| Additional related information may be found at: |
| Neuropsychopharmacology: The Fifth Generation of Progress |
A Critical Analysis of Neurochemical Methods for Monitoring Transmitter Dynamics
in the Brain
Janet M. Finlay and Gwenn S. Smith
Chemical techniques were first applied to the study of the nervous system in a systematic way in the late 19th Century. At that time "neurochemistry" meant describing the inorganic and organic constituents of neuronal tissue, an endeavor referred to as "chemical statics" by the parent of neurochemistry, J.W.L. Thudichum. There was some appreciation for the fact that the major task of the brain, the processing of information, was dynamic rather than static in nature. Yet for many decades the study of communication between neurons was mainly within the domain of electrophysiology, a tool which seemed to offer what neurochemistry could not: anatomical specificity and high temporal resolution. However, by the middle of the 20th century it had become clear that the great majority of the information transfer between neurons involved chemical signals, even in the brain, and neurochemical approaches began to be developed to monitor this process. Moving from bioassays to the use of high performance liquid chromatography, from whole brain analyses to microdissection, and from assays of postmortem tissue to assays of extracellular fluid sampled in vivo, continuing improvements in technology have made neurochemistry an ever more important partner in our efforts to understand the activity of the nervous system.
In this chapter we explore the neurochemical methods available to monitor the activity of specific neuronal systems in the central nervous system (CNS). We comment briefly on older methods, because some papers utilizing such methods remain an important part of the literature. However, our focus is on approaches that have undergone major development over the past decade (e.g., in vivo microdialysis and voltammetry) and on certain more established methods that are still in frequent use (e.g., release from tissue slices) (for additional reviews on these and related subjects see refs. 32, 53, 66, 75, 84). In considering these approaches we will critically evaluate them, noting where appropriate the distinction between the neurobiological variables of interest and what is actually measured in the laboratory. Often the examples chosen involve the catecholamines, norepinephrine (NE) and dopamine (DA), or the indoleamine, serotonin [5-hydroxytryptamine (5-HT)]. This is both because these biogenic amines are the transmitters with which the authors are most familiar and because at present a disproportionate amount of information is available regarding them. However, our focus on these examples should not detract from the generality of the themes that we wish to emphasize in this review: First, for any neurotransmitter there are many variables that one could hope to assess; each of these is presumably of biological significance, yet only a few are directly related to the two processes that are usually of greatest interest — transmitter release and the resulting postsynaptic response. Second, the plethora of feedback loops that participate in the synthesis, release, and degradation of a neurotransmitter or in its transduction to a postsynaptic event can be expected to complicate the task of monitoring transmitter dynamics because methods that influence these homeostatic processes can alter the variables under investigation, often making the proper interpretation of observations difficult.
Steady State Conditions
A major concern in post-mortem studies is the impact of death and postmortem delay on the variables under study (see Processes Underlying Sleep Regulation, Short-and-Long-Term Psychopharmacological Treatment Strategies, Brain Imaging in Mood Disorders, The Effects of Neuroleptics on Plasma Homovanillic Acid, Atypical Antipsychotic Drugs, The Pharmacotherapy of Acute Anxiety: A Mini-Update, Pharmacological Treatment of Depression in Late Life, and Parkinson's Disease, this volume). We will focus on another critical issue: the biological significance of measurements in postmortem tissue. In most such experiments animals are killed, either by decapitation or anesthesia, tissues are dissected, and the samples are homogenized in strong acid to break down cellular compartments and liberate the compounds of interest. These, in turn, are isolated and measured using one of many possible methods for quantification.
Transmitter content
The analysis of the amount of neurotransmitter present in tissue represented the earliest attempt to investigate the activity of chemically-identified neuronal systems. At first large regions of the CNS were examined; later improvements in dissection and in analysis permitted the measurement of increasingly smaller subdivisions of brain. Yet the assumption that changes in transmitter content would tell us much about the dynamics of cell-to-cell communication has proven incorrect, at least in most instances. Instead, a key aspect of transmitter dynamics appears to be "synthesis-secretion coupling," with the maintenance of constant transmitter stores in the face of changing rates of utilization (Fig. 1).
There are exceptions. For example, because the biosynthesis of neuropeptides occurs primarily in the cell body and preterminal axon, there is little capacity for rapid adjustments to changing rates of release (12, see Electrophysiology of Serotonin Receptor Subtypes and Signal Transduction Pathways, this volume). In addition, whereas acetylcholine (ACh) is synthesized in the nerve terminal at a rate that usually matches the rate of ACh utilization, ACh content in the neostriatum is inversely correlated with release (71). In general, however, the concentration of a given transmitter has no value as a measure of transmitter dynamics. Indeed, transmitter content is such a stable property of a given class of neurons that it has been used as an index of the number of neuronal elements present in the tissue.
Transmitter metabolites
Whereas the level of transmitter in postmortem neuronal tissue has not generally proven to be very useful, in the case of the monoamines considerable insight has been provided by measurement of transmitter metabolites (see Standard Antidepressant Pharmacotherapy for the Acute Treatment of Mood Disorders, this volume). For example, although stimulation of DA neurons within a physiological range does not deplete tissue of its transmitter stores, it does increase the level of the metabolite, dihydroxyphenylacetic acid (DOPAC) (47). As a result numerous insights have been gained from studies using tissue levels of DOPAC as an indirect measure of DA release.
