Estrogen is neuroprotective against a variety of insults, including β-amyloid peptide (Aβ); however, the underlying mechanism(s) is not fully understood. Here, we report that 17β-estradiol (E2) selectively regulates neuronal expression of the Bcl-2 family (bcl-2, bcl-x, bcl-w, bax, bak, bad, bik, bnip3, bid, and bim). In primary cerebrocortical neuron cultures under basal conditions, we observe that E2 upregulates expression of antiapoptotic Bcl-w and downregulates expression of proapoptotic Bim in an estrogen receptor (ER)-dependent manner. In the presence of toxic levels of Aβ, we observe that E2 attenuates indices of neuronal apoptosis: c-Jun N-terminal kinase (JNK)-dependent downregulation of Bcl-w and upregulation of Bim, mitochondrial release of cytochrome c and Smac, and cell death. These neuroprotective effects of E2 against Aβ-induced apoptosis are mimicked by the JNK inhibitor SP600125 (anthra[1,9-cd]pyrazol-6(2H)-one). In addition, E2 attenuates Aβ-induced JNK phosphorylation in an ER-dependent manner, but does not affect basal levels of JNK phosphorylation. These results suggest that E2 may reduce Aβ-induced neuronal apoptosis at least in part by two complementary pathways: (1) ER-dependent, JNK-independent upregulation of Bcl-w and downregulation of Bim under basal conditions, and (2) ER-dependent inhibition of Aβ-induced JNK activation and subsequent JNK-dependent downregulation of Bcl-w and upregulation of Bim, resulting in mitochondrial release of cytochrome c and Smac and eventual cell death. These data provide new understanding into the mechanisms contributing to estrogen neuroprotection, a neural function with potential therapeutic relevance to Alzheimer's disease.
Estrogen is an established modulator of neuron viability. Elevated estrogen levels are associated with decreased neuron death in specific sexually dimorphic nuclei during development (Forger, 2006) and in adulthood after toxic challenge (Wise et al., 2001). In cell culture models, 17β-estradiol (E2) is neuroprotective against a variety of insults, including serum deprivation (Green et al., 1997), oxidative stress (Behl et al., 1995), excitotoxicity (Goodman et al., 1996; Singer et al., 1996), and β-amyloid peptide (Aβ) (Behl et al., 1995; Goodman et al., 1996; Green et al., 1996; Mook-Jung et al., 1997; Pike, 1999). The mechanisms underlying estrogen neuroprotection are not fully understood; however, several candidates have been identified. One putative mechanism of estrogen neuroprotection is regulation of the Bcl-2 family, pivotal regulators of apoptosis that include both proteins that promote cell survival (e.g., Bcl-2, Bcl-xL, and Bcl-w) and others that antagonize it (e.g., Bax, Bak, Bad, Bik, BNIP3, Bid, and Bim) (Antonsson and Martinou, 2000). Upregulation of antiapoptotic proteins, such as Bcl-2 (Garcia-Segura et al., 1998; Singer et al., 1998; Dubal et al., 1999; Nilsen and Diaz Brinton, 2003) and Bcl-xL (Patrone et al., 1999; Pike, 1999; Koski et al., 2004), or downregulation of proapoptotic Bad (Gollapudi and Oblinger, 1999) may contribute to the protective effects of estrogen. However, the relationship between estrogen and neuronal expression of Bcl-2 family members remains incompletely defined.
If members of the Bcl-2 family are important mediators of estrogen neuroprotection, then elucidating the underlying mechanism(s) will require identification of not only estrogen-regulated Bcl-2 family members but also the upstream and downstream signaling components in this pathway. One putative upstream component of regulation by estrogen of Bcl-2 family expression is c-Jun N-terminal kinase (JNK) signaling. JNK signaling is linked to transcriptional regulation of many genes (Ip and Davis, 1998), including members of the Bcl-2 family (Harris and Johnson, 2001; Bae and Song, 2003). Interestingly, JNK activation is observed in cultured neurons after Aβ exposure, and its inhibition significantly attenuates Aβ toxicity (Bozyczko-Coyne et al., 2001; Morishima et al., 2001; Troy et al., 2001; Yao et al., 2005). The mitochondrial localization of Bcl-2 proteins suggests that a downstream component of this pathway includes regulation of the proapoptotic molecules released from mitochondria, such as cytochrome c (Liu et al., 1996) and second mitochondrion-derived activator of caspase (Smac/DIABLO) (Du et al., 2000; Verhagen et al., 2000). Previous studies indicate that Aβ causes mitochondrial release of both cytochrome c (Wang et al., 2001; Agostinho and Oliveira, 2003) and Smac (Yao et al., 2005) in cultured neurons. Whether the mechanism of estrogen neuroprotection against Aβ involves regulation of JNK activation and its downstream effectors is unclear.
In this study, we examined the regulatory effects of E2 on neuronal expression of members of the Bcl-2 family under basal conditions and after insult with Aβ. Furthermore, we investigated both the upstream (e.g., JNK signaling) and downstream (e.g., cytochrome c and Smac release) components of this antiapoptotic pathway.
