Abstract
Accumulating evidence indicates that astrocytes are actively involved in the physiological and pathophysiological functions of the brain. Intracellular Ca2+ signaling, especially Ca2+ release from the endoplasmic reticulum (ER), is considered to be crucial for the regulation of astrocytic functions. Mice with genetic deletion of IP3R type 2 (IP3R2) are reportedly devoid of astrocytic Ca2+ signaling, and thus widely used to explore the roles of Ca2+ signaling in astrocytic functions. While functional deficits in IP3R2-knockout (KO) mice have been found in some reports, no functional deficit was observed in others. Thus, there remains a controversy regarding the functional significance of astrocytic Ca2+ signaling. To address this controversy, we re-evaluated the assumption that Ca2+ release from the ER is abolished in IP3R2-KO astrocytes using a highly sensitive imaging technique. We expressed the ER luminal Ca2+ indicator G-CEPIA1er in cortical and hippocampal astrocytes to directly visualize spontaneous and stimulus-induced Ca2+ release from the ER. We found attenuated but significant Ca2+ release in response to application of norepinephrine to IP3R2-KO astrocytes. This IP3R2-independent Ca2+ release induced only minimal cytosolic Ca2+ transients but induced significant Ca2+ increases in mitochondria that are frequently in close contact with the ER. These results indicate that ER Ca2+ release is retained and is sufficient to increase the Ca2+ concentration in close proximity to the ER in IP3R2-KO astrocytes.
Introduction
Accumulating evidence indicates that astrocytes play not only trophic and supportive roles but also active roles in the physiological and pathophysiological functions of the brain (Volterra and Meldolesi, 2005; Barres, 2008). Therefore, clarifying the mechanisms for regulation of astrocytic functions has become a crucial issue in neuroscience. Intracellular Ca2+ signaling has attracted attention because astrocytes show robust spontaneous and stimulus-induced Ca2+ transients (Verkhratsky et al., 1998; Bazargani and Attwell, 2016).
Ca2+ release from the endoplasmic reticulum (ER) via inositol 1,4,5-trisphosphate receptor (IP3R) has been recognized as a major component of spontaneous and Gq-coupled receptor-induced Ca2+ signaling in astrocytes. Therefore, to explore the significance of astrocytic Ca2+ signaling, mice with genetic deletion of IP3R type 2 (IP3R2) (Li et al., 2005), which is enriched in astrocytes (Sharp et al., 1999; Holtzclaw et al., 2002; Zhang et al., 2014), have been widely used. In IP3R2-knockout (KO) mice, Ca2+ transients in astrocytes were considered to be absent based on an imaging study of cytosolic Ca2+ transients by fluorescent Ca2+ indicator dyes (Petravicz et al., 2008). Physiological and pathophysiological phenotypes have been observed in IP3R2-KO mice, including plasticity and learning (Takata et al., 2011; Chen et al., 2012; Navarrete et al., 2012; Perez-Alvarez et al., 2014; Padmashri et al., 2015; Kim et al., 2016; Monai et al., 2016; Yang et al., 2016), homeostasis of K+ (Wang et al., 2012a, 2012b), pathogenesis of stroke, traumatic brain injury and neurodegenerative disease (Dong et al., 2013; Kanemaru et al., 2013; Li et al., 2015; Rakers and Petzold, 2016; Saito et al., 2018), and depression-like behaviors (Cao et al., 2013). However, several studies have reported no changes in basal synaptic activity and behavior (Petravicz et al., 2008, 2014; Agulhon et al., 2013), hippocampal synaptic plasticity (Agulhon et al., 2010), and neurovascular coupling (Nizar et al., 2013; Takata et al., 2013; Bonder and McCarthy, 2014) in IP3R2-KO mice. These contradictory results have caused confusion about the functional significance of Ca2+ signaling in astrocytes (Hamilton and Attwell, 2010; Fiacco and McCarthy, 2018; Savtchouk and Volterra, 2018).