However, in interpreting changes in tissue levels of neurotransmitter metabolites, several issues should be taken into consideration (see also ref. 85). First, there often exists more than one catabolic pathway with different pathways associated with different cellular compartments. In fact, metabolites can even be formed in intracellular compartments and need not reflect changes in transmitter release. For example, DOPAC can be formed from newly synthesized DA that has not been released, as well as from DA that has been released and taken back up into the terminal. Second, the metabolites that are formed, as well as their biological significance, can change with brain region as well as with the experimental animal. Third, neurotransmitter release and metabolism need not always co-vary. This is because transmitter that is released can be taken back up and reutilized without being subject to metabolism and furthermore, there can be a change in the formation and clearance rates of a metabolite without any change in the rate of transmitter release. For example, by blocking high affinity DA uptake as well as monoamine oxidase activity, amphetamine greatly reduces the rate of formation of DOPAC while increasing extracellular DA concentrations. Fourth, changes in metabolite concentrations may be difficult to detect under conditions in which there is a high background level of metabolite already present. For example, changes in choline and acetic acid concentrations that occur during the release and catabolism of ACh may not be readily detected against the high background of preexisting choline and acetic acid present in CNS. The same may be true of glutamine formed from glutamate and succinic semialdehyde formed from gamma-aminobutyric acid (GABA). Finally, because the transmitter content of a nerve terminal is thought to be relatively constant (see above), metabolic data are often presented as a ratio of metabolite to transmitter concentrations in the tissue. However, such ratios can be misleading if the transmitter content of the tissue is affected by the treatment conditions. Thus, transmitter and of metabolite concentrations should also be provided separately.
In summary, a large body of useful information has been obtained by measurement of transmitter metabolites in postmortem tissue, particularly in the case of the biogenic amines. In interpreting such data, however, it is important to remember that the variable being measured is quite removed from release per se, and that the significance of particular metabolic routes may vary as a function of the specific transmitter system, the species, and the experimental conditions.
Conversion of Radiolabeled Precursors to a Transmitter
Since the 1960s, transmitter synthesis has been studied extensively by introducing a radiolabeled precursor into the CNS and measuring the accumulation of a radiolabeled product (35). Because transmitter release usually is coupled with transmitter synthesis, investigators often monitored synthesis and use this information to predict changes in the rate of release. As with all of the methods discussed in the present chapter, interpretation of data generated using this approach requires knowledge of the limitations of the method itself (see also refs. 18, 91). In particular, it is important to recognize that the radiolabeled precursor may not be handled in a manner identical to that of the natural precursor. In fact, each of the known transmitters is synthesized from a precursor that itself can be derived from multiple endogenous sources. For example, tyrosine used in the synthesis of catecholamines can be taken up from plasma, synthesized from phenylalanine, or formed through the degradation of protein. In addition, the amount of a precursor used for synthesis of a transmitter may be small relative to that used in other biochemical pathways. As a result, the rates of these other biochemical pathways may have a significant influence on the availability of a precursor for transmitter synthesis. For example, a fraction of the available tyrosine is used to form catecholamines, with the majority of tyrosine being incorporated into protein or transaminated into p-hydroxyphenylpyruvic acid. Finally, because newly synthesized transmitter not only accumulates, but also is utilized, synthesis rate can only be estimated from a determination of the accumulation of radiolabeled product during the earliest stages of an experiment, or through the use of multiple data points and complex equations.
Transmitter-synthesizing enzymes
In some cases, changes in transmitter release are accompanied by changes in the activity of one or more transmitter-synthesizing enzymes. An example of this is the increase in tyrosine hydroxylase (TH) activity that often occurs in catecholaminergic neurons during conditions of increased impulse flow (see Long-and-Short-Term Regulation of Tyrosine Hydroxylase, this volume). These changes in enzyme activity can occur rapidly, as a result of modification of existing TH molecules, or more slowly due to a change in the rate of formation, degradation and/or delivery of TH molecules. Such changes can be detected by measuring the kinetics of the reaction facilitated by a specific enzyme and the amount of enzyme protein. In addition, if the concentration of biosynthetic enzymes can be altered during changes in transmitter synthesis, the concentration of the relevant mRNAs might be expected to be altered as well, and can sometimes be of value in detecting longer-term changes in transmitter synthesis. Moreover, because specific mRNAs usually can be monitored by in situ hybridization, histochemical techniques can be used to provide a neuroanatomical precision that is not otherwise available in neurochemical studies.
However, the presence or absence of changes in transmitter synthesizing enzymes and mRNAs must be interpreted with caution. There are ways in which transmitter biosynthesis might be altered that are less readily monitored; for example, the delivery of some constituent of the reaction could change as could the availability of an inhibitor or an activator of the biosynthetic process. Thus, the absence of a change in the isolated form of a biosynthetic enzyme itself need not indicate that transmitter synthesis is unchanged. An example of this phenomenon involves the synthesis of the transmitter ACh. In this case, the synthesizing enzyme choline acetyltransferase does not appear to be subject to short-term covalent modulation. On the other hand, the high-affinity transporter that provides choline, a precursor for ACh, often is responsive to the rate at which the axon terminal is depolarized, and choline transport can sometimes be used to monitor ACh turnover (72; although see also ref. 71).
