Sublethal Doses of β-Amyloid Peptide Abrogate DNA-dependent Protein Kinase Activity*

Background: Accumulation of DNA damage and deficiency in DNA repair may contribute to neuronal loss in Alzheimer disease. Results: Sublethal concentrations of aggregated β-amyloid peptides inhibit DNA-PK kinase activity in PC12 cells. Conclusion: DNA-PK inhibition may contribute to neurodegeneration by impairing DNA repair capability, inducing DNA damage accumulation. Significance: This represents a novel mechanism by which Aβ exerts its neurotoxic effects in Alzheimer disease. Accumulation of DNA damage and deficiency in DNA repair potentially contribute to the progressive neuronal loss in neurodegenerative disorders, including Alzheimer disease (AD). In multicellular eukaryotes, double strand breaks (DSBs), the most lethal form of DNA damage, are mainly repaired by the nonhomologous end joining pathway, which relies on DNA-PK complex activity. Both the presence of DSBs and a decreased end joining activity have been reported in AD brains, but the molecular player causing DNA repair dysfunction is still undetermined. β-Amyloid (Aβ), a potential proximate effector of neurotoxicity in AD, might exert cytotoxic effects by reactive oxygen species generation and oxidative stress induction, which may then cause DNA damage. Here, we show that in PC12 cells sublethal concentrations of aggregated Aβ(25–35) inhibit DNA-PK kinase activity, compromising DSB repair and sensitizing cells to nonlethal oxidative injury. The inhibition of DNA-PK activity is associated with down-regulation of the catalytic subunit DNA-PK (DNA-PKcs) protein levels, caused by oxidative stress and reversed by antioxidant treatment. Moreover, we show that sublethal doses of Aβ(1–42) oligomers enter the nucleus of PC12 cells, accumulate as insoluble oligomeric species, and reduce DNA-PK kinase activity, although in the absence of oxidative stress. Overall, these findings suggest that Aβ mediates inhibition of the DNA-PK-dependent nonhomologous end joining pathway contributing to the accumulation of DSBs that, if not efficiently repaired, may lead to the neuronal loss observed in AD.

including Alzheimer disease (AD) 2 (1)(2)(3)(4)(5), Huntington disease (6 -9), Parkinson disease (10 -13), and amyotrophic lateral sclerosis (14 -17). The extent of DNA damage determines whether apoptotic cascades initiate, causing progressive neurodegeneration, or the DNA repair mechanisms become active to reverse the damage (18). Accumulation of DNA damage is thought to be particularly deleterious in post-mitotic cells, which cannot be replaced through cell division (19). For this reason, neurons are particularly susceptible to the toxic effects of reactive oxygen species (ROS), the primary mediators of oxidative stress, which may cause damages at the main macromolecular systems, including nucleic acids (20).
Oxidative stress and neuronal DNA damage are common features of neurodegenerative diseases (21)(22)(23) associated with misfolded proteins that accumulate as intracellular and/or extracellular amyloid or amyloid-like deposits (24), such as AD. Neuropathologically, AD is characterized by extensive loss of synapses and neurons, accumulation of intracellular tangles, and extracellular/intracellular ␤-amyloid peptide (A␤) deposits (25). During the aggregation process, monomeric A␤ peptides form soluble oligomeric intermediates, which further associate in high molecular weight assemblies (protofibrils) and eventually generate amyloid fibrils and plaques. Soluble nonfibrillar forms of A␤ are proximate effectors of neurotoxicity and synaptotoxicity (26,27) in AD. Studies performed on mammalian cell lines, primary cortical neurons, and AD brains suggest that A␤ peptide exerts neurotoxic effects through ROS production (28 -30), generated during the early stages of protein aggregation when only protofibrils or soluble oligomers are present. DNA damage caused by ROS includes altered bases, abasic sites, and single and double strand breaks (DSBs). Interestingly, increased levels of DNA breaks and alkali-labile sites were detected in the cerebral cortex of AD patients (2), and in situ labeling methods showed the presence of single and double strand breaks in different AD brain regions (3,31,32). Moreover, a decreased capacity for DNA repair in fibroblasts or lymphocytes from patients with familial AD was also reported (33).
