|Home | About | Journals | Submit | Contact Us | Français|
To identify interventions for brain aging, we must first identify the processes in which we hope to intervene. Brain aging is a period of decreasing functional capacity and increasing vulnerability, which reflect a reduction in morphological organization and perhaps degeneration. Since life is ultimately dependent upon the ability to maintain cellular organization through metabolism, this review explores evidence for a decline in neural metabolic support during aging, which includes a reduction in whole brain cerebral blood flow, and cellular metabolic capacity. Capillary density may also decrease with age, although the results are less clear. Exercise may be a highly effective intervention for brain aging, because it improves the cardiovascular system as a whole, and increases regional capillary density and neuronal metabolic capacity. Although the evidence is strongest for motor regions, more work may yield additional evidence for exercise-related improvement in metabolic support in non-motor regions. The protective effects of exercise may be specific to brain region and the type of insult. For example, exercise protects striatal cells from ischemia, but it produces mixed results after hippocampal seizures. Exercise can improve metabolic support and bioenergetic capacity in adult animals, but it remains to be determined whether it has similar effects in aging animals. What is clear is that exercise can influence the multiple levels of support necessary for maintaining optimal neuronal function, which is unique among proposed interventions for aging.
To intervene in brain aging, we must first identify the primary factors that underlie it. Brain aging is a period of decreasing functional capacity and increasing vulnerability, which reflect a reduction in morphological organization and perhaps degenerative changes. The termination of this period is death, which is defined as a point at which cell degradation is irreversible. In contrast to death, life is dependent upon metabolism for the reversal of movement toward disorder (Schrodinger, 1944). Through this lens, brain aging is a period in which neuronal metabolism may decline, leading to a reduction in function and anatomical integrity. Therefore, interventions directed at maintenance and restoration of neuronal metabolic capacity may target fundamental features of cellular aging. Here, we propose that exercise is one of the most effective interventions for aging, because it can simultaneously slow age-related decline in neuronal metabolic capacity and maintain cardiovascular support for neuronal metabolism. We will begin with a discussion of the relatively high metabolic demand of neurons and the fundamental importance of maintaining metabolic capacity for optimal neuronal function. We then review evidence for an age-related decline in both neuronal cardiovascular support and cellular metabolism, each followed by a review of evidence that exercise enhances these processes. We briefly review experiments conducted to test whether exercise can protect the brain from metabolic challenges that are more frequent during aging. We end by discussing the gaps in research, and the need for more direct tests of the effects of exercise on these systems in aging animals.
In the brain, the demand for oxygen and glucose to drive cellular metabolism necessary for the maintenance and restoration of ionic gradients is ever present, but the function of these processes diminishes with age. And yet, the ability to continually reverse cellular entropy (movement toward disorder) is the fundamental process that maintains and prolongs the organized systems we consider as living (Schrodinger, 1944). Failure of this process causes death, first of cells, then organs, and eventually, the life of the individual. The end point is particularly notable for neurons, because neurons store information that is required over a lifetime. Paradoxically, their information processing and storage function underlie a high demand for energy that makes neurons particularly vulnerable to energy deficits and impaired metabolic support. Neurons’ high demand for energy is driven by the need to maintain ionic gradients across the large membrane surface area of dendrites. In addition to the large membrane surface area of neurons, synaptic action across the dendritic membrane continually degrades the ionic gradients. As a result, the brain uses 20% of the oxygen and 25% of glucose consumed despite making up only 2% of body weight. Oxygen and glucose are used to produce ATP through oxidative phosphorylation, a process that accounts for 95% of ATP production in the brain (Erecinska and Silver, 1989). In the anesthetized brain 40–60% of ATP is used by the sodium potassium pump to restore ionic gradients (Astrup et al., 1981; Erecinska and Silver, 1989).
