Estradiol Replacement Enhances Working Memory in Middle-Aged Rats When Initiated Immediately after Ovariectomy But Not after a Lon
http://www.100md.com
《内分泌学杂志》
Department of Psychology, University of New Orleans, New Orleans, Louisiana 70148
Abstract
The goal of the present study was to explore the effects of long-term hormone deprivation on the ability of subsequent estrogen replacement to affect cognition. Female rats, 12 months of age, underwent ovariectomies (n = 30) or sham surgeries (n = 10). Intact rats and 20 ovariectomized rats received cholesterol implants. Ten ovariectomized rats received implants containing 25% estradiol. Five months later, implants were replaced. Half of the ovariectomized rats with cholesterol implants received estradiol implants and half received new cholesterol implants. Rats with estradiol implants received new estradiol implants. Intact rats were ovariectomized and given estradiol implants. Beginning 1 wk later, working memory performance was assessed in an eight-arm radial maze across 24 d of acquisition and during eight additional trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices. Estradiol replacement initiated immediately after ovariectomy at either 12 or 17 months of age significantly improved performance during acquisition and delay trials, compared with control treatment. When estradiol replacement was initiated at 17 months of age, 5 months after ovariectomy, no enhancements were evident. Uteri of rats that experienced delayed estradiol replacement weighed significantly more than uteri of ovariectomized controls but significantly less than uteri of rats that received immediate estradiol replacement. Uterine weight negatively correlated with mean errors during acquisition. These results indicate that whereas chronic estradiol replacement regimens positively affect working memory in middle-aged animals when initiated immediately after ovariectomy, estradiol replacement is not effective when initiated after long-term hormone deprivation.
Introduction
THERE IS A GREAT deal of evidence in the basic and clinical literature supporting a role for estrogens in the modulation of cognitive function in mammals (for review, see Ref.1). However, the direction and extent of those effects remain controversial. Whereas many early studies indicated that hormone replacement therapy (HRT) could delay or alleviate age-related cognitive decline in postmenopausal women (for review, see Ref.2), recent results of the large Women’s Health Initiative Memory Study conducted by the National Institutes of Health indicated that replacement regimens consisting of chronic conjugated equine estrogens (3, 4) or chronic conjugated equine estrogens plus medroxyprogesterone (5, 6) did not improve cognition and actually increased risk of dementia. In attempts to reconcile these findings with those of previous studies, various factors need to be considered including regimen of replacement, the age of the subjects, and the number of years of estrogen deficiency at the time of initiation of HRT (7). For example, the women included in the Women’s Health Initiative Memory Study were at least 65 yr of age at the time of intervention and the majority had never used previous hormone therapy. Thus, these women were well beyond the age at which HRT is typically initiated to alleviate postmenopausal symptoms and most had experienced more than a decade of ovarian hormone deprivation before beginning HRT. It has been hypothesized that long-term ovarian hormone deprivation affects the efficacy of subsequent HRT in postmenopausal women (8).
In addition to the clinical studies, basic research has been conducted to elucidate the effects of estrogens on cognitive function. Studies using rodent models have demonstrated that estradiol positively affects hippocampal structure and function (9, 10, 11) and enhances performance on many hippocampal-dependent tasks (12, 13, 14, 15, 16, 17, 18, 19, 20, 21). However, most of the studies assessing the effects of estradiol on cognitive function using rodent models have used young adult females. Studies using middle-aged and older animals have yielded mixed results. For example, 1 wk of chronic estradiol replacement in 13-month-old rats had no effect on the acquisition of a T-maze active avoidance task (22) and had no effect on spatial memory in aged female rats as measured in a radial-arm maze (23). In contrast, 5 wk of chronic oral estradiol in ovariectomized middle-aged mice improved object recognition (24) and daily injections of estradiol to aged mice improved spatial reference memory, compared with controls (25). There is evidence of an interaction of age and dose of estradiol used in chronic replacement paradigms that may contribute to inconsistent results (26). In addition, there is preliminary evidence in the basic literature to support the hypothesis that estrogen replacement may lose its ability to positively affect cognition after a long-term period of hormone deprivation. Chronic estradiol replacement initiated six months after ovariectomy in middle-aged rats was not effective at ameliorating working memory deficits in a delayed-nonmatching-to-sample task in a T maze unless primed with cyclic injections of estradiol (27). It is unclear whether the chronic replacement regimen was largely ineffective due to the age of the rats or the long-term hormone deficiency at the time of initiation of the replacement. In another study, continuous estradiol replacement initiated within 3 months, but not 10 months after ovariectomy, improved performance of aged rats in a delayed-matching-to-position spatial memory task (19). However, there was no effect of hormonal status on trials after acquisition when the intertrial interval was increased, suggesting that the hormone-induced enhancement during acquisition of the delayed-matching-to-position task may not have been due to a selective enhancement of spatial memory. Additionally, the duration of estradiol replacement differed between groups. Thus, further exploration into the effects of ovarian hormone deprivation on the ability of future estrogen replacement to affect cognitive function is warranted based on the results of these studies as well as the need to address the inconsistencies in the clinical literature regarding the efficacy of HRT.
The goal of the current study was to determine the effects of long-term hormone deprivation on the ability of subsequent estradiol replacement to positively affect working memory and control for other factors that may affect performance, such as age and duration of estradiol exposure. We determined whether the efficacy of estradiol replacement in middle-aged animals as related to working memory is affected by the age at which it is initiated, the duration of exposure, and importantly the length of hormone deprivation before replacement. We previously demonstrated that a chronic replacement paradigm delivering constant physiological levels of estradiol enhanced acquisition of a working memory task in a radial-arm maze in early middle-aged rats when initiated immediately after ovariectomy (28). In the present experiment, we compared the ability of physiological levels of estradiol replacement initiated at different time points to affect working memory performance in a radial-arm maze in 17-month-old ovariectomized rats. The effects of estradiol replacement initiated immediately after ovariectomy at either 12 or 17 months of age or initiated at 17 months of age, 5 months after ovariectomy, were determined. We assessed the effects of our treatments during acquisition of the maze task. In addition, because the effects of estradiol on working memory may be most apparent when the working memory load increases (14), we also assessed the effects of our treatments on delay trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices.
Materials and Methods
Subjects
Female Long-Evans rats, retired breeders, were purchased from Harlan Sprague Dawley, Inc. (Indianapolis, IN). Rats were housed individually in a temperature-controlled vivarium under a 12-h light, 12-h dark cycle (lights on at 0700 h) and were allowed free access to water and food for 2 wk. Animal care was in accordance with guidelines set by the National Institutes of Health Guide for the Care and Use of Laboratory Animals (1996).
Treatments
At approximately 12 months of age, females underwent either ovariectomies (n = 30) or sham surgeries (n = 10) while under anesthesia induced by injections of ketamine (100 mg/kg ip; Bristol Laboratories, Syracuse, NY) and xylazine (7 mg/kg ip; Miles Laboratories, Shawnee, KS). At the time of the surgeries, 5-mm SILASTIC brand capsules (0.58-in. inner diameter and 0.077-in. outer diameter; Dow Corning, Midland, MI) containing cholesterol (Sigma Chemical Co., St. Louis, MO) were implanted sc on the dorsal aspect of the neck in 20 of the rats that had undergone ovariectomies and all 10 of the rats that had undergone sham surgeries. Capsules containing 25% 17-estradiol (Sigma) diluted with cholesterol were implanted in 10 of the ovariectomized rats. Capsules of these dimension and estradiol concentration maintain circulating estradiol titers at 20–25 pg/ml, a low level typical of diestrus (19, 29, 30).
These treatments continued for a 5-month period during which time animals were weighed and handled once per week. Additionally, capsules were manually manipulated each week to prevent the build-up of scar tissue. We did not replace the capsules during this 5-month period because in a previous experiment in our laboratory, capsules of the same dimensions and estradiol concentration remained active for a similar time period (Daniel, J. M., unpublished data). During the final 2 wk of the 5-month period, daily vaginal smears were collected by lavage from each of the animals to confirm neuroendocrine status.
When rats were approximately 17 months of age, they all underwent a second set of surgeries while under anesthesia induced by injections of ketamine and xylazine. Animals that had previously undergone sham surgeries were ovariectomized and previously ovariectomized rats underwent sham surgeries. Additionally, all SILASTIC brand capsules were removed. Visual inspection of these capsules revealed that all had maintained their integrity. New capsules were implanted. Ovariectomized rats that had previously received estradiol capsules were implanted with new estradiol capsules. Half of the originally ovariectomized animals that had received cholesterol capsules were implanted with new cholesterol capsules. The other half of the originally ovariectomized animals that had received cholesterol capsules were implanted with estradiol capsules. Finally, the 10 newly ovariectomized rats that had originally received cholesterol capsules were implanted with estradiol capsules. Surgeries and treatments for each of the four groups are summarized as follows: Ch/Ch group (12 months of age: ovariectomy; cholesterol implant; 17 months of age: sham surgery; cholesterol implant); E/E group (12 months of age: ovariectomy; estradiol implant; 17 months of age: sham surgery; estradiol implant); Ch/E group (12 months of age: ovariectomy; cholesterol implant; 17 months of age: sham surgery; estradiol implant); Intact/E group (12 months of age: sham surgery; cholesterol implant; 17 months of age: ovariectomy; estradiol implant).
