Estrogen Can Act via Estrogen Receptor and ? to Protect Hippocampal Neurons against Global Ischemia-Induced Cell Death
http://www.100md.com
内分泌学杂志 2005年第7期
Department of Obstetrics, Gynecology and Women’s Health (N.R.M.), Division of Reproductive Endocrinology and Infertility; Department of Neuroscience (T.J., R.S.Z., A.M.E.); and Department of Epidemiology and Population Health (H.W.C.), Albert Einstein College of Medicine, Bronx, New York 10461
Address all correspondence and requests for reprints to: Anne M. Etgen, Ph.D., Albert Einstein College of Medicine, Department of Neuroscience, 1300 Morris Park Avenue, Forchheimer 113, Bronx, New York 10461. E-mail: etgen@aecom.yu.edu.
Abstract
Estradiol at physiological concentrations intervenes in apoptotic death cascades and ameliorates neuronal death in experimental models of focal and global ischemia. The cellular targets that mediate estradiol protection of hippocampal neurons in global ischemia are, however, unclear. The present study examined the hypothesis that estradiol protects hippocampal neurons in ovariectomized rats via estrogen receptor (ER) and/or ?. Estradiol (14 d pretreatment) afforded robust protection of CA1 neurons against global ischemia-induced death. The broad-spectrum ER antagonist ICI 182,780 (intracerebroventricularly, 0 and 12 h after ischemia) abolished estrogen protection, consistent with a role for ERs. To evaluate the potential roles of ER vs. ER? in estrogen protection, we administered subtype-selective agonists for 14 d before and 7 d after ischemia. The ER-selective agonist propyl pyrazole triol (PPT, 10 mg/kg) and ER?-selective agonist WAY 200070–3 (1 mg/kg) produced nearly complete protection of CA1 neurons in approximately 50% of the animals. PPT, but not WAY 200070–3, at doses used for protection, elicited lordosis, induced negative feedback inhibition of LH release, and reduced weight gain. These findings establish the efficacy of the PPT dose in neuroendocrine assays and specificity of WAY 200070–3 for ER?. We also examined the ability of estradiol and neuronal injury to regulate ER and ER? expression. Both estradiol and global ischemia markedly increased ER, but not ER?, protein in CA1. These data indicate that estradiol can act via ER and ER? to protect CA1 neurons from global ischemia-induced death and that both estradiol and global ischemia modulate ER expression in hippocampal CA1.
Introduction
THE BENEFITS OF postmenopausal estrogen replacement therapy for protecting the brain against the neuronal death associated with neurological diseases and disorders remain controversial. Global ischemia, arising during cardiac arrest or surgery in humans or induced experimentally in animals, elicits selective, delayed neuronal death; pyramidal neurons of the hippocampal CA1 are particularly vulnerable (1, 2). The ischemia-induced loss of CA1 neurons is associated with cognitive deficits. 17?-Estradiol (E2), the primary estrogen produced and secreted by the ovaries, and other estrogens afford neuroprotection in experimental models of global and focal ischemia (3, 4, 5, 6, 7, 8, 9). Recent work from this laboratory demonstrates that E2 at levels considered physiological for postmenopausal women prevents the activation of apoptotic signaling cascades and ameliorates global ischemia-induced neuronal loss in male gerbils (7). However, the cellular targets that mediate the protective actions of E2 on hippocampal neurons are, as yet, unclear.
Estrogen exerts a number of physiological actions via interaction with intracellular estrogen receptors (ERs), which serve as ligand-activated transcription factors (reviewed in Refs. 3 , 6 , 10 , and 11). Both of the known ER subtypes, ER and ER?, are expressed in hippocampal neurons (12, 13, 14, 15). E2 has widespread actions on the brain. In the hippocampus, E2 regulates spine density (16, 17), synapse number (18), the synthesis of neurotrophic factors (3, 6, 19), and N-methyl-D-aspartate receptor expression (18, 20). Studies involving mice with targeted deletions of ER and ER? indicate that ER mediates E2 protection of cortical neurons in an animal model of focal ischemia (21). However, ER? is implicated in the neuroprotective actions of E2 in mice subjected to global ischemia (22), and agonists selective for both ER and ER? can partially protect gerbil CA1 neurons from global ischemia-induced cell death (9).
The purpose of the present study was: 1) to determine whether the neuroprotective effects of E2 in a global ischemia model, which produces selective, delayed death of CA1 pyramidal neurons in rats, is mediated by ERs; 2) to identify which ER mediates neuroprotection in this model; and 3) to determine the ability of E2 and neuronal injury (global ischemia) to regulate expression of ER and ER? in the hippocampal CA1. We show that the broad-spectrum ER antagonist ICI 182,780 abolishes the neuroprotective effects of E2 in ovariohysterectomized (ovx) female rats subjected to global ischemia. To validate the dose of the ER-selective agonist propyl pyrazole triol (PPT) in eliciting ER-mediated neuroendocrine responses and the specificity of the ER? WAY 200070–3, we performed independent physiological tests. At the doses chosen for the ischemia studies, PPT, but not WAY 200070–3, facilitated female reproductive behavior, suppressed LH release, and minimized body weight gain. Both the ER and ER? agonists afforded protection against global ischemia-induced loss of hippocampal CA1 neurons in approximately 50% of animals. Estrogen pretreatment and global ischemia increased ER, but not ER?, protein expression in the CA1. These findings indicate that ERs mediate the survival of CA1 neurons afforded by E2 in global ischemia and suggest a role for both ER and ER? in protection against global ischemia-induced cell death.
Materials and Methods
Animals and experimental treatments
All procedures were performed in accordance with the National Institutes of Health (NIH) Guidelines for the Care and Use of Laboratory Animals and were approved by the Albert Einstein College of Medicine Institutional Animal Care and Use Committee. Rodents were maintained in a temperature- and light-controlled environment with 14-h light, 10-h dark cycle and had access to standard rat chow and water ad libitum.
Ovariohysterectomy and E2 pellet placement
Twenty-one-day-old female Sprague Dawley rats (Charles River Laboratories, Inc., Wilmington, MA) were ovx on day zero, under halothane (3% for induction, 1% for maintenance) anesthesia. On the same day, pellets containing E2 (50 μg; 21-d sustained release; Innovative Research of America, Sarasota, FL) or placebo were inserted sc beneath the dorsal surface of the neck. These pellets are designed to maintain serum hormone levels for 21 d. Experiments were carried out so that animals weighed approximately 100–120 g at the time global ischemia was induced, because the four-vessel occlusion (see below) produces the most reliable ischemia in rats of this size.
ER-selective agonists and ER antagonist compounds
For experiments using ER-selective compounds, ovx rats received daily sc injections of: 1) the ER agonist PPT (provided by Dr. Istvan Merchenthaler, Wyeth-Ayerst Laboratories, Inc., Collegeville, PA) at a dose of 10 mg/kg; 2) an ER? agonist WAY 200070–3 (provided by Dr. Heather Harris, Wyeth-Ayerst Laboratories, Inc.) at a dose of 1 mg/kg; or 3) vehicle, 25% dimethylsulfoxide in 0.9% saline, for a total vol of 200 μl. Agonists were always injected between 0700 and 0800 h. On the day of ischemia induction, the injections occurred between 2 and 4 h before surgery. For intracerebroventricular (ICV) injections: 1) 100 μg of the nonselective ER antagonist ICI 182,780 (Tocris Cookson, Inc., Ellisville, MO), or 2) vehicle (50% dimethylsulfoxide, in 0.9% saline) was injected into the right lateral ventricle, in a vol of 5 μl per injection, with a Hamilton syringe (Fisher Scientific, Pittsburgh, PA) using standard stereotaxic methods. The structures of these compounds are shown in Fig. 1.
FIG. 1. Structure of the ER-selective agonists and antagonist. A, ER-selective agonist, PPT (23 ). B, ER?-selective agonist, WAY-200070–3 (WAY) (34 ). C, ER antagonist ICI 182,780 (49 ).
ER-selective agonist treatment
For experiments using ER-selective agonists, PPT (ER agonist), WAY 200070–3 (ER? agonist), or vehicle was administered once daily from 24 h after ovx until d 21, when rodents were euthanized for histology. Body weights were measured at the time of ovx and then every 2–3 d to determine the correct dosage of the ER-selective agonists and to assess ER agonist effects on weight gain. PPT is 410-fold selective for ER relative to ER? (23). WAY 200070–3 is 110-fold selective for ER?, relative to ER, as assessed by radioligand binding assays, and it fails to stimulate uterine weight gain when administered sc at doses as high as 50 mg/kg (H. Harris, personal communication). Although the ability of WAY 200070–3 to cross the blood-brain barrier is not well characterized, upon sc administration of 3 mg/kg WAY 200070–3, a concentration of 12 ng/g brain tissue is achieved by 1 h (H. Harris, personal communication).
Global ischemia
On d 13, using halothane anesthesia, the vertebral arteries were exposed through a midline occipital-suboccipital incision and coagulated with bipolar cauterization between the first and second cervical vertebral bodies. This procedure by itself has no effect on cerebral blood flow but prevents collateral circulation to the forebrain during transient carotid artery occlusion on d 14. Twenty-four hours later, transient global ischemia was accomplished by bilateral occlusion of the common carotid arteries for 10 min followed by reperfusion. A rectal probe was inserted to monitor core temperature and was maintained at 36.5–37.5 C using a heat lamp during ischemia. Sham-operated rats had their vertebral arteries coagulated on d 13.
ICV injections
In the first experiment, immediately after ischemia (24 h after coagulating the vertebral arteries for sham-treated rats) and again 12 h later, rats were positioned in a stereotaxic apparatus and ICV injections performed under halothane anesthesia. The position of the right lateral ventricle was calculated based on the position of bregma (anterior/posterior, 0.92 mm; medial/lateral, 1.2 mm; dorsal/ventral, 3.6 mm), and then the nonselective ER antagonist ICI 182,780 or vehicle was infused with a 28-gauge needle, in a vol of 5 μl per infusion, over 1 min. The injection needle was left in place for an additional 1 min before being withdrawn.
Histological analysis
We first examined the impact of transient global ischemia on the volume of the hippocampal CA1. Rats were subjected to global ischemia or sham operation (n = 6 per group) and, at 7 d after surgery, were killed by transcardiac perfusion with 4% paraformaldehyde under deep anesthesia. Brains were rapidly removed, and the volume of the CA1 from ischemic and sham rats was assessed by stereological methods. Toluidine blue-stained coronal sections (20 μm) were collected through the entire dorsal hippocampus (bregma –2.3 to –4.5 mm). Digital images of every tenth section from each animal (100 sections per brain) were captured and used to trace the outline of the CA1. The mean area ± SEM of the entire CA1 was calculated and multiplied by the section thickness times the number of sections along the rostral/caudal axis. Although the CA1 pyramidal neurons were nearly ablated at 7 d after ischemia, no statistically significant difference in volume of the CA1 was detected for ischemic (2.43 ± 0.6 x 108 μm3; n = 6) vs. sham-operated rats (2.42 ± 0.5 x 108 μm3; n = 6). Based on this information, counts of pyramidal neurons were assessed in a 250-μm length of the CA1 pyramidal cell layer, in 3–4 sections per animal, at 7 d after ischemia or sham operation as described below.
Seven days after global ischemia (d 21 after ovx), animals were euthanized by deep anesthesia, and blood was collected by cardiac puncture for analysis of serum E2 levels. Each rat was transcardially perfused using 0.9% saline solution with heparin (150 ml; 15 min) followed by ice-cold 10% buffered formalin phosphate (200 ml; 20 min; Fisher Scientific, Pittsburgh, PA). Brains were then removed, placed in formalin at 4 C overnight, fixed in 30% sucrose in PBS at 4 C for 48 h, and then frozen at –40 C. The dorsal hippocampus was coronally sectioned on a cryotome into 15-μm slices, and every fourth section was collected, mounted, and stained with toluidine blue. Using x60 magnification, the CA1 region of the left and right hippocampi was photographed. Surviving pyramidal neurons in a 250-μm length of the stratum pyramidale of the left and right CA1 (see arrows in Fig. 2B, panel A) were counted. Cell counts are expressed as the average number of surviving neurons per side, counted from four sections per animal. More than 95% of the cells in the stratum pyramidale are pyramidal neurons; most glia and the cell bodies of inhibitory interneurons are localized to the stratum radiatum or stratum oriens. It is well established that a brief (10 min) episode of global ischemia affords delayed, selective death of CA1 pyramidal neurons. Inhibitory interneurons of the CA1 and all neurons of the nearby CA2 or transition zone, CA3, and dentate gyrus survive (2).