Materials and Methods
Primary cultures of cerebrocortical neurons were prepared from embryonic (gestational day 18) Sprague Dawley rat pups, with minor modifications of a previously described protocol (Nguyen et al., 2005). In brief, dissected cerebral cortices were incubated 5 min in 0.125% trypsin at 37°C, followed by trypsin quenching with 1 vol of DMEM containing 20% fetal bovine serum. Cell suspensions were centrifuged (5 min at 200 × g), resuspended in serum-free DMEM, mechanically dissociated by repeated passage through a fire-polished Pasteur pipette, and then filtered through a sterile 40 μm nylon mesh (Falcon, Franklin Lakes, NJ). Cells were plated on to poly-l-lysine (0.05 mg/ml)-coated multiwell plates (Nunc, Naperville, IL) at either 5 × 104 cells/cm2 (cell viability analysis) or 1.5 × 105 cells/cm2 (Western blot and RT-PCR) in serum-free, phenol red-free DMEM buffered with 26 mm bicarbonate, 20 mm HEPES, and supplemented with 100 μg/ml transferrin, 5 μg/ml insulin, 100 μm putrescine, and 30 nm selenium. Cultures were maintained in a humidified incubator at 37°C with room air supplemented to 5% CO2. These studies were conducted under a protocol approved by the Institutional Animal Care and Use Committee of the University of Southern California.
Experimental treatment of cultures.
Cortical neuron cultures were used for experimentation 3–6 d in vitro after plating. Cultures were treated with 0.001–1000 or 10 nm E2 (solubilized in 100% ethanol) (Sigma, St. Louis, MO), which in some experiments was followed 1 h later by exposure to 25 μm aggregated Aβ25–35 (Biochem, Torrance, CA) prepared as described previously (Pike et al., 1993). Estrogen receptor (ER) antagonist 7α,17β-[9-[(4,4,5,5,5-pentafluoropentyl)sulfinyl]nonyl]estra-1,3,5(10)-triene-3,17-diol (ICI 182,780) (1 μm; solubilized in 100% ethanol) (Tocris, Ellisville, MO) was added to cultures 1 h before E2. The JNK inhibitor anthra[1,9-cd]pyrazol-6(2H)-one (SP600125) (100 nm; solubilized in DMSO) (Calbiochem, La Jolla, CA) was added to cultures 1 h before Aβ or E2. Final concentrations of drug vehicles were ≤0.1%; vehicle controls were added as appropriate.
Assessment of cell viability.
Cell viability was assessed using calcein-AM and ethidium homodimer fluorescent staining (Invitrogen, Eugene, OR) as described previously (Pike, 1999). Briefly, live cells were counted in four fields per well, four to six wells per condition, in three or more independent culture preparations. The number of live cells counted per well in vehicle-treated controls ranged from 200 to 300. Cell viability is presented graphically as a percentage of live cells in the vehicle-treated, control condition.
RT-PCR was performed using a standard protocol, as described previously (Yao et al., 2005). In brief, total cellular RNA was isolated using Trizol reagent (Invitrogen) and reverse transcribed into the cDNA using the Superscript first-strand synthesis system (Invitrogen). Next, 1 μl of reverse transcription product was mixed with 2.5 U of JumpStart TaqDNA polymerase (Sigma), 20 pmol each of sense and antisense primers in a buffer containing 10 mm Tris-HCl, pH 8.3, 50 mm KCl, 2.5 mm MgCl2, and 0.2 mm of each dNTP in a volume of 50 μl. The primers used in this experiment were as follows: 5′-CCGGGAGAACAGGGTATGAT-3′, 5′-CAGGTATGCACCCAGAGTGA-3′ for bcl-2; 5′-AGGCTGGCGATGAGTTTGAA-3′, 5′-CGGCTCTCGGCTGCTGCATT-3′ for bcl-x; 5′-AGCCTCAACCCCAGACACAC-3′, 5′-AAGGCCCCTACAGTTACCAG-3′ for bcl-w; 5′-TCAGCCCATCTTCTTCCAGATGGT-3′, 5′-CCACCAGCTCTGAACAGATCATGA-3′ for bax; 5′-ACTGCGATGAGGCCCTGTCT-3′, 5′-GGCCCAACAGAACCACACCA-3′ for bak; 5′-ATGGGAACCCCAAAGCAGCC-3′, 5′-TCACTGGGAGGGAGTGGAGC-3′ for bad; 5′-ATTTCATGAGGTGCCTGGAG-3′, 5′-GGCTTCCAATCAAGCTTCTG-3′ for bik; 5′-GAATCTGGACGAAGCAGCTC-3′, 5′-AACATTTTCTGGCCGACTTG-3′ for bnip3; 5′-ACTCTGAGGTCAGCAACGGT-3′, 5′-CTAACCAAGTCCCTCACGTA-3′ for bid; 5′-GCCCCTACCTCCCTACAGAC-3′, 5′-CAGGTTCCTCCTGAGACTGC-3′ for bim; and 5′-AGCCATGTACGTAGCCATCC-3′, 5′-CTCTCAGCTGTGGTGGTGAA-3′ for β-actin (internal control). Primers were chemically synthesized (Integrated DNA Technologies, Coralville, IA). The PCR cycles consisted of initial incubation at 94°C for 1 min; denaturation at 94°C for 30 s; annealing at 52°C for 30 s; and extension at 72°C for 1 min, for 30 cycles, and final extension at 72°C for 3 min. RT-PCR products were electrophoresed on 1.7% agarose gels and visualized under UV light after ethidium bromide staining. Analysis of RT-PCR products for the bcl-x and bim primer sets was limited to the 337 bp bcl-xL band and the 319 bp bimEL band.
