The astroglial cells, or astrocytes, form a gap junction linked and electrically coupled syncytium. The single cells have long processes, some of which establish contacts with blood vessels and form part of the blood-brain barrier. Other processes extend toward and encapsulate synapses, especially the glutamate (Glu) synapses and also the varicosities, from which monoamines are released. Even neuronal cell bodies, neuronal processes and the brain surface are encapsulated by astroglial processes.

The astroglial cell mass constitutes a prominent part of the total brain cell number and volume (93). The anatomy of these cells, which "float" in the extracellular space of the brain, caused Virchow to propose, more than 100 years ago, that they have a metabolic and structurally supportive role for the neurons. Since then and until the last 15–20 years, this supportive role has been considered as passive and solely dependent on the requirements of the degree of neuronal activity. Very little attention has been paid to the astroglial cells for decades, mostly owing to the difficulties of studying them in the intact nervous system. Another important reason for this problem has been the lack of markers for the identification of the cells. Neurons were relatively easy to identify using electrophysiological techniques, mostly owing to their action potentials. Astroglial cells, on the other hand, are small, with cell bodies measuring 10-15 mm in diameter. The cell membrane can be depolarised but no action potential can be elicited. Furthermore, due to the syncytial arrangement, with the cells being electrically coupled to each other, it has been difficult to use conventional electrophysiology to register ion currents from the cells (see below). Therefore, it was not until the early 1970s that Eng, Bignami and co-workers (30) identified and isolated the glial fibrillary acidic protein (GFAP) as a true marker for the cells.

At the same time, methods were developed for the cultivation of astroglial cells. Earlier, cell lineages of tumor-derived glial cells had been cultivated, but a problem with those cells was the relatively low degree of differentiation, which made it difficult to extrapolate results from these experiments to the in vivo situation. Using different primary cultures, astroglial cells could be enriched and characterized in large numbers (11,42,79). Since then, astroglial research has developed very rapidly. In addition to their utility functions in relation to the neurons (e.g., energy supply, removal of extracellular K+ and Glu, and synthesis and release of neurotrophic factors) the cells have the capacity to monitor synaptic activity, to sense the composition of the extracellular space and the blood serum, to integrate the information obtained, and to influence neuronal activity by regulating the extracellular ion and amino acid homeostasis and the concentration and composition of trophic substances (Fig. 1).

In this chapter, we will provide a brief review of astroglial pharmacology. Owing to limits of space, we have only taken examples of a few membrane receptors, signal transduction systems and ion channels identified on astroglial cells. We briefly describe transport carriers for the amino acid neurotransmitter Glu, touch upon the important fields of astroglial energy metabolism and astroglial volume control, and we touch very briefly upon Ca2+ based astroglial excitability (Fig. 2), all these systems being possible targets for pharmacological manipulation. However, most results obtained to date are from experiments using various culture systems or in vitro preparations, and it is of utmost importance for future work to validate the properties, even using in vivo preparations, and to identify selective ways of interacting with them.

Astrocytes are the most frequently occurring cells in the brain, constituting over 50% of the total cell number in the cerebral cortex. Their relative number is especially high in humans and other highly developed mammals. They are recognized as star-shaped cells whose processes extend into the surrounding neuropil, and they are extensively coupled in a network—the astrocyte syncytium. They are also intimately connected to the neurons both structurally and functionally, indicating important roles for the astroglial cells in brain function.

Many astrocytes have processes that contact the surfaces of blood vessels with their "end-feet," forming part of the blood-brain barrier together with the capillary endothelium. Other processes extend to the neuronal cell bodies, and the astrocytes thereby serve as a connecting link between the neurons and the blood circulation. Furthermore, other processes have been shown to very closely approach the synaptic regions and to ensheathe the synaptic clefts. In view of the fact that astrocytes express a variety of ion channels, a large number of neurotransmitter receptors and several active release and uptake mechanisms for neuroactive compounds, this surrounding of the synapse enables the astrocyte to communicate with the neurons and to regulate the extraneuronal milieu. In addition to these vital extensions, astrocyte processes also reach the ependymal cells, connecting them with the cerebral ventricular system, while other processes extend to the brain surface to form expansions that constitute the glial limiting membrane.

Classically, two main types of astrocytes can be identified with light microscopy using the Golgi-Rio Hortega metallic impregnation (17). These are the fibrous and protoplasmatic types, divided according to morphology and their location in white or gray matter, respectively (93). Fibrous astrocytes are predominantly found in the myelinated areas and have a star-like morphology with thin, usually unbranched, processes spreading out symmetrically from the cell body. The processes, rich in intermediate filaments, extend over long distances and frequently form the end-feet on blood vessels. Protoplasmatic astrocytes have shorter and highly branched processes of varying dimensions that ensheathe neuronal cell bodies and their processes. They form the end-feet on blood vessels and they also make contact with the pial surface.

