Activation of the c-Jun N-terminal kinase signaling cascade mediates the effect of amyloid-beta on long term potentiation and cell death in hippocampus: a role for interleukin-1beta?

Amyloid-beta (Abeta) is a major constituent of the neuritic plaque found in the brain of Alzheimer's disease patients, and a great deal of evidence suggests that the neuronal loss that is associated with the disease is a consequence of the actions of Abeta. In the past few years, it has become apparent that activation of c-Jun N-terminal kinase (JNK) mediates some of the effects of Abeta on cultured cells; in particular, the evidence suggests that Abeta-triggered JNK activation leads to cell death. In this study, we investigated the effect of intracerebroventricular injection of Abeta(1-40) on signaling events in the hippocampus and on long term potentiation in Schaffer collateral CA1 pyramidal cell synapses in vivo. We report that Abeta(1-40) induced activation of JNK in CA1 and that this was coupled with expression of the proapoptotic protein, Bax, cytosolic cytochrome c, poly-(ADP-ribose) polymerase cleavage, and Fas ligand expression in the hippocampus. These data indicate that Abeta(1-40) inhibited expression of long term potentiation, and this effect was abrogated by administration of the JNK inhibitor peptide, D-JNKI1. In parallel with these findings, we observed that Abeta-induced changes in caspase-3 activation and TdT-mediated dUTP nick-end labeling staining in neuronal cultured cells were inhibited by D-JNKI1. We present evidence suggesting that interleukin (IL)-1beta plays a significant role in mediating the effects of Abeta(1-40) because Abeta(1-40) increased hippocampal IL-1beta and because several effects of Abeta(1-40) were inhibited by the caspase-1 inhibitor Ac-YVAD-CMK. On the basis of our findings, we propose that Abeta-induced changes in hippocampal plasticity are likely to be dependent upon IL-1beta-triggered activation of JNK.

One of the pathological hallmarks of Alzheimer's disease (AD) 1 is an accumulation of plaques consisting predominately of amyloid-␤ (A␤) peptide, which is processed from amyloid precursor protein by the action of ␤and ␥-secretase (1). Neuronal cell loss is one feature of AD, and evidence from analysis of changes in cultured cells suggests that A␤ acts as the executioner. Thus, neuronal cultures exposed to A␤ demonstrate signs of apoptosis (2)(3)(4), and previous evidence from this laboratory has revealed that cultured cortical neurons exposed to A␤  exhibited increased expression of the tumor suppressor p53; increased activation of caspase-3, a marker of apoptotic cell death; and increased TUNEL reactivity (5). The evidence is consistent with the idea that activation of the stress-activated protein kinase, c-Jun N-terminal kinase (JNK) played a significant role, because depletion of JNK1 following exposure to antisense oligonucleotide prevented the effects of A␤ (5). Similarly, Morishima et al. (6) reported that A␤ increased phosphorylation of JNK and c-Jun in cultured cortical neurons and that these changes were associated with expression of the death inducer Fas ligand (FasL). Others have reported findings that support a role for JNK activation in mediating at least certain effects of A␤. For instance, A␤-induced parallel increases in JNK activation and TUNEL reactivity in PC12 cells (7), whereas activation of JNK was shown to be localized to amyloid deposits in 7-and 12-month-old mice that overexpress amyloid precursor protein (8).
It has emerged in several experimental models that increased JNK phosphorylation is associated with deficits in synaptic function; for instance, increased activation of JNK has been reported in the hippocampi of aged rats (9, 10), rats exposed to whole body irradiation (11), and rats injected with the proinflammatory cytokine, interleukin (IL)-1␤ (12) or lipopolysaccharide (13), and in all cases glutamate release was decreased. In each of these experimental conditions, long term potentiation (LTP), a model of synaptic plasticity, was markedly impaired, and this impairment was coupled with an increased hippocampal concentration of IL-1␤.
A number of groups have reported that A␤ administration exerts an inhibitory effect on LTP. For instance, A␤ peptides (14 -16) and naturally secreted A␤ oligomers (17) inhibited LTP in the CA1 region in vivo, and A␤ peptides also inhibit LTP in dentate gyrus and the CA1 in vitro (18 -21). Similarly, a deficit in LTP was reported in aged mice that overexpress amyloid precursor protein and in which deposition of A␤ was observed (22). In this study, we investigated the signaling events induced by A␤  that might explain its impact on LTP and report that activation of JNK is a pivotal event in A␤-induced inhibition of LTP and in A␤-induced cell death.  (BioSource International) was made up as a 1 mM stock solution in sterile water and allowed to aggregate for 48 h at 30°C as described previously (5). For treatment of cortical neurons, aggregated A␤  was diluted to a final concentration of 2 M in prewarmed neurobasal medium (NBM; Invitrogen). For analysis of signaling events stimulated by A␤, aggregated A␤  at 37°C was injected intracerebroventricularly (5 l; 1nmol in sterile water). This A␤ preparation (and concentration) was adopted because it was shown to produce consistent, reliable, and reproducible results in a number of markers, suggesting that cell death occurred in cultured cells (5).

