The central nervous system (CNS) is a highly diverse cellular environment, where neurons are surrounded by a multitude of cell types, including astrocytes, microglia, oligodendrocytes, and ependymal cells, which are collectively known as glial cells. For many years, brain function and behavioral output were thought to depend exclusively on neuronal circuit activity, while glial cells were merely regarded as supportive cells. However, it is now recognized that glial cells actively communicate with neurons and modulate neuronal activity.

Among glial cells, astrocytes are the most abundant and are known to be indispensable for correct brain function, as they actively participate in synapse formation, plasticity and function, as well as controlling blood-brain barrier permeability and blood flow, providing metabolic support to neurons and modulating neuroinflammation (Sofroniew and Vinters, 2009). Indeed, astrocytes possess a highly branched, “star-shaped” morphology, enabling them to contact thousands to millions of synapses via peripheral astrocytic processes (Semyanov and Verkhratsky, 2021; Holt, 2023; Salmon et al., 2023). The discovery that astrocytes and neurons communicate with one another at the synapse to modulate synaptic transmission led to the concept of the tripartite synapse (Parpura et al., 1994; Pasti et al., 1997; Bezzi et al., 1998; Newman and Zahs, 1998; Araque et al., 1999).

Upon release from the neuronal pre-synaptic element, neurotransmitters activate neurotransmitter receptors expressed by astrocytes, including G protein-coupled receptors (GPCRs). A rise in astrocytic intracellular Ca²⁺ typically follows, which can then lead to the subsequent (local) release of gliotransmitters and other neuroactive molecules capable of modulating synaptic activity and plasticity (Perea et al., 2009). Importantly, astrocytes can also communicate with each other via gap junctions. This enables locally induced Ca²⁺ signals to propagate to neighbouring astrocytes, allowing these cells to also coordinate the activity of otherwise distant synapses (Giaume et al., 2021).

Given the wide array and range of astrocytic functions, particularly at the tripartite synapse, it is not surprising that astrocytic dysfunction has long been implicated in the pathogenesis of several CNS diseases, including neurodevelopmental, neuropsychiatric and neurodegenerative diseases, making these cells attractive therapeutic targets. In this review, we will first provide a general overview of the involvement of astrocytes in synapse formation, activity and plasticity in development and adulthood. Next, we will shed light on how chemogenetics, a technique using genetically engineered GPCRs to modulate astrocytic activity, induces astrocytic Ca²⁺ signalling and gliotransmitter release, mimicking, to a certain degree, endogenous signalling pathways. Finally, we will finish by discussing how this technique might reveal new molecular pathways that can be exploited therapeutically in the future.

The generation of fully functional neuronal circuits capable of receiving, integrating, and responding to a wide variety of intrinsic and extrinsic stimuli largely depends on the establishment of proper synaptic connections. Synaptogenesis begins when the terminal bouton of a neuron comes into close contact with tiny protrusions, known as spines, on the dendrites of another neuron. Both the shape and size of spines are vital to ensuring adequate synaptic transmission. Thinner spines are considered more unstable and are associated with a more immature or silent state, as they often lack the post-synaptic machinery necessary for synaptic transmission. Once physical contact has been established, spines can undergo maturation. During this process, the spine head enlarges, spines acquire a post-synaptic density (PSD) and neurotransmitter receptors accumulate in the post-synaptic membrane, forming larger mushroom-like spines, and leading to increased synaptic potency (Yoshihara et al., 2009; Xu et al., 2020). During post-natal brain development, critical periods are pivotal to uniquely shaping the CNS. During these periods of heightened experience-dependent circuit remodeling, stable synapses are formed at a high rate. Experience acts by strengthening the relevant ones, which are eventually integrated into the circuit. On the other hand, redundant spines weaken over time, due to lack of relevant stimulation, and are eventually eliminated. This maturation process changes both the cell-cell connections and functional output of an excitatory network producing a more stable and mature circuit. Astrocytes have been identified as key regulators of critical period closure, ensuring proper brain wiring (Ackerman et al., 2021; Ribot et al., 2021). In adulthood, synapses are much more stable but synaptic remodelling and plasticity still occur to enable circuit adaptations to new experiences, driving learning and memory formation, and recovery from CNS injury and disease (Pascual-Leone et al., 2005; Hübener and Bonhoeffer, 2014). Therefore, synaptic assembly and circuit refinement must be tightly regulated as aberrant synapse formation is thought to contribute, for example, to the emergence of neurodevelopmental diseases (Washbourne, 2015).

