lable at ScienceDirect

Neurobiology of Aging 36 (2015) 680e692

Contents lists avai

Neurobiology of Aging

journal homepage: www.elsevier .com/locate/neuaging

Ab and NMDAR activation cause mitochondrial dysfunctioninvolving ER calcium release

Ildete Luísa Ferreira a,b, Elisabete Ferreiro a,b, Jeannette Schmidt a,b,c, João M. Cardoso d,Cláudia M.F. Pereira a,b,e, Ana Luísa Carvalho a,b,f, Catarina R. Oliveira a,b,e,A. Cristina Rego a,b,e,*

aCenter for Neuroscience and Cell Biology (CNC), University of Coimbra, Coimbra, Portugalb Institute for Interdisciplinary Research of the University of Coimbra (IIIUC), Coimbra, Portugalc PhD Programme in Experimental Biology and Biomedicine (PDBEB), Center for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugald Instrumentation Center, Physics Department, University of Coimbra, Coimbra, Portugale Faculty of Medicine, University of Coimbra, Coimbra, PortugalfDepartment of Life Sciences, University of Coimbra, Coimbra, Portugal

a r t i c l e i n f o

Article history:Received 6 November 2013Received in revised form 26 August 2014Accepted 2 September 2014Available online 6 September 2014

Keywords:Alzheimer’s diseaseAmyloid-b oligomersMitochondriaCalciumN-methyl-D-aspartate receptorGluN2B subunit

ILF and EF contributed equally for this study.* Corresponding author at: Center for Neuroscience

ulty of Medicine, University of Coimbra, Rua Larga, 3Tel.: þ351 239 820190; fax: þ351 239 822776.

E-mail address: a.cristina.rego@gmail.com (A.C. Re

0197-4580/$ e see front matter � 2015 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.neurobiolaging.2014.09.006

a b s t r a c t

Early cognitive deficits in Alzheimer’s disease (AD) seem to be correlated to dysregulation of glutamatereceptors evoked by amyloid-beta (Ab) peptide. Ab interference with the activity of N-methyl-D-aspartatereceptors (NMDARs) may be a relevant factor for Ab-induced mitochondrial toxicity and neuronaldysfunction. To evaluate the role of mitochondria in NMDARs activation mediated by Ab, we followed insitu single-cell simultaneous measurement of cytosolic free Ca2þ ðCa2þi Þ and mitochondrial membranepotential in primary cortical neurons. Our results show that direct exposure to Ab þ NMDA largelyincreased Ca2þi and induced immediate mitochondrial depolarization, compared with Ab or NMDA alone.Mitochondrial depolarization induced by rotenone strongly inhibited the rise in Ca2þi evoked by Ab orNMDA, suggesting that mitochondria control Ca2þ entry through NMDARs. However, incubation withrotenone did not preclude mitochondrial Ca2þ (mitCa2þ) retention in cells treated with Ab. Ab-inducedCa2þi and mitCa2þ rise were inhibited by ifenprodil, an antagonist of GluN2B-containing NMDARs.Exposure to Abþ NMDA further evoked a higher mitCa2þ retention, which was ameliorated in GluN2B�/�

cortical neurons, largely implicating the involvement of this NMDAR subunit. Moreover, pharmacologicinhibition of endoplasmic reticulum (ER) inositol-1,4,5-triphosphate receptor (IP3R) and mitCa2þ uni-porter (MCU) evidenced that Ab þ NMDA-induced mitCa2þ rise involves ER Ca2þ release through IP3Rand mitochondrial entry by the MCU. Altogether, data highlight mitCa2þ dyshomeostasis and subsequentdysfunction as mechanisms relevant for early neuronal dysfunction in AD linked to Ab-mediatedGluN2B-composed NMDARs activation.

� 2015 Elsevier Inc. All rights reserved.

1. Introduction

Alzheimer’s disease (AD) is the most common age-relatedneurodegenerative disease and the major cause of dementia inthe elderly. Clinically, this disorder is characterized by globalcognitive dysfunction, especially memory loss, and behavior andpersonality changes. The neuropathologic hallmarks of AD include

and Cell Biology (CNC), Fac-004-504 Coimbra, Portugal.

go).

ll rights reserved.

extracellular deposits of amyloid-b (Ab) peptide (produced after themetabolic processing of the amyloid precursor protein by b- and g-secretases), intracellular neurofibrillary tangles (consisting ofabnormally hyperphosphorylated tau protein), dystrophic neurites,and amyloid angiopathy. These histopathologic lesions were shownto be restricted to selective brain regions involved in memory andlanguage, namely, the hippocampus and the cerebral cortex, whichappear atrophic in AD patients (reviewed by Ferreira et al., 2010).Despite this, reduced dendritic spine density and synaptic integrityseem to better correlate with memory loss than histopathologicmarkers, which is consistent with the recognition of AD as a “syn-aptic disease.” Indeed, studies in AD postmortem brain tissuesamples and AD animal models support a role for disruption of

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 681

synaptic Ca2þ regulation in the neurotoxic action of Ab (reviewedby Camandola and Mattson, 2011; Mota et al., 2014), which mayserve as a trigger for synaptic deterioration driving the cognitiveloss in AD. Interestingly, Ab was shown to be produced in highamounts in synaptic terminals of the hippocampal dentate gyrus oftransgenic mice and deposited in extracellular plaques (Lazarovet al., 2002). Thus, impaired cognitive function and memory lossobserved in AD patients, associated to synaptic dysfunction, may bebecause of perturbed synaptic Ca2þ handling in response to over-activation of glutamate receptors, namely, the N-methyl-D-aspar-tate receptors (NMDARs) (Mota et al., 2014).

The most widely expressed NMDARs contain two obligatoryGluN1 subunits plus GluN2B or GluN2A or a mixture of the two.NMDARs are concentrated on postsynaptic spines of neuronaldendrites being subjected to particularly high levels of Ca2þ influx,adenosine triphosphate (ATP) demand, and oxidative stress(Mattson et al., 1998). Concordantly, memantine, an uncompetitiveantagonist of NMDARs, is used in AD patients and tested in severalclinical trials as a neuroprotective and clinical stabilizer compound(http://www.clinicaltrials.gov), supporting the involvement ofNMDARs on Ab-induced neuronal damage. In accordance, NMDARsmay be direct or indirect targets for Ab, affecting the activity ofthese receptors (reviewed in Malinow, 2012). Ab oligomers (5 mM)were reported to cause Ca2þ entry in neurons through NMDARsactivation and consequent mitochondrial Ca2þ (mitCa2þ) overload(Alberdi et al., 2010; Sanz-Blasco et al., 2008). Indeed, duringexcitotoxicity, mitCa2þ overload is one of the most important fac-tors linked to mitochondrial dysfunction, occurring along withdecreasedmitochondrial membrane potential (DJm) (e.g., Nicholls,2009). We previously demonstrated that NMDARs activation leadsto transient mitCa2þ loading (e.g., Ward et al., 2000), which may beaccompanied by oxidative damage to plasma membrane Ca2þ

extrusion pathways and inhibition of mitochondrial respiration.In AD postmortem brain and cellular and animal AD models,

mitochondrial dysfunction can be triggered by Ab (Pagani andEckert, 2011). Interestingly, synaptic failure is strongly associ-ated with synaptic mitochondrial dysfunction in AD, which occursbefore changes in nonsynaptic mitochondria (for review,Dragicevic et al., 2010; Du et al., 2012). Studies in a transgenicmouse brain overexpressing the human mutant form of amyloidprecursor protein and Ab revealed that synaptic mitochondriaundergo age-dependent accumulation of Ab and mitochondrialalterations (Du et al., 2010). Ab-induced mitochondrial dysfunc-tion was reported to be related to the interaction of Ab withdifferent mitochondrial proteins, including proteins of the outermitochondrial membrane, intermembrane space, inner mito-chondrial membrane, and the matrix, impairing oxidative phos-phorylation and mitochondrial dynamics and increasing reactiveoxygen species (ROS) production (Pagani and Eckert, 2011). In thiscontext, interaction of cyclophilin D (CypD, which is part of themitochondrial permeability transition pore) with mitochondrialAb potentiated mitochondrial, neuronal, and synaptic stress;conversely, CypD-deficient cortical mitochondria were resistant toAb- and Ca2þ-induced mitochondrial swelling and permeabilitytransition, along with increased calcium buffering capacity andproduction of fewer mitochondrial ROS (Du et al., 2008). Morerecently, genetic deletion of CypD suppressed Ab-mediated acti-vation of p38 mitogen-activated protein kinase signaling pathway,reversed axonal mitochondrial trafficking abnormalities, andimproved synaptic function (Guo et al., 2013).

