INTRODUCTION

Gender differences in anxiety and depressive disorders favor males. Women are more likely than men to develop anxiety disorders (Seeman, 1997). Women, compared to men, are twice as likely to experience depression and their depressive episodes are longer-lasting and recur more often (Earls, 1987; Nolen-Hoeksema, 1987). These gender differences suggest that sex-linked factors, such as gonadal steroids, may influence their etiology and/or expression (Young, 1998; Young and Korszun, 2002).

The ovarian hormone estradiol (E2) may influence incidence of, or negative symptomology associated with, mood disorders. Gender differences in depression emerge at puberty and disappear after menopause (Kessler and Walters, 1998; Lewinsohn et al, 1998; Weissman and Olfson, 1995). In contrast, the incidence of generalized anxiety disorder is doubled in older, postmenopausal women (Bebbington, 1978; Jenkins, 1987; Weissman and Klerman, 1977; Wittchen and Hoyer, 2001). There are reports of negative symptomology associated with anxiety and depression being greater perimenstrually, postpartum, and after menopause or oophorectomy, when women's E2 levels would be expected to be lower (Torizuka et al, 2000). Among depressed women, plasma E2 levels are significantly lower than is observed in nondepressed women (Young et al, 2000). E2-replacement therapy to perimenopausal women with depression may provide mood benefits (Cohen et al, 2003; Schmidt et al, 2000; Soares et al, 2001). As well, postmenopausal women taking E2-replacement therapy had lower depression scores on the Profile of Mood States task than did age-matched women who were not on E2-replacement therapy (Miller et al, 2002). However, some women with anxiety disorders report less anxiety when E2 levels are low and/or stable (Schmidt et al, 1998). Although these data suggest that changes in E2 levels may influence anxiety and/or depression, E2's role is not completely understood.

Evidence from animal models also suggests that E2 has variable effects on anxiety or depressive behavior of rodents. Rats show less anxiety and depressive behavior during the proestrous phase of the estrous cycle, when E2 levels are high, compared to the diestrous phase, when E2 levels are low (Diaz-Veliz et al, 1997; Frye et al, 2000; Marcondes et al, 2001; Mora et al, 1996). Ovariectomy (ovx), removal of the primary source of E2, the ovaries, increases anxiety and depressive behavior, and acute, subchronic, or chronic physiological E2-replacement reduces anxiety and depressive behavior of rodents (Bernardi et al, 1989; Bowman et al, 2002; Diaz-Veliz et al, 1997; Estrada-Camarena et al, 2003; Frye and Walf, 2004a; Hilakivi-Clarke, 1996; Luine et al, 1998; Marcondes et al, 2001; Mora et al, 1996; Nomikos and Spyraki, 1988; Rachman et al, 1998; Slater and Blizard, 1976). However, longer E2-replacement regimen and/or a higher E2 dosages may not reduce anxiety and/or depressive behavior (Diaz-Veliz et al, 1997, 2000; Galea et al, 2002; Martinez-Mota et al, 2000; Mora et al, 1996; Morgan and Pfaff, 2001, 2002; Stoffel and Craft, 2003). Thus, E2's effects on anxiety and/or depressive behavior of rodents may depend upon the E2 dosage and/or regimen.

Estradiol may have effects to alter the function of the hypothalamic–pituitary–adrenal axis (HPA). Proestrous rats have increased basal and stress-induced plasma corticosterone levels compared to rats in other stages of the estrous cycle (Carey et al, 1995; Figueiredo et al, 2002; Frye and Bayon, 1999; Raps et al, 1971; Viau and Meaney, 1991). Administration of E2 to ovx rats may reduce basal and/or stress-induced corticosterone levels, but administration of higher E2 dosages may increase corticosterone levels (Burgess and Handa, 1992; Carey et al, 1995; Dayas et al, 2000; Kitay, 1963; McCormick et al, 2002; Redei et al, 1994). Differences in HPA reactivity may underlie some affective responses of rodents (Conrad et al, 1997; Maccari et al, 2003; Welberg et al, 2000), but whether E2's effects on anxiety and/or depressive behavior may require the HPA is not established.

Given E2's influence on affective behavior and the HPA, whether E2 may mitigate these behavioral processes in part through actions involving the HPA was of interest. We investigated this by testing the following hypotheses. First, if E2 mediates anxiety (open field, elevated plus maze) and/or depressive (forced swim test) behavior, then E2 should dose-dependently alter central entries in the open field, open arm time in the plus maze, and/or decrease immobility in the forced swim test. Second, if E2's effects on anxiety or depressive behavior are mediated by the HPA, then increasing HPA activity with restraint stress before behavioral testing should alter effects of E2 for anxiety or depressive behavior and corticosterone levels. Third, if E2's effects on anxiety and depressive behavior require HPA feedback, then adrenalectomy (ADX) should attenuate E2's dose-dependent effects on anxiety and depressive behavior. Fourth, if corticosterone mediates E2's effects on anxiety and depressive behavior, then corticosterone replacement would be expected to shift effects of E2 in ADX rats. To test these hypotheses, two experiments were performed. In experiment 1, ovx rats were administered different dosages of E2 (0, 2, 5, 10, 20, or 50 μg) 48 h before testing in anxiety or depression tasks. Some rats were restraint stressed before testing to determine the effects of activating the HPA on anxiety and depressive behavior, plasma corticosterone and E2 levels. In experiment 2, ovx rats were sham surgerized or ADX and replaced with vehicle, a low, or high dose of corticosterone in their drinking water and administered 0, 10, or 50 μg E2 48 h before testing. Effects on anxiety and depressive behavior, plasma corticosterone and E2 levels were examined.

MATERIALS AND METHODS

These methods were preapproved by the Institutional Animal Care and Use Committee at SUNY, Albany.

Animals and Housing

Female Long–Evans rats (N=284), approximately 55 days old, were obtained from the breeding colony at SUNY, Albany (original stock from Taconic Farms, Germantown, NY). Rats were group housed (4–5 per cage) in polycarbonate cages (45 × 24 × 21 cm) in a temperature-controlled room (21±1°C) in The Laboratory Animal Care Facility. The rats were maintained on a 12/12 h reversed light cycle (lights off 0800 hours) with continuous access to Purina Rat Chow and tap water in their cages. To prevent wasting, ADX and sham ADX rats also had continuous access to 0.09% saline in their cages.