Synaptogenesis greatly increases following astrocyte differentiation (Freeman, 2010) and the expansion in astrocyte numbers that occurs during the first post-natal week (Ge et al., 2012). Furthermore, synaptogenesis also appears concomitant to astrocyte structural maturation (Clavreul et al., 2019). Both astrocyte-secreted molecules and astrocyte-expressed cell adhesion proteins appear to be important factors in the process (Figure 1A). While progress has been made based on studies focusing on factors mediating excitatory synaptogenesis, it is likely that many other factors are still to be identified. In contrast, inhibitory synapse formation is not yet well understood (Um, 2017). Synapse formation can generally be regarded as a two-stage process. First, structural synapse formation takes place which is then followed by functional synapse maturation.

To induce the structural assembly of glutamatergic synapses, astrocytes secrete the pro-synaptogenic factors thrombospondin 1 and 2 (TSP1 and TSP2) (Christopherson et al., 2005) and hevin (Figure 1A) (Kucukdereli et al., 2011; Risher et al., 2014; Singh et al., 2016). TSP1 and TSP2 act by binding the neuronal receptor α2δ-1 (Eroglu et al., 2009), while hevin promotes synapse assembly by bridging neuronal neurexin-1α (Nrx1α) and neuroligin-1B (NL1B) (Singh et al., 2016). These synapses usually possess synaptic vesicles, active release sites and PSD. However, they are functionally inactive because, even though they possess post-synaptic N-methyl-D-aspartate glutamate receptors (NMDAR), they lack α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid glutamate receptors (AMPAR) and the release of synaptic vesicles is suboptimal (Christopherson et al., 2005).

To produce functionally mature synapses, astrocytes then secrete factors, such as the heparan sulfate proteoglycans glypican 4 and glypican 6 (Gpc4 and Gpc6), which increase the expression of the GluA1 subunit of AMPAR at the post-synaptic terminal (Figure 1A) (Allen et al., 2012). Additionally, astrocyte-secreted Chordin-like 1 (Chrdl1) has been shown to increase the levels of GluA2-containing AMPAR to the synapse, leading to Ca²⁺ impermeability of this ionotropic receptor and contributing to the maturation of excitatory glutamatergic synapses (Brill and Huguenard, 2008; Blanco-Suarez et al., 2018). Astrocytic tumor necrosis factor-alpha (TNF-α) can also recruit AMPAR to excitatory synapses, while decreasing GABA_A receptor density at inhibitory synapses, thus regulating overall neuronal circuit activity (Beattie et al., 2002; Stellwagen et al., 2005; Stellwagen and Malenka, 2006). Astrocyte-derived cholesterol is also crucial for proper synaptic maturation by regulating pre-synaptic vesicle exocytosis (Figure 1A). These effects have been described both in vitro and in vivo (Mauch et al., 2001; Pfrieger, 2003; van Deijk et al., 2017). For instance, in the hippocampus of mice with reduced cholesterol production, the total vesicle pool and the number of synaptic vesicles ready for release at the pre-synapse are reduced, which is accompanied by a reduction in the levels of SNAP-25, a protein necessary for vesicle fusion (van Deijk et al., 2017).

In addition to the factors which have a positive impact on synaptogenesis, astrocytes can also release secreted protein acidic and rich in cysteine (SPARC), which is an antagonist of hevin (Figure 1A). Thus, SPARC negatively affects synapse development by counteracting hevin-mediated synaptogenesis, likely by competitively interacting with the same neuronal proteins as hevin (Kucukdereli et al., 2011). Furthermore, SPARC has been shown to prevent the overaccumulation of AMPAR receptors at the excitatory post-synaptic membrane by destabilizing β3-integrin complexes (Jones et al., 2011), which are important regulators of AMPAR stability at the synapse (Cingolani et al., 2008).

Synapse development and stability are further controlled by astrocyte-neuron adhesion proteins (Figure 1A), which have been extensively reviewed (Tan and Eroglu, 2021). For instance, cultured embryonic retinal ganglion cells seem to require direct contact with astrocytes to form mature synapses, suggesting coordinated actions between secreted and contact-mediated signals in driving synaptogenesis (Barker et al., 2008). One of the most common examples of contact-mediated synaptogenesis is the one mediated by γ-protocadherins. This family of cell adhesion proteins is expressed at the tripartite synapse by neurons and astrocytes alike (Phillips et al., 2003; Garrett and Weiner, 2009) and has been shown to be essential for excitatory and inhibitory synapse formation in vitro and in vivo (Garrett and Weiner, 2009). Astrocyte development and, consequently, synaptogenesis are also controlled by the interaction of astrocytic neuroligins (NL) with neuronal neurexins (Nrx): knockdown of astrocytic NL2 decreases excitatory synapse formation and function while promoting inhibitory synaptic function (Stogsdill et al., 2017). These dynamics are likely very complex as astrocyte-neuron co-cultures only fully mature when both cell types are from the same brain area, implying a degree of regional specialization in astrocyte-neuronal interactions (Morel et al., 2017).

