Neuroscience 278 (2014) 354–366

Ab PROMOTES VDAC1 CHANNEL DEPHOSPHORYLATION IN NEURONAL LIPID RAFTS. RELEVANCE TO THE MECHANISMS OF NEUROTOXICITY IN ALZHEIMER’S DISEASE C. FERNANDEZ-ECHEVARRIA, a M. DI´AZ, b I. FERRER, c A. CANERINA-AMARO a AND R. MARIN a*

phosphatase inhibitors. VDAC1 dephosphorylation was corroborated in lipid rafts of AD brains. These results demonstrate that Ab is involved in alterations of the phosphorylation state of VDAC in neuronal lipid rafts. Modulation of this channel may contribute to the development and progression of AD pathology. Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved.

a Laboratory of Cellular Neurobiology, Department of Physiology, University of La Laguna, School of Medicine, Sta. Cruz de Tenerife, Spain b Laboratory of Membrane Physiology and Biophysics, Department of Animal Biology, Faculty of Biology, University of La Laguna, Sta. Cruz de Tenerife, Spain c Institute of Neuropathology, Bellvitge University Hospital, University of Barcelona, IDIBELL, CIBERNED, Hospitalet de Llobregat, Barcelona, Spain

Key words: voltage-dependent anion channel, amyloid precursor protein, amyloid beta, Alzheimer’s disease, lipid rafts, tyrosine phosphatase inhibitor.

Abstract—Voltage-dependent anion channel (VDAC) is a mitochondrial protein abundantly found in neuronal lipid rafts. In these membrane domains, VDAC is associated with a complex of signaling proteins that trigger neuroprotective responses. Loss of lipid raft integrity may result in disruption of multicomplex association and alteration of signaling responses that may ultimately promote VDAC activation. Some data have demonstrated that VDAC at the neuronal membrane may be involved in the mechanisms of amyloid beta (Ab)-induced neurotoxicity, through yet unknown mechanisms. Ab is generated from amyloid precursor protein (APP), and is released to the extracellular space where it may undergo self-aggregation. Ab aggregate deposition in the form of senile plaques may lead to Alzheimer’s disease (AD) neuropathology, although other pathological hallmarks (such as hyper-phosphorylated Tau deposition) also participate in this neurodegenerative process. The present study demonstrates that VDAC1 associates with APP and Ab in lipid rafts of neurons. Interaction of VDAC1 with APP was observed in lipid rafts from the frontal and entorhinal cortex of human brains affected by AD at early stages (I–IV/0-B of Braak and Braak). Furthermore, Ab exposure enhanced the dephosphorylation of VDAC1 that correlated with cell death. Both effects were reverted in the presence of tyrosine

INTRODUCTION Alzheimer’s disease (AD) is a neurodegenerative disease mostly related to aged brain, which is clinically characterized by a progressive cognitive decline, eventually leading to dementia. In this disease, there is abundant extracellular aggregation of amyloid-beta (Ab) peptide that forms senile plaques, and intracellular hyper-phosphorylated tau protein aggregation that forms neurofibrillary tangles (Duyckaerts and Dickson 2003; Lowe et al., 2008). Ab formation consists of the sequential cleavage of full-length amyloid precursor protein (APP) by b-secretase (BACE1) and c-secretase complex. The amyloidogenic processing results in the release of Ab peptides of between 39 and 42 residues in length, depending upon the precise cleavage site. APP is a transmembrane protein abundantly found at the neuronal membrane. There is consensus that part of APP processing takes place in particular membrane microstructures named lipid rafts (Bhattacharyya et al., 2013). Lipid rafts are plasma membrane microdomains with a distinct structure and lipid composition where numerous signaling proteins are integrated to participate in different intracellular processes (Lingwood and Simons, 2010). In particular in the case of Ab processing, APP translocates to lipid rafts following palmitoylation (Bhattacharyya et al., 2013), where its intracellular domain interacts with raft-integrated flotillin-1 to initiate amyloigenic processing (Chen et al., 2006; Schneider et al., 2008). Furthermore, a subset of BACE1 and c-secretase components also partition into lipid rafts, suggesting that Ab formation occurs in these structures (Ehehalt et al., 2003; Kalvodova et al., 2005). Indeed, recent data have demonstrated that early alterations of membrane microenvironment in lipid rafts from AD brains,

*Corresponding author. Address: Laboratory of Cellular Neurobiology, Department of Basic Medical Sciences, University of La Laguna, School of Health Sciences, 38071 Sta. Cruz de Tenerife, Spain. Tel: +34-922-319411; fax: +34-922-319397. E-mail address: [email protected] (R. Marin). Abbreviations: Ab, b-amyloid peptide; AD, Alzheimer’s disease; APP, amyloid precursor protein; BACE1, b-secretase; BSA, bovine serum albumin; CI, cell index; DMEM, Dulbecco’s Modified Eagle’s Medium; DMSO, Dimethyl sulfoxide; EDTA, ethylenediaminetetraacetic acid; ERa, estrogen receptor alpha; FBS, fetal bovine serum; IEF, isoelectrofocusing; IGF-1R, insulin-growth factor-1 receptor; PBS, phosphate-buffered saline; PMSF, phenyl methyl sulfonyl fluoride; pI, isoelectric points; PTP, protein tyrosine phosphatase; pTyr, phosphotyrosine; ROI, regions of interest; SDS–PAGE, sodium dodecyl-sulfate–polyacrylamide gel electrophoresis; Tyr, tyrosine; VDAC, voltage-dependent anion channel. http://dx.doi.org/10.1016/j.neuroscience.2014.07.079 0306-4522/Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved. 354

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

promotes the accumulation of BACE1 and increases its interaction with APP (Diaz et al., 2014; Fabelo et al., 2014). Overall, these data point to a dynamic system of Ab production with the participation of different lipid raft modulators. Another protein highly abundant in neuronal lipid rafts is the voltage-dependent anion channel (VDAC). Although originally characterized as a mitochondrial porin, VDAC present at the plasma membrane appears to play different roles. Thus, this versatile porin has been related to extrinsic apoptotic pathways, cellular ATP release, calcium and metabolite transport, volume control, redox homeostasis and NADH:ferricyanide reductase activity (Akanda et al., 2008; De Pinto et al., 2010; Park et al., 2010). Moreover, in particular in neurons, some studies are starting to reveal the importance of VDAC in the events related to AD pathology. Thus, VDAC1 associates with c-secretase in lipid rafts (Hur et al., 2012), suggesting the involvement of the porin in Ab processing. Furthermore, VDAC1 has been related to Ab-induced toxicity in cultured neurons (Marin et al., 2007). In line with this, immunohistochemical data in, both, APP/PS1 transgenic mouse and human AD brains have shown VDAC intimately related to senile plaques and neurofibrillary tangles (Ferrer, 2009; Ramirez et al., 2009). In AD brains, VDAC expression and modulation appear to be altered, showing an increase in VDAC nitration with the progression of the disease (Sultana et al., 2006). Activation and opening of VDAC at the plasma membrane may imply the post-translational modification of the channel phosphorylation state (Diaz et al., 2001). At the neuronal membrane, VDAC1 is present in at least three different isoforms phosphorylated at different residues that are conserved in, both, mouse and human brain tissue (Ferrer, 2009; Yoo et al., 2001). Interestingly, alterations in VDAC phosphorylation modulate its activation and gating, as observed by the extracellular application of anti-estrogens for a short period of time (65 years) without any apparent neurological, psychological or neuropathological disorders were obtained from the Institute of Neuropathology Brain Bank (Bellvitge University Hospital) following legal and ethical guidelines for Biomedical Research involving human subjects and approval of the local Ethics Committee. The postmortem delay was between 3 and 12 h. The frontal cortex (area 8), entorhinal cortex and cerebellum were dissected free of white matter and used for lipid raft isolation. In some cases, lipid raft fractions from the hippocampus of control subjects were also used to run on two-dimensional gel electrophoresis. Five cases were used for each experimental group comprising AD-related pathology stages I–II/0-A; AD-related pathology stage III–IV/A–B, and age-matched controls. Isolation of lipid rafts and microsomes Lipid raft isolation was performed following protocols described previously (Mukherjee et al., 2003) with minor modifications. The frontal cortex, entorhinal cortex, hippocampus and cerebellum were homogenized in buffer A (50 mM Tris–HCl pH 8.0, 10 mM MgCl2, 20 mM NaF, 1 mM Na3VO4, 5 mM b-mercaptoethanol, 1 mM phenyl methyl sulfonyl fluoride PMSF) in a glass homogenizer grinder, and processed for sucrose gradients differential centrifugation (Ramirez et al., 2009). Six fractions of 2 ml were obtained. Characterization of lipid raft fractions was performed by immunoblotting with lipid raft and nonraft protein markers: flotillin-1 and caveolin-1 for lipid raft fractions, and a1-subunit Na+/K+ ATPase and cytosolic Hsp90 for non-raft fractions (data not shown).