Surgery

All rats were ovariectomized (ovx) under Rompun (12 mg/kg; Bayer Corp., Shawnee Mission, KS) and Ketaset (80 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA) anesthesia 1 week prior to testing. Some rats (n=118) were also bilaterally ADX, 7–10 days post ovx and 1 week before testing.

ESTRADIOL-PRIMING

In experiment 1, 48 h before testing, ovx rats were subcutaneously administered the following dosages of 17β-E2 (Sigma Chemical Co., St Louis, MO) in sesame oil vehicle: 0, 2, 5, 10, 20, or 50 μg E2. In experiment 2, ovx rats were subcutaneously administered 0, 10, or 50 μg E2 48 h before testing. This regimen of E2 48 h prior to testing was used because it mimics the gradual increase in plasma E2 levels during the estrous cycle and is commonly employed to induce sexual receptivity in rats (Frye et al, 1998). We have also used this regimen with success in the past to examine E2's modulation of affective behavior of ovx rats (Frye and Walf, 2004a; Walf and Frye, 2003; Walf, Rhodes, and Frye, 2004)

Restraint Stress

In experiment 1, immediately before behavioral testing in the open field, elevated plus maze, or forced swim task, some rats from each E2 condition were placed in a plexiglas animal restrainer under a 60-Watt light for 20 min or were not stressed before testing (Frye and Orecki, 2002a, 2002b).

ADX Conditions and Corticosterone Replacement

To determine completeness of ADX surgery, rats were restrained for 20 min and tail blood was collected from rats 24 h following ADX. Plasma corticosterone levels were determined for each rat. Rats with plasma corticosterone levels above typical basal levels (greater than 2.0 μg/dl) were considered to have incomplete ADX and comprised the sham surgerized condition. Rats were considered to be completely ADX if their plasma corticosterone levels were less than 2.0 μg/dl. These ADX rats were randomly assigned to receive vehicle (0.09 % sodium chloride in water and <0.01% ethyl alcohol), 25 μg/ml, or 250 μg/ml corticosterone in the same vehicle in their drinking water. These dosages were based on previous findings that replacement of 25 μg/ml corticosterone in drinking water produces basal levels of plasma corticosterone and 250 μg/ml corticosterone produces stress-like plasma corticosterone levels (Glavas et al, 2001; McCormick et al, 2001). Daily average intake of corticosterone were determined for each cage of rats, and basal plasma corticosterone levels were determined in rats 4 days after replacement began, to determine effectiveness of replacement regimen used. ADX rats administered vehicle (ADX+veh group) had corticosterone levels (mean±SD) below basal levels typically observed in intact rats (0.7±1.1 μg/dl). ADX rats administered 25 μg/ml corticosterone (ADX+25 CORT) had corticosterone levels similar to the sham ADX group (4.3±9.3 and 5.2±9.4 μg/dl, respectively) and ADX rats administered 250 μg/ml corticosterone (ADX+250 CORT) had plasma corticosterone levels of 20.5±16.8 μg/dl.

Behavioral Testing

Most rats in experiment 1 were tested in all three of the following tasks in the same order once a week for 3 weeks. There were no observable differences in behavior of rats that were only tested in one task, compared to those that were tested in all three tasks. In experiment 2, all rats were tested once a week for 3 weeks so that each rat was tested in the open field, elevated plus maze, and forced swim task.

Open field

Rats were placed in a brightly lit open field (76 × 57 × 35 cm) and observed for 5 min, while the number of central and peripheral squares entered was recorded (Frye et al, 2000; McCarthy et al, 1995). The percentage of central entries compared to total entries made were determined for each rat (# of central entries/# of total entries × 100). Central entries are considered an index of antianxiety behavior.

Elevated plus maze

Rats were placed in a brightly lit elevated plus maze and the time spent in the closed and open arms of this maze was recorded for 5 min (Frye et al, 2000; Pellow and File, 1986). The plus maze consisted of four arms (50 × 10 cm) that are elevated 50 cm above the floor; two of these arms are enclosed by 37 cm high walls. Open arm duration is considered an index of antianxiety behavior.

Forced swim task

Rats were tested in a modified version of the forced swim test (Frye and Walf, 2002; Porsolt et al, 1977). Rats were placed in a cylindrical container filled with 30 cm of 30°C water. The amount of time the rats spent swimming, struggling, and immobile was recorded for 10 min. Struggling was characterized as the rats' movement in the water, which was typically quick movement of the front paws that break the surface of the water and appear to be attempts to escape the chamber. Swimming was recorded when the rats were engaged in movement of the front and back paws without the front paws breaking the surface of the water. Immobility was defined as an absence of any movement other than those necessary to keep the head and nose above the surface of the water (eg floating). Immobility is considered an index of depressive behavior.

Tissue Collection and Radioimmunoassay

Tissue collection

In experiment 1, blood was collected from all rats once—either immediately after removal from home cage (no restraint stress and not tested), or after behavioral testing with restraint, or testing with no prior restraint. In experiment 2, blood was collected from all rats on three occasions—24 h after ADX, 4 days after corticosterone-replacement, and immediately after elevated plus maze testing. Blood, collected from rats via a small nick in the tip of the tail, was placed in EDTA-containing chilled eppendorfs, centrifuged (3000g for 20 min), and plasma was stored at −20°C until radioimmunoassay (Tomie et al, 2002; Walf and Frye, 2003).

