Maintaining brain function during aging is very important for mental and physical health. Recent studies showed a crucial importance of communication between two major types of brain cells: neurons transmitting electrical signals, and glial cells, which maintain the well-being and function of neurons. Still, the study of age-related changes in neuron–glia signalling is far from complete. We have shown previously that cortical astrocytes are capable of releasing ATP by a quantal soluble N-ethylmaleimide-sensitive factor-attachment protein receptor (SNARE) complex-dependent mechanism. Release of ATP from cortical astrocytes can be activated via various pathways, including direct UV-uncaging of intracellular Ca2+or G-protein-coupled receptors. Importantly, release of both ATP and glutamate from neocortical astrocytes was not observed in brain slices of dominant-negative SNARE (dnSNARE) mice, expressing dnSNARE domain selectively in astrocytes. We also discovered that astrocyte-driven ATP can cause significant attenuation of synaptic inhibition in the pyramidal neurons via Ca2+-interaction between the neuronal ATP and γ-aminobutyric acid (GABA) receptors. Furthermore, we showed that astrocyte-derived ATP can facilitate the induction of long-term potentiation of synaptic plasticity in the neocortex. Our recent data have shown that an age-related decrease in the astroglial Ca2+ signalling can cause a substantial decrease in the exocytosis of gliotransmitters, in particular ATP. Age-related impairment of ATP release from cortical astrocytes can cause a decrease in the extent of astroglial modulation of synaptic transmission in the neocortex and can therefore contribute to the age-related impairment of synaptic plasticity and cognitive decline. Combined, our results strongly support the physiological relevance of glial exocytosis for glia–neuron communications and brain function.
Interaction between two cellular communities, the electrically excitable neuronal network and the glial cells, which provide structural and metabolic support to neurons, is very important for brain function [1–3]. Glia cells, in particular astrocytes, not only control neuronal homoeostasis but also actively participate in brain signalling [2–4]. Astroglial membrane contains neurotransmitter receptors which allow astrocytes to sense chemical signals from neurons [2–8]. In response, astrocytes can release glutamate, D-serine, ATP and other transmitter molecules, which mediate signalling between astrocytes and neurons [2–4,9–12]. Numerous astroglial processes enwrap cerebral arterioles and capillaries and thereby can couple neuronal activity to the local microcirculation [1,4,13]. The astroglial network can thereby function as a ‘brain hub’ which receives and integrates the signals from neurons and responds by increasing the metabolic support of neurons and modulating neuronal activity via release of gliotransmitters [2–4]. In the present paper, we focus mainly on signals mediated by ATP and glutamate.
Role for glutamate and ATP for glia–neuron communication
Bidirectional signalling between astrocytes and neurons involves two important components transmitted by glutamate and ATP. Glutamatergic transmission is ubiquitous in the central nervous system (CNS), being primarily responsible for higher brain function such as learning and memory. There is also growing evidence for ATP as an important neuro- and glio-transmitter in the brain [4,12,14,15]. ATP can be released with glutamate from presynaptic nerve terminals and they bind to a variety of specific ionotropic and metabotropic receptors on neuronal and astroglial membranes [14–16], so astrocytes with privileged access to synapses could detect both ATP and glutamate simultaneously. The importance of ATP and glutamate receptors for function of astrocytes in many brain areas has been widely reported. In particular, we have found that astrocytes of brain neocortex express glutamate and ATP receptors, which are distinguished by their pharmacological and functional properties, and subunit composition from neuronal receptors [5–8,15]. First, astroglial N-methyl-D-aspartate (NMDA) receptors are very weakly affected by Mg2+ block and can therefore be activated without strong depolarization of glial membrane. Secondly, astroglial NMDA and P2X receptors have a subunit composition different from that in those receptors expressed in neurons; this can confer different pharmacological sensitivity. For instance, NMDA receptors expressed in the neocortical astrocytes have much higher sensitivity to the GluN2C/D-specific antagonist UBP141 than neuronal receptors . Neocortical astrocytes predominantly express P2X1 and P2X5 subunits of ionotropic ATP receptors, whereas P2X4 subunits make a significant contribution to purinergic signalling in neurons [14–17]. This provides an ample opportunity for selective modulation of astrocytic as compared with neuronal signalling using subunit-specific antagonists of ATP and glutamate receptors. Also, we have discovered that astroglial P2X and NMDA receptors have high affinity for ATP and glutamate [6,16,17]. So P2X and NMDA receptors, activated by both spill-over and transient release of neurotransmitters from synapses, can make a significant contribution to the electrical and Ca2+ signals in astrocytes [5–8,15,17].