Design and transfection of small interfering RNAs.
Small interfering RNA (siRNA) that targets bim was designed using the target finder and design tool (Ambion, Austin, TX). The target mRNA sequence of the siRNA is 5′-AAGAUCUUCUCUGCUGUCCCG-3′, corresponding to nucleotides 240–260 of bim gene. As a negative control, a scrambled siRNA was designed consisting of the same nucleotide composition as the specific bim siRNA but lacking significant homology to the genome. As an additional negative control, a mismatched siRNA was used in which two bases in the specific bim siRNA were modified to make them noncomplementary to the target mRNA. The antisense and sense template DNA oligonucleotides for each siRNA, plus T7 promoter 5′-CCTGTCTC-3′ to the 3′ end, were chemically synthesized (Integrated DNA Technologies) and were as follows: siRNA targeting bim (sibim), 5′-AAGATCTTCTCTGCTGTCCCG-3′, 5′-AACGGGACAGCAGAGAAGATC-3′; scrambled siRNA (ncbim), 5′-AATGCCTCCGCTTGTCATCTG-3′, 5′-CAGATGACAAGCGGAGGCA-3′; mismatched siRNA (mmbim), 5′-AAGATCTTCTCGTCTGTCCCG-3′, 5′-AACGGGACAGACGAGAAGATC-3′. The synthesized template DNA was in vitro transcribed into double-strand siRNA using the Silencer siRNA construction kit (Ambion). siRNA transfection with siPORT Amine (Ambion) was performed according to the manufacturer's instructions.
Total cell lysates and mitochondrial and cytosolic extracts (prepared using a mitochondria/cytosol fractionation kit; Biovision, Mountain View, CA) were processed for Western blots using a standard protocol described previously (Pike, 1999). Briefly, lysate and extract samples were diluted into reducing sample buffer, electrophoresed for ∼1.5 h at 120 V in 15% polyacrylamide gels, and then transferred onto a polyvinylidene difluoride membrane (Millipore, Medford, MA) at constant voltage (100 V) for 1 h. After blocking of nonspecific binding (1 h incubation in 10 mm Tris, 100 mm NaCl, 0.1% Tween, 3% bovine serum albumin), membranes were incubated with primary antibody, which included goat anti-Bcl-w (Santa Cruz Biotechnology, Santa Cruz, CA), goat anti-Bim (Santa Cruz Biotechnology), mouse anti-cytochrome c (Santa Cruz Biotechnology), goat-anti-Smac (Santa Cruz Biotechnology), or mouse anti-phospho-JNK (Thr183/Tyr185) (Cell Signaling Technology, Beverly, MA). After rinsing (5 min; six times; in 10 mm Tris, 100 mm NaCl, 0.1% Tween 20), membranes were incubated in the appropriate horseradish peroxidase-conjugated secondary antibody, followed by enhanced chemiluminescence detection (Amersham Biosciences, Arlington Heights, IL). To detect total JNK or verify equal loading of protein across conditions, membranes were stripped (5 min in 100 mm glycine, pH 2.5; and then 5 min in 62.5 mm Tris, 2% SDS, 0.7% 2-mercaptoethanol, pH 6.7, at 60°C) and reprobed with rabbit anti-JNK (Cell Signaling Technology) or mouse anti-β-tubulin (Chemicon, Temecula, CA) antibody. Blots were quantified by band densitometry of scanned films using NIH Image 1.61 software. For JNK blots, the ratio of phospho:total JNK was determined and normalized to the vehicle control condition. For anti-Bim blots, the ∼24 kDa band corresponding to BimEL was quantified. Data are presented graphically as a percentage of control values.
All experiments were repeated at least three times using independent culture preparations. Quantitative data were statistically analyzed by one-way ANOVA, followed by between-group comparisons using Fisher's least significant difference test. Statistical significance was concluded with a value of p < 0.01 for all analyses.
Estrogen reduces Aβ-induced neuronal death
To confirm the established neuroprotective effect of E2 against Aβ-induced neuronal death in our experimental system (Pike, 1999; Cordey et al., 2003; Cordey and Pike, 2006), primary cerebrocortical neuron cultures were pretreated with increasing concentrations (0.001–1000 nm) of E2 for 60 min, followed by exposure to 25 μm aggregated Aβ25–35 for 6–48 h. Cell viability assays revealed that nanomolar levels of E2 significantly reduced neuronal death induced by Aβ25–35 in a dose-dependent (Fig. 1A) and time-dependent (Fig. 1B) manner. The 10 nm dose of E2 yielded maximal protection and thus was the concentration used in subsequent experiments.
