Amyloid beta peptide 1–42 disturbs intracellular calcium homeostasis through activation of GluN2B-containing N-methyl-d-aspartate receptors in cortical cultures

Cell Calcium 51 (2012) 95–106

Contents lists available at SciVerse ScienceDirect

Cell Calcium journal homepage: www.elsevier.com/locate/ceca

Amyloid beta peptide 1–42 disturbs intracellular calcium homeostasis through activation of GluN2B-containing N-methyl-d-aspartate receptors in cortical cultures I.L. Ferreira a , L.M. Bajouco a , S.I. Mota a , Y.P. Auberson c , C.R. Oliveira a,b , A.C. Rego a,b,∗ a

CNC-Center for Neuroscience and Cell Biology, University of Coimbra, Portugal Faculty of Medicine, University of Coimbra, Coimbra, Portugal c Novartis Institutes of Biomedical Research, Novartis Pharma AG, CH-4002 Basel, Switzerland b

a r t i c l e

i n f o

Article history: Received 18 April 2011 Received in revised form 4 November 2011 Accepted 17 November 2011 Available online 15 December 2011 Keywords: Alzheimer’s disease Amyloid beta peptide (A␤) Calcium N-Methyl-d-aspartate receptor GluN2A subunit GluN2B subunit

a b s t r a c t Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that leads to debilitating cognitive deficits. Recent evidence demonstrates that glutamate receptors are dysregulated by amyloid beta peptide (A␤) oligomers, resulting in disruption of glutamatergic synaptic transmission which parallels early cognitive deficits. Although it is well accepted that neuronal death in AD is related to disturbed intracellular Ca2+ (Ca2+ i ) homeostasis, little is known about the contribution of NMDARs containing GluN2A or GluN2B subunits on A␤-induced Ca2+ i rise and neuronal dysfunction. Thus, the main goal of this work was to evaluate the role of NMDAR subunits in dysregulation of Ca2+ i homeostasis induced by A␤ 1–42 preparation containing both oligomers (in higher percentage) and monomers in rat cerebral cortical neurons. The involvement of NMDARs was evaluated by pharmacological inhibition with MK-801 or the selective GluN2A and GLUN2B subunit antagonists NVP-AAM077 and ifenprodil, respectively. We show that A␤, like NMDA, increase Ca2+ i levels mainly through activation of NMDARs containing GluN2B subunits. Conversely, GluN2A-NMDARs antagonism potentiates Ca2+ i rise induced by a high concentration of A␤ (1 ␮M), suggesting that GluN2A and GluN2B subunits have opposite roles in regulating Ca2+ i homeostasis. Moreover, A␤ modulate NMDA-induced responses and vice versa. Indeed, pre-exposure to A␤ (1 ␮M) decrease NMDA-evoked Ca2+ I rise and pre-exposure to NMDA decrease A␤ response. Interestingly, simultaneous addition of A␤ and NMDA potentiate Ca2+ I levels, this effect being regulated by GluN2A and GluN2B subunits in opposite manners. This study contributes to the understanding of the molecular basis of early AD pathogenesis, by exploring the role of GluN2A and GluN2B subunits in the mechanism of A␤ toxicity in AD. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Alzheimer’s disease (AD) is the leading cause of dementia in western countries and the most prevalent neurodegenerative disease in the elderly population, affecting 26.6 million people worldwide [1]. Age-related forms of dementia lead to sporadic AD. Conversely, less than 10% of cases are associated with familial AD, due to mutations in either amyloid precursor protein, presenilin-1 or presenilin-2 genes. AD hallmarks include atrophy in the cortex, hippocampus and amygdala [2]. Neuropathologically, AD is characterized by senile plaques, composed of extracellular deposits of

∗ Corresponding author at: CNC-Center for Neuroscience and Cell Biology University of Coimbra, Largo Marquês de Pombal, 3004-517 Coimbra, Portugal. Tel.: +351 239 820190; fax: +351 239 822776. E-mail addresses: [email protected], [email protected], [email protected] (A.C. Rego). 0143-4160/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ceca.2011.11.008

amyloid-beta peptides (A␤) and intracellular neurofibrillary tangles formed by hyperphosphorylated tau [2,3]. A␤ is produced by proteolytic cleavage of APP by sequential activity of ␤- and ␥secretases, producing A␤1–42 and A␤1–40 [4,5]. Neurodegeneration and synaptic dysfunction induced by A␤ involves overactivation of the N-methyl-d-aspartate (NMDA) receptors (NMDARs) resulting in the elevation of intracellular Ca2+ i levels, a process named excitotoxicity [6–10]. Activation of NMDARs was hypothesized to occur at late-stage AD, when plaque formation is expected. However, recent reports strongly suggest that glutamate receptors are dysregulated by A␤ accumulation in the initial stages of AD, resulting in disruption of glutamatergic synaptic transmission, which parallels early cognitive deficits [11]. Thus, early phases of AD (characterized by the presence of A␤ monomers and oligomers) are linked to NMDAR-induced synaptic dysfunction, which appears to precede neurodegeneration [7,12–14]. Accordingly, recent studies suggest that A␤ oligomers are the main neurotoxic species involved early in AD [15,16].

96

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

Indeed, oligomeric species were shown to be more toxic to cortical neurons than fibrillar forms, the main components of senile plaques [17]. In addition, it has been reported that A␤1–42 oligomers are much more prone to aggregation and are more neurotoxic than those composed by the A␤1–40 peptide [5,18]. NMDAR functional downregulation is thought to take place during the initial stages of AD. Supporting this view, many studies have suggested that A␤ may reduce surface GluN1 subunit of NMDARs, impairing its function [19–21], leading to the depressed synaptic glutamatergic transmission observed in AD. Three families of genes (GluN1, GluN2 and GluN3) have been identified that encode NMDAR subunits [22]. Functional NMDARs are heterotetramers composed of two glycine-binding GluN1 subunits and two glutamate-binding GluN2 (GluN2A-GluN2D) subunits, or in some cases GluN3 (GluN3A and/or GluN3B) subunits, the latter replacing the GluN2 subunits [23]. The most widely expressed NMDARs contain the obligatory subunit GluN1 plus either GluN2B or GluN2A or a mixture of the two. GluN2B and GluN2D are expressed at high levels in early developmental stages (prenatally), whereas GluN2A and GluN2C expression is first detected near birth [24]. In adults, GluN2A is ubiquitously expressed in the brain, GluN2B is mostly restricted to the forebrain, GluN2C is restricted to the cerebellum, and GluN2D is expressed in small numbers of cells in selected brain regions [25]. It has been recently proposed that GluN2A- and GluN2Bcontaining NMDAR are linked to different intracellular cascades, participating in different functions, from synaptic plasticity to pathological conditions [26–28]. Although it is well accepted that neuronal death in AD is related to disturbed Ca2+ i homeostasis involving the NMDAR [6,10,29,30] and that early NMDAR dysregulation occur in AD, little is known about the contribution of the GluN2A or GluN2B subunits on A␤-induced neuronal dysfunction. In this work we evaluated the contribution of the NMDARs subunits GluN2A and GluN2B, on Ca2+ i dysregulation induced by direct exposure to A␤1–42. This study supports the hypothesis that GluN2A and/or GluN2B subunits of the NMDARs are important in the initial stages of AD pathogenesis and are involved in neuronal dysfunction induced by A␤, leading to cortical neurodegeneration.

2. Materials and methods 2.1. Materials Neurobasal medium and B27 supplement were purchased from GIBCO (Paisley, UK). Ifenprodi, resazurin and anti-␣-tubulin was from Sigma Chemical Co. (St. Louis, MO, USA). NMDA was obtained from Tocris (Cookson, UK) and synthetic amyloid-beta 1–42 peptide from Bachem (Bubendorf, Switzerland). (+)-5Methyl-10,11-dihydro-5H-dibenzo [a,d] cyclohepten-5,10-imine maleate (MK-801) was obtained from Calbiochem (Darmstadt, Germany); [(R)-[(S)-1-(4-bromophenyl)-ethylamino]-(2,3-dioxo1,2,3,4 tetrahydroquinoxalin-5-yl)-me thyl]-phosphonic acid (NVP-AAM077) was a generous gift from Novartis Pharma AG, Basel, Switzerland. Sulfo-NHS-SS-biotin and NeutrAvidinTM were obtained from Pierce (Rockford, IL, USA). The antibody against the denatured form of GluN1 subunit was obtained from Cell Signaling, whereas anti-GluN2A and GluN2B were from Millipore Chemicon (Peniecula, CA, USA). Secondary antibodies conjugated to alkaline phosphatase (anti-mouse and anti-rabbit) were purchased from Amersham Biosciences (Buckinghamshire, UK). The fluorescence probe Fura-2/AM and the antibody against transferrin receptor were obtained from Molecular Probes (Invitrogen, USA). All other reagents were of analytical grade.

