Astrocytic IP3Rs: Beyond IP3R2

Astrocytes are sensitive to ongoing neuronal/network activities and, accordingly, regulate neuronal functions (synaptic transmission, synaptic plasticity, behavior, etc.) by the context-dependent release of several gliotransmitters (e.g., glutamate, glycine, D-serine, ATP). To sense diverse input, astrocytes express a plethora of G-protein coupled receptors, which couple, via Gi/o and Gq, to the intracellular Ca2+ release channel IP3-receptor (IP3R). Indeed, manipulating astrocytic IP3R-Ca2+ signaling is highly consequential at the network and behavioral level: Depleting IP3R subtype 2 (IP3R2) results in reduced GPCR-Ca2+ signaling and impaired synaptic plasticity; enhancing IP3R-Ca2+ signaling affects cognitive functions such as learning and memory, sleep, and mood. However, as a result of discrepancies in the literature, the role of GPCR-IP3R-Ca2+ signaling, especially under physiological conditions, remains inconclusive. One primary reason for this could be that IP3R2 has been used to represent all astrocytic IP3Rs, including IP3R1 and IP3R3. Indeed, IP3R1 and IP3R3 are unique Ca2+ channels in their own right; they have unique biophysical properties, often display distinct distribution, and are differentially regulated. As a result, they mediate different physiological roles to IP3R2. Thus, these additional channels promise to enrich the diversity of spatiotemporal Ca2+ dynamics and provide unique opportunities for integrating neuronal input and modulating astrocyte–neuron communication. The current review weighs evidence supporting the existence of multiple astrocytic-IP3R isoforms, summarizes distinct sub-type specific properties that shape spatiotemporal Ca2+ dynamics. We also discuss existing experimental tools and future refinements to better recapitulate the endogenous activities of each IP3R isoform.


INTRODUCTION
Over the last three decades, Ca 2+ imaging has revealed new roles for astrocytes. Indeed, astrocytic Ca 2+ signaling was shown to regulate synaptic transmission, synaptic plasticity, and to influence behavior Park and Lee, 2020). Inositol 1,4,5-trisphosphate receptors (IP 3 Rs) mediated Ca 2+ signaling (IP 3 R-Ca 2+ ) is regarded a primary generator of astrocytic Ca 2+ signaling. Upon activation of G q -GPCRs, the main input pathway of astrocytes, phospholipase C breaks down PIP 2 into DAG and IP 3 , activating IP 3 R predominantly located on the membrane of endoplasmic reticulum (ER) resulting in Ca 2+ release (Bootman et al., 2001). This IP 3 R-mediated Ca 2+ signaling is considered to trigger the activity-dependent and selective release of chemical transmitters (gliotransmitters) such as glutamate, D-serine, and ATP, which have distinct influences over neuronal activity. Initially, IP 3 R subtype 2 (IP 3 R2) was the only recognized Ca 2+ channel in astrocytes; However, advanced Ca 2+ imaging techniques have since identified novel Ca 2+ sources, including mitochondria (Agarwal et al., 2017), transient receptor potential ankyrin 1 (Shigetomi et al., 2012(Shigetomi et al., , 2013b, L-type voltage gated Ca 2+ channels (Letellier et al., 2016), sodium/calcium exchanger (Kirischuk et al., 1997;Boddum et al., 2016;Rose et al., 2020), and transient receptor potential canonical (Shiratori-Hayashi et al., 2020) amongst others, thereby expanding the known Ca 2+ signaling toolkit of astrocytes. Doubtlessly additional Ca 2+ channels and sources will emerge in the future.
While the field's focus has moved on from understanding IP 3 Rs to identifying new Ca 2+ sources, understanding IP 3 R signaling in astrocytes remains highly relevant. Indeed, IP 3 Rs are the primary target for manipulating astrocytic activity, and such manipulations have proven to be very consequential in many studies. Since most of these manipulations indiscriminately influence all IP 3 R subtypes, this could reflect the key role played by IP 3 R subtypes other than IP 3 R2, namely IP 3 R1 and IP 3 R3, which were mostly overlooked. In this review, we summarize the evidence for different subtypes of IP 3 R and discuss how we can better study the role of IP 3 R-Ca 2+ signaling in astrocytes which is one of the core issues in understanding astrocyte physiology.