Synaptic pruning is also an essential part of neuronal circuit formation. Astrocytes can balance out their synaptogenic properties by modulating synapse elimination, thus preventing excessive accumulation of superfluous synapses. Astrocytes are phagocytic cells and express the receptors MEGF10 and MERTK. These receptors recognize phosphatidylserines at the surface of target synapses as opsonic signals, leading to their degradation (Chung et al., 2013; Scott-Hewitt et al., 2020). Furthermore, astrocyte-microglia crosstalk also contributes to synaptic pruning. For example, astrocytic IL-33 has been shown to induce microglia-mediated synaptic pruning, although the downstream mechanisms that trigger the microglial response are still unclear (Vainchtein et al., 2018).

It is likely that not all astrocytes have the same secretory phenotype, as they have been described as a rather heterogeneous cell population (Ben Haim and Rowitch, 2017; Khakh and Deneen, 2019; Batiuk et al., 2020). Astrocytic factors have specialized functions to ensure correct circuit maturation and different factors seem to be necessary for the formation and maturation of specific subtypes of synapses. Therefore, astrocytes in different (sub)regions of the brain could be specialized in secreting specific factors which are crucial for synapse formation in that region. For example, astrocyte-secreted hevin appears to be crucial for thalamocortical excitatory synapse formation (Risher et al., 2014). Chrdl1 expression is enriched particularly in upper cortical layers and in striatal astrocytes, indicating that its actions may be restricted to these brain areas (Blanco-Suarez et al., 2018). Astrocytes originating from the dorsal or ventral spinal cord have different gene expression profiles and both seem to be essential to synaptogenesis (Tsai et al., 2012; Molofsky et al., 2014). For example, the elimination of ventral astrocytes expressing Sema3a compromises motor and sensory neuron circuit formation (Molofsky et al., 2014). Hence, synaptogenesis appears to require a complex interplay of astrocytic molecules and signals that must be tightly coordinated to control circuit formation and refinement in vivo, which is far from being completely understood (Holt, 2023). Phagocytic capacity may also vary between astrocyte populations and brain regions. More detailed knowledge about these aspects of astrocyte physiology and function will be necessary to be able to phenocopy or boost astrocyte function as a therapy for brain disease in the future.

Functional, mature synapses continuously transfer information between neurons via the release of neurotransmitters, neuropeptides and neuromodulators. The remarkable discovery that astrocytes are also active participants in synaptic transmission, responding to and controlling neuronal excitability and synaptic plasticity through various mechanisms established the concept of the tripartite synapse (Araque et al., 1999) (Figure 1B).

Astrocytes express a wide variety of neurotransmitter transporters, including glutamate transporters, such as the glutamate transporter 1 (GLT-1) and glutamate/aspartate transporter (GLAST) (Figure 1B), and GABA transporters, such as GAT1 and GAT3, which take up their respective neurotransmitters from the synaptic cleft and extrasynaptic space following synaptic transmission. This restricts neurotransmitter action, preventing excitotoxicity which would lead to synaptic dysfunction (Ishibashi et al., 2019; Mahmoud et al., 2019).

Astrocytes also express a wide array of neurotransmitter receptors (Figure 1B). Neurotransmitters thus not only act by stimulating or inhibiting the pre- and post-synaptic neuronal elements but also control astrocytic activity. Most astrocytic neurotransmitter receptors consist of GPCRs, including metabotropic glutamate receptors (mGluR), purinergic receptors (P2Y), and the GABA_B receptor (Liu et al., 2021). Activation of these receptors typically triggers Ca²⁺ signals in astrocytes via the phospholipase C (PLC)-IP₃ pathway. Briefly, upon activation by their ligand, many GPCRs, such as Gq-coupled mGluRs, stimulate PLC to form IP₃ which then induces Ca²⁺ release from the endoplasmic reticulum via activation of IP₃ receptors (IP₃R) (Volterra et al., 2014; Shigetomi et al., 2016). Depending on their nature and intensity, Ca²⁺ signals can then propagate to neighbouring cells via gap junctions, thus highlighting the complexity of Ca²⁺ signals in astrocytes.