Estradiol radioimmunoassay

In experiment 1, to determine plasma E2 levels, plasma samples were pooled from some rats in the same E2 dosage and stress (no stress and not tested, no stress and tested, and restraint stress and tested), which produced 3–8 data points per group. From these samples, there was at least 1 data point for each of the three behavioral tasks so that we could analyze if testing itself altered plasma E2 levels of rats. In experiment 2, rats were not stressed, which resulted in larger plasma samples such that E2 could be measured in each rat, yielding 7–12 data points per group. E2 was extracted twice with ether, by snap freezing. Following chromatographic separation, pellets were reconstituted in PBS (pH=7.4). The standard curve was prepared in duplicate (12.5–1000 pg/0.1 ml). Standards were added to PBS, with E2 antibody (Dr Niswender, #244, Colorado State University, Fort Collins, CO), and [3H]E2 (NET-317, 51.3 ci/mmol; New England Nuclear, Boston, MA, 8000 dpm/100 ml; Frye and Bayon, 1999; Frye et al, 1996). Assay tubes are incubated at room temperature for 50 min. Dextran-coated charcoal was used to separate bound and free following a 10-min incubation on ice and centrifugation at 3000g for 10 min. Supernatant was pipetted into a glass scintillation vial with scintillation cocktail and counted using a Tri-Carb 2000CA Liquid Scintillation Analyzer. Unknowns were interpolated from the standard curve using Assay Zap, a program for radioimmunoassay analyses. The inter- and intra-assay coefficients of variance were 0.08 and 0.10, respectively.

Corticosterone radioimmunoassay

Corticosterone was measured in each sample collected. Corticosterone was extracted from plasma by heating at 60°C for 30 min. Samples were incubated for 60 min at room temperature with [3H] corticosterone (NET 182: specific activity=48.2 ci/mmol; New England Nuclear) and a 1 : 20 000 dilution of antibody (Endocrine Sciences:#B3-163; Frye and Bayon, 1999; Frye et al, 1996). Dextran-coated charcoal was used to separate bound and free following a 15-min incubation on ice and centrifugation at 3000 g for 10 min. Supernatant was decanted into a glass scintillation vial with scintillation cocktail and then counted using a Tri-Carb 2000CA Liquid Scintillation Analyzer. Unknowns were interpolated from the standard curve using Assay Zap. The inter- and intra-assay reliability coefficients were 0.05 and 0.08, respectively.

Statistical Analyses

In experiment 1, two-way analyses of variance (ANOVAs) were utilized to examine effects of E2 dosage and stress condition (no stress vs restraint stress) on open field, elevated plus maze, and forced swim test behavior. Two-way ANOVAs were used to examine effects of E2 dosage and stress condition (no stress and not tested, no stress and tested, or restraint stressed and tested) on plasma E2 and corticosterone levels. In experiment 2, two-way ANOVAs were also used to examine effects of E2 dosage and ADX condition (sham, ADX+veh, ADX+25 corticosterone, and ADX+250 corticosterone) on open field, elevated plus maze, and forced swim test behavior, plasma corticosterone and E2 levels. Where appropriate, Fisher's post hoc tests were utilized to determine group differences. The α level for statistical significance was p⩽0.05, and for a trend was p⩽0.10. In all figures and text, data are expressed as the mean and standard deviation (SD).

RESULTS

Plasma Estradiol Levels

E2-primed ovx rats had significantly higher plasma E2 concentrations, such that 0,2<5<10<20<50 μg E2 produced higher plasma E2 levels (see Table 1). There were no observable differences in E2 levels based upon which tasks the rats were tested in. There were no effects of restraint stress on plasma E2 levels.

Table 1 Experiment 1: Plasma Estradiol Levels (±SD)
Full size table

As in the first experiment, administration of 0, 10, and 50 μg E2 produced statistically different plasma concentrations of E2 in these rats (F(2,106)=90.23, p<0.01; see Table 2). There were no differences produced by ADX condition on these plasma E2 levels.

Table 2 Experiment 2: Plasma Estradiol Levels (±SD)
Full size table

Effects of E2 Dosage and Stress on Anxiety Behavior

Dosages of E2 (5 or 10 μg) that produced moderate physiological plasma levels of E2 decreased anxiety behavior in the open field compared to vehicle and all other E2 dosages administered. There was a significant effect of E2 dosage on percentage of central entries in the open field (F(5,115)=17.12, p<0.01; see Figure 1, left) and duration spent by rats on the open arms of the plus maze (F(5,107)=5.28, p<0.01; see Figure 2, left). Post hoc tests revealed that 5 or 10 μg E2, which produced plasma E2 levels within a moderate physiological range akin to levels observed during behavioral estrus (Frye and Bayon, 1999), significantly increased central entries in the open field compared to vehicle (0 μg), lower (2 μg), and higher (20 and 50 μg) E2 dosages (all post hoc comparisons were p<0.01). There was a similar pattern in the plus maze, such that 5 or 10 μg E2 increased duration spent by rats on the open arms compared to 0, 2, 20, and 50 μg E2 (all post hoc comparisons were p<0.01). There was a main effect of E2 dosage on total entries in the open field (F(5,115)=5.32, p<0.01). Post hoc tests revealed that 2 (124.4±54.8), 5 (142.0±55.6), or 10 (138.3±50.2) μg E2 increased total entries compared to 0 (85.0±39.0; vs 2 μg p<0.01; vs 5 μg p<0.01; vs 10 μg p<0.01), 20 (81.6±35.2; vs 2 μg p=0.04; vs 5 μg p=0.01; vs 10 μg p<0.01), and 50 (90.7±47.1; vs 2 μg p<0.01, vs 5 μg p<0.01; vs 10 μg p<0.01) μg E2. However, there were no differences among groups for total entries in the plus maze (data not shown).

Figure 1
figure 1

Effects of E2 dosage and restraint stress on open field behavior. Rats were not stressed and administered 0 (n=9), 2 (n=11), 5 (n=9), 10 (n=12), 20 (n=9), or 50 (n=10) μg E2 or stressed and administered 0 (n=11), 2 (n=11), 5 (n=12), 10 (n=11), 20 (n=11), or 50 (n=11) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. *Above grouped bars indicates that restraint stress is significantly different from no stress condition.

Full size image
Figure 2
figure 2

Effects of E2 dosage and restraint stress on elevated plus maze behavior. Rats were not stressed and administered 0 (n=9), 2 (n=9), 5 (n=8), 10 (n=7), 20 (n=11), or 50 (n=10) μg E2 or stressed and administered 0 (n=10), 2 (n=12), 5 (n=12), 10 (n=12), 20 (n=9), or 50 (n=11) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. *Above grouped bars indicates that restraint stress is significantly different from no stress condition.