Estrogen upregulates Bcl-w and downregulates Bim expression
To investigate the contributions of the Bcl-2 family to estrogen neuroprotection, we first assessed the effect of E2 on expression of bcl-2, bcl-x, bcl-w, bax, bak, bad, bik, bnip3, bid, and bim under basal culture conditions (i.e., in the absence of Aβ challenge). Exposure of neuron cultures to 10 nm E2 for 6–72 h did not induce detectable changes in mRNA levels of several Bcl-2 family members, including bak, bad, bik, and bid. Modest alterations in mRNA levels were consistently observed for four genes 12–72 h after E2 exposure: bcl-x and to a lesser extent bcl-2 were mildly increased, and bax and bnip3 were slightly decreased. The most robust findings were that E2 increased expression of antiapoptotic bcl-w and decreased expression of proapoptotic bim at 12–72 h (Fig. 2A). Because bcl-w and bim were clearly the most strongly E2-regulated Bcl-2 family genes in our system, all subsequent studies focused on these two genes.
To confirm E2 regulation of bcl-w and bim gene expression at the protein level, Bcl-w and Bim were analyzed by Western blot. Results revealed a similar time-dependent upregulation of Bcl-w (Fig. 2B, top panel) and downregulation of Bim (Fig. 2B, middle panel). Densitometry measures indicated that 48 h after E2 treatment Bcl-w was increased to ∼175% of basal level (Fig. 2C), whereas Bim was decreased to ∼35% (Fig. 2D).
Estrogen attenuates Aβ-induced Bcl-w downregulation and Bim upregulation
After determining the regulatory effects of E2 on expression of Bcl-w and Bim in neurons under basal conditions, we extended our investigation to evaluate E2 effects under Aβ challenge. First, in agreement with our recent observations (Yao et al., 2005), we found that exposure of neuron cultures to 25 μm Aβ25–35 in the absence of E2 resulted in significant, time-dependent decrease of bcl-w mRNA (Fig. 3A) and modest increase of bim mRNA (Fig. 3B). Next, to explore whether E2 affects the Aβ-induced changes in bcl-w and bim mRNAs, neuron cultures were pretreated with 10 nm E2 for 60 min, followed by exposure to 25 μm Aβ25–35 for 6–48 h. The results showed that E2 pretreatment attenuated both Aβ-induced bcl-w downregulation (Fig. 3A) and bim upregulation (Fig. 3B). These mRNA observations were confirmed at the protein level 48 h after Aβ exposure by Western blot using antibodies that detect Bcl-w (Fig. 3C, top panel) and Bim (Fig. 3C, middle panel). Densitometric analyses of Western blots showed that Aβ significantly reduced Bcl-w expression to ∼45% of basal levels (Fig. 3D) and increased Bim expression to ∼125% of basal levels (Fig. 3E) (p < 0.01 in comparison with vehicle-treated control group). Importantly, E2 pretreatment significantly attenuated the effects of Aβ on expression of Bcl-w (Fig. 3D) and Bim (Fig. 3E) (p < 0.01 relative to Aβ25–35-treated condition).
Reduction of Bim expression attenuates Aβ toxicity
If regulation by estrogen of Bcl-w and Bim expression is important to estrogen neuroprotection against Aβ, then levels of both Bcl-w and Bim should significantly contribute to Aβ toxicity. We recently demonstrated that Aβ toxicity is reduced by increased Bcl-w expression and potentiated by decreased Bcl-w expression (Yao et al., 2005). Previous research indicates that the proapoptotic Bim can contribute to Aβ toxicity (Yin et al., 2002). To confirm this finding in our culture paradigm, we evaluated the prediction that decreasing Bim expression using specific siRNA should reduce Aβ toxicity. Cultures were transfected for 1 or 3 d with siRNA directed against bim or with mismatched or scrambled bim siRNA. RT-PCR analyses show that bim mRNA was significantly reduced by the specific siRNA but by neither of the control siRNAs (Fig. 4A). Western blot analysis revealed similar effects of the siRNA on Bim protein expression (Fig. 4B). Quantitation of Western blots indicated that bim siRNA decreased Bim levels by 76% 1 d after transfection and by up to 81% 3 d after transfection in comparison with scrambled siRNA control (Fig. 4C). These results demonstrate that the designed bim siRNA had strong inhibitory effects on Bim expression at the mRNA and protein levels.
To examine the effect of Bim suppression on Aβ-induced neuron death, cultures were treated with 25 μm Aβ25–35 1 d after siRNA transfection, and cell viability was assessed 2 d later. In comparison with both the mismatched and scrambled siRNA conditions, cultures transfected with the bim siRNA showed significantly decreased Aβ-induced cell death (p < 0.01, compared with scrambled siRNA) (Fig. 4D). This finding suggests that Bim plays a significant role in regulating apoptosis pathways involved in Aβ toxicity.
