with fasting, and, second, at determining to what extent postexercise glucose supplementation could influence this regulation process. To this end, we used an original technique developed in our laboratory (32), associating brain microdialysis to determine glucose concentrations in the extracellular compartment (5) and NMR spectroscopy to monitor extra- and intracellular compartments in the brain (4, 15).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Male Wistar rats weighing 250-300 g from Janvier (Le Genest-Saint-Isle, France) were used in this experiment. They were individually housed in a temperature-controlled room (20 ± 1°C) illuminated from 8:00 AM to 8:00 PM. Food and water were available ad libitum (except during the food-deprivation experiment).

Surgery. Eight to ten days before the experiment, the animals were anesthetized with ketamine chlorydrate (150 mg/kg ip) and then placed in a stereotaxic frame (DKI model 900). Head skin was incised, temporal muscles were retracted on both sides of the head, and the cranium was cleaned of all organic material with a hydrogen peroxide solution. A 1-mm-diameter hole was drilled in the parietal bone, and a microdialysis stainless guide probe (Carnegie Medicin, Stockholm, Sweden) was placed horizontally in the frontoparietal cortex, after removal of dura matter, and sealed with dental cement. The coordinates used, according to the Paxinos and Watson atlas (30), were 2.6 mm behind the bregma and 1.5 mm below the horizontal zero plane. A triangular head-holding device was then chronically implanted to house the NMR coil and to allow the head to be attached to the NMR probe. This system and the microdialysis guide were covered by a polyethylene protection.

1

1

Physiological monitoring. The femoral artery catheter was used to monitor mean arterial blood pressure, PO₂, PCO₂, and pH. Respiratory rate and volume were adjusted (volume = 2.5 ml; frequency = 70 counts/min) to keep arterial pH at 7.35 ± 0.05, PCO₂ at 38 ± 3 Torr, and PO₂ at 108 ± 3 Torr. The electrocardiogram was monitored, and body temperature was maintained close to 38 ± 0.5°C throughout the experiment.

Microdialysis procedure. The microdialysis membrane was 4 mm long, with a 500-µM diameter and a molecular cutoff weight of 20 (CMA 20 microdialysis probe model, Carnegie Medicin). In vitro calibration tests completed before each experiment showed a relative recovery of ~14% for glucose.

Dialysates analysis. The dialysates were analyzed by reverse-phase HPLC with fluorometric detection. The system consisted of a 10-µl sample loop leading to an Eco-Cart Lichrospher RP-18 (5-µm) column. The mobile phase consisted of a 0.2 M Tris buffer (pH = 8) with EDTA. The enzymatic reagent was a hexokinase plus glucose-6-phosphate dehydrogenase combination. Detection was done at a 260-nm excitation wavelength and a 460-nm emission wavelength with a Jasco FP-920 fluorometric detector.

NMR spectroscopy. NMR experiments were carried out by using an AM 400 Wide Bore Bruker spectrometer (Oxford Instruments) equipped with a 9.4-T, 89-mm-bore vertical magnet and a custom-made probe for in vivo studies. A one-turn elliptical surface radio-frequency coil (11 × 8 mm) was used to acquire the ¹H-NMR signal. Two-dimensional (2D) correlation spectroscopy (COSY) spectra of glucose and glutamate/glutamine pool were obtained in module mode by using the SUPERCOSY sequence as previously described (2, 31). Briefly, the SUPERCOSY pulse scheme had two spin-echo delays that were introduced symmetrically with respect to the second /2 pulse. The delay might be adjusted for a specific range of scalar coupling constant J. This delay was set to 86 ms for 2D COSY ¹H spectra of small cerebral metabolites. It corresponded to 2 × 0.3/J, with J = 7 Hz (mean scalar coupling value of freely rotating rotors in open chains). The delay also allowed further reduction of the remaining water signal and suppression of phospholipid and protein signals by T2 filtering. The quadrature in dimension 1 was obtained by phase modulation. Taking the T2 relaxation of small metabolites into account, we used 86 ms for the spin-echo time. This delay also enabled a further reduction of the remaining water signal and the suppression through T2 filtering of the phospholipid and protein signal. Acquisition of eight free induction decays in the time domain T2 for each of the 128 increments in the time domain T1 led to a total acquisition time of 30 min. The spectral width in the corresponding frequency domain in F1 and F2 dimensions was 10 ppm. Before the 2D Fourier transform, the data were multiplied by an unshifted sine bell function in the two dimensions and were zero filled to obtain a 1,024 × 1,024 data point matrix. In vivo 2D chemical shifts were expressed in comparison to water resonance, assigned at 4.75 ppm in the two dimensions. Cerebral metabolite levels were measured on the 2D spectra according to the volume of their respective cross-correlation peaks by means of a specific software (Aurelia).

