Amyloid-β Oligomers Induce Differential Gene Expression in Adult Human Brain Slices*

Background: Soluble Aβ oligomers (AβOs) have been increasingly proposed as the cause of synapse failure and cognitive dysfunction in Alzheimer disease. Results: Sublethal AβO concentrations induce changes in gene expression in adult human brain slices. Conclusion: AβOs impact transcription in important neuronal pathways preceding neurodegeneration. Significance: Results establish early mechanisms involved in AβO-triggered neuronal dysfunction in a novel human-derived experimental model. Cognitive decline in Alzheimer disease (AD) is increasingly attributed to the neuronal impact of soluble oligomers of the amyloid-β peptide (AβOs). Current knowledge on the molecular and cellular mechanisms underlying the toxicity of AβOs stems largely from rodent-derived cell/tissue culture experiments or from transgenic models of AD, which do not necessarily recapitulate the complexity of the human disease. Here, we used DNA microarray and RT-PCR to investigate changes in transcription in adult human cortical slices exposed to sublethal doses of AβOs. The results revealed a set of 27 genes that showed consistent differential expression upon exposure of slices from three different donors to AβOs. Functional classification of differentially expressed genes revealed that AβOs impact pathways important for neuronal physiology and known to be dysregulated in AD, including vesicle trafficking, cell adhesion, actin cytoskeleton dynamics, and insulin signaling. Most genes (70%) were down-regulated by AβO treatment, suggesting a predominantly inhibitory effect on the corresponding pathways. Significantly, AβOs induced down-regulation of synaptophysin, a presynaptic vesicle membrane protein, suggesting a mechanism by which oligomers cause synapse failure. The results provide insight into early mechanisms of pathogenesis of AD and suggest that the neuronal pathways affected by AβOs may be targets for the development of novel diagnostic or therapeutic approaches.


Cognitive decline in Alzheimer disease (AD) is increasingly attributed to the neuronal impact of soluble oligomers of the amyloid-␤ peptide (A␤Os)
. Current knowledge on the molecular and cellular mechanisms underlying the toxicity of A␤Os stems largely from rodent-derived cell/tissue culture experiments or from transgenic models of AD, which do not necessarily recapitulate the complexity of the human disease. Here, we used DNA microarray and RT-PCR to investigate changes in transcription in adult human cortical slices exposed to sublethal doses of A␤Os. The results revealed a set of 27 genes that showed consistent differential expression upon exposure of slices from three different donors to A␤Os. Functional classification of differentially expressed genes revealed that A␤Os impact pathways important for neuronal physiology and known to be dysregulated in AD, including vesicle trafficking, cell adhesion, actin cytoskeleton dynamics, and insulin signaling. Most genes (70%) were down-regulated by A␤O treatment, suggesting a predominantly inhibitory effect on the corresponding pathways. Significantly, A␤Os induced down-regulation of synaptophysin, a presynaptic vesicle membrane protein, suggesting a mechanism by which oligomers cause synapse failure. The results provide insight into early mechanisms of pathogenesis of AD and suggest that the neuronal pathways affected by A␤Os may be targets for the development of novel diagnostic or therapeutic approaches.
Alzheimer disease (AD), 5 the most common form of dementia in the elderly, currently affects more than 35 million people worldwide (1). AD is characterized clinically by early memory dysfunction followed by progressive cognitive impairment (1,2). Soluble oligomers of the amyloid-␤ peptide (〈␤⌷s) are currently considered central players in the pathogenesis of AD (3)(4)(5)(6). A␤Os initially were found to act as potent cell surface ligands, instigating rapid inhibition of long-term potentiation and alterations in neuronal cell signaling (7). Those findings have been followed by reports on a variety of neurotoxic effects of A␤Os, including Tau hyperphosphorylation (8), calcium dysregulation and oxidative stress (9 -13), blockade of fast axonal transport (14,15), altered turnover of neuronal receptors involved in synaptic plasticity (e.g. 16 -20), and synapse loss (17,19).