Astrocytes seem to discriminate activity originating not only from different brain regions but also from different neuronal subtypes within the same region. As revealed by astrocytic Ca²⁺ imaging, hippocampal astrocytes from the stratum oriens of the CA1 respond to cholinergic but not glutamatergic inputs originating in the alveus (Araque et al., 2002), whereas the same astrocytes respond to glutamate if the signal originates from the Schaffer collateral (Perea and Araque, 2005). In the barrel cortex, astrocytes in layer 2/3 increase intracellular Ca²⁺ levels in response to glutamatergic stimuli from layer 4, but not from layer 2/3 (Schipke et al., 2008). A single astrocyte simultaneously contacts thousands of synapses, which may be excitatory or inhibitory, making it highly likely that astrocyte processes associated with different types of synapses express different receptors (Holt, 2023). Hence, astrocytes can respond to neuronal activity in a highly intricate way that is cell-, region-, and pathway-specific.

Astrocytes are thought to respond to neuronal activity by releasing small neuroactive molecules (gliotransmitters), including glutamate, GABA, D-serine, and ATP, which can in turn modulate synaptic activity (Figure 1B). A variety of mechanisms for gliotransmitter release have been proposed. Vesicular gliotransmitter release has been widely proposed and is thought to be controlled, at least in part, by synaptically-evoked increases in intracellular astrocyte Ca²⁺ (Perea et al., 2009). However, alternative mechanisms have also been proposed, not all of which are Ca²⁺-dependent, such as diffusion through conductance pores opened following P2X₇ activation by ATP and swelling-activated anion channels, or even diffusion through gap junction hemichannels (Hamilton and Attwell, 2010). Once released, gliotransmitters can contribute to regulate neuronal excitability and synaptic plasticity by inducing long-term potentiation (LTP) or long-term depression (LTD). LTP results in synaptic strengthening by increasing synaptic responses, while LTD weakens synapses due to decreasing synaptic responses (Turrigiano, 2008). In hippocampal dentate granule cells, astrocyte-released glutamate activates NMDARs in the afferent neurons, prolonging excitatory synaptic transmission (Jourdain et al., 2007), and induces AMPAR-mediated spontaneous excitatory synaptic current in CA1 pyramidal neurons (Fiacco and McCarthy, 2004). Additionally, astrocytic glutamate can activate pre-synaptic mGluRs, which enhances NMDAR-mediated currents in CA1 hippocampal neurons, but inhibits CA3 hippocampal neurons (Grishin et al., 2004). In addition to glutamate, NMDAR activation also requires D-serine binding. Astrocyte-released D-serine in CA1 induces NMDAR-dependent LTP, controlling the synaptic plasticity of neighboring excitatory synapses (Henneberger et al., 2010). Furthermore, ATP and its degradation product adenosine interact with A₁ and A_2A pre-synaptic receptors, either enhancing or inhibiting neuronal excitability. In hippocampal CA1, astrocyte activation via mGluR5 leads to ATP release, activating A_2A receptors and increasing synaptic transmission and LTP (Panatier et al., 2011). On the other hand, in hippocampal slices, GABA_B receptor-mediated astrocyte activation induces both glutamate and ATP release. An initial excitatory response is driven by glutamate, followed by the ATP response which induces synaptic depression (Covelo and Araque, 2018). These studies provide evidence that gliotransmitters may facilitate or inhibit neuronal excitability depending on the brain (sub)region and type of receptors expressed and activated. This might be further influenced by the temporally segregated co-release of different gliotransmitters by the same, or even different astrocytes, thus emphasizing the intricate nature of astrocyte-neuron communication.

Proper synaptic function and plasticity are vital to drive animal behavior and support complex brain processes. The role of astrocytes in, for example, memory consolidation has been highlighted by observations that in IP₃R2 knock-out mice, in which Ca²⁺ signalling in astrocytes is impaired, synaptic plasticity is compromised (Takata et al., 2011; Chen et al., 2012; Navarrete et al., 2012). Furthermore, Halassa and colleagues showed that blockage of astrocyte vesicular release affects memory formation (Halassa et al., 2009), while Stehberg and colleagues revealed that administration of a mixture of gliotransmitters rescued memory loss (Stehberg et al., 2012). Even though the exact roles astrocytes play in such processes still remain elusive, these studies point towards a path in which specifically manipulating astrocytic activity holds promise to allow in-depth characterization of astrocyte roles, not only in learning and memory but also in other high-order brain functions and behaviors.

