[1] 11w Amyloid Beta Deregulates Astroglial mGluR5-Mediated Calcium Signaling via Calcineurin and NF-kB

[1] 187w The amyloid hypothesis of Alzheimer's disease (AD) suggests that soluble amyloid b (Ab) is an initiator of a cascade of events eventually leading to neurodegeneration. Recently, we reported that Ab deranged Ca 2þ homeostasis specifically in hippocampal astrocytes by targeting key elements of Ca 2þ signaling, such as mGluR5 and IP 3 R1. In the present study, we dissect a cascade of signaling events by which Ab deregulates glial Ca 2þ : (i) 100 nM Ab leads to an increase in cytosolic calcium after 4-6 h of treatment; (ii) mGluR5 is increased after 24 h of treatment; (iii) this increase is blocked by inhibitors of calcineurin (CaN) and NF-kB. Furthermore, we show that Ab treatment of glial cells leads to de-phosphorylation of Bcl10 and an increased CaN-Bcl10 interaction. Last, mGluR5 staining is augmented in hippocampal astrocytes of AD patients in proximity of Ab plaques and co-localizes with nuclear accumulation of the p65 NF-kB subunit and increased staining of CaNAa. Taken together our data suggest that nanomolar [Ab] deregulates Ca 2þ homeostasis via CaN and its downstream target NF-kB, possibly via the cross-talk of Bcl10 in hippocampal astrocytes.

[1] 57w E xtracellular soluble amyloid b (Ab) oligomers trigger the so called ''amyloid cascade'' of events in early Alzheimer's disease (AD) pathogenesis that, with disease progression, leads to neuronal death with concomitant cognitive disturbances (Hardy and Selkoe, 2002). Yet, the exact mechanism by which Ab initiates cellular deregulation is still a question of debate (Hardy and Selkoe, 2002).

[2] 124w Deregulation of cellular Ca 2þ homeostasis has been proposed to play a role in the initial steps of disease progression (Thibault et al., 2007). According to the ''calcium hypothesis'' of AD, a long lasting overload of the cytoplasm and of the endoplasmic reticulum (ER) with Ca 2þ induces activation of mechanisms leading to cell death (Supnet and Bezprozvanny, 2010). The exact mechanism by which this occurs is still a matter of debate (Demuro et al., 2010;Kagan and Thundimadathil, 2010;Kuchibhotla et al., 2008;Tu et al., 2006). A number of effectors downstream of Ca 2þ in AD have been proposed, among which the calcium/calmodulin-dependent phosphatase calcineurin (CaN) and its direct downstream target nuclear factor of activated T cell (NFAT) (Abdul et al., 2011;Reese and Taglialatela, 2011).

[3] 98w Glia is an established partner of neurons in the execution and regulation of brain functions, including synaptic transmission. Recent evidence indicates that it may be an important player in AD pathogenesis (Parpura et al., 2012). Reactive astrocytes, present also in normal aging, are found in AD postmortem brains, as well as in animal models, around amyloid plaques (DeWitt et al., 1998). Furthermore, Ab provokes multiple alterations in glial homeostasis, including View this article online at wileyonlinelibrary.com. DOI: 10.1002/glia.22502 Published online Apr 25, 2013 in Wiley Online Library (wileyonlinelibrary.com). Received Jan 9, 2013, Accepted for publication Mar 5, 2013.

[4] 24w Address correspondence to Armando A. Genazzani, Department of Pharmaceutical Sciences, Universit a degli Studi del Piemonte Orientale ''Amedeo Avogadro,'' Novara 28100, Italy. E-mail: armando.genazzani@pharm.unipmn.it

[5] 38w From the 1 Department of Pharmaceutical Sciences, Universit a degli Studi del Piemonte Orientale ''Amedeo Avogadro,'' Novara, Italy; 2 Department of (Neuro)Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; 3 Stichting Epilepsie Instellingen Nederland (SEIN), Heemstede,

[6] 39w The Netherlands; 4 Swammerdam Institute for Life Sciences, Center for Neuroscience, University of Amsterdam, Amsterdam, The Netherlands.. deregulation of Ca 2þ signaling (Kuchibhotla et al., 2009), transcriptional changes (Peters et al., 2009), and inflammatory responses (Rubio-Perez and Morillas-Ruiz, 2012).

