Astrocyte-Derived Transforming Growth Factor-? Mediates the Neuroprotective Effects of 17?-Estradiol: Involvement of Nonclassical Genomic Si
http://www.100md.com
内分泌学杂志 2005年第6期
Institute of Molecular Medicine and Genetics, Program in Developmental Neurobiology, and Department of Neurology, Medical College of Georgia, Augusta, Georgia 30912
Address all correspondence and requests for reprints to: Dr. Darrell W. Brann, Institute of Molecular Medicine and Genetics, Program in Developmental Neurobiology, 1120 15th Street, Medical College of Georgia, Augusta, Georgia 30912. E-mail: dbrann@mail.mcg.edu.
Abstract
17?-Estradiol (E2) and selective estrogen receptor modulators (SERMs), such as tamoxifen, mediate numerous effects in the brain, including neurosecretion, neuroprotection, and the induction of synaptic plasticity. Astrocytes, the most abundant cell type in the brain, influence many of these same functions and thus may represent a mediator of estrogen action. The present study examined the regulatory effect and underlying cell signaling mechanisms of E2-induced release of neurotropic growth factors from primary rat cortical astrocyte cultures. The results revealed that E2 (0.5, 1, and 10 nM) and tamoxifen (1 μM) increased both the expression and release of the neuroprotective cytokines, TGF-?1 and TGF-?2 (TGF-?), from cortical astrocytes. The stimulatory effect of E2 was attenuated by the estrogen receptor (ER) antagonist, ICI182,780, suggesting ER dependency. The effect of E2 also appeared to involve mediation by the phosphotidylinositol 3-kinase (PI3K)/Akt signaling pathway, because E2 rapidly induced Akt phosphorylation, and pharmacological or molecular inhibition of the PI3K/Akt pathway prevented E2-induced release of TGF-?. Additionally, the membrane-impermeant conjugate, E2-BSA, stimulated the release of TGF-?, suggesting the potential involvement of a membrane-bound ER. Finally, E2, tamoxifen, and E2-BSA were shown to protect neuronal-astrocyte cocultures from camptothecin-induced neuronal cell death, effects that were attenuated by ICI182,780, Akt inhibition, or TGF-? immunoneutralization. As a whole, these studies suggest that E2 induction of TGF-? release from cortical astrocytes could provide a mechanism of neuroprotection, and that E2 stimulation of TGF-? expression and release from astrocytes occurs via an ER-dependent mechanism involving mediation by the PI3K/Akt signaling pathway.
Introduction
THE OVARIAN STEROID hormone, 17?-estradiol (E2), exerts a wide array of actions in the central nervous system (CNS). Although most work to date has focused on the hypothalamic control of reproduction, recent evidence suggests that E2 may also influence diverse processes, such as the regulation of synaptic plasticity and neuroprotection (1, 2, 3, 4). Although potentially beneficial, E2 possesses distinct limitations, including an increased risk of breast and uterine cancers, which may decrease its clinical utility. Because of such negative side effects, there is growing interest in the development and potential therapeutic use of SERMS [selective estrogen receptor (ER) modulators]. SERMs are a class of steroidal or nonsteroidal drugs that may exert E2-like agonistic or antagonistic effects depending on the tissue type, the ER isoforms present in the cell, and the coactivators/corepressors recruited to the receptor-DNA complex (5). Tamoxifen (TMX), a first generation SERM prescribed to approximately 700,000 women, is employed as an ER antagonist in the treatment of E2-dependent breast cancer. However, TMX is also neuroprotective in animal models of ischemic stroke and methamphetamine toxicity (6, 7, 8), and promotes synaptic plasticity in the hippocampus (9) in a manner similar to E2, suggesting an agonistic function of TMX (and potentially other SERMs) in the brain (10, 11, 12). In contrast, antagonistic effects of TMX are also reported in some parts of the CNS (13, 14), implying the potential for cell type and regional specificity.
Despite the well documented role of E2 in neuroprotection and synaptic plasticity, the cellular and molecular mechanisms underlying these actions remain poorly understood. Physiological concentrations of E2 may directly protect cultured neurons (15, 16, 17, 18), although others have failed to confirm a direct protection with these concentrations (19, 20, 21, 22, 23, 24), suggesting that an alternative or parallel neuroprotective pathway may exist, possibly involving a nonneuronal cell type. This hypothesis is supported by the observation that physiological concentrations of E2 are neuroprotective in organotypic cortical explant cultures, which maintain an intact cellular and tissue architecture and contain multiple cell types (25, 26), as well as in astrocyte-neuronal cocultures (22, 27). Astrocytes, once assigned a solely supportive role in the CNS, are now regarded as active participants of brain function, including the regulation of synaptic plasticity and neuroprotection. Astrocytic ER expression is detected in the hippocampus in vivo, suggesting that astrocytes may be physiological targets of steroid hormone actions (28). The present study sought to study this possibility by examining whether E2/TMX influence the release of glial-derived neuroprotective growth factors using rat cortical astrocytes as a model. A second goal was to determine the possible mechanism(s) underlying such a regulation and establish the functional relevance of these regulatory effects.
Materials and Methods
Reagents
All cell culture reagents, sera, and media were obtained from Invitrogen (Carlsbad, CA). LY294002 and PD98059 were purchased from Promega Corp. (Madison, WI). U0126 and Akt inhibitor [IL-6-hydroxymethyl-chiro-inositol-2-(R)-2-O-methyl-3-O-octadecylcarbonate] were purchased from Calbiochem (San Diego, CA). E2, TMX, 4-hydroxytamoxifen, BSA-conjugated E2 (E2-BSA), and BSA were purchased from Sigma-Aldrich Corp. (St. Louis, MO). ICI182,780 was obtained from Tocris (Ballwin, MO). ER antibodies (rabbit polyclonal anti-ER-, H-184; goat polyclonal anti-ER-?, Y-19) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The ER antibodies are specific and do not cross-react with the other ER subtype. The pan-specific TGF-?, which immunoneutralizes all three mammalian isoforms of TGF-?, and anti-TGF-?1-neutralizing antibodies, which specifically neutralize only TGF-?1, were purchased from R&D Systems, Inc. (Minneapolis, MN).
Tissue and cell culture
C6 glial cells were cultured as previously described (29). Primary astrocyte cultures were obtained from the cerebral cortex of 2- to 3-d-old Holtzman rats (Harlan, Indianapolis, IN) by the method of McCarthy and deVellis (30), with minor modifications (22). Briefly, astrocytes were grown in a humidified cell culture incubator under an atmosphere of 5% CO2-95% O2 at 37 C for 10 d, at which point cultures were confluent. Complete culture medium was comprised of DMEM supplemented with 10% fetal bovine serum and antibiotics. Cell cultures were shaken overnight to remove contaminating oligodendrocytes, microglia, and neurons. Astrocytes were recovered using 0.1 M EDTA and used for experiments after the first passage. Cultures were routinely more than 95% pure astrocytes, as assessed by glial fibrillary acidic protein immunostaining. Purified neuronal cultures were established from the cortices of embryonic d 18 embryos and plated at 1 x 106 cells/ml in poly-D-lysine-coated, 24-well plates. Neurons were cultured in phenol red-free Neurobasal medium (Invitrogen) supplemented with B27, 0.5 mM L-glutamine, and antibiotics. Neurons were used for experiments between DIV 10–12. Mixed cortical cultures were prepared as described previously (22). Briefly, embryonic d 18 rat cerebral cortices were dissected and plated in 24-well plates at 1 x 106 cells/well using a plating medium of Neurobasal medium containing 10% FBS, 2% B27 supplement, 0.5 mM glutamine, and antibiotics. Cultures were kept at 37 C in a humidified 5% CO2 incubator. After 2 d in vitro, nonneuronal cell division was halted by exposure to the cytosine arabinoside (5 μM) in Neurobasal medium. This procedure resulted in cultures containing 15–20% glial cells. Cultures were used between days in vitro 10–12.
Treatments
Astrocytes were plated at 4 x 105 cells/ml in six-well plates and cultured in complete culture medium until reaching approximately 70% confluence. At this time, medium was removed, and cells were gently washed in PBS, then treated in Opti-MEM I Reduced Serum Medium (Invitrogen) (which contains low concentrations of phenol red) or Opti-MEM I containing combinations of 10 nM E2, 1 μM SERMs, or inhibitors at the doses detailed in the figure legends. Cells were cultured for 6–72 h, then supernatants were collected and stored at –80 C until growth factor assays. To verify an equal number of viable cells per well after treatments, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) reduction assays were performed. For all experiments, equal numbers of viable cells were present in all wells used for the assays.
TGF-? measurements
TGF-? levels in astrocyte-conditioned medium (ACM) were determined by enzyme immunoassay (EIA) using commercially available kits (Promega Corp.). Ninety-six-well plates were coated with monoclonal TGF-? antibodies, which bind soluble TGF-? from solution. Captured TGF-? was bound by a polyclonal antibody specific for TGF-? and, after washing, was detected using antirabbit IgG conjugated to horseradish peroxidase. Absorbance of samples was read at 450 nm using a plate reader (Multiskan MCC/340, Labsystems, Chicago, IL). Total TGF-? was measured in acid-treated 100-μl samples with a sensitivity of 32 pg/ml and are expressed as picograms per milliliter. Equal amounts of total protein and cell counts were verified after treatments to ensure that the increases in release were not due to a proliferative effect of the treatments. For each treatment group, n = 6. The data presented are the mean ± SEM of three independent cultures for verification of results.
Brain-derived neurotropic factor (BDNF) measurement
BDNF was measured in ACM using the BDNF Emax Immunoassay System, as recommended by the manufacturer (Promega Corp.). This system uses a horseradish peroxidase-conjugated secondary antibody and a single component 3,3',5,5'-tetramethylbenzidine substrate for the final chromogenic detection of bound BDNF. The assay has a range of 7.8–500 pg/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
Glial-derived neurotropic factor (GDNF) measurement
GDNF was quantitated in ACM using a GDNF Emax ImmunoAssay System, as recommended by the manufacturer (Promega Corp.). GDNF was detected with an antibody sandwich format, using a horseradish peroxidase-conjugated secondary antibody and a single component 3,3',5,5'-tetramethylbenzidine substrate for the final chromogenic detection of bound GDNF. GDNF was measured over a linear range between 15.6 and 1000 pg/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
Plasminogen activator inhibitor-1 (PAI-1) measurement
PAI-1 levels in ACM were quantified by an IMUBIND plasma EIA for human PAI-1 (American Diagnostics, Stamford, CT). This assay detects both active and inactive PAI-1 as well as PAI-1 coupled to tissue plasminogen activator and urokinase-type plasminogen activator. PAI-1 was measured in cell culture supernatants according to the manufacturer’s recommendations and expressed as nanograms per milliliter, with a lower limit of detection of 1 ng/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
RNA isolation and RT-PCR
Total RNA was isolated from cultured rat cortical astrocytes using RNeasy (Qiagen, Valencia, CA) according to the manufacturer’s recommendations. RT-PCR for ERs was performed as described previously (22).
Protein lysates and Akt phosphorylation
Phosphorylation of Akt on Ser473 (AktSer473) was determined using a commercially available kit (Biosource International, Camarillo, CA). Briefly, cells were grown to approximately 70% confluence in 10-cm tissue culture dishes and were serum-starved overnight. Cells were then treated with 10 nM E2, 1 μM TMX, or compounds after a 45-min pretreatment with 1 μM ICI182,780. After the treatments, medium was removed, cells were gently washed with Dulbecco’s PBS, then lysates were collected in complete RIPA buffer [1x PBS, 1% IGEPAL (Sigma), 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate supplemented with phenylmethylsuflonylfluoride, aprotinin, and sodium orthovanadate]. Lysates were quantified using a modified Lowry assay, and aliquots were frozen at –80 C until use. For the determination of phosphorylated AktSer473 and total Akt protein, 8 μg protein were used for each assay, which was performed as recommended by the manufacturer.
Quantitative RT-PCR
Real-time RT-PCR of TGF-?1 and TGF-?2 was performed on a Smart Cycler (Cepheid, Sunnyvale, CA) using the RNA Amplification SYBR Green I kit (Roche, Indianapolis, IN), according to the manufacturer’s protocol. Primer pairs were as follows: TGF-?1: forward primer, 5'-TGC TTC AGC TCC ACA GAG AA-3'; reverse primer, 5'-TGG TTG TAG AGG GCA AGG AC-3'); and TGF-?2: forward primer, 5'-CTC CAC ATA TGC CAG TGG TG-3'; reverse primer, 5'-CTA AAG CAA TAG GCG GCA TC-3'. Product specificity was confirmed by melting curve analysis and visualization of a single band of the appropriate product size on a 2% agarose gel. Expression levels were quantified by constructing a standard curve using cDNA dilutions, and gene levels were normalized to glyceraldehyde dehydrogenase (forward primer, 5'-ATG GGA AGC TGG TCA TCA AC-3'; reverse primer, 5'-GTG GTT CAC ACC CAT CAC AA-3'), which did not change after treatments, to control for differences in starting RNA. Data are expressed as the fold change compared from vehicle-treated cultures, using three per group. Experiments were performed in triplicate for verification of results.
