News Physiol Sci 16: 178-184, 2001;
1548-9213/01 $5.00
News in Physiological Sciences, Vol. 16, No. 4, 178-184,
August 2001
© 2001 Int. Union Physiol. Sci./Am. Physiol. Soc.
Synaptic Transmission with the Glia
Sabino Vesce,
Paola Bezzi and
Andrea Volterra
Department of Pharmacological Sciences, University of Milan, Via Balzaretti 9-20133, Milan, Italy
 |
Abstract
|
|---|
For decades, scientists thought that all of the missing secrets of brain function resided in neurons. However, a wave of new findings indicates that glial cells, formerly considered mere supporters and subordinate to neurons, participate actively in synaptic integration and processing of information in the brain.
|
Introduction
|
|---|
Glia are the most numerous cells in the central nervous system, with well-established roles in providing structural, metabolic, and trophic support to neurons. However, the classic view that glial cells are subservient to neurons has been challenged by recent findings indicating that perisynaptic glial cells (such as astrocytes and oligodendrocytes in the central nervous system and Schwann cells in the peripheral nervous system) can be recruited for neurotransmission and exert a modulatory action on synaptic function. This new vision of "tripartite synapses," composed of perisynaptic glia in addition to pre- and postsynaptic terminals (1, 5), certainly makes this one of the most exciting discoveries in current neurobiology. This review will focus mainly on the rapid reciprocal communication between neurons and glia involving glutamate as the signaling molecule. However, this form of communication does not appear to be restricted to glutamatergic districts and may imply the release of transmitters other than glutamate from either the neurons or the glia (11).
|
Bidirectional signaling between neurons and glia during synaptic activity
|
|---|
In addition to playing a primary role in the clearance of synaptic glutamate, astrocytes express a number of neurotransmitter receptors, including functional 
-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA)/kainate and metabotropic glutamate receptors (mGluRs). Accumulating evidence indicates that such receptors may be a target for glutamate released from the presynaptic terminals. Indeed, electrical stimulation of neuronal afferents in hippocampal slices was found to induce elevations in intracellular calcium concentration ([Ca2+]i) in the surrounding astrocytes (6, 18, 17). The glial response was abolished in the presence of the sodium channel blocker tetrodotoxin (which blocks propagation of the action potentials along the axon and the consequent exocytosis) or the mGluR antagonist -methyl-4-carboxyphenylglycine (MCPG), therefore proving to be a consequence of the synaptic release of glutamate. These data strongly support the view that neurotransmitters liberated into the cleft during synaptic activity can not only stimulate their postsynaptic receptors but can also, by spilling over, act on receptors located on the nearby astrocyte membranes. Indeed, from a morphological standpoint, a good proportion of central synapses appear to be finely enwrapped up to the edge of the cleft by the distal processes of astrocytes. Even more surprisingly, a recent study indicates that direct synaptic connections may exist between neurons and glia, implying that the receptors on glial cells may represent the primary synaptic target of transmitter released from neuronal terminals (3). Indeed, in the hippocampal slice, stimulation of neuronal afferents induces inward currents, blocked by a selective AMPA receptor inhibitor, in a population of oligodendrocyte precursor cells (OPCs). The quantal nature of such responses, as well as their fast kinetics, can only correlate with glutamate released from a presynaptic terminal directly facing the glial receptors. Electron microscopy images support the existence of such neuron-OPC synapses.
What sort of events can follow the stimulation of glial cells by synaptically released transmitter? Do neurons talk to glia unilaterally, or can a reply be expected? The latter possibility was first suggested at about the same time by two studies performed on mixed neuron-astrocyte cultures. The authors independently observed that [Ca2+]i elevations induced in astrocytes were often followed by [Ca2+]i elevations in the neighboring neurons (13, 15). The two studies, however, proposed different mechanisms to explain their observation.
Parpura et al. (15) induced [Ca2+]i rises in astrocytes by several means, in particular by stimulating ligand (bradykinin) receptors coupled to calcium release from the internal stores. Such stimulation was found to cause release of glutamate from the astrocytes, and, since the [Ca2+]i elevations in neurons could be prevented by application of glutamate receptor antagonists, the authors concluded that the astrocyte-to-neuron signaling was glutamate mediated.
