Sensing and Regulating Synaptic Activity by Astrocytes at Tripartite Synapse

Abstract

Astrocytes are recognized as more important cells than historically thought in synaptic function through the reciprocal exchange of signaling with the neuronal synaptic elements. The idea that astrocytes are active elements in synaptic physiology is conceptualized in the Tripartite Synapse concept. This review article presents and discusses recent representative examples that highlight the heterogeneity of signaling in tripartite synapse function and its consequences on neural network function and animal behavior.

Introduction

Astrocytes are known to play important homeostatic roles in brain function, providing trophic, structural and metabolic support for neurons [1–3]. They have additionally been shown to display a calcium-based excitability and to be able to act as sensors and modulators of synaptic transmission and plasticity. Through the expression of a wide variety of membrane receptors expressed, astrocytes sense the synaptic activity by responding to different synaptically released neurotransmitters, which generally leads to the elevation of the astrocyte calcium levels. These calcium elevations stimulates the release of gliotransmitters, which acting on neuronal receptors, regulate synaptic transmission and plasticity [4–8]. Thus, in addition to the classical information flow between the pre- and postsynaptic neuronal elements of the synapse, there is a signaling exchange between these neuronal elements and the adjacent astrocytes. This bidirectional communication between astrocytes and the neuronal elements led to the concept of the tripartite synapse, which epitomizes the idea that synaptic function results from the interaction of three synaptic elements, the presynaptic terminal, the postsynaptic cell and the surrounding astrocyte. In this review, we will present and discuss recent paradigmatic examples that highlight the heterogeneity of signaling in tripartite synapse function.

Astrocytes Sense Synaptic Activity

Astrocytes are known to express receptors for a large plethora of neurotransmitters, such as glutamate, GABA, endocannabinoids, dopamine, serotonin, ATP/Adenosine, acetylcholine or opioids [9–12]. Such diversity of receptors illustrates the ability of astrocytes to sense multiple neuronal signals, which can be integrated in a non-linear manner [13] to confer a high variability of the astrocytic responses (Fig. 1). Many of the neurotransmitter receptors expressed by astrocytes are G protein-coupled receptors (GPCRs), which, upon activation, lead to intracellular calcium elevations [14–19]. Gq GPCRs activate phospholipase C that generates diacylglycerol and inositol 1,4,5-trisphosphate (IP3) which, ultimately, induces the release of calcium from the endoplasmic reticulum through activation of IP3 receptors. Type 2 IP3 receptors (IP3R2) have been shown to be the main responsible of the GPCR-mediated calcium mobilization in astrocytes [20–23]. Accordingly, in the IP3R2 knockout mice, which lack IP3R2, astrocyte calcium elevations in the soma are unaffected by neurotransmitters [24]. The relatively small calcium activity observed in restricted regions of the processes have been shown to have a mitochondrial origin [25–27]. In addition to GPCRs-mediated calcium mobilization, ionotropic receptors to glutamate and ATP are involved in astroglial Ca²⁺-signaling and neuron-glia communication [28–31]. Therefore, the role of ionotropic component of Ca2+ mobilization might help to explain some contradictory results obtained in astrocytic IP3R2 knockout-mice.

Fig. 1.

Neuron-astrocyte signaling and lateral communication. (1) the presynaptic terminal releases neurotransmitter. (2) Binding of neurotransmitter to the GPCRs (mGluRs) in the postsynaptic neuron. (3) Increase the postsynaptic calcium via PLC and release of retrograde messenger eCBs which (4) bind to the GPCRs in the astrocyte, CB1R. (5) Induction of calcium release from the ER and astrocyte calcium increase. (6) The exocytosis of endosoms-containing glitransmitters through SNARE complex induces the gliotransmitters release and, in turn, the interaction with GPCRs in the presynaptic terminal of heteroneuronal synapse. (7) Later, it is triggered the modulation of neurotransmitter release and (8) the Induction of slow inward currents dependent on extrasynaptic NMDARs. (Glu glutamate; eCBs endocannabinoids; DA dopamine; NE norepinephrine; Ach acetylcholine)

Notably, activation of Gi/o GPCRs also leads to calcium elevations in astrocytes, although the intracellular signaling pathways activated remains to be fully elucidated (see [32], for a discussion of the potential mechanisms involved). The fact that Gq GPCR activation led to cellular activation of both neurons and astrocytes, whereas the Gi/o GPCR activation led to cellular inhibition in neurons and cellular activation in astrocytes has led to suggest that inhibition is a specific property of neurons and may be fundamentally different between neurons and astrocytes [32].

Astrocytes Regulate Synaptic Transmission and Plasticity Through the Release of Gliotransmitters

The astrocyte calcium signal stimulates the release of gliotransmitters through calciumand SNARE protein-dependent processes, probably involving vesicle exocytosis [8, 33–38]. While astrocytes are known to be able to release different neuroactive substances, glutamate, d-serine and ATP/adenosine are the gliotransmitters most clearly identified [39–42]. The existence of different gliotransmitters that can distinctly impact neurotransmission in different brain areas and circuits represents clear evidence of the heterogeneity of astrocyte-induced synaptic regulation at tripartite synapse.

