Akt-dependent Expression of NAIP-1 Protects Neurons against Amyloid-β Toxicity*

Neurotrophins are a family of growth factors that attenuate several forms of pathological neuronal cell death and may represent a putative therapeutic approach to neurodegenerative diseases. In Alzheimer disease, amyloid-β (Aβ) is thought to play a central role in the neuronal death occurring in brains of patients. In the present study, we evaluate the ability of neurotrophin-3 (NT-3) to protect neurons against the toxicity induced by aggregated Aβ. We showed that in primary cultures of cortical neurons, NT-3 reduces Aβ-induced apoptosis by limiting caspase-8, caspase-9, and caspase-3 cleavage. This neuroprotective effect of NT-3 was concomitant to an increased level of Akt phosphorylation and was abolished by an inhibitor of the phosphatidylinositol-3 kinase (PI-3K), LY294002. In parallel, NT-3 treatment reduced Aβ induced caspase-3 processing to control levels. In an attempt to link PI-3K/Akt to caspase inhibition, we evaluated the influence of the PI-3K/Akt axis on the expression of a member of the inhibitors of apoptosis proteins (IAPs), the neuronal apoptosis inhibitory protein-1. We demonstrated that NT-3 induces an up-regulation of neuronal apoptosis inhibitory protein-1 expression in neurons that promotes the inhibition of Aβ-induced neuronal apoptosis. Together, these findings demonstrate that NT-3 signaling counters Aβ-dependent neuronal cell death and may represent an innovative therapeutic intervention to limit neuronal death in Alzheimer disease.

Alzheimer disease (AD) 1 is a neurodegenerative disease characterized by the accumulation of a 40 -42-amino acid-long peptide termed amyloid-␤ (A␤) into amyloid deposits and by the hyperphosphorylation of the protein tau, leading to neurofibrillary tangles (1)(2)(3). A␤ is thought to play a central role in the neuronal death occurring in brains of AD patients. The importance of A␤ is notably supported by the effects of genetic mutations that cause familial AD (4), all of which predispose to amyloid deposition. Moreover, fibrillar A␤ is specifically neu-rotoxic in primary cortical neurons, whereas soluble monomeric A␤ is not (5,6). Neuronal death induced by fibrillar A␤ exposure displays characteristic features of apoptosis including plasma membrane blebbing, nuclear condensation, DNA fragmentation (7), and caspase activation. Caspases are cysteine aspartate proteases regulating the entry of the cell into programmed cell death by two principal pathways: the Fas/TNF receptor-mediated pathway associated with caspase-8 and the mitochondria-dependent pathway linked with caspase-9 (8). The caspases 8 and 9, also called initiator caspases, trigger the activation of the effector caspases. Following A␤ treatment, activation of several caspases has been identified in dying neurons (9 -14), and caspase inhibitors block A␤-induced cell death (11,12,14,15), highlighting the contribution of apoptosis to A␤-induced toxicity. During brain development, where apoptosis plays a fundamental role in regulating cell fate (16), neurotrophins (NTs) antagonize naturally occurring programmed cell death and thus promote neuronal survival in many mammalian species (17). Four related members of the neurotrophin growth factor family have been identified: nerve growth factor, brain-derived neurotrophic factor, neurotrophin-3 (NT-3) and neurotrophin-4/5 (NT-4/5). They act on a set of high affinity tyrosine kinase receptors (18) (i.e. TrkA, TrkB, and TrkC). When activated, Trk receptors are autophosphorylated within tyrosine residues located at the cytoplasmic tail of Trk receptors, which serve as docking sites for kinases such as phosphatidylinositol 3-kinase (PI-3K). Because of their survival promoting properties during development, a widely held concept has emerged: neurotrophins may represent a potential candidate for therapeutic treatment of neurobiological disorders (19). Indeed, experimentally NTs have been shown to attenuate the neuronal death induced by various types of insults (20,21). Recently, a clinical trial with nerve growth factor for the treatment of early onset Alzheimer disease began in 2001 at the University of California (22).
