From Pathology to Physiology of Calcineurin Signalling in Astrocytes

Introduction

Calcineurin (CaN), a Ca2+/calmodulin (CaM)-activated serine/threonine phosphatase, serves as a master-regulator of development- and differentiation-associated cellular remodelling. CaN signalling is mostly studied in peripheral immune cells (Fric et al., 2012; Pan et al., 2013), muscle cells(Mallinson et al., 2009; Schiaffino, 2010; Hudson & Price, 2013), and developing neurons (Carafoli et al., 1999; Genazzani et al., 1999; Kramer et al., 2003). In differentiated neurons, changes of CaN activity are essential for establishment of long-term depression (Mulkey et al., 1994) and emotional memory (Baumgärtel et al., 2008; Baumgärtel & Mansuy, 2012), maintenance of neuronal excitability and metaplasticity (Abraham & Bear, 1996). In neuronal pathology, increased CaN signalling is implicated both in neuroinflammation and in memory impairment, the latter being likely to be caused by a combination of factors including decreased synaptic plasticity, synaptic loss, neuronal apoptosis and neuroinflammation (Reese & Taglialatela, 2010; Reese & Taglialatela, 2011). Regarding astroglial cells, CaN signalling is studied principally in pathological conditions (Reese & Taglialatela, 2010; Reese & Taglialatela, 2011; Furman & Norris, 2014) in most of which, including Alzheimer’s disease (AD), the augmented expression/hyperactivation of CaN was found. Activation of CaN is implicated in disease-associated inflammation and reactive astrogliosis both in chronic neurodegeneration (Furman & Norris, 2014) and in acute brain pathology (Kaminska et al., 2004). Neuroinflammation is a complex polyfactorial phenomenon in which astrocytes share the centre stage with microglial cells (Heneka et al., 2015; Heppner et al., 2015; DiSabato et al., 2016). The role of astrocytes in such an inflammatory duet, in which astrocytes can exert both neuroprotective and neurotoxic effects, is not yet precisely defined and is a matter of intensive investigations(Pekny & Pekna, 2014; Ben Haim et al., 2015). Astroglial CaN, in this context, acts as a molecular switch for setting-up astroglial reactivity and for amplification of inflammatory cascades (Schubert et al., 2000; Furman & Norris, 2014), yet it is required for the resolution of inflammation (Fernandez et al., 2007; Fernandez et al., 2012). Inflammatory aspects of CaN signalling in astrocytes have been recently reviewed in several comprehensive contributions addressing different aspects of this complex issue (Reese & Taglialatela, 2010; Reese & Taglialatela, 2011; Furman & Norris, 2014).

The present contribution is focused on the mechanistic aspects of CaN signalling in astrocytes. We analyse the literature reporting the Ca2+ signals, necessary for the activation of CaN and its downstream targets, and the signalling in astrocytesmediated by the direct CaN target, NFAT, as well as complex and intricate relationships with the signalling mediated by NF-kB.

Calcineurin

Molecular principles of CaN signalling

Among the plethora of processes and enzymes at work in the cell, which are regulated by spatio-temporally defined changes in intracellular Ca2+ concentration ([Ca2+]i), the serine/threonine phosphatase CaN is the only phosphatase directly activated by Ca2+ and CaM binding. This unique feature places CaN on the apex of Ca2+ regulated signalling cascades which require protein dephosphorylation, and it is difficult to overestimate its role in cellular (patho)physiology. In the field of brain pathology, there is growing interest for CaN as it is thought to be the main promoter of neuroinflammation and of transition to reactivity in astrocytes. Before discussing CaN signalling in astrocytes, we shall briefly describe the molecular principles of CaN activation and the modulation of its activity by Ca2+ signals.

CaN was discovered around four decades ago as an inhibitor of the CaM-dependent cyclic nucleotide phosphodiesterase (Wang & Desai, 1976; Klee et al., 1979). It was first characterized in neural tissue and then was found as a largely distributed protein in all tissues. Structurally, CaN is a heterodimer of a   ̴60 kDa catalytic A subunit (also known as calcineurin A or CaNA) and it is tightly associated with a  ̴19 kDa regulatory B subunit (also known as calcineurin B or CaNB). CaNA possesses the catalytic site at the N-terminal portion. In addition it has two binding domains, one recognized by CaNB and the other one able to bind CaM. In mammals three isoforms of CaNA have been identified: α, β and γ(Goto et al., 1986a; Klee et al., 1988; Kuno et al., 1992; Cohen et al., 1996). The isoforms α and β are ubiquitously distributed, while the γ isoform is specifically expressed in testis (Muramatsu & Kincaid, 1992).

The CaNB subunit possesses four distinct helix-loop-helix motifs known as EF-hands able to sense the Ca2+ fluctuations which are indispensable for activating a large group of Ca2+-binding proteins (Yanez et al., 2012). In CaNB, the first 2 EF-hands at the N-terminal lobe serve both as low affinity Ca2+-sensing motifs, while the other 2 EF-hands at the C-terminal lobe bind Ca2+ with high affinity with a constant of dissociation (Kd) in the nanomolar range (Kakalis et al., 1995). As suggested in different studies, the two pairs of motifs have different roles: the low affinity motifs have a regulatory role serving as sensors of the calcium fluctuations, whereas the high affinity motifs on the C-terminal lobe serve for stabilisation of the heterodimer (Stemmer & Klee, 1994; Klee et al., 1998).

The activation of CaN relies on a sequential structural rearrangement in which the first step is represented by the binding of two calcium ions to both low affinity EF-hands at the N-terminal lobe of CaNB. This event induces the opening of the CaM binding region that is then available for the recognition by CaM, which, in turn, promotes the displacement of the autoinhibitory domain from the catalytic site of CaNA (Stemmer & Klee, 1994; Yang & Klee, 2000). CaNB is indispensable for the catalytic activity of CaNA, and its genetic deletion has been widely used to knock-out CaN in different tissues and organs (Zeng et al., 2001; Chang et al., 2004; Heit et al., 2006; Parsons et al., 2007; Maillet et al., 2010).

