Astrocytes make up approximately half of the volume of the adult mammalian brain and are considered mainly as supportive cells for neurons [1]. It has been known for a long time that astrocytes play a role in maintaining extracellular homeostasis in the brain. Astrocytes maintain extracellular K⁺ and pH levels by expressing transporters for these ions on their membranes [2,3]. Likewise, astrocytes express transporters of neurotransmitters such as glutamate, which allow them to uptake extra-synaptic glutamate spill-over [4,5]. In addition, astrocytes are involved in maintaining structural integrity of the blood-brain barrier (BBB) [6,7].

Besides such traditional role of astrocytes for structural integrity and brain homeostasis, during the last decade it was discovered that astrocytes are crucially involved in the regulation of synaptic transmission. Astrocytes usually project their fine cytoplasmic processes in vivo and contact synapses and blood vessels in their vicinity in the brain. According to a study by Allen and colleagues, a single astrocyte can contact up to approximately 140,000 synapses, and more than 99% of the cerebrovascular surface is ensheathed by astrocyte end-feet [8]. Therefore, astrocytes are in a strategic position to monitor the microenvironment of neuronal synapses as well as brain microvessels. It has found that astrocytes express various neurotransmitter receptors including metabotropic (mGluR) and ionotropic (AMPAR) glutamate receptors, ATP receptors, and GABA receptors [5,8,9,10]. This finding suggests that neurotransmitters released from the presynaptic terminal not only activate postsynaptic neurons, but may also transmit information to peri-synaptic astrocytes. Indeed, it has been shown that activation of synaptic transmission induces intracellular calcium increase in peri-synaptic astrocytes [11,12]. In addition to having neurotransmitter receptors, astrocytes are also able to produce and release neurotransmitters by themselves such as glutamate [13], D-serine [14], ATP [15], and GABA [16], which are collectively called "gliotransmitters". Therefore, astrocytes are not merely listening to neuronal synaptic activity but also appear to be capable of modulating neuronal activity. Studies have shown that such gliotransmitters released from astrocytes are able to regulate the excitability of pre- and post-synaptic neurons [1,8,17,18,19]. Based on the evidence of bidirectional communication between astrocytes and pre/post-synaptic neurons, a novel concept of the tripartite synapse has been proposed (Fig. 1) [1].

Moreover, astrocytes are connected with each other via gap junctions composed of connexin channels and form extensive communication networks on their own [20]. The pore size of the connexin channels are large enough to be permeable to intracellular second messenger molecules such as cAMP and calcium [21,22]. As a result, astrocytes can communicate with each other via these signaling molecules through open connexin channels. It is well documented that intracellular calcium increase in a single astrocyte can be propagated to other astrocytes resulting in Ca⁺⁺ waves [20,23,24,25,26]. Thus far, how astrocytes release gliotransmitters has not been conclusively resolved. Data supporting vesicular as well as channel-mediated release have both been documented [10,17,26,27]. Intracellular calcium increase has been suggested as a common signal driving gliotransmitter release [1,8,11]. Assuming that calcium signals propagated via astrocyte gap junctions are capable of releasing gliotransmitters from other astrocytes, para-synaptic regulation and synchronized modulation of information processing by astrocytes are also conceivable.

The numerous studies showing astrocyte regulation of synaptic transmission suggested that astrocytes may also play a role in higher brain functions, which was substantiated by recent studies using astrocyte-specific genetically modified mice. For instance, Theis and colleagues reported that astrocyte-specific connexin-deficient mice showed enhanced locomotory activity [20]. These data indicated that the astrocyte network, possibly due to astrocyte calcium waves via gap junctions, is critical for such higher brain functions. In addition, conditional gene deletion of leptin receptors in hypothalamic astrocytes resulted in decreased feeding and altered synaptic inputs onto hypothalamic arcuate nucleus neurons [28]. This study suggests a putative role of hypothalamic astrocytes in synaptic transmission of hypothalamic neurons and subsequent feeding control. In addition, astrocytes were shown to modulate sleep behavior and the cognitive decline associated with sleep loss [29]. Likewise, astrocytes in the brainstem were shown to sense oxygen level in the blood via interstitial pH drop, and regulate respiration [30]. Based on these studies, it is nowadays indisputable that astrocytes play critical roles in physiological brain functions, although the underlying mechanisms still wait to be elucidated.