2.2. Primary cortical cultures Primary cultures of rat cerebral cortex were prepared as described previously [31] with some minor modifications. Briefly, frontal cerebral cortices free of meninges, were dissected out from Wistar fetal rats at embryonic 16 day and collected in Ca2+ , Mg2+ free Krebs medium (120 mM NaCl, 4.33 mM KCl, 1.2 mM KH2 PO4 , 25.5 mM NaHCO3 , 13 mM glucose, 10 mM Hepes, pH 7.4), containing 0.3% fatty acid-free BSA. Tissues were then treated with 0.035% trypsin plus 0.004% deoxyribonuclease I in BSA-Krebs medium, for 7 min at 37 ◦ C, followed by addition of 0.038% trypsin inhibitor in order to block enzymatic digestion. Cells were resuspended in Krebs medium and centrifuged at 1000 rpm for 5 min and then plated at a density of 0.63 × 106 cells/cm2 in both poly-d-lysine coated 96- or 6-well plates for intracellular Ca2+ measurements and Western Blotting, respectively, in Neurobasal Medium supplemented with 2% B27 supplement, 0.5 mM glutamine and 50 ␮g/ml gentamicin. Cells were cultured for 8 days, in a humidified incubator chamber with 95% air and 5% CO2 at 37 ◦ C. Cortical cultures at 8 days in vitro (DIV) contained few glial cells (less than 10%) as assessed using antibodies against the neuronal marker microtubule associated protein-2 (MAP-2) and the marker of astrocytic proliferation glial fibrillary acidic protein (GFAP) [32]. In cells cultured for 15 days, half medium was changed at day 8 in culture. All animal experiments were carried out following the Guide for laboratory animal practice of the Center for Neuroscience and Cell Biology, University of Coimbra, with care to minimize the number of animals and their suffering. 2.3. Preparation of amyloid-ˇ peptide A peptide preparation containing a high percentage of A␤ oligomers and monomers, previously described as ADDLs (Abetaderived diffusible ligands), was made from synthetic A␤1–42 peptide (henceforward referred to as A␤), as previously described [5,17]. Briefly, synthetic A␤ peptide was dissolved in 1,1,1,3,3,3hexafluoro-2-propanol (HFIP) to a final concentration of 1 mM. HFIP was then removed in a Speed Vac (Ilshin Lab. Co. Ltd., Ede, The Netherlands), and dried HFIP film was stored at −20 ◦ C. The peptide film was resuspended to make a 5 mM solution in anhydrous dimethyl sulfoxide. A␤ peptides were further prepared by diluting the solution in phenol red-free Ham’s F-12 medium without glutamine to a final concentration of 100 ␮M and incubated overnight at 4 ◦ C. The preparation was centrifuged at 15,000 × g for 10 min at 4 ◦ C to remove insoluble aggregates, and the supernatant containing soluble oligomers and monomers was transferred to clean tubes and stored at 4 ◦ C. Protein concentrations of A␤ were then determined using the Bio-Rad protein dye assay reagent. Samples containing 10 ␮g of protein were diluted (1:2) with sample buffer (containing: 40% glycerol, 2% SDS, 0.2 M Tris–HCl, pH 6.8 and 0.005% Coomassie G-250). The presence of different assembly peptide forms (monomers, oligomers and/or fibrils) in the preparation was evaluated by 4–16% Tris–Tricine SDS-PAGE gel electrophoresis and further stained with Coomassie blue. 2.4. Western blot for GluN1, GluN2A and GluN2B subunits Cellular extracts were performed in the embryo frontal cortex at day 0 (directly from the cortical tissue) or in cells cultured for 8 or 15 DIV, using the Ripa buffer (containing 150 mM NaCl, 50 mM Tris, 5 mM EGTA, 1% Triton X-100, 0.5% DOC, 0.1% SDS), supplemented with 1 mM DTT, 1 mM PMSF and 1:1000 protease inhibitor cocktail (chymostatin, pepstatin A, leupeptin and antipain). Protein content was determined by BioRad method, and the samples were denaturated with 6 times concentrated denaturating buffer at 95 ◦ C, for 5 min. Equivalent amounts of protein (15 ␮g) were separated

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

on a 7% SDS-PAGE gel electrophoresis and electroblotted onto polyvinylidene difluoride (PVDF) membranes. The membranes were further blocked with 5% fat-free milk and incubated with antibodies directed against GluN1 (1:1000), GluN2A (1:1000) or GluN2B (1:500) subunits. To control for loading of the gels, an anti-tubulin antibody was used. Immunoreactive bands were visualized by alkaline phosphatase activity after incubation with ECF reagent on a BioRad Versa Doc 3000 Imaging System. 2.5. Cell surface biotinylation Cortical cells cultured for 8 DIV were subjected to surface biotinylation in order to evaluate cell surface receptors, according to Kurup and collaborators [33] with some minor modifications. Cells were rinsed twice in ice-cold PBS and further incubated in PBS containing 1.5 mg/ml EZ-LinkTM sulfo-NHS-SS-biotin, for 20 min, at 4 ◦ C. The non-bound biotin was removed by washing the cells in PBS, and cell lysates were prepared in PBS containing protease inhibitors, 0.1% SDS and 1% Triton X-100. The lysates were sonicated for 30 s and centrifuged for 5 min at 20,800 × g (4 ◦ C); the supernatant was further incubated with 50 ␮l of NeutrAvidinTM plus beads, for 2 h, in a rotary shaker with gentle agitation, at 4 ◦ C. Biotinylated proteins were washed 3 times at 3000 × g for 3 min (at 4 ◦ C) and then eluted with denaturating buffer at 95 ◦ C for 5 min and centrifuged again at 20,800 × g, for 5 min, by using spinX centrifuge tube (0.45 ␮m filter). Samples were then processed by Western Blotting for analysis of surface expression of GluN2A and GluN2B, as described above, and the transferrin receptor was used as a loading control. 2.6. Alamar blue reducing assay Cortical neurons were exposed to 0.5 or 1 ␮M soluble A␤ 1–42 preparation described above for 6 h, in conditioned culture

97

medium. When indicated, neurons were exposed to 100 ␮M NMDA plus 20 ␮M glycine in Mg2+ -free Na+ medium (to selectively activate the NMDARs), containing (in mM) 140 NaCl, 5 KCl, 1 CaCl2 , 10 Glucose, 10 Hepes, pH 7.4/NaOH, for 15 min. In some experiments, after acute exposure to treatments, cells were incubated for further 18 h (‘recovery period’) in previously collected conditioned culture medium without the injury stimuli, in order to evaluate the delayed effects of A␤ or NMDA treatments. Pharmacological inhibition was achieved by incubating the cells in the absence or presence of NMDA receptor antagonists, namely ifenprodil (10 ␮M), which selectively inhibits GluN2B-NMDARs, NVP-AAM077 (50 nM), a selective antagonist of GluN2A-NMDARs or MK-801 (10 ␮M), an antagonist of all NMDAR subtypes. In the concentration used in this study, NVP-AAM077 is predicted to impair activation of GluN2A-NMDARs, but not GluN2B-NMDARs [34,35], whereas ifenprodil selectively inhibits GluN2B-NMDARs [36]. In these cases, 5 min pre-incubation with NMDARs antagonists was performed and maintained during treatment with NMDA. Alamar blue, a cell viability indicator that uses the natural reducing ‘power’ of living cells to convert resazurin to the fluorescent molecule resorufin, was added to treated or untreated cortical cells in Na+ medium at a final concentration (1:100), for 1 h at 37 ◦ C. The absorbance was detected at 570 nm (reference: 600 nm) by using a microplate reader Spectra Max Plus 384 (Molecular Devices, USA) and results were normalized to the percentage of control (untreated cells). 2.7. Intracellular Ca2+ recording Cortical cultures were incubated in fresh Neurobasal medium without added B27 supplement in the presence of the fluorescent probe Fura-2/AM (10 ␮M) in the incubator chamber, at 37 ◦ C for 1 h. After a washing step, Ca2+ i levels were measured in cell populations subjected to NMDA (10 or 100 ␮M) or A␤ (0.5 or 1 ␮M)

Fig. 1. Expression of NMDA receptor subunits in cerebral cortical cultures. (A) Time-dependent total expression levels of GluN1 (120 kDa), GluN2A (180 kDa) and GluN2B (180 kDa) subunits of the NMDA receptors in cultured cortical cells isolated from 16 day-old rat embryos (0 days in vitro, DIV) or cultured for 8 or 15 days. (B) Representative Western Blotting analysis of cell surface expression of GluN2A and GluN2B subunits labeled by biotinylation, followed by precipitation with neutravidin beads in cortical cells cultured for 8 DIV. The transferrin receptor (Transferrin R, 90 kDa) was used as loading control. Data are the mean ± S.E.M. of 3 independent experiments. Statistical analysis: ** p < 0.01, *** p < 0.01 significantly different when compared with subunit expression at DIV 0 (Tukey’s post hoc test).

