To determine the impact of 17β-estradiol (E₂) on Aβ neurotoxicity, 10 day old rat primary hippocampal neurons with elaborate neurite networks were treated with E₂ (10 ng/mL) for 48 hr prior to exposure to fibrillar Aβ_1–42 (1.5 μM) for 72 hr in the continued presence of E₂. The concentration of 10 ng/ml of E₂ was used based on multiple prior analyses indicating that this concentration induces maximal neuroprotection and the least degree of variability between experiments 20262728. CalceinAM and ethidium homodimer staining, which reflect neuronal survival and death, respectively, indicated neuronal death following 72 hr exposure to Aβ_1–42 (1.5 μM) compared to vehicle control-treated cultures (Fig. 1). Aβ_1–42 exposure significantly decreased neuronal survival (Fig. 1B) as indicated by a decrease in the calcein signal (Fig. 1B; p < 0.01 as compared to control; n = 4) and an increase in the amount of ethidium homodimer signal (Fig. 1C; p < 0.01 as compared to control; n = 4). E₂ (10 ng/mL) had no effect on neuronal survival in the absence of Aβ_1–42 (Figure 1A,B &1C). Pretreatment of hippocampal neurons with E₂ (10 ng/mL) significantly reduced the amount of Aβ_1–42-induced neurotoxicity (Fig. 1A), as indicated by the significant attenuation of the Aβ_1–42-induced neuronal death as indicated by increased Calcein fluorescence (Fig. 1B; p < 0.01 as compared to Aβ_1–42 alone; n = 4) and decreased ethidium homodimer fluorescence (Fig. 1C; p < 0.01 as compared to Aβ_1–42 alone; n = 4). Following 72 hr of Aβ exposure, neurites were thickened and beaded and there was an increase in the number of phase bright cells, indicative of neuronal death compared to vehicle control and E₂-treated neurons (data not shown).

To determine if E₂-induced alterations in neuronal survival were mediated via regulation of apoptosis, TUNEL staining was performed to identify apoptotic nuclei in response to Aβ_1–42 in the presence and absence of E₂ (Fig. 2A). Primary hippocampal neurons were pretreated with E₂ (10 ng/mL) or vehicle control for 48 hr prior to Aβ_1–42 (1.5 μM) exposure for 3 days in the continued presence of E₂ or vehicle control. Aβ_1–42 exposure resulted in a significant increase in the number of TUNEL-positive neurons (Fig. 2B; p < 0.05 as compared to control; n = 4). E₂ treatment prior to Aβ_1–42 exposure resulted in a significantly lower percentage of TUNEL-positive neurons than observed in response to Aβ_1–42 exposure in vehicle control treated neurons, indicating that E₂-treatment significantly reduced the amount of Aβ_1–42-induced apoptosis (Fig. 2A &2B; p < 0.05 as compared to Aβ_1–42 alone; n = 4). E₂ alone had no significant effect on the percentage TUNEL-positive neurons (Fig. 2).

One potential mechanism of Aβ_1–42-induced neurotoxicity is dysregulation of calcium homeostasis, which leads to an increased cytosolic calcium load and can result in neuronal dysfunction and death 4927. We have previously shown that estrogen-induced neuroprotection against glutamate excitotoxicity is mediated via regulation of neuronal calcium homeostasis 22 and that 17β-estradiol can both promote and restore calcium homeostasis in neurons derived from middle aged and aged rat hippocampi 27. To determine if E₂-induced neuroprotection against Aβ_1–42 is likewise due to regulation of calcium homeostasis, we conducted Fura2 calcium imaging to assess the effect of Aβ_1–42 on basal neuronal calcium loads in the presence and absence of E₂ (Fig. 3). Hippocampal neurons were pretreated with E₂ (10 ng/mL) or vehicle control for 48 hr prior to Aβ_1–42 exposure for 24 hr. Aβ_1–42 exposure resulted in a significant increase in the resting cytosolic calcium concentration (Fig. 3; p < 0.01 as compared to control; n = 4). E₂ pretreatment significantly attenuated the Aβ_1–42-induced rise in resting calcium concentration Fig. 3; p < 0.01 as compared to Aβ_1–42 alone; n = 4). E₂ by itself had no significant effect on resting calcium concentration (Fig. 3).