Preparation of A␤-A␤
Animals-Groups of young male Wistar rats (200 -300 g; Bio Resources Unit, Trinity College, Dublin 2, Ireland), maintained at an ambient temperature of 22-23°C under a 12 h light-dark schedule, were used in this experiment. The rats were anesthetized by intraperitoneal administration of urethane (1.5 mg/kg) and were injected intracerebroventricularly with either sterile water (5 l) or A␤  . 6 h post-injection, the rats were killed by decapitation, the brains were rapidly removed on ice, and area CA1 was dissected free. The tissue was cross-chopped (350 ϫ 350 m) and frozen in Krebs solution containing 10% Me 2 SO as previously described (23) until required for analysis.
Analysis of JNK Phosphorylation, c-Jun Phosphorylation, Cytosolic Cytochrome c Expression, Bax Expression, FasL Expression, and PARP Cleavage-JNK phosphorylation, c-Jun phosphorylation, and expression of Bax, cytosolic cytochrome c, PARP, and FasL were analyzed in samples prepared from CA1 tissue using a method previously described (13). In the case of JNK, c-Jun, FasL, and PARP, tissue homogenates were diluted to equalize for protein concentration, and aliquots (100 l, 2 mg/ml) were added to 100 l of sample buffer (0.5 mM Tris-HCl, pH 6.8, 10% glycerol, 10% SDS, 5% ␤-mercaptoethanol, 0.05% bromphenol blue (w/v)), boiled for 5 min, and loaded onto 10% SDS-PAGE gels. In the case of cytochrome c, cytosolic fraction was prepared by homogenizing slices of hippocampus in lysis buffer (20 mM HEPES, pH 7.4, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 5 g/ml pepstatin A, 2 g/ml leupeptin, 2 g/ml aprotinin), incubating for 20 min on ice, and centrifuging (15,000 ϫ g for 10 min at 4°C). The supernatant (i.e. cytosolic fraction) was suspended in sample buffer to a final concentration of 300 g/ml, boiled for 3 min, and loaded (6 g/lane) onto 12% SDS-PAGE gels. The pellet (i.e. mitochondrial fraction) was resuspended in sample buffer to a final concentration of 300 g/ml, boiled for 3 min, and loaded (6 g/lane) onto 12% SDS-PAGE gels. Bax expression was assessed in the mitochondrial fraction. In all experiments the proteins were separated by application of 32 mA constant current for 25-30 min, transferred onto nitrocellulose strips (225 mA for 90 min), and immunoblotted with the appropriate antibody. For JNK phosphorylation, the proteins were immunoblotted with an antibody that specifically targets phosphorylated JNK (1:300 in Tris-buffered saline (TBS)-Tween (0.05% Tween 20) containing 0.1% BSA; Santa Cruz Biotechnology Inc.) for 2 h at room temperature. The blots were stripped and stained for total JNK. Nitrocellulose strips were probed with a mouse monoclonal IgG 1 antibody (1:200; Santa Cruz Biotechnology Inc.) raised against a recombinant protein corresponding to amino acids 1-384 representing fulllength JNK1 of human origin. To assess phosphorylation of c-Jun, we immunoblotted with a mouse monoclonal IgG 1 antibody (1:400 in PBS-Tween (0.1% Tween 20) containing 2% nonfat dried milk) raised against the peptide corresponding to a short amino acid sequence of phosphorylated c-Jun of human origin (Santa Cruz Biotechnology Inc.). To assess cytoplasmic cytochrome c, a rabbit polyclonal antibody (1:250 in PBS-Tween containing 2% nonfat dried milk; Santa Cruz Biotechnology Inc.) raised against recombinant protein corresponding to amino acids 1-104 of cytochrome c was used. In the case of FasL, we immunoblotted with a rabbit polyclonal antibody (1:500 in TBS-Tween containing 1% BSA; Santa Cruz Biotechnology Inc.) raised against a peptide corresponding to a short amino acid sequence at the N terminus of FasL of human origin. Bax expression was assessed in the mitochondrial fraction using a mouse monoclonal IgG 1 antibody (1:200 in TBS-Tween containing 1% BSA; Santa Cruz Biotechnology Inc.). To assess the cleavage of PARP, we immunoblotted with an antibody (1:500 in PBS-Tween (0.1% Tween 20) containing 2% nonfat dried milk) raised against the epitope corresponding to amino acids 764 -1014 of PARP of human origin (Santa Cruz Biotechnology Inc.). All of the nitrocellulose strips were reprobed for actin expression to ensure equal loading of protein on all SDS-PAGE gels. Actin expression was assessed using a mouse monoclonal IgG 1 antibody (1:300 in PBS-Tween containing 2% nonfat dried milk) corresponding to an amino acid sequence mapping at the C terminus of actin of human origin (Santa Cruz Biotechnology Inc.). Immunoreactive bands were detected as follows: peroxidase-conjugated anti-mouse IgG (Sigma) and Supersignal chemiluminescence (Pierce) for JNK, c-Jun, Bax, and actin and peroxidase-conjugated anti-rabbit IgG (Sigma) and Supersignal (Pierce) for cytochrome c, FasL, and PARP.
