Astrocytes are an abundant non-neuronal cellular type in the brain [1
]. They exert fundamental housekeeping and homeostatic functions in the central nervous system (CNS) and are also involved in the pathogenesis of many neurological diseases. Astrocytes are non-excitable cells as they are largely unable to generate action potentials in response to electrical or chemical stimulation. Conversely, astrocytes respond to extracellular stimuli by generating intracellular calcium signals by exploiting two main mechanisms: (i) activation of metabotropic receptors on the plasma membrane, leading to liberation of calcium ions from internal calcium stores; and (ii) a receptor/store-operated mechanism of calcium entry from the extracellular milieu through the plasma membrane [1
]. Astroglial calcium signals are thought to have a number of implications for CNS pathophysiology, including modulation of synaptic release [2
], synchronization of neuronal activity [3
], regulation of frequency of spontaneous α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid AMPA receptor currents [4
] and participation in vesicular glutamate release [5
]. Although the physiological role of calcium signals in astrocytes is still a matter of debate [6
], there is broad consensus about their role and their alterations in brain pathology [9
Calcineurin (CaN) is a calcium/calmodulin-activated serine-threonine phosphatase, which is highly expressed in the brain [16
]. In neurons, CaN regulates neuronal excitability and synaptic transmission [17
]. Moreover, CaN activation is associated with long-term depression [18
], while CaN inactivation is required for establishment of aversive memory [19
]. In astrocytes, calcineurin is principally involved in setting up reactive gliosis and neuroinflammation in a number of neuropathological conditions [20
]. Conversely, activation of CaN in astrocytes has so far not been documented under physiological conditions and its role in housekeeping and homeostatic functions of astrocytes is currently unknown.
The present work is designed as an in vitro proof of principle of astroglial CaN activation in response to neuronal activity. We show that long-term potentiation (LTP)-like neuronal activity robustly activates CaN in adjacent astrocytes and that store-operated Ca2+ entry through the astroglial plasma membrane is required for this process to occur.
In this report, we provide an in vitro proof of principle of activation of astroglial CaN by neuronal activation. The main findings are: (1) in mixed neuron-astroglial hippocampal primary cultures, cLTP induction protocol, which specifically stimulates neuronal activity, induced intracellular Ca2+ signals and robust CaN activation in astrocytes, and (2) astroglial Ca2+ signals and CaN activation required extracellular Ca2+ entry via the SOCE mechanism.
Although, to our knowledge, there is no data published to date that neuronal activity may result in CaN activation in astrocytes, it has been reported that the increase in neuronal activity is able to induce CaN activation and nuclear translocation of NFAT in pericytes in cortical slices [34
]. This landmark contribution suggests that neuronal activity may, in fact, activate CaN in non-neuronal cells. Now, we demonstrate that, in an in vitro setting, CaN may be activated also in astrocytes.
Calcineurin is activated by a specific pattern of calcium signaling which is characterized by low and sustained (minutes) elevations of baseline cytosolic calcium levels [35
], while the specificity of downstream CaN targets activation can be further achieved by a specific temporal pattern of Ca2+
]. Stimulation of neuronal activity is known to produce calcium signals in astrocytes in vivo [37
], including awake animals [46
], in brain slice preparations [34
] and in mixed neuron-astrocyte primary cultures [52
]. Most of these signals have been registered as short single or oscillatory transients with duration from several milliseconds to seconds [38
]. Some of them, however, lasted long enough (tens of seconds to minutes) [41
] to speculate that they would be sufficient for CaN activation. Recent experiments employing fast 3D calcium imaging suggest that the spatio-temporal pattern of Ca2+
signals in astrocytes is extremely complex, and depends on the nature of Ca2+
-related receptors/channels expressed in a particular subdomain of astroglial plasma membrane [55
]. Accordingly, it can be speculated that localized CaN activation may be necessary to achieve spatial specificity of processes controlled by the astrocytes. Further experiments are needed to demonstrate and characterize in vivo activity-dependent CaN activation in astrocytes.