There is a strong relationship between metabolic demand and the density of excitatory synapses over space and over developmental stages (Kageyama and Wong-Riley, 1982; Mjaatvedt and Wong-Riley, 1988). ATP production is coupled to the activity of cytochrome oxidase (Wong-Riley, 1989), for which a chemical reaction allows visualization of cellular energy consumption across nuclei, and across cellular compartments (Kageyama and Wong-Riley, 1982; Mjaatvedt and Wong-Riley, 1988). Within dendritic regions, which make up 78–86% of gray matter tissue volume in the hippocampus (Sousa et al., 1998; Tata et al., 2002; Tata and Anderson, 2010), the layers with small dendritic processes and higher synaptic density have higher metabolic demand than areas with large dendrites and smaller synaptic density (Kageyama and Wong-Riley, 1982; Mjaatvedt and Wong-Riley, 1988). Both have greater metabolic demand than layers with cell bodies, which tend to have low metabolic demand (Kageyama and Wong-Riley, 1982) and make up less than 20% of volume. The high metabolic demand of the dendritic layers underlies the disproportionate need for oxygen, glucose and ATP production in the brain. Depriving cells of primary afferent input causes a dramatic down-regulation of cytochrome oxidase activity, whereas restoration of afferent input up-regulates cytochrome oxidase activity (Wong-Riley, 1989). If metabolic support decreases subtly with age, dendritic retraction could effectively reduce energy demand, thereby sparing cells. So it may not be surprising that age can be associated with dendritic atrophy (Markham et al., 2005), and is not always related to cell loss (Rapp and Gallagher, 1996; Yates et al., 2008).
Despite the high and constant neural demand for ATP, it is used within minutes of being produced, leaving the brain highly vulnerable to any decline in metabolic support. ATP production is reduced when there is too little glucose or oxygen, or there is damage to cellular metabolic machinery. Aging is associated with a reduction in cytochrome oxidase reactivity (Ojaime et al., 1999), which would appear to place neurons at risk in periods of high metabolic demand. With too little ATP, the sodium potassium pumps may fail to restore resting ion gradients, resulting in excess depolarization. An increase in cell membrane depolarization allows excess calcium influx that, in turn, can cause cell death by initiating apoptosis, or necrosis. If age is related to damage to cellular metabolic machinery, then there is an increase in neuronal vulnerability to excitotoxicity.
As much as life is dependent upon metabolism, it is also dependent upon a second fundamental feature, collaboration (Dupre and O'Malley, 2009). This collaborative process is illustrated best by the dependence of cellular metabolism on the delivery of oxygen and glucose, which is carried out by the cardiovascular system. Neurons store little glucose, the fundamental source of energy, or oxygen, which is necessary for glucose metabolism. Since little ATP is stored, it must be produced within minutes of being used. Not surprisingly, neuronal activity triggers an increase in the delivery of oxygen and glucose; cerebral glucose utilization and cerebral metabolic rate for oxygen are coupled to neuronal activity (Magistretti and Pellerin, 1996). Similarly, across regions and dendritic layers, cytochrome oxidase activity correlates with glucose metabolism and capillary density (Borowsky and Collins, 1989). The dynamic and exquisite coupling between neuronal activity, cellular metabolism and oxygen and glucose delivery illustrates the fundamental role collaboration plays in supporting neuronal function, and therefore cognition.
The consequences of insufficient oxygen and glucose are illustrated by the sensory and cognitive deficits resulting from low glucose (Bie-Olsen et al., 2009; Gonder-Frederick et al., 2009) and hypoxia at high altitudes (Kramer et al., 1993) as well as mitochondrial dysfunction (Lauritzen et al., 2010). As we will see, aging can represent a period of slow decline in cellular metabolism, cardiovascular support or in the coupling between these systems, and therefore this decline may play a role in age-related cognitive deficits. As we consider targets for brain aging intervention, processes critical for metabolism must be considered. Thus, the most basic interventions in cognitive aging may require interventions in metabolic decline that span cellular to systems levels of support.