Through the identification of GFAP (22), a cell-specific marker for astrocytes was detected, as both protoplasmatic and fibrous astrocytes had earlier been shown to contain gliofibrils. Other astrocyte markers beside the intermediate filaments have also been used, including antibodies against the enzyme glutamine synthetase and the calcium binding protein S100, both found in protoplasmatic and fibrous astrocytes. The intermediate filament vimentin has sometimes been used as a marker, especially in the early development of the CNS, where this is the only filament expressed, although it is not as cell specific as the other markers.

Techniques for maintaining highly enriched astrocytes in dissociated primary culture have been successfully developed during the last 25 years (11,79). These preparations consist of actively proliferating cells, isolated from the brains of neonatal or postnatal animals, and using different culture conditions, they can be enriched in one specific cell type. Using astroglial cell primary cultures, originally from the optic nerve, two morphologically and antigenically distinct astroglial cell types have been observed (96). One is a polygonally shaped flat cell having few processes and forming a confluent bed layer in the culture dish. These cells have been called type 1 astrocytes, and the overwhelming majority of astrocytes in primary culture are of this type (Fig. 3). A second cell type, exhibiting a central body with numerous long processes, grows on top of the bed layer cells. These have been called type 2 astrocytes and are most commonly detected immunocytochemically by A2B5, a monoclonal antibody against an extracellular ganglioside antigen (80).

The development of cultivation conditions for astrocytes during the last two decades has been the prerequisite for our present knowledge of the biochemistry and pharmacology of astrocytes. However, it should be emphasised that cell and tissue cultures are artificial model systems, and whatever is discovered in culture must also be demonstrated in vivo. The current trend in astroglial research is therefore to strive for changes in methods towards more in vivo-like systems. For these purposes, techniques have been developed using acutely isolated astrocytes from intact brains or brain slice preparations, and these will probably be of increasing importance in future astroglial studies.

The mGluRs are both functionally and pharmacologically different from the iGluRs. The mGluRs are linked to G proteins. Their stimulation generates the formation of second messengers and/or regulates ion channel function by mediating intracellular signal transduction. Cloning experiments reveal that there are at least eight subtypes, mGluR1-mGluR8 (94), which have been classified into three classes: Class I comprises mGluR1 and mGluR5. MGluR1 is primarily coupled to phosphoinositide (PI) hydrolysis/Ca2+ signal transduction, as are also mGluR5a and b. Class II comprises mGluR2 and mGluR3, and class III comprises mGluR4, mGluR6-8. Classes II and III inhibit the formation of cyclic AMP. They differ, however, in their pharmacological profiles in relation to specific agonists. MGluR4a is potently activated by L-1-amino-4-phosphonobutanoic acid (L-AP4), which suggests that the mGluR4a receptor is a possible candidate for L-AP4 on astrocytes (43,125). MGluR3 and mGluR5 have been detected by antibody staining and by in situ hybridization in hippocampal astrocytes (102). MGluR1 has been found on glia of the optic nerve (57).

It is well known that astrocytes are involved in the termination of inhibitory transmission via GABA uptake mechanisms. Receptors for GABA are also known to be present on astrocytes. Both GABAA and GABAB receptor subtypes are expressed (35). Activation of GABAA receptors results in an efflux of Cl- through the intrinsic Cl--specific channels (75) and also to an elevation of [Ca2+]i (89). GABAA-receptors are modulated by barbiturates and benzodiazepines and are therefore very similar pharmacologically to their neuronal counterparts. There is considerably less evidence for the presence of the GABAB receptors on the astrocytes. The GABAB receptor subtype in neurons is linked to K+ and Ca2+ channels by second messenger systems which include GTP-binding proteins (35). In astrocytes, GABAB receptor stimulation is negatively coupled to the inositol phosphate second messenger cascade (35). GABAB stimulation can also hyperpolarise astrocytes in culture (35) and evoke transient rises in [Ca2+]i (89).

Several different subtypes of 5-HT receptors have been recognized with different effector pathways, activated by the various 5-HT receptors. Most results are derived from neuronal preparations (54). To date, some 5-HT receptor subtypes have been identified on astroglial cells: the 5-HT1A receptor (3,134), the 5-HT2A receptor (formerly 5-HT) [26,47], the 5-HT5A receptor (19) and the 5-HT7 receptor (51). The existence of 5-HT1A receptors on astrocytes is still a matter of controversy (50,61).