Our findings suggest that neuronal activity induces Ca2+
entry in astrocytes via SOCE mechanism. SOCE is one of the fundamental mechanisms of astroglial Ca2+
] and is involved in the generation of Ca2+
oscillations, refilling of the endoplasmic reticulum with Ca2+
], astroglial cytokine production [58
] and astroglial metabolism [59
]. Spontaneous Ca2+
oscillations in vivo in fine astroglial processes were shown to involve Ca2+
entry through the plasma membrane, probably through store-operated channels [60
]. Concerning pathological conditions, SOCE is involved in the invasion of human glioblastoma [62
], and is augmented in primary [30
] astrocytes from an AD mouse model. Our results provide the physiological rationale for SOCE activation in astrocytes by neuronal activity. We also show that ionotropic NMDA and metabotropic mGluR5 receptors are involved in SOCE generation by LTP-like neuronal activity. Although it was impossible to discriminate the cell type on which NMDA or mGluR5 reside in the current setting, our previous observations [29
] and direct stimulation of mixed cultures either with NMDA or DHPG (Figure 6
), suggest that NMDAR are expressed in neurons while mGluR5 is located in astrocytes. Furthermore, cultured astrocytes are somewhat more sensitive than neurons to DHPG, as 20 μM DHPG is enough to induce Ca2+
increase in astrocytes, but not in neurons [29
] (Figure 6
), while 200 μM DHPG were used to elicit mGluR5-dependent Ca2+
transients in neurons [63
]. Last, astroglial mGluR5 receptors are required to sustain long lasting Ca2+
entry in cultures astrocytes that was efficiently blocked by SOCE inhibitors [30
SOCE is known to activate CaN/NFAT axis and modulate gene transcription in a number of cell types, including T-lymphocytes and mast cells [64
], cardiomyocytes [65
], skeletal muscle cells [66
] and in neural progenitor cells [67
]. We show now that SOCE is also required for neuronal activity-induced CaN activation in astrocytes in vitro while future experiments will show if SOCE activates CaN in astrocytes also in intact brain. Regarding the nature of SOCE channels, our pharmacological survey suggests that Orai1-containing channels are operative in astrocytes (efficiently inhibited by Orai1-blocking Pyr3 and Pyr6, [32
]), while participation of TRPC3 is questionable, since Pyr10, a specific TRPC3 inhibitor [32
]), was not as efficient as other SOCE blockers. In line with this, in our previous report, Pyr10 failed to inhibit the DHPG-induced after-peak Ca2+
entry, while it was efficiently inhibited by Pyr6 [30
In neurons, CaN is involved in long-term changes during neuronal plasticity, e.g., forebrain neuronal deletion of CaN specifically affects bidirectional synaptic plasticity and episodic-like working memory [18
], and inactivation of CaN is essential for the onset of transcriptional remodeling during long-term plasticity and memory formation [17
]. Our results suggest that also in astrocytes activity-induced CaN activation may be involved in long-term transcriptional remodeling leading to structural, biochemical and functional astroglial plasticity [68
]. Further studies are necessary to confirm this hypothesis.
The present report is a proof of principle in vitro study, and is not devoid of limitations, the two principal of which are: (1) the in vitro setup, which, obviously, only in part replicates the complex LTP phenomenon occurring in intact brain or even in brain slices; and (2) chemical instead of electrical LTP induction. We have consciously accepted these limitations for the following reasons. The in vitro setting proved to be simple and highly reproducible both in terms of lentiviral NY infection and in terms of cLTP and NY nuclear translocation. Thereby, it allowed us to rule out astrocytes as a primary target of cLTP protocol. Yet, we used hippocampal primary cultures and the effect may be different in cultures prepared from the other brain regions. Regarding cLTP induction, there are several protocols which are basically the modifications of two principal variants: the first consists in the application of a cocktail containing protein kinase A (PKA) activators, like forskolin and rolipram [28
], which recruits downstream targets of PKA-dependent phosphorylation involved in LTP induction. For obvious reasons, this protocol is not specific to neurons and would induce PKA-dependent phosphorylation also in astrocytes. The second variant consists in the relieve of GABA-dependent inhibition by blocking ionotropic GABA(A) receptors with bicuculline and strychnine and in facilitation of NMDA receptors with glycine and Mg2+
-free buffer [73
]. We have chosen this second protocol because it minimally interferes with astrocyte biochemistry and does not result in NY translocation in pure astroglial cultures. Relieve of GABA-dependent inhibition alone is known to increase neuronal firing rate [77
] and, notably, it was sufficient to induce nuclear translocation of NY and, hence, CaN activation, although significantly later than it was achieved by cLTP. Similar effect was achieved by Mg2+
withdrawal, which indicates that different strategies to facilitate neuronal activity may lead to CaN activation in astrocytes, although with different temporal pattern.