Non-steady State Conditions
As discussed previously, postmortem analysis indicates that transmitter levels often remain constant even in the face of changing rates of utilization. However, transmitter levels have been used to monitor transmitter dynamics under conditions where a pharmacological intervention is coupled with the measurement of transmitter level. For example, transmitter synthesis might be blocked and the rate of disappearance of transmitter determined. Alternatively, catabolism might be inhibited and transmitter accumulation monitored. Whenever steady state conditions are disrupted, one must be concerned about the physiological significance of the resulting observation. One reason for this derives from the existence of numerous feedback mechanisms, which can cause a treatment that modifies one step in neurotransmission (e.g., release) to induce changes in other steps as well (e.g., synthesis). Because this problem often cannot be avoided, it is important to use more than one method to estimate the variable of interest, determining whether approaches that disturb the system in different ways nonetheless yield similar results. Under conditions where feedback mechanisms have been disrupted, measurements of transmitter dynamics may be more useful for rough, qualitative estimates than for more precise quantitative conclusions. A second potential source of problems results from the fact that newly synthesized transmitter often appears to be released preferentially.
The first attempts to overcome the problems imposed by synthesis-secretion coupling by pharmacological means involved the biogenic amines. Several approaches were taken, including inhibiting transmitter metabolism and monitoring the accumulation of the amine under study (61); inhibiting biogenic amine synthesis at the initial, rate limiting step and determining the rate of transmitter disappearance (11); and inhibiting amine synthesis at the second step, decarboxylation of dopa or 5-hydroxytryptamine, and measuring the accumulation of precursor (15). Despite the problems with non-steady-state techniques, these approaches proved to be an important advance. For example, using α-methyl-p-tyrosine, an inhibitor of tyrosine hydroxylase, it was soon shown that certain stressors altered the rate of disappearance of NE, although stress-induced changes in NE content had not been detected without drug pretreatment. Several of these pharmacological approaches remain in use today.
An alternative to post-mortem analysis is to use an in vitro preparation in which the processes of synthesis and release are ongoing and can be monitored. One of the first such preparations was the tissue slice, introduced in the 1930s as a method for studying the respiration of central neural tissue. Over the years other in vitro preparations have been developed, including explants, isolated nerve ending fractions or synaptosomes, and cultured cells. Although, in this section we will focus on the use of brain slices, many of our comments are relevant to these other preparations as well.
The methods for preparing and incubating brain slices have been reviewed elsewhere (49). In most cases, the slices are prepared using a mechanical device, such as a tissue chopper or "vibratome," although in earlier experiments slices were prepared free hand or with a simple guide. Slices are generally 300-500 um in thickness and they are incubated in a buffered iso-osmotic salt solution. Transmitter release can be triggered by depolarizing the slice and the incubation medium collected for analysis of transmitter content.
Initial studies involved monitoring the efflux of radiolabeled tracers. As in the case of studies in which tracers are administered in vivo (see above) care must be taken in interpreting such results. Whereas radiolabeled tracers are still used in some studies, in other cases, analytical techniques have advanced to the point where one can measure the endogenous, non-radiolabeled transmitter directly. In some cases these analytical methods are combined with the use of drugs that inhibit the inactivation of the transmitter in question. For example, acetylcholinesterase inhibitors might be used to block the degradation of ACh or an inhibitor of the DA transporter might be employed to inhibit the reuptake of DA. Whereas the objective in such cases is to increase the sensitivity of the method, the danger is that which is inherent in any study involving a disturbance of the usual life cycle of a transmitter - the possibility of interfering with a normal homeostatic process. Thus, inhibitors of ACh hydrolysis or DA reuptake will increase extracellular concentration of the respective transmitter and may thereby influence synthesis and/or release modulating autoreceptors.
A variety of stimuli can be used to examine the response of the brain slice to depolarization. The most common being elevated K+ concentrations; others include veratridine and ouabain. Although each of these conditions causes depolarization by a different mechanism, they share a common drawback. Whereas a piece of neuronal membrane is normally in a depolarized state for less than 10% of the time (assuming a 1-msec action potential and a frequency of less than 100 Hz), each of these approaches cause a continuous state of depolarization. Although it is never possible to mimic the physiological state with regard to the characteristics of the stimulus or the number of neurons affected, a major advance in this direction is provided by the use of electrical field depolarization, an approach which permits the use of a wide variety of frequencies, pulse durations, and patterns. On the other hand, any procedure that involves exposing an entire slice to depolarizing conditions will produce a highly unphysiological mix of transmitters in the extracellular compartments, and thus interactions that bear little resemblance to the normal state. Despite these problems, the use of in vitro preparations such as the tissue slice has provided many important insights that were not available from biochemical analyses of tissue. For the first time it was possible to measure the transmitter content of an extracellular compartment. Moreover, serial measurements could be made on the same tissue sample, increasing both the power and the efficiency of experiments.
Although in vitro preparations can provide insights not readily available from analysis of post-mortem samples, they utilize tissue that has had both neural and vascular connections transacted and is incubated in a fluid that is at best only a rough approximation of extracellular fluid. For a more accurate index of transmitter dynamics under physiological conditions, one must move to an in vivo preparation. In this section we will focus on several approaches that have been used to monitor extracellular levels in intact animals (Fig. 2). Particular attention will be paid to two methods whose use has grown dramatically over the past several years: microdialysis and voltammetry.
Cortical Cup
The first in vivo preparation developed for sampling the extracellular environment was the cortical cup (Fig. 2a). As early as 1953, the cortical cup was used to demonstrate that ACh release was related to the spontaneous electrical activity of the cortex (51). With this method the skull overlaying the cortical area of interest is removed and a cylinder is placed in tight contact with the cortical surface (for details of the method see ref. 57). The cylinder is then filled with a buffered solution comparable to that used with in vitro preparations. The fluid is collected (either by continuous flow or sequential washes), and it is analyzed for the compounds of interest. This method has been used to study factors regulating neurotransmitter efflux from cortex of both anesthetized and unanesthetized preparations (59). In many ways the procedure is analogous to the in vitro slice preparation except that the superfusion is performed in the intact living organism. However, the method has limited applications because it can only be applied to the cortical surface.