A high number of studies demonstrated the importance ofNMDARs on Ab toxicity, but the downstream interplay betweenendoplasmic reticulum (ER) and mitochondria is yet to be revealed.Recently, we observed the involvement of ER stress linked tonicotinamide adenine dinucleotide phosphate oxidaseemediated

superoxide production in hippocampal neurons exposed to Aboligomers in a process involving the activation of NMDARs (Costaet al., 2012). To maintain normal cellular homeostasis, mitochon-dria and the ER directly communicate through the mitochondria-associated membranes (MAMs), which include inositol-1,4,5-triphosphate receptor (IP3R) on the ER and the voltage-dependent anion channel in mitochondria (reviewed in Ferreiroet al., 2012). The Ca2þ released from ER, through IP3R, can enterdirectly into mitochondria through the voltage-dependent anionchannel in the outer mitochondrial membrane and through themitCa2þ uniporter (MCU) in the inner mitochondrial membrane.Dysfunction of these contact sites and mitCa2þ overload can lead tothe opening of permeability transition pore, dissipation of the DJm,and activation of apoptotic cell death (Celsi et al., 2009). Interest-ingly, these receptors are involved in Ab-induced release of Ca2þ

from the ER, leading to depolarization of the mitochondrial mem-brane, release of cytochrome c on Bax translocation to mitochon-dria, and activation of apoptosis (Ferreiro et al., 2004, 2006, 2008).ER-mitochondria communication can thus play a fundamental rolein AD pathogenesis.

Prolonged exposure to high concentrations of Ab oligomers wasshown to induce Ca2þ entry through NMDAR and AMPA receptors,mitochondrial dysfunction, and oxidative stress (Alberdi et al.,2010). More recently, we also demonstrated that immediate expo-sure to Ab increases cytosolic free Ca2þ ðCa2þi Þ through the activa-tion of NMDARs containing GluN2B subunits (Ferreira et al., 2012)and further demonstrated that Ab can directly interact with theextracellular domain of GluN1 and GluN2B subunits and evoke ERstress downstream of GluN2B-composed NMDARs (Costa et al.,2012). Thus, the aim of this study was to evaluate the immediateeffect of Ab linked to NMDARs activation on mitochondrial functionand the intricate role of mitochondria and ER interplay in Ca2þ

homeostasis. Using in situ single-cell synchronizedmeasurement ofCa2þi and DJm in rat primary brain cortical neurons, we show thatsimultaneous exposure to Ab and NMDA affects the response toGluN2B-composed NMDARs, largely causing mitochondrial depo-larization and mitCa2þ retention. Moreover, mitCa2þ accumulationvia the MCU involved the release of ER Ca2þ through the IP3Rs.These deleterious effects triggered by the simultaneous exposure toAb and activation of NMDARs may thus potentiate the neurode-generative process in AD and contribute to cognitive impairmentand dementia.

2. Methods

2.1. Primary neuronal cultures

2.1.1. Rat cortical culturesPrimary neuronal cultures of rat cerebral cortex were prepared

as described previously (Agostinho and Oliveira, 2003) with someminormodifications. Briefly, frontal cerebral corticeswere dissectedfrom 16-day Wistar rat embryos and collected in Ca2þ Mg2þ-freeKrebs medium containing (in millimolar) 120 NaCl, 4.33 KCl,1.2 KH2PO4, 25.5 NaHCO3, 13 glucose, 10 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), pH 7.4, plus 0.3% (w/v)fatty acidefree bovine serum albumin (BSA). Tissues were treatedwith 0.035% (w/v) trypsin in BSA-Krebs medium, for 7 minutes at37 �C, followed by addition of 0.038% (w/v) trypsin inhibitor andcentrifugation at 140� g for 5 minutes. After a washing step, cellswere resuspended in neurobasal medium (Gibco, Life Technologies,Paisley, UK) supplemented with 2% (v/v) B27 (Gibco), 0.5 mMglutamine, and 50 mg/mL gentamicin and plated in poly-D-lysi-neecoated glass coverslips, at a density of 0.2� 106/cm2. Cells werecultured for 8e9 days in vitro, in a humidified incubator chamberwith 95% air and 5% CO2 at 37 �C.

Fig. 1. N-methyl-D-aspartate (NMDA)einduced and amyloid-beta (Ab)einduced changes in cytosolic free Ca2þ ðCa2þi Þ and mitochondrial membrane potential (DJm). Corticalneurons were stimulated with NMDA (100 mM) or Ab (0.5 mM). (A) Representative F340/F380 or rhodamine 123 (Rh123) fluorescence (arbitrary unit) traces from individual cells. (B)Peak amplitude of Fura-2 or Rh123 fluorescence after NMDA or Ab, before and after FCCP (2 mM) exposure. Data are the mean � standard error of the mean of 3e6 independentexperiments, comprising a total of 163e270 cells per experimental group. Statistical analysis: ***p < 0.001 significantly different compared with the control (one-way analysis ofvariance with Bonferroni multiple comparison test). M, MK-801; F, FCCP; and O, oligomycin.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692682

2.1.2. Mice cortical cultures from GluN2Bþ/þ and GluN2B�/�

embryosBecause GluN2B(�/�) mice die shortly after birth, GluN2B(þ/�)

mice (Kutsuwada et al., 1996) were mated to obtain GluN2B(�/�)and littermate control GluN2Bþ/þ embryos, which were used toculture cortical neurons. Genotyping of pups and embryos wasperformed according to a previously described protocol (Tovaret al., 2000). Briefly, DNA was extracted with phenol/choloroform/isoamyl alcohol (Sigma, St Louis, MI, USA) after tissue digestionwithproteinase K (0.1 mg/mL; Invitrogen, Life Technologies, Paisley, UK).polymerase chain reaction amplification using the Supreme NZY-Taq 2� Green Master Mix (Nzytech, Lisboa, Portugal) was per-formed with specific primers for GluN2B (50 ATG AAG CCC AGC GCAGAG TG 30 and 50 AGG ACT CAT CCT TAT CTG CCA TTA TCA TAG 30)and for the neomycin cassette (50 GGC TAC CTG CCC ATT CGA CCACCA AGC GAA AC 30). GluN2Bþ/þ and GluN2B�/� cortical neuronalcultures were prepared from 17 to 18 days embryonic mice. Corticeswere dissected and maintained in Hibernate E (Brain Bits, Spring-field, IL, USA) supplemented with NeuroCult SM1 (Stemcell, Gre-noble, France) at 4 �C overnight while genotyping was performed.

Tissues from the same genotype were pooled together and disso-ciated with papain (20 units/mL, 10 minutes, 37 �C; WorthingtonBiochemical Corporation, Lakewood, NJ, USA) and deoxyribonu-clease I (0.2 mg/mL, Invitrogen). Cortical neurons were plated at adensity of 0.2 � 106/cm2 on coverslips coated with poly-D-lysine inminimum essential medium supplemented with 10% horse serum.After w3 hours in culture, the plating medium was replaced byneurobasal medium (Gibco) supplemented with 2% NeuroCult SM1,0.5 mM glutamine (Sigma), 0.125 mg/mL gentamicin (Gibco), andinsulin (20 mg/mL; Sigma). Neurons were maintained at 37 �C in ahumidified incubator of 5% CO2. Cultures were used after 8e9 daysin vitro.