In contrast, Nedergaard (13) stimulated single astrocytes electrically, so as to induce [Ca2+]i elevations. Thereby, a wave of calcium was triggered that propagated to connected astrocytes and, when the wave front reached an overlaying neuron, [Ca2+]i in the neuron was also augmented. This phenomenon was abrogated in the presence of octanol, a gap junction channel uncoupler, so the author attributed the signaling mechanism to the existence of gap junctions that could diffuse the calcium signal from astrocytes to neurons. However, octanol is a quite nonspecific agent, so the ability of astrocytes to excite neurons via direct gap junctional communication awaits further confirmation. Moreover, excitation of neurons following the passage of calcium waves in neighboring astrocytes was observed in experimental conditions similar to Nedergaard's and shown to be reduced by administration of glutamate receptor antagonists (9).
Together, the above studies indicate that: 1) neuronal activity can trigger [Ca2+]i increases in astrocytes and 2) [Ca2+]i elevations in astrocytes can start signaling to neurons. Therefore, they indicate the existence of a bidirectional flow of information between neurons and astrocytes during synaptic activity (Fig. 1
). Such a continuous reciprocal signaling was subsequently revealed in acutely isolated hippocampal slices (4, 17). Thus not only did astrocytes respond to electrical stimulation of Schaffer collaterals with [Ca2+]i oscillations sensitive to the mGluR antagonist MCPG but, in conditions in which the synaptic glutamate release was abolished with tetanus toxin, their [Ca2+]i oscillatory responses were reproduced by the mGluR agonist 1-aminocyclopentane-trans-1,3,-dicarboxylic acid (t-ACPD), which also elevated [Ca2+]i in the surrounding neurons. Importantly, the neuronal [Ca2+]i rises were found to be only in part the result of t-ACPD action on its neuronal receptors, being significantly attenuated by blockers of ionotropic glutamate receptors. Therefore, the neuronal calcium response comprised an indirect component mediated by glutamate, most likely released from the astrocytes (17).
View larger version (83K):
[in this window]
[in a new window]
|
FIGURE 1. Rapid reciprocal signaling between neurons and astrocytes during excitatory synaptic transmission. A good proportion of central synapses is enwrapped by astrocyte processes. As result of nerve terminal depolarization, glutamate (Glu) is released from the presynapse to stimulate its receptors on the postsynapse and continue communication in the neuronal network (1). However, under certain circumstances, a collateral signaling pathway is activated: enough transmitter spills over from the cleft to reach receptors located on the surrounding astrocytes (2). As a result, intracellular calcium concentration ([Ca2+]i) in these cells is elevated with the start of regulated glutamate release. According to recent observations in hippocampal cultures (see Ref. 1), glutamate released from the astrocytes can reach its neuronal receptors located at either pre- or postsynaptic sites (or even extrasynaptically; not shown) to induce modulatory actions on synaptic transmission (3).
|
|
The above conclusion was reinforced by the direct demonstration that astrocytes release glutamate in response to stimulation of mGluRs [and, more efficiently, in response to coincident stimulation of mGluRs and AMPA/kainate glutamate receptors (4)]. A novel signaling pathway was found to underlie this form of receptor-activated, calcium-dependent transmitter release from glial cells (Fig. 2). Thus Bezzi et al. (4) showed that stimulation of glutamate receptors rapidly enhances formation of cyclooxygenase eicosanoids (prostaglandins, in particular PGE2) and that these boost the [Ca2+]i elevations that trigger the actual glutamate release process. With the same experimental approach used for t-ACPD, these authors demonstrated the crucial role of PGE2 in mediating astrocyte-to-neuron signaling. Thus administration of PGE2 to slices with inhibited synaptic release produced [Ca2+]i elevations in astrocytes, followed by [Ca2+]i elevations in neurons. The neuronal effect was abolished by a combination of ionotropic glutamate receptor antagonists and was therefore indirect and mediated by astrocyte-released glutamate.
View larger version (89K):
[in this window]
[in a new window]
|
FIGURE 2. Glutamate stimulates calcium-dependent glutamate release from astrocytes: underlying signal-transduction events. Astrocytes express on their plasma membranes both ionotropic and metabotropic glutamate receptors. Activation of the metabotropic subtype coupled to breakdown of phosphoinositides (mGLUR5) together with the ionotropic AMPA receptor (AMPAR) results in [Ca2+]i elevation and potent glutamate release. Interestingly, [Ca2+]i rise stimulates a phospholipase A2 (PLA2) to produce arachidonic acid (AA), subsequently transformed by cyclooxygenases (COXs) into prostaglandins. One such agent, PGE2, induces [Ca2+]i elevation and glutamate release either by direct intracellular action or via autocrine/paracrine stimulation of the astrocytes. The potent inhibitory action exerted by COX blockers (indomethacin and aspirin) on glutamate release demonstrates that prostaglandins crucially control this process (see Ref. 4). Sensitivity of the release to tetanus neurotoxin (TeNT) and bafilomycin A1 (BAF-A1) indicates that it occurs via a process resembling neuronal exocytosis.
