Extracellular plaques containing aggregated amyloid-β (Aβ) peptides and intracellular neurofibrillary tangles of hyperphosphorylated tau protein are the two salient pathological hallmarks of the Alzheimer's disease (AD) brain. Aβ is generated by consecutive cleavages of amyloid precursor protein (APP) by β-secretase (β-site APP cleaving enzyme 1, BACE 1) and the γ-secretase protein complex of which presenilin is the catalytic subunit (Hardy, 2006). Familial mutations in APP or presenilin lead to increased generation and/or aggregation of Aβ peptides and cause early-onset AD (Hardy, 2006). Aggregated Aβ interferes with synaptic plasticity and causes neuronal cell death (Lambert et al., 1998; Townsend et al., 2006; Kayed et al., 2004; Wogulis et al., 2005). This and other evidence supports the amyloid hypothesis which postulates that toxicity exerted by aggregated Aβ initiates AD, with synaptic dysfunction, neurofibrillary pathology and oxidative injury constituting downstream events (Hardy, 2006).
A number of studies, however, suggest that oxidative damage occurs in mild cognitive impairment (MCI) and early stages of sporadic AD, before widespread plaque and tangle development (Sayre et al., 1997; Butterfield et al., 2006; Nunomura et al., 2006; Williams et al., 2006; Lovell & Markesbery, 2008). Stress-activated BACE 1 and γ-secretase have been proposed to contribute to the deposition of Aβ peptides in sporadic AD (Tamagno et al., 2008). Supporting this idea, expression and activity of BACE 1 is elevated in the brains of sporadic AD patients (Fukumoto et al., 2002; Holsinger et al., 2002). In addition, BACE 1 is a demonstrated stress-induced protease that is up-regulated in response to oxidative stress (Tamagno et al., 2005; Tong et al., 2005), energy depletion (Velliquette et al., 2005), traumatic brain injury (Blasko et al., 2004) or cerebral ischemia (Wen et al., 2004; Tesco et al., 2007), all of which are events associated with increased sporadic AD risk. These findings suggest that oxidative stress is upstream of Aβ in AD and that Aβ might be generated as a compensatory response in neurons attempting to attenuate oxidative stress (Smith et al., 2002; Lee et al., 2006). In addition, neurofibrillary degeneration may exacerbate the oxidative damage and elevation of Aβ (Yan et al., 1995). In this context, mutations in APP that cause familial AD or in tau that cause frontotemporal dementias could contribute to neurodegeneration in part by increasing neuronal vulnerability to oxidative stress (Marques et al., 2003; Dias-Santagata et al., 2007).
Intracellular Aβ has been linked to secretory pathway dysfunction and apoptosis (LeBlanc, 1995; Gasparini et al., 1997; Zhang et al., 1997; Ohyagi et al., 1999; Gouras et al., 2005) and its elevation following oxidative stress has been demonstrated in neuroblastoma cells (Misonou et al., 2000; Paola et al., 2000; Tamagno et al., 2005, 2008; Zheng et al., 2006), rat cortical neurons (Hasegawa et al., 2005), astrocytes (Busciglio et al., 2002), mixed fetal guinea pig brain cells (Ohyagi et al., 1999) and mammalian lenses (Frederikse et al., 1996). Since Aβ peptides largely accumulate in the extracellular spaces of the AD brain, we were interested in determining whether oxidative stress by exposure to H2O2 influences secretion of Aβ peptides from neurons derived from embryonic chick brain. The human and chicken Aβ peptide sequences are identical and chick neurons ubiquitously secrete Aβ, making this a relevant and readily accessible model for investigating the modulation of endogenous APP processing in neurons (Esselmann et al., 2004; Carrodeguas et al., 2005).
Cell cultures were prepared from 7-days-in-ovo embryonic chick tecta and were characterized as ~97% neuronal after 8 days in vitro, estimated by the proportion of cells positively labeled by the neuron-specific markers beta tubulin-III and tau protein (Fig. 1A–C) (detailed methods provided in Supporting information). At least 92 ± 13% of cells were viable following treatment with up to 50 µm H2O2 for 20 h (mean ± standard deviation, n ≥ 150 cells) (Fig. 2A,B). A sharp decline in viability was observed at H2O2 concentrations > 50 µm (63 ± 13% and 10 ± 13% at 125 µm and 250 µm H2O2, respectively; mean ± standard deviation, n ≥ 150 cells) (Fig. 2A). Aβ peptides were immunoprecipitated from culture medium after 20 h exposure to sublethal concentrations of 10–20 µm H2O2 (Fig. 2C,D). H2O2 caused a 2.4 ± 0.6-fold increase in secreted Aβ (mean fold increase ± standard error of mean, n = 12 H2O2 treated and n = 10 control, Student's t-test: p < 0.05; Fig. 2E). No changes were observed in total endogenous APP (Fig. 2F) or tau protein levels in the corresponding cell lysates (Fig. 2D), but heat-shock protein 90 (Hsp90) levels were elevated 2.5 ± 0.3-fold (mean fold increase ± standard error of mean, n = 9 H2O2 treated and n = 9 control, Student's t-test: p < 0.01; Fig. 2D,G). Notably, heat-shock proteins are up-regulated in AD brain (Hamos et al., 1991; Dickey et al., 2007).
In summary, we demonstrate that exposure to nonlethal doses of peroxidative stress for 20 h significantly increases the accumulation of extracellular Aβ in primary chick neuronal cultures. This is accompanied by increased cellular Hsp90. These results suggest that Aβ, in addition to causing oxidative stress through generation of reactive oxygen species (Behl et al., 1994), may be generated by neurons in response to increased oxidative stress during aging (Lee et al., 2006; Nunomura et al., 2006). Chick brain neurons secrete a highly conserved pattern of Aβ peptides analogous to that normally found in human cerebrospinal fluid (Aβ1-37/38/39/40/42) (Esselmann et al., 2004). Here, we observe an overall increase in total Aβ levels induced by H2O2. Further studies are needed to determine whether any Aβ isoforms in particular are specifically increased by oxidative insults, what mechanism this might involve and whether oxidative stress also influences the phosphorylation and/or distribution of tau protein in these neurons. The conserved properties of APP trafficking and processing (Goldsbury et al., 2006), as well as the analogous sequence and phosphorylation patterns of tau protein in chick neurons (Yoshida & Goedert, 2002), make this primary culture system a promising model for delineating the cell biology and pathological modulation of AD-related proteins.