Although it is now generally accepted that A␤Os cause synapse dysfunction, the molecular and cellular mechanisms underlying their toxicity are still poorly understood. In part, this is because most studies of A␤O toxicity have utilized rodent-derived cell/tissue cultures or transgenic models of AD, which do not necessarily recapitulate the complexity of the human disease. Previous studies have indicated changes in neuronal gene expression in AD brain (e.g. [21][22][23] and in transgenic mouse models of AD (e.g. 24,25). However, a central issue that remains to be determined is whether the changes in gene expression detected in those studies represent a late consequence of cerebral inflammation and neurodegeneration or a direct and early effect of A␤Os on neuronal gene expression.
Here, we report an investigation of changes in gene expression induced by A␤Os using human adult cortical slices as a novel experimental model to study the neuronal impact of A␤Os. Using a combination of DNA microarray analysis and quantitative reverse transcriptase PCR (qRT-PCR), we found that sublethal concentrations of A␤Os induce changes in expression of genes classified into several pathways important for neuronal physiology, including synaptophysin, a presynaptic vesicle membrane protein that has been implicated in synaptic failure in AD (26 -28). A detailed understanding of the mechanisms by which A␤Os affect neuronal gene expression and function may uncover early biomarkers of AD pathology and lead to the development of novel and effective therapeutics to prevent or halt the progression of AD.

EXPERIMENTAL PROCEDURES
Adult Human Cortical Slices-Adult human brain cortical fragments were obtained from patients with pharmacoresistant temporal lobe epilepsy submitted to hippocampectomy and anterior temporal lobectomy for removal of epileptic foci detected by magnetic resonance imaging (MRI). Only healthy cortical tissue removed from the area approached as surgical access to the target hippocampus was used for culture purposes. The health of cortical tissues was determined based on the structural/anatomical integrity revealed by MRI (supplemental Fig. 1), macroscopic appearance, and lack of correlation with epileptic activity in scalp video-EEG monitoring. Donors (six men and four women) were 34 Ϯ 9 years old and gave written informed consent for use of brain tissue that would otherwise have been discarded. All procedures were approved and regulated by the National Committee for Research Ethics (CONEP) of the Brazilian Ministry of Health (Protocol 0069.0.197.000-05).
Tissue was placed immediately in transport medium (50% (v/v) Hanks' balanced salt solution containing 10 mM HEPES, 3 mg/ml glucose, and 50 g/ml gentamicin diluted in Neurobasal A medium supplemented previously with 2% B27 and 0.5 mM glutamine) and transported to the laboratory within 30 min of removal. Under sterile conditions, meninges were carefully removed, and 0.3-mm slices were prepared using a McIlwain tissue chopper. Slices were plated on 24-well plates (2-3 slices/ well) in Neurobasal A/B27 (Invitrogen) medium with antibiotics. One-third of the medium was replaced after 3 days in culture. After 5 days at 37°C in a 5% CO 2 atmosphere, cultures were treated with vehicle or A␤Os (500 nM) and further incubated at 37°C for 12 h. Cell viability in cultured slices was evaluated by Live/Dead and MTT reduction assays as described below.
Dissociated Hippocampal Neuronal Cultures-Hippocampal neuronal cultures were prepared from 18-day-old rat embryos as described previously (29,30). Briefly, hippocampi were dissected in PBS-glucose and mechanically dissociated, and cells were plated onto poly-L-lysine-coated wells at a density of 1.5 ϫ 10 6 cells/well (for 35-mm wells) in Neurobasal/B27 medium with antibiotics. All procedures were approved by and followed the guidelines of the Institutional Animal Care and Utilization Committee of the Federal University of Rio de Janeiro (Protocol IBQM 022). After 21 days at 37°C in a 5% CO 2 atmosphere, cultures were treated with vehicle or A␤Os and further incubated at 37°C for 12 or 24 h.
A␤O Preparation-A␤Os were prepared in PBS (31) with minor modifications as described previously (32) using A␤ 42 peptide from Bachem (Torrance, CA). The preparation was centrifuged at 14,000 ϫ g for 10 min at 4°C to remove any insoluble aggregates, and the supernatant containing soluble A␤Os was transferred to clean tubes and stored at 4°C. Protein concentration was determined using the BCA assay (Pierce). Oligomer solutions were used within 24 h of preparation. Routine characterization of oligomer preparations was performed by size-exclusion chromatography and Western blot and, occasionally, by transmission electron microscopy (supplemental Fig. 2). Collectively, the results indicate that our preparation comprises soluble oligomeric species including dimers, trimers, tetramers, and higher molecular mass oligomers of ϳ50 -180 kDa, ranging in diameter from ϳ1.5 to 3.5 nm.