Considering the biological relevance of CaN, it is easy to deduce that it can interact with a wide broad of proteins. In the years, many substrates, involved in a plethora of signal transduction pathways, were characterized, highlighting the key role of this enzyme in cellular homeostasis (Ingebritsen & Cohen, 1983; Goto et al., 1985; Pallen & Wang, 1985; Klee et al., 1988; Jain et al., 1993; Lieberman & Mody, 1994; Coghlan et al., 1995; Wang et al., 1999; Bhattacharyya et al., 2009). Among the CaN interactors, the most studied is the nuclear factor of activated T-cells (NFAT). The NFAT proteins are a family of transcriptional factors largely expressed in different cells of the immune system, where they were first identified (Shaw et al., 1988). Subsequently, it was found that NFAT is present in other cell types, including those of the nervous system (Ho et al., 1994). The NFAT family is formed by five members: NFATc1 (NFAT2, NFATc), NFATc2 (NFAT1, NFATp), NFATc3 (NFAT4, NFATx), NFATc4 (NFAT3) and NFAT5 (TonEBP: tonicity element binding protein or OREB: osmotic response element binding protein) (Im & Rao, 2004). The NFATc1-4 proteins are dephosphorylated at the N-terminal portion by activated CaN (Clipstone & Crabtree, 1992; Klee et al., 1998). This dephosphorylation induces a structural rearrangement of the proteins leading to the exposure of the nuclear localization sequences (NLSs) allowing the nuclear import of NFAT. The nature of the transcriptional modulation induced by NFAT could be either activating or repressive depending on the transcription factors which interact with NFAT in DNA binding, suggesting a context-dependent regulation. Indeed, NFAT activates the transcription when it cooperates with activator protein-1 (AP-1; an heterodimer formed by Fos and Jun), GATA proteins, MEF-2 and histone deacetylases (HDACs) (Blaeser et al., 2000; McKinsey et al., 2000; Avni et al., 2002; Hogan et al., 2003; Lee et al., 2003), but it also acts as a transcriptional repressor when the complex with AP-1 is disrupted (Macián et al., 2002).

The dephosphorylation of NFATs and their activation could be prevented by immunosuppressive drugs: cyclosporine A (CsA) and FK506 (tacrolimus). Both CsA and FK506 form complexes with the immunophilins cyclophilin A and FKBP12, respectively. These drug–immunophilin complexes then bind to and inhibit CaN (Griffith et al., 1995; Kissinger et al., 1995; Huai et al., 2002). Because both CsA and FK506 inhibit the phosphatase CaN activity directed to all the targets, their effect is not specific to NFAT (Martinez-Martinez & Redondo, 2004). For the selective inhibition of the signalling prompted by NFAT, specific synthetic peptides, able to compete with NFAT for CaN binding and to inhibit NFAT dephosphorylation, were developed (Aramburu et al., 1998; Garcia-Cozar et al., 1998). NFAT has two CaN binding sites: the PxIxIT motif at the N-terminal and the LxVP motif at the C-terminal region of the protein. The first characterized PxIxIT motif in NFATc2 was the PRIEIT peptide, whose overexpression prevents NFAT nuclear localization by competing with NFAT itself for the binding with CaN (Aramburu et al., 1998). Thereafter, a randomized PxIxIT motif library was generated leading to the identification of an optimized peptide called PVIVIT that has a 50-fold higher affinity for CaN than PRIEIT. Aramburu et colleagues demonstrated that the replacement of the native PRIEIT sequence with PVIVIT in NFAT induces the constitutive CaN-dependent dephosphorylation of NFAT and its permanent translocation inside the nucleus (Aramburu et al., 1999). When the PVIVIT peptide alone is overexpressed in the cell, it blocks specifically the NFAT-dependent pathway. In fact, it has become a powerful tool to inhibit NFAT-dependent processes both in vitro and in vivo(Furman et al., 2012; Furman et al., 2016).  While the PxIxIT motif is considered to be the primary CaN-binding site for all members of the NFAT family, the LxVP motif seems to act as a regulator which modulates the strength of the CaN/NFAT-binding (Martínez-Martínez et al., 2006). Although the studies centred on the characterization of the synthetic peptides demonstrate the involvement of both PxIxIT and LxVP in NFAT binding to CaN, the functional interaction between these two sites remains to be fully understood. 

Ca2+ signals and CaN activation

As suggested by the name, CaN is a Ca2+-dependent enzyme, and it is therefore important  to understand the sources and spatio-temporal properties of Ca2+ signals implicated in its activation. The experiments dissecting the Ca2+ signals necessary for CaN-dependent gene transcription have mostly been performed in lymphocytes by either stimulating T-cell (or B-cell) receptor or by using inhibitors of the sarco-endoplasmic reticulum Ca2+ ATPase (SERCA) to induce [Ca2+]i elevation. A receptor mediated Ca2+ signal, which results from the inositol-1,4,5-trisphosphate (InsP3)-mediated Ca2+ release from the internal stores, has the shape of a transient rise which decays to the baseline after reaching a maximum of amplitude. For long-lasting Ca2+ signals, monitored as a monotonic plateau or Ca2+ oscillations, specific mechanisms have to be engaged,such as Ca2+-induced Ca2+ release (CICR) and/or store-operated Ca2+ entry(SOCE) (see below). Activation of the T cell receptor (TCR) may result in either a single Ca2+ transient or a transient followed by a long-lasting plateau or oscillations. Using the β-galactosidase reporter gene (lacZ) under the control of NFAT element of the interleukin-2 (IL-2) enhancer, Negulescu and colleagues demonstrated that CaN/NFAT-mediated gene expression depends on a long-lasting elevation of [Ca2+]i but not on a short Ca2+ transient (Negulescu et al., 1994). In this study it was elegantly demonstrated that T cells stimulated with immobilized anti-CD3 antibody exhibited a sustained Ca2+ signal that lasted for more than 25 min and promoted lacZ expression. Using the SERCA blocker thapsigargin (TG) to fully activate SOCE and clamping extracellular Ca2+ at different [Ca2+], the authors discovered that a sustained elevation of [Ca2+]i ranging between 0.2 and 1.6 μM for at least 10 min is necessary for CaN/NFAT to be activated (Negulescu et al., 1994). CaN activates several distinct signalling cascades, the most important of which are mediated by the transcription factors NFAT and NF-kB. An outstanding question is how the Ca2+ signals selectively discriminate between these two CaN targets. Dolmetsch and Lewis with their colleagues were the first to address this question. They found that, in B lymphocytes, the amplitude and the duration of Ca2+ signals differentially control the activation of NFAT versus NF-kB. NFAT was specifically activated by a low, sustained Ca2+ plateau, while NF-kB required a large Ca2+ transient (Dolmetsch et al., 1997). While Ca2+/CaN is the only way to activate NFAT, NF-kB-dependent gene transcription is the result of a multistep signalling cascades (see below), in which CaN can intervene at different steps. The refinement of the Ca2+ sensitivity of NFAT and NF-kB may be achieved by modulating the frequency and the amplitude of Ca2+ oscillations. Thus, low frequency oscillations activate preferentially NF-kB, while rapid oscillations activate both the transcription factors. Since Ca2+ oscillations have many more parameters to be modulated by specific mechanisms than the monotonic Ca2+ plateau (Dragoni et al., 2011; Smedler & Uhlén, 2014), they are widely used by different cell types, including astrocytes, in order to provide the ultimate specificity for Ca2+-regulated processes.