FIG. 2. ER antagonist ICI 182,780 treatment abolishes E2 neuroprotection after global ischemia. All groups were treated with E2 or placebo (P) for 2 wk before undergoing ischemia or sham surgery. ICI 182,780 (ICI) or vehicle (V) was then injected ICV immediately after ischemia and repeated 12 h later (for sham groups, ICI 182,780 or vehicle was injected 24 h and 36 h after sham operation). A, E2 treatment affords robust neuroprotection in the CA1 region of the hippocampus after global ischemia (E2+V vs. P+V in ischemic animals; #, P < 0.002). Injection of the ER antagonist ICI 182,780 after E2 treatment and ischemia abolished this neuroprotection (E2+ICI vs. E2+V in ischemic animals: , P < 0.002). Treatment with the ER antagonist did not affect the number of surviving neurons in any of the other treatment groups; therefore, for statistical analysis, sham-operated vehicle and ICI 182,780-treated groups were combined. Rats that underwent global ischemia had significantly fewer surviving neurons than the sham-operated groups (*, P < 0.001). Data (n = 3–10 per treatment group) are represented as medians and analyzed using Kruskal Wallis with Dunn’s post hoc analysis. B, Histologic evaluation of the dorsal hippocampus after global ischemia or sham operation in low (x4, top) and high power (x60, bottom) magnification. DG, Dentate gyrus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Placebo (A and B) and E2 (C and D)-treated rats that underwent sham operation similarly have many surviving neurons in the CA1 region of the hippocampus. Placebo-treated rats that underwent global ischemia (E and F) have selective, delayed neuronal death in CA1. G and H, E2-treated rats that underwent global ischemia have selective, delayed neuronal death in CA1, but approximately 50% of the neurons survived. I and J, E2-treated rats that underwent global ischemia and then received ICI 182,780 have few surviving neurons in the CA1 region. C, Rats treated with E2 have higher serum E2 levels than the placebo (P)-treated group. Sera were collected 15 d (E2 15d) or 21 d (E2 21d) after E2 pellets were placed. Statistically significant differences between hormonal treatments and time are shown (E2 15d vs. P: *, P < 0.001; E2 21d vs. P: , P < 0.002; E2 15d vs. E2 21d: *, P < 0.001). Data (n = 20–33 per treatment group) are represented as mean ± SE and analyzed using one-way ANOVA.
Behavioral testing
Adult female Sprague Dawley rats, weighing 200–350 g (Taconic Farms, Inc., Germantown, NY), were ovx and maintained in a reverse light-dark cycle (14-h dark, 10-h light, with lights off at 1100 h). Two weeks later, rats (n = 8) were randomly assigned to receive E2 benzoate (2 μg in 100 μl peanut oil) or an ER-selective agonist sc daily for 2 d. The PPT dose was 10 mg/kg·d, and the WAY 200070–3 dose was 1 mg/kg·d. On the third day, progesterone (500 μg in 100 μl peanut oil) was administered, and behavior testing was performed 4 h later. All tests were performed in the dark, under red light illumination, between 1300 and 1700 h. Each female rat was placed into a cage with a sexually active male rat, and female sexual receptivity was recorded. The lordosis quotient (LQ) was determined by counting the number of times the female assumed the lordosis posture, in response to 10 mount attempts by the male, and then multiplied by 100. One week later, the same female rats were injected with a different ER-selective agonist or E2 benzoate, and behavior testing was repeated. Testing was continued until all female rats were exposed to each treatment.
Serum E2 assay
Tubes containing whole blood were placed on ice for 20 min and centrifuged at 300 x g for 5 min, and sera were collected and stored (–20 C) until analyzed. Serum hormone levels were measured by fluoroimmunoassay using the DELPHIA E2 assay (PerkinElmer Life Sciences, Turku, Finland). All assays were performed in duplicate, and the mean value was reported. The sensitivity of detection was 13 pg/ml. The inter- and intraassay coefficients of variance were 10.1% and 4.1%, respectively.
LH assay
Serum LH levels were measured by enzyme immunoassay (Biotrak; Amersham Biosciences, Buckinghamshire, UK). Assays were performed in duplicate, and the mean value was reported. The sensitivity of detection is 0.1 ng/ml. The cross-reactivity of LH with rat FSH was less than 0.016%. The inter- and intraassay coefficients of variance are 6.2% and 2.4%, respectively.
Western blot
For quantification of protein in the CA1 hippocampal region, immunoblots were performed as described (7). The ovx animals received placebo or E2 pellets as described above and were subjected to sham operation or 10-min global ischemia 2 wk later. Animals were deeply anesthetized and euthanized by decapitation at 24 h and 48 h after ischemia or 24 h after sham operation. Before decapitation, blood samples were collected by cardiac puncture from animals killed 24 h after ischemia or sham operation, for later analysis of serum E2 levels on d 15. Hippocampi were quickly dissected and cut into 1-mm transverse slices on a McIllwain tissue chopper (Vibratone Co., St. Louis, MO). The entire CA1 region from these slices of the left and right hippocampus was separated quickly by microdissection, placed in ice-cold PBS containing the protease inhibitor phenylmethylsulfonylfluoride (1 mM; Sigma- Aldrich Corp., St. Louis, MO), and stored at –20 C until used. Protein concentration of samples was determined using the BCA protein assay kit (Pierce Chemical Co., Rockford, IL). Laemmli’s buffer was combined with 30 μg protein and loaded onto 10% polyacrylamide minigels (Bio-Rad Laboratories, Inc., Richmond, CA) and then subjected to electrophoresis.
Protein bands were transferred onto a polyvinylidene fluoride membrane (Millipore Corp., Bedford, MA). After blocking for 30 min with 25 mM Tris-HCl (pH 8.0) containing 125 mM NaCl, 0.1% Tween 20, and 4% skim milk, membranes were incubated with primary antibody, washed, and then incubated with secondary antibody. After reaction, membranes were treated with enhanced chemiluminescence reagents (Amersham Biosciences, Arlington Heights, IL) and apposed to x-ray film (Eastman Kodak Co., Rochester, NY).
Antibodies for Western blot
Primary antibodies were: 1) a mouse monoclonal antibody directed to human ER (ER 6F11; Vector Laboratories Ltd., Burlingame, CA) at a concentration of 1:100 and incubated at 4 C overnight, and 2) a polyclonal rabbit antibody directed to amino acids 64–82 of rat ER? (ER? antibody; provided by Dr. Stephen Alves, Merck & Co., Inc., Rahway, NJ) at a concentration of 1:30,000 and incubated at room temperature for 16 h. Both of these antibodies have been validated for use in immunoblotting of rat brain tissue. Secondary antibodies were: 1) a horseradish peroxidase-conjugated antimouse IgG for ER (Cell Signaling Technology, Inc., Beverly, MA) at a concentration of 1:1000 and incubated for 2 h at room temperature; and 2) horseradish peroxidase-conjugated antirabbit IgG for ER? (Cell Signaling Technology, Inc.) at a concentration of 1:2,000 and incubated for 1.5 h at room temperature.
A Scan Jet 4-C computing densitometer (Hewlett-Packard Co., Palo Alto, CA) was used with NIH Image 1.61 software to quantitate protein abundance. Optical densities of ER bands were first normalized to the amount of protein loaded, assessed by stripping and reblotting for type III ?-tubulin in each sample. These values were then expressed relative to normalized values obtained for samples from ovx, sham-operated rats on the same membrane to allow for comparison of band densities of immunoblots exposed to different pieces of film.
Statistical analysis
Data analysis was performed using SPSS 12.0 (SPSS, Inc., Chicago, IL). One-way ANOVA was used to compare group means of normally distributed data (serum E2, serum LH, rat weight gain, and immunoblots) with Newman-Keuls post hoc analysis. Friedman’s test with Dunn’s multiple-comparison post hoc analysis was performed to analyze lordosis behavior tests. Kruskal Wallis testing was used for comparing neuron counts after E2 and/or ICI 182,780 treatment, followed by Dunn’s post hoc analysis. Other nonparametric methods were used to analyze pyramidal neuron counts after ER subtype-selective agonist treatment, because the data were bimodal. Subjects were divided into two categories based on the number of surviving neurons. Animals with greater than 40 surviving neurons per 250 μm were considered to be protected, and animals with less than 30 surviving neurons per 250 μm were considered not to be protected. Then 2 and Fisher’s exact test were used to analyze whether there were treatment-related differences in the number of animals protected and not protected and in the median number of surviving neurons. Differences were considered significant at P 0.05.
Results
To examine the cellular targets that mediate the neuroprotective actions of estrogen on hippocampal neurons, we treated rats with E2 or placebo for 14 d, subjected animals to sham operation or global ischemia, and acutely administered the broad-spectrum ER antagonist ICI 182,780 or vehicle at 0 and 12 h after reperfusion. E2 and placebo pellets remained in place until rats were killed for histological evaluation 7 d after ischemia. Global ischemia induced a substantial loss of hippocampal CA1 neurons in vehicle-treated rats at 7 d post ischemia (loss of 80% of neurons in the CA1 pyramidal cell layer; Fig. 2A). Long-term E2 treatment afforded robust neuroprotection (50% neuronal survival in the CA1) in nine of nine treated rats. The ER antagonist ICI 182,780 alone (injected at 0 and 12 h after reperfusion) did not detectably alter the number of surviving neurons in sham-operated or ischemic rats (Fig. 2A) but abrogated E2-elicited neuroprotection (Fig. 2B). These findings indicate that E2 acts via the classical ERs to elicit protection of hippocampal neurons.
To confirm that estrogen pellet implantation effectively released E2, we monitored serum E2 levels. Groups treated with E2 exhibited significantly higher serum E2 levels than did placebo-treated rats, evident at 15 and 21 d after pellet implantation (Fig. 2C). The mean serum E2 level on d 15 (64 ± 7.0 pg/ml), measured in animals killed 24 h after ischemia for the ER immunoblotting study (see below), was slightly above the peak physiological level seen during proestrus. By d 21, the mean serum E2 level (33 ± 2.9 pg/ml) was in the midphysiological range. The change between d 15 and d 21 most likely reflects the increasing body weight of the animals (see Fig. 3B) and may also indicate that, by the end of the experiment, the 21-d pellets have released most of their E2.
FIG. 3. ER-selective agonist treatment and physiologic responses. A, Treatment with the ER-selective agonist PPT induces lordosis behavior. Two weeks after ovx, the ER agonist (PPT), the ER? agonist (WAY), or E2 benzoate (E2) was administered for two sequential days. Progesterone was given on the third day, and behavioral testing was performed 4 h later. WAY vs. E2: *, P < 0.001; WAY vs. PPT: , P < 0.05. Data are expressed as medians (n = 8) and analyzed using Friedman test. B, Treatment with the ER-selective agonist PPT reduces rat weight gain. Total weight gain determined after 21 d of daily ER agonist or vehicle treatment. PPT vs. V or WAY: *, P < 0.001. Data (n = 17–19 per treatment group) are expressed as mean ± SE and analyzed using ANOVA. C, ER agonist treatment causes feedback inhibition of LH. Sera were collected after 21 d of daily hormone administration. PPT vs. V: , P < 0.02; PPT vs. WAY: *, P < 0.001). Data (n = 16–19 per treatment group) are expressed as mean ± SE and analyzed using ANOVA.
To evaluate the potential roles of ER vs. ER? in mediating estrogen protection of hippocampal neurons, we administered ER- and ER?-selective agonists to ovx rats for 14 d before and 7 d after ischemia. To validate the doses of the agonists used for protection, and to evaluate the specificity of the agonists for the ER subtypes, we tested known actions of estrogen mediated entirely (lordosis behavior) or at least in part (suppression of LH release and weight gain) at the level of the brain. In a pilot study, the ER agonist PPT at 2 mg/kg·d did not elicit a lordosis response; rats exhibited a median LQ of zero (interquartile range: 0, 15; n = 8). PPT at a higher dose (10 mg/kg) elicited a high median LQ similar to that produced by administration of E2 benzoate (Fig. 3A). This finding verifies that PPT at 10 mg/kg crosses the blood-brain barrier and achieves a concentration in brain tissue sufficient to produce neuroendocrine responses mediated by ER at the level of the hypothalamus within 1–2 d after administration. The ER?-selective agonist WAY 200070–3 did not elicit lordosis at either 1 mg/kg (Fig. 3A) or 10 mg/kg (median LQ = 10, interquartile range: 5, 15; n = 8), suggesting that it does not cross-react detectably with ER to elicit sexual behavior in this dose range.
The ER-selective agonist PPT at 10 mg/kg·d, but not WAY 200070–3 (1 mg/kg·d), also reduced weight gain by young rats during the 3-wk regimen used in the ischemia studies (Fig. 3B). Control (vehicle-treated) rats gained an average of 92 g during the 21 d, whereas PPT-treated rats (10 mg/kg·d) gained an average of only 59 g. These PPT-treated rats differed significantly from both the vehicle- and WAY 200070–3-treated groups (P < 0.01). Rats treated with 2 mg/kg·d PPT also gained significantly more weight than did those treated with 10 mg/kg PPT (P < 0.002). The weight gained by rats treated with the lower ER agonist dose (data not illustrated) did not differ significantly from either that of the vehicle- or of the ER?-agonist-treated groups. Within a given agonist treatment group, there was no significant difference in weight gain for ischemic vs. sham-operated rats.
Estrogen also elicits negative feedback inhibition of LH release, at least in part via ER (24). Administration of the ER-selective agonist PPT (2 mg/kg·d) reduced the median LH serum level in ovx rats from approximately 15 ng/ml to 4.6 ng/ml (interquartile range: 2.4, 9.0; P < 0.02 vs. vehicle-treated rats), consistent with negative feedback inhibition of LH release. Increasing the dose of the ER agonist to 10 mg/kg·d reduced the mean serum LH level further (Fig. 3C). In contrast, WAY 200070–3 did not detectably alter serum LH levels relative to ovx, vehicle-treated females (Fig. 3C).
Both the ER- (10 mg/kg·d) and ER?- (1 mg/kg·d) selective agonists elicited pronounced neuroprotection in approximately 40–50% of treated ischemic rats. Both groups exhibited bimodal distributions in the number of surviving neurons in the hippocampal CA1. Approximately 40–50% of the rats treated with each agonist showed virtually complete survival of CA1 pyramidal neurons; the remainder of the rats exhibited approximately10% surviving neurons, a result statistically indistinguishable from that of the vehicle-treated ischemia group (Fig. 4). The behaviorally ineffective dose of PPT, 2 mg/kg, elicited modest protection in only two of seven rats subjected to global ischemia (data not shown). Interestingly, these two rats gained less weight during the experiment (43.5 ± 9.1 g) than did the four rats that showed extensive loss of CA1 pyramidal neurons (83.6 ± 11.3 g). Raising the dose of WAY 200070–3 to 10 mg/kg did not significantly improve the outcome relative to that of the lower dose (one of five rats showed increased survival of CA1 neurons relative to vehicle controls). Neither ER agonist induced detectable pyramidal cell loss in CA1 when given alone to sham-operated rats.