Live/Dead Viability Assay-Cell viability in cultured slices from three different donors was determined using the Live/ Dead assay (Invitrogen). At different days in vitro, slices were exposed to dyes (4 M calcein and 2 M ethidium homodimer) for 40 min, washed twice with ice-cold PBS, and cut into 12-m sections for imaging on a Nikon Eclipse TE300 microscope. Live or dead neurons were identified by green calcein fluorescence or red ethidium fluorescence, respectively. Percentages of live neurons are expressed relative to the total number of cells (determined by DAPI staining) in each slice.
Immunohistochemistry-After 4 days in culture, brain slices were fixed in 4% paraformaldehyde and double stained for mature neurons (mouse anti-NeuN, Chemicon; 1:100) and astrocytes (rabbit anti-GFAP, Dako; 1:200). Alexa 488 and Alexa 594 conjugates were used as secondary antibodies (Molecular Probes; 1:200). DAPI (Sigma) was used to visualize nuclei. Antibody incubations were performed using free floating sections (treated previously with 0.1 M citrate buffer, pH 6, at 60°C for 5 min and blocked with 5% BSA, 5% normal goat serum, and 1% Triton X-100 for 6 h at room temperature) in PBS containing 1% Triton X-100. Primary antibodies were diluted in blocking solution and incubated for 48 h at 4°C, followed by incubation with Alexa-conjugated secondary antibodies for 24 h at 4°C. Tissue autofluorescence was quenched by incubation with 0.06% potassium permanganate for 10 min at room temperature.
MTT Assay-Hippocampal slices were incubated for 24 h in the presence of A␤Os at different concentrations, and cell viability was analyzed by the MTT reduction assay carried out as described previously (13). Briefly, slices were incubated with a 0.5 mg/ml solution of MTT at 37°C for 4 h to allow reduction to formazan blue by metabolically active cells. Cells were then lysed, and formazan crystals were solubilized by incubation in 0.01 N HCl containing 10% SDS for 24 h under agitation at room temperature. Optical density at 570 nm was measured in a ThermoMax Microplate reader.
RNA Extraction and Amplification-Total RNA from hippocampal cultures was extracted with TRIzol (Invitrogen) following the manufacturer's instructions. One milliliter of TRIzol was used to extract RNA from 1.5 ϫ 10 6 cells. The purity and integrity of RNA preparations were checked by the 260/280 nm absorbance ratio and by agarose gel electrophoresis. Only preparations with 260/280 nm ratios Ն 1.8 and no signs of RNA degradation were used. mRNA from human cortical slices was obtained using an RNA amplification protocol. After treatment with A␤Os, slices were washed twice with PBS and kept in RNAlater (Ambion) at Ϫ20°C until use. For total RNA extraction, tissue was ground using an electric homogenizer, and RNA was purified using an RNeasy mini kit (Qiagen). Total RNA was then used as a template in a two-round linear amplification procedure based on T7-driven amplification as described previously (33). The quality of amplified RNAs was checked by agarose gel electrophoresis upon visualization with ethidium bromide (supplemental Fig. 3). All samples presented a smear between 300 and 700 bp and no signs of RNA degradation.
Microarray Analysis-Ten g of amplified RNA from each sample (three different donors, each used for both control and A␤O-treated conditions, corresponding to six paired samples in total) were used as template for cDNA synthesis. cDNA samples were labeled indirectly with Alexa Fluor 555 or Alexa Fluor 647 reactive dye (Invitrogen) and IMPROM II reverse transcriptase (Promega) in a reverse transcriptase reaction and purified as described (34). Hybridization reactions were performed in duplicate with dye swapping using a human universal 4,800 chip (35), with excellent agreement between duplicates. After washing, slides were scanned on a confocal laser scanner (ScanArray Express, PerkinElmer Life Sciences), and data were extracted with ScanArray Express software. Self-self hybridization was utilized to define cutoff limits for differential gene expression (36). The percentile used was 0.95.
A list of differentially expressed genes induced by A␤O treatment was used to identify the affected biological pathways using the Gene Set Analysis Toolkit V2 (37). The gene list used as input for the Kegg pathway analysis comprised both up-and down-regulated genes present at the intersections between at least two samples. The parameters used were as follows: organism, Homo sapiens; ID type, gene_symbol; reference set, Entrez Gene; significance level, 0.05; statistics test, hypergeometric; multiple test correction, Benjamini and Hochberg; minimum genes per category, 2.