At the level of the whole brain, an age-related decline in metabolic support is clearly established. Resting cerebral blood flow has been found to be lower in normal healthy elderly with a variety of methods, which include continuous arterial spin labeling (CASL) magnetic resonance (MR) imaging (Bertsch et al., 2009), PET (Frackowiak et al., 1980), and SPECT (Devous et al., 1986). Cerebral blood flow was significantly less in elderly (M =71years old) than in young individuals (M=27years old) (Stoquart-ElSankari et al., 2007). Within the elderly sample, the decline was linearly related to age. At the regional level, neurovascular coupling is impaired. In aged compared to young subjects, the BOLD signal, which is dependent upon neurovascular coupling, has a higher signal to noise ratio, and can take longer to reach half the maximum value (D'Esposito et al., 2003). It is unclear whether these reductions reflect a reduction in neural activity driving a reduction in metabolism and need for oxygen and glucose, or conversely, whether impaired metabolic support reduces the capacity for neural activity. The initiating factor may vary across individuals.
Lifestyle correlates with cardiovascular health and cerebral blood flow. Within aging populations, adults who chose to continue working at retirement age, and adults who retired but remained active maintained cerebral blood flow 3years past retirement relative to the decline seen in adults who became physically inactive (Rogers et al., 1990). It is unclear to what extent the reduction in blood flow is a cause or reflection of age-related changes in cognition and tissue volume. Individuals with active lifestyles have tighter cerebral regulation for cerebral blood perfusion (Formes et al., 2010). Although exercise is well-known to maintain heart health, reduce blood pressure and elevate high density lipoproteins (National Institutes of Health, 1996), all effects that should support brain function, more studies need to address whether exercise can directly maintain cerebral blood flow. The importance of cardiovascular function for cognitive function in aging humans is consistent with this possibility (Kramer et al., 1993, 1999; Colcombe et al., 2004; McAuley et al., 2004).
Total cerebral blood flow and regional cerebral blood flow are dependent upon maintenance of arteries, arterioles and capillaries. The importance of the health of the cerebral vascular network may be best illustrated by evidence that vascular damage is the primary cause of cognitive deficits reported after tumor irradiation (Brown et al., 2005, 2007). Capillary density determines the distance over which oxygen and glucose must diffuse to reach cells, and can respond to changes in cellular energy demands. Capillary density varies with energy demand across space and developmental phases (Gross et al., 1987; Borowsky and Collins, 1989; Tuor et al., 1994), and is affected by conditions such as hypoxia (Mironov et al., 1994). An uncoupling between demand and density would be expected to alter basal neuronal metabolic capacity, which in turn may affect cognition. The risk of vascular occlusion increases with age so that a low density leaves more cells vulnerable to deficient oxygen and glucose when small vessels are occluded. Recognizing the importance of cerebral capillary density, numerous studies of capillary density in aged humans and non-human animals have been conducted (Riddle et al., 2003).
In humans, there are numerous reports that capillary density is lower in autopsy tissue from aged subjects relative to young subjects (Hunziker et al., 1979; Bell and Ball, 1981, 1985, 1990; Mann et al., 1986; Abernethy et al., 1993). Capillary density has also been reported to be lower in aged animals (Buchweitz-Milton and Weiss, 1987; Jucker and Meier-Ruge, 1989; Jucker et al., 1990; Amenta et al., 1995). Although, a number of studies have also reported no age-related differences in capillary density in rats and humans (Hunziker et al., 1979; Knox and Oliveira, 1980; Meier-Ruge et al., 1980; Hughes and Lantos, 1987; Meier-Ruge and Schulz-Dazzi, 1987). The reductions reported in humans are typically 16%, and in rats as much as 30%. Nearly all of these studies labeled capillaries with alkaline phosphatase histochemistry, which, while allowing labeling of capillaries in post-mortem tissue, has disadvantages. There is often variability in staining quality within samples, requiring investigators to forego random sampling. Capillary density estimates with this method have been reported to be 30% lower than the other methods in brain (Gobel et al., 1990), and muscle tissue (Hansen-Smith et al., 1992). Since alkaline phosphatase can be inhibited by a number of factors, a reduction in labeled capillaries may represent more than just a reduction in density. Using electron microscopy, age was not associated with capillary density in white matter, but age was associated with differences in capillary integrity (Farkas et al., 2006), which has been explored for its role in Alzheimer's disease. The consistent reports of age-related declines in cerebral blood flow, taken with the alkaline phosphatase measures of capillary density support the need for more research to test the relationship between capillary density and age. If capillary density decreases with age, an intervention that maintains capillary density or induces angiogenesis should be beneficial.