One of the important factors that, alongside the classical Ca2+ signalling toolkit, shapes the intracellular Ca2+ signals is represented by mitochondria, which are also denominated as the cellular “Ca2+ firewall” (Straub et al., 2000; Walsh et al., 2009). Mitochondria act as a high capacity and low affinity Ca2+ buffer which absorbs high quantities of Ca2+ ions generated locally in the proximity of Ca2+ channels of both the ER membrane and the plasma membrane. By their fast Ca2+ uptake mitochondria smooth the steep Ca2+ overshoot and prevent the channels from inactivating, thus generating the prolonged elevation of Ca2+ required for NFAT activation in T-cells (Hoth et al., 2000). Whether a similar mechanism is also used by astrocytes remains to be investigated.

In pathology, the maladaptive remodelling of the Ca2+ signalling machinery may lead to exaggerated and uncontrolled Ca2+ signals and to non-specific activation of Ca2+-dependent proteases which can transform physiologic CaN signalling to a pathologic one (Carafoli, 2004; Brini et al., 2013). As reported in a series of studies, CaN may be constitutively activated via cleavage by the Ca2+-dependent protease calpain (Tallant et al., 1988; Wang et al., 1989; Lakshmikuttyamma et al., 2004). This, in fact, occurs in a number of pathological conditions in which Ca2+-dependent proteases are activated (Wu et al., 2007), including ischemic myocardium (Lakshmikuttyamma et al., 2003; Burkard et al., 2005) and experimental glaucoma (Huang et al., 2010). In brain pathology, calpain cleavage of CaN has been reported in brain ischemia (Shioda et al., 2006), excitotoxic neuronal death (Wu et al., 2004), and in Alzheimer’s disease (Liu et al., 2005; Mohmmad Abdul et al., 2011). However, it is not clear from these reports whether the calpain-mediated activation of CaN occurs only in neurons or also in astrocytes. This latter hypothesis is supported by the notion that calpain is expressed in astrocytes (Perlmutter et al., 1990; Lee et al., 2000; König et al., 2003) and that both calpain and CaN inhibition protect astroglial cells form Ca2+-reperfusion-induced apoptotic death (Takuma et al., 1999; Takuma et al., 2004).

Over the past decade, different fluorescence CaN reporters have been developed (Newman & Zhang, 2008; Lodygin et al., 2013; Mehta et al., 2014), none of which has, however, been tested in astrocytes. Using fluorescence resonance energy transfer (FRET)-based CaN sensors directed to different sub-cellular locations Zhang and colleagues have demonstrated the crucial role played by CaN distribution in CaN activation (Mehta et al., 2014).

Astroglial Ca2+signalling and Ca2+ routs for CaN activation

Astrocytes are the major cellular type in the brain outnumbering neurons by 2-8 folds(Verkhratsky, 2007).They perform fundamental housekeeping functions, including the maintenance of structural, metabolic and ionic homeostasis, and constitute an essential component of the neuro-vascular unit (Verkhratsky, 2007). Astrocytes have an extremely complex sponge-like morphology with a huge number of very fine processes (Bushong et al., 2002; Nedergaard et al., 2003). The unprecedented complexity of protoplasmic astrocytes does not only involve their morphological aspects (Bushong et al., 2002; Nedergaard et al., 2003), but it is becoming clear that the Ca2+ signalling architecture in these cells is extremely complex as well (Volterra et al., 2014; Shigetomi et al., 2016). Astrocytic processes ensheath synaptic structures forming the so called tripartite synapse – a structure composed of pre- and postsynaptic terminals and of astroglial processes, which insulate the synapse, providing the micro-environment for optimal synaptic functioning (Araque et al., 1999; Verkhratsky & Nedergaard, 2014). The tripartite structure provides a tool for communication between neurons and astrocytes. It has been hypothesized that astrocytes can play an active role in bidirectional astrocyte-neuron signalling by modulating the availability of neurotransmitters, thereby participating in information processing (De Pittà et al., 2016). The gliotransmission and the role of astroglial Ca2+ signals in modulation of synaptic transmission is a subject of intensive debates (Kimelberg, 2007; Agulhon et al., 2008; Fiacco et al., 2009; Pereira & Furlan, 2010; Agulhon et al., 2012; Benneyworth et al., 2012; Nedergaard & Verkhratsky, 2012; Ota et al., 2013; Araque et al., 2014;Fujita et al., 2014; Perea et al., 2014; Sloan & Barres, 2014; Volterra et al., 2014;Oliveira et al., 2015). Elevations of [Ca2+]i in astrocytes, which may be associated with CaN activation, are frequently  found in acute and chronic neurodegenerative pathologies.In this section, we discuss the mechanisms of generation of Ca2+ signals in astrocytes in the context of their possible role in CaN activation.

Calcium is a universal second messenger which regulates a vast spectrum of cellular processes (Fedrizzi et al., 2008; Brini et al., 2013). The specificity of the signal, which is also defined as information encoding, is assured by the organ/cell-specific composition of the so called Ca2+signalling toolkit (Berridge et al., 2003; Verkhratsky et al., 2012a) which generates a stimulus-specific rise of [Ca2+]i precisely defined in time and space. The changes of [Ca2+]i in astrocytes have been associated with virtually all physiological functions ranging from release of gliotransmitters (Parpura & Haydon, 2000) to astroglial plasticity (Pirttimaki & Parri, 2013) and responses to sensory stimulation (Lind et al., 2013).