98

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

direct stimulation in Na+ medium without added MgCl2 and in the presence of 20 ␮M glycine (to drive the selective activation of NMDARs), by using a microplate reader Spectrofluorometer Gemini EM (Molecular Devices, USA) (340/380 nm excitation wavelength, 510 nm emission wavelength). Experiments were performed in the absence or in the presence of NMDA receptor antagonists MK-801 (10 ␮M), ifenprodil (10 ␮M) or NVP-AA077 (50 nM). Unless indicated, all plotted values were normalized for baseline levels. Data were analyzed by using Excel (Microsoft, Seattle, WA, USA) and Prism (GraphPad Software, San Diego, CA, USA) softwares. 2.8. Statistical analysis Data were expressed as the mean ± S.E.M. of the number of experiments indicated in the figure legends. Comparisons among multiple groups were performed by one-way ANOVA, followed by Tukey’s post hoc test. Student’s t-test was also performed for comparison between two Gaussian populations, as described in figure legends. Significance was accepted at p < 0.05. 3. Results 3.1. Expression of NMDA receptor subunits in cerebral cortical cultures In this study we used cultured rat cerebral cortical cells at 7–8 DIV to evaluate the role exerted by NMDARs, particularly those composed by GluN2A and/or GluN2B subunits, in the regulation of Ca2+ i homeostasis. Since the expression of GluN2A- and GluN2B-NMDARs depends on the developmental stage of the neurons [37,38], it was critical to first examine whether both subtypes of NMDARs were present in these cells at different time in culture. Our results demonstrate that the GluN1 subunit is already expressed in E16 embryos (0 days in vitro) and that total expression levels do not change with maturation time points (Fig. 1A). Total protein levels of GluN2A and GluN2B subunits of the NMDA receptor are highly expressed at 8–15 days in culture (Fig. 1A), a time near synapse formation. Although the results obtained after 15 DIV were slightly greater, data were not significantly different between 8 and 15 DIV (Fig. 1A). The presence of GluN2A and GluN2B subunits at the surface of plasma membrane of cortical cells maintained for 8 DIV was also determined by biotinylation assay and protein expression levels analyzed by Western Blotting. Our results demonstrate that at this stage both NMDAR subunits are present at the membrane surface in cortical cultures (Fig. 1B). Reduced GluN2A surface levels compared to GluN2B surface levels may be accounted for by the predominant presence of GluN2A at synapses, whereas GluN2B appear to be mostly extrasynaptic (e.g. [51,52]). 3.2. Analysis of preparation of Aˇ and cytotoxicity in cortical cultures exposed to NMDA and Aˇ A␤ preparation was composed by a large percentage (∼60%) of low-n oligomers (with ∼17 and 24 kDa, corresponding to the assembly of 4 or 6 A␤ peptides) and monomers (by about 40%). A representative gel shows that fresh A␤1–42 preparation are enriched in oligomers and monomers and do not contain fibrils (Fig. 2A). Results depicted in Fig. 2B demonstrate that exposure to 0.5 ␮M A␤ for 6 h did not affect cell viability, even when evaluated after the recovery period in the presence of a higher concentration of A␤ (1 ␮M). Conversely, when neurons were exposed for 15 min to 100 ␮M, but not 10 ␮M NMDA, a significant decrease in cell viability was observed (Fig. 2C). When the toxic effects were evaluated after the recovery period, a significant decrease in cell viability was observed for both concentrations tested (Fig. 2C). The toxic effect

Fig. 2. Effect of NMDA and A␤ on cell viability. (A) Representative gel of two independent A␤ samples prepared from synthetic A␤1–42 as described in Materials and methods (MS, molecular weight standard). (B) Cells were incubated with 0.5 or 1 ␮M A␤ in culture medium for 6 h and immediately assessed for cell viability. Alternatively, 1 ␮M A␤-treated cells returned to recovery period of 18 h in conditioned cell culture medium, as described in material and methods, in order to evaluate delayed toxicity. (C) Cortical cells were incubated with 10 or 100 ␮M NMDA for 15 min in Na+ medium and immediately assessed for cell viability. Alternatively, cells returned to recovery period of 18 h in conditioned cell culture medium in order to evaluate delayed toxicity. (D) Cells were incubated with 100 ␮M NMDA for 15 min in the absence or presence of 10 ␮M ifenprodil, 50 nM NVP-AAM077 or 10 ␮M MK-801, followed by a recovery period, as described for (B). Data are the mean ± S.E.M. of 3 independent experiments performed in duplicates. Statistical analysis: *** p < 0.01 significantly different when compared with control conditions (Tukey’s post hoc test).

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

99

Fig. 3. Antagonist modulation of NMDA-induced Ca2+ i rise. Cortical cells were stimulated with 10 ␮M or 100 ␮M NMDA in the absence or presence of 10 ␮M ifenprodil, 50 nM NVP-AAM077 or 10 ␮M MK-801. (i) Representative tracings. (ii) Results were plotted as the difference between the maximum value achieved and the basal value before NMDA addition. Data are the mean ± S.E.M. of 3–14 independent experiments performed in duplicates to quadruplicates. Statistical analysis: * p < 0.05, *** p < 0.001 significantly different when compared with NMDA alone (Tukey’s post hoc test).

exerted by 100 ␮M NMDA, observed after the recovery period, was counteracted by the GluN2B subunit antagonist ifenprodil (10 ␮M), whereas treatment with the GluN2A subunit-preferring antagonist NVP-AAM077 did not significantly influence the decrease of cortical neurons viability. As expected, MK-801 (10 ␮M) fully prevented the toxic effect exerted by 100 ␮M NMDA (Fig. 2D). 3.3. Involvement of NMDARs in regulating Ca2+ i homeostasis Cerebral cortical cultures (8 DIV) were stimulated with 10 ␮M or 100 ␮M NMDA in the absence or presence of a NMDA receptor antagonist, namely ifenprodil, NVP-AAM077 or MK-801. Selective stimulation of NMDARs (with NMDA plus glycine in Mg2+ -free medium) induced a concentration-dependent increase in Ca2+ i levels (Fig. 3). Ifenprodil (10 ␮M) partially, but significantly, inhibited the NMDA-induced Ca2+ i rise caused by 10 ␮M or 100 ␮M NMDA by about 38.8% (p < 0.001) and 49.6% (p < 0.001), respectively, suggesting the involvement of GluN2B subunits in the NMDA-induced Ca2+ response. Conversely, blockade of GluN2A-NMDARs (Fig. 3), achieved with NVP-AAM077 (50 nM), failed to prevent NMDAinduced Ca2+ i rise upon stimulation with 10 ␮M or 100 ␮M NMDA. Interestingly, GluN2A antagonism by NVP-AAM077 significantly potentiated Ca2+ i rise induced by 100 ␮M NMDA (p < 0.05). Moreover, NMDA-induced responses were almost completely blocked in the presence of MK-801 (10 ␮M), decreasing Ca2+ i levels by 88.5% and 91.5% (p < 0.001), as compared with the effect of 10 ␮M and 100 ␮M NMDA alone (Fig. 3). Our results clearly show that activation of NMDARs induces an increase in Ca2+ i levels in a concentration-dependent manner, and that this effect is mostly mediated by GluN2B-containing NMDARs.