Full size image

Restraint stress increased anxiety behavior of ovx rats compared to no restraint stress before testing. There was a significant effect of stress on total (F(1,115)=33.96, p< 0.01) and percentage of central (F(1,115)=61.91, p<0.01; see Figure 1, right) entries in the open field and time spent in the open arms of the elevated plus maze (F(1,107)=5.79, p<0.01; see Figure 2, right). Post hoc tests revealed that 20 min of restraint stress immediately before behavioral testing significantly reduced central entries in the open field (p<0.01), and tended to decrease open arm time (p=0.07), compared to no restraint stress before testing. Restraint stress significantly reduced the total number of entries in the open field (87.1±11.8) compared to no restraint stress before testing (133.7±16.3; p<0.01). Restraint stress did not alter total entries in the plus maze (data not shown).

There was an interaction of E2 dosage and restraint stress on behavior in the open field. In the open field, restraint stress significantly decreased total (F(5,115)=2.28, p<0.05) and central (F(5,115)=6.91, p<0.01) entries of rats administered 5 or 10 μg E2, but not other E2 dosages, compared to no prior restraint stress. Rats that were restraint stressed prior to testing and administered 5 (90.1±44.2), 10 (103.2±45.7), but not 0 (72.9±28.0), 2 (102.7±58.4), 20 (95.4±25.5), or 50 (72.8±28.0) μg E2, made significantly fewer total entries in the open field compared to rats that were not restraint stressed prior to testing and administered 5 (198±67.0), or 10 (173.3±54.6) μg E2, but not 0 (97.1±49.9), 2 (146.0±51.2), 20 (117.7±45.4), and 50 (108.6±66.2) μg E2 (all post hoc comparisons were p<0.01).

In experiment 2, effects of E2 dosages on anxiety behavior replicated findings from experiment 1. There were significant effects of E2 dosage on percentage of central entries (F(2,106)=10.43, p<0.01; see Figure 3) and the duration of time spent on the open arms of the plus maze (F(2,106)=16.42, p<0.01; see Figure 4). Post hoc tests revealed that 10 μg E2, which produced moderate physiological plasma E2 concentrations, significantly increased central entries in the open field and duration spent by rats on the open arms of the plus maze compared to vehicle (0 μg) or 50 μg E2 (all post hoc comparisons were p<0.01). There was a main effect of E2 dosage on total entries in the open field (F(2,106)=7.70, p<0.01) and in the elevated plus maze (F(2,106)=6.16, p<0.01). In the open field, 10 (79.0±31.8) or 50 (66.0±40.4) μg E2 significantly increased total entries compared to vehicle (56.3±25.4; all post hoc comparisons were p<0.01). In the plus maze, 10 (5.8±3.4) μg E2 significantly increased total entries in the elevated plus maze compared to 0 (3.8±1.6) and 50 (4.5±2.0) μg E2 (all post hoc comparisons were p<0.01).

Figure 3
figure 3

Effects of E2 dosage and ADX on open field behavior. Rats were sham surgerized and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 , ADX+veh and administered 0 (n=8), 10 (n=9), or 50 (n=9) μg E2, ADX+25 CORT and administered 0 (n=10), 10 (n=9), or 50 (n=11) μg E2, or ADX+250 CORT and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. # above a single bar indicates that there is a tendency for 50 μg E2 to be different from 0 μg E2 condition. +Above grouped bars indicates that ADX condition is significantly different than ADX+veh condition. *Above grouped bars indicates that ADX condition is significantly different than ADX+veh and ADX+250 CORT condition.

Full size image
Figure 4
figure 4

Effects of E2 dosage and ADX on elevated plus maze behavior. Rats were sham surgerized and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 , ADX+veh and administered 0 (n=8), 10 (n=9), or 50 (n=9) μg E2, ADX+25 CORT and administered 0 (n=10), 10 (n=9), or 50 (n=11) μg E2 , or ADX+250 CORT and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. *Above grouped bars indicates that ADX condition is significantly different than ADX+veh and ADX+250 CORT condition.

Full size image

In experiment 2, ADX altered anxiety behavior of rats. There were significant effects of ADX on percentage of central entries (F(3,107)=5.37, p<0.01; see Figure 3) and the duration of time spent on the open arms of the plus maze (F(3,107)=6.24, p<0.01; see Figure 4). Post hoc tests revealed that ADX rats administered 25 μg/ml corticosterone in their drinking water had significantly increased percentage of central entries in the open field compared to ADX rats administered vehicle or 250 μg/ml corticosterone in their drinking water (all post hoc comparisons were p<0.01). In the plus maze, post hoc tests revealed that sham ADX rats and ADX rats administered 25 μg/ml corticosterone in their drinking water spent significantly more time on the open arms of the plus maze compared to rats administered vehicle or 250 μg/ml corticosterone (all post hoc comparisons were p<0.01). Although there was no main effect of ADX on total entries in the open field (data not shown), there was a difference among groups for total entries in the plus maze (F(3,106)=3.67, p<0.01). Post hoc tests revealed that ADX rats administered 25 μg/ml corticosterone in their drinking water made more total entries (6.0±2.7) in the plus maze than did ADX rats administered vehicle (3.8±1.6; p=0.01) or 250 μg/ml corticosterone (4.2±2.3; p<0.01) in their drinking water.

In the open field, there was a tendency for an interaction between E2 dosage and ADX condition. Although these data did not reach statistical significance, results suggest that 10 μg E2, compared to vehicle or 50 μg E2, increased percentage of central entries in sham and ADX rats administered 25 μg/ml corticosterone, but not ADX rats administered vehicle or 250 μg/ml corticosterone. Furthermore, there was a tendency for 50 μg E2 to increase central entries compared to rats administered 0 μg E2 in the group of ADX rats administered 25 μg/ml corticosterone only.

A similar pattern of effects was observed in the elevated plus maze. There was a significant interaction between E2 dosage and ADX condition (F(6,106)=3.16, p<0.01). Post hoc tests revealed that, compared to vehicle or 50 μg E2, 10 μg E2 increased time spent on the open arms of the plus maze in sham and ADX rats administered 25 μg/ml corticosterone, but not ADX rats administered vehicle or 250 μg/ml corticosterone (see Figure 4; all post hoc comparisons were p<0.01).

