ERK1/2 Activation Mediates Aβ Oligomer-induced Neurotoxicity via Caspase-3 Activation and Tau Cleavage in Rat Organotypic Hippocampal Slice Cultures*

In this study, we investigated the molecular basis for the altered signal transduction associated with soluble amyloid β-protein (Aβ) oligomer-mediated neurotoxicity in the hippocampus, which is primarily linked to cognitive dysfunction in Alzheimer disease (AD). As measured by media lactate dehydrogenase levels, and staining with propidium iodide, acute exposure to low micromolar concentrations of the Aβ1-42 oligomer significantly induced cell death. This was accompanied by activation of the ERK1/2 signal transduction pathway in rat organotypic hippocampal slices. Notably, this resulted in caspase-3 activation by a process that led to proteolytic cleavage of Tau, which was recently confirmed to occur in AD brains. Tau cleavage likely occurred in the absence of overt synaptic loss, as suggested by the preserved levels of synaptophysin, a presynaptic marker. Moreover, among the pharmacological agents tested to inhibit several kinase cascades, only the ERK inhibitor significantly attenuated Aβ1-42 oligomer-induced toxicity concomitant with the reduction of activation of ERK1/2 and caspase-3 to a lesser extent. Importantly, the caspase-3 inhibitor also decreased Aβ oligomer-induced cell death, with no appreciable effect on the ERK signaling pathway, although such treatment was effective in reducing caspase-3 activation and Tau cleavage. Therefore, these results suggest that local targeting of the ERK1/2 signaling pathway to reduce Tau cleavage, as occurs with the inhibition of caspase-3 activation, may modulate the neurotoxic effects of soluble Aβ oligomer in the hippocampus and provide the rationale for symptomatic treatment of AD.

osition into dense senile plaques, the formation of neurofibrillary tangles composed of hyperphosphorylated Tau, and the loss of neurons and synapses in the affected brain region leading to the progressive loss of cognitive function (1,2). Recent evidence suggests that soluble, prefibrillar, and oligomeric forms of A␤ are acutely toxic (3) and can interfere with synaptic plasticity in the brain, suggesting that this form of the peptide may be responsible for episodic memory deficits, an early symptom of AD, which is linked to hippocampal pathology (4). In fact, soluble A␤ oligomers, also referred as A␤-derived diffusible ligands, strikingly elevated in the AD brain (5)(6)(7)(8) are also described in human amyloid precursor protein transgenic mice AD models (9) and can inhibit the long term potentiation (LTP) of synaptic efficiency (10,11). The pathogenic relevance of this form of A␤ has been substantiated by a newly identified Arctic familial AD mutation, which has an increased propensity to oligomerize (12,13), and by in vitro data demonstrating potent neurotoxicity upon exposure to soluble oligomer or protofibrils (3,14,15). Moreover, the ultrastructural localization of A␤ oligomer within neuritic processes and at synaptic terminals in AD brains (16,17) further supports the hypothesis that soluble A␤ oligomers may play a crucial role as the earliest effector that causes synaptic dysfunction and early memory loss associated with dementia in AD (4,18). Hence, considerable interest is now focused on understanding the neurotoxic mechanisms elicited by soluble A␤ oligomer in order to generate effective therapeutic strategies targeting this molecule for the treatment of AD.
Among the mitogen-activated kinase (MAPK) family, a role for the extracellular signal-regulated kinase (ERK)1/2 signaling pathway, in neuronal death, has been proposed recently in AD (19 -21). Indeed, fibrillar A␤1-42 can induce ERK activation, and sustained activation of the ERK1/2 signaling pathway, mediated by fibrillar A␤, can lead to abnormal phosphorylation of Tau, the generation of dystrophic neurites, and progressive neuronal degeneration (22,23). However, the signaling events associated with neurotoxicity triggered by soluble A␤ oligomer in the AD brain are largely unknown.
On the other hand, a potential role for the proteolytic cleavage of Tau, the microtubule-stabilizing protein, in fibrillar A␤-induced neuronal degeneration, has also been implicated. A previous study showed that truncated Tau was present in AD brains but not in age-matched controls (24). In addition, the cleavage of Tau, by caspases, has been demonstrated both in neurons treated with aggregated A␤ and in AD brains (25)(26)(27)(28). Moreover, these studies provided evidence supporting a key role of truncated Tau either by caspases or calpain in fibrillar A␤-induced neurodegeneration and enhancement of pathological Tau filament assembly in neurons. Interestingly, A␤-mediated Tau cleavage appears to occur as an early event that may precede hyperphosphorylation, in the evolution of the AD tangle pathology (27). These findings imply that aberrant Tau truncation, in addition to Tau hyperphosphorylation, may represent novel cellular mechanisms functionally linking amyloid deposition and neurofibrillary tangles in AD. However, despite such extensive studies of the fibrillar form of A␤, as a key neuropathological factor in AD, there is little known about the precise role of the soluble A␤ oligomer in the modulation of neuronal loss and the molecular basis for the altered signal transduction associated with neurotoxicity in response to this molecule in the hippocampus, a region that is particularly vulnerable to the ravages of AD and that underlies the type of memory deficit detected in early AD (29). Furthermore, few studies have tried to understand how A␤ oligomer affects Tau cleavage and how the A␤ oligomer may modulate this process at the molecular level.
