Address correspondence and reprint requests to Seigo Tanaka, Laboratory of Molecular Clinical Chemistry, Institute for Chemical Research, Kyoto University, Uji, Kyoto 611–0011, Japan. E-mail: firstname.lastname@example.org
Non-amyloid beta (Aβ) component of Alzheimer's disease (AD) amyloid (NAC) coexists with Aβ protein in senile plaques. After exposure to NAC fibrils, cortical neurons of rat brain primary culture became apoptotic, while astrocytes were activated with extension of their processes. NAC fibrils decreased the activity of reducing 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in cortical neurons more markedly (IC50 = 5.6 µm) than in astrocytes (IC50≈ 50 µm). The neuron-specific toxicity of NAC fibrils was indicated also by an increased release of lactate dehydrogenase from the cells. Neuronal apoptosis was suppressed by pre-treatment with the antioxidants, propyl gallate (PG) and N-t-butyl-phenylnitrone (BPN), or overexpression of human Bcl-2. Exposure to NAC fibrils enhanced generation of reactive oxygen species (ROS) in neurons and less efficiently in astrocytes, as demonstrated by oxidation of 2′,7′-dichlorofluorescin. The site of ROS generation was shown to be mitochondria by oxidation of chloromethyl-tetramethyl rosamine. Exposure to NAC fibrils increased also the nuclear translocation of nuclear factor kappa B (NF-κB) and enhanced its DNA-binding activity, which was inhibited by PG and BPN more efficiently in neurons than in astrocytes. These results suggest that NAC fibrils increase mitochondrial ROS generation and activate NF-κB, thereby causing a differential change in gene expression between neurons and astrocytes in the AD brain.
non-Aβ component of Alzheimer's disease amyloid precursor
nuclear factor kappa B
reactive oxygen species.
Amyloid deposition in senile plaque cores is one of the histopathological changes characteristic of Alzheimer's disease (AD). The AD amyloid consists of Aβ protein and many minor substances, including the non-Aβ component of AD amyloid (NAC; Uéda et al. 1993). Immunohistochemical studies have shown that NAC is located more abundantly in the central portion than the peripheral portion of senile plaque cores, whereas Aβ is distributed homogeneously throughout cores (Masliah et al. 1996). NAC has been demonstrated to bind to Aβ and stimulate its aggregation invitro (Yoshimoto et al. 1995). NAC is a hydrophobic peptide that displays an amyloidogenic β-sheet structure (Iwai et al. 1995a), and consists of at least 35 amino acids derived from a precursor protein, termed NACP (Uéda et al. 1993). NACP was later identified as α-synuclein, a pre-synaptic protein of the synuclein family (Iwai et al. 1995b). This family comprises α-, β- and γ-synucleins, among whichonly α-synuclein has the NAC domain. The mechanism of NAC production from α-synuclein remains to be elucidated.
Oxidative stress has been claimed to play a central role in the pathogenesis of AD (Smith et al. 2000), and a number of sources of ROS have been suggested, including activated glial cells (Colton and Gilbert 1987), accumulated metals (Smith et al. 1997), and mutated mitochondria (Davis et al. 1997). The Aβ protein is another potential source of ROS; Aβ itself generates ROS in vitro (Hensley et al. 1994) and also induces intracellular production of ROS (Behl et al. 1994). The idea of ROS involvement in the neurotoxicity of Aβ protein is supported by the protective action of Bcl-2 (Sailléet al. 1999) which decreases the formation of ROS and suppresses oxidative stress in neuronal cell lines (Kane et al. 1993).
Oxidative stress is known to activate the transcription factor nuclear factor kappa B (NF-κB) in a neuronal cell line (Behl et al. 1994). This factor is ubiquitously expressed in both neurons and glial cells in the brain (O'Neill and Kaltschmidt 1997), and its activation has been demonstrated in neurons and astrocytes around the senile plaques in AD brains (Kaltschmidt et al. 1997). The activation was also observed in cultured neurons, astrocytes (Akama et al. 1998), and microglia (Bonaiuto et al. 1997) after exposure to Aβ protein.