Behavioral testing
Radial maze acquisition.
Rats were allowed 1 wk of recovery time after the second set of surgeries before training on the radial maze began. Beginning 3 d before training, all rats were placed on diets to maintain body weights at 90% of their free-feeding weights. Animals were trained to obtain food rewards (Froot Loops; Kellogg Co., Battle Creek, MI) from the arms of an elevated eight-arm radial maze. Each day during the week before training, several Froot Loops were placed in the cage of each rat. One day before the beginning of training, each rat was placed in the maze for a 15-min acclimation period. Froot Loops were sprinkled throughout the maze and the rat was allowed to travel freely throughout the maze with all arms opened. The maze was purchased from SciPro, Inc. (Sanborn, NY) and consisted of gray metal floors and clear Plexiglas walls. The arms (15 cm wide x 45 cm long x 23 cm high) extended out from an octagonal center (48 cm across). Food receptacles, which were opaque in color to block the view of the food reward, were attached to the ends of each arm. The maze was located in the center of a 3 x 3 m room. Extramaze cues including lighting fixtures, a door, and electrical outlets as well as geometric shapes attached to walls were visible from the maze.
To begin each trial, the rat was placed in the center compartment with all doors leading to the arms opened. Doors remained opened throughout the trial. The rat was allowed to enter any of the eight arms. The experimenter, who was seated in the room at a fixed location approximately 1 m from the maze, recorded arm choices. An arm choice was scored if the rat traveled halfway down the length of an arm. The animal was allowed to choose arms in any order until all arms were visited or until 5 min elapsed. An error was scored if a rat reentered an arm previously visited. Arm-choice accuracy was measured by the number of errors made in the first eight visits. If a rat did not enter eight arms in a trial, arms never visited were counted as errors. However, after the first block of trials, this rarely occurred. A record of the arm-choice order was kept for each rat and analyzed by the experimenter on a daily basis to determine whether any rats were using response algorithms such as adjacent-arm strategies or other patterned responses across trials to solve the maze. No such patterns were revealed. As a measure of the speed at which an animal traversed the maze, total number of arms entered per minute (total number of arms entered in a trial/time to complete the trial) was calculated. Maze training took place at least 5 d/wk. Each animal received one trial per day across 24 d of acquisition.
Delay trials.
After the 24-d acquisition period, animals received eight additional daily trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices. During the delay trials, the animal was removed from the maze after it had made four correct arm choices and placed in a holding cage for period of 2.5 h. It then was returned to the center compartment of the maze with all arms opened and allowed to choose arms in any order until all arms were visited or until 5 min had elapsed. Errors made during the predelay and postdelay periods were scored separately. Postdelay errors were broken down into retroactive errors and proactive errors (31). A retroactive error was the first reentry into an arm already visited before the delay. A proactive was a reentry into an arm already visited in the postdelay period.
Vaginal cytology, uterine weight
To identify neuroendocrine status during the second set of hormone treatments, daily vaginal smears were collected from all animals during a 2-wk period between the completion of behavioral testing and time of being killed. These samples, as well as those that were collected at the end of the initial 5-month hormone treatment period, were stained with Harris hematoxylin solution and eosin and microscopically examined. The following criteria were used to identify estrous stage: estrus was characterized by a predominance of cornified cells; diestrus was characterized by the predominance of leukocytes; and proestrus was characterized by the predominance of nucleated epithelial cells (32, 33).
Two weeks after the completion of behavioral testing, animals were anesthetized with ketamine and xylazine. After animals were killed, uteri were removed and weighed (see Fig. 1 for a summary of treatment and testing schedule).
Statistical analyses
One-way ANOVAs, with treatment as the factor, were used to test for differences in body weight and uterine weight. Data collected from acquisition trials and from delay trials were analyzed by two-way ANOVAs (treatment x 4-d block) with repeated measures on block. Data from individual blocks during acquisition were analyzed by one-way ANOVA (treatment). If a significant main effect was revealed in any of these analyses (P < 0.05), then Duncan’s multiple range post hoc test (P < 0.05) was applied.
Results
Several of the rats died during the course of the experiment. Data from animals that survived at least through completion of the acquisition trials were included in the analyses of body weight data and acquisition data (Ch/Ch, n = 9; E/E, n = 8; Ch/E, n = 8; Intact/E, n = 10). One additional animal died (Ch/Ch) before the completion of the delay trials and another died (Intact/E) before measurement of uterine weight.
Behavioral testing results
Acquisition.
As illustrated in Fig. 2, estradiol replacement initiated in rats immediately after ovariectomy either at 12 or 17 months of age significantly improved performance during acquisition of a working memory task in a radial-arm maze, compared with ovariectomized controls. However, when estradiol replacement was initiated at 17 months of age after 5 months of hormone deprivation, no such enhancement was revealed. A significant main effect of treatment was revealed for number of incorrect choices in the first eight arm visits (F3,31 = 3.033, P < 0.05). Post hoc analyses revealed that females ovariectomized at 12 months that received immediate, continuous estradiol replacement (E/E) and those ovariectomized at 17 months that received immediate, continuous estradiol replacement (Intact/E) achieved significantly fewer errors than those control animals ovariectomized at 12 months that received continuous cholesterol treatment (Ch/Ch; Fig. 2A). Females ovariectomized at 12 months but whose estradiol treatment was delayed until 1 wk before training when they were 17 months of age (Ch/E) did not significantly differ from any of the other groups. There was a significant effect of block (F5,155 = 22.79, P < 0.0001), indicating that the accuracy of all groups improved over time. There was no interactive effect of treatment and block. Figure 2B illustrates the effects of various estradiol treatments on acquisition over 24 d, presented in blocks of 4 d.
A significant effect of treatment for number of incorrect choices for individual blocks of trials emerged only for the final block (F3,31 = 3.88, P < 0.05), indicating that the effects of our treatments were most apparent at the end of the acquisition period when the animals were familiar with the demands of the task. Post hoc analyses revealed that E/E and Intact/E females achieved significantly fewer errors in the last block of trials than Ch/Ch control animals. Ch/E females did not significantly differ from any of the other groups.
There was a significant effect of block (F5,155 = 29.45, P < 0.0001) but no significant effect of treatment or interactive effect of treatment and block for mean number of arms entered per minute (data not shown), indicating that whereas arm entry rate increased for all groups across trials, this effect did not differ across treatment groups.
Delay trials.
As illustrated in Fig. 3, estradiol replacement initiated at either 12 or 17 months of age immediately after ovariectomy but not when initiated at 17 months of age after long-term hormone deprivation resulted in significantly fewer postdelay retroactive errors, compared with control treatment. There were no main or interactive effects for the predelay period because very few errors were made by any of the groups (data not shown). There was a significant main effect of treatment for the number of postdelay retroactive errors (F3,30 = 3.14, P < 0.05). Post hoc analyses revealed that E/E and Intact/E females made significantly fewer retroactive errors than Ch/Ch controls (Fig. 3A). Ch/E females did not significantly differ from any of the other groups. The effect of block approached the level of statistical significance (F3,30 = 4.089, P = 0.052). There was no significant interactive effect of treatment and block for retroactive errors. The effect of treatment on number of postdelay proactive errors approached but did not reach the level of statistical significance (F3,30 = 2.68, P = 0.065; Fig. 3B). There was neither a significant main effect of block nor an interactive effect of block and treatment for number of proactive errors.
Efficacy of initial hormone treatments
The efficacy of the initial 5 months of hormone treatments was assessed by two measures, body weight and examination of vaginal cytology. There was a significant effect of hormone treatment on body weight (F2,34 = 3.38, P < 0.05). Post hoc analyses revealed that at the end of the initial hormone treatments, the mean weight of the ovariectomized rats with cholesterol implants (347 g ± 9.3 SEM) was significantly more than the mean weight of the ovariectomized rats with estradiol implants (316 g ± 8.4 SEM). The mean weight of the gonadally intact rats (321 g ± 10.6 SEM) did not significantly differ from either of the other groups.