Western blotting
Western blotting was performed as described previously by our laboratory (31). Astrocytes were plated in 10-cm cell cultures dishes and were grown to approximately 70% confluence. Cell lysates were collected in complete RIPA buffer and stored at –80 C until SDS-PAGE. After SDS-PAGE, proteins were transferred to polyvinylidene fluoride membranes (Millipore Corp., Bedford, MA) and probed with antibodies for ERs, as described previously (22). Proteins were visualized using enhanced chemiluminescence (ECL Plus, Amersham Biosciences, Piscataway, NJ). Blots were exposed, and molecular weight determinations were performed using a digital imaging system (Alpha Innotech, San Leandro, CA).
Transient transfections of cortical astrocytes
Astrocytes were plated at approximately 50% confluence overnight in complete culture medium without antibiotics. One microgram of dominant negative Akt (dnAkt), constitutively active Akt (myrAkt), or empty vector plasmids (Upstate Biotechnology, Inc., Lake Placid, NY) were complexed with Lipofectamine 2000 (Invitrogen), dissolved in Opti-MEM I Reduced Serum Medium, and added to cells overnight, as described previously (32) and as recommended by the manufacturers. After a 24-h incubation, cells were returned to original culture medium for an additional 12 h. Cells were then treated with vehicle or E2 and processed for TGF-? release or mRNA expression, as detailed in the figure legends.
Statistical analysis
For all experiments, six wells per treatment group was used. All studies were performed in three independent experiments for verification of results. Data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by the Student-Newman-Keuls post hoc test. P < 0.05 was considered significant. Different superscripts are used to denote significant differences in figures.
Results
Cortical astrocytes express ERs
To determine whether astrocytes express ERs, the presence of ER- and ER-? was investigated. Both ER- and ER-? were identified in cultured rat cortical astrocytes at the mRNA and protein levels, as determined by RT-PCR and Western blotting (Fig. 1).
FIG. 1. Rat cortical astrocytes express ERs. RT-PCR and Western blot analysis of ER- (A) and ER-? (B) expression in rat cortical astrocytes. Rat hypothalamus was included as a positive control for ER- and ER-?. Mk, 100-bp marker; Hyp, rat hypothalamus; RCA, rat cortical astrocytes.
Estrogen and SERMs increase the astrocytic release of TGF-? via a membrane-bound ER
Incubation of cultured rat cortical astrocytes with E2 or TMX induced the release of both TGF-?1 (Fig. 2, A–C) and TGF-?2 (Fig. 2, D–F) 6 h after treatment, an effect that persisted for 36 h. The stimulatory effect was eliminated by 48 and 72 h after treatment, suggesting a transient increase (data not shown). In contrast to that observed for E2, the stimulatory effect of TMX on TGF-?2 was lost after 36 h of treatment, implying the possibility of differential regulation between the two isoforms. A stimulatory effect on TGF-? release was also observed for 4-hydroxytamoxifen, the bioactive metabolite of TMX, in vivo. Additionally, both E2 and TMX enhanced the release of TGF-? from cultured hippocampal astrocytes, suggesting that this effect is not restricted to cerebral cortical astrocytes (data not shown). In contrast to the stimulatory effect of E2 and TMX on TGF-?, no such stimulatory effect was observed for BDNF or GDNF from rat cortical astrocytes (data not shown). However, after a 36-h exposure, both E2- and TMX-treated astrocytes maintained a higher release of a known TGF-?-regulated neuroprotective factor, PAI-1, compared with the vehicle control (Fig. 3).
FIG. 2. Temporal pattern of E2- and TMX-induced TGF-? release in rat cortical astrocytes. Rat cortical astrocytes were cultured in the presence of 0.5–10 nM E2, 1 μM TMX, or 1 μM 4-hydroxytamoxifen (4-HO-TMX) for 6 h (A and D), 18 h (B and E), or 36 h (C and F). After treatments, supernatants were collected and assayed for total TGF-?1 (A–C) or total TGF-?2 (D–F) content using a specific EIA. For all experiments, there were six wells per treatment group, and experiments were repeated in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
FIG. 3. Effects of E2 and TMX on PAI-1 release from rat cortical astrocytes. Treatment of cultured rat cortical astrocytes with E2 (1 and 10 nM) or TMX (1 μM) significantly increased the release of PAI-1 after a 36-h treatment. There were six wells per treatment group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
To determine whether the release of TGF-? was ER mediated, ICI182,780, a potent and specific ER antagonist, was used. ICI182,780 (1 μM), which equally inhibited both ER- and ER-?, completely blocked the ability of E2 and TMX to stimulate TGF-?1 and TGF-?2 release from astrocytes after an 18-h exposure (Fig. 4, A and B). Additional work demonstrated that the membrane-impermeable E2-BSA conjugate was equally effective at inducing TGF-?1 and TGF-?2 release in both rat cortical astrocytes (Fig. 4, C and D) and C6 glial cells (data not shown), effects that were prevented by pretreatment with ICI182,780. An equimolar treatment of cells with BSA failed to elicit a stimulatory effect, demonstrating the specificity of the E2-BSA conjugate effect (data not shown).
FIG. 4. Dependence of ER on the ability of E2 to induce TGF-? release in rat cortical astrocytes. Rat cortical astrocytes were treated with 10 nM E2 or 1 μM TMX in the presence or absence of ICI182,780 (1 μM). Cell supernatants were collected after an 18-h treatment and assayed for TGF-?1 release (A) or TGF-?2 release (B). In both panels, different superscripts denote significant differences between groups (P < 0.05). Treatment of rat cortical astrocytes with E2-BSA also significant stimulated both TGF-?1 (C) and TGF-?2 (D) release (P < 0.05) in an ICI182,780-dependent manner. For all experiments, there were six wells per treatment group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
E2 induction of TGF-? release is dependent on phosphatidylinositol 3-kinase (PI3K) activation
The role of nonclassical genomic signaling pathways in the E2-induced stimulation of TGF-? release was next investigated. Treatment of astrocytes with LY294002 or wortmannin, specific PI3K inhibitors, or Akt inhibitor, which directly prevents Akt activation, completely blocked the induction of TGF-?1 release after an 18-h treatment with either E2 or E2-BSA (Fig. 5A). Assays of cell death did not reveal a loss of cell viability due to LY294002 or Akt inhibitor I pretreatment (data not shown), suggesting that this effect is not due to passive growth factor release after cell death-induced membrane disruption. In contrast, the MAPK kinase inhibitors, PD98059 and U0126, were ineffective at inhibiting E2- or E2- and BSA-induced TGF-?1 release (Fig. 5B). To implicate Akt in the induction of TGF-? release by E2, astrocytes were transfected with a dnAkt, which contains a SerAla mutation at amino acid 473, rendering Akt incapable of activation and thereby preventing its actions downstream. Overexpression of dnAkt completely inhibited the ability of E2 or E2-BSA to induce TGF-? release and resulted in TGF-?1 levels equivalent to those in vehicle-treated controls (Fig. 5C). Conversely, overexpression of myrAkt increased TGF-?1 to levels significantly higher than those in E2- or E2- plus BSA-treated cultures. Empty vector transfection had no effect on TGF-?1 release after E2 or E2-BSA treatment, confirming the specificity of the dnAkt and myrAkt effects (Fig. 5C) and implicating the PI3K pathway in TGF-?1 release.
FIG. 5. Effect of PI3-K/Akt on E2-induced TGF-? release in rat cortical astrocytes. A, Rat cortical astrocytes were cotreated with E2 in the presence or absence or the PI3K inhibitors, LY294002 (20 μM) and wortmannin (200 nM), or Akt inhibitor prevented the induction of TGF-?1 release by E2 after an 18-h treatment. B, Cotreatment of rat cortical astrocytes with E2 in the presence of MAPK kinase inhibitors, PD98059 (30 μM) and U0126 (10 μM), was without effect on E2-induced TGF-?1 release. C, Overexpression of a dnAkt construct prevented E2-induced release of TGF-?1 in rat cortical astrocytes. Conversely, myrAkt significantly increased TGF-?1 release compared with that by empty vector-transfected astrocytes. For all studies, there were six wells per group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
Transient induction of Akt phosphorylation by E2 in cortical astrocytes
Given that the stimulation of TGF-? release by E2 was prevented by PI3K/Akt inhibition, the effect of E2 on Akt phosphorylation was studied. E2 and E2-BSA induced a rapid and transient phosphorylation of AktSer473 beginning 5 min after treatment and returning to baseline by 6 h after treatment, an effect that was prevented by ICI182,780, suggesting that a membrane-associated ER mediates the effect on Akt phosphorylation (Fig. 6, A and B). This effect was also blocked by pretreatment with LY294002, demonstrating a role for PI3K in the E2- and E2-BSA-mediated induction of Akt phosphorylation (Fig. 6, C and D).
FIG. 6. Effect of E2 on Akt phosphorylation in rat cortical astrocytes. A, Temporal pattern of phosphorylation of AktSer-473 in rat cortical astrocytes by E2 (10 nM) or E2-BSA (100 nM) in the presence or absence of the ER antagonist, ICI182,780 (1 μM). Cultures were serum-starved overnight before hormone exposure. Data represent the means of three independent trials. B, Temporal pattern of phosphorylation of AktSer-473 in rat cortical astrocytes by E2 or E2-BSA in the presence or absence of the PI3K inhibitor, LY294002 (20 μM). In all panels, data represent the ratio of phosphorylated AktSer473 to total Akt protein levels. Independent experiments were performed in triplicate for verification of results.
E2 up-regulates TGF-?1 gene expression via Akt
To determine whether the increased release of TGF-? was mediated via genomic regulation, the effects of E2 and TMX on TGF-?1 mRNA levels were investigated. Treatment of rat cortical astrocytes with E2 or TMX significantly increased TGF-?1 and TGF-?2 mRNA levels within 9 h, an effect that was inhibited by the ER antagonist, ICI182,780 (Fig. 7, A and B). As was observed with TGF-? release, the membrane-impermeant E2-BSA conjugate was also stimulatory with respect to TGF-?1 mRNA levels. Similarly, LY294002 and Akt inhibitor abrogated the E2-induced increase in TGF-?1 mRNA (Fig. 7C), as did overexpression of dnAkt, which completely eliminated the E2-induced increase in TGF-?1 mRNA (Fig. 7D).
FIG. 7. Effect of E2 on TGF-? mRNA expression rat cortical astrocytes. Rat cortical astrocytes were treated for 9 h with 10 nM E2, 1 μM TMX, or E2-BSA in the presence of absence of 1 μM ICI182,780. RNA was then extracted, and quantitative RT-PCR was performed for TGF-?1 (A) and TGF-?2 (B). Quantitation was performed using serial cDNA dilutions of the target genes. All samples were normalized to glyceraldehyde dehydrogenase to control for equal starting RNA. C, Inhibition of PI3-K/Akt using LY294002 (20 μM) or Akt inhibitor I attenuated the stimulation of TGF-?1 by 10 nM E2, 1 μM TMX, or E2-BSA after a 9-h treatment. D, Overexpression of a dnAkt construct inhibited the ability of 10 nM E2 or 100 nM E2-BSA to stimulate TGF-?1 mRNA after a 9-h exposure. For all studies, there were six wells per group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
Neuroprotection by E2 and TMX is mediated by TGF-?
To determine whether the E2-induced increase in TGF-? influenced neuroprotection, the effect of TGF-? immunoneutralization on E2-mediated neuroprotection was studied. Using a neuronal-glial coculture model, pretreatment with physiological levels of E2, therapeutic levels of TMX, or E2-BSA rescued cultures from camptothecin, a neuronal selective, apoptosis inducer (Fig. 8A). This effect required E2/TMX pretreatment, because cotreatment at the time of camptothecin exposure failed to prevent cell death (Fig. 8B). These effects did not occur directly at the level of the neuron, as neither E2, TMX, nor E2-BSA reversed camptothecin- induced cell death in purified cortical neuronal cultures (which contain >97% neurons and <2% astrocytes), suggesting that glial cells may be involved in the neuroprotective effect (Fig. 8C). The neuroprotective effects also were dependent, at least in part, on TGF-?1 release, because immunoneutralization of TGF-? using a pan-specific TGF-? antibody or a TGF-?1-specific immunoneutralization antibody, reversed the observed protection by E2/TMX/E2BSA (Fig. 8D). In contrast, basic fibroblast growth factor immunoneutralization had no effect on E2- or TMX-induced neuroprotection, demonstrating the specificity of the TGF-? effect (data not shown). The protective effect of E2/TMX was also ER and PI3K mediated, because both ICI182,780 and LY294002 inhibited neuroprotection, implicating the E2-induced activation of Akt and subsequent release of TGF-? in neuroprotection (Fig. 8E).
FIG. 8. Effect of astrocyte-derived TGF-? on E2- and TMX-mediated neuroprotection from camptothecin (CPT) in mixed cortical cultures. A, Effects of E2, TMX, and E2-BSA on cell death induced by CPT in mixed glial-neuronal cultures. Mixed cultures were pretreated for 24 h with 10 nM E2, 1 μM TMX, or 100 nM E2-BSA before 10 μM CPT treatment for 24 h before determination of cell viability. B, Treatment of glial-neuronal mixed cultures with E2, TMX, or E2-BSA at the time of CPT treatment does not affect cell death 24 h later. C, Pretreatment of purified cortical neurons with E2, TMX, or E2-BSA does not reverse the cell death induced by 24 h 10 μM CPT. D, Mixed cultures were pretreated with E2, TMX, or E2-BSA in the presence or absence of a pan-specific TGF-? isoform-neutralizing antibody (-pan-TGF?) or a TGF-?1-specific neutralizing antibody (-TGF-?1), followed by a 24-h CPT exposure. Cell viability was determined 24 h after treatment with CPT. E, Effects of the ER antagonist, ICI182,780, and the PI3-K inhibitor, LY294002, on E2- and TMX-mediated rescue. Mixed cultures were pretreated with E2, TMX, or E2-BSA in the presence or absence of 1 μM ICI182,780 or 20 μM LY294002 for 24 h. Cultures were then exposed to CPT for another 24 h, followed by determination of cell viability. For all studies, cellular viability was determined using the MTT assay. Vehicle-treated cultures were considered to be 100% viable, and all treatment groups were compared with these control cultures. Viability was also confirmed using lactate dehydrogenase release assays (data not shown). In all panels, data are expressed as the mean ± SEM, and groups with different superscripts are significantly different from each other (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test). For all studies, there were six wells per treatment group and experiments were performed in three independent sets of cultures for verification of results.