FIG. 4. ER (PPT, 10 mg/kg·d) and ER? (WAY, 1 mg/kg·d) agonists elicit neuroprotection in some rats after global ischemia. Rats were injected daily with vehicle (V), PPT, or WAY for 2 wk before undergoing global ischemia or sham surgery. Hormonal treatments were continued until rats were euthanized 7 d after ischemia or sham operation. Sham-operated rats are represented by open symbols, and ischemic rats are represented by filled symbols. The median number of surviving pyramidal cells for each treatment group is demonstrated by a horizontal line, and differences between groups were analyzed by Fisher’s exact test. (ischemia+V vs. sham+V: ***, P < 0.001; PPT+ischemia vs. PPT+sham: *, P < 0.04; WAY+ischemia vs. WAY+sham: **, P < 0.02). Differences in the number of rats per group with more than 40 surviving neurons after drug treatment and ischemia were analyzed using 2 (ER?+ischemia vs. V+ischemia: , P = 0.04; ER+ischemia vs. V+ischemia: #, P = 0.05).
To assess whether the bimodal distribution of CA1 pyramidal neuron survival was related to individual differences in accessibility of brain tissues to 10 mg/kg PPT, we examined serum LH levels in protected (n = 5) and nonprotected (n = 8) rats. Mean LH levels were essentially identical in the two groups (protected, 2.35 ± 0.8 ng/ml vs. nonprotected, 2.25 ± 0.8 ng/ml). Likewise, there was no significant difference between LH levels and pyramidal cell survival in the CA1 of rats treated with 1 mg/kg WAY 200070–3 (protected, 27.0 ± 15.6 ng/ml, n = 5; nonprotected, 19.4 ± 7.2 ng/ml, n = 6). We also detected no difference in weight gain during the 3 wk of agonist administration in PPT-treated females that were protected (47.8 ± 14.6 g) and not protected (57.8 ± 7.1 g). The same was true for WAY 200070–3 (protected, 90.2 ± 22.1 g vs. nonprotected, 93.3 ± 20.3 g). Moreover, linear regression analysis revealed no correlation between plasma LH levels or weight gain and pyramidal neuron counts in CA1 of PPT- or WAY 200070–3-treated rats (data not illustrated).
Previous studies indicate that neuronal injury, including focal ischemia, can modify ER expression levels (25, 26). Therefore, we determined the abundance of ER and ER? protein in the CA1 region of the hippocampus of ovx rats treated with placebo or E2 pellets for 14 d before global ischemia or sham operation. Global ischemia induced a significant increase in ER protein expression, evident at 24 and 48 h after ischemia in the placebo-treated groups relative to placebo-treated, sham-operated rats (Fig. 5). E2 pretreatment alone also increased ER protein levels in CA1 in sham-operated rats. The combination of E2 pretreatment and ischemia did not increase ER protein levels above that produced by either manipulation alone. Immunoblots for ER in other hippocampal subfields indicated that neither ischemia nor estrogen changed ER levels in the dentate gyrus (data not illustrated). In CA3, ischemia, but not estrogen, modestly increased ER at both 24 and 48 h. Estrogen prevented the ischemia-induced increase at 24 h but not 48 h (data not illustrated). Although global ischemia induced a trend toward down-regulation of ER? protein in CA1, this effect was not significant at either time (Fig. 6; ANOVA, P < 0.06). Likewise, E2 pretreatment did not significantly alter ER? protein abundance in the CA1 of either ischemic females or sham-operated rats.
FIG. 5. ER protein expression in ovx rats is up-regulated in CA1 by E2 and after global ischemia. Rats were treated with E2 pellets or placebo (P) for 2 wk before undergoing global ischemia or sham surgery. Rats were euthanized 24 h (24 ) or 48 h (48 ) after ischemia (I) or 0 h for sham operation (S). A, Representative immunoblots of ER (top) and ?-tubulin (bottom) in hippocampal CA1 from rats treated under each experimental condition. B, Quantitative analysis of ER after ischemia. ER is up-regulated at 24 h and 48 h after global ischemia in placebo-treated groups (P+24 and P+48 vs. C: **, P < 0.001). ER is also up-regulated in all E2-treated groups (E2+24 and E2+S vs. S: *, P < 0.01; E2+48 vs. C: **, P < 0.001). For each treatment, ER antibody is normalized to ?-tubulin and expressed as the fold stimulation over that of placebo-treated, sham-operated rats. Values represent mean ± SEM (n = 6) analyzed using ANOVA.
FIG. 6. ER? protein expression after global ischemia. Rats were treated with E2 pellets or placebo (P) for 2 wk before undergoing global ischemia or sham surgery. Rats were euthanized 24 h (24 ) or 48 h (48 ) after ischemia (I) or at 0 h for sham operation (S). A, Representative immunoblots of ER? (top) and ?-tubulin (bottom) from rats treated under each experimental condition. B, Quantitative analysis of ER? protein levels in CA1 after ischemia in placebo- and E2-treated female rats. For each treatment, ER? immunoblots are normalized to ?- tubulin and expressed as the fold stimulation over that of placebo-treated, sham-operated rats. Values represent mean ± SEM (n = 6) and were analyzed using ANOVA.
Discussion
The ability of estrogen to afford protection in animal models of focal and global ischemia is well established (3, 4, 5, 6, 7, 8, 9). Our demonstration that E2 affords robust protection in estrogen-deprived female rats is consistent with findings of others, i.e. that estrogen affords protection against global ischemia-induced neuronal death in proestrous female rats (5) and in male and female gerbils (7, 8, 27, 28, 29) and mice (30). Our study extends those findings in that it identifies cellular targets that mediate the protective actions of E2 on rat hippocampal neurons in vivo. We demonstrate that the broad-spectrum ER antagonist ICI 182,780 (given ICV at 0 and 12 h after ischemia) abolished E2 protection, arguing against a role for ER-independent mechanisms in estrogen protection in this model. Because the ER antagonist was infused directly into the brain ventricular system, this finding also implicates brain ERs as the mediators of estrogen neuroprotection in transient global ischemia. These data extend previous work in female gerbils showing that systemic administration of the selective estrogen response modifiers, raloxifene and tamoxifen, can attenuate estrogen’s protective actions on global ischemia-induced cell death in CA1 neurons (8, 9).
We also report the novel finding that pretreatment with agonists selective for both ER and ER? can preserve a large majority of hippocampal CA1 pyramidal neurons from global ischemia-induced death in some rats. The ER-selective agonist PPT and the ER?-selective agonist WAY 200070–3 afforded marked protection of CA1 neurons in approximately 40–50% of estrogen-deprived female rats. WAY 200070–3, administered at doses that afforded neuroprotection, did not elicit lordosis behavior, negative feedback inhibition of LH release, or a reduction in weight gain, providing independent validation that the drug does not cross-react with ER in vivo. By contrast, the 10-mg/kg dose of PPT elicited all of these actions of E2 that are known to be mediated, at least in part, by brain ER receptors. We further show that global ischemia regulates ER expression in hippocampal neurons. There was a marked increase in ER protein and a small, but not significant, decrease in ER? protein in CA1 at both 24 and 48 h after ischemia. Estrogen pretreatment mimicked the injury-induced increase in ER protein expression, and the effects of ischemia and E2 pretreatment were not additive. These data indicate that E2 can act via ER and ER? to protect CA1 pyramidal neurons from global ischemia-induced death and that both E2 and global ischemia modulate ER expression in the hippocampal CA1.
The finding that estrogen can act via either ER or ER? to elicit protection of hippocampal neurons is of interest in that the cellular targets that mediate estrogen neuroprotection remain controversial. Studies involving transgenic mice with targeted deletion of either ER or ER? indicate that ER is critical to estrogen protection of cortical and striatal neurons in focal ischemia (21). Interestingly, estrogen does not protect hippocampal neurons in that model, presumably because permanent occlusion of the middle cerebral artery is a catastrophic insult. There is one report that estrogen does not afford protection against focal ischemia-induced neuronal death in mice with targeted deletion of ER (31), although ICI 182,780 administration increased infarct volume in this model (32). These investigators concluded that E2 most likely exerts neuroprotection through the activation of ER?, not ER (31, 32). Interestingly, ER has recently been implicated in the protective actions of estradiol in a model of ischemic liver injury (33).
There are also conflicting reports regarding the role of ER subtypes in neuroprotection after global ischemia. A recent pharmacological study showed that the ER?-selective agonist diaryl propiolnitrile is more effective than PPT (2 mg/kg) in protecting hippocampal neurons against global ischemia-induced death (22). However, our data clearly show that a PPT dose of 2 mg/kg·d is not sufficient to fully activate ER-mediated responses to estrogen, such as lordosis behavior, LH suppression, and reduction of weight gain, in most female rats. Our finding that PPT at a higher dose (10 mg/kg) is as effective as the selective ER? agonist WAY 200070–3 in affording neuroprotection suggests a role for both ER and ER? in estrogen protection. There is also a preliminary report that PPT and unidentified ER?-selective agonists are partially neuroprotective in gerbil CA1 after global ischemia (9). These authors used in situ hybridization signals for neurogranin mRNA to quantify CA1 cell survival, and the duration and dose of agonist treatment were not always specified, so it is somewhat difficult to make direct comparisons between that report and our study. However, in agreement with our results using 2 mg/kg·d PPT, the latter authors detected no protection in ovariectomized gerbils when PPT was used at a dose of 3 mg/kg.
ER-selective agonist treatment and neuroprotection
In our experiments, ER- and ER?-selective agonists produced a bimodal distribution of neuroprotection in rats subjected to global ischemia, with nearly complete survival of CA1 pyramidal neurons in 38 and 45% of animals, respectively, and less than 10% neuronal survival in the remainder of animals (Fig. 4). This pattern of neuroprotection was in contrast to that of E2, which elicited approximately 50% neuronal survival in all animals (Fig. 2). It is notable that, in our hands, a high dose of PPT (10 mg/kg) was necessary to prime lordosis, to suppress weight gain, and to suppress LH completely. The dose of PPT needed to produce biological responses when administered in vivo is in contrast to its high affinity in binding to ER, as assessed by radioligand assays in vitro (23). Thus, it is possible that the ability of these drugs to promote neuronal survival in only a subset of ischemic animals reflects either their low intrinsic efficacy or individual differences in drug disposition and metabolism. However, the bimodal distribution in pyramidal neuron survival is unlikely to be entirely due to individual differences in penetration of drugs through the blood-brain-barrier or to their differential action at peripheral ERs, because neither serum LH levels nor body weight gain differed in rats that exhibited preservation vs. those that exhibited loss of CA1 pyramidal neurons. Others have reported that PPT at 2 mg/kg·d prevented ovx-induced weight gain and maintained uterine weight and total and trabecular bone mineral density for as long as 6 wk in adult rats but that a higher dose (15 mg/kg·d) was required to prevent hot flashes (34). Thus, a relatively high dose of PPT may be required to elicit neural responses mediated by ER. We were unable to independently validate the ER? agonist dose used in our protection studies as physiologic, because there is currently no known behavioral or neuroendocrine response that is mediated exclusively via ER?.
Although our findings suggest that either ER or ER? can protect hippocampal neurons from ischemia-induced cell death, this conclusion must be tempered by the fact that we could not independently verify that WAY 200070–3 actually activated ER? or that PPT failed to do so in the brain or in peripheral tissues. Likewise, we cannot rule out the possibility that both receptors must be activated to rescue CA1 pyramidal cells. Although our data suggest that WAY 200070–3 did not activate brain ER, it is impossible to determine whether the high dose of PPT acted as an agonist at ER?. We attempted to use ER knockout mice to clarify the role of ER subtypes in global ischemia; however, using the standard two-vessel occlusion model of global ischemia in mice, we were unable to reliably produce selective loss of CA1 pyramidal neurons in the wild-type strains upon which the ER knockouts were produced (data not shown). Therefore, the questions raised by our work may only be resolved by the development of highly selective ER and ER? agonist and antagonist ligands that easily penetrate the blood-brain-barrier and mimic the biological activity of E2 at brain ERs.
ER-selective agonists and physiologic responses mediated by brain ERs
The present study also shows that the ER agonist PPT, but not the ER? agonist WAY 200070–3, at doses used for protection, elicits lordosis behavior and negative feedback inhibition of LH release and attenuates weight gain. These studies establish an effective dose for PPT and validate the selectivity of WAY 200070–3 in vivo. However, they do not rule out the possibility that PPT and WAY 200070–3 action on nonneuronal target tissues may also contribute to the neuroprotection. Estrogen negative feedback and suppression of body weight gain involve estrogen action at peripheral (e.g. pituitary, adipose tissue) as well as neural sites. Likewise, we cannot rule out the possibility that the drugs acted on vascular tissues to improve postischemia reperfusion (e.g. see Ref. 5). This explanation is, however, less likely, in that physiological levels of E2 protect neurons against global ischemia-induced cell death even when postischemic cerebral blood flow is strictly controlled (35).
Transgenic mice with targeted deletion of ER or ER? have been used to delineate the differential actions of estrogen mediated by each receptor subtype. These studies implicate ER in E2-dependent sexual receptivity (36, 37, 38), negative feedback regulation of LH (24, 39), and suppression of ovariectomy-induced increases in food intake and weight gain (40). In contrast, female ER? null mutants exhibit normal lordosis responses (37) and normal serum LH levels (41) and are able to reproduce and lactate (42). Our findings that PPT elicited the lordosis response, produced negative feedback inhibition of serum LH levels, and reduced the rate of weight gain in young, ovx female rats provide confirmation of a role for ER in responses to estrogen that are mediated, at least in part, by brain receptors.