It was found early on that the 5-HT2A receptor induced the formation of PI, leading to an increase in inositol (1,4,5)-trisphosphate (IP3) [26,47]. In other cell systems, activation of the G protein-coupled 5-HT2A receptor results in phospholipase C (PLC)-mediated PI hydrolysis, which liberates the second messengers diacylglycerol (DAG) and IP3 (21). The 5-HT-evoked Ca2+ transients were pertussis toxin (PTX)-insensitive and were suppressed by the PLC inhibitor neomycin. These data indicate that, in cultivated astroglia, the 5-HT2A receptors are linked to Gq proteins coupled to PLC (39). In the regional expression obtained from cell culture studies, it was found that the receptors were more abundant in the cerebral cortex and brain stem than in the hippocampus and striatum (47).

Neurotransmitter and hormone signals are propagated via G protein- and non-G protein receptors. G proteins located in the plasma membrane interact with the cytoplasmic loops of receptor proteins, thereby transferring information to effector molecules such as AC, PLC, and ion channels, as well as protein kinases and phosphatases. The molecules generate various second messengers which, in turn, induce a range of cellular physiological responses integrating the information from numerous signaling pathways. This requires cross-talk among different pathways, often simultaneously with mobilization of intracellular and extracellular Ca2+. Various possible models for cross-talk in cellular signal transduction have been suggested (16). Either a single receptor can activate two alternative signaling pathways through two different G proteins, different receptors that use the same G protein to produce different signals, or cross-talk can occur between two pathways through the interaction of the second messengers that are generated. Given the complexity of the regulation of cellular processes by the second messengers, cyclic AMP, Ca2+, or DAG, it is not surprising that extensive cross-regulation of their levels in cells has been observed. A number of receptors in a range of cells and tissues are known to generate more than one second messenger when activated by the appropriate physiological stimuli.

Cross-talk between second messenger systems certainly plays a crucial role in the regulation of multiple signal transduction mechanisms within the cell. Cyclic AMP and PI pathways appear to share positive as well as negative modulatory interactions, leading to a final cell response (25). The G protein-linked receptors are intimately involved in modulating both direct and indirect responses to ion channels including Ca2+ and K+ channels (7,38). This leads to activation of PLC, PLA2, PLD, PKA and PKC (7). Nearly all intracellular pathways are influenced by activation of PKA and/or PKC.

Evidence is now accumulating that both 5-HT and NA can facilitate and modulate neuronal activity and transmission by interaction with astroglia. At least two mechanisms might be involved: 1) interaction of 5-HT with adrenoceptor signal transduction systems, probably leading to modulatory effects on astroglial energy metabolism and energy supply to the neurons, 2) 5-HT induced hyperpolarisation of the astroglia in some cell types, with resultant possibilities for facilitation of the astroglial K+ and Glu clearance of the extracellular space, at least in those brain regions or in those situations where the astroglial network is hyperpolarised by 5-HT. Furthermore, it might be tempting to postulate that the 5-HT and NA induced Ca2+ signaling within the astroglial network serves to integrate the neuronal-astroglial activity level in one brain region with other parts not primarily reached by release of monoamines. This might be one way in which neuronal activity level is integrated over larger brain areas.

The recently discovered wealth of receptors and ion channels in the astroglial cell membrane has severely challenged the theories related to information exchange in the brain (5,113,133). Most of the ion channels found in neurons (voltage, ligand and mechanically-activated channels) also have been confirmed as existing in astrocytes (5,113). Although the functional significance of the ion channels is not fully understood, data indicate that they have important functions. For example, ion channels constitute a prerequisite for astroglial extracellular buffering, potential transmission responses and intercellular communication. Some important and interesting classes of ion channels in this context are the potassium, calcium and sodium channels in the astrocytes. They are expressed in different categories, have several subclasses and exhibit different characteristics.

Numerous studies have recently demonstrated that voltage-dependent K+ channels can be expressed in astroglial cells in vitro and in situ (114). Using the whole-cell patch-clamp technique, it has been possible to characterize certain potassium-channel phenotypes (I–IV). 1) Whole-cell currents mediated by delayed rectifier potassium channels (Kd) show a characteristic outward rectification, with higher conductances at potentials more positive than -50 mV. These channels are largely inactive at rest or at more negative potentials. They are equally inhibited by both tetraethylammoniumchloride (TEA) and 4-aminopyridine (4-AP) [91]. 2) Transient A-type potassium channels (Ka) are rapidly inactivating channels that mediate transient currents requiring brief hyperpolarisation prior to activation. A-currents are more sensitive to 4-AP than TEA (91,114). 3) Inwardly rectifying potassium channels (Kir) are characterized by a large open probability close to and more negative than the resting potential, while open probability is virtually absent at more depolarised potentials (114). These channels are sensitive to external Cs+ and Ba2+, and conductance depends critically on extracellular K+, showing increased conductance with increased [K+]e. Astroglial Kir probably plays a dominant role in setting the resting membrane potential, since it has a large open probability near the equilibrium potential for K+. Furthermore, it has been demonstrated that Kir channels are largely involved in K+ homeostasis. 4) Two major subtypes of calcium-activated potassium channels (KCa) have been identified in astrocytes (8,131): the maxi KCa channel (BK) which is characterised by large conductance (200–300 pS) and sensitivity to charybdotoxin, and the mini KCa channels (SK) which have conductances of only 10–14 pS and are sensitive to apamin (8). The KCa channels are also sensitive to changes in intracellular Ca2+ buffering.