Recently, the neurotoxicity of NAC and α-synuclein was reported (El-Agnaf et al. 1998; Liu and Schubert 1998); however, the molecular mechanism(s) involved have yet to be elucidated. Here we report that NAC as well as Aβ fibrils induce mitochondrial ROS generation and lead to apoptosis of neurons, whereas both fibrils activate astrocytes in rat brain primary cultures. It is of interest that, although ROS could be a potent activator of NF-κB in both neurons and astrocytes, the ultimate cellular responses are different between the two cell types.
Materials and methods
Synthesis of amyloid peptides
NAC peptide of 35 amino acids [61–95 of α-synuclein/NACP sequence (Uéda et al. 1993)] and reversed NAC peptide (95–61) were synthesized by the Fmoc-solid-phase synthesis method on a Rink amide resin with a peptide synthesizer (PSSM-8; Shimadzu, Kyoto, Japan). The protected peptide resin of NAC or reversed NAC was treated with trifluoroacetic acid containing 5% 1,2-ethanedithiol for 1.5 h at room temperature, and purified by high-performance liquid chromatography (HPLC, D-6500; Hitachi, Tokyo, Japan) on an ODS column (Cosmosil 5C18-ARII; Waters, Milford, MA, USA). The fidelity of the amino acid sequence produced was ascertained by time-of-flight mass spectrometry (TOFMS, Voyager-DE STR; Applied Biosystems, Foster City, CA, USA). Aβ1–40 peptide was prepared by the same method.
Amyloid fibril formation
Synthesized NAC and Aβ1–40 peptides were dissolved in borate buffer (pH 9.1) or distilled water, respectively, at the concentration of 500 µm, followed by dilution with phosphate-buffered saline (PBS) as specified. These peptide solutions were incubated for 24 h at 37°C to allow for amyloid fibril formation, that was confirmed by electron microscopy (H-7000; Hitachi) as previously reported (Naiki and Nakakuki 1996).
Primary cell cultures of rat brain
Primary cell cultures of rat brain were prepared from Wister rat embryos (embryonic day 15). Cerebral cortices were dissected in cold Leibovitz's L-15 medium (Gibco-BRL, Rockville, MD, USA) and digested with papain (10 units/mL) of Earle's balanced salt solution (Gibco-BRL) containing 20 mm glucose for 20 min at 30°C. Then, the cortical tissues were treated with 1% DNase I, dissociated by repeated pipettings, and filtered through a Falcon cell strainer (Becton Dickinson Labware, Franklin Lakes, NJ, USA). The cells were plated at a density of 1 × 106/mL on polyethylenimine-coated dishes, and maintained in Dulbecco's modified Eagle medium 2 (Nissui Seiyaku, Tokyo, Japan) supplemented with 10% fetal bovine serum, penicillin (100 units/mL), streptomycin (0.1 mg/mL), 0.5 mm glutamine and 0.5 × Serum Extender (Becton Dickinson Labware) containing 10 mm glucose. For developing neuronal cell cultures, the cells were incubated with 4 µm cytosine arabinoside (araC; Sigma, St Louis, MO, USA) for 7 days to inhibit the growth of glial or endothelial cells. Before NAC or Aβ1–40 treatment, the medium was replaced with araC-free medium. For developing glial cell cultures, the cells were incubated in the araC-free medium for 10 days.
Morphological analysis of neurons and astrocytes
Primary cell cultures were incubated with NAC or Aβ1–40 fibrils for 2–36 h at 37°C. After fixation with acetone/methanol (1 : 1), cells were double-stained with the fluorescent DNA-binding dye, Hoechst 33258 (Sigma), and mouse monoclonal anti-synaptophysin antibody (Dako, Copenhagen, Denmark) for neurons or rabbitpolyclonal anti-glial fibrillary acidic protein (GFAP) antibody (Dako) for astrocytes, followed by detection with fluorescein isothiocyanate (FITC)-labeled rabbit polyclonal anti-mouse IgG antibody (Dako) or rhodamine-labeled swine polyclonal anti-rabbit IgG antibody (Dako), respectively. Nuclear morphology was observed under a fluorescent microscope (Axioskop; Carl Zeiss, Oberkochen, Germany) with a UV wide-range filter. Cells containing large nuclei with uniformly stained chromatin were considered to be live, while cells with condensed chromatin or fragmented nuclei were considered to be apoptotic. The number of live or apoptotic cells among synaptophysin- or GFAP-positive cells (total 200–300 cells) was counted in at least three different high-power (×200) fields for each experiment, and the percentage of apoptotic cells was calculated.