2.2. Experimental conditions

Experiments were performed in cells subjected to direct stim-ulation with 100-mM NMDA (Tocris, Bristol, UK) and/or 0.5 mM Ab(American Peptide, Sunnyvale, CA, USA). Pre-exposure of cells toAb during 2 hours, before NMDA treatment, was also tested.Ab1e42 preparation containingw60% of low n oligomers andw40%

Fig. 2. Influence of ifenprodil on amyloid-beta (Ab)einduced cytosolic free Ca2þ ðCa2þi Þ and mitochondrial membrane potential (DJm). Cortical neurons were stimulated with N-methyl-D-aspartate (NMDA) (100 mM) or Ab (0.5 mM) in the absence or presence of ifenprodil (selective GluN2B subunit antagonist, 10 mM). (A) Representative F340/F380 orrhodamine 123 (Rh123) fluorescence (arbitrary unit) traces from individual cells. (B) Peak amplitude of Fura-2 fluorescence after NMDA or Ab stimulation in the absence or presenceof ifenprodil, before and after FCCP (2 mM) exposure. For single-cell imaging studies, data are the mean � standard error of the mean of 3e6 independent experiments, comprising atotal of 118e230 cells per experimental group. Statistical analysis: ***p < 0.001 significantly different compared with NMDA; ###p < 0.001 significantly different compared with Ab(one-way analysis of variance with Bonferroni multiple comparison test). M, MK-801; F, FCCP; and O, oligomycin.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 683

of monomers was obtained, as previously described (Ferreira et al.,2012). To assess the involvement of NMDARs in regulating Ca2þihomeostasis, cells were preincubated for 5 minutes with ifenprodil(10 mM; Sigma). To modulate mitochondrial function, cells wereincubated for 3 minutes with 2 mM rotenone before the

stimulation with NMDA and/or Ab. To inhibit mitCa2þ accumula-tion and to evaluate the role of the MCU, cells were incubated with10 mM ruthenium 360 (Ru360; Calbiochem, Darmstadt, Germany)during rhodamine 123 (Rh123) and Fura-2/AM loading and withfresh Ru360 during the experiment. To evaluate the contribution

Fig. 3. Modulation of mitochondrial function by rotenone pre-exposure on N-methyl-D-aspartate (NMDA) or amyloid-beta (Ab)einduced cytosolic free Ca2þ ðCa2þi Þ and mito-chondrial membrane potential (DJm). Cortical cells were stimulated with 0.5 mM Ab or 100 mM NMDA in the absence or in the presence of rotenone (complex I inhibitor, 2 mM). (A)Magnification of the rotenone effect on DJm. (B) Representative traces of Fura-2 or rhodamine 123 (Rh123) fluorescence after NMDA, Ab, or rotenone pre-exposure. (C) Peakamplitude of Fura-2 and Rh123 fluorescence after NMDA or Ab stimulation, before and after FCCP (2 mM) stimulation. Data are the mean � standard error of the mean of 3e4independent experiments, comprising a total of 123e199 cells per experimental group. Statistical analysis: ***p < 0.001 significantly different compared with NMDA alone, #p < 0.05and ###p < 0.001 significantly different compared with Ab (one-way analysis of variance with Bonferroni multiple comparison test). M, MK-801; F, FCCP; and O, oligomycin.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692684

of ER Ca2þ to the mitCa2þ content, cells were preincubated during1 hour with 1 mM of xestospongin C (XeC; Tocris).

2.3. Monitoring dynamic changes in Ca2þi and DJm

Cortical neurons, plated on glass coverslips, were washed withNaþ medium containing (in millimolar) 140 NaCl, 5 KCl, 1 CaCl2, 1MgCl2, 10 glucose, 10 HEPES, and pH 7.4/NaOH and loaded with Naþ

medium supplemented with the DJm-sensitive probe (1.3 mMRh123, quench mode) and 0.1% BSA for 10 minutes at room tem-perature. Then, glass coverslips were transferred to Naþ medium

containing 0.2% (w/v) pluoronic acid, 0.1% (w/v) BSA, 1.3 mM Rh123,and 5 mM Fura-2/AM for 30 minutes at 37 �C. The fluorescenceprobes Fura-2/AM and Rh123 were obtained from MolecularProbes, Life Technologies (Eugene, OR, USA). To directly assess therole of mitochondria, glycolysis was inhibited by adding 2-deoxy-D-glucose (2 mM) and by replacing glucose with pyruvate (10 mM) inNaþ medium (pyruvate-based medium) (e.g., Oliveira et al., 2006;Rego et al., 2001). After rinsing with fresh buffer, coverslips wereassembled in a nonperfused chamber filled with 500 mL of Mg2þ-free pyruvate medium containing 20 mM glycine, at room temper-ature. Simultaneous Ca2þi levels and Rh123 fluorescence were

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 685

measured by using an inverted fluorescence microscope Axiovert200 (Zeiss, Jena, Germany) equipped with a dual-band path emis-sion filter (510/40 and 600/60 nm) and a Lambda DG4 apparatus(Sutter Instruments Company, Novato, CA, USA). In situ calibrationof Ca2þi responses was performed at the end of every individualexperiment, by determining maximal 340/380 ratio for each indi-vidual cell, after the addition of 2 mM ionomycin (Sigma). Maximalmitochondrial depolarization (DJm collapse) was also performedin every individual experiment by adding a protonophore (2-mMcarbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP);Sigma), which was preceded by oligomycin (2 mg/mL; Sigma) toprevent ATP synthase reversal. The determinations of fluorescencetime courses were calculated on the basis of 1 microscopic field percoverslip, containing w75 cells. Moreover, individual whole cellswere identified as regions of interest.

2.4. Data analysis

Data analysis of single-cell imaging was performed with sys-tematic custom scripts of Mathworks Matlab 2007a (32 bits) forfluorescence time-course parametrization. For each experimentcondition, a cell response normalization procedure to the time-course basal level (first value obtained) was performed. The peakfrom the first stimulus amplitude histogram of Fura-2 (340/380ratios) time course was used to interactively set a cut-off level,adjusted for each experiment, to remove low-amplitude timecourses from nonresponsive cells. Therefore, only data fromresponsive cells were considered for peak amplitude measure-ments. These were performed by means of the highest 340/380ratio difference in the vicinity of the registered peak in analysis.Notice that the cut-off exclusion criteria were not applied in theexperiments involving rotenone and ifenprodil pre-exposure. Eachexperimental condition was assayed at least in three differentcoverslips from three independent culture preparations.

Data were expressed as mean � standard error of the mean ofthe number of experiments indicated in Figs. 1e6. Comparisonsamong multiple groups were performed by one- or two-wayanalysis of variance, followed by Bonferroni post hoc test or Stu-dent t test, as described in the figure legends. Significance wasaccepted at p < 0.05.

3. Results

In this work, we detailed the single-cell simultaneous changes inCa2þi and DJm in brain cortical cells stimulated with NMDA (100mM), Ab (0.5 mM), simultaneous exposure to Ab plus NMDA (Ab þNMDA), or incubation with NMDA in cells pre-exposed for 2 hoursto Ab. These changes were followed in basal conditions or aftermaximal mitochondrial depolarization achieved by the mitochon-drial uncoupler FCCP in conditions of ATP synthase inhibition, in thepresence of oligomycin, to prevent ATP hydrolysis.

3.1. Ab induces immediate Ca2þi rise and depolarization ofmitochondrial membrane

Our results show that NMDA and, to a less extent, Ab cause arobust increase in Ca2þi levels (Fig. 1A and Bi) associated with aslight decrease in DJm (Fig. 1Bii), compared with the maximumdepolarization obtained with FCCP (see representative traces inFig. 1A). Complete mitochondrial depolarization induced by FCCP incells treated with NMDA increased Rh123 fluorescence comparablewith the control, indicating unaltered DJm w20 minutes after theinitial stimuli; Ab-treated cells showed reduced mitochondrialaccumulation of Rh123 after incubation with FCCP, indicatingdecreased DJm (Fig. 1A and Biv). Importantly, the higher increase

in Ca2þi levels achieved by NMDA and Ab was associated with anincrease in mitCa2þ, as represented by the rise in mitochondrialFCCP-releasable Ca2þ pool (Fig. 1A and Biii). Our results show thatimmediate stimulation of cortical cells with Ab induces a significantincrease in Ca2þi levels and mitochondrial membrane depolariza-tion, associated with mitCa2þ retention.