Electrophysiological recordings from astrocytes show expression of voltage-activated Na+ currents. These currents are transient, show rapid activation and inactivation, reverse at the Na+ equilibrium potential and can be inhibited by the Na+ channel specific toxin tetrodotoxin (TTX). These channels may be involved in the neuron-glial communication (115).

Another group of plasma membrane ion channels are not voltage-activated. Instead, these channels open in response to certain ligands, cytoplasmic agents such as calcium, or deformation of the plasma membrane (132). Examples of astroglial ligand-gated channels include Glu and g-aminobutyric acid- (GABA) gated ion channels. They are present in astrocytes from various preparations in different model systems. Glu i opens ligand-gated Ca2+ channels, while GABA predominantly opens Cl- channels (75,88). A specific group of channels in this context are the voltage-independent calcium channels (VICCs). Although the classification of these channels is far from clear, it is possible to categorize some of them: second-messenger-operated calcium channels (SMOCCs), receptor-operated calcium channels (ROCCs), depletion-operated calcium channels or calcium-release activated channels (DOCCs or CRACs) and finally, calcium channels that are sensitive to plasma membrane deformation, called stretch-activated calcium channels (SACCs). Little is known about the existence of these different calcium channel types in astrocytes. However, there are data supporting the presence of ROCCs and SACCs in astrocytes (5,55)

Synaptic transmission is largely terminated by high affinity, sodium-dependent transport of neurotransmitters from the synaptic cleft (56). Glu transporters help to terminate the postsynaptic action of the neurotransmitter Glu and keep the extracellular Glu concentration at low levels (1–3 mM). The process is capable of concentrating intracellular Glu up to 10,000-fold compared to the extracellular environment (86). The Glu transport is a sodium-potassium coupled process which proceeds across the plasma membranes of neurons and glial cells via high and low-affinity transport systems (44,58,107). Both astroglial cells and neurons possess similar, although not identical, Glu uptake carriers on their plasma membranes. Three distinct, complementary DNAs encoding excitatory amino acid transporters have been cloned and sequenced. One of the transporters, termed the excitatory amino acid carrier (EAAC1), is expressed by neuronal cells, including pyramidal cells of the hippocampus (58). The remaining two transporters termed the Glu aspartate transporter (GLAST) [121] and the Glu transporter (GLT-1) [95] have characteristics more consistent with those of the glial Glu transporters. The GLT-1 transporter is expressed at high concentrations in the hippocampus, lateral septum, cerebral cortex and striatum (70). The GLAST transporter, on the other hand, has a different pattern of expression, with the highest abundance in the molecular layer of the cerebellum in the olfactory bulb. Both transporter proteins are expressed throughout the brain with similar distributions in some regions, e.g., the cerebral hemispheres and the brainstem. Areas with a high density of Glu transporters, such as the hippocampus, are paralleled by dense glutamatergic innervation. The distribution of mRNA for GLT-1 is in general agreement with the distribution of the transporter proteins GLT-1 and GLAST (121,129). Mechanistic studies on the electrophysiology of glial Glu transport have suggested that the transport is electrogenic with three Na+ (or two Na+ and one H+) ions co-transported into the cell, and one K+ ion is transported out of the cell for each transported amino acid (14).

The expression of astroglial Glu carriers is most prominent in those brain regions with the most extensive Glu transmission. Transmitter uptake in astrocytes is acutely regulated by surface receptors in both a stimulatory and an inhibitory way (45). Other Glu uptake systems, in addition to the well-known Na+-dependent uptake, have been described. A Cl--dependent transport mechanism, and a Ca2+-dependent transport system have been characterised in astroglia in primary culture (34).

However, the capacity of neurons to take up Glu seems to be less than that of glia, even though the anatomy of the synaptic cleft might favour a neuronal removal of Glu after its release from the presynaptic region. The uptake capacity of Glu by astroglia, however, is considered to be sufficient to account for all Glu released by neurons. Furthermore, the astroglial carriers could be important in maintaining defined levels of extracellular Glu. In addition, glial Glu uptake plays an integral role in the glutamine cycle, which has been proposed to generate terminal Glu via glial glutamine synthetase and diffusion of the glutamine so formed to the terminals for hydrolysis to Glu.