Bcl-2 is well established as an anti-apoptotic protein in neurons that can avert neuronal death and protect mitochondria against toxin-induced dysfunction 2930. It has been shown that Aβ down regulates Bcl-2 in human primary neurons 31. Further, Bcl-2 is an estrogen-responsive protein 72232333435. It has also been suggested that Bcl-2 is involved in maintenance of cellular calcium homeostasis 2930. To determine if mitochondrial expression of this key regulator of apoptosis was involved in E₂-induced neuroprotection, we determined the effect of E₂ and Aβ_1–42 on Bcl-2 expression in the mitochondrial fraction of rat primary hippocampal neurons. Neurons were pretreated with E₂ (10 ng/mL) or vehicle control for 48 hr prior to exposure to Aβ_1–42 (1.5 μM) for 24 hr. Crude mitochondrial fractions were assessed for Bcl-2 expression by Western blot analysis. Exposure to Aβ_1–42 resulted in a significant decrease (~60%) in Bcl-2 expression in the mitochondrial fraction (Fig. 4; p < 0.01 as compared to control; n = 4). As expected, E₂ by itself significantly increased (~60%) mitochondrial Bcl-2 expression (Fig. 4A; p < 0.05 as compared to control). Pretreatment of neurons with E₂ completely prevented Aβ_1–42-induced decrease in mitochondrial Bcl-2 (Fig. 4; p < 0.01 as compared to Aβ_1–42 alone; p < 0.05 as compared to control; n = 4).

Neurotoxicity resulting from apoptosis often results from the induction of Bax translocation from the cytosol to the mitochondria where it can mediate the release of apoptotic factors such as cytochrome c 36373839. To determine if the neuroprotective effect of E₂ is associated with regulation of Bax translocation or cytochrome c release, primary hippocampal neurons were pretreated with E₂ (10 ng/mL) or vehicle control for 48 hr prior to Aβ_1–42 (1.5 μM) exposure for 24 hr and assessed for Bax localization by Western blot and immunocytochemical analyses.

Aβ_1–42 exposure resulted in a significant increase in Bax immunoreactivity in the mitochondrial fraction (Fig. 5; p < 0.05 as compared to control; n = 4). E₂ pretreatment significantly attenuated the Aβ_1–42-induced increase in Bax immunoreactivity in the mitochondrial fraction (Fig. 5A; p < 0.05 as compared to Aβ_1–42 alone; n = 4). In contrast to the mitochondrial fraction there was no change in Bax immunoreactivity in the cytosolic fraction (Fig. 5A; n = 4). To confirm the shift in subcellular localization of Bax, we performed immunoflourescent staining for Bax. In control neurons, the Bax immunofluorescence signal was diffuse throughout the cell body and neurites (Fig. 5B, left panel). In Aβ_1–42-treated cells the Bax immunofluorescence signal was punctate, consistent with a shift to mitochondrial localization (Fig. 5B middle panel). Pretreatment of hippocampal neurons with E₂ prevented Aβ_1–42-induced translocation of Bax, resulting in a diffuse cytosolic staining pattern with slight areas of clustering consistent with partial mitochondrial translocation (Fig. 5B, right panel).

To determine the consequences of regulation of Bcl-2 and translocation of Bax, we determined the impact of E₂ pretreatment and Aβ_1–42 exposure on cytochrome C release from mitochondria. Aβ_1–42 exposure resulted in a significant increase in cytochrome c immunoreactivity in the cytosolic fraction (Fig. 6A; p < 0.05 as compared to control; n = 4). E₂ pretreatment significantly attenuated the Aβ_1–42-induced increase in cytochrome c immunoreactivity in the cytosolic fraction while having no effect alone (Fig. 6A; p < 0.05 as compared to Aβ_1–42 alone; n = 4). There was a corresponding significant decrease in cytochrome c immunoreactivity in the mitochondrial fraction in response to Aβ_1–42 (Fig. 6A; p < 0.05 as compared to control; n = 4) that was prevented by E₂ pretreatment (Fig. 6A; p < 0.05 as compared to Aβ_1–42 alone; n = 4). To confirm the subcellular localization of cytochrome c in our conditions, neurons were double stained for cytochrome c and mitochondria (Mitotracker Red CMXRos). In control neurons cytochrome c immunofluorescence (green) exhibited a punctate staining pattern that overlapped (yellow) with the mitochondrial localization signal (red) (Fig. 6B). Aβ_1–42 exposure resulted in a diminution of the punctuate labeling of cytochrome c immunofluorescence (green) particularly in the soma while the sparse neuritic labeling for cytochrome C labeling remained. However, the neuritic labeling for cytochrome C was not co-localized with the mitochondrial localization signal (red) indicating release of cytochrome c from the mitochondria (Fig. 6B). The cytochrome c immunofluorescence (green) in the E₂ pretreated and Aβ_1–42 exposed neurons exhibited a punctate staining pattern that overlapped (yellow) with the mitochondrial localization signal (red), as in the control cultures (Fig. 6B).