Push-Pull Cannula
Development of the push-pull cannula method in the early 1960s provided the first opportunity to evaluate neurochemical events in discrete structures deep within the intact CNS (30, 31). Present forms of the method developed from the merging of two independent push-pull cannula techniques; one designed to monitor cerebrospinal fluid (7, 8) and the other to sample the chemical environment of subcutaneous tissue (28).
Over the years, many refinements of this method have appeared. In general however, the push-pull cannula consists either of two concentric tubes, as originally designed (30; Fig. 2b), or of two parallel tubes (21). The cannula, is implanted into the brain so that the tip is located in the brain structure of interest. A perfusion solution is forced down one of the tubes and pulled up the other tube using two pumps working in synchrony as pushing and pulling devices. The push-pull method can be performed in an acute anesthetized preparation or in freely moving animals, either by implantation of the push-pull cannula directly or by implantation of the guide cannula and subsequent implantation of the push-pull cannula. Although the method is conceptually straight-forward, its proper use requires very close attention to details associated with the perfusion itself (32). For example, the distance between the tips of the inner and outer tubes of the concentric cannula generates a siphon and must be accurately adjusted in order to obtain a constant flow rate and avoid having tissue being pulled into the perfusion medium. In addition, the rate of push and pull of the perfusion solution must be carefully adjusted in order to reach a constant flow. The perfusion solution, once collected, can be analyzed by any suitable analytical method, providing a way of continuously sampling the microenvironment of the brain under in vivo conditions.
One major disadvantage of the push-pull method is the tissue perturbation produced by directly infusing a perfusion solution into the area from which neurotransmitter efflux is being determined (32). A related problem is the damage produced by the cannula itself, disrupting as it does a large region of brain in the very center of the area of measurement. Nonetheless, the method has been of great value in the study of transmitter dynamics and continues to be used by a number of research groups to examine the release of endogenous ACh, catecholamines, amino acids, and peptides, as well as efflux of radiolabeled transmitters following administration of labeled precursors.
Microdialysis
Several years after the introduction of the push-pull cannula technique, it was proposed that the principles of dialysis could be applied to a method for sampling the extracellular fluid of brain, thereby circumventing the problems associated with having a perfusion solution coming into direct contact with brain tissue. This method involves introducing a dialysis membrane between the perfusion solution and the extracellular fluid. Molecules that are sufficiently small to pass through the dialysis membrane will then diffuse across the membrane from an area of high concentration to an area of low concentration. In a first attempt to dialyze the extracellular fluid, Bito et al. (9) implanted dialysis sacs into the cerebral hemispheres of dogs; after several days, these sacs were removed and the amino acid levels of the dialysate determined. A few years later, Delgado and collaborators (22, 23) developed the dialytrode, a push-pull cannula with a small permeable bag attached to the end. Relative to the dialysis sac, this method offered the advantage of providing an opportunity for continuous sampling of the extracellular environment. These early attempts to dialyze the extracellular fluid of brain laid the foundation for subsequent work in this area. However, popularization of this method can be attributed to the seminal work of Ungerstedt and colleagues (78, 79).
Ungerstedt began by inserting dialysis tubing transversely through brain tissue and measuring radiolabeled DA efflux (80). Subsequently, endogenous DA also was measured (81, 90), primarily due to the parallel development of an analytical method with sufficient sensitivity for detection of endogenous catecholamines, high pressure liquid chromatography with electrochemical detection developed by Adams and colleagues (46). When first introduced, the microdialysis probe consisted of a hollow tube of dialysis membrane inserted transversely through a region of brain (Fig. 2c). Such transverse probes are simple to make, but cause severe damage to muscle and skull, as well to an extensive area of CNS through which they pass. The development of the loop probe provided a means of reducing the extent of surgically induced injury. This probe consists of a loop of dialysis membrane which is implanted vertically into the brain via a single hole in the skull (Fig. 2d). Still less damage is produced by a vertical concentric style dialysis probe (Fig. 2e). This probe consists of a single piece of dialysis tubing blocked off at one end with glue; the inlet and/or outlet portions of the probe pass down into the dialysis tubing. Whereas these probes are technically more difficult to make, their smaller overall diameter (typically about 200 um) and subsequent reduction in tissue damage has resulted in their widespread use (3).
The advantages of microdialysis over earlier in vivo methods have been summarized (60, 78, 88). Most of these advantages are directly related to the presence of the dialysis membrane at the tip of the microdialysis probe. Because the membrane prevents the perfusion solution from directly contacting the tissue, this provides a barrier to turbulence and infection. As with any invasive technique, mechanical injury to CNS tissue occurs during surgical implantation of a microdialysis probe. Results of a recent study indicate that careful surgical procedure can be a critical factor in ensuring the success of this method (20). Nonetheless, several hours after implantation of a microdialysis probe, efflux of a number of transmitters appears to be almost entirely due to physiological processes as demonstrated by its sensitivity to manipulation of extracellular Ca2+ and blockade of voltage-sensitive Na+ channels (87, however, see also 77). In contrast, this has not been established for the use of push-pull cannulae. Indeed, the blockade of Na+ channels only partly inhibits the release of DA from striatum as measured by push-pull cannula and furthermore stimulates DA release from the substantia nigra (62). This suggests that mechanical disruption of the tissue, such as that produced by turbulence, plays a continuing role in producing the transmitter efflux measured by the push-pull method.