Being electrically non-excitable cells, in the sense that they do not trigger a voltage-dependent action potential, which is a characteristic of neurons, astrocytes respond to extracellular stimuli by employing mainly metabotropic Ca2+signalling and the membrane of the endoplasmic reticulum (ER) as an excitable media (Verkhratsky et al., 2012a).  Astrocytes express a variety of metabotropic G-protein-coupled receptors (GPCRs) (Verkhratsky, 2007), of which, perhaps, the best studied are the group I metabotropic glutamate receptors (mGluRs) and the purinergic metabotropic receptors of the P2Y family. The binding of a ligand (e.g. glutamate or ATP) to a receptor triggers a signalling cascade which initiates with GTP-dependent dissociation of the trimeric G-protein. The αq/11 subunit of the G-protein binds to and activates PLCb, which in turn, hydrolyses phosphatidylinositol 4,5-bisphosphate (PIP2) in the plasma membrane, generating two second messengers, diacylglycerol (DAG) and InsP3. While DAG is involved in protein kinase C (PKC) signalling, InsP3, a soluble and diffusible second messenger, binds to and activates the InsP3 receptors (InsP3Rs) in the membrane of the ER, inducing release of Ca2+ from the ER lumen into the cytosol. Of the three known InsP3R isotypes, type 2 is considered to be the principal InsP3R in astrocytes in intact tissue(Holtzclaw et al., 2002; Petravicz et al., 2008), although presence and activity of InsP3R1 in cultured astrocytes has also been documented (Grolla et al., 2013b; Grolla et al., 2013a).

The InsP3-induced increase of [Ca2+]i, measured as a local Ca2+ transient, is the initial event which may be terminated without propagation (Verkhratsky et al., 2012a), but could be amplified within the same cell. Furthermore, the increase of [Ca2+]i may also propagate to the neighbouring astrocytes, generating inter-cellular Ca2+ waves (Charles, 2005; Scemes & Giaume, 2006), which are mainly mediated by InsP3 diffusion through gap junctions (GJ), alongside with the paracrine action of ATP (Cornell-Bell et al., 1990; Dani et al., 1992).

The amplification and propagation of Ca2+ signals within the astrocyte occur mainly via two mechanisms: i) the CICR and ii) the SOCE. Ca2+ ions, massively released from the ER through InsP3Rs, generate a local domain of high [Ca2+], the so called “hot spot” or “high Ca2+ microdomain” (Rizzuto et al., 1999; Parekh, 2008; Contreras et al., 2010). Diffusion of Ca2+ ions through the cytosol is severely restricted (Santella et al., 2003) due to potent buffering mechanisms (cytosolic Ca2+ buffers, Ca2+ uptake by juxtaposed mitochondria, Ca2+ transporters). However, Ca2+ ions are able to act at the neighbouring InsP3Rs, sensitizing them to a lower [InsP3], generating a Ca2+ wave, which propagates along the ER membrane (Verkhratsky & Kettenmann, 1996; Verkhratsky, 2007).

Another additional important mechanism to achieve a long-lasting Ca2+ elevation in the cytosol of astrocytes is SOCE, also known as capacitative Ca2+ entry (CCE) (Verkhratsky & Parpura, 2014). SOCE is triggered when Ca2+ is released from the ER after activation of InsP3Rs and the ER [Ca2+] drops down (Putney, 2007; Moccia et al., 2015). Stromal interacting protein 1 (Stim 1), an integral protein of the ER membrane, senses the decrease of intra-ER [Ca2+] and transmits the signal to a channel on the plasma membrane called Orai, of which three isoforms exist (Orai 1, 2, 3) (Putney, 2007; Moccia et al., 2015). Although it is postulated that the activation of Orai is essential for SOCE, another group of cation-permeable channels, the so called Transient Receptor Potential(TRP) channels may be involved in CCE(Salido et al., 2009; Cheng et al., 2013). In astrocytes, SOCE generates a long-lasting (minutes) Ca2+ elevation which is recorded as a plateau which follows the peak of the InsP3R-mediated Ca2+ transient (Ronco et al., 2014; Verkhratsky & Parpura, 2014). SOCE is a common feature present in all types of astrocytes (Tuschick et al., 1997; Pivneva et al., 2008; Verkhratsky & Parpura, 2014).Although all the components of the SOCE molecular machinery, including STIM, Orai and TRP channels (Pizzo et al., 2001; Grimaldi et al., 2003; Golovina, 2005; Barajas et al., 2008; Malarkey et al., 2008; Ronco et al., 2014; Verkhratsky et al., 2014), are present in astrocytes, experiments suggest that astroglial cells employ mainly TRP channels to generated SOCE (Verkhratsky & Parpura, 2014) and to coordinate ion signalling and astroglial excitability (Verkhratsky et al., 2014). Intriguingly, SOCE is the main mechanism for CaN activation in peripheral immune cells and lymphocytes (Gwack et al., 2007; Oh-hora & Rao, 2008).

Another form of Ca2+ signals which is common for both the individual astrocytes and the astroglial networks are Ca2+ oscillations (Zonta & Carmignoto, 2002; Koizumi, 2010). The mechanisms of CICR and SOCE have been generally implicated in setting up the oscillatory Ca2+ dynamics (Carafoli et al., 2001; Uhlen & Fritz, 2010; Dupont et al., 2011; Parekh, 2011). Specifically in astrocytes, TRP channels are involved in glutamate-induced Ca2+ oscillations (Pizzo et al., 2001). Ca2+-dependent release of ATP, instead, is implicated in synchronized Ca2+ oscillations in astroglial networks (Koizumi, 2010).

Ca2+ entry from the extracellular medium via ligand-gated and voltage-gated Ca2+channels may generate Ca2+ signals in astrocytes (James & Butt, 2002; Verkhratsky, 2007; Palygin et al., 2010; Verkhratsky et al., 2012b). The expression of Ca2+ ion channels in astrocytes is subtype- and region-specific. It has been suggested that most types of astrocytes express L-type and T-type Ca2+ currents (Verkhratsky &Steinhauser, 2000). Whilethe expression of different types of Ca2+channels was found in a population of immature mouse hippocampal astrocytes, apparent absence of Ca2+ currents was found in adult rat hippocampal and visual cortex astrocytes (reviewed in Verkhratsky& Steinhauser, 2000). Ca2+ permeable AMPA receptors/channels have been found in oligodendrocytes and NG2 glia (Lin& Bergles, 2002). Astrocytes in cortex and spinal cord express NMDA receptors (Verkhratsky & Kirchhoff, 2007).Stimulation of nicotinic acetylcholine receptors (nAChRs) either with nicotine or with acetylcholine has been shown to elevate astroglial [Ca2+]i (Teaktong et al., 2003; Oikawa et al., 2005). A series of reports suggest that astroglial nAchRs can be target of Aβ in AD (reviewed in Sadigh-Eteghad et al., 2016). Culturedastrocytes express different types of purinergic P2X receptors (P2X1-5,7). In intact tissue, P2X receptors have been found in astrocytes of many brain regions including cortex, hippocampus, cerebellum and brainstem (reviewed in Illes et al., 2012). Ionotropic receptors in astrocytes are mainly localized in fine processes adjacent to the post-synaptic terminals and it is not known whether they participate in the propagation of Ca2+ signals through the astrocyte and in CaN activation. In addition, Ca2+ permeability of ligand-gated channels in astrocytes is relatively low with respect to Na+(Kirischuk et al., 2012), which makes them more suitable for controlling Na+ homeostasis.

