Abstract
The key pathologic entities driving the destruction of synaptic function and integrity during the evolution of Alzheimer's disease (AD) remain elusive. Astrocytes are structurally and functionally integrated within synaptic and vascular circuitry and use calcium-based physiology to modulate basal synaptic transmission, vascular dynamics, and neurovascular coupling, which are central to AD pathogenesis. We used high-resolution multiphoton imaging to quantify all endogenous calcium signaling arising spontaneously throughout astrocytic somata, primary processes, fine processes, and capillary endfeet in the brain of awake APP/PS1 transgenic mice (11 male and 6 female mice). Endogenous calcium signaling within capillary endfeet, while surprisingly as active as astrocytic fine processes, was reduced ∼50% in the brain of awake APP/PS1 mice. Cortical astrocytes, in the presence of amyloid plaques in awake APP/PS1 mice, had a cell-wide increase in intracellular calcium associated with an increased frequency, amplitude, and duration of spontaneous calcium signaling. The cell-wide astrocytic calcium dysregulation was not directly related to distance to amyloid plaques. We could re-create the cell-wide intracellular calcium dysregulation in the absence of amyloid plaques following acute exposure to neuronally derived soluble Abeta from Tg2576 transgenic mice, in the living brain of male C57/Bl6 mice. Our findings highlight a role for astrocytic calcium pathophysiology in soluble-Abeta mediated neurodegenerative processes in AD. Additionally, therapeutic strategies aiming to protect astrocytic calcium physiology from soluble Abeta-mediated toxicity may need to pharmacologically enhance calcium signaling within the hypoactive capillary endfeet while reducing the hyperactivity of spontaneous calcium signaling throughout the rest of the astrocyte.
SIGNIFICANCE STATEMENT Astrocytic calcium signaling is functionally involved in central pathologic processes of Alzheimer's disease. We quantified endogenous calcium signaling arising spontaneously in the brain of awake APP/PS1 mice, as general anesthesia suppressed astrocytic calcium signaling. Cell-wide astrocytic calcium dysregulation was not related to distance to amyloid plaques but mediated in part by neuronally derived soluble Abeta, supporting a role for astrocytes in soluble-Abeta mediated neurodegeneration. Spontaneous calcium signaling is largely compartmentalized and capillary endfeet were as active as fine processes but hypoactive in the presence of amyloid plaques, while the rest of the astrocyte became hyperactive. The cell-wide calcium pathophysiology in astrocytes may require a combination therapeutic strategy for hypoactive endfeet and astrocytic hyperactivity.
Introduction
Alzheimer's disease (AD) is a devastating neurologic illness and one of the top 10 causes of death in the United States (Heron, 2021). It is expected that, by 2060, >13 million individuals in the United States will be living with a diagnosis of AD (Rajan et al., 2021). These estimates highlight the clinical urgency for disease-modifying therapies for AD. The most significant known pathologies of AD are synaptic loss and neuronal cell death, but the pathologies driving the destruction of the critical components of synaptic function and integrity remain incompletely understood (DeKosky and Scheff, 1990; Terry et al., 1991; Selkoe, 2002). Astrocytes have a profound influence over local synaptic transmission and can upregulate basal synaptic transmission using calcium-dependent mechanisms (Di Castro et al., 2011; Panatier et al., 2011). Astrocytic calcium signaling is intricately coupled with synaptic and vascular dynamics, neurovascular coupling, glutamate homeostasis, potassium release, and sleep (Parpura and Haydon, 2000; Girouard et al., 2010; Khakh and McCarthy, 2015; Bazargani and Attwell, 2016; Mishra et al., 2016; Papouin et al., 2017; Bojarskaite et al., 2020). These calcium-dependent physiological functions of astrocytes have been reported to be dysfunctional within the human AD brain and in mouse models of AD (DeKosky and Scheff, 1990; Masliah et al., 1996; Takano et al., 2007; Hefendehl et al., 2016; Boscia et al., 2017; Tarantini et al., 2017). Previous in vivo studies using the bulk loading of calcium indicator dyes into the brain of anesthetized APP/PS1 mice revealed a wide-spreading calcium dysregulation throughout the astrocytic network in the presence of amyloid plaques (Kuchibhotla et al., 2009; Delekate et al., 2014; Reichenbach et al., 2018). However, use of general anesthesia can greatly suppress astrocytic calcium signaling, and these commonly used calcium indicator dyes can limit quantification of astrocytic calcium signaling to within somata and proximal processes, leaving most (∼90%) of the astrocyte unsampled (Reeves et al., 2011; Thrane et al., 2012). Astrocytic fine processes occupy most cellular volume (∼75% of cellular volume), exhibit the most calcium signaling events, and use calcium-dependent mechanisms to influence thousands of local synapses, neighboring astrocytes, and possibly other cell types (Bushong et al., 2002; Nedergaard and Verkhratsky, 2012; Bindocci et al., 2017; Papouin et al., 2017; Kelly et al., 2018; Kiyoshi et al., 2020). Therefore, the intracellular calcium signaling occurring within astrocytic somata cannot serve as a surrogate measure for the intracellular calcium activity within the rest of the astrocyte (Reeves et al., 2011; Srinivasan et al., 2015; Bazargani and Attwell, 2016). Interestingly, astrocytic calcium signaling varies greatly throughout astrocytes with respect to spatiotemporal dynamics, magnitude, and frequency (Di Castro et al., 2011; Kanemaru et al., 2014; Srinivasan et al., 2015; Bindocci et al., 2017). Astrocytes exhibit highly localized intracellular calcium signaling within their fine processes (Shigetomi et al., 2013; Bindocci et al., 2017), vascular endfoot processes (Shigetomi et al., 2013; Delekate et al., 2014; Bindocci et al., 2017; Lind et al., 2018), as well as asynchronous localized intracellular calcium events within individual primary processes (Reeves et al., 2011; Volterra et al., 2014). We used our ratiometric genetically encoded calcium indicator, yellow Cameleon gfa2.yc3.6, to quantify all intracellular and intercellular calcium signaling arising spontaneously throughout cortical astrocytic somata, primary processes, fine processes, and vascular endfeet in the presence of amyloid plaques in awake APP/PS1 mice, without use of general anesthesia. We found a cell-wide calcium dysregulation in cortical astrocytes unrelated to distance to amyloid plaques in vivo. Capillary endfeet are hypoactive in the presence of amyloid plaques, while calcium signaling occurred with increased frequency, amplitude, and duration throughout the rest of the astrocyte. Neuronally derived soluble Abeta evoked cell-wide astrocytic calcium dysregulation in the absence of amyloid plaques highlighting a role for astrocytes in soluble-Abeta mediated neurodegeneration in AD. Therapies restoring astrocytic calcium signaling may need a combination therapy to enhance calcium signaling within capillary endfeet, while separately therapeutically protecting the rest of the astrocyte from hyperactivity.