Light microscopic examination of vaginal smears collected during the final 2 wk of the initial hormone treatment period revealed that the hormone treatments were effective. The smears of all ovariectomized animals that had cholesterol implants during the initial 5-month treatment period were characterized by a predominance of leukocytes. Smears of the ovariectomized animals that had estradiol implants were characterized by a predominance of cornified cells with some nucleated epithelial cells. All of the smears of the gonadally intact rats showed changes across days, although there was evidence of disruptions in the cycles. The length of the cycles, as determined by the number of days between consecutive estrous stages (22), ranged from 5 to 8 d. The cycles of all but one of the intact rats included a proestrous stage.
Efficacy of final hormone treatments
The efficacy of the final period of hormone treatments was assessed by two measures, examination of vaginal cytology collected during the final 2 wk before the animals were killed and uterine weights. Examination of vaginal cytology indicated that the hormone treatments were effective. The vaginal smears of all of the ovariectomized animals that had cholesterol implants for both periods of hormone treatments (Ch/Ch) were characterized by a predominance of leukocytes. The smears of all of the ovariectomized animals that had estradiol implants for both periods of hormone treatments (E/E) were characterized by a predominance of cornified cells with some nucleated epithelial cells. The smears of all of the animals that were not ovariectomized until the beginning of the second period of hormone treatments at which time they received estradiol implants (Intact/E) were also characterized by a predominance of cornified cells with some nucleated epithelial cells. The smears of one of the ovariectomized animals that were treated with cholesterol for 5 months before the initiation of estradiol treatment (Ch/E) were characterized by a predominance of leukocytes. However, estradiol treatment was considered effective in this rat based on our other measure, uterine weight. The rest of the smears from this group were characterized by a predominance of cornified cells with some nucleated epithelial cells (see Fig. 4 for representative smears from each group).
There was a main effect of hormone treatment on uterine weight (F3,32 = 28.19, P < 0.001). Post hoc analyses revealed that the uteri of ovariectomized rats that had cholesterol implants for both periods of hormone treatments (Ch/Ch) weighed significantly less than the other three groups (see Fig. 5). Additionally, the uteri of the rats that experienced 5 months of hormone deprivation before initiation of estradiol treatment (Ch/E) weighed significantly less than the two groups of rats that had estradiol treatments initiated immediately after ovariectomies (E/E and Intact/E), indicating that estradiol replacement initiated after a long period of deprivation was less effective at increasing uterine weight than was replacement initiated immediately after ovariectomy. In light of this finding, correlation analysis was conducted to determine whether there was a relationship between the ability of our various treatments to affect uterine weight and cognitive function. Results revealed a significant negative correlation (r = –0.414, P < 0.01) between uterine weight and the mean number of errors committed during trials in the acquisition period (see Fig. 6).
Finally, because there were differences in body weights between the E/E and Ch/E animals at the beginning of the second treatment period, we wanted to control for the possibility that the attenuated effect of estradiol in the Ch/E group, compared with the other estradiol groups, was due to lower levels of circulating estradiol as a result of increased body weight. We therefore did correlation analyses including the three groups that received estradiol treatment to determine whether there was a relationship between body weight at the end of the first treatment period and mean number of errors during acquisition and/or between body weight at the end of the first treatment period and uterine weight at the time the animals were killed. No significant relationships were revealed between body weight and these measures, indicating that differences in body weight did not affect the efficacy of the various estradiol treatments.
Discussion
The primary finding of the present study was that estradiol replacement loses its ability to positively affect working memory in middle-aged rats when it is initiated after a long period of ovarian hormone deprivation. Whereas chronic estradiol replacement was effective in ameliorating ovariectomy-induced deficits in working memory when initiated immediately after ovariectomy at either 12 or 17 months of age, it was not effective when initiated after a 5-month period of hormone deprivation. Interestingly, factors such as the length of estradiol replacement before training and the point during the middle-age period at which the animals were ovariectomized did not influence the efficacy of estradiol replacement as related to its effects on working memory. Replacement paradigms initiated 5 months or 1 wk before acquisition training were equally effective at enhancing working memory performance if initiated immediately after ovariectomy. In contrast, estradiol replacement initiated 1 wk before acquisition but 5 months after ovariectomy had no effect, compared with control treatment. Thus, the present study reveals that the ability of estradiol replacement to affect working memory performance in ovariectomized middle-aged animals is affected by the timing of its initiation. Future studies will be needed to determine whether it is necessary to initiate estradiol replacement immediately after ovariectomy or, as suggested by a prior study in which estradiol replacement improved performance on a delayed-matching-to-position task when initiated 3 months but not 10 months after ovariectomy (19), there is a window of opportunity after the loss of ovarian function in which estradiol replacement must be initiated to be effective. Additionally, the length of this window may vary with the regimen of hormone treatment (27).
The possibility exists that in the present study, group differences in performance during the acquisition trials as well as the delay trials were not due to direct effects of our treatments on working memory but were due to effects of our treatments on other requirements of the tasks. However, our data suggest otherwise. There were no differences in the rates at which each of the groups traversed the maze, suggesting similar motivation and attention levels among the groups. Additionally, there were no differences in performance in the early blocks of trials of the acquisition period during which time the animals were learning the requirements of the task. Group differences did emerge in the final block of trials when the requirements of the tasks were familiar, and therefore differences were more likely due to specifically to differences in working memory.
The mechanism by which estradiol loses its ability to affect working memory after a period of hormone deprivation is unknown. However, there is evidence that estradiol differentially affects brain areas important for cognition in young and older animals. For example, a series of studies indicate that there are different mechanisms of hippocampal plasticity in young and aged female rats. In young animals, estradiol regulates hippocampal morphology as evidenced by increases in CA1 spine and synapse density (34) and concurrent increases in N-methyl-D-aspartate (NMDA) receptor binding levels (9, 11). There is a blunted response to estradiol replacement in the hippocampus of aged female rats, compared with young animals. In aged animals, estradiol failed to increase hippocampal spine density, although it did up-regulate synaptic NMDA receptor subunit 1 (NR1), the obligatory subunit of the NMDA receptor (35). In contrast to the up-regulation of NR1 in the young animals, which was characterized by increases that were in proportion to new spines, the estradiol-induced up-regulation of NR1 in the aged animals was characterized by increases in NR1 per synapse. This decreased responsiveness of the aged animals to the effects of estradiol may be related to a decrease in CA1 estrogen receptor- levels that is evident in aged animals (36). As in CA1, the response of the dentate gyrus to estradiol also changes with age. In aged female rats, acute estradiol exposure results in increased spine density of dentate granule cells, even after months of hormonal deprivation (37). No such effects are evident in young adult female rats.
It will be important to determine whether the differences in the hippocampal response to estradiol in young and aged animals are due to changes in the hormonal milieu experienced during aging or are due to the aging process independent of the effects of hormonal changes. There have been attempts to differentiate the influence of age and changes in hormonal status on the effects of estradiol in the hippocampus. The results of a study that analyzed the effects of postovariectomy interval and estradiol replacement on levels of hippocampal NMDA receptor mRNA in young, middle-aged, and aged rats revealed age and length of ovarian hormone deprivation had significant impacts on NMDA mRNA levels with little impact of estradiol replacement (38). However, the replacement paradigms did not begin until at least 1 month after ovariectomy.
Future studies need to be completed to determine whether estrogen replacement initiated in middle-age when decreases in estrogen levels first occur would prevent the age-related change in the response to estrogen in aged animals. Consistent with this possibility is a report that long-term ovarian hormone deprivation has negative effects on basal forebrain cholinergic neurons that go beyond the effects of aging alone (39). Decreases in both choline acetyltransferase and trkA mRNA in the medial septum and nucleus basalis magnocellularis were evident in middle-aged rats killed 6 months, but not 3 months, after ovariectomy relative to age-matched, gonadally intact controls. In contrast to estradiol-induced increases in levels of choline acetyltransferase and trkA mRNA in these areas exhibited by ovariectomized young animals (40, 41), short-term estradiol replacement initiated 6 months after ovariectomy in middle-aged animals had no effect.
In the present study, the finding that our different estradiol treatments differentially affected uterine weight is intriguing. Although all three regimens of estradiol replacement significantly increased uterine weight compared with control treatment, the replacement paradigm that followed a long period of deprivation resulted in significantly less increases than the two paradigms that were initiated immediately after ovariectomy. Additionally, uterine weight was negatively correlated with the number of errors made during the acquisition period, indicating a relationship between the ability of estradiol to affect performance on measures of learning and memory and its ability to affect estradiol-sensitive peripheral tissues. Thus, although the mechanism remains to be determined, the effect of long-term deprivation on the efficacy of estradiol replacement may extend beyond its effects on the brain and cognition.
In conclusion, the results of the present study support the hypothesis that long-term ovarian hormone deprivation negatively affects the ability of subsequent estradiol replacement to enhance performance on tasks of learning and memory. The mechanism by which the brain becomes less susceptible to the effects of estradiol after ovarian hormone deprivation remains to be determined. However, these results have implications toward reconciling the inconsistencies in the clinical literature regarding the effects of HRT on cognitive function and suggest that the timing of the initiation of HRT may be a crucial factor in determining its efficacy.
