Peroxisomal Proliferation Protects from β-Amyloid Neurodegeneration*

Alzheimer disease is a neurodegenerative process that leads to severe cognitive impairment as a consequence of selective death of neuronal populations. The molecular pathogenesis of Alzheimer disease involves the participation of the β-amyloid peptide (Aβ) and oxidative stress. We report here that peroxisomal proliferation attenuated Aβ-dependent toxicity in hippocampal neurons. Pretreatment with Wy-14.463 (Wy), a peroxisome proliferator, prevent the neuronal cell death and neuritic network loss induced by the Aβ peptide. Moreover, the hippocampal neurons treated with this compound, showed an increase in the number of peroxisomes, with a concomitant increase in catalase activity. Additionally, we evaluate the Wy protective effect on β-catenin levels, production of intracellular reactive oxygen species, cytoplasmic calcium uptake, and mitochondrial potential in hippocampal neurons exposed to H2 O2 and Aβ peptide. Results show that the peroxisomal proliferation prevents β-catenin degradation, reactive oxygen species production, cytoplasmic calcium increase, and changes in mitochondrial viability. Our data suggest, for the first time, a direct link between peroxisomal proliferation and neuroprotection from Aβ-induced degenerative changes.

Peroxisomes are subcellular organelles found in most animal cells that perform diverse metabolic functions, including detoxification of reactive oxygen species (ROS) 2 through their matrix enzyme catalase (1)(2)(3) and regulation of the oxidative balance and fatty acid oxidation (4 -6). Peroxisomes are present in the cell bodies, dendrites, and presynaptic axon terminals of neuronal cells (7,8) as well as in growing neurites (9). Tau overexpression inhibits kinesin-dependent transport of peroxisomes, neurofilaments, and Golgi-derived vesicles into neurites (10), and it has been suggested that a loss of peroxisomes apparently makes neurons more vulnerable to oxidative stress (10). Peroxisome proliferators (PPs) are a class of structurally dissimilar industrial and pharmaceutical chemicals that were originally identified as inducers of peroxisome proliferation in rat and mouse hepatocytes (11,12). Several PPs have shown to bind to peroxisome proliferator-activated receptors (PPARs), these include Wy-14.643 (Wy), which binds with great affinity to PPAR␣ and induces a strong activation of this receptor (11). 4-Phenyl butyric (4-PB) is a PP that, in contrast to other PPs, is able to induce human peroxisome proliferation (13); however, the mechanism of peroxisome proliferation remains to be elucidated. According to Liu et al. (14), 4-PB activates PPARs in astrocytes; nevertheless, they suggested that peroxisome proliferation may be independent of PPAR␣ activation.
Alzheimer disease (AD) is characterized by a progressive neurodegeneration associated with extracellular deposits of amyloid ␤-peptide (A␤) in the form of senile plaques (15,16). A␤ peptide acquires neurotoxic properties when it forms homo-oligomeric species (17) or heterooligomeric species with molecules associated with mature senile plaques and/or A␤ deposits in cerebral blood vessels (18), such as the enzyme acetylcholinesterase (AChE), a protein that forms stable A␤⅐AChE complexes (19), which are more neurotoxic than A␤ alone (20).
There is evidence relating the etiopathology of AD with free radicals in AD brain patients (21)(22)(23)(24) as well as in in vitro experiments, which suggest that A␤ neurotoxicity operates by an oxidative mechanism (25)(26)(27). Evidence suggests that H 2 O 2 at a high concentration produces toxic reactions; however, at a moderate concentration, H 2 O 2 may functions as a second messenger mediating a variety of signal transduction pathways (28), including tyrosine phosphatase 1B, Jun N-terminal kinase, protein kinase C, and transcription factors like AP-1 and NF-B (29). Several studies have also suggested that H 2 O 2 mediates A␤-cellular toxicity and increases the production of both A␤ and the amyloid precursor protein; hence, the redox interaction between A␤, amyloid precursor protein, and metals may be a pathological positive feedback system between A␤ amyloidosis and oxidative stress (30,31). Moreover, several studies report neuroprotection by antioxidants against A␤-mediated cytotoxicity (32)(33)(34)(35).
In the present work, we have studied the possible neuroprotective properties of peroxisomes, when these organelles were induced to proliferate by specific PPs, in rat hippocampal neurons. Catalasespecific activity, bulk-rise calcium, and ␤-catenin levels, a key protein in the Wnt signaling pathway, were evaluated on hippocampal neurons pretreated with PPs and then challenged with A␤ in order to find possible neuroprotective effects of peroxisomes in these cells. We report here that a PPAR␣ agonist induces the peroxisomal proliferation and attenuated A␤-dependent neurotoxicity, decreases the intraneuronal ROS production, protects from ␤-catenin degradation, and prevents the cytoplasmic calcium influx induced by H 2 O 2 and the A␤ peptide. These results suggest that the activation of PPAR␣ prevented A␤-dependent neurotoxicity by a mechanism that involves the induction of the peroxisome proliferation. Moreover, PPs prevents the neurotoxic changes induced by oxidative stress, an event that is involved in the A␤ neuropathologic effects.

EXPERIMENTAL PROCEDURES
Reagents-Synthetic A␤-  peptide corresponding to the human A␤ wild type sequence was obtained from Chiron Corp. Inc. (Emeryville, CA) and Calbiochem (Postfach, Germany).
Primary Rat Hippocampal Neuron Cultures-Hippocampal neurons were obtained from 18-day-old Sprague-Dawley rat embryos (36). Hippocampi were aseptically dissected and trypsinized for 20 min. After centrifugations for 1 min, cells were seeded in phenol red-free Dulbecco's modified Eagle's medium plus 10% horse serum into 1% poly-Llysine-coated plates. After 120 min, medium was removed, and neurobasal medium was added containing 1% B27 supplement from Invitrogen, plus streptomycin and penicillin. On day 3 of culture, hippocampal neurons were treated with 2 M 1-␤-D-arabinofuranosylcytosine (AraC) for 24 h to reduce the number of proliferating nonneuronal cells (36). On day 3-4, neurons were incubated with Wy or 4-PB. Neurons of 5-day-old cultures were used for various experiments; the average number of neurons in each experiment was around 95% of the total cells present in the cultures.