If exercise can increase capillary density, then it may spare neuronal function by maintaining adequate glucose and oxygen availability to neurons. In humans, MR angiograms offer a method for measuring arteries in live subjects, but are limited with regard to measurement of capillaries. Using this method, exercise was shown to increase the number of small vessels ranging from 0.25 to 0.6mm radius (Bullitt et al., 2009). No studies have tested relationships between fitness levels in humans and the density of capillaries. In rats, exercise at ages that are equivalent to human adulthood and middle age has been reported to increase capillary density in motor areas of the brain. One month of voluntary or forced (1h/day at 10m/min) exercise in middle-aged female rats increased capillary density in the cerebellar molecular layer (Isaacs et al., 1992). Wheel running also increased capillary density and integrins in the motor cortex, as well as capillary perfusion (Kleim et al., 2002; Swain et al., 2003). In the striatum, treadmill walking for 3weeks at 30min/day increased capillary density in 3- and 22-month-old rats (Ding et al., 2004, 2006b). In the substantia nigra, treadmill training reversed age-related declines in vessels labeled with endothelial cell antigen (Villar-Cheda et al., 2009). Exercise in monkeys (5days a week for 5months) increased vascular volume fraction in the motor cortex. The effect did not persist after 3months of sedentary conditions (Rhyu et al., 2010), suggesting that exercise must be maintained. The exercise-induced increases in vascular volume fraction were restricted to 15- to 17-month-old (mature) monkeys, and were not seen in 10- to 12-year-old (middle-aged) monkeys. This is notable because it is one of the few studies that test the effects of exercise across ages and including older subjects. In our lab, treadmill training in rats with the parameters used by Isaacs et al. (1992) did not increase capillary density or glucose transporter mRNA in the hippocampus (McCloskey and Anderson, 2005), although exercise is reported to increase endothelial cell proliferation in the hippocampus (Ekstrand et al., 2008). There is strong evidence that exercise can increase capillary density in motor regions of the brain obtained with methods that are less subject to artifacts. Unfortunately, few studies have tested for the effects of exercise in non-motor regions. Therefore more work is necessary to test whether exercise-related changes in regional vascularization are widespread, and whether these effects generalize to aging animals.
If exercise can induce capillary growth, angiogenesis, it may do so through an increase in the expression of vascular endothelial growth factor (VEGF). Exercise (1 session of treadmill running at 24m/min, 10° incline for 1h) in mice increases VEGF transcription, mRNA and protein in the hippocampus, but not in the cerebellum or striatum (Tang et al., 2010). In our lab, 3weeks of voluntary exercise in adult male mice also increased hippocampal VEGF protein, measured using ELISA (unpublished observations). In the substantia nigra pars compacta, both VEGF mRNA and vascularization are reduced in aging, and these reductions are reversed by exercise (Villar-Cheda et al., 2009). The correspondence between vessel density and VEGF in this study is notable because it is the only study in which both factors are studied in the same animals. The lack of correspondence between exercise-induced VEGF expression and exercise-induced angiogenesis within other brain areas needs to be further explored by addressing both variables in the same species with the same exercise conditions, and in middle-aged and aging animals.
Increasingly vascular pathology is being explored for its role in the etiology of AD (Farkas and Luiten, 2001), which increases in risk with age. Basement membrane thickness and fibrosis are pathological features in AD that increase in the brain with age (Farkas and Luiten, 2001; Farkas et al., 2006). Thickened microvessels may represent impairments in nutrient, glucose, and oxygen extraction as well as impaired elimination of waste. Furthermore, they are predicted to affect elasticity, and therefore the potential for a reduction in neurovascular coupling. The increase in basement membrane thickening in rats begins to appear at 20months, but is visible by 30months (de Jong et al., 1992). The age-related changes in humans are accelerated by risk factors such as hypertension (Farkas et al., 2000). Few studies have addressed whether exercise affects basement membrane thickness in cerebral capillaries. However, in muscle tissue, 9months of endurance training decreased skeletal muscle capillary basement membrane thickening in older humans (M=63years old), producing a return to values similar to those seen in young subjects (M =28years old) (Williamson et al., 1996). It is possible these effects generalize to the brain since pathology in peripheral vasculature correlates with pathology in central nervous system arteries (Farkas and Luiten, 2001). If exercise can maintain capillary integrity, it may not only protect or slow age-related deficits in cognition, but also provide partial protection from vascular dementia and Alzheimer's disease. A stronger relationship between age and capillary density needs to be established, but age clearly affects capillary integrity, which in turn may underlie some age-related neuropathologies.