An important question which remains unanswered in this work is the nature of the mediator which is released by neurons to induce the astroglial response. Assuming that astrocytic mGluR5 receptors are involved in astrocyte activation, it is plausible to speculate that the messenger is glutamate released during neuronal activity. This has been demonstrated in vitro [78
] and in situ [79
]. However, in vivo it has been shown that mGluR5 is downregulated during postnatal development and no longer active in adult astrocytes [80
]. Therefore, the mechanisms of the activity-induced generation of Ca2+
signals in astrocytes, as well as of CaN activation, may differ between ex vivo and in vivo preparations, and may also depend on age at which the preparation is made.
In summary, we propose a model (Figure 7
) in which LTP-like neuronal activity, possibly through activation of NMDA receptors and glutamate release form neuronal terminals, induces Ca2+
entry through the plasma membrane, possibly implicating astrocytic mGluR5 receptors. The Ca2+
entry occurs through Orai1-containing SOCE channels and results in activation of CaN inside the astrocytes, which, in turn, leads to activation and nuclear translocation of CaN sensor NY. CaN activation in astrocytes may lead to transcriptional remodeling and long-term changes analogous to what occur during neuronal plasticity. In this model, we provide a framework for future investigation of astroglial CaN activation during neuronal activity and plasticity in physiology and pathology of CNS.
4. Materials and Methods
C57Bl/6 mice have been purchased from Charles River Laboratories (Calco (Lecco), Italy). The animals have been given food and water ad libitum; light/dark cycle has been automatically controlled in respect with the natural circadian rhythm and the temperature has been thermostatically regulated. Animals were managed in accordance with European directive 2010/63/UE and with Italian low D.l. 26/2014. The procedures were approved by the local animal-health and ethical committee (Università del Piemonte Orientale) and were authorized by the national authority (Istituto Superiore di Sanità; authorization number N. 22/2013, 3 February 2013).
4.2. Primary Hippocampal Mixed and Astroglial Cultures
To prepare mixed neuron-astroglial hippocampal primary cultures, new-born mice (less than one day old) were used. Pups were sacrificed by decapitation and hippocampi were dissected in cold HBSS (Sigma, Darmstadt, Germany, Cat. H6648). After dissection, hippocampi were digested in 0.1% Trypsin (Sigma, Cat. T4049) for 20 min in 37 °C water bath. Then, trypsin was neutralized by addition of Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma, Cat. D5671) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Life Technologies, Monza, Italy, Cat. 10270-106) and tissue was spun at 300× g for 5 min. Tissue pellet was resuspended in HBSS supplemented with 10% FBS and dissociated by 30 strokes of a 1000 μL automatic pipette. After pipetting, the tissues were left for 5 min to allow sedimentation of un-dissociated tissue, and cell suspension was transferred to a new tube and centrifuged at 250× g for 5 min. Hippocampal cells were resuspended in Neurobasal-A medium supplemented with 2% B-27 and with 2 mg/mL glutamine, 10 U/mL penicillin and 100 µg/mL streptomycin, and plated onto Poly-l-lysine-coated coverslips (0.1 mg/mL). For NY translocation time-course, the cells were plated onto 13 mm round coverslips in 24 well plates, 2 × 104 cells per well. For Fura-2 imaging, the cells were spotted (2 × 104 cells per spot) on 24 mm round coverslips in 6 well plates. Half of medium was changed every 5 days. Cultures were used for treatments and experiments after 12 days in vitro (DIV12). At DIV12-DIV14 the neuron/astrocyte ratio was 1.29 ± 0.49 as was counted in 34 coverslips from at least 8 independent cultures.