Among these growth factors, NT 3 has raised a particular interest as a potential therapeutic tool for limiting neuronal death during AD. Opposite to brain-derived neurotrophic factor, whose expression is up-regulated in the brains of AD patients, NT-3 expression decreases during adult life (16), and treatment with recombinant NT-3 might offer protection to neurons from the proapoptotic environment observed in AD. In the present study, we investigate how and by which mechanism recombinant NT-3 may protect neurons against A␤-induced apoptotic cell death. 4-isoxazolepropionate, 6-cyano-7-nitroquinoxaline-2,3-dione, kainate, and (ϩ)-5-methyl-10,11-dihydro-5H-dibenzo(a,b)cyclohepten-5,10-imine maleate were from Tocris (Bristol, UK). Laminin was from Invitrogen. Mitogen-activated protein kinase/extracellular signal-regulated kinase kinase inhibitor U0126 was from Promega France (Charbonnieres, France), and recombinant human neurotrophin-3 was obtained from R&D Systems, Inc. (Minneapolis, MN).
Primary Cell Cultures-Mouse cortical cultures of neurons were prepared from 14 -15-day-old embryos as previously described (23), plated in 2-cm 3 dishes previously coated with poly-D-lysine and 0.5 mg/ml laminin. Cultures were performed in Dulbecco's modified Eagle's medium supplemented with 2 mM of glutamine in the presence of 5% fetal bovine serum and 5% horse serum. After 3 days in vitro, neurons were treated with 10 M AraC to inhibit proliferation of nonneuronal cells. Experiments were realized on pure neuronal cultures (Ͼ98% microtubule-associated protein-2 immunoreactive cells) after 12-14 days in vitro.
Estimation of Neuronal Cell Death-Toxicity was determined by estimating the activity of lactate dehydrogenase released in the media by dying neurons and by counting of apoptotic neurons eliciting either nuclear DNA condensation or DNA fragmentation by stained by 4Ј,6Јdiamidino-2-phenylindole dihydrochloride (DAPI) (1 g/ml). Application of DAPI was realized onto paraformaldehyde-fixed neurons for 30 min, and then neuronal cultures were washed three times in phosphate-buffered saline. Apoptotic and total neurons were blind counted, and the ratio between neurons displaying morphological nuclear changes and total neurons stained was performed to assess apoptotisinduced toxicity.
DNA-Agarose Electrophoresis-Genomic DNA was harvested by using the Wizard® Genomic DNA kit (Promega, Charbonnières, France) as described by the manufacturer. DNA amounts were estimated by spectrometry and loaded into 3% agarose electrophoresis gels containing 0.1% ethidium bromide. Gels were visualized by UV illumination and acquired with a CCD camera.
Western Blotting-Cells were harvested in a lysis solution containing 50 mM Tris-HCl (pH 7.6), 1% Nonidet P-40 (Sigma), 150 mM NaCl, 2 mM EDTA with 1 mM phenylmethylsulfonyl fluoride in the presence of a protease inhibitor mixture (Sigma). Cell lysates were centrifuged for 10 min at 12,000 rpm, supernatants were isolated, and the corresponding pellet was resuspended with the protease inhibitor-containing lysis buffer to extract membrane-bound proteins. Plasma membranes were solubilized in lysis buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% Triton X-100, 1 mM EGTA, 0.1% SDS, 1% deoxycholate, 1 mM of phenylmethylsulfonyl fluoride) in the presence of protease inhibitors (Sigma). Once resuspended, membrane lysates were subjected to centrifugation, and the soluble fraction was removed for electrophoresis analysis. Protein amounts were determined by the Bradford protein assay (BCA protein assay; Pierce) and normalized to 20 -30 g of protein/sample. Electrophoreses were done on 12.5-20% SDS-polyacrylamide Tris/glycine gels. Thereafter, gels were transferred to a polyvinylidene difluoride membrane (Polyscreen® membrane; PerkinElmer Life Sciences), membranes were blocked in nonfat milk containing 4% bovine serum albumin and probed with appropriate antibodies. Blots were finally developed with an ECL Western blotting detection system (Western Lightning™ Chemiluminescence Reagent Plus (Enhanced Luminol); PerkinElmer Life Sciences).
Immunoprecipitations-Aliquots (30 g) of cell extracts were diluted to 100 l with dilution buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl) and incubated with 1 l of 22C11 antibody or the R7 antiserum. The mixture was incubated overnight at 4°C and mixed with 20 l of Protein G-Sepharose, Fast Flow® (Amersham Biosciences) for 1 h. The beads were washed twice in Buffer A (50 mM Tris-HCl (pH 7.4), 300 mM NaCl, 0.1% Triton X-100) and twice with Buffer B (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1% Triton X-100), and then proteins were eluted in 50 l of loading SDS-PAGE buffer by boiling.