Finally, just like during brain development, synapses in the adult brain are also subject to ongoing pruning by astrocytes, using similar molecular machinery to help maintain circuit homeostasis and facilitate processes like learning and memory formation (Lee et al., 2021).

Given the significance of astrocytes in synapse formation and function (including plasticity), there has been a growing interest to manipulate their activity to modulate neuronal activity and positively impact CNS health. Chemogenetics is a widely used technique, based on genetically engineered proteins, which have been specifically modified to respond to otherwise inert synthetic molecules, instead of their endogenous ligands. Since GPCRs comprise the main group of receptors activating astrocytes, chemogenetics commonly uses engineered GPCRs known as Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) (Armbruster et al., 2007).

Target genes, such as DREADDs, can be expressed as transgenes in genetically manipulated mouse lines, by using vectors [such as those based on adeno-associated virus (AAV) or lentivirus (LV)] injected into target brain regions or intravenously (e.g., retro-orbital and tail injections), or even by resorting to in utero (IUE) or post-natal (PALE) electroporation. By combining any of these approaches with astrocyte-specific promoters, expression of the gene of interest can be restricted to this cell population. When choosing between one of these methods to study astrocyte function, one should not only take the goal of the experiment into account but also the advantages and disadvantages inherent to each approach (Table 1). Many transgenic mouse lines, mostly tamoxifen-inducible Cre lines (Cre/ERT2), are particularly suitable to study the impact of astrocytes on brain-wide function. However, high rates of efficiency and specificity are sometimes difficult to accomplish since astrocyte gene expression highly depends on several factors such as developmental stage and brain region, and genes regarded as astrocytic markers may also be expressed in other cell types, including neural progenitor cells (Yu et al., 2020a). The Aldh1l1-Cre/ERT2 and the Fgfr3-Cre/ERT2 mouse lines are, to date, some of the mouse lines which have achieved the highest rates of efficiency and specificity (Young et al., 2010; Srinivasan et al., 2016; Winchenbach et al., 2016; Yu et al., 2021), while others, such as the Slc1a3-Cre/ERT2 (Slezak et al., 2007; Srinivasan et al., 2016), S100β-Cre (Tanaka et al., 2008), and GFAP-Cre/ERT2 (Casper et al., 2007; Park et al., 2018), have been described to target fewer astrocytes and to have more off-target effects than the Aldh1l1-Cre/ERT2 mouse line. Importantly, by crossing GFAP-Cre/ERT2 mice with Cre-responsive Rosa-CAG-lox-hM3Dq, an inducible DREADD mouse line was successfully created in which the DREADD construct (hM3Dq) was specifically expressed in the soma and processes of Gfap-positive glial cells (Sciolino et al., 2016). Intravenous injection of AAV-PHP.eB, containing the GfaABC ₁ D promoter, consistently and specifically targets high amounts of astrocytes across the entire brain, thus providing a valuable alternative to the use of transgenic mouse lines (Chan et al., 2017; Challis et al., 2019). While is it true that brain-wide specific astrocyte targeting can provide valuable information about the contribution of these cells to global brain function, the emerging evidence that astrocytes show inter- and intra-regional heterogeneity, and appear to be matched to local circuits responsible for generating particular behaviors, makes it increasingly important to tackle the role of astrocyte (subpopulations) in distinct brain areas (Nagai et al., 2021). Given the limited diffusion capacity of certain AAV and LV vectors (Scheyltjens et al., 2015), region-specific astrocyte targeting is mostly accomplished by performing intracranial viral vector injections. Namely, AAV2/5 and AAV2/9, containing the GfaABC ₁ D promoter, are commonly used as this has been described to achieve relatively high rates of efficiency (Nagai et al., 2019: 84%; Yu et al., 2018: 89%) and specificity (Vagner et al., 2016; Octeau et al., 2018; Yu et al., 2018; Nagai et al., 2019; Yu et al., 2021). Similarly, LV, such as the Mokola pseudotyped LV (Droguerre et al., 2019), can also be used to locally target astrocytes with the advantage that they can transport more genetic material than AAVs (Cannon et al., 2011; Droguerre et al., 2019). However, LVs have been associated with the risk of mutagenesis since they integrate into the genome. IUE provides yet another alternative to express genes of interest in brain cells and has been successfully implemented in several animal models such as mice, rats, ferrets and cats (Yamashiro et al., 2022; Edwards-Faret et al., 2023). As astrocytes start developing in the CNS shortly after birth, PALE, an adaptation of IUE, can be used to specifically target astrocytes (García-Marqués and López-Mascaraque, 2013; Gee et al., 2015; Stogsdill et al., 2017; Kittock and Pilaz, 2023). In this technique, plasmids, upon injection into the parenchyma of newborn mouse pups (P0-P1), are delivered into the cells due to the action of electrical impulses which disrupt the cell membrane, allowing the passage of the DNA. Besides being less damaging, since the manipulation is performed early in life, and allowing the transfection of bigger constructs than the viral vector approaches, by precisely controlling the injection site and/or the position of the electrodes, both IUE and PALE allow region specific transgene expression (Yamashiro et al., 2022; Kittock and Pilaz, 2023). IUE and PALE have been used to express DREADDs in neurons (Hurni et al., 2017; Muthusamy et al., 2017). Although this has yet to be accomplished for astrocytes, the positive results obtained for neurons hold great promise to also take advantage of this technique to target DREADDs to astrocytes when using the appropriate promoters. It is interesting to note that the above-mentioned approaches can also be combined for better results regarding specificity and expression levels. For instance, astrocyte-specific Cre lines, such as Aldh1l1-CreERT2, generally make use of full-length promoters and are, therefore, generally considered to more faithfully recapitulate the endogenous gene expression profile. Combining the use of such Cre-lines with injections of viral vectors (or plasmids) that express DREADDs in a Cre-inducible manner, allows transgene expression under the control of strong ubiquitous promoters, such as the cytomegalovirus (CMV) promoter, which often results in higher expression levels (although potentially enhanced toxicity cannot be discounted). Additionally, combining IUE/PALE with the use of DREADDs could hold the potential to unravel the roles of astrocytes in non-model organisms, as well as species with a more complex brain structure than mice. However, none of the approaches mentioned above take the matter of intra-regional diversity into account for which vectors and/or transgenic lines containing promoters for specific astrocyte subpopulations should be used. Even though the existence of such vectors and transgenic lines is not yet a reality, the recent generation of single-cell/nucleus RNA sequencing data sets might help propel the development of such tools and even the use of intersectional genetics (Beckervordersandforth et al., 2010). This will be crucial to determine the role of specific subpopulations in defined brain regions (Huang et al., 2020). As described below, to date, DREADDs have been successfully expressed in astrocytes, under the control of the GFAP promoter, in both transgenic mouse models and AAV systems, and it is likely that such techniques can be modified to use promoters targeting astrocyte subpopulations (Bang et al., 2016; Shen et al., 2021).