A␤ Fibril Formation-The amyloid fibrils were obtained as previously described (37). Specifically, stock solutions were prepared by dissolving freeze-dried aliquots of A␤-  in Me 2 SO. Peptide stock aliquots were diluted in 0.1 M Tris-HCl (pH 7.4) to a final concentration of 100 M A␤. The solutions were stirred continuously (210 rpm) at room temperature for 48 h. The quality of amyloid was verified by different criteria, such as turbidometry, Congo red binding, and thioflavin T spectrofluorometry (19,(37)(38). Aliquots of the homogenate were transferred to a denaturating buffer and subjected to Tris-Tricine gels to quantify the amounts of A␤ peptide contained in the fibrils.
ThT-based Fluorometric Assay-Aliquots of A␤ peptide at the indicated concentrations were incubated at different times in phosphatebuffered saline, pH 7.2, at room temperature. To quantify the amyloid formation, the ThT fluorescence method was used (39). ThT binds specifically to amyloid, and such binding produces a shift in its emission spectrum and an increase in the fluorescent signal, which is proportional to the amount of amyloid formed (19,38). Following incubation, A␤ alone in 50 mM sodium phosphate buffer, pH 6.0, and 0.1 mM ThT in a final volume of 2 ml were added. Fluorescence was monitored at excitation of 450 nm and emission of 485 nm using a Shimadzu spectrofluorometer, as described previously by Inestrosa et al. (19).
Electron Microscopy-Fresh aliquots of samples were diluted 1:3 in water, and 5 l were placed on Parlodion/carbon-coated 300-mesh copper grids for 1 min. Excess sample was removed, and 15 l of 2% aqueous uranyl acetate was placed onto the grid for 30 s, followed by the removal of excess staining solution with filter paper and air drying. Observations were carried out using a Phillips Tecnai 12 electron microscope, as described previously (19). The quantification of amyloid aggregates and fibrils was made using Sigma Scan Pro software (19,38).
Neuronal Viability (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) Reduction)-MTT assays were performed as previously described (40). Hippocampal neurons (1.1 ϫ 10 5 cells/100-l well) were assayed in serum-and phenol red-free medium. Cells were previously incubated with or without Wy, a rodent peroxisome proliferator, for 48 h at 37°C, and after this incubation, the neurons were treated with A␤ or with A␤ plus 3-AT for 30 min before the addition of A␤. Cell viability was measured by the MTT method. Following the addition of MTT, the reaction was ended using MTT solubilization solution (1% Triton X-100 in acidic isopropyl alcohol) (0.1 N HCl). MTT reduction was determined in an ELISA spectrophotometer METER-TECH E960 at 540 and 650 nm.
Immunofluorescence Studies-Hippocampal neurons were plated on polylysine-coated coverslips (40,000 neurons/cover). After 5 days in Neurobasal/B27 medium, neurons were washed and exposed to A␤ in the presence of the PPAR␣ agonist. Neurons were fixed with 4% paraformaldehyde and permeabilized with 90% methanol. Immunostaining was done with polyclonal antiperoxisomal membrane proteins (41) and secondary antibody labeled with fluorescein isothiocyanate (Affinity Bio Reagents Inc., Golden, CO), as described by Santos et al. (42,43). The number of peroxisomes was quantified following the procedure described by Wei et al. (13), and the number of neurites was evaluated using Image-Pro Plus software (Media Cybernetic, Silver Spring, MD). To detect ␤-catenin, an anti-␤-catenin antibody (1:200) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), was used.
Reverse Transcription-PCR of PPAR␣-Neurons were harvested, and the total cellular RNA was extracted with TRIZOL TM reagent (Invitrogen) and reverse-transcribed using poly(dT) primers and the Superscript II enzyme. Amplification by PCR was performed using the following primers for PPAR␣: 5Ј-GTG CCT GTC CGT CGG GAT GT-3Ј and 5Ј-GTG AGC TCG GTG ACG GTC TC-3Ј (fragment size 364 bp). As a semiquantitative control Quantum RNA TM ␤-actin from Ambion was used. The reaction volume was 25 l, and the products were visualized in 1.5% agarose/TAE gels and stained with ethidium bromide. PCR products were analyzed by electrophoresis on 1% agarose gel poured in TAE buffer. The band intensity for PPAR␣ was measured with the software MATRIX and normalized with ␤-actin band intensity.
Catalase Activity-Hippocampal neurons were seeded at 2 ϫ 10 5 cells/well and treated with the PPs as indicated before. Neurons were washed with cold phosphate-buffered saline, scraped, and sonicated. Catalase activity was measured using a specific substrate (42,44).
ROS Measurements-Hippocampal neurons were incubated with the fluorescent probe 2,7-dichlorofluorescein at 10 M for 10 min (45). Then wells were washed three times with phosphate-buffered saline solution, and 10 M A␤ was added to the wells and incubated for 4 h. Cells were photographed in a fluorescent microscope. Photographs were scanned and analyzed by the SCION Image Program (SCION Co.). Results in arbitrary units were expressed as a percentage of the fluorescence signal of untreated cell (control) set at 100%.