Therefore, this study investigated the intracellular signaling mechanisms triggered by soluble A␤ oligomer, and determined how they correlate with the A␤ oligomer-mediated neurotoxic effect on the hippocampus. By using a rat organotypic hippocampal slice culture, which maintains the cytoarchitecture of the intact brain, direct toxicity of A␤1-42 oligomer was determined by measuring the propidium iodide (PI) uptake and lactate dehydrogenase (LDH) release. We also assessed whether caspase-3 activation mediates A␤1-42 oligomer-induced neuronal cell death in the hippocampus. Finally, the A␤ oligomertriggered signaling cascade, and specifically ERK1/2 activation, was assessed to determine its relation to A␤ oligomer-induced apoptotic cell death. To this end, we analyzed the effect of the A␤ oligomer on the proteolysis of Tau, recently identified as a downstream substrate of caspase-3 (26,27), and we evaluated the modulation of Tau proteolysis by the ERK inhibitor in comparison with the caspase-3 inhibitor. Our data constitute the first evidence that A␤ oligomer-induced toxicity in the hippocampus occurs through the activation of the ERK1/2 signaling cascade. In addition, this study posits a possible link between A␤ oligomer-mediated toxicity and Tau cleavage, which occurs through caspase-3 activation upon stimulation of the ERK1/2 signaling cascade, in AD pathology.
Preparation of Homogeneous Populations of Amyloid Peptide Monomer and Oligomer-A␤1-42 peptide was synthesized as described previously (3). Soluble A␤1-42 monomers and oligomers were prepared as described previously (30). Briefly, 1.0 mg of A␤1-42 peptide was dissolved in 400 l of hexafluoroisopropanol for 10 -20 min at room temperature, and 100 l of the resulting seedless solution was added to 900 l of doubledistilled water in a siliconized Eppendorf tube. After a 10 -20min incubation at room temperature, the samples were centrifuged for 15 min at 14,000 ϫ g, and the supernatant fraction (pH 2.8 -3.5) was transferred to a new tube and subjected to a gentle stream of N 2 for 10 min to evaporate the hexafluoroisopropanol. Vehicle was prepared the same way except with no peptide. The monomer solutions were used immediately after evaporation of hexafluoroisopropanol. For soluble oligomers, the samples were then stirred at 500 rpm using a Teflon-coated micro stir bar for 48 h at 22°C. A␤1-42 oligomer was confirmed by dot blot assay using oligomer-specific A11 antibody (3) and was further verified by Western blotting using 6E10 antibody as described below.
Hippocampal Slice Cultures-All experimental procedures were carried out under protocols approved by the University of California, Irvine, Institutional Animal Care and Use Committee as described previously (31). Briefly, hippocampal slice cultures were prepared from 10-day-old Sprague-Dawley rat pups (Charles River Breeding Laboratories, Inc., Wilmington, MA). Slices were cut at 400 m on a McIlwain tissue chopper, transferred to Millicell (Millipore Corp., Bedford, MA) membrane inserts (0.4 m), and placed in 6-well culture plates. The upper surfaces of the slices were exposed to a humidified 37°C atmosphere containing 5% CO 2 . Slice culture media consisted of basal Eagle's medium with Earle's balanced salt solution, 20% heat-inactivated horse serum, and the following supplements: 20 mM NaCl, 5 mM NaHCO 3 , 1.7 mM MgSO 4 , 0.2 mM CaCl 2 , 26.7 mM HEPES, 26.6 mM L-glutamine, 48 mM L-(ϩ)-glucose, 100 units/ml penicillin, and 100 mg/liter streptomycin, pH 7.25. The medium was changed every other day. Slices were examined periodically for viability, and any dark or abnormal slices were discarded.