The dialysis membrane also acts as a filter to prevent the diffusion of large molecules from extracellular fluid into the perfusion medium. This provides certain advantages for the analysis of transmitter content in the dialysate. First, the membrane can prevent large molecules such as enzymes from entering the perfusion solution and thereby halt the continuous enzymatic degradation of neurotransmitters once they have entered the perfusion solution. Also, by virtue of its ability to exclude molecules from the perfusion solution, the membrane partially purifies samples prior to their analysis. The presence of the dialysis membrane also provides certain technical advantages. For example, the size of the perfused area can be controlled by limiting the active surface of the membrane. In addition, the perfusion is greatly simplified because it is not necessary to adjust inflow and outflow accurately in order to prevent build up of pressure or clogging of the cannula as with the push-pull technique.
Of course, implantation of the dialysis probe results in several reactions within the CNS tissue. Knowledge of the time course of these events is critical in determining the interval during which microdialysis experiments can be performed with minimal interference from tissue reactions (2,3,66). In general it is thought that dialysis experiments should not be performed either very soon (£ 10 h) or very long (several days to weeks) after probe implantation and that the optimal interval for performing microdialysis experiments is approximately 16-48 h after implantation of the dialysis probe. Efforts have been made to develop methods whereby sampling can be carried out over many days in a single subject using either chronic implantation of a dialysis probe or implantation of a guide cannula followed by multiple insertions of a probe over days. However, these have generally been unsuccessful (65). In addition to the post-operative time at which samples are collected, other important variables include the ionic composition of the dialysate, and the rate of perfusion. It is clear that the composition of the dialysate should mimic the ionic constituents of extracellular fluid in brain as closely as possible (4,56,63,86), and that the flow rate of the dialysate should be as low as possible (usually 0.1-1.5 ul/min) (3).
In vivo microdialysis undoubtedly provides information about qualitative changes in the extracellular concentration of a neurochemical. In addition, methods for establishing quantitative information regarding extracellular concentrations have recently been developed. Originally, extracellular concentrations of molecules were estimated based on the in vitro recovery of the dialysis probe, that is, the concentration of a compound that appeared in dialysate as a percentage of the concentration present in a beaker into which the probe had been placed (81). However, because the mass transport of substances in brain tissue is very different from that in an aqueous solution this led to errors (5). Subsequently, many approaches have been developed to generate quantitative data using the microdialysis technique. These approaches can be divided into those based primarily on a theoretical approach and those employing an empirical approach (43). Recently, empirical approaches to estimating in vivo concentrations have been developed that involve extrapolation to zero flow (40) or calculation of the equilibration concentration, also referred to as the method of no net flux (50); these approaches have been applied extensively to the study of extracellular DA levels in CNS (41).
In vivo microdialysis is itself only a sampling technique. The ability to measure compounds within dialysate is entirely dependent upon the sensitivity of an appropriate analytical method. As in the case of in vitro preparations, under conditions when the analytical method is not sufficiently sensitive to detect dialysate levels of a particular compound, several approaches have been taken. Pharmacological tools have been used to compensate for inadequate sensitivity by increasing the level of substance to be analyzed. For example, acetylcholinesterase inhibitors have been used to enable detection of ACh in dialysate (19, although see also ref. 20). Another approach has been to prelabel neurons by infusing isotopes of the transmitter or precursors and assaying the radiolabeled compounds (48). The problems with these approaches have been previously discussed. In some instances, such as the study of endogenous DA in extracellular fluid of certain brain regions, it has become fairly routine to detect dialysate levels of neurotransmitter using high-pressure liquid chromatography coupled with electrochemical detection. This analytical method generally requires the collection of several microlitres of dialysate, and therefore it cannot be used to answer questions about rapid changes in extracellular fluid concentrations of a compound. However, recent advances in microbore liquid chromatography and capillary electrophoresis are enabling investigators to monitor transmitters in nanolitre volumes, dramatically increasing the temporal resolution of this method (17, 44). Finally, in cases of compounds such as GABA and glutamate, where both neurotransmitter and non-transmitter functions are being subserved, one must interpret results with particular caution (77).
Despite the several concerns that have been noted, use of microdialysis in neuroscience has increased rapidly since its development. Its popularity can be attributed to the fact that, relative to earlier methods, microdialysis does offer several distinct advantages, including (a) an opportunity to examine changes in the concentration of compounds in extracellular fluid derived from anatomically discrete brain regions with minimal disruption of the tissue and the blood-brain barrier and (b) the ability to sample extracellular fluid on a continuous basis. Microdialysis is now used extensively for the study of several transmitters in the CNS. In addition to its widespread use for measurement of endogenous catecholamines, laboratories are using this approach to monitor amino acid neurotransmitters (3, 77), ACh (20), serotonin (67), peptide neurotransmitters (52), and neurotrophic factors (39). The method also has been used to examine the efflux of cyclic adenosine monophosphate in extracellular fluid to provide a measure of in vivo receptor function (27). In addition to its extensive use in experimental animals, the method has been used to a limited extent in clinical neuroscience to monitor extracellular fluid in human brain during focal brain ischemia and seizures (36, 38).