Astrocyte Proteome
Several studies explored IP 3 Rs using immunohistochemistry which provided a consensus over the expression of IP 3 R2 in hippocampal/cortical astrocytes and Bergmann glia (Sharp et al., 1999;Holtzclaw et al., 2002;Hertle and Yeckel, 2007;Takata et al., 2011;Chen et al., 2012). While there are some conflicting reports over the immunoreactivity of IP 3 R3 in astrocytes and Bergmann glia Yamamoto-Hino et al., 1995;Hamada et al., 1999;Sharp et al., 1999;Holtzclaw et al., 2002), IP 3 R1 immunoreactivity was not initially observed in glia (Nakanishi et al., 1991;Dent et al., 1996;Hamada et al., 1999;Sharp et al., 1999;Holtzclaw et al., 2002;Hertle and Yeckel, 2007). These findings supported the view that IP 3 R2 is the predominant astrocytic IP 3 R. However, these results may also reflect limitations of the available IP 3 R antibodies or the difficulty of accurately assigning proteins located within ultrathin astrocyte processes, which are below the resolution limit of conventional microscopy (Panatier et al., 2014;Arizono et al., 2020) and buried amongst neuronal dendrites.
IP 3 R1 immunoreactivity was recently detected in spinal dorsal horn astrocytes (Shiratori-Hayashi et al., 2020) and, albeit with low stringency, in isolated astrocytes from adult mice (Chai et al., 2017). Using a state of the art TurboID construct to biotinylate proteins in the immediate proximity of tripartite synapses, Takano et al. (2020) report enrichment of IP 3 R1 protein in the peri-synaptic astrocytic compartment (Takano et al., 2020). This finding, however, should be interpreted with some caution as identified proteins were assigned to astrocytes based on published mRNA datasets (Zhang et al., 2014(Zhang et al., , 2016. Nevertheless, IP 3 R1 protein enrichment in fine astrocytic processes is consistent with Ca 2+ imaging studies (Sherwood et al., 2017) and could account for the poor detection in various assays which favor detection in large subcellular compartments, i.e., major processes and soma of astrocytes. Notably, IP 3 R2, which is reported to be in the soma and main branches (Chen et al., 2012), was not enriched in peri-synaptic astrocytic compartments (Takano et al., 2020), likely reflecting the different subcellular distribution of IP 3 R1 and IP 3 R2 ( Figure 1A and Table 1).

Astrocyte Transcriptome
IP 3 R1, IP 3 R2, and IP 3 R3 are encoded by the respective genes ITPR1, ITPR2, and ITPR3. Notably, mRNA for all three genes have been detected in astrocytes isolated from young and aged mouse brain (Cahoy et al., 2008;Zhang et al., 2014;Chai et al., 2017;Clarke et al., 2018) as well as humans (Zhang et al., 2016). Furthermore, ITPR1 and ITPR2 are actively translated in astrocytes of adult mice (Srinivasan et al., 2016;Chai et al., 2017;Clarke et al., 2018;Yu et al., 2018) indicating that their proteins are produced in astrocytes. In astrocytes the isoform transcript abundance is generally ITPR2 > ITPR1 >>> ITPR3 (ITPR3 mRNA is present in very small amounts and may be negligible). However, these transcripts are developmentally and differentially regulated across brain regions and in some instances ITPR1 mRNA appears to be more abundant than ITPR2 mRNA (Yu et al., 2018; Table 1).
2-APB was introduced as an antagonist of IP 3 Rs (Maruyama et al., 1997) and has been widely used to investigate the contribution of IP 3 Rs to cellular Ca 2+ signaling. 2-APB appears to preferentially block IP 3 R1 and IP 3 R3, whereas cells predominantly expressing IP 3 R2 seem largely insensitive (Kukkonen et al., 2001;Bootman et al., 2002;Saleem et al., 2014). Thus, the fact that 2-APB reduces astrocyte Ca 2+ amplitude and responsiveness in diverse brain regions (Sul et al., 2004;Young et al., 2010;Tamura et al., 2014;Tang et al., 2015;Arizono et al., 2020) and spinal dorsal horn (Shiratori-Hayashi et al., 2020) may support the presence of functional IP 3 R1 and IP 3 R3 in astrocytes. The functional role of non-IP 3 R2 in astrocytic Ca 2+ signaling remains to be determined, it is possible that some membrane receptors are functionally coupled solely to IP 3 R1 or IP 3 R3 and generate local Ca 2+ events, which may or may not be involved in triggering IP 3 R2 dependent Ca 2+ waves (Petravicz et al., 2008).