3.2. Ab-induced Ca2þi rise and mitochondrial depolarization areprevented by selective blockade of GluN2B subunits

To characterize the involvement of NMDARs, in particular theGluN2B subunit, in Ca2þi changes evoked by NMDA or Ab, as previ-ously demonstratedbyus (Costa et al., 2012; Ferreira et al., 2012), cellswere stimulated with NMDA or Ab in the absence or presence of theselective GluN2B subunit antagonist ifenprodil (Fig. 2). Our resultsdemonstrate that Ca2þi increase exerted by NMDA or Ab was signifi-cantly counteracted by ifenprodil (Fig. 2A and Bi). Interestingly, thedisruption of DJm induced by NMDA and Abwas not affected by theNMDARs antagonist (Fig. 2A and Bii). Moreover, the increase inmitCa2þ retention elicited by NMDA and by Ab was also partiallydecreased (approximately by half) in the presence of ifenprodil(Fig. 2A and Biii). These data show a fundamental role of GluN2Bsubunits in defining Ab-evoked Ca2þi rise andmitCa2þ accumulation.Maximal Rh123 cytosolic release achieved by oligomycinplus FCCP incells exposed to NMDA was shown to be slightly increased in thepresence of ifenprodil (Fig. 2A and Biv), suggesting a rise in DJm.

3.3. Rotenone-induced mitochondrial depolarization modulates Ab-induced changes in Ca2þi and DJm

To evaluate the contribution of functional mitochondria incontrolling Ca2þi levels, cells were incubated for 3 minutes with2 mM rotenone (complex I inhibitor and depolarizing agent) beforestimulation with NMDA or Ab. These were added under conditionsof mild depolarization evoked by rotenone, that is, before completerotenone depolarization (that was achieved after w15 minutes in-cubation), to detect synergistic effects of NMDA or Ab. Resultsdepicted in Fig. 3A demonstrate that rotenone per se induced asignificant decrease in DJm, as observed by the large increase inRh123 fluorescence. Moreover, mitochondrial dysfunction inducedby rotenone attenuated the rise in Ca2þi induced by immediateexposure to NMDA or Ab (Fig. 3B and Ci) and enhanced mito-chondrial depolarization in response to NMDA and Ab stimulation(Fig. 3B and Cii). Complete mitochondrial depolarization, achievedby oligomycin plus FCCP in cells preincubated with rotenone, evi-denced a decrease in mitCa2þ retention in cells exposed to NMDA.Conversely, a very slight but significant increase in mitCa2þ reten-tion was observed in Ab-treated cells pre-exposed to rotenone(Fig. 3B and Ciii), which may suggest that Ab-evoked mitCa2þ

retention is not primarily dependent on the maintenance of DJm.Moreover, oligomycin-evoked mitochondrial depolarization wasobserved in cells treated with rotenone (Fig. 3B), indicative ofmitochondrial ATP hydrolysis (e.g., Ward et al., 2000). Despite this,and as expected, exposure to FCCP evidenced a complete mito-chondrial depolarization (Fig. 3B and Civ).

These results suggest that mitochondrial function modulatesCa2þi rise after exposure to NMDA and Ab. In contrast with NMDA,Ab-evoked mitCa2þ accumulation does not seem to be accountedfor by preserved mitochondrial function.

3.4. Increased mitCa2þ accumulation and decreased DJm afterexposure to Ab þ NMDA

Impairment in glutamatergic neurotransmission has been sug-gested to occur concomitantly with the presence of extracellular

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692686

levels of Ab oligomers (Parameshwaran et al., 2008). Thus, the effectof simultaneous addition of NMDA and Ab or NMDA in cells pre-exposed to Ab for 2 hours was then evaluated. Interestingly, Ab þNMDA stimulation produced a significant rise in Ca2þi levels,comparedwith NMDAor Ab alone (Fig. 4B and Ci), and preincubationwith Ab for 2 hours significantly decreased Ca2þi levels in response toNMDA stimulation, compared with cells not previously exposed toAb (Fig. 4B and Ci). Both pre- and simultaneous exposure to Abaffected the response to NMDAR activation, largely causing anadditional immediate mitochondrial depolarization, compared withAb or NMDA alone (Fig. 4B and Cii). Notably, mitCa2þ retention ca-pacity was largely increased on addition of Ab þ NMDA, comparedwith NMDA or Ab alone, or even after NMDA stimulation in cells pre-exposed to Ab (Fig. 4B and Ciii). Moreover, cortical cultures treatedwith Ab þ NMDA or NMDA on pre-exposure to Ab showed reducedretention of Rh123, compared with NMDA or Ab alone, evidencingincreased mitochondrial depolarization (Fig. 4B and Civ).

These data suggest the activation of different cellular mecha-nisms after addition of Ab þ NMDA, compared with previousexposure with Ab followed by NMDA exposure. Indeed, incubationfor 2 hours with Ab per se caused a slight but significant rise in Ca2þiand mitochondrial depolarization, in comparison with untreatedconditions (Fig. 4A).

3.5. Decreased Ca2þi rise and mitochondrial depolarization inGluN2B(�/�) cortical neurons

To confirm the involvement of the GluN2B subunit in both Ca2þilevels and mitochondrial membrane potential, we tested corticalcultures derived from GluN2B(�/�) mice treated with NMDA, Ab,and Ab þ NMDA (Fig. 5). Our results demonstrated that Ca2þi in-crease exerted by the 3 stimuli was significantly reduced inGluN2B(�/�) cortical cultures, compared with GluN2Bþ/þ cells(Fig. 5A and Bi) and closely associated with lower mitochondrialdepolarization (Fig. 5A and Bii). The extent of Ca2þi rise and mito-chondrial depolarization was higher after exposure to Ab þ NMDA(as verified in Fig. 4Ci and ii), compared with NMDA or Ab stimu-lation alone. Interestingly, mitCa2þ content was significantly higherin GluN2B(�/�) compared with GluN2Bþ/þ cells, in the absence(control) or presence of NMDA (Fig. 5A and Biii). Exposure to Ab þNMDA in GluN2Bþ/þ cells was associated with increased mitCa2þ

(Fig. 5Biii), similarly as observed in rat cortical cells (Fig. 4Ciii).However, exposure to Ab or Ab þ NMDA in cells from GluN2B(�/�)mice largely reduced the capacity of mitochondria to retain Ca2þ

(Fig. 5Biii). This lower mitCa2þ retention after Ab exposure waslinked to increased mitochondrial membrane potential inGluN2B(�/�)ederived cells (Fig. 5A and Biv), confirming an impor-tant role of this NMDAR subunit in facilitating mitCa2þ accumula-tion and the consequent mitochondrial depolarization.

3.6. Increased mitCa2þ accumulation induced by Ab þ NMDA ismodulated by IP3R and MCU

Previous studies showed that Ab activates the IP3-generatingphospholipase C and triggers ER Ca2þ release via IP3R-sensitivechannels (Resende et al., 2008) and that ER-to-mitochondria Ca2þ

transfer is involved in neuronal dysfunction induced by Ab (Ferreiroet al., 2008). To clarify the mechanism involved in enhancedmitCa2þ

retention that occurs in cells exposed to Abþ NMDA (Fig. 4Ciii), bothCa2þi entry and the IP3R-mediated transfer of Ca2þ from the ER tomitochondria were evaluated and compared with Ab treatment. Forthis purpose, cortical cells were preincubated with Ru360, an MCUinhibitor, or with XeC, an inhibitor of the IP3R (Fig. 6).

Our results demonstrate that inhibition of Ca2þ release throughthe IP3R significantly attenuated the increase in Ca2þi levels (Fig. 6i)

and reduced mitochondrial depolarization (Fig. 6i) evoked bysimultaneous exposure to Ab þ NMDA (as shown in Fig. 4B and4Ci and ii). Importantly, a reduction in mitochondria Ca2þ retentionwas observed when the cells were pre-exposed to Ru360 andtreated with Ab þ NMDA (Fig. 6ii), suggesting that under thisexperimental condition, MCU contributes for mitCa2þ retention. Inthese cells, a significant decrease in Ca2þ uptake by mitochondria(Fig. 6iii) and a rescue in mitochondrial membrane potential(Fig. 6iv) were also observed when the release of Ca2þ through IP3Rwas blocked. These data demonstrate that simultaneous exposureto AbþNMDA causes the release of Ca2þ through IP3R and the entryto mitochondria through the MCU.