To determine whether E2 acts directly upon astrocytes or in concert with astrocyte-derived factors to elicit neuroprotection, astrocytes were exposed to E2, TMX, or E2-BSA for 12 h. Conditioned medium was then collected (termed E2-CM, TMX-CM, or E2BSA-CM, respectively) and tested for neuroprotection in rat cortical neurons compared with vehicle-treated ACM (Fig. 9A). Although ACM exerted a small neuroprotective effect, this effect was significantly enhanced by prior exposure of astrocytes to E2, TMX, or E2BSA. This effect was prevented by pretreatment of astrocytes with ICI182,780, delineating an important role for an ER in mediating the observed increase in neuroprotection. The protective effect of E2-CM, TMX-CM, or E2BSA-CM may be independent from residual steroids in the ACM, because the addition of ICI182,780 to the ACM after the 12-h treatments did not block the observed neuroprotection (Fig. 9B), although a neuroprotective contribution of an ICI182,780 insensitive receptor cannot be ruled out.
FIG. 9. Effect of ACM from E2-treated astrocytes on cell death. A, Rat purified neuronal cultures were treated for 24 h with ACM (12 h) that was previously incubated with vehicle (VEH-CM), 10 nM E2 (E2-CM), 1 μM TMX (TMX-CM), or 100 nM E2-BSA (E2BSA-CM). A second set of conditioned media was included in which astrocytes were incubated in the presence of estrogenic compounds after a 45-min pretreatment with 1 μM ICI182,780. Cultures were then exposed to 10 μM CPT for 24 h, followed by assessment of cell viability. B, Neuronal cultures were pretreated for 24 h with VEH-CM, E2-CM, TMX-CM, or E2BSA-CM in the presence or absence of 1 μM ICI182,780 to determine whether residual E2 in the E2-CM influences the observed neuroprotection. Cultures were then exposed to 10 μM CPT for 24 h, followed by assessment of cellular viability. For all studies, cellular viability was determined using the MTT assay. Vehicle-treated cultures were considered to be 100% viable, and all treatment groups were compared with these control cultures. Viability was also confirmed using lactate dehydrogenase release assays (data not shown). In all panels, data are expressed as the mean ± SEM and groups, with different superscripts indicating values that are significantly different from each other (P < 0.05, one-way ANOVA, Student-Newman-Keuls post hoc test). For all studies, there were six wells per treatment group, and experiments were performed in three independent sets of cultures for verification of results.
Discussion
E2 is an ovarian steroid hormone that is neuroprotective in animal models of neurological injury, including ischemic stroke and Alzheimer’s pathology (33, 34, 35, 36, 37, 38). A direct neuroprotective effect of E2 has been reported (15, 16, 17, 18), although a number of studies failed to replicate direct protection with physiological concentrations of E2 (19, 21, 22, 24, 27), suggesting that an alternative or parallel pathway of protection may exist under certain conditions. This possibility is supported by the observation that E2 was consistently protective when neurons were cultured in the presence of glial cells, either in dissociated neuronal cultures (22, 27) or in organotypic explant cultures (25, 26), suggesting that glia may mediate some of the actions of E2. A role for glia-derived TGF-? in E2 protection from ?-amyloid was recently suggested by Sortino and co-workers (27). Our study extends this E2-induced protection to the apoptotic agent, camptothecin, and reveals the underlying signaling cascades/mechanisms responsible for mediating E2 effects on TGF-? expression and release. Specifically, E2 induced TGF-?1 and TGF-?2 gene expression and release from cortical astrocytes in an ER-dependent manner. Furthermore, E2 rapidly stimulated activation of the PI3K-Akt pathway, an effect that was shown to be functionally relevant, because blockade of the PI3K-Akt pathway significantly attenuated the ability of E2 to enhance TGF-? expression/release from astrocytes. Additionally, our study showed the efficacy of E2-BSA and the SERM, tamoxifen, in regulation of TGF-? in cortical astrocytes, suggesting a potential cell membrane site of action for E2 and demonstrating that the E2 effect extends to clinically relevant SERMs, such as tamoxifen.
The neuroprotective effect of E2 in organotypic explants required an ICI182,780-sensitive ER and pretreatment, suggesting a possible role for genomic regulation (26). Cultured rat cortical astrocytes expressed both ER- and ER-? (39). However, in vivo, only ER-? has been demonstrated in hippocampal astrocytes under basal conditions (28). In contrast, ER- is specifically induced in astrocytes after brain injury in rats, nonhuman primates, and humans (40, 41, 42). Given the reported widespread expression of ER-?, but not ER-, in the normal adult cerebral cortex and hippocampus (43, 44, 45), ER-? probably mediates at least some of the physiological actions of E2 in the cerebral cortex. In contrast, astrocytic ER- may be more important in the steroidal response to brain injury.
Astrocytes may mediate at least some of the neuroprotective actions of E2 and TMX (22, 23, 27, 46); thus, understanding the estrogenic regulation of astrocyte-derived neuroprotective factors is important. TGF-? superfamily members, which represent potential mediators of these effects, influence neurosecretion, synaptic plasticity, and neuroprotection (27, 31, 47, 48, 49). Specifically, astrocyte-derived TGF-? is neuroprotective both in vitro (48) and in vivo (49, 50, 51, 52). In the present study, TGF-? release from cortical astrocytes was enhanced by physiological and therapeutic concentrations of E2 and TMX, respectively. Similarly, the release of PAI-1, another astrocyte-derived neuroprotective factor (53, 54, 55, 56), was maintained by E2. However, the role of PAI-1 in E2 neuroprotection is unclear, because its release was delayed compared with that of TGF-?, and the release occurred after a period when E2 first exerts its neuroprotection (e.g. E2 protection is observed as early as 24 h). Thus, additional work is needed to determine the importance of the elevation/maintenance of higher PAI-1 levels by E2. In contrast to the stimulatory effect of E2 and TMX on TGF-? isoforms and PAI-1 in cortical astrocytes, no such stimulation was observed for GDNF or BDNF (data not shown). BDNF, which is regulated by E2 in vivo (57, 58), is weakly expressed in astrocytes under basal conditions (59), suggesting that the E2-induced increase may occur in cell types other than astrocytes. The lack of E2 regulation of GDNF, another TGF-? superfamily member, also demonstrated the specificity for TGF-? isoforms.
The induction of TGF-? release from cortical astrocytes by E2/TMX was dependent upon an ICI182,780-sensitive ER. The ER isoform mediating this effect was not determined, although a stimulatory effect on TGF-? release was observed in C6 astroglial cells, which express ER-?, but not ER- (our unpublished observations). Similarly, E2-BSA also was stimulatory with respect to TGF-?, suggesting the possible involvement of a membrane-associated ER. However, it is important to point out that some E2 may have dissociated from the BSA, and this free E2 could have contributed to the observed stimulatory effect of the E2-BSA conjugate. Nonetheless, both ER- and ER-? are associated with the membrane fraction in astrocytes, with ER-? the predominant membrane-associated isoform (39). Additionally, ER-? mediated the rapid CNS actions of E2 in vivo (60). The involvement of ER-? in the activation of PI3K/Akt and TGF-? release was not investigated; however, both E2 and E2-BSA, at the identical concentrations used in the present study, rapidly mobilized calcium release in cortical astrocytes, an effect dependent upon phospholipase C (39), which has also been implicated in PI3K/Akt activation in astrocytoma cells (61). These possibilities, coupled with the documented neuroprotective activity of specific ER-? agonists in vivo (62), imply that astrocytic, membrane-bound ER-? may be involved in the observed neuroprotective effects of E2, although a role for ER- cannot be excluded, because this isoform is induced in astrocytes after brain injury and thus could mediate a steroidal response to injury.
E2 and TMX activate both classical and nonclassical genomic signaling pathways. Both compounds increased the gene expression of TGF-?, which lacks a consensus estrogen response element, supporting the idea that nonclassical genomic signaling may be involved in this regulation. Pathways activated by E2 include Akt and ERK, which are neuroprotective in organotypic cortical explant cultures and cultured cortical neurons (15, 16, 17, 18, 23, 25). These same signaling cascades regulate TGF-? in astrocytes after treatment with a metabotropic glutamate agonists, also demonstrating the potential for these pathways to mediate the E2 regulation of TGF-? (52). In the current studies, inhibition of PI3K/Akt, but not ERK, prevented TGF-? induction. Furthermore, Akt phosphorylation was induced in an ICI182,780-dependent manner by E2, confirming a role for this pathway in this effect. Given the importance of Akt activation in E2-induced neuroprotection and the potential involvement of glia in mediating these actions, these findings may provide a conceptual framework for a novel mechanism of action for E2 in CNS function.
Clinical trials using hormone replacement therapy in postmenopausal women have not replicated the potent neuroprotective effects of E2 from animal studies and, in fact, indicated that E2 may actually increase stroke risk (63, 64). Importantly, the natural steroid hormone, E2, was not used in these trials; rather, patients were administered Premarin (Wyeth, Madison, NJ), a formulation comprised of E2 in addition to dozens of equine-specific estrogens that are not produced by humans, compounds possessing androgenic and glucocorticoid activities, which exacerbate ischemic injury, as well as a large number of compounds with unknown biological activity. Thus, the relevance of this trial in understanding the physiological effects of E2 should be interpreted with caution. Nonetheless, these findings coupled with the known deleterious effects of unopposed estrogens have spurred the development of SERMs, which ideally possess the documented beneficial actions of E2 without the associated risk factors. Several SERMs are protective against ischemic stroke, including TMX and raloxifene analogs (6, 65), suggesting the possible utility of these compounds in the treatment of neurological disease. TMX and 4-hydroxytamoxifen, the primary metabolite of TMX in vivo, exhibited ER agonistic activity with respect to TGF-? release, suggesting that TMX may share a common cellular mechanism of neuroprotective actions with E2.
Gender differences in susceptibility to neurological damage have contributed greatly to the understanding of mechanisms of neuroprotection. It is widely appreciated that females possess a distinct neuroprotective advantage over males; however, the molecular and cellular bases underlying these differences remain largely unresolved. Animal studies indicated that E2 conveys this endogenous protection, an effect that is lost at menopause. Based on the present studies, we postulate that physiological E2 or clinically therapeutic concentrations of the SERM, TMX, protect the brain, at least in part, via astrocytic mediation. This potential parallel pathway may act in conjunction with the direct neuroprotective effects of E2 and help to explain how E2 globally protects the cerebral cortex, striatum, and hippocampus despite the scattered neuronal expression of ERs in these regions. Conversely, TGF-? receptors are ubiquitously expressed in the brain, suggesting that E2 may also protect ER–/TGF-? receptor+ neurons, via the astrocytic induction of TGF-?. Together, these findings describe a novel mechanism of neuroprotection by E2 and a SERM, effects that involve communication among multiple cell types.
