Gliotransmission: A global view
Gliotransmission is the complex process by which astrocytes communicate with neurons and other glial cells to modulate synapses by releasing ‘gliotransmitters’ (for example ATP, glutamate and D-serine), chemicals with the ability to modulate neurons by interacting with pre- or post-synaptic receptors to control their output and modulate neuronal network activity. Research in the past 20 years has portrayed Glial cells as active players in the modulation of neural networks and synapses as opposed to just a structural component of the brain.
The electrical capacities of neuronal cells are not paralleled by astrocytes (a form of glia) which are highly compartmentalised cells with slow transmission abilities and thus poor long-distance communication abilities. Therefore, defining methods of bidirectional neuron-glial communication has been a key focus of research, with elevated calcium transients as a widely accepted mechanism and some other unknown mechanisms, research in this area is hotly debated, with some studies even suggesting that gliotransmission does not exist. Most studies suggest that, in response to elevated intracellular calcium levels of glia, gliotransmitters, are released and act on pre/post synaptic receptors of neurons, or even on axons or outside of the synapse.
Over the recent years gliotransmission has sparked controversy across the glia biology community. In this essay I will review the literature on gliotransmission to determine their role in the brain and within neural networks, discussing the bidirectional communication, mechanisms by which this occurs, Gliotransmitters involved and opposing views to its existence.
Exocytosis in Astrocytes involving SNARE and VAMP
Ca(2+)-regulated exocytosis in astrocytes is vital for the release of gliotransmitters such as ATP or for modulating the surface expression of glutamate transporters and uptake by astrocytes . This process occurs differently in astrocytes, on a slower time scale involving Small synaptic-like vesicles in astrocytes carry vesicle-associated vSNARE proteins, VAMP3 (cellubrevin) and VAMP2 (synaptobrevin 2). VAMP3 was found to be expressed in mouse cortical astrocytes, which is an isoform of VAMP2, with these VAMP3 vesicles undergoing Ca(2+)-independent cycling at the plasma membrane modulated by cAMP, which regulates membrane glutamate transporters . Although some sceptics refute studies involving SNARE and VAMP2 due to small quantities compared to neuronal vesicle numbers it is thought that quick recycling of vesicles occur, thus large quantities of vesicles are not required.
Exocytosis in astrocytes is a vital process contributing to synaptic plasticity supported by the finding that glutamate exocytosis, mediated by NMDA ionotropic glutamate receptors in the rat hippocampus enhanced synaptic strength at excitatory synapses between path afferents and granule cells . Furthermore, NMDA receptor 2B subunits were found on the extra-synaptic part of nerve terminals with receptor distribution spatially related to glutamate-containing synaptic-like micro-vesicles in astrocytic processes . This begins to paint a picture of the complexity of gliotransmission as multiple pathways are intertwined and defined spatiotemporally. This pathway is endogenously activated by neuronal activity–dependent stimulation of purinergic P2Y1 receptors on the astrocytes which provides functional and ultrastructural evidence for astrocyte-mediated control of synaptic activity via exocytosis of glutamate .
Ca2+signaling in LTP at hippocampal CA3-CA1 synapses
The simultaneous publication of two papers reporting contradictory results on the role of astrocytic Ca2+signaling in LTP at hippocampal CA3-CA1 synapses sparked controversy in the glial biology community. In one study, LTP was disrupted by addition of Ca2+ chelator into CA1 astrocytes, concluding that Ca2+-dependent d-serine release was the astrocyte mechanism necessary for LTP  .However, the opposing study used transgenic disruption or activation of the astrocyte Ca2+ signalling pathway in their experiment, observing no difference in LTP . This conflicting evidence was widely accepted by the community of sceptics as proof of non-existence of gliotransmission, however closer methodological examination reveals an incomplete understanding of sources and mechanisms of Ca2+ signalling in astrocytes and suggests diversity in the astrocytic contribution to LTP depending on the circuit and form of LTP studied . Transgenic disruption studies were often not efficacious in their results, producing “negative” evidence, for example transcriptome analysis data for vGlut mRNA expression being low/undetectable as this could be due to only small amounts of expression needed to produce a response .
It appears that astrocytes do not passively readout synaptic activity level, rather they display selective responsiveness to specific synaptic inputs, showing calcium-based excitability that displays nonlinear input–output relationships, and have intrinsic properties that support nonlinearity, demonstrating the complexity of neuron-astrocyte communication and presenting astrocytes as synaptic information processors .