Another route for generating intracellular Ca2+ signals, which appears to be specific for astrocytes, is represented by the Na+/Ca2+ exchanger (NCX). NCX is an antiporter that exploits the energy of the electrochemical gradient of Na+ to extrude Ca2+ out from the cell. To expel one Ca2+ ion, NCX allows three Na+ ions to enter (Annunziato et al., 2004; DiPolo & Beauge, 2006). In the majority of cellular types NCX constitutes one of the most important mechanisms of extruding Ca2+ from the cells representing low affinity and high capacity Ca2+ transport system. It is, therefore, different from the plasma membrane Ca2+ ATPase pump (PMCA) which has high Ca2+ affinity but relatively low transport capacity. In astrocytes, NCX functions in the reverse mode,transporting Na+ out and allowing Ca2+ to flow into the cell (Takuma et al., 1994; Reyes et al., 2012). This occurs due to low reversal potential for NCX (~ -87 mV), which is just slightly more negative than the resting membrane potential (-80 to -85 mV). Increases of [Na+]i (e.g., after the activation of Na+-permeableionotropic receptors, voltage-gated Na+channels andNa+-dependent glutamate transporters)reduce the electrochemical Na+ gradient, therebyallowing Ca2+ to enter the cell (Annunziato et al., 2004; DiPolo & Beauge, 2006; Verkhratsky et al., 2012a). However, at present it is not known whether NCX-mediated Ca2+ signals contribute to CaN activation in astrocytes.

Taken together, astrocytes are endowed with versatile and stimulus-specific mechanisms to encode the signalling information within Ca2+ signals (Figure 1). In the following sections we will discuss how these signals are decoded by CaN and by its downstream targets to achieve specific transcriptional and functional outputs.

Figure 1. Astroglial Ca2+ routes for CaN activation The principal mechanism of CaN activation in astrocytes is represented by the activation of the seven transmembrane domains G-protein coupled receptors (GPCR), which generates the second messenger InsP3 that in turn activates the InsP3Rs on the ER membrane (depicted as thin green arrows). Once activated, the InsP3Rs induce the store-operated Ca2+ entry, which involves the interaction between the STIMs molecules onto the ER and the Orai channels located into the plasma membrane. That recruits the transient receptor potential (TRP) channels to induce a massive Ca2+ entry (depicted as bold green arrow). Ca2+-induced Ca2+ release (depicted as light-green thin arrows) may also be involved in the amplification of Ca2+ signal necessary for CaN activation, whereas the roles of mitochondrial Ca2+ uptake (Mit), ligand gated Ca2+ channels (LGCC) and sodium/calcium exchanger (NCX) is not clear. Note that IP3R1 has been studied in cultured astrocytes while in intact tissue IP3R2 is the principal isoform

Calcineurin signalling in astrocytes

Ca2+, CaN and NFAT

The term “calcineurin” was coined by Claude Klee and co-workers in 1979 “on the basis of its Ca2+-binding properties and its specificity for the nervous system” (Klee et al., 1979). In later years, CaN was found to be expressed in many (if not all) cellular types, including astrocytes and microglia. Curiously, initial examination of the brain using anti-CaN antibody failed to detect CaN in astrocytes (Goto et al., 1986b), although experiments using pharmacological inhibitors of CaN suggested its presence (Randriamampita & Tsien, 1995; Hosoi et al., 1997; Matsuda et al., 1998). Thereafter, CaN protein was found in cultured astrocytes and in rodent brain in situ(Hashimoto et al., 1998), and presence of both the catalytic CaNA and the regulatory CaNB subunits was confirmed (Vinadé et al., 1997; Matsuda et al., 1998). As determined by microarray and RNA-sequencing, at mRNA level, astrocytes express high levels of the two principal isoforms of the catalytic subunit, CaNAα and CaNAβ, and the CaNB1 isoform of the regulatory subunit, while CaNAγ and CaNB2 mRNA are present at much lower levels (Cahoy et al., 2008; Orre et al., 2014; Zhang et al., 2014). Both CaNAα and CaNAβ proteins are expressed in cultured astrocytes (Canellada et al., 2008; Lim et al., 2013).

As it was mentioned above, NFAT is the direct and, at the same time, the most studied target of CaN. Numerous publications reported the expression of NFAT in astrocytes and astroglial cell lines (Jones et al., 2003; Canellada et al., 2008). At the mRNA level, both developing and adult astrocytes express all four NFAT isoforms although their overall amount is lower than that of CaN (Cahoy et al., 2008; Orre et al., 2014; Zhang et al., 2014). At the protein level, NFATc3 has been detected in most studies (Jones et al., 2003; Filosa et al., 2007; Canellada et al., 2008; Serrano-Pérez et al., 2011; Neria et al., 2013; Yan et al., 2014; Furman et al., 2016),while the expression of NFATc1 (Pérez-Ortiz et al., 2008) and NFATc2 (Furman et al., 2016) has also been reported.

CaN and, consequently, NFAT in astrocytes may be activated by a variety of stimuli including neuro(glio)transmitters (glutamate and ATP) (Jones et al., 2003; Filosa et al., 2007; Canellada et al., 2008; Pérez-Ortiz et al., 2008), inflammatory agents (IL-1β) (Sama et al., 2008), experimental manipulations using a Ca2+ ionophore alone (Jones et al., 2003; Pérez-Ortiz et al., 2008) or in combination with phorbol myristate acetate (PMA), an activator of PKC and a wide range of its downstream kinases (Neria et al., 2013).

Treatment of cultured astrocytes with β-amyloid (Aβ) peptide, which is thought to be the main cause of AD (Selkoe & Hardy, 2016), induced the rapid translocation of NFAT into the nucleus (Abdul et al., 2009). In the human astrocytoma U373MG cell line, Aβ-induced activation of the CaN/NFATc3 axis has been shownto elicit the expression of beta-site amyloid precursor protein (APP) cleaving enzyme 1 (BACE1) (Jin et al., 2012), which is involved in the over-production of toxic Aβ(1-42) (Aβ42). In this context, it has been demonstrated that rat hippocampal astrocytes in culture express APP itself and all the enzymatic machinery to process APP and to generate and release β-amyloid peptide in a cell autonomous manner (Grolla et al., 2013b).