Phenotypic Comparison
Comparison of WT and total IP 3 R2KO mice has revealed some important physiological roles of IP 3 R2 signaling in astrocytes, i.e., motor learning (Padmashri et al., 2015) modulating depressivelike behaviors (Cao et al., 2013) but this remains controversial (Petravicz et al., 2014). The possible roles of astrocytic IP 3 R1 and IP 3 R3 may be gleaned by comparing WT and IP 3 R2KOs with manipulations that impair all IP 3 R subtypes. For example, while IP 3 R2KO has no impact on sleep (Cao et al., 2013) overexpressing a transgene for the IP 3 hydrolyzing enzyme, IP 3 -5-phosphatase, in astrocytes, suppresses Ca 2+ release from all IP 3 R subtypes and disrupts sleep (Foley et al., 2017). Similarly, although IP 3 R2KO had no impact on hippocampal LTP (Agulhon et al., 2010), we showed that loading a single astrocyte with the membraneimpermeable pan-IP 3 R antagonist, heparin, blocks long-term synaptic potentiation (Sherwood et al., 2017). These studies, taken together, indicate that non-IP 3 R2 IP 3 Rs regulate sleep and hippocampal LTP.

Summary
The expression of multiple IP 3 R subtypes in astrocytes has a wide-reaching implication in the field. Studies using IP 3 R2KO as a model for blocked IICR would need to be re-assessed to include the possibility of IICR mediated by IP 3 R1 and IP 3 R3.
The difference in IP 3 R subtype properties is further characterized by various binding partners, including kinase and phosphatases, which can further fine-tune Ca 2+ profiles. Interestingly, while there are many interacting partners common to all three IP 3 R subtypes, the nature of their regulation can be subtype-specific. For instance, protein kinase C, depending on the IP 3 R subtype, can either be stimulatory or inhibitory; this difference likely reflects isoform-specific phosphorylation sites. Further detailed biochemical study (Hamada et al., 2017) is required to investigate how various IP 3 R binding molecules specifically regulate each IP 3 R isoform (Hamada and Mikoshiba, 2020).

Different Distribution and Role of IP 3 R Subtypes Within Various Tissues
One important feature that defines the subtype-specific role of IP 3 Rs in vivo is the tissue distribution patterns. While IP 3 R1 is mostly expressed in the central nervous system, IP 3 R2 and IP 3 R3 are broadly expressed in various organs such as the heart, pancreas, liver, and salivary glands (Hisatsune and Mikoshiba, 2017). This distribution pattern is tightly linked to the physiological role of IP 3 R subtypes. Reflecting its rich expression in Purkinje cells, mice lacking IP 3 R1 exhibit severe cerebellar ataxia, a seizure−like posture, impaired cerebellar LTD. (Inoue et al., 1998), and die within 3-4 weeks of birth (Matsumoto et al., 1996). The dysregulation of IP 3 R1 is linked with other brain disorders such as Huntington's disease and Alzheimer's disease (Hisatsune and Mikoshiba, 2017). IP 3 R2 is associated with sweating (Klar et al., 2014), bone formation (Kuroda et al., 2008), and heart hypertrophy (Nakayama et al., 2010;Drawnel et al., 2012;Vervloessem et al., 2015). IP 3 R3 plays a role in taste sensing (Hisatsune et al., 2007) and hair cycle (Sato-Miyaoka et al., 2012). IP 3 R2 and IP 3 R3 together are associated with the heart's development (Uchida et al., 2010(Uchida et al., , 2016 and secretion of saliva and tears (Futatsugi, 2005;Inaba et al., 2014).