Microglia are another type of glia that have unique characteristics compared to other glial cells. Unlike other cell types in the brain, they are derived from the yolk sac during development and differentiate into brain-resident innate immune cells [31,32]. Like other innate immune cells in peripheral organs, they constantly monitor their microenvironment for infection or tissue damage with their fine processes [33,34]. Upon infection or tissue damage, they are immediately activated and express genes involved in inflammatory responses and wound healing. In addition, they obtain phagocytic functions and clear pathogens and cellular/tissue debris [35,36,37,38]. Due to their role in brain tissue inflammation, the function of microglia has been studied mainly in the context of neurological disorders for the past several decades.

However, a series of recent studies indicate that microglia are also actively involved in synapse remodeling. It was initially reported by Schafer and colleagues that microglia are involved in synapse pruning during brain development [35]. Later, it was found that microglia eliminate synapses in the adult brain as well [39,40,41]. Of interest, microglia utilize the same molecular machinery for synapse pruning as they utilize for phagocytic clearance of pathogens or cell/tissue debris; they utilize complement receptors to eliminate synapses [39] as well as for elimination of complement-coated pathogens. This discovery suggests that microglia may play an active role in regulating synaptic plasticity in the adult brain by regulating synapse remodeling. Synaptic plasticity is a key element of higher brain functions such as learning and memory. Therefore, it can be reasoned that microglia activity may play certain roles in such higher brain functions. This possibility was supported by the study by Parkhurst wherein microglia-depleted mice showed a significant reduction in learning and memory [42]. In this study, the authors showed that BDNF is released from activated microglia and affects synaptic plasticity by regulating synapse formation.

Taken together, these studies substantiated the active role of glia in physiological synaptic transmission, and support the putative role of glial cells in higher brain functions such as cognitive and affective brain functions. However, thus far, the in vivo function of glial cells in higher-order brain functions has not been clearly elucidated, mainly due to the lack of experimental techniques to manipulate glial cell activity in vivo in behaving animals. In this regard, recently developed optogenetic techniques may provide valuable tools to dissect in vivo glial cell functions.

Optogenetic manipulation of astrocytes was first reported by Gourine and colleagues [30]. In the study, they showed that illumination of ChR2-expressing astrocytes leads to a pH decrease followed by ATP release [30]. Using this optogenetic astrocyte manipulation system, they found that ATP released from activated astrocytes in the brainstem activated chemoreceptor neurons and regulated respiration. Optogenetic astrocyte manipulation was later shown to modulate astrocyte activity in the visual cortex and regulate response selectivity of visual cortex neurons [3]. In this study, light-stimulation of ChR2-expressing visual cortex astrocytes enhanced both excitatory and inhibitory synaptic transmission in neurons nearby. This regulatory effect of astrocytes was dependent on the metabotropic glutamate receptor (mGluR1) implicating glutamate as a gliotransmitter in the response upon optogenetic stimulation. Glutamate release by astrocyte ChR2 stimulation was directly demonstrated by Sasaki and colleagues [44]. In this study, optogenetic ChR2 stimulation of cerebellar astrocytes, including Bergmann glia, activated AMPA receptors on Purkinje cells and induced long-term depression of parallel fiber-to-Purkinje cell synapses. More recently, Yamashita and colleagues [45] reported that optogenetic astrocyte Ch2R activation in the anterior cingulate cortex during neuropathic pain can modulate wakefulness and sleep. Similarly, ChR2 stimulation of astrocytes in the posterior hypothalamus dramatically induced sleep during the active phase of the sleep-wake cycle [46] affirming the involvement of astrocytes in sleep-wake regulation. These studies demonstrate that astrocyte ChR2 stimulation successfully regulates astrocyte activity. Of note, the final outcome of optogenetic astrocyte activation is brain region-specific. Optogenetic Ch2R stimulation of brainstem astrocytes results in ATP release and respiratory regulation, while Ch2R activation of cerebellar astrocytes results in glutamate release and perturbation of cerebellar modulated motor behaviour [44].