Fluorescent Immunocytochemistry-Primary cortical neurons cultures were fixed with paraformadehyde and washed, and nonspecific sites were blocked in phosphate-buffered saline containing 4% bovine serum albumin and 0.1% Tween 20 (Sigma) for 30 min and incubated overnight with the primary antibody in phosphate-buffered saline plus 1% bovine serum albumin and 0.1% Tween 20. Cells were then washed and incubated for 1 h with the appropriate secondary biotin-conjugated antibody, secondary Alexa Fluor 488-or Alexa Fluor 555-conjugated antibodies (Molecular Probes Europe, Leiden, The Netherlands).
Antibodies-The following primary antibodies were used: Asp175 Densitometric Analysis-Agarose gels or blots from three independent experiments were acquired by a CCD camera and saved in a resolution of 600 dpi for software analysis. PCR products and Western blot signals were quantified by two-dimensional densitometry analysis with OptiQuant® software (Packard Instrument Co.).
Statistical Analysis-Results are expressed as mean Ϯ S.D. Statistical analysis were performed with StatView (Abacus, Berkeley, CA) by

FIG. 1. NT-3 rescues neurons from A␤-(25-35)-induced apoptotic cell death.
A, neuronal cell death was estimated by measuring the activity of lactate dehydrogenase in the bathing media and by counting apoptotic cells eliciting nuclear alterations revealed by a DAPI staining. Dark bars, lactate dehydrogenase values normalized to neuronal death induced by prolonged NMDA application (at 100 M); empty bars, normalized number of neurons stained with DAPI displaying the characteristic nuclear alterations induced by apoptosis. Statistical analysis was realized by ANOVA (n ϭ 16, p Ͻ 0.001) followed by Bonferroni-Dunn's test. *, p Ͻ 0.01, compared with controls; #, p Ͻ 0.001, compared with A␤ at the maximal doses. B, dose-response effect of NT-3 against A␤-(25-35)-induced apoptosis. Dark bars, A␤-(25-35)treated neurons; light gray bars, co-application of CHX (at 1 g/ml); hatched bars, co-treatment with recombinant human NT-3 (0.1-10 ng/ ml). Normalized number of neurons stained with DAPI displaying nuclear alterations characteristic features of apoptosis. Statistical analysis was realized by ANOVA (n ϭ 16, p Ͻ 0.001) followed by Bonferroni-Dunn's test. *, p Ͻ 0.001, compared with controls; #, p Ͻ 0.01, compared with A␤ at the maximal dose. C, microphotographs of neuronal nuclei stained with DAPI following a 24-h application of A␤ in the presence of NT-3 (at 10 ng/ml) or CHX (at 1 g/ml). Scale bar, 10 m.
one-way variance analysis (ANOVA) followed by Bonferroni-Dunn's test or Student's t test.
NT-3 Attenuates A␤-induced Caspase Cleavage-As previously described (9 -12, 14), A␤ exposure in neurons led to the processing of inactive procaspase-8 and -9 into their active forms. Similarly a 3-4-fold increase of activated caspase-3 (p20) was also detected following A␤ treatment. CHX, a protein synthase inhibitor characterized for its antiapoptotic activity abolished caspase activation induced by A␤ (Fig. 2B). Similarly, NT-3 treatment at 10 ng/ml reduced both A␤-induced neuronal cell death and the activation of caspase-8, caspase-9, and caspase-3 (Fig. 2C). Results are the mean of three independent experiments performed in triplicate. C, densitometry analysis of ERK activation in cultured neurons exposed to A␤-(25-35) (dark bars), with NT-3 (hatched dark bars), or exposed to NT-3 alone (gray bars) for 15 min, 1, 12, and 24 h. Results are the mean of three independent experiments performed in triplicate. Statistical analysis was realized by ANOVA (n ϭ 12, p Ͻ 0.003) followed by Bonferroni-Dunn's test. *, p Ͻ 0.005, compared with controls; double asterisk, p Ͻ 0.001, compared with controls. No statistical differences were observed for neurons subjected to A␤, A␤ ϩ NT-3, or NT-3 alone. D, estimation of neuronal cell death in neurons exposed to A␤-(25-35) (20 M for 24 h) or NT-3 (10 ng/ml) in the presence of PI-3K pathway inhibitor LY294002 (at 1 M). LY294002 was added as 30 min before A␤ incubation and co-applied with A␤ and NT-3 for 24 h. Statistical analysis was realized by ANOVA (n ϭ 12, p Ͻ 0.002) followed by Bonferroni-Dunn's test. *, p Ͻ 0.001, compared with controls; #, p Ͻ 0.002, compared with A␤; double asterisk, p Ͻ 0.001, compared with A␤ ϩ NT-3).
