Brain Microdialysis in Exercise Research

Transcription

1 REVIEW ARTICLE Sports Med 2001; 31 (14): /01/ /$22.00/0 Adis International Limited. All rights reserved. Brain Microdialysis in Exercise Research Romain Meeusen, M. Francesca Piacentini and Kenny De Meirleir Department of Human Physiology and Sports Medicine, Faculty of Physical Education and Physiotherapy, Vrije Universiteit Brussel, Brussels, Belgium Contents Abstract Brain Microdialysis Principle of Microdialysis Features of Microdialysis Concerns with Using Microdialysis Characteristics of Microdialysis Sampling Applications of Microdialysis Microdialysis and Behaviour Methodological Aspects Behavioural Studies Stress Activity State Intracerebral Microdialysis During Exercise Conclusion Abstract During the last 5 to 10 years, the microdialysis technique has been used to explore neurotransmitter release during exercise. Microdialysis can collect virtually any substance from the brains of freely moving animals with a limited amount of tissue trauma. It allows the measurement of local neurotransmitter release in combination with ongoing behavioural changes such as exercise. Several groups examined the effect of treadmill running on extracellular neurotransmitter levels. Microdialysis probes were implanted in different brain areas to monitor diverse aspects of locomotion (striatum, hippocampus, nucleus accumbens, frontal cortex, spinal cord), food reward (hypothalamus, hippocampus, cerebral cortex), thermoregulation (hypothalamus). Some studies combined microdialysis with running on a treadmill to evaluate motor deficit and improvement following dopaminergic grafts in 6-hydroxydopamine lesioned rats, or combined proton nuclear magnetic resonance spectroscopy and cortical microdialysis to observe intra- plus extracellular brain glucose variations. This method allows us to understand neurotransmitter systems underlying normal physiological function and behaviour. Because of the growing interest in exercise and brain functioning, it should be possible to investigate increasingly subtle behavioural and physiological changes within the central nervous system,

2 966 Meeusen et al. and to link this with neuroendocrinological data, to better understand a stress called exercise. There is now compelling evidence that regular physical activity is associated with significant physiological, psychological and social benefits in the general population. [1] In contrast with our knowledge about the peripheral adaptations to exercise, studies relating exercise to brain neurotransmitter levels are scarce. [2] It is of interest to examine the effect of short and long term exercise on neurotransmitter release, since movement initiation and control of locomotion have been shown to be related to striatal neurotransmitter function, and one of the possible therapeutic modalities in movement, and mental disorders is exercise therapy. Until very recently most experimental studies on brain chemistry were conducted with postmortem tissue. However, in part because of shortcomings with postmortem methods, and in part because of the desire to be able to directly relate neurochemistry to behaviour, there has been considerable interest in the development of in vivo neurochemical methods. Because total tissue levels may easily mask small but important neurochemical changes related to activity, it is important to sample directly in the extracellular compartment of nervous tissue in living animals. Since the chemical interplay between cells occurs in the extracellular fluid, there was a need to access this compartment in the intact brain of living and freely moving animals. Estimation of the transmitter content in this compartment is believed to be directly related to the concentration at the site where these compounds are functionally released: in the synaptic cleft. As measurements in the synapse are not yet possible, in vivo measurements in the extracellular fluid appear to provide the most directly relevant information currently available. [3] This article provides an overview of the in vivo microdialysis technique as a method for measuring in the extracellular space, and its application in exercise science. Although this technique has been used in different tissues such as brain, adipose tissue, spinal cord and muscle, in animals as well as humans, we will focus on the use of this in vivo method in brain tissue. Recently two excellent reviews on the application of microdialysis in human experiments especially in subcutaneous tissue have been published, [4,5] and we refer the interested reader to these articles. 1. Brain Microdialysis The idea of using the principles of dialysis to sample extracellular fluid in brain is 35 years old. Indeed, in 1966 Bito et al. [6] implanted dialysis sacs into the subcutaneous tissue of the neck and into the parenchyma of the cerebral hemispheres of dogs. Later, Delgado et al. [7] developed a diatrode which was very similar to the present microdialysis probes. The method has been continuously refined [8-10] and today it is used in a large variety of experiments for bioanalytical sampling of substances from the brain and other tissues. As a sampling technique, microdialysis is not directly coupled to any particular method of chemical analysis, but to being able to analyse the sometimes very small amounts of chemicals in the dialysate, it is necessary to have a sufficiently sensitive analytical method. [11] It is evident that analytical chemistry is of great impact for the microdialysis technique, as the limits of the method are usually set by the sensitivity of the assay technique. [11] Highly sensitive assays have now been developed for most of the neurotransmitters. The vast majority of these studies suggest that high pressure liquid chromatography (HPLC) in conjunction with electrochemical or fluorescence detection provides a very sensitive analytical method that is convenient to use with dialysates. [12] Because blood cells, proteins and other large molecules cannot enter the dialysate, the sample is automatically cleaned up and can be injected without further purification into the HPLC system.