Materials and Methods
AAV-GFA2-YC3.6 cloning and production
The AAV-GFA2-YC3.6 vector backbone was constructed by cloning the ratiometric-based calcium indicator yellow Cameleon YC3.6 (Nagai et al., 2004) from our initial AAV-CBA-YC3.6 plasmid (Arbel-Ornath et al., 2017) in place of the GFP reporter initially present in the AAV-GFA2-GFP plasmid generously provided by Norris from the University of Kentucky College of Medicine (Furman et al., 2012). The astrocyte promoter GFA2 was initially characterized by Michael Brenner's laboratory at the University of Alabama–Birmingham (Lee et al., 2008). Briefly, the YC3.6 sequence was amplified by PCR (forward primer: 5′-CCGCGGGCCCGGGATCCACCGGGTACAAGTAAAGCGGCCGCCACCATGGTGAGCAAGGGCGAGGAGCTGTTCACCGGGGTGG-3′; reverse primer: 5′-CAATTGTTGTTGTTAACTTGTTTATTGCAGCTTATAATGGTTACAAATAAAGCAA-3′), the PCR product was digested by SacII and MfeI and cloned in place of the GFP reporter after a similar cut of the plasmid AAV-GFA2-GFP. The identity of the ITR sequences in the vector backbone was from an AAV2. After full sequencing of the vector backbone, high titers of AAV serotype 5 were produced using the triple transfection protocol by the Massachusetts Eye and Ear infirmary Gene Transfer Vector Core. The virus was titrated by quantitative PCR, and the final concentrations of these AAV viral stocks reached 4 × 1012 vg/ml.
Animals
APP/PS1 transgenic mice on a C57BL/6J genetic background (male and female; MMRRC Strain #034832-JAX; The Jackson Laboratory) (Jankowsky et al., 2001) were 12- to 17-month-old, age-matched nontransgenic littermates, C57BL/6J mice (male; 3- to 6-month-old; strain #000664; The Jackson Laboratory) and Tg2576 mice (male; 2- to 3-month-old; model #1349; Taconic Farms) were socially housed with food and water available ad libitum. An automated lighting system created 12 h light/12 h dark cycles, and temperature was controlled within the facility. Male Tg2576 mice were mated with age-matched nontransgenic female mice for preparation of transgenic cortical neurons and nontransgenic cortical neurons. All surgical procedures and postoperative care for each animal were conducted in accordance with the Massachusetts General Hospital Institutional Animal Care and Use Committee (IACUC)-approved protocol.
Preparation of cultured neuronally derived bioactive soluble Abeta
Primary neuronal cultures and conditioned media were prepared as described previously (Arbel-Ornath et al., 2017; Calvo-Rodriguez et al., 2020). Mouse embryos (E14-E16; male Tg2576 mice mated with female nontransgenic mice) were collected and genotyped. Transgenic primary neurons and nontransgenic primary neurons were cultured from the embryonic cerebral cortices. Cortical neurons were maintained within Neurobasal/B-27 media for 14 d without renewal of the media; and finally, the media was collected (conditioned media [CM]). For preparation of immunodepleted- transgenic conditioned media (TgCM), Abeta was removed within TgCM by overnight incubation with 6E10 antibody (BioLegend catalog #803004) and protein G Sepharose beads (Sigma-Aldrich). The protein G beads were washed in cold Neurobasal medium, incubated with 1 ml TgCM and centrifuged. The supernatant was incubated overnight with 100 µl prewashed protein G beads and 9 µg of 6E10 antibody. The concentration of Aβ40 within the TgCM, nontransgenic conditioned media (ntgCM), and immunodepleted-transgenic conditioned media (immunodepleted-TgCM) was determined by ELISA (Wako).
Stereotaxic intracortical injections of AAV-GFA2.YC3.6 under anesthesia
Aged APP/PS1 transgenic mice (n = 10), age-matched nontransgenic mice (n = 9), and young C57BL/6J mice (n = 5-7 mice per group) each received bihemispheric intracortical injections of AAV2/5.gfa2.yc3.6 into the somatosensory cortex, under inhaled general anesthesia of 1.5% isoflurane and oxygen. Briefly, once the animal reached a stable plane of anesthesia, with loss of pedal withdrawal reflexes, the mouse was secured within a heated, stereotactic frame. Sterile ophthalmic ointment (Dechra) was carefully applied to the animal's eyes to prevent corneal desiccation. Preoperative preparation of the animal's scalp included hair removal and a series of alternating washing steps with povidone-iodine scrub (Betadine Surgical Scrub) followed by 70% isopropyl alcohol. A small burr hole was drilled through the skull at 1 mm posterior and 1 mm distance lateral from bregma to permit the delivery of 3 µl AAV2/5.gfa2.yc3.6 at a depth of −0.6 mm into the somatosensory cortex at 0.15 µl/minute (Spires et al., 2005). The animal's scalp was sutured and covered in antibiotic ointment (Curad). IACUC-approved analgesics and postoperative care were provided for 3 d after surgery.
Cranial window surgery for awake in vivo multiphoton imaging
Aged APP/PS1 transgenic mice and age-matched nontransgenic mice, at 3 weeks following gfa2.yc3.6 intracortical injections, were anesthetized to a stable plane of surgical anesthesia (5% isoflurane for induction and ∼1.5% isoflurane and oxygen for maintenance). A ∼6 mm circular portion of the skull, over gfa2.yc3.6 intracortical injection sites, was surgically removed, and an 8 mm sterile coverslip with PBS secured to the animal's skull using a combination of dental cement and glue (Skoch et al., 2005). A custom-made 10 mm headpost (Ponoko) was attached to the animal's skull for awake multiphoton imaging (Van Veluw et al., 2020). Analgesics and postoperative care were provided for 3 d after surgery, following an IACUC-approved protocol.
In vivo imaging of cortical astrocytes exposed to neuronally derived soluble Abeta (TgCM), topically applied onto the brain of anesthetized C57BL/6J mice
C57BL/6J mice, at 3 weeks following gfa2.yc3.6 intracortical injections, were anesthetized (5% isoflurane for induction and 1% isoflurane and oxygen to maintain a stable surgical plane of anesthesia). Animals were secured within a heated stereotactic frame and sterile ophthalmic ointment (Dechra) applied to eyes. A ∼6 mm circular skull portion, above gfa2.yc3.6 injection sites, and dura mater were surgically removed. A sterile 8 mm coverslip with PBS was secured to the skull by a combination of dental cement (Lang Dental) and glue (Skoch et al., 2005). For in vivo multiphoton imaging, we used our Olympus FluoView FV1000MPE multiphoton laser-scanning system mounted on an Olympus BX61WI microscope with a 25× objective (Olympus; NA = 1.05). A Mai Tai DeepSee Ti:sapphire laser (Spectra-Physics) generated two-photon excitation at 860 nm, and three photomultiplier tubes (Hamamatsu) collected emitted light in the cyan (460-500 nm), yellow (520-560 nm), and red (575-630 nm) spectral channels. The cerebral vasculature was fluorescently labeled by intravenous administration of 150 µl Texas Red dextran (70 kDa; 12.5 mg/ml in PBS; Invitrogen) before baseline in vivo multiphoton imaging (512 × 512, 2 µm slices, depth of 200-300 µm; acquiring multiple cortical volumes containing an average of 70 gfa2.yc3.6-expressing astrocytes per animal; total imaging duration ∼1 h). The coverslip above the craniotomy was surgically removed and replaced with a sterile coverslip containing ∼50 µl of TgCM, ntgCM, or immunodepleted-TgCM. After 1 h the cortical volumes acquired during baseline imaging were reimaged. Photomultiplier tube settings (PMTs) remained the same throughout all imaging sessions while the laser power was adjusted as required. After the final imaging session, each animal was sacrified immediately in accordance with the Massachusetts General Hospital IACUC-approved protocol.