DSBs are considered the most lethal form of DNA damage that, if unrepaired, might cause cell death (34). In mammalian cells, DSB repair is achieved by two highly efficient mechanisms, homologous recombination (35,36) and nonhomologous end joining (NHEJ). Although there is increasing evidence that these pathways compete for DSB repair (37), NHEJ is considered the predominant pathway in higher eukaryotes (38). It requires the DNA-dependent protein kinase holoenzyme (DNA-PK), which is formed by a 470-kDa catalytic subunit, DNA-PKcs, and a heterodimer of 70-and 80-kDa polypeptides, known as Ku, which binds to DNA strand breaks, recruiting and activating the DNA-PKcs (39 -41). DNA-PKcs is a serine/threonine kinase belonging to the PI3K-like family of kinases, which includes ataxia telangiectasia-mutated kinase and Rad3-related kinase. End joining activity and protein levels of DNA-PKcs are significantly lower in AD brains compared with normal controls. The amount of end joining activity correlates with the expression of DNA-PKcs and is dependent on DNA-PK catalytic activity (4). In addition, immunohistochemical analysis of AD temporal cortex showed a decrease, although not significant, of DNA-PKcs expression both in neurons and astrocytes with increasing Braak stages (42). Overall, these findings suggest that repair of DNA DSBs may be deficient in AD, although the molecular candidate causing the NHEJ impairment has yet to be identified.
Because a low amount of A␤ is likely present in AD brain for extended periods prior to neuronal cell death, we investigated whether sublethal doses of A␤ might inhibit DNA-PK function, thus compromising the repair of DSBs. Addressing this issue, we found that sublethal concentrations of aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) inhibit DNA-PK kinase activity in PC12 cells. This inactivation was associated with down-regulation of the DNA-PKcs protein levels, caused by oxidative stress and resolved by the antioxidant NAC. We also found that exogenous A␤  oligomers enter the nucleus of PC12 cells, accumulate as insoluble oligomeric species, and inhibit DNA-PK kinase activity. Finally, we demonstrated that A␤-mediated DNA-PK kinase activity inhibition renders proliferating PC12 cells susceptible to nonlethal oxidative injury and attenuates DSB repair in NGF-differentiated PC12 cells.
Antioxidant NAC (Sigma) was used at final concentrations of 0.5 mM for 28 h (4-h pretreatment followed by a 24-h co-incubation with or without A␤). Hydrogen peroxide (Sigma) was added at final concentrations of 100 M for 30 min to induce sublethal oxidative stress. In both conditions, RPMI 1640 medium without serum was used as incubation buffer. DNA-PK inhibitor NU7026 was dissolved in DMSO at 7.1 mM and used at a final concentration of 10 M in complete medium for 24 h (44,45). DSB inducer doxorubicin (DOX) was used on proliferating and differentiated PC12 cells at 1 M for 8 h. Analysis of dose-response (0.01-10 M) and time-response (4 -24 h) curves revealed that this experimental condition had maximal DNA damage response in proliferating and NGF-differentiated PC12 cells, evaluated by histone H2AX phosphorylation (␥H2AX), without significant cell death.
The effects of A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) and A␤(1-42) on plasma membrane integrity were evaluated by a LDH assay (Promega). PC12 cells, cultured and treated in the same experimental conditions used for MTT assay, were centrifuged at 250 ϫ g for 10 min at room temperature (RT), and then 50 l of supernatant was added to an isovolume of Substrate Mix and incubated for 30 min at RT and light-covered. After adding the Stop Solution, absorbance readings were measured by a spectrophotometer microplate reader (wavelength, 490 nm, reference wavelength 630 nm). Results were expressed as % maximum LDH release, obtained by adding 0.9% Triton X-100 for 60 min, subtracted by values corresponding to base-line LDH release (i.e. untreated cells). Data are representative of 4 -6 independent experiments.
Nuclei were stained with 0.5 g/ml Hoechst 33342 for 4 min at RT in distilled water, and condensed and/or fragmented nuclei were counted as apoptotic nuclei.
At least 500 cells for each coverslip were counted in both assays, and data were representative of three independent experiments.