To prepare purified astroglial cultures, one to three-day old pups were sacrificed by decapitation, hippocampi were rapidly dissected and placed in cold HBSS. Tissue was digested in 0.25% trypsin (37 °C, 20 min), washed in complete culture medium ((DMEM), supplemented with 10% FBS, and with 2 mg/mL glutamine, 10 U/mL penicillin and 100 µg/mL streptomycin (all from Sigma) and resuspended in cold HBSS supplemented with 10% FBS. After 30 strokes of dissociation with an automatic pipette, cell suspension was centrifuged (250× g
, 5 min), pellet was resuspended in complete medium and plated in 100 mm Falcon culture dishes, pretreated with 0.1 mg/mL Poly-l
-lysine (Sigma). At sub-confluence (DIV5-10), cells were detached with trypsin and microglial cells were removed by magnetic-activated cell sorting (MACS) negative selection using anti-CD11b conjugated beads and MS magnetic columns (Miltenyi Biotech, Bologna, Italy, Cat. 130-093-634). After MACS, astrocytes were counted and plated for experiments as described above. Virtually no microglial cells have been detected in purified astroglial cultures after MACS procedure as was assessed by immunostaining with anti-Iba1 antibody (1:500, D.B.A., Segrate, Italy, Cat. 019-19741) [81
4.3. Lentivirus Production and Transduction
mCherry-H2Bc-ΔNFAT-EYFP-expressing pFUW (pFUW-NY) third generation lentiviral vector was a kind gift from Prof. Alexander Flügel, Göttingen [23
]. The production of infectious lentiviral particles was done using polyethylene glycol (PEG) method as describe elsewhere [82
]. Briefly, HEK293T cells were transfected with four plasmids (2.75 μg pMDLg.pRRE, 1 μg pRSV.Rev, 1.5 μg pMD2.VSVG, and 6 μg pFUW-NY per 100 mm dish). After 48–72 h medium was collected and viral particles were precipitated by adding 1/5 of 5× PEG solution (200 g PEG (Sigma, Cat. n. 89510), 12 g NaCl, 1 mL Tris 1 M, pH 7.5, in ddH2
O to a final volume of 500 mL) overnight. On the next day the precipitate was concentrated by centrifugation (2800× g
at 4 °C for 30 min) and resuspended in HBSS (1/100 of the original medium volume, aliquoted and stored at −80 °C. All manipulations with lentiviral vector were conducted in biosafety level 2 environment in accordance with Italian ministry of health-approved protocol.
DIV10-12 mixed neuron-astroglial cultures or 50% confluent purified astrocytes were transduced by adding (50 μL/mL) concentrated NY-expressing viral particles. Experiments were conducted 2–4 days after infection. For each condition and time-point, at least 2 coverslips were used (technical replicates) in three independent culture preparations (biological replicates).
4.4. Induction of cLTP and NY Translocation Quantification
cLTP was induced by a cLTP cocktail containing 20 μM bicuculline, 1 μM strychnine, 200 μM glycine in Mg2+-free Krebs-Ringer buffer (KRB, 135 mM NaCl, 5 mM KCl, 0.4 mM KH2PO4, 5.5 mM glucose, 20 mM HEPES, pH 7.4). First the cells were rinsed by Mg2+-free KRB along to wash out Mg2+ ions. Then, cLTP cocktail was applied for 4 min to induce LTP (cLTP induction phase). After this, the cLTP cocktail was changed to Mg2+- and Ca2+-containing KRB (KRB + Mg + Ca) and kept until the cells were fixed or imaged (cLTP development phase). At indicated time-points after cLTP induction, KRB + Mg + Ca was quickly removed and cells were fixed with 4% formalin in PBS (20 min, room temperature (RT)). Fixed cells were washed 3 times with PBS and mounted on microscope slides using SlowFade® Gold Antifade mountant (Life Technologies, Monza, Italy).