Experimental Treatment-On days 10 -11 post-dissection, treatment of the slices was started. All reagents were added to serum-free medium (no horse serum), which was equilibrated at 37°C, 5% CO 2 before their addition to the slices. Monomeric or oligomeric A␤1-42 was added to slice cultures as described previously (31) with some modification. Briefly, slices were pretreated with various pharmacological agents, as described in the text. All concentrations were selected on the basis of the maximal effects of these drugs on their specified targets. Peptide was then added to cultures in serum-free medium at various concentrations up to 2 M as noted in the text. Vehicles were treated the same way except with no peptide. The effect of vehicle alone on cell viability was not detectable. At the indicated times after treatment initiation, slices were rinsed twice in 1ϫ phosphate-buffered saline, then harvested by removing the Millicell membrane insert after freezing samples on dry ice, and processed for immunoblotting as described below.
Assessment of Neuronal Cell Death-To visualize neuronal cell death, hippocampal slices were stained by adding PI into the culture medium at a concentration of 5 g/ml (5 g/ml of media) throughout the A␤ or vehicle treatment, and the degree of hippocampal neuronal death was evaluated by microscopic observation of PI uptake as described previously (32,33). PI is a polar compound that is not toxic to neurons and is impermeable to an intact cell membrane, but it penetrates damaged cell membranes of dying cells and binds to nuclear DNA to generate a bright red fluorescence. PI fluorescence images were captured at different time points after treatment as indicated. The PI images were recorded with an Axiovert 200 inverted microscope (Carl Zeiss Light Microscopy, Göttingen, Germany) with an Axio-Cam (Zeiss) digital camera controlled by AxioVision program (Zeiss). Images were analyzed with the KS 300 analysis program (Zeiss) and neuronal cell death, in the different cultures, was expressed as an arbitrary unit of PI uptake.
Lactate Dehydrogenase Activity-LDH enzymatic activity in the culture medium was used to evaluate the extent of cellular damage produced in cultured slices subjected to the different treatments as described previously (34). Briefly, culture medium was collected after incubation of slices, as indicated in the figures, and from the vehicle-treated control cultures. 100-l aliquots of culture media were taken for determination of LDH activities using Tox-7 (Sigma) according to the manufacturer's directions. The activity was expressed as the relative percentage of neuronal death using respective values for vehicle-or A␤1-42-treated slice cultures as 100%.
Dot Blot Assay-The monomeric and oligomeric A␤ peptides were prepared as described above and applied to nitrocellulose membrane. Dot blot assay was performed as described previously (3) with minor modifications. Briefly, the membrane was blocked with 3% nonfat milk in TBS-T (137 mM NaCl, 20 mM Tris base, and 0.1% Tween 20; pH 7.4) at room temperature for 1 h. The membrane was washed with TBS-T and probed with anti-oligomer A11 antibody (1:5,000) or 6E10 (1: 5,000) in 1% nonfat milk for 1 h at room temperature. After washing, it was probed with anti-rabbit horseradish peroxidase-or antimouse horseradish peroxidase-conjugated antibody solution (1:3,000; Jackson ImmunoResearch, West Grove, PA) for 1 h at room temperature. The blots were developed by using ECL chemiluminescence system (Amersham Biosciences).
Caspase-3 Activity Assay-Caspase-3 activity was measured using the fluorometric caspase-3 activity assay kit (Calbiochem) according to the manufacturer's instructions. The fluorescence was measured after cleavage of the caspase-3 substrate (DEVD) labeled with a fluorescent molecule, 7-amino-4-trifluoromethyl coumarin (AFC), to AFC by caspase-3. Briefly, hippocampal slices, treated with 2 M A␤1-42 oligomer as described above, were harvested in extraction buffer and incubated on ice for 20 min. After centrifugation at 16,000 ϫ g for 10 min, the supernatant was incubated with the caspase-3 substrate (DEVD-AFC) for 1 h at 37°C. The fluorescence was assessed using a fluorescent plate reader with a 400 nm excitation and a 505 nm emission according to the manufacturer's protocol. Caspase-3 activity is expressed as a relative percentage of vehicle-or A␤1-42-treated samples as 100%.
Tau in Vitro Cleavage Assay-For experiments using purified caspase-3 (Sigma), total lysates were prepared from hippocampal slices and then centrifuged for 10 min at 16,000 ϫ g. The supernatants (20 l) were then incubated with recombinant caspase-3 (Sigma) for 1 h at 37°C, as described previously (35). The reactions were terminated by adding 1 volume of 2ϫ SDS sample buffer and boiled for 5 min. Samples were subjected to Western blot analysis using Tau-5 antibody that recognizes total Tau.
Protease Inhibitor Treatment-To inhibit caspase activation, the caspase inhibitors, Z-VAD-fmk or Ac-DEVD-CHO were added to the medium of hippocampal slice cultures 1 h prior to and for the duration of the A␤ treatment. ⌻he supernatants, prepared as described above, were subjected to Western blot analysis using Tau antibodies.