measured in the absence or presence of NMDAR antagonists. Our results show that, similarly to stimulation with NMDA, A␤ induced a concentration-dependent increase in intracellular Ca2+ levels (Fig. 4). Stimulation with 0.5 ␮M A␤ increased Ca2+ i levels below those achieved by stimulation with a sub-toxic concentration (10 ␮M) of NMDA (Fig. 4i–ii). Moreover, there was a clear trend to return to baseline values. In contrast, after stimulation with 1 ␮M A␤ this tendency was not observed, and in fact, Ca2+ i levels increased similarly to the stimulation observed with the toxic concentration of 100 ␮M NMDA (Fig. 4). In the presence of ifenprodil (10 ␮M), A␤-induced Ca2+ i rise was again partially, but significantly, reduced by 36.5% or 27.6% (p < 0.05) after stimulation with 0.5 ␮M or 1 ␮M A␤, respectively. Interestingly, as for 100 ␮M NMDA, stimulation of cortical cultures with 1 ␮M A␤ in the presence of the GluN2A subunit antagonist NVP-AAM077 (50 nM) produced a significant rise in intracellular Ca2+ i , by about 48.3% (p < 0.01) (Fig. 4). Conversely, the NMDAR antagonistic effect of MK-801 (10 ␮M) resulted in a strong inhibition of intracellular Ca2+ i levels in response to A␤ stimulation (Fig. 4). Because an increase in intracellular Ca2+ levels was observed upon incubation with NMDA or A␤ in the presence of GluN2A subunit antagonist (NVP-AAM077), and in order to exclude a possible effect exerted by the antagonist itself, we analyzed the basal levels of intracellular Ca2+ recordings in the absence or in the presence of NVP-AAM077, ifenprodil or MK-801. As shown in Table 1, NVP-AAM077 or ifenprodil alone did not account for any effect in intracellular Ca2+ levels, whereas MK-801 slightly, but significantly, decreased basal Ca2+ levels. This observation implicates basal glutamate release and is probably due to the potent inhibitory effect of MK-801, which is independent of NMDAR composition.

3.4. Aˇ induces changes in intracellular Ca2+ levels through NMDAR activation Next, we examined whether activation of NMDARs contributes to A␤-induced loss of Ca2+ i homeostasis. For this purpose, we exposed neurons to A␤ at non-toxic micromolar concentrations (0.5 ␮M or 1 ␮M) (Fig. 2B). In order to assess whether GluN2A/B-NMDARs subunits are involved in A␤ peptide-induced loss of Ca2+ i homeostasis, cerebral cortical neurons were stimulated with A␤, and Ca2+ i levels

3.5. Effect of pre-exposure to Aˇ in intracellular Ca2+ response elicited by NMDA In order to determine the influence of A␤ on NMDA-evoked responses, we then investigated the influence of A␤ pre-exposure on NMDA-induced Ca2+ i changes. Primary cortical cells were preexposed (for 180 s) to A␤ (0.5 or 1 ␮M), followed by a subsequent stimulation with NMDA (10 or 100 ␮M); the involvement of

100

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

Fig. 4. Influence of NMDAR antagonists in Ca2+ i changes induced by direct effects of A␤. Cortical neurons were exposed to 0.5 ␮M or 1 ␮M A␤ in the absence or in the presence of 10 ␮M ifenprodil, 50 nM NVP-AAM077 or 10 ␮M MK-801. (i) Representative tracings. (ii) Results were plotted as the difference between the maximum value achieved and the basal value before A␤ addition. Data are the mean ± S.E.M. of 3–9 independent experiments performed in duplicates to quadruplicates. Statistical analysis: * p < 0.05, ** p < 0.01, *** p < 0.001 significantly different when compared with A␤ alone (Tukey’s post hoc test); t p < 0.05 significantly different when compared with A␤ alone (Student’s t-test analysis).

GluN2A and GluN2B-NMDAR subunits was assessed by using NVPAAM077 and ifenprodil, respectively. Interestingly, when NMDA (10 or 100 ␮M) stimulation was secondary to pre-exposure to 0.5 ␮M A␤, NMDA-evoked increase in Ca2+ i was not significantly different from that observed with NMDA alone (Fig. 5A), which may account for by the fact that 0.5 ␮M A␤-evoked Ca2+ i responses were transient. As expected, ifenprodil (10 ␮M) partially decreased the Ca2+ i rise triggered by 10 ␮M NMDA (p < 0.05) or 100 ␮M NMDA (p < 0.001) following A␤ (0.5 ␮M) exposure by about 42.8% and 69.1%, respectively (Fig. 5A). After blockade of GluN2A-NMDARs with NVP-AAM077 (50 nM), Ca2+ i rise did not differ significantly from that observed in the absence of the antagonist. Notably, the effect of NVP-AAM077 was not largely different from that observed upon primary NMDA stimulation (e.g. in the absence of A␤), as shown in Fig. 3. Conversely, when neurons were pre-exposed to 1 ␮M A␤ (180 s), the secondary stimulation with 10 and 100 ␮M NMDAevoked Ca2+ i was significantly reduced by 43.5% (p < 0.01) and 62.8% (p < 0.001), respectively, when compared with the effect of a primary stimulation with the same concentration of NMDA (Fig. 5B), indicating that at this concentration A␤ largely interferes with NMDARs. Notably, 1 ␮M A␤-induced elevation of Ca2+ i was not transient as observed after exposure to 0.5 ␮M A␤. Ifenprodil (10 ␮M) further reduced the secondary stimulation evoked by 10 and 100 ␮M NMDA-induced Ca2+ i rise by 29.3% and 53.4% (p < 0.05), respectively (Fig. 5B). Interestingly, ifenprodil inhibited by the same extent NMDA-evoked Ca2+ i in the absence (Fig. 3) or following exposure to 1 ␮M A␤ (Fig. 5 B), suggesting a similar contribution of GluN2B subunits in both conditions.

Conversely, the antagonism of GluN2A-NMDARs (Fig. 5B) induced by NVP-AAM077 (50 nM) increased NMDA-induced Ca2+ i by 238% after secondary exposure to 100 ␮M NMDA (p < 0.001). NVP-AAM077-induced Ca2+ i rise under these conditions was higher compared to the primary stimulation with 100 ␮M NMDA (Fig. 3), suggesting an important role of GluN2A subunits in controlling Ca2+ i increases under stress stimuli. These results indicate that 0.5 ␮M A␤ does not greatly affect Ca2+ i rise evoked by NMDA stimulation. In contrast, a higher concentration of A␤ (e.g. 1 ␮M) largely reduces the Ca2+ i response triggered by NMDA, which is alleviated upon blockade of GluN2A subunits, supporting the notion that NMDAR subunits may differently modulate Ca2+ i particularly under stress conditions. 3.6. Influence of NMDA pre-exposure on Ca2+ i rise evoked by Aˇ To investigate the hypothesis that A␤ increase in Ca2+ i directly interfered with NMDARs and thus was affected by prior NMDAR stimulation, cortical neurons were pre-exposed (180 s) to NMDA (10 or 100 ␮M) and subsequently stimulated with A␤ (0.5 or 1 ␮M). Our results show that pre-exposure to NMDA in non-toxic (10 ␮M) and sub-maximal or toxic (100 ␮M) concentrations significantly reduces the increase in Ca2+ i induced by A␤ (0.5 ␮M or 1 ␮M). Indeed, we observed that 10 ␮M NMDA reduced 0.5 and 1 ␮M A␤ A␤-evoked Ca2+ i rise by 59.8% (p < 0.01) and 87.2% (p < 0.001), respectively, when compared with the effect of primary stimulation with the same concentration of the soluble peptide (Fig. 6A). Following pre-exposure to 100 ␮M NMDA, and similarly to 10 ␮M, the increase in Ca2+ i due to secondary exposure to 0.5

Table 1 Effect of NMDA receptor antagonists on intracellular basal Ca2+ levels. Control

Ifenprodil

NVP-AAM077

MK-801

1.23 ± 0.013 (n = 12)

1.23 ± 0.016 (n = 7)

1.23 ± 0.011 (n = 6)

1.13 ± 0.014*** (n = 4)

Cortical cells were incubated in the absence (control) or in the presence of 10 ␮M ifenprodil, 50 nM NVP-AAM077 or 10 ␮M MK-801. Data are the mean ± S.E.M. of 4–12 independent experiments performed in duplicates to quadruplicates. *** Statistical analysis: p < 0.001 significantly different when compared to condition in the absence of antagonists (Tukey’s post hoc test).