Calcium Measurement-Neuronal calcium changes were measured by confocal microscopy. Neurons were grown on poly-L-lysine-coated glass coverslips and cultured for 5 days. The cells were then loaded for 30 min with 5 M Fluo3-AM in Krebs-Ringer-Hepes glucose containing 0.02% pluronic acid and then washed and allowed to equilibrate for 30 min. Coverslips were then mounted in a chamber on the stage of a confocal laser-scanning microscope (model Pascal; Carl Zeiss Ltd.). The fluorescence changes determined by Fluo3-AM represent the cytoplasmatic calcium changes (46). Images were acquired using a 488-nm argon laser to excite Fluo3-AM fluorescence. The signals were collected at 505-530 nm for detecting Fluo3-AM. Signals from control neurons and neurons treated with H 2 O 2 were compared using identical settings for laser power, confocal thickness, and detector sensitivity (40). The images were analyzed, and the mean of the Fluo3-AM signal was measured per live cell. Estimation of fluorescence intensity of Fluo3-AM was presented like the pseudoratio (⌬F/F) indicated by the following formula: ⌬F/F ϭ (F Ϫ F base )/(F base Ϫ B), where F is the measured fluorescence intensity of the indicator, F base is the fluorescence intensity before the stimulation, and B is the background signal determined from the average of areas adjacent to the cells (47,48).
Mitochondrial Potential Staining-Mitochondria membrane potential was estimated by specific mitochondrial probes tetramethylrhodamine ethyl ester perchlorate and Mitotracker Orange, detected in a confocal microscope. Neurons were grown on poly-L-lysine-coated glass coverslips and cultured for 5 days. The cells were then loaded for 30 min with Mitotracker Orange plus Calsein AM, or 30 nM TMRM ϩ in Krebs-Ringer-Hepes (KRH)-glucose containing 0.02% pluronic acid and then washed and allowed to equilibrate for 30 min. Coverslips were then mounted in a chamber on the stage of a confocal laser-scanning microscope (LSM Pascal Zeiss model 510; Carl Zeiss Ltd.). The fluorescence changes determined by Mitotracker Orange and TMRM ϩ fluorescence indicated the mitochondria potential staining (49). Images were acquired using a 543-nm helium-neon laser to excite TMRM ϩ or Mitotracker Orange, and the signals were collected at 570 nm (49). Signals from control neurons and neurons incubated with different treatments were compared using identical settings for laser power, confocal thickness, and detector sensitivity (40). The images were analyzed, and the mean TMRM ϩ and Mitotracker Orange signals were measured per live cell. Estimation of fluorescence intensity of TMRE was presented like the pseudoratio (⌬F/F) indicated by the following formula: where F is the measured fluorescence intensity of the indicator, F base is the fluorescence intensity before the stimulation, and B is the background signal determined from the average of areas adjacent to the cells (27).
Statistical Analysis-Data were expressed as the mean Ϯ S.E. of the values from the number of experiments as indicated in the corresponding figures. Data were evaluated statistically by using Student's t test, with p Ͻ 0.05 considered significant. An analysis of variance test was used to compare n differences between experiments.

RESULTS
Formation and Characterization of Amyloid Aggregates-A␤ peptide is neurotoxic for different neuron types and induced a series of neuro- pathologic alterations. There are controversies about the A␤ species involved in the neurodegenerative process. The A␤ fibrils induced neurodegenerative changes, including neuronal cell death, apoptosis, oxidative stress, and calcium deregulation (16). Other A␤ species, like oligomers, that did not present fibril structure induced preferentially their effect at synaptic regions (49). In order to investigate the A␤ species formed in our studies, we characterized our A␤ fibril preparations (Fig.  1). The A␤ peptide was aggregate for 24 h under constant stirring at room temperature. Then aliquots of the total fraction (T) were taken, and the amyloid fibrils were separated by centrifugation into a pellet (P) and a supernatant (S). The amount of A␤ fibrils formed was established under Tris-Tricine SDS-PAGE (Fig. 1A). Under these conditions, the A␤ peptide formed practically 100% of fibril species, indicating the absence of A␤ soluble species in the supernatant of our preparations. Additionally, the quantity of A␤ fibrils formed was studied by thioflavin T assay on the different fractions after 24 h of the aggregation process (Fig. 1B). In both the total fraction and the pellet, the A␤ species formed correspond to fibrils. Finally, we studied the form and shape of the A␤ aggregates formed over time under the electron microscope (Fig. 1C).
Pictures indicated typical amyloid fibrils in the total fraction after A␤ aggregation. After 72 h, the A␤ fibrils are still stable, showing the typical fibrilar structure reported previously (15,19). These results show that under our experimental conditions, a stable fraction of A␤ fibrils was used in our studies.
The Induction of Peroxisomal Proliferation Protects from the Neuronal Cell Death Induced by A␤ in Hippocampal Neurons-Peroxisome proliferators were originally identified as inducers of peroxisome proliferation in different cell types (11,50). Therefore, we exposed hippocampal neurons to the peroxisome proliferator Wy and determined its potential neuroprotective action. Additionally, we preincubated the hippocampal cultures with 3-AT, an inhibitor of catalase activity. After 24 h of preincubation, neurons were challenged with increasing concentrations of A␤ peptide (Fig. 2, A and B). Neurons exposed to A␤ show a progressive dose-dependent death with a significantly impaired cellular metabolism (ϳ50% at 5 M A␤) ( Fig. 2A). To corroborate the effect of peroxisome proliferation involved in the neuronal protection against A␤-induced toxicity, we tested 10 mM 3-AT by inhibiting endogenous catalase activity present in the peroxisomes. Among the neuronal pop- ulation tested, the viability value dropped to half when 0.1 M A␤ was used, and the decay became even more pronounced with 5 and 10 M A␤ (Fig. 2B), suggesting that catalase plays an important role in the neuroprotection against A␤. Moreover, hippocampal neurons preincubated with 100 M Wy ( Fig. 2A) protect neurons by improvement of their metabolic activity by around 30 -40% between 5 and 10 M A␤ (p Ͻ 0.001). We also used 4-PB, which induces human peroxisome proliferation (13) and protects from hypoxia damage in neuroblastoma cells (51). In this case, the preincubation with 5 mM 4-PB prevents the neuronal cell death induced by A␤, and the survival was very similar to the one obtained with the PPAR␣ agonist Wy (data not show). In order to corroborate the neuronal cell death protection induced by Wy, we investigated the mitochondrial potential and the neuronal body staining with Mitotracker Orange and Calsein AM, respectively (Fig. 2C). Control neurons show an almost intact neuronal cell body and a defined pattern of red fluorescence that indicated mitochondrial potential staining in the neuronal process (Fig. 2C, a (white arrows)) in comparison with A␤-treated neurons (Fig. 2C, b). A␤-treated neurons lack mitochondrial potential staining in the neurites, and there is clear evidence of damage in their neuronal body represented by green fluorescence staining of Calsein AM (Fig. 2C, b, white arrows). However, the preincubation with Wy prevent the mitochondrial potential loss and the morphological damage observed in hippocampal neurons exposed to A␤ (Fig. 2C, d, white arrows). Besides, the treatment for 24 h with Wy displays a similar mitochondrial potential staining and neuronal morphology as the untreated cells (Fig. 2C, c). Moreover, 3-AT induces a decrease in the mitochondrial potential staining, concomitant with neuronal morphology alterations (Fig. 2C, e). The simultaneous treatment with 3-AT and A␤ peptide exacerbated the mitochondrial potential loss and induced a severe shortened in the neuronal process (Fig. 2C, f, white arrows). These results suggest that drugs that induce peroxisome proliferation actually prevent the neuronal cell death and mitochondrial collapse induced by A␤.