356

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

For microsomal extraction, cells were homogenized in RSB buffer (10 mM Tris–HCl, pH 8.0, 20 mM NaCl, 25 mM EDTA, complete protease inhibitor cocktail (Roche Diagnosts, Barcelona, Spain), and phosphatase inhibitor cocktail (Cultek, Barcelona, Spain) at 4 °C. Homogenates were exposed to serial centrifugation: first at 900g for 15 min to remove nuclei, second at 10,000g for 15 min to separate mitochondria, followed by a third centrifugation at 100,000g to sediment microsomal fractions. The precipitated pellet contained the microsomal fraction. Microsomes were diluted in loading buffer (625 mM Tris–HCl, 1% sodium dodecyl sulfate, 10% glycerol, 5% b-mercaptoethanol and 0.001% bromophenol blue, pH 6.8) for 12.5% SDS–PAGE electrophoresis.

gels (17% acrylamide–bisacrilamide (37:1), 400 mM Tricine buffer pH8.45, 0.2% SDS and 30% glycerol), as described previously (Scha¨gger, 2006). Gels were fixed in 50% ethanol and 30% acetic acid, and stained with Sypro Ruby (Sigma) according to the manufacturer’s instructions. Gels were analyzed in VersaDoc MP4000 (Bio-Rad, Madrid, Spain). Immunoprecipitation assays

Septal (SN56) and hippocampal (HT22) murine cells were routinely cultured in standard medium (DMEM supplemented with 10% fetal bovine serum (FBS), 3.7 mg/ml NaHCO3, 60 U/ml penicillin-G and 100 lg/ml streptomycin sulfate), and maintained at 37 °C under a 95% air/5% CO2 atmosphere. The human neuroblastoma SH-SY5Y cell line was maintained in DMEM/Ham’s F12. Culture medium was changed every two days and, at 90% confluency, cells were trypsinized with 0.25% trypsin-EDTA mixture, at a density of 4  105/ml on either 16-well plates (for cell proliferation studies) or T75 flasks (for protein expression studies).

To study the potential interactions between VDAC, Ab and APP, we performed immunoprecipitation assays using specific anti-VDAC1 monoclonal antibody (Abcam pl). Protein lysates from cortical areas of, both, AD patients and control subjects, as well as SN56, HT22 and SHSY-5Y neuronal cultures were resuspended in cold immunoprecipitation buffer [50 mM Tris–HCl pH 7.4; 150 mM NaCl; 10% glycerol; 1% Nonidet-P 40; 1 mM PMSF; complete proteases inhibitor cocktail]. Samples were processed using dynabeads manufacturer’s instructions form immunoprecipitation. Briefly, 5–10 lg of anti-VDAC1 antibody and anti-APP antibody per 100 lg of protein amount were incubated together with the protein extracts overnight at 4 °C with gentle agitation. Then, 30 ll of dynabeads were added for 2 h at room temperature to pull down antibodyprotein complexes. VDAC1 or APP immunoprecipitated elutes were loaded onto precast Criterion for SDS– PAGE analyses of co-precipitated proteins. No unspecific protein binding was observed in magnetic beads used for pre-adsorption (not shown).

Amyloid beta preparation and treatment

SDS–PAGE and two-dimensional gel electrophoresis

Ab1–42 was prepared as 1 mM stock solution in cold hexafluoroisopropanol (HFIP), dried and dissolved in Dimethyl sulfoxide (DMSO) at 5 mM. Then, the peptide was diluted 50-fold in Dulbecco’s Modified Eagle’s Medium (DMEM), and aggregation was induced by incubation at 4 °C for 24 h, as previously reported (Klein, 2002). Separation of insoluble aggregates and soluble toxic oligomers was performed by centrifugation at 14,000g for 10 min at 4 °C. Amyloid beta 25–35 aminoacid fragment (Ab25–35) was resuspended in 2.5 mM stock solution in deionized distilled water, and immediately stored at 80 °C until use. Different experimental concentrations were prepared in cultured medium (DMEM with 10% FBS), and added to cell plates. The formation of toxic globular oligomers was monitored by quantification of Ab binding to thioflavin T (T-3516, Sigma–Aldrich), a classical marker for deposits of b-sheet amyloid (Schlenzig et al., 2009; To˜ugu et al., 2009) that fluoresces at 536 nm. Incubation of the peptide with thioflavin T resulted in a kinetic progression of Ab oligomerization in a dose-dependent manner (data not shown). This method has been previously characterized by different approaches (Levine, 1993; reviewed in Bensey-Cases et al., 2012). Furthermore, distinct morphological formations were also characterized in Tris-Tricine gel electrophoresis. Eight micrograms of Ab1–42 fibril formations, insoluble aggregates and oligomers was run on 17.5% Tricine-SDS–PAGE preparation

Lipid raft fractions containing approximately 30 lg of protein were resuspended in DeStreak Rehydration Solution (GE Healthcare) in a total volume of 125 lL. Samples were applied to immobilized pH gradient 7-cmlength strips (pH 7–11 or pH 3–10), and processed for isoelectrofocusing (IEF) up to 20,000 V/h. Then IEF strips were reduced in equilibration buffer containing 2% (w/v) dithiothreitol (DTT) for 10 min at room temperature, followed by alkylation in equilibration buffer containing 2.5% (w/v) iodoacetamide. Strips were loaded onto 12.5% SDS–PAGE for second dimension resolution. In other set of experiments, protein extracts were resuspended in loading buffer (625 mM Tris–HCl, 1% sodium dodecyl sulfate, 10% glycerol, 5% b-mercaptoethanol and 0.001% bromophenol blue, pH 6.8) for 12.5% SDS–PAGE electrophoresis.

Cell culture conditions and treatments

Immunoblotting Proteins resolved by SDS–PAGE were transferred to polyvinyl (PVDF) membranes using the Trans-Blot Turbo rapid western blotting transfer system (Bio-Rad, Madrid, Spain). PVDF membranes were incubated overnight at 4 °C with different primary antibodies (antiVDAC1, anti-Ab, raft marker flotillin-1 and membrane marker a1 Na+/K+ ATPase monoclonal antibodies, and anti-APP and anti-pTyr residues polyclonal antibodies). All antibodies were diluted 1:1000 in BLOTTO, except