ER regulation by E2 and after global ischemia
Our observation that ER protein levels are significantly higher in CA1 when ovx rats receive E2 replacement extends to the in vivo situation the recent finding that E2 increases ER protein levels in cultured hippocampal neurons (43). There is now evidence that E2 may be synthesized de novo in hippocampal neurons of both adult and developing rats (44, 45) and that it is important for maintaining hippocampal synaptic connections (46). Therefore, E2 may be an essential physiological regulator of hippocampal ER expression and synaptic connectivity. Neuronal injury can also regulate expression of a number of genes, including those encoding ERs (9, 25, 26). Our finding that global ischemia up-regulates ER protein expression in the hippocampal CA1 is consistent with findings that focal ischemia up-regulates ER mRNA expression in neocortex (25). In contrast, global ischemia does not increase ER expression in adult macaque monkeys, whereas it markedly increases glial ER? immunostaining in the CA1 of this species (47). Our finding that neither global ischemia nor estrogen significantly alters ER? expression in CA1 thus stands in contrast to the reported actions of global ischemia in monkeys and to previous findings that focal ischemia causes a 50% decrease in ER? mRNA in the rat cerebral cortex, a response that was prevented by exogenous estrogen (25). Hence, changes in receptor expression in response to neuronal injury are unlikely to account for the all-or-none effect of the ER-selective agonists in affording neuroprotection. Moreover, we cannot rule out the possibility that the recently described ER-X, whose levels also increase after ischemia (26), or other molecular entities that bind E2 and ICI 182,780 or that modulate the efficacy of ligand-induced transcriptional activation (48) also contribute to the neuroprotective actions of E2 in vivo.
Summary
It is well established that estrogen affords neuroprotection in animal models of stroke and global ischemia. The present study shows that the neuroprotective actions of estrogen can be mediated by either ER or ER? in a rat model of transient global ischemia. Our finding that the ER antagonist ICI 182,780 administered ICV abolishes protection by estrogen implicates brain ERs as the cellular mediators of estrogen neuroprotection and argues against a role for other, receptor-independent mechanisms. The ability of ICI 182,780 to abolish estrogen protection when given ICV only at 0 and 12 h after ischemia, despite the continued presence of estrogen, also demonstrates that E2 action in the early postischemic period is critical. The up-regulation of ER protein expression in CA1 observed at 24 and 48 h after ischemia, times at which histological evidence of cell death is not yet observed in ischemic rats, suggests that this may be a protective response by the injured brain. Delineation of the cellular targets that mediate neuroprotection by estrogen contributes to our understanding of how estrogen promotes neuronal survival and may help in the prevention and/or treatment of global ischemia arising during cardiac arrest.
Acknowledgments
Special thanks to Istvan Merchenthaler, Ph.D., and Heather Harris, Ph.D., from Wyeth-Ayerst Laboratories, Inc., for supplying the ER-selective agonists for this study and to Stephen Alves, Ph.D., from Merck, for supplying the ER? antibody. Dr. Patrick Hof of Mount Sinai School of Medicine provided invaluable assistance with the stereological estimates of CA1 volume. We also acknowledge Dr. Nanette Santoro for her thoughtful scientific input and comments on the manuscript. We also thank Dr. Alice Shu, Ms. Monique Bryan, and Dr. Tovaghgol Adel for excellent technical assistance.
References
Tanaka H, Grooms SY, Bennett MV, Zukin RS 2000 The AMPAR subunit GluR2: still front and center-stage. Brain Res 886:190–207
Zukin RS, Jover T, Yokota H, Calderone A, Simionescu M, Lau CG 2004 Molecular and cellular mechanisms of ischemia-induced neuronal death. In: Mohr JP, Choi DW, Grotta JC, Weir B, Wolf PA, eds. Stroke: pathophysiology, diagnosis, and management. Philadelphia: The Curtis Center; 829–854
Behl C 2002 Oestrogen as a neuroprotective hormone. Nat Rev Neurosci 3:433–442
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
He Z, He YJ, Day AL, Simpkins JW 2002 Proestrus levels of estradiol during transient global cerebral ischemia improves the histological outcome of the hippocampal CA1 region: perfusion-dependent and -independent mechanisms. J Neurol Sci 193:79–87
Garcia-Segura LM, Azcoitia I, DonCarlos LL 2001 Neuroprotection by estradiol. Prog Neurobiol 63:29–60
Jover T, Tanaka H, Calderone A, Simionescu M, Yokota H, Etgen AM, Zukin RS 2002 Estrogen protects against global ischemia-induced neuronal death and prevents activation of apoptotic signaling cascades in the hippocampal CA1. J Neurosci 22:2115–2124
Shughrue PJ, Merchenthaler I 2003 Estrogen prevents the loss of CA1 hippocampal neurons in gerbils after ischemic injury. Neuroscience 116:851–861
Merchenthaler I, Dellovade TL, Shughrue PJ 2003 Neuroprotection by estrogen in animal models of global and focal ischemia. Ann NY Acad Sci 1007:89–100
Toran-Allerand CD 2004 Minireview: a plethora of estrogen receptors in the brain: where will it end? Endocrinology 145:1069–1074
Blaustein JD 2004 Minireview: neuronal steroid hormone receptors: they’re not just for hormones anymore. Endocrinology 145:1075–1081
Perez SE, Chen EY, Mufson EJ 2003 Distribution of estrogen receptor and ? immunoreactive profiles in the postnatal rat brain. Brain Res Dev Brain Res 145:117–139
Mitra SW, Hoskin E, Yudkovitz J, Pear L, Wilkinson HA, Hayashi S, Pfaff DW, Ogawa S, Rohrer SP, Schaeffer JM, McEwen BS, Alves SE 2003 Immunolocalization of estrogen receptor ? in the mouse brain: comparison with estrogen receptor . Endocrinology 144:2055–2067
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
Merchenthaler I, Lane MV, Numan S, Dellovade TL 2004 Distribution of estrogen receptor and ? in the mouse central nervous system: in vivo autoradiographic and immunocytochemical analyses. J Comp Neurol 473:270–291
Gould E, Woolley CS, Frankfurt M, McEwen BS 1990 Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. J Neurosci 10:1286–1291
Murphy DD, Segal M 1996 Regulation of dendritic spine density in cultured rat hippocampal neurons by steroid hormones. J Neurosci 16:4059–4068
Woolley CS, McEwen BS 1994 Estradiol regulates hippocampal dendritic spine density via an N-methyl-D-aspartate receptor-dependent mechanism. J Neurosci 14:7680–7687
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
Gazzaley AH, Weiland NG, McEwen BS, Morrison JH 1996 Differential regulation of NMDAR1 mRNA and protein by estradiol in the rat hippocampus. J Neurosci 16:6830–6838
Dubal DB, Zhu H, Yu J, Rau SW, Shughrue PJ, Merchenthaler I, Kindy MS, Wise PM 2001 Estrogen receptor , not ?, is a critical link in estradiol-mediated protection against brain injury. Proc Natl Acad Sci USA 98:1952–1957
Carswell HV, Macrae IM, Gallagher L, Harrop E, Horsburgh KJ 2004 Neuroprotection by a selective estrogen receptor ? agonist in a mouse model of global ischemia. Am J Physiol Heart Circ Physiol 287:H1501–H1504
Stauffer SR, Coletta CJ, Tedesco R, Nishiguchi G, Carlson K, Sun J, Katzenellenbogen BS, Katzenellenbogen JA 2000 Pyrazole ligands: structure-affinity/activity relationships and estrogen receptor--selective agonists. J Med Chem 43:4934–4947
Dorling AA, Todman MG, Korach KS, Herbison AE 2003 Critical role for estrogen receptor in negative feedback regulation of gonadotropin-releasing hormone mRNA expression in the female mouse. Neuroendocrinology 78:204–209
Dubal DB, Shughrue PJ, Wilson ME, Merchenthaler I, Wise PM 1999 Estradiol modulates bcl-2 in cerebral ischemia: a potential role for estrogen receptors. J Neurosci 19:6385–6393
Toran-Allerand CD, Guan X, MacLusky NJ, Horvath TL, Diano S, Singh M, Connolly Jr ES, Nethrapalli IS, Tinnikov AA 2002 ER-X: a novel, plasma membrane-associated, putative estrogen receptor that is regulated during development and after ischemic brain injury. J Neurosci 22:8391–8401
Chen J, Xu W, Jiang H 2001 17?-Estradiol protects neurons from ischemic damage and attenuates accumulation of extracellular excitatory amino acids. Anesth Analg 92:1520–1523
Jover T, Simionescu M, Yokota H, Etgen AM, Zukin RS 2002 Estrogen acts via estrogen receptors and insulin-like growth factor-I receptors to protect against global ischemia-induced neuronal death. Soc Neurosci Abstr 28:6
Sudo S, Wen TC, Desaki J, Matsuda S, Tanaka J, Arai T, Maeda N, Sakanaka M 1997 ?-Estradiol protects hippocampal CA1 neurons against transient forebrain ischemia in gerbil. Neurosci Res 29:345–354
Horsburgh K, Macrae IM, Carswell H 2002 Estrogen is neuroprotective via an apolipoprotein E-dependent mechanism in a mouse model of global ischemia. J Cereb Blood Flow Metab 22:1189–1195
Sampei K, Goto S, Alkayed NJ, Crain BJ, Korach KS, Traystman RJ, Demas GE, Nelson RJ, Hurn PD2000 Stroke in estrogen receptor--deficient mice. Stroke 31:738–743; discussion 744
Sawada M, Alkayed NJ, Goto S, Crain BJ, Traystman RJ, Shaivitz A, Nelson RJ, Hurn PD 2000 Estrogen receptor antagonist ICI182,780 exacerbates ischemic injury in female mouse. J Cereb Blood Flow Metab 20:112–118
Harada H, BharwaniI S, Pavlick KP, Korach KS, Grisham MB 2004 Estrogen receptor-, sexual dimorphism and reduced-size liver ischemia and reperfusion injury in mice. Pediatr Res 55:450–456
Harris HA, Katzenellenbogen JA, Katzenellenbogen BS 2002 Characterization of the biological roles of the estrogen receptors, ER and ER?, in estrogen target tissues in vivo through the use of an ER-selective ligand. Endocrinology 143:4172–4177
Wang Q, Santizo R, Baughman VL, Pelligrino DA, Iadecola C 1999 Estrogen provides neuroprotection in transient forebrain ischemia through perfusion-independent mechanisms in rats. Stroke 30:630–637
Ogawa S, Eng V, Taylor J, Lubahn DB, Korach KS, Pfaff DW 1998 Roles of estrogen receptor- gene expression in reproduction-related behaviors in female mice. Endocrinology 139:5070–5081
Ogawa S, Chan J, Chester AE, Gustafsson JA, Korach KS, Pfaff DW 1999 Survival of reproductive behaviors in estrogen receptor ? gene-deficient (?ERKO) male and female mice. Proc Natl Acad Sci USA 96:12887–12892
Rissman EF, Early AH, Taylor JA, Korach KS, Lubahn DB 1997 Estrogen receptors are essential for female sexual receptivity. Endocrinology 138:507–510
Wersinger SR, Haisenleder DJ, Lubahn DB, Rissman EF 1999 Steroid feedback on gonadotropin release and pituitary gonadotropin subunit mRNA in mice lacking a functional estrogen receptor . Endocrine 11:137–143
Geary N, Asarian L, Korach KS, Pfaff DW, Ogawa S 2001 Deficits in E2-dependent control of feeding, weight gain, and cholecystokinin satiation in ER- null mice. Endocrinology 142:4751–4757
Couse JF, Yates MM, Walker VR, Korach KS 2003 Characterization of the hypothalamic-pituitary-gonadal axis in estrogen receptor (ER) null mice reveals hypergonadism and endocrine sex reversal in females lacking ER but not ER?. Mol Endocrinol 17:1039–1053
Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, Sar M, Korach KS, Gustafsson JA, Smithies O 1998 Generation and reproductive phenotypes of mice lacking estrogen receptor ?. Proc Natl Acad Sci USA 95:15677–15682
Prange-Kiel J, Wehrenberg U, Jarry H, Rune GM 2003 Para/autocrine regulation of estrogen receptors in hippocampal neurons. Hippocampus 13:226–234
Amateau SK, Alt JJ, Stamps CL, McCarthy MM 2004 Brain estradiol content in newborn rats: sex differences, regional heterogeneity, and possible de novo synthesis by the female telencephalon. Endocrinology 145:2906–2917
Hojo Y, Hattori TA, Enami T, Furukawa A, Suzuki K, Ishii HT, Mukai H, Morrison JH, Janssen WG, Kominami S, Harada N, Kimoto T, Kawato S 2004 Adult male rat hippocampus synthesizes estradiol from pregnenolone by cytochromes P45017 and P450 aromatase localized in neurons. Proc Natl Acad Sci USA 101:865–870
Kretz O, Fester L, Wehrenberg U, Zhou L, Brauckmann S, Zhao S, Prange-Kiel J, Naumann T, Jarry H, Frotscher M, Rune GM 2004 Hippocampal synapses depend on hippocampal estrogen synthesis. J Neurosci 24:5913–5921
Takahashi N, Tonchev AB, Koike K, Murakami K, Yamada K, Yamashima T, Inoue M 2004 Expression of estrogen receptor-? in the postischemic monkey hippocampus. Neurosci Lett 369:9–13
Simons Jr S 2003 The importance of being varied in steroid receptor transactivation. Trends Pharmacol Sci 24:253–259
Sun J, Huang YR, Harrington WR, Sheng S, Katzenellenbogen JA, Katzenellenbogen BS2002 Antagonists selective for estrogen receptor . Endocrinology 143:941–947(Nora R. Miller, Teresa Jo)
Address all correspondence and requests for reprints to: Anne M. Etgen, Ph.D., Albert Einstein College of Medicine, Department of Neuroscience, 1300 Morris Park Avenue, Forchheimer 113, Bronx, New York 10461. E-mail: etgen@aecom.yu.edu.