Subcellular Fractionation-The presence of oligomeric A␤  in the nuclear compartment of PC12 cells was verified by nuclear and cytosolic fractionation. Cells (ϳ17 ϫ 10 5 ), treated for 24 h with 50 M oligomeric A␤ , were washed twice with ice-cold PBS, harvested, and homogenized on ice by Dounce disruption in 10 mM Tris-HCl, pH 7.5, and 5 mM EDTA containing 1 mM PMSF and mixture protease inhibitors (Sigma). Cell integrity was microscopically checked using trypan blue (Sigma) staining. The solution was brought up to 320 mM sucrose and centrifuged four times at 1000 ϫ g for 10 min at 4°C to obtain the nuclear fraction. The supernatant was then centrifuged at 14,000 ϫ g at 4°C for 10 min to separate soluble (supernatant) and insoluble (pellet) cytosolic proteins. Isolated nuclei were lysed and centrifuged as described in "DNA-PK kinase activity assays" paragraph to separate soluble (supernatant) and insoluble (pellet) nuclear proteins. Insoluble proteins of nuclear and cytosolic compartments were extracted by incubation for 10 min at RT with Novex Tricine SDS sample buffer (Invitrogen), boiled for 10 min, and then spun 2 min at 14,000 ϫ g. Soluble and insoluble proteins were then separated by 10 -20% Novex Tricine gel (Invitrogen) and analyzed by Western blot for A␤  oligomers.
To assess the effect of DOX and A␤(25-35) on Ku70/Ku86 compartmentalization, treated proliferating PC12 cells (ϳ5 ϫ 10 6 ) were washed twice with ice-cold PBS, harvested, spun at 4°C for 5 min at 600 ϫ g, and lysed on ice with NE-PER nuclear and cytoplasmic extraction reagents (Thermo Scientific) following the manufacturer's instructions. For each fraction, 40 g of proteins were separated by SDS-PAGE and analyzed by Western blot.
Protein Carbonylation Analysis and ROS Detection-Oxy-Blot protein oxidation detection kit (Chemicon) was used to detect carbonyl groups introduced into proteins by oxidative reaction. Briefly, ϳ5 ϫ 10 6 cells were grown onto poly-L-lysinecoated 100-mm polystyrene dishes and treated with A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) or A␤(1-42) as indicated. Cells were washed twice with icecold PBS, harvested, centrifuged at 4°C for 5 min at 600 ϫ g, and lysed on ice with RIPA buffer containing protease and phosphatase inhibitors. Five g of each sample was denatured by adding an isovolume of 12% SDS (w/v) before the derivatization reaction, where carbonyl groups of the protein side chain were derivatized into 2,4-dinitrophenylhydrazone by reaction with 2,4-dinitrophenylhydrazine. Each sample was mixed with an isovolume of 2,4-dinitrophenylhydrazine and incubated at RT for 15 min. The reaction was blocked by adding the Neutralization Solution and mixing with 2-mercaptoethanol to achieve a final concentration of 0.74 M. Proteins were loaded onto 12% polyacrylamide gel, separated by SDS-PAGE, and transferred to a nitrocellulose membrane for 1 h at 100 V at 4°C. After blocking with 10% milk in PBST (PBS containing 0.05% Tween 20), blots were probed 1 h at RT with rabbit anti-dinitrophenyl antibody (1:150). Immunoreactive bands were detected with goat anti-rabbit IgG (HRP-conjugated) (1:300) and visualized by enhanced chemiluminescence detection system. The effective correct isoloading of samples was checked by staining acrylamide gel with Coomassie Brilliant Blue R-250 solution (Bio-Rad) for 1 h with gentle shaking at RT. The solution was drained off, and gel was washed at RT with destaining solution (40% (v/v) methanol and 10% (v/v) glacial acetic acid) until the protein bands were visible without background. At the end of the washings, gel was soaked in 2% (v/v) glycerol for 15 min and dried. Blots are representative of three independent experiments.
ROS intracellular accumulation was evaluated by the conversion of dihydroethidium (DHE) to ethidium. Briefly, 10 M DHE (Invitrogen) were added to 2 ϫ 10 5 PC12 cells grown onto poly-L-lysine-coated coverslips placed at the bottom of 35-mm dishes and incubated at 37°C for 30 min to allow for the conversion of the DHE. Cells were then washed twice with PBS, fixed with 4% paraformaldehyde, and counterstained for 4 min at RT with 0.5 g/ml Hoechst 33342.
At least 500 cells for each coverslip were counted and analyzed by fluorescence microscopy. Data are representative of 3 independent experiments.