One of the possible explanations that can resolve this controversy is that IP3R2-independent Ca2+ signaling exists and contributes to the functions of IP3R2-KO astrocytes. In fact, recent studies using genetically encoded Ca2+ indicators (GECIs) have clearly shown the presence of Ca2+ signals in IP3R2-KO astrocytes (Kanemaru et al., 2014; Srinivasan et al., 2015; Rungta et al., 2016; Agarwal et al., 2017). These IP3R2-independent Ca2+ transients were thought to be mediated by Ca2+ influx via the plasma membrane (Shigetomi et al., 2012; Srinivasan et al., 2015; Rungta et al., 2016) or Ca2+ release from mitochondria (Agarwal et al., 2017). These studies suggest that increases in the concentration of Ca2+ originating from sources other than the ER may regulate the functions of astrocytes. However, it remains elusive whether ER Ca2+ release is completely abolished in IP3R2-KO astrocytes.
We thus examined whether IP3R2-independent Ca2+ release from the ER is present in IP3R2-KO astrocytes. Because cytosolic Ca2+ transients can be generated by Ca2+ derived from sources other than the ER, we directly detected ER Ca2+ release as a decrease in the Ca2+ concentration within the ER ([Ca2+]ER). We recently developed a series of calcium-measuring organelle-entrapped protein indicators (CEPIAs) that allow visualization of Ca2+ dynamics within the ER and mitochondria with a high signal-to-noise ratio (Suzuki et al., 2014, 2016; Okubo et al., 2015). In this study, we expressed the ER Ca2+ indicator G-CEPIA1er in cortical and hippocampal astrocytes of IP3R2-KO mice. We found attenuated but significant Ca2+ release from the ER. We also used the mitochondrial Ca2+ indicator CEPIA2mt and successfully observed significant Ca2+ transfer into mitochondria from the ER in IP3R2-KO astrocytes. These findings allowed us to reinterpret the results from IP3R2-KO mice and shed new light on the significance of astrocytic Ca2+ signaling.
Materials and Methods
Animals
All animal experiments were carried out in accordance with the regulations and guidelines of the Institutional Animal Care and Use Committee at The University of Tokyo and were approved by the Institutional Review Committee of the Graduate School of Medicine, The University of Tokyo. C57BL/6 mice were used as wild-type (WT) mice. IP3R2-KO mice (Li et al., 2005) were obtained from J. Chen (University of California at San Diego). Mice were kept under a 12 h light/dark cycle with ad libitum access to food and drinking water.
Preparation of viral vectors
To generate adeno-associated viruses (AAVs) for astrocyte-specific expression of GECIs, the cytomegalovirus promoter of pAAV-MCS (AAV Helper Free Expression System, Cell Biolabs, Inc., San Diego, CA) was replaced with the gfaABC1D astrocyte-specific promoter derived from pZac2.1-gfaABC1D-Lck-GCaMP3 (Shigetomi et al., 2013). G-CEPIA1er, CEPIA2mt, GCaMP6f, and ER-YFP were inserted downstream of the gfaABC1D promoter to generate pAAV-gfaABC1D-G-CEPIA1er, pAAV-gfaABC1D-CEPIA2mt, pAAV-gfaABC1D-GCaMP6f, and pAAV-gfaABC1D-ER-YFP. AAV5 vectors were packaged using the AAV Helper Free Expression System. The packaging plasmids (pAAV-RC5 and pHelper) and transfer plasmid (pAAV-gfaABC1D-G-CEPIA1er, pAAV-gfaABC1D-CEPIA2mt, pAAV-gfaABC1D-GCaMP6f, or pAAV-gfaABC1D-ER-YFP) were transfected into HEK293T cells using the calcium phosphate method. The medium was replaced at 18 h after transfection with fresh medium, and the cells were incubated for 48 h. Harvested cells were lysed by repeated freezing and thawing, and a crude cell extract containing AAV5 vector particles was obtained. AAV5 vector particles were purified by ultracentrifugation with cesium chloride. The purified particles were dialyzed against PBS and then concentrated by ultrafiltration using an Amicon 10K MWCO filter (Merck Millipore, Darmstadt, Germany). The copy number of the viral genome (vg) was determined by real-time quantitative PCR.