3 Brain Microdialysis During Exercise Principle of Microdialysis The microdialysis technique employs the dialysis (from Greek: to separate) principle in which a membrane, permeable to water and small solutes, separates 2 fluid compartments. The principle is based on the kinetic dialysis principle: the membrane is continuously flushed on one side with a solution that lacks the substances of interest, whereas the other side is in contact with the extracellular space. Provided that no osmolarity (pressure) and electrical potential differences exist across the membrane, solutes transport between the 2 compartments is governed only by diffusion, that is, concentration gradients. A concentration gradient is created causing diffusion of substances from the extracellular space into the dialysis probe (and vice versa). A microdialysis probe was constructed to create a convenient way of introducing a dialysis tube into the tissue. [13] A dialysis membrane is connected with in- and outlet tubings, allowing fluid to enter and leave the microdialysis probe. The continuous flow through the probe carries substances to the sampling site for further analysis. Figure 1 shows a schematic drawing of a microdialysis probe. 3. Features of Microdialysis Ungerstedt [13] stated that the idea of microdialysis is to mimic the passive function of a capillary blood vessel by perfusing a thin dialysis tube implanted into the tissue. Microdialysis has several features (table I). It samples the extracellular fluid as distinct from the whole tissue collected by biopsies, punches or dissections. It can be performed locally in almost every organ and tissue of the body in intact tissue of living, awake and freely moving animals, distinguishing it from other preparations such as slices and synaptosomes. Microdialysis makes it possible to sample continuously for hours or days in a single animal, which in addition to other advantages, decreases the number of animals needed in an experiment. Because the microdialysis probe may remain implanted in a single animal over a full experimental period, sampling via microdialysis can be used to reduce intersubject variation. This method Inlet Outlet Steel shaft Membrane Inlet Outlet Fig. 1. Schematic representation of the microdialysis probe. A microdialysis probe is a specially designed cannula with a semipermeable membrane at its tip (insert). The outer diameter of the dialysis probe is 0.25 to 0.52mm. The length of the semipermeable membrane varies depending on the tissue of interest. Usually, the membrane length is between 1 and 3mm. The inlet of the probe is connected to a microinjection pump. At the outlet tubing the microdialysis sample (dialysate) is collected.

4 968 Meeusen et al. Table I. Features of microdialysis It samples the extracellular fluid Almost every organ and tissue of the body can be sampled Intact tissue of living, awake and freely moving animals can be sampled Continuous samples can be collected Intersubject variation is reduced Endogenous and exogenous substances in the tissue can be recovered and/or introduced A representative sample of all substances in the extracellular fluid can be collected Samples are relatively clean and ready for analysis Tissue damage is minimal can be used for recovering and/or introducing endogenous and exogenous substances in the tissue. Microdialysis collects a representative sample of all substances in the extracellular fluid (provided that they pass the membrane), and carries them out of the body for further analysis. The samples are relatively clean and ready for analysis, because large molecules (enzymes) do not pass the membrane due to their molecular cut-off weight. Several experimental studies [14-17] have shown that the damage to the blood brain-barrier is minimal. The created tissue damage is minimal compared with other intracerebral sampling techniques, because there is no direct contact between the liquid flowing inside the membrane and the cell of the tissue. 4. Concerns with Using Microdialysis Of course, there are some concerns with the use of microdialysis. Since the sample volume collected during one experiment depends on the duration of collection and the flow rate of the perfusion fluid, only a small sample (e.g. 20 to 40µl) is collected. The sampling time is often determined by the detection limit of the compound. [11] Because of this need for sufficient volume, the time-resolution is usually between 5 to 20 minutes. This means that the amount of fluid recovered in one sample, contains the amount of substances (transmitters) harvested during total collection time. In case of behavioural studies it is very important to register any disturbances that occur during the ongoing experiment to create a full picture of all the events (sometimes also unwanted effects such as disturbance of the animal) that occur during sampling. Thus, sampling efficiency depends on several factors such as membrane properties, perfusion flow rate, possible interactions between the membrane and specific substances. Readers interested in further methodological details are referred to reviews by Kehr, [18] and Benveniste. [19] 5. Characteristics of Microdialysis Sampling According to Zetterström et al., [17] the ratio between the concentration of a substance in the outflow solution and the concentration of the same solution outside the probe is defined as recovery. Because the dialysis probe is continuously flushed, the levels of substances in the dialysate (outflow solution) are only reflections of the true extracellular fluid. The recovery of substances from the extracellular fluid depends on the length of the dialysis membrane, the flow of the perfusion fluid, the speed of diffusion of the substance through the extracellular fluid and the properties of the membrane. For small molecules such as the monoamine transmitters, the factor limiting recovery is usually the speed of diffusion through the extracellular fluid, not the diffusion through the membrane. [13] This means that 2 substances with different diffusion characteristics may appear in the perfusate in a concentration ratio that is different from that in the tissue. [18] One of the great advantages of microdialysis is that experiments on awake animals are relatively easy. However, working on awake animals implies that the animals are susceptible to all kinds of influences ranging from conceivable pain of the implantation and restraint by tubing and wires, to reactions in response to a new environment. [13] When doing experiments on awake animals it is of utmost importance to control the setup of the experiment. It is also important to consider the diurnal rhythm of the animals and therefore to perform the experiments during the same period of the day. Transfer-