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

101

Fig. 5. A␤ modulation of secondary Ca2+ i response elicited by NMDA stimulation. Cortical cells were stimulated with (A) 0.5 ␮M or (B) 1 ␮M A␤ followed by a second stimulation (after 180 s) with 10 ␮M or 100 ␮M NMDA in the absence or in the presence of 10 ␮M ifenprodil or 50 nM NVP-AAM077. (i) Representative tracings. (ii) Results were plotted as the difference between the maximum value achieved and the basal value before NMDA addition. The open bars (NMDA) were re-plotted from Fig. 2 in order to facilitate graph interpretation. Data are the mean ± S.E.M. of 3–5 independent experiments performed in duplicates to quadruplicates. Statistical analysis: *** p < 0.001 significantly different when compared to stimulation with NMDA alone (Tukey’s post hoc test); tt p < 0.01 significantly different when compared with NMDA alone (Student’s t-test analysis); # p < 0.05, ### p < 0.001 significantly different when compared with the effect of stimulation with NMDA following pre-exposure to 0.5 or 1 ␮M A␤ (Tukey’s post hoc test); t p < 0.05 significantly different when compared with the effect of stimulation with NMDA following pre-exposure to 0.5 or 1 ␮M A␤ (Student’s t-test analysis).

and 1 ␮M A␤ was reduced by 74.5% (p < 0.01) and 83.9% (p < 0.001), respectively, when compared with the effect of primary stimulation with the same concentration of A␤ (Fig. 6B). GluN2B-NMDAR subunit antagonism achieved by ifenprodil (10 ␮M) did not affect any of the previously described responses (Fig. 6). These data indicate that activation of NMDARs by NMDA (10 ␮M or 100 ␮M) almost completely inhibited the intracellular Ca2+ i rise evoked by A␤. However, the antagonism of GluN2A-NMDARs induced by NVPAAM077 (50 nM) in these conditions reverted the inhibitory effect of 0.5 or 1 ␮M A␤-induced Ca2+ i rise in cells pre-exposed to 10 ␮M (Fig. 6A) or 100 ␮M (Fig. 6B) NMDA.

3.7. Intracellular Ca2+ modifications evoked by simultaneous exposure to Aˇ and NMDA Finally, we analyzed the effects in intracellular Ca2+ levels triggered by simultaneous exposure to A␤ and NMDA. The involvement of specific NMDAR subunits was further determined by using ifenprodil and NVP-AAM077. Stimulation of cortical cultures with 0.5 ␮M A␤ plus 10 ␮M NMDA potentiated the increase in Ca2+ i levels by 172% or 50% (p < 0.001), respectively, when compared with those achieved upon stimulation with 0.5 ␮M A␤ or 10 ␮M NMDA alone (Fig. 7A).

102

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

Fig. 6. NMDA modulation of secondary Ca2+ i response elicited by A␤. Cortical neurons were stimulated with (A) 10 ␮M or (B) 100 ␮M NMDA for 180 s, followed by a second stimulation with 0.5 ␮M or 1 ␮M A␤ in the absence or in the presence of 10 ␮M ifenprodil or 50 nM NVP-AAM077. (i) Representative tracings. (ii) Results were plotted as the difference between the maximum value achieved and the basal value before A␤ addition. The black bars (A␤) were re-plotted from Fig. 3 in order to facilitate graph interpretation. Data are the mean ± S.E.M. of 3–4 independent experiments performed in duplicates to quadruplicates. Statistical analysis: ** p < 0.01, *** p < 0.001 when compared with the effect of A␤ alone; ## p < 0.01, ### p < 0.001 significantly different when compared with the effect of stimulation with A␤ following pre-exposure to 10 or 100 ␮M NMDA (Tukey’s post hoc test).

Moreover, simultaneous exposure to 0.5 ␮M A␤ plus 100 ␮M NMDA increased Ca2+ i by 300% when compared to 0.5 ␮M A␤ (p < 0.001) and by 158% when compared to 100 ␮M NMDA (p < 0.001) alone (Fig. 7A). Ifenprodil (10 ␮M) partially reduced Ca2+ i rise triggered by simultaneous addition of 0.5 ␮M A␤ and NMDA (10 ␮M or 100 ␮M) to levels similar to those evoked by NMDA alone, suggesting the involvement of GluN2B-NMDARs in A␤ response (Fig. 7A). In the presence of NVP-AAM077 (50 nM), Ca2+ i rise elicited by 0.5 ␮M A␤ plus NMDA (10 ␮M or 100 ␮M) was not significantly different from that observed when the antagonist was absent. Interestingly, Ca2+ i rise achieved upon stimulation of cortical neurons with 1 ␮M A␤ plus 10 ␮M NMDA was not significantly different from that achieved upon exposure to 1 ␮M A␤ alone.

However, Ca2+ i was 50% higher (p < 0.001) than that observed in the presence of 10 ␮M NMDA only (Fig. 7B). Simultaneous stimulation with 1 ␮M A␤ plus 100 ␮M NMDA enhanced Ca2+ i (Fig. 7B) by 54% higher than that observed in response to 1 ␮M A␤ alone, and 41% higher than upon stimulation with 100 ␮M NMDA alone (p < 0.05). Blockade of GluN2B-NMDARs with ifenprodil (10 ␮M) partially decreased Ca2+ i rise triggered by 1 ␮M A␤ plus NMDA (10 ␮M or 100 ␮M) (Fig. 7B). Again, the opposite effect was observed in the presence of NVP-AAM077 (50 nM). Indeed, we observed a significant increase in Ca2+ i levels by about 136% when the cells were stimulated with 1 ␮M A␤ plus 10 ␮M NMDA (p < 0.05) and by 160% after exposure to 1 ␮M A␤ plus 100 ␮M NMDA (p < 0.001) in the presence of NVP-AAM077 (Fig. 7B).

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

103

Fig. 7. Simultaneous exposure to A␤ oligomers and NMDA potentiate Ca2+ i rise. Cortical neurons were simultaneously stimulated with (A) 0.5 or (B) 1 ␮M A␤ plus 10 or 100 ␮M NMDA, in the absence or in the presence of 10 ␮M ifenprodil and 50 nM NVP-AAM077. (i) Representative tracings. (ii) Results were plotted as the difference between the maximum value achieved and the basal value before simultaneous addition of A␤ oligomers and NMDA. The open (NMDA) and black (A␤) bars were re-plotted from Figs. 2 and 3 in order to facilitate graph interpretation. Data are the mean ± S.E.M. of 3–6 independent experiments performed in duplicates to quadruplicates. Statistical analysis: * p < 0.05, ** p < 0.01, *** p < 0.001 significantly different when compared with A␤ or NMDA (Tukey’s post hoc test); # p < 0.05, ### p < 0.001 significantly different when compared with A␤ plus NMDA (Tukey’s post hoc test); t p < 0.05 significantly different when compared with A␤ plus NMDA (Student’s t-test analysis).

Together, these results show that A␤ and NMDA, when present simultaneously, potentiate Ca2+ i rise, probably as a result of an overactivation of NMDARs. Moreover, data suggest that GluN2Aand GluN2B-NMDARs exhibit conflicting roles in mediating the underlying mechanisms.

4. Discussion Abnormal homeostasis of Ca2+ i has been observed in both elderly and AD subjects [10,39,40]. Several studies have recently demonstrated that A␤ directly interacts with cell function, either