AAV injection
Male C57BL/6 mice or IP3R2-KO mice (postnatal day 56–120) were anesthetized with isoflurane (induction at 5%, maintenance at 1–2%, vol/vol, MK-A100, Muromachi, Kyoto, Japan). The mice were placed in a stereotaxic frame (SR-5M-HT, Narishige, Tokyo, Japan). The skull was thinned (about 1 mm in diameter) above the right parietal cortex using a burr powered by a high speed drill (ULTIMATE XL-D, NSK, Kanuma, Japan). AAV5-gfaABC1D-G-CEPIA1er (0.98 or 1.3 × 1013 vg ml−1), AAV5-gfaABC1D-CEPIA2mt (0.98 × 1013 vg ml−1), AAV5-gfaABC1D-GCaMP6f (1.1 × 1013 vg ml−1), or pAAV-gfaABC1D-ER-YFP (3.0 × 1013 vg ml−1) was unilaterally injected into the cortex (1.5–2 mm posterior to the bregma, 1–1.5 mm lateral to the midline, and 300 µm from the surface) or hippocampus (2 mm posterior to the bregma, 1.5 mm lateral to the midline, and 1.6 mm from the surface) through glass pipettes. A viral solution (1 µL) was delivered at a rate of 200 nL min−1 using a micropump (Legato 130, KD scientific, Holliston, MA). Glass pipettes were left in place for at least 10 min. Mice were sacrificed at 14–28 days after AAV injection for imaging and immunohistochemistry.
Preparation of brain slices and imaging
Coronal cortical or hippocampal slices (300 µm in thickness) were prepared as described previously (Edwards et al., 1989). Slices were prepared in ice-cold artificial cerebrospinal fluid (ACSF) bubbled with 95% O2 and 5% CO2 using a vibrating slicer (PRO7, Dosaka, Kyoto, Japan). Slices were incubated in a holding chamber containing ACSF bubbled with 95% O2 and 5% CO2 at 35°C for 1 h and then returned to 23°C. ACSF contained (in mM) 125 NaCl, 2.5 KCl, 2 CaCl2, 1 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, and 20 glucose. Imaging was carried out with a two-photon microscope (TSC MP5, Leica, Wetzlar, Germany) equipped with a water immersion objective (×25, NA 0.95, HCS IR APO, Leica) and Ti:sapphire laser (MaiTai DeepSee; Spectra Physics, Santa Clara, CA). Slices were transferred to a recording chamber under a microscope and continuously perfused with ACSF bubbled with 95% O2 and 5% CO2. Tetrodotoxin (1 µM) was added to ACSF to inhibit neuronal activities throughout imaging. Norepinephrine (NE, 10 µM) and cyclopiazonic acid (CPA, 50 µM) were dissolved in ACSF and administered through the perfusion system of the recording chamber. The excitation wavelength was 900–920 nm. Emitted fluorescence was filtered by a 500–550 nm barrier filter and detected with photomultiplier tubes. Data were acquired in time lapse XY-scan mode (0.2–0.5 Hz). Experiments were carried out at room temperature (22–24°C).
Immunohistochemistry
After anesthetization with pentobarbital, mice were fixed by perfusion of PBS containing 4% paraformaldehyde (PFA). The isolated brain was postfixed with 4% PFA for 1 h at 4°C and cryoprotected with 20% sucrose overnight. Brain sections (20 µm) prepared with a cryomicrotome (CM1200; Leica) were permeabilized and blocked with PBS containing 0.2% Triton X-100, 10% FBS, and 1% BSA for 1 h at room temperature. After incubation with antibodies for green fluorescent protein (monoclonal; GF090R, 1:100; NACALAI TESQUE) and glial fibrillary acidic protein (GFAP, polyclonal; G9269, 1:20; Sigma) overnight at 4°C, the sections were stained with Alexa Fluor 546- or 647-labeled goat anti-mouse or anti-rabbit IgG antibody (1:200; Invitrogen) for 1 h at room temperature. Sections were imaged using a confocal microscope (LEICA TCS SP8; Leica) equipped with an oil-immersion objective (63×, NA = 1.4).