Voltammetry
Adams and his colleagues (1,45) published the first report of an in vivo electrochemical method which allows the direct monitoring of a specific subset of molecules within extracellular fluid of brain. Voltammetry makes use of the fact that certain compounds are readily oxidizable. The method typically employs a working electrode, a reference electrode, and an auxiliary electrode. These electrodes are positioned in electrical continuity with one another, the working electrode being positioned in the brain structure of interest. A controlled potential is then applied between the working and reference electrode and the resultant current that flows from the working electrode provides a measure of the amount of electroactive material in the solution.
In most in vivo voltammetric experiments, the applied voltage is not held at a constant value but rather consists of a waveform generated by pulsing a voltage (chronoamperometry), applying a voltage sweep (cyclic voltammetry), or a combination of the two (differential pulse voltammetry). Those methods which alter the applied potential during the detection interval (i.e., cyclic voltammetry) provide information about the chemical identity of the compound being oxidized, whereas those that employ a single voltage pulse during the detection interval do not (i.e., chronoamperometry) (42,54). In addition, the various voltammetric methods differ in terms of their sensitivity and time resolution.
The design of voltammetric electrodes has changed considerably since the first use of this method leading to reduction in size and improvements in selectivity. A major technical advance has been the development of electrodes that range in size from 1-30 um diameter, and provide the best available spatial resolution of any method for studying neurochemical events in vivo. In contrast, improving the selectivity of voltammetric electrodes has been a much more difficult and less successful objective. Voltammetry makes use of the fact that several neurotransmitters, including DA, NE and 5-HT, are readily oxidizable at potentials at which most other compounds in the extracellular fluid do not oxidize. However, even within this range of potentials there are other molecules that will oxidize and thereby complicate the detection of these neurotransmitters. For example, oxidation of DOPAC and ascorbate occur within the same voltage range as the catecholamines, and under basal conditions both DOPAC and ascorbate are present in extracellular fluid at much higher concentrations than the catecholamines. In methods involving the collection of extracellular fluid prior to analysis, it is possible to separate molecules with similar oxidation potentials from one another by high-pressure liquid chromatography before they reach the working electrode. However, separation of this sort cannot be achieved with voltammetry alone, and therefore the molecular specificity of a particular in vivo voltammetric technique must be demonstrated convincingly in order to validate the method (89).
At present, the best methods for improving the selectivity of in vivo voltammetry involve modification of the electrode surface itself (55). For example, coating electrodes with the negatively charged polymer Nafion increases the selectivity of the electrode for cations (such as DA and NE) relative to anions (such as ascorbate and DOPAC) by reducing the likelihood that the latter compounds will gain access to the surface of the working electrode (34). Improved selectivity of voltammetric methods has also been achieved by other means. First, because the neurochemical content of different regions of the brain vary considerably, a certain amount of selectivity can be achieved by selection of an appropriate target site for the voltammetric electrode. For example, by placing an electrode in the striatum, which has a high DA content but very little NE, one can usually be confident that most of the current generated is not due to oxidation of NE. Second, pharmacological manipulations of compounds in the extracellular environment have also been used to enhance selectivity. For example, pargyline pre-treatment has been used to prevent formation of DOPAC. This permits the detection of DA without interference from its principal metabolite, although it creates the problems inherent in pharmacological pretreatments that have already been discussed. Improved selectivity may also be attained by restricting the in vivo voltammetric determination of a compound to an interval during which the axons of the neurons of interest are directly stimulated. Under these conditions, the immediate increase in current should be due to release of the parent neurotransmitter. However, the approach is limited by the fact that transmitter release can only be determined under conditions of electrical stimulation of axons or cell bodies, requires the presence of a discrete site for the selective stimulation of a specific set of neurons, and may not reflect the type of response elicited by more physiological increases in neuronal firing.
Comparison of Dialysis and Voltammetry
Microdialysis and voltammetry each makes a unique contribution to this field of study as a result of differing temporal and spatial resolution (Fig. 3), chemical resolution, and sensitivity (88). For example, voltammetric measurements of compounds in the extracellular fluid have been made at intervals as short as 100 ms, whereas dialysis measurements represent an integration of changes that have occurred over a 5- to 20-min interval. As a result, voltammetry and microdialysis are best suited to examination of momentary versus long lasting changes in the content of extracellular fluid, respectively. Recent advances in analytical chemistry, allowing for the analysis of small volumes, have greatly improved the temporal resolution of microdialysis (see above). However, these methods have not yet seen widespread use. Voltammetric electrodes sample from a very small area as their size ranges from 1-30 um in diameter. In fact, it has been demonstrated that a 10-um carbon fiber electrode can be used to sample DA from a distance of less than 10 um from the electrode surface (42). In contrast, dialysis probes are 200-300 um in diameter or larger. Moreover, due to continuous removal of the substances from the extracellular as a result of perfusion through the probe, concentrations gradients are set up that may extend several millimeters into the tissue (13). As indicated previously, the chemical resolution of voltammetric electrodes has improved greatly in recent years such that a number of compounds can be detected with confidence. However, to date, the use of voltammetry has been limited to the detection of readily oxidizable species. In contrast, the range of chemicals that can be examined using in vivo microdialysis is limited only by the ability of a given compound to pass through the dialysis membrane and the availability of an appropriate analytical technique. As a result, in vivo microdialysis has seen a wider range of applications than in vivo voltammetry. The sensitivity of in vivo voltammetry varies depending upon the electrode and the voltammetric method. However, in general, this method is not sufficiently sensitive to measure basal extracellular levels of neurotransmitters and thus as a result, cannot be used to examine the impact of treatments thought to decrease extracellular neurotransmitter levels. In contrast, basal levels of catecholamines are readily detectable using in vivo microdialysis. In summary, it is clear that these methods are capable of providing complementary rather than redundant information about the behavior of molecules in the extracellular fluid.