Acute effects of TgCM were evaluated also in vitro. Briefly, WJE cells (Morikawa et al., 2005) maintained in Advanced DMEM (Invitrogen) supplemented with 10% FBS (Atlanta Biologicals) and 1% Glutamax (Invitrogen) in a humidified 37°C incubator with 5% CO2 were transfected at 70%-80% confluence with 4 µg gfa2.yc3.6 plasmid using Lipofectamine 2000 (Invitrogen), as instructed by the manufacturer. At 24 h later, YFP/CFP ratios were acquired for cells, before and following acute exposure to either ntgCM or TgCM for 1 h, using an Olympus FV3000RS Confocal Laser Scanning Microscope equipped with CO2/heating units (Tokai-Hit STX-Co2 Digital CO2 Gas Mixing System, STFX model). Two-photon excitation was used at 860 nm, and emitted light was collected in two channels in the range 520-560 nm (yellow) and 460-500 nm (cyan).
In vivo imaging of endogenous astrocytic calcium signaling in awake mice
For 1 week before awake in vivo multiphoton imaging, each APP/PS1 mouse (n = 10) and nontransgenic mouse (n = 9) was habituated to being awake while head-fixed within the custom-made stereotactic frame used for awake imaging (Gao et al., 2017). On the day before in vivo imaging, each mouse received 300 µl intraperitoneal injection of fluorescent amyloid-binding dye methoxy X04 (Glixx laboratory; ∼5 mg/kg dissolved in Cremophor EL [Sigma-Aldrich] and PBS) (Klunk et al., 2002). Cerebral vasculature was fluorescently labeled before awake in vivo multiphoton imaging by a 150 µl intravenous injection of Texas Red dextran (70 kDa; 12.5 mg/ml in PBS; Invitrogen). For awake in vivo multiphoton imaging, we used our multiphoton laser-scanning system, generated two-photon excitation at 860 nm, and collected emitted light in the cyan (460-500 nm), yellow (520-560 nm), and red (575-630 nm) spectral channels. PMTs remained the same throughout each awake in vivo multiphoton imaging session, and laser power was kept to a minimum. An average of 7 cortical volumes was acquired for each awake mouse (Z series, 512 × 512 pixels; 4× magnification; imaging acquisition beginning at first appearance of fluorescent cerebral vasculature in mouse cortex and continuing at 2 µm step size for up to 130 slices). The endogenous astrocytic intracellular and intercellular calcium signaling occurring spontaneously within the awake mouse brain was acquired using high-magnification time-lapse imaging (300 s acquired 700 frames at ∼2.3 Hz, resolution of 256 × 256 at 2 µs/pixel; 8× magnification) for 129 YC3.6-expressing cortical astrocytes (60 astrocytes from awake nontransgenic mice [n = 9] and 69 astrocytes from awake APP/PS1 mice [n = 10], imaging session lasting ∼2 h). Upon completion of awake in vivo imaging, each animal was sacrified in accordance with the Massachusetts General Hospital IACUC-approved protocol.
In vivo image analysis
Public domain ImageJ/FIJI software (Schindelin et al., 2012) was used to manually draw ROIs around all astrocytic cellular compartments (somata, primary processes, and capillary endfeet). Corresponding YFP/CFP ratios determined using custom-written macro for ImageJ/FIJI software. In vivo multiphoton images were pseudo-colored using custom-written MATLAB script (version R2016a; The MathWorks). Astrocytic calcium signaling events associated with mouse movement were analyzed separately. We used multiple custom-written MATLAB scripts to quantify the baseline ratio, frame number, amplitude, and duration of all calcium events occurring spontaneously within and between cortical astrocytes in 129 in vivo time-lapses within the brain of awake mice [10 awake APP/PS1 mice (69 astrocytes) and 9 awake nontransgenic mice (60 astrocytes)]. All spontaneous calcium events (greater than a predetermined threshold of 3 SDs greater than baseline) were manually confirmed by visual inspection of each pseudo-colored time-lapse image.
Confirming astrocytic specificity of AAV2/5.GFA2.YC3.6 within murine brain in vivo by colocalization with sulforhodamine 101 (SR-101)-labeled astrocytes
A 5-month-old C57BL/6J mouse was prepared for the surgical delivery of 3 µl AAV2/5.gfa2.yc3.6 into both hemispheres of the somatosensory cortex, as described in Materials and Methods. At least 3 weeks after surgery, the mouse received inhaled general anesthesia at an induction rate of 5% isoflurane in oxygen until a stable anesthetic plane, which was maintained at 1.5% isoflurane in oxygen. The mouse skull was prepared for the surgical etching of a ∼6 mm craniotomy over the somatosensory cortex. The skull portion was carefully removed before removal of the murine dura to permit the topical application of a filtered (50 μm) solution of Sulforhodamine 101 (SR-101; Sigma-Aldrich) dissolved in PBS (Nimmerjahn et al., 2004; Rasmussen et al., 2016) onto the surface of the living mouse brain. The SR-101 solution remained on the surface of the living brain for ∼10 min before thorough washing of the surface of the brain with PBS. The mouse remained under 1.5% isoflurane and oxygen during the multiphoton imaging of AAV2/5.gfa2.yc3.6-expressing astrocytes in the mouse cortex, which colocalized with SR-101-labeled astrocytes.
Immunohistochemical examination of AAV2/5.gfa2.yc3.6-expressing astrocytes in frozen mouse brain sections
A 7-month-old C57BL/6J mouse received a surgical bihemispheric intracortical injection of AAV2/5.gfa2.yc3.6 into the somatosensory cortex under general isoflurane, with induction at 5% isoflurane/oxygen and maintenance at 1.5% isoflurane/oxygen. Following suturing of the scalp and IACUC-approved postoperative care, the animal was killed at least 3 weeks later with carbon dioxide inhalation. The animal received an intracardiac perfusion of PBS followed by 4% PFA with removal of the whole brain that was fixed overnight within 4% PFA solution. The PFA-fixed whole brain was transferred to 30% sucrose solution in PBS for at least 3 d and was subsequently snap-frozen in optimal cutting temperature compound in dry ice and sectioned in 20-μm-thick sections using a cryostat (Microm550), with sections collected into cryoprotect solution. Brain sections with cortical astrocytes expressing AAV2.5.gfa2.yc3.6 were used for immunohistochemistry. Briefly, brain sections were washed with TBS (1×), permeabilized using 0.5% Triton X-100 in TBS for 20 min, washed in TBS for 5 min 3 times, blocked in 10% NGS at room temperature for 1 h, and subsequently incubated with GFAP (1:500, Sigma-Aldrich G3893), glutamine synthetase (1:500, Abcam ab73593), or NeuN (1:500, Mab377) overnight at 4 degrees. Sections were washed in TBS 1× for 5 min 3 times, incubated for 1 h at room temperature with either goat anti-mouse 555 (1:250, Invitrogen, A28180) or goat anti-rabbit 555 (1:250, Invitrogen, A27039), and washed with TBS for 5 min 3 times. Fluorescent mounting media was applied to the sections, and a coverslip was securely attached to each slide. All sections were examined using fluorescent microscopy (Zeiss Imager.Z2).