Immunofluorescence Analysis-The apoptotic effects of DNA-PK kinase activity inhibition under oxidative stress conditions were assessed in proliferant PC12 cells. Cells (ϳ7 ϫ 10 5 ) were grown onto poly-L-lysine-coated coverslips placed at the bottom of 35-mm culture dishes for 24 h and treated as indicated. At the end of treatments, cells were rinsed twice with ice-cold PBS and incubated for 20 min at Ϫ20°C with ice-cold 70% acetone, 30% methanol solution. Cells were then re-hydrated for 10 min in PBS at RT and washed twice with PBS, and nuclei were stained with 0.5 g/ml Hoechst 33342 in distilled water for 4 min at RT. Condensed and/or fragmented nuclei were counted as apoptotic nuclei.
To analyze the effects of A␤(25-35) on DNA damage and DSB repair activity, proliferating and NGF-differentiated PC12 cells were analyzed for ␥H2AX foci formation and repair. Treated cells were rinsed and fixed as described above and incubated with 1 g/ml anti-phospho-H2AX (Ser-139) monoclonal antibody (Millipore) in 0.2 mg/ml BSA in PBS for 2 h at RT. After three washes in PBS, PBST (0.05% (v/v) Tween 20 in PBS), and PBS, Alexa Fluor 594 goat anti-mouse IgG antibody, 1:1000 (Invitrogen), was used as secondary antibody by incubation at RT for 25 min. Cells having more than 10 foci/nucleus were scored as positive. Nuclei counterstaining was performed as described above.
At least 500 cells for each coverslip were examined with Nikon Eclipse TE2000-U fluorescence microscope (ϫ60 immersion oil objective) equipped with a CCD camera. Data are representative of 3-4 independent experiments.
DNA-PK Kinase Activity Assays-PC12 cells (ϳ5 ϫ 10 6 ) were grown onto poly-L-lysine-coated 100-mm polystyrene dishes for 24 h and treated as indicated. Cells were then washed twice with ice-cold PBS, harvested, and centrifuged at 4°C for 5 min at 600 ϫ g. Resulting pellet was resuspended on ice with a high salt whole cell extract (WCE) buffer (20 mM Hepes, pH 7.6, 450 mM NaCl, 25% glycerol, 0.2 mM EDTA, 0.2 mM DTT) contain-ing 1 mM PMSF, mixture protease inhibitors, and 50 mM sodium fluoride and subjected to five freeze/thaw cycles (ethanol and dry ice, 30°C). After centrifugation at 15,000 ϫ g for 10 min at 4°C, protein concentration in the supernatants (WCE) was determined. Forty g of proteins were used for Western blot analysis of DNA-PK complex protein levels as described under "SDS-PAGE and Western Blot Analysis." DNA-PK from cellular lysates was isolated by two different protocols, dsDNAcellulose (Sigma) pulldown and immunoprecipitation with anti-DNA-PKcs Ab-4 mixture.
200 -400 g of WCE were incubated with 40 l of preswollen dsDNA-cellulose for 30 min at 4°C. After a centrifugation at 500 ϫ g for 1 min at 4°C, the supernatant was collected, and the dsDNA-cellulose was washed six times with ice-cold low salt buffer (20 mM Hepes, pH 7.6, 0.2 mM EDTA, 0.2 mM DTT) containing 0.5 mM PMSF, mixture protease inhibitors, and 50 mM sodium fluoride.
200 -400 g of WCE were precleared by incubation with 40 l of protein G bead slurry (50%) (Pierce) for 60 min at 4°C with end over end rotation. After centrifugation at 1000 ϫ g at 4°C for 1 min, the protein G was separated from the supernatant and discarded. Cleared lysate was incubated for 14 -16 h at 4°C in slow rotation with 2.5 g of DNA-PKcs Ab-4, and proteinantibody complex was precipitated by adding 40 l of protein G beads and incubation for 2-4 h at 4°C under slow rotary agitation. At the end of incubation time, the sample was centrifuged at 1000 ϫ g at 4°C for 1 min, and bound proteins were washed as described above.
Samples obtained from both protocols were then processed with SignaTECT TM DNA-dependent protein kinase assay system (Promega) following the manufacturer's instructions. Kinase reactions were conducted with 20-l aliquots of the resuspended DNA-PK-absorbed cellulose/protein G beads for 30 min at 30°C and were performed in both the presence and absence of a biotinylated DNA-PK p53-derived substrate. Both assays were performed in triplicate.