4. Discussion

The present study uncovers a fundamental mechanism un-derlying Ab-induced synaptic dysfunction, namely, the link be-tween NMDARs activation and mitochondrial dysregulation,particularly their capacity to handle Ca2þ in cultured corticalneurons. By using in situ single-cell simultaneous measurementof Ca2þi levels and DJm, we show that cells challenged simul-taneously with Ab oligomericeenriched preparation, the mainsynaptotoxic species involved in AD (Haass and Selkoe, 2007;Klein, 2006) and to the NMDARs agonist NMDA exhibitelevated Ca2þi levels, leading to mitCa2þ retention and depolari-zation through a pathway that involves both ER IP3R and theMCU. The data evidence the dynamics between NMDARs activityand ER-mitochondria interplay in Ca2þ homeostasis and theresulting mitochondrial (dys)function in response to an imme-diate effect of Ab.

Sustained changes in Ca2þ homeostasis that occur during agingwere earlier described to be in the origin of the development ofneurodegenerative diseases (Gibson and Peterson, 1987), namelyAD. Importantly, NMDARs dysregulation evoked by Ab and theconsequent loss of Ca2þ homeostasis are thought to be related tothe early cognitive deficits observed in this disorder. Severalmechanisms have been suggested for Ab-mediated Ca2þi rise,namely: (1) formation of Ab Ca2þ-permeable pores in the plasmamembrane (Arispe et al., 1993); (2) activation of pre-existing ionchannels, which include the NMDARs (De Felice et al., 2007;Ferreira et al., 2012); and/or (3) induction of Ca2þ release from in-ternal stores (Ferreiro et al., 2008; Resende et al., 2008; reviewed inSupnet and Bezprozvanny, 2010). By either being a direct or indirecttarget of Ab, NMDARs play a fundamental role in Ab toxicity(reviewed in Malinow, 2012). Ab was reported to bind to or be inclose proximity to NMDARs, triggering neuronal damage throughNMDARs-dependent Ca2þ influx (Costa et al., 2012; De Felice et al.,2007). Indeed, we demonstrated that, in rat hippocampal neurons,blockade of extracellular domains of GluN1 or GluN2B, but notGluN2A, subunits reduces the binding of Ab oligomers (Costa et al.,2012). Conversely, Ab-induced activation of GluN2A-containingNMDARs in non-neuronal cells, namely in Xenopus laevis oocytes(Texidó et al., 2011) and HEK293cells (Domingues et al., 2007), waspreviously reported. In the AD brain and human cortical neurons,excitatory synapses containing the GluN2B subunit of the NMDARsappear to be the main sites of oligomer accumulation (Deshpandeet al., 2009). Recently, we showed that Ab increased Ca2þi levelsthrough the activation of NMDARs containing GluN2B subunits,shown to be present at the membrane surface in cortical neurons(Ferreira et al., 2012), which may result in microtubule disassemblyand reduced neurite length (Mota et al., 2012). One possibledestination for Ca2þ that enters through the NMDARs are the Ca2þ

internal stores, namely, the mitochondria and the ER. Recently, weshowed that Ab triggers ER stress by an NMDARs-dependentmechanism, leading to neuronal dysfunction (Costa et al., 2012).

Fig. 4. Single-cell analysis of free cytosolic Ca2þ ðCa2þi Þ and mitochondrial membrane potential (DJm) on N-methyl-D-aspartate (NMDA), amyloid-beta (Ab), Ab þ NMDA stimulationor NMDA in cells pre-exposed to Ab. Cortical cells were stimulated with 100 mM NMDA, 0.5 mM Ab, Ab þ NMDA, or NMDA in cells pre-exposed for 2 hours to Ab. (A) Basal levels ofFura-2 or rhodamine 123 (Rh123) fluorescence after 2 hours of exposure to Ab. (B) Representative F340/F380 or Rh123 fluorescence (arbitrary unit) traces from individual cells.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 687

Fig. 5. Role of GluN2B subunits on single-cell analysis of Ca2þi and mitochondrial membrane potential (DJm). GluN2B�/� versus GluN2Bþ/þ mouse cortical neurons were stimulatedwith 100 mM N-methyl-D-aspartate (NMDA), 0.5 mM amyloid-beta (Ab), or Ab þ NMDA. (A) Representative F340/F380 or rhodamine 123 (Rh123) fluorescence (arbitrary unit) tracesfrom individual cells. (B) Peak amplitude of Fura-2 or Rh123 fluorescence after NMDA, Ab, or Ab þ NMDA stimulation before or after FCCP (2 mM) stimulation. Data are the mean �standard error of the mean of 3e4 independent experiments, comprising a total of 69e358 cells per experimental group. Statistical analysis: ***p < 0.001 when GluN2B�/� cellsare compared with GluN2Bþ/þ cells; ##p < 0.01 and ###p < 0.001 compared with control conditions; &p < 0.05, &&p < 0.01, and &&&p < 0.001 compared with NMDA alone; and$$$p < 0.001 compared with Ab alone (two-way analysis of variance with Bonferroni multiple comparison test). M, MK-801; F, FCCP; and O, oligomycin.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692688

The present results demonstrate that Ab preparation enriched inoligomeric forms induces Ca2þi rise and immediate mitochondrialdepolarization, similarly to the selective activation of NMDARs withNMDA and that part of Ca2þi is further accumulated within mito-chondria. Our experiments were performed in the presence ofglycine, which favors the activation of extrasynaptic GluN2B-containing NMDARs. These observations are in accordance withthe findings in the CA1 region of the hippocampus demonstrating

(C) Peak amplitude of Fura-2 or Rh123 fluorescence after NMDA, Ab, Ab þ NMDA stimulstimulation. Data are the mean � standard error of the mean of 10e15 independent experim*p < 0.05, ***p < 0.001 significantly different compared with the control (A, by Student t testAb (one-way analysis of variance with Bonferroni multiple comparison test). M, MK-801; F

=

that synaptic and extrasynaptic NMDARs are gated by differentendogenous co-agonists, respectively, D-serine and glycine (Papouinet al., 2012).We confirmed that Ca2þi changeswere largelymediatedby GluN2B subunits because both Ca2þi and mitCa2þ accumulationwere prevented by ifenprodil, a selective inhibitor of GluN2B-containing NMDARs. This inhibition of Ca2þ entry throughNMDARs did not affect DJm as determined immediately after theNMDA or Ab stimuli and slightly induced mitochondrial Rh123

ation or NMDA in cells pre-exposed for 2 hours to Ab, before or after FCCP (2 mM)ents, comprising a total of 591e676 cells per experimental group. Statistical analysis:

) or NMDA exposure (C); ##p < 0.01, ###p < 0.001 significantly different compared with, FCCP; and O, oligomycin.

Fig. 6. Modulation by inositol-1,4,5-triphosphate receptor (IP3R) and mitochondrial Ca2þ uniporter of Ca2þi and mitochondrial membrane potential (DJm) after exposure toamyloid-beta (Ab) or Ab þ N-methyl-D-aspartate (NMDA). Cortical cells were stimulated with 0.5 mM Ab or Ab þ NMDA in the absence or presence of Ru360 (mitCa2þ uniporterinhibitor, 10 mM) or xestospongin C (XeC) (IP3R inhibitor, 1 mM). The results are expressed as the percentage of the respective control (Ab or Ab þ NMDA), in the absence of theinhibitors; data are the mean � standard error of the mean of 3e5 independent experiments, comprising a total of 115e366 cells per experimental group. Statistical analysis: ***p <

0.001 significantly different compared with the respective control (one-way analysis of variance with Bonferroni multiple comparison test).

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 689

retention (as shown in NMDA-treated cells), suggesting that theobserved decrease in NMDARs-coupled Ca2þ entry is not sufficientfor largely modifying the DJm in cultured neurons. However, theclose relationship between mitochondrial function and NMDARsactivation is evident in rotenone-treated cells. Thus, ourdata suggestthat Ab-induced Ca2þi dyshomeostasis is, at least in part, because ofexcessive activation of NMDARs composed by GluN2B subunits.