MTT reduction assay and LDH release assay
Cell viability was estimated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay (Mosmann 1983). After incubation of neuronal cell cultures with NAC or Aβ1–40 fibrils for 2–36 h, the culture medium was changed, and fresh medium without amyloid fibrils was added, followed by the addition of 0.1 volume of 5 mg/mL MTT (Sigma) solution. After another incubation for 1 h at 37°C, an equal volume of isopropanol containing 40 mm HCl was added to dissolve the MTT formazan products. The absorbance was measured by a spectrophotometer (UV-1200, Shimadzu) at 570 nm. For the dose–response analysis of cytotoxicity, neuronal as well as glial cell cultures were incubated with NAC or Aβ1–40 fibrils at various concentrations for 24 h and subjected to the MTT reduction assay. Reversed NAC peptide was also incubated for 24 h at 37°C, and added to the cell cultures for comparison.
Cytotoxicity was also evaluated by measuring the release of lactate dehydrogenase (LDH) into the medium (Decker and Lohmann-Matthes 1988). Neuronal and glial cell cultures were incubated with amyloid fibrils for 24 h in medium containing 5% fetal bovine serum. The LDH activity in the culture medium was determined using a colorimetric LDH-Cytotoxicity Assay Kit (BioVision, Mountain View, CA, USA) according to the manufacturer's protocol. The assay is based on coupled enzymatic reactions: LDH oxidizes lactate to pyruvate with conversion of NAD+ to NADH, which then reacts with iodonitrotetrazolium to form formazan. The absorbance was measured by a microplate reader (model 550; Bio-Rad Laboratories, Hercules, CA, USA) at 490 nm.
Evaluation of antioxidants and quantification of ROS
The role of ROS in the cytotoxic action of amyloid was examined by pre-treatment of cells with a water-soluble antioxidant, propyl gallate (PG; Sigma), or a lipophilic spin trapping agent, N-t-butyl-phenylnitrone (BPN; Sigma). Five micromols of PG or 50 µm of BPN were added to the culture medium 1 h before the addition of amyloid fibrils. After a 24-h incubation with 10 µm NAC or Aβ1–40 fibrils, cells were stained with Hoechst 33258, and the number of live or apoptotic cells was counted as above.
Intracellular generation of ROS was measured using a fluorogenic probe, 2′,7′-dichlorofluorescin diacetate (also known as 2′,7′-dichlorodihydrofluorescein diacetate; DCFH–DA; Molecular Probes, Eugene, OR, USA; Royall and Ischiropoulos 1993). DCFH–DA is intracellularly deacetylated to 2′,7′-dichlorofluorescin (DCFH) and then oxidized by hydrogen peroxide to a fluorescent compound, 2′,7′-dichlorofluorescein (DCF). After a 4-h exposure to NAC or Aβ1–40 fibrils, neurons and astrocytes (5 × 106) were incubated with 10 µm DCFH–DA in phenol red-free medium for 1 h at 37°C. Cells were washed twice with PBS, collected by scraping, and lysed in sonication buffer [50 mm potassium phosphate, 0.1 mm EDTA, and 0.1% CHAPS (pH 7.0)]. The cell suspension was sonicated on ice five times for 5 s each with 30-s intervals by an ultrasonic homogenizer (Astrason W-385; Misonix, Farmingdale, NY, USA). The cell lysate was centrifuged at 12 000 g for 10 min at 4°C. The concentration of DCF in the supernatant was measured by a fluorescence spectrophotometer (F-2000; Hitachi) with excitation at 502 nm and emission at 523 nm.