Co-immunoprecipitation Experiments-PC12 cells (ϳ5 ϫ 10 6 ) were grown on poly-L-lysine-coated 100-mm polystyrene dishes for 24 h and treated for 24 h with 50 M oligomeric A␤ . Cells were then washed twice with ice-cold PBS, harvested, and centrifuged at 4°C for 5 min at 600 ϫ g. Resulting pellet was homogenized on ice by Dounce disruption as described under "Subcellular Fractionation" to isolate soluble and insoluble proteins of nuclear and cytoplasmic compartments. Insoluble proteins were extracted by incubation for 10 min with 50 mM Tris, pH 8.0, 2% SDS, 10% glycerol, followed by sonication for 20 s. After dilution (1:20 v/v) with 50 mM Tris-Cl, pH 8.0, 150 mM NaCl containing mixture protease inhibitors, samples were precleared by incubation with 25 l of protein G beads for 30 min at 4°C with end over end rotation. After cen-trifugation at 1000 ϫ g at 4°C for 1 min, the protein G was separated from the supernatant and stored. Cleared lysate was incubated for 14 -16 h at 4°C in slow rotation with 2.5 g of DNA-PKcs Ab-4, and the protein-antibody complex was precipitated by adding 25 l of protein G beads and incubated for 4 h at 4°C under slow rotary agitation. At the end of incubation time, the sample was centrifuged at 1000 ϫ g at 4°C for 1 min, and bound proteins were washed five times with 50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% Nonidet P-40 and twice with 5 mM Tris, pH 8.0. Immunoprecipitates were then separated by 10 -20% Novex Tricine gel (Invitrogen) and analyzed with 4G8 antibody by Western blot for A␤(1-42) oligomer detection. The same protocol was applied for immunoprecipitation of A␤(1-42) oligomers by using 2 g of 4G8 antibody, and DNA-PKcs Ab-4 antibody was used to reveal DNA-PKcs protein.
Statistical Analysis-Data were expressed as means Ϯ S.E. Results were analyzed using one-way analysis of variance followed, where appropriate, by Tukey's post hoc test and were considered significant when p Ͻ 0.05.
Overall, these results show that A␤ peptides, up to 50 M and within 24 h, inhibit mitochondrial metabolism without disrupting membrane integrity, triggering apoptotic cell death or causing DSBs. Therefore, 1-50 M A␤(25-35) and A␤(1-42) concentrations were defined as sublethal and used in all subsequent experiments.
DNA-PK kinase activity is regulated by different mechanisms, including modifications in the catalytic subunit and/or the regulative subunits Ku70 and Ku86 protein levels (60,61). To evaluate whether the A␤-mediated decrease in DNA-PK kinase activity was dependent on protein level modifications, we performed Western blot analysis on protein extracts derived from A␤(25-35)and A␤(1-42)-treated PC12 cells. We found that DNA-PKcs protein levels were down-regulated in a concentration-dependent manner (Fig. 2C, left panel) with the greatest effect of 51.4 Ϯ 7.2% reduction after treatment for 24 h with 50 M A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35). Densitometric analysis of immunoreactive bands (Fig. 2, D and E) showed that the effect was detectable as early as 8 h after A␤ exposure (20 Ϯ 2.5% decrease). On the contrary, A␤(1-42) did not modify DNA-PKcs protein levels even at maximal dose and after 24 h treatment (Fig. 2C, right  panel). Ku70 and Ku86 protein levels also remained unchanged after both A␤ peptide treatments (Fig. 2C). We then analyzed whether treatment of PC12 cells with A␤(25-35) was able to mobilize Ku70 and Ku86 to the nucleus by Western blot analysis on nuclear and cytoplasmic PC12 cell extracts. We found that 50 M A␤(25-35) treatment induces a 95% increase of Ku86 in the nucleus as compared with untreated cells, whereas Ku70 compartmentalization remained unmodified (Fig. 2F). Treatment of PC12 cells with 1 and 10 M DOX, a well established DSB inducer (62), induced a similar nuclear translocation of Ku86 (108% with 1 M and 67% with 10 M of increase) and a nonstatistically significant accumulation of Ku70 with only 10 M DOX (Fig. 2G). These results indicate that sublethal doses of A␤(25-35) induce Ku86 nuclear translocation even in the absence of DSBs. It is worth noting that DNA-PK kinase activity, DNA-PK complex protein levels, and compartmentalization were not significantly affected by exposure to the reversed sequence peptides A␤  and A␤(42-1) (data not shown).