Different Distribution and Role of IP 3 R Subtypes Within a Cell
Some cells express multiple IP 3 R subtypes, enabling each subtype to uniquely contribute to Ca 2+ profiles and cellular functions. For instance, in HeLa cells, knock-down of IP 3 R1 terminates Ca 2+ oscillations, whereas knock-down of IP 3 R3 results in more robust and long-lasting Ca 2+ oscillations (Hattori et al., 2004). The specific contribution of IP 3 R subtypes can also depend on their distinct subcellular distribution, as seen in pancreatic acinar cells , COS cells (Pantazaka and Taylor, 2011), and DT40 cells (Bartok et al., 2019). In Bergmann glia, knocking out IP 3 R2 resulted in decreased agonist-induced Ca 2+ release while knocking out IP 3 R1 and IP 3 R3 resulted in delaying the peak of agonist-induced Ca 2+ release specifically in astrocytic processes (Tamamushi et al., 2012), suggesting their subtype-specific distribution. It is likely that astrocytes, which express multiple IP 3 R subtypes, also take advantage of subtypespecific distribution.

Summary
While IP 3 R subtypes are regulated by IP 3 and Ca 2+ and have many common interacting partners, they differ in how they are affected by these regulators. Such differences enrich the diversity of spatio-temporal Ca 2+ profiles created by IP 3 Rs. The IP 3 R subtype expression pattern, in vivo, is tissue specific and their subcellular localization is highly variable and dependent on cell types, and this carries important functional implications. Together with recent reports showing the distinct role of IP 3 R1 and IP 3 R3 in Bergmann glia and astrocytes, these facts support the view that IP 3 R isoforms 1 and 3 are unique Ca 2+ channels that need to be addressed independently of IP 3 R2.

TOOLS TO DISSECT THE ROLE OF IP 3 R ISOFORMS Experimental and Analytical Tools
To understand the role of the various Ca 2+ signals in astrocyte physiology, it will be necessary to make quantitative measurements (Neher, 2008). Progress in this direction has been frustrated by the unique astrocyte morphology and difficulties in interpreting recorded Ca 2+ -dependent fluorescent signals (Rusakov, 2015). To accurately capture Ca 2+ dynamics in sub-cellular compartments, there is a need to adopt imaging techniques with improved resolution and to develop tools for efficient analysis in three-dimensional (Bindocci et al., 2017;Romanos et al., 2019). Because of our poor understanding of functional compartments, analysis of astrocytic Ca 2+ dynamics has been moving toward state-of-the-art event-based analysis (Romanos et al., 2019;Wang et al., 2019;Bjørnstad et al., 2021), nevertheless ROI (regionof-interest) based analysis, informed by cellular anatomy (functional compartments) and molecular architecture, has been critical for understanding neuron physiology (e.g., spines and boutons). To this end, the identification of morphologically distinct subcellular compartments are promising targets for classical ROI based analysis, i.e., "glial microdomains" on Bergmann glial processes (Grosche et al., 1999) and "astrocytic compartments" on major branches (Panatier et al., 2011), both revealed using confocal microscopy, and astrocytic nodes and shafts within the spongiform structure visualized using live STED microscopy (Arizono et al., 2020). The ultimate goal of extracting quantitative Ca 2+ dynamics from fluorescent data is non-trivial but achievable using realistic biophysical cell models (Rusakov, 2015;Denizot et al., 2019), a task simplified by the recent development of the open-source flexible model builder ASTRO (Savtchenko et al., 2018). For an accurate understanding of Ca 2+ dynamics, it will be necessary to constrain models further using empirically determined details, e.g., receptor kinetics, expression patterns, endogenous Ca 2+ buffering, etc.