5 Brain Microdialysis During Exercise 969 ring an animal from a cage with several rats to the isolated environment of the test cage has a profound influence which can be controlled by habituating the animals to the test cage and to handling before the experiment. [13] To avoid that the trauma of probe implantation interferes with the results from the experiment a guide cannula is implanted before probe implantation. The probe is inserted once the animal has recovered from surgery (figure 2). The skull is exposed and a guide cannula is implanted through a borehole in the brain area of interest according to a stereotaxic atlas. The guide (or probe) is secured by dental cement around the guide and small screws fixed to the skull bone. Care must be taken not to fix the screws so deep that they reach the brain. When doing behaviour studies it is usually advised to wait 24 to 48 hours after the guide implantation to allow the animals to recover from surgery before the guide is replaced by the microdialysis probe. The microdialysis probe is connected to a micro infusion pump and perfused with a perfusion fluid at a constant flow rate. The syringe is connected via wires to the microdialysis probe; a swivel allows the animal to move freely without getting caught in the tubing. In our exercise experiments, [20-23] the animal is transferred to the treadmill once the probe is implanted. This is always 1 day before the actual experiment takes place. On the morning of the experiment, the first 3 to 4 hours, baseline samples are collected before starting the (exercise) intervention. The counter balance arm used, is designed in such a way that it can follow the animal while it is running on the treadmill (figure 3). The composition of the perfusion fluid should be as close as possible to that of the extracellular fluid. The Ca 2+ concentration of the perfusion fluid varies among studies from 1.2 to 3.4 mmol/l. The basal output of neurotransmitters is strongly Ca 2+ dependent and the neurotransmitter output is nearly linearly related to Ca 2+ concentrations. However, it is not clear whether 1.2 mmol/l Ca 2+ is essential in brain dialysis. Sarre [24] stated that each investigator must select a Ca 2+ concentration on the basis of the minimal one which results in an impulse dependent release, acceptable on the basis of the a b Fig. 2. Before insertion of the microdialysis probe into the brain area of interest, the animal is anaesthetised and placed on a stereotaxic frame. A guide cannula is inserted into the brain. Two days later, when the animal has fully recovered, the guide is replaced by the microdialysis probe. Next day the experiment can start. a = Guide canula secured with dental cement; b = stereotaxic frame.

6 970 Meeusen et al. a b c d e Fig. 3. In exercise experiments the animal will be placed on a treadmill. A specially designed counterbalance arm with a swivel allows the animal to move freely on the treadmill without getting caught in the tubing. a = microinfusion pump; b = counterbalance arm; c = swivel; d = microdialysis probe in brain of running rat; e = treadmill. current criteria for in vivo neurotransmitter release. [24] In most of the behavioural experiments, aca 2+ concentration of 1.1 or 1.2 mmol/l will be the best choice. 6. Applications of Microdialysis Analysing the biochemical functions of the body has commonly involved the sampling of blood or dissection of tissue. However, blood is a distant reflection of the many events taking place in cells and organs and the dissected tissue represents a static picture. Although microdialysis is mostly used in brain research, there are other possible applications for this technique. A variety of other organs and tissues have been studied with microdialysis. Several studies have been published using microdialysis in adipose tissue, blood, adrenals, skin, bile duct, eye, heart, liver, muscle and tumours. In general, microdialysis is simpler to perform in these large organs because of their large size making it possible to use larger membranes. [13] Studies in human brain [25-29] are limited because of obvious ethical reasons; however, data are present regarding intrahippocampal microdialysis in epilepsy, [29] during neurosurgery, [26,30-33] or cerebral ischaemia and trauma. [26,33,34] Several laboratories have already used the microdialysis technique in human muscle, [35-39] human subcutaneous tissue [40-48] or peritendinous tissue. [49] The following sections discuss possible applications of intracerebral microdialysis, such as the use of this technique in behavioural research. 7. Microdialysis and Behaviour 7.1 Methodological Aspects A primary advantage of in vivo microdialysis is that it can be used to examine neurochemistry in behaving animals. Several kinds of behaviour have been studied using microdialysis. Ideally, samples of extracellular fluid should be collected without any disruption in behaviour. Similarly, behaviour should not disrupt the collection of samples. However, at times these effects will occur and data will

7 Brain Microdialysis During Exercise 971 be lost. [50] In contrast to pharmacological manipulation (which mostly create huge increases in extracellular neurotransmitter concentrations), behavioural studies sometimes only create small disturbances of transmission. Therefore, several studies have incorporated reuptake inhibitors, releasers, or other compounds to increase the extracellular neurotransmitter concentration, when performing behavioural studies. [12] Many behaviourally induced changes in neurochemical output are subtle, that is, in the order of 20 to 50% of baseline levels. In contrast, pharmacologically induced changes typically range from 100 to 500% or more. To reliably interpret the results from behavioural studies it becomes important to establish very stable baseline levels upon which the effects of experimental manipulations can be assessed. Obtaining stable baseline measures is especially important when studying motivated behaviours. [51,52] Baseline levels of serotonin (5-hydroxytryptamine; 5-HT) and dopamine (DA) are known to be affected by several disturbances such as food deprivation [53] and of course circadian rhythms. [54,55] A central question is whether neurotransmitters sampled by microdialysis are reflecting true synaptic release or a more nonspecific overflow from synaptic and nonsynaptic sources. One has to be especially careful in interpreting the results from amino acid neurotransmitters since the origin of extracellular γ-aminobutyric acid (GABA) and glutamate (GLU) levels might represent synaptic release, carrier-mediated release or glial metabolism. [56] It is generally believed that dialysis probes sample the extracellular fluid directly, but it is unclear how the concentrations of neurotransmitters in this compartment are related to the amount released in the synaptic cleft. [12,57] Neurotransmitters are released from neurons during a brief period following depolarisation. To reach the microdialysis membrane, neurotransmitters, released from the nerve terminals have to travel a relatively long distance in the extracellular space. [18,58,59] The extracellular space is not simply a uniform volume of distribution of released neurotransmitters. Many events, such as efflux, the encounter with catabolic enzymes, uptake and binding sites, take place between the release of the transmitter substance and the recovery via the probe. These processes will affect the free level of the neurotransmitter, and therefore the amount harvested through the probe. Neurotransmitters are only present in small amounts in the extracellular fluid, primarily due to the efficiency of clearing processes. When these clearing mechanisms change during ongoing behaviour (e.g. fluctuations in blood flow), they might cause changes in the extracellular content of neurotransmitters. [12] Estimations of the absolute levels of neurotransmitters in the extracellular fluid have been made, but these calculations are of limited value as they have little relation to the level in the synaptic cleft. [12,59] Therefore, we agree with Westerink [12] when he states that in vitro calibrations of probes have little significance for behavioural studies, while the relative changes in the dialysate content of neurotransmitters are of much greater interest. When measured during behavioural expressions they may indicate the relative changes in neurotransmitter content (averaged over time and space) that occurs in the synaptic cleft. Therefore, it is recommended to express dialysis data as a percentage of controls. [12] 7.2 Behavioural Studies The different types of behaviour that have been investigated is almost endless. Microdialysis provides a useful method for the study of extracellular neurotransmitters of conscious rats during active behaviour. [60,61] Several experiments indicate a role for neurotransmitter pathways in the mediation of a wide spectrum of behaviours, such as feeding and drinking, [62-65] orientation to sensory stimuli, [66] locomotion, [67-69] exploratory behaviour and stress (for review Westerink [12] ). In vivo monitoring of extracellular levels of transmitters such as DA, noradrenaline (NA) and serotonin by means of intracerebral microdialysis allows detailed analysis of how transmitter release is regulated in the terminal fields of these transmitters during ongoing behavi-