[7] 142w Recently, using an in vitro Tat-Pro-ADAM10 model of AD (Marcello et al., 2007), we have demonstrated that in astrocytes Ab produces alterations of Ca 2þ homeostasis by over-expressing mGluR5 and IP 3 R1 at the transcriptional level (Grolla et al., 2013). In the present work, we dissect the molecular mechanism by which Ab 42 leads to mGluR5 up-regulation in glia and, as a consequence, Ca 2þ -deregulation. We now propose that Ab 42 leads to cytosolic calcium increases, this leads to CaN activation, which in turn (possibly via Bcl10), activates NF-kB-dependent transcription of mGluR5. IP 3 R2 appears to be controlled in a similar manner. We also provide evidence that mGluR5 staining is augmented in hippocampal astrocytes of AD patients in proximity of Ab plaques and is co-localized with nuclear accumulation of the p65 NF-kB subunit and with increased staining of CaNAa.

[1] 99w Ab 42 -Induced Up-regulation of Astroglial Calcium Signaling is Calcineurin and NF-kB-Dependent In the first instance, we performed Fura-2 measurements of Ca 2þ transients induced by the mGluR agonist DHPG (20 lM) in astrocytes pretreated with the CaN inhibitor FK506 and stimulated with Ab 42 for 48-60 h. As shown in Fig. 1, Ab 42 significantly augmented the amplitude of the DHPGinduced Ca 2þ transient (Fig. 1A(b)) as compared to control cells (Fig. 1A(a)). Pre-treatment with FK506 reduced significantly the peak of the Ca 2þ transient and reduced significantly also the fraction of cells responding to DHPG (Fig. 1B).

[2] 123w Recent publications have described CaN-dependent activation of NF-kB in the immune system (Frischbutter et al., 2011;Palkowitsch et al., 2011). Furthermore, analysis of the promoter regions of the mGluR5 gene revealed no presence of NFAT binding site, a direct classical downstream CaN target, but of an NF-kB binding site, present in the second of three active promoter regions (Corti et al., 2003). We thus performed identical experiments using two inhibitors of NF-kB nuclear translocation, namely CAPE and JSH. Pretreatment with CAPE and JSH did not reduce significantly the Ca 2þ peak amplitude, but dramatically reduced the fraction of responding astrocytes (Fig. 1A(d),B). Thus, it is plausible that alterations of mGluR5 signaling, induced by Ab 42 , are controlled by CaN but also by NF-kB.

[3] 220w Ab 42 -Induced CaN-Mediated Up-regulation of mGluR5 and IP 3 R2 is NF-kB-Dependent, but Up-regulation of IP 3 R1 Is Not Treatment of astrocytes with Ab 42 induced a significant increase in mGluR5 expression (Fig. 2A). Pretreatment with inhibitors of both CaN and NF-kB abrogated Ab 42 -induced up-regulation of mGluR5 at both the mRNA (Fig. 2A(a)) and protein levels (Fig. 2A(b)). Previously, we demonstrated that both neuronal and glial splice variants of IP 3 R1 were upregulated in glial cells in a CaN-dependent manner (Grolla et al., 2013). Thus, we investigated whether Ab 42 -induced over-expression of IP 3 R1 was also dependent on NF-kB activation. Figure 2B(a) shows that both FK506 and CsA restored the effect of Ab 42 on IP 3 R1 mRNA levels, but both inhibitors of nuclear NF-kB translocation, CAPE and JSH, failed to do so. The data therefore support the notion that this receptor is regulated by the direct CaN target NFAT (Graef et al., 1999;Groth and Mermelstein, 2003). In astrocytes, however, IP 3 R2 is the dominant isoform of the IP 3 receptors (Sharp et al., 1999). Thus, we decided to analyze IP 3 R2 expression by real-time PCR. IP 3 R2 was up-regulated by Ab 42 treatment and inhibitors of both CaN and NF-kB were able to abolish this up-regulation (Fig. 2B(b)).