Aggregated A␤(25-35)-induced Oxidative Stress
Causes the Impairment of DNA-PK Kinase Activity-DNA-PK activity can be modulated by oxidative stress. Previous studies demonstrated that ROS levels are inversely correlated with DNA-PK kinase activity upon exposure to chemotherapeutic agents, and treatment with antioxidant reversed this inhibition (63,64). Thus, we asked whether the induction of A␤-mediated oxidative stress could induce the impairment of DNA-PK kinase activity. To this aim, we first verified whether aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) and oligomeric A␤(1-42) were able to induce oxidative stress in PC12 cells by evaluating protein carbonylation levels and ROS intracellular accumulation by the conversion of dihydroethidium (DHE) to ethidium. As shown in Fig. 3A   increase was evident as early as 8 h. On the contrary, oligomeric A␤(1-42) treatment did not increase protein carbonylation levels in the same experimental conditions (Fig. 3A, right  panel). Interestingly, ROS production was already observed 4 h after A␤(25-35) treatment, whereas it required 24 h of A␤(1-42) treatment (data not shown). Exposure to reversed sequences A␤(35-25) and A␤(42-1) did not induce oxidative stress (data not shown). To assess the contribution of oxidative stress on the alteration of DNA-PK kinase activity, we performed a DNA-PK kinase assay with protein extracts derived from PC12 cells treated with 50 M A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) or A␤(1-42) for 24 h, in the presence or not of 0.5 mM NAC, a glutathione precursor with established antioxidant properties (65). A preliminary dose-response experiment with different NAC concentrations established that this concentration did not interfere with basal DNA-PK kinase activity (data not shown).
To establish whether ROS were able to directly inhibit DNA-PK activity, we performed a DNA-PK kinase activity assay using commercially available purified human DNA-PK incubated with increasing concentrations of H 2 O 2 . DNA-PK was preincubated with H 2 O 2 for 10 min at 30°C before the assay of kinase activity. H 2 O 2 impaired DNA-PK kinase activity in a dose-dependent manner, reaching a maximal effect at 100 These results demonstrate that A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) can exert its inhibitory effect via the production of ROS, whereas the impairment of DNA-PK kinase activity by A␤(1-42) is not dependent on oxidative stress induction, at least at sublethal concentrations. Since 4 h of A␤(25-35) treatment on PC12 cells strongly impaired DNA-PK kinase activity while the correspondent protein levels were unmodified (compare Fig. 2, B and E), it is reasonable to conclude that the early inhibition of DNA-PK kinase activity by A␤(25-35) caused ROS production.
Exogenous A␤  Oligomers Enter the Nucleus of PC12 Cells-It has been previously demonstrated that under both normal and oxidative DNA damage conditions, ␤-amyloid peptides may localize in the nuclear compartment (66 -69). Importantly, A␤(1-42) immunoreactivity has been found in both the cytosol and nuclei of some degenerating neurons in transgenic mice and AD brains (68,69). Therefore, we asked whether in PC12 cells exogenous A␤(1-42) oligomers, the most neurotoxic A␤ species observed in AD, can enter the nucleus where DNA-PK exerts its DNA repair activity. Because it is well known that different kinds of A␤ species have distinct biological actions, we first analyzed by Western blot our A␤(1-42) preparation (43). As illustrated in Fig. 4B, A␤(1-42) preparation was enriched in large and low molecular weight oligomers. In particular, we found a larger amount of A␤(1-42) monomer and dimer as compared with trimer, tetramer, and larger oligomeric assemblies. Fig. 4B also shows that the aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) preparation was enriched in low molecular weight oligomers (tetramer and decamer), although a minor amount of high molecular weight species was also present.
To assess the presence of A␤  in the nucleus, we then carried out a biochemical fractionation on proliferating PC12 cells exposed to 50 M A␤(1-42) oligomers for 24 h. Soluble and insoluble nuclear and cytoplasmic proteins were extracted and analyzed for the presence of A␤(1-42) oligomers by Western blot. We found that A␤(1-42) oligomers (mainly monomer, dimer, and trimer) accumulate in the insoluble fractions of both nuclear and cytoplasmic compartments, although a slight amount of A␤(1-42) dimer was present in the cytoplasmic soluble pool (Fig. 4A). These results demonstrate that exogenous A␤(1-42) oligomers enter the nucleus of PC12 cells and accumulate as insoluble species.