Excitotoxicity is normally precluded by astrocytic glutamatetransporters, an event that is counteracted by Ab because of thedownregulation of the astrocytic glutamate uptake capacity (Matoset al., 2008). Concomitantly with AD progression, an imbalance inglutamatergic neurotransmission, because of the increase inglutamate levels at the synaptic cleft, may occur along with extra-cellular Ab oligomer accumulation, which may potentially exacer-bate NMDARs activation and promote Ca2þi dyshomeostasis. Ourprevious work demonstrated that simultaneous addition of Ab þNMDA leads to a large increase in Ca2þi levels compared with theeffect of NMDA or Ab alone and that this effect is prevented byGluN2B-containing NMDARs (Ferreira et al., 2012). The presentstudy further demonstrates that enhanced mitCa2þ evoked by Ab þNMDA is partially precluded in GluN2B�/� cortical neurons, impli-cating this NMDAR subunit in facilitating mitCa2þ accumulationand the consequent mitochondrial depolarization. Moreover, sus-tained incubation (for 2 hours) with Ab induced a slight but sig-nificant decrease in Ca2þi levels evoked by further stimulation withNMDA, along with a significant increase in mitochondrial depo-larization, compared with that occurring in response to NMDAalone. Although NMDA stimulation in cells subjected to Ab pre-exposure induced similar levels of mitCa2þ retention, resembling

those obtained with NMDA alone, mitochondria were largelydepolarized, which may be accounted for by the prolonged pre-exposure to Ab. This indicates that mitochondrial depolarizationoccurred during Ab exposure (before incubation with NMDA),impairing Ca2þi increase, which appears to be in accordance withthe effect of rotenone on NMDARs modulation. Exposure for 2hours with high concentration of Ab1e42 oligomers (5 mM) waspreviously shown to cause mitochondrial depolarization andoxidative stress, dependent on the activation of NMDA and AMPAreceptors (Alberdi et al., 2010).

Interestingly, in the presentwork,we show that Ab andNMDAco-stimulation elevate Ca2þi levels and that this increase in Ca2þ isassociated with enhanced mitCa2þ retention and mitochondrial de-polarization. In accordance with our data, inhibiting mitCa2þ clear-ance accelerated mitochondrial depolarization (Stanika et al., 2012).Indeed, mitCa2þ overload, along with mitochondrial depolarization(DJm dissipation), is strongly linked to mitochondrial dysfunction(e.g., Nicholls, 2009), reinforcingmitochondria as a target of Ab in ADpathology (for review, Pagani and Eckert, 2011). Intraneuronal Abaccumulation occurs early in the neuropathologic phenotype of AD,before neurofibrillary tangle formation and plaque deposition(Gouras et al., 2000). Intracellular Abwas suggested to occur becauseof internalization of extracellular Ab after the interaction of prese-creted Ab with membrane transporters and receptors, including a-7nicotinic acetylcholine receptors, advanced glycation end productsreceptors, and NMDARs (reviewed by Caldeira et al., 2013). Abuptake was completely blocked by NMDAR antagonists, evidencingan involvement of this receptor in the re-uptake of the peptide(Bi et al., 2002). Moreover, it was previously demonstrated that Ab is

Fig. 7. Proposed molecular pathways for amyloid-beta (Ab) or Ab þ N-methyl-D-aspartate (NMDA) modulation of Ca2þi and mitochondrial membrane potential (DJm). Ab induces anincrease in Ca2þi and an immediate mitochondrial depolarization in cortical neurons, mediated by GluN2B-containing NMDA receptor (NMDAR) activation. Ca2þ entry throughNMDAR is taken up by the endoplasmic reticulum (Costa et al., 2012), which may be released by the inositol-1,4,5-triphosphate receptor (IP3R) (Resende et al., 2008). After exposureto Ab þ NMDA, Ca2þi increase is potentiated, inducing mitochondrial depolarization and enhanced mitochondrial Ca2þ retention through a pathway that involves the release of Ca2þ

through IP3R and the entry to the mitochondria by the mitCa2þ uniporter. (For interpretation of the references to color in this figure, the reader is referred to the web version of thisarticle.)

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692690

transported into mitochondria via the translocase of the outermembrane machinery and is located in mitochondria cristae(Hansson Petersen et al., 2008). Ab interferes with oxidative phos-phorylation and ROS production within mitochondria and causesdecreased DJm, complex IV (cytochrome c oxidase) activity, and ATPproduction (Hauptmann et al., 2009). By using isolated rat brainmitochondria treated with Ab, both mitochondrial transmembranepotential and the mitochondrial capacity to accumulate Ca2þ wereshown to be decreased and to cause complete uncoupling of respi-ration (Moreira et al., 2001). More recently, hippocampal and corticalmitochondria of cognitive impaired AD transgenic mice showed highlevels of dysfunction, whereas striatal and amygdala mitochondriawere less affected (Dragicevic et al., 2010), reinforcing increasedmitochondrial dysfunction in selective brain areas mostly affected inAD. In addition, transcription regulators related with mitochondrialbiogenesis and antioxidant defenses are altered in different ADmodels (reviewed by Caldeira et al., 2013), implicatingmitochondrialdysfunction in AD.

In our model, rotenone-induced high mitochondrial depolari-zation reduced Ca2þi increase promoted by Ab or NMDA, implicatingthe fundamental role of functional mitochondria in the activity ofionotropic glutamate receptors. This occurred concomitantly withmitochondrial ATP hydrolysis (as evidenced by oligomycin-evokeddepolarization in rotenone-treated cells); notably, no significantchanges in total ATP/adenosine diphosphate levels were observedin cells exposed to Ab, NMDA, or Ab þ NMDA for w20 minutes,using the same protocol of single-cell analysis up to oligomycinaddition (data not shown). Additionally, in cells incubated withrotenone and then exposed to NMDA, the decrease in Ca2þi levelswas accompanied by a decrease in mitCa2þ retention. In the pre-vious studies, blockade of mitCa2þ sequestration induced by oli-gomycin plus rotenone in cortical brain slices caused a reduction inCa2þ influx through the NMDAR channel after glutamate treatment(Kannurpatti et al., 2000). Although this may be hypothesized forNMDA-exposed cells, an alternative rationale should be made forAb. In fact, although pre-exposure to rotenone inhibited Ab-induced Ca2þi increase, mitochondria were still able to accumulateCa2þ. Thus, oligomeric Ab may not only interact with NMDARs butalso interfere with mitochondria function/activity, as describedpreviously.

Mitochondria and the ER directly communicate to maintainnormal cellular homeostasis. Recently, itwas shown that presenilin-1and -2, components of the g-secretase complex, are enriched in theMAMs (Area-Gomez et al., 2009) and that mutations in presenilin-1and -2 upregulate MAM function and increase ER-mitochondriaconnectivity (Area-Gomez et al., 2012). We have previously shownin cortical neurons that Ab1e40, in a fibrillar state, induces the releaseof ERCa2þ through IP3RandRyRchannels (Ferreiroet al., 2004),whichis implicated in mitochondrial depolarization, release of cytochromec, translocation of Bax to mitochondria, and apoptosis (Ferreiro et al.,2006, 2008). Our results largely suggest that the potentiation ofmitCa2þ retention thatoccurs incells exposedtoAbþNMDAis relatedto ER-mitochondria crosstalk. Indeed, by inhibiting IP3Rwith XeC,weshowdecreasedmitCa2þ retention incells exposed toAbþNMDA inaprocess involving theMCU. Interestingly, along with the inhibition ofER Ca2þ transfer, XeC also prevented mitochondrial depolarizationoccurring in response to Ab þ NMDA stimulation. Nevertheless, wecannot exclude that XeC can somehow affect NMDARs function. Infact, in cells pre-exposed to XeC and treated with Ab þ NMDA, weobserved a decrease in Ca2þ entry to the cytosol, further preventingmitochondrial depolarization. Also, we cannot exclude the possibleinteractions of Abwith metabotropic receptors leading to IP3 forma-tion and ER Ca2þ release; however, studies linking the activation ofmetabotropic receptorswith ER stress in neuronal cells exposed to Aboligomers are presently missing.