For example, it has been shown to depress excitatory synaptic transmission through activation of presynaptic A1 receptors in the hippocampus [43, 44] and is also responsible for the developmental regulation of the spike timing-dependent depression in the hippocampus [45]. The synaptic regulation of astrocytic ATP/adenosine has also recently been found in the nucleus accumbens, a key brain area involved in reward and addiction. Astrocytes in this nucleus respond to dopaminergic inputs from the ventral tegmental area with calcium elevations that stimulates ATP/adenosine release the consequent activation of presynaptic A1 receptors and the depression of the excitatory synaptic transmission, which is a crucial synaptic event in brain reward signaling [46].

Dr. Parpura et al., in 1994 described how glutamate released from astrocytes induced neuronal calcium elevation in astrocyte-neuron co-cultures that, however, did not occur in solitary neurons [47]. Moreover, Drs. Parpura and Haydon later described that the astrocytic calcium stimulates glutamate release to modulate adjacent neurons at physiological levels [48]. Furthermore, the modulation of synaptic transmission by astrocytic glutamate has been widely documented in the hippocampus [49–52]. Moreover, it has also been reported that glutamate mediates the spike timing-dependent depression in the barrel cortex [53] and the synaptic potentiation in the dorsal striatum [54]. In nucleus accumbens, metabotropic glutamate receptor 5 (mGluR5) in astrocytes induces Ca2+ elevations with correlated NMDAR-dependent slow inward currents which increase the excitation, raising the astrocytes as potential intermediary in neuronal adaptation [55]. In the hippocampus, the astrocytic glutamate induces a short-term potentiation of the synaptic efficacy through activation of neuronal group I metabotropic glutamate receptors (mGluRs). While endocannabinoids (eCBs) released by neurons lead to direct homosynaptic depression by activating presynaptic type 1 eCB receptors (CB1Rs), they also activate these receptors in astrocytes, leading to the astrocyte-mediated synaptic potentiation of adjacent synapses, a process termed lateral regulation of synaptic transmission [52]. Moreover, astrocytic glutamate is also involved in some forms of synaptic plasticity. Indeed, the astrocyte-induced glutamate-mediated transient potentiation can become long-term potentiation when the nitric oxide is released by postsynaptic neuron [56]. In addition, cholinergic-induced long-term potentiation (LTP) has been shown to be mediated by glutamate released by astrocytes activated by cholinergic inputs [57].

Finally, d-serine, acting as a co-agonist of the N-methyl-d-aspartate receptors (NMDARs), has been found to regulate synaptic transmission and plasticity the hippocampus and barrel cortex [58–60]. In barrel cortex, d-serine improves the cholinergic plasticity induced by whisker stimulation through astrocytic muscarinic acetylcholine receptors (mAchRs) and mediated by NMDARs [58]. In the hippocampus, d-serine has been shown to be crucial for the LTP and object recognition memory task [59, 60]. However, the main source of the d-serine is a matter of debate. While some studies point to the neurons as origin of d-serine, the astrocytes have emerged as main machinery of d-serine release [42, 61]. Nevertheless, further investigations are needed to provide the balanced point of view.

Through binding to neuronal receptors, gliotransmitters have been shown to modulate both excitatory and inhibitory synaptic transmission in many brain areas [43, 46, 50, 52–54, 62–64]. Whether different gliotransmitters are released by different astrocytes or whether single astrocytes can release different gliotransmitters is a relevant question for our understanding of the heterogeneity of both the astrocyte properties and the astrocyte-mediated synaptic regulation. We have recently investigated this issue by stimulating either single interneurons signaling to astrocytes or single astrocytes and monitoring astrocyte-mediated regulation of the excitatory synaptic transmission in the hippocampus. We found that single hippocampal astrocytes can release both glutamate and ATP/adenosine, producing a temporally distinct biphasic regulation of synaptic transmission, which consists in an initial glutamate-mediated synaptic potentiation and a delayed adenosine-mediated synaptic depression [43]. Moreover, the distinct gliotransmitter release was found to be controlled by the neuronal firing activity, suggesting that astrocytes decode neuronal signaling to produce specific regulatory consequences [43].