Next we investigated which survival-promoting pathways activated by NT-3 are responsible for the observed neuroprotection. The Ras/mitogen-activated protein kinase pathway, including extracellular-signal regulated kinases 1/2 (ERK1/2) and ERK5, has been identified as one of the survival-promoting pathways used by NTs in the central nervous system (18, 24 -27). To determine the contribution of the ERK family members to NT-3-induced neuroprotection against A␤ neurotoxicity, we used the ERK1/2 inhibitor U0126 (at 10 M), which also inhibits ERK5 activation (26). The co-treatment of NT-3 with U0126 did not abolish the neuroprotective effect of the neurotrophin, indicating that ERKs do not contribute to this process (Fig. 3A).
To confirm these pharmacological data, we estimated ERK1/2 activation by performing a Western blotting of the proteins recovered from neurons exposed to A␤ at 15 min, 1 h, 12 h, and 24 h in the presence of NT-3. As previously reported, NT-3 alone induced a transient phosphorylation of ERKs 15 min and 1 h after its incubation (Fig. 3, B and C). A␤ treatment alone led to an activation of ERKs, which was slightly enhanced by co-application of NT-3 without, however, being significant. Whereas phospho-ERKs were markedly detected during the first 1 h of treatment, they declined to basal levels at either 12 or 24 h (Fig. 3C).
NTs have also been described as a putative activator of the PI-3K/Akt pathway. This pathway has been the subject of emerging interest over the past years for its ability to reduce A␤-induced apoptosis in PC12 cells or in neuroblastoma cell lines (28,29). To determine whether NT-3 mediates its protective effect against A␤-induced neurotoxicity in primary cultured cortical neurons through the activation of the PI-3K/Akt pathway, we first tested the influence of the PI-3K pathway inhibitor, LY294002. Because several studies have shown that the LY294002 alone induces apoptosis in primary neuronal culture, we chose a concentration of LY294002 that did not show any proapoptotic activity on its own (at 10 M). The co-application of LY294002 reversed the neuroprotection promoted by NT-3, whereas LY294002 alone did not exert any detrimental effect (Fig. 3D). Accordingly, Western blotting revealed that Akt is rapidly phosphorylated and remained activated under NT-3 exposure alone or in the presence of A␤ for 12 or 24 h (Fig. 4A). The addition of LY294002 (pretreatment plus co-treatment) prevented NT-3-induced phosphorylation of Akt (Fig. 4B). Finally, neurons exposed to A␤ and LY294002 did not exhibit phospho-Akt levels different from controls.
Under the same experimental conditions, active caspase-3 levels (p20) were reduced in the conditions when phospho-Akt Results are representative of three independent experiments performed in triplicate. Densitometry analysis of phospho-AKT/AKT in cultured neurons exposed to A␤-(25-35) (dark bars), co-treated with NT-3 (hatched dark bars), or exposed to NT-3 alone (gray bars) for 15 min, 1 h, 12 h, and 24 h. Results are the mean of three independent experiments performed in triplicate. Statistical analysis was realized by ANOVA (n ϭ 12, p Ͻ 0.003) followed by Bonferroni-Dunn's test. *, p Ͻ 0.005, compared with controls; double asterisk, p Ͻ 0.001, compared with controls. No statistical differences were observed in neurons subjected to A␤, A␤ ϩ NT-3, or NT-3 alone. B, Western blot of phosphorylated Akt, total Akt, and activated caspase-3 (p20) in cultured neurons exposed to A␤-(25-35) (20 M for 24 h) in the presence of either CHX (1 g/ml) or NT-3 (10 ng/ml). Results are representative of three independent experiments performed in triplicate. Densitometry analysis of cleaved caspase-3 levels in cultured neurons exposed to A␤-(25-35) with NT-3 in the presence of LY294002 or exposed to NT-3 alone for 24 h. Results are the mean of three independent experiments performed in triplicate. Statistical analysis was realized by ANOVA (n ϭ 12, p Ͻ 0.001) followed by Bonferroni-Dunn's test. *, p Ͻ 0.005, compared with controls; #, p Ͻ 0.001, compared with A␤. C, immunodetection of phosphorylated Akt (red fluorescence) in cultured neurons exposed to A␤-(25-35) (20 M for 24 h) in the absence or the presence of NT-3 (10 ng/ml). Neuronal nuclei were counterstained with DAPI (blue fluorescence). Overlaid images are presented in upper panels, and phospho-Akt immunoreactivity is displayed below to better appreciate its subcellular localization. Scale bar, 25 m.