Glu is thought of as the key mediator of excitotoxic brain damage, e.g., cerebral ischemia and hypoglycemia (20). This makes the Glu transport systems very important for neuronal survival and protection against excitotoxic insults. Furthermore, disturbed Glu transport may be associated with amyotrophic lateral sclerosis (ALS) [99]. This suggestion was originally based on the theory that impaired Glu transport could cause neurodegeneration. In fact, pharmacological inhibition of all Glu transporters can cause neurotoxicity (98), in particular motorneuron toxicity (100). Studies on sporadic ALS have suggested that the functional Glu transport is impaired in this disease (99,109). Whether the loss of astroglial GLT-1 transporters could theoretically account for neuronal toxicity remains speculative. Antisense oligonucleotide experiments have suggested that GLT-1 and GLAST are responsible for more than 80% of total Glu transport, and that loss of glial Glu transport produces neuronal toxicity (101). An almost complete and selective loss of the astroglial GLT-1 protein, despite complete preservation of the glial population found in sporadic ALS patients, suggests that either a selective GLT-1 positive population of astroglial cells is lost or damaged or there is a selective loss of the protein from glial cells in the affected regions of the brain (101).

Astrocytes are involved in the regulation of electrolyte concentrations and water volume in the extracellular space of the CNS (63). The cells have a well- developed capacity to change their volume. One main reason for volume changes is to keep the osmolarity at a constant level both outside and inside the cells. Cell swelling may occur both after a brain injury and under normal conditions, e.g., under intense neuronal activity with changes in the composition of neuroactive substances in the extracellular milieu as a result. Initially, the cells swell at the expense of the extracellular space. The fraction of brain volume that is extracellular space amounts to between 20–27% of the total brain volume (64,87). As astrocytes constitute a substantial part of the brain volume—estimated to be some 20%—changes in astroglial cell volume are important for the water redistribution in brain (106).

Cell swelling can be due to an uptake of ions such as Na+, Cl-, K+ or Glu from the extracellular to the intracellular space. The function of glial cells as a potassium buffering system suggests that excess K+ produced in the extracellular space during neuronal activity is transported away from the neurons into the glial cell syncytium (60). Another important mechanism leading to swelling could be transmitter-stimulated carbonic anhydrase activities, such that H+ and HCO3- are created by hydration of CO2 and transported out of the cells via the Na+/H+ and Cl-/HCO3- carriers. This would lead to an accumulation of NaCl and therefore to a net increase in osmolarity, drawing water into the cell. In fact, there seems to be an intimate relationship between cell volume control, ion fluxes and intracellular pH (59). Transport models for pH and CO2-driven swelling of astrocytes have been reviewed and discussed by Kimelberg (63).

Fatty acids and free radicals can lead to swelling, presumably because they can cause a breakdown of the selective permeability of the membranes, leading to a non-selective ion influx. This influx principally involves Na+ and Cl- because of the high extracellular concentration of these ions (13).

The excitatory neurotransmitter Glu is rapidly accumulated by astrocytes against its steep intracellular/extracellular concentration gradient. Glu-induced cell swelling has been observed in primary astroglial cultures as well as in C6 glioma cells (43,106). This swelling is gradual and begins immediately when the cells are incubated in Glu. Calcium is believed to play a fundamental role in cell volume regulation (81). The swelling of most cells leads to transport changes that seem designed to give rise to shrinkage of the cells back to control levels, a process termed regulatory volume decrease (RVD) [52].

After a stroke or brain trauma, an initial energy deficiency occurs, with decreased ATP levels at the cellular level. This decrease in the available metabolic energy could lead to reduced activity of energy-requiring ion pumps, especially the Na+-K+-ATPase (60). The result is partial membrane depolarisation and an increase in extracellular K+ concentration ([K+]e). Furthermore, membrane depolarisations lead to Glu release, with increased extracellular Glu concentrations ([Glu]e). Within minutes after the impairment of the blood flow after a stroke or brain trauma, cellular swelling is seen. Astroglial processes, especially those directed towards blood vessels, increase in volume, followed by the cell body swelling (66). The result is a decrease in the volume of the extracellular space. This volume is comparatively small in brain, with the cell membranes of different cells being only some 200 nm apart from each other under physiologic conditions. Therefore, even a slight increase in astroglial cell volume results in a prominent reduction in extracellular volume, with further increased [K+]e and [Glu]e as a result.