In the hippocampus astrocyte calcium elevations and subsequent glutamate release lead to the synchronous excitation of clusters of pyramidal neurons, suggesting that gliotransmission may contribute to neuronal synchronization . A single gliotransmitter is shown to exert multiple effects on the neuronal network, an example being astrocyte Ca2+elevations induced by activation of PAR-1 receptors, but not P2Y1 receptors, stimulated glutamate release to evoke NMDA receptor-mediated SICs (slow currents) in pyramidal neurons of the hippocampus .
Astrocytes were also shown to enhance synaptic strength through activation of presynaptic group I metabotropic glutamate receptors in CA3–CA1 synapses to facilitate synaptic transmission through activation of presynaptic kainate receptors in interneurons, increasing neurotransmission through activation of presynaptic NMDA receptors in dentate granule cells, which modulated inhibitory synaptic transmission between interneurons and CA1 pyramidal cells and depressed inhibitory transmission through activation of group II/III metabotropic glutamate receptors . In the CA1 area of the hippocampus, glutamate released from Schaffer collaterals excites postsynaptic pyramidal neurons as well as inhibitory interneurons, which activates astrocytes elevating their intracellular Ca2+, stimulating the release of ATP, leading to synaptic depression of adjacent excitatory synapses, showing that heterosynaptic depression is mediated by the coordinated activation of successive intercellular signalling events . These mechanisms of synchronous activity and activation of pre-synaptic receptors at individual cells may strongly affect hippocampal function, demonstrating that astrocytes are a key player in modulation of hippocampal synapses.
BESTROPHIN-1 AND TREK-1
As noted above, Astrocytes release glutamate upon activation of various GPCRs to exert important roles in synaptic functions, although the mechanism of release has been debated. Studies have suggested two kinetically distinct modes of non-vesicular, channel-mediated glutamate release. The fast mode requires activation of G(αi), dissociation of G(βγ), and subsequent opening of glutamate-permeable potassium channel TREK-1. The slow mode is Ca(2+) dependent, requiring G(αq) activation and opening of glutamate-permeable, Ca(2+)-activated anion channel Best1 . These distinct modes of gliotransmitter release are important for communicating different signals resulting in tonic or phasic activity.
Neuronal inhibition has recently drawn much attention; however, the mechanisms involved in tonic release of and the cellular source of the neurotransmitter involved, γ-aminobutyric acid (GABA), have been difficult to define. The tonic release of GABA in the cerebellum occurs through the Best-1 anion channel of astrocytes and Bergmann glial cells confirming that glia can serve as a source of GABA for tonic inhibition of neurons . Ultrastructural analyses demonstrate that TREK-1 is preferentially localized at cell body and processes, whereas Best-1 is mostly found in microdomains of astrocytes near synapses . This suggests that localisation of certain channels can aid our understanding of localisation of different forms of transmission in brain areas as these two distinct sources of astrocytic glutamate that can differentially influence neighbouring neurons.
Modulation of mammalian spinal CPG for locomotion
CPG networks prove a useful tool for studying gliotransmission as they are a more simplistic pathway than those found higher brain regions providing more clear-cut results. The contribution of glia in the modulation of the mammalian spinal CPG for locomotion was studied using the agonist TFLLR to activate protease-activated receptor-1 (PAR1), an endogenous GPCR expressed by spinal glia and found that this resulted in a decreased frequency of locomotor-related bursting. This was proven to be glial specific as with the application of gliotoxins, and TFLLR had no effect . This pathway was blocked by Theophylline, a non-selective adenosine receptor antagonist and the A1 antagonist but not the A2a antagonist which shows that extracellular adenosine upon glial stimulation results in A1 receptor mediates inhibition of neuronal activity in locomotor networks.
The next phase of the experiment involved blocking ectonucleotidase inhibitors ARL67156 and network modulation was found to be blocked indicating that glial cells release ATP, subsequently converted to adenosine by ectonucleotidases which go on to modulate neuronal activity. This adenosine has been shown to act via inhibitory circuitry components to modulate output. Several further studies have detected Ca2+-dependent release of ATP from glia, which is subsequently degraded to adenosine and resulting in neuronal A1 or A2Aadenosine receptor activation . This experiment demonstrated the exact modulation of the CPG by astrocytes by breaking down each step of the pathway to ensure glial cells were responsible for the effects seen.