8 972 Meeusen et al. our. Such studies can also help us to get more insight into the daily variations in spontaneous transmitter release and help us to interpret different results. Kalen et al. [60] examined hippocampal NA and serotonin release over 24 hours. NA and serotonin systems have been implicated in the regulation of sleep and vigilance states; the extracellular levels of these transmitters in awake animals can, therefore, be expected to vary substantially over time. The NA and serotonin output showed pronounced fluctuations during the 24-hour period. Mean NA and serotonin output over the dark period of the day was 43 and 38% higher, respectively, than during the light period. Several studies were performed to examine the neurotransmitter release in different brain regions during feeding behaviour. [62,70] Hernandez et al. [70] for instance examined hypothalamic serotoninintreatmentsforfeedingdisordersand depression. The combination of the administration of drugs and manipulation of food intake was performed to examine the involvement of serotonin in appetite and mood. Hypothalamic serotonin increased significantly relative to controls in response to tryptophan loading, in the context of dietary control of serotonin. Serotonin also increased significantly following the smell of food and eating a meal, in the context of food anticipation, satiety and circadian rhythm. [70] Long term food deprivation and subsequent bodyweight loss have been found to increase drug-seeking behaviour and voluntary drug intake in animals. Long term food deprivation may alter basal DArelease in the nucleus accumbens. [71] There is considerable evidence suggesting an involvement of nucleus accumbens in reward-related behaviours. [51,72-75] The data of some studies demonstrate the sensitivity of the mesolimbic DA system to conditioned stimuli associated with ingestion, but also suggest that its activity is modifiable in a manner consistent with the affective valence of a conditioned stimulus. [51] 7.3 Stress The activation of central catecholaminergic systems by stressful stimuli has been the focus of many studies. Several studies have shown that brain increases in monoamines following stress are dependent on the form of stress used. [76] It is well established that exposure to stress increases the monoaminergic [76-81] and glutamatergic [81-83] activity in several brain areas. Many acute stressors increase the extracellular levels of NA and DA in a number of brain regions, indicating that stress can elicit widespread activation of catecholaminergic neurons. Cortical NA and DA projections may represent a component of a globally activated catecholaminergic system that is responsible for stress-induced anxiety. [78,81] NA release in the hippocampus has been reported to increase during immobilisation stress or intermittent tail shocks, as well as during application of mildly stressful stimuli, such as gentle handling of the animals or tail pinch. [84] These authors made a systematic comparison of changes in DA overflow in several brain regions during exposure to a wider variety of stimuli other than stress. They compared mild stressors such as handling or tail pinch with nonstressful, rewarded behaviour, such as eating not preceded by food deprivation. Their results support the idea that mesocortical and mesostriatal DA systems are differentially regulated during ongoing behaviour and that they respond quite differently to stressful and rewarding stimuli, and that NA release can be selectively activated in different cortical and striatal terminal subfields. [84] Animals are also confronted with repeated or prolonged exposure to stressors, and it is clear that exposure to long term stress can alter the response of an animal to subsequent stressors. [75,85,86] Following the short term administration of amphetamine there is a marked dose-dependent increase in motor activity that is accompanied by a dose-dependent increase in the extracellular levels of DA in both the nucleus accumbens and caudate nucleus. [87-90] The repeated, intermittent administration of cocaine produces a progressive enhancement in the behavioural response to subsequent ex-