A vicious spiral arises, and if the cell swelling progresses and is not inhibited, the extracellular space collapses. Later in the development of the brain injury, swelling of neuronal dendrites can be observed, but swelling of the neuronal cell soma is more limited and swelling of oligodendrocytes is hardly seen. Other factors leading to the disturbed extracellular milieu are the formation of arachidonic acid and free radicals, which have been shown to inhibit the astroglial Glu uptake capacity and thus further increase [Glu]e. There will be a decrease in pH. Cytokines are formed, some of which lead to astroglial depolarisation with a resultant further decrease in the Glu uptake capacity, acting in the same direction as the above-mentioned events, all with increased [Glu]e and [K+]e as a result.

These events are well characterised and occur minutes to hours after a brain injury. Later on, disturbance and even destruction of the blood-brain-barrier occurs, and water passes from the blood to the extracellular space, leading to vasogenic edema.

Gap junctions are membrane specializations providing a pathway for intercellular exchange of small molecules and ions between cells. Gap junctional communication is thought to facilitate co-ordination of cellular activity in several physiological processes such as tissue homeostasis, cardiac and smooth muscle contraction and development (6). Gap junctional channels are formed by members of closely related membrane proteins called connexins. Each channel is formed by two sets of six connexin molecules spanning the membrane of each connecting cell, forming the connexon; the functional gap junction.

It has been suggested that, in vivo, astrocytes form a syncytium (84) that mediates spatial buffering of potassium ions (36). The connexin-43 molecule expressed in astroglial cells shows unitary conductances in the range of 50–60 pS (27).

Recently, it has become evident that some cell types, including astroglial cells, communicate intercellularly via complex spatio-temporal calcium fluxes (22). These waves can be induced either mechanically or pharmacologically. While the physiological function of the waves is still largely unknown, somewhat more is known about the cellular mechanisms of these calcium fluxes (105).

Glu induces intracellular calcium increases that can propagate as calcium waves between astrocytes in cultures or organotypic slices (22,24). The metabotropic receptors are responsible for initiating fast propagating calcium spikes involving IP3, and the ionotropic receptors are responsible for smoothly propagating regenerative waves involving Na+/Ca2+ exchange (62).

Mechanical stimulation of a single astrocyte in a confluent culture induces intercellular calcium waves that propagate with constantly decreasing velocities (9) [Fig. 5]. Experimental data (10) and theoretical models (112) indicate that it is the second messenger, IP3, rather than calcium itself that passes from cell to cell through gap junctions, releasing calcium from intracellular stores via the action on IP3-receptors (IP3Rs).

The gap junction conductivity in astroglial cells is subjected to complex regulation at several levels including, at least, regulation of the expression including transcriptional regulation, and regulation of mRNA stability. Regulation of the fraction of connexins forming functional connexons vs. the fraction of non-connexon forming connexins. Furthermore, the function of connexons is regulated by phosphorylation, which is the most common pathway for fast receptor-mediated regulation. Gap junction conductivity is down-regulated by PKC activation, increased intracellular calcium and lowered pH, and up-regulated by cyclic AMP, membrane depolarisation and Glu (for review, see ref. 37).

We have recently shown that 5-HT decreases propagation area and increases the wave rate of mechanically induced calcium waves (9). When hyperpolarising the cells with valinomycin or zero [K+]e, there is a decrease in propagation area similar to that seen in 5-HT-treated cells. Furthermore, the excitatory neurotransmitter Glu increased the propagation area. This was also seen after the cells were depolarised with 56 mM [K+]e. These data suggest that the membrane potential could be important for communication through gap junctions in the astroglial syncytium.

Increased astroglial [Ca2+]i attributable to a propagating wave could influence the action of astroglial K+ channels, Ca2+ channels, Ca2+-sensitive release and uptake of neuromodulators and neurotransmitters, metabolism, and cell structure, with resultant changes in extracellular composition (111). Modulation of the extracellular composition (concentration of ions and neuromodulators) suggests that astrocytes might play an active part in neurotransmission. An effect of astroglial Ca2+ waves on neuronal function has recently been shown; calcium waves in astroglial cells of neuronal-glial co-cultures increased [Ca2+]i and depolarised and induced bursts of action potentials in neurons lying on astrocytes that were passed by the wave (48).

The Ca2+ waves may serve to synchronize and coordinate different cell functions such as the previously proposed glial potassium buffering. Our finding that prolonged, high extracellular Glu concentration can increase the propagation of mechanically induced Ca2+ waves in cultured hippocampal astrocytes indicates that Glu could extend the range of influence of astrocytes on neurons.

Two very important and unanswered questions are how far these propagating waves can travel in the intact brain and how tightly regulated they are. If Ca2+ waves are present and regulated in a controlled and directed way in the brain, the astrocytes could represent yet another dimension in information processing.