Data analysis
Data were analyzed using ImageJ software. Astrocytes were analyzed in layer 2/3 of the cortex or in the CA1 stratum radiatum of the hippocampus. Fluorescence intensities were corrected for background fluorescence by measuring a non-fluorescent area and normalized by the average of the first 10 or 20 frames to calculate the fractional changes in fluorescence intensity (ΔF/F0). Spontaneous G-CEPIA1er responses were manually detected and analyzed. Fluorescence changes with a rate of decrease faster than –0.015 ΔF/F0 s-1 and amplitude larger than –0.05 ΔF/F0 were defined as responses. Synchronous decreases in G-CEPIA1er fluorescence throughout a single astrocyte were defined as global responses. Localized fluorescence decreases with about 10–15 µm in diameter at the process region were defined as process-localized responses. The amplitude of spontaneous G-CEPIA1er responses was measured after smoothing time courses of ΔF/F0 by a moving average with five frame windows. The amplitude of NE-induced G-CEPIA1er, CEPIA2mt, GCaMP6f, or ER-YFP responses was defined as the average of ΔF/F0 within the 3-min time window starting from the start of NE application. The amplitude of the CPA-induced G-CEPIA1er fluorescence decrease was defined as the ΔF/F0 value in a steady state.
Results
G-CEPIA1er detects Ca2+ release from the ER in cortical astrocytes of IP3R2-KO mice
We used a serotype 5 AAV carrying the minimal astrocyte-specific gfaABC1D promoter to express G-CEPIA1er in astrocytes of the adult mouse cortex. Immunohistochemical analysis indicated co-localization of G-CEPIA1er and the astrocyte marker GFAP (Fig. 1A), confirming the astrocyte specificity of AAV5-gfaABC1D-mediated expression (Shigetomi et al., 2013). At subcellular levels, expression of G-CEPIA1er was observed throughout the soma and processes, although localization at the tips of ramified processes was unclear (Fig. 1A). This G-CEPIA1er distribution was consistent with a report on the distribution pattern of the ER in astrocytes (Patrushev et al., 2013).
To analyze Ca2+ release from the ER, G-CEPIA1er-expressing layer 2/3 astrocytes in acute cortical slice preparations were imaged using a two-photon microscope. Application of CPA, a sarco/endoplasmic reticulum Ca2+ -ATPase (SERCA) inhibitor, was used to deplete Ca2+ within the ER. CPA-induced Ca2+ depletion resulted in a marked decrease in the G-CEPIA1er fluorescence intensity of both WT and IP3R2-KO astrocytes, confirming that G-CEPIA1er reports ER Ca2+ levels (Fig. 1B). There was no significant difference in the amplitudes of the CPA-induced G-CEPIA1er fluorescence decrease between WT and IP3R2-KO astrocytes, indicating that the basal [Ca2+]ER in IP3R2-KO astrocytes was similar to that in WT astrocytes (Fig. 1B). However, the rate of decrease in [Ca2+]ER was greater in WT astrocytes than in IP3R2-KO astrocytes (Fig. 1B). In WT astrocytes, [Ca2+]ER was often decreased in a stepwise manner as shown in Fig. 1B. Each step appeared to reflect spontaneous Ca2+ release events via IP3R2 as described in the following.