Induction of LTP in CA1 in Vivo-Male Wistar rats (175-200 g; Biomedical Facility, University College, Dublin, Ireland) were anesthetized with urethane (1.5 mg/kg), placed in a stereotaxic frame, and assessed for LTP as described previously (16). Small holes were drilled in the skull to allow insertion of a guide cannula to facilitate intracerebroventricular injection and to allow insertion of the reference, stimulating, and recording electrodes. The recording electrode was positioned in the stratum radiatum of area CA1 (3 mm posterior and 2 mm lateral to bregma), and a bipolar stimulating electrode was placed in the Schaffer collateral/commissural pathway distal to the recording electrode (4 mm posterior and 3 mm lateral to bregma). The cannula was positioned above the lateral ventricle in the opposite hemisphere to that of the electrodes (1 mm posterior and 1.2 mm lateral to bregma). Test shocks (0.033 Hz) were delivered to the Schaffer collateral/commissural pathway, and base-line excitatory postsynaptic potentials (EPSPs), recorded at 35-40% of maximal response, were sampled for at least 30 min to allow the response to stabilize. Rats were then injected intracerebroventricularly with either A␤  (1nmol in 5 l), the membrane soluble JNK inhibitor D-JNKI1 (1 nmol in 5 l), combined A␤  and D-JNKI1 (1nmol of each in 5 l) or vehicle (5 l sterile water); and base-line recordings were monitored for a further 60 min before delivery of a series of high frequency stimuli (HFS; 3 ϫ 10 trains of 10 stimuli at 200Hz; intertrain interval, 20 s). Responses to test shock stimulation were recorded for a further 5 h post-HFS, and deep body temperature was maintained at 36.5 Ϯ 0.5°C using heating pads. Paired pulse facilitation with an interstimulus interval of 50 ms was also examined preinjection, 1 h post-injection of drug/vehicle (prior to HFS), and 5 h following HFS. Deep body temperature was maintained at 36.5 Ϯ 0.5°C using heating pads. Extracellular field potentials were amplified (ϫ100), filtered at 5 kHz, digitized, and recorded using MacLab software. The EPSP slope was used to measure synaptic efficacy. EPSPs are expressed as percentages of the mean initial slope measured during the first 10 min of the base-line recording period.
Analysis of IL-1␤ Concentration-IL-1␤ concentration was analyzed in homogenate prepared from CA1 by enzyme-linked immunosorbent assay (R & D Systems) and in supernatants prepared from cultured cells as described below. Antibody-coated (100 l; final concentration, 1.0 g/ml; diluted in PBS, pH 7.3; goat anti-rat IL-1␤ antibody) 96-well plates were incubated overnight at room temperature, washed several times with PBS containing 0.05% Tween 20, blocked for 1 h at room temperature with 300 l of blocking buffer (PBS, pH 7.3, containing 5% sucrose, 1% BSA, and 0.05% NaN 3 ), and washed. IL-1␤ standards (100 l; 0 -1000 pg/ml in PBS containing 1% BSA) or samples (homogenized in Krebs solution containing 2 mM CaCl 2 ) were added, and incubation proceeded for 2 h at room temperature. Secondary antibody (100 l; final concentration, 350 ng/ml in PBS containing 1% BSA and 2% normal goat serum; biotinylated goat anti-rat IL-1␤ antibody) was added and incubated for 2 h at room temperature. The wells were washed, and detection agent (100 l; horseradish peroxidase-conjugated streptavidin; 1:200 dilution in PBS containing 1% BSA) was added and incubated continued for 20 min at room temperature. Substrate solution (100 l; 1:1 mixture of H 2 O 2 and tetramethylbenzidine) was added and incubated at room temperature in the dark for 1 h, after which time the reaction was stopped using 50 l of 1 M H 2 SO 4 . Absorbance was read at 450 nm, and the values were corrected for protein (24) and expressed as pg IL-1␤/mg protein.