[4] 273w Ab 42 -Induced Calcineurin Activation in Astrocytes is Calcium Dependent Next, we investigated possible mechanisms by which Ab 42 may activate CaN and then NF-kB. Several possibilities have been proposed, including the incorporation of the Ab 42 peptide in the plasma membrane and the formation of a pore permeable to cations (Kawahara and Kuroda, 2000) or the activation by Ab 42 of calpain, which in turn can cleave and activate CaN (Abdul et al., 2011). Furthermore, Ab was shown to produce Ca 2þ oscillations in astrocytes in mixed neuronal/glial cultures (Abramov et al., 2004). We therefore investigated whether Ab could lead to cytosolic calcium increases in glial cells. In our hands, acute Ab 42 treatment (up to 1 lM) in primary glial cultures did not produce any notable change in cytosolic Ca 2þ concentrations in the first hour of treatment (data not shown). Yet, after 4-6 h of treatment, cells treated with 100 nM Ab displayed a statistically significant increase in free cytosolic Ca 2þ (Fig. 3A(a)). As shown in Fig. 3A(b), this significant increase was not homogeneous, but was given by a sub-group of astrocytes which increased basal Ca 2þ to 100-150 nM. Pre-incubation with the calcium chelator BAPTA-AM abolished the elevation of cytosolic [Ca 2þ ] (data not shown). Next we investigated the effect of BAPTA on Ab 42 -induced up-regulation of mGluR5 mRNA. One hour pre-incubation with BAPTA-AM completely abolished mGluR5 mRNA up-regulation induced by Ab 42 (Fig. 3B(a)). The same result was obtained for IP 3 R1 and IP 3 R2 genes (Fig. 3B(b,c)), indicating that elevation of cytosolic [Ca 2þ ] is required for the activation of CaN.

[5] 95w We next used a range of Ca 2þ channel blockers to investigate whether we could pin-point the source of the Ca 2þ -increase. Neither nifedipine (100 nM) nor DNQX (30 lM) had a significant effect (data not shown), while 2-APB (100 lM) blocked such up-regulation for all three genes, mGluR5, IP 3 R1, and IP 2 R2 (Fig. 3B), indicating that TRP channels may be implicated in Ab 42 -induced elevation of cytosolic [Ca 2þ ]. Yet, 2-APB has been shown to be a rather non-specific inhibitor, and it is therefore difficult to draw conclusions.

[6] 249w Calcineurin Interacts with and De-phosphorylates Bcl10 in Ab 42 -Treated Glial Cells In the immune system, it has recently been reported that CaN interacts with and de-phosphorylates B cell lymphoma 10 (BCL10) and this leads to activation of NF-kB (Frischbutter et al., 2011;Palkowitsch et al., 2011). Therefore, we investigated whether the two major CaN isoforms, CaNAa and CaNAb, interact with Bcl10 and whether this interaction results in de-phosphorylation of Bcl10. Figure 4A,B shows that both CaNAa and CaNAb interact with Bcl10 when CaNAa or CaNAb were immunoprecipitated using isoform-specific antibodies. Moreover, densitometric analysis revealed that CaNA/Bcl10 interaction is augmented in Ab 42treated glial cultures. Interestingly, both CaN inhibitors, FK506 and CsA, significantly attenuated the interaction of Bcl10 with CaNAa, the most expressed isoform of CaNA, in Ab 42 -treated cells, while neither FK506 nor CsA had any significant effect on Bcl10 interaction with CaNAb. Similar results were obtained when glial lysates were immunoprecipitated with anti-Bcl10 antibody and precipitates were probed with anti-CaNAa or anti-CaNAb antibodies (Fig. 4C). To assess de-phosphorylation of Bcl10 by CaN, we first immunoprecipitated the lysates with anti-Bcl10 antibody and then probed precipitates with antibody recognizing phosphorylated serine residues (p-Ser). To normalize intensities of the bands the ratio Bcl10/p-Ser was used for statistical tests. As shown in Fig. 4D, Ab 42 significantly increased Bcl10/p-Ser ratio indicating augmented of Bcl10 de-phosphorylation. This effect was completely reversed by FK506 and attenuated by CsA. These data indicate that in glial cells CaN may activate NF-kB through de-phosphorylation of Bcl10.

[7] 254w Calcineurin Mediates Ab 42 -Induced NF-kB Activation and p65 Nuclear Translocation in Astrocytes Next, we investigated whether Ab 42 -induced activation of CaN may result in nuclear translocation of NF-kB. Figure 5A shows that the NF-kB-Luc reporter gene was significantly activated by Ab 42 treatment. The effect was abolished when FIGURE 2: Ab 42 -induced and CaN-mediated up-regulation of mGluR5 and IP 3 R2 is NF-kB-dependent, but up-regulation of IP 3 R1 is not. (A) Real-time PCR (a) and WB analysis (b) of mGluR5 in primary glial cultures stimulated with Ab 42 for 24 h or 48 h, respectively. (B) Real-time PCR of IP 3 R1 and IP 3 R2 in primary glial cultures stimulated with Ab 42 for 24 h. Data were normalized to S18 ribosomal protein subunit mRNA or actin and expressed as mean 6 SD reported to control. Data are from at least five independent cultures performed in triplicate (RT-PCR) or from three independent cultures (WB). *P < 0.05; **P < 0.01; ***P < 0.001. cells were pretreated with FK506, CsA, CAPE, or JSH. Then, we performed ICC to visualize accumulation of p65 NF-kB subunit in the nuclear compartment. As shown in Fig. 5B,C, stimulation with Ab 42 for 5-9 h resulted in a clear nuclear localization of p65 in 42.7 6 8.4% of astrocytes which is significantly more than in control cells (11.3 6 3.01%, P < 0.001). Blockers of both CaN and NF-kB strongly inhibited translocation of p65 to the nucleus, confirming the results obtained using the NF-kB-luc reporter.