Estrogen regulation of Bcl-w and Bim expression is ER dependent
To determine whether the regulatory effects of E2 on Bcl-w and Bim expression are dependent on ERs, neuron cultures were pretreated with ICI 182,780, an ER antagonist (Wakeling et al., 1991). ICI 182,780 (1 μm; effective concentration determined in Fig. 9A) exposure had no effect on basal expression of either bcl-w (Fig. 5A) or bim (Fig. 5C), but blocked E2 regulation of both bcl-w (Fig. 5A) and bim expression (Fig. 5C). Furthermore, ICI 182,780 also blocked the inhibitory effects of E2 on Aβ-induced bcl-w downregulation (Fig. 5B) and bim upregulation (Fig. 5D). These mRNA observations were confirmed at the protein level 48 h after Aβ exposure by Western blot using antibodies directed against Bcl-w (Fig. 5E, top panel) and Bim (Fig. 5E, middle panel). Quantitative analysis of blots confirmed that ICI 182,780 not only blocked E2 regulation of basal Bcl-w (Fig. 5F) and Bim expression (Fig. 5G), but also E2 inhibition of Aβ-induced Bcl-w downregulation (Fig. 5F) and Bim upregulation (Fig. 5G).
Effect of JNK inhibition on estrogen regulation of Bcl-w and Bim expression
JNK signaling is an established signaling pathway in the regulation of Bcl-2 family expression (Harris and Johnson, 2001; Bae and Song, 2003; Yao et al., 2005). To begin investigating the role of JNK signaling in E2 regulation of Bcl-2 family members, we assessed the effect of the specific JNK inhibitor SP600125 (Bennett et al., 2001) on E2 induced bcl-w upregulation and bim downregulation under nonchallenged conditions. Neuron cultures were pretreated for 60 min with 100 nm SP600125 followed by treatment with 10 nm E2 for 24 and 48 h. RT-PCR analyses showed that basal mRNA levels of bcl-w (Fig. 6A) and bim (Fig. 6C) were not affected by pharmacological inhibition of JNK. Similarly, E2-induced bcl-w upregulation (Fig. 6A) and bim downregulation (Fig. 6C) also were not altered by JNK inhibition, suggesting that the regulation of E2 on basal expression of bcl-w and bim is not dependent on JNK signaling.
Our recent data indicate that JNK signaling contributes to the mechanism by which Aβ regulates expression of Bcl-2 family members (Yao et al., 2005). We confirm here that inhibition of JNK signaling by pretreatment with 100 nm SP600125 mostly prevents Aβ-induced downregulation of bcl-w (Fig. 6B) and upregulation of bim (Fig. 6D). Although the regulation of E2 on basal expression of bcl-w and bim is not dependent on JNK signaling (Fig. 6A,C), we investigated whether JNK signaling may contribute to the inhibitory effect of E2 on Aβ-induced regulation of bcl-w and bim. Neuron cultures were pretreated with 100 nm SP600125 for 60 min, followed by treatment with 10 nm E2 for 60 min and exposure to 25 μm Aβ25–35 for 48 h. RT-PCR analyses showed that the inhibitory effects of E2 on Aβ-induced downregulation of bcl-w (Fig. 6B) and upregulation of bim (Fig. 6D) appeared to be enhanced by cotreatment with SP600125. These mRNA observations were evaluated at the protein level with Bcl-w (Fig. 6E) and Bim (Fig. 6F) Western blots. Densitometric analyses of blots confirmed that JNK inhibition attenuated Aβ-induced downregulation of Bcl-w (Fig. 6G) (*p < 0.01 relative to Aβ treatment) and upregulation of Bim (Fig. 6H) (*p < 0.01 relative to Aβ treatment). However, the effect of estrogen and SP600125 cotreatment was not significantly greater than the effect of SP600125 alone for either Aβ-induced downregulation of Bcl-w (Fig. 6G) or upregulation of Bim (Fig. 6H). These data suggest the possibility that the inhibitory effect of E2 on Aβ-induced changes of Bcl-w and Bim may involve not only a JNK-independent regulatory action on Bcl-w and Bim expression, but also inhibition of Aβ-induced JNK signaling.
Estrogen reduces Aβ-induced JNK activation
To explore whether E2 regulates JNK signaling, we first determined the effect of E2 on activation of JNK under basal conditions. Neuron cultures were treated with 10 nm E2 for 5 min to 48 h and whole-cell extracts were analyzed by Western blot using antibodies that recognize total JNK and phosphorylated JNK, an indicator of activation. Results showed that two isoforms of JNK (46 and 54 kDa) were detected. E2 treatment affected neither phosphorylated (Fig. 7A, top panel) nor total (Fig. 7A, bottom panel) JNK protein levels, indicating that E2 does not significantly affect JNK signaling under nonchallenge conditions (Fig. 7B).