Abstract
Estradiol at physiological concentrations intervenes in apoptotic death cascades and ameliorates neuronal death in experimental models of focal and global ischemia. The cellular targets that mediate estradiol protection of hippocampal neurons in global ischemia are, however, unclear. The present study examined the hypothesis that estradiol protects hippocampal neurons in ovariectomized rats via estrogen receptor (ER) and/or ?. Estradiol (14 d pretreatment) afforded robust protection of CA1 neurons against global ischemia-induced death. The broad-spectrum ER antagonist ICI 182,780 (intracerebroventricularly, 0 and 12 h after ischemia) abolished estrogen protection, consistent with a role for ERs. To evaluate the potential roles of ER vs. ER? in estrogen protection, we administered subtype-selective agonists for 14 d before and 7 d after ischemia. The ER-selective agonist propyl pyrazole triol (PPT, 10 mg/kg) and ER?-selective agonist WAY 200070–3 (1 mg/kg) produced nearly complete protection of CA1 neurons in approximately 50% of the animals. PPT, but not WAY 200070–3, at doses used for protection, elicited lordosis, induced negative feedback inhibition of LH release, and reduced weight gain. These findings establish the efficacy of the PPT dose in neuroendocrine assays and specificity of WAY 200070–3 for ER?. We also examined the ability of estradiol and neuronal injury to regulate ER and ER? expression. Both estradiol and global ischemia markedly increased ER, but not ER?, protein in CA1. These data indicate that estradiol can act via ER and ER? to protect CA1 neurons from global ischemia-induced death and that both estradiol and global ischemia modulate ER expression in hippocampal CA1.
Introduction
THE BENEFITS OF postmenopausal estrogen replacement therapy for protecting the brain against the neuronal death associated with neurological diseases and disorders remain controversial. Global ischemia, arising during cardiac arrest or surgery in humans or induced experimentally in animals, elicits selective, delayed neuronal death; pyramidal neurons of the hippocampal CA1 are particularly vulnerable (1, 2). The ischemia-induced loss of CA1 neurons is associated with cognitive deficits. 17?-Estradiol (E2), the primary estrogen produced and secreted by the ovaries, and other estrogens afford neuroprotection in experimental models of global and focal ischemia (3, 4, 5, 6, 7, 8, 9). Recent work from this laboratory demonstrates that E2 at levels considered physiological for postmenopausal women prevents the activation of apoptotic signaling cascades and ameliorates global ischemia-induced neuronal loss in male gerbils (7). However, the cellular targets that mediate the protective actions of E2 on hippocampal neurons are, as yet, unclear.
Estrogen exerts a number of physiological actions via interaction with intracellular estrogen receptors (ERs), which serve as ligand-activated transcription factors (reviewed in Refs. 3 , 6 , 10 , and 11). Both of the known ER subtypes, ER and ER?, are expressed in hippocampal neurons (12, 13, 14, 15). E2 has widespread actions on the brain. In the hippocampus, E2 regulates spine density (16, 17), synapse number (18), the synthesis of neurotrophic factors (3, 6, 19), and N-methyl-D-aspartate receptor expression (18, 20). Studies involving mice with targeted deletions of ER and ER? indicate that ER mediates E2 protection of cortical neurons in an animal model of focal ischemia (21). However, ER? is implicated in the neuroprotective actions of E2 in mice subjected to global ischemia (22), and agonists selective for both ER and ER? can partially protect gerbil CA1 neurons from global ischemia-induced cell death (9).
The purpose of the present study was: 1) to determine whether the neuroprotective effects of E2 in a global ischemia model, which produces selective, delayed death of CA1 pyramidal neurons in rats, is mediated by ERs; 2) to identify which ER mediates neuroprotection in this model; and 3) to determine the ability of E2 and neuronal injury (global ischemia) to regulate expression of ER and ER? in the hippocampal CA1. We show that the broad-spectrum ER antagonist ICI 182,780 abolishes the neuroprotective effects of E2 in ovariohysterectomized (ovx) female rats subjected to global ischemia. To validate the dose of the ER-selective agonist propyl pyrazole triol (PPT) in eliciting ER-mediated neuroendocrine responses and the specificity of the ER? WAY 200070–3, we performed independent physiological tests. At the doses chosen for the ischemia studies, PPT, but not WAY 200070–3, facilitated female reproductive behavior, suppressed LH release, and minimized body weight gain. Both the ER and ER? agonists afforded protection against global ischemia-induced loss of hippocampal CA1 neurons in approximately 50% of animals. Estrogen pretreatment and global ischemia increased ER, but not ER?, protein expression in the CA1. These findings indicate that ERs mediate the survival of CA1 neurons afforded by E2 in global ischemia and suggest a role for both ER and ER? in protection against global ischemia-induced cell death.
Materials and Methods
Animals and experimental treatments
All procedures were performed in accordance with the National Institutes of Health (NIH) Guidelines for the Care and Use of Laboratory Animals and were approved by the Albert Einstein College of Medicine Institutional Animal Care and Use Committee. Rodents were maintained in a temperature- and light-controlled environment with 14-h light, 10-h dark cycle and had access to standard rat chow and water ad libitum.
Ovariohysterectomy and E2 pellet placement
Twenty-one-day-old female Sprague Dawley rats (Charles River Laboratories, Inc., Wilmington, MA) were ovx on day zero, under halothane (3% for induction, 1% for maintenance) anesthesia. On the same day, pellets containing E2 (50 μg; 21-d sustained release; Innovative Research of America, Sarasota, FL) or placebo were inserted sc beneath the dorsal surface of the neck. These pellets are designed to maintain serum hormone levels for 21 d. Experiments were carried out so that animals weighed approximately 100–120 g at the time global ischemia was induced, because the four-vessel occlusion (see below) produces the most reliable ischemia in rats of this size.
ER-selective agonists and ER antagonist compounds
For experiments using ER-selective compounds, ovx rats received daily sc injections of: 1) the ER agonist PPT (provided by Dr. Istvan Merchenthaler, Wyeth-Ayerst Laboratories, Inc., Collegeville, PA) at a dose of 10 mg/kg; 2) an ER? agonist WAY 200070–3 (provided by Dr. Heather Harris, Wyeth-Ayerst Laboratories, Inc.) at a dose of 1 mg/kg; or 3) vehicle, 25% dimethylsulfoxide in 0.9% saline, for a total vol of 200 μl. Agonists were always injected between 0700 and 0800 h. On the day of ischemia induction, the injections occurred between 2 and 4 h before surgery. For intracerebroventricular (ICV) injections: 1) 100 μg of the nonselective ER antagonist ICI 182,780 (Tocris Cookson, Inc., Ellisville, MO), or 2) vehicle (50% dimethylsulfoxide, in 0.9% saline) was injected into the right lateral ventricle, in a vol of 5 μl per injection, with a Hamilton syringe (Fisher Scientific, Pittsburgh, PA) using standard stereotaxic methods. The structures of these compounds are shown in Fig. 1.
FIG. 1. Structure of the ER-selective agonists and antagonist. A, ER-selective agonist, PPT (23 ). B, ER?-selective agonist, WAY-200070–3 (WAY) (34 ). C, ER antagonist ICI 182,780 (49 ).
ER-selective agonist treatment
For experiments using ER-selective agonists, PPT (ER agonist), WAY 200070–3 (ER? agonist), or vehicle was administered once daily from 24 h after ovx until d 21, when rodents were euthanized for histology. Body weights were measured at the time of ovx and then every 2–3 d to determine the correct dosage of the ER-selective agonists and to assess ER agonist effects on weight gain. PPT is 410-fold selective for ER relative to ER? (23). WAY 200070–3 is 110-fold selective for ER?, relative to ER, as assessed by radioligand binding assays, and it fails to stimulate uterine weight gain when administered sc at doses as high as 50 mg/kg (H. Harris, personal communication). Although the ability of WAY 200070–3 to cross the blood-brain barrier is not well characterized, upon sc administration of 3 mg/kg WAY 200070–3, a concentration of 12 ng/g brain tissue is achieved by 1 h (H. Harris, personal communication).
Global ischemia
On d 13, using halothane anesthesia, the vertebral arteries were exposed through a midline occipital-suboccipital incision and coagulated with bipolar cauterization between the first and second cervical vertebral bodies. This procedure by itself has no effect on cerebral blood flow but prevents collateral circulation to the forebrain during transient carotid artery occlusion on d 14. Twenty-four hours later, transient global ischemia was accomplished by bilateral occlusion of the common carotid arteries for 10 min followed by reperfusion. A rectal probe was inserted to monitor core temperature and was maintained at 36.5–37.5 C using a heat lamp during ischemia. Sham-operated rats had their vertebral arteries coagulated on d 13.
ICV injections
In the first experiment, immediately after ischemia (24 h after coagulating the vertebral arteries for sham-treated rats) and again 12 h later, rats were positioned in a stereotaxic apparatus and ICV injections performed under halothane anesthesia. The position of the right lateral ventricle was calculated based on the position of bregma (anterior/posterior, 0.92 mm; medial/lateral, 1.2 mm; dorsal/ventral, 3.6 mm), and then the nonselective ER antagonist ICI 182,780 or vehicle was infused with a 28-gauge needle, in a vol of 5 μl per infusion, over 1 min. The injection needle was left in place for an additional 1 min before being withdrawn.
Histological analysis
We first examined the impact of transient global ischemia on the volume of the hippocampal CA1. Rats were subjected to global ischemia or sham operation (n = 6 per group) and, at 7 d after surgery, were killed by transcardiac perfusion with 4% paraformaldehyde under deep anesthesia. Brains were rapidly removed, and the volume of the CA1 from ischemic and sham rats was assessed by stereological methods. Toluidine blue-stained coronal sections (20 μm) were collected through the entire dorsal hippocampus (bregma –2.3 to –4.5 mm). Digital images of every tenth section from each animal (100 sections per brain) were captured and used to trace the outline of the CA1. The mean area ± SEM of the entire CA1 was calculated and multiplied by the section thickness times the number of sections along the rostral/caudal axis. Although the CA1 pyramidal neurons were nearly ablated at 7 d after ischemia, no statistically significant difference in volume of the CA1 was detected for ischemic (2.43 ± 0.6 x 108 μm3; n = 6) vs. sham-operated rats (2.42 ± 0.5 x 108 μm3; n = 6). Based on this information, counts of pyramidal neurons were assessed in a 250-μm length of the CA1 pyramidal cell layer, in 3–4 sections per animal, at 7 d after ischemia or sham operation as described below.
Seven days after global ischemia (d 21 after ovx), animals were euthanized by deep anesthesia, and blood was collected by cardiac puncture for analysis of serum E2 levels. Each rat was transcardially perfused using 0.9% saline solution with heparin (150 ml; 15 min) followed by ice-cold 10% buffered formalin phosphate (200 ml; 20 min; Fisher Scientific, Pittsburgh, PA). Brains were then removed, placed in formalin at 4 C overnight, fixed in 30% sucrose in PBS at 4 C for 48 h, and then frozen at –40 C. The dorsal hippocampus was coronally sectioned on a cryotome into 15-μm slices, and every fourth section was collected, mounted, and stained with toluidine blue. Using x60 magnification, the CA1 region of the left and right hippocampi was photographed. Surviving pyramidal neurons in a 250-μm length of the stratum pyramidale of the left and right CA1 (see arrows in Fig. 2B, panel A) were counted. Cell counts are expressed as the average number of surviving neurons per side, counted from four sections per animal. More than 95% of the cells in the stratum pyramidale are pyramidal neurons; most glia and the cell bodies of inhibitory interneurons are localized to the stratum radiatum or stratum oriens. It is well established that a brief (10 min) episode of global ischemia affords delayed, selective death of CA1 pyramidal neurons. Inhibitory interneurons of the CA1 and all neurons of the nearby CA2 or transition zone, CA3, and dentate gyrus survive (2).
FIG. 2. ER antagonist ICI 182,780 treatment abolishes E2 neuroprotection after global ischemia. All groups were treated with E2 or placebo (P) for 2 wk before undergoing ischemia or sham surgery. ICI 182,780 (ICI) or vehicle (V) was then injected ICV immediately after ischemia and repeated 12 h later (for sham groups, ICI 182,780 or vehicle was injected 24 h and 36 h after sham operation). A, E2 treatment affords robust neuroprotection in the CA1 region of the hippocampus after global ischemia (E2+V vs. P+V in ischemic animals; #, P < 0.002). Injection of the ER antagonist ICI 182,780 after E2 treatment and ischemia abolished this neuroprotection (E2+ICI vs. E2+V in ischemic animals: , P < 0.002). Treatment with the ER antagonist did not affect the number of surviving neurons in any of the other treatment groups; therefore, for statistical analysis, sham-operated vehicle and ICI 182,780-treated groups were combined. Rats that underwent global ischemia had significantly fewer surviving neurons than the sham-operated groups (*, P < 0.001). Data (n = 3–10 per treatment group) are represented as medians and analyzed using Kruskal Wallis with Dunn’s post hoc analysis. B, Histologic evaluation of the dorsal hippocampus after global ischemia or sham operation in low (x4, top) and high power (x60, bottom) magnification. DG, Dentate gyrus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Placebo (A and B) and E2 (C and D)-treated rats that underwent sham operation similarly have many surviving neurons in the CA1 region of the hippocampus. Placebo-treated rats that underwent global ischemia (E and F) have selective, delayed neuronal death in CA1. G and H, E2-treated rats that underwent global ischemia have selective, delayed neuronal death in CA1, but approximately 50% of the neurons survived. I and J, E2-treated rats that underwent global ischemia and then received ICI 182,780 have few surviving neurons in the CA1 region. C, Rats treated with E2 have higher serum E2 levels than the placebo (P)-treated group. Sera were collected 15 d (E2 15d) or 21 d (E2 21d) after E2 pellets were placed. Statistically significant differences between hormonal treatments and time are shown (E2 15d vs. P: *, P < 0.001; E2 21d vs. P: , P < 0.002; E2 15d vs. E2 21d: *, P < 0.001). Data (n = 20–33 per treatment group) are represented as mean ± SE and analyzed using one-way ANOVA.
