INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

FOR THE PAST DECADE, we have been examining the effects of hypobaric hypoxia on regional cerebral glucose metabolic rates (rCMR_glc) in humans. In our first study, we assessed rCMR_glc in Quechuas living at high altitude in the Andes (11). On their arrival at sea level, rCMR_glc was depressed relative to rCMR_glc in lowlanders. Reductions were most evident in frontal cortex and the angular gyrus. After acclimating for 3 wk to sea level normoxia, rCMR_glc changed modestly but still did not fall within the lowlander range. Later studies with Sherpas from the Himalayas, a group living at high altitude for a longer period than the Quechuas, showed regional patterns and rCMR_glc similar to lowland normals. Moreover, these values of rCMR_glc were unchanged after 3 wk of low-altitude acclimation (13). Our rCMR_glc findings were consistent with medical studies of the Sherpas, reporting an absence of symptoms that are typically associated with hypobaric hypoxia exposure in lowlanders (5, 6). Therefore we interpreted these data to indicate that, probably due to longer time, the Sherpas expressed a more complete metabolic adaptation than the Quechuas. If this hypothesis is correct, one would anticipate that lowlanders exposed to high altitudes for an extended time period may exhibit major changes in brain metabolism.

The opportunity to test this hypothesis arose when a US Marine training center in California agreed to provide six volunteers before and after a 63-day high-altitude training program. Because we were relying on the largess of the US Marine Corp, we could not control for confounding factors such as nutrition, training regimen, and travel stress. However, given the logistics, ethical concerns, and expense of such a study, we decided to take advantage of the opportunity. We found that after the prolonged high-altitude exposure the brain's metabolic patterns changed significantly, with five regions having significant decreases in metabolism and two regions having significant increases in rCMR_glc. Although we used each individual as his own control (pretraining vs. posttraining comparisons) and were not able to include a control group who experienced similar conditions with the exception of high-altitude exposure, the magnitude of the changes and the bidirectionality are completely inconsistent with changes associated with test/retest studies of rCMR_glc (34). Thus we tentatively conclude that the effects that we observed are largely or solely due to prolonged exposure to hypobaric hypoxia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. The subjects for the present study were six male US marines of Caucasian (European) ancestry stationed at a training center in California (altitude ~6,000 ft). Their mean age was 29.7 yr (SD = 3.6 yr, range 23.0-33.0). Before the program started, all six subjects were exercise tested using bicycle ergometry (Table 1). The group then underwent an extensive training period for high-altitude operations at three different sites, ranging from ~10,500 to 14,400 ft. After this training, they successfully reached the 20,320-ft summit of Denali (also known as Mt. McKinley, AK) over a 16-day period. The total time of high-altitude training and scaling the summit of Denali was 63 days. Brain metabolic features of all subjects were assessed before training and 1 day after the descent of Denali. For these positron emission tomography (PET) assessments, the subjects were admitted as inpatients to the University of Wisconsin General Clinical Research Center. Medical and neurological examinations were negative before and after high-altitude exposure. On completion of the brain PET studies, a number of exercise parameters were again determined and none were found to have changed drastically throughout the 63-day program. Most significantly, the average body weight was unchanged at ~80 kg, and the maximum aerobic metabolic rate (O_{2 max}) tended to increase, but the values also were statistically similar before and after the high-altitude program (~50 ml O₂ · kg¹ · min¹) (Table 1). The O_{2 max} values in fact are similar to those for other mountaineers (16, 20). All experimental procedures were approved by the Human Ethics Committees at the Universities of British Columbia and Wisconsin. Informed consent was obtained for all nonmilitary aspects of this study.