We next investigated Gq-coupled receptor-induced Ca2+ release from the ER in astrocytes. NE, which is released from the axons of locus coeruleus neurons, induces significant ER Ca2+ release through activation of α1-adrenergic receptors in astrocytes in vivo (Ding et al., 2013; Paukert et al., 2014; Srinivasan et al., 2015). Although previous studies have shown that NE-induced Ca2+ release is absent in IP3R2-KO astrocytes, we re-evaluated this result using G-CEPIA1er. Bath application of NE (10 µM) induced a large decrease in [Ca2+]ER of WT astrocytes, indicating robust Ca2+ release from the ER (Fig. 2A and B). Following washout of NE, there was slow recovery of [Ca2+]ER due to Ca2+ reuptake by the ER. The time course was consistent with ER Ca2+ dynamics induced by IP3R-mediated Ca2+ release in other cell types shown in our previous studies (Suzuki et al., 2014; Okubo et al., 2015). NE-induced [Ca2+]ER responses showed almost the same amplitude and time course in the soma and processes, indicating induction of cell-wide Ca2+ releases in response to bath application of NE (Fig. 2A, B and D). To our surprise, NE application induced a smaller but significant decrease in [Ca2+]ER of IP3R2-KO astrocytes throughout the soma and processes (Fig. 2A, B and D). This observation clearly indicates that ER Ca2+ release is not completely abolished in IP3R2-KO astrocytes.
Most GECIs including G-CEPIA1er are sensitive to pH. To exclude the possibility that the observed changes in the G-CEPIA1er fluorescence intensity was due to a change in pH within the ER, we expressed ER-localized YFP that is sensitive to pH but insensitive to Ca2+ (Suzuki et al., 2014). ER-YFP showed no response to NE in either WT or IP3R2-KO astrocytes, confirming that NE-induced G-CEPIA1er responses reflected decreases in [Ca2+]ER (Fig. 2C and D).
G-CEPIA1er detects spontaneous Ca2+ release in WT astrocytes
During our observation of [Ca2+]ER in WT astrocytes, we found spontaneous decreases in [Ca2+]ER (Fig. 3). There were two types of spontaneous events: a “global” response that showed a synchronous decrease in [Ca2+]ER throughout the astrocyte (Fig. 3A and D), and a “process-localized” response that was localized at processes with diameters of about 10–15
µm (Fig. 3B and D). These Ca2+ release dynamics may at least partly correspond to cell-wide and localized Ca2+ signals imaged by cytosolic Ca2+ indicators in previous studies (Kanemaru et al., 2014; Srinivasan et al., 2015), although precise correlations require further analyses. However, we did not find spontaneous [Ca2+]ER responses in IP3R2-KO astrocytes (Fig. 3C). Thus, in IP3R2-KO astrocytes, spontaneous Ca2+ release events, if any, appeared to be attenuated to a level undetectable by G-CEPIA1er.
Cytosolic Ca2+ transients are mediated by IP3R2-independent Ca2+ release
To clarify the significance of IP3R2-independent Ca2+ release in the intracellular Ca2+ signaling of astrocytes, we investigated changes in the cytosolic Ca2+ concentration ([Ca2+]Cyt) in response to NE application (Fig. 4). Cytosolic Ca2+ indicator GCaMP6f (Chen et al., 2013) was expressed in astrocytes using an AAV, showing a diffuse distribution throughout the cytosol (Fig. 4A). Application of NE induced large and cell-wide [Ca2+]Cyt responses in WT astrocytes (Fig. 4B and C). These NE-induced [Ca2+]Cyt responses showed no significant subcellular distribution, but were synchronized throughout the cell. Responses were abolished after treatment with CPA, indicating a critical contribution of ER Ca2+ release to NE-induced [Ca2+]Cyt transients (Fig. 4B and C). We also observed spontaneous and spatially confined [Ca2+]Cyt responses that probably correspond to previously reported “Ca2+ twinkle” or “microdomain” Ca2+ signals (Kanemaru et al., 2014; Srinivasan et al., 2015). These localized [Ca2+]Cyt signals hardly contributed to ΔF/F0 traces of GCaMP6f averaged within the entire cell.