Further studies on the adenosine pathway have revealed increasing complexity as a cross-over with dopamine-like receptors was found which resulted in a negative-feedback loop with astrocytes acting as the inhibitor. Endogenous adenosine was found to act as antagonist for Dopamine-like receptor 1 (D1LR) and thus decreased locomotor CPG activity, whereas the D1LR agonist SKF38393 increased the frequency of locomotor related bursting induced by 5htp and n-methyl-d aspartate, showing that in this network gliotransmission acts as a negative feedback mechanism decreasing locomotor-related bursting which is increased by the dopamine pathway . Furthermore, modulation of locomotor-related activity by adenosine was abolished following ablation of glia, demonstrating that glial cells are the principal source of modulatory adenosine in murine spinal motor networks .
These studies provide evidence that adenosine produced upon stimulation of astrocytes acts via A1Rs to inhibit signalling through D1LRs. Astrocytes are proposed as the principal source of modulatory adenosine in spinal motor networks adding a layer of complexity to neuron-astrocyte cross talk in spinal motor networks providing a mechanism by which a second-order neuromodulator refines the effects of a first-order neuromodulator with a behaviourally relevant network output.
Interpretation by a gliotransmission sceptic
A major controversy persists within the field of glial biology concerning whether, under physiological conditions, neuronal activity leads to Ca2+-dependent release of neurotransmitters from astrocytes, known as gliotransmission. From the perspective of a sceptic, the results of ‘gliotransmission’ seen are a result of techniques to cause an effect, with evidence gathered using astrocyte-specific and more physiological approaches suggesting that gliotransmission is a pharmacological phenomenon rather than a physiological process .
Evidence against gliotransmission can be found in studies using approaches involving stimulation of Gq-GPCRs expressed only in astrocytes, as well as removal of the primary proposed source of astrocyte Ca2+ responsible for gliotransmission. These approaches contrast with those supportive of gliotransmission, which include mechanical stimulation, strong astrocytic depolarization using whole-cell patch-clamp or optogenetics, uncaging Ca2+ or IP3, chelating Ca2+ and bath application of agonists to receptors. Their argument is that these techniques are not subtle and therefore are not supportive of recent suggestions that gliotransmission requires very specific and delicate temporal and spatial requirements, thus further studies must be conducted.
Other evidence found, includes lack of propagating Ca2+waves between astrocytes in healthy tissue and lack of expression of vesicular release machinery, provides additional evidence against gliotransmission, according to sceptics with their data suggesting that Ca2+-dependent release of neurotransmitters is the duty of neurons, and not astrocytes under physiological conditions .
The contributions of astrocytes to synaptic plasticity has been highlighted in many studies as a result of their unique properties, including astrocytic-receptor and channels allowing dynamic interactions with neurons. Connexin 30, one of the two main astroglia gap-junction subunits, is thought to be involved in behavioural and basic cognitive processes. Although the underlying cellular and molecular mechanisms are unknown, it has been shown that in mice, connexin 30 controls hippocampal excitatory synaptic transmission through modulation of astrocytic glutamate transport, directly altering synaptic glutamate levels . Connexin 30 regulated cell adhesion, migration and modulation of glutamate transport, occurring independently of its channel function, mediated by morphological changes controlling insertion of glial processes into synaptic clefts. By setting excitatory synaptic strength, connexin 30 was found to play an important role in long-term synaptic plasticity and in hippocampus-based contextual memory . These results portray connexin 30 as a valuable regulator of synaptic strength by controlling the synaptic location of glial processes .
Further behaviour evidence comes from studies of neuron-glia interactions mediated by gap-junctional coupling between glial processes, with connexin-43 (Cx43) as a major channel component abundantly expressed. It was found that Cx43 knockout mice and the transgenic line expressing GFAP:Cre exhibited a significant loss of Cx43 in astrocytes in the barrel cortex with further behavioural tests showing that the ability of Cx43 KO mice to sense the environment with their whiskers decreased . This suggests that Cx43-mediated gap-junctional coupling between astrocytes is important in the neuron-glia interactions required for whisker-related sensory functions and plasticity . This study was useful in demonstrating a behavioural change in response to disruption of gap junctions of glial cells highlighting their importance in mammalian behaviour.
Earlier it was thought that astrocytic signalling was simplistic with a global elevation in calcium being responsible for output, however a recent study used 3D imaging to study the endogenous calcium activity in vivo in awake mice, finding mostly asynchronous signalling, which was spatially uncoupled, occurring in frequent fast local transients, especially in cell peripheries or gliapil . This study shows that at present there is still much to be discovered about the mechanisms by which calcium results in different outputs, although previous studies have highlighted that one large calcium elevation often has a smaller output than fast local small elevations, future studies may be able to further define these relations.