We next assessed the direct interaction between DNA-PKcs and oligomeric A␤  in PC12 cells by co-immunoprecipitation experiments. To this aim, we immunoprecipitated DNA-PKcs from both soluble and insoluble fractions (nuclear and cytoplasmic compartments), followed by Western blot with anti-A␤ antibody. Surprisingly, we did not find the presence of A␤ oligomers bound to DNA-PKcs either in soluble or insoluble fractions. Immunoprecipitation of A␤ followed by Western blot analysis with anti-DNA-PKcs antibody confirmed that A␤ and DNA-PKcs do not co-immunoprecipitate in our experimental conditions (data not shown). The same results were obtained carrying out immunoprecipitation experiments with aggregated A␤ (data not shown) (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35).
DNA-PK Kinase Activity Inhibition Mediated by A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) Attenuates DNA Repair Activity in NGF-differentiated PC12 Cells-Post-mitotic cells adopt mainly NHEJ to repair DSBs (76,77). Hence, to evaluate the effect of A␤-mediated DNA-PK kinase activity inhibition on DSB repair, we used NGF-differentiated PC12 cells treated with 1 M DOX (62). The ability of cells to repair DSBs was assessed by counting ␥H2AX foci. Cells having more than 10 foci/nucleus were scored positive.  As shown in Fig. 6B, untreated control cells showed ϳ4% of nuclei positive to ␥H2AX as well as cells treated with aggregated 50 M A␤(25-35) for 24 h (data not shown). Eight hours of exposure to DOX induced foci accumulation in 38% of nuclei without any further effect after 24 h of treatment (data not shown). Following 24 h of recovery, only 3% of nuclei showed a positive immunostaining for ␥H2AX, indicating that DSB repair was completed. Preincubation of differentiated PC12 cells with aggregated 50 M A␤(25-35) for 24 h significantly decreased the ability to repair DNA damage such that, after 24 h of recovery, 17% of nuclei remained positive to ␥H2AX. Parallel analysis of apoptotic cells, assessed by Hoechst staining, revealed that 8 h of DOX application alone did not induce apoptosis (2.4% DOX-treated cells versus 1.4% control cells, evaluated after 24 h), whereas pretreatment with aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) induced 19% of apoptotic cells (Fig. 6B). It is worth mentioning that, in this condition, most of ␥H2AX-positive cells showed also apoptotic nuclear morphology supporting a functional association between not repaired DNA damage and apoptotic phenotype.

DISCUSSION
Unrepaired DNA lesions may trigger apoptosis, and the consequent accumulation of DNA damage potentially contributes to the progressive neuronal loss observed in neurodegenerative diseases, including AD (4,5,78).
Here, we found that aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) inhibits DNA-PK kinase activity, the key complex for DSB repair in mammalian cells in a dose-dependent manner with effects already evident after 4 h of treatment. This inactivation was associated with the generation of ROS and down-regulation of DNA-PKcs protein levels, and it is resolved by treatment with antioxidant NAC. Sublethal doses of oligomeric A␤  were also able to inhibit DNA-PK kinase activity, although to a minor extent and only after 24 h of treatment, without affecting DNA-PK complex protein levels. This down-regulation was not mediated by oxidative stress and could not be reversed by NAC.
The induction of oxidative stress mediated by aggregated A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) may cause the impairment of DNA-PK activity through two mechanisms. The first mechanism is that ROS production induces the degradation of DNA-PKcs protein and consequently down-regulates its kinase activity. Accordingly, we showed that down-regulation of DNA-PKcs protein is paralleled by A␤(25-35)-induced protein carbonylation, an irreversible and not repairable oxidative protein modification, which may render proteins more prone to proteolytic degradation by proteasomes (79,80). Although it is well known that A␤ oligomers and fibrils can inhibit the proteasome activity (81,82), it has been demonstrated that soluble A␤(1-40) is able to induce the degradation of post-synaptic density-95, specifically by proteasomes (83). In addition, naturally secreted A␤ oligomers cause the reduction of neuronal EphB2 protein, and co-incubation with lactacystin, a specific proteasome inhibitor, rescues this effect with a concomitant increase of ubiquitinated EphB2 (84). Hence, A␤ may have opposite effects on proteasome-mediated degradation, and also in the case of DNA-PKcs, it is plausible to hypothesize an involvement of proteasomes in the reduction of DNA-PKcs protein levels.
The second mechanism is that ROS directly inhibit DNA-PK kinase activity. Indeed, H 2 O 2 impairs DNA-PK kinase activity in a cell-free assay, and this may explain the early inhibition of kinase activity in PC12 cells (4 h after A␤(25-35) treatment) regardless of down-regulation of DNA-PKcs protein levels.