Statistics
GraphPad Prism software was used for all statistical comparisons. Parametric t tests or nonparametric Mann–Whitney U tests were used to statistically compare astrocytic intracellular calcium within the brain of awake APP/PS1 transgenic mice versus nontransgenic mice. The amplitude and duration of all spontaneous calcium events were statistically compared between groups using two-way ANOVA with Tukey's multiple comparison test. The comparison between ntgCM, tgCM, and immunodepleted-CM was done by one-way ANOVA with Tukey's multiple comparison test, or by paired two-tailed t test when comparing before and after CM application (calcium overload). p < 0.05 was considered statistically significant.
Results
Cell-wide astrocytic calcium dysregulation in APP/PS1 mice is re-created in absence of amyloid plaques following acute exposure to soluble Abeta in vivo
Use of our ratiometric genetically encoded calcium indicator, gfa2.yc3.6, permitted quantification of astrocytic intracellular calcium within 700-800 astrocytic somata, 1100-1600 primary processes, and 860-1000 astrocytic endfeet ROIs at 510-550 cerebral capillaries in the awake brain of 12- to 17-month-old APP/PS1 transgenic mice with amyloid plaques (n = 10) compared with age-matched awake nontransgenic mice (n = 9; Fig. 1A–D). We confirmed that cortical gfa2.yc3.6-expressing astrocytes in mouse brain did colocalize with the astrocytic immunohistochemical markers GFAP or glutamine synthetase (Fig. 2). Thereby, we have observed that the cortical astrocytes within the somatosensory cortex of awake APP/PS1 mice had significantly higher levels of intracellular calcium within somata (Fig. 1E–G; p = 0.0109; parametric t test), primary processes (Fig. 1F; p = 0.0172; Mann–Whitney), and capillary endfeet (Fig. 1G; p = 0.0116; parametric t test) compared with age-matched nontransgenic mice. We sought to identify any potential pathologic driver(s) of cell-wide intracellular calcium dysfunction throughout cortical astrocytes within the brain of awake APP/PS1 mice. We first tested whether astrocytes in close proximity to amyloid plaques were more vulnerable to an elevated cell-wide intracellular calcium within the awake brain of APP/PS1 mice. We manually measured the distance from the somata, primary processes, and capillary endfeet of 799 cortical astrocytes to the edge of the nearest amyloid plaque in three dimensions in the brain of 10 awake APP/PS1 mice (Fig. 1H–J). There were 179 of 799 (22%) cortical astrocytic somata, 356 of 1597 (22%) primary processes, and 194 of 1024 (19%) ROIs at capillary endfeet within 50 µm of nearest amyloid plaque in the brain of awake APP/PS1 mice. There was no significant difference in the intracellular calcium within astrocytic somata (Fig. 1H; p = 0.6845), primary processes (Fig. 1I; p = 0.4987), and capillary endfeet (Fig. 1J; p = 0.3724) in close proximity to amyloid plaques compared with astrocytes quantified at a distance >50 µm from amyloid plaques the brain of awake APP/PS1 mice. gfa2.yc3.6-expressing astrocytes did not qualitatively show any obvious morphologic changes consistent with astrocytic reactivity when observed in close proximity to amyloid plaques in vivo. Our findings suggest the cell-wide elevation in intracellular calcium throughout astrocytes is not strongly influenced by distance to amyloid plaques within the somatosensory cortex of awake APP/PS1 mice.
Cell-wide astrocytic intracellular calcium dysregulation unrelated to distance to amyloid plaques in awake brain of APP/PS1 mice. A, Maximum projection images showing network of gfa2.yc3.6-expressing astrocytes (green) and cerebral vasculature (red) in the brain of an awake nontransgenic mouse. Corresponding pseudo-colored image showing the YFP/CFP ratio of each astrocyte as low (blue) to high (red; B) and age-matched APP/PS1 transgenic mouse, with gfa2.yc3.6-expressing astrocytes, vasculature, and methoxy-labeled amyloid plaque (C, cyan; corresponding pseudo-colored image, D). Astrocytic intracellular calcium (YFP/CFP ratios) is significantly elevated within astrocytic somata (E; p = 0.0109; parametric t test), astrocytic primary processes (F; p = 0.0172; Mann–Whitney test), and capillary endfeet (G; p = 0.0116; parametric t test) in awake brain of APP/PS1 mice (n = 10) compared with awake brain of nontransgenic mice (n = 9). The cell-wide intracellular calcium dysregulation within astrocytic somata (p = 0.6845; H), primary processes (p = 0.4987; I), and capillary endfeet (p = 0.3724; J) was not influenced by distance to amyloid plaques in the brain of awake APP/PS1 mice. Data are mean ± SEM. Scale bar, 25 µm.
gfa2.yc3.6-expressing astrocytes colocalize with SR-101-positive astrocytes within living mouse brain and immunohistochemical astrocyte markers glutamine synthetase and GFAP in mouse brain sections. At the top, AAV2/5.gfa2.yc3.6-expressing astrocytes (A) and Sulforhodamine 101 (SR-101) fluorescently labeled astrocytes (B) colocalized within the living cortex of a 6-month-oldÜ57BL/6J mouse in vivo (C). At the bottom, Immunohistochemical staining of gfa2.yc3.6-expressing astrocytes within 8-month-old C57BL/6J mouse brain sections shows colocalization with either glutamine synthetase-labeled astrocytes (D–F) or GFAP-labeled astrocytes (G–I), but not NeuN-positive neurons (A–C). Scale bars: 25 μm, 50 μm.