Statistical Analysis-Data were expressed as the means Ϯ S.E. of the values from the number of experiments indicated in the corresponding figures. Differences between groups were examined for statistical significance using one-way analysis of variance with an unpaired Student's t test. A p value less than 0.05 denoted the presence of a statistically significant difference.

Treatment of Cultured Hippocampal Slices with A␤1-42 Oligomer Increases PI Uptake and LDH Release-Homogeneous
populations of A␤ monomer and prefibrillar A␤ oligomers were prepared as described above. As shown in Fig. 1A, the oligomeric A␤ preparation was confirmed by dot blot assay using a conformationally directed antibody against amyloid oligomers (larger than tetramer), A11, which is highly specific for ERK1/2 Activation Mediates A␤ Oligomer-induced Neurotoxicity toxic oligomer subspecies as characterized (3,36). The monomeric preparation contained a small amount of detectable A␤ oligomers. This A␤ oligomer preparation was shown to have an approximate molecular mass of ϳ90 kDa by size exclusion chromatography, contained very little material of lower molecular mass, and consisted of spherical vesicles with diameters of ϳ3-5 nm (30). Furthermore, Western blot analysis revealed that this A␤ oligomer was detected as a stable, SDS-resistant band, corresponding to an ϳ24-mer, in the culture medium from A␤ oligomer-treated slices after 24 h of incubation (Fig. 1B).
In order to investigate the direct toxic effect of the A␤1-42 oligomer on neuronal viability, in the organotypic hippocampal slices, we first examined the degree of cell death by measuring the PI uptake and LDH release. PI uptake has been confirmed as an accurate quantitative measure of neuronal death in hippocampal slice cultures with other measures of neuronal death, including release of LDH (32,33). The organotypic hippocampal slice culture is a good model system to study the mechanisms of neurodegeneration. This is because several features of hippocampal circuitry, including the maturation of synapses, receptors, and intrinsic fiber pathways, are preserved in vivo, and the preparation is well suited for prolonged pharmacological treatment and recovery, which would be difficult to perform in an intact animal system (37,38). As shown in Fig. 2, the amount of LDH in the media containing hippocampal slices treated with the A␤1-42 oligomer was significantly increased in a dose-and time-dependent manner above control levels, indicating that cell death had occurred. Exposure to 1 M A␤1-42 oligomer for 24 h increased the percentage of LDH release to approximately 3.5-fold over the vehicle-treated controls ( Fig. 2A), and there was approximately a 6-fold increase in the degree of cell death when the concentration of A␤1-42 oligomer was increased to 2 M (Fig. 2B). Consistent with the LDH results, A␤1-42 oligomer caused a time-dependent increase in PI uptake in the hippocampus, which was first detectable within 10 h and continued to increase at 24 h (Fig. 3, B and C). Exposure to 0.5 M A␤1-42 oligomer for 24 h induced PI uptake, which was evident in the hippocampal CA1, a region important for memory formation but not seen in vehicle-treated slices. The degree of PI uptake clearly increased, and the area of staining extended to other hippocampal regions when the concentration of A␤1-42 oligomer was increased to 2 M (Fig. 3, A and C). In parallel, slices were also exposed to the A␤1-42 monomer under the same experimental conditions, but there was no observed increase in cell death with this peptide as measured by either PI uptake or LDH release ( Fig. 2 and  Fig. 3). These results suggest that the A␤1-42 oligomer had a greater neurotoxicity than the A␤1-42 monomer in organotypic hippocampal slices. These findings are consistent with other recent studies (14,39).
A␤1-42 Oligomer-induced Cell Death Was Accompanied by Caspase-3 Activation-To further clarify the mechanism by which A␤1-42 oligomer causes cell death in the hippocampal slices, we evaluated the activation of caspase-3, a key executor in apoptosis, by using the caspase activation assay. The results indicate that A␤1-42 oligomer induced caspase-3 activity, whereas A␤1-42 monomer had no appreciable effect (Fig. 4A).  This activation pattern of caspase-3 is closely correlated with the pattern of cell death obtained by both the LDH assay and the PI assay presented above (Fig. 2 and Fig. 3). In addition, the potential role of caspase-3 in the A␤1-42 oligomer-induced toxicity was also confirmed by the caspase inhibitors Z-VAD-fmk and Ac-DEVD-CHO. Pretreatment of Z-VAD-fmk, which is relatively specific for caspase-1, -3, and -4, or Ac-DEVD-CHO, a specific caspase-3 inhibitor, almost completely prevented the A␤1-42 oligomer-induced caspase-3 activation compared with A␤1-42 oligomer-treated cultures (Fig. 4B). Furthermore, these caspase-3 inhibitors, partially but significantly, reduced A␤1-42 oligomerinduced cell death as evaluated by LDH release (Fig. 4C). These data indicate that A␤1-42 oligomer-induced cell death was accompanied by activation of caspase-3 and that the cell death seen in the hippocampal slices, treated with relatively low concentrations of A␤1-42 oligomer, appeared, at least in part, to be due to apoptosis. In addition, partial protection from A␤1-42 oligomer-induced toxicity, observed with the caspase inhibitors, suggests that multiple mechanisms, including necrosis as well as apoptosis, may be involved in the hippocampal toxicity elicited by the A␤1-42 oligomer.