Quantitative RT-PCR Assays-Quantitative expression analysis of genes of interest was performed by quantitative real-time PCR on an Applied Biosystems 7500 real-time PCR system with the Power SYBR kit (Applied Biosystems). For human tissue, cDNA converted from amplified RNA was used as the template. This procedure has been found not to introduce any bias on relative gene expression measurement (38). Three reference human genes, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), hypoxanthine phosphoribosyltransferase 1 (HPRT), and ␤-actin (ACTB), were used to generate a single normalization factor as described (39).
For rat neuronal cultures, template cDNA was prepared from total RNA (1 g) using 50 pmol of oligo dT 20 and the Superscript III First Strand cDNA kit (Invitrogen), and ACTB or GAPDH was used as the endogenous control. Cycle threshold (Ct) values were used to calculate -fold changes in gene expression using the 2 Ϫ⌬⌬Ct method (40). In all cases, reactions were performed in 20-l reaction volumes.
Western Blotting-Mature rat embryonic hippocampal cultures were treated with A␤Os (500 nM) or vehicle for 12 or 24 h, rinsed with PBS, and lysed in buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1.5 mM MgCl 2 , 1.5 mM EDTA, 1% Triton X-100, 10% glycerol, and Halt TM protease inhibitors mixture (Thermo Fisher Scientific, Rockford, IL). Human brain slice extracts were prepared using the same buffer in the ratio of 50 l/slice. The protein content in the extracts was determined using the BCA TM protein assay kit (Thermo Fisher Scientific). Extracts (100 g of protein/lane) were resolved on 12% SDS-PAGE. Proteins were transferred to nitrocellulose membranes (Hybond TM -C Extra, Amersham Biosciences) for 90 min at 100 V. Membranes were blocked with 3% BSA in Tris-buffered saline/Tween 20 (TBS-T: 10 mM Tris, pH 7.2, 150 mM NaCl, and 0.1% Tween 20), followed by overnight incubation with anti-synaptophysin monoclonal antibody (5 g/ml; Sigma-Aldrich) and polyclonal anti-cyclophilin B (0.01 g/ml; Abcam) at 4°C. After washing with TBS-T, immunoreactivity was visualized using peroxidase-conjugated anti-rabbit IgG secondary antibody (1:10,000 dilution; Zymed Laboratories Inc., Carlsbad, CA) for cyclophilin B detection (used as a loading control) and anti-mouse IgG secondary antibody (1:50,000 dilution; Amersham Biosciences) for synaptophysin detection. SuperSignal West Pico chemiluminescent detection (Thermo Fisher Scientific) was used for visualization. Densitometric scanning and quantification were carried out using NIH ImageJ (Windows version).

Exposure to a Sublethal A␤O Concentration Alters Gene Expression in Human Cortical
Slices-Organotypic cultures of post-mortem human cortex have been previously prepared (41). Here, starting from ex vivo human cortical tissue, we developed a novel experimental model to investigate the neuronal impact of A␤ oligomers. Slices from healthy cortical tissue obtained from adults submitted to surgical removal of hippocampal epileptic foci were maintained successfully in culture for up to 25 days with cell viability greater than 50% (Fig. 1). Semiquantitative analysis of cell types in slices from three different donors revealed an average of 60% neurons and 21% GFAP-positive cells in cortical human slices cultured for 4 days. Previous studies have shown that exposure of dissociated neuronal cultures (for up to 24 h) to A␤Os at submicromolar concentrations causes neuronal dysfunction in the absence of cell death (8,17) (reviewed in Ref. 5). Consistent with those findings, no change in cell viability was detected in human cortical slices exposed to A␤Os (500 nM) for 24 h compared with control, vehicle-treated slices (supplemental Fig. 4).