Behavioral testing
Adult female Sprague Dawley rats, weighing 200–350 g (Taconic Farms, Inc., Germantown, NY), were ovx and maintained in a reverse light-dark cycle (14-h dark, 10-h light, with lights off at 1100 h). Two weeks later, rats (n = 8) were randomly assigned to receive E2 benzoate (2 μg in 100 μl peanut oil) or an ER-selective agonist sc daily for 2 d. The PPT dose was 10 mg/kg·d, and the WAY 200070–3 dose was 1 mg/kg·d. On the third day, progesterone (500 μg in 100 μl peanut oil) was administered, and behavior testing was performed 4 h later. All tests were performed in the dark, under red light illumination, between 1300 and 1700 h. Each female rat was placed into a cage with a sexually active male rat, and female sexual receptivity was recorded. The lordosis quotient (LQ) was determined by counting the number of times the female assumed the lordosis posture, in response to 10 mount attempts by the male, and then multiplied by 100. One week later, the same female rats were injected with a different ER-selective agonist or E2 benzoate, and behavior testing was repeated. Testing was continued until all female rats were exposed to each treatment.
Serum E2 assay
Tubes containing whole blood were placed on ice for 20 min and centrifuged at 300 x g for 5 min, and sera were collected and stored (–20 C) until analyzed. Serum hormone levels were measured by fluoroimmunoassay using the DELPHIA E2 assay (PerkinElmer Life Sciences, Turku, Finland). All assays were performed in duplicate, and the mean value was reported. The sensitivity of detection was 13 pg/ml. The inter- and intraassay coefficients of variance were 10.1% and 4.1%, respectively.
LH assay
Serum LH levels were measured by enzyme immunoassay (Biotrak; Amersham Biosciences, Buckinghamshire, UK). Assays were performed in duplicate, and the mean value was reported. The sensitivity of detection is 0.1 ng/ml. The cross-reactivity of LH with rat FSH was less than 0.016%. The inter- and intraassay coefficients of variance are 6.2% and 2.4%, respectively.
Western blot
For quantification of protein in the CA1 hippocampal region, immunoblots were performed as described (7). The ovx animals received placebo or E2 pellets as described above and were subjected to sham operation or 10-min global ischemia 2 wk later. Animals were deeply anesthetized and euthanized by decapitation at 24 h and 48 h after ischemia or 24 h after sham operation. Before decapitation, blood samples were collected by cardiac puncture from animals killed 24 h after ischemia or sham operation, for later analysis of serum E2 levels on d 15. Hippocampi were quickly dissected and cut into 1-mm transverse slices on a McIllwain tissue chopper (Vibratone Co., St. Louis, MO). The entire CA1 region from these slices of the left and right hippocampus was separated quickly by microdissection, placed in ice-cold PBS containing the protease inhibitor phenylmethylsulfonylfluoride (1 mM; Sigma- Aldrich Corp., St. Louis, MO), and stored at –20 C until used. Protein concentration of samples was determined using the BCA protein assay kit (Pierce Chemical Co., Rockford, IL). Laemmli’s buffer was combined with 30 μg protein and loaded onto 10% polyacrylamide minigels (Bio-Rad Laboratories, Inc., Richmond, CA) and then subjected to electrophoresis.
Protein bands were transferred onto a polyvinylidene fluoride membrane (Millipore Corp., Bedford, MA). After blocking for 30 min with 25 mM Tris-HCl (pH 8.0) containing 125 mM NaCl, 0.1% Tween 20, and 4% skim milk, membranes were incubated with primary antibody, washed, and then incubated with secondary antibody. After reaction, membranes were treated with enhanced chemiluminescence reagents (Amersham Biosciences, Arlington Heights, IL) and apposed to x-ray film (Eastman Kodak Co., Rochester, NY).
Antibodies for Western blot
Primary antibodies were: 1) a mouse monoclonal antibody directed to human ER (ER 6F11; Vector Laboratories Ltd., Burlingame, CA) at a concentration of 1:100 and incubated at 4 C overnight, and 2) a polyclonal rabbit antibody directed to amino acids 64–82 of rat ER? (ER? antibody; provided by Dr. Stephen Alves, Merck & Co., Inc., Rahway, NJ) at a concentration of 1:30,000 and incubated at room temperature for 16 h. Both of these antibodies have been validated for use in immunoblotting of rat brain tissue. Secondary antibodies were: 1) a horseradish peroxidase-conjugated antimouse IgG for ER (Cell Signaling Technology, Inc., Beverly, MA) at a concentration of 1:1000 and incubated for 2 h at room temperature; and 2) horseradish peroxidase-conjugated antirabbit IgG for ER? (Cell Signaling Technology, Inc.) at a concentration of 1:2,000 and incubated for 1.5 h at room temperature.
A Scan Jet 4-C computing densitometer (Hewlett-Packard Co., Palo Alto, CA) was used with NIH Image 1.61 software to quantitate protein abundance. Optical densities of ER bands were first normalized to the amount of protein loaded, assessed by stripping and reblotting for type III ?-tubulin in each sample. These values were then expressed relative to normalized values obtained for samples from ovx, sham-operated rats on the same membrane to allow for comparison of band densities of immunoblots exposed to different pieces of film.
Statistical analysis
Data analysis was performed using SPSS 12.0 (SPSS, Inc., Chicago, IL). One-way ANOVA was used to compare group means of normally distributed data (serum E2, serum LH, rat weight gain, and immunoblots) with Newman-Keuls post hoc analysis. Friedman’s test with Dunn’s multiple-comparison post hoc analysis was performed to analyze lordosis behavior tests. Kruskal Wallis testing was used for comparing neuron counts after E2 and/or ICI 182,780 treatment, followed by Dunn’s post hoc analysis. Other nonparametric methods were used to analyze pyramidal neuron counts after ER subtype-selective agonist treatment, because the data were bimodal. Subjects were divided into two categories based on the number of surviving neurons. Animals with greater than 40 surviving neurons per 250 μm were considered to be protected, and animals with less than 30 surviving neurons per 250 μm were considered not to be protected. Then 2 and Fisher’s exact test were used to analyze whether there were treatment-related differences in the number of animals protected and not protected and in the median number of surviving neurons. Differences were considered significant at P 0.05.
Results
To examine the cellular targets that mediate the neuroprotective actions of estrogen on hippocampal neurons, we treated rats with E2 or placebo for 14 d, subjected animals to sham operation or global ischemia, and acutely administered the broad-spectrum ER antagonist ICI 182,780 or vehicle at 0 and 12 h after reperfusion. E2 and placebo pellets remained in place until rats were killed for histological evaluation 7 d after ischemia. Global ischemia induced a substantial loss of hippocampal CA1 neurons in vehicle-treated rats at 7 d post ischemia (loss of 80% of neurons in the CA1 pyramidal cell layer; Fig. 2A). Long-term E2 treatment afforded robust neuroprotection (50% neuronal survival in the CA1) in nine of nine treated rats. The ER antagonist ICI 182,780 alone (injected at 0 and 12 h after reperfusion) did not detectably alter the number of surviving neurons in sham-operated or ischemic rats (Fig. 2A) but abrogated E2-elicited neuroprotection (Fig. 2B). These findings indicate that E2 acts via the classical ERs to elicit protection of hippocampal neurons.
To confirm that estrogen pellet implantation effectively released E2, we monitored serum E2 levels. Groups treated with E2 exhibited significantly higher serum E2 levels than did placebo-treated rats, evident at 15 and 21 d after pellet implantation (Fig. 2C). The mean serum E2 level on d 15 (64 ± 7.0 pg/ml), measured in animals killed 24 h after ischemia for the ER immunoblotting study (see below), was slightly above the peak physiological level seen during proestrus. By d 21, the mean serum E2 level (33 ± 2.9 pg/ml) was in the midphysiological range. The change between d 15 and d 21 most likely reflects the increasing body weight of the animals (see Fig. 3B) and may also indicate that, by the end of the experiment, the 21-d pellets have released most of their E2.
FIG. 3. ER-selective agonist treatment and physiologic responses. A, Treatment with the ER-selective agonist PPT induces lordosis behavior. Two weeks after ovx, the ER agonist (PPT), the ER? agonist (WAY), or E2 benzoate (E2) was administered for two sequential days. Progesterone was given on the third day, and behavioral testing was performed 4 h later. WAY vs. E2: *, P < 0.001; WAY vs. PPT: , P < 0.05. Data are expressed as medians (n = 8) and analyzed using Friedman test. B, Treatment with the ER-selective agonist PPT reduces rat weight gain. Total weight gain determined after 21 d of daily ER agonist or vehicle treatment. PPT vs. V or WAY: *, P < 0.001. Data (n = 17–19 per treatment group) are expressed as mean ± SE and analyzed using ANOVA. C, ER agonist treatment causes feedback inhibition of LH. Sera were collected after 21 d of daily hormone administration. PPT vs. V: , P < 0.02; PPT vs. WAY: *, P < 0.001). Data (n = 16–19 per treatment group) are expressed as mean ± SE and analyzed using ANOVA.
To evaluate the potential roles of ER vs. ER? in mediating estrogen protection of hippocampal neurons, we administered ER- and ER?-selective agonists to ovx rats for 14 d before and 7 d after ischemia. To validate the doses of the agonists used for protection, and to evaluate the specificity of the agonists for the ER subtypes, we tested known actions of estrogen mediated entirely (lordosis behavior) or at least in part (suppression of LH release and weight gain) at the level of the brain. In a pilot study, the ER agonist PPT at 2 mg/kg·d did not elicit a lordosis response; rats exhibited a median LQ of zero (interquartile range: 0, 15; n = 8). PPT at a higher dose (10 mg/kg) elicited a high median LQ similar to that produced by administration of E2 benzoate (Fig. 3A). This finding verifies that PPT at 10 mg/kg crosses the blood-brain barrier and achieves a concentration in brain tissue sufficient to produce neuroendocrine responses mediated by ER at the level of the hypothalamus within 1–2 d after administration. The ER?-selective agonist WAY 200070–3 did not elicit lordosis at either 1 mg/kg (Fig. 3A) or 10 mg/kg (median LQ = 10, interquartile range: 5, 15; n = 8), suggesting that it does not cross-react detectably with ER to elicit sexual behavior in this dose range.
The ER-selective agonist PPT at 10 mg/kg·d, but not WAY 200070–3 (1 mg/kg·d), also reduced weight gain by young rats during the 3-wk regimen used in the ischemia studies (Fig. 3B). Control (vehicle-treated) rats gained an average of 92 g during the 21 d, whereas PPT-treated rats (10 mg/kg·d) gained an average of only 59 g. These PPT-treated rats differed significantly from both the vehicle- and WAY 200070–3-treated groups (P < 0.01). Rats treated with 2 mg/kg·d PPT also gained significantly more weight than did those treated with 10 mg/kg PPT (P < 0.002). The weight gained by rats treated with the lower ER agonist dose (data not illustrated) did not differ significantly from either that of the vehicle- or of the ER?-agonist-treated groups. Within a given agonist treatment group, there was no significant difference in weight gain for ischemic vs. sham-operated rats.
Estrogen also elicits negative feedback inhibition of LH release, at least in part via ER (24). Administration of the ER-selective agonist PPT (2 mg/kg·d) reduced the median LH serum level in ovx rats from approximately 15 ng/ml to 4.6 ng/ml (interquartile range: 2.4, 9.0; P < 0.02 vs. vehicle-treated rats), consistent with negative feedback inhibition of LH release. Increasing the dose of the ER agonist to 10 mg/kg·d reduced the mean serum LH level further (Fig. 3C). In contrast, WAY 200070–3 did not detectably alter serum LH levels relative to ovx, vehicle-treated females (Fig. 3C).
Both the ER- (10 mg/kg·d) and ER?- (1 mg/kg·d) selective agonists elicited pronounced neuroprotection in approximately 40–50% of treated ischemic rats. Both groups exhibited bimodal distributions in the number of surviving neurons in the hippocampal CA1. Approximately 40–50% of the rats treated with each agonist showed virtually complete survival of CA1 pyramidal neurons; the remainder of the rats exhibited approximately10% surviving neurons, a result statistically indistinguishable from that of the vehicle-treated ischemia group (Fig. 4). The behaviorally ineffective dose of PPT, 2 mg/kg, elicited modest protection in only two of seven rats subjected to global ischemia (data not shown). Interestingly, these two rats gained less weight during the experiment (43.5 ± 9.1 g) than did the four rats that showed extensive loss of CA1 pyramidal neurons (83.6 ± 11.3 g). Raising the dose of WAY 200070–3 to 10 mg/kg did not significantly improve the outcome relative to that of the lower dose (one of five rats showed increased survival of CA1 neurons relative to vehicle controls). Neither ER agonist induced detectable pyramidal cell loss in CA1 when given alone to sham-operated rats.