U0126 Partially but Significantly Reduced the Hippocampal Toxicity in Response to the A␤1-42 Oligomer-To unravel the signaling pathway, upstream of caspase-3 activation, leading to cell death observed upon exposure of the A␤1-42 oligomer to the hippocampal slices, we investigated the results of blocking various signaling pathways with pharmacological agents. By using inhibitors of certain kinase cascades to test their involvement in hippocampal signaling, we first determined their effects on cell death in response to oligomeric A␤1-42 by using the LDH release assay. Among the MAPK inhibitors tested, U0126, known to selectively block the activity of MAPK kinase kinase (MEK1/2), activator of ERKs (ERK1/2), partially but significantly decreased A␤1-42 oligomer-mediated toxicity (Fig. 5). However, SB202580, a specific inhibitor of p38 MAPK, and SP600125, an inhibitor of Jun NH 2 -terminal  JULY 21, 2006 • VOLUME 281 • NUMBER 29 JOURNAL OF BIOLOGICAL CHEMISTRY 20319 kinase (JNK), failed to reduce the cell death because of A␤1-42 oligomer treatment. Likewise, SB216763, an inhibitor of GSK-3␤, and LY294002, an inhibitor of phosphatidylinositol 3-kinase, an activator of Akt, showed only minimal effects. Based on the evidence presented in previous papers (40 -42) where these inhibitors were used, the concentrations employed in the current experiment should have been enough to inhibit their respective kinases without significantly affecting cell viability.

A␤1-42 Oligomer Consistently Induced ERK1/2 Activation in Hippocampal Slices Undergoing Apoptotic Cell Death-To
further delineate the potential involvement of ERK signaling in A␤1-42 oligomer-induced toxicity in hippocampal slices, we analyzed whether ERK signal transduction is activated. Consistent with the pharmacological study presented above, hippocampal slice cultures treated acutely with A␤1-42 oligomer resulted in the sustained activation of ERK1/2 in a dose-and time-dependent manner as compared with the vehicle-treated control (Fig. 6, A and B). As expected, U0126 consistently suppressed A␤1-42 oligomer-mediated activation of ERK1/2 (Fig.  7A). U0126 also significantly down-regulated caspase-3 activity induced by A␤1-42 oligomer; this was concomitant with decreased ERK1/2 activation (Fig. 7C). By contrast, Ac-DEVD-CHO, which efficiently blocked A␤1-42 oligomer-induced activation of caspase-3 as presented above (Fig. 4B), did not exhibit the down-regulation of A␤1-42 oligomer-induced activation of ERK1/2 (Fig. 7B). These results are consistent with the idea that A␤1-42 oligomer-induced apoptotic cell death, in the hippocampus, is in part mediated through the activation of the ERK signaling pathway, which in turn activates caspase-3.