Acknowledgments
We thank Misha Rehage, Kelvin Jenkins, and Elizabeth Daray for their assistance in the collection of behavioral data.
Footnotes
This work was supported by National Science Foundation Grant 0423331 (to J.M.D.).
First Published Online October 20, 2005
Abbreviations: HRT, Hormone replacement therapy; NMDA, N-methyl-D-aspartate; NR1, NMDA receptor subunit 1.
Accepted for publication October 10, 2005.
References
Dohanich GP 2002 Gonadal steroids, learning and memory. In: Pfaff DW, Arnold AP, Etgen AM, Farbach SE, Rubin RT, eds. Hormones, brain and behavior. Vol 1. San Diego: Academic Press; 265–327
Sherwin BB 2003 Estrogen and cognitive functioning in women. Endocr Rev 24:133–151
Espeland MA, Rapp SR, Shumaker SA, Brunner R, Manson JE, Sherwin BB, Hsia J, Margolis KL, Hogan PE, Wallace R, Dailey M, Freeman R, Hays J 2004 Conjugated equine estrogens and global cognitive function in postmenopausal women: Women’s Health Initiative Memory Study. JAMA 291:2959–2968
Shumaker SA, Legault C, Kuller L, Rapp SR, Thal L, Lane DS, Fillit H, Stefanick ML, Hendrix SL, Lewis CE, Masaki K, Coker LH 2004 Conjugated equine estrogens and incidence of probable dementia and mild cognitive impairment in postmenopausal women: Women’s Health Initiative Memory Study. JAMA 291:2947–2958
Rapp SR, Espeland MA, Shumaker SA, Henderson VW, Brunner RL, Manson JE, Gass ML, Stefanick ML, Lane DS, Hays J, Johnson KC, Coker LH, Dailey M, Bowen D 2003 Effect of estrogen plus progestin on global cognitive function in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289:2663–2672
Shumaker SA, Legault C, Rapp SR, Thal L, Wallace RB, Ockene JK, Hendrix SL, Jones III BN, Assaf AR, Jackson RD, Kotchen JM, Wassertheil-Smoller S, Wactawski-Wende J 2003 Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289:2651–2662
Gleason CE, Carlsson CM, Johnson S, Atwood C, Asthana S 2005 Clinical pharmacology and differential cognitive efficacy of estrogen preparations. Ann NY Acad Sci 1052:93–115
Pinkerton JV, Henderson VW 2005 Estrogen and cognition, with a focus on Alzheimer’s disease. Semin Reprod Med 23:172–179
Woolley CS, Weiland NG, McEwen BS, Schwartzkroin PA 1997 Estradiol increases the sensitivity of hippocampal CA1 pyramidal cells to NMDA receptor-mediated synaptic input: correlation with dendritic spine density. J Neurosci 17:1848–1859
Gabor R, Nagle R, Johnson DA, Gibbs RB 2003 Estrogen enhances potassium-stimulated acetylcholine release in the rat hippocampus. Brain Res 962:244–247
Daniel JM, Dohanich GP 2001 Acetylcholine mediates the estrogen-induced increase in NMDA receptor binding in CA1 of the hippocampus and the associated improvement in working memory. J Neurosci 21:6949–6956
Daniel JM, Fader AJ, Spencer AL, Dohanich GP 1997 Estrogen enhances performance of female rats during acquisition of a radial arm maze. Horm Behav 32:217–225
Luine VN, Richards ST, Wu VY, Beck KD 1998 Estradiol enhances learning and memory in a spatial memory task and effects levels of monoaminergic neurotransmitters. Horm Behav 34:149–162
Bimonte HA, Denenberg VH 1999 Estradiol facilitates performance as working memory load increases. Psychoneuroendocrinology 24:161–173
Fader AJ, Johnson PE, Dohanich GP 1999 Estrogen improves working but not reference memory and prevents amnestic effects of scopolamine of a radial-arm maze. Pharmacol Biochem Behav 62:711–717
Fader AJ, Hendricson AW, Dohanich GP 1998 Estrogen improves performance of reinforced T-maze alternation and prevents the amnestic effects of scopolamine administered systemically or intrahippocampally. Neurobiol Learn Mem 69:225–240
Daniel JM, Winsauer PJ, Brauner IN, Moerschbaecher JM 2002 Estrogen improves response accuracy and attenuates the disruptive effects of 9-THC in ovariectomized rats responding under a multiple schedule of repeated acquisition and performance. Behav Neurosci 116:989–998
Gibbs RB 1999 Estrogen replacement enhances acquisition of a spatial memory task and reduces deficits associated with hippocampal muscarinic receptor inhibition. Horm Behav 36:222–233
Gibbs RB 2000 Long-term treatment with estrogen and progesterone enhances acquisition of a spatial memory task by ovariectomized aged rats. Neurobiol Aging 21:107–116
Luine VN, Jacome LF, MacLusky NJ 2003 Rapid enhancement of visual and place memory by estrogens in rats. Endocrinology 144:2836–2844
Sandstrom NJ, Williams CL 2001 Memory retention is modulated by acute estradiol and progesterone replacement. Behav Neurosci 115:384–393
Savonenko AV, Markowska AL 2003 The cognitive effects of ovariectomy and estrogen replacement are modulated by aging. Neuroscience 119:821–830
Luine V, Rodriguez M 1994 Effects of estradiol on radial arm maze performance of young and aged rats. Behav Neural Biol 62:230–236
Fernandez SM, Frick KM 2004 Chronic oral estrogen affects memory and neurochemistry in middle-aged female mice. Behav Neurosci 118:1340–1351
Frick KM, Fernandez SM, Bulinski SC 2002 Estrogen replacement improves spatial reference memory and increases hippocampal synaptophysin in aged female mice. Neuroscience 115:547–558
Foster TC, Sharrow KM, Kumar A, Masse J 2003 Interaction of age and chronic estradiol replacement on memory and markers of brain aging. Neurobiol Aging 24:839–852
Markowska AL, Savonenko AV 2002 Effectiveness of estrogen replacement in restoration of cognitive function after long-term estrogen withdrawal in aging rats. J Neurosci 22:10985–10995
Daniel JM, Olvet KR, Dohanich GP 1999 Effects of prior and continuous estrogen exposure on performance during acquisition of a radial arm maze by young adult and middle-aged female rats. Soc Behav Neuroendocr (Abstract)
Luine VN 1997 Steroid hormone modulation of hippocampal dependent spatial memory. Stress 2:21–36
Singh M, Meyer EM, Millard WJ, Simpkins JW 1994 Ovarian steroid deprivation results in a reversible learning impairment and compromised cholinergic function in female Sprague-Dawley rats. Brain Res 644:305–312
Chappell J, McMahan R, Chiba A, Gallagher M 1998 A re-examination of the role of basal forebrain cholinergic neurons in spatial working memory. Neuropharmacology 37:481–487
Montes GS, Luque EH 1988 Effects of ovarian steroids on vaginal smears in the rat. Acta Anat (Basel) 133:192–199
Marcondes FK, Bianchi FJ, Tanno AP 2002 Determination of the estrous cycle phases of rats: some helpful considerations. Braz J Biol 62:609–614
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
Adams MM, Shah RA, Janssen WG, Morrison JH 2001 Different modes of hippocampal plasticity in response to estrogen in young and aged female rats. Proc Natl Acad Sci USA 98:8071–8076
Adams MM, Fink SE, Shah RA, Janssen WG, Hayashi S, Milner TA, McEwen BS, Morrison JH 2002 Estrogen and aging affect the subcellular distribution of estrogen receptor-alpha in the hippocampus of female rats. J Neurosci 22:3608–3614
Miranda P, Williams CL, Einstein G 1999 Granule cells in aging rats are sexually dimorphic in their response to estradiol. J Neurosci 19:3316–3325
Adams MM, Oung T, Morrison JH, Gore AC 2001 Length of postovariectomy interval and age, but not estrogen replacement, regulate N-methyl-D-aspartate receptor mRNA levels in the hippocampus of female rats. Exp Neurol 170:345–356
Gibbs RB 1998 Impairment of basal forebrain cholinergic neurons associated with aging and long-term loss of ovarian function. Exp Neurol 151:289–302
Gibbs RB 1998 Levels of trkA and BDNF mRNA, but not NGF mRNA, fluctuate across the estrous cycle and increase in response to acute hormone replacement. Brain Res 810:294
Gibbs RB 1996 Fluctuations in relative levels of choline acetyltransferase mRNA in different regions of the rat basal forebrain across the estrous cycle: effects of estrogen and progesterone. J Neurosci 16:1049–1055(Jill M. Daniel, Jerielle L. Hulst and Je)
Abstract
The goal of the present study was to explore the effects of long-term hormone deprivation on the ability of subsequent estrogen replacement to affect cognition. Female rats, 12 months of age, underwent ovariectomies (n = 30) or sham surgeries (n = 10). Intact rats and 20 ovariectomized rats received cholesterol implants. Ten ovariectomized rats received implants containing 25% estradiol. Five months later, implants were replaced. Half of the ovariectomized rats with cholesterol implants received estradiol implants and half received new cholesterol implants. Rats with estradiol implants received new estradiol implants. Intact rats were ovariectomized and given estradiol implants. Beginning 1 wk later, working memory performance was assessed in an eight-arm radial maze across 24 d of acquisition and during eight additional trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices. Estradiol replacement initiated immediately after ovariectomy at either 12 or 17 months of age significantly improved performance during acquisition and delay trials, compared with control treatment. When estradiol replacement was initiated at 17 months of age, 5 months after ovariectomy, no enhancements were evident. Uteri of rats that experienced delayed estradiol replacement weighed significantly more than uteri of ovariectomized controls but significantly less than uteri of rats that received immediate estradiol replacement. Uterine weight negatively correlated with mean errors during acquisition. These results indicate that whereas chronic estradiol replacement regimens positively affect working memory in middle-aged animals when initiated immediately after ovariectomy, estradiol replacement is not effective when initiated after long-term hormone deprivation.