104

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

by insertion into the membrane to form a cation-conducting pore [41], activation of cell surface receptors coupled to Ca2+ influx [30] or induction of oxidative stress, leading to deregulation of mitochondrial homeostasis [9]. Moreover, A␤ oligomers, previously described to be more toxic than monomers or fibrils [13,17,42], were also described to disrupt the integrity of both plasma and intracellular membranes, leading to cell death in a channel-independent way [42]. A␤ oligomers were also reported to co-immunoprecipitate with extracellular domains of the GluN1 subunit, suggesting a direct interaction with NMDA receptors [9]. Indeed, these authors have also shown that ADDLs bind to or are in close proximity to NMDARs, triggering neuronal damage through NMDAR-dependent Ca2+ influx. Although much attention has been given to the toxic profile of A␤ oligomers, recent findings from Jan and collaborators [43] evidence that crude A␤42 preparations, consisting of a monomeric and heterogeneous mixture of A␤42 oligomers, were more toxic than purified monomeric, protofibrillar fractions, or fibrils in different cell lines and primary neurons. Moreover, these authors observed that selective removal of monomeric A␤42 from these preparations, using insulin-degrading enzyme, reversed the toxicity of A␤42 protofibrils, demonstrating that A␤42 toxicity is not linked to specific prefibrillar aggregate(s) but rather to the ability of these species to grow and undergo fibril formation, which depends on the presence of monomeric A␤42. Thus, A␤42 neurotoxicity is mediated by ongoing nucleated polymerization process rather than by individual A␤42 species [43]. Our data show that similarly to NMDA (10–100 ␮M), exposure to A␤ preparation (0.5–1 ␮M) induces a concentration-dependent rise in Ca2+ i levels in cerebral cortical cultures expressing equivalent total levels of GluN1 as well as GluN2A and GluN2B subunits, and also surface levels of both GluN2A and GluN2B subunits. These effects were almost completely prevented by the non-selective NMDAR antagonist MK-801, suggesting an activation of NMDARs by A␤. Involvement of GluN2A- and GluN2B-NMDAR subtypes was further evaluated through pharmacological blockade by competitive antagonists of the NMDAR subunits, namely NVP-AAM077 and ifenprodil. Our results indicate that Ca2+ i increase upon stimulation with NMDA or A␤ is mainly due to GluN2B-NMDARs. Indeed, ifenprodil was able to partially reduce, but not fully inhibit, the Ca2+ i influx induced by either NMDA or A␤. This could be attributed to the fact that ifenprodil only blocks pure GluN1/GluN2B NMDARs by approximately 80% [35,36,44]. Incomplete blockade of GluN2Bcontaining receptors may be due to the fact that other NMDAR subunits, namely GluN2D or GluN2C, are present in the cortical cell cultures and thus may also contribute to the rise in Ca2+ i . When using the higher concentration of A␤ (1 ␮M) or NMDA (100 ␮M), but not the lower concentrations as the stimulating agent, NVPAAM077 antagonism of GluN2A-NMDARs potentiated Ca2+ i rise compared to control conditions (in the absence of NVP-AAM077). These observations suggest that GluN2A-NMDARs restrain Ca2+ influx under particular stress stimuli. Our results are supported by recent studies where application of GluN2B antagonist, but not GluN2A-containing receptor antagonist NVP-AAM077, prevented both apoptosis and necrosis induced by NMDA, indicating the critical involvement of GluN1/GluN2B, but not GluN2A-containing NMDAR subtypes in mediating neuronal death. These findings are also in accordance with previous reports showing that excitotoxicity is triggered by the selective activation of NMDARs containing the GluN2B subunit [27,38], whereas GluN2A-containing NMDARs promote survival [27]. Indeed, GluN2A- and GluN2B-NMDARs appear to have opposite roles in regulating Ca2+ i in the presence of oligomeric A␤ 1–42. These findings support the concept that dysregulation of Ca2+ i homeostasis is induced by a possible interaction of A␤ with NMDARs, particularly of the GluN2B subtype. In contrast, in a recent publication, A␤ oligomers were

shown to directly activate NMDA receptors, particularly those composed of GluN2A subunit heterologously expressed in Xenopus laevis oocytes [45], which may be related to the expression of NMDARs subunits in non-neuronal cells. Indeed, in a previous study we showed that fibrillary Abeta 1–40 mediated necrotic cell death through changes in Ca2+ homeostasis in HEK293 cells selectively expressing GluN1/GluN2A subunits, but not GluN1/GluN2B subunits [46]. An alternative hypothesis for A␤-mediated NMDAR activation was recently described to involve A␤-evoked glutamate release [47]. In line with the findings reported here, a recent study by Alberdi et al. [48] has shown that A␤ oligomers induce inward currents, Ca2+ i increase, through a mechanism requiring NMDA and AMPA receptors activation in both rat cortical neurons and hippocampal organotypic slices [48]. Accordingly, it was demonstrated that in the AD brain and human cortical neurons, excitatory synapses containing the GluN2B subunit of the NMDA receptor appear to be the main sites of oligomer accumulation. A␤ oligomers co-localize with synaptic markers, and this effect is counteracted by the NMDA antagonists ifenprodil and memantine. The latter is an uncompetitive NMDAR antagonist that blocks the ion channel formed by NMDARs [14]. Of note, memantine was shown to improve cognitive functions in patients with moderate to severe forms of AD, also preventing oxidative stress and calcium influx produced by A␤ oligomers in hippocampal neuronal cultures [9]. Further evidence suggesting that A␤ may interact with NMDARs was given by studying the effects of A␤ on NMDA-evoked Ca2+ responses. We observed that pre-exposure to 1 ␮M A␤ (which caused sustained elevation of Ca2+ i ), but not 0.5 ␮M A␤, significantly reduced the Ca2+ i response triggered by NMDA. Interestingly, ifenprodil inhibited NMDA-evoked Ca2+ i by the same extent, in the absence or subsequent exposure to 1 ␮M A␤, suggesting a similar contribution of GluN2B subunits. Conversely, NVP-AAM077 increased Ca2+ i rise under these conditions when compared to the primary stimulation with 100 ␮M NMDA, suggesting an important role of GluN2A subunits in controlling Ca2+ i increase under stress stimuli. In agreement with our findings following 1 ␮M A␤ pre-exposure, A␤ oligomers application was shown to mimic a state of partial NMDAR blockade, reducing NMDAR activity and NMDAR-dependent calcium influx [13]. Moreover, neurons from a genetic mouse model of AD were found to express reduced amounts of surface GluN1 subunit of NMDARs [19], and A␤ 1–42 was also found to reduce surface expression of the GluN1 subunit, in both cortical and hippocampal neurons [19–21]. Accordingly, A␤ can interact with NMDAR either through Abeta-induced modulation of NMDA binding sites and/or A␤mediated protein conformational changes, leading to decreased NMDA-evoked responses. Data from the current study also showed that A␤-induced increase in Ca2+ i was affected by prior receptor stimulation with NMDA, providing additional evidence that A␤ directly interferes with NMDARs. Thus, selective activation of NMDARs inhibited the Ca2+ i rise evoked by A␤, suggesting that A␤-induced Ca2+ i responses require the presence of an unstimulated NMDAR. Furthermore, our results demonstrate that simultaneous exposure to A␤ and NMDA induces overactivation of NMDARs and potentiates the increase in Ca2+ i levels when compared to stimulation with NMDA or A␤ alone. Interestingly, this effect was greater for 0.5 ␮M than 1 ␮M A␤, suggesting different molecular target sites at the extracellular domain of the receptor. In line with data observed in the presence of 0.5 ␮M A␤, recent studies have reported that treatment with A␤ oligomers potentiates NMDA-evoked firing and induces a rapid and transient increase in intracellular Ca2+ levels that is blocked by memantine in mature hippocampal neurons [9,49]. Moreover, Ca2+ influx elicited by NMDA (10 ␮M or 100 ␮M) plus A␤ (0.5 ␮M or 1 ␮M) in both concentrations used was partially