NAIP-1 Protects Neurons against A␤
was detected (Fig. 4B). Conversely, immunocytological studies reveal that neurons eliciting phospho-Akt473 immunoreactivity did not exhibit characteristic apoptotic nuclear alterations (Fig. 4C), suggesting that NT-3-mediated Akt activation rescued cortical neurons from A␤-induced apoptosis. In addition, we determined by immunoblotting the level of activated initiator caspases in neurons exposed to A␤ in the presence of NT-3 and LY294002 (Fig. 5). NT-3 treatment (at 10 ng/ml) lowered both levels of activated caspase-8 and -9, whereas co-application of LY294002 restored levels of the proapoptotic forms of caspases to those observed in neurons exposed to A␤ alone.
Activation of the PI-3K/Akt Pathway Induces NAIP-1 Expression-PI-3K/Akt signaling pathway has been described as a putative activator of proteins involved in regulating cell survival in various cell systems (30). Among these potential targets of the PI-3K/Akt pathway, members of the inhibitors of apoptosis proteins (IAPs), a family of proteins that directly inhibit caspase function (i.e. XIAP and HIAP (human IAP), and NAIP (neuronal apoptosis inhibitor protein)) have raised particular interest as a potential target for antiapoptotic drug treatment (31). Indeed, it has been shown in different cell systems that Akt activates the expression of survival IAP genes (32)(33)(34)(35)(36). Thus, we evaluated the influence of PI-3K/Akt activation on the expression of the neuronal member of IAPs, NAIP-1 (Fig. 6). First, we performed double immunochemistry experi-ments in primary cortical neurons exposed to A␤ that revealed a co-localization of NAIP-1 with phospho-Akt-immunoreactive neurons when treated NT-3 (Fig. 6A). To confirm this link between the PI-3K/Akt pathway and NAIP up-regulation, we estimated the level of expression of NAIP-1 by Western blotting (Fig. 6B). We observed a ϳ2-2.5-fold increase in NAIP-1 expression induced by NT-3 alone. This up-regulation of NAIP-1 expression induced by NT-3 was also observable when experiments were performed in the presence of A␤. This effect was abolished by the co-application of LY294002, an inhibitor of the PI-3K pathway.
Because the antiapoptotic activity of IAP proteins is inhibited by a mitochodria-derived activator of caspase, the protein SMAC (also called Diablo), we performed immunoprecipitations with an antibody raised against NAIP-1, and we revealed the blots with an antibody raised against SMAC to determine whether NT-3 influences its interaction with SMAC (Fig. 6C). We found that A␤ application resulted in the formation of a Smac-NAIP-1 complex in the cytosol, whereas SMAC was nearly undetectable in controls. Co-treatment of NT-3 prevented formation of such a complex. Of note, NT-3-induced NAIP-1 up-regulation was confirmed in these conditions. In addition, co-incubation with LY294002 restored the formation of the deleterious assembly of NAIP-1-SMAC. To resume, these results suggest that the antiapoptotic effect of NT-3 on A␤- Results are representative of three independent experiments performed in triplicate. Densitometry analysis of NAIP-1/actin (p42) ratio in cultured neurons exposed to A␤-(25-35) (dark bars) co-treated with NT-3 (hatched bars) in the presence or not of LY294002 or exposed to NT-3 alone for 24 h (empty bar). Results are the mean of three independent experiments performed in triplicate. Statistical analysis was realized by ANOVA (n ϭ 9, p Ͻ 0.004) followed by Bonferroni-Dunn's test. *, p Ͻ 0.004, compared with A␤; #, p Ͻ 0.002, compared with A␤ ϩ NT-3). C, immunoprecipitation of NAIP-1 from the same protein extracts used in B revealed either with an antibody raised against SMAC or NAIP-1. The immunoprecipitation of NAIP-1 revealed with NAIP-1 confirmed results presented in B, showing an increased expression of NAIP-1 under NT-3 treatment with or without A␤- (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35).