FIG. 4. ER (PPT, 10 mg/kg·d) and ER? (WAY, 1 mg/kg·d) agonists elicit neuroprotection in some rats after global ischemia. Rats were injected daily with vehicle (V), PPT, or WAY for 2 wk before undergoing global ischemia or sham surgery. Hormonal treatments were continued until rats were euthanized 7 d after ischemia or sham operation. Sham-operated rats are represented by open symbols, and ischemic rats are represented by filled symbols. The median number of surviving pyramidal cells for each treatment group is demonstrated by a horizontal line, and differences between groups were analyzed by Fisher’s exact test. (ischemia+V vs. sham+V: ***, P < 0.001; PPT+ischemia vs. PPT+sham: *, P < 0.04; WAY+ischemia vs. WAY+sham: **, P < 0.02). Differences in the number of rats per group with more than 40 surviving neurons after drug treatment and ischemia were analyzed using 2 (ER?+ischemia vs. V+ischemia: , P = 0.04; ER+ischemia vs. V+ischemia: #, P = 0.05).
To assess whether the bimodal distribution of CA1 pyramidal neuron survival was related to individual differences in accessibility of brain tissues to 10 mg/kg PPT, we examined serum LH levels in protected (n = 5) and nonprotected (n = 8) rats. Mean LH levels were essentially identical in the two groups (protected, 2.35 ± 0.8 ng/ml vs. nonprotected, 2.25 ± 0.8 ng/ml). Likewise, there was no significant difference between LH levels and pyramidal cell survival in the CA1 of rats treated with 1 mg/kg WAY 200070–3 (protected, 27.0 ± 15.6 ng/ml, n = 5; nonprotected, 19.4 ± 7.2 ng/ml, n = 6). We also detected no difference in weight gain during the 3 wk of agonist administration in PPT-treated females that were protected (47.8 ± 14.6 g) and not protected (57.8 ± 7.1 g). The same was true for WAY 200070–3 (protected, 90.2 ± 22.1 g vs. nonprotected, 93.3 ± 20.3 g). Moreover, linear regression analysis revealed no correlation between plasma LH levels or weight gain and pyramidal neuron counts in CA1 of PPT- or WAY 200070–3-treated rats (data not illustrated).
Previous studies indicate that neuronal injury, including focal ischemia, can modify ER expression levels (25, 26). Therefore, we determined the abundance of ER and ER? protein in the CA1 region of the hippocampus of ovx rats treated with placebo or E2 pellets for 14 d before global ischemia or sham operation. Global ischemia induced a significant increase in ER protein expression, evident at 24 and 48 h after ischemia in the placebo-treated groups relative to placebo-treated, sham-operated rats (Fig. 5). E2 pretreatment alone also increased ER protein levels in CA1 in sham-operated rats. The combination of E2 pretreatment and ischemia did not increase ER protein levels above that produced by either manipulation alone. Immunoblots for ER in other hippocampal subfields indicated that neither ischemia nor estrogen changed ER levels in the dentate gyrus (data not illustrated). In CA3, ischemia, but not estrogen, modestly increased ER at both 24 and 48 h. Estrogen prevented the ischemia-induced increase at 24 h but not 48 h (data not illustrated). Although global ischemia induced a trend toward down-regulation of ER? protein in CA1, this effect was not significant at either time (Fig. 6; ANOVA, P < 0.06). Likewise, E2 pretreatment did not significantly alter ER? protein abundance in the CA1 of either ischemic females or sham-operated rats.
FIG. 5. ER protein expression in ovx rats is up-regulated in CA1 by E2 and after global ischemia. Rats were treated with E2 pellets or placebo (P) for 2 wk before undergoing global ischemia or sham surgery. Rats were euthanized 24 h (24 ) or 48 h (48 ) after ischemia (I) or 0 h for sham operation (S). A, Representative immunoblots of ER (top) and ?-tubulin (bottom) in hippocampal CA1 from rats treated under each experimental condition. B, Quantitative analysis of ER after ischemia. ER is up-regulated at 24 h and 48 h after global ischemia in placebo-treated groups (P+24 and P+48 vs. C: **, P < 0.001). ER is also up-regulated in all E2-treated groups (E2+24 and E2+S vs. S: *, P < 0.01; E2+48 vs. C: **, P < 0.001). For each treatment, ER antibody is normalized to ?-tubulin and expressed as the fold stimulation over that of placebo-treated, sham-operated rats. Values represent mean ± SEM (n = 6) analyzed using ANOVA.
FIG. 6. ER? protein expression after global ischemia. Rats were treated with E2 pellets or placebo (P) for 2 wk before undergoing global ischemia or sham surgery. Rats were euthanized 24 h (24 ) or 48 h (48 ) after ischemia (I) or at 0 h for sham operation (S). A, Representative immunoblots of ER? (top) and ?-tubulin (bottom) from rats treated under each experimental condition. B, Quantitative analysis of ER? protein levels in CA1 after ischemia in placebo- and E2-treated female rats. For each treatment, ER? immunoblots are normalized to ?- tubulin and expressed as the fold stimulation over that of placebo-treated, sham-operated rats. Values represent mean ± SEM (n = 6) and were analyzed using ANOVA.
Discussion
The ability of estrogen to afford protection in animal models of focal and global ischemia is well established (3, 4, 5, 6, 7, 8, 9). Our demonstration that E2 affords robust protection in estrogen-deprived female rats is consistent with findings of others, i.e. that estrogen affords protection against global ischemia-induced neuronal death in proestrous female rats (5) and in male and female gerbils (7, 8, 27, 28, 29) and mice (30). Our study extends those findings in that it identifies cellular targets that mediate the protective actions of E2 on rat hippocampal neurons in vivo. We demonstrate that the broad-spectrum ER antagonist ICI 182,780 (given ICV at 0 and 12 h after ischemia) abolished E2 protection, arguing against a role for ER-independent mechanisms in estrogen protection in this model. Because the ER antagonist was infused directly into the brain ventricular system, this finding also implicates brain ERs as the mediators of estrogen neuroprotection in transient global ischemia. These data extend previous work in female gerbils showing that systemic administration of the selective estrogen response modifiers, raloxifene and tamoxifen, can attenuate estrogen’s protective actions on global ischemia-induced cell death in CA1 neurons (8, 9).
We also report the novel finding that pretreatment with agonists selective for both ER and ER? can preserve a large majority of hippocampal CA1 pyramidal neurons from global ischemia-induced death in some rats. The ER-selective agonist PPT and the ER?-selective agonist WAY 200070–3 afforded marked protection of CA1 neurons in approximately 40–50% of estrogen-deprived female rats. WAY 200070–3, administered at doses that afforded neuroprotection, did not elicit lordosis behavior, negative feedback inhibition of LH release, or a reduction in weight gain, providing independent validation that the drug does not cross-react with ER in vivo. By contrast, the 10-mg/kg dose of PPT elicited all of these actions of E2 that are known to be mediated, at least in part, by brain ER receptors. We further show that global ischemia regulates ER expression in hippocampal neurons. There was a marked increase in ER protein and a small, but not significant, decrease in ER? protein in CA1 at both 24 and 48 h after ischemia. Estrogen pretreatment mimicked the injury-induced increase in ER protein expression, and the effects of ischemia and E2 pretreatment were not additive. These data indicate that E2 can act via ER and ER? to protect CA1 pyramidal neurons from global ischemia-induced death and that both E2 and global ischemia modulate ER expression in the hippocampal CA1.
The finding that estrogen can act via either ER or ER? to elicit protection of hippocampal neurons is of interest in that the cellular targets that mediate estrogen neuroprotection remain controversial. Studies involving transgenic mice with targeted deletion of either ER or ER? indicate that ER is critical to estrogen protection of cortical and striatal neurons in focal ischemia (21). Interestingly, estrogen does not protect hippocampal neurons in that model, presumably because permanent occlusion of the middle cerebral artery is a catastrophic insult. There is one report that estrogen does not afford protection against focal ischemia-induced neuronal death in mice with targeted deletion of ER (31), although ICI 182,780 administration increased infarct volume in this model (32). These investigators concluded that E2 most likely exerts neuroprotection through the activation of ER?, not ER (31, 32). Interestingly, ER has recently been implicated in the protective actions of estradiol in a model of ischemic liver injury (33).
There are also conflicting reports regarding the role of ER subtypes in neuroprotection after global ischemia. A recent pharmacological study showed that the ER?-selective agonist diaryl propiolnitrile is more effective than PPT (2 mg/kg) in protecting hippocampal neurons against global ischemia-induced death (22). However, our data clearly show that a PPT dose of 2 mg/kg·d is not sufficient to fully activate ER-mediated responses to estrogen, such as lordosis behavior, LH suppression, and reduction of weight gain, in most female rats. Our finding that PPT at a higher dose (10 mg/kg) is as effective as the selective ER? agonist WAY 200070–3 in affording neuroprotection suggests a role for both ER and ER? in estrogen protection. There is also a preliminary report that PPT and unidentified ER?-selective agonists are partially neuroprotective in gerbil CA1 after global ischemia (9). These authors used in situ hybridization signals for neurogranin mRNA to quantify CA1 cell survival, and the duration and dose of agonist treatment were not always specified, so it is somewhat difficult to make direct comparisons between that report and our study. However, in agreement with our results using 2 mg/kg·d PPT, the latter authors detected no protection in ovariectomized gerbils when PPT was used at a dose of 3 mg/kg.
ER-selective agonist treatment and neuroprotection
In our experiments, ER- and ER?-selective agonists produced a bimodal distribution of neuroprotection in rats subjected to global ischemia, with nearly complete survival of CA1 pyramidal neurons in 38 and 45% of animals, respectively, and less than 10% neuronal survival in the remainder of animals (Fig. 4). This pattern of neuroprotection was in contrast to that of E2, which elicited approximately 50% neuronal survival in all animals (Fig. 2). It is notable that, in our hands, a high dose of PPT (10 mg/kg) was necessary to prime lordosis, to suppress weight gain, and to suppress LH completely. The dose of PPT needed to produce biological responses when administered in vivo is in contrast to its high affinity in binding to ER, as assessed by radioligand assays in vitro (23). Thus, it is possible that the ability of these drugs to promote neuronal survival in only a subset of ischemic animals reflects either their low intrinsic efficacy or individual differences in drug disposition and metabolism. However, the bimodal distribution in pyramidal neuron survival is unlikely to be entirely due to individual differences in penetration of drugs through the blood-brain-barrier or to their differential action at peripheral ERs, because neither serum LH levels nor body weight gain differed in rats that exhibited preservation vs. those that exhibited loss of CA1 pyramidal neurons. Others have reported that PPT at 2 mg/kg·d prevented ovx-induced weight gain and maintained uterine weight and total and trabecular bone mineral density for as long as 6 wk in adult rats but that a higher dose (15 mg/kg·d) was required to prevent hot flashes (34). Thus, a relatively high dose of PPT may be required to elicit neural responses mediated by ER. We were unable to independently validate the ER? agonist dose used in our protection studies as physiologic, because there is currently no known behavioral or neuroendocrine response that is mediated exclusively via ER?.
Although our findings suggest that either ER or ER? can protect hippocampal neurons from ischemia-induced cell death, this conclusion must be tempered by the fact that we could not independently verify that WAY 200070–3 actually activated ER? or that PPT failed to do so in the brain or in peripheral tissues. Likewise, we cannot rule out the possibility that both receptors must be activated to rescue CA1 pyramidal cells. Although our data suggest that WAY 200070–3 did not activate brain ER, it is impossible to determine whether the high dose of PPT acted as an agonist at ER?. We attempted to use ER knockout mice to clarify the role of ER subtypes in global ischemia; however, using the standard two-vessel occlusion model of global ischemia in mice, we were unable to reliably produce selective loss of CA1 pyramidal neurons in the wild-type strains upon which the ER knockouts were produced (data not shown). Therefore, the questions raised by our work may only be resolved by the development of highly selective ER and ER? agonist and antagonist ligands that easily penetrate the blood-brain-barrier and mimic the biological activity of E2 at brain ERs.
ER-selective agonists and physiologic responses mediated by brain ERs
The present study also shows that the ER agonist PPT, but not the ER? agonist WAY 200070–3, at doses used for protection, elicits lordosis behavior and negative feedback inhibition of LH release and attenuates weight gain. These studies establish an effective dose for PPT and validate the selectivity of WAY 200070–3 in vivo. However, they do not rule out the possibility that PPT and WAY 200070–3 action on nonneuronal target tissues may also contribute to the neuroprotection. Estrogen negative feedback and suppression of body weight gain involve estrogen action at peripheral (e.g. pituitary, adipose tissue) as well as neural sites. Likewise, we cannot rule out the possibility that the drugs acted on vascular tissues to improve postischemia reperfusion (e.g. see Ref. 5). This explanation is, however, less likely, in that physiological levels of E2 protect neurons against global ischemia-induced cell death even when postischemic cerebral blood flow is strictly controlled (35).
Transgenic mice with targeted deletion of ER or ER? have been used to delineate the differential actions of estrogen mediated by each receptor subtype. These studies implicate ER in E2-dependent sexual receptivity (36, 37, 38), negative feedback regulation of LH (24, 39), and suppression of ovariectomy-induced increases in food intake and weight gain (40). In contrast, female ER? null mutants exhibit normal lordosis responses (37) and normal serum LH levels (41) and are able to reproduce and lactate (42). Our findings that PPT elicited the lordosis response, produced negative feedback inhibition of serum LH levels, and reduced the rate of weight gain in young, ovx female rats provide confirmation of a role for ER in responses to estrogen that are mediated, at least in part, by brain receptors.
