Address correspondence and reprint requests to Christian J. Pike, University of Southern California, Andrus Gerontology Center, 3715 McClintock Ave., Los Angeles, CA 90089–0191, USA. E-mail: email@example.com
It has been previously demonstrated that estrogen can protect neurons from a variety of insults, including β-amyloid (Aβ). Recent studies have shown that estrogen can rapidly modulate intracellular signaling pathways involved in cell survival. In particular, estrogen activates protein kinase C (PKC) in a variety of cell types. This enzyme plays a key role in many cellular events, including regulation of apoptosis. In this study, we show that 17β-estradiol (E2) rapidly increases PKC activity in primary cultures of rat cerebrocortical neurons. A 1 h pre-treatment with E2 or phorbol-12-myristate-13-acetate (PMA), a potent activator of PKC, protects neurons against Aβ toxicity. Protection afforded by both PMA and E2 is blocked by pharmacological inhibitors of PKC. Further, depletion of PKC levels resulting from prolonged PMA exposure prevents subsequent E2 or PMA protection. Our results indicate that E2 activates PKC in neurons, and that PKC activation is an important step in estrogen protection against Aβ. These data provide new understanding into the mechanism(s) underlying estrogen neuroprotection, an action with therapeutic relevance to Alzheimer's disease and other age-related neurodegenerative disorders.
17β-Estradiol (E2), Aβ23–35 and insulin-like growth factor I (IGF-I) were purchased from Sigma Chemicals (St. Louis, MO, USA). Aβ1–42 peptide was obtained from US Peptide (Rancho Cucamonga, CA, USA) and as a generous gift from Dr C. Glabe (UC Irvine, CA, USA). Phorbol-12-myristate-13-acetate (PMA), 4α-phorbol-12,13-didecanoate (4α-phorbol), GF 109203X (bisindolylmaleimide I), Gö 6983 and bisindolylmaleimide V were purchased from Calbiochem (La Jolla, CA, USA). zVAD-fmk (zVAD) was obtained from Enzyme Systems (Dublin, CA, USA). Antibodies included PKCα (1 : 1000), PKCθ (1 : 250), PKCι (1 : 250) (from PKC sampler kit, Transduction Laboratories, Lexington, KY, USA), β-tubulin (1 : 500; Boehringer-Mannheim Biochemicals, Indianapolis, IN, USA), HRP-conjugated anti-rabbit (1 : 10000; Pierce, Rockford, IL, USA) and anti-mouse (1 : 3000; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Cell culture reagents were purchased from Invitrogen (Buffalo, NY, USA). All other chemicals were acquired from Fisher Scientific (Pittsburgh, PA, USA) unless otherwise stated.
Neuron-enriched (about 95% neuronal) cultures were generated with slight modifications of standard techniques previously described (Pike et al. 1993). Briefly, cerebral cortices of gestational day 17 Sprague–Dawley rat pups (n = 5 pups per preparation) were dissected then dissociated, both enzymatically (0.125% trypsin) and mechanically, before filtering through a 40 µm cell strainer (Falcon, Franklin Lakes, NJ, USA). The cell suspension was diluted in Neurobasal medium containing the serum-free, antioxidant-free supplement B27 and 0.5 mm l-glutamine, then plated at a cell density of 5 × 104 cells/cm2 in either 48- or six-well plates (cell viability and western blots, respectively), or at 1 × 105 cells/cm2 in 10 cm dishes (PKC activity assay) previously coated with poly-l-lysine (0.05 mg/mL). Cells were maintained at 37°C in a humidified incubator with 95% room air/5% CO2.
Cultures were used for experimentation 3–5 days in vitro after plating. For cell viability experiments, cultures were exposed to E2, PMA, IGF-I or zVAD beginning 1 h before administration of aggregated Aβ prepared as previously described (Pike et al. 1993). PKC inhibitors were added 1 h prior to exposure to E2, PMA, IGF-I or zVAD. E2 was solubilized in 100% ethanol and IGF-I was solubilized in 0.1 m acetic acid. All other drugs were solubilized in dimethylsulfoxide (DMSO) and diluted to a final vehicle concentration of ≤ 0.2%. Control conditions were treated with the appropriate amount of vehicle.
Cell viability was assessed 20–24 h after administration of Aβ, using calcein-AM and ethidium homodimer fluorescent staining (Molecular Probes, Eugene, OR, USA) as previously described (Pike 1999). Briefly, live cells were counted in four fields per well, three 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. Raw data were analyzed statistically by anova followed by pairwise comparisons using Fisher's LSD test (p < 0.05). Cell viability is presented graphically as a percentage of the number of live cells in the vehicle-treated control condition.