induced neurotoxicity could be due to both an increased expression of NAIP-1 and a reduced interaction between NAIP-1 and SMAC by a mechanism dependent on PI-3K/Akt activation.

NT-3 Protects Neurons from A␤-(1-42)-induced
Neurotoxicity-In the brain of AD patients, A␤-(1-42) has been identified as a potential source of neuronal dysfunction and toxicity. To test whether the NT-3 induced neuroprotective effect against the toxicity induced by the artificial peptide A␤- (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) was also present in neurons exposed to A␤-(1-42), we evaluated by Western blotting levels of active caspase-3 (Fig. 7, A and B) and nuclear morphology (Fig. 7B), following the incubation of aggregated A␤-(1-42) (at 20 M) in the presence of 10 ng/ml NT-3. Both methods confirmed that NT-3 application reduced the neurotoxicity of A␤- . Indeed, incubation of NT-3 lowered the level of active caspase-3 and the presence of neuronal nuclear DNA alterations. As previously described with A␤- (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), treatment with LY294002 abolished the neuroprotective activity of NT-3, whereas the application of U0126 was not able to reverse NT-3 neuroprotective effects (Fig. 7C). DISCUSSION AD is a neurodegenerative disease characterized by the abnormal accumulation of small aggregative peptides (A␤s) in brain tissue that leads to severe cortical dystrophy (4). To determine the biochemical pathways involved on the neurotoxic action of aggregated A␤, studies on primary cell culture systems have been extensively used. Various deleterious mechanisms have been identified to support A␤-induced cell toxicity, and there is now a large consensus in the scientific community that application of aggregated A␤ induces neuronal apoptosis in primary neuronal cultures (37)(38)(39). However, no valid therapeutic approaches have emerged from these studies. Here, we describe that a growth factor, NT-3, is able to reduce the proapoptotic action of aggregated A␤ on primary cortical neurons. This effect is mediated by the activation of PI-3K/Akt pathway and results in the reduction of A␤-induced caspases activation. In addition, we identified the mechanism by which NT-3 limits caspase activation; NT-3 induces both an increased expression of NAIP-1, a neuronal form of IAPs that can directly inhibit caspases, and a reduced interaction between NAIP and its endogenous inhibitor SMAC. Because IAPs has been shown to reduce activation of caspase-3 and -9 in many systems, it is likely that the effect of NT-3 is downstream of caspase-8 activation.
Significant progress has been made in identifying the molecular mechanisms causing neuronal cell death in AD. Rare genetic cases of AD that develop an early onset have provided mechanistic insights relevant to most common sporadic cases. This has led to the hypothesis that familial and probably sporadic AD is initially caused by an abnormal accumulation of A␤. Although the role of oligomeric forms of A␤ has been recently reappraised (40,41), it is generally believed that aggregated forms of A␤ are responsible for the neurotoxic actions observed in AD. Here we showed that aggregated A␤ triggers caspase-dependent apoptosis in primary cortical neurons. Similar results were obtained by others in primary culture systems (9 -11, 13, 14). More recently, analyses from post-mortem tissue from patient, transgenic animal and cell culture system experiments have confirmed the putative implication apoptosis in the pathophysiology of AD.