9 Brain Microdialysis During Exercise 973 posure to cocaine, a phenomenon known as behavioural sensitisation. As with psychostimulants, previous exposure to environmental stress, whether given repeatedly or as a single exposure, results in both behavioural sensitisation, and in enhanced DA neurotransmission. [91] The striking similarities in the changes that are produced by stress and psychostimulant-induced sensitisation has led to studies demonstrating behavioural and neurochemical crosssensitisation between these stimuli. [91] Until now, most of the studies on motor behaviour were conducted to measure the effect of amphetamine or other stimulants on the animals locomotor activity. In these cases, intracerebral transmitter release was provoked by drugs and the subsequent increase (or decrease) in locomotor behaviour was measured. There is a possible interaction between DA and GLU in the nucleus accumbens in the initiation of locomotion. [67,92-95] The relationship between other transmitters and locomotion were also examined. [68,69,96] Direct injection of excitatory transmitter agonists into several brain regions increases locomotion. [95] 7.4 Activity State There are substantial fluctuations in neurotransmitter release in freely moving animals which correlate with the activity state of the animals. [60,97] Because animals are more active during the dark period than during the day, it is possible that the observed changes in neurotransmitter overflow are associated with the level of arousal or vigilance, or with gross body movements. [97] Systemic administration of a variety of serotonergic drugs, including precursors, agonists and releasers produces a motor syndrome in rats. [98] Its most conspicuous signs are hyperactivity, head shakes or wet dog shakes, hyperactivity, tremor, rigidity, hind limb abduction, Straub tail, lateral head weaving and reciprocal forepaw treading. This so-called serotonin syndrome has served as an important research tool because it represents one of the few pure behavioural signs of central serotonin activity. [99] Because brain serotonin is thought to be important in a wide variety of behavioural (aggression, feeding, sleep), and physiologic (thermoregulation, cardiovascular control, glucoregulation) processes, several studies examined serotonin neuronal activity under a diversity of conditions. Jacobs [99] and his co-workers performed several studies to examine the electrophysiological activity of serotonin neurons in a number of behavioural and environmental or physiological challenges. In most studies, these challenges did not significantly activate serotonin neurons above a level seen during an undisturbed active waking state. In later investigations, microdialysis experiments were coupled to simultaneous monitoring of the behavioural state of the animal. As predicted, extracellular serotonin increased and decreased in parallel with behavioural state. [100] The results from these studies indicated that during periods of low motor activity, the activity of the serotonin neurons may be suppressed. [99] Since in some forms of depression there is a deficit in serotonin neurotransmission, it could be beneficial for these patients to train their tonic motor activity. One of the possibilities is to engage in some form of repetitive locomotion such as walking, riding a bicycle and jogging. These arguments are in agreement with the (sometimes anecdotal) research data on the effects of exercise on depression. Jacobs [99] stated that since obsessive compulsive disorders are a form of psychopathology where serotonin appears to play an important role, and reuptake inhibitors are often effective in the treatment of these disorders (and also of depression). Repetitive motor acts increase serotonin neuronal activity, that is, exercise or locomotion might activate brain serotonin in a nonpharmacological manner. 8. Intracerebral Microdialysis During Exercise Over the last 5 to 10 years, the microdialysis technique has also been used to investigate neurotransmitter release during exercise (table II). Two studies [104,106] found that 20 minutes of exercise on a treadmill significantly increased DArelease in rat striatum. In their first study, Hattori et al. [104] com-

10 974 Meeusen et al. bined microdialysis with running to evaluate motor deficit and improvement following dopaminergic grafts in 6-hydroxydopamine (6-OHDA) lesioned rats. DAand its metabolites [3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid] significantly increased during the treadmill exercise in control (unlesioned) animals. In another study, [101,116] the walking behaviour in animals that were 6-OHDA depleted as neonates was examined (walking speed: 3 m/min). There was no significant change in extracellular DA compared with baseline during treadmill walking in 6-OHDA control animals. In a follow-up study, Hattori et al. [106] showed that there is a threshold speed (between 3 and 6.6 m/min) above which striatal DA release increases, which could explain the results from the Castaneda et al. [101,116] study. Once above this (low) speed it seems that extracellular DA increases during, and about 40 to 60 minutes following, treadmill exercise. [106] Saboletal. [102] used a circular treadmill (speed of the treadmill 10 m/min) to let the animals turn and walk in place for sucrose water reward. The rats were fixed by their tails and walked in place for 24 minutes. Extracellular levels of DA and DOPAC were measured in the nucleus accumbens/medial striatum and lateral striatum. An increase in DA and DOPAC release in the lateral striatum was found. The authors were unable to conclude whether these changes were caused by motor activity, the act of drinking, the tail pinch stress or the amount of fluid consumed. Similar effects were presented by McCullough and Salamone, [103] who stimulated locomotor activity by periodic food presentation. The locomotor activity correlated weakly with the rise in extracellular DA in the nucleus accumbens, but again it was difficult to distinguish food presentation from exercise per se. Changes in brain extracellular GLU during movement stress were studied using in vivo microdialysis. [113] Male Long-Evans rats were placed in a clear cylinder designed to elicit behavioural activation while undergoing microdialysis sampling from either the hippocampus, striatum or sensorimotor cortex. Dialysate GLU levels were significantly enhanced in the samples taken while the rat was in the cylinder compared with samples taken before or after exposure to the cylinder. In a second study, [113] rats were implanted with bilateral probes in the forelimb sensorimotor cortex, and one forelimb was immobilised by means of a plaster of paris cast. GLU, aspartate, serine and taurine levels were quantified in casted animals. In casted animals, dialysate GLU levels were lower on the side contralateral to the immobilised limb during both quiescence and movement stress. Aspartate and taurine, but not serine levels increased during movement stress in both the side contralateral and the side ipsilateral to the immobilised limb. These results suggest that there is extracellular overflow of GLU and other neuroactive amino acids during spontaneous movement, and long term disuse can suppress extracellular GLU levels. [113] It has been shown that extracellular acetylcholine (ACh), regardless of the brain area sampled, responds by an increased output when the rat awakes or initiates locomotor output. [12] A correlation was established between the release of ACh and locomotor activity. Kurosawa et al. [105] used a treadmill that was manipulated manually at a low speed (2.3 m/min). The animals were restrained in a metal harness which was fixed to the chest and abdomen of the rat with plaster. Extracellular ACh, NA and serotonin in the parietal lobe of the cerebral cortex was examined. Walking for 5 minutes produced an increase of all 3 neurotransmitters. Recently, it was shown that treadmill running for 60 minutes significantly increased extracellular serotonin levels in the hippocampus of trained rats. [20,110] Pagliari and Peyrin [108,109] examined the effect of exercise on the in vivo cerebral release and turnover of NA in trained rats running on a treadmill for 60 minutes. The authors used a long term probe implantation in the frontal cortex. NA turnover and release increased during exercise and was even further increased when exercise time was prolonged to 2 hours of running. In their second study, [109] rats were trained during 2 weeks to run on a treadmill. Prior physical conditioning greatly