A␤1-42 Oligomer-induced Tau Cleavage Is Consistent with Caspase-3 Activation-Recent
studies provide direct evidence supporting a key role for cleaved Tau in the mechanisms leading to fibrillar A␤-induced neurodegeneration and tangle formation in AD (25)(26)(27). Therefore, we reasoned that Tau would be a downstream target of caspase-3, which is activated by ERK1/2 signal transduction in the hippocampus in response to oligomeric A␤1-42. To test this premise, hippocampal slices were treated with the A␤1-42 oligomer, and the level of Tau was monitored. Analysis with Tau-5 antibody that recognizes total Tau revealed significant breakdown of this protein in a dose-and time-dependent manner (Fig. 8, A and B). The intact Tau appears in all lanes as 69-to 55-kDa bands. However, levels of Tau breakdown products were increased significantly and progressively elevated during the time course of A␤1-42 oligomer treatment. This pattern of Tau cleavage corresponded to that of caspase-3 activity, which is induced by the ERK1/2 signaling pathway as described (Fig. 4 and Fig. 6). By contrast,

ERK1/2 Activation Mediates A␤ Oligomer-induced Neurotoxicity
synaptophysin levels, which correlate closely with synapse number and hence are commonly used to assay for loss of synapses (35,43,44), were not significantly altered, and cleavage of synaptophysin was not clearly seen under the same experimental conditions. To compare A␤1-42 oligomer-mediated Tau cleavage pattern in hippocampal slices to that generated by caspase-3, we subjected the total lysate of hippocampal slices to in vitro caspase-3 digestion. Indeed, in vitro caspase-3 digestion prominently generated the Tau cleavage pattern observed in the hippocampal slices exposed to A␤1-42 oligomer, whereas caspase-3 failed to cleave synaptophysin (Fig. 8C), which was truncated by calpain-1 generating a breakdown product with a mass of ϳ30 kDa. 3 Finally, in order to further confirm whether caspase-3 activation triggered by A␤1-42 oligomer treatment could induce the proteolysis of Tau in the hippocampal slices, we treated them with cell-permeable inhibitors of caspase-3 or calpain-1 prior to A␤1-42 oligomer incubation. The caspase-3 inhibitor Ac-DEVD-CHO significantly prevented Tau cleavage induced by the A␤1-42 oligomer, whereas the calpain-1 inhibitor, calpeptin, did not block Tau cleavage (Fig. 8D). These results imply that the A␤1-42 oligomer can induce proteolytic cleavage of Tau, a downstream target for the caspase-3 that is activated, at least in part, by the ERK signaling pathway in hippocampal slices undergoing A␤1-42 oligomer-induced apoptotic cell death. Our results also suggest that synaptophysin is not a primary target for caspase-3.

DISCUSSION
The results presented in this paper verify a critical role for the soluble A␤ oligomer in inducing neurotoxicity in rat organo-typic hippocampal slice cultures. This A␤-induced neurotoxicity may be mediated, at least in part, through activation of the ERK1/2 signal transduction pathway, which results in caspase-3 activation leading to proteolytic cleavage of Tau. Several lines of observation support this conclusion. First, A␤1-42 oligomer conferred a potent neurotoxicity in the hippocampus, which was accompanied by increased, sustained activation of the ERK1/2 pathway leading to caspase-3 activation in a dose-and time-dependent manner. Second, among the pharmacological agents used to inhibit several kinase cascades, only MEK-ERK1/2 inhibitor significantly attenuated A␤1-42 oligomer-induced toxicity concomitant with the reduction of activation of ERK1/2 and caspase-3 to a lesser extent. Third, cleavage of Tau was detected in the hippocampal slice cultures exposed to the A␤1-42 oligomer that causes caspase-3 activation. Finally, the caspase-3 inhibitors, which were effective in reducing both caspase-3 activation and Tau cleavage, significantly attenuated A␤1-42 oligomer-induced cell death without appreciable effect on the ERK1/2 signaling pathway. Therefore, these data imply that one mechanism, by which the soluble A␤ oligomer promotes neurotoxicity, at least in the hippocampus, is via a sustained increase in activation of ERK1/2 and caspase-3 and by cleavage of Tau. Remarkably, even at the relatively low concentrations of soluble A␤ oligomer, selective region-specific toxicity was evident in the hippocampal CA1, a division important for cognition and memory, which is in good agreement with the selective neurodegeneration distinctly manifest in AD (45).
Recent studies clearly demonstrated that specific oligomeric forms of A␤, including the ϳ12-mer as well as ϳ24-mer, are highly associated with cognitive decline in HuAPPsw (Tg2576) and 3xTg-AD mice models (36,46,62). Trimers and hexamers were excluded as components of toxic A␤ oligomer because they were present before memory impairment (36). In our study, we demonstrated that ϳ24-mer along with lesser quantity of ϳ12-mer of A␤ oligomers were detected as stable, SDSresistant species in the culture medium from A␤1-42 oligomer-treated slices even after 24 h of incubation. These findings together provide collective evidence that these high molecular weight A␤ oligomers may represent the primary pathogenic structure responsible for mediating A␤ toxicity in AD. This idea is further strengthened by the recent reports showing that intracerebroventricular injection of oligomers inhibits LTP (10,11) and specifically disrupts cognitive function (47) and that the concomitant injection of the anti-A␤ antibodies 6E10 or A11 with A␤ oligomers neutralizes the oligomer-induced LTP dysfunction and A␤ pathology (48, 62). The presence of ϳ24-mer along with a lower mass of oligomer species in the culture media as shown in this study may indicate the possibility of dissociation of toxic oligomers in some situations. Alternatively, some of the A␤ oligomers in our preparation are not yet SDS-resistant. Thus, further biochemical experiments will be necessary to better address these conformational changes during incubation time. Importantly, soluble A␤ oligomers have been isolated from post-mortem brains of AD patients (5)(6)(7)(8) and are particularly found within the neuritic processes and at synaptic terminals in AD brains (16,17), strongly supporting in vivo the relevance of this study.