Growth factors are polypeptides synthesized by most tissues that contribute to the control of cell developmental processes like proliferation, differentiation, survival, migration, and maturation. They act usually by binding to membrane receptors, and many of them have tyrosine kinase activity in their intracellular domain (e.g., fibroblast growth factors; FGFs) [32]. FGFs are a family consisting of several members with a group of many transmembrane high-affinity receptors (92). Basic FGF (bFGF) is released from astrocytes (2). At least some FGFs in astrocytes are altered and upregulated in several neurodegenerative disorders such as Alzheimer's disease, Huntington's disease and Parkinson's disease (128). Other growth factors are also composed of subgroups, such as epidermal growth factor (EGF), transforming growth factor b (TGFb), insulin-like growth factor (IGF) and platelet-derived growth factor (PDGF), and they have effects on astrocytes (for review, see ref. 67). Glial cells express relatively low levels of neurotrophic factors such as nerve growth factor (NGF) or ciliary neurotrophic factor (CNTF) under normal conditions. Following injury, synthesis increases dramatically and transiently in the area around the wound site and become mainly localized in reactive astrocytes forming the glial scar (127). The glial-cell-line-derived neurotrophic factor (GDNF), a member of the family TGFb, has been shown to be active in midbrain dopaminergic neurons and to increase the high-affinity uptake of dopamine. Therefore, this compound might be of importance in the treatment of Parkinson's disease (73). There has been increased appreciation of the role of glial-derived immune and neurotrophic cytokines, which are produced in response to neuronal dysfunction or death (69,83,135).

Astrocytes (Fig. 6) are regarded as processing and storage plants for brain energy. Their high tolerance for cellular stress such as hypoxia and hypoglycaemia is unique among the cells of the brain. Accumulating data indicate that cerebral energy metabolism is under detailed regulation of neurotransmitters acting on these glial cells (40,76,130).

Neuron-glial interaction is vital to the energy metabolism of the brain. It is very likely that there is an astrocytic export of "fuel" substrates such as lactate to neurons. Metabolic inhibition of glial cells reduces and modulates synaptic transmission, and the astrocytes have a verified neuroprotective effect under conditions of metabolic stress. A considerable amount of the energy generated is, however, used within the astrocyte itself to operate the membrane pumps necessary for vital brain functions. Ion buffering, Glu uptake and cell-volume regulation are the highest metabolic priorities. Energy status also plays a role in regulating intercellular communication (49,124,130).

Glycogen is the largest energy reserve in brain and is almost exclusively stored in astrocytes. The enzymes necessary for synthesis, storage and catabolism of glycogen are practically unique to astroglia and are regulated by phosphorylation cascades under the control of intracellular Ca2+ and cAMP. The majority of scientific data published on pharmacological regulation of astroglial metabolism concern glycogen turnover. Adenosine, arachidonic acid, ATP, Glu, histamine, insulin, NA, vasoactive intestinal peptide and 5-HT are known to stimulate synthesis and/or breakdown of glycogen in primary astrocyte cultures from rat and mouse and in mouse brain slices. The pharmacological effects vary in different model systems and receptor subtypes. In addition to acting directly on enzymes through second messengers, the mechanisms may also involve gene expression (28,40,49,76,117,124,130).

Several other pathways in astroglial energy metabolism are regulated via receptor stimulation, e.g., glucose uptake, glycolysis, lactate production, formation of acetyl CoA from pyruvate, uptake of Glu/glutamine and their metabolism. Pharmacologically induced elevation of cytosolic [Ca2+] leads to a rise of intramitochondrial [Ca2+] with activation of the three regulating enzymes of the TCA cycle and stimulation of oxidative phosphorylation in astrocytes. NA stimulates activity in the TCA cycle in primary cultures from mouse, and Glu increases oxygen consumption in rat cultures. 5-HT, on the other hand, inhibits oxygen-consumption in rat brain in vivo (45,49,53,76,123).

Upon release of Glu from the presynaptic terminal of Glu-ergic terminals, Glu interacts with the synaptic membranes and with astroglial mGluRs and iGluRs. The activation of the mGluRs induces the formation of IP3 and a Ca2+ transient in the astroglial cytoplasm, giving rise to Ca2+ oscillations or waves (22). Similarly, there is a massive K+ release attributable to neuronal depolarisation. K+ and Glu released into the extracellular space are actively taken up by the astroglia (44,107) and induce astroglial depolarisation, which spreads rapidly within the gap junction coupled astroglial cells. These substances also induce a volume increase (swelling) of the astroglial cell (43,60,63), with the result that the astroglial and neuronal cell membranes approach each other.

The propagation of the Ca2+ waves and alterations in membrane potential within the astroglial network make it possible to integrate signals and information from synapses, from the blood serum and from the extracellular milieu within the astroglial syncytium to modulate energy metabolism, ion channels, membrane properties, second messengers, amino acid uptake carriers and the synthesis and release of neurotrophic and neuroactive substances (see ref. 111).