Detection of ROS in mitochondria
The generation of ROS in mitochondria was determined using reduced chloromethyl–tetramethyl rosamine (MitoTracker® Red; Molecular Probes). This compound accumulates in the mitochondria and is oxidized by hydrogen peroxide to its fluorescent form (Whitaker et al. 1991). After a 4-h exposure to NAC or Aβ1–40 fibrils, cells were incubated with 500 nm of reduced rosamine in phenol red-free medium for 45 min at 37°C, washed twice with PBS, and fixed with acetone/methanol (1 : 1). Images of cellular fluorescence were acquired using a confocal laser-scanning microscope (LSM 510; Carl Zeiss) with a krypton/argon laser and a 590-nm bandpass filter. The intensity of the laser beam and the sensitivity of the photodetector were held constant to allow quantitative comparisons of relative fluorescence intensity of oxidized rosamine.
Adenovirus-mediated gene transduction of human Bcl-2
The Cre/loxP system (Kanegae et al. 1995) was used for the adenovirus-mediated gene transduction of human Bcl-2. The recombinant adenovirus, AxCALNL-hBcl-2, was produced as reported previously (Shinoura et al. 1999). The on/off switching unit CALNL-hBcl-2 consisted of the CAG promoter, the neor gene, and polyadenylic acid sequence flanked by a pair of loxP sites, the human Bcl-2 cDNA, and another polyadenylic acid sequence. In the presence of NCre, that contains the Cre recombinase, the neor gene was excised at the loxP sites, and then Bcl-2 cDNA was expressed under the control of the CAG promoter. The AxCALNL–hBcl-2 was co-infected with AxNCre at a multiplicity of infection (MOI) ratio of 4 : 1. Expression of human Bcl-2 was observed by immunostaining with mouse monoclonal anti-Bcl-2 antibody (Transduction Laboratories, Lexington, KY, USA), followed by detection with FITC-labeled rabbit polyclonal anti-mouse IgG antibody (Dako). The number of live or apoptotic cells among Bcl-2-positive or -negative cells (total 200–300 cells) was counted in at least three different high-power (×200) fields for each experiment, and the percentage of apoptotic cells was calculated.
Electrophoretic mobility shift assay of NF-κB
Nuclear extracts were prepared from HeLa cells after a 3-h exposure to 10 ng/mL tumor necrosis factor (TNF)-α (Roche Diagnostics, Basel, Switzerland) and also from cortical neurons and astrocytes after a 24-h exposure to 20 µm NAC or Aβ1–40 fibrils, according to previously published method (Dignam et al. 1983) with some modifications. Treated cells (5 × 106) were washed twice with cold PBS and scraped into cold lysis buffer [10 mm HEPES, 10 mm KCl, and 1.5 mm MgCl2 ( pH 7.9)]. Cells were pelleted by centrifugation, and lysed by suspending in 20 µL of lysis buffer containing 0.1% Nonidet P-40 for 10 min on ice. The cell lysate was vigorously mixed and centrifuged at 12 000 g for 5 min at 4°C. The nuclear pellet was washed once with the lysis buffer without Nonidet P-40, and centrifuged again at 12 000 g for 5 min at 4°C. The pellet was resuspended in 20 µL of protein extraction buffer [420 mm NaCl, 20 mm HEPES, 1.5 mm MgCl2, 0.2 mm EDTA, and 25% glycerol (pH 7.9)] and placed on ice for 10 min. The suspension was vigorously mixed and centrifuged as above. The supernatant was saved and centrifuged again. The supernatant (nuclear extract) was diluted with 30 µL of dilution buffer [50 mm KCl, 20 mm HEPES, 0.2 mm EDTA, and 20% glycerol (pH 7.9)], and the protein concentration was measured using the BCA protein assay kit (Pierce Chemical, Rockford, IL, USA). All buffers described above were supplemented with a mixture of protease inhibitors [0.5 mm (for lysis or extraction) or 0.2 mm (for dilution) phenylmethylsulfonyl fluoride (PMSF), and 10 µg/mL leupeptin] and 0.5 mm dithiothreitol (DTT). The extracted nuclear proteins were stored at − 80°C until use for assay.