In contrast, release of ATP and glutamate (with its cofactor D-serine) from astrocytes can modulate the activity of neuronal excitatory and inhibitory synapses [3,10,12,15,18,19], long-term synaptic plasticity [9,15,20–22] and vascular coupling [3,4,13]. Thus ATP- and glutamate-mediated signalling is widely seen as an integral component of glia–neuron communication [1–4]. A very important aspect of this communication is a mechanism by which gliotransmitters are released.
Role for glial exocytosis in the brain function
From the early days of research into glia–neuron interaction, a concept of fast vesicular release of chemical transmitters from astrocytes attracted lots of attention and, lately, lots of controversy [2–4,11,23–25]. This idea was attractive because capability of astrocytes to release the similar repertoire of transmitters, in the similar time scale as neurons, implied the equal importance of astrocytes for synaptic physiology. That is why exocytosis of gliotransmitters was embedded in the popular concept of tripartite synapse [2–4].
There is a large body of evidence for the presence of soluble N-ethylmaleimide-sensitive factor-attachment protein (SNARE) complex-dependent and Ca2+-dependent exocytotic machinery in the astrocytes. Ca2+-dependent exocytosis of glutamate and ATP, mainly from cultured hippocampal astrocytes, has been reported [2–4,25]. There are also accumulating reports of physiological roles for SNARE-dependent glial exocytosis . In particular, exocytosis of ATP followed by its conversion into adenosine is involved in the regulation of long-term potentiation (LTP) in the hippocampus and sleep homoeostasis in the hypothalamus . The key element of latter's work was development of dominant-negative SNARE (dnSNARE) transgenic mice with inducible inhibition of exocytosis, selectively in astrocytes [18,22]. There is a growing evidence that exocytotic pathways of gliotransmitters release can involve lysosomes or smaller synaptic-like vesicles [2,12,23].
Still, vesicular exocytosis is not the only pathway of gliotransmitter release. Several mechanisms of non-exocytotic release from undamaged glial cells have been identified in the last decade, in particular concentration gradient-driven diffusion through large conductance channels such as gap junction hemichannels, anion channels, dilated P2X7 receptors and TREK-1 potassium channels [11,14,15,26]. It should be noted that the majority of data of Ca2+-dependent astroglial exocytosis have been obtained from experiments on astrocytes in culture, thus fuelling debate on the importance of this mechanism for gliotransmission in intact tissue . Physiological relevance of Ca2+-dependent exocytosis of gliotransmitters in situ and in vivo has also been disputed  on a ground that alteration of astroglial InsP3-mediated Ca2+ signalling did not have a significant effect on glutamatergic synaptic transmission in the hippocampal slices [24,25]. Still, more recent in situ and in vivo data demonstrated an effect of astroglial InsP3-mediated Ca2+ signalling on cholinergic modulation of synaptic plasticity in hippocampus and neocortex [20,21].
In our recent experiments, we observed that vesicular release of ATP from cortical astrocytes in situ can be triggered by the Ca2+ transients, attainable under physiological conditions [10,12]. Release of ATP from cortical astrocytes occurred, most probably, from synaptic-like microvesicles and was impaired in the transgenic mice expressing dnSNARE protein selectively in astroglial cells. We also found out that vesicular release of ATP from astrocytes can directly activate excitatory signalling in the neighbouring neurons, operating through purinergic P2X receptors. Activation of P2X receptors by astrocyte-derived ATP down-regulated the inhibitory synaptic signalling in the neocortical neurons.
Thus there is emerging consensus that Ca2+-dependent exocytosis of gliotransmitters can play an important role in communication between astrocytes and neurons in physiological conditions.