Age-related decline in metabolism is apparent at the cellular level. At autopsy, there is a relationship between age and cytochrome coxidase activity in the frontal cortex, temporal cortex, superior temporal gyrus, cerebellum and putamen (Ojaimi et al., 1999). Similarly, 26-month-old rats had a decrease in cytochrome oxidase subunits in the cerebral cortex (Nicoletti et al., 1995). Cytochrome oxidase increases after exercise in motor areas of the rat brain. Six months of voluntary wheel running in rats increases cytochrome oxidase reactivity in limb representations of the striatum and motor cortex, but not the hippocampus (McCloskey et al., 2001, see Figure Figure1).1). The lack of an effect in the hippocampus may suggest that the relationship between exercise and metabolic capacity is less clear beyond motor regions. For example, Molteni et al. (2002) found that 7days of exercise produced marginal increases in gene expression for cytochrome oxidase subunits IV, V, VI, VIII, 4 of the 13 protein subunits necessary for a functional enzyme. Yet, Tong et al. (2001) found that subunits I, II, and III (3 of 13) were decreased in expression in male rats that exercised for 3weeks. Although there is evidence that the regulation of gene expression for the 13 subunits is coordinated when there is dramatic increase or reduction in afferent input (Liang et al., 2006), it seems possible that exercise may produce a less dramatic increase in neural activity in the hippocampus, and therefore may be less capable of initiating the coordinated up-regulation necessary for a functional enzyme. Exercise has also been shown to influence anti-oxidants (Radak et al., 2001, 2008a,b), which should protect mitochondrial DNA, and in turn maintain the capacity to produce functional cytochrome oxidase enzyme. The demonstration that exercise can affect cellular metabolic capacity, at least in motor regions, is a key demonstration that any restoration in cardiovascular support can be utilized by cells.
Neuron loss may not always occur in healthy aging (Rapp and Gallagher, 1996; Yates et al., 2008), but neurons are at risk from the rising probability of an imbalance between metabolic capacity and energy demands as age increases. Age is associated with an increased risk for epilepsy (Hauser, 1992) and stroke. During ischemia, the energy demand of basal neural activity exceeds the diminished metabolic support, whereas during seizures, neural activity is high and produces metabolic demand that exceeds available supply. In both, the energy demands of neural activity exceed supply leading, in turn, to excess depolarization and potential cell death. Animal models of ischemia and seizures provide the opportunity to test for neuroprotective effects of exercise.
If exercise increases capillary density and the cellular metabolic capacity, then exercise should protect cells when they face metabolic challenges that increase in frequency with age. The risk for ischemia increases with age, but exercise pre-conditioning protects some cell populations from ischemia (Stummer et al., 1994, 1995; Sim et al., 2005). Exercise pre-conditioning reduced mortality from a 20-min occlusion of the carotid artery (Stummer et al., 1994). Gerbils that ran for 2weeks in a running wheel prior to a 15-min occlusion of the carotid artery had only 50% cell loss in the limb representations of the striatum relative to 90% cell loss in control gerbils. Similarly, treadmill training for 30min a day over 1–3weeks in rats prior to middle cerebral artery occlusion reduced inflammatory injury in the striatum and reduced infarct volume (Ding et al., 2005, 2006a). It is as of yet unclear whether the greater metabolic capacity and capillary density protect against damage from capillary occlusion since these studies tested for protection after occlusion of major arteries. After MCAO occlusions, the protective effects of exercise are at least partly due to a reduction in inflammatory processes during reperfusion (Li et al., 2004). Exercise pre-conditioning increased TNF-alpha expression, which reduced NFκB expression. Reperfusion injury, mediated by TNF-alpha binding to NFκB, was reduced because of the exercise-related reduction in NFκB. Consequently, reperfusion injury, mediated by inflammatory processes, was reduced in exercising animals (Ding et al., 2006c), but angiogenesis also plays a role (Ding et al., 2004). Exercise was not protective when MCAO occlusion was permanent. Together, the data suggest that exercise protects striatal cells from metabolic challenge, in part through an increase in metabolic support.