|
|
The notion that prostaglandins play a role in rapid brain signaling and neuromodulation is new and important, because these agents are often thought of only as mediators of slow and pathology-related processes (such as the inflammatory reaction). Indeed, the brain levels of prostaglandins are under the control of inducible synthetic enzymes such as cyclooxygenase-2, which appear to be dramatically upregulated in a number of brain pathologies. In view of the above findings, enhancement of prostaglandin production in pathological conditions might result in an altered control of glial glutamate release and of its modulatory effects on synaptic function.
|
Glia-to-neuron signaling modulates neuronal excitability and synaptic transmission
|
|---|
Revealing the existence of rapid signaling circuits between neurons and glia raises the critical question as to how they affect neuronal excitation and synaptic transmission. Recent in vitro data obtained in mixed neuron-glia hippocampal cultures by the group of Haydon (see Ref. 1) provide some of the first evidence that astrocyte activation indeed has profound effects on synaptic function. First of all, these authors, upon mechanical stimulation of individual astrocytes, were able to record responses in neighboring neurons in the form of slow inward currents (SICs). Such currents were abolished by preventing [Ca2+]i rises in the astrocytes with the calcium chelator 1,2-bis-(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) or in the presence of ionotropic glutamate receptor antagonists, suggesting that they depended on calcium-dependent glutamate release from the astrocytes. When the authors evoked synaptic activity while stimulating a surrounding astrocyte, they often observed a clear reduction of the postsynaptic currents (PSC) compared with responses obtained in the absence of astrocyte stimulation. The same phenomenon was observed at both excitatory and inhibitory synapses.
In a second study (1), the authors monitored miniature postsynaptic currents (mPSCs; the currents elicited by spontaneous quantal release events) while inducing [Ca2+]i elevations in the surrounding astrocytes by mechanical stimulation. [Ca2+]i rises in the astrocytes were accompanied by an increased frequency (not amplitude) of the mPSCs, reflecting a higher number of spontaneous release events at both inhibitory and excitatory synaptic terminals. Modulation of mPSC frequency was shown to be a direct consequence of glial cell [Ca2+]i increase because it was prevented by microinjection of BAPTA into the astrocytes and reproduced by ultraviolet photolysis of a calcium cage similarly microinjected into the astrocytes. The authors ascribed the two distinct modulatory effects of astrocyte activation on evoked and spontaneous synaptic currents to the release of glutamate from the astrocytes upon [Ca2+]i elevation, followed by interaction with neuronal receptors of distinct type and localization in the two cases: presynaptic mGluRs for reduction of evoked PSCs and extrasynaptic N-methyl-D-aspartate (NMDA) receptors for enhancement of mPSC frequency.
These studies indicate that glial glutamate inputs to synapses indeed have the potential to modulate neurotransmission and to produce a range of effects, also depending on the type of ongoing activity and the neuronal receptors stimulated. However, a reservation concerning the above data obtained in an in vitro preparation is that such preparation does not entirely reflect the structural organization and the specific morphological relations occurring between neurons and glia in the intact brain tissue.
|
Modulatory glial loops during synaptic transmission in the intact nervous tissue
|
|---|
Three recent studies performed in different preparations in situ provide the first compelling demonstration of the active participation of glial cells in synaptic transmission. The first study, performed in the intact frog neuromuscular junction, reveals the modulatory role of perisynaptic Schwann cells (19). In this preparation, high-frequency stimulation of the axons typically causes depression of neuromuscular transmission. Such depression has often been attributed to an activity-dependent depletion of vesicles in the nerve terminal. However, Robitaille (19) was able to reproduce the same response by microinjecting GTP
S, a nonhydrolizable GTP analog that stimulates G protein-dependent signaling, specifically into the perisynaptic Schwann cells. Furthermore, microinjection into the Schwann cells of GDBßS, a functional antagonist of GTPS, blocked synaptic depression induced by high-frequency stimulation. These data suggest that, during high-frequency stimulation, presynaptic terminals send a signal to the associated Schwann cell, which activates internal G protein signaling, resulting in a feedback inhibitory response on neurotransmission between the nerve and the muscle.