Cells were processed for western blots using a standard protocol previously described (Pike 1999). Briefly, cell lysates in reducing sample buffer were electrophoresed for approximately 1.5 h at 120 V in 10% polyacrylamide gels, then transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Medford, MA, USA) at constant voltage (100 V) for 1 h. Following blocking (10 mm Tris, 100 mm NaCl, 0.1% Tween, 3% bovine serum albumin), membranes were sequentially incubated with PKC antibody, then horseradish peroxidase-conjugated secondary antibody, followed by enhanced chemiluminescence detection (Amersham, Arlington Heights, IL, USA). To verify equal loading of protein among conditions, membranes were stripped (5 min 100 mm glycine pH 2.5, then 5 min 62.5 mm Tris, 2% sodium dodecyl sulfate (SDS), 0.7% 2-β-mercaptoethanol, pH 6.7 at 60°C) and re-probed with β-tubulin antibody.
PKC activity assay
Cultures were treated with vehicle, 10 nm E2 or 10 ng/mL PMA for the indicated time periods (0–30 min) then assayed for PKC activity as previously described (Gopalakrishna et al. 1992). This assay specifically measures PKC activity, but does not discriminate between PKC isozymes. Briefly, after cells were collected in homogenization buffer (20 mm Tris-HCl, 1 mm EDTA, 0.5 mm phenylmethylsulfonyl fluoride (PMSF), 20 mm leupeptin, 0.15 mm pepstatin A, pH 7.5), PKC was isolated from both soluble and 1% NP-40 detergent-solubilized membrane fractions using a small (0.5 mL) diethylaminoethyl (DEAE) cellulose column. The calcium- and phosholipid-stimulated PKC activity (proform) was eluted using 0.1 m NaCl. PKC reaction samples containing 20 mm Tris-HCl buffer, pH 7.4, 10 mm MgCl2, 0.33 mm CaCl2, 0.1 m ATP, 0.1 mg/mL histone H1 and 0.04 µm leupeptin were incubated in a 96-well filtration plate at 30°C for 5 min. Transfer of 32P to histone H1 was determined after filtration of these samples. PKC activity was expressed as the difference in the activity that was observed in the presence of phosphatidylserine (20 µg/mL)/diolein (0.8 µg/mL) and the basal activity that was observed in the presence of EGTA (4 mm). Protein was measured using a dye-binding method (Bradford 1976). PKC activity, determined in duplicate samples from three independent culture preparations, is expressed as units per milligram of protein, where one unit of enzyme transfers one nanomole of phosphate to histone H1 per minute at 30°C. Data were analyzed statistically by an unpaired t-test (p < 0.05).
E2 activates PKC in primary cortical neurons
In order to determine the role of PKC in estrogen neuroprotection, we first evaluated the ability of E2 to activate PKC in cortical neuron cultures. We used PMA, a potent and chronic activator of PKC (Nishizuka 1992), as a positive control. In comparison to vehicle-treated controls, PMA consistently induced a robust, approximately threefold increase in PKC activity in the membrane fraction (Fig. 1). In comparison, treatment with 10 nm E2 also significantly increased PKC activity in the membrane fraction, but at a lower level than observed with PMA (Fig. 1). E2 activation of PKC was rapid, occurring within 5–15 min, and transient, returning to control levels after 30 min. PMA typically caused a concomitant decrease in cytosolic PKC activity, indicating translocation to the membrane. In contrast, E2 did not significantly decrease cytosolic PKC activity (data not shown).
PKC activation is neuroprotective
To determine whether activation of PKC pathways protects neurons against Aβ, we treated cultures with increasing concentrations of PMA beginning 1 h before and maintained during a 24 h exposure to 25 µm aggregated Aβ25–35 (Fig. 2a). Our results show a dose-dependent reduction in Aβ toxicity following PMA treatment. The inactive 4α-phorbol had no effect on Aβ toxicity. Following the same paradigm, E2 also significantly reduced Aβ toxicity in a dose-dependent manner, but to a lesser extent than PMA (Fig. 2b, see also Figs 3a, b and c). E2 (10 nm) or 10 ng/mL PMA provided maximal protection, doses used in the subsequent experiments. Treatment of cultures with PMA or E2 alone did not significantly affect cell viability (data not shown).