Exercise also protects against ischemia in the hippocampus. Fifty percent of neurons in the CA3 subregion of the hippocampus of exercising gerbils survived after a 15-min occlusion of the carotid artery, whereas only 10% of neurons survived in the control gerbils (Stummer et al., 1994). However, protection was restricted to CA3, and was not seen in CA1, a subregion of the hippocampus particularly vulnerable to ischemia. Treadmill training (30min/day for 10days) in gerbils decreased apoptosis and caspase-3 expression in the dentate gyrus after 5min occlusion of both carotid arteries (Sim et al., 2004). One month of treadmill training (30min/day for four consecutive weeks) attenuated functional deficits (Sim et al., 2004, 2005). Altogether, there is substantial evidence that exercise can spare neurons from death following ischemia. However, there is less evidence that exercise can enhance metabolic support in the hippocampus, so it remains unclear whether metabolic factors account for the protective effects of exercise in the hippocampus or enhanced expression of growth factors.
Epilepsy and acute unprovoked seizures are predominant in patients older than 65years and younger than 12months of age (Hauser and Beghi, 2008). Although the onset of seizures is often the result of an antecedent event, 50% of new epilepsy cases in the elderly have no known antecedent event (Hauser, 1992; Hauser et al., 1993; Li et al., 1997; Leppik, 2006). Late onset epilepsy is correlated with a decline in vascular factors, suggesting that age-related declines increase vulnerability to epilepsy (Li et al., 1997). Seizures are events in which energy demand is high, and at times exceed the cell's metabolic capacity. Since exercise can up-regulate cellular bioenergetic capacity and oxygen and glucose availability, it may modulate the progression of seizures and cell loss. However, exercise regulation of cell energy metabolism has not been reported in the hippocampus, which is the focus of models of temporal lobe epilepsy. In our own studies, 4–7weeks of wheel running reduced the proportion of rats progressing to continuous generalized seizures (status epilepticus, SE) after systemic injections of kainic acid (10mg/kg, see Figure Figure2).2). When rats ran for only 1month, the protection from status epilepticus was restricted to rats that ran over the median number of wheel rotations (unpublished observations). High runners (above the median split) had a significant protection from SE relative to the combined group of control and low runners (Yates χ2 (1) =9.235, p <0.05). SE causes a dramatic loss of CA3 pyramidal cells, whereas moderate seizures cause cell loss in only small patches, if at all. As a consequence, exercise, by protecting rats from progressing to continuous generalized seizures, provides significant protection against cell death. In another study, voluntary wheel running had no effect on seizure severity (based on the Racine scale) when the dose was below 10mg/kg, but reduced seizure severity when kainate was given at a dose above 10mg/kg (Reiss et al., 2009). These studies suggest that exercise moderates seizure progression when kainic acid is administered systemically.
Consistent with previous studies, exercise raised the threshold for kindling induced seizures (Arida et al., 1998), although this effect was restricted to exercise in adulthood; exercise during development did not protect adult animals (Arida et al., 2007). After 21days of voluntary wheel running, seizure ratings were lower after ventricular infusion of low, but not moderate doses of kainic acid (Reiss et al., 2009). We found that injections of 0.55μg kainic acid into the lateral ventricles at a site most anterior to the tip of the hippocampus yielded no difference in the proportion of animals progressing to SE, despite using 6 rather than 3weeks of exercise. Only 21% of the animals entered SE, which is consistent with the low dose.