Peroxisome Proliferators Protect Hippocampal Neuronal Morphology in Cells Exposed to Antibody-In additional experiments, the neuronal morphology of hippocampal neurons exposed to either A␤ or Wy was examined by confocal microscopy using an antibody against the heavy neurofilament subunit (NF-200). Control hippocampal neurons formed an extensive network of well spread shapes and a profusion of long neurites (Fig. 3A). When neurons were treated with 100 M Wy for 48 h, they maintain their well spread shape with long neurites (Fig. 3C). Upon exposure of hippocampal neurons to 5 M A␤ during 24 h, neuronal cells showed clear morphological alterations, including somatic shrinkage and axon and dendritic dystrophy (Fig. 3B , white arrows). Longer exposures to A␤ result in further neurodegeneration, including the complete loss of neurites and cell death (35,40,52). Additionally, axonal swelling and accumulation of kinesin and cargo proteins are found in transgenic models that overexpress amyloid precursor protein (53). This impaired in axonal transport can promote aberrant A␤ generation and provide an appropriate substrate for the formation of senile plaques (53). However, when neurons pre-exposed to A␤ were co-incubated with either 100 M Wy or 5 mM 4-PB, they preserved their normal cell body shape as well as their long neurites (Fig. 3, D and E). On the other hand, when ␤-amyloid-treated neurons were co-incubated with 10 mM 3-AT, an increase in neuronal damage was observed, with an almost complete loss of neurites (Fig. 3F , white arrows). In further studies, a quantification of the neurites per neuron was carried out in all of the above conditions (TABLE ONE). Results are consistent with the immunofluorescence analysis, and we want to emphasize the protection of the neuronal integrity elicited by Wy in the presence of A␤, in which more than 75% of the neurites/neuron were maintained (TABLE ONE) in comparison with A␤-treated neurons alone, in which only 15% of the neurites/neuron were preserved (TABLE ONE). As a whole, the results presented in this section are consistent with the idea that the activation of the PPAR␣ increased the neuronal viability and prevented the neurite collapse of hippocampal neurons exposed to A␤.
Increasing the Number of Peroxisomes Protects Hippocampal Neurons from A␤ Neurotoxicity-PPAR␣ induces the expression of several peroxisomal membrane proteins in different cell types (54). In order to correlate the protective effect of the Wy in A␤-treated neurons, we decided to observe the presence and distribution of peroxisomes in hippocampal neurons. To accomplish this, we used a specific antibody against the peroxisomal membrane proteins with indirect immunofluorescence (42,43). This method allows estimating the total peroxisome number in different cell cultures (41). It has been shown that peroxisomes of neuronal cells are present in cell bodies, axons, dendrites, and presynaptic axon terminals (7,9). As shown in control cells, peroxisomes are detected throughout the neurons, at the soma but also at in neuronal processes (Fig. 4A). This finding is consistent with reports showing the presence of peroxisomes in neuronal processes and in the astrocytic processes (55). The treatment with Wy during 48 h induced a severe increase in the peroxisome number, with a preferential localization in all neurite processes (Fig. 4C). In fact, the number of peroxisomes FIGURE 3. Peroxisome proliferation prevents the morphological changes induced by A␤ in hippocampal neurons. A, control hippocampal neurons. C, neurons exposed to Wy. Neurons exposed to A␤ alone (5 M for 10 h) displayed a somatic shrinkage plus dendritic dystrophy (B) (white arrows). On the other hand, hippocampal neurons exposed to A␤ plus either 100 M Wy or 5 mM 4-PB showed an almost intact morphology, with a well developed branching of neurites (D and E, respectively). However, neurons exposed to A␤ plus 10 mM 3-AT showed, besides a somatic shrinkage, a dramatic reduction in the number of neurites per hippocampal neuron (white arrows) (F). Bar, 10 m.  4C and TABLE TWO). In addition, in PC12 cells, peroxisomes observed in neurites were mostly found in the thicker parts at varicosities, at branch points, and close to or within growth cones (9). This localization has a possible function in the exocytosis (56), suggesting an active metabolic process associated with peroxisomes. Therefore, the peroxisomal localization pattern induced by Wy may be involved in the neuroprotection mechanism against the A␤ toxicity (53). Moreover, when neurons were challenged with 10 M A␤ for 8 h, they showed clear changes in the total peroxisome number ( Fig. 4B and TABLE TWO). The peroxisome number severely decreases, and the neurons showed evident signs of neurite loss and axon dystrophy (Fig. 4B). However, neurons previously treated with 100 M Wy for 48 h increased the number of peroxisomes, and the A␤-induced damage was prevented ( Fig. 4D and TABLE TWO). Additionally, the human peroxisomal proliferator, 4-PB, also shows protective effects when it is used before A␤; in fact, neurons treated with 5 mM 4-PB plus A␤ preserve the number of peroxisomes and morphology (Fig. 4E). On the other hand, the effect of A␤ plus 10 mM 3-AT was dramatic, showing a clear decrease in the number of peroxisomes present in the neuronal cultures (Fig. 4F). These results suggest that the peroxisomal proliferators used induced peroxisome proliferation, and this event may be involved in the protection of hippocampal neurons from the A␤ neurotoxic action. Induction of PPAR␣ and Catalase Activity Protect Neurons from A␤ Toxicity-In order to determine whether the peroxisome proliferators have a positive effect on the expression of the PPAR␣ gene, we ran reverse