PKC inhibitors block E2 protection
If the effects of E2 on neuronal viability are mediated by PKC, then blocking PKC activation should reduce neuroprotection afforded by E2. In order to investigate this possibility, we pre-treated our cultures for 1 h with two specific PKC inhibitors, GF 109203X (2 µm) or Gö 6983 (2 µm). At the concentration used, Gö 6983 inhibits all PKC isozymes except PKCµ (Gschwendt et al. 1996). GF 109203X exhibits high affinity for the conventional PKCs (α, β, γ) as well as the novel isozymes PKCδ and PKCε (Toullec et al. 1991). These experiments yielded several interesting findings. First, these inhibitors had no significant effect on either cell viability or Aβ toxicity after 24 h (Fig. 4a). Second, PMA protection was completely blocked by GF 109203X and Gö 6983 (Fig. 4b). Finally, GF 109203X and Gö 6983 completely blocked E2 protection against Aβ25–35 (Fig. 4c) as well as full length Aβ1–42 (Fig. 4d, see also Figs 3c and d), providing pharmacological evidence of PKC involvement in the pathway of estrogen neuroprotection. Bisindolylmaleimide V, an inactive structural analog of GF 109203X and Gö 6983, had no effect on either PMA or estrogen protection (Figs 4b, c and d).
To examine the relationship between PKC inhibition and modulation of neuroprotection further, we evaluated the effect of PKC inhibitors on neuroprotection afforded by two other known protective agents, zVAD, a broad-spectrum caspase inhibitor, and IGF-I (Dore et al. 1997). Like estrogen, both zVAD and IGF-I provided protection against Aβ toxicity in our system. However, unlike estrogen, protection by IGF-I and zVAD was not blocked by the PKC inhibitor Gö 6983 (Fig. 5), indicating that inhibition of PKC pathways does not indiscriminately attenuate neuroprotection.
PKC depletion blocks E2 protection
One interesting characteristic of phorbol esters is that when applied chronically, they down-regulate the expression of conventional and novel PKCs (Szallasi et al. 1994). We exploited this property to evaluate estrogen neuroprotection in PKC-depleted cultures. As expected, we observed that a 24 h exposure to 100 ng/mL PMA depleted conventional (α and β only, since γ was not detected by immunoblotting in our cultures) and novel (δ, ε and θ) isozymes, but not atypical PKCs (ι). Representative western blot results for a conventional (PKCα), novel (PKCθ) and atypical (PKCι) PKC isozyme are shown in Fig. 6. In contrast to PMA, the same dose of 4α-phorbol did not affect protein levels of any PKC isozyme. As a control, we verified that neither PMA nor 4α-phorbol treatment affected β-tubulin levels (Fig. 6). Following this 24 h exposure to 100 ng/mL PMA or 4α-phorbol, we treated the cultures with 10 nm E2 or 10 ng/mL PMA for 1 h before adding 25 µm Aβ25–35. Cell viability was assessed 20–24 h later. PKC depletion with 100 ng/mL PMA did not prevent Aβ toxicity, but blocked PMA protection (Fig. 7a). Similarly, E2 protection was blocked by PKC depletion (Fig. 7b). Exposure to 100 ng/mL PMA or 4α-phorbol for 48 h did not significantly affect neuronal viability compared with vehicle-treated control (Fig. 7a).
In this study, we provide evidence that E2 activates PKC in primary cultures of cerebrocortical neurons, and that this activation mediates estrogen protection against Aβ toxicity in vitro. We observed that inhibition of PKC pathways by either of two separate strategies blocked estrogen protection. First, we found that antagonism of PKC with two specific inhibitors prevented the protective actions of estrogen. Second, PKC depletion induced by chronic PMA exposure also prevented estrogen protection. These data clearly support a role for PKC activation in the mechanism of estrogen neuroprotection. This interpretation is validated by the observation that inhibition and depletion of PKC blocked protection afforded by PMA, a potent activator of PKC.
This study provides the first evidence indicating that estrogen can rapidly activate PKC in primary neuron cultures. We found that estrogen elicited a transient but substantial increase in PKC activity in the membrane fraction of neurons. The fact that we did not observe a decrease in cytosolic activity is consistent with a prior observation of PKC activation by estrogen in non-neuronal cells (Sylvia et al. 1998), perhaps suggesting activation of PKC already present at the membrane. Our data complement previous work showing that estrogen treatment increases PKC activity in the preoptic area of female rats (Ansonoff and Etgen 1998), and that PKC inhibition blocks estrogen attenuation of µ-opioid-induced hyperpolarization in hypothalamic slices (Kelly et al. 1999). How estrogen might activate PKC in neurons is not clear. In non-neuronal cells, a rapid estrogen stimulation of PKC activity has been shown to be caused by non-genomic actions involving G-protein coupled phospholipase C activation in chondrocytes (Sylvia et al. 2000) and inositol triphosphate in HEPG2 cells (Marino et al. 1998). Such non-genomic actions are believed to be mediated at least in part by estrogen binding to putative membrane-associated receptors (Toran-Allerand et al. 2002). A similar type of rapid PKC activation involving membrane receptor-elicited non-genomic actions was reported with another steroid 1,25-dihydroxyvitamin D3 (Sylvia et al. 1996). The potential role of estrogen receptor subtypes in mediating estrogen activation of PKC, as well as identification of the upstream components of the signaling cascade, are the subject of ongoing and future studies.