Simulation of acute brain pathology by injection of kainic acid and mechanical lesion triggers NFATc3 over-expression in vivo in a subset of reactive astrocytes (Serrano-Pérez et al., 2011; Neria et al., 2013). In another in vivo model, the ischemia-reperfusion, Rcan1-4 was induced in GFAP-positive astrocytes in the site of lesion (Sobrado et al., 2012). Regulation of Rcan1-4 and COX2 gene expression in astrocytes has been shown to be controlled by CaN/NFATc3 in cultured(Canellada et al., 2008) and in situ (Sobrado et al., 2012) astrocytes.

Using traumatic brain injury (TBI) with controlled cortical impact (CCI), Furman et al (Furman et al., 2016) observed over-expression of NFATc3 and its increased DNA binding in hippocampal lysates at 1 week after CCI TBI. Immunohistochemical analysis confirmed that the over-expressed NFATc3 was localized in astrocytes. Furthermore, adeno-associated virus (AAV)-mediated astrocyte-specific expression of VIVIT, a peptide which inhibits NFAT-CaN interaction thereby precluding NFAT activation (Aramburu et al., 1999) (see above), significantly reduced TBI-induced nuclear translocation of NFATc3 (Furman et al., 2016). However, using the same CCI TBI protocol, Yan et al (Yan et al., 2014) found down-regulation of NFATc3 at all the time-points (from 6 hours to 4 weeks) after CCI TBI in both cytosolic and nuclear hippocampal fractions.Immunohistochemical analysis showed astroglial localization of NFATc3 and it is concluded that the astroglial NFATc3 signalling is down-regulated after CCI TBI (Yan et al., 2014). The reasons for such a discrepancy are not clear. These are the first reports that use the CCI TBI protocol of brain injury to monitor NFAT activation and further experiments are warranted to confirm the CCI TBI-induced NFATc3 activation in astrocytes.

Among the transcriptional targets of CaN/NFAT signalling in astrocytes there are Rcan1-4 (Canellada et al., 2008; Serrano-Pérez et al., 2011; Neria et al., 2013), COX2 (Blanco et al., 2008; Canellada et al., 2008; Serrano-Pérez et al., 2011), TNFα, GM-CSF, S100β, Vimentin (Sama et al., 2008), IL-6 (Serrano-Pérez et al., 2011), BACE1 (Jin et al., 2012), and MMP3 (Neria et al., 2013). All these genes were up-regulated after activation of NFAT. Instead, GLT-1 (EAAT2) was down-regulated when NFAT was activated by either IL-1β or β-amyloid oligomers (Sama et al., 2008; Abdul et al., 2009). Ca2+ entry appears to be essential for the above effects since Ca2+ channel blockers (Canellada et al., 2008; Sama et al., 2008) or a Ca2+ chelator (Jin et al., 2012) abolished NFAT activation in cultured astrocytes. In other cases, to achieve CaN/NFAT activation, Ca2+ entry was forced directly by an ionophore (Canellada et al., 2008; Neria et al., 2013) or indirectly by kainic acid lesion (Serrano-Pérez et al., 2011) (See Table 1).

Table 1. CaN regulation of NFAT/NF-kB-mediated signaling in astrocytes.

A microarray approach was undertaken by Norris et al (Norris et al., 2005) to find CaN-dependent genes in mixed neuron-astrocyte cultures in which astrocytes were transduced with adenovirus (Ad) encoding a truncated  ~ 45 kD constitutively active CaN fragment (ΔCaN). Of 3403 genes analyzed across the three conditions (non-transduced, Ad-LacZ and Ad-ΔCaN), 719 were differentially expressed between the Ad-LacZ and the Ad-ΔCaN groups. The more stringent group of 298 genes was selected in the Ad-ΔCaN set which differs from both non-transduced and Ad-LacZ to eliminate probable impact of viral infection on gene expression (Norris et al., 2005). Among the most overrepresented gene ontology (GO) terms in this dataset there were: morphogenesis, ion homeostasis, immune response and cell adhesion for up-regulated genes; while the GO terms lipid biosynthesis, tricarboxylic acid cycle and intracellular signalling cascade were overrepresented in down-regulated genes (Norris et al., 2005). Somewhat surprising is that among the aforementioned CaN/NFAT transcriptional targets, only two genes, namely the renowned astroglial markers S100β and vimentin, were differentially regulated by the overexpression of ΔCaN. This may be tentatively explained by the differential recruitment of CaN to stimulus-specific pathways. For example, PMA plus ionophore treatment, which produce robust Ca2+ signals and phosphorylation, induces Rcan1-4 and COX2, but not InsP3R1 expression (Canellada et al., 2008). Instead, Aβ42-induced and CaN-mediated up-regulation of InsP3R1 in astrocytes (Grolla et al., 2013b; Grolla et al., 2013a) require low and sustained elevation of [Ca2+]i.

Ca²⁺, CaN and NF-kB

NF-kB is a family of closely related transcription factors with an established role in innate immunity and inflammation (Newton & Dixit, 2012; Hoesel & Schmid, 2013). The family consists of hetero- and homodimers composed of five monomers, RelA (p65) (encoded by gene Rela), RelB (Relb), cRel (Rel), p50 and p52. The most important heterodimers are RelA:p50, cRel:p50 and RelB:p52 (Shih et al., 2011). In addition to transcriptional regulation of monomer’s availability, p50 and p52 are also regulated by processing of the precursor proteins p105 (encoded by Nfkb1) and p100 (encoded by Nfkb2), respectively.

The signalling cascade responsible for NF-kB activation may be triggered by a number of stimuli including pro-inflammatory agents like TNFα, IL-1β and LPS, and by stimulation of T-cell or B-cell receptors (TCR and BCR) (Paul & Schaefer, 2013). The molecular cascades leading to NF-kB activation are complex and cell-specific, yet somewhat non-linear. The canonical NF-kB pathway, in response to TCR activation, converges on a trimeric molecular complex composed of CARMA1, MALT1 and Bcl10 proteins (CBM complex) (Rosebeck et al., 2011). The formation of CBM complex, through intermediate passages, results in phosphorylation of IKKβ, which phosphorylates the inhibitor of kBα (IkBα), whose consequent ubiquitination and proteasomal degradation enables nuclear translocation of a p65/p50 heterodimer, which interacts with partner transcription factors to drive the expression of numerous genes. It has long known that, under experimental conditions, the addition of a Ca²⁺ ionophore significantly facilitates NF-kB activation, thus establishing a role for  Ca²⁺ in NF-kB activation. In fact, it is a general laboratory practice to induce NF-kB activation in a cell culture by stimulation with PMA (which stimulates PKC and downstream cascades) in combination with the Ca²⁺ ionophores A23187 or ionomycin. Ca²⁺-dependent steps in TCR-triggered NF-kB activation include initial stabilization of CMB complex by CaMKII-mediated phosphorylation of CARMA1 and Bcl10 with consequent dissociation and IKKβ phosphorylation (Paul & Schaefer, 2013). CaN-mediated dephosphorylation of Bcl10 is required for CBM complex stability in canonical signalling from TCR to NF-kB (Frischbutter et al., 2011; Palkowitsch et al., 2011).