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

antibodies purchased from Santa Cruz biotech that were diluted 1:200 according to the manufacturer’s indications. The anti-pTyr antibody was diluted in 5% bovine serum albumin (BSA). Membranes were washed three times for 5 min in Tris-buffered saline (TBS) with 0.5% Tween-20. Proteins were detected using the corresponding peroxidase-conjugated anti-mouse or anti-rabbit secondary antibody diluted 1:5000 for 1 h at room temperature. Specific bands were detected with the ECL Plus chemiluminescent kit (GE Healthcare, Madrid, Spain). Chemiluminescent signals were scanned and quantified by densitometry, using GS-800 calibrated densitometer (Bio-Rad, Madrid, Spain). Double-labeling immunofluorescence analysis To immunolocalize the different protein markers at the neuronal membrane, cultured neurons grown in 8-well cultured slides (Falcon, Madrid, Spain) were fixed under non-permeabilized conditions, using 2% qformaldehyde, 0.5% glutaraldehyde and 120 mM sucrose in phosphate-buffered saline (PBS, pH 7.4). Experiments were performed in the absence of detergent to avoid intracellular antibody leaking. For lipid raft visualization, cells were chilled on ice for 10 min, and then washed twice with ice-cold PBS. Then, cells were incubated with 2-lg/ml cholera-toxin B conjugated to Alexa Fluor 555 (Molecular Probes, Madrid, Spain) in 0.1% BSA–PBS for 20 minutes at 4 °C, followed by fixation under non-permeabilized conditions. For VDAC and APP double immunofluorescence staining, after fixation, cells were incubated overnight at 4 °C with antiVDAC monoclonal antibody (diluted 1:200 in PBS) and anti-APP polyclonal antibody (diluted 1:500 in PBS) in the presence of 1:200 normal serum. The day after, cells were washed with PBS, and incubated with the secondary biotinylated anti-rabbit antibody and with cyanin-3-coupled anti-mouse antibody, both diluted 1:200 in PBS for 2 h at room temperature. Slides were washed three times in PBS and exposed to cyanine-2conjugated streptavidin (diluted 1:500 in PBS) for 30 min at room temperature. After washing in PBS, slides were cover-slipped and processed for fluorescent detection in a laser scanning confocal imaging Fluoview 1000 system (Olympus, Barcelona, Spain). The confocal settings for image acquisition were the following: 60 oil immersion objective, laser Argon wavelength at 488 nm for cyanine-2 dye, laser Argon wavelength at 514 for cyanine-3 dye, and laser Helium–Neon at 543 nm for Alexa 555 dye. For colocalization quantification, we used the Bio-Rad lasersharp software. Microscope settings were constant throughout the assays, and local background light was subtracted. Image acquisition parameters were identical for the different channels. Points resolution to avoid over-sampling was performed according to Nyquist sampling theorem in terms of acquisition parameters according to pixel size and optical resolution (Oppenheim et al., 1983). To avoid immunosignal cross-talk, acquisition was obtained by sequential methods. Regions of interest (ROI) were selected, and co-localization percentage quantified by Pearson’s coefficient (Bolte and

357

Cordelie`res, 2006). At least a hundred ROIs were selected for each quantification. Real-time cell proliferation assays The xCELLigence impedance-based technology (Roche Diagnostics Corporation) was used to ascertain the cell death index of, both, SN56 and HT22 cells in the presence of Ab (Ke et al., 2011). Real-time cell proliferation studies were performed using xCELLigence electrical impedance-based system that allows the measurement of real-time cell proliferation (Ke et al., 2011). Cells were cultivated for 24 h before subculture on E 16-well xCELLigence plates at a density of 12–20  103 cells/ well. After 24 h, cells were treated with Bipy, a potent inhibitor of tyrosine (Tyr) phosphatase at 200 nM for 30 min, followed by addition of Ab at two different doses (5 and 20 lM) for 72 h. The impedance was recorded at 15-min intervals throughout the experiment, and the experimental values converted to cell index (CI), which is a measure of the degree of cellular adhesion to the multi-electrode array. The greater the CI values the greater the total adhesion of cells present in the wells (higher growth and survival). Generally, cell number directly correlates with output CI reading until confluency is achieved (CImax). Determination of cell growing rate in the different experiments was calculated from the slopes in the linear part of curves before reaching CImax. Statistical analyses Data were analyzed by a one-way analysis of variance (ANOVA) followed by Student–Newman–Keuls t-test were appropriate. Numerical data were represented as mean ± SEM. Statistical significance in figures is indicated from P < 0.05.

RESULTS VDAC interacts with Ab and APP in neuronal lipid rafts Previous evidence has demonstrated that VDAC1 is highly abundant in lipid rafts, where it may be affected by the mechanisms related to Ab-induced toxicity (Marin et al., 2007; Ramirez et al., 2009). To determine whether APP and/or Ab may interact with VDAC, we performed immunoprecipitation experiments with anti-VDAC1 monoclonal antibody, anti-APP polyclonal antibody, and anti-Ab monoclonal antibody, using lipid raft fractions purified from SN56, HT22 and SHSY-5Y neuronal lines. Immunoprecipitation of VDAC1 in these fractions resulted in the co-precipitation of APP and Ab, revealing by immunoblotting 6 kDa Ab and 100 kDa full-length APP (Fig. 1, IP VDAC). Moreover, immunoprecipitation assays using anti-Ab antibody produced the coprecipitation of VDAC1 and APP, thus confirming the association between these proteins (Fig. 1, IP Ab). This interaction was predominant in lipid raft fractions at least for SN56 and HT22 cells, since VDAC was not significantly represented in non-raft membrane fractions, as previously demonstrated (Marin et al., 2007).

358

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

Examination of double-immunofluorescence overlapping in murine SN56 and HT22 revealed a high degree of colocalization (Pearson’s coefficient 0.68 ± 0.02 and 0.77 ± 0.03, respectively) at the cell surface, most probably in discrete raft microstructures (Fig. 2B). Similar results were observed for SHSY-5Y cells (data not shown). Overall, these data are in agreement with the immunoprecipitation assays of VDAC1 and fulllength APP. VDAC1 increases its interaction with APP in lipid rafts of human brains at early stages of AD-related pathology

Fig. 1. VDAC1 interacts with full-length APP and Ab in lipid rafts from murine septal SN56 and hippocampal HT22 cells, and human neuroblastoma SHSY-5Y cells. IP VDAC indicates immunoprecipitation with VDAC1 antibody (monoclonal), and co-precipitation of APP and Ab by immunoblotting with corresponding anti-APP and anti-Ab specific antibodies (IB APP and IB Ab). IP Ab indicates immunoprecipitation with Ab antibody in the same lipid raft protein extracts, and co-precipitation by immunoblotting with anti-VDAC and anti-APP specific antibodies (IB VDAC and IB APP). Total lipid raft (LR) and non-raft (NR) protein extracts were also immunoblotted with the three antibodies as a control (INPUT). Five assays per group.

Interaction of APP and VDAC1 in lipid rafts was corroborated by immunocytochemistry under nonpermeabilizing conditions, in order to visualize VDAC– APP interaction at the plasma membrane. First, the presence of VDAC in lipid rafts was observed by double fluorescence staining using cholera-toxin B (ChTx) conjugated to Alexa Fluor 555 to visualize lipid rafts, followed by immunocytochemistry using anti-VDAC antibody. The results shown in Fig. 2A illustrate a high double staining of, both, ChTx and VDAC in SN56, HT22 and SHSY-5Y cultures associated with raft microdomains. Colocalization calculated by Pearson’s coefficients were 0.65 ± 0.04, 0.73 ± 0.05 and 0.71 ± 0.02 for SN56, HT22s and SHSY-5Y cells, respectively. These results corroborate that most VDAC localizes in lipid rafts. Furthermore, to confirm VDAC–APP interaction at the plasma membrane, cultures were co-incubated with monoclonal anti-VDAC and polyclonal anti-APP specific antibodies, followed by incubation with corresponding secondary antibodies (see Experimental procedures).

We next conducted another set of experiments in raft extracts obtained from the frontal and entorhinal cortices of human brains, two areas observed to be differentially affected during AD progression. The entorhinal cortex is first involved at stages I–II whereas the frontal cortex is not significantly affected by neurofibrillary tangle pathology until stage V, which has not been included in the present series. Regarding amyloid b deposition, senile plaques are found at stage A mainly in the temporal lobes but the frontal cortex is not involved until stage B. Cerebellar microdomains of the same subjects were used for comparative purposes, as the cerebellum is not affected by neurofibrillary tangles, and only a few b-amyloid plaques can be seen at terminal stages of the disease (Braak and Braak, 1999; Braak et al., 2006). In lipid raft fractions from these areas, we observed that VDAC amount was increased in ADI-/II/0-A and ADIII– IV/A-B cortical areas, as compared with healthy subjects (Fig. 3A). On the contrary, VDAC immunoblotting in cerebellar raft samples did not show any differences between the experimental groups (Fig. 3B). Reincubations with anti-APP polyclonal antibody and anti-flotillin 1 monoclonal antibody were used as controls. Bar diagram representation of VDAC immunosignals normalized to APP showed that, in particular in ADI–II/0-A, VDAC1 increased its level in about 40% as compared to controls in both cortical areas (Fig. 3C). These results demonstrate that VDAC is highly abundant in lipid rafts of cortical areas of AD brains at the first stages of the disease. In order to analyze whether VDAC1 interacts with APP and Ab in lipid rafts of human brains, we performed immunoprecipitation assays using VDAC-specific monoclonal antibody in lipid raft fractions of human frontal and entorhinal cortices. The resultant precipitated material was then immunoblotted with anti-APP polyclonal antibody and with anti-Ab monoclonal antibody, showing increased association of Ab and fulllength APP in ADI–II/0-A and ADIII–IV/A–B extracts (Fig. 3D). No significant APP or Ab co-precipitation was detected in control samples. Interestingly, VDAC was highly associated with oligomeric Ab, suggesting that this protein may be involved in Ab aggregates generation. Similar immunoprecipitation experiments in raft cerebellar samples failed to co-precipitate neither APP nor Ab using anti-VDAC antibody for immunoprecipitation, indicating that the immunocomplex is not present in this brain area (Fig. 3E). These results demonstrate that VDAC interaction with APP and Ab in