Preparation of Cultured Cortical Neurons-Primary cortical neurons were isolated and prepared from 1-day-old Wistar rats (BioResources Unit, Trinity College, Dublin 2, Ireland) and maintained in NBM as previously described (5). The rats were decapitated, the cerebral cortices were dissected, and the meninges were removed. The cortices were incubated in PBS with trypsin (0.25 g/ml) for 25 min at 37°C. The cortical tissue was then triturated in PBS containing soy bean trypsin inhibitor (0.2 g/ml) and DNase (0.2 mg/ml) and gently filtered through a sterile mesh filter (40 m). The suspension was centrifuged at 2000 ϫ g for 3 min at 20°C, and the pellet was resuspended in warm NBM, supplemented with heat inactivated horse serum (10%), penicillin (100 units/ml), streptomycin (100 units/ml), and glutamax (2 mM). The sus-pended cells were plated at a density of 0.25 ϫ 10 6 cells on circular 10-mm diameter coverslips, coated with poly-L-lysine (60 g/ml), and incubated in a humidified atmosphere containing 5% CO 2 :95% air at 37°C for 2 h prior to being flooded with prewarmed NBM. After 48 h, 5 ng/ml cytosine-arabinofuranoside was added to the culture medium to suppress the proliferation of non-neuronal cells. The media were exchanged for fresh media every 3 days, and the cells were grown in culture for up to 7 days prior to treatment. In one set of experiments the neurons were incubated in the absence/presence of A␤  (2 M in NBM) for 72 h with or without caspase-1 inhibitor (100 nM in NBM; Ac-YVAD-CMK; Calbiochem) or D-JNKI1 (1 M in NBM; Alexis Biochemicals). In the case of A␤-treated neurones, the supernatant was removed at 20 h, and IL-1␤ concentration was assessed. At 72 h, the cells were rinsed in TBS and fixed in 4% paraformaldehyde in TBS for immunohistochemical assessment of JNK phosphorylation, caspase-3 activation, and DNA fragmentation. The cells were incubated in A␤   Analysis of Bax mRNA and caspase-3 mRNA-Total RNA was extracted from cortical neurones using TRI reagent (Sigma). cDNA synthesis was performed on 1 g of total RNA using oligo(dT) primer as per the manufacturer's instructions (Superscript reverse transcriptase; Invitrogen). The RNA was treated with RNase-free DNase I (Invitrogen) at 1 unit/g of RNA for 15 min at 30°C. Equal amounts of cDNA were used for PCR amplification for a total of 28 cycles. Primers were pretested through an increasing number of amplification cycles to obtain reverse transcriptase-PCR products in the exponential range. In the case of Bax mRNA expression following A␤ treatment primers used were as follows: rat Bax, sense 5Ј-GCAGAGAGGATGGCTGGGGAGA-3Ј, and antisense 5Ј-TCCAGACAAGCAGCCGCTCACG-3Ј (25); rat ␤-actin, sense 5Ј-GAAATCGTGCGTGACATTAAAGAGAAGCT and antisense 5Ј-TCAGGAGGAGCAATGATGATCTTGA-3Ј. The cycling conditions were 95°C for 5 min followed by cycles of 95°C for 75 s, 52°C for 75 s, and 72°C for 90 s. A final extension step was carried out at 70°C for 10 min. These primers generated Bax PCR products of 352 base pairs and ␤-actin PCR product of 360 base pairs. In the case of caspase-3 mRNA expression following treatment with A␤  and Bax mRNA expression following treatment with IL-1␤, multiplex PCR was performed using the Quantitative PCR Cytopress detection kit (Rat Apoptosis Set 2; BioSource International) generating caspase-3 PCR products of 320 base pairs, Bax PCR products of 352 base pairs and glyceraldehyde-3-phosphate dehydrogenase PCR product of 532 base pairs. The cycling conditions were 94°C for 1 min and 58°C for 2 min. A final extension step was carried out at 70°C for 10 min. The PCR products were analyzed by electrophoresis on 1.5% agarose gels, photographed, and quantified using densitometry. The target genes were normalized to expression of ␤-actin or glyceraldehyde-3-phosphate dehydrogenase housekeeping genes. No observable change in ␤-actin or glyceraldehyde-3-phosphate dehydrogenase mRNA was observed in any of the treatment conditions TUNEL Staining-Apoptotic cell death was assessed using the DeadEnd colorimetric apoptosis detection system (Promega) according to the manufacturer's instructions. Briefly, cultured cortical neurones were prepared from neonatal rats as described above and maintained in NBM for 12 days before incubating in the absence/presence of A␤  (1 M in NBM) for 72 h with or without caspase-1 inhibitor (100 nM in NBM) or D-JNKI1 (1 M in NBM). Biotinylated nucleotide was incorporated at 3Ј-OH DNA ends by incubating cells with terminal deoxynucleotidyl transferase for 30 min at 37°C. The washed cells were incubated in horseradish peroxidase-labeled streptavidin and then incubated in diaminobenzidine chromogen solution, and TUNEL-positive cells were calculated as a proportion of the total cell number.