There are also implications for the formation of nitric oxide (NO) in astrocytes after Glu stimulation. NO has been suggested as a candidate for a neuronal-astroglial-blood-vessel mediator and might increase local blood flow in response to increased neuronal activation. Therefore, a basis exists for astroglial/neuronal control of the local microcirculation.

There are now strong indications that astroglia modulate synaptic excitability by influencing the ion and amino acid homeostasis in the extracellular space. Astroglial depolarisation with a K+ outflow near the pre-synaptic site can lead to pre-synaptic depolarisation and increased transmitter release. Similarly, astroglial depolarisation at the post-synaptic site can induce a relative depolarisation of the post-synaptic membrane and lead to increased excitability. A decrease in the [Ca2+]e concentration can diminish neurotransmitter release from the pre-synaptic site and also decrease excitability of the post-synaptic membrane (4,68). Astroglial swelling decreases extracellular volume and increases the concentration of ions and amino acids, with consequent changes in neuronal excitability.

The picture of the astrocyte that evolves here is of a cell that can sense multiple stimuli (via both membrane receptors in the vicinity of synapses and long-range signaling within the astroglial syncytium) and produce integrated responses (see ref. 46).

The above mentioned possibilities are just some examples of the role of astroglia in normal synaptic physiology. Many of the events described, and thus the astroglial influence on neuronal excitability, physiology and pathology, are influenced by membrane receptor activation. Using selective drugs, it might thus be possible to influence neuronal excitability via astroglial membrane receptor interaction. This may mean that a new arena for drug interactions in the CNS must be taken into account.

The last few decades have produced a vast amount of knowledge about the function of the nervous system, mostly concerning the function of neurons. Astroglial cells have recently been highlighted in several areas including normal physiology and pathological conditions. The existence of inter-astrocytic calcium communication may be one of the most exciting areas, demonstrating a novel extraneuronal communication system, possibly with information processing capacity. Communication within the glial network might include key features such as synthesis and release of trophic factors, redistribution of energy supply, as well as the regulation of glial K+ and Glu uptake to facilitate synaptic transmission. Future developments concerning calcium signaling might include the possibility that the glial network provides a super-regulatory system might be affected by selective pharmacological agents. One plausible mechanism of action is selective modulation of astroglial membrane receptors or gap-junctional communication. This system might provide novel and unique therapeutic possibilities concerning diseases such as stroke, epilepsy and Parkinson's disease (Fig. 7).

During the last few years, research in astroglial pharmacology and physiology has increased dramatically. It is now well known that these cells have features that were previously thought to be specific to neurons, such as membrane receptors for most neurotransmitters and neuromodulators with functional signal transduction systems, amino acid neurotransmitter carriers and ion channels. This means that astroglial cells have the prerequisites that enable them to monitor synaptic activity and to sense the composition of the extracellular milieu and the blood serum. The cells form a gap junction linked syncytium with a newly discovered Ca2+ signaling system from cell to cell within the astroglial network. Furthermore, the cells are electrically coupled, with the degree of coupling as well as the Ca2+ communication being possible targets for pharmacological intervention. The information obtained from the environment of the cells may thus be integrated within the astroglial syncytium. It seems that the cells are in a position to alter the ionic and amino acid content of the extracellular milieu and the volume of this space, and thereby to influence the neuronal excitability level. The astroglia support the neurons metabolically. Glucose is taken up into the astroglia from the blood, transformed to glycogen and stored. When required, owing to neuronal activity, the cells release lactate and other energy-rich compounds to be used as metabolic fuel for the neurons. The astroglial cells also synthesize and release neuroactive and neurotrophic substances under the control of membrane receptors. This means that the astroglial supportive and modulatory functions in relation to the neurons can be regulated pharmacologically. One problem at present, in relation to our understanding of astroglial receptors, is that expression is not selective. Rather, there are similar receptors on neurons and other brain cells, and interactions with these receptors may excite effects other than those mediated by the astroglia. A next step forward would therefore be to identify subclasses of receptors selectively expressed on astroglia, affect them, and thereby, have new ways of modulating neuronal activity and protecting neurons from functional disturbances and pathological disintegration.

This research was supported by grants from the Swedish Medical Research Council (project no 12X-06812 and no 14X-06005), the Swedish Work Environment Fund (grant no 94-0214), the Swedish Council for Work Life Research (grant no 95-0231), the Faculty of Medicine, Göteborg University, Göteborgs Läkaresällskap, Hjärnfonden, John and Brit Wennerströms Foundation for Neurological Research and Edit Jacobsson's Foundation. The skillful technical assistance of Barbro Eriksson and Armgard Kling is highly appreciated.