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

359

Fig. 2. Presence of VDAC in neuronal lipid rafts and co-localization of VDAC with APP at the neuronal membrane in SN56, HT22 and SHSY-5Y cells detected by immunocytochemistry. (A) Lipid rafts were visualized by incubation with cholera-toxin B conjugated to Alexa Fluor 555 (2 lg/ml) followed by immunostaining with anti-VDAC antibody under non-permeabilized conditions. To visualize cell shape, transmission images are shown. To better clarify co-localization, panels on the right show double immunofluorescence spots only (merge). (B) Co-localization of VDAC and APP by incubation with anti-VDAC and anti-APP specific antibodies in non-permeabilized SN56, SHSY-5Y and HT22 cells. (C and D) Colocalization analysis: fluorescence cytograms (2D histograms of pixel fluorescence intensities), showing a cluster of signals at xy diagonal line, indicating a high degree of colocalization (green vs red channel). Four assays per group. Bar = 50 lM. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

lipid rafts in the frontal and entorhinal cortices takes place since the first stages of AD-related pathology. VDAC1 is dephosphorylated at the plasma membrane following Ab exposure VDAC at the neuronal membrane is mainly phosphorylated at Tyr residues, as previously demonstrated in SN56 and HT22 cells (Herrera et al., 2011a). To determine whether the presence of Ab may affect VDAC phosphorylation at the plasma membrane, cultured SN56, HT22 and SHSY5Y neurons were exposed to 20 lM Ab (1–42 fragment for SN56 and SHSY-5Y cells, and 25–35 fragment for HT22 cells). First, the different morphological populations of Ab species were resolved by Tris–Tricine electrophoresis, to determine the proportion of fibril formation, oligomers and insoluble aggregates (Fig. 4A). Fractions enriched in toxic oligomeric Ab were used for cell treatment, since these fractions have been reported to induce neurotoxicity (Lambert et al., 1998). Lipid raft samples extracted from cells exposed to Ab oligomers were run on SDS–PAGE, and immunoblotted with a pTyr polyclonal antibody, followed by reblotting with anti-VDAC1 monoclonal antibody. As a loading control, an antibody against membrane marker a1

subunit of Na+/K+ ATPase was also used to reblot with the same membranes. A significant reduction in Tyrphosphorylated VDAC1 was observed after a short exposure (5 min) to Ab25–35 and Ab1–42 (Fig. 4B). These results indicate a rapid dephosphorylation of this porin at the plasma membrane in the presence of Ab. The post-translational modification of VDAC1 following Ab treatment was further examined by twodimensional gel electrophoresis, using lipid raft extracts of neuronal cultures that were immunoblotted with antiVDAC-1 antibody (Fig. 4C). As previously demonstrated in SN56 and HT22 cells, VDAC1 IEF is resolved in three isoforms with isoelectric points (pI) of approximately 8.4, 9.5 and 10.0 in SN56 and HT22 cells. Spot migrating at 9.5 pI corresponded to VDAC phosphorylated at least in a Tyr residue (Herrera et al., 2011a), and spot migrating at 10.0 pI corresponded to unmodified VDAC. Ab treatment (1–42 fragment for SN56 and SHSY-5Y cells, and 25–35 fragment for HT22 cells) at 20 lM reduced the intensity of this spot, indicating VDAC1 dephosphorylation subsequent to the treatment (Fig. 4C). Moreover, the presence of a specific tyrosine phosphatase inhibitor (Bipy, 200 nM for 30 min) prior to Ab exposure prevented this modification, showing a significant increase in the intensity of 9.5 pI spot

360

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

Fig. 3. Increased levels of VDAC were observed in the lipid rafts from the frontal cortex (area 8) and entorhinal cortex of the post-mortem brain from subjects with different stages of AD progression (ADI–II/0-A and ADIII–IV/A–B). (A) Immunoblotting analysis of VDAC in lipid rafts from frontal and entorhinal cortices (FC and ERC, respectively). Age-matched samples with no neurological and neuropathological lesions were used as controls (C). The same membranes were re-blotted with anti-APP antibody and anti-Flotillin 1 antibody, as an integral raft protein. (B) Similar immunoblotting assays using lipid raft samples from cerebellum (CB) as non-pathological controls. (C) Densitometric values of VDAC band staining relative to APP bands. (D) Lipid rafts from frontal cortex (FC) and entorhinal cortex (ERC) from different AD stages (ADI–II/0-A and ADIII–IV/A–B) were immunoprecipitated with anti-VDAC specific antibody (IP VDAC). Raft protein fractions from healthy subjects were used as controls (C). Coprecipitation was detected by immunoblotting with anti-APP antibody (IB:VDAC) and anti-Ab antibody (IB:Ab). This antibody also recognized oligomeric Ab aggregates co-precipitated with VDAC. (E) Similar VDAC immunoprecipitation assays were performed in cerebellar samples of the same subjects (IP VDAC), demonstrating that most of APP and Ab staining was detected in the supernatant of VDAC immunoprecipitates (Not IP). ⁄ p < 0.05 compared to controls. Five assays per brain area.

(Fig. 4C). These results indicate that VDAC1 at the neuronal membrane is dephosphorylated by activation of dephosphatases as a result of Ab treatment, and this effect is inhibited by tyrosine phosphatase inhibition. Inhibition of VDAC1 dephosphorylation at the plasma membrane reduces Ab toxicity Phosphorylation of VDAC at the neuronal membrane maintains the channel inactivated whereas its dephosphorylation and gating participates in cell death by Ab (Diaz et al., 2001; Marin et al., 2007; Herrera et al., 2011b). To explore whether VDAC dephosphorylation in the presence of Ab may contribute to enhance neurotoxicity, neuronal cultures treated with Tyr phosphatase inhibitor Bipy prior to Ab were quantified for real-time cell proliferation and viability using xCELLigence impedance-based technique. We first determined that 5- and 20-lm concentrations of Ab (1–42 fragment for SN56

and 25–35 fragment for HT22) for 5 min provoked a reduction in CI, indicating a progressive decrease in the number of adhered cells correlating with cell number decrease (Fig. 5A). HT22 cells are particularly sensitive to Ab25–35 fragment at low micromolar concentrations (Dargusch and Schubert, 2002; Marin et al., 2007), whereas SN56 cells are preferentially sensitive to Ab1–42 fragment (Marin et al., 2003). This is in agreement with previous reports demonstrating that cell lines may show different susceptibilities to amyloid aggregates (Cecchi et al., 2005). Quantification of differences in cell impedance monitored by xCELLigence technique reflected cell toxicity induced by Ab in a dose-dependent manner, showing 10–20% CI reduction at 5 lM and 30–40% at 20 lM of Ab (Fig 5B). Maximal CI was obtained for the control cells (cells exposed to 0.01% DMSO as vehicle). Moreover, cultures treated with Bipy (200 nM for 30 min) significantly reduced CI decay when exposed to the lower dose of Ab (5 lm, Fig 5C). Maximal CI calculated as