To conclude, the field of glial biology has grown in the past 20 years with the discovery that astrocytes possess a much more complex role than earlier thought, contributing to synaptic plasticity and LTP, modulating neural networks and enabling synchronisation of neurons or networks. Due to the intricate nature of the astrocyte that their full potential cannot be studied with todays technology and finer tools are needed, especially with the finding that smaller faster calcium transients in the periphery and gliapil hold high importance and output. Alternative technologies aimed at other astrocytic intracellular signals such as IP3, cGMP and cAMP hold potential to provide insights, possibly revealing novel intracellular and intercellular signalling mechanisms. Perhaps what we can gain from this review is that gliotransmission is a far more complex process than previously thought with interpretation of data being difficult due to intertwined networks, therefore defining causality becomes difficult as there are so many aspects to account for, thus studies on murine spinal networks are useful in providing a clear-cut link between a pathway and its effect.
Astrocytic cellular subtypes may differ phenotypically and functionally, dependent on the neuron–glia network and characterising phenotypic and functional properties will prove insightful, for example glutamate-mediated signalling is a widespread phenomenon, however in different brain regions it likely has differing physiological meaning in neuronal network function.
Research to date uses mice however it is likely that astrocytes in humans could be responsible for the complexity of the higher brain enabling higher thinking as they are large with thousands of connections to synapses and possibly different mechanisms to the ones found at present.
 D. Li, H. Karine, K. Zylbersztejn, M. A. Lauterbach, M. Guillon, M. Oheim, and N. Ropert, “Astrocyte VAMP3 vesicles undergo Ca 2 + -independent cycling and modulate glutamate transporter trafficking,” vol. 13, pp. 2807–2832, 2015.
 P. Jourdain, L. H. Bergersen, K. Bhaukaurally, P. Bezzi, M. Santello, M. Domercq, C. Matute, F. Tonello, V. Gundersen, and A. Volterra, “Glutamate exocytosis from astrocytes controls synaptic strength,” vol. 10, no. 3, 2007.
 C. Henneberger, T. Papouin, S. H. R. Oliet, and A. Dmitri, “Europe PMC Funders Group Long term potentiation depends on release of D-serine from astrocytes,” vol. 463, no. 7278, pp. 232–236, 2010.
 C. Agulhon, T. A. Fiacco, and K. D. Mccarthy, “Hippocampal Short- and Long-Term Plasticity Are Not Modulated by Astrocyte Ca 2+ Signaling,” vol. 09601, no. March, pp. 1250–1255, 2010.
 I. Savtchouk and A. Volterra, “Gliotransmission: Beyond Black-and-White,” J. Neurosci., vol. 38, no. 1, pp. 14–25, 2018.
 A. Araque and M. Navarrete, “Glial cells in neuronal network function,” pp. 2375–2381, 2010.
 D. Acton and G. B. Miles, “Stimulation of Glia Reveals Modulation of Mammalian Spinal Motor Networks by Adenosine,” pp. 1–17, 2015.
 E. Ben Jacob, G. Wallach, J. Lallouette, N. Herzog, M. De Pitta, H. Berry, and Y. Hanein, “Glutamate Mediated Astrocytic Filtering of Neuronal Activity,” vol. 10, no. 12, 2014.
 E. Witts, K. Panetta and G. Miles, “Glial-derived adenosine modulates spinal motor networks in mice”, Journal of Neurophysiology, vol. 107, no. 7, pp. 1925-1934, 2012.
 S. Lee, B. Yoon, K. Berglund, S. Oh, H. Park, H. Shin, G. Augustine and C. Lee, “Channel-Mediated Tonic GABA Release from Glia”, Science, vol. 330, no. 6005, pp. 790-796, 2010.
 U. Pannasch, D. Freche, G. Dallérac, G. Ghézali, C. Escartin, P. Ezan, M. Cohen-Salmon, K. Benchenane, V. Abudara, A. Dufour, J. Lübke, N. Déglon, G. Knott, D. Holcman and N. Rouach, “Connexin 30 sets synaptic strength by controlling astroglial synapse invasion”, Nature Neuroscience, vol. 17, no. 4, pp. 549-558, 2014.
 Y. Han, H. Yu, M. Sun, Y. Wang, W. Xi and Y. Yu, “Astrocyte-restricted disruption of connexin-43 impairs neuronal plasticity in mouse barrel cortex”, European Journal of Neuroscience, vol. 39, no. 1, pp. 35-45, 2013.
 E. Bindocci, I. Savtchouk, N. Liaudet, D. Becker, G. Carriero and A. Volterra, “Three-dimensional Ca2+imaging advances understanding of astrocyte biology”, Science, vol. 356, no. 6339, p. eaai8185, 2017.