Several investigators have also addressed the possibility that exercise protects from seizure-induced cell death. Following systemic injections of domoic acid, mice trained to run on a treadmill for 1km/day had less cell loss in the hilus of the hippocampus (Carro et al., 2001). In contrast, when kainic acid was infused directly into the hippocampus, wheel running in female rats exacerbated, rather than protected against cell loss in the CA3 region (Ramsden et al., 2003). It is important to note that Ramsden infused kainic acid in animals anesthetized with a drug used to control seizures. We carried out a series of experiments to further test the effects of exercise on cell loss. Like Ramsden et al. (2003), we used kainic acid infusion into the lateral ventricles, but in contrast to their methods, we infused kainic acid into awake male rats. When we restrict our analysis to cell loss in animals that did not progress to SE, we find, like Ramsden et al. (2003), that exercise exacerbates damage. Together, these latter studies leave us to conclude that exercise does not protect CA3 cells from moderate seizures, but instead increases neuron vulnerability. The discrepancy in exercise effects on cell vulnerability in the hippocampus may not be surprising given that there is little evidence for exercise-related up-regulation of cellular metabolic capacity or capillary density in this region. The two reports that exercise protects the striatum from ischemic damage parallel exercise-related enhancement of metabolic support in that region. Whereas these studies support a beneficial effect of exercise in the aging striatum, future studies need to directly test that hypothesis in aging animals..
It is important to note that the majority of evidence for the ability of exercise to influence regional vasculature and cellular metabolic machinery come from motor systems like the striatum and motor cortex. Exercise also affects the hippocampus, but those effects may be through growth factor expression rather than through an increase in capillary density or metabolic capacity (Kesslak et al., 1998; Gomez-Pinilla, 2008). For example, VEGF, which is increased in the hippocampus by exercise, can decrease seizure-like activity there (McCloskey et al., 2005), and protect hippocampal neurons from seizures (Nicoletti et al., 2008), independent of its ability to increase vasculature. This may explain why the evidence for cell protection is mixed in the hippocampus.
Many of the studies reviewed here utilized animal models, which allow greater control over exercise conditions as well as the application of methods that can directly test the biological mechanisms of aging and exercise. Exercise manipulations vary in these animal studies. The use of voluntary exercise is common because activity levels can be increased without inducing stress. When voluntary exercise is used, animals can run in wheels any time over the 24-h day, although running has a circadian rhythm. Voluntary exercise in rats may be analogous to humans with physically active lives, and jobs with manual labor. In contrast, forced exercise on a treadmill usually constitutes forced running for 1h a day for 5–7days a week, while rats are sedentary for the remaining 23h a day. This may be analogous to gym workouts by individuals with desk jobs. Therefore the two types of exercise used in animal studies serve as models for very different physical activity patterns in humans. Nevertheless, both wheel running and treadmill training increase capillary density in motor regions. Wheel running has been shown to increase cellular metabolic capacity in these regions, and treadmill training has not been tested.
Here, we propose that the functional deficits that occur with age are at least partially related to the decline in cellular function resulting from a decline in metabolism. This decline involves cardiovascular function at the level of the organism and brain, in cerebral blood flow, in cytochrome oxidase activity, and possibility capillary density. There is substantial evidence that exercise can influence metabolic support at a number of levels, including cardiovascular function, regional capillary density, and cellular bioenergetic capacity. We hypothesize that exercise can counter the age-related decline in metabolic support (see Figure Figure3).3). Unfortunately, there is little evidence to support this hypothesis, because evidence that exercise can influence vasculature and cellular metabolism is primarily limited to adult and middle-aged animals. If exercise can induce similar effects in aging animals, then it may be a powerful intervention through its influence on the multiple levels of support necessary for cellular metabolism. It is important to note that most of the studies demonstrating exercise effects on human cognition have studied aging rather than young or adult samples. Accordingly, the evidence that exercise improves cognition in elderly humans provides incentive for understanding the neurobiological mechanisms, and supports the possibility that effects established in adult animals may generalize to aging animals. To bridge this gap, studies of exercise effects on regional vasculature and cellular metabolism should be extended to aging animals, and should include analyses of both motor and non-motor structures.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
W. Burghardt Turner Fellowship and the NSF Bridges to the Doctorate Fellowship and Turner Summer Research Fellowship to Shayri J. Greenwood. We thank Sara Mellers for helpful suggestions on manuscript.