Following treatment with A␤⌷, mRNA was extracted from slices and hybridization was performed on a customized cDNA platform containing 4608 human genes (35). Microarray analysis revealed significant alterations in gene expression induced by A␤Os compared with paired vehicle-treated slices from the same donors (Fig. 2). Considering all of the genes identified as up-or down-regulated that were present in at least two independent donors, a total of 345 genes were found to be differentially expressed upon A␤O treatment (72% of which were down-regulated ( Fig. 2A)), corresponding to ϳ7% of the total number of genes represented on the chip. When the intersec-tion between the three independent donors was taken into account, 27 genes showed consistent alteration in expression (Fig. 2B). Again, most of those genes (70%) were down-regulated (Fig. 2C). Complete lists of up-and down-regulated genes, with -fold change values for each independent donor, are shown in supplemental Tables 1 and 2, respectively.
Functional Classification of Differentially Expressed Genes in A␤O-treated Slices-To assign biological functions to the list of A␤O-induced differentially expressed genes, we mapped both up-and down-regulated genes present in hybridizations from at least two donors to specific biological Kegg pathways (42), an approach known to add confidence to the results found for expression of individual genes (43). This analysis revealed 18 pathways that were significantly over-represented (p Յ 0.05) in A␤O-treated tissue ( Fig. 3 and supplemental Table 3). Significantly, six of those pathways ("SNARE interaction in vesicular transport," "Alzheimer's disease," "axon guidance," "long-term depression," "regulation of actin cytoskeleton," and "focal adhesion") have been implicated in the mechanisms of pathogenesis in the central nervous system. Of considerable interest was finding the "insulin-signaling pathway" among the over-represented pathways, as impaired neuronal insulin signaling has been linked recently to AD pathogenesis and, specifically, to the impact of A␤Os (18,19). Another over-represented pathway previously associated with energy metabolism and AD pathogenesis is the "adipocytokine-signaling pathway", as leptin, an adipocytokine involved in fatty acid metabolism, has recently been reported to control A␤ levels (44). It is noteworthy that, among the 41 distinct genes mapped to the 18 over-represented pathways (representing 12% of the total number of differen-tially expressed genes), 63% were down-regulated by A␤Os. Functional classification analysis thus suggests that A␤Os down-regulate expression of genes crucial for proper neuronal function and plasticity, establishing organotypic cultures of adult human brain cortex as a novel experimental model that effectively recapitulates the association between molecular abnormalities and clinical manifestations of AD.
Validation of A␤O-induced Changes in Gene Expression-Selected differentially expressed genes identified by microarray analysis were validated by qRT-PCR using the same set of cortical slice samples plus four extra pairs of A␤O-and vehicletreated samples obtained from independent donors not included in the microarray analysis. The basic criterion used to select genes for qRT-PCR analysis was that they should be present at the intersection between the three donors used in the microarray analysis ( Fig. 2C and supplemental Tables 1 and 2). mRNA levels of 4 of 6 up-regulated annotated genes and 11 of 17 down-regulated annotated genes were determined directly by qRT-PCR (Fig. 4), corresponding to 65% of the differentially expressed genes identified at the intersection among three donors in the microarray data (Fig. 2C). Two additional genes tested (rbed and prlr) rendered inconclusive results because of very low basal expression levels. Primer sequences for all of the genes can be found in supplemental Table 4. Differential expression was confirmed by qRT-PCR for five down-regulated genes and one up-regulated gene (Fig. 4).

Synaptophysin Down-regulation at mRNA and Protein Levels in A␤O-treated Hippocampal Neuronal Cultures and Human
Brain Slices-Because synaptophysin has been implicated in synapse loss in AD (26, 45, 46) and its expression was markedly down-regulated by A␤Os (ϳ50% decrease compared with vehicle-treated slices (Fig. 4)), we further investigated the effect of short-term exposure to A␤Os (500 nM) on synaptophysin levels using mature rat hippocampal neuronal cultures. Consistent with the finding in human brain slices, exposure of neuronal cultures to A␤Os for 12 h caused a significant reduction in synaptophysin mRNA level (Fig. 5A). Although synaptophysin protein levels were not affected by A␤Os at 12 h of exposure, a longer (24 h) incubation caused a significant reduction in synaptophysin (Fig. 5B), indicating that A␤O-induced down-regulation of synaptophysin mRNA is followed by a comparable reduction in the protein level. Importantly, corroborating our microarray/qPCR findings, we also detected down-regulation of synaptophysin protein levels in A␤O-treated human brain slices obtained from three new independent donors (Fig. 5C).