The PKC family has been implicated in the regulation of cell death in non-neuronal as well as neuronal cells. PKCs are classically divided into conventional (α, βI, βII, γ), novel (δ, ε, η/Δ, θ, µ) and atypical (ζ, λ/ι) isozymes. The conventional (c)PKCs require Ca2+ and diacylglycerol (DAG) for activation, whereas the novel (n)PKCs only require DAG; Ca2+ and DAG are not involved in activation of the atypical isozymes (aPKCs). These differences are responsible for the selective activation of cPKCs and nPKCs by phorbol esters, which act at the DAG-binding site of the kinase (Nishizuka 1992). Therefore, PMA-induced neuroprotection in our system may be mediated by either one or several cPKC or nPKC isozymes. Similarly, our data indicate a role for cPKC and/or nPKC isozymes in estrogen neuroprotection, as protection was blocked not only by PMA-induced down regulation of cPKCs and nPKCs, but also by general (Gö 6983) and cPKC/nPKC-specific (GF 109203X) PKC inhibitors. The identification of the specific PKC isozyme(s) mediating estrogen neuroprotection is the subject of ongoing studies.
The use of phorbol esters as PKC activators has led to contradictory results regarding the role of PKC in apoptosis. Indeed, emerging data suggest that cPKCs and aPKCs antagonize apoptotic pathways of cell death, whereas some nPKCs can promote apoptosis (Deacon et al. 1997). It is highly unlikely that PKC activation contributes to Aβ toxicity in this case, since we observed that both pharmacological inhibition and depletion of PKC pathways did not prevent Aβ-induced cell death. Rather, our data indicate that PKC activation by PMA and E2 triggers neuroprotective pathways. Similar protective roles of PKC activation against Aβ toxicity have been reported (Behrens et al. 1999; Xie et al. 2000). It is noteworthy that PMA, in contrast to DAG, is metabolically stable and characterized by more robust and prolonged PKC activation than is typically observed with physiological regulators such as estrogen (Jaken 1996). The observed higher level of neuroprotection afforded by PMA in comparison with estrogen parallels these differences in enzyme activation. Thus, it appears that while maintained activation of PKC yields strong attenuation of cell death, even the observed transient activation of PKC by estrogen is sufficient to induce significant neuroprotection.
How PKC activation leads to neuroprotection has yet to be clearly defined in neurons. Studies in non-neuronal cells indicate that PKC activation can prevent apoptosis via two main survival pathways: the anti-apoptotic protein Bcl-2 and the mitogen-activated protein kinase (MAPK)/extracellular regulated kinase (ERK) cascades. In fact, PKCα can phosphorylate Bcl-2 at a site that increases its anti-apoptotic function (Ruvolo et al. 1998), and overexpression of PKCε results in increased expression of Bcl-2 (Gubina et al. 1998). Furthermore, MAPK/ERK cascades, which have been shown to inhibit apoptosis in a number of systems, can be activated by PKC. For example, PKCα phosphorylates and activates raf-1, an upstream kinase in the MAPK/ERK pathway (Kribben et al. 1993). Pharmacological inhibition of MAPK/ERK signaling blocks phorbol ester-induced protection of neuronal cells against glutamate toxicity (Maher 2001), but not against serum deprivation (Behrens et al. 1999).
Clinical reports suggest that estrogen replacement therapy decreases the risk of developing AD (Yaffe et al. 1998). The protective actions of estrogen are theorized to involve not only promotion of neuronal survival, but also reduction of Aβ burden by enhancement of non-amyloidogenic processing of the amyloid precursor protein (APP) (Jaffe et al. 1994; Xu et al. 1998). Interestingly, many (Gandy et al. 1993; Hung et al. 1993; Savage et al. 1998) though not all (LeBlanc et al. 1998) reports suggest that phorbol ester-induced activation of PKC is one of several cell signaling pathways (Gouras et al. 1998; Petanceska and Gandy 1999) that can drive non-amyloidogenic APP metabolism. Together with our data, these observations suggest that PKC activation may be a common mechanistic feature underlying the key protective actions of estrogen against AD.
In conclusion, our results provide novel insight into the mechanism of estrogen neuroprotection and implicate PKC as a mediator of rapid effects of estrogen in the brain. Extensive understanding of estrogen actions is particularly important for the development of therapeutic strategies against neurodegenerative disorders such as AD.
This study was supported by a grant from the NIA (AG15961).