POSTSYNAPTIC MEASURES OF NEUROTRANSMITTER DYNAMICS
Thus far, our discussion has focused on methods for studying the presynaptic components of neurotransmitter dynamics including transmitter metabolism, release, and uptake. However, chemical neurotransmission also involves the interaction of a transmitter with a receptor protein. Under many conditions presynaptic measures of chemical transmission provide a reliable indicator of the level of transmitter-receptor interaction however, this may not always be the case. For example, in vivo voltammetry and microdialysis have been used extensively to monitor extracellular neurotransmitter concentrations. The extra-synaptic transmitter that can be detected using these methods may provide a measure of intra-synaptic transmitter concentrations and, in turn, the level of transmitter-receptor interaction (Fig. 4a). Increased release of transmitter will result in increased transmitter-receptor interactions in the synapse and increased concentrations of transmitter in the extracellular space as detected using voltammetry and microdialysis (Fig. 4b). However, extra-synaptic and intra-synaptic transmitter concentrations may not always co-vary, particularly under conditions of pharmacologically induced changes in transmitter dynamics. For example, d-amphetamine elicits robust increases in extrasynaptic dopamine concentrations due to its action in reversing the dopamine transporter. However, synaptic concentrations may be relatively unaffected or perhaps even decreased, as a result of depletion of vesicular dopamine in the nerve terminal (Fig. 4c). Thus, studies of neurotransmitter dynamics must also invoke methods for evaluating the impact of a transmitter on its target.
The postsynaptic consequences of neurotransmitter release have been assessed on the basis of alterations in measures such as electrophysiological activity, receptor occupancy and number, specific gene products, second messengers, and deoxyglucose utilization. Below we discuss two such approaches, the use of positron emission tomography (PET) to study transmitter-induced changes in receptor occupancy and the use of c-fos as a measure of transmitter-induced alterations in gene expression.
Positron Emission Tomography
Due to the invasive nature of the methods described above, the study of neurotransmitter dynamics has largely been confined to animal models. However, recently an approach has been developed to measure changes in transmitter-receptor interactions in the living non-human primate and human brain. This approach uses positron emission tomography (PET) to quantify binding of radiotracers that have been developed to occupy specific receptor proteins in brain. In vivo, endogenous transmitter can compete with the radiotracer for receptor occupancy and as a result, radiotracer binding may also provide an indirect measure of transmitter release. Specifically, an increase in transmitter levels should decrease radiotracer binding and a decrease in transmitter levels should increase radiotracer binding (see also Chapter 76, this volume).
To date, research using this method has focused primarily on central DA-containing neurons. This is in part, due to the availability of [11C]-raclopride, a suitable radiotracer for the DA D2/D3 receptor. This radiotracer exhibits several of the properties required for use in PET imaging studies: (1) a high ratio of specific to non-specific binding in the target brain regions, (2) a high degree of selectivity for a specific receptor system, (3) a regional pattern of binding that corresponds with the known receptor localization, (4) reversible binding to receptors, (5) low test-retest variability, and (6) an affinity for the receptor site is comparable to that of the neurotransmitter itself. Results of preclinical studies indicate that [11C]-raclopride binding in striatum meets these criterion. In addition, preclinical studies confirmed that drug-evoked increases in extracellular DA concentrations are correlated with a decrease in striatal [11C]-raclopride binding, suggesting that radiotracer binding can be used as a measure of extracellular transmitter concentrations (25, 82).
This approach to studying transmitter dynamics has been extended to clinical studies. The results of such studies indicate that in schizophrenic subjects, administration of d-amphetamine produces a greater decrease in striatal [11C]-raclopride binding than in control subjects (10). The latter observations suggests that the amphetamine-induced increase in extracellular DA is potentiated in schizophrenic subjects. However, administration of methylphenidate to detoxified cocaine abusers produces a smaller decrease in [11C]-raclopride binding than in control subjects, suggesting that prior experience with cocaine attenuates the amphetamine-evoked increase in extracellular DA concentrations (83). This method has also been applied to the study of interactions between transmitter systems in human brain. For example, a recent study examined the effects of a 5-HT releasing agent and uptake inhibitor, fenfluramine, on [11C]-raclopride binding in human control subjects (73). Fenfluramine administration resulted in decreased striatal [11C]-raclopride binding, suggesting that the drug-evoked increase in 5-HT concentrations resulted in a secondary increase in extracellular DA concentrations (Fig. 5). A similar approach has also been used to examine GABA-DA and ACh-DA interactions in neostriatum of human subjects (24, 26).
Whereas brain imaging provides novel information about transmitter dynamics in nigrostriatal DA neurons in the human brain, several factors hinder the ability to image other transmitter systems. The limited spatial resolution of this approach makes it difficult to image transmitter dynamics in small nuclei such as the cell body region of most chemically-defined pathways. In addition, the level of sensitivity is not adequate for visualizing brain regions that have a low density of receptors, as is the case for DA D2/D3 receptors in cortex. Finally, the method can only be applied to the study of transmitter systems for which well-characterized radiotracers are available. As a relatively new and very powerful method for study of transmitter dynamics, considerable emphasis has been placed on further development of PET imaging. For example, efforts are being made to develop a DA D1 radiotracer that is sensitive to changes in DA concentrations with the hope that this ligand can be used to visualize transmitter dynamics in both striatum and cortex. In addition, a recent study has established the test-retest variability of the radiotracer [F18]-altanserin as a first step in using this radiotracer to monitor extracellular 5-HT concentrations in human brain (74).