Next, we determined the effect of E2 on Aβ-induced JNK activation. Neuron cultures were pretreated with 10 nm E2 for 60 min, followed by exposure to 25 μm Aβ25–35. Western blot results show that Aβ25–35 triggers JNK phosphorylation that was detectable as early as 3 h after Aβ treatment and persisted at least through 48 h (Fig. 7C). Both the 46 and 54 kDa JNK bands showed an Aβ-induced increase in phosphorylation, with the 54 kDa band showing the more robust increase (Fig. 7D). E2 pretreatment incompletely blocked the Aβ-induced increase in JNK activation at most time points (Fig. 7C, top panel; D) but did not alter total JNK protein levels (Fig. 7C, bottom panel). E2 similarly inhibited Aβ-induced phosphorylation of the 46 and 54 kDa JNK bands. To assess the role of ER in this effect, neuron cultures were pretreated with ER antagonist ICI 182,780 (1 μm) for 60 min, followed by treatment with 10 nm E2 for 60 min, and then exposed to 25 μm Aβ25–35 for 6 h. Western blots indicated that ICI 182,780 blocked the inhibitory effect of E2 on Aβ-induced JNK phosphorylation (Fig. 7E,F), demonstrating that the regulatory effect of E2 on Aβ-induced JNK activation occurs in an ER-dependent manner. For comparison, we confirmed our previous finding (Yao et al., 2005) that the JNK inhibitor SP600125 both reduces basal levels of JNK phosphorylation and blocks Aβ-induced JNK phosphorylation of the 46 and 54 kDa JNK bands (Fig. 7G,H).
Estrogen attenuates JNK-dependent mitochondrial cytochrome c and Smac release induced by Aβ
Bcl-2 family members exert their effects in part through regulating mitochondrial release of cytochrome c and Smac into the cytosol (Kuwana and Newmeyer, 2003), an event that is a general feature of the mitochondrial pathway of apoptosis (Liu et al., 1996; Du et al., 2000; Verhagen et al., 2000). In this paradigm, we previously reported that Bcl-w overexpression effectively inhibited Aβ-induced Smac release from mitochondria, whereas knock-down of Bcl-w expression increased Smac release (Yao et al., 2005). On the contrary, Bim can increase mitochondrial cytochrome c (Putcha et al., 2001; Whitfield et al., 2001) and Smac (Yin et al., 2002) release. Because we observed that E2 increases Bcl-w and decreases Bim expression, E2 may regulate Aβ-induced mitochondrial cytochrome c and Smac release. To investigate this possibility and the potential involvement of JNK signaling, we analyzed by Western blot mitochondrial and cytosolic extracts of neuron cultures treated with 25 μm Aβ25–35 in the presence and absence of either 100 nm SP600125 or 10 nm E2. We observed that, 48 h after Aβ25–35 treatment, cytochrome c (Fig. 8A) and Smac (Fig. 8B) levels decreased in the mitochondrial fraction and increased in the cytosolic fraction. Pretreatment with JNK inhibitor SP600125 mostly prevented Aβ-induced cytochrome c and Smac release, suggesting a JNK-dependent mechanism. Similarly, E2 also attenuated Aβ-induced mitochondrial release of both cytochrome c (Fig. 8C) and Smac (Fig. 8D). Densitometric analyses of blots showed that E2 significantly reduced Aβ-induced depletion of cytochrome c (Fig. 8E) and Smac (Fig. 8F) from mitochondria and accumulation of cytochrome c (Fig. 8G) and Smac (Fig. 8H) in cytosol (*p < 0.01 relative to 24 h Aβ treatment; #p < 0.01 relative to 48 h Aβ treatment).
Estrogen neuroprotection against JNK-dependent Aβ toxicity is mediated by estrogen receptor
The above studies suggest that E2 antagonizes Aβ-induced apoptosis signaling by an ER-dependent mechanism that operates at least in part by inhibiting JNK activation. Next, we investigated the effects of ER antagonism on the ability of E2 to attenuate Aβ-induced neuron death. To verify that the neuroprotective effect of E2 on Aβ toxicity is mediated by ER, neuron cultures were pretreated with increasing concentrations (0.01–10 μm) of the ER antagonist ICI 182,780, followed by 48 h exposure to 25 μm Aβ25–35 in the presence or absence of 10 nm E2. ICI 182,780 alone had no effect on cell viability under basal conditions and after Aβ challenge (Fig. 9A). However, at concentrations of 1 μm and above, ICI 182,780 almost completely blocked E2-mediated neuroprotection against Aβ toxicity (p < 0.01 in comparison with E2 plus Aβ treatment) (Fig. 9A).
We observed that E2 inhibits JNK signaling and that JNK signaling contributes to Aβ-induced changes in Bcl-w and Bim expression and mitochondrial release of cytochrome c and Smac,. Here, we extended these observations to neuronal survival. Neuron cultures were exposed for 48 h to 25 μm Aβ25–35 with or without pretreatment with 100 nm SP600125 and or 10 nm E2. Cell viability assays showed that independent treatment with either SP600125 or E2 significantly attenuated neuronal death induced by Aβ25–35 (Fig. 9B). Cotreatment with SP600125 and E2 resulted in a modest increase in neuroprotection that was significantly greater than E2 treatment alone but not significantly greater than SP600125 treatment alone (Fig. 9B).