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

due to the activation of GluN2B-NMDARs, as demonstrated by ifenprodil antagonism. Meanwhile, the rising effect of NVP-AAM077 was only observed in the presence of 1 ␮M A␤, again suggesting that GluN2A-NMDARs act by controlling Ca2+ influx, particularly under stress stimuli. Consistent results of NVP-AAM077 in the presence of 1 ␮M A␤ suggest that GluN2A and GluN2B exert differential and opposite roles in regulating Ca2+ i homeostasis. In addition, NMDAR-dependent Ca2+ influx triggers a number of intracellular events that convey the signal downstream to either confer neuroprotection or to trigger cell death. Thus, GluN2A- and GluN2B NMDARs may have antagonist roles in activating pro-survival and pro-death signaling pathways. There is a growing body of evidence that NMDAR activity has the potential to promote survival or death in neurons of the central nervous system [50], which may be related to differences in synaptic versus extrasynaptic NMDAR signaling. The molecular basis for the apparent differences may be due to distinct locations in GluN2A (synaptic)- and GluN2B (extrasynaptic)-containing NMDARs. As demonstrated by recent reports that focus on NMDAR trafficking, the subunit composition in synaptic versus extrasynaptic membranes depends upon the phosphorylation state of specific tyrosine residues of GluN2B [51]. GluN2A is incorporated into synaptic NMDARs by a mechanism involving the cytoplasmic C-terminus [52]. GluN2A may thus become enriched at synapses, compared to extrasynaptic locations. Therefore, GluN2A-enriched synaptic NMDARs appear to selectively promote survival. Nevertheless, Liu et al. reported that GluN2A-containing NMDARs promote survival, whereas GluN2BNMDARs promote death independent of their location [27]. In contrast, pro-death NMDAR signaling was reported to be mediated by GluN2A-NMDARs [53], showing that the subunit composition is not important in determining excitotoxicity. Indeed, the concept that GluN2A partitions near exclusively into synaptic locations has been challenged recently by findings that GluN2A can end up at extrasynaptic locations in cultured neurons [54] and that the subunit composition of synaptic and extrasynaptic NMDARs is similar in 3 week-old acutely dissociated hippocampal slices [55]. Moreover, it was recently shown that in developing hippocampal neurons, GluN2B-NMDARs are capable of mediating antagonistic signaling to survival or death, as well as synaptic potentiation, depending on the stimulus employed. This indicates that in immature hippocampal neurons, the subunit composition of the NMDAR may not account for dichotomous NMDAR signaling [56]. Another contributing factor explaining the differences between synaptic and extrasynaptic pools could be the way they are activated: brief saturating activation in the case of synaptic NMDARs, compared with chronic, low level activation of extrasynaptic NMDARs by bath application of glutamate. Differences in the properties of intracellular Ca2+ transients evoked by these different stimuli may differentially affect signaling, even if the overall Ca2+ load is similar [25,57]. Therefore, with low doses of NMDA, synaptic NMDA receptors are activated far more strongly than extrasynaptic receptors, enabling pro-survival synaptic signaling to dominate and activate protein kinase B (Akt), extracellular regulated kinases 1 and 2 (Erk 1/2) and cAMP response element-binding (CREB) proteins. The scenario is very different with toxic doses of NMDA; action potential firing is suppressed and the Ca2+ influx is attributable to the direct activation of both synaptic and extrasynaptic NMDA receptors. Hence, extrasynaptic signaling is dominant and is coupled to CREB shut-off pathway [57]. As such, the question of whether the GluN2 subunits influence NMDARs signal to survival or death (controlling for the amount of Ca2+ that fluxes into cell) remains unresolved and further studies are required. The possible role of other GluN2 subunits, namely GluN2C and GluN2D, also needs to be clarified. GluN2C- or GluN2Dcontaining receptors appear to give rise to ‘low-conductance’ openings with a lower sensitivity to extracellular Mg2+ , which may

105

affect the Ca2+ influx generated by synaptic activation of NMDAR [22]. Another subject that can be targeted in future investigations concerns the mechanism underlying the GluN2A-induced modulation of Ca2+ rise observed in the present study. In summary, this work demonstrates that A␤ preparation composed by oligomers and monomers disturbs Ca2+ homeostasis through NMDAR activation. In this regard, A␤ was shown to modulate the effects of NMDAR activation. We also show that GluN2A and GluN2B subtypes of NMDARs have opposing roles in regulating Ca2+ i homeostasis. However, the question of whether activation of each of these NMDAR subtypes signals to pro-survival or pro-death pathways is yet to be answered. Overall, this work contributes to the understanding of the molecular basis of AD and thus provides insight into potential therapeutic targets that may be used to prevent or ameliorate early neuronal dysfunction in AD, commonly attributed to A␤ oligomers. Conflict of interest None declared. Acknowledgments This work was supported by Fundac¸ão para a Ciência e a Tecnologia (FCT), project reference PTDC/SAU-NEU/71675/2006 and by Lundbeck Foundation. The authors thank researchers at the Center for Neuroscience and Cell Biology, University of Coimbra: Márcio Ribeiro for technical assistance with fluorimetric measurements, Dr. Teresa Oliveira for figure editing, and Dr Rosa Resende and Rui Costa for technical expertise with A␤ preparation. References [1] R. Brookmeyer, E. Johnson, K. Ziegler-Graham, H.M. Arrighi, Forecasting the global burden of Alzheimer’s disease, Alzheimers Dement 3 (2007) 186–191. [2] S. Oddo, A. Caccamo, M. Kitazawa, B.P. Tseng, F.M. LaFerla, Amyloid deposition precedes tangle formation in a triple transgenic model of Alzheimer’s disease, Neurobiol. Aging 24 (2003) 1063–1070. [3] J. Hardy, D.J. Selkoe, The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics, Science 297 (2002) 353–356. [4] D.J. Selkoe, Presenilin, Notch, and the genesis and treatment of Alzheimer’s disease, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 11039–11041. [5] K.N. Dahlgren, A.M. Manelli, W.B. Stine Jr., L.K. Baker, G.A. Krafft, M.J. LaDu, Oligomeric and fibrillar species of amyloid-beta peptides differentially affect neuronal viability, J. Biol. Chem. 277 (2002) 32046–32053. [6] M.P. Mattson, B. Cheng, D. Davis, K. Bryant, I. Lieberburg, R.E. Rydel, betaAmyloid peptides destabilize calcium homeostasis and render human cortical neurons vulnerable to excitotoxicity, J. Neurosci. 12 (1992) 376–389. [7] B.L. Kelly, A. Ferreira, beta-Amyloid-induced dynamin 1 degradation is mediated by N-methyl-d-aspartate receptors in hippocampal neurons, J. Biol. Chem. 281 (2006) 28079–28089. [8] F. Pellistri, M. Bucciantini, A. Relini, et al., Nonspecific interaction of prefibrillar amyloid aggregates with glutamatergic receptors results in Ca2+ increase in primary neuronal cells, J. Biol. Chem. 283 (2008) 29950–29960. [9] F.G. De Felice, P.T. Velasco, M.P. Lambert, et al., Abeta oligomers induce neuronal oxidative stress through an N-methyl-d-aspartate receptor-dependent mechanism that is blocked by the Alzheimer drug memantine, J. Biol. Chem. 282 (2007) 11590–11601. [10] I. Bezprozvanny, M.P. Mattson, Neuronal calcium mishandling and the pathogenesis of Alzheimer’s disease, Trends Neurosci. 31 (2008) 454–463. [11] K. Parameshwaran, M. Dhanasekaran, V. Suppiramaniam, Amyloid beta peptides and glutamatergic synaptic dysregulation, Exp. Neurol. 210 (2008) 7–13. [12] P.N. Lacor, M.C. Buniel, P.W. Furlow, et al., Abeta oligomer-induced aberrations in synapse composition, shape, and density provide a molecular basis for loss of connectivity in Alzheimer’s disease, J. Neurosci. 27 (2007) 796–807. [13] G.M. Shankar, B.L. Bloodgood, M. Townsend, D.M. Walsh, D.J. Selkoe, B.L. Sabatini, Natural oligomers of the Alzheimer amyloid-beta protein induce reversible synapse loss by modulating an NMDA-type glutamate receptor-dependent signaling pathway, J. Neurosci. 27 (2007) 2866–2875. [14] A. Deshpande, H. Kawai, R. Metherate, C.G. Glabe, J. Busciglio, A role for synaptic zinc in activity-dependent Abeta oligomer formation and accumulation at excitatory synapses, J. Neurosci. 29 (2009) 4004–4015. [15] A. Deshpande, E. Mina, C. Glabe, J. Busciglio, Different conformations of amyloid beta induce neurotoxicity by distinct mechanisms in human cortical neurons, J. Neurosci. 26 (2006) 6011–6018.