transcription-PCR experiments to clarify this issue. Hippocampal neurons were treated with increasing concentrations of Wy for 12 h. The incubation with Wy induced a consistent increase in the PPAR␣ mRNA levels (Fig. 5). This change is independent of the ␤-actin mRNA levels, used as an internal control (Fig. 5A). The maximal induction of PPAR␣ was observed at 100 M Wy, with a 2-fold expression over the basal levels (Fig. 5A). In order to correlate the induction of PPAR␣ with catalase, a specific enzyme present in peroxisomes, their specific activity was measured (Fig. 5B). As compared with control neurons (Fig. 5B, first  bar), the treatment with Wy shows an increase of 2.6-fold in catalasespecific activity (Fig. 5B, second bar). The neurons challenged with A␤ show a partial decrease in the activity levels (Fig. 5B, third bar). However, in A␤-treated neurons plus Wy, an increase in more than 30% of catalase activity was observed as compared with the A␤-treated neurons alone (Fig. 5B, fourth bar) (p Ͻ 0.05). Hippocampal neurons previously induced by the PPAR␣ agonist further treated by 10 mM 3-AT show a 96% decrease in catalase activity in comparison with neurons exposed to the agonist alone (Fig. 5B, fifth bar compared with second bar). These studies are supported by the studies of Schriner et al. (57), in which a transgenic mouse was generated that overexpressed human catalase. These authors observed an increase in lifespan time and the fact that cardiac pathology and cataract development were delayed; moreover, oxidative damage and H 2 O 2 production was reduced (57). Our results suggest that PPAR␣ induction controlled the increase in catalase activity, which improves the neuronal antioxidant capacity, in order to protect the hippocampal neurons from the A␤ neurotoxic action.
Peroxisome Proliferators Prevent the Oxidative Stress Induced by A␤-In order to test whether the peroxisomal proliferation is responsible for the neuroprotection against A␤ toxicity mediated by the induction of ROS, the intraneuronal content of ROS was quantified using the fluorescent probe 2,7-dichlorofluorescein (45) (Fig. 6). As expected, the intraneuronal ROS levels were increased 3.5-fold over the control in neurons exposed to 10 M A␤ (Fig. 6, B and F (second bar)). As a control, we evaluated the ROS levels in hippocampal neurons exposed to H 2 O 2 . The treatment with 50 M H 2 O 2 showed a 2.5-fold increase in the amount of intraneuronal ROS (Fig. 6, C and F (third bar)). The neurons previously treated with Wy showed ROS levels similar to those of untreated neurons (Fig. 6, D and F (fourth bar)). Moreover, the preincubation with Wy prevented the ROS production induced by 10 M A␤ (Fig. 6E). These neurons displayed a fluorescent level in a range of control cells (Fig. 6, E and F (fifth bar)). Fig. 6F shows the quantification of the intraneuronal ROS levels observed in different representative experiments. These results indicate that the peroxisome proliferator protects from A␤ neurotoxicity by a mechanism that involves control of the generation of oxidative stress.
Peroxisome Proliferators Prevent Changes in the Wnt Signaling Pathway Triggered by Oxidative Stress-Defects in the Wnt signaling pathway have been postulated to contribute to the pathogenesis of AD, and some of these events may be related to the modulation of ␤-catenin and glycogen synthase kinase 3␤ (GSK-3␤) (52,58). Regulation of ␤-catenin stability is a crucial control point in Wnt signaling pathways (59). A␤ peptide increases the production of intraneuronal ROS and stimulates the H 2 O 2 levels through metal ion reduction (30). In vitro experiments also suggest that the A␤ neurotoxic effect is mediated by free radical mechanisms (23,25) and alteration of Ca 2ϩ homeostasis (35). Therefore, we investigated whether the protective effect of Wy on A␤-induced neurodegeneration involved ␤-catenin stabilization in H 2 O 2 -  Effect of Wy on peroxisome number of hippocampal neurons exposed to A␤ Neuronal peroxisomes were detected by indirect immunofluorescence using an antiserum to detect peroxisomal membrane protein (42,43). Data are mean Ϯ S.E. from four separate experiments performed in duplicate. treated neurons. Immunofluorescence studies were carried out to localize ␤-catenin in hippocampal neurons exposed to H 2 O 2 (Fig. 7A,  a-d). Under control conditions, ␤-catenin displayed a predominantly cytoplasmic localization (Fig. 7A, a). However, after treatment with 50 M H 2 O 2 for 24 h, a decrease in the cytoplasmic ␤-catenin level became apparent in hippocampal neurons (Fig. 7A, b). On the other hand, the incubation with 100 M Wy increases the intraneuronal ␤-catenin levels (Fig. 7A, c), and the 24 h preincubation with Wy prevented ␤-catenin destabilization induced by 50 M H 2 O 2 , within the cytoplasmic compartment and the neuronal processes (Fig. 7A, d). The immunofluorescence studies were corroborated with Western blots that show the ␤-catenin stabilization in neurons incubated with Wy in the presence of 50 M H 2 O 2 (Fig. 7B). The changes in the ␤-catenin levels are independent of the intraneuronal ␤-tubulin control levels (Fig. 7B). As a whole, these results suggest that the peroxisome proliferators induced a stabilization of total ␤-catenin levels, an event that may be involved in the neuroprotective effect against A␤ and H 2 O 2 toxicity. Peroxisomal Proliferation Prevents Changes in the Wnt Signaling Pathway Triggered by A␤ Peptide-We also investigated whether the protective effect of Wy on A␤-induced neurodegeneration involved ␤-catenin stabilization. Double immunofluorescence studies were carried out to localize ␤-catenin (green fluorescence) and neurofilament protein (red fluorescence) in hippocampal neurons (Fig. 8A, a-d).