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

361

Fig. 4. The pattern of VDAC phosphorylation is modified in the presence of Ab. (A) Example of a Tris-Tricine gel stained with Sypro Ruby after electrophoresis of Ab1–42 preparations that were processed as indicated in Experimental Procedures. The different morphological populations, i.e. fibrils, oligomers and insoluble aggregates, are indicated. (B) Microsomal extracts from SN56, HT22 and SHSY-5Y cultures treated with different Ab fragments (Ab1–42 for SN56 and SHSY-5Y cells, and Ab25–35 for HT22 cells) for 5 min were immunoblotted with an antibody that specifically recognized tyrosine phosphate groups (pTyr). Equal amounts of the same samples were loaded in parallel, and immunoblotted with anti-VDAC antibody (VDAC). Membranes were re-blotted with an antibody that specifically recognized a1 subunit of Na+/K+ ATPase (ATPase), as a loading control. Bar diagrams represent normalized densitometric values of pTyr amount relative to total VDAC [pTyr/VDAC (au)]. (C) Lipid raft samples from cultured SN56 and HT22 neurons treated with Ab (20 lM) or with Bipy tyrosine phosphatase protein inhibitor (200 nM, for 30 min) followed by treatment with 20 lM Ab were electrophoresed in two-dimensional gels using pH 7–11, nonlinear gradient strips. Relative isoelectric points of the different VDAC spots are indicated. Second dimensional SDS–PAGE was immunoblotted with anti-VDAC antibody. Bar diagrams represent variations in densitometric values of 9.5 spot (corresponding to VDAC phosphoisoform) relative to 10.0 pI spot. ⁄p < 0.05. Four assays per brain area.

percentage of cell toxicity reflected a partial inhibition (approximately 30% for SN56 and 40% for HT22) of 5 lM Ab-induced toxicity in the presence of Bipy. Tyrosine phosphatase inhibitor alone did not cause any cell toxicity (data not shown). These data reveal a strong correlation between VDAC dephosphorylation and Ab-induced neurotoxicity at the neuronal membrane, and suggest that a tyrosine phosphatase activation affecting VDAC protein participates in the mechanism of neurotoxicity. The pattern of VDAC1 phosphoisoforms is altered in lipid rafts of AD brains The potential alterations of VDAC1 phosphorylation were also investigated in plasma membrane microdomains of cortical areas of human brains. Lipid raft fractions of different brain areas of subjects with ADI/II and subjects of similar age without apparent lesions were used to double run on two-dimensional electrophoresis using pH 7–11 non-linear gradient strips, and immunoblotted with anti-VDAC1 monoclonal antibody or, alternatively, with anti-pTyr polyclonal antibody. We observed by twodimensional gels of lipid rafts isolated from the frontal cortex and hippocampus that VDAC showed three isoforms with similar pI than those observed in cultured neurons and the mouse frontal cortex (Herrera et al., 2011b). In addition, there was a fourth isoform focused at approximately 7.9 pI (Fig 6A). This pattern was similar in lipid raft extracts from entorhinal cortex (data not shown), indicating a highly conserved post-translational signature of VDAC1 at the neuronal membrane. Indeed, no differences in the pattern of VDAC1 isoforms were observed between men and women (data not shown). Two-dimensional gels performed in lipid raft extracts from the frontal cortex were processed for immunoblotting with

anti-pTyr antibody in order to determine the presence of pTyr residues migrating at VDAC level. The results demonstrated the presence of two spots reacting with antipTyr antibody at 10.0 and 8.4 pI overlapping with VDAC1 isoforms, suggesting that these spots may correspond to Tyr-phosphorylated forms of VDAC. No phospho-serine residues were detected at this level (data not shown). In raft fractions from AD cortex, there was a reduction in the number of VDAC1 isoforms, and anti-pTyr antibody showed a faint staining as compared with samples without lesions. As a control, the same samples were run on similar two-dimensional gels using pH 3–10 nonlinear gradient strips, and immunoblotted with anti-flotillin 1 antibody as a scaffolding protein of lipid rafts. We detected a main large spot focused at approximately 7.0 pI that corresponds to flotillin-1 (Fig. 6C). These data suggest that the level of VDAC dephosphorylation in AD brain lipid rafts correlates progressively with the severity of the disease progression most probably due to enhanced interaction with Ab.

DISCUSSION Our results confirm the importance of lipid rafts and their components in relation to Ab-induced mechanisms of neuropathogenesis. It is presently accepted that lipid rafts play a main role in the formation and aggregation of amyloid peptides (Cordy et al., 2006; Rushworth and Hooper, 2011) and, therefore, are considered a crucial domain where many events related to AD progression take place (Schengrund, 2010). Thus, the main molecular actors of amyloidogenic processing, APP, BACE and presenilin are targeted to these microdomains by palmitoylation (Cheng et al., 2009; Vetrivel et al., 2009; Bhattacharyya et al., 2013), whose trafficking and

362

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

Fig. 5. Dephosphorylation of VDAC caused by incubation of cells with Ab contributes to neurotoxicity. (A) Representative real-time courses of cell index profiles (as monitored with xCELLigence technique) obtained with the different cell lines exposed to Ab at different doses (5 and 20 lM), and with or without Bipy inhibitor (200 nM for 30 min) prior to Ab exposure (Bipy + Ab). Cells were cultivated on 16-well plates and quantified for realtime cell proliferation by xCELLigence impedance-based technology, which quantifies electrical impedance of cells interacting with an electrode located onto the well. (B) Bar diagrams summarizing the results of normalized cell index changes in response to exposure to Ab1–42 or Ab25–35 (both at 5 and 20 lM for 5 min) in three different experiments. Values were compared with control cultures exposed to vehicle (C, 0.05% DMSO). ⁄ p < 0.05 compared to vehicle. Three assays per group. (C) Bar diagrams of cell toxicity obtained in cultures exposed to the presence or absence of Bipy ( Bipy, +Bipy, respectively) before exposure to Ab for 5 min. Control cultures were exposed to vehicle (C, 0.05% DMSO). ⁄p < 0.05 compared to the absence of Bipy. Three assays per group.

corralling into these microstructures enhance production of Ab (Das et al., 2013; Fabelo et al., 2014). On the contrary, non-amyloidogenic processing by a-secretase takes place mainly in non-raft fractions (Ehehalt et al., 2003). Besides canonical proteins associated with amyloid generation, this study shows that VDAC1 is involved in the mechanisms related to Ab toxicity. This porin is highly abundant in neuronal lipid rafts from different origins, and is associated with caveolin-1, a representative scaffolding protein of lipid rafts (Marin et al., 2007; Ramirez et al., 2009). Here, we show for the first time that VDAC1 associates with full-length APP and Ab (monomers and oligomers) in neuronal lipid rafts. The presence of Ab induces changes in the pattern

of VDAC1 phosphorylation, and this modification is very likely involved in the mechanisms of amyloid-induced neurotoxicity. These results have been obtained in cultured neurons and different human brain areas, supporting the fact that it may be a general phenomenon in different systems. Dephosphorylation in tyrosine residues of VDAC1 in the presence of Ab in cultured neurons was correlated with similar alteration of VDAC1 post-translational pattern in lipid rafts extracted from cortical areas of AD brains at ADI–II/0-A and ADIII–IV/A–B stages of the disease. These data indicate that VDAC activity as a result of membrane impairment may play a critical role at early stages of AD. Also, dephosphorylation of the porin in lipid rafts might be a parameter associated with the progression of neuronal impairment.

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

363

Fig. 6. Post-translational pattern of VDAC is modified in lipid rafts from different areas of AD brains. (A) Lipid rafts (LR) from the healthy frontal cortex and hippocampus were electrophoresed in two-dimensional gels using pH 7–11 nonlinear gradient strips, followed by second-dimensional 12.5% SDS–PAGE. Proteins transferred to membranes were immunoblotted with a specific antibody against VDAC. Numbers on top represent relative pI values. (B) Lipid rafts from the frontal cortex of late stages of AD (AD) were concomitantly run in double identical two-dimensional gels (using pH 7–11 nonlinear gradient strips) followed by 12.5% SDS–PAGE. Identical protein amounts of lipid rafts from the healthy frontal cortex were used as controls (C). Proteins transferred to membranes were immunoblotted with either anti-VDAC antibody (VDAC) of anti-tyrosine phosphate antibody (pTyr). Another set of two-dimensional gels with the same raft samples were immunoblotted with anti-flotillin 1 antibody as a raft control. Five assays per group.