Pharmacological Tools
It is difficult to disentangle the physiological roles of IP 3 R subtypes in cells that typically express complex mixtures of homo-and hetero-tetrameric IP 3 Rs. There are no ligands that usefully distinguish among IP 3 R subtypes (Saleem et al., 2013a,b) and nor are there effective antagonists that lack serious side effects (Michelangeli et al., 1995). Of the available antagonists, heparin is currently the most useful. Heparin is a membrane impermeant pan-IP 3 R inhibitor that may be selectively loaded into astrocytes using a whole-cell patch-pipette (Sherwood et al., 2017). Recent developments report small impermeant competitive antagonists of IP 3 R1, which, compared to heparin, are likely to have fewer off-targets (Konieczny et al., 2016). Well-characterized function-blocking monoclonal antibodies are powerful tools to specifically inhibit IP 3 R subtypes (Miyazaki et al., 1992;Inoue et al., 1998;Nishiyama et al., 2000;Gerasimenko et al., 2009). This technology has not yet been applied to astrocytes.

Activation of IP 3 Induced Ca 2+ Release in Astrocytes Pharmacogenetics
DREADDs (designer receptor exclusively activated by designer drug) enable the selective activation of GPCR-IP 3 R-Ca 2+ signaling in astrocytes. The most used DREADDs are the excitatory Gq or inhibitory G i -coupled receptors, hM3Dq and hM4Di, respectively (derived from human M3/M4 muscarinic receptor). Both receptors are activated by a pharmacologically inert but bioavailable ligand clozapine-N-oxide (CNO) while being non-responsive to endogenous GPCR ligands (Agulhon et al., 2013). hM3Dq has been used to demonstrate an astrocytic role in behaviors such as food intake (Yang et al., 2015), fear response , and memory recall (Adamsky et al., 2018). While the DREADDs enables selective activation of astrocytes, they have two major drawbacks: Firstly, the exogenous receptors have not been targeted to specific signaling domains and are likely to be spatially uncoupled from signaling nanodomains critical to IP 3 R physiology (Bootman et al., 2001); Secondly, because of the sustained (hour-long) activation by exogenous ligands (Iwai et al., 2021), temporal features of astrocyte signaling are lost. While perhaps mimicking global Ca 2+ surges, the available DREADDs are unlikely to recapitulate many of the local Ca 2+ transients typically observed within fine astrocytic processes (Shigetomi et al., 2013a;Arizono et al., 2020).

Summary -Future Developments
While having great potential for controlling astrocytic activation, a central question is to what extent do the chimeras mimic the signaling of wild-type receptors. GPCRs can have multiple signaling axis, e.g., multiple G-protein axes, β-arrestins, wntfrizzled, or the hedgehog-smoothened axes (Bailes and Lucas, 2013;Tichy et al., 2019), and the signaling axes bias is often not characterized but can have profound side effects on physiology (Agulhon et al., 2013;Tichy et al., 2019). Indeed, the functional outcome of activating G q in astrocytes using different exogenous receptor (i.e., hM3Dq and MrgA1) is not reproducible (Agulhon et al., 2010;Adamsky et al., 2018). While multiple signaling axis could confuse the role of IICR, they may be required to obtain an optimal IP 3 R response (Konieczny et al., 2017). Another primary concern is that DREADDs and optoXRs likely activate IP 3 Rs from cellular compartments distinct from those of the endogenous receptors, limiting their ability to recreate physiologically relevant Ca 2+ profiles. To address this, nextgeneration activation tools are being engineered to mimic the subcellular distribution of endogenous receptors (Oh et al., 2010;Masseck et al., 2014;Spoida et al., 2014;Tichy et al., 2019).
In the last decade, substantial progress has been made revealing diverse spatio-temporal Ca 2+ signaling in astrocytes. Understanding the subtleties of these signals will require detailed knowledge of the astrocytic Ca 2+ signaling toolbox along with the generation and characterization of more sophisticated tools to control and accurately recapitulate the physiologically relevant Ca 2+ signals.