References
McEwen B 2002 Estrogen actions throughout the brain. Recent Prog Horm Res 57:357–384
McEwen B, Akama K, Alves S, Brake WG, Bulloch K, Lee S, Li C, Yuen G, Milner TA 2001 Tracking the estrogen receptor in neurons: implications for estrogen-induced synapse formation. Proc Natl Acad Sci USA 98:7093–7100
Wise PM, Dubal DB, Wilson ME, Rau SW, Bottner M 2002 Minireview: neuroprotective effects of estrogen: new insights into mechanisms of action. Endocrinology 142:969–973
Murphy S, McCullough L, Littleton-Kearney M, Hurn P 2003 Estrogen and selective estrogen receptor modulators: neuroprotection in the Women’s Health Initiative era. Endocrine 21:17–26
Lonard DM, Smith CL 2002 Molecular perspectives on selective estrogen receptor modulators (SERMs): progress in understanding their tissue-specific agonist and antagonist actions. Steroids 67:15–24
Mehta S, Dhandapani KM, DeSevilla LM, Webb RC, Mahesh VB, Brann DW 2003 Tamoxifen, a selective estrogen receptor modulator, reduces ischemic damage caused by middle cerebral artery occlusion in the ovariectomized female rat. Neuroendocrinology 77:44–50
Mickley KR, Dluzen DE 2004 Dose-response effects of estrogen and tamoxifen upon methamphetamine-induced behavior responses and neurotoxicity of the nigrostriatal dopaminergic system in female mice. Neuroendocrinology 79:305–316
Kimelberg HK, Feustel PJ, Jin Y, Paquette J, Boulos A, Keller Jr RW, Tranmer BI 2000 Acute treatment with tamoxifen reduces ischemic damage following middle cerebral artery occlusion. Neuroreport 11:2675–2679
Silva I, Mello LE, Freymuller E, Haidar MA, Baracat EC 2000 Estrogen, progestogen, and tamoxifen increase synaptic density of the hippocampus of ovariectomized rats. Neurosci Lett 291:183–186
McMillan PL, LeMaster AM, Dorsa DM 2002 Tamoxifen enhances choline acetyltransferase mRNA expression in rat basal forebrain cholinergic neurons. Brain Res Mol Brain Res 103:140–145
Cyr M, Thibault C, Morissette M, Landry M, Di Paolo T 2001 Estrogen-like activity of tamoxifen and raloxifene on NMDA receptor binding and expression of its subunits in rat brain. Neuropsychopharmacology 25:242–257
Cyr M, Morissette M, Landry M, Di Paolo T 2001 Estrogenic activity of tamoxifen and raloxifene on rat brain AMPA receptors. Neuroreport 12:535–539
Wilson ME, Mook D, Graves F, Felger J, Bielsky IF, Wallen K 2003 Tamoxifen is an estrogen antagonist on gonadotropin secretion and responsiveness of the hypothalamic-pituitary-adrenal axis in female monkeys. Endocrine 22:305–315
Sanchez-Criado JE, Guelmes P, Bellido C, Gonzalez M, Hernandez G, Aguilar R, Garrido-Gracia JC, Bellow AR, Alonso R 2002 Tamoxifen but not other selective estrogen receptor modulators antagonizes estrogen actions on luteinizing hormone secretion while inducing gonadotropin-releasing hormone self-priming in the rat. Neuroendocrinology 76:203–213
Honda K, Shimohama S, Sawada H, Kihara T, Nakamizo T, Shibasaki H, Akaike A 2001 Nongenomic antiapoptotic signal transduction by estrogen in cultured cortical neurons. J Neurosci Res 64:466–475
Honda K, Sawada H, Kihara T, Urushitani M, Nakamizo T, Akaike A, Shimohama S 2000 Phosphatidylinositol-3-kinase mediates neuroprotection by estrogen in cultured cortical neurons. J Neurosci Res 60:321–327
Singer CA, Figueroa-Masot XA, Batchelor RH, Dorsa DM 1999 The mitogen-activated protein kinase pathway mediates estrogen neuroprotection after glutamate toxicity in primary cortical neurons. J Neurosci 19:2455–2463
Singer CA, Rogers KL, Strickland TM, Dorsa DM 1996 Estrogen protects primary cortical neurons from glutamate toxicity. Neurosci Lett 212:13–16
Bae YH, Hwang JY, Kim YH, Koh JY 2000 Anti-oxidant neuroprotection by estrogens in mouse cortical cultures. J Korean Med Sci 15:327–336
Moosmann B, Behl C 1999 The antioxidant neuroprotective effects of estrogens and phenolic compounds are independent from their estrogenic properties. Proc Natl Acad Sci USA 96:8867–8872
Regan RF, Guo Y 1997 Estrogens attenuate neuronal injury due to hemoglobin, chemical hypoxia, and excitatory amino acids in murine cortical cultures. Brain Res 764:133–140
Dhandapani KM, Brann DW 2003 Neuroprotective effects of estrogen and tamoxifen in vitro: a facilitative role for glia? Endocrine 21:59–66
Dhandapani KM, Brann DW 2002 Protective effects of estrogen and SERMs in the brain. Biol Reprod 67:1379–1385
Behl C, Skutella T, Lezoualc’h F, Post A, Widmann M, Newton CJ, Holsboer F 1997 Neuroprotection against oxidative stress by estrogens: structure-activity relationship. Mol Pharmacol 51:535–541
Wilson ME, Liu Y, Wise PM 2002 Estradiol enhances Akt activation in cortical explant cultures following neuronal injury. Brain Res Mol Brain Res 102:48–54
Wilson ME, Dubal DB, Wise PM 2000 Estradiol protects against injury-induced cell death in cortical explant cultures: a role for estrogen receptors. Brain Res 873:235–242
Sortino MA, Chisari M, Merlo S, Vancheri C, Caruso M, Nicoletti F, Canonico PL, Copani A 2004 Glia mediate the neuroprotective action of estradiol on ?-amyloid-induced neuronal death. Endocrinology 145:5080–5086
Azcoitia I, Sierra A, Garcia-Segura LM 1999 Localization of estrogen receptor ?-immunoreactivity in astrocytes of the adult rat brain. Glia 26:260–267
Dhandapani KM, Wade FM, Mahesh VB, Brann DW 2002 Basic fibroblast growth factor induces TGF-? release in an isoform and glioma-specific manner. Neuroreport 13:239–241
McCarthy KD, de Vellis J 1979 The regulation of adenosine 3':5'-cyclic monophosphate accumulation in glia by -adrenergic agonists. Life Sci 24:639–649
Buchanan CD, Mahesh VB, Brann DW 2000 Estrogen-astrocyte-luteinizing hormone-releasing hormone signaling: a role for transforming growth factor-?1. Biol Reprod 62:1710–1721
Dalby B, Cateo S, Harris A, Ohki EC, Tilkins ML, Price PJ, Ciccarone VC 2004 Advanced transfection with Lipofectamine 2000 reagent: primary neurons, siRNA, and high-throughput applications. Methods 33:95–103
Simpkins JW, Rajakumar G, Zhang YQ, Simpkins CE, Grenwald D, Yu CJ, Bodor N, Day AL 1997 Estrogens may reduce mortality and ischemic damage caused by middle cerebral artery occlusion in the female rat. J Neurosurg 87:724–730
Rusa R, Alkayed NJ, Crain BJ, Traystman RJ, Kimes AS, London ED, Klaus JA, Hurn PD 1999 17?-Estradiol reduces stroke injury in estrogen-deficient female animals. Stroke 30:1665–1670
Dubal DB, Kashon ML, Pettigrew LC, Ren JM, Finklestein SP, Rau SW, Wise PM 1998 Estradiol protects against ischemic injury. J Cereb Blood Flow Metab 18:1253–1258
Alkayed NJ, Murphy SJ, Traystman RJ, Hurn PD, Miller VM 2000 Neuroprotective effects of female gonadal steroids in reproductively senescent female rats. Stroke 31:161–168
Dubal DB, Wise PM 2001 Neuroprotective effects of estradiol in middle-aged female rats. Endocrinology 142:43–48
Cholerton B, Gleason CE, Baker LD, Asthana S 2002 Estrogen and Alzheimer’s disease: the story so far. Drugs Aging 19:405–427
Chaban VV, Lakhter AJ, Micevych P 2004 A membrane estrogen receptor mediates intracellular calcium release in astrocytes. Endocrinology 145:3788–3795
Lu YP, Zeng M, Hu XY, Xu H, Swaab DF, Ravid R, Zhou JN 2003 Estrogen receptor -immunoreactive astrocytes are increased in the hippocampus in Alzheimer’s disease. Exp Neurol 183:482–488
Garcia-Ovejero D, Veiga S, Garcia-Segura LM, Doncarlos LL 2002 Glial expression of estrogen and androgen receptors after rat brain injury. J Comp Neurol 450:256–271
Blurton-Jones M, Tuszynski MH 2001 Reactive astrocytes express estrogen receptors in the injured primate brain. J Comp Neurol 433:115–123
Zhang JQ, Cai WQ, Zhou de S, Su BY 2002 Distribution and differences of estrogen receptor ? immunoreactivity in the brain of adult male and female rats. Brain Res 935:73–80
Deitch HR, Mershon JL, Clark KE 2001 Estrogen receptor ? is the predominant isoform expressed in the brain of adult and fetal sheep. Am J Obstet Gynecol 184:1077–1079
Shughrue PJ, Lane MV, Merchenthaler I 1997 Comparative distribution of estrogen receptor- and -? mRNA in the rat central nervous system. J Comp Neurol 388:507–525
Dhandapani KM, Brann DW 2002 Estrogen-astrocyte interactions: implications for neuroprotection. BMC Neurosci 3:6
Zwain IH, Arroyo A, Amato P, Yen SS 2002 A role for hypothalamic astrocytes in dehydroepiandrosterone and estradiol regulation of gonadotropin-releasing hormone (GnRH) release by GnRH neurons. Neuroendocrinology 75:375–383
Dhandapani KM, Hadman M, De Sevilla LM, Wade MF, Mahesh VB, Brann DW 2003 Astrocyte protection of neurons: role of TGF-? signaling via a c-Jun-AP-1 protective pathway. J Biol Chem 278:43329–43339
Dhandapani KM, Brann DW 2003 TGF-?: a neuroprotective factor in cerebral ischemia. Cell Biochem Biophys 39:13–22
Ruocco A, Nicole O, Docagne F, Ali C, Chazalviel L, Komesli S, Yablonsky F, Roussel S, MacKenzie ET, Vivien D, Buisson A 1999 A transforming growth factor-? antagonist unmasks the neuroprotective role of this endogenous cytokine in excitotoxic and ischemic brain injury. J Cereb Blood Flow Metab 19:1345–1353
Bruno V, Battaglia G, Casabona G, Copani A, Caciagli F, Nicoletti F 1998 Neuroprotection by glial metabotropic glutamate receptors is mediated by transforming factor-?. J Neurosci 18:9594–9600
D’Onofrio M, Cuomo L, Battaglia G, Ngomba RT, Storto M, Kingston AE, Orzi F, De Blasi A, Di Iorio P, Nicoletti F, Bruno V 2001 Neuroprotection mediated by glial group-II metabotropic glutamate receptors requires the activation of the MAP kinase and the phosphatidylinositol-3-kinase pathways. J Neurochem 78:435–445
Docagne F, Nicole O, Gabriel C, Fernandez-Monreal M, Lesne S, Ali C, Plawinski L, Carmeliet P, MacKenzie ET, Buisson A, Vivien D 2002 Smad3-dependent induction of plasminogen activator inhibitor-1 in astrocytes mediates neuroprotective activity of transforming growth factor-?1 against NMDA-induced necrosis. Mol Cell Neurosci 21:634–644
Gabriel C, Ali C, Lesne S, Fernandez-Monreal M, Docagne F, Plawinski L, MacKenzie ET, Buisson A, Vivien D 2003 Transforming growth factor - induced expression of type 1 plasminogen activator inhibitor in astrocytes rescues neurons from excitotoxicity. FASEB J 17:277–279
Treichel JA, Reddington M, Kreutzberg GW 1998 Regulation of plasminogen activator inhibitor-1 mRNA accumulation by basic fibroblast growth factor and transforming growth factor-?1 in cultured rat astrocytes. J Neurochem 71:1944–1952
Lesne S, Blanchet S, Docagne F, Liot G, Plawinski L, MacKenzie ET, Auffray C, Buisson A, Pietu G, Vivien D 2002 Transforming growth factor-?1-modulated cerebral gene expression. J Cereb Blood Flow Metab 22:1114–1123
Sohrabji F, Miranda RC, Toran-Allerand CD 1995 Identification of a putative estrogen response element in the gene encoding brain-derived neurotrophic factor. Proc Natl Acad Sci USA 92:11110–11114
Solum DJ, Handa RJ 2002 Estrogen regulates the development of brain-derived neurotrophic factor mRNA and protein in the rat hippocampus. J Neurosci 22:2650–2659
Zafra F, Lindholm D, Castren E, Hartikka J, Thoenen H 1992 Regulation of brain-derived neurotrophic factor and nerve growth factor mRNA in primary cultures of hippocampal neurons and astrocytes. J Neurosci 12:4793–4799[Abstract]
Abraham IM, Han SK, Todman MG, Korach KS, Herbison AE 2003 Estrogen receptor ? mediates rapid estrogen actions on gonadotropin releasing hormone neurons in vivo. J Neurosci 23:5771–5777
Tang X, Batty IH, Downes CP 2002 Muscarinic receptors mediate phospholipase C-dependent activation of protein kinase B via Ca2+, ErbB3, and phosphoinositide 3-kinase in 1321N1 astrocytoma cells. J Biol Chem 277:338–344
Carswell HW, Macrae IN, Gallagher L, Harrop E, Horsburgh KJ 2004 Neuroprotection by a selective oestrogen receptor b agonist in a mouse model of global ischemia. Am J Physiol 287:H1501–H1504
Wassertheil-Smoller S, Hendriz SL, Limacher M, Heiss G, Kooperberg C, Baird A, Kotchen T, Curb JD, Black H, Rossouw JE, Aragaki A, Safford M, Stein E, Laowattana S, Mysiw WJ, WHI Investigators 2003 Effect of estrogen plus progestin on stroke in postmenopausal women: the Women’s Health Initiative: a randomized trial. J Am Med Assoc 289:2673–2684
Anderson GL, Limacher M, Assaf AR, Bassford T, Beresfold SA, Black H, Bonds D, Brunner R, Brzyski R, Caan B, Chlebowski R, Curb D, Gass M, Hays J, Heiss G, Hendrix S, Howard BV, Hsia J, Hubbell A, Jackson R, Johnson KC, Judd H, Kotchen JM, Kuller L, LaCroix AZ, Lane D, Langer RD, Lasser N, Lewis CE, Manson J, Margolis K, Ockene J, O’Sullivan MJ, Phillips L, Prentice RL, Ritenbaugh C, Robbins J, Rossouw JE, Sarto G, Stefanick ML, Van Horn L, Wactawski-Wende J, Wallace R, Wassertheil-Smoller S, Women Health Initiative Steering Committee 2004 Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women’s Health Initiative randomized controlled trial. J Am Med Assoc 291:1701–1712
Rossberg MI, Murphy SJ, Traystman RJ, Hurn PD 2000 LY353381.HCl, a selective estrogen receptor modulator, and experimental stroke. Stroke 31:3041–3046(Krishnan M. Dhandapani, F)
Address all correspondence and requests for reprints to: Dr. Darrell W. Brann, Institute of Molecular Medicine and Genetics, Program in Developmental Neurobiology, 1120 15th Street, Medical College of Georgia, Augusta, Georgia 30912. E-mail: dbrann@mail.mcg.edu.