We sought to identify the potential driver(s) of elevated astrocytic intracellular calcium within the amyloid plaque environment. Accordingly, we cultured low molecular weight soluble Abeta assemblies from transgenic cortical neurons. We exposed gfa2.yc3.6-expressing cortical astrocytes, in the absence of amyloid plaques, in the brain of young C57BL/6J mice to media collected from cultured neurons from ntgCM, TgCM, or immunodepleted-TgCM mice. Cortical astrocytes exposed to TgCM exhibited a cell-wide elevation in intracellular calcium throughout astrocytic somata, primary processes, and astrocytic endfeet (Fig. 3). The relative change in astrocytic YFP/CFP ratio, before and following topical application of ntgCM, TgCM, or immunodepleted-TgCM, was calculated and statistically compared (Fig. 3D). Additionally, calcium overload was defined previously as YFP/CFP ratio ≥2 SDs greater than YFP/CFP ratio mean of nontransgenic mice (Kuchibhotla et al., 2008). The topical application of TgCM onto the living brain of young C57BL/6J evoked a cell-wide intracellular calcium overload within cortical astrocytes (Fig. 3C,E). We used longitudinal in vivo imaging to repeatedly image gfa2.yc3.6-expressing astrocytes with calcium overload in the brain of awake APP/PS1 mice. Longitudinal in vivo repeated imaging of cortical astrocytes with calcium overload in the brain of awake APP/PS1 mice showed astrocytes with calcium overload persisted chronically and robustly over months (Fig. 4). Furthermore, application of TgCM also evoked an increase in intracellular calcium in astroglial WJE cells in vitro (Fig. 5), in absence of neurons or other cell types. These data suggest that the observed increase in astroglial calcium due occurs in part through the direct effect of soluble Abeta into astrocytes. Together, these findings suggest astrocytic cell-wide elevated intracellular calcium is chronic, in the absence of amyloid plaques, by acute exposure to soluble Abeta. These findings suggest that the elevated intracellular calcium in astrocytes is pathologically involved in the soluble-Abeta mediated neurodegenerative processes in vivo.
Neuronally derived soluble Abeta evokes a cell-wide calcium overload in cortical astrocytes, in the absence of amyloid plaques in the brain of young C57BL/6J mice, during anesthesia by isoflurane. A, Schematic representation of the experimental procedure to determine the effects of neuronally derived soluble Abeta on astroglial resting calcium in the healthy mouse brain in vivo. B, Representative in vivo images of gfa2.yc3.6 expressing astrocytes (green) and fluorescently labeled cerebral blood vessels (red) in brain of young C57BL/6J. Astrocytes within white box with dotted lines shown at higher magnification for baseline in vivo imaging and at 1 h following topical application of ntgCM, TgCM, or Aβ-immunodepleted transgenic conditioned media. C, Histograms of astrocytic calcium frequency distribution (YFP/CFP ratio) for the three conditions before (basal) and after application of CM (1 h). The percentage of astrocytic cellular compartments exceeding the threshold for calcium overload (black dotted line; 2 SDs greater than the YFP/CFP mean determined for the baseline) is noted on the graphs. D, Relative increase in astroglial [Ca2+] (ΔR/R0) after conditioned media application for each volume acquired. E, Percentage of calcium overload (YFP/CFP ratios .2 SDs above mean of YFP/CFP ratios for nontransgenic mice) before and after application of conditioned media for every mouse analyzed. Acute exposure to TgCM significantly increased calcium overload within astrocytic somata (from 2.2% to 11.8%, p < 0.001), primary processes (from 1.9% to 7.7%, p < 0.05), and astrocytic endfeet (from 1.8% to 8.0%, p < 0.05). Error bars indicate mean ± SEM. ***p < 0.001. **p < 0.005. n = 5-7 mice per group. Scale bars: 20 µm; insets, 5 µm.
Chronic astrocytic calcium overload in awake brain of APP/PS1 mice. Representative pseudo-colored in vivo image showing high YFP/CFP ratio (red) in gfa2.yc3.6-expressing astrocyte (white arrowhead) adjacent to a methoxy-X04-positive amyloid plaque (blue; A) longitudinally imaged in brain of awake APP/PS1 mice each month (B,C) revealing chronic elevated calcium within robust cortical astrocytes.
Neuronally derived soluble Abeta increases astrocytic intracellular calcium in vitro. A, Representative images of astrocytic cell line WJE before and following acute 1 h exposure to either ntgCM cultured from neurons from nontransgenic mice or TgCM cultured from neurons from Tg2576 transgenic mice. B, Quantification of YFP/CFP ratios for every WJE cell before and after either nTgCM or TgCM. C, Averaged YFP/CFP ratios before (basal) and after acute exposure to nTgCM or TgCM per experiment. ****p < 0.0001; **p < 0.005; two-tailed t test. n = 3 or 4 experiments. Scale bar, 50 μm.
Hyperactive calcium signaling within cortical astrocytes occurs concomitantly with hypoactive capillary endfeet in awake brain of APP/PS1 transgenic mice
We quantified all endogenous calcium signaling events occurring spontaneously within high resolution time-lapse imaging of 129 gfa2.yc3.6-expressing astrocytes (somata, primary processes, fine processes, and capillary endfeet) in 10 awake APP/PS1 mice and 9 awake nontransgenic mice. In contrast to astrocytic intracellular calcium (“resting calcium”), which we quantified with similar levels within somata and corresponding primary processes (Fig. 6), the spontaneous calcium signaling is highly compartmentalized in astrocytes within the awake mouse brain (Fig. 7). Compartmentalized spontaneous calcium signaling was observed occurring solely within capillary endfeet (Fig. 7A–C), individual primary processes (Fig. 7D-F), or astrocytic fine processes (Fig. 7G–J) independently of the rest of the astrocyte in the brain of awake nontransgenic mice and APP/PS1 mice. We also observed the compartmentalized nature of spontaneous calcium “waves” as they traveled along capillary endfeet, spread into an adjacent astrocytic somata and across the cortex (for a total ∼255 μm over 7-8 s; ∼33 μm/s; Movie 1) in awake brain of nontransgenic mouse (Fig. 7C, black arrowheads). Interestingly, capillary endfeet were found to be as active as astrocytic fine processes in the awake brain of nontransgenic mice (compare Fig. 7K and Fig. 7M), but capillary endfeet are hypoactive in APP/PS1 mice (reduced from 0.2-1.2 events/minute in nontransgenic mice to 0.2-0.4 events/minute in awake APP/PS1 mice; Fig. 7K). Unlike capillary endfeet, spontaneous calcium signaling occurring within individual astrocytic primary processes did not become hypoactive (primary processes in nontransgenic mice 0.2-0.6 events/minute compared with 0.2-0.8 events/minute in APP/PS1 mice; Fig. 7L). Similarly, the compartmentalized calcium signaling within astrocytic fine processes retained their activity and showed a trend toward hyperactivity in the brain of awake APP/PS1 mice (0.2-1.8 events/minute) compared with 0.2-1.4 events/minute in nontransgenic mice (Fig. 7M). Together, our findings highlight capillary endfeet as the site of very active spontaneous calcium signaling and involved in spontaneous calcium waves but pathologically hypoactive while the rest of the astrocyte is becoming more hyperactive in the presence of amyloid plaques in the brain of awake APP/PS1 mice. We sought to determine whether amyloid plaques exacerbated hypoactivity within the cortical astrocytic network (Fig. 8). However, we found a similar proportion of cortical astrocytes exhibiting cell-wide quiescence within the brain of awake APP/PS1 mice versus nontransgenic mice. Spontaneous calcium signaling is strongly suppressed by use of general anesthesia, and we found (2% isoflurane) largely suppressed the compartmentalized spontaneous calcium signaling within capillary endfeet, primary processes, and fine processes in the living mouse brain (Fig. 9). Use of isoflurane anesthesia during the in vivo study of conditioned media on astrocytic calcium strongly suppressed spontaneous calcium signaling. Accordingly, we examined the effects of conditioned media on astrocytic intracellular calcium, which is not