Aging-related remodelling of neuron–glia communication in the neocortex
Mechanisms of glia–neuron interaction, in particular release of gliotransmitters, can gain a new importance in the context of brain aging. Maintaining brain function during the aging process is very important for an individual and population as a whole. Owing to changing demographics, age-related cognitive decline and neurological disorders are contributing to rising costs for medical and social care. There is a growing evidence of a positive link between an enriched brain activity, and physical health and positive effect of active lifestyle and physical exercise on the function of aging brain [27,28]. Still, the fundamental mechanisms of brain longevity are not fully understood. Working as a ‘brain hub’, astroglial cells are ideally placed to couple the enriched mental and physical activity to the synaptic function and metabolic support of neurons [1–4]. Thus it is conceivable that efficient communication between neurons and astroglia can be very important for brain longevity. Despite accumulating evidence of the benefits of exercise and environmental enrichment on neurogenesis and synaptic plasticity in aged brain [27–29], their impact on the function of glial cells is not as clear. Although enriched environment has been reported to induce the structural changes in hippocampal astrocyte and affect astrocytic control of extracellular γ-aminobutyric acid (GABA) and glutamate concentrations , the impact of an active lifestyle on astroglial signalling and release of gliotransmitters in old age is yet to be studied.
It is universally acknowledged that impairment of glia function can strongly affect the well-being of neurons and could therefore contribute to the pathogenesis of many age-related disorders, such as Alzheimer's disease and Parkinson's disease [2–4]. There are also scattered data suggesting that physiological aging can induce marked changes in the morphology and function of astrocytes in various brain regions [31–36]. It has been shown that the level of gliotransmitter D-serine, which is essential for induction of synaptic plasticity in hippocampus, decreases with aging . Still, the age-related changes in neuron–glia interactions, and physiology of maturing and aging astrocytes are virtually unexplored. Most information about communication between neuronal and glial cells has been derived from the experiments on newborn or young animals. We have shown for the first time that maturation and aging of the brain of mice (1–24 months) affected the density of ATP and glutamate receptors, and glutamate transporters in astrocytes and their contribution to the synaptically activated glial signalling . Following changes in the density of P2X and NMDA receptors, synaptically activated Ca2+ signalling was maximal at 3–6 months and then declines steeply . The age-related decline in the astroglial Ca2+ signalling could potentially impair release of gliotransmitters and thereby affect the ability of astrocytes to modulate neuronal activity.
In our recent experiments, we assessed the age-related changes in vesicular release of gliotransmitters, using the experimental paradigm successfully applied previously [10,12]. To verify the presence of ATP-storing vesicles, we used activity-dependent immunostaining of living astrocytes with antibodies to vesicular nucleotide transporter 1 (VNUT1) . We assessed release of ATP from acutely dissociated neocortical astrocytes using a ‘sniffer-cell’ approach (Figure 1). Our preliminary data indicate that astrocytes from neocortex of older mice retain exocytotic machinery, but vesicular release of ATP declines with age (Figure 1). This is manifest in the significant reduction of frequency and net charge transmitted by pulsatile currents activated in the ‘sniffer-cells’, contacting astrocytes isolated from neocortex of old mice. Interestingly, the quantal size of the sniffer-cell response exhibited a considerable decrease in old age, suggesting a lower content of ATP in the astrocytic vesicles.
Age-related changes in the release of ATP from cortical astrocytes
Role for glial exocytosis in modulation of synaptic plasticity in neocortex
The Ca2+-dependent release of ATP  and D-serine  from astrocytes has been shown previously to regulate synaptic plasticity in the hippocampus. Importantly, the former work demonstrated the crucial importance of vesicular SNARE-dependent release of gliotransmitters. The role of vesicular release of gliotransmitters in the modulation of synaptic plasticity in the neocortex remains almost unexplored. We have shown recently that astrocyte-derived ATP can down-regulate postsynaptic GABA receptors in the neocortical pyramidal neurons . The ability of neocortical astrocytes to release ATP in response to elevated neuronal activity  and ability of neuronal ATP receptors to modulate neural excitability through attenuation of phasic and tonic GABA conductance could underlie the impact of astroglia on synaptic plasticity in the neocortex.