The heterogeneity of astrocyte-induced synaptic regulation is not only manifested by the release of different gliotransmitters, but also by the same gliotransmitter acting on different neuronal receptors in specific synapses. In the amygdala, astrocytic release of ATP/Adenosine has been shown to distinctly impact excitatory and inhibitory synaptic transmission in neurons of the same amygdala subnucleus, the centromedial amygdala. Astrocyte activation stimulates the release of ATP/adenosine that leads to the A1-mediated potentiation of inhibitory synapses and A2A-mediated depression of excitatory synapses [64]. This differential regulation of synaptic transmission is translated into the decrease of firing rate of neurons of the centromedial amygdala and a decrease of the fear responses of mice subjected to a fear conditioning paradigm. Furthermore, ATP released by astrocytes induces a short-term depression of the inhibitory synaptic transmission through postsynaptic and extrasynaptic GABA_AR down regulation in neocortex [65]. In addition, ATP-derived astrocytes down regulates AMPA by P2XRs and induces depression of field potential in CA1 of hippocampus [66] and downregulates NMDARs trafficking in excitatory terminals with an important role in the induction of LTP [67].

Age-dependent regulation of synaptic transmission by astrocytes represents an important biological variable that has been relatively understudied and deserves further attention to understand the full impact of astrocytes on brain function. Yet, recent studies have addressed the impact of signaling at tripartite synapse during brain development and aging. During development, critical period of synapse function maturation plays a significant role in the establishment of properly efficient neuronal circuits. The group of Rodriguez-Moreno has elegantly shown that the spike timing-dependent plasticity (STDP) in the hippocampus is temporally regulated. The spike timing-dependent depression shown by young animals become spike timing-dependent potentiation in adult mice, a maturation phenomenon that depends on adenosine of astrocytic origin [45, 68]. In aging, neuron-astrocyte signaling has been found to be largely preserved across the lifespan of mice [69], although the decline in the astrocyte-neuron network has been observed. It has been shown a decrease in the P2X, AMPA and NMDARs-mediated miniature glial synaptic currents in old mice [70] and astrocytic Ca2+ signaling age-related decrease that underlies to the synaptic transmission modulation [71]. Together all these changes define a remodeling of synaptic strength and information processing, contributing to a cognitive impairment. In fact, synaptic plasticity is severely compromised in an Alzheimer’s disease (AD) mouse animal model deficient of astrocyte IP3R2-mediated calcium signal at early stages of the disease, indicating that astrocytes and their calcium signaling play crucial roles in the AD pathology, accelerating the progression of synaptic plasticity dysfunction [69].

Astrocytes Regulate Network Function and Animal Behavior

While much effort has been done to investigate the mechanisms underlying astrocytic regulation of synaptic transmission and plasticity and the specific their specific properties in certain synapses and brain regions, their impact on network function and animal behavior have only been initially explored. While recent reviews provide a comprehensive discussion on these issues (see e.g., [72, 73]), we will describe here some recent specific examples of the contribution of astrocytes to neural network function and behavior.

Network function results not only from the activity of glutamatergic excitatory and GABAergic inhibitory signals but also from the activity of neuromodulators like acetylcholine, dopamine or norepinephrine and cannabinoids. Astrocytes have been recently found to respond to these neuromodulators, suggesting that they can mediate their actions in the control of network activity. For example, in vivo cholinergic-induced regulation of LTP in hippocampus and cortex has been shown to be mediated by astrocytes [57, 58]. Han et al., in 2012 showed that exogenous cannabinoid induces LTD in vivo which depends on astroglial CB1R expression with an impairment in the spatial working memory as consequence of this down regulation [74]. Norepinephrine (NE) also signal to astrocytes, and the NE release associated with locomotor activity enhances the astrocyte calcium signaling as a detector of neuronal activity in different brain areas [75]. Finally, dopamine has been recently shown to activate astrocytes in the nucleus accumbens and regulate glutamatergic excitatory inputs in that region thus mediating the behavior effects of the psychostimulant amphetamine [46].

Cortical network function has been found to be regulated by astrocytes [76–80]. More recently, astrocytes in the somatosensory cortex have been shown to respond with calcium elevations to sensory stimulation in vivo that were associated with cortical gamma activity [81]. Sensory stimuli elicit a surge of neuronal network activity in the gamma range that was followed by a delayed astrocyte activity that dampens the steady-state of this activity. This sensory-evoked gamma activity increase is enhanced in IP3R2 knockout mice, in which astrocyte calcium signaling is impaired, and is decreased by pharmacogenetic stimulation of astrocytes with “designer receptor exclusively activated by designer drugs” (DREADDS), indicating that cortical astrocytes respond to sensory inputs and regulate sensory-evoked neuronal network activity maximizing its dynamic range [81]. Astrocytes in the medial prefrontal cortex, a key region involved in goal-directed behavior, have also found to alter the firing properties of cortical neurons and gamma oscillations by modulating the inhibition/excitation balance in that region [82]. Disrupting the astrocyte signaling in this network activity is manifested as working memory deficits [82].

Concluding Remarks

Since the decade of 1990, accumulating evidence of new roles of astrocytes transformed the idea of synapse function, establishing the tripartite synapsis concept that changed the view of classical neuron-neuron communication to include astrocytes as additional important underlying network activity and brain function through the concerted signaling with neurons.