compartmentalized in the living mouse brain. Endogenous calcium signaling is best studied in the brain of awake mice. Interestingly, we observed that animal movement was sufficient to elicit an immediate visible movement in the imaging plane accompanied by a global synchronized increase in astrocytic calcium signaling throughout the somatosensory cortex in the brain of awake nontransgenic mice and APP/PS1 mice (Fig. 10). Each animal in our study could walk, groom, and move as much as desired on a passive treadmill while head-fixed during in vivo imaging. We separated all mouse movement-associated global calcium signaling events from our analyses and found a similar number of movement events between groups (Fig. 10C). The amplitude of movement-associated calcium signaling events was not significantly different between groups (Fig. 10D; p = 0.6983; two-way ANOVA). Interestingly, the duration of movement-induced events is impacted by hyperactivity within fine processes in the brain of awake APP/PS1 mice (Fig. 10E; p = 0.0010; two-way ANOVA). In addition to compartmentalized calcium signaling within astrocytic cellular compartments (Fig. 4) and movement-associated events (Fig. 10), we also quantified all spontaneous “multicompartmental” calcium signaling events occurring during acquisition of our in vivo time-lapse imaging of cortical astrocytes within the cortical network in the awake mouse brain (Fig. 11). Multicompartmental events occurred spontaneously and simultaneously occupied multiple cellular compartments within the awake mouse brain (Fig. 11A-C). The frequency of multicompartmental events was not different in APP/PS1 mice compared with nontransgenic mice (0.2-0.6 events/minute in both groups; Fig. 11D). Multicompartmental events occurred much less frequently than compartmentalized events within individual cellular compartments and may therefore need to be considered as a separate calcium signaling entity. Somatic calcium dysregulation within cortical astrocytes of APP/PS1 mice offset somatic calcium signaling during multicompartmental events resulting in ∼64% increase in amplitude in APP/PS1 mice (Fig. 11E; p = 0.0066; two-way ANOVA). Similar to tight regulation of calcium signaling duration, lasting ∼10 s during movement, we also found that multicompartmental events were also tightly regulated and lasted ∼8-9 s within somata, primary processes, and endfeet of nontransgenic mice but ∼13 s within fine processes (Fig. 11F). Similarly, multicompartmental calcium signaling lasted ∼8-9 s within somata, primary processes, and endfeet of awake APP/PS1 mice but were increased to ∼18 s within fine processes (Fig. 11F; p = 0.0089; two-way ANOVA). To determine whether multicompartmental calcium signaling events may be neuronally driven via fine processes, somatic derived or emanating from the vasculature via astrocytic endfeet, we quantified their first appearance within cortical astrocytes in the brain of awake nontransgenic mice compared with awake APP/PS1 mice (Fig. 11G,H). Spontaneous multicompartmental calcium signaling events can first appear within any of the astrocytic cellular compartments (somata, primary processes, capillary endfeet, and fine processes); but in the presence of amyloid plaques, they are more likely to emanate from astrocytic primary processes (Fig. 11H). This may be a result of increased hyperactivity within astrocytic primary processes in the brain of awake APP/PS1 mice compared with nontransgenic mice (Fig. 12). We observed the proportion of spontaneous calcium signaling occurring within fine processes increased from 12% in nontransgenic mice to 22% in awake APP/PS1 mice (Fig. 12A,B). Concomitantly, compartmentalized calcium signaling within capillary endfeet was hypoactive and decreased from 28% in nontransgenic mice to only 10% in APP/PS1 mice (Fig. 12A,B). We also observed the proportion of spontaneous calcium signaling occurring within capillary endfeet in nontransgenic mice (28%) was similar to endogenous compartmentalized calcium signaling within fine processes (33%) of nontransgenic mice, while only 12% calcium signaling was compartmentalized within primary processes (Fig. 12A). This further supports that spontaneous calcium signaling within capillary endfeet is as active as astrocytic fine processes, but capillary endfeet become hypoactive in the presence of amyloid plaques while fine processes and primary processes are showing hyperactivity. Next, we quantified all spontaneous intercellular calcium signaling events spreading throughout all cellular compartments of neighboring cortical astrocytes in awake mouse brain (Fig. 13A–C). We observed multiple intercellular calcium “waves” occurring at a different time during our 300 s of in vivo time-lapse acquisition in the brain of awake nontransgenic mice (Fig. 13D,E) and awake APP/PS1 transgenic mice (Fig. 13F,G). We traced the trajectory of propagation of these spontaneous multicellular calcium waves (white numbers indicate position in astrocyte and the spread of calcium wave in numerical sequence across neighboring astrocytes; Fig. 13D–G). The second spontaneous calcium wave occurring during 5 min acquisition can first appear at a different astrocytic cellular compartment than the first spontaneous calcium wave and follow a different trajectory in a seemingly concentric wave in the brain of nontransgenic and APP/PS1 transgenic mice (Fig. 13D–G). These spontaneous concentric calcium waves were hyperactive within astrocytic somata and primary processes in the brain of awake APP/PS1 mice (Fig. 13H; p = 0.01; two-way ANOVA). Together, our findings collectively support a cell-wide calcium dysregulation involving hyperactive calcium signaling within somata, primary processes, and fine processes affecting a complex hierarchy of local and global calcium signaling arising spontaneously within the awake brain of APP/PS1 mice but occurring with hypoactivity within capillary endfeet.
Astrocytic intracellular “resting” calcium not compartmentalized throughout somata and primary processes in somatosensory cortex of awake mice. Astrocytic intracellular calcium within somata is strongly correlated with the average intracellular calcium of all corresponding primary processes of the cell in the somatosensory cortex of awake 12- to 17-month-old nontransgenic mice (96 astrocytes; n = 2 mice; r = 0.7905, p < 0.0001; Pearson correlation coefficient two-tailed) and awake APP/PS1 mice (249 astrocytes; n = 5 mice; r = 0.7110, p < 0.0001; nonparametric Spearman correlation two-tailed).
Compartmentalized spontaneous calcium signaling hypoactive within capillary endfeet alongside astrocytic hyperactivity in awake APP/PS1 mice. A–I, Representative time-lapse in vivo images of gfa2.yc3.6-expressing astrocytes (green), vasculature (red), and amyloid plaques (cyan) acquired within the somatosensory cortex of awake mice. Aligned image, using custom-written MATLAB script, demonstrating positioning of manually drawn ROIs (B,E,H) and corresponding spatiotemporal dynamics of all endogenous compartmentalized calcium signaling events occurring within ROIs during 300 s time-lapse indicated by black arrows (C,F,J). Spontaneous calcium wave traveling along capillary endfeet at ∼33 μm/s and spreading into adjacent astrocytic somata and across the cortex in concentric wave (C, black arrowheads; Movie 1). Qualitative pseudo-colored time-lapse in vivo image showing high YFP/CFP ratio occurring solely within astrocytic fine processes (I) and corresponding calcium signaling event indicated by black arrow (J). K, Compartmentalized calcium signaling is very active within capillary endfeet (0.2-1.2 events/min) in awake nontransgenic mice but hypoactive within the brain of awake APP/PS1 mice (0.2-0.4 events/min). L, Endogenous calcium signaling compartmentalized within astrocytic primary processes is less active than capillary endfeet in awake nontransgenic mice (0.2-0.6 events/min) and awake APP/PS1 mice (0.2-0.8 events/min). M, Spontaneous calcium signaling compartmentalized within astrocytic fine processes is very active within nontransgenic mice (0.2-1.4 events/min) and APP/PS1 mice (0.2-1.8 events/min). Scale bar, 20 µm.