Altogether, these data demonstrate that Ab induces an imme-diate increase in Ca2þi and mitochondrial depolarization, partiallymediated through GluN2B-containing NMDARs, highlighting therole of NMDARs in Ca2þ dyshomeostasis in AD. This study furtherevidences the intricate role of mitochondria in retaining Ca2þ in Ab-treated cells. Moreover, when extracellular accumulation of Aboccurs concomitantly with a direct activation of GluN2B-containingNMDARs by glutamate in the synaptic cleft, mitCa2þ retention isexacerbated involving ER Ca2þ release through the IP3R and theMCU (Fig. 7). These data deepen the understanding of the mecha-nisms underlying Ab-induced neuronal and synapse dysfunctioninvolving NMDARs overactivation and ER-mitochondria crosstalk;this may be exacerbated during aging and further contribute for ADneurodegeneration, emphasizing mitochondria and GluN2B sub-units as therapeutic targets in early stages of AD pathogenesis.

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692 691

Disclosure statement

The authors have no conflict of interest to declare.

Acknowledgements

Theworkwas supported by Fundo Europeu de DesenvolvimentoRegional (FEDER) through “Programa Operacional Factores deCompetitividade (COMPETE),” “Fundação para a Ciência e a Tec-nologia,” project references PEst-C/SAU/LA0001/2013-2014 andPTDC/SAU-NEU/71675/2006, postdoctoral fellowships SFRH/BPD/43536/2008 and SFRH/BPD/86551/2012, Quadro de ReferênciaEstratégico Nacional 2007e2013 project “Desenvolvimento e Oper-acionalização da Investigação de Translação,” Lundbeck Founda-tion, and the Center for Neuroscience and Cell Biology, University ofCoimbra, Coimbra, Portugal.

References

Agostinho, P., Oliveira, C.R., 2003. Involvement of calcineurin in the neurotoxic ef-fects induced by amyloid-beta and prion peptides. Eur. J. Neurosci. 17,1189e1196.

Alberdi, E., Sánchez-Gómez, M.V., Cavaliere, F., Pérez-Samartín, A., Zugaza, J.L.,Trullas, R., Domercq, M., Matute, C., 2010. Amyloid beta oligomers induce Ca2þ

dysregulation and neuronal death through activation of ionotropic glutamatereceptors. Cell Calcium 47, 264e272.

Area-Gomez, E., de Groof, A.J., Boldogh, I., Bird, T.D., Gibson, G.E., Koehler, C.M.,Yu, W.H., Duff, K.E., Yaffe, M.P., Pon, L.A., Schon, E.A., 2009. Presenilins areenriched in endoplasmic reticulum membranes associated with mitochondria.Am. J. Pathol. 175, 1810e1816.

Area-Gomez, E., Del Carmen Lara Castillo, M., Tambini, M.D., Guardia-Laguarta, C.,de Groof, A.J., Madra, M., Ikenouchi, J., Umeda, M., Bird, T.D., Sturley, S.L.,Schon, E.A., 2012. Upregulated function of mitochondria-associated ER mem-branes in Alzheimer disease. EMBO J. 31, 4106e4123.

Arispe, N., Pollard, H.B., Rojas, E., 1993. Giant multilevel cation channels formed byAlzheimer disease amyloid b-protein [AbP-(1e40)] in bilayer membranes. Proc.Natl. Acad. Sci. U.S.A. 90, 10573e10577.

Bi, X., Gall, C.M., Zhou, J., Lynch, G., 2002. Uptake and pathogenic effects of amyloidbeta peptide 1-42 are enhanced by integrin antagonists and blocked by NMDAreceptor antagonists. Neuroscience 112, 827e840.

Caldeira, G.L., Ferreira, I.L., Rego, A.C., 2013. Impaired transcription in Alzheimer’sdisease: key role in mitochondrial dysfunction and oxidative stress.J. Alzheimers Dis. 34, 115e131.

Camandola, S., Mattson, M.P., 2011. Aberrant subcellular neuronal calcium regula-tion in aging and Alzheimer’s disease. Biochim. Biophys. Acta 1813, 965e973.

Celsi, F., Pizzo, P., Brini, M., Leo, S., Fotino, C., Pinton, P., Rizzuto, R., 2009. Mito-chondria, calcium and cell death: a deadly triad in neurodegeneration. Biochim.Biophys. Acta 1787, 335e344.

Costa, R.O., Lacor, P.N., Ferreira, I.L., Resende, R., Auberson, Y.P., Klein, W.L.,Oliveira, C.R., Rego, A.C., Pereira, C.M., 2012. Endoplasmic reticulum stress oc-curs downstream of GluN2B subunit of N-methyl-d-aspartate receptor inmature hippocampal cultures treated with amyloid-b oligomers. Aging Cell 11,823e833.

De Felice, F.G., Velasco, P.T., Lambert, M.P., Viola, K., Fernandez, S.J., Ferreira, S.T.,Klein, W.L., 2007. Abeta oligomers induce neuronal oxidative stress through anN-methyl-D-aspartate receptor-dependent mechanism that is blocked by theAlzheimer drug memantine. J. Biol. Chem. 282, 11590e11601.

Deshpande, A., Kawai, H., Metherate, R., Glabe, C.G., Busciglio, J., 2009. A role forsynaptic zinc in activity-dependent Abeta oligomer formation and accumula-tion at excitatory synapses. J. Neurosci. 29, 4004e4015.

Domingues, A., Almeida, S., da Cruz e Silva, E.F., Oliveira, C.R., Rego, A.C., 2007.Toxicity of beta-amyloid in HEK293 cells expressing NR1/NR2A or NR1/NR2B N-methyl-D-aspartate receptor subunits. Neurochem. Int. 50, 872e880.

Dragicevic, N., Mamcarz, M., Zhu, Y., Buzzeo, R., Tan, J., Arendash, G.W.,Bradshaw, P.C., 2010. Mitochondrial amyloid-beta levels are associated with theextent of mitochondrial dysfunction in different brain regions and the degree ofcognitive impairment in Alzheimer’s transgenic mice. J. Alzheimers Dis. 20(Suppl 2), S535eS550.

Du, H., Guo, L., Fang, F., Chen, D., Sosunov, A.A., McKhann, G.M., Yan, Y., Wang, C.,Zhang, H., Molkentin, J.D., Gunn-Moore, F.J., Vonsattel, J.P., Arancio, O., Chen, J.X.,Yan, S.D., 2008. Cyclophilin D deficiency attenuates mitochondrial and neuronalperturbation and ameliorates learning and memory in Alzheimer’s disease. Nat.Med. 14, 1097e1105.

Du, H., Guo, L., Yan, S.S., 2012. Synaptic mitochondrial pathology in Alzheimer’sdisease. Antioxid. Redox Signal. 16, 1467e1475.

Du, H., Guo, L., Yan, S., Sosunov, A.A., McKhann, G.M., Yan, S.S., 2010. Early deficits insynaptic mitochondria in an Alzheimer’s disease mouse model. Proc. Natl. Acad.Sci. U. S. A. 107, 18670e18675.

Ferreira, I.L., Bajouco, L.M., Mota, S.I., Auberson, Y.P., Oliveira, C.R., Rego, A.C., 2012.Amyloid beta peptide 1-42 disturbs intracellular calcium homeostasis throughactivation of GluN2B-containing N-methyl-D-aspartate receptors in corticalcultures. Cell Calcium 51, 95e106.

Ferreira, I.L., Resende, R., Ferreiro, E., Rego, A.C., Pereira, C.F., 2010. Multiple defectsin energy metabolism in Alzheimer’s disease. Curr. Drug Targets 11, 1193e1206.

Ferreiro, E., Baldeiras, I., Ferreira, I.L., Costa, R.O., Rego, A.C., Pereira, C.F.,Oliveira, C.R., 2012. Mitochondrial- and endoplasmic reticulum-associatedoxidative stress in Alzheimer’s disease: from pathogenesis to biomarkers. Int.J. Cell Biol. 2012, 735206.

Ferreiro, E., Oliveira, C.R., Pereira, C., 2004. Involvement of endoplasmic reticulum Ca2þ

release through ryanodine and inositol 1,4,5-triphosphate receptors in the neuro-toxic effects induced by the amyloid-beta peptide. J. Neurosci. Res. 76, 872e880.