Immunohistochemical Staining for Phosphorylated JNK and Activated Caspase-3-Cultured cortical neurones were prepared from neonatal rats as described previously (5) and maintained in NBM for 12 days before incubating in the absence/presence of A␤    2 with lanes 1)). Total JNK (b) was assessed to ensure equal protein loading, and no significant difference was observed between groups. c-Jun blots were stripped and reprobed for actin to ensure equal protein loading (see second sample immunoblot in c).
in TBS, mounted with an aqueous mounting medium (Vectastain; Vector Laboratories), and sealed. The slides were examined under a Zeiss fluorescence microscope with the appropriate filter (fluorescein isothiocyanate: excitation, 495 nm, and emission, 525 nm; L-rhodamine: excitation, 540 and 574 nm, and emission, 602 nm).
Statistical Analysis-The data are expressed as the means Ϯ S.E. A one-way analysis of variance (ANOVA) was performed to determine whether there were significant differences between conditions. When this analysis indicated significance (at the 0.05 level), post hoc Student Newmann-Keuls test analysis was used to determine which conditions were significantly different from each other. A repeated measures ANOVA was used to compare mean EPSP slopes at different time points in the electrophysiological experiments. When comparisons were being made between two treatments, an unpaired Student's t test for independent means was performed. Fig. 1a shows a sample immunoblot in which a marked increase in p-JNK was observed following intracerebroventricular injection of A␤  ; assessment of the mean data obtained from densitometric analysis revealed a statistically significant increase in JNK phosphorylation induced by A␤ (p Ͻ 0.001; Student's t test for independent means; n ϭ 5). In contrast to the change in JNK phosphorylation, total JNK expression was similar in A␤-treated and control rats as demonstrated in the sample immunoblot and the mean data (Fig.  1b). The A␤-induced increase in JNK phosphorylation was paralleled by the change in c-Jun phosphorylation; thus, the sample immunoblot shown in Fig. 1c and the mean data obtained from densitometric analysis indicated that A␤  induced a marked increase in c-Jun phosphorylation (p Ͻ 0.05; Student's t test for independent means; n ϭ 5). Protein loading

FIG. 3. D-JNKI1 reverses the A␤-induced inhibition of LTP.
Mean percentage EPSP slopes in the 5-min period immediately following tetanic stimulation (a) and in the first 5-min periods in each subsequent hour (b-f) were significantly lower in A␤-treated rats compared with controls (***, p Ͻ 0.001; ANOVA). This inhibition was reversed by D-JNKI1 treatment from 2 h so that there was a significant difference between mean EPSP slopes in A␤-treated and A␤ϩD-JNKI1-treated rats (ϩϩϩ, p Ͻ 0.001; ANOVA). was checked by reprobing immunoblots for actin, and the data indicate that its expression was similar in samples prepared from control and A␤-treated rats.
We argued that this A␤-induced increase in JNK activation may contribute to the previously reported A␤-induced inhibi-tion of LTP (16), and to assess this, rats were injected intracerebroventricularly with A␤  alone or in combination with the peptide inhibitor, D-JNKI1. Fig. 2a shows that, in control rats tetanic stimulation led to an immediate and persistent increase in EPSP slope (p Ͻ 0.001; ANOVA); treatment with D-JNKI1 (1nmol) did not significantly affect this change. In contrast, intracerebroventricular injection of A␤  inhibited LTP (p Ͻ 0.001; ANOVA, Fig. 2b); the effect was observed immediately such that the mean percentage change in the EPSP slope in the 5 min immediately following tetanic stimulation was significantly reduced in A␤-treated compared with control rats (p Ͻ 0.01; ANOVA; Fig. 3a). The A␤-associated change persisted so that the mean percentage changes in EPSP slopes in the final 5-min period of each hour were also significantly reduced in A␤-treated rats compared with control animals (***, p Ͻ 0.001 in all cases; ANOVA; Fig. 3, b-f). Coinjection of D-JNKI1 and A␤  reversed the inhibitory effect of A␤   (Fig. 2b), but this effect was not apparent until 2 h after tetanic stimulation (ϩϩϩ, p Ͻ 0.001; ANOVA; Fig. 3, b-f). The mean percentage EPSP slopes at 0 -5 min post-tetanus were significantly reduced in rats treated with A␤ and D-JNKI1 compared with control rats (p Ͻ 0.001; ANOVA; Fig. 2b). The results from paired pulse facilitation experiments found that there was no significant change in paired pulse facilitation observed either between groups of animals or preinjection compared with 1 h post-injection or 5 h following HFS. This indicates that at the

FIG. 4. Effect of A␤ on IL-1␤ concentration in hippocampus.