We investigated the LTP of the field excitatory postsynaptic potentials (EPSPs) in layer II/III of the somatosensory cortex of wild-type and dnSNARE mice. The EPSPs were evoked by the stimulation of the neuronal afferents descending from layers IV/V, as described previously [10,12]. Potentiation of EPSPs was induced using conventional protocol for neocortex where LTP is induced by several episodes of θ-burst stimulation (TBS). In the young wild-type mice, five episodes of TBS induced robust LTP in all 15 trials (Figure 2A). The extent of LTP of cortical EPSPs was dramatically reduced in the dnSNARE mice (Figure 2A). To elucidate a putative roles for glia-derived ATP, we tried to rescue the LTP in dnSNARE mice by substitution of ATP with its exogenous analogues. The induction of LTP in the dnSNARE mice was rescued by application of non-hydrolysable selective agonist of P2X receptors adenosine 5′-[γ-thio]triphosphate (ATPγS) (Figure 2B). These results indicate the importance of vesicular release of ATP from astrocytes for synaptic plasticity in the neocortex. One of the putative mechanisms of ATP action could be lowering the threshold of LTP induction via down-regulation of neuronal GABA receptors caused by activation of postsynaptic P2X receptors.
Release of ATP from astrocytes modulates synaptic plasticity in the neocortex
Since release of gliotransmitters, in particular ATP, from astrocytes can decline with aging (Figure 1), one could expect this mechanism to contribute to age-related impairment of synaptic plasticity. In agreement with this notion, the magnitude of LTP was much lower in the neocortex of 12-month-old mice than in 3-month-old mice (Figure 2C). The LTP in the neocortex of old mice could be rescued either by application of exogenous ATP and its non-hydrolysable analogues (Figure 2C) or by additional activation of astrocytic Ca2+ signalling by glia-specific  protease-activated receptor 1 (PAR1) (Figure 2D).
Our observation of facilitatory effect of astroglia-driven ATP on long-term synaptic potentiation in the neocortex is in good agreement with previous observations of modulation of hippocampal LTP by ATP released from astrocytes via a SNARE-dependent mechanism . Furthermore, our results give a new insight into the mechanisms of action of ATP as gliotransmitter; in addition to catabolism of ATP to adenosine and modulation of synaptic transmission via presynaptic adenosine receptors [18,22], ATP can enhance neuronal excitability by down-regulating the phasic and tonic inhibition acting via postsynaptic P2X receptors. As our previous results show [10,12,37], down-regulation of postsynaptic GABA and NMDA receptors by Ca2+ influx via purinergic P2X receptors may provide an efficient mechanism of regulation of signalling within tripartite synapse.
There is growing evidence that Ca2+-dependent exocytosis of gliotransmitters plays an important physiological role in communication between astrocytes and neurons. It becomes evident now that release of just two gliotransmitters, ATP and glutamate, can activate a multitude of pre- and post-synaptic regulatory cascades, which can change synaptic efficacy in opposite ways. Taking into account that exocytosis of gliotransmitters from astrocytes may be triggered by Ca2+ elevation created by both metabotropic and ionotropic receptors, one can expect more complex behaviour of tripartite synapse than it was previously assumed. Moreover, bidirectional signalling between neurons and astrocytes can undergo considerable age-related decline, which can contribute to impairment of cognitive function.
Astrocytes in Health and Neurodegenerative Disease: A joint Biochemical Society/British Neuroscience Association Focused Meeting held at Institute of Child Health, University College London, London, U.K., 28–29 April 2014. Organized and Edited by Jon Cooper (Institute of Psychiatry, King's College London, U.K.), Diane Hanger (King's College London, U.K.), Wendy Noble (King's College London, U.K.), Michael Sofroniew (University of California Los Angeles, U.S.A.), Alexei Verkhratsky (University of Manchester, U.K.) and Brenda Williams (Institute of Psychiatry, King's College London, U.K.).
This work was supported by the Biotechnology and Biological Sciences Research Council [grant number BB/K009192/1].