11 Brain Microdialysis During Exercise 975 Table II. Summary of studies that have used microdialysis techniques to investigate neurotransmitter release during exercise Reference Animal model Exercise Training Brain area Results Castaneda et al. [101] Long-Evans rats, Treadmill walking 1h at Stratium DA, DOPAC, HVA no n=20 3m/min increase Sabol et al. [102] Circular disk Lateral striatum DOPAC, DA= Treadmill: 24 min at 10 m/min, water reinforcement Medial striatum + nucleus accumbens, frontal accumbens McCullough & Wistar rats After food presentation nucleus accumbens DA Salomone [103] Hattori et al. [104] Kurosawa et al. [105] Hattori et al. [106] Meeusen et al. [21,22] Wistar rats, n = 34, TTL (4 groups) Wistar rats, male ( g), n=12 Wistar rats, n=8 Wistar rats, male ( g), n=10 Treadmill: 20 min at 18 m/min Treadmill: 2.3 m/min manually moved Treadmill: 20 min different speeds Treadmill: 20 min at 12 m/min Gerin et al. [107]a S-D rats, n = 7 Treadmill: 60 min at 16 m/min 5% incline Pagliari & Peyrin [108] S-D rats, n = 8-12 Pagliari & Peyrin [109] S-D rats, n=56 Wilson & Marsden [110] Meeusen et al. [20] Meeusen et al. [23] Gerin et al. [57] Gerin & Privat [111,112]a Lister-hooded rats, male ( g), n=12 Wistar rats, male ( g), n=24 Wistar rats, male ( g), n=16 S-D rats, male, n=6 Treadmill: 60 min at 25 m/min, 3% incline Treadmill: 1 or 2h at 25 m/min, 3% incline Treadmill: 60 min at 20 m/min Treadmill: 60 min at 12 m/min Treadmill: 60 min at 26 m/min Treadmill: 6 min at 26.6 m/min 7d accommodation 20 min, 18 m/min Striatum Parietal lobe of cerebral cortex DOPAC, DA = DA, DOPAC, HVA ACh, NA, serotonin 7d accommodation Striatum DA, DOPAC, HVA in relation to speed 2wk accommodation: 2 d/wk: speed and run time gradually increased Striatum 5 d/wk gradually Ventral funiculus increasing speed and time spinal cord 2wk, 5 days/week, 1h/d,25m/min, 3% incline 2wk, 5 d/wk, 1 h/d, 25 m/min, 3% incline 4wk accommodation: speed and run time gradually increased 2wk accommodation: 2 d/wk: speed and run time gradually increased 6wk training (30 sessions) 5 d/wk gradually increasing time and speed last week 80minat26m/min Accommodation 5 d/wk Frontal cortex Frontal cortex Hippocampus Hippocampus Striatum T: DA ; NA ; GLU ; GABA = Ventral horn, spinal cord Bland et al. [113] Long-Evans rats Walking cylinder Hippocampus, GLU striatum, motor cortex Gomez-Merino et al. [114] Gomez-Merino et al. [115] Wistar rats (300g) Wistar rats (300g) Treadmill: 120 min at 25 m/min Treadmill: 120 min at 25 m/min. Saline or valine injection DA, NA, serotonin, GLU, GABA= Serotonin, 5-HIAA, DA, MHPG NA, MHPG-DHPG NA, DHPG, MHPG Serotonin =in good runners Serotonin, 5-HIAA C: DA ; NA ; GLU =; GABA= No serotonin increase during exercise Accommodation 5-6 times Hippocampus; ventral serotonin and 5-HIAA gradually increased speed cortex in both brain areas Accommodation 5-6 times Hippocampus gradually increased speed serotonin and 5-HIAA, TRP, valine suppressed exercise-induced serotonin release a Method used was microdialysis (chronic implantation). ACh = acetylcholine; C = control animals; DA = dopamine; DHPG = 3,4-dihydroxyphenylethylene glycol; DOPAC = 3,4-dihydroxyphenylacetic acid; GABA = γ-aminobutyric acid; GLU = glutamate; HVA = homovanillic acid; MHPG = 3-methoxy-4-hydroxyphenylglycol; NA = noradrenaline; S-D = Sprague-Dawley rats; T = trained animals; TRP = tryptophan; TTL = total; 5-HIAA = 5-hydroxy indole acetic acid; indicates significant increase; = indicates no significant difference.