[8] 115w mGluR5 is Up-regulated in Astrocytes Located Close to Ab Plaques in AD Patient's Hippocampus In adult control hippocampus, mGluR5 is expressed throughout the CA pyramidal cells. No detectable immunoreactivity (IR) was observed in glial cells (Fig. 6A,B). In AD hippocampus, strong mGluR5 IR was detected throughout the different hippocampal regions, particularly around Ab deposits associated with pTAU positive dystrophic neurites (Fig. 6C-F). Double labeling experiments confirmed the increased expression of mGluR5 in astrocytes (GFAP positive cells) of AD patients in proximity of Ab plaques (Fig. 6D-H). Only occasionally co-localization with a microglial marker (HLA-DR) was observed (not shown). Double labeling with p65 and CaNAa showed co-localization with mGluR5 in astrocytes of AD hippocampus (Fig. 6I-J).

[1] 30w In the present study, we have investigated the effects of Ab 42 oligomers on the astroglial signaling cascade, which includes CaN, mGluR5, and IP 3 receptors. Our principal findings are:

[2] 343w (1) in hippocampal astrocytes, activation of CaN by Ab 42 requires elevation of cytosolic [Ca 2þ ] via Ca 2þ entry from the extracellular milieu; (2) CaN activation leads to nuclear translocation of the transcription factor NF-kB, possibly via dephosphorylation of Bcl10, that up-regulates expression of both mGluR5 and IP 3 R2; (3) a similar pathway involving CaN, but not NF-kB, controls IP 3 R1 expression; and (4) these transcriptional changes have repercussions on mGluR5 activation and calcium homeostasis in glial cells. We also provide evidence that this pathway may be relevant to AD: in the hippocampus of AD patients, mGluR5 is found to be over-expressed in concomitance with over-expression of CaNAa and with p65 NF-kB subunit in GFAP-positive astrocytes located in proximity to Ab aggregates. These data A, B) Representative results and densitometric analysis of the effect of Ab 42 , FK506, or CsA on the interaction of Bcl10 with CaNAa and with CaNAb. Lysates of primary glial cultures treated with Ab 42 for 5 h were immunoprecipitated with anti-CaNAa and anti-CaNAb primary antibody. Precipitates were probed with anti Bcl10 antibodies (upper bands) and with anti-CaNA (lower bands). Data in histograms are expressed as ratio of Bcl10/CaN, expressed as mean 6 SD and reported to control. The differences are significant at P < 0.05 for three independent IP for each condition. (C) Representative results and densitometric analysis of the effect of Ab 42 , FK506, or CsA on the de-phosphorylation of Bcl10 and interaction with CaNAa and with CaNAb. After 5 h of incubation with Ab 42 , cells were lysed and immunoprecipitated with primary anti-Bcl10 antibody. The precipitates were probed with anti-phospho serine (p-Ser), anti-Bcl10, anti-CaNAa, and anti CaNAb primary antibodies. (D) The Bcl10/p-Ser ratios are expressed as mean 6 SD and reported to control. Differences are significant at P < 0.05 for 6 IPs from three independent glial cultures. are in line with previous report by Norris et al. (2005) describing CaN-immunoreactive astrocytes surrounding amyloid plaques in APP/PS1 Tg mice. We therefore propose the signaling cascade illustrated in Fig. 7.

[3] 104w It should be acknowledged that while our claim is that the data presented is relevant to AD, there is still controversy on whether Ab is the key driver of the disease, or whether it represents a corollary phenomenon, for example of a more generalized protein mis-processing (Aguzzi and Haass, 2003). Furthermore, it remains to be ascertained whether Ab induces a specific astrocyte activation or whether other astrocyte insults are able to induce similar effects. Therefore, a rearrangement of the calcium signaling machinery via CaN could be a specific phenomenon linked to AD or might have a more general function in reactive astrocytes or astrogliosis.