Depressive Behavior

Dosages of E2 (5 or 10 μg) that produced moderate physiological plasma levels of E2 decreased depressive behavior in the forced swim test compared to vehicle and all other E2 dosages. There was a significant effect of E2 dosage on immobility (F(5,154)=19.92, p<0.01), struggling (F(5,154)=6.56, p<0.01), and swimming (F(5,154)=9.10, p<0.01) in the forced swim test. Post hoc tests revealed that 5 or 10 μg E2 significantly decreased duration of immobility in the forced swim test compared to vehicle or all other E2 dosages (see Figure 5, left; all post hoc comparisons were p<0.01). Struggling and swimming were increased by administration of 5 (155.2±43.7 and 223.5±54.8) or 10 (138.3±35.5 and 236.0±62.5) μg compared to 0 (110.0±40.7 and 159.3±45.0), 2 (117.0±32.6 and 187.8±58.1), 20 (131.5±40.7 and 159.1±65.0), or 50 (103.5±29.9 and 170.7±57.1) μg E2 (all post hoc comparisons were p<0.01).

Figure 5
figure 5

Effects of E2 dosage and restraint stress on forced swim test behavior. Rats were not stressed and administered 0 (n=15), 2 (n=14), 5 (n=16), 10 (n=17), 20 (n=17), or 50 (n=13) μg E2 or stressed and administered 0 (n=13), 2 (n=12), 5 (n=13), 10 (n=11), 20 (n=13), or 50 (n=12) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. *Above grouped bars indicates that restraint stress is significantly different from no stress condition.

Full size image

Restraint stress increased depressive behavior of ovx rats compared to no restraint stress before testing. There was a significant effect of restraint stress on immobility (F(1,154)=67.63, p<0.01), struggling (F(1,154)=5.67, p<0.01), and swimming (F(1,154)=45.99, p<0.01). Post hoc tests revealed that restraint stress significantly increased immobility in the forced swim test compared to no restraint stress (see Figure 5, right; p<0.01). Duration of struggling and swimming were also decreased in rats that were restraint stressed (118.6±33.9 and 158.8±59.2) compared to nonstressed rats (133.1±40.4 and 219.8±54.9; all post hoc comparisons were p<0.01).

There was an interaction of E2 dosage and restraint stress on immobility and swimming behavior in the forced swim test. Restraint stressed rats administered 5 or 10 μg E2 had significantly reduced duration of immobility (F(5,154)=2.31, p<0.04; see Figure 5) and increased swimming (F(5,154)=1.05, p<0.04) compared to rats administered vehicle or 2, 20, or 50 μg E2. For swimming behavior, rats that were restraint stressed and administered 5 (181.1±59.3) or 10 (183.4±72.8) μg E2, but not 0 (119.4±46.8), 2 (163.7±45.7), 20 (144.6±74.7), or 50 (160.8±56.2) μg E2, had decreased swimming compared to nonstressed rats administered 5 (265.9± 50.4) or 10 (288.5±52.1) μg E2, but not 0 (199.1±43.2), 2 (211.3±70.4), 20 (173.5±55.3), or 50 (180.5±58.0) μg E2 (all post hoc comparisons were p<0.01).

In experiment 2, E2 altered forced swim test behavior and replicated the pattern of results observed in experiment 1. There was a significant effect of E2 dosage on immobility (F(2,106)=38.14, p<0.01), struggling (F(2,106)=4.48, p<0.01), and swimming (F(2,106)=7.65, p<0.01) in the forced swim test. Post hoc tests revealed that 10 μg E2 significantly decreased duration of immobility in the forced swim test compared to vehicle or 50 μg E2 (see Figure 6; p<0.01). E2 (50 μg) significantly reduced duration of immobility compared to vehicle administration (p<0.01), 10 μg (141.0±50.3 and 200.1±66.9) similarly increased duration of struggling and swimming in the forced swim test compared to that observed in rats administered vehicle (112.6±51.4 and 143.9±54.5) or 50 (125.6±43.3 and 165.2±69.8) μg E2 (all post hoc comparisons were p<0.01).

Figure 6
figure 6

Effects of E2 dosage and ADX on forced swim test behavior. Rats were sham surgerized and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 , ADX+veh and administered 0 (n=8), 10 (n=9), or 50 (n=9) μg E2, ADX+25 CORT and administered 0 (n=10), 10 (n=9), or 50 (n=11) μg E2 , or ADX+250 CORT and administered 0 (n=10), 10 (n=11), or 50 (n=10) μg E2 before testing. *Above a single bar indicates that E2 dosage is significantly different from all other E2 dosages. *Above grouped bars indicates that ADX condition is significantly different than ADX+veh and ADX+250 CORT condition.

Full size image

ADX condition altered forced swim test behavior of rats. There was a significant effect of ADX on immobility (F(3,106)=19.12, p<0.01), struggling (F(3,106)=2.85, p<0.04), and swimming (F(3,106)=4.42, p<0.01) in the forced swim test. Post hoc tests revealed that sham ADX rats and ADX rats administered 25 μg/ml corticosterone had significantly decreased duration of immobility in the forced swim test compared to vehicle or 250 μg/ml corticosterone-administered ADX rats (see Figure 6; all post hoc comparisons were p<0.01). Struggling behavior was significantly decreased by vehicle (106.4±37.6) administration to ADX rats compared to 25 (137.8±55.3) or 250 (134.4±55.3) μg/ml corticosterone in their drinking water (all post hoc comparisons were p<0.01). Post hoc tests revealed the same pattern of effects for swimming behavior, as were observed for immobility. Sham (184.5±62.8) and ADX rats administered 25 (188.7±57.6) μg/ml corticosterone had significantly increased swimming behavior compared to ADX rats administered vehicle (141.7±59.1) or 250 (164.0±75.3) μg/ml corticosterone in their drinking water (all post hoc comparisons were p<0.01).