Our data demonstrating that A␤1-42 oligomers elicited greater neurotoxicity than the A␤1-42 monomer, in the organotypic hippocampal slices, are consistent with the recent studies demonstrating in vitro that the oligomeric A␤ conformation is ϳ10-fold more neurotoxic than the plaque-forming fibrillar A␤ assembly and ϳ40-fold more toxic than the monomeric A␤ (14,39). Given that A␤1-42 oligomer-induced cell death was accompanied by the activation of caspase-3, the cell death seen in the hippocampal slices treated with A␤1-42 oligomer appeared, at least in part, to be due to apoptosis. This is consistent with earlier studies showing evidence of apoptosis in AD brains (49) and apoptotic cascades triggered by A␤ in a neuronal cell culture model (50). However, almost complete inhibition of caspase-3 activity by inclusion of Z-VAD-fmk or Ac-DEVD-CHO, the caspase-3 inhibitors, led to partial protection from oligomeric A␤1-42-induced toxicity. This strongly suggests that there are multiple mechanisms underlying A␤1-42 oligomer toxicity in the hippocampus. Several studies have reported that both apoptosis and necrosis may occur in parallel during neurodegeneration (51). In addition, biphasic cell death, beginning with acute necrosis followed by a delayed apoptosis, the hybrid forms of neuronal death, and/or switch from apoptosis to necrosis strongly support this view (52,53). Further experiments are necessary to clearly establish the connection between apoptosis and necrosis induced by the A␤ oligomer in the hippocampus.
This study constitutes the first report demonstrating an intracellular pathway by which the A␤1-42 oligomer triggers ERK1/2 activation, caspase activation, Tau cleavage, and apoptotic cell death in the hippocampus, suggesting a key role for this molecule in AD pathology. The fact that the MEK-ERK1/2 inhibitor U0126 precluded ERK1/2 activation, significantly ameliorating A␤1-42 oligomer-induced cell death, indicates that this event is initially triggered by MEK-ERK1/2. Importantly, our finding that suppression of caspase-3 activation by U0126 was only partial indicates the involvement of multiple signaling events in A␤1-42 oligomer-induced apoptosis. Few studies have explored the signaling mechanism underlying A␤ oligomer-mediated neurotoxicity, although there is one study that provides evidence for the involvement of the protein kinase C signaling pathway (54). On the other hand, a recent study implies that A␤ oligomer-evoked inhibition of LTP is mediated via activation of the kinases JNK, Cdk5, and p38 MAPK in hippocampal slices (11). However, our study demonstrating that A␤ oligomer-mediated toxicity in the hippocampus was only reduced by U0126, not by other protein kinase inhibitors, suggests a lack of direct involvement of JNK, p38 MAPK, or GSK-3␤. This diversity of signaling pathways triggered by the soluble A␤ oligomer might be due to the physical state of the peptides and the time of exposure to the stimulus, as suggested by others (55), or possibly from the culture conditions and treatment schemes. Indeed, our results may be only relevant to the high molecular weight A␤ oligomers such as ϳ24-mer present in our oligomer preparation because A␤-derived diffusible ligands or low molecular weight oligomers did not appear to cause neurotoxicity through the same mechanism (11,54). Nevertheless, our findings are consistent with a recent study that demonstrated activation of the ERK signaling pathway following acute A␤1-42 oligomer treatment of cultured hippocampal slices, although the resulting toxicity was not examined (55). Moreover, the potential role of sustained activation of the ERK signaling pathway in A␤ oligomer-mediated toxicity in the hippocampus, as seen in this study, is strongly supported by high levels of activated ERK1/2 detected around amyloid deposits in  Fig. 6. The lower panels correspond to the quantification of ERK1/2 activation normalized to the total amount of ERK1/2 of A and B, respectively. C, the effect of U0126 on A␤1-42 oligomer-induced activation of caspase-3. Slices were treated with U0126 at the concentrations indicated for 1 h before the incubation with A␤1-42 oligomer for 24 h. Caspase-3 activity assays were performed as described in Fig.  4. Values are expressed as a percentage of the A␤1-42 oligomer-treated slices. Each number represents the mean Ϯ S.E. from two to three different experiments. *, p Ͻ 0.05, different from the A␤1-42 oligomer-treated samples.