In IP3R2-KO astrocytes, NE induced very small [Ca2+]Cyt responses (Fig. 4B and C). Only very few IP3R2-KO astrocytes showed clear but relatively small responses. These [Ca2+]Cyt responses in IP3R2-KO astrocytes were also abolished after treatment with CPA (Fig.4B and C). These results indicate that IP3R2-independent Ca2+ release, which was detected by G-CEPIA1er, did not always increase [Ca2+]Cyt to a level detectable by GCaMP6f.
Significant Ca2+ transients in mitochondria are mediated by IP3R2-independent Ca2+ release
We next examined whether IP3R2-independent Ca2+ release contributed to astrocytic Ca2+ signaling other than cell-wide cytosolic Ca2+ transients. To this end, we focused on mitochondria. The ER and mitochondria often form close contacts where Ca2+ is efficiently transferred from the ER to mitochondria without a global increase in [Ca2+]Cyt (Rizzuto et al., 1998; de Brito and Scorrano, 2010; Hirabayashi et al., 2017). We thus investigated changes in the Ca2+ concentration within mitochondria ([Ca2+]Mito) using CEPIA2mt, a Ca2+ indicator for the mitochondrial matrix (Suzuki et al., 2014) (Fig. 5). AAV-mediated expression of CEPIA2mt in astrocytes resulted in a punctate distribution of fluorescence, which was consistent with the morphology of mitochondria (Fig. 5A). Application of NE to release Ca2+ from the ER induced an increase in [Ca2+]Mito of WT astrocytes (Fig. 5B and C). A fraction of WT astrocytes did not show a response as observed in HeLa cells in our previous study (Suzuki et al., 2014) (Fig. 5B and C). The NA-induced [Ca2+]Mito response was abolished after treatment with CPA to deplete the ER (Fig. 5B and C). These results indicate that Ca2+ enters mitochondria following Ca2+ release from the ER in WT astrocytes. We did not observe subcellular heterogeneity in the NE-induced [Ca2+]Mito responses.
Of note, NE also induced [Ca2+]Mito responses in IP3R2-KO astrocytes (Fig. 5B and C). Similar to WT astrocytes, a fraction of IP3R2-KO astrocytes showed no responses. Compared with the minimal [Ca2+]Cyt responses, [Ca2+]Mito showed attenuated but significant responses in IP3R2-KO astrocytes in response to NE. These [Ca2+]Mito responses in IP3R2-KO astrocytes were also abolished after treatment with CPA (Fig. 5B and C). These results indicate that IP3R2-independent Ca2+ release efficiently induces an increase in [Ca2+]Mito, and suggest that IP3R2-independent Ca2+ release increases the local Ca2+ concentration in the vicinity of the ER.
IP3R2-independent Ca2+ release in hippocampal astrocytes
Regional differences in the properties of astrocytes have been reported (Chai et al., 2017). To investigate whether IP3R2-independent Ca2+ release could be observed in astrocytes of brain regions other than the cortex, we visualized Ca2+ release from the ER in hippocampal astrocytes (Fig. 6). G-CEPIA1er was expressed in hippocampal astrocytes using an AAV. G-CEPIA1er-expressing astrocytes in the CA1 stratum radiatum of acute hippocampal slice preparations were imaged by two-photon microscopy. “Global” and “process localized” spontaneous G-CEPIA1er responses were observed in WT astrocytes, but not in IP3R2-KO astrocytes, which was consistent with the results in cortical astrocytes (Fig. 6A–D). Application of NE induced large G-CEPIA1er responses in WT astrocytes, whereas attenuated but significant G-CEPIA1er responses were observed in IP3R2-KO astrocytes (Fig. 6E). Therefore, IP3R2-independent Ca2+ release was commonly observed in cortical and hippocampal astrocytes of IP3R2-KO mice.