Amyloid plaque environment did not appear to influence cell-wide quiescence in cortical astrocytes in brain of awake APP/PS1 mice. Representative in vivo time-lapse image of gfa2.yc3.6-expressing astrocyte within awake brain of nontransgenic mouse (A) and manually drawn ROIs for quantification of spontaneous calcium signaling throughout entire cell (B). C, An example of cell-wide quiescence with no spontaneous calcium events occurring in any of the cellular compartments during 300 s time-lapse. D, In vivo time-lapse image showing quiescent cortical astrocyte in awake brain of APP/PS1 mouse, with no spontaneous calcium events occurring within any of the manually drawn ROIs (E) over 300 s time-lapse (F). G, Approximately 37% of cortical astrocytes within the somatosensory cortex of nontransgenic mice show cell-wide quiescence over 300 s compared with 41% cortical astrocytes with cell-wide quiescence in awake APP/PS1 mice. Scale bar, 20 µm.
Use of isoflurane (2%) during in vivo multiphoton imaging largely suppressed all types of compartmentalized spontaneous calcium events within cortical astrocytic processes in the living brain of 4- to 6-month-old C57BL/6J mice. A, Isoflurane use silenced compartmentalized spontaneous calcium events within 23 of 25 ROIs at endfeet at 19 cerebral capillaries in the brain of C57BL/6 mice. Compartmentalized spontaneous calcium events were also silenced within 28 of 30 primary processes (B) and fine processes of 21 cortical astrocytes (C) in the anesthetized brain of C57BL/6 mice (n = 6 mice).
Mouse movement-associated global calcium signaling occurring with hyperactivity within astrocytic fine processes in awake APP/PS1 mice. Representative in vivo time-lapse images of gfa2.yc3.6-expressing astrocytes (green) and vasculature (red) within the cortical network of awake nontransgenic mice (A) and APP/PS1 transgenic mice (B). C, Mice were habituated to being awake during in vivo multiphoton imaging, and movement events associated with global astrocytic calcium signaling were similar between groups. D, Global astrocytic calcium signaling associated with mouse movement occurred with similar amplitude throughout all astrocytic cellular compartments in awake APP/PS1 mice and nontransgenic mice (p = 0.6983; two-way ANOVA). E, Astrocytic calcium signaling within fine processes was hyperactive during movement-induced calcium signaling in awake APP/PS1 mice compared with nontransgenic mice (p = 0.0010; two-way ANOVA). Data are mean ± SEM. Scale bar, 50 µm.
Astrocytic calcium signaling involving multiple cellular compartments can emanate from any cellular compartment and appear hyperactive in awake APP/PS1 mice. A, Representative in vivo time-lapse images showing gfa2.yc3.6-expressing astrocyte (green) and fluorescently labeled vasculature (red) in awake mouse brain. Manually drawn ROIs (B) and corresponding calcium signaling for each ROI show spontaneous calcium signaling occurring simultaneously within somata and primary processes (C). D, The number of multicompartmental events occurring per minute (0.2-0.6 events/s) was not different between groups. E, There was significantly greater amplitude of calcium signaling within astrocytic somata in APP/PS1 mice compared with nontransgenic mice (p = 0.0066; two-way ANOVA). F, Duration of calcium signaling was significantly greater within astrocytic fine processes in APP/PS1 mice compared with nontransgenic mice (p = 0.0089; two-way ANOVA). Multicompartmental calcium signaling events in astrocytes can first appear within any cellular compartment (G) but most commonly occur within primary processes in awake APP/PS1 mouse brain (H). Data are mean ± SEM. Scale bar, 20 µm.
Hyperactive spontaneous intracellular calcium activity within astrocytic processes and concomitant hypoactivity within astrocytic endfeet in the amyloid plaque environment in the brain of awake APP/PS1 mice. Proportional segments of 98 spontaneous intracellular calcium events occurring throughout astrocytes within brain of awake WT mice (A) and 111 spontaneous intracellular calcium events occurring throughout astrocytes within brain of awake APP/PS1 mice (B). Compartmentalized calcium activity within capillary endfeet (28%) as active as fine processes (33%) in brain of awake nontransgenic mice (A) but reduced to 10% in brain of APP/PS1 mice while spontaneous calcium signaling within primary processes show hyperactivity (B).
Quantification of spontaneous “multicellular” calcium events occurring throughout cortical astrocytes within the brain of awake mice. A, Two cortical astrocytes adjacent to fluorescently labeled small blood vessels (red) within the awake brain of nontransgenic mouse. B, Aligned time-lapse image with manually drawn ROIs. C, YFP/CFP ratio traces, for 700 frames and 300 s, showing two spontaneous multicellular calcium events that occupy all cellular compartments of both astrocytes in A, B. D–G, Spatiotemporal tracking of spontaneous multicellular events, with sequence of event propagation highlighted by numerical order, showing heterogeneity between first and second occurrence of a multicellular event during 300 s of time-lapse acquisition within the awake mouse brain. H, Spontaneous multicellular calcium events occur with greater amplitude within astrocytic somata and primary processes in cortical astrocytes within the awake brain of APP/PS1 mice compared with nontransgenic mice (p = 0.01; two-way ANOVA). Data are mean ± SEM. Scale bar, 20 µm.
Spontaneous intercellular calcium wave propagating throughout astrocytic endfeet, somata, and across cortex in a concentric manner at ∼33 μm/s in the awake mouse brain in vivo. Pseudo-colored in vivo time-lapse acquisition of 700 frames for 300 s (∼2.3 Hz) showing two cortical astrocytes with endfoot processes adjacent to fluorescently labeled capillaries in the awake brain of a 12- to 17-month-old nontransgenic mouse. A compartmentalized calcium signaling event (red) occupying a portion of capillary endfeet is evident at 6 s. Additionally, an intercellular calcium wave (red) is evident at 10 s that travels along capillary endfeet and enters astrocytic somata before traveling across the cortex in a concentric manner within the awake brain of a nontransgenic mouse in vivo.