Number of peroxisomes
Under control conditions ␤-catenin displayed a predominantly cytoplasmic and nuclear localization (Fig. 8A, a); however, in the presence of ␤-amyloid, a decrease in the nuclear and cytoplasmatic ␤-catenin level became apparent in hippocampal neurons (Fig. 8A, b). On the other hand, preincubation for 24 h with 100 M Wy prevented ␤-catenin destabilization induced by A␤ within the nuclear compartment and the neuronal processes (Fig. 8A, d). On the other hand, Wy did not induce any apparent changes in the intraneuronal ␤-catenin levels (data not shown); however, the co-incubation with catalase inhibitor, 3-AT, and A␤, produces a severe decrease in the ␤-catenin levels (Fig. 8A, c), indicating a possible role of the neuronal redox mechanism in the Wnt pathway activation. An analysis of the effect of Wy on the ␤-catenin content of neurons exposed to A␤ was studied by Western blot (Fig. 8B). Results show clearly that Wy prevents the decrease in ␤-catenin levels in hippocampal neurons exposed to A␤ (Fig. 8B). This change is independent of the variations in the total neuronal protein amount, indicated by the ␤-actin levels (Fig. 8B). The effect of the Wy in the ␤-catenin stabilization is represented by a graph that indicates the result of three independent experiments (Fig. 8C). Interestingly, the effect of Wy in the ␤-catenin decrease induced by A␤ has a similar pattern that the one observed in hippocampal neurons treated by H 2 O 2 (Fig. 7C). This observation suggests that the oxidative stress induces by A␤ treatment is perhaps responsible for the changes in the Wnt signaling pathway. Additionally, these results are supported by studies by Shin et al. (60), who found that H 2 O 2 causes the reduction of ␤-catenin level by regulating GSK-3␤ activity.

Peroxisome Proliferation Prevents the Calcium Deregulation Induced by Oxidative Stress and ␤-Amyloid Peptide in Primary Hippocampal
Neurons-The cytoplasmic calcium increase induced by oxidative stress and A␤ has been reported previously (40,61), and it seems to be involved in the apoptosis and mitochondrial collapse in different cell types (61). In order to evaluate the role of the peroxisome proliferators in the neuronal calcium perturbations, we studied the cytoplasmic calcium changes in neurons loaded with Fluo3-AM and preincubated with Wy in the presence of H 2 O 2 and A␤ peptide (Fig. 9). Fluo3-AM detects the cytoplasmic calcium changes in different cell types (46). 50 M H 2 O 2 produces an increasing entry of cytoplasmic calcium during the neuronal treatment (Fig. 9, A (2) and B (black circles)). In neurons preincubated for 24 h with 100 M Wy, the pretreatment prevents the calcium increase in hippocampal neurons exposed to H 2 O 2 (Fig. 9, A (4) and B (white circles)). The neurons preincubated with Wy show a very similar fluorescence level as compared with untreated neurons loaded with Fluo3-AM (data not shown). Quantified relative units at 30 min show that H 2 O 2 increased calcium entry around 100% over control neurons; however, cells previously treated with Wy show a significant decrease when compared with H 2 O 2 alone (Fig. 9C). Moreover, we performed calcium imaging experiments using the Fluo3-AM indicator in neurons  DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 exposed to A␤ peptide and in cells preincubated for 24 h with Wy (Fig.  9D). Preincubation with the peroxisome proliferator, Wy, previous to A␤ treatment prevented the cytoplasmic calcium increase induced by the peptide (Fig. 9, D (4) and E (black circles)) in hippocampal neurons treated with the peptide alone (Fig. 9, D (2) and E (white circles)). In Fig.  9F, we show the quantification of the changes in fluorescence 30 min after the addition of the treatments. These observations show a statistically significant difference between the effect of the A␤ peptide and the co-treatment with Wy (Fig. 9F). The preincubation with Wy avoided the increase in fluorescence, keeping the fluorescence intensities at a control level (Fig. 9F, third bar). These results indicate that a peroxisome proliferator prevents the cytoplasmic calcium increase induced by oxidative stress and A␤, an event that may be involved in a common toxicity mechanism in hippocampal neurons.

Effect of Wy in the Mitochondrial Potential Loss Produced by H 2 O 2 and A␤ Peptide in Primary Hippocampal
Neurons-Additionally, we determined the possible role of the peroxisomal proliferation in the mitochondrial potential loss induced by the oxidative stress in hippocampal neurons (Fig. 10). Here, we used the probe TMRM that detects the mitochondrial potential changes in vivo by confocal microscopy (27). In these experiments, we observed that Wy prevents the mitochondrial potential loss induced by the H 2 O 2 treatment (Fig. 10A). The mitochondrial potential alterations denote a severe decrease in the red fluorescence in hipocampal neurons treated with H 2 O 2 (Fig. 10, A  (2) and B (black circles)). These perturbations are prevented in hippocampal neurons pretreated for 24 h with Wy and then exposed to 50 M H 2 O 2 (Fig. 10, A (4) and B (white circles)). The graph in Fig. 10C represents a quantification of three independent experiments, where Wy prevents the mitochondrial potential loss induced by oxidative stress. We also studied the mitochondrial potential changes in hippocampal neurons exposed to A␤ and preincubated with Wy (Fig. 10D).