Introduction
THERE IS A GREAT deal of evidence in the basic and clinical literature supporting a role for estrogens in the modulation of cognitive function in mammals (for review, see Ref.1). However, the direction and extent of those effects remain controversial. Whereas many early studies indicated that hormone replacement therapy (HRT) could delay or alleviate age-related cognitive decline in postmenopausal women (for review, see Ref.2), recent results of the large Women’s Health Initiative Memory Study conducted by the National Institutes of Health indicated that replacement regimens consisting of chronic conjugated equine estrogens (3, 4) or chronic conjugated equine estrogens plus medroxyprogesterone (5, 6) did not improve cognition and actually increased risk of dementia. In attempts to reconcile these findings with those of previous studies, various factors need to be considered including regimen of replacement, the age of the subjects, and the number of years of estrogen deficiency at the time of initiation of HRT (7). For example, the women included in the Women’s Health Initiative Memory Study were at least 65 yr of age at the time of intervention and the majority had never used previous hormone therapy. Thus, these women were well beyond the age at which HRT is typically initiated to alleviate postmenopausal symptoms and most had experienced more than a decade of ovarian hormone deprivation before beginning HRT. It has been hypothesized that long-term ovarian hormone deprivation affects the efficacy of subsequent HRT in postmenopausal women (8).
In addition to the clinical studies, basic research has been conducted to elucidate the effects of estrogens on cognitive function. Studies using rodent models have demonstrated that estradiol positively affects hippocampal structure and function (9, 10, 11) and enhances performance on many hippocampal-dependent tasks (12, 13, 14, 15, 16, 17, 18, 19, 20, 21). However, most of the studies assessing the effects of estradiol on cognitive function using rodent models have used young adult females. Studies using middle-aged and older animals have yielded mixed results. For example, 1 wk of chronic estradiol replacement in 13-month-old rats had no effect on the acquisition of a T-maze active avoidance task (22) and had no effect on spatial memory in aged female rats as measured in a radial-arm maze (23). In contrast, 5 wk of chronic oral estradiol in ovariectomized middle-aged mice improved object recognition (24) and daily injections of estradiol to aged mice improved spatial reference memory, compared with controls (25). There is evidence of an interaction of age and dose of estradiol used in chronic replacement paradigms that may contribute to inconsistent results (26). In addition, there is preliminary evidence in the basic literature to support the hypothesis that estrogen replacement may lose its ability to positively affect cognition after a long-term period of hormone deprivation. Chronic estradiol replacement initiated six months after ovariectomy in middle-aged rats was not effective at ameliorating working memory deficits in a delayed-nonmatching-to-sample task in a T maze unless primed with cyclic injections of estradiol (27). It is unclear whether the chronic replacement regimen was largely ineffective due to the age of the rats or the long-term hormone deficiency at the time of initiation of the replacement. In another study, continuous estradiol replacement initiated within 3 months, but not 10 months after ovariectomy, improved performance of aged rats in a delayed-matching-to-position spatial memory task (19). However, there was no effect of hormonal status on trials after acquisition when the intertrial interval was increased, suggesting that the hormone-induced enhancement during acquisition of the delayed-matching-to-position task may not have been due to a selective enhancement of spatial memory. Additionally, the duration of estradiol replacement differed between groups. Thus, further exploration into the effects of ovarian hormone deprivation on the ability of future estrogen replacement to affect cognitive function is warranted based on the results of these studies as well as the need to address the inconsistencies in the clinical literature regarding the efficacy of HRT.
The goal of the current study was to determine the effects of long-term hormone deprivation on the ability of subsequent estradiol replacement to positively affect working memory and control for other factors that may affect performance, such as age and duration of estradiol exposure. We determined whether the efficacy of estradiol replacement in middle-aged animals as related to working memory is affected by the age at which it is initiated, the duration of exposure, and importantly the length of hormone deprivation before replacement. We previously demonstrated that a chronic replacement paradigm delivering constant physiological levels of estradiol enhanced acquisition of a working memory task in a radial-arm maze in early middle-aged rats when initiated immediately after ovariectomy (28). In the present experiment, we compared the ability of physiological levels of estradiol replacement initiated at different time points to affect working memory performance in a radial-arm maze in 17-month-old ovariectomized rats. The effects of estradiol replacement initiated immediately after ovariectomy at either 12 or 17 months of age or initiated at 17 months of age, 5 months after ovariectomy, were determined. We assessed the effects of our treatments during acquisition of the maze task. In addition, because the effects of estradiol on working memory may be most apparent when the working memory load increases (14), we also assessed the effects of our treatments on delay trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices.
Materials and Methods
Subjects
Female Long-Evans rats, retired breeders, were purchased from Harlan Sprague Dawley, Inc. (Indianapolis, IN). Rats were housed individually in a temperature-controlled vivarium under a 12-h light, 12-h dark cycle (lights on at 0700 h) and were allowed free access to water and food for 2 wk. Animal care was in accordance with guidelines set by the National Institutes of Health Guide for the Care and Use of Laboratory Animals (1996).
Treatments
At approximately 12 months of age, females underwent either ovariectomies (n = 30) or sham surgeries (n = 10) while under anesthesia induced by injections of ketamine (100 mg/kg ip; Bristol Laboratories, Syracuse, NY) and xylazine (7 mg/kg ip; Miles Laboratories, Shawnee, KS). At the time of the surgeries, 5-mm SILASTIC brand capsules (0.58-in. inner diameter and 0.077-in. outer diameter; Dow Corning, Midland, MI) containing cholesterol (Sigma Chemical Co., St. Louis, MO) were implanted sc on the dorsal aspect of the neck in 20 of the rats that had undergone ovariectomies and all 10 of the rats that had undergone sham surgeries. Capsules containing 25% 17-estradiol (Sigma) diluted with cholesterol were implanted in 10 of the ovariectomized rats. Capsules of these dimension and estradiol concentration maintain circulating estradiol titers at 20–25 pg/ml, a low level typical of diestrus (19, 29, 30).
These treatments continued for a 5-month period during which time animals were weighed and handled once per week. Additionally, capsules were manually manipulated each week to prevent the build-up of scar tissue. We did not replace the capsules during this 5-month period because in a previous experiment in our laboratory, capsules of the same dimensions and estradiol concentration remained active for a similar time period (Daniel, J. M., unpublished data). During the final 2 wk of the 5-month period, daily vaginal smears were collected by lavage from each of the animals to confirm neuroendocrine status.
When rats were approximately 17 months of age, they all underwent a second set of surgeries while under anesthesia induced by injections of ketamine and xylazine. Animals that had previously undergone sham surgeries were ovariectomized and previously ovariectomized rats underwent sham surgeries. Additionally, all SILASTIC brand capsules were removed. Visual inspection of these capsules revealed that all had maintained their integrity. New capsules were implanted. Ovariectomized rats that had previously received estradiol capsules were implanted with new estradiol capsules. Half of the originally ovariectomized animals that had received cholesterol capsules were implanted with new cholesterol capsules. The other half of the originally ovariectomized animals that had received cholesterol capsules were implanted with estradiol capsules. Finally, the 10 newly ovariectomized rats that had originally received cholesterol capsules were implanted with estradiol capsules. Surgeries and treatments for each of the four groups are summarized as follows: Ch/Ch group (12 months of age: ovariectomy; cholesterol implant; 17 months of age: sham surgery; cholesterol implant); E/E group (12 months of age: ovariectomy; estradiol implant; 17 months of age: sham surgery; estradiol implant); Ch/E group (12 months of age: ovariectomy; cholesterol implant; 17 months of age: sham surgery; estradiol implant); Intact/E group (12 months of age: sham surgery; cholesterol implant; 17 months of age: ovariectomy; estradiol implant).