To drive astrocyte-specific opsin expression in vivo, two main approaches have been employed. First, a recombinant adeno-associated virus was used for expression of opsin constructs in the central nervous system [3,30,45,46]. Viral vectors encoding opsin genes under the astrocyte-specific promoter (e.g. GFAP) have been utilized to drive astrocyte-specific ChR2 expression in the mouse brain [3,30,46]. On the other hand, viral vectors encoding opsin genes preceded by a floxed-stop codon sequence can be transduced into astrocyte-specific Cre transgenic mice [45]. Alternatively, astrocyte-specific opsin transgenic or knock-in mice can also be used [44,47]. Both Cre/loxP and tetO-tTA systems can be used for astrocyte-specific opsin-transgenic mice, which can be generated by mating astrocyte-specific Cre or tTA mice with floxed-stop-opsin knock-in mice or tetO-opsin transgenic mice. Using this tetO-tTA system, Tanaka and colleagues [47] generated astrocyte specific ChR2 transgenic mice that express sufficient amounts of ChR2 in astrocytes for optogenetic astrocyte manipulation. In these bigenic mice (Mlc1-tTA::tetO-ChR2), they showed that optogenetic astrocyte ChR2 stimulation triggers glutamate release in vivo [44]. These prior studies support that astrocyte gliotransmitter release and subsequent synaptic transmission can be regulated by optogenetic manipulation, which can be a valuable tool to uncover the physiological role of astrocytes in higher-order brain functions.

Although optogenetics is a highly valuable tool, it should be mentioned that it has several limitations applying it to astrocyte manipulation. For instance, the intracellular calcium increase in astrocytes after optogenetic ChR2 activation does not exactly recapitulate physiological astrocyte calcium elevation that is mostly due to GPCRs activation. While optogenetic ChR2 stimulation supposedly increase calcium level by calcium influx [30], physiological calcium signals by GPCR activation triggers Ca²⁺ release from ER [48,49,50,51]. Furthermore, recent studies indicate that astrocytic Ca²⁺ responses are distinct depending on the subcellular compartments due to the heterogeneous expression of receptors and Ca²⁺ sources. Therefore, it is possible that astrocytic Ca²⁺ signaling induced by ChR2 may not represent the physiological astrocyte activation regulating synaptic functions. In this regard, development and utilization of other optogenetic tools such as optoXR [52] or OptoSTIM1 [53] might be required to dissect and manipulate physiological astrocyte function.

Unlike astrocytes, the possibility of optogenetic microglia manipulation has not yet been reported. As previously mentioned, microglia can regulate synapse formation/elimination by releasing BDNF. In addition, microglia-derived BDNF was shown to regulate synaptic transmission by disinhibiting inhibitory interneurons [54,55]. BDNF release can be evoked by an intracellular calcium increase in microglia. In this regard, it is conceivable that optogenetic stimulation that accompanies intracellular calcium increase may lead to BDNF release from the activated microglia, and thereby regulate synaptic transmission and plasticity [55,56]. Meanwhile, it has also been documented that microglia activation is accompanied by inward chloride conductance, which contributes to microglial phagocytic activity and NO production [57]. Chloride influx can be recapitulated by the introduction and activation of NpHR, a chloride transporter. In this regard, it will be interesting to test if optogenetic NpHR activation of microglia under yellow light (540 nm) illumination can regulate microglial phagocytic activity that is an underlying mechanism of microglial synaptic pruning. Taken together, it is theoretically possible that optogenetic ChR2 or NpHR stimulation of microglia may affect their ability to physiologically regulate synaptic activity, which should be tested in future studies.

Besides BDNF release, microglia activation is often accompanied by the expression and release of other proinflammatory genes including cytokines and chemokines. Some of these genes, namely TNF-α and ROS, are also implicated in the regulation of neuronal excitability [58,59], and thus have a potential to affect synaptic transmission. Intracellular calcium increase is a common signaling pathway leading to expression of these proinflammatory genes, which can potentially be regulated by optogenetic microglia stimulation. Although intracellular calcium increase after optogenetic ChR2 stimulation is detected in neurons and astrocytes, it has not been confirmed in microglia. Therefore, it is possible that ChR2 activation may not be sufficient to induce microglia activation (e.g. BDNF release). If this is the case, the optically regulated G-protein coupled receptor (GPCR) gene (opto-XR) [52] can be utilized instead. Since most microglial intracellular calcium signaling is dependent on the Gq-PLC pathway, optic stimulation of optoXR will be most likely to induce microglia activation. In addition, Heo and colleagues recently invented an ingenious light-responsive intracellular calcium signal activator called optoSTIM1 that activates the endogenous store-operated calcium channel CRAC by light-induced oligomerization of STIM1 [53]. Applying this novel genetically modified intracellular calcium signal activator to microglia will be a promising technique to provide more efficient tools to manipulate microglia activation and dissect their in vivo roles in brain functions in the future.