Although abundant evidence has established estrogen as a neuroprotective factor that is relevant across the life span, elucidation of its protective mechanisms has proven challenging. One compelling mechanism involves regulation of the Bcl-2 family. Findings from this study not only further define regulation by estrogen of Bcl-2 family members, they also identify two separate pathways of regulation: (1) increased expression of antiapoptotic Bcl-w and decreased expression of proapoptotic Bim under basal conditions, and (2) attenuation of JNK-dependent changes in expression of Bcl-2 family that occurs under apoptotic challenge (Fig. 10).
Consistent with an emerging literature, we found that, in the absence of toxic challenge, estrogen regulates Bcl-2 family members in a manner that favors antagonism of apoptosis. Our most robust findings are the novel observations that E2 significantly increases basal expression of antiapoptotic Bcl-w and decreases basal expression of proapoptotic Bim. Bcl-w is widely expressed in mammalian tissues, including CNS (Print et al., 1998; Hamnér et al., 1999; O'Reilly et al., 2001), and functions as a negative regulator of neuronal apoptosis (Gibson et al., 1996; Hamnér et al., 2001; Middleton et al., 2001). Neural expression of Bcl-w is highest in the mature brain (Hamnér et al., 1999), suggesting that Bcl-w function may be particularly important in adulthood. In a previous study, we reported that Aβ-induced neuronal death was reduced by Bcl-w overexpression and potentiated by Bcl-w suppression (Yao et al., 2005), findings that suggest a critical role for Bcl-w in Aβ-induced apoptosis.
In contrast to Bcl-w, Bim is a proapoptotic protein that exists in several isoforms: BimEL, BimL, BimS, and BimAD (O'Connor et al., 1998; U et al., 2001; Marani et al., 2002). In the CNS, Bim expression is localized primarily in neurons (O'Reilly et al., 2000) and is upregulated in a variety of neuron death paradigms (Putcha et al., 2001; Whitfield et al., 2001; Becker et al., 2004; Biswas et al., 2005). Inhibition of Bim by antisense and genetic knock-out approaches can significantly reduce neuronal apoptosis (Whitfield et al., 2001; Becker et al., 2004). In cultured cerebral endothelial cells, Aβ toxicity is associated with increased Bim expression and is suppressed by Bim knock-down (Yin et al., 2002). Consistent with our observations, these results implicate Bim in the mechanism of Aβ-induced apoptosis. Bim produces its proapoptotic effects by interacting with and neutralizing antiapoptotic Bcl-2 family proteins such as Bcl-w (O'Connor et al., 1998; Puthalakath et al., 1999; Wilson-Annan et al., 2003; Shinoda et al., 2004). Thus, E2-induced upregulation of Bcl-w and downregulation of Bim should result in complementary inhibition of neuronal apoptosis induced by toxins such as Aβ.
In addition to E2 regulation of Bcl-w and Bim, we observed relatively modest elevations in mRNA levels of antiapoptotic Bcl-2 family members bcl-2 and bcl-x, and small but consistent reductions in mRNA levels of proapoptotic bax and bnip3. Consistent with our findings are previous results in neuronal cultures showing that estrogen increases basal expression of Bcl-2 (Singer et al., 1998; Honda et al., 2001; Nilsen and Diaz Brinton, 2003; Zhao et al., 2004) and Bcl-xL (Patrone et al., 1999; Pike, 1999; Koski et al., 2004) and decreases expression of proapoptotic bnip2 (Belcredito et al., 2001). Similar results have been observed in rodent brain under nonchallenge conditions. For example, accumulating evidence links developmental sex differences in E2 exposure to differences in Bcl-2 family expression and neuron survival in sexually dimorphic brain regions (for review, see Forger, 2006). In adult female rats, E2 positively regulates expression of Bcl-2 in hypothalamus (Garcia-Segura et al., 1998) and bcl-x in hippocampus (Stoltzner et al., 2001).
One interpretation of the current data is that estrogen functions as an endogenous, homeostatic regulator of apoptosis signaling. Under normal conditions in estrogen-responsive brain regions, estrogen may help maintain long-term neuronal viability by regulating the expression of several Bcl-2 family members. Because apoptosis is affected by the net interactions of proapoptotic and antiapoptotic Bcl-2 members, we speculate that neuroprotective actions of estrogen reflect additive regulatory effects on Bcl-w, Bim, and other Bcl-2 proteins. This “E2 maintenance pathway” (Fig. 10) appears to be ER dependent, but it is unclear whether the mechanism involves predominantly classic genomic pathways (i.e., interaction of activated ER with estrogen response elements on target genes) (Pike, 1999; Perillo et al., 2000) and/or indirect genomic pathways activated by cell signaling pathways [e.g., CREB (cAMP response element-binding protein) signaling] (Honda et al., 2001; Wu et al., 2005). Although activation of JNK signaling is linked to regulation of Bcl-2 family members (Harris and Johnson, 2001; Bae and Song, 2003), our data do not implicate JNK signaling in the E2 maintenance pathway, because under basal conditions E2 did not affect JNK phosphorylation and JNK inhibition did not affect E2 regulation of Bcl-w and Bim.