PET scanning. After an overnight fast, a venous cannula was placed in the left antecubital vein and a right hand vein. The right hand was placed in a warming device to induce hyperemia. Blood samples taken from this hand were considered to be adequately representative of arterial blood. [¹⁸F]-2-deoxy-2-fluro-D-glucose (FDG; 7 mCi) was injected into the antecubital cannula as a bolus, and blood samples were taken at a progressively decreasing rate from the right hand over the next 70 min. A dynamic study of myocardial uptake (data not shown) was performed during the next 46-min uptake period. The subject was then positioned for brain imaging in a GE advanced positron tomograph (GE Medical Systems, Waukesha, WI) at the Wisconsin PET Imaging Center. The 14.4-cm axial field of view was positioned to encompass the entire cerebrum and cerebellum. Image data were acquired for 30 min in conventional 2-D mode followed by a 10-min transmission scan using the postemission mode offered by the manufacturer. Transmission data were used to correct the emission data for attenuation. The time course of radioactivity in plasma was derived from the blood sample sequence using conventional counting techniques. Additional blood samples were used to determine the plasma glucose concentrations at three time points over the course of the procedure. Plasma metabolite concentrations were determined using standard clinical chemistry laboratory methods. Plasma glucose concentrations were stable under both study conditions, 5.13 and 5.03 µmol/ml of plasma, respectively.

The acquired maps of radioactivity concentration were converted pixel by pixel to glucose uptake rates using the Sokoloff et al. equation (33) as modified by Huang et al. (19). Standard gray matter rate constants (31) and a lumped constant of 0.48 were used. Regional estimates for 25 cortical and subcortical areas were derived by visually placing circular regions of interest (ROI; area = 1.1 cm² in a single axial plane; 4.25 mm axial plane separation) to conform to a standard PET template of neuroanatomical structures. The rCMR_glc estimates for the 25 ROIs were averaged over subjects. The changes in mean metabolic rate for each region pre- and posttraining were compared using a paired t-test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The means, SD, and resulting t-values for the 25 ROIs are given in Table 2, with illustrative PET images presented in Fig. 1. For five regions, significant decreases in rCMR_glc were found, whereas for two regions, the right and left cerebellum, significant increases were observed. The rCMR_glc values for most other regions seemed the same in pre- and postacclimatization states. The total number of significant differences (7 of 25) significantly exceeded chance expectancy, whereas the fact that some regions increased while others decreased means that the findings are not the result of statistical artifact or scalar drift. For the five regions where decreases were observed, three were in frontal cortex [average change 11.1, 7.4, and 10.5%, respectively (Table 2; Fig. 1)]. In addition, the left occipital cortex sustained an average decrease in rCMR_glc of 9.0%, whereas the right thalamus sustained a 16.9% decrease. The left thalamus also displayed a decrease in rCMR_glc, but this change only approached the accepted statistical significance level (P < 0.1). In contrast to the other brain regions, the cerebellum sustained a significant increase in rCMR_glc. For the right and left cerebellum, the increases were 17.8 and 16.6%, respectively (Table 2).