There were significant interactions between E2 dosage and ADX condition for behavior in the forced swim test. There were significant interactions between E2 dosage and ADX condition on immobility (F(6,106)=8.53, p<0.01), struggling (F(6,106)=3.10, p<0.01), and swimming (F(6,106)=2.35, p<0.04) in the forced swim test. In all, 10 μg decreased duration spent immobile in sham and ADX rats administered 25 μg/ml corticosterone, but not ADX rats administered vehicle or 250 μg/ml corticosterone, compared to vehicle or 50 μg E2 (see Figure 6; all post hoc comparisons were p<0.01). Rats administered 50 μg E2 had significantly decreased immobility, compared to rats administered 0 μg E2, in the group of ADX rats administered 25 μg/ml corticosterone only (all post hoc comparisons were p<0.01). For swimming behavior, 10 μg, but not vehicle or 50 μg, E2 increased duration spent swimming in sham (240.7±7.35) and ADX rats administered 25 (228.4±71.7) μg/ml corticosterone, but not ADX rats administered vehicle (112.5±57.9) or 250 (170.3±46.5) μg/ml corticosterone (all post hoc comparisons were p<0.01).

Plasma Corticosterone Levels

Dosages of E2 that produced moderate physiological plasma levels of E2 decreased plasma corticosterone levels. There was a significant main effect of E2 dosage on corticosterone levels (F(5,348)=46.77, p<0.01). E2 (5 or 10 μg) significantly decreased plasma corticosterone levels compared to vehicle (0 μg), lower (2 μg), and higher (20 and 50 μg) E2 dosages (see Table 3); 20 and 50 μg E2 increased corticosterone levels compared to all other dosages, as revealed by post hoc tests (all post hoc comparisons were p<0.01).

Table 3 Experiment 1: Plasma Corticosterone Levels (±SD)
Full size table

Restraint stress, compared to the no stress and no testing condition, increased plasma corticosterone, but not E2, levels of rats. There was a significant effect of stress condition on plasma corticosterone levels (F(2,348)=2.67, p<0.01; see Table 3), such that rats that were exposed to testing alone or with restraint had significantly higher plasma corticosterone levels than did rats that were not exposed to restraint stress or tested (p<0.01).

Dosages of E2 (5 or 10 μg) that produced moderate physiological plasma levels of E2 decreased basal, post-testing, and post-testing and restraint plasma corticosterone levels (F(10, 348)=2.57, p<0.01; see Table 3). E2 (5 or 10 μg) significantly decreased corticosterone levels of rats that were not stressed or tested, as well as corticosterone levels after testing without prior restraint stress compared to vehicle (0 μg), lower (2 μg), and higher (20 and 50 μg) E2 dosages (all post hoc comparisons were p<0.01). Compared to vehicle or lower E2 dosages, 20 or 50 μg E2 significantly increased basal, post-testing, and post-testing and restraint stress corticosterone levels (all post hoc comparisons were p<0.01).

Although there were no main effects of E2 dosage on plasma corticosterone levels of sham and ADX rats, there were effects of ADX condition on plasma corticosterone levels (F(3,106)=24.73, p<0.01; see Table 4). Post hoc tests revealed that sham surgerized rats and ADX rats replaced with 25 μg/ml corticosterone had significantly higher corticosterone levels than did ADX rats administered vehicle, and significantly lower corticosterone levels compared to ADX rats administered 250 μg/ml corticosterone (all post hoc comparisons were p<0.01).

Table 4 Experiment 2: Plasma Corticosterone Levels (±SD)
Full size table

DISCUSSION

The present results support the hypotheses that E2 may have antianxiety and antidepressive effects, which may depend on E2 dosage, and be modulated by activity of the HPA. Ovx rats administered E2 dosages (5 or 10 μg) that produced moderate physiological levels of E2 entered significantly more central entries in the open field, spent significantly more time on the open arms of the plus maze, spent significantly less time immobile in the forced swim test, and had lower plasma corticosterone levels than did rats with lower or higher plasma E2 levels. Restraint stress before testing attenuated the antianxiety and antidepressive behavior of rats in these tasks and increased plasma corticosterone levels. In rats without an intact HPA, the effects of 10 μg E2 to reduce anxiety and depressive behavior were not observed, unless ADX rats were replaced with corticosterone that produced low plasma levels of corticosterone. Together, these data suggest that moderate E2 levels, like that produced by 5 or 10 μg E2 administration, may decrease anxiety and depressive behavior and corticosterone levels and that these effects may be modulated by the tone of the HPA.

The present results, that dosages of E2 that produced physiological levels of plasma E2 reduced anxiety behavior, confirm previous reports of E2's effects for affective behavior of rodents. Similar E2 regimen (5–10 μg; SC), which likely also produced physiological plasma E2 levels, increased open field activity and time spent on the open arms of the plus maze, and decreased freezing in response to a footshock (Slater and Blizard, 1976; Frye and Walf, 2004a). However, some E2 regimen of lower or higher E2 dosages, or a longer duration of E2 exposure, produce no effect on anxiety behavior, decrease anxiety behavior, or may indeed increase anxiety behavior of female rodents in these tasks (Diaz-Veliz et al, 1997; Luine et al, 1998; Martinez-Mota et al, 2000; Mora et al, 1996; Morgan and Pfaff, 2001, 2002; Nomikos and Spyraki, 1988; Stoffel and Craft, 2003). The present results that low dosages of E2 (<5 μg) were without effect, and that high dosages of E2 (>10 μg) also did not have antianxiety effects, are consistent with this previous research. Together, these data from the present report suggest that E2's effects on anxiety behavior may be sensitive to the E2 dosages and regimen employed, such that E2 dosages that produce physiological E2 levels (5 or 10 μg) produce antianxiety behavior in the open field and elevated plus maze tasks compared to lower (0 or 2 μg) or higher (20 or 50 μg) E2.