Behavioral testing
Radial maze acquisition.
Rats were allowed 1 wk of recovery time after the second set of surgeries before training on the radial maze began. Beginning 3 d before training, all rats were placed on diets to maintain body weights at 90% of their free-feeding weights. Animals were trained to obtain food rewards (Froot Loops; Kellogg Co., Battle Creek, MI) from the arms of an elevated eight-arm radial maze. Each day during the week before training, several Froot Loops were placed in the cage of each rat. One day before the beginning of training, each rat was placed in the maze for a 15-min acclimation period. Froot Loops were sprinkled throughout the maze and the rat was allowed to travel freely throughout the maze with all arms opened. The maze was purchased from SciPro, Inc. (Sanborn, NY) and consisted of gray metal floors and clear Plexiglas walls. The arms (15 cm wide x 45 cm long x 23 cm high) extended out from an octagonal center (48 cm across). Food receptacles, which were opaque in color to block the view of the food reward, were attached to the ends of each arm. The maze was located in the center of a 3 x 3 m room. Extramaze cues including lighting fixtures, a door, and electrical outlets as well as geometric shapes attached to walls were visible from the maze.
To begin each trial, the rat was placed in the center compartment with all doors leading to the arms opened. Doors remained opened throughout the trial. The rat was allowed to enter any of the eight arms. The experimenter, who was seated in the room at a fixed location approximately 1 m from the maze, recorded arm choices. An arm choice was scored if the rat traveled halfway down the length of an arm. The animal was allowed to choose arms in any order until all arms were visited or until 5 min elapsed. An error was scored if a rat reentered an arm previously visited. Arm-choice accuracy was measured by the number of errors made in the first eight visits. If a rat did not enter eight arms in a trial, arms never visited were counted as errors. However, after the first block of trials, this rarely occurred. A record of the arm-choice order was kept for each rat and analyzed by the experimenter on a daily basis to determine whether any rats were using response algorithms such as adjacent-arm strategies or other patterned responses across trials to solve the maze. No such patterns were revealed. As a measure of the speed at which an animal traversed the maze, total number of arms entered per minute (total number of arms entered in a trial/time to complete the trial) was calculated. Maze training took place at least 5 d/wk. Each animal received one trial per day across 24 d of acquisition.
Delay trials.
After the 24-d acquisition period, animals received eight additional daily trials in which a 2.5-h delay was imposed between the fourth and fifth arm choices. During the delay trials, the animal was removed from the maze after it had made four correct arm choices and placed in a holding cage for period of 2.5 h. It then was returned to the center compartment of the maze with all arms opened and allowed to choose arms in any order until all arms were visited or until 5 min had elapsed. Errors made during the predelay and postdelay periods were scored separately. Postdelay errors were broken down into retroactive errors and proactive errors (31). A retroactive error was the first reentry into an arm already visited before the delay. A proactive was a reentry into an arm already visited in the postdelay period.
Vaginal cytology, uterine weight
To identify neuroendocrine status during the second set of hormone treatments, daily vaginal smears were collected from all animals during a 2-wk period between the completion of behavioral testing and time of being killed. These samples, as well as those that were collected at the end of the initial 5-month hormone treatment period, were stained with Harris hematoxylin solution and eosin and microscopically examined. The following criteria were used to identify estrous stage: estrus was characterized by a predominance of cornified cells; diestrus was characterized by the predominance of leukocytes; and proestrus was characterized by the predominance of nucleated epithelial cells (32, 33).
Two weeks after the completion of behavioral testing, animals were anesthetized with ketamine and xylazine. After animals were killed, uteri were removed and weighed (see Fig. 1 for a summary of treatment and testing schedule).
Statistical analyses
One-way ANOVAs, with treatment as the factor, were used to test for differences in body weight and uterine weight. Data collected from acquisition trials and from delay trials were analyzed by two-way ANOVAs (treatment x 4-d block) with repeated measures on block. Data from individual blocks during acquisition were analyzed by one-way ANOVA (treatment). If a significant main effect was revealed in any of these analyses (P < 0.05), then Duncan’s multiple range post hoc test (P < 0.05) was applied.
Results
Several of the rats died during the course of the experiment. Data from animals that survived at least through completion of the acquisition trials were included in the analyses of body weight data and acquisition data (Ch/Ch, n = 9; E/E, n = 8; Ch/E, n = 8; Intact/E, n = 10). One additional animal died (Ch/Ch) before the completion of the delay trials and another died (Intact/E) before measurement of uterine weight.
Behavioral testing results
Acquisition.
As illustrated in Fig. 2, estradiol replacement initiated in rats immediately after ovariectomy either at 12 or 17 months of age significantly improved performance during acquisition of a working memory task in a radial-arm maze, compared with ovariectomized controls. However, when estradiol replacement was initiated at 17 months of age after 5 months of hormone deprivation, no such enhancement was revealed. A significant main effect of treatment was revealed for number of incorrect choices in the first eight arm visits (F3,31 = 3.033, P < 0.05). Post hoc analyses revealed that females ovariectomized at 12 months that received immediate, continuous estradiol replacement (E/E) and those ovariectomized at 17 months that received immediate, continuous estradiol replacement (Intact/E) achieved significantly fewer errors than those control animals ovariectomized at 12 months that received continuous cholesterol treatment (Ch/Ch; Fig. 2A). Females ovariectomized at 12 months but whose estradiol treatment was delayed until 1 wk before training when they were 17 months of age (Ch/E) did not significantly differ from any of the other groups. There was a significant effect of block (F5,155 = 22.79, P < 0.0001), indicating that the accuracy of all groups improved over time. There was no interactive effect of treatment and block. Figure 2B illustrates the effects of various estradiol treatments on acquisition over 24 d, presented in blocks of 4 d.
A significant effect of treatment for number of incorrect choices for individual blocks of trials emerged only for the final block (F3,31 = 3.88, P < 0.05), indicating that the effects of our treatments were most apparent at the end of the acquisition period when the animals were familiar with the demands of the task. Post hoc analyses revealed that E/E and Intact/E females achieved significantly fewer errors in the last block of trials than Ch/Ch control animals. Ch/E females did not significantly differ from any of the other groups.
There was a significant effect of block (F5,155 = 29.45, P < 0.0001) but no significant effect of treatment or interactive effect of treatment and block for mean number of arms entered per minute (data not shown), indicating that whereas arm entry rate increased for all groups across trials, this effect did not differ across treatment groups.
Delay trials.
As illustrated in Fig. 3, estradiol replacement initiated at either 12 or 17 months of age immediately after ovariectomy but not when initiated at 17 months of age after long-term hormone deprivation resulted in significantly fewer postdelay retroactive errors, compared with control treatment. There were no main or interactive effects for the predelay period because very few errors were made by any of the groups (data not shown). There was a significant main effect of treatment for the number of postdelay retroactive errors (F3,30 = 3.14, P < 0.05). Post hoc analyses revealed that E/E and Intact/E females made significantly fewer retroactive errors than Ch/Ch controls (Fig. 3A). Ch/E females did not significantly differ from any of the other groups. The effect of block approached the level of statistical significance (F3,30 = 4.089, P = 0.052). There was no significant interactive effect of treatment and block for retroactive errors. The effect of treatment on number of postdelay proactive errors approached but did not reach the level of statistical significance (F3,30 = 2.68, P = 0.065; Fig. 3B). There was neither a significant main effect of block nor an interactive effect of block and treatment for number of proactive errors.
Efficacy of initial hormone treatments
The efficacy of the initial 5 months of hormone treatments was assessed by two measures, body weight and examination of vaginal cytology. There was a significant effect of hormone treatment on body weight (F2,34 = 3.38, P < 0.05). Post hoc analyses revealed that at the end of the initial hormone treatments, the mean weight of the ovariectomized rats with cholesterol implants (347 g ± 9.3 SEM) was significantly more than the mean weight of the ovariectomized rats with estradiol implants (316 g ± 8.4 SEM). The mean weight of the gonadally intact rats (321 g ± 10.6 SEM) did not significantly differ from either of the other groups.