View larger version (61K): [in this window] [in a new window]   Fig. 1.   Representative positron emission tomography (PET) images of regional cerebral glucose metabolic rate (rCMRglc) in an individual subject pre- and postacclimation to high-altitude regimen. Color coding (red through yellow to green) covers an rCMRglc range from ~0.6 to 0.3 µmol glucose · g1 · min1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One of the main problems with many studies of humans at high altitude is that it is frequently impossible to isolate hypobaric hypoxia as the only variable parameter. In our case, for example, we were unable to rigorously control nutritional intake, exercise training regimen, and, in final stages of the program, any metabolic artifacts due to air travel. However we feel that these factors did not contribute to observed pre- and postexposure brain metabolic differences for several reasons. In the first place, it is worth emphasizing that air travel preceded both sets of PET measurements and moreover that all subjects underwent medical examination on arrival and on departure from the General Clinical Research Center. No problems or differences between the first and second visits were noted and an impression of fit individuals, all in good health, predominated our discussion with the medical personnel performing these examinations. Second, during the actual 63-day training for altitude operations these subjects were "on duty" in a professional organization with an interest in maintaining factors such as fitness, physical training regimens, and nutritional status at constant high levels. That is why we anticipated, and indeed found, that there were essentially no statistically significant changes in several physical and physiological parameters measured before and after the 63-day high-altitude program (Table 1). When measured at sea level, there were modest or no effects on body weight, blood pressure, maximum exercise-induced heart rate, maximum respiration rate, tidal volume, or maximum ventilation. The modestly higher O_{2 max} and exercise times to fatigue observed after the 63-day high-altitude program are consistent with the work of others on high-altitude effects on exercise (see Refs. 16 and 20 for more literature on hypoxia exposure and exercise capacities). These data persuade us that parameters such as nutrition and exercise training regimens did not introduce significant artifacts to our data. Thus we assume that, as in many other studies of this type (16, 20, 28, 29), the dominant environmental parameter that differed in the two states we examined was the long exposure to hypobaric hypoxia. Any effects of the 63-day program presumably would be predominantly due to hypobaric hypoxia. We therefore tentatively conclude that in lowlanders of Caucasian lineages the acclimation response of the central nervous system to prolonged exposure to high-altitude hypoxia is complex and includes stabilizing rCMR_glc in many regions of the brain, decreasing rCMR_glc in five regions of the brain (the frontal cortex, the left occipital cortex, and the thalamus), and increasing rCMR_glc in a few other specific regions (especially the cerebellum).

Unfortunately there are few studies of hypoxia acclimation effects on mammalian brain against which to compare these results. In rats, hypoxia acclimation has been shown to lead to a global increase in glucose uptake and metabolism (7, 8, 23), but the metabolic data from the rat studies were collected while the animals were in the hypoxic state. Although there was a 20% increase in rCMR_glc during the sustained hypoxia (e.g., Ref. 8), there could well have been a decrease in rCMR_glc when the animals were returned to normoxia for the FDG uptake measurements. In fact there are indirect indications of downregulation of cerebral energy metabolism in rats after hypoxia acclimation. For example, hypoxia acclimation is associated with depressed cytochrome oxidase activity in both rat (22) and mouse (2) brain, and decreased neutrophil mitochondrial volume densities have been noted (35). LaManna and coworkers (21, 24) also have proposed that small changes in rCMR_glc may represent changes in the ratio of glycolytic to oxidative pathways that serve to balance tissue acid-base balance induced by the ventilatory response to hypoxia. The hypothesis proposes that ventilatory-induced mild alkalosis is reversed by glycolysis-associated H⁺ production to restore tissue intracellular pH balance. This hypothesis probably does not apply to our data in this study, because our PET measurements were done under normoxic conditions with no hyperventilation-induced alkalosis, nor to our knowledge are there any other data that would bear directly on this issue in human brain metabolism. In fact, given the large interest in human hypoxia responses, the dearth of studies of hypoxia effects on rCMR_glc in humans is surprising. To be sure, Cohen et al. (4) did examine the effects of hypoxia on total brain metabolic rates, but perhaps because their work predated the advent of the PET/FDG technique, they found no differences in hypoxic vs. normoxic states. However, they posited the possibility of regional adjustments in rCMR_glc. Interestingly our findings are consistent with their speculation. Other studies of region-specific metabolic rates in the brain of humans have been directed to issues other than hypoxia acclimation (for examples, see Refs. 5, 6, 19, 31, 34).

An important issue that requires explanation is the magnitude of rCMR_glc differences, both region specific and due to high-altitude acclimatization. If we assume that essentially all of the glucose metabolized is fully oxidized, yielding 36 mol ATP/mol glucose, then in region-by-region comparisons, the left frontal cortex displays the highest metabolic rates noted (23.0 µmol ATP · g¹ · min¹), whereas the lowest rCMR_glc was observed in the left cerebellum (14.8 µmol ATP · g¹ · min¹). After high-altitude acclimatization, the metabolic rate in the left frontal cortex decreased from 23.0 to 19.1 µmol ATP · g¹ · min¹, whereas the left cerebellum increased from 14.8 to 17.6 µmol ATP · g¹ · min¹. To remain in energy balance, ATP demand or energetic efficiency (or a combination of both) presumably had to change by equivalent amounts. These are substantial shifts in energy supply; for reference, it may be worth mentioning that resting muscle turns over ATP at ~1 µmol ATP · g¹ · min¹ (16, 26).