[4] 92w Furthermore, while the obvious context of our experiments would be placed around an increase in Ab from neurons, we and others have previously shown that Ab can be produced by astroglia (Bettegazzi et al., 2011;Grolla et al., 2013;Rossner et al., 2005). In this context, our experiments could also support a primary pivotal role of astroglia in the pathogenesis of the disease, as suggested by others (Kulijewicz et al., 2012;Yeh et al., 2011). Indeed, early astrocyte atrophy has been reported in a mouse model of AD (Rodriguez and Verkhratsky, 2011;Verkhratsky et al., 2012).

[5] 333w Independently of these considerations, there is an increasing evidence that CaN is involved in AD pathogenesis (Reese and Taglialatela, 2011). Activation of CaN specifically in astrocytes has been proposed to have a role in AD and in other pathological conditions (Abdul et al., 2009;Fernandez et al., 2012;Furman et al., 2012;Grolla et al., 2013;Jin et al., 2012;Norris et al., 2005;Sama et al., 2008) although the mechanisms of its activation remain largely unknown. CaN could be directly activated by calcium entering via channel-like structures formed by Ab (Kagan and Thundimadathil, 2010), via Ca 2þ -permeable channels on the plasma membrane modulated by Ab (Pellistri et al., 2008), or indirectly via the cleavage of the CaNA auto-inhibitory domain by calpain (Abdul et al., 2011). In our experiments, we used nanomolar [Ab 42 ], compatible with the [Ab] found in the cerebro-spinal fluid of AD patients (Mehta et al., 2001). We report that the acute treatment of cultured astrocytes with 100 nM Ab 42 did not change baseline [Ca 2þ ] for up to 1 h of recording. However, when cells were taken for Ca 2þ imaging 4-6 h after Ab 42 addition, a fraction of cells had elevated [Ca 2þ ]cyt in a range of 100-150 nM, which is in line with [Ca 2þ ]cyt necessary to activate CaN in cerebellar granule neurons (Guerini et al., 1999). Abdul et al. (2009) reported nuclear translocation of NFAT-EGFP reporter already 15 min after addition of similar [Ab], indicating that in their experiment Ab-induced CaN activation occurred significantly earlier than in our experiments. At least two factors may account for this discrepancy: (i) different procedures of preparation of Ab which could affect effective concentration oligomeric Ab; and (ii) different protocols of preparations primary astroglial cultures. Furthermore, all CaN-dependent changes in mRNA expression were abolished by clamping cytosolic Ca 2þ , and indicate that involvement of cation channels is possible in the Ab 42 -induced Ca 2þ entry. Last, we find that 2-APB, a nonspecific cation blocker abolishes the effect.

[6] 266w CaN inhibitors have been widely used as immunosuppressants and most of their actions can be reconciled with inhibition of NFAT-dependent transcription (Lee and Burckart, 1998). Astroglial CaN-NFAT signaling has been characterized by several groups (Canellada et al., 2008;Furman et al., 2010;P erez-Ortiz et al., 2008;Sama et al., 2008). Recently, it was proposed that CaN is an essential for activation of NF-kB in immune cells (Frischbutter et al., 2011;Palkowitsch et al., 2011) where Bcl10, a member of the so called CBM complex, interacts with and is de-phosphorylated by CaN to recruit other two CBM members, mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1) and caspase recruitment domain membrane-associated guanylate kinase protein 1 (CARMA1) in a ternary complex, whose downstream effects are degradation of IkB and nuclear translocation of NF-kB. This may therefore constitute a parallel pathway that mediates CaN's actions in the immune system. Thus, we decided to investigate whether CaN interacts with Bcl10 also in glial cells. We found that, in fact, this was the case. Both CaNAa and CaNAb co-immunoprecipitated with Bcl10 in glial cultures. Moreover, in our experiments, stimulation with Ab 42 significantly augmented this interaction in which Bcl10 was also dephosphorylated. As the effect of CaN and NF-kB inhibitors overlap in our experiments for mGluR5 and IP 3 R2, we propose this pathway as being responsible. Yet, it has been proposed also that, in astrocytes, CaN activates NF-kB via interaction with Forkhead box O (FoxO) transcription factor 3 (Foxo3) in TNFa-stimulated cells (Fernandez et al., 2012). Interestingly, Foxo3 may also be activated downstream of Bcl10 and Ikk signaling cascade (Luron et al., 2012).