IL-1␤ concentration was significantly increased in hippocampus of A␤treated rats compared with control (Con) rats (***, p Ͻ 0.001; Student's t test for independent means; n ϭ 5). doses used here, A␤  and D-JNKI1 do not alter neurotransmitter release, suggesting a postsynaptic site for the modulation of LTP observed in our experiments.
The findings of several studies have indicated that enhanced JNK phosphorylation in the hippocampus is coupled with enhanced IL-1␤ concentration and impaired LTP (9,12,25); therefore we considered that the effect of A␤  might be mediated by IL-1␤. Fig. 4 shows that the concentration of IL-1␤ in hippocampus was significantly increased in hippocampus of A␤-treated rats compared with control rats (p Ͻ 0.001; Student's t test for independent means; n ϭ 5).
Our findings from a previous study (11) linked activation of JNK with translocation of cytochrome c from mitochondria, suggesting that the patency of the mitochondrial membrane was affected by JNK activation. In an effort to address this question, we assessed expression of the pro-apoptotic protein Bax in a mitochondrial fraction and cytochrome c in a cytosolic preparation obtained from control and A␤-treated rats. The sample immunoblot shown in Fig. 5a indicates that A␤  increased expression of Bax, and densitometric analysis revealed that the mean value in samples prepared from A␤treated rats was significantly enhanced compared with that in control rats (p Ͻ 0.05; Student's t test for independent means; n ϭ 5). The sample immunoblot and mean data in Fig. 5b indicate that cytochrome c expression in cytosolic samples prepared from A␤-treated rats was enhanced compared with samples from control rats; comparison of mean values indicated that the difference was statistically significant (p Ͻ 0.05; Student's t test for independent means; n ϭ 5). A similar pattern was observed in terms of expression of the intact fragment (116 kDa) of PARP; Fig. 5c shows one sample immunoblot and the mean data indicating that A␤  significantly decreased expression of 116-kDa PARP in hippocampal tissue prepared from A␤-treated rats (p Ͻ 0.05; Student's t test for independent means; n ϭ 5). It is considered that, in addition to increased Bax expression, cytosolic cytochrome c expression, and PARP cleavage, expression of FasL is indicative of apoptotic cell death, and, in parallel with the increases in Bax expression, cytosolic cytochrome c expression and PARP cleavage, A␤  significantly enhanced expression of FasL (p Ͻ 0.01; Student's t test for independent means; n ϭ 5; Fig. 5d), and in the case of Bax, cytochrome c, PARP, and FasL, equal protein loading was confirmed by reprobing for actin.
The A␤-associated increase in Bax protein was mirrored by an A␤-induced increase in Bax mRNA in cultured cortical neurons as shown by the sample gel and by analysis of the mean data indicating that the difference between expression in A␤treated and control cultured cells reached statistical significance (p Ͻ 0.05; Student's t test for independent means; n ϭ 5; Fig. 6a). Fig. 6c shows that IL-1␤ induced a similar increase in Bax mRNA; the increase shown in the sample gel reflected the mean changes, which revealed a significant IL-1␤-induced increase (p Ͻ 0.05; Student's t test for independent means; n ϭ 5). Consistent with the evidence that A␤  induces cell death is the finding that it also increased caspase-3 mRNA in cortical cells (Fig. 6b).
We addressed the question of whether IL-1␤ might mediate some of the actions of A␤ by investigating the effect of A␤  on IL-1␤ concentration and also on JNK phosphorylation, caspase-3 activation, and TUNEL staining in cortical neuronal cells in the presence or absence of the caspase-1 inhibitor, FIG. 6. Effects of A␤ and IL-1␤ on Bax and caspase-3 mRNA in vitro. The cells were treated with A␤ (2 M), total RNA was extracted, and expression of Bax (a) and caspase-3 (b) mRNA was examined by PCR. A␤ evoked a significant increase in Bax and caspase-3 mRNA expression. Representative images of the ethidium bromide-stained gels are shown in the insets. The results are the means Ϯ S.E. for six independent observations. In cells treated with IL-1␤ (5 ng/ml) (c), a significant increase in Bax mRNA expression was observed. A representative image of the ethidium bromide-stained gel is shown as an inset. The results are the means Ϯ S.E. for six independent observations (*, p Ͻ 0.05). Con, control.