In these experiments, we observed a small decrease in the mitochondrial membrane potential fluorescence in A␤-treated neurons (Fig. 10, D (2) and E (white circles)). However, the preincubation with Wy totally prevented the mitochondrial potential loss induced by A␤ (Fig. 10, D (4) and E (black circles)). This effect is partial and of lower magnitude than the changes observed in neurons exposed to 50 M H 2 O 2 (Fig. 10, compare B and E). The quantification of three independent experiments shows that Wy prevents the decrease in the TMRM fluorescence induced by A␤ (Fig. 10F, third bar), results that are in agreement with Fig. 2, where we showed that Wy protects mitochondrial viability against A␤ toxicity using another mitochondrial probe. These results suggest that peroxisome proliferation prevents the cytotoxicity effects induced by A␤ and that these events are related with the generation of oxidative stress in neurons exposed to A␤ peptide.

DISCUSSION
Multicellular organisms have evolved complex homeostatic mechanisms to sense and respond to a diverse range of exogenous and endogenous signals (28,62). There is evidence showing that PPAR␣ and the genes under its control play a role in the evolution of oxidative stress excesses observed in aging (63). The administration to aged mice of agents capable of activating the ␣ isoform of PPAR was found to restore the cellular redox balance (64). Additionally, a PPAR␣ activator, fenofibrate, protects against cerebral injury by antioxidant and anti-inflammatory mechanisms (65). This evidence suggests that PPAR␣ could be a new pharmacological target to control the neurodegenerative changes induced by oxidative stress. A␤-mediated toxicity has been associated with the production of free radicals that are possibly involved in the pathogenesis and activation of the apoptotic pathway in AD (23). In order to study whether a similar pathophysiological mechanism applied to hippocampal neurons, we treated such neurons with A␤ in the absence or presence of Wy, a specific PPAR␣ agonist and peroxisome proliferator (3,11).
A␤ is neurotoxic to neurons incubated with increasing concentrations, as shown by a reduction of cell viability and loss of neuronal processes (35,40,52). Here we showed that neurons pretreated with Wy prevent the neuronal cell death induced by A␤ peptide. Additionally, this compound produced a 2-3-fold increase in the number of peroxisomes and in the specific activity of the peroxisomal matrix protein, catalase. The increase in the peroxisome number can explain A␤ neuroprotection against A␤, because the co-treatment with Wy and A␤ partially induced neuronal cell survival (data not show). These observations discard an antioxidant mechanism in the peroxisome proliferator actions. Neurons exposed to A␤ pretreated with Wy clearly showed neuroprotection as shown by an increase in the survival rate, mitochondrial viability, and the recovery of the number of neurites. Hippocampal neurons, which under control conditions form an extensive network of healthy and long neurites, after A␤ treatment showed clear morpholog- ical alteration and axonal and dendritic dystrophy (35,40,52). Wy and 4-PB drugs that increase the number of peroxisomes protect hippocampal neurons by way of preserving their normal shape, axonal processes, and neurites. The inhibition of catalase by 3-AT increased A␤ neurotoxicity, and this suggests that catalase is involved in the neuroprotection exerted against the ␤-amyloid peptide. Besides, catalase detoxification is one of the major protective defenses to oxidative stress in different cell types (2). Moreover, the expression of Bcl-2, an antiapoptotic protein, enhanced both the expression and the activity of catalase in different neuronal cell types (66). This evidence indicates an important role of catalase in the neuronal protection mechanism. Alternative pathways of H 2 O 2 detoxification by other systems like glutathione peroxidase that play some role in neuronal cell lines (67) apparently do not play a major role in the hippocampal neurons (68).
PPAR␣ activation results in the transcriptional up-regulation of many genes, including those involved in peroxisomal and mitochondrial fatty acid ␤-oxidation, certain isozymes of the cytochrome P450 family, and antioxidant enzymes (12,69). Additionally, the PPARs have been demonstrated to antagonize signaling on important pathways (STATs, AP-1, and NF-B) or cross-talk with the Wnt pathway to protect neurons from A␤ cytotoxicity (40,64). The Wnt signaling pathway plays an important role in a wide range of embryonic development, cell fate, and neurodegeneration, where ␤-catenin plays a central role in signaling (59). Besides, it has been demonstrated that H 2 O 2 down-regulated Tcf/ Lef-dependent transcription activity by degradation of ␤-catenin and deregulated the Wnt signaling pathway (70). In our studies, we show that Wy, a PPAR␣ agonist, stabilizes ␤-catenin levels in hippocampal neurons exposed to H 2 O 2 and A␤. These effects for both treatments are very similar, an observation that strongly involved the ROS production as a possible factor responsible for Wnt pathway deregulation induced by A␤. Prevention of decrease in the ␤-catenin levels by A␤ or oxidative stress is similar to that observed with lithium, which stabilizes the intraneuronal ␤-catenin levels by inhibiting the GSK-3␤ activity (52,59). Additionally, we found that hippocampal neurons exposed to Wy increase the total ␤-catenin levels (see Fig. 8D). These results suggest a cross-talk between the Wnt signaling pathway and the PPAR␣ activation.
We also tested the level of intracellular ROS in different conditions and showed that the increased level of ROS produced by incubation of neurons with A␤ or H 2 O 2 is dramatically reduced in neurons subjected to peroxisomal proliferation. These experiments illustrate that the increased number of peroxisomes protect from the oxidative damage   DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 induced by A␤. These studies are consistent with the idea that the A␤ neurotoxicity is mediated by the production of H 2 O 2 (30,31,35). Previous studies suggest that the PPAR␣ activation may enhance neuronal cell death (71); however, those neurons were previously treated with high K ϩ concentrations and then exposed to Wy. Under these conditions, the cell death effect may correspond to the activation of other pathways, because an increase in the extracellular K ϩ concentrations induced a very rapid loss of the neuronal energy and a large production of ROS by the mitochondria (72).