To determine the relationship between cytochrome c release and Bax translocation, hippocampal neurons were assessed for Bax and cytochrome c expression by double immunofluorescence (Fig. 7). In control neurons the Bax immunofluorescence signal (red) was diffuse and, as above, the cytochrome c immunofluorescence signal (green) was punctate (Fig. 7), consistent with cytosolic Bax and mitochondrial cytochrome c. In Aβ_1–42-treated cells the Bax immunofluorescence signal (red) was punctate, consistent with a shift to mitochondrial localization, and, as above, the cytochrome c immunofluorescence signal (green) was diminished with a diffuse localization (Fig. 7). Pretreatment of hippocampal neurons with E₂ prevented Aβ_1–42-induced translocation of Bax and release of cytochrome C from mitochondria (Fig. 7).

We sought to determine if the E₂ treatment that prevented the Aβ_1–42-induced alterations in mitochondrial-associated proteins of primary hippocampal neurons affected brain mitochondria in vivo. An increased mitochondrial calcium capacity should coincide with an enhanced ability to withstand calcium load and in parallel a sustained mitochondrial respiratory function. Whole brain mitochondria were isolated and assessed for respiratory function in the presence of glutamate and malate as respiratory substrates following a 2 min challenge with 100 μM Ca²⁺. Oxygen consumption during State 4 (resting) respiration was not affected by the presence or absence of calcium (100 μM) in control mitochondria (Fig. 8A). Although mitochondrial calcium overload can result in an uncoupling effect that would be expected to increase State 4 respiration, this was relatively unchanged in these experiments, consistent with previously published results 30. Oxygen consumption during State 3 (ADP stimulated) respiration was significantly decreased following Ca²⁺ challenge in control brain mitochondria relative to Ca²⁺ naive mitochondria (Fig. 8A; p < 0.05; n = 7). There was a likewise decrease (~45%) in the respiratory control ratio (RCR) following Ca²⁺ challenge of control mitochondria (Fig. 8C; p < 0.05; n = 7). RCR values determined using atractyloside to induce State 4 respiration following ADP depletion were indistinguishable from those calculated using basal State 4 respiration readings (data not shown).

To determine if E₂-treatment serves a protective function against increased calcium loads, mitochondrial respiratory function was assessed following a 2 min calcium exposure in the presence of glutamate and malate. Ovariectomized adult rats were administered E₂ (30 μg/kg in sesame oil) or vehicle control (sesame oil) 24 hr prior to mitochondrial isolation from whole brain tissue. Brain mitochondria from E₂-treated rats displayed significantly greater oxygen consumption following Ca²⁺ challenge in State 3 respiration and a significantly higher RCR than mitochondria from control rats (Fig. 8C; p < 0.05; n = 7). Further there was a smaller decrease in the RCR values following Ca²⁺ challenge in the brain mitochondria from E₂-treated rats than control (Fig. 8C; p < 0.05; n = 7).

Culture materials were from Gibco BRL (Rockville, MD). Chemicals were from MP Biomed (Irvine, CA) unless otherwise noted. Steroids were dissolved in ethanol and diluted in culture medium with final ethanol concentration <0.001%. Fura2-AM, Calcein AM and ethidium homodimer-1 were from Molecular Probes (Eugene, OR). Amyloid beta_1–42 was from American Peptide Company (Sunnyvale, CA).