For the binding assay, nuclear extracts (5 µg each as protein) were suspended in 20 µL of reaction buffer [25 mm Tris (pH 7.5), 50 mm NaCl, 1 mm DTT, 2 mm EDTA, 5% (v/v) glycerol, 3 mm MgCl2, 0.05 mg/mL poly(dI-dC), 0.01% Nonidet P-40 and 50 µg/mL bovine serum albumin]. For the supershift assay, nuclear extracts were incubated on ice for 1 h with 1 µL of antibody against NF-κB subunit p50, p65 (Rel A; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or DNA-PK Ku antigen (Sigma) before suspension in the reaction buffer. All sample mixtures were preincubated for 15 min at room temperature before incubation with a 32P-labeled double-stranded oligonucleotide shift probe (approximately 50 000 cpm; Promega, Madison, WI, USA) for 30 min at room temperature. The shift probe contained a tandem repeat of the consensus sequence for the NF-κB DNA binding site, GGGGACTTTCC. After incubation, all samples were electrophoresed in 6% non-denaturing polyacrylamide gel at 150 V for 2 h. Electrophoresis was carried out at 4°C in TGE buffer [50 mm Tris–HCl (pH 8.5), 380 mm glycine, and 2 mm EDTA]. Gels were dried and subjected to autoradiography. Images of exposed films were acquired by Photoshop 5.0 (Adobe, San Jose, CA, USA) using an image scanner (ES-2000; Epson, Tokyo, Japan). The density of bands was measured using NIH Image 1.61.
Induction of neuronal apoptosis and astrocytic activation by NAC fibrils
Primary cell cultures of rat brain were used to examine the effects of NAC fibrils on neurons and astrocytes. After incubation with 4 µm araC for 7 days, almost all cells extended neurites and were positively stained with anti-synaptophysin antibody. On the other hand, after a 10-day incubation in the araC-free medium, a majority, 80–90%, of the cells were identified as astrocytes by immunostaining with antibody against GFAP, an intermediate filament protein specific for astrocytes (Raff 1989). Both type-1 astrocytes, flat and polygonal without processes, and type-2 astrocytes, stellate-shaped with a number of processes, existed (data not shown).
After incubation of NAC peptide solution for 24 h at 37°C, NAC aggregates emerged, showing birefringence with Congo red staining under polarized light and green fluorescence with thioflavin T staining, which is characteristic of amyloid. Electron microscopy of the aggregates revealed formation of fibrils (data not shown). After the addition of 1 or 10 µm NAC fibrils thus formed to the culture medium, rat cortical neurons showed features of apoptosis, i.e. chromatin condensation and nuclear fragmentation, as revealed by Hoechst 33258 staining (Fig. 1a). Signs of apoptosis first emerged at 8 h and reached the maximum by 24 h. Exposure of cortical neurons to 10 µm NAC fibrils initiated apoptosis in 31.5% of the cells (Fig. 1c). On the other hand, few astrocytes, either type-1 or type-2, became apoptotic even after a 24-h incubation (Figs 1b and c), indicating that induction of apoptosis by NAC fibrils is specific for neurons. Of the two types of astrocytes, type-2 astrocytes was activated and extended their processes. The effects of NAC fibrils on neurons and astrocytes resembled both qualitatively and quantitatively the effects of Aβ fibrils (Fig. 1a–c).
Neurotoxic effects of NAC fibrils
NAC fibrils decreased the MTT-reducing activity of cortical neurons in a time- and dose-dependent manner, whereas reversed NAC peptide did not (Figs 2a and b). The effect became evident after a 2-h exposure to NAC fibrils and reached a maximum level in 24 h (IC50 = 5.6 µm). NAC fibrils also decreased the MTT-reducing activity of astrocytes, but much less efficiently (IC50 ≈ 50 µm). Neuron-specific toxicity of NAC fibrils was further demonstrated by the LDH release assay. NAC fibrils induced a release of LDH from cortical neurons, but not astrocytes, into medium in a dose-dependent manner (Fig. 2c). These findings were comparable with neuron-specific induction of apoptosis (see above).