The modified human M3 and M4 muscarinic receptors coupled to Gq or Gi proteins (hM3Dq and hM4Di, respectively) are the most frequently used DREADDs. Owing to two-point mutations in the ligand binding domain, both receptors can be easily activated upon administration of synthetic compounds of which the most widely used is clozapine-N-oxide (CNO) (Armbruster et al., 2007; Wess et al., 2013). Pioneering studies from Fiacco and colleagues, using transgenic mice expressing the GPCR A1 (MrgA1) in astrocytes, showed that stimulation of MrgA1 elicited Ca²⁺ waves in astrocytes (Fiacco et al., 2007). However, this did not seem to influence neuronal excitability and synaptic plasticity. This lack of effect on neuronal function was met with disappointment by the field and MrgA1-based studies were generally discontinued in favor of those using DREADDs expressed in astrocytes, which were shown to modulate neuronal activity and hence impact animal physiology and behavior (Bang et al., 2016; Shen et al., 2021).

In GFAP-hM3Dq transgenic mice (Agulhon et al., 2013) and in mice injected in the visual cortex with AAV-GFAP-hM3Dq (Bonder and McCarthy, 2014), hM3Dq stimulation with CNO increases intracellular Ca²⁺ levels in astrocytes. The same response has also been described in hM3Dq-activated astrocytes in the hippocampus (Chai et al., 2017; Adamsky et al., 2018; Durkee et al., 2019), striatum (Chai et al., 2017), and nucleus accumbens core (Bull et al., 2014; Corkrum et al., 2020). In contrast, the exact signalling outcome of hM4Di-mediated astrocyte manipulation seems to be less evident as some studies have reported increased astrocytic Ca²⁺ concentrations (Chai et al., 2017; Durkee et al., 2019; Nagai et al., 2019), while others have shown a decrease (Yang et al., 2015; Xin et al., 2019), or even no difference (Nam et al., 2019). These distinct effects have been described between different brain regions (Chai et al., 2017) but also within the same region, like the hippocampus (Chai et al., 2017; Durkee et al., 2019; Kol et al., 2020). Astrocytes are known to be a molecularly and functionally heterogeneous cell population both between and within brain regions (Ben Haim and Rowitch, 2017; Khakh and Deneen, 2019; Pestana et al., 2020; Holt, 2023). Hence, these different responses could be potentially linked to the activation of molecularly distinct astrocyte subtypes. Indeed, work by Chai and colleagues revealed that hM3Dq expression and activation in both hippocampal and striatal astrocytes produces roughly equivalent increases in intracellular Ca²⁺, while the effects of hM4Di on Ca²⁺ were significantly higher in striatal astrocytes, again pointing to specific differences in intracellular signalling pathways (Chai et al., 2017).