Ac-YVAD-CMK. Fig. 7a indicates that incubation of cells in the presence of A␤  significantly increased IL-1␤ in supernatant (p Ͻ 0.05; ANOVA; n ϭ 4), and this effect was inhibited by co-incubation in the presence of Ac-YVAD-CMK. The cultured cells were stained with an antibody that identified the phosphorylated form of JNK, and Fig. 7b shows that A␤  induced a marked increase in the number of cells staining positively for phosphorylated JNK; significantly, this effect of A␤  was abolished by co-incubation of cultures in the presence of A␤  and the caspase-1 inhibitor. Similarly, A␤  markedly increased the number of cells that stained positively for activated caspase-3 (Fig. 7c) and for TUNEL (Fig.  7d), and these changes were abolished by co-incubation of cells in the presence of A␤  and the caspase-1 inhibitor, Ac-YVAD-CMK. In an effort to establish whether JNK activation contributed to the effects induced by A␤  , neuronal cells were treated with A␤  alone or in combination with D-JNKI1. Fig. 7 (c and d) indicates that inhibition of JNK abolished the A␤-associated increase in the number of cells that stained positively for activated caspase-3 and TUNEL.
To confirm the finding that IL-1␤ mimics at least some of the effects of A␤  , cultured cortical neurons were incubated with IL-1␤ in the presence or absence of D-JNKI1 and the cell lysates assessed for activation of c-Jun and for expression of FasL. Fig. 8a indicates that IL-1␤ enhanced c-Jun phosphorylation, and analysis of the mean data indicated that the IL-1␤induced effect was statistically significant (p Ͻ 0.05; ANOVA; n ϭ 2); the increase in c-Jun phosphorylation was abrogated by co-incubation in the presence of D-JNKI1. Similarly, IL-1␤ significantly increased expression of FasL ( Fig. 8b; p Ͻ 0.05; ANOVA; n ϭ 4), and this effect was also abrogated by coincubation in the presence of D-JNKI1. In both cases, equal protein loading was confirmed by reprobing blots for actin expression.

DISCUSSION
We report that A␤  induces an increase in IL-1␤ in hippocampal tissue and in neuronal cell cultures and that this increase, in combination with enhanced activation of JNK, mediates the inhibitory effects of A␤  on LTP in CA1 and the A␤-induced activation of cell death events. Intracerebroventricular injection of A␤  led to a marked increase in phosphorylation of JNK in the hippocampus, coupled with a parallel increase in c-Jun phosphorylation. These observations, which we believe are the first such findings in vivo, bear a marked similarity to those described by Morishima et al. (6), who reported that A␤ induced activation of JNK and c-Jun in cortical neurons. A number of previous studies have indicated that an inverse correlation exists between JNK phosphoryla-tion and LTP expression; for example impaired LTP is coupled with JNK phosphorylation in the hippocampus of aged rats (9), rats treated with lipopolysaccharide (13,26) or IL-1␤ (12), and rats exposed to ␥-irradiation (11). Consistently, LTP is restored when the increase in JNK phosphorylation is blocked, for example by , by inhibition of caspase-1 (13), or by treatment with eicosapentaenoic acid (11). The present findings demonstrate that LTP in area CA1 of the hippocampus was profoundly inhibited by A␤  administration and that this inhibition was abrogated by D-JNKI1, providing another example of the inverse correlation between JNK activation and LTP. The A␤-induced inhibition of LTP supports previous reports in CA1 in vivo (15)(16)(17) and in dentate gyrus in vitro (20,21,27) but provides the first demonstration of an effect of A␤  on LTP that is dependent on JNK activation. The evidence presented indicates that the effect of D-JNKI1 is not immediate but rather becomes evident after 2 h. The mechanism by which A␤ inhibits LTP may derive from the ability of A␤ to induce cell death in hippocampus, and this is supported by the finding that DJNKI1 inhibits these changes and, in parallel, suppresses A␤-induced inhibition of LTP. These findings also indicate a pivotal role for JNK activation in the events triggered by A␤. Thus, we present several findings indicating that the A␤-stimulated increase in JNK phosphorylation is paralleled by several changes that are hallmarks of cell death. For example, A␤ treatment enhanced phosphorylation of c-Jun in the hippocampus, which is a downstream consequence of JNK activation and which has been shown to play a significant role in triggering neuronal apoptosis in a variety of cells in vitro (28 -30). Similarly, increased Bax expression, cytosolic expression of cytochrome c and PARP cleavage, as well as caspase-3 activation and Fas ligand expression were observed in tissue treated with A␤, whereas DJNKI1 prevented all of these actions, suggesting that sequential activation of JNK and c-Jun triggers apoptotic changes in hippocampus. Increased Bax translocation to mitochondria has been identified as an important factor in triggering A␤-induced changes (31)(32)(33)(34), because it reduces the patency of the mitochondrial membrane and leads to the release of proteins normally contained within the intermembrane space, like cytochrome c (35,36); in turn, the presence of cytochrome c in cytoplasm initiates caspase-3 activation, which results in apoptosis (37). Interestingly cell death induced by treatment of neuroblastoma cell lines with A␤  was associated with activation of JNK and caspase-3, and because the apoptotic changes were attenuated by caspase-3 inhibition or by overexpression of a dominant-negative mutant of SEK1, it was concluded that activation of both caspase-3 and JNK significantly contributed to A␤-induced apoptosis in these cells (38). The present findings concur with these data because D-JNKI1 prevented all A␤-induced apoptotic changes investigated, and they are also consistent with several findings indicating that JNK phosphorylation is a pivotal event in the induction of A␤-stimulated cell death (6,7,39).