Intracellular generation of ROS by NAC fibrils
Among various chemicals tested, antioxidants, such as PG and BPN, were found to protect cortical neurons against the toxicity of NAC fibrils. While the effect of BPN was marginal, PG suppressed significantly the induction of neuronal apoptosis (Fig. 3a), indicating a role of ROS in NAC neurotoxicity. In contrast, antioxidants had no effects on the NAC-induced activation of astrocytes (data not shown). The intracellular generation of ROS was measured using the fluorogenic probe, DCFH–DA. The intracellular production of DCF, that is a measure of ROS generation, was less in cortical neurons than in astrocytes before the addition of amyloid fibrils. After NAC treatment, however, ROS generation was more abundant in neurons than in astrocytes. Exposure to 10 µm NAC fibrils for 4 h resulted in a 4.4-fold increase of ROS generation in neurons, but only a 1.6-fold increase in astrocytes (Fig. 3b). The amount of ROS generated by the 10 µm NAC treatment was greater than that caused by 100 µm hydrogen peroxide in neurons (data not shown). Aβ1–40 fibrils were slightly more potent in ROS generation than NAC fibrils in both neurons and astrocytes.
ROS generation by NAC fibrils in mitochondria
To further specify the subcellular location of ROS generation, cells were stained with a mitochondria-specific probe, chloromethyl–tetramethyl rosamine. The confocal images showed that the exposure to 10 µm NAC fibrils or Aβ1–40 fibrils enhanced the generation of ROS in the mitochondria of both cortical neurons and astrocytes (Fig. 4). These findings are consistent with the results of ROS generation estimated with DCFH–DA probe (see above) and support the view that the mitochondria are the site of ROS generation.
Inhibition of NAC-induced apoptosis by Bcl-2
The protective effect of Bcl-2 against oxidative stress has been reported for neuronal cell death under various conditions, including exposure to Aβ proteins (Sailléet al. 1999). Therefore, the effect of overexpression of human Bcl-2 gene on NAC-induced apoptosis was investigated. Cortical neurons were infected with AxCALNL-hBcl-2 at a MOI of 10–100. Transduction efficiency was about 30% at a MOI of 100, as determined by immunostaining with anti-Bcl-2 antibody. After a 24-h incubation with 10 µm NAC fibrils, more than 90% of Bcl-2-positive cells remained intact, whereas about 40% of the Bcl-2-negative cells were apoptotic (Figs 5a and b), indicating a protective effect of Bcl-2 against the neurotoxic action of NAC fibrils.
Activation of NF-κB by NAC fibrils
Based on the finding that transcription factor NF-κB was activated by oxidative stress (Behl et al. 1994), the effect of NAC treatment on its activation was evaluated by EMSA. At least five shifted bands were detected (Fig. 6a, lane 1), and three of them were identified. Band II was supershifted by both antibodies against NF-κB subunits p50 and p65 (Rel A; Fig. 6a, lanes 2 and 3), suggestive of a heterodimer of p50 and p65. Band I was particularly dense in neurons (Fig. 6b), and identical with a ‘neuronal κB-binding factor’ that is distinct from NF-κB (Moerman et al. 1999). Band IV was supershifted by anti-Ku antibody (Fig. 6a, lane 4), supporting a known fact that Ku, a regulatory subunit of DNA-dependent protein kinase, binds to double-stranded DNA termini with high affinity and produces non-specific bands in the EMSA (Klug 1997). Of the two other dense bands, band III, but not band V, was abolished by the addition of an excess amount of cold non-specific probe to the assay (data not shown), suggesting that these bands represented unknown proteins that bind to the NF-κB consensus oligonucleotide probe in a sequence non-specific and specific manner, respectively.