AD brains (20), as well as the sustained activation of ERK in hippocampal neurons in response to fibrillar A␤-induced neurotoxicity (23,25,56). In addition, our observation that the changes in ERK1 phosphorylation, indicative of kinase activation, were not as robust as ERK2 closely corresponds to earlier reports that ERK2 is implicated as a critical factor for hippocampal synaptic plasticity or learning and memory (57,58).
The present study provides evidence that the A␤1-42 oligomer can induce cell toxicity via proteolytic cleavage of Tau, a downstream target for the caspase-3 that is activated by the ERK signaling pathway. Indeed, two proteases proposed to play a role in the pathogenesis of AD are caspase-3 and calpain-1 (26,27,59), which are families of cysteine proteases that have important roles in the initiation, regulation, and execution of cell death (60,61). Our finding that caspase-3, rather than calpain-1, plays a primary role in Tau cleavage, which is linked to the toxic action of A␤1-42 oligomer in the hippocampus, is consistent with previous reports that Tau, the dendritic and axonal constituent, is considered the preferred substrate for caspases leading to the generation of dystrophic neurites in response to fibrillar A␤ treatments (25)(26)(27)(28)59). In rat primary cortical neurons, Tau cleavage was observed by 10 M fibrillar A␤ (26), which is a 10-fold higher dose than A␤ oligomer used in this study. In addition, caspase-3-cleaved Tau was recently confirmed to occur in AD brains as an early event, and thus is likely to be in part associated with the development of tangle pathology within AD brains (26,27), and could be a marker for hippocampal neuronal cell death.
Several studies have demonstrated that fibrillar A␤ peptides promote pathological Tau filament assembly in neurons by triggering caspase cleavage of Tau, and generating a proteolytic product that assembles more rapidly and more extensively into Tau filaments in vitro than wild-type Tau (25)(26)(27)(28). This supports the hypothesis that amyloid deposition, neurofibrillary tangles, and caspase activation could share a common pathway. Furthermore, in AD brains, activated ERK1/2 and phosphorylated Tau colocalize with amyloid-rich areas (20). A recent study has demonstrated that a single intrahippocampal injection of a specific oligomeric antibody efficiently cleared A␤ pathology and, more importantly, Tau pathology in an in vivo model of AD (62). This finding along with our results strongly implicates a link between A␤ oligomerization and Tau pathology. Accordingly, the interference of A␤ oligomerization might be a valid therapeutic target for AD treatment (48,62,63). In addition, our finding that hippocampal slice cultures treated acutely with  . A schematic of the proposed role of A␤1-42 oligomer in hippocampal toxicity by the ERK1/2 signaling pathway leading to caspase-3-activation and Tau cleavage. This scheme presents a molecular mechanism by which low micromolar concentrations of A␤1-42 oligomer may cause a significant increase in the activation of the ERK signaling pathway, which in turn activates caspase-3 in the hippocampus. In the experimental paradigm employed here, the role of Ca 2ϩ as an upstream effector responsible for activation of ERK in the machinery of apoptotic cell death remains unsettled despite our previous knowledge that dysregulation of Ca 2ϩ ions, through cytotoxic pores created by A␤ oligomer, might play a crucial role in A␤ oligomer-mediated toxicity (3,30). Tau truncation by active caspase-3 might be associated with A␤1-42 oligomer-induced cell death in the hippocampus by leading to cytoskeletal damage and neuronal/synaptic dysfunction, which is associated with memory deficits as seen in early AD patients. JULY 21, 2006 • VOLUME 281 • NUMBER 29 A␤ oligomer did not exhibit synaptophysin cleavage was clearly confirmed by the lack of synaptophysin proteolysis by caspase-3 in vitro. These observations suggest that Tau cleavage might occur independently from synapse loss in the hippocampus; this might precede proteolytic cleavage of synaptophysin as described recently (35,64).

ERK1/2 Activation Mediates A␤ Oligomer-induced Neurotoxicity
In conclusion, we propose that an apoptotic route, in hippocampal slices after A␤1-42 oligomer treatment, might involve Tau cleavage, possibly leading to cell structure deregulation, upon caspase-3 activation through the ERK signaling pathway as summarized in Fig. 9. The data generated from this study will help elucidate how soluble A␤ oligomer affects the signal pathway in the hippocampus, possibly shedding light on how elevated A␤ oligomer causes hippocampal dysfunction leading to the episodic memory deficits associated with early AD. Thus, control of ERK signaling or use of specific caspase proteolytic inhibitors might be useful in preventing, at least partially, A␤ oligomer-induced neurotoxicity in the hippocampus.