Ferreiro, E., Oliveira, C.R., Pereira, C.M., 2008. The release of calcium from theendoplasmic reticulum induced by amyloid-beta and prion peptides activatesthe mitochondrial apoptotic pathway. Neurobiol. Dis. 30, 331e342.

Ferreiro, E., Resende, R., Costa, R., Oliveira, C.R., Pereira, C.M., 2006. An endoplasmic-reticulum-specific apoptotic pathway is involved in prion and amyloid-betapeptides neurotoxicity. Neurobiol. Dis. 23, 669e678.

Gibson, G.E., Peterson, C., 1987. Calcium and the aging nervous system. Neurobiol.Aging 8, 329e343.

Gouras, G.K., Tsai, J., Naslund, J., Vincent, B., Edgar, M., Checler, F., Greenfield, J.P.,Haroutunian, V., Buxbaum, J.D., Xu, H., Greengard, P., Relkin, N.R., 2000. Intra-neuronal Abeta42 accumulation in human brain. Am. J. Pathol. 156, 15e20.

Guo, L., Du, H., Yan, S., Wu, X., McKhann, G.M., Chen, J.X., Yan, S.S., 2013. CyclophilinD deficiency rescues axonal mitochondrial transport in Alzheimer’s neurons.PLoS One 8, e54914.

Haass, C., Selkoe, D.J., 2007. Soluble protein oligomers in neurodegeneration: les-sons from the Alzheimer’s amyloid beta-peptide. Nat. Rev. Mol. Cell Biol. 8,101e112.

Hansson Petersen, C.A., Alikhani, N., Behbahani, H., Wiehager, B., Pavlov, P.F.,Alafuzoff, I., Leinonen, V., Ito, A., Winblad, B., Glaser, E., Ankarcrona, M., 2008.The amyloid beta-peptide is imported into mitochondria via the TOM importmachinery and localized to mitochondrial cristae. Proc. Natl. Acad. Sci. U. S. A.105, 13145e13150.

Hauptmann, S., Scherping, I., Dröse, S., Brandt, U., Schulz, K.L., Jendrach, M.,Leuner, K., Eckert, A., Müller, W.E., 2009. Mitochondrial dysfunction: an earlyevent in Alzheimer pathology accumulates with age in AD transgenic mice.Neurobiol. Aging 30, 1574e1586.

Kannurpatti, S.S., Joshi, P.G., Joshi, N.B., 2000. Calcium sequestering ability ofmitochondria modulates influx of calcium through glutamate receptor channel.Neurochem. Res. 25, 1527e1536.

Klein, W.L., 2006. Synaptic targeting by Abeta oligomers (ADDLS) as a basis formemory loss in early Alzheimer’s disease. Alzheimers Dement. 2, 43e55.

Kutsuwada, T., Sakimura, K., Manabe, T., Takayama, C., Katakura, N., Kushiya, E.,Natsume, R., Watanabe, M., Inoue, Y., Yagi, T., Aizawa, S., Arakawa, M.,Takahashi, T., Nakamura, Y., Mori, H., Mishina, M., 1996. Impairment of sucklingresponse, trigeminal neuronal pattern formation, and hippocampal LTD inNMDA receptor epsilon 2 subunit mutant mice. Neuron 16, 333e344.

Lazarov, O., Lee, M., Peterson, D.A., Sisodia, S.S., 2002. Evidence that synapticallyreleased beta-amyloid accumulates as extracellular deposits in the hippocam-pus of transgenic mice. J. Neurosci. 22, 9785e9793.

Malinow, R., 2012. New developments on the role of NMDA receptors in Alzheimer’sdisease. Curr. Opin. Neurobiol. 22, 559e563.

Matos, M., Augusto, E., Oliveira, C.R., Agostinho, P., 2008. Amyloid-beta peptide de-creases glutamate uptake in cultured astrocytes: involvement of oxidative stressand mitogen-activated protein kinase cascades. Neuroscience 156, 898e910.

Mattson, M.P., Partin, J., Begley, J.G., 1998. Amyloid beta-peptide induces apoptosis-related events in synapses and dendrites. Brain Res. 807, 167e176.

Moreira, P.I., Santos, M.S., Moreno, A., Oliveira, C., 2001. Amyloid b-peptide promotespermeability transition pore in brain mitochondria. Biosci. Rep. 2, 789e800.

Mota, S.I., Ferreira, I.L., Pereira, C., Oliveira, C.R., Rego, A.C., 2012. Amyloid-betapeptide 1-42 causes microtubule deregulation through N-methyl-D-aspartatereceptors in mature hippocampal cultures. Curr. Alzheimer Res. 9, 844e856.

Mota, S.I., Ferreira, I.L., Rego, A.C., 2014. Dysfunctional synapse in Alzheimer’s dis-ease - a focus on NMDA receptors. Neuropharmacology 76 (Pt A), 16e26.

Nicholls, D.G., 2009. Mitochondrial calcium function and dysfunction in the centralnervous system. Biochim. Biophys. Acta 1787, 1416e1424.

Oliveira, J.M., Chen, S., Almeida, S., Riley, R., Gonçalves, J., Oliveira, C.R., Hayden, M.R.,Nicholls, D.G., Ellerby, L.M., Rego, A.C., 2006. Mitochondrial-dependent Ca2þ

handling in Huntington’s disease striatal cells: effect of histone deacetylaseinhibitors. J. Neurosci. 26, 11174e11186.

Pagani, L., Eckert, A., 2011. Amyloid-Beta interaction with mitochondria. Int. J. Alz-heimers Dis. 2011, 925050.

Papouin, T., Ladépêche, L., Ruel, J., Sacchi, S., Labasque, M., Hanini, M., Groc, L.,Pollegioni, L., Mothet, J.P., Oliet, S.H., 2012. Synaptic and extrasynaptic NMDAreceptors are gated by different endogenous coagonists. Cell 150, 633e646.

Parameshwaran, K., Dhanasekaran, M., Suppiramaniam, V., 2008. Amyloid betapeptides and glutamatergic synaptic dysregulation. Exp. Neurol. 210, 7e13.

Rego, A.C., Ward, M.W., Nicholls, D.G., 2001. Mitochondria control ampa/kainatereceptor-induced cytoplasmic calcium deregulation in rat cerebellar granulecells. J. Neurosci. 21, 1893e1901.

Resende, R., Ferreiro, E., Pereira, C., Oliveira, C.R., 2008. Neurotoxic effect of oligo-meric and fibrillar species of amyloid-beta peptide 1-42: involvement of

I.L. Ferreira et al. / Neurobiology of Aging 36 (2015) 680e692692

endoplasmic reticulum calcium release in oligomer-induced cell death.Neuroscience 155, 725e737.

Sanz-Blasco, S., Valero, R.A., Rodríguez-Crespo, I., Villalobos, C., Núñez, L., 2008.Mitochondrial Ca2þ overload underlies Abeta oligomers neurotoxicityproviding an unexpected mechanism of neuroprotection by NSAIDs. PLoS One3, e2718.

Stanika, R.I., Villanueva, I., Kazanina, G., Andrews, S.B., Pivovarova, N.B., 2012.Comparative impact of voltage-gated calcium channels and NMDA re-ceptors on mitochondria-mediated neuronal injury. J. Neurosci. 32,6642e6650.

Supnet, C., Bezprozvanny, I., 2010. The dysregulation of intracellular calcium inAlzheimer disease. Cell Calcium 47, 183e189.

Texidó, L., Martín-Satué, M., Alberdi, E., Solsona, C., Matute, C., 2011. Amyloid bpeptide oligomers directly activate NMDA receptors. Cell Calcium 49, 184e190.

Tovar, K.R., Sprouffske, K., Westbrook, G.L., 2000. Fast NMDA receptor-mediatedsynaptic currents in neurons from mice lacking the epsilon2 (NR2B) subunit.J. Neurophysiol. 83, 616e620.

Ward, M.W., Rego, A.C., Frenguelli, B.G., Nicholls, D.G., 2000. Mitochondrial mem-brane potential and glutamate excitotoxicity in cultured cerebellar granule cells.J. Neurosci. 20, 7208e7219.