Gene Expression
The interaction of a transmitter with its receptor can elicit a cascade of biochemical events in the intracellular compartment of the postsynaptic neuron. For example, physiologic and pharmacologic stimuli have been shown to increase the expression of immediate early genes in the central nervous system (see also Cholecystokinin, this volume). Much of this work has focused on the immediate early gene c-fos, making use of immunostaining and in situ hybridization to measure changes in c-fos protein and mRNA, respectively (for a review of immunostaining and in situ hybridization methods, see Cytology and Circuitry of this volume). One advantage of this approach is that changes in c-fos expression can be used to identify individual neurons as well as neural circuits that are activated by a particular stimulus (37).
The limitations of using gene expression as a measure of chemical neurotransmission have been summarized in a recent review paper (16). One important consideration is that c-fos is not expressed in all neurons. As a result, the absence of c-fos expression does not necessarily indicate that neurons have not been activated by a pharmacologic or physiologic stimulus. Interpretation of c-fos data is also complicated by the fact that numerous transmitters as well as hormones can alter c-fos expression, making it difficult to determine the precise relationship between extracellular stimulation and expression of c-fos. For example, activation of DA and NE receptors has been observed to induce c-fos expression (64, 76). As a result, in brain regions innervated by DA and NE fibers, c-fos expression may reflect increases in either or both transmitters. An approach that has been used recently to overcome the latter limitation, is to combine in vivo monitoring of extracellular transmitter concentrations with the evaluation of c-fos expression. Investigators have used in vivo microdialysis in conjunction with c-fos immunohistochemistry to examine the relationship between drug-evoked alterations in extracellular DA and c-fos expression (14). Approaches such as this, involving both pre- and post-synaptic measures of neurotransmitter dynamics in a single preparation, can be technically difficult. None the less, there is continued interest in bringing methods such as in vivo voltammetry and microdialysis together with electrophysiology, gene expression, and PET imaging. The outcome of such efforts should greatly enhance our understanding of transmitter dynamics in chemically-defined neuronal circuits.
In this chapter we have discussed the advantages and limitations of a number of methods for monitoring transmitter dynamics in the central nervous system. It is not our view that any single method is best suited to the study of chemical neurotransmission but rather, we maintain that a synthesis of the data generated using multiple approaches provides the most accurate account of this complex process. Given this, it is important to reinforce that data generated using any of these methods must be interpreted within the limitations of the methods themselves. Therefore, our goal has been to focus on issues relevant to the chemical specificity, spatial resolution, and temporal resolution of methods used to monitor transmitter dynamics.
Although the focus of this chapter has been to critically evaluate specific neurochemical methods, we now raise an additional issue that must be considered by investigators studying transmitter dynamics using any of these methods - this is the importance of interpreting functional measures of chemical neurotransmission within the context of what is known about the structural elements involved in the process. As a result of recent advances in anatomical methods, investigators are now able to examine the subcellular localization of proteins involved in chemical transmission. By integrating neurochemical and anatomical findings, a more complete understanding of chemical transmission can be obtained than is possible with either method alone. For example, studies performed using in vivo voltammetry and microdialysis suggested that DA diffuses a greater distance in the prelimbic cortex than in the neostriatum of rat (33, 70). Because the high-affinity DA transporter is known to be a critical determinant in regulating the duration of action and sphere of influence of DA in extracellular space (see also The Dopamine Transporter: Potential Involvement in Neuropsychiatric Disorders, this volume), it was proposed that regional differences in this protein may explain differences in transmitter dynamics in these two brain regions. Subsequently, the results of an anatomical study indicated that relative to nigrostriatal DA axons, mesocortical DA axons have a lower density of DA transporter and the transporter is located a greater distance from the synaptic release sites (69; Fig. 6). Together, these findings suggest that DA in the prefrontal cortex may diffuse from the site of release and influence targets at a distance. While such paracrine-like actions cannot be excluded (see ref. 29), several additional structural features of this system suggest that it is unlikely. First, as we have already discussed, the process of chemical transmission is not complete until the transmitter has interacted with its receptor proteins. If DA is to act at a distance, it might be expected that DA receptors would be diffusely distributed. However, subcellular localization of DA receptors suggests that they are relatively specific in their neuronal distribution (6, 58). Furthermore, such a diffuse action is inconsistent with the observation that a significant number of DA varicosities form conventional synapses (68, see also Electron Microscopy of Central Dopamine Systems, this volume).
Methods for monitoring neurotransmitter dynamics in the central nervous systems have come a long ways from the earliest days in which all that could be measured were concentrations of a small number of transmitters in large samples of brain tissue. At present, methods are available for monitoring transmitter release with a spatial resolution of microns and a temporal resolution of milliseconds. Furthermore, with the development of brain imaging methods it is now possible to study transmitter dynamics in humans. In the next generation of progress, advances in a number of fields including not only neurochemistry, but also electrophysiology, brain imaging, and anatomy should bring us even closer to understanding the process responsible for communication between neurons in the nervous system.
Acknowledgments: Our thanks to Adrian Michael and Alan F. Sved for helpful comments on an earlier version of this chapter. This work was supported in part by U.S. Public Health Service Grants MH29670, MH45156, MH49936, MH57078, and MH01621.
published 2000