ER regulation by E2 and after global ischemia
Our observation that ER protein levels are significantly higher in CA1 when ovx rats receive E2 replacement extends to the in vivo situation the recent finding that E2 increases ER protein levels in cultured hippocampal neurons (43). There is now evidence that E2 may be synthesized de novo in hippocampal neurons of both adult and developing rats (44, 45) and that it is important for maintaining hippocampal synaptic connections (46). Therefore, E2 may be an essential physiological regulator of hippocampal ER expression and synaptic connectivity. Neuronal injury can also regulate expression of a number of genes, including those encoding ERs (9, 25, 26). Our finding that global ischemia up-regulates ER protein expression in the hippocampal CA1 is consistent with findings that focal ischemia up-regulates ER mRNA expression in neocortex (25). In contrast, global ischemia does not increase ER expression in adult macaque monkeys, whereas it markedly increases glial ER? immunostaining in the CA1 of this species (47). Our finding that neither global ischemia nor estrogen significantly alters ER? expression in CA1 thus stands in contrast to the reported actions of global ischemia in monkeys and to previous findings that focal ischemia causes a 50% decrease in ER? mRNA in the rat cerebral cortex, a response that was prevented by exogenous estrogen (25). Hence, changes in receptor expression in response to neuronal injury are unlikely to account for the all-or-none effect of the ER-selective agonists in affording neuroprotection. Moreover, we cannot rule out the possibility that the recently described ER-X, whose levels also increase after ischemia (26), or other molecular entities that bind E2 and ICI 182,780 or that modulate the efficacy of ligand-induced transcriptional activation (48) also contribute to the neuroprotective actions of E2 in vivo.
Summary
It is well established that estrogen affords neuroprotection in animal models of stroke and global ischemia. The present study shows that the neuroprotective actions of estrogen can be mediated by either ER or ER? in a rat model of transient global ischemia. Our finding that the ER antagonist ICI 182,780 administered ICV abolishes protection by estrogen implicates brain ERs as the cellular mediators of estrogen neuroprotection and argues against a role for other, receptor-independent mechanisms. The ability of ICI 182,780 to abolish estrogen protection when given ICV only at 0 and 12 h after ischemia, despite the continued presence of estrogen, also demonstrates that E2 action in the early postischemic period is critical. The up-regulation of ER protein expression in CA1 observed at 24 and 48 h after ischemia, times at which histological evidence of cell death is not yet observed in ischemic rats, suggests that this may be a protective response by the injured brain. Delineation of the cellular targets that mediate neuroprotection by estrogen contributes to our understanding of how estrogen promotes neuronal survival and may help in the prevention and/or treatment of global ischemia arising during cardiac arrest.
Acknowledgments
Special thanks to Istvan Merchenthaler, Ph.D., and Heather Harris, Ph.D., from Wyeth-Ayerst Laboratories, Inc., for supplying the ER-selective agonists for this study and to Stephen Alves, Ph.D., from Merck, for supplying the ER? antibody. Dr. Patrick Hof of Mount Sinai School of Medicine provided invaluable assistance with the stereological estimates of CA1 volume. We also acknowledge Dr. Nanette Santoro for her thoughtful scientific input and comments on the manuscript. We also thank Dr. Alice Shu, Ms. Monique Bryan, and Dr. Tovaghgol Adel for excellent technical assistance.
References
Tanaka H, Grooms SY, Bennett MV, Zukin RS 2000 The AMPAR subunit GluR2: still front and center-stage. Brain Res 886:190–207
Zukin RS, Jover T, Yokota H, Calderone A, Simionescu M, Lau CG 2004 Molecular and cellular mechanisms of ischemia-induced neuronal death. In: Mohr JP, Choi DW, Grotta JC, Weir B, Wolf PA, eds. Stroke: pathophysiology, diagnosis, and management. Philadelphia: The Curtis Center; 829–854
Behl C 2002 Oestrogen as a neuroprotective hormone. Nat Rev Neurosci 3:433–442
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
He Z, He YJ, Day AL, Simpkins JW 2002 Proestrus levels of estradiol during transient global cerebral ischemia improves the histological outcome of the hippocampal CA1 region: perfusion-dependent and -independent mechanisms. J Neurol Sci 193:79–87
Garcia-Segura LM, Azcoitia I, DonCarlos LL 2001 Neuroprotection by estradiol. Prog Neurobiol 63:29–60
Jover T, Tanaka H, Calderone A, Simionescu M, Yokota H, Etgen AM, Zukin RS 2002 Estrogen protects against global ischemia-induced neuronal death and prevents activation of apoptotic signaling cascades in the hippocampal CA1. J Neurosci 22:2115–2124
Shughrue PJ, Merchenthaler I 2003 Estrogen prevents the loss of CA1 hippocampal neurons in gerbils after ischemic injury. Neuroscience 116:851–861
Merchenthaler I, Dellovade TL, Shughrue PJ 2003 Neuroprotection by estrogen in animal models of global and focal ischemia. Ann NY Acad Sci 1007:89–100
Toran-Allerand CD 2004 Minireview: a plethora of estrogen receptors in the brain: where will it end? Endocrinology 145:1069–1074
Blaustein JD 2004 Minireview: neuronal steroid hormone receptors: they’re not just for hormones anymore. Endocrinology 145:1075–1081
Perez SE, Chen EY, Mufson EJ 2003 Distribution of estrogen receptor and ? immunoreactive profiles in the postnatal rat brain. Brain Res Dev Brain Res 145:117–139
Mitra SW, Hoskin E, Yudkovitz J, Pear L, Wilkinson HA, Hayashi S, Pfaff DW, Ogawa S, Rohrer SP, Schaeffer JM, McEwen BS, Alves SE 2003 Immunolocalization of estrogen receptor ? in the mouse brain: comparison with estrogen receptor . Endocrinology 144:2055–2067
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
Merchenthaler I, Lane MV, Numan S, Dellovade TL 2004 Distribution of estrogen receptor and ? in the mouse central nervous system: in vivo autoradiographic and immunocytochemical analyses. J Comp Neurol 473:270–291
Gould E, Woolley CS, Frankfurt M, McEwen BS 1990 Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. J Neurosci 10:1286–1291
Murphy DD, Segal M 1996 Regulation of dendritic spine density in cultured rat hippocampal neurons by steroid hormones. J Neurosci 16:4059–4068
Woolley CS, McEwen BS 1994 Estradiol regulates hippocampal dendritic spine density via an N-methyl-D-aspartate receptor-dependent mechanism. J Neurosci 14:7680–7687
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
Gazzaley AH, Weiland NG, McEwen BS, Morrison JH 1996 Differential regulation of NMDAR1 mRNA and protein by estradiol in the rat hippocampus. J Neurosci 16:6830–6838
Dubal DB, Zhu H, Yu J, Rau SW, Shughrue PJ, Merchenthaler I, Kindy MS, Wise PM 2001 Estrogen receptor , not ?, is a critical link in estradiol-mediated protection against brain injury. Proc Natl Acad Sci USA 98:1952–1957
Carswell HV, Macrae IM, Gallagher L, Harrop E, Horsburgh KJ 2004 Neuroprotection by a selective estrogen receptor ? agonist in a mouse model of global ischemia. Am J Physiol Heart Circ Physiol 287:H1501–H1504
Stauffer SR, Coletta CJ, Tedesco R, Nishiguchi G, Carlson K, Sun J, Katzenellenbogen BS, Katzenellenbogen JA 2000 Pyrazole ligands: structure-affinity/activity relationships and estrogen receptor--selective agonists. J Med Chem 43:4934–4947
Dorling AA, Todman MG, Korach KS, Herbison AE 2003 Critical role for estrogen receptor in negative feedback regulation of gonadotropin-releasing hormone mRNA expression in the female mouse. Neuroendocrinology 78:204–209
Dubal DB, Shughrue PJ, Wilson ME, Merchenthaler I, Wise PM 1999 Estradiol modulates bcl-2 in cerebral ischemia: a potential role for estrogen receptors. J Neurosci 19:6385–6393
Toran-Allerand CD, Guan X, MacLusky NJ, Horvath TL, Diano S, Singh M, Connolly Jr ES, Nethrapalli IS, Tinnikov AA 2002 ER-X: a novel, plasma membrane-associated, putative estrogen receptor that is regulated during development and after ischemic brain injury. J Neurosci 22:8391–8401
Chen J, Xu W, Jiang H 2001 17?-Estradiol protects neurons from ischemic damage and attenuates accumulation of extracellular excitatory amino acids. Anesth Analg 92:1520–1523
Jover T, Simionescu M, Yokota H, Etgen AM, Zukin RS 2002 Estrogen acts via estrogen receptors and insulin-like growth factor-I receptors to protect against global ischemia-induced neuronal death. Soc Neurosci Abstr 28:6
Sudo S, Wen TC, Desaki J, Matsuda S, Tanaka J, Arai T, Maeda N, Sakanaka M 1997 ?-Estradiol protects hippocampal CA1 neurons against transient forebrain ischemia in gerbil. Neurosci Res 29:345–354
Horsburgh K, Macrae IM, Carswell H 2002 Estrogen is neuroprotective via an apolipoprotein E-dependent mechanism in a mouse model of global ischemia. J Cereb Blood Flow Metab 22:1189–1195
Sampei K, Goto S, Alkayed NJ, Crain BJ, Korach KS, Traystman RJ, Demas GE, Nelson RJ, Hurn PD2000 Stroke in estrogen receptor--deficient mice. Stroke 31:738–743; discussion 744
Sawada M, Alkayed NJ, Goto S, Crain BJ, Traystman RJ, Shaivitz A, Nelson RJ, Hurn PD 2000 Estrogen receptor antagonist ICI182,780 exacerbates ischemic injury in female mouse. J Cereb Blood Flow Metab 20:112–118
Harada H, BharwaniI S, Pavlick KP, Korach KS, Grisham MB 2004 Estrogen receptor-, sexual dimorphism and reduced-size liver ischemia and reperfusion injury in mice. Pediatr Res 55:450–456
Harris HA, Katzenellenbogen JA, Katzenellenbogen BS 2002 Characterization of the biological roles of the estrogen receptors, ER and ER?, in estrogen target tissues in vivo through the use of an ER-selective ligand. Endocrinology 143:4172–4177
Wang Q, Santizo R, Baughman VL, Pelligrino DA, Iadecola C 1999 Estrogen provides neuroprotection in transient forebrain ischemia through perfusion-independent mechanisms in rats. Stroke 30:630–637
Ogawa S, Eng V, Taylor J, Lubahn DB, Korach KS, Pfaff DW 1998 Roles of estrogen receptor- gene expression in reproduction-related behaviors in female mice. Endocrinology 139:5070–5081
Ogawa S, Chan J, Chester AE, Gustafsson JA, Korach KS, Pfaff DW 1999 Survival of reproductive behaviors in estrogen receptor ? gene-deficient (?ERKO) male and female mice. Proc Natl Acad Sci USA 96:12887–12892
Rissman EF, Early AH, Taylor JA, Korach KS, Lubahn DB 1997 Estrogen receptors are essential for female sexual receptivity. Endocrinology 138:507–510
Wersinger SR, Haisenleder DJ, Lubahn DB, Rissman EF 1999 Steroid feedback on gonadotropin release and pituitary gonadotropin subunit mRNA in mice lacking a functional estrogen receptor . Endocrine 11:137–143
Geary N, Asarian L, Korach KS, Pfaff DW, Ogawa S 2001 Deficits in E2-dependent control of feeding, weight gain, and cholecystokinin satiation in ER- null mice. Endocrinology 142:4751–4757
Couse JF, Yates MM, Walker VR, Korach KS 2003 Characterization of the hypothalamic-pituitary-gonadal axis in estrogen receptor (ER) null mice reveals hypergonadism and endocrine sex reversal in females lacking ER but not ER?. Mol Endocrinol 17:1039–1053
Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, Sar M, Korach KS, Gustafsson JA, Smithies O 1998 Generation and reproductive phenotypes of mice lacking estrogen receptor ?. Proc Natl Acad Sci USA 95:15677–15682
Prange-Kiel J, Wehrenberg U, Jarry H, Rune GM 2003 Para/autocrine regulation of estrogen receptors in hippocampal neurons. Hippocampus 13:226–234
Amateau SK, Alt JJ, Stamps CL, McCarthy MM 2004 Brain estradiol content in newborn rats: sex differences, regional heterogeneity, and possible de novo synthesis by the female telencephalon. Endocrinology 145:2906–2917
Hojo Y, Hattori TA, Enami T, Furukawa A, Suzuki K, Ishii HT, Mukai H, Morrison JH, Janssen WG, Kominami S, Harada N, Kimoto T, Kawato S 2004 Adult male rat hippocampus synthesizes estradiol from pregnenolone by cytochromes P45017 and P450 aromatase localized in neurons. Proc Natl Acad Sci USA 101:865–870
Kretz O, Fester L, Wehrenberg U, Zhou L, Brauckmann S, Zhao S, Prange-Kiel J, Naumann T, Jarry H, Frotscher M, Rune GM 2004 Hippocampal synapses depend on hippocampal estrogen synthesis. J Neurosci 24:5913–5921
Takahashi N, Tonchev AB, Koike K, Murakami K, Yamada K, Yamashima T, Inoue M 2004 Expression of estrogen receptor-? in the postischemic monkey hippocampus. Neurosci Lett 369:9–13
Simons Jr S 2003 The importance of being varied in steroid receptor transactivation. Trends Pharmacol Sci 24:253–259
Sun J, Huang YR, Harrington WR, Sheng S, Katzenellenbogen JA, Katzenellenbogen BS2002 Antagonists selective for estrogen receptor . Endocrinology 143:941–947(Nora R. Miller, Teresa Jo)