12 976 Meeusen et al. influenced central NA response: 1-hour trained rats experienced 2 hours running as extremely stressful, whereas the 2-hour trained animals exhibited a progressive sustained NA efflux. [109] Gerin et al. [57] used an interesting approach. To investigate the effects of exercise on spinal cord serotonin, these authors chronically implanted a microdialysis probe in the ventral horn of the lumbar spinal cord of rats. The probe was kept in place during 40 days. In the ventral horn, extracellular release of serotonin did not increase during 60 minutes of exercise. In a follow-up study [107] the dialysis probes were chronically implanted in the ventral funiculus of the spinal cord and significant increases of serotonin, DAand their metabolites [5- hydroxy indole acetic acid (5-HIAA), 3-methoxy- 4-hydroxyphenylglycol (MHPG)] were found during locomotion. To precisely define the relationship between descending monoaminergic systems and the motor system,thesameauthors [112] measured the variations of extracellular levels of serotonin, 5-HIAA, DA and MHPG in the ventral horn of the spinal cord of adult rats. Measurements were performed during rest, endurance running on a treadmill, and a postexercise period, with microdialysis probes implanted permanently for 45 days. They found a slight decrease in both serotonin and 5-HIAA during locomotion with a more marked decrease during the post-exercise period compared with the mean of rest values. In contrast, the levels of DAand MHPG increased slightly during the exercise and decreased thereafter. These results, when compared with those of a previous study, [107] which measured monoamines in the spinal cord white matter, highlight the complex regulation of the release of monoamines that occurs in the ventral horn. Ishide et al. [117,118] determined whether the effects of administering L-arginine into the rostral (RVLM) and caudal (CVLM) ventrolateral medulla on cardiovascular responses elicited during static muscle contraction, were mediated via an alteration of localised GLU levels. Results suggested that nitric oxide within the RVLM and CVLM plays opposing roles in modulating cardiovascular responses during static exercise via decreasing and increasing, respectively, extracellular GLU levels. [117] They further determined whether this RVLM-mediated opioidergic modulation of cardiovascular responses was associated with localised changes in extracellular levels of GLU. The results suggested that an opioidergic receptor-mediated mechanism within the RVLM attenuates cardiovascular responses during static exercise via modulating extracellular levels of GLU in the RVLM. [118] It is well established that in the absence of photic cues, the circadian rhythms of rodents can be readily phase-shifted and entrained by various nonphotic stimuli that induce increased levels of locomotor activity. Kalsbeek et al. [119] investigated whether restricted daytime feeding (RF) induces a phase shift of the melatonin rhythm, and found that prior entrainment to a nonphotic stimulus such as RF may phase lock the circadian oscillator and in that way hinder resynchronisation after a phase shift. [119] Another approach was used by Bequet et al., [120] who combined proton nuclear magnetic resonance spectroscopy and cortical microdialysis to observe intra- plus extracellular and extracellular brain glucose variations. These results confirmed the existence of a link between glucose level variations in peripheral and cerebral areas, but also showed that exercise increased extracellular brain glucose levels despite peripheral hypoglycaemia, suggesting a specific regulation mechanism of cerebral glucose metabolism during exercise. [120] Hasegawa et al. [121] examined the role of monoamines and amino acids in thermoregulation. They measured their levels in the preoptic area and anterior hypothalamus (PO/AH) in exercising rats, using an in vivo microdialysis technique. The data indicated that DAbreakdown processes in the PO/AH are activated during exercise. DA in the PO/AH maybeinvolvedintheheatlossmechanismsfor thermoregulation when body temperature rises during exercise. [121] Recently, a series of experiments were conducted using microdialysis for monitoring brain neuro-