In a second study performed in the rat eyecup preparation, activation of glial cells (retinal astrocytes and Müller cells) was shown to modulate neuronal firing activity evoked by light in the retinal ganglion cell layer (14). Calcium waves were initiated in astrocytes and Müller cells by mechanical stimulation of the retinal surface. At the moment the wave front reached the neurons, changes in their spiking activity were recorded: inhibition in most cases, excitation in a few cases. The inhibition was reduced both by -aminobutyric acid (GABA) receptor antagonists and by glutamate receptor antagonists, suggesting that it is mediated by inhibitory interneurons stimulated by glutamate release from glial cells. A third study revealed that activity-dependent potentiation of inhibitory synaptic transmission in the hippocampal slice depends on a modulatory glial loop (11). When GABAergic interneurons synaptically connected to CA1 pyramidal cells are activated repeatedly, the number of synaptic failures in spike-evoked inhibitory PSC decreases. The authors found that this effect depends on activation of GABAB receptors and is mimicked by stimulation of the surrounding astrocytes. Their observations led them to conclude that repeated interneuronal firing causes recruitment of GABAB receptors in astrocytes, followed by [Ca2+]i elevation in these cells, which in turn potentiates inhibitory transmission, possibly via release of glutamate (see also Ref. 1). Indeed, they inhibited synaptic potentiation by buffering [Ca2+]i in the astrocytes as well as by blocking either GABAB or ionotropic glutamate receptors. The results of this latter study not only support the concept that glial cells are integral components of the functional hippocampal circuits but also indicate that they are involved in activity-dependent adaptation of synaptic strength. Therefore, they call attention to the potential, and so far totally unexplored, role that these cells might play in synaptic plasticity phenomena in this brain area.
|
How is information transferred between neurons and glia?
|
|---|
Because glial cells are progressively being recognized as active components of synaptic circuits, a number of critical questions have arisen concerning the modalities of their modulatory cross-talk with neurons. The first concerns the conditions for recruitment of glial cells to neurotransmission: do glia respond to any level of neuronal activity or does a certain threshold exist? A second question concerns the mechanisms allowing for the transfer of coded information from neurons to glia and vice versa. Glial cells do not use electrical coding, and therefore glial loops in brain circuits cannot resemble normal synaptic relays, in which chemical (transmitter) signals are translated into electrical signals to be retranslated into chemical signals at the next station and in which the glial cell would play a role similar to an interneuron. Indeed, the available data indicate that glial cells use a different code for translation and elaboration of the signals they receive based on their calcium signaling properties (see below). For this code the definition of "calcium excitability" has been proposed, which distinguishes glial cells from the electrically excitable neurons but at the same time highlights their active properties.
Today we are just starting to comprehend the fascinating communication between neurons and glia in the brain, given the very limited number of studies that have, directly or indirectly, addressed the issue. According to these initial data, recruitment of glial cells into synaptic transmission could be graded and correlate with the level of ongoing neuronal activity. Thus Porter and McCarthy (18), in their study on hippocampal slices, not only showed that electrical stimulation of neuronal afferents triggers [Ca2+]i elevations in astrocytes, inhibited by the mGluR antagonist MCPG, but observed that, in the presence of the K+ channel antagonist 4-aminopyridine (which prolongs neuronal terminal depolarization, promoting greater glutamate release), the astrocyte calcium responses are higher and can be blocked only by a combination of mGluR and ionotropic glutamate receptor antagonists. In line with these observations, Pasti et al. (17) found that an increase in the intensity or frequency of stimulation of the neuronal afferents promotes higher and more frequent [Ca2+]i oscillations in the astrocytes, with a clear quantitative correlation between the two events. In both studies, [Ca2+]i responses in astrocytes were elicited by trains of spikes evoked in neurons rather than by single action potentials, which suggests that sporadic neuronal activity may not be sufficient to trigger a glial response.
The data obtained by Robitaille (19) at the frog neuromuscular junction are in line with this idea. Perisynaptic Schwann cells activated by neuronal signals exert a feedback inhibitory effect on transmitter release from motor nerve endings. However, Schwann cell modulation is seen only when the motor neuron fiber is stimulated at high frequency (10 Hz), whereas it is absent during low-frequency stimulation (0.2 Hz). This observation suggests that signals originating from neurons might need to exceed a threshold to activate glial cells nearby. However, the situation may be significantly different at different synapses, and other factors, in addition to the level of neuronal activity and the corresponding amount of transmitter released, can control the activation of glial cells. The geometry of the synaptic space and the specific synapse-glia morphological relations may themselves play a critical role. For example, glial cell activation is the expected consequence of synaptic transmitter release at direct neuron-OPC connections in the hippocampus, whereas glial activation might be significantly less predictable when it depends on spillover of transmitter from nearby neuronal synapses. Efficient glutamate transporters, located in both neuronal and astrocyte membranes, bind and remove the transmitter from the extracellular space. The type, number and localization of these transporters may vary profoundly at different synapses, and so may their influence on transmitter availability. In particular, little is known about the reciprocal location of glutamate carriers and glial receptors. In any case, it makes sense to predict that the more neurotransmitter is released, the more it will diffuse out of the cleft and escape the buffering capacity of transporters, with enhanced chances to reach and activate receptors on the neighboring glial processes.