Light microscopic examination of vaginal smears collected during the final 2 wk of the initial hormone treatment period revealed that the hormone treatments were effective. The smears of all ovariectomized animals that had cholesterol implants during the initial 5-month treatment period were characterized by a predominance of leukocytes. Smears of the ovariectomized animals that had estradiol implants were characterized by a predominance of cornified cells with some nucleated epithelial cells. All of the smears of the gonadally intact rats showed changes across days, although there was evidence of disruptions in the cycles. The length of the cycles, as determined by the number of days between consecutive estrous stages (22), ranged from 5 to 8 d. The cycles of all but one of the intact rats included a proestrous stage.
Efficacy of final hormone treatments
The efficacy of the final period of hormone treatments was assessed by two measures, examination of vaginal cytology collected during the final 2 wk before the animals were killed and uterine weights. Examination of vaginal cytology indicated that the hormone treatments were effective. The vaginal smears of all of the ovariectomized animals that had cholesterol implants for both periods of hormone treatments (Ch/Ch) were characterized by a predominance of leukocytes. The smears of all of the ovariectomized animals that had estradiol implants for both periods of hormone treatments (E/E) were characterized by a predominance of cornified cells with some nucleated epithelial cells. The smears of all of the animals that were not ovariectomized until the beginning of the second period of hormone treatments at which time they received estradiol implants (Intact/E) were also characterized by a predominance of cornified cells with some nucleated epithelial cells. The smears of one of the ovariectomized animals that were treated with cholesterol for 5 months before the initiation of estradiol treatment (Ch/E) were characterized by a predominance of leukocytes. However, estradiol treatment was considered effective in this rat based on our other measure, uterine weight. The rest of the smears from this group were characterized by a predominance of cornified cells with some nucleated epithelial cells (see Fig. 4 for representative smears from each group).
There was a main effect of hormone treatment on uterine weight (F3,32 = 28.19, P < 0.001). Post hoc analyses revealed that the uteri of ovariectomized rats that had cholesterol implants for both periods of hormone treatments (Ch/Ch) weighed significantly less than the other three groups (see Fig. 5). Additionally, the uteri of the rats that experienced 5 months of hormone deprivation before initiation of estradiol treatment (Ch/E) weighed significantly less than the two groups of rats that had estradiol treatments initiated immediately after ovariectomies (E/E and Intact/E), indicating that estradiol replacement initiated after a long period of deprivation was less effective at increasing uterine weight than was replacement initiated immediately after ovariectomy. In light of this finding, correlation analysis was conducted to determine whether there was a relationship between the ability of our various treatments to affect uterine weight and cognitive function. Results revealed a significant negative correlation (r = –0.414, P < 0.01) between uterine weight and the mean number of errors committed during trials in the acquisition period (see Fig. 6).
Finally, because there were differences in body weights between the E/E and Ch/E animals at the beginning of the second treatment period, we wanted to control for the possibility that the attenuated effect of estradiol in the Ch/E group, compared with the other estradiol groups, was due to lower levels of circulating estradiol as a result of increased body weight. We therefore did correlation analyses including the three groups that received estradiol treatment to determine whether there was a relationship between body weight at the end of the first treatment period and mean number of errors during acquisition and/or between body weight at the end of the first treatment period and uterine weight at the time the animals were killed. No significant relationships were revealed between body weight and these measures, indicating that differences in body weight did not affect the efficacy of the various estradiol treatments.
Discussion
The primary finding of the present study was that estradiol replacement loses its ability to positively affect working memory in middle-aged rats when it is initiated after a long period of ovarian hormone deprivation. Whereas chronic estradiol replacement was effective in ameliorating ovariectomy-induced deficits in working memory when initiated immediately after ovariectomy at either 12 or 17 months of age, it was not effective when initiated after a 5-month period of hormone deprivation. Interestingly, factors such as the length of estradiol replacement before training and the point during the middle-age period at which the animals were ovariectomized did not influence the efficacy of estradiol replacement as related to its effects on working memory. Replacement paradigms initiated 5 months or 1 wk before acquisition training were equally effective at enhancing working memory performance if initiated immediately after ovariectomy. In contrast, estradiol replacement initiated 1 wk before acquisition but 5 months after ovariectomy had no effect, compared with control treatment. Thus, the present study reveals that the ability of estradiol replacement to affect working memory performance in ovariectomized middle-aged animals is affected by the timing of its initiation. Future studies will be needed to determine whether it is necessary to initiate estradiol replacement immediately after ovariectomy or, as suggested by a prior study in which estradiol replacement improved performance on a delayed-matching-to-position task when initiated 3 months but not 10 months after ovariectomy (19), there is a window of opportunity after the loss of ovarian function in which estradiol replacement must be initiated to be effective. Additionally, the length of this window may vary with the regimen of hormone treatment (27).
The possibility exists that in the present study, group differences in performance during the acquisition trials as well as the delay trials were not due to direct effects of our treatments on working memory but were due to effects of our treatments on other requirements of the tasks. However, our data suggest otherwise. There were no differences in the rates at which each of the groups traversed the maze, suggesting similar motivation and attention levels among the groups. Additionally, there were no differences in performance in the early blocks of trials of the acquisition period during which time the animals were learning the requirements of the task. Group differences did emerge in the final block of trials when the requirements of the tasks were familiar, and therefore differences were more likely due to specifically to differences in working memory.
The mechanism by which estradiol loses its ability to affect working memory after a period of hormone deprivation is unknown. However, there is evidence that estradiol differentially affects brain areas important for cognition in young and older animals. For example, a series of studies indicate that there are different mechanisms of hippocampal plasticity in young and aged female rats. In young animals, estradiol regulates hippocampal morphology as evidenced by increases in CA1 spine and synapse density (34) and concurrent increases in N-methyl-D-aspartate (NMDA) receptor binding levels (9, 11). There is a blunted response to estradiol replacement in the hippocampus of aged female rats, compared with young animals. In aged animals, estradiol failed to increase hippocampal spine density, although it did up-regulate synaptic NMDA receptor subunit 1 (NR1), the obligatory subunit of the NMDA receptor (35). In contrast to the up-regulation of NR1 in the young animals, which was characterized by increases that were in proportion to new spines, the estradiol-induced up-regulation of NR1 in the aged animals was characterized by increases in NR1 per synapse. This decreased responsiveness of the aged animals to the effects of estradiol may be related to a decrease in CA1 estrogen receptor- levels that is evident in aged animals (36). As in CA1, the response of the dentate gyrus to estradiol also changes with age. In aged female rats, acute estradiol exposure results in increased spine density of dentate granule cells, even after months of hormonal deprivation (37). No such effects are evident in young adult female rats.
It will be important to determine whether the differences in the hippocampal response to estradiol in young and aged animals are due to changes in the hormonal milieu experienced during aging or are due to the aging process independent of the effects of hormonal changes. There have been attempts to differentiate the influence of age and changes in hormonal status on the effects of estradiol in the hippocampus. The results of a study that analyzed the effects of postovariectomy interval and estradiol replacement on levels of hippocampal NMDA receptor mRNA in young, middle-aged, and aged rats revealed age and length of ovarian hormone deprivation had significant impacts on NMDA mRNA levels with little impact of estradiol replacement (38). However, the replacement paradigms did not begin until at least 1 month after ovariectomy.
Future studies need to be completed to determine whether estrogen replacement initiated in middle-age when decreases in estrogen levels first occur would prevent the age-related change in the response to estrogen in aged animals. Consistent with this possibility is a report that long-term ovarian hormone deprivation has negative effects on basal forebrain cholinergic neurons that go beyond the effects of aging alone (39). Decreases in both choline acetyltransferase and trkA mRNA in the medial septum and nucleus basalis magnocellularis were evident in middle-aged rats killed 6 months, but not 3 months, after ovariectomy relative to age-matched, gonadally intact controls. In contrast to estradiol-induced increases in levels of choline acetyltransferase and trkA mRNA in these areas exhibited by ovariectomized young animals (40, 41), short-term estradiol replacement initiated 6 months after ovariectomy in middle-aged animals had no effect.
In the present study, the finding that our different estradiol treatments differentially affected uterine weight is intriguing. Although all three regimens of estradiol replacement significantly increased uterine weight compared with control treatment, the replacement paradigm that followed a long period of deprivation resulted in significantly less increases than the two paradigms that were initiated immediately after ovariectomy. Additionally, uterine weight was negatively correlated with the number of errors made during the acquisition period, indicating a relationship between the ability of estradiol to affect performance on measures of learning and memory and its ability to affect estradiol-sensitive peripheral tissues. Thus, although the mechanism remains to be determined, the effect of long-term deprivation on the efficacy of estradiol replacement may extend beyond its effects on the brain and cognition.
In conclusion, the results of the present study support the hypothesis that long-term ovarian hormone deprivation negatively affects the ability of subsequent estradiol replacement to enhance performance on tasks of learning and memory. The mechanism by which the brain becomes less susceptible to the effects of estradiol after ovarian hormone deprivation remains to be determined. However, these results have implications toward reconciling the inconsistencies in the clinical literature regarding the effects of HRT on cognitive function and suggest that the timing of the initiation of HRT may be a crucial factor in determining its efficacy.