[7] 60w The Ca 2þ -CaN-NF-kB pathway activation, in our system, leads to a profound remodeling of Ca 2þ -handling capacity of astrocytes. As such, it might be suggested that, in AD brains, this would lead to profound changes in astroglial Ca 2þ -dependent processes, including gliotransmission and reactive inflammation. As such, these could participate actively in the pathogenesis of the disease.

[8] 37w In support of the occurrence of these phenomena in AD, we also show that in hippocampi from human AD brains, mGluR5 was over-expressed with CaNAa in GFAPpositive astrocytes in proximity to amyloid plaques and colocalized with p65.

[1] 95w Primary hippocampal astroglial cell cultures were prepared from postnatal days 1-3 (P1-P3) rat hippocampi as described previously (Fresu et al., 1999). Hippocampal glial cells were seeded in Dulbecco's Modified Eagle's Medium, supplemented with 10% fetal bovine serum, 2 mg/mL glutamine, 10 U/mL penicillin, and 100 lg/mL streptomycin (Sigma, Milan, Italy). Cells were grown until confluence (2-4 days) and then were re-plated for the experiments on plates coated with 0.1 mg/mL poly-L-lysine. The purity of cultures was assayed by immunostaining with anti-MAP2 (neuronal marker) and anti-GFAP (glial marker). No neurons were detected in glial primary cultures.

[2] 101w Ab 1-42 (Ab 42 ) was purchased either from Bachem (Bubendorf, Switzerland) or from Innovagen (Lund, Sweden). Ab 42 oligomers were prepared as described by Giuffrida et al. (2009) with some modifications. Briefly, the peptide was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP), Fluka Cat. 52512 to 1 mg/mL, monomerized by 1 h incubation at 37 C, lyophilized, resuspended in dimethylsulfoxide (DMSO) at 5 mM, diluted in ice-cold minimum essential medium (MEM) to 100 lM and oligomerized for 24 h at 4 C. The peptide was snap frozen and kept at À80 C. Unless otherwise stated, final concentration of Ab 42 was 100 nM.

[3] 63w About 1 Â 10 5 cells/well were plated in 24-well plates and 12-24 h after plating were transfected with an NF-kB-Luc reporter plasmid (Clontech) using Lipofectamine 2000 (Life Technologies, Milan). Twenty-four hours after transfection, cells were treated with inhibitors and with Ab 42 for 12-15 h. Luciferase activity was assayed using Bright-Glo Luciferase Assay System (Promega, Milan, Italy) according to the manufacturer's instruction.

[4] 222w Cells were loaded with Fura-2 AM as described by Grolla et al. (2013). After de-esterification (30 min at room temperature (RT)) the coverslip was mounted in an acquisition chamber and placed on the stage of a Leica epifluorescent microscope equipped with a S Fluor Â40/1.3 objective. Cells were excited alternatively with 340/ 380 nm using a monochromator Polichrome V (Till Photonics, Munich, Germany), and the fluorescence light, filtered through a bandpass 510 nm filter was collected by cooled CCD camera (Hamamatsu, Japan) and acquired by MetaFluor software. To quantify the differences in the peaks of Ca 2þ transients the ratio values were normalized using the formula (F i À F o )/F o (referred to as normalized Fura-2 ratio, norm. ratio). The cells with norm. ratio above 0.2 were considered as responders. For baseline Ca 2þ measurements, cells were treated with Ab 42 for 3 h, then 2 lM Fura-2 was added for 30 min directly to culture medium and the cells were transferred to room temperature to avoid Fura-2 compartmentalization. At the end of each recording, cells were perfused with a solution containing 10 lM ionomycin with either 10 mM Ca 2þ or 25 mM EGTA. Concentrations of Ca 2þ were calculated according to the work by Grynkiewicz et al. (1985). Data were analyzed using GraphPad Prism Software (San Diego, CA).