DISCUSSION
Changes in gene expression have been reported as part of AD pathology, but a general consensus is still lacking as to which genes or cellular pathways are primarily affected, especially at early stages of the disease. Previous studies have relied largely on the analysis of post-mortem material, making it difficult to distinguish whether changes in transcription were elicited directly by A␤⌷ toxic signaling or were a consequence of events taking place at later stages of the disease, including brain FIGURE 2. Differentially expressed genes in A␤O-treated human cortical slices. A, a total of 345 genes (72% down-regulated) were found to be differentially expressed in A␤O-treated slices. B, Venn diagrams showing the total number of up-and down-regulated genes in samples from three different donors (labeled S1, S2, and S3). C, heat map for genes found to be up-(red) or down-regulated (green) by A␤Os in samples from the three independent donors (S1-S3). The color code (see scale bar) reflects -fold change in expression.  . qRT-PCR analysis of A␤O-induced differential gene expression. cDNA from the three donors used in the microarray analysis plus four samples from additional independent donors were used. Expression levels were determined by relative expression, calculated using the gnorm method (see "Experimental Procedures"). Three different normalizer genes were used (ACTB, HPRT, and GAPDH). Each data point (gene or donor) reflects the mean value from an independent determination (for an individual donor) performed at least in triplicate. Down-regulated genes (indicated by microarray analysis) are represented by green symbols: synaptophysin (SYP), cullin 3 (CUL3), vaccinia-related kinase 3 (VRK3), heterogeneous nuclear ribonucleoprotein A1 (HNRPA1), C2 and WW domain-containing E3 ubiquitin protein ligase 1 (HECW1), casein kinase 2 (CSNK2A2), asparaginyl tRNA synthetase 2 (NARS2), Bcl-2-modifying factor (BMF), HEAT repeat containing 5A (HEATR5A), neurochondrin (NCDN), and adaptor-related protein complex 3, beta 2 subunit (AP3B2). Up-regulated genes (indicated by microarray analysis) are represented by red symbols: cytoplasmic FMR1-interacting protein 1 (CYF1P1), cell division cycle 14 (CDC14A), taste receptor 2 (TAS2R49), and vesicle-associated membrane protein 3 (VAMP3). Medians are denoted by horizontal solid lines.
inflammation and neurodegeneration. In the present study, we present the first demonstration that A␤⌷s (used at a sublethal concentration) markedly affect gene expression in the absence of overt neurodegeneration in adult human brain slices in culture.
Analyses of changes in gene expression in post-mortem AD brain have yielded controversial results. For example, up-regulation of tumor suppressor genes (21) and down-regulation of retromer trafficking complex genes (22) have each been reported independently as the most significant transcriptional alteration in AD hippocampus. Similarly, in cortical tissue from AD patients, expression of both immunity-related MHC II (47) and calcium-signaling genes (48) were found to be dysregulated. More recently, an elegant study employing cultivated slices from post-mortem AD brain (23) described significant alterations in the expression of genes related to synaptic activity and ␤-amyloid processing in prefrontal cortex in the presymptomatic stage, when plaque pathology was not detected. In addition to post-mortem AD brain, transgenic mouse models of AD have been used to detect gene expression alterations relevant to AD pathogenesis. Again, there is no consensus among transcriptional alterations reported in those studies. For example, Reddy et al. (25) found up-regulation of genes involved in mitochondrial energy metabolism in animals expressing human APP harboring an AD-associated mutation (Tg2576 mice), whereas the down-regulation of genes important for memory consolidation was reported in a study using a double transgenic mouse model (APP ϩ PS1 human transgenes) (24). In addition to technical considerations (e.g. different mouse models or DNA microarray platforms employed), it is likely that differences between specific molecular and cellular events underlying AD pathogenesis and findings in animal models or in post-mortem tissue contributed to the differences in outcome in the studies mentioned above.
Animal models do not mirror the full range of AD symptoms (49,50), and studies with post-mortem human brain samples are often hampered by difficulties in determining the effect of factors such as the post-mortem interval, brain pH, age, disease status, and other neurological conditions affecting tissue quality (51, 52). As a novel approach to gain insight into clinically relevant disease mechanisms that cannot be properly addressed using post-mortem or animal model brain tissue, here we introduced the use of cultured adult human brain slices as a model to investigate A␤⌷-induced AD pathology. To date, a single study has investigated the effects of A␤⌷ treatment on transcription (53) and revealed down-regulation of genes involved in protein modification and degradation. However, the possible physiopathological significance of those findings appears questionable, as the concentration of A␤Os used (50 M) was orders of magnitude higher than the concentrations found in the diseased human brain. Moreover, that study made use of an immortalized cell line (rather than neurons), further complicating correlation of the findings with human brain neuropathology.