Our present results, that 5 or 10 μg E2, but not lower or higher E2 dosages, decreases immobility in the forced swim test support previous findings that E2 reduces depressive behavior of rodents. Similar to effects of E2 on anxiety behavior, there is evidence for dose-dependent effects of E2 on depressive behavior in the forced swim test. E2 regimen similar to the effective one utilized in this study (5–10 μg) reduced immobility and increased swimming in the forced swim test of ovx rodents (Bernardi et al, 1989; Estrada-Camarena et al, 2003; Hilakivi-Clarke, 1996; Rachman et al, 1998). However, E2 to ovx voles, in a high dosage (0.75 μg) that is sufficient to induce a lordosis response, but is not considered a supraphysiological dosage, did not reduce depressive behavior in the forced swim test (Galea et al, 2002). Moreover, administration of high dosages of E2 (50 μg) to rats for 4 days in an experimentally induced postpartum state (ie, withdrawal from high E2 and progesterone levels akin to pregnancy) reduces immobility compared to vehicle administration in one, but not another, study (Galea et al, 2001; Stoffel and Craft, 2003). Although a low dosage of E2 (2 μg) administered 48 h before testing did not significantly increase E2 levels of ovx rats at the time of testing and did not reduce immobility in the present study and another study, which utilized a similar regimen (2.5 μg/rat 1 or 48 h before testing; Estrada-Camarena et al, 2003), a longer exposure (2 weeks) to similar low dosages of E2 (0.3–3.0 μg/rat/day) to ovx rats decreased immobility in the forced swim test (Okada et al, 1997). These data, and the present results, suggest that E2's effects to reduce depressive behavior in the forced swim test of female rodents may be mitigated by the dosage and/or duration of E2 regimen utilized. Thus, E2 regimen that produced physiological E2 levels (5 or 10 μg, 48 h before testing), but not regimen that produced lower (0, 2 μg) or higher (20, 50 μg) E2 levels, reduced depressive behavior in the forced swim.

The results from the present study extend previous research of E2's effects for anxiety and depressive behavior by suggesting that these effects of E2 may be modulated by HPA responding. Restraint stress attenuated the antianxiety and antidepressive effects of 5 and 10 μg E2 to ovx rats. Moreover, plasma corticosterone levels of rats were lowest in ovx rats administered 5 or 10 μg E2, irrespective of stressors, such as testing and/or restraint. These data support previous findings that 10 μg E2 for 7 days reduced immobility in the forced swim task and reduced stress responses (c-fos immunoreactivity in the hippocampus and amygdala, brain areas integral for affective behavior; Rachman et al, 1998). However, it must be noted that in the present study, the dosages of E2 that reduced immobility were administered in a more acute regimen of 48 h before testing. Furthermore, stress (or HPA activation) may alter rodents' response to gonadal hormones. This has been examined in an animal model of gestational stress. In addition to having a different pattern of corticosterone secretion compared to nongestationally stressed rats (Weinstock, 2001), gestationally stressed female rats may have reduced responsiveness to E2. For instance, female rats that are gestationally stressed and then tested as adults have decreased sexual behavior towards a male when naturally receptive or when E2-primed (Frye and Orecki, 2002a, 2002b). Effects of E2 replacement to reduce depressive behavior are attenuated in female gestationally stressed rats compared to controls (Frye and Wawrzycki, 2003). As well, gestationally stressed female, but not gestationally stressed male, rats have a reduced number of hippocampal granule cells, compared to effects observed in nongestationally stressed controls (Schmitz et al, 2002). Thus, the present findings that E2's effects to reduce anxiety and depressive behavior and corticosterone levels are attenuated by an acute stressor (restraint) substantiate these effects of gestational stress on behavioral and hormonal responses in female rodents.

The effects observed in this study for physiological E2 concentrations to reduce anxiety and depressive behavior, which may be influenced by a prior stressor, are analogous to recent studies of E2's biphasic effects on cognitive behavior among female rodents. Moderate, but not low or high, levels of E2 may improve cognitive performance (Shors and Leuner, 2003). Stressor exposure produces performance deficits among female rodents and this may be due to stress-induced elevations in plasma E2 levels (Shors et al, 1998, 1999; Shors and Leuner, 2003; Wood and Shors, 1998; Wood et al, 2001). However, chronic exposure of female rats, but not males, to a stressor has enhancing effects on cognitive performance (Bowman et al, 2001, 2002). Although the stress paradigm that we used in the present study did not increase E2 levels, effects of E2 were sensitive to the dosage administered and restraint stress reduced the antianxiety and antidepressive effects of moderate, physiological E2 levels.

The present results characterize some of the effects of different dosages of E2 to ovx rats and manipulation of the HPA with pretesting restraint stress and ADX for anxiety and depressive behavior. We found an effect of moderate dosages (5–10 μg), but not lower (0–2 μg) or higher (20–50 μg) dosages, of E2 to reduce basal corticosterone levels and levels following behavioral testing with or without prior restraint. When the HPA was disrupted in rats with ADX, the ability of the moderate dose of E2 (10 μg) to reduce anxiety and depressive behavior was abrogated; however, E2's antianxiety and antidepressive effects were reinstated with replacement of low (25 μg/ml), but not high (250 μg/ml), dosages of corticosterone. Unlike in experiment 1, the dose-dependent effects of E2 to reduce anxiety and depressive behavior of rats did not have the same pattern of effects on plasma corticosterone levels. In experiment 2, 10 μg E2 reduces anxiety and depressive behavior of ADX rats replaced with 25 μg/ml corticosterone in their drinking water, but corticosterone levels of rats administered 0, 10, or 50 μg E2 were not different. This suggests that E2's effects on affective behavior cannot completely be explained by corticosterone levels.