Peroxisome Proliferation Protects from A␤
All of the observed effects support the view that increasing the number of peroxisomes can overcome vulnerability to oxidative stress. On the basis of these data, one could imagine two potential causes for neuronal damage due to A␤ cytotoxicity: 1) Production of intracellular ROS, which cause general neuronal damage and 2) retention of peroxisomes in the cell body by alterations of their axonal transport induced by A␤ (10). Loss of peroxisomes makes neurons vulnerable to oxidative stress and has serious consequence for growth and survival, which eventually leads to degeneration. Neurons lacking peroxisomes in neurites have shorter processes and become highly sensitive to oxidative condi-tions like H 2 O 2 (10). In endothelial cells, an increase in permeability and protein kinase C activity has been observed when H 2 O 2 was administrated exogenously; in this effect, the ERK1/ERK2 signaling is involved with an increase in cellular tyrosine kinase activity (73). Other studies indicated that H 2 O 2 stimulates the internalization of endothelial cadherins and cadherin disorganization away from lateral cell-cell junctions combined with a loss of ␤-catenin/cytoskeletal association (74). Exogenous catalase can provide substantial protection to such oxidative damage (74,75).
There is increasing experimental evidence suggesting the involvement of peroxisomes into the metabolism of ROS, including radical and nonradical derivatives of O 2 (8,68). A strain of mice that bear a null mutation in PPAR␣ (PPAR␣ Ϫ/Ϫ ) experimentally demonstrate that normal PPAR␣ function is necessary to effectively maintain a balance in cellular REDOX state; these animals express indicators of oxidative stress much earlier in their lifespan that wild-type mice (76). Besides, analysis of mice lacking the PEX2 peroxisome assembly gene, in which peroxisomal function is disrupted, reveals abnormal cerebellar histogenesis due to the disturbance of multiple cellular processes within neurons and a severe defect in the neuronal migration process and cortex deficiency architecture (63,77). These features are common in pathologies like Zellweger syndrome and adrenoleukodystrophy (63, 78 -80). Similar results were obtained using young PPAR␣ ϩ/ϩ and PPAR␣ Ϫ/Ϫ mice rendered redox-imbalanced; a decline in cellular PPAR␣ expression was observed to occur with normal aging, with reduced levels of acyl-CoA oxidase and catalase mRNA expression (50). However, the treatment of H4 cells with fenofibrate, a PPAR␣ agonist, increased the A␤ 42 production without significant alteration in the A␤ 40 levels (81). Elevations in A␤ 42 production are comparable with A␤ 42 levels observed in AD patients, although the incubation with fenofibrate did not modify the cell viability as observed by lactate dehydrogenase and MTT assays (81). Additionally, in Tg2567 transgenic mice, fenofibrate, did not significantly alter the A␤ 42 levels in vivo (81).
The preincubation with Wy prevents the cytoplasmic calcium increase induced by H 2 O 2 and A␤. Additionally, Wy treatment protects from mitochondrial stress induced by A␤ and H 2 O 2 , indicating a common mechanism of neurotoxicity. However, the mitochondrial potential change induced by A␤ is less drastic that the ones observed with H 2 O 2 , indicating in A␤ neurotoxicity there is an additional toxicity mechanism besides ROS generation. The calcium observations are in agreement with previous evidence, where we found that the PPAR␥ activation prevents the calcium deregulation induced by A␤ (40). Besides, the PPAR␥ agonists prevent the neuronal cell death and the morphological changes induced by A␤ treatment (40). Additionally, the PPAR␥ agonist pioglitazone was effective in relaxing the pressurized arteries, suggesting that the vasodilation of arteries could be explained by the inhibition of calcium entry through L-type voltage-dependent calcium channels (82). Besides, diverse peroxisome proliferators produced changes in [Ca 2ϩ ] i in hepatocytes, through the redistribution of different internal Ca 2ϩ pools (83). Therefore, the novel neuroprotective role of the PPARs against A␤ toxicity is very significant, and it may be of therapeutic value.
In conclusion, we demonstrate here that the activation of PPAR␣ prevents ␤-amyloid-induced neuronal cell death and the corresponding morphological changes. These events are mediated by the increase in the number of peroxisomes and in catalase activity. Our results justify further study of the novel role of the PPARs in the prevention of the  (1). Wy protects from mitochondrial potential loss in hipocampal neurons exposed to H 2 O 2 (4). B shows a time course of fluorescence recordings obtained for TMRM in neurons exposed to H 2 O 2 (black circles) and Wy-preincubated neurons exposed to H 2 O 2 (white circles). Quantification of mitochondrial potential fluorescence as relative units at 30 min (C) shows the control neurons (first bar), and neurons treated with H 2 O 2 showed a severe mitochondrial potential disruption (second bar). Neurons pretreated with Wy plus H 2 O 2 show similar mitochondrial potential fluorescence levels as control neurons (third bar). Data are mean Ϯ S.E. (bars) from three separate experiments performed in duplicate. *, p Ͻ 0.05 by nonpaired Student's t test. D, confocal microscopy photographs of mitochondrial potential staining at control neurons (1) and neurons treated with A␤ for 30 min (2). Neurons previously treated for 12 h with Wy are shown (control cells (3)), and neurons treated with Wy plus A␤ for 30 min are shown (4). A␤ induces a partial decrease in the mitochondrial potential fluorescence (2), in comparison with control neurons (1). Wy protects from mitochondrial potential loss triggered by 5 M A␤ treatments (4). E shows a time course of fluorescence recordings obtained for TMRM in neurons exposed to A␤ (black circles) and Wy-preincubated neurons exposed to A␤ (white circles). Quantification of mitochondrial potential fluorescence as relative units at 30 min (F) shows the control neurons (first bar), and neurons treated with A␤ showed a partial mitochondrial potential decrease (second bar). Neurons pretreated with Wy plus A␤ show similar mitochondrial potential fluorescence levels as control neurons (third bar), indicating the prevention effect. Data are mean Ϯ S.E. (bars) from three separate experiments performed in duplicate. *, p Ͻ 0.05 by nonpaired Student's t test.