13 Brain Microdialysis During Exercise 977 chemistry during exercise. In the first study, [21,22] rats exercised for 20 minutes (12 m/min) on a treadmill and striatal extracellular levels of DA, NA, serotonin, GABA and GLU were measured. Dialysates were analysed with HPLC with electrochemical detection for monoamines and GABA, and fluorescence detection for GLU. During the 20 minutes of exercise DA, NA, serotonin and GLU levels rose significantly above baseline. Extracellular GABA levels did not significantly change and fluctuated just above baseline values during the entire experiment. All neurotransmitter levels returned to baseline values at the end of the experiment (160 minutes after exercise). These findings indicated that exercise induces release of some neurotransmitters such as DA, NA, serotonin and GLU in the rat striatum, and has no significant effect on the release of GABA. Furthermore, it seems that there is an interaction between various neurotransmitters during exercise in regulating their release. [21,22] In a second study, [23] the in vivo microdialysis sampling technique was used to register extracellular levels of neurotransmitters in the striatum of trained and untrained rats. The influence of 1 hour of exercise on striatal release of DA, NA, GLU and GABA was investigated in trained and untrained rats. Male Wistars were randomly assigned to a training or control group. The exercise training consisted of running on a treadmill for 6 weeks, 5 days per week, running time and speed gradually increased from 30 minutes at 19 m/min during the firstweekto80minutesat26m/minduringthe final training week. Control animals were placed on the treadmill twice a week, and received a total of 4 adaptation sessions in which they exercised 15 to 45 minutes at 26 m/min. To evaluate the effect of the exercise training, soleus muscle citrate synthase activity was determined. Brain dialysates were analysed with microbore liquid chromatography with electrochemical detection (monoamines and GABA) and fluorescence detection (GLU). Soleus citrate synthase significantly differed between trained and control animals (respectively: 29.4 ± 7.0 units/g wet weight for trained and 14.3 ± 5.9 units/g wet weight for control animals). Basal levels of DA, NA and GLU were significantly (p < 0.05) lower for the trained compared with control animals. An hour of exercise significantly increased extracellular DA, NA and GLU levels in both control and trained animals (p < 0.05). There was no statistically significant difference in the exercise-induced increase between trained and control animals. There was also no statistical difference in basal or exercise-induced GABA levels between trained and control animals. The results indicated that exercise training appears to result in diminished basal activity of striatal neurotransmitters, while maintaining the necessary sensitivity for responses to acute exercise. [23] Another study [20] examined whether exerciseelicited increases in brain tryptophan availability (and in turn serotonin synthesis) altered serotonin release in the hippocampus of food-deprived rats. To this end, the respective effects of acute exercise, administration of tryptophan, and the combination of both treatments upon extracellular serotonin and 5-HIAA levels were compared. All rats received adaptation sessions to run on a treadmill before implantation of the microdialysis probe and 24 hours of food deprivation. Short term exercise (12 m/min for 1 hour) increased in a time-dependent manner extracellular serotonin levels, these levels returning to their baseline levels within the first hour of the recovery period. Acute administration of a tryptophan dose (50 mg/kg intraperitoneal) that increased extracellular 5-HIAA (but not serotonin) levels in fed rats, increased extracellular serotonin and 5- HIAA levels in food-deprived rats within 60 minutes. While serotonin levels returned toward their baseline levels within 160 minutes following tryptophan administration, extracellular 5-HIAAlevels rose throughout the experiment. Lastly, treatment with tryptophan (60 minutes beforehand) before short term exercise led to marked increases in extracellular serotonin (and 5-HIAA) levels throughout the 240 minutes that followed tryptophan administration.

14 978 Meeusen et al. This study indicated that exercise stimulates serotonin release in the hippocampus of fasted rats, and that pretreatment with tryptophan (at a dose increasing extracellular serotonin levels) amplifies exercise-induced serotonin release. It should be noted that in this study, none of the animals showed any sign of fatigue during the exercise session, although extracellular serotonin levels increased markedly, especially in the tryptophan and exercise trial. Furthermore, since it was shown that during exercise serotonin, DA, NA and GLU release increased in the striatum, [21,22] as well as serotonin release in the hippocampus, [113] without affecting the running capacity of the animals, the direct relationship between increased serotonin release and fatigue could not be established. These results were recently confirmed by Gomez- Merino et al. [114,115] who first examined the impact of short term intensive treadmill running (2 hours at 25 m/min, which is close to exhaustion time), on extracellular serotonin and 5-HIAA levels in 2 different brain areas in rats, the ventral hippocampus and the frontal cortex. [114] Hippocampal and cortical serotonin levels increased significantly after 90 minutes of exercise and were maximal in the first 30 minutes of recovery. Thereafter, cortical serotonin levels followed a rapid and significant decrease when hippocampal levels were still maximal. During exercise, changes in extracellular 5-HIAA levels paralleled serotonin changes, but showed no difference between the 2 brain areas during recovery. Thus, an intensive exercise induces a delayed increase in brain serotonin release but recovery seems to display site-dependent patterns. [114] In a second study, these authors [115] examined the effect of pre-exercise L-valine administration on serotonin metabolism in the ventral hippocampus of rats submitted to an acute intensive treadmill running. The data clearly demonstrated that exercise induces serotonin release in the rat hippocampus; in the control group, hippocampal serotonin levels increased from ± 6.4% at the end of exercise to ± 6.4% after 60 minutes of recovery. Moreover, 2 hours of intensive running induced significant increases in both extracellular tryptophan levels (from 120 minutes of exercise to 30 minutes of recovery) and 5-HIAA levels (from 90 minutes of exercise to 90 minutes of recovery). Pre-exercise administration of L-valine significantly prevented the exercise-induced serotonin release; serotonin levels were maintained to baseline during exercise and recovery. With regard to the competitive effect of L-valine with tryptophan, a treatment-induced decrease in brain tryptophan levels (from 120 minutes of exercise to the end of recovery) was observed. L-valine did not prevent an exercise-induced increase in 5-HIAA levels. This study demonstrated that a short term intensive exercise stimulates serotonin metabolism in the rat hippocampus, and that pre-exercise administration of L-valine prevents, via a limiting effect on serotonin synthesis, exerciseinduced serotonin release. In both studies there was no report of premature fatigue in exercising animals although serotonin levels increased significantly. [115] As it has been shown that exercise and training influences neurotransmitter release in various brain nuclei. It appears important to further link these findings with peripheral measures. Exercise training appears to result in diminished basal activity of striatal neurotransmitters, while maintaining the necessary sensitivity for responses to acute exercise. These observations raise the possibility that there could exist an exercise-induced change in receptor sensitivity. Apossible disregulation at this level could play a key role in the maladaptation to the stress of exercise, training and overtraining, which should also be explored in the future. 9. Conclusion There are numerous levels at which central neurotransmitters can affect motor behaviour, from sensory perception and sensory-motor integration to motor effector mechanisms. Until now, most studies were performed on homogenised tissue, which provided no indication of the dynamic release of neurotransmitters in the extracellular space of