Discussion
In this study, we showed Ca2+ release from the ER in IP3R2-KO astrocytes for the first time. This IP3R2-independent Ca2+ release was effective to increase the Ca2+ concentration within mitochondria that make close contacts with the ER. These results provide new insights into the functional significance of ER Ca2+ release in astrocytes. The controversy derived from the assumption that ER Ca2+ release is abolished in IP3R2-KO astrocytes should be re-evaluated considering the presence of IP3R2-independent Ca2+ release.
Highly sensitive and specific detection of ER Ca2+ release using G-CEPIA1er
Although IP3R2-independent Ca2+ release from the ER in IP3R2-KO astrocytes was detectable by G-CEPIA1er responses, the resulting changes in [Ca2+]Cyt were only barely detected by GCaMP6f responses. The difficulty in detecting the [Ca2+]Cyt response is consistent with previous reports that failed to detect evoked [Ca2+]Cyt responses in IP3R2-KO astrocytes (Petravicz et al., 2008, 2014, Agulhon et al., 2010, 2013; Nizar et al., 2013; Takata et al., 2013). These results suggest that the amount of Ca2+ release sufficient to change free [Ca2+]ER is often insufficient to change free [Ca2+]Cyt. There appear to be some potential reasons. First, the volume occupied by the smooth ER is about 10% of the cytoplasm (Paumgartner et al., 1981). Therefore, Ca2+ released from the ER is diluted in the cytosol. Another possible reason is the Ca2+ buffering capacity k (a change in the total Ca2+ concentration in a subcellular compartment divided by the corresponding change in free Ca2+ concentration in the compartment). In many cell types, cytosolic k is estimated to be 10–1000 (Neher, 1995; Mogami et al., 1999). In pancreatic acinar cells, k of the ER is estimated to be two orders of magnitude smaller than that of cytosol (Mogami et al., 1999). If k is similarly lower in the ER than in the cytosol, this would also reduce the change in [Ca2+]Cyt. Therefore, G-CEPIA1er is expected to report ER Ca2+ release with higher sensitivity than cytosolic Ca2+ indicators.
Furthermore, G-CEPIA1er can specifically detect ER Ca2+ release, whereas cytosolic Ca2+ indicators do not specify the source of Ca2+.
Ca2+ channels that mediate IP3R2-independent Ca2+ release
Although IP3R2 is the major ER Ca2+ release channel in astrocytes, expression of IP3R1 and/or IP3R3 in astrocytes has been indicated by transcriptome analyses (Cahoy et al., 2008; Zhang et al., 2014; Chai et al., 2017) and immunohistochemical studies (Yamamoto-Hino et al., 1995; Sharp et al., 1999). Furthermore, comparison between IP3R2-KO mice and IP3R2/IP3R3-double KO mice indicate a functional contribution of IP3R3 to astrocytic Ca2+ signaling, albeit small (Tamamushi et al., 2012; Sherwood et al., 2017). These previous studies suggest the presence of IP3-induced Ca2+ release via IP3R1 and/or IP3R3 in IP3R2-KO astrocytes.
In addition, expression of ryanodine receptor type3 (RyR3) in astrocytes has been reported (Matyash et al., 2002; Chai et al., 2017). Furthermore, the functional significance of RyR3 was reported in both astrocytic Ca2+ signaling and motility (Matyash et al., 2002). Therefore, it also appears possible that Ca2+ -induced Ca2+ release via RyR3 enhances Ca2+ release from the ER in collaboration with IP3R1 and/or IP3R3-dependent Ca2+ release in astrocytes.