The present work constitutes the first description of the early effects of A␤Os at low concentrations on gene expression in adult human neurons under conditions in which cell viability is not impaired. The investigation of the impact of A␤Os on human brain tissue may provide insight into mechanisms that are directly relevant to AD pathogenesis but might not be apparent in studies based on transgenic rodent models or dissociated cell cultures.
Quantitative RT-PCR was used to validate microarray results. The percentage of validation of differential expression was higher in the set of down-regulated genes, consistent with the finding that A␤Os predominantly induced down-regulation rather than up-regulation of gene expression ( Fig. 2A). Moreover, down-regulation of synaptophysin (SYP) and vaccinia-related kinase 3 (VRK3), two of the three genes with the highest level of down-regulation in the microarray analysis (Fig.  2C), was validated by qPCR (Fig. 4), confirming that the most significant alterations detected in the microarray analysis indeed reflect gene expression changes triggered by A␤O treatment. qPCR data also validated the up-regulation of cytoplasmic FMR1-interacting protein 1 (CYFIP1), initially detected in the microarray analysis (Fig. 2C). CYFIP1, along with EIF4E, forms a complex with Fragile X mental retardation protein (FMR1), a key regulator of translation in dendritic spines known to be down-regulated in the Fragile X syndrome (54). It has been proposed recently that dementia in AD and mental FIGURE 5. A␤Os induce down-regulation of synaptophysin protein levels in rat hippocampal cultures and human brain slices. A, down-regulation of synaptophysin mRNA was observed in hippocampal cultures exposed to A␤Os (500 nM). qPCR analysis was performed using two different gene normalizers (ACTB and GAPDH). *, p ϭ 0.09; **, p Ͻ 0.05, Student's t test. B, synaptophysin (Syp) down-regulation was also detected at the protein level (determined by Western blot analysis) after a 24-h treatment with A␤Os (*, p Ͻ 0.05 compared with control, indicated by the dashed line). qPCR and Western blot results are presented as means Ϯ S.E. from at least three independent cultures. C, human brain slices were exposed to 500 nM A␤Os for 12 h and protein levels (synaptophysin and cyclophilin (Cyc)) were determined by Western blots. Significant down-regulation (*, p Ͻ 0.02, Student's t test) of synaptophysin was verified compared with vehicle. Results are means Ϯ S.E. from experiments using material from three independent donors. A representative blot is shown at the top. retardation in Fragile X syndrome share common pathogenic features and may both be consequences of synaptic dysfunction caused by loss of or decreased FMR1 activity (6,55). Therefore, it is possible that A␤O-induced up-regulation of CYFIP1 is related to disruption of FMR1-mediated control of translation in AD.
Synaptophysin loss has been extensively associated with AD pathology. Synaptophysin mRNA and protein levels are reduced in the post-mortem AD brain (45,46), and protein levels are also found to be down-regulated in animal model brains (56,57). Synaptophysin was the most significantly downregulated gene in our analysis, establishing a clear correlation between our findings and a typical AD-type pathology. Interestingly, whereas previous studies have investigated synaptophysin as a marker of AD pathology rather than a direct target of A␤Os, we now show that significant reductions in synaptophysin mRNA and protein levels are induced by short-term A␤O treatment (Fig. 5). These results implicate synaptophysin down-regulation as an early abnormality triggered by A␤O toxic signaling and synapse loss.
Synaptophysin plays an important role in intraneuronal vesicle transport, namely in the processes of presynaptic vesicle membrane fusion and neurotransmitter release in the synaptic cleft (58). Interestingly, current results indicated that "SNARE interactions in vesicle transport" was the most significantly overrepresented pathway in A␤O-treated slices (supplemental Table 3). Two additional important SNARE proteins in which gene expression was affected by A␤Os are VAMP and syntaxin (Fig. 3), both involved in synaptic vesicle fusion to the plasma membrane and neurotransmitter release (59). Along with previous reports of AD-associated alterations in neuronal vesicle transport machinery at the mRNA level (22), current data indicate that A␤Os promote a rapid and specific attack on the neuronal vesicle transport system, which likely culminates in synaptic dysfunction.