A possible therapeutic approach of AD is to develop strategies to block the apoptotic trigger or to develop antiapoptotic strategies. Because neurotrophins have been identified as one of the most powerful antiapoptotic agents in the CNS, it was tempting to test the influence of NT-3 against A␤-induced apoptosis. We report that NT-3-induced neuroprotection against A␤-driven apoptosis requires the activation of the PI-3K/Akt pathway, which induces an up-regulation of the expression of NAIP-1, a protein that directly inhibits caspase-3 and -7 (42). The PI-3K/Akt has already been identified as second messenger system providing antiapoptotic signal to various cell types (30). Upon activation of this signaling pathway, the induction of the IAP family members, including XIAP and HIAP (also called ITA) by PI-3K/Akt, has been reported in different cell types (32,43,44). Recently, Zhang et al. (45) have demonstrated that p75 NTR -induced neuroprotection against extracellular A␤ toxicity was dependent on PI-3K but not Akt in primary human neurons. Their data and ours strengthen the importance of the PI-3K/Akt pathway in A␤-induced neurotoxicity in primary neuronal cultures.
Finally, we reported that whereas A␤ application promotes the formation of a Smac-NAIP-1 complex enhancing caspasedependent apoptotic pathways, NT-3 prevents this association. Thus, by enhancing NAIP-1 expression, NT-3 treatment decreases caspase activation and reduces A␤-induced neurotoxicity. The NAIP gene encodes for a protein of 156 kDa with a  (1-42). A, Western blot analysis of cleaved caspase-3 (p20) expression in cultured neurons exposed to aggregated A␤-(1-42) (20 M for 24 h) in the presence of NT-3 (10 ng/ml) with PI-3K pathway inhibitor LY294002 (1 M). ␤-tubulin levels are displayed above as loading control. Results are the mean of three independent experiments performed in triplicate. B, immunolabeling of cleaved caspase-3 (p20) (green fluorescence) in cultured neurons exposed to aggregated A␤-(1-42) (20 M for 24 h) in the presence of NT-3 (10 ng/ml) with PI-3K pathway inhibitor LY294002 (1 M). Neuronal nuclei were counterstained with propidium iodine (red fluorescence). Scale bar, 50 m. C, estimation of apoptotic cells in cultured neurons shown in B. Statistical analysis was realized by ANOVA (n ϭ 12, p Ͻ 0.001) followed by Bonferroni-Dunn's test. *, p Ͻ 0.001, compared with controls; #, p Ͻ 0.002, compared with A␤). strong homology to the other inhibitor of apoptosis proteins (42). An antiapoptotic effect of NAIP and other members of the human IAP family has been shown in many cell culture systems. However, the mode of action of NAIP is not as well understood as its family homologues. MacKenzie's group (46) recently pointed out that NAIP overexpressed in HeLa cells was able to bind to caspase-9 in the presence of ATP in an SMAC-independent manner. Although we do not know whether NT-3-induced up-regulation of NAIP-1 led to a binding with caspase-9, under our conditions, we report that endogenous NAIP-1 is able to complex with SMAC in primary cortical neurons subjected to A␤ fibrils. To summarize, our results provide a strong support for a model in which NAIP is playing a critical role in regulation of A␤-induced neuronal apoptosis by directly inhibiting effector caspases.
In addition to the regulation of NAIP-1 expression by the PI-3K/Akt axis, this same signaling pathway can also affect other proteins that are able to inhibit apoptosis, including the FOXO family of Forkhead transcription factors (47,48). Indeed, once activated, Akt can directly phosphorylate all three of FOXO members and trigger their translocation out of the nucleus, preventing these factors from inducing the transcription of proapoptotic genes (49,50). Indeed, the links of both apoptosis-inhibitory pathways to Akt activation could converge and potentially be additive in preventing A␤-induced neuronal death.
Nerve growth factor and brain-derived neurotrophic factor expression are increased in brains of AD patients (51)(52)(53). Such up-regulation may be related to physiopathology of AD and may correspond to a stress response from brain cells. The recent discovery that a neurotrophin precursor protein and its proteolytically processed products may activate pro-and antiapoptotic cellular responses questioned the beneficial consequences of such increased expression (54). Opposite to nerve growth factor and brain-derived neurotrophic factor, NT-3 is mainly expressed during development and is not present in the adult (51,(55)(56)(57). In the present study, we reduced neuronal vulnerability to aggregated A␤ by activating NT-3 signaling through a molecular mechanism involving NAIP increased expression. Thus, we postulate that the stimulation of NAIP expression into neuronal cell exposed to proapoptotic stimuli may represent an innovative therapeutic strategy to limit the extent of neuronal lesion occurring in the brain of an AD patient.