A number of works reported that in astrocytes NF-kB activation by LPS, IL-1β or TNFα: i) requires Ca²⁺/CaM-dependent enzymes; and that ii) this requirement may be context-specific. For examples, Liu and colleagues (Liu et al., 2000) reported the involvement of Ca²⁺ signalling in IL-1β-induced NF-kB-dependent transcription of TNFα and IL-6 in cultured human fetal astrocytes, whereas Schwaninger and colleagues (Schwaninger et al., 1999) found that Ca²⁺/CaM dependent protein kinases are essential for bradykinin-induced NF-kB activation and IL-6 induction in murine astrocytes. As reported by (Nath et al., 1999), NF-kB-dependent induction of IL-6 but not IL-1β production was promoted by HIV-1 Tat protein independently on extracellular Ca²⁺, which however does not rule out the requirement for Ca²⁺ released from internal stores. In fact, El-Hage et al showed that NF-kB-mediated production of MCP-1 and IL-6 by HIV Tat is highly Ca²⁺-dependent and that opiates potentiate Tat-induced NF-kB activation via a Ca²⁺-dependent pathway (El-Hage et al., 2008).

Although the entire cascade responsible for NF-kB activation in astrocytes is not yet fully identified, available literature suggests that CaN can intervene in NF-kB signalling at several points (Figure 2): i) CaN dephosphorylates Bcl10, which positively regulated NF-kB signalling in hippocampal cultured astrocytes(Lim et al., 2013); ii) CaN dephosphorylates IkBα, thus precluding its degradation which inhibits NF-kB nuclear translocation in culture(Pons & Torres-Aleman, 2000); iii) CaN, in complex with NF-kB and FOXO3, is required for TNFα-induced NF-kB nuclear translocation and activation of transcriptionin vitro and in vivo(Fernandez et al., 2012). In the first case, the co-immunoprecipitation assay suggests constitutive interaction of both CaNA isoforms, α and β, with Bcl10 in cultures of hippocampal astrocytes. Importantly, an amount of co-precipitated Bcl10 increased upon treatment with 100 nM Aβ₄₂ oligomers, while the phosphorylation status of Bcl10 decreased (Lim et al., 2013). At the same time, Aβ₄₂ induced the increase of [Ca²⁺]i followed by CaN/NF-kB-dependent up-regulation of key components of the astroglial Ca²⁺signalling toolkit, such as mGluR5 and InsP₃R2 (Lim et al., 2013; Ronco et al., 2014), suggesting a role for CaN/NF-kB axis in Aβ-induced deregulation of astroglial Ca²⁺signalling in AD (Lim et al., 2014; Lim et al., 2016). Interestingly, in hippocampal astroglial cultures, pro-inflammatory agents such as TNFα, IL-1β and LPS, induced NF-kB-dependent down-regulation of mGluR5, InsP₃R1 and InsP₃R2 mRNA as well as of DHPG-elicited Ca²⁺ signals (Ronco et al., 2014). However, in these case TNFα, IL-1β or LPS-induced degradation of IkBα was CaN-independent (Ronco et al., 2014) (Table 2). mGluR5 has previously been shown to be down-regulated by TNFα, IL-1β and LPS in astrocytes in culture(Aronica et al., 2005; Tilleux et al., 2007; Berger et al., 2012).

Figure 2. Schematic illustration of specific points in the NF-kB astroglial signalling cascade in which CaN can intervene  (1) CaN dephosphorylates Bcl10, which potentiates NF-kB signalling in hippocampal cultured astrocytes (Lim et al., 2013); (2) CaN dephosphorylates IkBα, thus precluding its degradation which results in the inhibition of NF-kB nuclear translocation (Pons & Torres-Aleman, 2000); (3) CaN, in complex with NF-kB and FOXO3, is required for TNFα-induced NF-kB nuclear translocation and activation of transcription both in vitro and in vivo(Fernandez et al., 2012)

A dual nature of CaN as activator or inhibitor of NF-kB/NFAT signalling in astrocytes has been accurately dissected in a series of reports from Torres-Aleman’s group (Pons & Torres-Aleman, 2000; Fernandez et al., 2007; Fernandez et al., 2012) (Figure 3). Such an ambivalent functional outcome of CaN activation is set up by the equilibrium between pro- or anti-inflammatory environments(Fernandez et al., 2012; Furman & Norris, 2014). While the canonical pro-inflammatory pathway, induced by TNFα or LPS, requires CaN to mediate the NF-kB/NFAT-dependent induction of iNOS and COX2, this effect is counteracted by the anti-inflammatory mediator IGF-I, which recruits CaN to inhibit NF-kB/NFAT mediated inflammation (Fernandez et al., 2007). In astrocytes, IGF-I-activated CaN dephosphorylates IkBα at serine 32, thus precluding its degradation and NF-kB activation. Interestingly, IGF-I-induced CaN activation in astrocytes requires neither PI3-kinase or Akt (Okubo et al., 1998; Pons & Torres-Aleman, 2000), nor p38 or PLCγ, although elevation of [Ca²⁺]i is essential for the effect since chelating intracellular Ca²⁺ with BAPTA-AM has the same inhibitory effect on IGF-I-induced CaN activation as inhibition of CaN itself with CsA (Pons & Torres-Aleman, 2000). The exact mechanism by which IGF-I generates intracellular signals in astrocytes to activate CaN is, however, yet to be elucidated. When focusing on the IGF-I-induced inhibition of NF-kB, a specific role of FOXO3 and PPARγ has been recently investigated. TNFα stabilizes the interaction of CaNA with FOXO3 and NF-kB generating a complex, which drives the transcriptional activity of NF-kB (Fernandez et al., 2012). Interestingly, both CaN and FOXO3 are required for TNFα-induced NF-kB activation and induction of inflammation-associated genes like COX2 (Fernandez et al., 2012). In this context, TNFα-induced CaN activation does not affect degradation of IkBα (Ronco et al., 2014). To counteract TNFα-induced NF-kB activation, in addition to the dephosphorylation of IkBα by CaN, IGF-I recruits PPARγ to displace FOXO3 from the CaNA/NF-kB complex (Fernandez et al., 2007). In the complex CaNA/NF-kB/PPARγ, PPARγ is sumoilated which results in trans-repression of NF-kB transcription (Glass & Ogawa, 2006; Fernandez et al., 2007). In according with data on cultured astrocytes, the interaction of CaN with NF-kB and FOXO3 occurs in vivo in the brain of the AD model APP/PS1 mice. Moreover, the complex is disrupted by IGF-I or by over-expression of ΔCaNA (Fernandez et al., 2012). The functional outcomes of the ΔCaNA-induced NF-kB inhibition are the resolution of neuroinflammation and the alleviation of AD pathology. Recently, Fernandez and colleagues developed a series of decoy compounds which disrupted the interaction of CaN with FOXO3 in astrocytes thereby causinginhibition of NF-kB activation, resolution of inflammation and protection against Aβ-induced neurotoxicity in APP/PS1 mouse AD model (Fernandez et al., 2016). 