Acknowledgments
We thank Misha Rehage, Kelvin Jenkins, and Elizabeth Daray for their assistance in the collection of behavioral data.
Footnotes
This work was supported by National Science Foundation Grant 0423331 (to J.M.D.).
First Published Online October 20, 2005
Abbreviations: HRT, Hormone replacement therapy; NMDA, N-methyl-D-aspartate; NR1, NMDA receptor subunit 1.
Accepted for publication October 10, 2005.
References
Dohanich GP 2002 Gonadal steroids, learning and memory. In: Pfaff DW, Arnold AP, Etgen AM, Farbach SE, Rubin RT, eds. Hormones, brain and behavior. Vol 1. San Diego: Academic Press; 265–327
Sherwin BB 2003 Estrogen and cognitive functioning in women. Endocr Rev 24:133–151
Espeland MA, Rapp SR, Shumaker SA, Brunner R, Manson JE, Sherwin BB, Hsia J, Margolis KL, Hogan PE, Wallace R, Dailey M, Freeman R, Hays J 2004 Conjugated equine estrogens and global cognitive function in postmenopausal women: Women’s Health Initiative Memory Study. JAMA 291:2959–2968
Shumaker SA, Legault C, Kuller L, Rapp SR, Thal L, Lane DS, Fillit H, Stefanick ML, Hendrix SL, Lewis CE, Masaki K, Coker LH 2004 Conjugated equine estrogens and incidence of probable dementia and mild cognitive impairment in postmenopausal women: Women’s Health Initiative Memory Study. JAMA 291:2947–2958
Rapp SR, Espeland MA, Shumaker SA, Henderson VW, Brunner RL, Manson JE, Gass ML, Stefanick ML, Lane DS, Hays J, Johnson KC, Coker LH, Dailey M, Bowen D 2003 Effect of estrogen plus progestin on global cognitive function in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289:2663–2672
Shumaker SA, Legault C, Rapp SR, Thal L, Wallace RB, Ockene JK, Hendrix SL, Jones III BN, Assaf AR, Jackson RD, Kotchen JM, Wassertheil-Smoller S, Wactawski-Wende J 2003 Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289:2651–2662
Gleason CE, Carlsson CM, Johnson S, Atwood C, Asthana S 2005 Clinical pharmacology and differential cognitive efficacy of estrogen preparations. Ann NY Acad Sci 1052:93–115
Pinkerton JV, Henderson VW 2005 Estrogen and cognition, with a focus on Alzheimer’s disease. Semin Reprod Med 23:172–179
Woolley CS, Weiland NG, McEwen BS, Schwartzkroin PA 1997 Estradiol increases the sensitivity of hippocampal CA1 pyramidal cells to NMDA receptor-mediated synaptic input: correlation with dendritic spine density. J Neurosci 17:1848–1859
Gabor R, Nagle R, Johnson DA, Gibbs RB 2003 Estrogen enhances potassium-stimulated acetylcholine release in the rat hippocampus. Brain Res 962:244–247
Daniel JM, Dohanich GP 2001 Acetylcholine mediates the estrogen-induced increase in NMDA receptor binding in CA1 of the hippocampus and the associated improvement in working memory. J Neurosci 21:6949–6956
Daniel JM, Fader AJ, Spencer AL, Dohanich GP 1997 Estrogen enhances performance of female rats during acquisition of a radial arm maze. Horm Behav 32:217–225
Luine VN, Richards ST, Wu VY, Beck KD 1998 Estradiol enhances learning and memory in a spatial memory task and effects levels of monoaminergic neurotransmitters. Horm Behav 34:149–162
Bimonte HA, Denenberg VH 1999 Estradiol facilitates performance as working memory load increases. Psychoneuroendocrinology 24:161–173
Fader AJ, Johnson PE, Dohanich GP 1999 Estrogen improves working but not reference memory and prevents amnestic effects of scopolamine of a radial-arm maze. Pharmacol Biochem Behav 62:711–717
Fader AJ, Hendricson AW, Dohanich GP 1998 Estrogen improves performance of reinforced T-maze alternation and prevents the amnestic effects of scopolamine administered systemically or intrahippocampally. Neurobiol Learn Mem 69:225–240
Daniel JM, Winsauer PJ, Brauner IN, Moerschbaecher JM 2002 Estrogen improves response accuracy and attenuates the disruptive effects of 9-THC in ovariectomized rats responding under a multiple schedule of repeated acquisition and performance. Behav Neurosci 116:989–998
Gibbs RB 1999 Estrogen replacement enhances acquisition of a spatial memory task and reduces deficits associated with hippocampal muscarinic receptor inhibition. Horm Behav 36:222–233
Gibbs RB 2000 Long-term treatment with estrogen and progesterone enhances acquisition of a spatial memory task by ovariectomized aged rats. Neurobiol Aging 21:107–116
Luine VN, Jacome LF, MacLusky NJ 2003 Rapid enhancement of visual and place memory by estrogens in rats. Endocrinology 144:2836–2844
Sandstrom NJ, Williams CL 2001 Memory retention is modulated by acute estradiol and progesterone replacement. Behav Neurosci 115:384–393
Savonenko AV, Markowska AL 2003 The cognitive effects of ovariectomy and estrogen replacement are modulated by aging. Neuroscience 119:821–830
Luine V, Rodriguez M 1994 Effects of estradiol on radial arm maze performance of young and aged rats. Behav Neural Biol 62:230–236
Fernandez SM, Frick KM 2004 Chronic oral estrogen affects memory and neurochemistry in middle-aged female mice. Behav Neurosci 118:1340–1351
Frick KM, Fernandez SM, Bulinski SC 2002 Estrogen replacement improves spatial reference memory and increases hippocampal synaptophysin in aged female mice. Neuroscience 115:547–558
Foster TC, Sharrow KM, Kumar A, Masse J 2003 Interaction of age and chronic estradiol replacement on memory and markers of brain aging. Neurobiol Aging 24:839–852
Markowska AL, Savonenko AV 2002 Effectiveness of estrogen replacement in restoration of cognitive function after long-term estrogen withdrawal in aging rats. J Neurosci 22:10985–10995
Daniel JM, Olvet KR, Dohanich GP 1999 Effects of prior and continuous estrogen exposure on performance during acquisition of a radial arm maze by young adult and middle-aged female rats. Soc Behav Neuroendocr (Abstract)
Luine VN 1997 Steroid hormone modulation of hippocampal dependent spatial memory. Stress 2:21–36
Singh M, Meyer EM, Millard WJ, Simpkins JW 1994 Ovarian steroid deprivation results in a reversible learning impairment and compromised cholinergic function in female Sprague-Dawley rats. Brain Res 644:305–312
Chappell J, McMahan R, Chiba A, Gallagher M 1998 A re-examination of the role of basal forebrain cholinergic neurons in spatial working memory. Neuropharmacology 37:481–487
Montes GS, Luque EH 1988 Effects of ovarian steroids on vaginal smears in the rat. Acta Anat (Basel) 133:192–199
Marcondes FK, Bianchi FJ, Tanno AP 2002 Determination of the estrous cycle phases of rats: some helpful considerations. Braz J Biol 62:609–614
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
Adams MM, Shah RA, Janssen WG, Morrison JH 2001 Different modes of hippocampal plasticity in response to estrogen in young and aged female rats. Proc Natl Acad Sci USA 98:8071–8076
Adams MM, Fink SE, Shah RA, Janssen WG, Hayashi S, Milner TA, McEwen BS, Morrison JH 2002 Estrogen and aging affect the subcellular distribution of estrogen receptor-alpha in the hippocampus of female rats. J Neurosci 22:3608–3614
Miranda P, Williams CL, Einstein G 1999 Granule cells in aging rats are sexually dimorphic in their response to estradiol. J Neurosci 19:3316–3325
Adams MM, Oung T, Morrison JH, Gore AC 2001 Length of postovariectomy interval and age, but not estrogen replacement, regulate N-methyl-D-aspartate receptor mRNA levels in the hippocampus of female rats. Exp Neurol 170:345–356
Gibbs RB 1998 Impairment of basal forebrain cholinergic neurons associated with aging and long-term loss of ovarian function. Exp Neurol 151:289–302
Gibbs RB 1998 Levels of trkA and BDNF mRNA, but not NGF mRNA, fluctuate across the estrous cycle and increase in response to acute hormone replacement. Brain Res 810:294
Gibbs RB 1996 Fluctuations in relative levels of choline acetyltransferase mRNA in different regions of the rat basal forebrain across the estrous cycle: effects of estrogen and progesterone. J Neurosci 16:1049–1055(Jill M. Daniel, Jerielle L. Hulst and Je)