In addition to regulating expression of Bcl-2 family members under basal conditions, our data show that E2 antagonizes proapoptotic changes in Bcl-2 family expression induced by toxic challenge. Consistent with our previous report (Yao et al., 2005), we found that Aβ-induced neuronal apoptosis involves JNK-dependent alterations in Bcl-2 family expression: downregulation of Bcl-w and upregulation of Bim. Other members of the Bcl-2 family may also play a significant role in Aβ toxicity. We observed that E2 significantly attenuated Aβ-induced changes in Bcl-w and Bim expression. Notably, E2 also significantly reduced Aβ-induced JNK phosphorylation, suggesting that E2 attenuation of Aβ-induced changes in Bcl-w and Bim expression involves inhibition of JNK activation.
Activation of JNK signaling has been closely linked to a variety of apoptotic stimuli, whereas inhibition of JNK signaling provides protection against neuronal apoptosis in multiple paradigms, including Aβ neurotoxicity (Bozyczko-Coyne et al., 2001; Morishima et al., 2001; Troy et al., 2001). It appears that JNK signaling promotes apoptosis at least in part via both transcriptional and posttranslational regulation of Bcl-2 family members (Sanchez and Yuan, 2001). Consequently, JNK inhibition can block alterations in Bcl-2 family expression induced during apoptosis (Harris and Johnson, 2001; Linseman et al., 2002; Schuster et al., 2002; Lei and Davis, 2003; Okuno et al., 2004; Yao et al., 2005; Papadakis et al., 2006). Our results show that inhibition of JNK phosphorylation by both the pharmacological inhibitor SP600125 and E2 attenuated Aβ-induced changes in Bcl-w and Bim expression, mitochondrial cytochrome c and Smac release, and neuron death. Consistent with our data are findings in non-neural cell types that E2 can reduce JNK activation, which results in attenuation of JNK-dependent gene expression (Srivastava et al., 1999, 2001) and/or increased cell survival (Razandi et al., 2000; Eckhoff et al., 2003). Interestingly, despite significantly attenuating indices of apoptosis, E2 provided only partial protection against Aβ-induced cell death. Incomplete neuroprotection in primary neuron cultures is commonly observed with not only estrogen (Singer et al., 1996; Pike, 1999; Harms et al., 2001; Honda et al., 2001) but also androgens (Ahlbom et al., 2001; Hammond et al., 2001; Pike, 2001) and progesterone (Nilsen and Brinton, 2002), suggesting that sex steroid hormones act as partial modulators of neuronal apoptosis pathways.
Thus, E2 may regulate Bcl-2 family members not only by an E2 maintenance pathway but also by an “E2 response pathway” (Fig. 10). In the latter case, E2 can respond to at least some toxic challenges by an ER-dependent pathway to inhibit activation of cell signaling pathways that induce proapoptotic changes in Bcl-2 family expression (e.g., JNK signaling). Importantly, the literature indicates several mechanisms that contribute to estrogen neuroprotection (Green and Simpkins, 2000). We speculate that the pathways identified in this study likely function in conjunction with other mechanisms of estrogen neuroprotection.
Recent observations in animal models are consistent with an E2 response pathway in which E2 responds to neural injury by regulating expression of Bcl-2 family members in a way that reduces apoptosis. For example, in ischemic brain injury models, E2 is neuroprotective and attenuates injury-induced downregulation of Bcl-2 (Dubal et al., 1999; Alkayed et al., 2001; Zhang et al., 2004) and, in some reports, upregulation of Bax (Won et al., 2006). Similarly, E2 neuroprotection is associated with increased Bcl-2 expression after traumatic brain injury (Soustiel et al., 2005) and bcl-2 and bcl-x levels after spinal cord injury (Yune et al., 2004). Such an E2 response pathway may be particularly important for resistance to neuronal death associated with stroke, Alzheimer's disease, and other age-related disorders.
Together, our results are consistent with the hypothesis that E2 attenuates Aβ-induced neuronal death, at least in part, by ER-dependent upregulation of Bcl-w and downregulation of Bim, as well as by inhibition of Aβ-induced JNK activation, subsequent JNK-dependent downregulation of Bcl-w and upregulation of Bim, and mitochondrial release of cytochrome c and Smac. Furthermore, our findings suggest that these neuroprotective actions of E2 reflect concurrent effects of at least two different, ER-dependent signaling pathways: (1) an E2 maintenance pathway of JNK-independent regulation of Bcl-2 family members under basal, nonchallenge conditions, and (2) an E2 response pathway of inhibition of JNK-dependent regulation of Bcl-2 family members associated with neural injury. These data provide new understanding into the mechanisms contributing to estrogen neuroprotection, a function with potential therapeutic relevance to Alzheimer's disease and other age-related neurodegenerative disorders.
This work was supported by National Institutes of Health Grants AG26752 and AG23739.
- Correspondence should be addressed to Dr. Christian J. Pike, Davis School of Gerontology, University of Southern California, 3715 McClintock Avenue, Los Angeles, CA 90089-0191.