It is worth to mention that mitochondrial membranes lack lipid rafts, and that these microdomains do not contain significant mitochondrial VDAC amount (Zheng et al., 2009). However, the plasmalemmal VDAC counterpart is expressed in the plasma membrane microdomains, and has been previously found associated with c-secretase (Hur et al., 2012). Further, the interaction with APP in lipid rafts found here, strongly support the involvement of plasmalemmal VDAC1 (pl-VDAC) in Ab formation. Moreover, VDAC1 gene silencing appears to affect Ab production although, on the contrary, overexpression of VDAC1 does not provoke enhancement of Ab levels (Hur et al., 2012). In line with our data, these findings reflect the requirement of VDAC to be located in raft microdomains in order to participate in APP processing. Interrelation between VDAC and Ab has also been found in APP transgenic mice that showed increased levels of VDAC1 in the hippocampus, as well as in the frontal cortex of AD patients (Cuadrado-Tejedor et al., 2011; Aso et al., 2012). Altogether, these data suggest that VDAC interacting with APP processing machinery may participate in the pathophysiological derangements taking place in AD as a consequence of membrane perturbations, ultimately enhancing cell exhaustion and extrinsic apoptosis (Ferrer, 2009; Thinnes 2011). In line with this, double immunohistochemistry in APP mouse and AD brain tissues has demonstrated that VDAC1 is highly abundant surrounding senile plaques and neurofibrillary tangles (Ferrer 2009; Ramirez et al., 2009; Aso et al., 2012). Also, in the human frontal cortex of AD patients, mitochondrial VDAC1 co-localizes with full-length APP and oligomeric Ab, thus confirming the relevance of this porin in AD progression (Manczak and Reddy, 2012). The presence of Ab may activate signal transduction related to modification of VDAC1 post-translational and phosphorylation pattern. Indeed, previous work has evidenced that regulation of VDAC1 phosphorylation at the plasma membrane correlated with Ab mechanisms

of toxicity. VDAC1 shows different consensus sites for tyrosine phosphorylation, one of which is an endogenous phospho-Tyr group most probably located in residue 62, at the caveolar binding domain of this channel, which is conserved in different species including human (Marin et al., 2008). Binding to this domain is compatible with a GxxxG motif located in positions 20–24 at N-terminal of VDAC1 sequence, which is an aggregation/membrane perturbation motif also found in Ab peptides at C-terminal. This sequence may be a consensus site of contact between VDAC and Ab peptides (Thinnes 2012, 2013). Preservation of phosphorylation maintains the channel inactivated under normal physiological conditions, whereas its activation by the loss of phosphate group(s) has been related to initiation of toxic signal transduction, including brain ischemia and hypoxia (Sabirov and Merzlyak, 2012). VDAC activity at the neuronal membrane enhanced by its dephosphorylation may be one of the main events caused by Ab that leads to neurotoxicity. In agreement with this, Bipy, a protein tyrosine phosphatase (PTP) inhibitor, modifies the pattern of VDAC phosphoisoforms resolved by twodimensional electrophoresis (Herrera et al., 2011a). This correlates with the inhibition of Ab-induced neurotoxicity shown here. Similar effects have also been observed in other cell types from different origins, where VDAC activation was promoted by PTP-mediated dephosphorylation (Toychiev et al., 2009). Moreover, physical blocking of the channel using specific antibodies to avoid VDAC opening has been shown to prevent cell death caused by Ab (Marin et al., 2007). Then, it is plausible that PTP regulation may be an essential step in toxic intracellular signaling events governed by Ab at the plasma membrane. In this sense, emerging data are starting to identify potential tyrosine phosphatases susceptible of Ab modulation, such as the striatal-enriched protein tyrosine phosphatase (STEP) whose expression is increased in the frontal cortex and hippocampus of APP/PS1 mice, a

364

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

familial model of AD (Xu et al., 2012; Zhang et al., 2013). Furthermore, PTPs may also participate in Ab translocation into lipid rafts. Thus, PTPs modulate in lipid rafts Fyn kinase, a member of Src tyrosine kinases family (Vacaresse et al., 2008). Fyn participates in oligomer Ab recruitment into these microdomains (Williamson et al., 2008). Loss of tyrosine phosphate groups in VDAC protein has also been observed in lipid rafts from the frontal cortex of AD brains at late stages, as demonstrated in this work. This is in agreement with previous studies in different brain regions of patients with AD and Down syndrome (that present senile plaques and neurofibrillary tangles in adults), where similar alterations in VDAC1 phosphoisoforms were observed in temporal, frontal and occipital cortices (Yoo et al., 2001). VDAC1 has also been associated with other multiprotein complexes found in lipid rafts that form signaling platforms (signalosomes) related to neuronal preservation (Marin, 2011; Hicks et al., 2012). Thus, VDAC1 in neuronal lipid rafts forms a complex with caveolin-1, estrogen receptor alpha (ERa) and insulin-growth factor-1 receptor (IGF-1R). ERa and IGF-1R are two receptors that play an important role in neuronal preservation against different injuries through interaction with different natural extracellular ligands, such as estradiol and IGF (Marin et al., 2009). Interestingly, estradiol bound to ERa in these microdomains preserves VDAC1 phosphorylation, and keeps the channel inactivated (Diaz et al., 2014). On the contrary, anti-estrogens such as tamoxifen dephosphorylate the channel and promote its gating (Diaz et al., 2001; Valverde et al., 2002), resulting in Ab-induced cell death (Herrera et al., 2011a). These data suggest that sustained VDAC phosphorylation may preserve against neuropathology. The fact that specific multimolecular interactions in lipid rafts play a pivotal role in, both, neurodegeneration and neuroprotection against injury, strengthens the importance of lipid raft physicochemical microenvironment for these associations. Lipid raft are characterized by particular lipid composition, where cholesterol and sphingolipids are enriched, while phospholipids contain a precise proportion of saturates over polyunsaturated fatty acids, which specifically interact with signaling platforms to trigger numerous intracellular responses (Pike 2003; Lingwood and Simons, 2010). An emerging concept is that lipid rafts may evolve toward alterations in composition with aging (Fabelo et al., 2012). The progression of neurodegenerative diseases related to aging may exacerbate these lipid alterations, thereby inducing profound modifications in raft physicochemical properties (Martı´ n et al., 2010; Diaz et al., 2012), that may largely affect protein interactions and trafficking. These evidences have been observed related to, both, AD and PD neuropathology, even at early stages of these diseases (Martı´ n et al., 2010; Fabelo et al., 2011, 2014). A premature aberrant lipid composition in these microstructures has also been detected in APP/PS1 double-transgenic mice (Fabelo et al., 2012; Diaz et al., 2012). These alterations also affect signaling platforms involved in neuroprotection, where associations of

VDAC with components of raft signalosomes are also modified as a consequence of raft lipid matrix impairment (Marin, 2011). Furthermore, alterations of raft structure may be also at the basis of progressive neuronal deterioration and enhancement of neuropathological hallmarks, such as that reported here for Ab/APP and VDAC interaction and channel dephosphorylation. Fig. 7 depicts the hypothetical events related to VDAC modulation following interaction with APP and Ab taking place in lipid rafts at early stages of AD-related pathology. Undoubtedly, elucidation of the underlying alterations in

Fig. 7. Illustration of the hypothetical model of interaction of VDAC1 with APP and Ab in lipid rafts of AD brains. In normal brains (Controls), the channel at the plasma membrane is constitutively phosphorylated (red circle) and located in lipid rafts, where it partially associates with caveolin-1 and with a fraction of palmitoylated APP (palAPP) also located in lipid rafts. Both, beta-secretase (BACE) and gamma-secretase complex are also partially found in raft fractions. In AD brains (Alzheimer’s disease), alterations of lipid raft structure, most probably by derangement of lipid composition, leads to a higher accumulation of VDAC1 and APP, thus increasing interactions with VDAC/APP. As recently demonstrated (Fabelo et al., 2014), BACE is also displaced toward AD rafts, thereby favoring interactions with APP, and the generation and accumulation of Ab upon cleavage by csecretase (event 1). Ab oligomers may be formed by an excess of amyloid production from different origins. Concomitantly with Ab oligomer formation, two parallel events are triggered: on one side, Ab aggregates interact with VDAC1 (event 2), forming toxic multicomplexes in the lipid raft. On the other hand, Ab may activate tyrosine phosphatase activity to induce dephosphorylation of the channel (event 3). Under these circumstances, loss of tyrosine phosphorylation activates VDAC1 and modifies its gating and opening, which may ultimately provoke dissipation of ion gradients, metabolic exhaustion and neuronal death, thus contributing to AD neuropathology. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

multimolecular complexes integrated in pathological lipid rafts may pave the way of novel strategies of early diagnosis (Marin et al., 2013).