With completion of these studies, firm databases are now established on human brain metabolic organization in three different human lineages, each varying in historical adaptation to hypoxia. Comparative biochemists have long appreciated the interplay between evolutionary time and adaptational options: the longer the time available the more profound or complete the adaptive response (9, 10, 14, 20). Thus it is not surprising that many workers in this field explicitly or implicitly consider Sherpas to be most exquisitely high-altitude adapted of all human lineages, they have been indigenous to high altitude for millennia (12, 14, 28, 29). The Quechuas are presumed to have been living in high-altitude environments for intermediate time periods (28, 29), and of course the lowlanders used in the present study represent about a 9-wk acclimatization to hypobaric hypoxia. A recent analysis of the evolution of hypoxia tolerance in our species (14) points out that the last time the Quechuas and Sherpas shared a common ancestor was a very long time ago, about one-third the age of our species, and the last time Europeans shared ancestors with both highlander groups was even further back in our phylogenetic history. In terms of rCMR_glc, Sherpas appear to show the fullest degree or most complete form of metabolic adaptation, i.e., comparable to normoxic lowlanders (13). The dominant brain metabolic defense strategy in the Quechuas (11) appears to be a mild suppression of rCMR_glc, presumably coincident with mild suppression of various energy-demanding processes (9, 10, 15, 25, 32). The lowlanders in the present study seem to represent an intermediate pattern of defense against chronic hypoxia, with rCMR_glc increasing, decreasing, or unchanging, depending on the brain region being interrogated. Parenthetically, it should be noted that the two earlier studies of Sherpas and Quechuas were done on a different PET scanner than the current study; although comparisons of relative patterns within each study series are valid, direct comparisons of absolute rates between the two groups of studies are not.

Perspectives

To evolutionary biologists familiar with numerous molecular and metabolic defense mechanisms that have evolved in hypoxia-tolerant species through phylogenetic time to protect cells and tissues against oxygen limitation (9, 10, 14, 15, 25, 32), the issues are somewhat different. To these workers, it would be particularly interesting if the observed changes in metabolic organization of the human brain after hypoxia acclimatization of lowlanders were in any way advantageous or "protective" against hypoxia. Unfortunately our data cannot be simply interpreted this way. Indeed, if our above speculation regarding frontal cortex function is correct, any such biologically protective mechanisms may not be behaviorally adaptive in the short run. Although further work is required to probe the issue of adaptive significance of rCMR_glc adjustments during prolonged hypoxia exposure, it is tempting to consider that it is the phylogenetically "older" portion of the human brain, the so-called mammalian/reptilian brain, not the human species-specific neocortex (frontal lobe), that may be the region preferentially "protected" during hypoxia acclimation, because it is the region that sustains an elevated rCMR_glc during week-long hypoxic exposure. That is, the coordination of basic biological functions such as locomotor abilities may supercede the more advanced neurological functions normally ascribed to the frontal cortex.

In any event, the fundamental insight that derives from these studies is that the brain of healthy human lowlanders responds to chronic hypoxia exposure with precise, region-specific fine tuning of rCMR_glc: downregulation occurs in the frontal cortical regions, whereas upregulation occurs in the cerebellum. Most other regions sustain stable metabolism even after prolonged acclimation to hypobaric hypoxia. Interestingly, rather striking region-specific metabolic responses to acute hypoxia in the rat brain already are known (27), but to our knowledge no one has yet applied late 20th century technology to the region-by-region monitoring of metabolic responses to acute hypoxia in the human brain.