Discussion
The insidious multidecadal accumulation of insoluble and soluble Abeta in the brain is one of the most significant pathologies in the etiology of AD (Selkoe, 2002). Previous in vivo multiphoton imaging of anesthetized APP/PS1 mice, using the bulk loading of commonly used calcium indicator dyes, revealed widespread astrocytic hyperactivity unrelated to distance to amyloid plaques in vivo (Kuchibhotla et al., 2009). To quantify intracellular calcium throughout astrocytic somata, primary processes, fine processes, and capillary endfeet in the presence of amyloid plaques in vivo, which is not usually possible with small molecular calcium indicators, we expressed our genetically encoded calcium indicator, gfa2.yc3.6, throughout cortical astrocytes in the brain of awake APP/PS1 mice. We report cortical astrocytes have a cell-wide elevation in astrocytic intracellular calcium throughout somata, primary processes, and capillary endfeet that is unrelated to distance to amyloid plaques in the awake mouse brain. We cultured bioactive soluble Abeta, from neurons from a transgenic mouse model of AD, at a concentration corresponding to levels in the CSF of patients with sporadic AD (Mehta et al., 2000; Arbel-Ornath et al., 2017). Synthetic preparations of soluble Abeta, at micromolar concentrations, can form assemblies with biophysical properties different from a naturally derived preparation (Selkoe, 2002). Soluble Abeta assemblies are cytotoxic and can (in the absence of amyloid plaques) initiate neurodegenerative processes leading to synaptic loss and memory deficits in vivo (Shankar et al., 2008; Wu et al., 2010; Arbel-Ornath et al., 2017; Yang et al., 2017). Topical application of neuronally derived soluble Abeta onto the living brain of C57BL/6J mice (in the absence of amyloid plaques) disrupted neuronal network(s) and was directly associated with advanced morphologic changes (Arbel-Ornath et al., 2017; Calvo-Rodriguez et al., 2020). Here, we found that acutely exposing cortical astrocytes to neuronally derived soluble Abeta, in the absence of amyloid plaques, evoked a transient cell-wide elevation in intracellular calcium in the living brain of C57BL/6J mice. Our findings demonstrate a direct pathophysiological effect of soluble-Abeta on astrocytes in vivo. Therefore, soluble Abeta disrupts neuronal and astrocytic networks concomitantly possibly highlighting a role for astrocytes in soluble-Abeta mediated neurodegeneration. Inhibiting astrocytic calcium/calcineurin/NFAT pathway in mixed hippocampal cultures (85%-90% astrocytes and neurons) exposed to soluble Abeta protected neurons from cell death (Abdul et al., 2009), further supporting a pathophysiological role for astrocytes in soluble-Abeta mediated neurodegeneration. We quantified the effect of soluble Abeta on astrocytic intracellular resting calcium, an important determinant for the generation of spontaneous and evoked calcium signaling in astrocytes in vivo (King et al., 2020). Additional studies are necessary to determine the effects of soluble Abeta on spontaneous astrocytic calcium signaling in the living brain of C57BL/6J mice as the use of general anesthesia in our present study of conditioned media greatly suppressed spontaneous calcium signaling in astrocytes in vivo. Although astrocytes in the human brain are ∼2.6-fold larger with ∼10 fold more primary processes than mouse astrocytes, they do share important similarities that support the study of mouse astrocytes as a sufficient model (Oberheim et al., 2009). Interestingly, both human astrocytes and mouse astrocytes can exhibit calcium signaling and intercellular calcium waves during the pharmacological silencing of local neuronal activity, thus highlighting astrocytic calcium signaling as not solely secondary to neuronal activity (Nedergaard, 1994; Nett et al., 2002; Kuchibhotla et al., 2009; Oberheim et al., 2009; Di Castro et al., 2011; Thrane et al., 2012; Navarrete et al., 2013; Delekate et al., 2014). Therefore, astrocytic calcium signaling dysregulation may be an important pathologic driver of soluble-Abeta neurodegeneration in the living mouse brain. Interestingly, our longitudinal in vivo multiphoton imaging revealed cortical astrocytes with elevated intracellular calcium (“calcium overload”) remained within the neurovascular unit of awake APP/PS1 mice for months, despite persisting calcium overload. This suggests astrocytic elevated calcium as a chronic phenotype; and, unlike neurons, cortical astrocytes may not be particularly vulnerable to the degeneration associated with calcium dysregulation (Kuchibhotla et al., 2008). Indeed, APP/PS1 transgenic mice with amyloid plaques have a similar number of astrocytes as nontransgenic mice, and an unbiased stereological analyses of human AD brain showed a similar number of astrocytes as age-matched nondemented human brain (Serrano-Pozo et al., 2013; Galea et al., 2015). Together, these findings suggest that the cell-wide intracellular calcium in cortical astrocytes may be a chronic phenotype, independent of amyloid plaques, and mediated, at least in part, by neuronally derived soluble Abeta. Surprisingly, our detailed quantification of spontaneous calcium signaling occurring solely within astrocytic cellular processes in awake mice revealed that capillary endfeet calcium signaling is similarly as active as fine processes in nontransgenic mice. This may be because astrocytes play an important role in neurovascular coupling, in which local increases in neuronal activity require local increases in blood flow to meet an increased metabolic demand (Attwell et al., 2010; Bazargani and Attwell, 2016). Cerebral capillaries contribute to >80% of blood flow, differ mechanistically from arterioles, and may respond to sensory stimulation before arterioles (Hall et al., 2014; Biesecker et al., 2016; Mishra et al., 2016). Interestingly, the buffering of astrocytic intracellular calcium reduced neuronal stimulation-evoked capillary dilation in brain slices (Mishra et al., 2016). APP/PS1 transgenic mice show deficits in neurovascular coupling and perivascular clearance in vivo, and these functional vascular deficits were not considered to be neuronally driven (Van Veluw et al., 2020). Further work is required to decode the complex hierarchy of astrocytic calcium signaling events into physiological function in health and AD. It is currently unclear whether the ∼50% functional deficit we observed in capillary endfeet calcium signaling in the awake brain of APP/PS1 mice plays a pathologic role in neurovascular coupling deficits and impacts global intercellular calcium waves. Global astrocytic calcium signaling associated with mouse movement did not occur with functional deficits in capillary endfeet but hyperactivity within fine processes in awake APP/PS1 mice. Astrocytic hyperactivity was also observed during multicompartmental and multicellular events in which somatic calcium dysregulation may have offset the amplitude of intracellular calcium signaling. Together, here we observed astrocytic calcium hyperactivity within somata, primary processes, and fine processes in the amyloid plaque environment but, unlike the rest of the astrocyte, capillary endfeet are hypoactive. Therefore, astrocytic-targeting therapies may need a combination therapeutic strategy to enhance capillary endfeet calcium signaling while concomitantly reducing astrocytic hyperactive calcium signaling. Additionally, as astrocytic calcium dysregulation may have an important role in soluble-Abeta mediated neurodegeneration, the therapeutic restoration of astrocytic cell-wide calcium physiology may protect neuronal structure and function from soluble Abeta-mediated neurodegeneration.
Footnotes
This work was supported by National Institutes of Health 5R01AG054598-02.
The authors declare no competing financial interests.
- Correspondence should be addressed to Brian J. Bacskai at bbacskai{at}mgh.harvard.edu