Acknowledgments—This work was supported by grants SAF2010-22114-C02-01/02 (Spanish Ministry of Science and Innovation), Tricontinental Atlantic Campus (CEI-Canarias), and Seventh Framework Programme of the European Commission, grant agreement 278486: DEVELAGE. CEF was hired from ACIISI, and ACA was hired from Fundacio´n CajaCanarias.

REFERENCES Akanda N, Tofighi R, Brask J, Tamm C, Elinder F, Ceccatelli S (2008) Voltage dependent anion channels (VDAC) in the plasma membrane play a critical role in apoptosis in differentiated hippocampal neurons but not in neural stem cells. Cell Cycle 7:3225–3234. Aso E, Lomoio S, Lo´pez-Gonza´lez I, Joda L, Carmona M, Ferna´ndezYagu¨e N, Moreno J, Juve´s S, Pujol A, Pamplona R, Portero-Otin M, Martı´ n V, Dı´ az M, Ferrer I (2012) Amyloid generation and dysfunctional immunoproteasome activation with disease progression in animal model of familial Alzheimer’s disease. Brain Pathol 22(5):636–653. Benseny-Cases N, Klementieva O, Cladera J (2012) In vitro oligomerization and fibrillogenesis of amyloid-beta peptides. Subcell Biochem 65:53–74. Bhattacharyya 1 R, Barren C, Kovacs DM (2013) Palmitoylation of amyloid precursor protein regulates amyloidogenic processing in lipid rafts. J Neurosci 33(27):11169–11183. Bolte S, Cordelie`res FP (2006) A guided tour into subcellular colocalization analysis in light microscopy. J. Microscopy 224:213–232. Braak H, Braak E (1999) Temporal sequence of Alzheimer’s diseaserelated pathology. In: Peters A, Morrison JH, editors. Cerebral cortex. Neurodegenerative and age-related changes in structure and function of cerebral cortex, vol. 14. New York, Boston, Dordrecht, London, Moscow: Kluwer Academic/Plenum Publishers. p. 475–512. Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del Tredici K (2006) Staining of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 112:389–404. Cecchi C, Baglioni S, Fiorillo C, Pensalfini A, Liguri G, Nosi D, Rigacci S, Bucciantini M, Stefani M (2005) Insights into the molecular basis of the differing susceptibility of varying cell types to the toxicity of amyloid aggregates. J Cell Sci 118(Pt 15):3459–3470. Chen TY, Liu PH, Ruan CT, Chiu L, Kung FL (2006) The intracellular domain of amyloid precursor protein interacts with flotillin-1, a lipid raft protein. Biochem Biophys Res Commun 342(1):266–272. Cheng H, Vetrivel KS, Drisdel RC, Meckler X, Gong P, Leem JY, Li T, Carter M, Chen Y, Nguyen P, Iwatsubo T, Tomita T, Wong PC, Green WN, Kounnas MZ, Thinakaran G (2009) S-palmitoylation of gamma-secretase subunits nicastrin and APH-1. J Biol Chem 284(3):1373–1384. Cordy JM, Hooper NM, Turner AJ (2006) The involvement of lipid rafts in Alzheimer’s disease. Mol Membr Biol 23:111–122. Cuadrado-Tejedor M, Vilarino M, Cabodevilla F, Del Rio J, Frechilla D, Perez-Mediavilla A (2011) Enhanced expression of the voltage-dependent anion channel 1 (VDAC1) in Alzheimer’s disease transgenic mice: an insight into the pathogenic effects of amyloid-beta. J Alzheimers Dis 23:195–206. Dargusch R, Schubert D (2002) Specificity of resistance to oxidative stress. J Neurochem 81:1394–13400. Das U, Scott DA, Ganguly A, Koo EH, Tang Y, Roy S (2013) Activityinduced convergence of APP and BACE-1 in acidic microdomains via an endocytosis-dependent pathway. Neuron 79(3):447–460.

365

De Pinto V, Messina A, Lane DJ, Lawen A (2010) Voltage-dependent anion selective channel (VDAC) in the plasma membrane. FEBS Lett 584:1793–1799. Diaz M, Bahamonde MI, Lock H, Mun˜oz FJ, Hardy SP, Posas F, Valverde MA (2001) Okadaic acid-sensitive activation of maxi-Clchannels by triphenylethylene antioestrogens in C1300 mouse neuroblastoma cells. J Physiol 536:79–88. Diaz ML, Fabelo N, Marin R (2012) Genotype-induced changes in biophysical properties of frontal cortex lipid raft from APP/PS1 transgenic mice. Front Physiol 3:454–458. Diaz M, Fabelo N, Ferrer I, Gomez T, Marin R (2014) Biophysical alterations in lipid rafts from human cerebral cortex associate with increased BACE1/AbPP interaction in early stages of alzheimer’s disease. J Alzheimers Dis. http://dx.doi.org/10.3233/JAD-141146. Duyckaerts C, Dickson DW (2003). In: Dickson DW, editor. Neurodegeneration: the molecular pathology of dementia and movement disorders. Basel: ISN Neuropath. p. 47–68. Ehehalt R, Keller P, Haass C, Thiele C, Simons K (2003) Amyloidogenic processing of the Alzheimer beta-amyloid precursor protein depends on lipid rafts. J Cell Biol 160:113–123. Fabelo N, Martı´ n V, Santpere G, Marı´ n R, Torrent L, Ferrer I, Dı´ az M (2011) Severe alterations in lipid composition of frontal cortex lipid rafts from Parkinson’s disease and incidental Parkinson’s disease. Mol Med 17(9–10):1107–1118. Fabelo N, Martin V, Marin R, Santper G, Aso E, Ferrer I, Diaz M (2012) Evidence for premature ‘‘Lipid raft aging’’ in APP/PS1 double transgenic mice, a familial model of Alzheimer’s disease. J. Neuroendocrinol Exp Neurol 71(10):868–881. Fabelo N, Martı´ n V, Marı´ n R, Moreno D, Ferrer I, Dı´ az M (2014) Altered lipid composition in cortical lipid rafts occurs at early stages of sporadic Alzheimer’s disease and facilitates APP/ BACE1 interactions. Neurobiol Aging 5(8):1801–1812. Ferrer I (2009) Altered mitochondria, energy metabolism, voltagedependent anion channel, and lipid rafts converge to exhaust neurons in Alzheimer’s disease. J Bioenerg Biomembr 41(5):425–431. Herrera JL, Diaz M, Hernandez-Fernaud JR, Salido E, Alonso R, Fernandez C, Morales A, Marin R (2011a) Voltage-dependent anion channel as a resident protein of lipid rafts, posttransductional regulation by estrogens and involvement in neuronal preservation against Alzheimer’s disease. J Neurochem 116. 830-827. Herrera JL, Fernandez C, Diaz M, Cury D, Marin R (2011b) Estradiol and tamoxifen differentially regulate a plasmalemmal voltagedependent anion channel involved in amyloid-beta induced neurotoxicity. Steroids 76(9):840–844. Hicks DA, Nalivaeva NN, Turner AJ (2012) Lipid rafts and Alzheimer’s disease: protein-lipid interactions and perturbation of signaling. Front Physiol 3:189–194. Hur JY, Teranishi Y, Kihara T, Yamamoto NG, Inoue M, Hosia W, Hashimoto M, Winblad B, Frykman S, Tjernberg LO (2012) Identification of novel c-secretase-associated proteins in detergentresistant membranes from brain. J Biol Chem 287:11991–12005. Kalvodova L, Kahya N, Schwille P, Ehehalt R, Verkade P, Drechsel D, Simons K (2005) Lipids as modulators of proteolytic activity of BACE, involvement of cholesterol, glycosphingolipids & anionic phospholipids in vitro. J Biol Chem 280:36815–36823. Ke N, Wang X, Xu X, Abassi YA (2011) The xCELLigence system for real-time and label-free monitoring of cell viability. Methods Mol Biol 740:33–43. Klein WL (2002) Ab toxicity in Alzheimer’s disease: globular oligomers (ADDLs) as new vaccine and drug targets. Neurochem Int 41:345–352. Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C, Finch CE, Krafft GA, Klein WL (1998) Diffusible, nonfibrillar ligands derived from Abeta1–42 are potent central nervous system neurotoxins. Proc Natl Acad Sci USA 95:6448–6453. LeVine 3rd H (1993) Thioflavine T interaction with synthetic Alzheimer’s disease beta-amyloid peptides: detection of amyloid aggregation in solution. Protein Sci 2(3):404–410.