Contribution of IP3R2-independent Ca2+ release to Ca2+ signaling in IP3R2-KO astrocytes
When the extent of the NE-induced increase in Ca2+ concentration was compared between WT and IP3R2-KO astrocytes, there was a much higher Ca2+ increase in mitochondria than in the cytoplasm of IP3R2-KO astrocytes (Figs. 4 and 5). These results suggest a mechanism that allows privileged transfer of Ca2+ from the ER to mitochondria. Indeed, it has been shown that the ER and mitochondria make close contacts (10–30 nm in distance), and that such contact sites are crucial for efficient Ca2+ transfer (Rizzuto et al., 1998; de Brito and Scorrano, 2010; Hirabayashi et al., 2017). This finding suggests that IP3R2-independent Ca2+ release is capable of increasing [Ca2+]Cyt within close proximity to the ER membrane, generating Ca2+ nanodomains around the ER (Rizzuto and Pozzan, 2006). Thus, in addition to extracellular and mitochondrial sources of Ca2+ for localized Ca2+ signaling (Srinivasan et al., 2015; Agarwal et al., 2017), IP3R2-independent Ca2+ release may contribute to Ca2+ signaling in IP3R2-KO astrocytes.
We also noted the spatial distribution of IP3R2-independent Ca2+ release, which was observed throughout the cell including the soma and processes in NE-stimulated IP3R2-KO astrocytes (Fig. 2). This compares with the subcellular-confined Ca2+ signals within the processes due to Ca2+ influx from the extracellular space or Ca2+ release from mitochondria in IP3R2-KO astrocytes in the absence of agonist stimulation (Kanemaru et al., 2014; Srinivasan et al., 2015; Agarwal et al., 2017). Therefore, IP3R2-independent Ca2+ release may have distinct roles, especially in the somatic region of IP3R2-KO astrocytes.
Reinterpretation of astrocytic functions in IP3R2-KO mice
It has been assumed that Ca2+ release from the ER is absent in IP3R2-KO astrocytes. Thus, the absence of functional deficits in IP3R2-KO mice (Petravicz et al., 2008, 2014, Agulhon et al., 2010, 2013; Nizar et al., 2013; Takata et al., 2013; Bonder and McCarthy, 2014) has been considered to indicate the absence of functional roles of intracellular Ca2+ release in astrocytes. The presence of IP3R2-independent Ca2+ release may necessitate reinterpretation of such results.
Ca2+ release from the ER may be inhibited by expressing the IP3 hydrolyzing enzyme, IP3 5-phosphatase (5ppase) (Kanemaru et al., 2007; Mashimo et al., 2010; Foley et al., 2017). Indeed, exogenous expression of 5ppase effectively suppresses IP3 production following IP3R-dependent Ca2+ release in astrocytes (Kanemaru et al., 2007; Mashimo et al., 2010). Because 5ppase is expected to suppress IP3R-induced Ca2+ release regardless of IP3R subtypes, the phenotypic difference between 5ppase-expressing mice and IP3R2-KO mice may clarify the role of IP3R2-independent Ca2+ release. In fact, a recent study has indicated that astrocytic expression of 5ppase disrupts sleep (Foley et al., 2017), whereas sleep disruption was not observed in IP3R2-KO mice (Cao et al., 2013). Although there remains the possibility of developmental adaptation or other compensatory mechanisms in these mice, the results suggest that residual Ca2+ signaling, including IP3R2-independent Ca2+ release in IP3R2-KO mice, accounts for the absence of functional deficits in IP3R2-KO mice.
In conclusion, the presence of IP3R2-independent Ca2+ release as well as Ca2+ signaling generated by sources other than the ER limit the use of IP3R2-KO mice to assess the role of Ca2+ signaling in astrocytic functions when functional deficits are not observed in the mutant mice. This notion may help to resolve the long-lasting controversy arising from the assumption that Ca2+ signaling is abolished in IP3R2-KO astrocytes. Furthermore, the present results facilitate the study of the functional roles of Ca2+ signaling in close proximity to the ER in astrocytes.
Acknowledgements
We thank Y. Kawashima for technical assistance and Dr. B. Khakh (University of California at San Francisco) for providing the pZac2.1-gfaABC1D-Lck-GCaMP3 plasmid. This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI [Grant Numbers JP16K08543 (Y.O.), JP15H05648 (K.K.), and JP21229004 and JP25221304 (M.I.)] and grants from the Tokyo Society of Medical Sciences (Y.O.).