[5] 103w Total mRNA was extracted from 7.0 Â 10 5 cells using QIAzol Lysis Reagent (Qiagen, Milan, Italy) according to manufacturer's instructions. First strand of cDNA was synthesised from 1 lg of total RNA using ImProm-II RT system (Promega). Real-time PCR was performed using GoTaq qPCR Master Mix (Promega) on an SFX96 Real-Time System (Biorad, Segrate, Italy). S18 ribosomal protein was used to normalize PCR product levels. Following oligonucleotide primers were used (from 5 0 to 3 0 ): S18 (NM_213557) forward (Forw)-TGCGAGTACTCAACACCAACA, reverse (Rev) CTGCTT-TCCTCAACACCACA; mGluR5 (NM_017012) Forw GCCATGG-TAGACATAGTGAAGAGA, Rev TAAGAGTGGGCGATGCAAAT; IP3R1 (NM_001007235) Forw GGCTACAGAGTGCCTGACCT, Rev CCATTCGTAGATCCCTCTGC; IP3R2 (NM_031046) Forw TCCAAAAGACGTTGGACACA, Rev TTCATCCCCTTCCTCT-GGAT.

[6] 106w Twenty-four hours before treatment, 5 Â 10 4 glial cells were plated onto 13 mm coverslips in 24-well plates. Treated cells were fixed in 4% formaldehyde in PBS for 15 min at RT, permeabilized for 7 min in PBS with 0.1% Triton X-100 and blocked for 30 min in 2% gelatine. Then primary (1 h, 37 C) and secondary (1 h, RT) antibody were applied in PBS with 2% gelatin. After washing (3Â5 min), nuclei were stained with 4 0 ,6-diamidino-2-phenylindole dihydrochloride (DAPI) for 15 min at RT. Fluorescence images were acquired using a Leica epifluorescent microscope equipped with S Fluor Â40/1.3 objective using MetaMorph software.

[7] 187w For immunoprecipitation (IP), 3-5 Â 10 5 glial cells were plated in 60-mm Petri dishes. On the day of experiment, cells were pretreated with inhibitors and then treated with Ab 42 for 6 h. Cells were then scraped on ice in 300 ll IP buffer (50 mM Tris-HCl, pH ¼ 7.4, 150 nM NaCl, 1% NP-40) supplemented with protein inhibitors cocktail (PIC), 0.1 mM phenylmethanesulfonylfluoride (PMSF), and phosphatase inhibitors cocktail (all from Sigma), quantified with Micro BCA Protein Assay Kit (Pierce, Rockford, IL). About 500 lg of total proteins were immunoprecipitated using A/G agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA) according to manufacturer's instructions with 2 lg primary anti-Bcl10, anti-CnAa, or anti-CnAb antibodies in 0.5 mL at 4 C O/N on a rotator shaker. After intensive washing, precipitates were dissolved in 100 ll of 1Â Laemmli sample buffer and 30 ll were used for WB. The raw densitometric data were expressed as ratio of Bcl10 to CaNA in Fig. 4A,B, and as ratio of Bcl10 to p-Ser band intensities in Fig. 4C,D. The ratios then were expressed as fold of increase as compared to control samples.

[8] 60w For total lysates, cells were treated with Ab 42 for 48-60 h, scraped in IP buffer and total proteins were quantified. About 30-50 lg of total proteins were resolved in 5-12% gradient sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and blotted onto nitrocellulose membrane (GE Healthcare, Milan, Italy). Densitometric analysis was performed with Quantity One v. 4.6 software (Bio-Rad, Hercules, CA).

[9] 122w The subjects included in this study were selected from the databases of the Departments of Neuropathology of the Academic Medical Center, University of Amsterdam (The Netherlands). Informed consent was obtained for the use of brain tissue and for access to medical records for research purposes. Tissue was obtained and used in a manner compliant with the Declaration of Helsinki. We included six hippocampal specimens of patients with AD (Braak stage V and VI) and six hippocampal specimens obtained at autopsy from controls (without evidence of degenerative changes, and lacking a clinical history of cognitive impairment; Table 1). All AD cases were pathologically staged according to Braak and Braak criteria (Braak et al., 2006). All autopsies were performed within 24 h after death.

[10] 66w One or two representative paraffin blocks per case (hippocampus) were sectioned, stained, and assessed. Formalin fixed, paraffin-embedded tissue was sectioned at 6 lm and mounted on pre-coated glass slides (Star Frost, Waldemar Knittel GmbH, Braunschweig, Germany). Sections of all specimens were processed for haematoxylin eosin (HE), luxol fast blue (LFB), and Nissl stains as well as for immunocytochemical stainings for a number of markers described below.