We evaluated the DNA-binding activity of NF-κB by the density of band II. This band was not detectable in neurons and slightly visible in astrocytes under unstimulated conditions (Figs 6b and c). Exposure to NAC or Aβ fibrils made it clearly detectable in both cell types, suggesting activation of NF-κB by translocation into the nucleus and enhancement of its DNA-binding activity. The antioxidants, PG and BPN, suppressed NF-κB activation in neurons by 75–90% (Fig. 6b), and in astrocytes by 10–50% (Fig. 6c), thus further confirming the role of ROS in NF-κB activation.
Amyloidogenic peptides, such as Aβ protein and prion protein, which form a β-pleated sheet structure, are generally cytotoxic to neurons (Yankner et al. 1990; Forloni et al. 1993; Lorenzo and Yankner 1994). Similarly, NAC peptide prone to form a β-pleated sheet structure (Uéda et al. 1993) is cytotoxic. The neurotoxicity of NAC was first reported by Liu and Schubert (1998) using rat brain tumor B12 cells, and by El-Agnaf et al. (1998) using human neuroblastoma SH-SY5Y cells. Bodles et al. (2001) showed that amino acid residues 8–16 of NAC are crucial for the cytotoxic action. We previously reported that NAC fibrils induce cell death in differentiated PC12 cells partly by necrosis and partly by apoptosis (Tanaka et al. 2001). The present study clearly showed induction of apoptosis by NAC fibrils in cortical neurons of rat brain.
We then examined the effects of NAC fibrils on astrocytes. Glial cell cultures comprised type-1 and type-2 astrocytes that can be distinguished by morphology and phenotypic expression of A2B5 antigen (Raff 1989). Exposure to NAC fibrils induced extension of processes or activation of type-2 astrocytes, in sharp contrast to induction of apoptosis in neurons. The neuron-specific cytotoxicity was also demonstrated by an increased release of LDH from neurons, but not from astrocytes. It is of interest that this difference between neuronal and glial cells closely resembles the situation in the AD brain; neurons are degenerated, whereas astrocytes and microglias are markedly activated in and around senile plaques (Pike et al. 1995).
We also evaluated cytotoxicity of NAC fibrils by the MTT reduction assay. Two mechanisms have been suggested for suppression of MTT-reducing activity by Aβ protein. One is inhibition of mitochondrial succinate dehydrogenase (Kaneko et al. 1995) that is a component of complex II of the electron transport system located in the inner mitochondrial membrane. The suppression of MTT-reducing activity by NAC fibrils in cortical neurons might reflect their deleterious effects on mitochondria. The other mechanism is acceleration of MTT formazan exocytosis, as reported for rat brain tumor B12 cells (Liu and Schubert 1997) and astrocytes (Abe and Saito 1998; Kerokoski et al. 2001). In the present study, an accelerated formation of needle-like MTT formazan crystals by NAC fibrils was observed on the cell surface of both neurons and astrocytes (data not shown), suggesting that the second mechanism is, in part, responsible.
The neuroprotective effects of antioxidants suggest a role of oxidative stress in NAC-induced neurotoxicity, whereas their negative effects on the morphological changes of astrocytes preclude the role of ROS on NAC-induced astrocytic activation. However, intracellular oxidation of DCFH in astrocytes as well as neurons indicated intracellular generation of ROS in both cell types. Oxidative stress in the AD brain has been suggested by several lines of indirect evidence, such as increased protein oxidation (Smith et al. 1991), increased lipid peroxidation (Lovell et al. 1995), and the presence of advanced glycation end products in senile plaques (Vitek et al. 1994). While DCFH–DA monitors ROS generation in all cell compartments (Royall and Ischiropoulos 1993), rosamine monitors it more specifically in mitochondria. The increased staining of amyloid fibril-treated cells with rosamine suggests the generation of ROS to be mitochondrial in origin. These results are in agreement with a generally accepted view that mitochondria are the primary source ofROS (Loschen et al. 1974). In this context, Hsu et al. (2000) reported that α-synuclein overexpression led to morphological changes and dysfunction of mitochondria accompanied by an increased level of ROS.