Fixed cells were imaged on a Leica DMI6000 epifluorescent microscope equipped with Polychrome V monochromator (Till Photonics, Graefelfing, Germany) and S Fluor ×40/1.3 objective (Leica, Buccinasco, Italy). The cells were alternatively excited by 488 and 546 nm and emission light was filtered through 520/20 and 600/40 nm bandpass filters, respectively, and collected by a cooled CCD camera (Leica, Hamamatsu, Japan). The fluorescence signals were acquired and processed using MetaMorph software (Molecular Device, Sunnyvale, CA, USA).
To quantify NY translocation, 5 random fields were photographed from each coverslip, and the astrocytes with clear expression of NY sensor were used for analysis (3–15 astrocytes per field). For each cell two regions of interest (ROIs) were placed inside the nucleus (which was evidenced by mCherry-H2Bc expression) and two in different sides of the cytosol close to the nucleus. For each cell, the fluorescence intensity of two nuclear (Nuc) and two cytosolic (Cyt) ROIs, respectively, measured in green channel, was averaged and Nuc/Cyt ratio was obtained by dividing the resulting Nuc fluorescence to Cyt fluorescence. The data are expressed as mean ± SEM for each cell analyzed form three independent culture preparations (e.g., Figure 3
4.5. Induction of cLTP and NY Translocation Quantification
For Fura-2 Ca2+ imaging experiments, the cells were first loaded with 2 μM Fura-2-AM in presence of 0.02% of Pluronic-127 (both from Life Technologies) and 10 μM sulfinpyrazone (Sigma) for 20 min at RT. Fura-2 loaded cells were washed in KRB + Mg + Ca and allowed to de-esterify for 20 min before cLTP induction.
After that, the coverslips were mounted into acquisition chamber and placed on the stage of a Leica DMI6000 epifluorescence microscope equipped with S Fluor ×40/1.3 objective. The probe was excited by alternate 340 and 380 nm using a Polychrome IV monochromator and the Fura-2 emission light was filtered through 520/20 bandpass filter and collected by a cooled CCD camera (Hamamatsu, Japan). The fluorescence signals were acquired and processed using MetaFluor software (Molecular Device, Sunnyvale, CA, USA). To quantify the differences in the amplitudes of Ca2+ transients the ratio values were normalized using the formula ΔF/F0 (referred to as normalized Fura-2 ratio, “Norm. Fura ratio”). At least two coverslips for each of three independent culture preparation were imaged for each condition. In mixed neuron-astroglial cultures, astrocytes were recognized as flat polygonal or star-like cells, while neurons were recognized by round bodies with few processes located on upper focal plane above the astrocytes. The cells with uncertain morphology were not taken in consideration.
4.6. Pharmacological Reagents
MK801 (Cat. 0924, stock 100 mM in DMSO), MTEP (Cat. 2921, stock 100 mM in DMSO), 2APB (Cat. 1224, stock 50 mM in DMSO), Xestospongin C (Cat. 1280, stock 2 mM in DMSO), U73122 (Cat. 1268, stock 10 mM in DMSO), NMDA (Cat. 0114, stock 100 mM in H2O) and DHPG (Cat. 0342, stock 10 mM in PBS) were from Tocris (Bristol, UK). Pyr3 (Cat. P0032, stock 10 mM in DMSO), Pyr6 (Cat. SML1241, stock 10 mM in DMSO) and Pyr10 (Cat. SML1243, stock 10 mM in DMSO) were from Sigma.
4.7. Statistical Analysis
Statistical analysis was performed using GraphPad Prism software v.7. For analysis of Nuc/Cyt NY ratio (Figure 3
) each dataset at indicated time-points was compared with respective control using a two-tailed unpaired Students’s t
-test. Differences were considered significant at p
< 0.05. Data are expressed as mean ± SEM.