366

C. Fernandez-Echevarria et al. / Neuroscience 278 (2014) 354–366

Lingwood D, Simons K (2010) Lipid rafts as a membrane-organizing principle. Science 327:46–50. Lowe J, Mirra SS, Hyman B, Dickson DW (2008) Ageing and dementia. In: Love S, Louis DN, Ellison DW, editors. Greenfield’s neuropathology. London: Hodder Arnold. p. 1031–1152. 1. Manczak M, Reddy HP (2012) Abnormal interaction of VDAC1 with amyloid beta and phosphorylated tau causes mitochondrial dysfunction in Alzheimer’s disease. Human Mol Genet 21:5131–5146. Marin R, Guerra B, Herna´ndez-Jime´nez JG, Kang XL, Fraser JD, Lo´pez FJ, Alonso R (2003) Estradiol prevents amyloid-beta peptide-induced cell death in a cholinergic cell line via modulation of a classical estrogen receptor. Neuroscience 121(4):917–926. Marin R, Ramı´ rez CM, Gonza´lez M, Gonza´lez-Mun˜oz E, Zorzano A, Camps M, Alonso R, Dı´ az M (2007) Voltage-dependent anion channel (VDAC) participates in amyloid beta-induced toxicity and interacts with plasma membrane estrogen receptor alpha in septal and hippocampal neurons. Mol Membr Biol 24(2):148–160. Marin R, Ramı´ rez C, Morales A, Gonza´lez M, Alonso R, Dı´ az M (2008) Modulation of Abeta-induced neurotoxicity by estrogen receptor alpha and other associated proteins in lipid rafts. Steroids 73:992–996. Marin R, Dı´ az M, Alonso R, Sanz A, Are´valo MA, Garcia-Segura LM (2009) Role of estrogen receptor alpha in membrane-initiated signaling in neural cells: interaction with IGF-1 receptor. J Steroid Biochem Mol Biol 114:2–7. Marin R (2011) Signalosomes in the brain: relevance in the development of certain neuropathologies such as Alzheimer’s disease. Front Physiol 2(23):1–4. Marin R, Rojo JA, Fabelo N, Fernandez CE, Diaz M (2013) Lipid raft disarrangement as a result of neuropathological progresses: a novel strategy for early diagnosis? Neuroscience 245:26–39. Martı´ n V, Fabelo N, Santpere G, Puig B, Marı´ n R, Ferrer I, Dı´ az M (2010) Lipid alterations in lipid rafts from Alzheimer’s disease human brain cortex. J Alzheimers Dis 19(2):489–502. Mukherjee A, Arnaud L, Cooper JA (2003) Lipid-dependent recruitment of neuronal Src to lipid rafts in the brain. J Biol Chem 278:40806–40814. Oppenheim AV, Willsky AS, Young IT (1983) Signals and systems. 2nd ed. Englewood Cliffs, New Jersey: Prentice Hall. Park E, Lee GJ, Choi S, Choi SK, Chae SJ, Kang SW, Pak YK, Park HK (2010) The role of glutamate release on voltage-dependent anion channels (VDAC)-mediated apoptosis in an eleven vessel occlusion model in rats. PLoS One 5:e15192. Pike LJ (2003) Lipid rafts: bringing order to chaos. J Lipid Res 44:655–667. Ramı´ rez CM, Gonza´lez M, Dı´ az M, Alonso R, Ferrer I, Santpere G, Puig B, Meyer G, Marin R (2009) VDAC and ERalpha interaction in caveolae from human cortex is altered in Alzheimer’s disease. Mol Cell Neurosci 42(3):172–183. Rushworth JV, Hooper NM (2011) Lipid rafts: linking Alzheimer’s amyloid-beta production, aggregation, and toxicity at neuronal membranes. In J Alzheimers Dis. http://dx.doi.org/10.4061/2011/ 603052. Sabirov RZ, Merzlyak PG (2012) Plasmalemmal VDAC controversies and maxi-anion channel puzzle. Biochim Biophys Acta 1818(6):1570–1580. Scha¨gger H (2006) Tricine-SDS–PAGE. Nat Prot 1:16–22. Schlenzig D, Manhart S, Cinar Y, Kleinschmidt M, Hause G, Willbold D, Funke SA, Schilling S, Demuth HU (2009) Pyroglutamate

formation influences solubility and amyloidogenicity of amyloid peptides. Biochemistry 48(29):7072–7078. Schengrund CL (2010) Lipid rafts: keys to neurodegeneration. Brain Res Bull 82(1–2):7–17. Schneider A, Rajendran L, Honsho M, Gralle M, Donnert G, Wouters F, Hell SW, Simons M (2008) Flotillin-dependent clustering of the amyloid precursor protein regulates its endocytosis and amyloidogenic processing in neurons. J Neurosci 28(11):2874–2882. Sultana R, Poon HF, Cai J, Pierce WM, Merchant M, Klein JB, Markesbery WR, Butterfield DA (2006) Identification of nitrated proteins in Alzheimer’s disease brain using a redox proteomics approach. Neurobiol Dis 22(76–87):37. Thinnes FP (2011) Apoptogenic interactions of plasmalemmal type-1 VDAC and Ab peptides via GxxxG motifs induce Alzheimer’s disease – a basic model of apoptosis? Wien Med Wochenscher 161:274–276. Thinnes FP (2012) Alzheimer disease controls cancer concerning the apoptogenic interaction of cell membrane-standing type-1 VDAC and amyloid peptides via GxxxG motifs. Mol Genet Metab 106(4):502–503. Thinnes FP (2013) New findings concerning vertebrate porin II–on the relevance of glycine motifs of type-1 VDAC. Mol Genet Metab 108(4):212–224. To˜ugu V, Karafin A, Zovo K, Chung RS, Howells C, West AK, Palumaa P (2009) Zn(II)- and Cu(II)-induced non-fibrillar aggregates of amyloid-beta (1–42) peptide are transformed to amyloid fibrils, both spontaneously and under the influence of metal chelators. J Neurochem 110(6):1784–1795. Toychiev AH, Sabirov RZ, Takahashi N, Ando-Akatsuka Y, Liu H, Shintani T, Noda M, Okada Y (2009) Activation of maxi-anion channel by protein tyrosine dephosphorylation. Am J Physiol Cell Physiol 297(4):C990–1000. Vacaresse N, Møller B, Danielsen EM, Okada M, Sap J (2008) Activation of c-Src and Fyn kinases by protein-tyrosine phosphatase RPTPalpha is substrate-specific and compatible with lipid raft localization. J Biol Chem 283(51):35815–35824. Valverde MA, Hardy SP, Dı´ az M (2002) Activation of Maxi Cl(-) channels by antiestrogens and phenothiazines in NIH3T3 fibroblasts. Steroids 67(6):439–445. Vetrivel KS, Meckler X, Chen Y, Nguyen PD, Seidah NG, Vassar R, Wong PC, Fukata M, Kounnas MZ, Thinakaran G (2009) Alzheimer disease Abeta production in the absence of Spalmitoylation-dependent targeting of BACE1 to lipid rafts. J Biol Chem 284:3793–3803. Williamson R, Usardi A, Hanger DP, Anderton BH (2008) Membranebound beta-amyloid oligomers are recruited into lipid rafts by a fyn-dependent mechanism. FASEB J 22(5):1552–1559. Xu J, Kurup P, Nairn AC, Lombroso PJ (2012) Striatal-enriched protein tyrosine phosphatase in Alzheimer’s disease. Adv Pharmacol 64:303–325. Yoo BC, Fountoulakis M, Cairns N, Lubec G (2001) Changes of voltage-dependent anion-selective channel proteins VDAC1 and VDAC2 brain levels in patients with Alzheimer’s disease and Down syndrome. Electrophoresis 22(1):172–179. Zhang L, Xie JW, Yang J, Cao YP (2013) Tyrosine phosphatase STEP61 negatively regulates amyloid b-mediated ERK/CREB signaling pathways via a7 nicotinic acetylcholine receptors. J Neurosci Res 91(12):1581–1590. Zheng YZ, Berg KB, Foster LJ (2009) Mitochondria do not contain lipid rafts, and lipid rafts do not contain mitochondrial proteins. J Lipid Res 50(5):988–998.

(Accepted 24 July 2014) (Available online 26 August 2014)