Measuring glycemia. Glycemia was measured by means of the peroxydase method (glucose enzymatique color kit, BIOTROL) with a spectrophotometric detection at 500 nm.

Sampling CSF. A direct sampling of CSF was used to determine the real concentration of glucose in the extracellular compartment. Sampling was through a puncture within the third ventricle, thus allowing for 100-µl samples of CSF, analyzed on line with HPLC (n = 5).

Experimental protocol. A first group of 10 rats was used to determine the values of glycemia and extracellular brain glucose in normal feeding and rest conditions. Then two groups of eight rats were randomized: the first group was deprived of food for 36 h, and the second was submitted to strenuous and prolonged exercise on a treadmill (running for 1 h 30 min at 30 m/min, 0% grade) just before the experiment, after 1 wk of training. Animals underwent the surgical procedure described in Surgery above after running or fasting, when equilibration of the dialysis membrane occurs. Dialysates and NMR data were simultaneously obtained every 30 min, with the first dialysate being sampled after the probe was equilibrated for 2 h. The protocol for glucose infusion was as follows: 30% glucose constantly infused for 2 h at a rate of 0.3 ml · h¹ · 100 g¹. Data acquisition was done postexercise, for 4 h 30 min, with a stable state of 1 h 30 min, a glucose infusion of 2 h, and then a return to stable state for 1 h. Blood and cerebral glucose were analyzed after sampling; dialysates were then kept frozen at 80°C for later analysis.

Data analysis. Data concerning glycemic variations and extracellular glucose level comparisons are expressed as means ± SE of concentrations. Data concerning the effect of infusion are expressed as means ± SE of percent increase or decrease vs. concentrations obtained before infusion. Percentage was used instead of concentration to allow a direct comparison of glucose variations in the three compartments.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of food deprivation, exercise, and glucose infusion on blood glucose metabolism. A mean value for glycemia in appropriate feeding and rest conditions was obtained from the group of rats kept under these conditions. This value was 10.07 ± 0.24 mM under anesthesia (n = 10). This group of rats was used as a control group throughout the study.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Variations in venous blood glycemia after fasting (n = 10 rats) and exercise (n = 10 rats) with glucose infusion between 0 and 120 min, compared with normoglycemic (N) value. * P < 0.05, ** P < 0.01, *** P < 0.001, compared with N value.

Effect of hypoglycemia induced by food deprivation and exercise on cerebral glucose metabolism. Figure 2 shows the glucose concentration in brain microdialysates for the two hypoglycemic situations vs. the normoglycemic one. Cerebral glucose decreased in line with the drop in glycemia after food deprivation (139.5 ± 9.1 µM; n = 8) but showed an unexpected high level after the treadmill workout (273.6 ± 22.1 µM; n = 8), compared with the normoglycemic situation (203.9 ± 18.7 µM; n = 10) (P < 0.05).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Glucose concentration in dialysates before glucose infusion in situation of fasting-induced hypoglycemia (after fasting; n = 8 rats) or exercise-induced hypoglycemia (after exercise; n = 8 rats), compared with control (n = 10 rats). * P < 0.05 compared with control; ### P < 0.001, after fasting vs. after exercise.

Effect of glucose infusion on cerebral glucose metabolism. In fasting conditions, brain glucose levels increased during infusion, in line with variations in glycemia, and reached a maximum of 342 ± 28% for total brain glucose, as detected by NMR, and 263 ± 20% for extracellular brain glucose, as noted by microdialysis (P < 0.001), compared with baseline. A decrease was then observed during all postinfusion periods in both compartments (Fig. 3A).

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Percent variations in glucose in the 2 brain compartments, compared with variations in glycemia (in mM) after fasting (n = 8 rats; A) and after exercise (n = 8 rats; B). Glucose infusion was between 0 and 120 min. C, control. * P < 0.05, ** P < 0.01, *** P < 0.001, compared with 0 time point.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Glucose infused after strenuous physical exercise or fasting resulted in significant hyperglycemia. However, the glucose increase was slightly greater after food deprivation than after exercise. This can be explained by an enhanced muscle glucose transfer across muscle membrane, as previously demonstrated (24, 33), caused by the phenomenon of nonsuppressive insulin-like activity and by enhanced sensitivity to insulin during and after exercise. This result highlights a metabolic difference existing between exercise and fasting in terms of peripheral glucose regulation.