[11] 150w Glial fibrillary acidic protein (GFAP; polyclonal rabbit, DAKO, Glostrup, Denmark; 1:4,000), neuronal nuclear protein (NeuN; mouse clone MAB377, IgG1; Chemicon, Temecula, CA; 1:1,000), human leukocyte antigen (HLA)-DP, DQ, DR (HLA-DR; major histocompatibility complex class II, MHC-II; mouse clone CR3/43; DAKO, Glostrup, Denmark, 1:400), Ab (Mouse clone 6F/3D; DAKO; 1:200), phosphorylated Tau (pTau; mouse clone AT8; Innogenetics, Alpharetta, GA; 1:5,000) were used in the routine immunocytochemical analysis. For the detection of mGluR5 we used two antibodies (polyclonal rabbit ab53090, from Abcam, 1:100; polyclonal rabbit from Upstate Biotechnology, Lake Placid, NY; 1:100). For the detection of p65 we used a polyclonal rabbit (sc-372; Santa Cruz; 1:50) and for the detection of CaNAa a polyclonal goat (sc-6123, Santa Cruz; 1:50) was used. Immunohistochemistry was carried out as previously described (Aronica et al., 2003). Single-label immunohistochemistry was developed using the Powervision kit (Immunologic, Duiven, The Netherlands) with 3,3-diaminobenzidine (Sigma, St. Louis, USA) as chromogen.

[12] 149w For double-labeling sections were incubated with Brightvision poly-alkaline phosphatase (AP)-anti-Rabbit (Immunologic, Duiven, The Netherlands) for 30 min at room temperature, and washed with PBS. Sections were washed with Tris-HCl buffer (0.1 M, pH 8,2) to adjust the pH. AP activity was visualized with the alkaline phosphatase substrate kit I Vector Red (SK-5100, Vector laboratories Inc., CA). To remove the first primary antibody sections were incubated at 121 C in citrate buffer (10 mM NaCi, pH 6.0) for 10 min. Incu-bation with the second primary antibody was performed overnight at 4 C. Sections with primary antibody other than rabbit were incubated with post antibody blocking from the Brightvisionþ system (containing rabbit-a-mouse IgG; Immunologic, Duiven, The Netherlands). AP activity was visualized with the AP substrate kit III Vector Blue (SK-5300, Vector laboratories, CA). Sections incubated without the primary Abs or with the primary antibodies, followed by heating treatment were essentially blank.

[13] 89w For the double-label immunofluorescent staining, after incubation with the primary antibodies overnight at 4 C, incubated for 2h at room temperature with Alexa FluorV R 568-conjugated anti-rabbit and Alexa FluorV R 488 anti-mouse IgG or anti-goat IgG (1:100, Molecular Probes, The Netherlands). Sections were mounted with Vectashield containing DAPI (targeting DNA in the cell nucleus; blue emission) and analyzed by means of a laser scanning confocal microscope (Leica TCS Sp2, Wetzlar, Germany). Sections were then analyzed by means of a laser scanning confocal microscope (Leica TCS Sp2, Wetzlar, Germany).

[1] 96w Primary antibodies to: mGluR5 (ab53090, for Western Blot (WB) 1:500) was from Abcam (Cambridge, UK). Anti-IP 3 R1 (WB 1:500) antibody was a kind gift from Dr. Colin Taylor (University of Cambridge). Antibodies to p65 (sc-372, for immunocytochemistry (ICC) 1:50), CaNAa (sc-6123, WB 1:200), CaNAb (sc-6124, WB 1:200), Bcl10 (sc-5611, WB 1:500) were from Santa Cruz (Santa Cruz, CA). Anti-phosphoserine (p-Ser, ALX-804-167, WB 1:1,000) was from Alexis (Enzo Life Sciences, Lausen, Switzerland); anti-b-actin (A1978, WB 1:4,000) was from Sigma. AlexaFluor 488 secondary antibodies were from Life Sciences (Milan, Italy), peroxidase-conjugated secondary antibodies from Pierce (Rockford, IL).

[2] 68w In all experiments, FK506, cyclosporine A, caffeic acid phenethyl ester (CAPE) and 4-methyl-N 1 -(3-phenylpropyl)benzene-1,2diamine (JSH-23, JSH) (all from Sigma, Milan, Italy) were used at 100 nM, 1 lM, 40 lM, and 20 lM, respectively, 1 h before treatment with Ab 42 . (S)-3,5-Dihydroxyphenylglycine (DHPG), 2-aminoethoxydiphenylborane (2-APB), nifedipine, 6,7-dinitroquinoxaline-2,3-dione (DNQX) (all from Tocris, Bristol, UK) were used at 20 lM, 100 lM, 100 nM, and 30 lM, respectively.