106

I.L. Ferreira et al. / Cell Calcium 51 (2012) 95–106

[16] S. Lesne, M.T. Koh, L. Kotilinek, et al., A specific amyloid-beta protein assembly in the brain impairs memory, Nature 440 (2006) 352–357. [17] R. Resende, E. Ferreiro, C. Pereira, O.C. Resende de, Neurotoxic effect of oligomeric and fibrillar species of amyloid-beta peptide 1–42: involvement of endoplasmic reticulum calcium release in oligomer-induced cell death, Neuroscience 155 (2008) 725–737. [18] Y. Yan, C. Wang, Abeta42 is more rigid than Abeta40 at the C terminus: implications for Abeta aggregation and toxicity, J. Mol. Biol. 364 (2006) 853–862. [19] E.M. Snyder, Y. Nong, C.G. Almeida, et al., Regulation of NMDA receptor trafficking by amyloid-beta, Nat. Neurosci. 8 (2005) 1051–1058. [20] S. Johansson, A.C. Radesater, R.F. Cowburn, J. Thyberg, J. Luthman, Modelling of amyloid beta-peptide induced lesions using roller-drum incubation of hippocampal slice cultures from neonatal rats, Exp. Brain Res. 168 (2006) 11–24. [21] Y. Goto, T. Niidome, A. Akaike, T. Kihara, H. Sugimoto, Amyloid beta-peptide preconditioning reduces glutamate-induced neurotoxicity by promoting endocytosis of NMDA receptor, Biochem. Biophys. Res. Commun. 351 (2006) 259–265. [22] S. Cull-Candy, S. Brickley, M. Farrant, NMDA receptor subunits: diversity, development and disease, Curr. Opin. Neurobiol. 11 (2001) 327–335. [23] S.G. Cull-Candy, D.N. Leszkiewicz, Role of distinct NMDA receptor subtypes at central synapses, Sci. STKE (2004), 2004, re16. [24] G. Kohr, NMDA receptor function: subunit composition versus spatial distribution, Cell Tissue Res. 326 (2006) 439–446. [25] G.E. Hardingham, Coupling of the NMDA receptor to neuroprotective and neurodestructive events, Biochem. Soc. Trans. 37 (2009) 1147–1160. [26] G. Krapivinsky, L. Krapivinsky, Y. Manasian, et al., The NMDA receptor is coupled to the ERK pathway by a direct interaction between NR2B and RasGRF1, Neuron 40 (2003) 775–784. [27] Y. Liu, T.P. Wong, M. Aarts, et al., NMDA receptor subunits have differential roles in mediating excitotoxic neuronal death both in vitro and in vivo, J. Neurosci. 27 (2007) 2846–2857. [28] M.J. Kim, A.W. Dunah, Y.T. Wang, M. Sheng, Differential roles of NR2A- and NR2B-containing NMDA receptors in Ras-ERK signaling and AMPA receptor trafficking, Neuron 46 (2005) 745–760. [29] M.P. Mattson, S.L. Chan, Neuronal and glial calcium signaling in Alzheimer’s disease, Cell Calcium 34 (2003) 385–397. [30] K.N. Green, F.M. LaFerla, Linking calcium to Abeta and Alzheimer’s disease, Neuron 59 (2008) 190–194. [31] P. Agostinho, C.R. Oliveira, Involvement of calcineurin in the neurotoxic effects induced by amyloid-beta and prion peptides, Eur. J. Neurosci. 17 (2003) 1189–1196. [32] E. Ferreiro, R. Resende, R. Costa, C.R. Oliveira, C.M. Pereira, An endoplasmicreticulum-specific apoptotic pathway is involved in prion and amyloid-beta peptides neurotoxicity, Neurobiol. Dis. 23 (2006) 669–678. [33] P. Kurup, Y. Zhang, J. Xu, et al., Abeta-mediated NMDA receptor endocytosis in Alzheimer’s disease involves ubiquitination of the tyrosine phosphatase STEP61, J. Neurosci. 30 (2010) 5948–5957. [34] Y.P. Auberson, H. Allgeier, S. Bischoff, K. Lingenhoehl, R. Moretti, M. Schmutz, 5-Phosphonomethylquinoxalinediones as competitive NMDA receptor antagonists with a preference for the human 1A/2A, rather than 1A/2B receptor composition, Bioorg. Med. Chem. Lett. 12 (2002) 1099–1102. [35] P.A. Frizelle, P.E. Chen, D.J. Wyllie, Equilibrium constants for (R)-[(S)-1-(4bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydroquino xalin-5-yl)methyl-phosphonic acid (NVP-AAM077) acting at recombinant NR1/NR2A and NR1/NR2B N-methyl-d-aspartate receptors: implications for studies of synaptic transmission, Mol. Pharmacol. 70 (2006) 1022–1032. [36] K. Williams, Ifenprodil discriminates subtypes of the N-methyl-d-aspartate receptor: selectivity and mechanisms at recombinant heteromeric receptors, Mol. Pharmacol. 44 (1993) 851–859.

[37] H. Monyer, N. Burnashev, D.J. Laurie, B. Sakmann, P.H. Seeburg, Developmental and regional expression in the rat brain and functional properties of four NMDA receptors, Neuron 12 (1994) 529–540. [38] M. Zhou, M. Baudry, Developmental changes in NMDA neurotoxicity reflect developmental changes in subunit composition of NMDA receptors, J. Neurosci. 26 (2006) 2956–2963. [39] M. Giacomello, I. Drago, P. Pizzo, T. Pozzan, Mitochondrial Ca2+ as a key regulator of cell life and death, Cell Death Differ. 14 (2007) 1267–1274. [40] F.M. LaFerla, Calcium dyshomeostasis and intracellular signalling in Alzheimer’s disease, Nat. Rev. Neurosci. 3 (2002) 862–872. [41] A. Demuro, I. Parker, Optical single-channel recording: imaging Ca2+ flux through individual ion channels with high temporal and spatial resolution, J. Biomed. Opt. 10 (2005) 11002. [42] A. Demuro, E. Mina, R. Kayed, S.C. Milton, I. Parker, C.G. Glabe, Calcium dysregulation and membrane disruption as a ubiquitous neurotoxic mechanism of soluble amyloid oligomers, J. Biol. Chem. 280 (2005) 17294–17300. [43] A. Jan, O. Adolfsson, I. Allaman, et al., Abeta42 neurotoxicity is mediated by ongoing nucleated polymerization process rather than by discrete Abeta42 species, J. Biol. Chem. 286 (2011) 8585–8596. [44] K.R. Tovar, G.L. Westbrook, The incorporation of NMDA receptors with a distinct subunit composition at nascent hippocampal synapses in vitro, J. Neurosci. 19 (1999) 4180–4188. [45] L. Texido, M. Martin-Satue, E. Alberdi, C. Solsona, C. Matute, Amyloid beta peptide oligomers directly activate NMDA receptors, Cell Calcium (2011). [46] A. Domingues, S. Almeida, E.F. da Cruz e Silva, C.R. Oliveira, A.C. Rego, Toxicity of beta-amyloid in HEK293 cells expressing NR1/NR2A or NR1/NR2B N-methyld-aspartate receptor subunits, Neurochem. Int. 50 (2007) 872–880. [47] J. Brito-Moreira, A.C. Paula-Lima, T.R. Bomfim, et al., Abeta oligomers induce glutamate release from hippocampal neurons, Curr. Alzheimer Res. (2011). [48] E. Alberdi, M.V. Sanchez-Gomez, F. Cavaliere, et al., Amyloid beta oligomers induce Ca2+ dysregulation and neuronal death through activation of ionotropic glutamate receptors, Cell Calcium 47 (2010) 264–272. [49] V. Szegedi, G. Juhasz, D. Budai, B. Penke, Divergent effects of Abeta1–42 on ionotropic glutamate receptor-mediated responses in CA1 neurons in vivo, Brain Res. 1062 (2005) 120–126. [50] S. Papadia, G.E. Hardingham, The dichotomy of NMDA receptor signaling, Neuroscientist 13 (2007) 572–579. [51] S.M. Goebel-Goody, K.D. Davies, R.M. Alvestad Linger, R.K. Freund, M.D. Browning, Phospho-regulation of synaptic and extrasynaptic N-methyl-daspartate receptors in adult hippocampal slices, Neuroscience 158 (2009) 1446–1459. [52] F. Steigerwald, T.W. Schulz, L.T. Schenker, M.B. Kennedy, P.H. Seeburg, G. Kohr, C-Terminal truncation of NR2A subunits impairs synaptic but not extrasynaptic localization of NMDA receptors, J. Neurosci. 20 (2000) 4573–4581. [53] J. von Engelhardt, I. Coserea, V. Pawlak, et al., Excitotoxicity in vitro by NR2Aand NR2B-containing NMDA receptors, Neuropharmacology 53 (2007) 10–17. [54] C.G. Thomas, A.J. Miller, G.L. Westbrook, Synaptic and extrasynaptic NMDA receptor NR2 subunits in cultured hippocampal neurons, J. Neurophysiol. 95 (2006) 1727–1734. [55] A.Z. Harris, D.L. Pettit, Extrasynaptic and synaptic NMDA receptors form stable and uniform pools in rat hippocampal slices, J. Physiol. 584 (2007) 509–519. [56] M.A. Martel, D.J. Wyllie, G.E. Hardingham, In developing hippocampal neurons, NR2B-containing N-methyl-d-aspartate receptors (NMDARs) can mediate signaling to neuronal survival and synaptic potentiation, as well as neuronal death, Neuroscience 158 (2009) 334–343. [57] F.X. Soriano, S. Papadia, F. Hofmann, N.R. Hardingham, H. Bading, G.E. Hardingham, Preconditioning doses of NMDA promote neuroprotection by enhancing neuronal excitability, J. Neurosci. 26 (2006) 4509–4518.