The correlation observed between patterns of neuronal electrical activity and types of [Ca2+]i response in glial cells (17, 18) strongly suggests that such a signaling underlies a regulated transfer of information from neurons to glia. Therefore, variations in the amplitude and frequency of the [Ca2+]i elevations in response to changes in neuronal activity most likely are key elements of the glial coding, into which the neuronal information is translated. They may also represent the basis for regulation of the feedback responses to neurons. Indeed, [Ca2+]i rises trigger a pathway of transmitter (glutamate) release in glial cells (4, 15). The extent and modalities of such a release might therefore depend on the specific features of the Ca2+ signals elaborated by the glial cell. A few experimental observations support this view. For example, on the one hand Porter and McCarthy (18) find that intensifying synaptic activity leads to higher [Ca2+]i elevations in astrocytes, which depend on activation of ionotropic glutamate receptors in addition to mGluR. On the other hand, Bezzi et al. (4) report that activation of AMPA/kainate together with mGluRs boosts astrocyte glutamate release. Further evidence correlating the extent of glutamate release to the amplitude of [Ca2+]i elevation in astrocytes is provided by Parpura and Haydon (16). These authors recorded SICs from single neurons plated onto microislands of cultured astrocytes in response to calibrated glial [Ca2+]i elevations. The same group showed previously that SICs are the result of stimulation of neuronal receptors by glutamate released from the astrocytes (1) and use them as quantitative detectors of the release process itself. Controlled rises of [Ca2+]i in the astrocytes are obtained by photolysis of the preloaded calcium cage o-nitrophenyl-EGTA via standardized ultraviolet pulses. The authors find that even modest astrocyte [Ca2+]i elevations (in the range of 80–140 nM) induce SICs in the neighboring neurons. When astrocyte [Ca2+]i is raised progressively, in a staircase-like fashion, with trains of ultraviolet pulses, some (even if not all) neurons show proportionally increased SICs.
The above study does not, however, address the role of temporal (frequency) coding of [Ca2+]i elevations. To explore this aspect, the group of Carmignoto (17a) has used, as sensors of astrocytic glutamate release, HEK cells transfected with NMDA receptors and plated onto the astrocytes. Upon induction of oscillatory [Ca2+]i responses in the astrocytes with the glutamate receptor agonist quisqualate, the authors observed corresponding pulsatile NMDA receptor-dependent [Ca2+]i responses in the HEK cells, which they ascribed to repetitive episodes of glutamate release. On the basis of these observations, they proposed that [Ca2+]i oscillations represent a frequency-encoded signaling system of astrocytes that dictates pulsatile release of glutamate. In this light, an enhanced frequency of the [Ca2+]i oscillations in astrocytes, as observed in response to intensified neuronal activity (17), would underlie a higher number of episodes of glial glutamate release.
Although preliminary, the above observations suggest that the bidirectional communication between neurons and glia is dynamically interrelated and finely regulated and therefore represents a dialogue of the utmost physiological significance.
|
Glial calcium signaling
|
|---|
As seen above, astrocytes finely sense the activity of neighboring neurons and respond accordingly. But what is the spatial and temporal range of their responses? Do they function as local modulatory loops, feeding back signals to the synaptic domains by which they received a stimulus? Or, given their syncytial organization, do they propagate information to more distal domains, extending communication to more distant synapses? Along this pathway, do they provide spatial-temporal integration of the signals they receive? The available data suggest the existence of both scenarios. Indeed, in monolayers of cultured astrocytes, stimulation of an individual cell (mechanically or by local glutamate application) was shown to generate two distinct and totally asynchronous calcium responses, one consisting of focal [Ca2+]i oscillations, the other of [Ca2+]i increases spreading in the glial network in the form of a slow, regenerative "calcium wave." As seen in cell culture, calcium waves travel, often with complex routes, for millimeters at a constant velocity of ~15–20 µm/s. Observations in the retina (14) suggest that similar waves exist in the intact brain tissue, although perhaps with more restricted domains of propagation. Importantly, glial calcium waves excite calcium transients in neurons when these are reached by the wave front, thereby causing modulation of the neuronal activity (13, 14). Thus long-range calcium signaling in the glial network could be a form of spatial transfer of the information and serve to coordinate distant synaptic circuits. This way, astrocytes would function as temporal-spatial integrators of brain activity that work beside and in collaboration with neurons. However, since they diffuse calcium signals at a speed considerably slower than propagation of action potentials, their role is clearly distinct.