Hippocampal neurons from embryonic day 18 (E18) rat fetuses were cultured as previously described and generated cultures 98% neuronal in phenotype 22. Briefly, embryonic rat hippocampi were dissociated by passage through fire-polish constricted Pasteur pipettes. Neurons plated on poly-d-lysine coated coverslips (22 mm round), chamberslides (Falcon; 4 well) or polyethylenimine coated 6-well plates were grown in Neurobasal medium supplemented with 5 U/ml penicillin, 5 mg/ml streptomycin, and B27 supplement at 37°C in humidified 5% CO₂ atmosphere for 10–12 days prior to experimentation.

Neurons were pretreated with 17β-estradiol (10 ng/mL) or vehicle control for 48 hr prior to exposure to fibrillar amyloid beta (1.5 μM) for the indicated times. Fibrillar Aβ_1–42 was prepared by solubilizing Aβ_1–42 in HCL (10 mM) at a concentration of 1 mM. Aβ_1–42 was diluted in PBS to 100 μM and incubated for 72 hr at 22°C 40. At the beginning of each experiment Aβ_1–42 was further diluted to 1.5 μM in complete Neurobasal medium.

Cell viability was measured by calcein/ethidium homodimer staining. Neurons were incubated in phosphate buffered saline (PBS) containing calcein AM (1 μM) and ethidium homodimer-1 (2 μM) for 30 min at room temperature. Fluorescent intensity was measured using a dual-wavelength fluorescent plate reader (GeniosPro; Molecular Devices) at 485/530 nm Ex/Em and 530/645 nm Ex/Em for calcein and ethidium, respectively. Data represents percent live or dead cells normalized to control fluorescent values. Data is presented as mean ± S.E.M. from 4 independent experiments with 8 wells per condition per experiment. Images were captured using the Marianas imaging system with Slidebook software (Intelligent Imaging Innovations, Inc., Santa Monica, CA) based on a Zeiss 200 M inverted microscope.

Neurons grown on chamber slides were treated as for Neuronal Survival above and fixed in 4% paraformaldehyde for 15 at room temperature and processed for TUNEL staining by the In Situ Cell Death Detection Kit, Fluorescein kit (Roche Applied Science; Indianapolis, IN) according to the manufacturer's instructions. Neurons were counterstained with DAPI to label nuclei. Three random fields per well were captured using the Marianas imaging system with Slidebook software (Intelligent Imaging Innovations, Inc., Santa Monica, CA) based on a Zeiss 200 M inverted microscope. Images were captured from 3 wells per condition per experiment (total of 9 fields per condition). Number of TUNEL positive neurons and number of nuclei (DAPI) were determined using Slidebook software and percent TUNEL positive neurons was calculated from the ratio of number of TUNEL positive neurons to number of total nuclei. Data represents the mean ± SEM from 4 independent experiments.

Hippocampal neurons grown on glass coverslips were pretreated with E₂ (10 ng/mL) or vehicle control for 48 hr prior Aβ_1–42 exposure for 24 hr. Neurons were loaded in the dark with Fura2-AM (2 μM) in Hank's Buffered Saline (HBS) (45 min.; 37°C). Cytosolic calcium concentrations were determined by comparing the 340/380 ratio to a standard curve as previously described 22. At least 10 neurons per coverslip were assessed with at least 3 coverslips per condition per experiment. Data represents the mean ± SEM from 4 independent experiments. Equal dye loading was determined as previously described 22.

Cytosolic and mitochondrial fractions were obtained using the Cytosol/Mitochondrial Fractionation Kit (Calbiochem; Sand Diego, CA) according to the manufacturer's instructions. Protein concentration was determined by the BCA assay. 25 μg of total protein were loaded per well and electorphoresed in a 2% SDS-PAGE gel. Protein was electrotransfered to PVDF membranes and probed with primary antibodies to Bcl-2 (1:250; BD Transduction Laboratories (610539); San Jose, CA), Bax (1:1000; Cell Signaling Technology (2772); Beverly, MA) and cytochrome c (1:1000; BD Transduction Laboratories (5564333); San Jose, CA) and respective horseradish peroxidase (HRP)- conjugated secondary antibodies (1:10,000; 1 hr room temperature). Porin (Mitosciences, Eugen, OR) and Actin (Santa Cruz, Santa Cruz CA) were used as loading controls for mitochondrial and cytosolic fraction, respectively (Data not shown). Bands were visualized with TMB peroxidase kit (Vector Laboratories) and quantitated by scanning and densitometry with Un-Scan-It software (Silk Scientific; Orem, UT). Data are presented as means ± S.E.M. from 4 independent experiments.