It should be noted that most of the work on the characterization of CaN signalling in the brain has been done in cultured astrocytes or in whole brain tissues in which neurons are also present. Highlighting that the results obtained in cultures provide a fundamental information on interactions between proteins making up signalling cascades, one should be aware that the expression of these signalling components may differ from that of intact astrocytes in the brain. As an example, it has been demonstrated that the procedure of preparation of acute hippocampal brain slices per se induces reactive gliosis (Takanoet al., 2014).

Figure 3. Activation vs inhibition of NF-kB-mediated signalling: role of CaN interacting proteins Upon stimulation with TNFα, CaN recruits and dephosphorylates FOXO3 to interact with and activate NF-kB, which promotes transcription of inflammatory genes. When co-stimulated with IGF-I, CaN recruits PPARγ to displace FOXO3 from the CaN/FOXO3/NF-kB complex generating the CaN/PPARγ/NF-kB complex, thus precluding NF-kB transcriptional activity (Fernandez et al., 2012)

Can astroglial CaN be activated during physiological brain activity?

CaN signalling in astrocytes has been principally described in pathological or pathology-related experimental conditions in which CaN mediates a spectrum of signalling and functional outcomes. Conversely, it is still unclearwhether and how CaN signalling play any role in astroglial physiology. CaN is a renowned Ca2+/calmodulin-dependent master-regulator of cellular remodelling. Therefore, it is plausible to speculate that CaN could participate in structural and functional plastic changes, known also as astroglial plasticity (Shao & McCarthy, 1994; Theodosis & Poulain, 1999; Theodosis et al., 2008; Pirttimaki & Parri, 2013; Cheung et al., 2015), to which astrocytes undergo during neuronal activity. The hint that this could be the case can be taken from the landmark contribution by Filosa and colleagues reporting the CaN-dependent nuclear translocation of NFATc3, in both astrocytes and pericytes in cortical brain slices from ten days old rats,in response to electrical field stimulation (EFS) (Filosa et al., 2007). This effect was reproduced by treating the slices with the agonist of mGluR5 metabotropic glutamate receptors,t-ACPD. Interestingly, NFATc3 translocation in pericytes required functional astrocytes (Filosa et al., 2007). This is the first work reporting the activity-dependent activation of CaN in non-neuronal cells including astrocytes. However, the translation of these results to the activity-dependent CaN activation in adult healthy brain astrocytes is not straightforward. The expression of theCa2+-linked glutamate receptor mGluR5 is developmentally down-regulated and its transcript is not detectable after 3 weeks in mouse cortical and hippocampal astrocytes and in adult human astrocytes (Sun et al., 2013). This finding is supported by the failure of specific agonists (t-ACPD, DHPG) to elicit detectable Ca2+ responses in adult mouse astrocytes (Sun et al., 2013). This suggests thatmGluR5-mediated Ca2+ signals are unlikely to participate in CaN activation in intact adult brain, thereby raising the question about the mechanisms by which the activity-induced or glutamate-mediated Ca2+ elevations in adult astrocytes are generated.Another concern regards the duration and the intensity of Ca2+ transients which are very strong in experimental conditions (both in vitro and in vivo, being of mechanical, electrical or chemical nature), likely involving CICR and SOCE. It is argued that the duration ofspontaneous or evoked Ca2+ transients in healthy astrocytes both in slices (Pivneva et al., 2008; Szokolet al., 2015; Tanget al., 2015) and in intact brain (Takano et al., 2007; Kuchibhotla et al., 2009)in most cases does not exceed one minute and this would not be enough to activate CaN. However, it should be also acknowledged that CaN is sensitive to Ca2+ elevations which are as low as just a doubling of the resting [Ca2+]i , i.e., from about 70 nM to  about 150 nM (Dolmetschet al., 1997; Lim et al., 2013). Yet, such Ca2+ transients might occur just in the population of astrocytes serving activated neurons. It could be that such low Ca2+ elevations are out of the performance range of the available Ca2+probes, or arerecorded as a part of noise.The tridimensional Ca2+ signalling architecture in single astrocytes in intact brain tissue is extremely complex (Volterraet al., 2014); therefore, localizedyet low amplitude Ca2+ transients may be overlooked. Finally, the CaN activation pattern within a cell appears to be not uniform (Mehtaet al., 2014),by exhibiting a steady increase in some locations and an oscillatory pattern in others, making it challenging to detect in intact brain astrocytes.

Table 2. Negative regulation of NFAT/NF-kB-mediated signaling by CaN in astrocytes.

Conclusions

Over the last two decades CaN came on the stage of the astroglial (patho)physiology as a molecular switch for transforming normal astroglial phenotype into reactive astrogliosis. Being a target of a variety of stimuli, from neurotransmitters to inflammatory mediators, CaN drives transcriptional remodelling in astrocytes mainly through activation/modulation of two transcription factors, NFAT and NF-kB. While CaN directly activates NFAT, the modulation of NF-kB pathway is intricate and may result in both activation and repression of transcription. Among the transcriptional targets of astroglial CaN, there are key components of the astrocytic Ca2+signalling toolkit, astroglial marker proteins, glutamate transporters, growth factors and pro-inflammatory mediators. CaN signalling has mainly been studied in disease, while very little is known on the role of CaN in brain physiology. Detailed investigation of CaN-inducing Ca2+ signals and CaN function in astrocyte physiology is challenging but it may give a clue to our understanding of how the brain works in health and how it becomes ill, favouring the identification of pharmacological targets and novel therapeutic approaches for neurodegenerative diseases.