Focal [Ca2+]i oscillations in astrocytes may serve a different function. Indeed, oscillations have often been found to occur asynchronously in different domains of the same cell, with frequencies higher at the distal processes apposed to synapses than in the cell soma (17). These rapid and spatially restricted calcium bursts most likely underlie quick responses to an input signal and help integration of the messages at a local level.
Recent work supports a dual function of glial calcium signaling, revealing novel aspects of the underlying structural and functional organization. Three-dimensional computer reconstruction based on electron microscopic images of the Bergmann glia cells in cerebellar slices highlights their highly ramified organization (7). Interestingly, the numerous branches all arise from a few stalks, departing orthogonally from the main fiber. Unlike dendrites in neuronal cells, glial branches arise from the stalks in an unpredictable and polymorphic way. However, a common structure can be recognized in each branch, consisting of a thin stalk that ends in a "cabbage-like" head. Each of these units, which the authors named "glial microdomains," is metabolically independent, since it contains a few mitochondria. Microdomain heads wrap around one or a few synapses, suggesting the possibility that individual microdomains interact with the neighboring synapse(s) independently of the rest of the cell. Indeed, stimulation of parallel fibers results in [Ca2+]i elevations in the surrounding Bergmann glia that are restricted to small areas of 100 µm2 or less, just about the size of a single microdomain. The authors concluded that a Bergmann glial cell may consist of hundreds of independent compartments capable of autonomous interactions with the particular group of synapses that they ensheath.
On the other hand, glial cells are extensively coupled to each other by gap junctions consisting of intercellular channels, the connexons, that allow passage of ions and small signaling molecules between connected cells. Gap junction channels have long been considered the molecular pathway for propagation of calcium signals in the glial network. However, given the regenerative nature of calcium waves, simple diffusion of calcium via connected cells cannot explain the phenomenon. In addition to calcium, inositol 1,4,5-trisphosphate has been proposed to diffuse through open gap junction channels, causing active propagation of the calcium release in the adjacent cells. However, as noticed by some authors, calcium waves in astrocyte cultures propagate even through areas free of cells (up to 120 µm long), implying the existence of an extracellular signaling component. Recently, ATP, released from one astrocyte to act at its purinergic receptors on the neighboring ones, has been identified as the key extracellular mediator (8). The ATP and the gap junction pathways appear to be highly interactive and cross-regulatory. In addition, gap junction openings are modulated by a number of chemicals, released from the neurons or the glia itself. Such a complex regulation possibly controls the domain of propagation of the Ca2+ signal, which could shrink or extend dynamically.
The above studies, although inconclusive, shed some light on the biochemical mechanisms underlying calcium waves, revealing an unexpected level of complexity. However, how is information actually transferred between glia and neurons by means of a calcium wave? A recently published study shows that, at least in vitro, release of glutamate accompanies the Ca2+ wave all along its path of propagation (10). The authors imaged extracellular glutamate indirectly by using its reaction with glutamate dehydrogenase that transforms NAD+ into the fluorescent NADH (see also Ref. 4). Thereby, they observed that stimuli evoking a calcium wave induce a parallel regenerative wave of NADH fluorescence that moves radially from the original source at an average speed of 26 µm/s, well correlating with the speed of the calcium wave. These data provide an indication that the route of a calcium wave might identify the pathway of chemical transmitter (glutamate) signaling to the neuronal circuits present at its interface. Thereby, information would be spatially transferred along defined glial circuits.