Several observations have contributed to the development of the idea that FasL expression and consequently Fas activation play a role in neurodegeneration, and we report that increased hippocampal expression of FasL accompanied the A␤-induced increases in JNK phosphorylation, cytochrome c translocation, PARP cleavage, and caspase-3 activation and that these changes were abrogated by D-JNKI1. These data suggest that in the hippocampus a causal relationship between these factors exists that has been shown previously in experimental ischemic injury (28,40,41), Parkinson's disease, and Down's syndrome (42)(43)(44). Specifically, activation of the JNK cascade has been shown to play a significant role in FasL expression that mediates cell death in cortical neurons (6) 3 and 1), which alone exerted no effect (compare lanes 4 and 1). The blots were stripped and reprobed for actin to ensure equal protein loading, and no significant differences in actin expression among the groups were observed (second sample immunoblot in a and b). Con, control. epithelial and lymphoid cells (45,46), whereas Fas-Fc, which prevents Fas binding to FasL, protects cells from apoptotic cell death (47,48). Significantly, increases in JNK and c-Jun phosphorylation and expression of FasL are found in association with apoptotic neurons that are detected in the AD brain (2, 42-44, 49, 50), suggesting that activation of the JNK-c-Jun-FasL signaling cascade may mediate A␤-induced neuronal cell death. Indeed the finding that JNK activation is detected in degenerating neurons in AD brains has led to the hypothesis that JNK activation plays a key role in neuronal loss in AD (51,52).
We demonstrate that A␤ treatment increased in IL-1␤ concentration in hippocampus in vivo and cortical neurons in vitro. Interestingly the A␤-induced increase in IL-1␤ concentration observed in cortical neurons was blocked by co-incubation of cells in the presence of the caspase-1 inhibitor Ac-YVAD-CMK. The A␤-induced increases in activation of JNK and caspase-3 and in TUNEL staining were also inhibited by Ac-YVAD-CMK, consolidating a role for IL-1␤ in A␤-induced changes. A link between A␤ and IL-1␤ has been described in other studies; for example, A␤ has been shown to stimulate production of proinflammatory cytokines, like IL-1␤ from differentiated human monocytes and from a microglial cell line (53), whereas IL-1␤ was induced in reactive astrocytes surrounding A␤-containing deposits in 14-month-old transgenic mice that overexpress human amyloid precursor protein (54). Interestingly, astrocytic overexpression of S100␤, a component of A␤-containing plaques, has been reported to be triggered by IL-1␤ (55). Consistently, several reports have provided evidence demonstrating a role for IL-1␤ in the etiology of AD based largely on the finding that IL-1␤ expression in different brain areas in AD and also in the cerebrospinal fluid of AD patients (56 -58), and it has been shown that a common polymorphism in IL-1B (the gene encoding IL-1␤) is associated with a 4-fold increase in IL-1␤ production and an associated increased risk of the disease (55). Of particular significance are our observations that the inhibitory effect of Ac-YVAD-CMK on A␤-induced changes is closely paralleled by similar effects of D-JNKI1 and that D-JNKI1 also inhibited IL-1␤-triggered changes in c-Jun phosphorylation and FasL expression.
Several studies have revealed that increased hippocampal IL-1␤ concentration, paralleled by increased JNK activation, exerts an inhibitory effect on LTP (9,12), and these changes, both of which are induced by A␤, undoubtedly contribute to the deficit in LTP observed here. Coupled with these changes are the downstream consequences of enhanced IL-1␤ and JNK activation on cell viability, and we therefore conclude that the A␤-induced deficit in LTP is a consequence of activation of cellular cascades induced by IL-1␤ and JNK that lead to cell death.