Abstract
17?-Estradiol (E2) and selective estrogen receptor modulators (SERMs), such as tamoxifen, mediate numerous effects in the brain, including neurosecretion, neuroprotection, and the induction of synaptic plasticity. Astrocytes, the most abundant cell type in the brain, influence many of these same functions and thus may represent a mediator of estrogen action. The present study examined the regulatory effect and underlying cell signaling mechanisms of E2-induced release of neurotropic growth factors from primary rat cortical astrocyte cultures. The results revealed that E2 (0.5, 1, and 10 nM) and tamoxifen (1 μM) increased both the expression and release of the neuroprotective cytokines, TGF-?1 and TGF-?2 (TGF-?), from cortical astrocytes. The stimulatory effect of E2 was attenuated by the estrogen receptor (ER) antagonist, ICI182,780, suggesting ER dependency. The effect of E2 also appeared to involve mediation by the phosphotidylinositol 3-kinase (PI3K)/Akt signaling pathway, because E2 rapidly induced Akt phosphorylation, and pharmacological or molecular inhibition of the PI3K/Akt pathway prevented E2-induced release of TGF-?. Additionally, the membrane-impermeant conjugate, E2-BSA, stimulated the release of TGF-?, suggesting the potential involvement of a membrane-bound ER. Finally, E2, tamoxifen, and E2-BSA were shown to protect neuronal-astrocyte cocultures from camptothecin-induced neuronal cell death, effects that were attenuated by ICI182,780, Akt inhibition, or TGF-? immunoneutralization. As a whole, these studies suggest that E2 induction of TGF-? release from cortical astrocytes could provide a mechanism of neuroprotection, and that E2 stimulation of TGF-? expression and release from astrocytes occurs via an ER-dependent mechanism involving mediation by the PI3K/Akt signaling pathway.
Introduction
THE OVARIAN STEROID hormone, 17?-estradiol (E2), exerts a wide array of actions in the central nervous system (CNS). Although most work to date has focused on the hypothalamic control of reproduction, recent evidence suggests that E2 may also influence diverse processes, such as the regulation of synaptic plasticity and neuroprotection (1, 2, 3, 4). Although potentially beneficial, E2 possesses distinct limitations, including an increased risk of breast and uterine cancers, which may decrease its clinical utility. Because of such negative side effects, there is growing interest in the development and potential therapeutic use of SERMS [selective estrogen receptor (ER) modulators]. SERMs are a class of steroidal or nonsteroidal drugs that may exert E2-like agonistic or antagonistic effects depending on the tissue type, the ER isoforms present in the cell, and the coactivators/corepressors recruited to the receptor-DNA complex (5). Tamoxifen (TMX), a first generation SERM prescribed to approximately 700,000 women, is employed as an ER antagonist in the treatment of E2-dependent breast cancer. However, TMX is also neuroprotective in animal models of ischemic stroke and methamphetamine toxicity (6, 7, 8), and promotes synaptic plasticity in the hippocampus (9) in a manner similar to E2, suggesting an agonistic function of TMX (and potentially other SERMs) in the brain (10, 11, 12). In contrast, antagonistic effects of TMX are also reported in some parts of the CNS (13, 14), implying the potential for cell type and regional specificity.
Despite the well documented role of E2 in neuroprotection and synaptic plasticity, the cellular and molecular mechanisms underlying these actions remain poorly understood. Physiological concentrations of E2 may directly protect cultured neurons (15, 16, 17, 18), although others have failed to confirm a direct protection with these concentrations (19, 20, 21, 22, 23, 24), suggesting that an alternative or parallel neuroprotective pathway may exist, possibly involving a nonneuronal cell type. This hypothesis is supported by the observation that physiological concentrations of E2 are neuroprotective in organotypic cortical explant cultures, which maintain an intact cellular and tissue architecture and contain multiple cell types (25, 26), as well as in astrocyte-neuronal cocultures (22, 27). Astrocytes, once assigned a solely supportive role in the CNS, are now regarded as active participants of brain function, including the regulation of synaptic plasticity and neuroprotection. Astrocytic ER expression is detected in the hippocampus in vivo, suggesting that astrocytes may be physiological targets of steroid hormone actions (28). The present study sought to study this possibility by examining whether E2/TMX influence the release of glial-derived neuroprotective growth factors using rat cortical astrocytes as a model. A second goal was to determine the possible mechanism(s) underlying such a regulation and establish the functional relevance of these regulatory effects.
Materials and Methods
Reagents
All cell culture reagents, sera, and media were obtained from Invitrogen (Carlsbad, CA). LY294002 and PD98059 were purchased from Promega Corp. (Madison, WI). U0126 and Akt inhibitor [IL-6-hydroxymethyl-chiro-inositol-2-(R)-2-O-methyl-3-O-octadecylcarbonate] were purchased from Calbiochem (San Diego, CA). E2, TMX, 4-hydroxytamoxifen, BSA-conjugated E2 (E2-BSA), and BSA were purchased from Sigma-Aldrich Corp. (St. Louis, MO). ICI182,780 was obtained from Tocris (Ballwin, MO). ER antibodies (rabbit polyclonal anti-ER-, H-184; goat polyclonal anti-ER-?, Y-19) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The ER antibodies are specific and do not cross-react with the other ER subtype. The pan-specific TGF-?, which immunoneutralizes all three mammalian isoforms of TGF-?, and anti-TGF-?1-neutralizing antibodies, which specifically neutralize only TGF-?1, were purchased from R&D Systems, Inc. (Minneapolis, MN).
Tissue and cell culture
C6 glial cells were cultured as previously described (29). Primary astrocyte cultures were obtained from the cerebral cortex of 2- to 3-d-old Holtzman rats (Harlan, Indianapolis, IN) by the method of McCarthy and deVellis (30), with minor modifications (22). Briefly, astrocytes were grown in a humidified cell culture incubator under an atmosphere of 5% CO2-95% O2 at 37 C for 10 d, at which point cultures were confluent. Complete culture medium was comprised of DMEM supplemented with 10% fetal bovine serum and antibiotics. Cell cultures were shaken overnight to remove contaminating oligodendrocytes, microglia, and neurons. Astrocytes were recovered using 0.1 M EDTA and used for experiments after the first passage. Cultures were routinely more than 95% pure astrocytes, as assessed by glial fibrillary acidic protein immunostaining. Purified neuronal cultures were established from the cortices of embryonic d 18 embryos and plated at 1 x 106 cells/ml in poly-D-lysine-coated, 24-well plates. Neurons were cultured in phenol red-free Neurobasal medium (Invitrogen) supplemented with B27, 0.5 mM L-glutamine, and antibiotics. Neurons were used for experiments between DIV 10–12. Mixed cortical cultures were prepared as described previously (22). Briefly, embryonic d 18 rat cerebral cortices were dissected and plated in 24-well plates at 1 x 106 cells/well using a plating medium of Neurobasal medium containing 10% FBS, 2% B27 supplement, 0.5 mM glutamine, and antibiotics. Cultures were kept at 37 C in a humidified 5% CO2 incubator. After 2 d in vitro, nonneuronal cell division was halted by exposure to the cytosine arabinoside (5 μM) in Neurobasal medium. This procedure resulted in cultures containing 15–20% glial cells. Cultures were used between days in vitro 10–12.
Treatments
Astrocytes were plated at 4 x 105 cells/ml in six-well plates and cultured in complete culture medium until reaching approximately 70% confluence. At this time, medium was removed, and cells were gently washed in PBS, then treated in Opti-MEM I Reduced Serum Medium (Invitrogen) (which contains low concentrations of phenol red) or Opti-MEM I containing combinations of 10 nM E2, 1 μM SERMs, or inhibitors at the doses detailed in the figure legends. Cells were cultured for 6–72 h, then supernatants were collected and stored at –80 C until growth factor assays. To verify an equal number of viable cells per well after treatments, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) reduction assays were performed. For all experiments, equal numbers of viable cells were present in all wells used for the assays.
TGF-? measurements
TGF-? levels in astrocyte-conditioned medium (ACM) were determined by enzyme immunoassay (EIA) using commercially available kits (Promega Corp.). Ninety-six-well plates were coated with monoclonal TGF-? antibodies, which bind soluble TGF-? from solution. Captured TGF-? was bound by a polyclonal antibody specific for TGF-? and, after washing, was detected using antirabbit IgG conjugated to horseradish peroxidase. Absorbance of samples was read at 450 nm using a plate reader (Multiskan MCC/340, Labsystems, Chicago, IL). Total TGF-? was measured in acid-treated 100-μl samples with a sensitivity of 32 pg/ml and are expressed as picograms per milliliter. Equal amounts of total protein and cell counts were verified after treatments to ensure that the increases in release were not due to a proliferative effect of the treatments. For each treatment group, n = 6. The data presented are the mean ± SEM of three independent cultures for verification of results.
Brain-derived neurotropic factor (BDNF) measurement
BDNF was measured in ACM using the BDNF Emax Immunoassay System, as recommended by the manufacturer (Promega Corp.). This system uses a horseradish peroxidase-conjugated secondary antibody and a single component 3,3',5,5'-tetramethylbenzidine substrate for the final chromogenic detection of bound BDNF. The assay has a range of 7.8–500 pg/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
Glial-derived neurotropic factor (GDNF) measurement
GDNF was quantitated in ACM using a GDNF Emax ImmunoAssay System, as recommended by the manufacturer (Promega Corp.). GDNF was detected with an antibody sandwich format, using a horseradish peroxidase-conjugated secondary antibody and a single component 3,3',5,5'-tetramethylbenzidine substrate for the final chromogenic detection of bound GDNF. GDNF was measured over a linear range between 15.6 and 1000 pg/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
Plasminogen activator inhibitor-1 (PAI-1) measurement
PAI-1 levels in ACM were quantified by an IMUBIND plasma EIA for human PAI-1 (American Diagnostics, Stamford, CT). This assay detects both active and inactive PAI-1 as well as PAI-1 coupled to tissue plasminogen activator and urokinase-type plasminogen activator. PAI-1 was measured in cell culture supernatants according to the manufacturer’s recommendations and expressed as nanograms per milliliter, with a lower limit of detection of 1 ng/ml. For each treatment group, n = 6. The data presented are the mean ± SEM from three independent cultures for verification of results.
RNA isolation and RT-PCR
Total RNA was isolated from cultured rat cortical astrocytes using RNeasy (Qiagen, Valencia, CA) according to the manufacturer’s recommendations. RT-PCR for ERs was performed as described previously (22).
Protein lysates and Akt phosphorylation
Phosphorylation of Akt on Ser473 (AktSer473) was determined using a commercially available kit (Biosource International, Camarillo, CA). Briefly, cells were grown to approximately 70% confluence in 10-cm tissue culture dishes and were serum-starved overnight. Cells were then treated with 10 nM E2, 1 μM TMX, or compounds after a 45-min pretreatment with 1 μM ICI182,780. After the treatments, medium was removed, cells were gently washed with Dulbecco’s PBS, then lysates were collected in complete RIPA buffer [1x PBS, 1% IGEPAL (Sigma), 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate supplemented with phenylmethylsuflonylfluoride, aprotinin, and sodium orthovanadate]. Lysates were quantified using a modified Lowry assay, and aliquots were frozen at –80 C until use. For the determination of phosphorylated AktSer473 and total Akt protein, 8 μg protein were used for each assay, which was performed as recommended by the manufacturer.
Quantitative RT-PCR
Real-time RT-PCR of TGF-?1 and TGF-?2 was performed on a Smart Cycler (Cepheid, Sunnyvale, CA) using the RNA Amplification SYBR Green I kit (Roche, Indianapolis, IN), according to the manufacturer’s protocol. Primer pairs were as follows: TGF-?1: forward primer, 5'-TGC TTC AGC TCC ACA GAG AA-3'; reverse primer, 5'-TGG TTG TAG AGG GCA AGG AC-3'); and TGF-?2: forward primer, 5'-CTC CAC ATA TGC CAG TGG TG-3'; reverse primer, 5'-CTA AAG CAA TAG GCG GCA TC-3'. Product specificity was confirmed by melting curve analysis and visualization of a single band of the appropriate product size on a 2% agarose gel. Expression levels were quantified by constructing a standard curve using cDNA dilutions, and gene levels were normalized to glyceraldehyde dehydrogenase (forward primer, 5'-ATG GGA AGC TGG TCA TCA AC-3'; reverse primer, 5'-GTG GTT CAC ACC CAT CAC AA-3'), which did not change after treatments, to control for differences in starting RNA. Data are expressed as the fold change compared from vehicle-treated cultures, using three per group. Experiments were performed in triplicate for verification of results.
Western blotting
Western blotting was performed as described previously by our laboratory (31). Astrocytes were plated in 10-cm cell cultures dishes and were grown to approximately 70% confluence. Cell lysates were collected in complete RIPA buffer and stored at –80 C until SDS-PAGE. After SDS-PAGE, proteins were transferred to polyvinylidene fluoride membranes (Millipore Corp., Bedford, MA) and probed with antibodies for ERs, as described previously (22). Proteins were visualized using enhanced chemiluminescence (ECL Plus, Amersham Biosciences, Piscataway, NJ). Blots were exposed, and molecular weight determinations were performed using a digital imaging system (Alpha Innotech, San Leandro, CA).