|
Regulated transmitter release from the glia
|
|---|
The presented evidence highlights the exchange of coded information between neurons and glia during synaptic transmission. To be apt to this function, glial cells must be able to translate their [Ca2+]i code into a regulated release process, which governs the output of chemical signals to neurons. Some of the properties of calcium-dependent glutamate release in response to neuroligands (4, 15) suggest that glia use a process resembling neuronal exocytosis. Indeed, such a release is blocked by either tetanus neurotoxin or bafilomycin A1 (2, 4). The former is a protease that selectively inactivates the vesicular protein, synaptobrevin; the latter inhibits the vacuolar H+-ATPase pump necessary for transmitter uptake into vesicles. The specificity of action of these agents is confirmed by their lack of effect on other release processes observed in astrocytes, such as calcium-independent release via glutamate transporters (4). To reinforce the above observations, a recent report shows that bafilomycin A1 and botulinum B toxin (which cleaves synaptobrevin like tetanus neurotoxin) inhibit the glutamate-mediated SICs in neurons elicited by astrocyte stimulation (2). In addition, a selective activator of neuronal exocytosis, the black spider venom toxin -latrotoxin, promotes glutamate release from astrocytes. Finally, in sniffer cells, stimulation of astrocyte glutamate release with the mGluR agonist t-ACPD induces ion current responses with fast kinetics resembling quantal release events (17a). Functional data are paralleled by immunochemical evidence that astrocytes express several of the proteins forming the exocytotic machinery in neurons. Some of these proteins, like synaptobrevin, have been ultrastructurally localized at the membranes of vesicular organelles (12).
Overall, the above data provide important preliminary evidence that glial cells might indeed possess a process resembling neuronal exocytosis. To assert the existence of a "glial exocytosis," however, more compelling demonstrations are required, including detection of the glutamate-containing vesicles and of their fusion with the cytoplasmic membrane, with ensuing release of the transmitter. Given that glia have physiological properties well distinct from neurons, glial exocytosis might differ significantly from its neuronal counterpart. For example, it is already clear that it would not rely on electrical excitation or depend critically on calcium influx via voltage-operated channels. Other central aspects, including whether or not the process would occur at specialized sites like in neurons, have not been explored yet. Although more conclusive evidence for vesicular release in glial cells and its fine properties is awaited, the present data support the view that glia is indeed endowed with regulated transmitter release properties, suggesting that this process is a fundamental component of its tight dialogue with neurons during synaptic function.
|
Conclusions
|
|---|
The wave of new information reviewed here is changing our view of the contribution of glial cells to information processing in the brain. The emerging picture is one of a continuous, dynamic, and integrated exchange of rapid signals between nerve cells and their surrounding glia during brain activity. Synaptic function itself, the core of integrative brain skills, apparently involves the active participation of glial cells. Thus glia not only receive coded information concerning neighboring synaptic events but also elaborate their own modulatory responses, which in turn regulate synaptic transmission. For decades, glia were considered passive supporters of nerve cell function, lacking the fundamental prerogatives of excitability and signal integration. In contrast, we now appreciate that glia exhibit "calcium excitability" and regulated transmitter release. We are certainly only at the beginning of understanding the modalities of the intense dialogue between neurons and glia and its role in brain activity. However, these first insights indicate that glial cells can no longer be excluded from the functional brain circuitry, calling for a reconsideration of several of the past and present theories on the physiology of synaptic transmission. Indeed, at the beginning of the 21st century, brain circuits look like synaptic networks of both neurons and glia (1, 5) (Fig. 3).
View larger version (23K):
[in this window]
[in a new window]
|
FIGURE 3. Synaptic networks of neurons and glia? This cartoon illustrates new hypotheses on signal integration in the brain with the active involvement of glial cells. From top to bottom: astrocytes, activated during synaptic transmission, can feed back a modulatory response to the original synaptic input (1) and/or to a different synapse of the same spatial domain (2). In this way, glial cells participate in local integration of the signals. Alternatively (or in addition), astrocytes can spread information in the glial network, either to contacting cells coupled via gap junction channels (3a) or to noncontacting cells, via "paracrine" secretion of transmitter (3b). Thereby, information travels spatially to eventually reach synapses located far from the original input (4). In this way, glial cells may function as temporal-spatial integrators of the synaptic activity, providing coordination between distant synaptic circuits.
|
|
|
Acknowledgments
|
|---|
We are supported by grants from the European Community (QLK6-CT1999-02203) and Ministero della Ricerca Scientifica e Tecnologica of Italy (MURST), Cofin 1998–1999 and 2000–2001 (to A.Volterra).