Overexpression of Bcl-2 suppressed ROS generation as well as apoptosis in neuronal cell lines (Kane et al. 1993). The regulation of ROS generation proved not to be the principal mechanism of its anti-apoptotic function, but a secondary effect of preserving mitochondrial membrane functions (Jacobson and Raff 1995; Shimizu et al. 1995). Bcl-2 is localized in the outer membrane of mitochondria and blocks the opening of pores or channels through which apoptotic factors such as cytochrome c, apoptosis-inducing factor (AIF) and Smac/DIABLO are released (Hengartner 2000). Our results indicating the inhibition of NAC- or Aβ-induced apoptosis by Bcl-2 overexpression are expected to result from preservation of the mitochondrial membrane integrity and down-regulation of ROS.
The transcription factor NF-κB was activated in both cortical neurons and astrocytes exposed to NAC or Aβ fibrils. This finding may explain why NF-κB was activated in neurons and astrocytes in the close vicinity of diffuse and primitive plaques in the AD brain (Kaltschmidt et al. 1997). Furthermore, it is likely that activation of NF-κB, in astrocytes and microglias, augments neurodegenerative changes in the AD brain by up-regulating expression of proinflammatory or cytotoxic factors such as IL-1β, IL-6 (Bales et al. 1998) and NO synthase (Akama et al. 1998). On the other hand, NF-κB is known to induce expression of anti-apoptotic factors such as c-IAP1 and c-IAP2 in damaged cells (Wang et al. 1998), suggesting a possibility that activation of NF-κB in neurons may help protection against the cytotoxic action of amyloid. This idea is supported by the finding that pretreatment with IκB (NF-κB-binding inhibitor) antisense oligonucleotide promoted NF-κB activation and inhibited the neurotoxic action of Aβ protein (Barger et al. 1995).
The mechanism how NAC fibrils activate NF-κB has not been clarified. Our present study suggested that mitochondrial ROS is responsible for not only NAC cytotoxicity but also NAC-induced NF-κB activation. A role of mitochondrial ROS in NF-κB activation was suggested in TNF-induced gene transcription (Schulze-Osthoff et al. 1993). We observed a gel shift with the nuclear extract of NAC-treated neurons, implying that mitochondrial ROS is involved in the release of the NF-κB p50/p65 complex from IκB, its nuclear translocation, and the binding to target sequences of DNA.
Recently, Culvenor et al. (1999) reported no immunoreactivity of NAC in senile plaques and argued against the presence of NAC thereabout. Hashimoto et al. (2000), who originally reported the immunoreactivity of NAC in senile plaques (Uéda et al. 1993; Masliah et al. 1996), ascribed the discrepancy to a difference in antibodies and/or methods of tissue preparation. In support of this view,we have experienced a marked difference in the avidityamong antibodies against NAC.
Our preliminary experiment suggested that NAC accelerates Aβ amyloid formation invitro and enhances its neurotoxicity in cultures (unpublished data). These findings suggest a possibility that NAC interacts with Aβ protein to form composite amyloid fibrils and such fibrils may contribute, partly or principally, to neuronal cell injury in the AD brain. In order to assess this possibility, we need knowledge how α-synuclein is processed to NAC in the brain. In view of the presence of α-synuclein in extracellular spaces, e.g. the cerebrospinal fluid (Borghi et al. 2000) and extracellular Lewy bodies (Togo et al. 2001), it seems plausible that α-synuclein is released from damaged neurites and extracellularly proteolysed to NAC. Thus, the elucidation of α-synuclein processing will shed light on the molecular mechanisms of amyloid formation and neurodegeneration in the AD brain, thereby giving a clue to preventive and therapeutic interventions.
We thank Dr Hironobu Naiki and Dr Takeo Suga for their help in the electron microscopic work. We also thank Dr Todd Stedeford for useful discussion, and Barbara Baehr for preparing the manuscript. This study was performed through Special Coordination Funds for Promoting Science and Technology from the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government, and also supported by grants from Japan Foundation for Applied Enzymology and Yamanouchi Foundation for Research on Metabolic Disorders.