There are complex effects of E2 and HPA interactions that may involve more than interactions of E2 and corticosterone. In our study, we grossly manipulated the HPA and replaced back corticosterone to determine whether E2's effects on antianxiety and antidepressive behavioral patterns require corticosterone. That E2's effects were absent in ADX rats suggest that E2's effects may be modulated by HPA feedback, but the nature of these effects were not completely revealed. That E2's antianxiety and antidepressive effects were reinstated in ADX rats with low dosage corticosterone-replacement suggests that some corticosterone may be necessary for E2's effects for these behaviors. Interestingly, there were differences in how much corticosterone-replacement was necessary for E2's behavioral effects. There were different effects of low vs high dose corticosterone replacement on E2's behavioral effects, which is consistent with variable reports of E2's effects on corticosterone secretion in intact female rodents. For instance, during behavioral estrus, when E2 levels are typically high in female rodents, there is evidence for increased basal and stress-induced corticosterone secretion (Carey et al, 1995; Figueiredo et al, 2002; Frye and Bayon, 1999; Raps et al, 1971; Viau and Meaney, 1991), albeit there is variability within and between reports. Despite some evidence of higher corticosterone levels, proestrous rats also have reduced anxiety and depressive behavior compared to rats in other stages of estrous (Diaz-Veliz et al, 1997; Frye et al, 2000; Marcondes et al, 2001; Mora et al, 1996). Some of these variable effects may be due to differences in the stress reactivity of these animals and the type of stressor paradigm utilized and how stressful the behavioral task utilized is (reviewed in Herman et al, 2003). Interestingly, the dose-dependent effects of E2 that we observed in sham surgerized rats were less robust in ADX, corticosterone-replaced rats. E2 (50 μg) produced similar effects as 10 μg E2 to reduce anxiety and depressive behavior in ADX rats that were replaced with a low dosage of corticosterone. Indeed, previous studies have demonstrated a complex relationship between E2 and the HPA. As such, it would be important to examine effects of different dosages of E2 on other indices of HPA activity, such as corticosteroid-binding globulin (CBG) and/or adrenocorticotropin (ACTH) levels, which also can be altered by gonadal hormones (Carey et al, 1995; Dayas et al, 2000; Figueiredo et al, 2003; McCormick et al, 2002; Viau and Meaney, 1991; Young et al, 2001), for these effects on anxiety and depressive behavior. Indeed, it has been suggested that E2 may sensitize adrenals to corticosterone secretion and have effects on HPA feedback via changing CBG and/or ACTH levels (Figueiredo et al, 2002, 2003). Thus, these effects further support that reductions in anxiety and depressive behavior may not be completely accounted for by changes in corticosterone levels per se.

E2 has been reported to enhance arousal of rodents, which may underlie some of the effects observed in the present study. For instance, E2 increases motor behavior of rodents (Colvin and Sawyer, 1969; Frye and Walf, 2004a; Morgan and Pfaff, 2001, 2002). However, profound effects of E2 on gross motor behavior of rats cannot completely explain our present results of decreased anxiety and depressive behavior. A consistent effect of E2 on total entries in the open field or elevated plus maze or struggling and swimming in the forced swim test was not observed, but the changes observed in these indices suggests that some of the effects observed may be due to changes in arousal/motor behavior following E2. Indeed, these alterations in arousal may be related to the effects of E2 on HPA activity, but the extent to which E2's modulation of arousal is related to effects on HPA processes needs further attention.

The present data provide justification for further examining the nature and mechanisms by which E2 and HPA activity interact for effects on anxiety and depressive behavior. First, although these data suggest involvement of the HPA at least for some of E2's effects on anxiety and depressive behavior, direct manipulation of this putative mechanism by pharmacological blockade of mineralocorticoid or glucocorticoid receptors would be of interest. Notably, E2 may have effects on distribution of glucocorticoid receptors in the brain (Patchev et al, 1995). Second, as discussed, it would also be of interest to examine the effects of E2 for altering other indices of HPA activity besides corticosterone. Further examination may reveal a more direct target of E2 on HPA activity, which would underlie the effects observed for anxiety and depressive behavior. For instance, the effects of E2 for the observed behavioral differences may be mediated by CBG, corticotropin releasing factor, and/or ACTH, which are also involved in modulating affective behavior (Almeida et al, 1997; Bale and Vale, 2003; Bale et al, 2000, 2002; Landgraf et al, 1999). Another possibility to consider is E2's effects on monoamines, such as serotonin, which also mitigate anxiety and depressive behavior and HPA activity, and may underlie the behavioral effects observed poststress and/or E2 administration (Bowman et al, 2003). Fourth, it is also important to consider the role that E2 has on progestins, which may underlie some of the effects observed. Progestins, particularly 5α-reduced progestins, such as 5α-pregnan-3α-ol-20-one (3α, 5α-THP), modulate antianxiety and antidepressive behavior of female rodents (Frye et al, 2000, 2004; Frye and Walf, 2002, 2004b; Picazo and Fernandez-Guasti, 1995; Martinez-Mota et al, 2000; Rhodes and Frye, 2001). E2 enhances 3α,5α-THP formation by increasing 5α-reductase activity and can enhance progestins' effects on behavior (Cheng and Karavolas, 1973; Frye and Duncan, 1996; Vongher and Frye, 1999). Notably, 5α-reduced progestins may dampen HPA activity (Patchev and Almeida, 1996). Thus, although the present data suggest that E2's antianxiety and antidepressive behavioral effects may be mitigated by the HPA tone of the animal, and not corticosterone levels alone, the other modulating factors are not presently known.

In summary, rats administered E2 dosages (5 or 10 μg), which produced physiological plasma E2 concentrations, had significantly reduced anxiety (open field central entries, open arm duration in the plus maze) and depressive (immobility in the forced swim test) behavior and corticosterone levels compared to rats administered vehicle, lower (2 μg), or higher (20 or 50 μg) E2 dosages. Restraint stress attenuated the antianxiety and antidepressive effects of E2. ADX abrogated the effects of E2 to reduce anxiety and depressive behavior, but these effects could be partially reversed by administration of 10 or 50 μg E2 in ADX rats replaced with a low, but not high, dosage of corticosterone. Thus, these data suggest that E2 may have dose-dependent effects on anxiety and depressive behavior of female rodents, in part through actions involving HPA responsiveness. Given that hormone-replacement therapy for women has recently come under fire (Rapp et al, 2003; Rossouw et al, 2002; Shumaker et al, 1998), these data that response to E2 may be different depending on stress/HPA are important to consider. There is evidence that E2 can enhance mood of some women; however, there are differences in reports (Bjorn et al, 2003; Cohen et al, 2003; Halbreich, 1997; Morrison et al, 2004; Sherwin 1991, 1996; Schmidt et al, 1998, 2000; Soares et al, 2001). Indeed, the variable reports on the efficacy of E2 replacement to influence anxiety or depression of women may in part involve individual differences in stress responsiveness related to past and/or current experiences.