Present address for S. Vesce: Buck Institute for Age Research, Novalto, CA 94945.
|
References
|
|---|
-
Araque A, Parpura V, Sanzgiri RP, and Haydon PG. Tripartite synapses: glia, the unacknowledged partner. Trends Neurosci 22: 208–215, 1999.[ISI][Medline]
-
Araque A, Li N, Doyle RT, and Haydon PG. SNARE protein-dependent glutamate release from astrocytes. J Neurosci 20: 666–673, 2000.[Abstract/Free Full Text]
-
Bergles DE, Roberts JDB, Somogyi P, and Jahr CE. Glutamatergic synapses on oligodendrocyte precursor cells in the hippocampus. Nature 405: 187–191, 2000.[Medline]
-
Bezzi P, Carmignoto G, Pasti L, Vesce S, Rossi D, Lodi-Rizzini B, Pozzan T, and Volterra A. Prostaglandins stimulate calcium-dependent glutamate release in astrocytes. Nature 391: 281–285, 1998.[Medline]
-
Bezzi P and Volterra A. A neuron-glia signalling network in the active brain. Curr Opin Neurobiol 11: 387–394, 2001.[ISI][Medline]
-
Dani JW, Chernjavsky A, and Smith SJ. Neuronal activity triggers calcium waves in hippocampal astrocyte networks. Neuron 8: 429–440, 1992.[ISI][Medline]
-
Grosche J, Matyash V, Moller T, Verkhratsky A, Reichenbach A, and Kettenmann H. Microdomains for neuron-glia interaction: parallel fiber signaling to Bergmann glial cells. Nat Neurosci 2: 139–143, 1999.[ISI][Medline]
-
Guthrie PB, Knappenberger J, Sega M, Bennett MVL, Charles A, and Kater SB. ATP released from astrocytes mediates glial calcium waves. J Neurosci 19: 520–528, 1999.[Abstract/Free Full Text]
-
Hassinger TD, Atkinson PB, Strecker GJ, Whalen LR, Dudek FE, Koseel AH, and Kater SB. Evidence for glutamate-mediated activation of hippocampal neurons by glial calcium waves. J Neurobiol 28: 159–170, 1995.[ISI][Medline]
-
Innocenti B, Parpura V, and Haydon PG. Imaging extracellular waves of glutamate during calcium signaling in cultured astrocytes. J Neurosci 20: 1800–1808, 2000.[Abstract/Free Full Text]
-
Kang J, Jiang L, Goldman SA, and Nedergaard M. Astrocyte-mediated potentiation of inhibitory synaptic transmission. Nat Neurosci 1: 683–692, 1998.[ISI][Medline]
-
Maienschein V, Marxen M, Volknandt W, and Zimmermann H. A plethora of presynaptic proteins associated with ATP-storing organelles in cultured astrocytes. Glia 26: 233–244, 1999.[ISI][Medline]
-
Nedergaard M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 263: 1768–1771, 1994.[Abstract/Free Full Text]
-
Newman EA and Zahs KR. Modulation of neuronal activity by glial cells in the retina. J Neurosci 18: 4022–4028, 1998.[Abstract/Free Full Text]
-
Parpura V, Basarsky TA, Liu F, Jeftinija K, Jeftinija S, and Haydon PG. Glutamate-mediated astrocyte-neuron signaling. Nature 369: 744–747, 1994.[Medline]
-
Parpura V and Haydon PG. Physiological astrocytic calcium levels stimulate glutamate release to modulate adjacent neurons. Proc Natl Acad Sci USA 97: 8629–8634, 2000.[Abstract/Free Full Text]
-
Pasti L, Volterra A, Pozzan T, and Carmignoto G. Intracellular calcium oscillations in astrocytes: a highly plastic, bi-directional form of communication between neurons and astrocytes in situ. J Neurosci 17: 7817–7830, 1997.[Abstract/Free Full Text]
-
Pasti L, zonta M, Pozzan T, Vicini S, and Carmignoto G. Cytosolic calcium oscillations in astrocytes may regulate exocytotic release of glutamate. J Neurosci 21: 477–484, 2001.[Abstract/Free Full Text]
-
Porter JT and McCarthy KD. Hippocampal astrocytes in situ respond to glutamate released from synaptic terminals. J Neurosci 16: 5073–5081, 1996.[Abstract/Free Full Text]
-
Robitaille R. Modulation of synaptic efficacy and synaptic depression by glial cells at the frog neuromuscular junction. Neuron 21: 847–855, 1998.[ISI][Medline]
This article has been cited by other articles:
|
|
|
|
 
E. A. Newman
Glial Cell Inhibition of Neurons by Release of ATP
J. Neurosci.,
March 1, 2003;
23(5):
1659 - 1666.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Introduction
Bidirectional signaling between...