Increasing evidence supports a recently proposed link between human brain insulin resistance and AD pathology (18, 60 -65). At the cellular level, A␤Os have been shown to induce insulin receptor internalization and the impairment of neuronal insulin signaling in hippocampal cultures (18,19). In line with those studies, we found here that the insulin-signaling pathway is one of the pathways targeted by A␤Os (Fig. 3). Interestingly, among the genes identified in the present study are two key components of insulin signaling, AMP-activated kinase (AMPK) and mTOR (Fig. 3, gene symbols prkag1 and frap1,  respectively). The possible involvement of mTOR in the pathophysiology of AD is still a controversial matter. For example, Spilman et al. (66) show that pharmacological inhibition of mTOR by rapamycin decreases A␤42 levels and rescues the cognitive function in a transgenic mouse model of AD, suggesting that mTOR activity may be increased in A␤-affected tissue. Conversely, Ma et al. (67), also using AD transgenic mice (albeit a different strain from that used by Spilman et al. (66)), found that mTOR signaling was inhibited both in cultured neurons and hippocampal slices from AD-transgenic mice and in wildtype hippocampal slices exposed to exogenous A␤1-42 and that this mTOR dysregulation correlated with impairment in synaptic plasticity. It is thus possible that different genetic backgrounds may lead to distinct responses in terms of mTOR activity in transgenic mice. The present observation of alterations in the expression of mTOR and other proteins involved in insulin signaling in A␤O-treated human tissue may reflect an early pathological response not apparent in transgenic mice (which exhibit exacerbated amyloid pathology) or AD postmortem samples. Taken together, the current results, along with recent and discrepant literature reports, warrant further studies to elucidate the role of mTOR and functionally associated proteins in AD-linked insulin resistance, particularly in the early stages of the disease. In addition, the results suggest that transcriptional dysregulation of genes responsive to insulin signaling may be part of the mechanism by which neuronal insulin resistance is induced by A␤Os.
Ubiquitin-mediated proteolysis, a pathway found to be altered by A␤O treatment (Fig. 3), has also been described as part of AD pathogenesis (22,68). One of the genes grouped in this pathway is cullin 3, which encodes an essential component of the ubiquitin E3 ligase complex. We found a significant down-regulation of cullin 3 in A␤O-treated slices, in accord with results reported using post-mortem AD hippocampal tissue (21). Down-regulation of the ubiquitin system at the protein level in AD has also been reported previously (69), strengthening the possibility that ubiquitin-mediated systems are specific targets of A␤O toxicity.
Cell adhesion molecules are well known players in the regulation of synaptic plasticity, learning, and memory (70) and have been suggested to be involved in A␤-mediated neurotoxicity (71,72). We found that the cell adhesion-related pathways "cell adhesion molecules" and "focal adhesion", as well as "regulation of actin cytoskeleton" (a crucial intracellular component of the cell adhesion machinery), were among the A␤⌷targeted pathways in human slices (Fig. 3). Furthermore, expression of the acting-binding protein CYFIP1, a component of the complex that regulates cytoskeleton remodeling in neurons (73), was up-regulated in A␤O-treated slices from five of the seven donors tested by qPCR (Fig. 4). These results are consistent with the notion that alterations in cell adhesion are involved in synaptic failure in AD.
A␤ has been shown to impede the reversal of long-term depression (LTD) in rat hippocampal slices (74,75). Here we found that A␤Os disrupt the expression of genes associated with LTD (Fig. 3). Facilitation of LTD has been proposed as a mechanism by which A␤Os cause synapse failure in AD (20,76), and it is possible that altered expression of genes involved in LTD may underlie the dysfunctional synaptic plasticity instigated by A␤Os.
In conclusion, we found that short-term exposure to a sublethal concentration of pathologically relevant A␤Os disrupts gene expression in adult human brain slices, causing, in most cases, the down-regulation of genes important for neuronal physiology. Changes in gene expression found in the present study shed light on mechanisms by which A␤Os affect neuronal gene expression and function and may be considered early events in AD pathology and, therefore, potential targets for the development of novel diagnostic or therapeutic approaches in AD.