We also found that A␤(1-42) oligomers accumulate in the nucleus of PC12 cells as insoluble species. Because the nucleus is the subcellular compartment where DNA-PK exerts its DNA repair function, this finding raises the question whether A␤ may directly interact with DNA-PKcs in the nucleus. Although we observed that A␤ peptides down-regulate DNA-PK activity in a cell-free assay, co-immunoprecipitation experiments did not support a direct interaction between A␤ and DNA-PKcs in PC12 cells.
A possible explanation of DNA-PK kinase activity inhibition by oligomeric A␤  is that specific modulators of DNA-PK mediate the inhibition of its kinase activity. Indeed, it has been reported that A␤, in addition to specific kinases, can also modulate the function of protein phosphatases. For example, fulllength A␤(1-40) and A␤  peptides, as well as A␤ (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), specifically inhibit protein phosphatase 1 (PP1) activity both in a cell-free assay and in cells (87,88). PP1 is an abundant neuronal serine/threonine-specific phosphatase that positively modulates the kinase activity of DNA-PK (89,90). Hence, it is possible to speculate that in our experimental conditions, PP1 activity might be inhibited by A␤ , thus leading to the down-regulation of DNA-PK kinase function. Another modulator of DNA-PK, casein kinase II (CKII) (91), might be involved in the inhibition of DNA-PK kinase activity. Accordingly, it has been reported that in AD brains CKII activity and protein levels are altered and that A␤ affects CKII activity (92,93). Further experiments to identify the mechanism(s) by which A␤  impairs DNA-PK kinase activity would be required.
We showed that A␤-mediated DNA-PK kinase activity inhibition renders PC12 cells susceptible to nonlethal oxidative injury leading to cell death. A similar result was obtained following exposure to the specific DNA-PK inhibitor NU7026, indicating that this effect was specifically mediated by the abrogation of DNA-PK activity. Because of the high rate of oxidative metabolism in the brain, neurons are continuously exposed to oxidative stimuli. Moreover, neurons have low levels of antioxidant enzymes, thus particularly susceptible to the damaging and highly toxic effects of ROS (94). DNA-PK exerts a protec-tive function under different cell death conditions, including oxidative stress and excitotoxicity (71,73). Hence, it is possible that exposure to nonlethal oxidative injuries in the presence of A␤, a hallmark of AD pathology, elicits neuronal cell death by suppression of DNA-PK anti-apoptotic function, contributing to the neurodegenerative process.
Finally, we showed that aggregated A␤ attenuates DNA repair activity in NGF-differentiated PC12 cells exposed to the DSB inducer DOX, suggesting an impairment of DNA-PK-mediated NHEJ. End joining activity and protein levels of DNA-PKcs are reduced in the midfrontal cortex of patients with AD (4). Similarly, a decrease of DNA-PKcs expression, although not significant, was observed in neurons and astrocytes of temporal cortices of AD cases with increasing Braak stages (42). Taken together, these results enable us to hypothesize that during AD pathology continuous accumulation of A␤ and ROS production would impair DNA-PK activity contributing to neurodegeneration through the inhibition of DNA-PK-mediated DSB repair pathway. Inhibition of NHEJ may indeed increase DSBs, potentially lethal lesions if unrepaired or not correctly restored.
Several human diseases, including xeroderma pigmentosum, Cockayne syndrome, Nijmegen breakage syndrome, ataxia telangiectasia, ataxia with occulomotor apraxia 1 (AOA1), and spinocerebellar ataxia with axonal neuropathy 1 (SCAN1) are caused by mutations in DNA repair genes and present important neurological implications (77). Moreover, it has been recently shown that also in Huntington disease (HD) DNA repair dysfunction is a critical factor of the pathological process (9). DNA repair dysfunction in HD is not genetically determined but is caused by the accumulation of neurotoxic species of mutant huntingtin protein (Htt). Indeed, mutant Htt interacts with Ku70 and diminishes DNA-PK kinase activity both in vitro and in vivo. The impairment of DNA damage repair contributes to the increase of DSBs in HD pathology.
In conclusion, based on these data, it is tempting to speculate that protein misfolding and aggregation of amyloidogenic proteins, such as A␤ in AD and Htt in HD, contribute to the DSB accumulation observed in neurodegenerative diseases via inhibition of NHEJ and eventually to neurodegeneration.