Similar to the effects generated by the activation of endogenous GPCRs, several studies have shown that DREADD-induced Ca²⁺ increases result in gliotransmitter release from astrocytes, with subsequent effects on synaptic plasticity and function and, consequently, in behavior and processes such as learning and memory formation. hM3Dq-activated astrocytes in the rat nucleus accumbens core were reported to release glutamate, which was suggested to impact synaptic plasticity (Bull et al., 2014; Scofield et al., 2015). Other groups have reported the release of ATP from hM3Dq-activated astrocytes in the medial basal hypothalamus (Yang et al., 2015) and the nucleus accumbens core (Corkrum et al., 2020). Kang et al. showed that astrocytes in the dorsomedial striatum, activated using hM3Dq, release ATP which, once metabolized into adenosine, induces neuronal activity and, consequently, a shift from habitual to goal-directed seeking behaviors (Kang et al., 2020). Release of ATP from hM3Dq-activated astrocytes in the amygdala was shown to activate A_1A receptors, inhibiting neuronal activity and reducing fear behavior (Martin-Fernandez et al., 2017). Moreover, in astrocytes from the hippocampal CA1, hM3Dq-mediated activation induced astrocyte secretion of D-serine, which enhanced synaptic plasticity and memory formation (Adamsky et al., 2018). In an independent study, Durkee and colleagues found that stimulation of astrocytes using hM3Dq in the same brain region resulted in glutamate release, increasing neuronal excitability via NMDAR activation (Durkee et al., 2019). The effects of hM4Di activation on gliotransmitter release have been less explored. Evidence suggests that hippocampal astrocyte activation via hM4Di leads to glutamate release and a consequent increase in neuronal activity (Durkee et al., 2019; Nam et al., 2019). Overall, these studies highlight the role of astrocytes in complex animal behavior and function, while showcasing the incredible potential of this technique to further expand our knowledge of astrocyte-neuron interplay during processes such as memory formation.

Overall, these studies confirm that using Gi- and Gq-DREADDs to manipulate astrocytes seems to recapitulate, at least to a degree, their responses to endogenous neurotransmitters, resulting in neuronal modulation and behavioral effects. Thus, their use presents a potential entry point for uncovering new molecular pathways to manage CNS function. However, despite their incredible potential, it is important to acknowledge that using DREADDs also comes with its challenges and limitations. For example, it is unlikely that DREADD activation faithfully recapitulates all aspects of the highly complex Ca²⁺ signalling elicited under standard physiological conditions (Semyanov et al., 2020). In addition, when using viral vector-based delivery systems, the genomic titer of the vectors, multiplicity of infection in cells and relative promoter strength may limit DREADD expression levels. These factors could account, at least in part, for the distinct effects of DREADD activation observed in different studies. Furthermore, it has been reported that systemically administered CNO does not easily penetrate the blood-brain barrier and back-metabolizes into clozapine (Jann et al., 1994; Gomez et al., 2017). Since clozapine itself is a muscarinic agonist, it can activate endogenous receptors, potentially leading to off-target effects, as it has been observed in rats, mice and humans (Jann et al., 1994; MacLaren et al., 2016; Gomez et al., 2017; Bærentzen et al., 2019). Considering these issues, an effort has been made to develop new chemical compounds, such as Compound 21 (C21) and perlapine (Chen et al., 2015; Thompson et al., 2018), JHU37152 and JHU37160 (Bonaventura et al., 2019), and deschloroclozapine (Nagai et al., 2020; Nentwig et al., 2022). These molecules showcase high affinity and selectivity for DREADDs and, so far, minimal off-target effects have been identified. However, these novel compounds are still poorly characterized, and therefore, despite its obvious limitations, CNO remains the most commonly used compound for DREADD activation.