Treated neurons grown on chamberslides were fixed in 4% paaraformaldehyde for 15 at room temperature 24 hr following Aβ exposure. For mitochondrial labeling, neurons were incubated in Mitotracker Red CMXRos for 10 min at 37°C for 10 min prior to fixation. Neurons were washed in PBS and permeabilized in PBS + 0.01% triton X-100 for 5 min prior to incubation in primary antibody (Bax: 1:250; Cell Signaling Technology (2772); Beverly, MA; cytochrome c: 1:250; Pharmingen (556432); San Jose, CA) for 2 hr at room temperature. Antibody-antigen complexes were visualized by incubating in FITC (1:250; Vector Laboratories; Burlingame, CA)- or CY3 (1:1000; Amersham; Piscataway, NJ)-conjugated secondary antibodies for 45 min at room temperature, coverslipping with DAPI containing mounting medium (Vector Laboratories; Burlingame, CA) and capturing images using the Marianas imaging system with Slidebook software (Intelligent Imaging Innovations, Inc., Santa Monica, CA).

Adult (4–6 months old) female ovariectomized Sprague-Dawley rats were injected subcutaneously with 17b-estradiol (30 μg/kg) in sesame oil 2 weeks following surgery. 24 hr later, rats were sacrificed and whole brain tissue was homogenized in mitochondrial isolation buffer (0.32 M sucrose, 1 mM EDTA and 10 mM Tris; pH 7.4). Homogenates were centrifuged at 1,330 × g for 5 min at 4°C. Pellets were re-homogenized and centrifuged. The two postnuclear supernantants were combined and centrifuged at 21,200 × g for 10 min at 4°C. The resulting crude mitochondrial pellets were resuspended in 15% Percoll and layered over a 23%/40% discontinuous Percoll gradient and centrifuged at 27, 000 × g for 10 min at 4°C. The fraction accumulating at the 23%/40% interface was collected and diluted 1:4 with isolation buffer and centrifuged at 16,000 × g for 10 min at 4°C.

The pellet was transferred to a 1.5 ml tube centrifuged at 7,300 × g for 10 min at 4°C. The pellet will be resuspended in 40 μL isolation buffer. Protein concentration will be determined by the BCA assay. Mitochondrial integrity was assessed assessing cytochrome c oxidase activity (Cytochrome c oxidase activity kit, Sigma) in intact and lysed mitochondrial samples. Cytochrome c oxidase in inaccessible in intact mitochondria and high activity in these samples relative to total activity (lysed samples) indicates poor mitochondrial integrity.

Aliquots of mitochondria (1 mg/mL) were used in measurements of respiratory activity using a Clark-type oxygen electrode (Hansatech Oxygraph) as previously described. Oxygen electrode buffer (130 mM KCl, 2 mM KH₂PO₄, 3 mM HEPES, 2 mM MgCl₂, 1 mM EGTA) was incubated for 1 min in a magnetically stirred chamber at 28°C. The respiratory substrates, glutamate (5 mM) and malate (2.5 mM), were added followed by the isolated mitochondria (100 μg). Basal respiration was first measured in the absence of ADP for 1 min, followed by Ca²⁺ (100 μM) or equivalent volume of buffer for an additional 2 min. Subsequently, State 3 respiration was measured in the presence of ADP, by addition of 20 mL ADP (0.027 M in 0.067 M NaPO₄) to determine the maximal rate of coupled ATP synthesis. Then State 4 respiration was induced by addition of the adenine nucleotide translocator inhibitor atractyloside (50 μM). The respiratory control ratio was calculated using the ratio of State 3 to State 4 respiratory rates.

Statistically significant differences were determined by one-way ANOVA followed by Student-Neuman Keuls post hoc analysis.