Transient transfections of cortical astrocytes
Astrocytes were plated at approximately 50% confluence overnight in complete culture medium without antibiotics. One microgram of dominant negative Akt (dnAkt), constitutively active Akt (myrAkt), or empty vector plasmids (Upstate Biotechnology, Inc., Lake Placid, NY) were complexed with Lipofectamine 2000 (Invitrogen), dissolved in Opti-MEM I Reduced Serum Medium, and added to cells overnight, as described previously (32) and as recommended by the manufacturers. After a 24-h incubation, cells were returned to original culture medium for an additional 12 h. Cells were then treated with vehicle or E2 and processed for TGF-? release or mRNA expression, as detailed in the figure legends.
Statistical analysis
For all experiments, six wells per treatment group was used. All studies were performed in three independent experiments for verification of results. Data are presented as the mean ± SEM. Data were analyzed using one-way ANOVA, followed by the Student-Newman-Keuls post hoc test. P < 0.05 was considered significant. Different superscripts are used to denote significant differences in figures.
Results
Cortical astrocytes express ERs
To determine whether astrocytes express ERs, the presence of ER- and ER-? was investigated. Both ER- and ER-? were identified in cultured rat cortical astrocytes at the mRNA and protein levels, as determined by RT-PCR and Western blotting (Fig. 1).
FIG. 1. Rat cortical astrocytes express ERs. RT-PCR and Western blot analysis of ER- (A) and ER-? (B) expression in rat cortical astrocytes. Rat hypothalamus was included as a positive control for ER- and ER-?. Mk, 100-bp marker; Hyp, rat hypothalamus; RCA, rat cortical astrocytes.
Estrogen and SERMs increase the astrocytic release of TGF-? via a membrane-bound ER
Incubation of cultured rat cortical astrocytes with E2 or TMX induced the release of both TGF-?1 (Fig. 2, A–C) and TGF-?2 (Fig. 2, D–F) 6 h after treatment, an effect that persisted for 36 h. The stimulatory effect was eliminated by 48 and 72 h after treatment, suggesting a transient increase (data not shown). In contrast to that observed for E2, the stimulatory effect of TMX on TGF-?2 was lost after 36 h of treatment, implying the possibility of differential regulation between the two isoforms. A stimulatory effect on TGF-? release was also observed for 4-hydroxytamoxifen, the bioactive metabolite of TMX, in vivo. Additionally, both E2 and TMX enhanced the release of TGF-? from cultured hippocampal astrocytes, suggesting that this effect is not restricted to cerebral cortical astrocytes (data not shown). In contrast to the stimulatory effect of E2 and TMX on TGF-?, no such stimulatory effect was observed for BDNF or GDNF from rat cortical astrocytes (data not shown). However, after a 36-h exposure, both E2- and TMX-treated astrocytes maintained a higher release of a known TGF-?-regulated neuroprotective factor, PAI-1, compared with the vehicle control (Fig. 3).
FIG. 2. Temporal pattern of E2- and TMX-induced TGF-? release in rat cortical astrocytes. Rat cortical astrocytes were cultured in the presence of 0.5–10 nM E2, 1 μM TMX, or 1 μM 4-hydroxytamoxifen (4-HO-TMX) for 6 h (A and D), 18 h (B and E), or 36 h (C and F). After treatments, supernatants were collected and assayed for total TGF-?1 (A–C) or total TGF-?2 (D–F) content using a specific EIA. For all experiments, there were six wells per treatment group, and experiments were repeated in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
FIG. 3. Effects of E2 and TMX on PAI-1 release from rat cortical astrocytes. Treatment of cultured rat cortical astrocytes with E2 (1 and 10 nM) or TMX (1 μM) significantly increased the release of PAI-1 after a 36-h treatment. There were six wells per treatment group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
To determine whether the release of TGF-? was ER mediated, ICI182,780, a potent and specific ER antagonist, was used. ICI182,780 (1 μM), which equally inhibited both ER- and ER-?, completely blocked the ability of E2 and TMX to stimulate TGF-?1 and TGF-?2 release from astrocytes after an 18-h exposure (Fig. 4, A and B). Additional work demonstrated that the membrane-impermeable E2-BSA conjugate was equally effective at inducing TGF-?1 and TGF-?2 release in both rat cortical astrocytes (Fig. 4, C and D) and C6 glial cells (data not shown), effects that were prevented by pretreatment with ICI182,780. An equimolar treatment of cells with BSA failed to elicit a stimulatory effect, demonstrating the specificity of the E2-BSA conjugate effect (data not shown).
FIG. 4. Dependence of ER on the ability of E2 to induce TGF-? release in rat cortical astrocytes. Rat cortical astrocytes were treated with 10 nM E2 or 1 μM TMX in the presence or absence of ICI182,780 (1 μM). Cell supernatants were collected after an 18-h treatment and assayed for TGF-?1 release (A) or TGF-?2 release (B). In both panels, different superscripts denote significant differences between groups (P < 0.05). Treatment of rat cortical astrocytes with E2-BSA also significant stimulated both TGF-?1 (C) and TGF-?2 (D) release (P < 0.05) in an ICI182,780-dependent manner. For all experiments, there were six wells per treatment group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
E2 induction of TGF-? release is dependent on phosphatidylinositol 3-kinase (PI3K) activation
The role of nonclassical genomic signaling pathways in the E2-induced stimulation of TGF-? release was next investigated. Treatment of astrocytes with LY294002 or wortmannin, specific PI3K inhibitors, or Akt inhibitor, which directly prevents Akt activation, completely blocked the induction of TGF-?1 release after an 18-h treatment with either E2 or E2-BSA (Fig. 5A). Assays of cell death did not reveal a loss of cell viability due to LY294002 or Akt inhibitor I pretreatment (data not shown), suggesting that this effect is not due to passive growth factor release after cell death-induced membrane disruption. In contrast, the MAPK kinase inhibitors, PD98059 and U0126, were ineffective at inhibiting E2- or E2- and BSA-induced TGF-?1 release (Fig. 5B). To implicate Akt in the induction of TGF-? release by E2, astrocytes were transfected with a dnAkt, which contains a SerAla mutation at amino acid 473, rendering Akt incapable of activation and thereby preventing its actions downstream. Overexpression of dnAkt completely inhibited the ability of E2 or E2-BSA to induce TGF-? release and resulted in TGF-?1 levels equivalent to those in vehicle-treated controls (Fig. 5C). Conversely, overexpression of myrAkt increased TGF-?1 to levels significantly higher than those in E2- or E2- plus BSA-treated cultures. Empty vector transfection had no effect on TGF-?1 release after E2 or E2-BSA treatment, confirming the specificity of the dnAkt and myrAkt effects (Fig. 5C) and implicating the PI3K pathway in TGF-?1 release.
FIG. 5. Effect of PI3-K/Akt on E2-induced TGF-? release in rat cortical astrocytes. A, Rat cortical astrocytes were cotreated with E2 in the presence or absence or the PI3K inhibitors, LY294002 (20 μM) and wortmannin (200 nM), or Akt inhibitor prevented the induction of TGF-?1 release by E2 after an 18-h treatment. B, Cotreatment of rat cortical astrocytes with E2 in the presence of MAPK kinase inhibitors, PD98059 (30 μM) and U0126 (10 μM), was without effect on E2-induced TGF-?1 release. C, Overexpression of a dnAkt construct prevented E2-induced release of TGF-?1 in rat cortical astrocytes. Conversely, myrAkt significantly increased TGF-?1 release compared with that by empty vector-transfected astrocytes. For all studies, there were six wells per group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
Transient induction of Akt phosphorylation by E2 in cortical astrocytes
Given that the stimulation of TGF-? release by E2 was prevented by PI3K/Akt inhibition, the effect of E2 on Akt phosphorylation was studied. E2 and E2-BSA induced a rapid and transient phosphorylation of AktSer473 beginning 5 min after treatment and returning to baseline by 6 h after treatment, an effect that was prevented by ICI182,780, suggesting that a membrane-associated ER mediates the effect on Akt phosphorylation (Fig. 6, A and B). This effect was also blocked by pretreatment with LY294002, demonstrating a role for PI3K in the E2- and E2-BSA-mediated induction of Akt phosphorylation (Fig. 6, C and D).
FIG. 6. Effect of E2 on Akt phosphorylation in rat cortical astrocytes. A, Temporal pattern of phosphorylation of AktSer-473 in rat cortical astrocytes by E2 (10 nM) or E2-BSA (100 nM) in the presence or absence of the ER antagonist, ICI182,780 (1 μM). Cultures were serum-starved overnight before hormone exposure. Data represent the means of three independent trials. B, Temporal pattern of phosphorylation of AktSer-473 in rat cortical astrocytes by E2 or E2-BSA in the presence or absence of the PI3K inhibitor, LY294002 (20 μM). In all panels, data represent the ratio of phosphorylated AktSer473 to total Akt protein levels. Independent experiments were performed in triplicate for verification of results.
E2 up-regulates TGF-?1 gene expression via Akt
To determine whether the increased release of TGF-? was mediated via genomic regulation, the effects of E2 and TMX on TGF-?1 mRNA levels were investigated. Treatment of rat cortical astrocytes with E2 or TMX significantly increased TGF-?1 and TGF-?2 mRNA levels within 9 h, an effect that was inhibited by the ER antagonist, ICI182,780 (Fig. 7, A and B). As was observed with TGF-? release, the membrane-impermeant E2-BSA conjugate was also stimulatory with respect to TGF-?1 mRNA levels. Similarly, LY294002 and Akt inhibitor abrogated the E2-induced increase in TGF-?1 mRNA (Fig. 7C), as did overexpression of dnAkt, which completely eliminated the E2-induced increase in TGF-?1 mRNA (Fig. 7D).
FIG. 7. Effect of E2 on TGF-? mRNA expression rat cortical astrocytes. Rat cortical astrocytes were treated for 9 h with 10 nM E2, 1 μM TMX, or E2-BSA in the presence of absence of 1 μM ICI182,780. RNA was then extracted, and quantitative RT-PCR was performed for TGF-?1 (A) and TGF-?2 (B). Quantitation was performed using serial cDNA dilutions of the target genes. All samples were normalized to glyceraldehyde dehydrogenase to control for equal starting RNA. C, Inhibition of PI3-K/Akt using LY294002 (20 μM) or Akt inhibitor I attenuated the stimulation of TGF-?1 by 10 nM E2, 1 μM TMX, or E2-BSA after a 9-h treatment. D, Overexpression of a dnAkt construct inhibited the ability of 10 nM E2 or 100 nM E2-BSA to stimulate TGF-?1 mRNA after a 9-h exposure. For all studies, there were six wells per group, and experiments were performed in three independent cultures. Different superscripts denote significant differences between groups (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test).
Neuroprotection by E2 and TMX is mediated by TGF-?
To determine whether the E2-induced increase in TGF-? influenced neuroprotection, the effect of TGF-? immunoneutralization on E2-mediated neuroprotection was studied. Using a neuronal-glial coculture model, pretreatment with physiological levels of E2, therapeutic levels of TMX, or E2-BSA rescued cultures from camptothecin, a neuronal selective, apoptosis inducer (Fig. 8A). This effect required E2/TMX pretreatment, because cotreatment at the time of camptothecin exposure failed to prevent cell death (Fig. 8B). These effects did not occur directly at the level of the neuron, as neither E2, TMX, nor E2-BSA reversed camptothecin- induced cell death in purified cortical neuronal cultures (which contain >97% neurons and <2% astrocytes), suggesting that glial cells may be involved in the neuroprotective effect (Fig. 8C). The neuroprotective effects also were dependent, at least in part, on TGF-?1 release, because immunoneutralization of TGF-? using a pan-specific TGF-? antibody or a TGF-?1-specific immunoneutralization antibody, reversed the observed protection by E2/TMX/E2BSA (Fig. 8D). In contrast, basic fibroblast growth factor immunoneutralization had no effect on E2- or TMX-induced neuroprotection, demonstrating the specificity of the TGF-? effect (data not shown). The protective effect of E2/TMX was also ER and PI3K mediated, because both ICI182,780 and LY294002 inhibited neuroprotection, implicating the E2-induced activation of Akt and subsequent release of TGF-? in neuroprotection (Fig. 8E).
FIG. 8. Effect of astrocyte-derived TGF-? on E2- and TMX-mediated neuroprotection from camptothecin (CPT) in mixed cortical cultures. A, Effects of E2, TMX, and E2-BSA on cell death induced by CPT in mixed glial-neuronal cultures. Mixed cultures were pretreated for 24 h with 10 nM E2, 1 μM TMX, or 100 nM E2-BSA before 10 μM CPT treatment for 24 h before determination of cell viability. B, Treatment of glial-neuronal mixed cultures with E2, TMX, or E2-BSA at the time of CPT treatment does not affect cell death 24 h later. C, Pretreatment of purified cortical neurons with E2, TMX, or E2-BSA does not reverse the cell death induced by 24 h 10 μM CPT. D, Mixed cultures were pretreated with E2, TMX, or E2-BSA in the presence or absence of a pan-specific TGF-? isoform-neutralizing antibody (-pan-TGF?) or a TGF-?1-specific neutralizing antibody (-TGF-?1), followed by a 24-h CPT exposure. Cell viability was determined 24 h after treatment with CPT. E, Effects of the ER antagonist, ICI182,780, and the PI3-K inhibitor, LY294002, on E2- and TMX-mediated rescue. Mixed cultures were pretreated with E2, TMX, or E2-BSA in the presence or absence of 1 μM ICI182,780 or 20 μM LY294002 for 24 h. Cultures were then exposed to CPT for another 24 h, followed by determination of cell viability. For all studies, cellular viability was determined using the MTT assay. Vehicle-treated cultures were considered to be 100% viable, and all treatment groups were compared with these control cultures. Viability was also confirmed using lactate dehydrogenase release assays (data not shown). In all panels, data are expressed as the mean ± SEM, and groups with different superscripts are significantly different from each other (P < 0.05, by one-way ANOVA, Student-Newman-Keuls post hoc test). For all studies, there were six wells per treatment group and experiments were performed in three independent sets of cultures for verification of results.