Some studies have previously described how glucose is transferred between blood and brain (12, 27), but the data presented here focus on the difference between blood and brain glucose metabolism after fasting and physical exercise. Results confirmed the existence of a direct link between glycemia variations and cerebral glucose concentrations. This means that, under such physiological conditions of glycemia, there is no limitation in glucose transport between the peripheral compartment and the brain across the BBB (18). During infusion, extra- and intracellular cerebral glucose increased more significantly in conditions of fasting than in conditions of exercise. Long-term regulation mechanisms might come into play: evidence was found that the number of GLUT-1 glucose transporters on the BBB (7) and of GLUT-3 transporters on the neuron membrane (36) were inversely proportional to blood glucose concentration. A 36-h fast would cause an increase in the quantity of GLUT-1 transporters on the BBB (7), contrary to what 1 h 30 min of running would do, where hypoglycemia is very brief. Kinetic data, which show faster cerebral glucose increases in the food deprivation situation, strongly support this hypothesis.

Coupling NMR spectroscopy with microdialysis allowed for the separation of data pertaining to intra- and extracellular compartments. Indeed, NMR addresses mainly both compartments, as the microvascular compartment represents a very slight percentage of the NMR volume and can be ignored and microdialysis only addresses the extracellular compartment. The data showed a greater increase inside cells, ~2.5 times the increase in CSF, as calculated above. This confirms that, when the cerebral glucose concentration increases, neurons are privileged compared with the rest of the brain. This, in turn, is due to the characteristic differences between GLUT-1 and GLUT-3, the latter being in charge of transferring glucose across the neuron membrane. As a matter of fact, the Michaelis-Menten constant of GLUT-3 is three times smaller than that of GLUT-1; therefore, when glucose increases in blood, it crosses the BBB by the exclusive means of GLUT-1 and is then redirected toward nervous system cells. The transfer across the BBB was noted as more important in the case of fasting. However, figures seemed to not be significant for cell membrane crossing. This tends to show that the long-term transporter regulation obtained for GLUT-1 does not apply to GLUT-3 in our fasting protocol, unlike results obtained in a previous study on the effect of chronic hypoglycemia on GLUT-3 (36). In other words, it seems that a slight and long-enough hypoglycemia is sufficient to induce GLUT-1 protein in rat brain neurons but not GLUT-3 protein. Our study did not entail the measure of GLUT transporters, but the way they operate was reflected and thus observed by the differences in glucose content in the two brain compartments during infusion. These results all seem to indicate that the effects of peripheral hyperglycemia are regulated in the brain by the glucose transport system across the BBB and the cell membranes, involving the transporters of the GLUT family. Despite previous observations of a speedy effect of glucose supplementation on the parameters of physical exercise (29), the results presented herein evidence a latency between the start of glucose infusion and the increase in cerebral glucose levels. However, our glucose supplementation occurred only after exercise, and the situation could be different when supplementation occurs before or during exercise.

Our most unexpected results involved brain glucose regulation in hypoglycemic situations. A decrease in brain glucose levels was expected in the two experimental conditions, but the extracellular concentrations noted after exercise were far more superior than those observed in the normoglycemic state or after fasting. Two main hypotheses may account for these results. First, exercise is known to require a reorganization of the vasodilatation/vasoconstriction system and an increase in heart beat rate, leading to a significant increase in cerebral blood flow (21, 23). This might explain the increase in cerebral glucose levels despite the peripheral decrease. Second, exercise is known to induce a significant rise in glucocorticoids, which are strongly involved in cerebral glycogen catabolism in brain astrocytes (17, 30). Exercise is likely to activate this metabolic cascade, thereby increasing the quantity of extracellular glucose. Moreover, fasting decreases glucocorticoid levels (22), and the differences observed between exercise and fasting results may be enhanced by these metabolic variations. Both hypotheses may partly explain the phenomenon, which could be a mechanism protecting brain tissue during exercise and/or the reflection of an induced central fatigue expressed through a link with some neurotransmitters.

Hypoglycemia is known to be a peripheral fatigue marker, reflecting the major role of glucose metabolism in physical performance. The results presented herein show that cerebral glucose could be a central fatigue index as well, probably related to another central intermediary such as serotonin (8, 37), GABA (1), or another neurotransmitter more directly involved than glucose in the appearance of fatigue.