Neuronal circuit activity is easily disrupted as a consequence of synaptic dysfunction, a common hallmark of several neurological disorders, ranging from neurodevelopmental to neurodegenerative. Astrocyte dysfunction has also been reported in many CNS disorders over the years and is thought to contribute to disease mechanisms. Therefore, manipulating astrocytic activity and function at an early stage could hold potential for the development of novel approaches to decrease the severity and progression of the disease (Figure 2). It is interesting to note that categories of disorders present phenotypic commonalities. While the strategy to manage, for example, neurodevelopmental diseases could mainly focus on promoting astrocyte-mediated synaptogenesis, for other disorders one could prioritize re-establishing homeostatic astrocyte-neuron signalling and controlling chronic inflammation.

Altogether, these studies highlight the central role of astrocyte dysfunction, and consequent synaptic dysfunction, in several CNS pathologies and thus the essential contribution of astrocytes to normal circuit development and function. Given the importance of GPCR signalling for astrocyte activity, we propose that expressing genetically engineered GPCRs in astrocytes could be a promising strategy to identify relevant signalling pathways that could ameliorate brain dysfunction occurring after injury or throughout disease. The potential application of DREADDs across the spectrum of conditions lies in the range of downstream effects triggered upon astrocyte stimulation, which most notably includes the release of distinct neuroactive molecules that differentially modulate synaptogenesis and neuronal activity. The positive impact of astrocyte manipulation using DREADDs on synaptic transmission, cognitive function and behavior, mainly by eliciting astrocytic Ca²⁺ waves and increasing the release of TSP1, ATP, D-serine or glutamate, has been demonstrated by several groups. Even though this holds potential to manage brain disease, direct evidence of DREADD-mediated astrocyte activation leading to the improvement of disease pathology is still scarce.

Most CNS pathologies share common features like loss of spine density, impaired glutamate clearance and synaptic transmission, and neuroinflammation. Still, the causative factor(s) for these diseases differ widely. Genetic mutations, the cellular and molecular environment characteristic of each disease, as well as the brain regions and specific neuronal circuits affected, differ between diseases. Therefore, manipulating astrocyte activity in these different contexts may very well lead to different functional consequences, which could either improve or worsen the disease state. Thus, future research should focus on assessing the effects of astrocyte activity modulation in each specific disease, and probably also needs to acknowledge the issue of disease time-course (see below). Furthermore, astrocyte heterogeneity is likely a contributing factor in determining the specific outcome of DREADD manipulation. The complexity of this topic increases even more when considering that new astrocyte subtypes characteristic for each particular disease, and even disease stage-specific astrocytes, usually arise. Identifying and differentiating the molecular profile of such disease-associated astrocytes may allow the design of astrocyte subtype-specific promoters to target DREADD vectors to maleficent astrocyte subpopulations. We also propose that conducting sequencing studies on DREADD-activated astrocytes could be highly relevant to fully understand the effects of this type of stimulation on the molecular fingerprint of the cells. This could bring insights into whether astrocytes respond by boosting endogenous signalling pathways, or by triggering new ones, generating insights into potential signalling pathways that can be exploited therapeutically. Another key aspect to take into consideration is that mouse and human astrocytes are morphologically, transcriptionally, and functionally different (Oberheim et al., 2006; Li et al., 2021). Using, for example, human cerebral organoids (Dezonne et al., 2017) would be a valuable strategy to test how human astrocytes are affected by DREADDs and help in the transition from the bench to the clinic.

It is interesting to note that very few studies so far have demonstrated inhibition of astrocytic Ca²⁺ waves following DREADD activation. Most studies also report an increase in TSP1 or gliotransmitter release, particularly glutamate, upon DREADD-mediated astrocyte stimulation. However, some pathologies display increased TSP1 release (Krencik et al., 2015), enhanced Ca²⁺ signalling in astrocytes or glutamate accumulation at the synapse, thus suggesting that use of the existent DREADD receptors might not be a good strategy to unravel possible ways to manage such diseases, due to undesirable, or even toxic, effects. For instance, excessive TSP1 release might induce the overproduction of new synapses, which may further disrupt neuronal circuits, while excessive glutamate release can lead to excitotoxicity. It is also important to point out that Ca²⁺ is likely not the only second messenger influenced by DREADDs. To what extent DREADDs are also acting on alternative, Ca²⁺-independent signalling pathways affecting, for example, cAMP levels remains unclear. Exploring the impact of DREADD-evoked astrocyte activation on other signalling pathways, all of which may impact synapse formation and function, most likely depends on the development of new and more sensitive tools. This will hopefully create a comprehensive understanding of the biological effects of DREADD activation on cells, leading to a deeper mechanistic understanding of cell function and insights into disease, which will ultimately allow the development of ‘next-generation’ therapeutics.