To determine whether E2 acts directly upon astrocytes or in concert with astrocyte-derived factors to elicit neuroprotection, astrocytes were exposed to E2, TMX, or E2-BSA for 12 h. Conditioned medium was then collected (termed E2-CM, TMX-CM, or E2BSA-CM, respectively) and tested for neuroprotection in rat cortical neurons compared with vehicle-treated ACM (Fig. 9A). Although ACM exerted a small neuroprotective effect, this effect was significantly enhanced by prior exposure of astrocytes to E2, TMX, or E2BSA. This effect was prevented by pretreatment of astrocytes with ICI182,780, delineating an important role for an ER in mediating the observed increase in neuroprotection. The protective effect of E2-CM, TMX-CM, or E2BSA-CM may be independent from residual steroids in the ACM, because the addition of ICI182,780 to the ACM after the 12-h treatments did not block the observed neuroprotection (Fig. 9B), although a neuroprotective contribution of an ICI182,780 insensitive receptor cannot be ruled out.
FIG. 9. Effect of ACM from E2-treated astrocytes on cell death. A, Rat purified neuronal cultures were treated for 24 h with ACM (12 h) that was previously incubated with vehicle (VEH-CM), 10 nM E2 (E2-CM), 1 μM TMX (TMX-CM), or 100 nM E2-BSA (E2BSA-CM). A second set of conditioned media was included in which astrocytes were incubated in the presence of estrogenic compounds after a 45-min pretreatment with 1 μM ICI182,780. Cultures were then exposed to 10 μM CPT for 24 h, followed by assessment of cell viability. B, Neuronal cultures were pretreated for 24 h with VEH-CM, E2-CM, TMX-CM, or E2BSA-CM in the presence or absence of 1 μM ICI182,780 to determine whether residual E2 in the E2-CM influences the observed neuroprotection. Cultures were then exposed to 10 μM CPT for 24 h, followed by assessment of cellular viability. For all studies, cellular viability was determined using the MTT assay. Vehicle-treated cultures were considered to be 100% viable, and all treatment groups were compared with these control cultures. Viability was also confirmed using lactate dehydrogenase release assays (data not shown). In all panels, data are expressed as the mean ± SEM and groups, with different superscripts indicating values that are significantly different from each other (P < 0.05, one-way ANOVA, Student-Newman-Keuls post hoc test). For all studies, there were six wells per treatment group, and experiments were performed in three independent sets of cultures for verification of results.
Discussion
E2 is an ovarian steroid hormone that is neuroprotective in animal models of neurological injury, including ischemic stroke and Alzheimer’s pathology (33, 34, 35, 36, 37, 38). A direct neuroprotective effect of E2 has been reported (15, 16, 17, 18), although a number of studies failed to replicate direct protection with physiological concentrations of E2 (19, 21, 22, 24, 27), suggesting that an alternative or parallel pathway of protection may exist under certain conditions. This possibility is supported by the observation that E2 was consistently protective when neurons were cultured in the presence of glial cells, either in dissociated neuronal cultures (22, 27) or in organotypic explant cultures (25, 26), suggesting that glia may mediate some of the actions of E2. A role for glia-derived TGF-? in E2 protection from ?-amyloid was recently suggested by Sortino and co-workers (27). Our study extends this E2-induced protection to the apoptotic agent, camptothecin, and reveals the underlying signaling cascades/mechanisms responsible for mediating E2 effects on TGF-? expression and release. Specifically, E2 induced TGF-?1 and TGF-?2 gene expression and release from cortical astrocytes in an ER-dependent manner. Furthermore, E2 rapidly stimulated activation of the PI3K-Akt pathway, an effect that was shown to be functionally relevant, because blockade of the PI3K-Akt pathway significantly attenuated the ability of E2 to enhance TGF-? expression/release from astrocytes. Additionally, our study showed the efficacy of E2-BSA and the SERM, tamoxifen, in regulation of TGF-? in cortical astrocytes, suggesting a potential cell membrane site of action for E2 and demonstrating that the E2 effect extends to clinically relevant SERMs, such as tamoxifen.
The neuroprotective effect of E2 in organotypic explants required an ICI182,780-sensitive ER and pretreatment, suggesting a possible role for genomic regulation (26). Cultured rat cortical astrocytes expressed both ER- and ER-? (39). However, in vivo, only ER-? has been demonstrated in hippocampal astrocytes under basal conditions (28). In contrast, ER- is specifically induced in astrocytes after brain injury in rats, nonhuman primates, and humans (40, 41, 42). Given the reported widespread expression of ER-?, but not ER-, in the normal adult cerebral cortex and hippocampus (43, 44, 45), ER-? probably mediates at least some of the physiological actions of E2 in the cerebral cortex. In contrast, astrocytic ER- may be more important in the steroidal response to brain injury.
Astrocytes may mediate at least some of the neuroprotective actions of E2 and TMX (22, 23, 27, 46); thus, understanding the estrogenic regulation of astrocyte-derived neuroprotective factors is important. TGF-? superfamily members, which represent potential mediators of these effects, influence neurosecretion, synaptic plasticity, and neuroprotection (27, 31, 47, 48, 49). Specifically, astrocyte-derived TGF-? is neuroprotective both in vitro (48) and in vivo (49, 50, 51, 52). In the present study, TGF-? release from cortical astrocytes was enhanced by physiological and therapeutic concentrations of E2 and TMX, respectively. Similarly, the release of PAI-1, another astrocyte-derived neuroprotective factor (53, 54, 55, 56), was maintained by E2. However, the role of PAI-1 in E2 neuroprotection is unclear, because its release was delayed compared with that of TGF-?, and the release occurred after a period when E2 first exerts its neuroprotection (e.g. E2 protection is observed as early as 24 h). Thus, additional work is needed to determine the importance of the elevation/maintenance of higher PAI-1 levels by E2. In contrast to the stimulatory effect of E2 and TMX on TGF-? isoforms and PAI-1 in cortical astrocytes, no such stimulation was observed for GDNF or BDNF (data not shown). BDNF, which is regulated by E2 in vivo (57, 58), is weakly expressed in astrocytes under basal conditions (59), suggesting that the E2-induced increase may occur in cell types other than astrocytes. The lack of E2 regulation of GDNF, another TGF-? superfamily member, also demonstrated the specificity for TGF-? isoforms.
The induction of TGF-? release from cortical astrocytes by E2/TMX was dependent upon an ICI182,780-sensitive ER. The ER isoform mediating this effect was not determined, although a stimulatory effect on TGF-? release was observed in C6 astroglial cells, which express ER-?, but not ER- (our unpublished observations). Similarly, E2-BSA also was stimulatory with respect to TGF-?, suggesting the possible involvement of a membrane-associated ER. However, it is important to point out that some E2 may have dissociated from the BSA, and this free E2 could have contributed to the observed stimulatory effect of the E2-BSA conjugate. Nonetheless, both ER- and ER-? are associated with the membrane fraction in astrocytes, with ER-? the predominant membrane-associated isoform (39). Additionally, ER-? mediated the rapid CNS actions of E2 in vivo (60). The involvement of ER-? in the activation of PI3K/Akt and TGF-? release was not investigated; however, both E2 and E2-BSA, at the identical concentrations used in the present study, rapidly mobilized calcium release in cortical astrocytes, an effect dependent upon phospholipase C (39), which has also been implicated in PI3K/Akt activation in astrocytoma cells (61). These possibilities, coupled with the documented neuroprotective activity of specific ER-? agonists in vivo (62), imply that astrocytic, membrane-bound ER-? may be involved in the observed neuroprotective effects of E2, although a role for ER- cannot be excluded, because this isoform is induced in astrocytes after brain injury and thus could mediate a steroidal response to injury.
E2 and TMX activate both classical and nonclassical genomic signaling pathways. Both compounds increased the gene expression of TGF-?, which lacks a consensus estrogen response element, supporting the idea that nonclassical genomic signaling may be involved in this regulation. Pathways activated by E2 include Akt and ERK, which are neuroprotective in organotypic cortical explant cultures and cultured cortical neurons (15, 16, 17, 18, 23, 25). These same signaling cascades regulate TGF-? in astrocytes after treatment with a metabotropic glutamate agonists, also demonstrating the potential for these pathways to mediate the E2 regulation of TGF-? (52). In the current studies, inhibition of PI3K/Akt, but not ERK, prevented TGF-? induction. Furthermore, Akt phosphorylation was induced in an ICI182,780-dependent manner by E2, confirming a role for this pathway in this effect. Given the importance of Akt activation in E2-induced neuroprotection and the potential involvement of glia in mediating these actions, these findings may provide a conceptual framework for a novel mechanism of action for E2 in CNS function.
Clinical trials using hormone replacement therapy in postmenopausal women have not replicated the potent neuroprotective effects of E2 from animal studies and, in fact, indicated that E2 may actually increase stroke risk (63, 64). Importantly, the natural steroid hormone, E2, was not used in these trials; rather, patients were administered Premarin (Wyeth, Madison, NJ), a formulation comprised of E2 in addition to dozens of equine-specific estrogens that are not produced by humans, compounds possessing androgenic and glucocorticoid activities, which exacerbate ischemic injury, as well as a large number of compounds with unknown biological activity. Thus, the relevance of this trial in understanding the physiological effects of E2 should be interpreted with caution. Nonetheless, these findings coupled with the known deleterious effects of unopposed estrogens have spurred the development of SERMs, which ideally possess the documented beneficial actions of E2 without the associated risk factors. Several SERMs are protective against ischemic stroke, including TMX and raloxifene analogs (6, 65), suggesting the possible utility of these compounds in the treatment of neurological disease. TMX and 4-hydroxytamoxifen, the primary metabolite of TMX in vivo, exhibited ER agonistic activity with respect to TGF-? release, suggesting that TMX may share a common cellular mechanism of neuroprotective actions with E2.
Gender differences in susceptibility to neurological damage have contributed greatly to the understanding of mechanisms of neuroprotection. It is widely appreciated that females possess a distinct neuroprotective advantage over males; however, the molecular and cellular bases underlying these differences remain largely unresolved. Animal studies indicated that E2 conveys this endogenous protection, an effect that is lost at menopause. Based on the present studies, we postulate that physiological E2 or clinically therapeutic concentrations of the SERM, TMX, protect the brain, at least in part, via astrocytic mediation. This potential parallel pathway may act in conjunction with the direct neuroprotective effects of E2 and help to explain how E2 globally protects the cerebral cortex, striatum, and hippocampus despite the scattered neuronal expression of ERs in these regions. Conversely, TGF-? receptors are ubiquitously expressed in the brain, suggesting that E2 may also protect ER–/TGF-? receptor+ neurons, via the astrocytic induction of TGF-?. Together, these findings describe a novel mechanism of neuroprotection by E2 and a SERM, effects that involve communication among multiple cell types.
References
McEwen B 2002 Estrogen actions throughout the brain. Recent Prog Horm Res 57:357–384
McEwen B, Akama K, Alves S, Brake WG, Bulloch K, Lee S, Li C, Yuen G, Milner TA 2001 Tracking the estrogen receptor in neurons: implications for estrogen-induced synapse formation. Proc Natl Acad Sci USA 98:7093–7100
Wise PM, Dubal DB, Wilson ME, Rau SW, Bottner M 2002 Minireview: neuroprotective effects of estrogen: new insights into mechanisms of action. Endocrinology 142:969–973
Murphy S, McCullough L, Littleton-Kearney M, Hurn P 2003 Estrogen and selective estrogen receptor modulators: neuroprotection in the Women’s Health Initiative era. Endocrine 21:17–26
Lonard DM, Smith CL 2002 Molecular perspectives on selective estrogen receptor modulators (SERMs): progress in understanding their tissue-specific agonist and antagonist actions. Steroids 67:15–24
Mehta S, Dhandapani KM, DeSevilla LM, Webb RC, Mahesh VB, Brann DW 2003 Tamoxifen, a selective estrogen receptor modulator, reduces ischemic damage caused by middle cerebral artery occlusion in the ovariectomized female rat. Neuroendocrinology 77:44–50
Mickley KR, Dluzen DE 2004 Dose-response effects of estrogen and tamoxifen upon methamphetamine-induced behavior responses and neurotoxicity of the nigrostriatal dopaminergic system in female mice. Neuroendocrinology 79:305–316
Kimelberg HK, Feustel PJ, Jin Y, Paquette J, Boulos A, Keller Jr RW, Tranmer BI 2000 Acute treatment with tamoxifen reduces ischemic damage following middle cerebral artery occlusion. Neuroreport 11:2675–2679
Silva I, Mello LE, Freymuller E, Haidar MA, Baracat EC 2000 Estrogen, progestogen, and tamoxifen increase synaptic density of the hippocampus of ovariectomized rats. Neurosci Lett 291:183–186
McMillan PL, LeMaster AM, Dorsa DM 2002 Tamoxifen enhances choline acetyltransferase mRNA expression in rat basal forebrain cholinergic neurons. Brain Res Mol Brain Res 103:140–145
Cyr M, Thibault C, Morissette M, Landry M, Di Paolo T 2001 Estrogen-like activity of tamoxifen and raloxifene on NMDA receptor binding and expression of its subunits in rat brain. Neuropsychopharmacology 25:242–257
Cyr M, Morissette M, Landry M, Di Paolo T 2001 Estrogenic activity of tamoxifen and raloxifene on rat brain AMPA receptors. Neuroreport 12:535–539
Wilson ME, Mook D, Graves F, Felger J, Bielsky IF, Wallen K 2003 Tamoxifen is an estrogen antagonist on gonadotropin secretion and responsiveness of the hypothalamic-pituitary-adrenal axis in female monkeys. Endocrine 22:305–315
Sanchez-Criado JE, Guelmes P, Bellido C, Gonzalez M, Hernandez G, Aguilar R, Garrido-Gracia JC, Bellow AR, Alonso R 2002 Tamoxifen but not other selective estrogen receptor modulators antagonizes estrogen actions on luteinizing hormone secretion while inducing gonadotropin-releasing hormone self-priming in the rat. Neuroendocrinology 76:203–213
Honda K, Shimohama S, Sawada H, Kihara T, Nakamizo T, Shibasaki H, Akaike A 2001 Nongenomic antiapoptotic signal transduction by estrogen in cultured cortical neurons. J Neurosci Res 64:466–475
Honda K, Sawada H, Kihara T, Urushitani M, Nakamizo T, Akaike A, Shimohama S 2000 Phosphatidylinositol-3-kinase mediates neuroprotection by estrogen in cultured cortical neurons. J Neurosci Res 60:321–327
Singer CA, Figueroa-Masot XA, Batchelor RH, Dorsa DM 1999 The mitogen-activated protein kinase pathway mediates estrogen neuroprotection after glutamate toxicity in primary cortical neurons. J Neurosci 19:2455–2463
Singer CA, Rogers KL, Strickland TM, Dorsa DM 1996 Estrogen protects primary cortical neurons from glutamate toxicity. Neurosci Lett 212:13–16
Bae YH, Hwang JY, Kim YH, Koh JY 2000 Anti-oxidant neuroprotection by estrogens in mouse cortical cultures. J Korean Med Sci 15:327–336
Moosmann B, Behl C 1999 The antioxidant neuroprotective effects of estrogens and phenolic compounds are independent from their estrogenic properties. Proc Natl Acad Sci USA 96:8867–8872
Regan RF, Guo Y 1997 Estrogens attenuate neuronal injury due to hemoglobin, chemical hypoxia, and excitatory amino acids in murine cortical cultures. Brain Res 764:133–140
Dhandapani KM, Brann DW 2003 Neuroprotective effects of estrogen and tamoxifen in vitro: a facilitative role for glia? Endocrine 21:59–66
Dhandapani KM, Brann DW 2002 Protective effects of estrogen and SERMs in the brain. Biol Reprod 67:1379–1385
Behl C, Skutella T, Lezoualc’h F, Post A, Widmann M, Newton CJ, Holsboer F 1997 Neuroprotection against oxidative stress by estrogens: structure-activity relationship. Mol Pharmacol 51:535–541
Wilson ME, Liu Y, Wise PM 2002 Estradiol enhances Akt activation in cortical explant cultures following neuronal injury. Brain Res Mol Brain Res 102:48–54
Wilson ME, Dubal DB, Wise PM 2000 Estradiol protects against injury-induced cell death in cortical explant cultures: a role for estrogen receptors. Brain Res 873:235–242
Sortino MA, Chisari M, Merlo S, Vancheri C, Caruso M, Nicoletti F, Canonico PL, Copani A 2004 Glia mediate the neuroprotective action of estradiol on ?-amyloid-induced neuronal death. Endocrinology 145:5080–5086
Azcoitia I, Sierra A, Garcia-Segura LM 1999 Localization of estrogen receptor ?-immunoreactivity in astrocytes of the adult rat brain. Glia 26:260–267
Dhandapani KM, Wade FM, Mahesh VB, Brann DW 2002 Basic fibroblast growth factor induces TGF-? release in an isoform and glioma-specific manner. Neuroreport 13:239–241
McCarthy KD, de Vellis J 1979 The regulation of adenosine 3':5'-cyclic monophosphate accumulation in glia by -adrenergic agonists. Life Sci 24:639–649
Buchanan CD, Mahesh VB, Brann DW 2000 Estrogen-astrocyte-luteinizing hormone-releasing hormone signaling: a role for transforming growth factor-?1. Biol Reprod 62:1710–1721
Dalby B, Cateo S, Harris A, Ohki EC, Tilkins ML, Price PJ, Ciccarone VC 2004 Advanced transfection with Lipofectamine 2000 reagent: primary neurons, siRNA, and high-throughput applications. Methods 33:95–103
Simpkins JW, Rajakumar G, Zhang YQ, Simpkins CE, Grenwald D, Yu CJ, Bodor N, Day AL 1997 Estrogens may reduce mortality and ischemic damage caused by middle cerebral artery occlusion in the female rat. J Neurosurg 87:724–730
Rusa R, Alkayed NJ, Crain BJ, Traystman RJ, Kimes AS, London ED, Klaus JA, Hurn PD 1999 17?-Estradiol reduces stroke injury in estrogen-deficient female animals. Stroke 30:1665–1670
Dubal DB, Kashon ML, Pettigrew LC, Ren JM, Finklestein SP, Rau SW, Wise PM 1998 Estradiol protects against ischemic injury. J Cereb Blood Flow Metab 18:1253–1258
Alkayed NJ, Murphy SJ, Traystman RJ, Hurn PD, Miller VM 2000 Neuroprotective effects of female gonadal steroids in reproductively senescent female rats. Stroke 31:161–168
Dubal DB, Wise PM 2001 Neuroprotective effects of estradiol in middle-aged female rats. Endocrinology 142:43–48
Cholerton B, Gleason CE, Baker LD, Asthana S 2002 Estrogen and Alzheimer’s disease: the story so far. Drugs Aging 19:405–427
Chaban VV, Lakhter AJ, Micevych P 2004 A membrane estrogen receptor mediates intracellular calcium release in astrocytes. Endocrinology 145:3788–3795
Lu YP, Zeng M, Hu XY, Xu H, Swaab DF, Ravid R, Zhou JN 2003 Estrogen receptor -immunoreactive astrocytes are increased in the hippocampus in Alzheimer’s disease. Exp Neurol 183:482–488
Garcia-Ovejero D, Veiga S, Garcia-Segura LM, Doncarlos LL 2002 Glial expression of estrogen and androgen receptors after rat brain injury. J Comp Neurol 450:256–271
Blurton-Jones M, Tuszynski MH 2001 Reactive astrocytes express estrogen receptors in the injured primate brain. J Comp Neurol 433:115–123
Zhang JQ, Cai WQ, Zhou de S, Su BY 2002 Distribution and differences of estrogen receptor ? immunoreactivity in the brain of adult male and female rats. Brain Res 935:73–80
Deitch HR, Mershon JL, Clark KE 2001 Estrogen receptor ? is the predominant isoform expressed in the brain of adult and fetal sheep. Am J Obstet Gynecol 184:1077–1079
Shughrue PJ, Lane MV, Merchenthaler I 1997 Comparative distribution of estrogen receptor- and -? mRNA in the rat central nervous system. J Comp Neurol 388:507–525
Dhandapani KM, Brann DW 2002 Estrogen-astrocyte interactions: implications for neuroprotection. BMC Neurosci 3:6
Zwain IH, Arroyo A, Amato P, Yen SS 2002 A role for hypothalamic astrocytes in dehydroepiandrosterone and estradiol regulation of gonadotropin-releasing hormone (GnRH) release by GnRH neurons. Neuroendocrinology 75:375–383
Dhandapani KM, Hadman M, De Sevilla LM, Wade MF, Mahesh VB, Brann DW 2003 Astrocyte protection of neurons: role of TGF-? signaling via a c-Jun-AP-1 protective pathway. J Biol Chem 278:43329–43339
Dhandapani KM, Brann DW 2003 TGF-?: a neuroprotective factor in cerebral ischemia. Cell Biochem Biophys 39:13–22
Ruocco A, Nicole O, Docagne F, Ali C, Chazalviel L, Komesli S, Yablonsky F, Roussel S, MacKenzie ET, Vivien D, Buisson A 1999 A transforming growth factor-? antagonist unmasks the neuroprotective role of this endogenous cytokine in excitotoxic and ischemic brain injury. J Cereb Blood Flow Metab 19:1345–1353
Bruno V, Battaglia G, Casabona G, Copani A, Caciagli F, Nicoletti F 1998 Neuroprotection by glial metabotropic glutamate receptors is mediated by transforming factor-?. J Neurosci 18:9594–9600
D’Onofrio M, Cuomo L, Battaglia G, Ngomba RT, Storto M, Kingston AE, Orzi F, De Blasi A, Di Iorio P, Nicoletti F, Bruno V 2001 Neuroprotection mediated by glial group-II metabotropic glutamate receptors requires the activation of the MAP kinase and the phosphatidylinositol-3-kinase pathways. J Neurochem 78:435–445
Docagne F, Nicole O, Gabriel C, Fernandez-Monreal M, Lesne S, Ali C, Plawinski L, Carmeliet P, MacKenzie ET, Buisson A, Vivien D 2002 Smad3-dependent induction of plasminogen activator inhibitor-1 in astrocytes mediates neuroprotective activity of transforming growth factor-?1 against NMDA-induced necrosis. Mol Cell Neurosci 21:634–644
Gabriel C, Ali C, Lesne S, Fernandez-Monreal M, Docagne F, Plawinski L, MacKenzie ET, Buisson A, Vivien D 2003 Transforming growth factor - induced expression of type 1 plasminogen activator inhibitor in astrocytes rescues neurons from excitotoxicity. FASEB J 17:277–279
Treichel JA, Reddington M, Kreutzberg GW 1998 Regulation of plasminogen activator inhibitor-1 mRNA accumulation by basic fibroblast growth factor and transforming growth factor-?1 in cultured rat astrocytes. J Neurochem 71:1944–1952
Lesne S, Blanchet S, Docagne F, Liot G, Plawinski L, MacKenzie ET, Auffray C, Buisson A, Pietu G, Vivien D 2002 Transforming growth factor-?1-modulated cerebral gene expression. J Cereb Blood Flow Metab 22:1114–1123
Sohrabji F, Miranda RC, Toran-Allerand CD 1995 Identification of a putative estrogen response element in the gene encoding brain-derived neurotrophic factor. Proc Natl Acad Sci USA 92:11110–11114
Solum DJ, Handa RJ 2002 Estrogen regulates the development of brain-derived neurotrophic factor mRNA and protein in the rat hippocampus. J Neurosci 22:2650–2659
Zafra F, Lindholm D, Castren E, Hartikka J, Thoenen H 1992 Regulation of brain-derived neurotrophic factor and nerve growth factor mRNA in primary cultures of hippocampal neurons and astrocytes. J Neurosci 12:4793–4799[Abstract]
Abraham IM, Han SK, Todman MG, Korach KS, Herbison AE 2003 Estrogen receptor ? mediates rapid estrogen actions on gonadotropin releasing hormone neurons in vivo. J Neurosci 23:5771–5777
Tang X, Batty IH, Downes CP 2002 Muscarinic receptors mediate phospholipase C-dependent activation of protein kinase B via Ca2+, ErbB3, and phosphoinositide 3-kinase in 1321N1 astrocytoma cells. J Biol Chem 277:338–344
Carswell HW, Macrae IN, Gallagher L, Harrop E, Horsburgh KJ 2004 Neuroprotection by a selective oestrogen receptor b agonist in a mouse model of global ischemia. Am J Physiol 287:H1501–H1504
Wassertheil-Smoller S, Hendriz SL, Limacher M, Heiss G, Kooperberg C, Baird A, Kotchen T, Curb JD, Black H, Rossouw JE, Aragaki A, Safford M, Stein E, Laowattana S, Mysiw WJ, WHI Investigators 2003 Effect of estrogen plus progestin on stroke in postmenopausal women: the Women’s Health Initiative: a randomized trial. J Am Med Assoc 289:2673–2684
Anderson GL, Limacher M, Assaf AR, Bassford T, Beresfold SA, Black H, Bonds D, Brunner R, Brzyski R, Caan B, Chlebowski R, Curb D, Gass M, Hays J, Heiss G, Hendrix S, Howard BV, Hsia J, Hubbell A, Jackson R, Johnson KC, Judd H, Kotchen JM, Kuller L, LaCroix AZ, Lane D, Langer RD, Lasser N, Lewis CE, Manson J, Margolis K, Ockene J, O’Sullivan MJ, Phillips L, Prentice RL, Ritenbaugh C, Robbins J, Rossouw JE, Sarto G, Stefanick ML, Van Horn L, Wactawski-Wende J, Wallace R, Wassertheil-Smoller S, Women Health Initiative Steering Committee 2004 Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women’s Health Initiative randomized controlled trial. J Am Med Assoc 291:1701–1712
Rossberg MI, Murphy SJ, Traystman RJ, Hurn PD 2000 LY353381.HCl, a selective estrogen receptor modulator, and experimental stroke. Stroke 31:3041–3046(Krishnan M. Dhandapani, F)