EFFECTS OF LONG-TERM TREATMENT WITH ESTRADIOL AND ESTROGEN RECEPTOR SUBTYPE AGONISTS ON SEROTONERGIC FUNCTION IN OVARIECTOMIZED RATS

Abstract

Acute estradiol treatment was reported to slow the clearance of serotonin via activation of estrogen receptors (ER)β and/or GPR30 and to block the ability of a selective serotonin reuptake inhibitor (SSRI) to slow serotonin clearance via activation of ERα. In this study, the behavioral consequences of longer-term treatments with estradiol or ER subtype-selective agonists and/or an SSRI were examined in the forced swim test (FST). Ovariectomized rats were administered for two weeks: estradiol, ERβ agonist (DPN), GPR30 agonist (G1), ERα agonist (PPT), and/or the SSRI sertraline. Similar to sertraline, longer-term treatment with estradiol, DPN or G1 induced an antidepressant- like effect. By contrast, PPT did not, even though it blocked the antidepressant-like effect of sertraline. Uterus weights, used as a peripheral measure of estrogenic activity, were increased by estradiol and PPT but not DPN or G1 treatment. A second part of this study investigated, using Western blot analyses in homogenates from hippocampus, whether these behavioral effects are accompanied by changes in the activation of specific signaling pathways and/or TrkB. Estradiol and G1 increased phosphorylation of Akt, ERK and TrkB. These effects were similar to those obtained after treatment with sertraline. Treatment with DPN increased phosphorylation of ERK and TrkB but it did not alter that of Akt. Treatment with PPT increased phosphorylation of Akt and ERK without altering that of TrkB. In conclusion, activation of at least TrkB and possibly ERK may be involved in the antidepressant-like effect of estradiol, ERβ and GPR30 agonists whereas Akt activation may not be necessary.

INTRODUCTION

Many of the effects of estrogen are mediated by estrogen receptors (ERs), primarily by ERα and ERβ, which are encoded by different genes [1, 2]. These receptors are regionally distributed throughout the body and brain [3]. Some of the regional differences in ERα versus ERβ distribution in brain may confer specific functional effects of estradiol (E₂) acting via these receptor isoforms. ERα is predominant in the ventromedial hypothalamus and is related to E₂'s facilitatory actions on female rodent sexual responding; ERβ is predominant in limbic and/or stress-responsive regions of brain, such as the hippocampus, and is related to the affective behaviors of E₂ [4]. The effects of E₂ in producing anxiolytic-like or antidepressant (AD)–like behaviors through activation of the ERβ subtype have been demonstrated using both ERβ knockdown and ERβ activation approaches [5–10]. Estrogen can act either by a genomic mechanism, using classical nuclear transcription factor receptors [11], or by a non-genomic, less well-characterized membrane receptor signaling mechanism [12]. The genomic pathway requires at least 30–60 minutes to be manifest and is associated with changes in protein synthesis. Rapid effects of estrogen occurring within seconds to minutes may be mediated by membrane receptors leading to activation of signaling pathways. E₂ has both rapid as well as slower effects on serotonin transporter (SERT) function, indicating the involvement of nongenomic as well as genomic mechanisms in its effects [10, 13].

Recently, a novel G protein-coupled ER, GPR30 (also known as GPER1), was identified [14]. GPR30 is a seven transmembrane-spanning G-protein coupled receptor that shows specific high affinity binding for estradiol and related estrogens and promotes rapid estrogen signaling in a variety of cell types [15]. The GPR30 specific agonist, G1, can reproduce many effects of estrogen [14]. In addition, acute activation of GPR30 by G1 mimicked the AD-like effects of E₂, similar to that observed with an ERβ agonist [10].

Using in vivo chronoamperometry in young adult ovariectomized (OVX) rats, acute systemic treatment as well as local administration of E₂ produced changes in SERT function, as shown by a decreased basal clearance of serotonin (5-HT); it also prevented the inhibitory effect of the selective serotonin reuptake inhibitor (SSRI) fluvoxamine on 5-HT clearance [13]. These two effects were mediated by different ER subtypes: the AD-like effect of estradiol was due to ERβ and/or GPR30 activation whereas its blockade of the ability of an SSRI to slow 5-HT clearance was mediated by activation of ERα [10].

Two major signaling pathways have been associated with the effects of E₂ in the brain; mitogen activated protein kinase (MAPK) / extracellular signal-regulated kinase (ERK1/2) and phosphatidylinositol-3-kinase (PI3K) / Akt (also known as protein kinase B) pathways. The mechanisms of E₂ signaling in the brain are complex, tissue specific and include independent as well as co-dependent effects through ERα, ERβ and GPR30 as well as interactions between ER and other membrane receptors such as insulin-like growth factor receptor (IGFIR), metabotropic glutamate receptor (mGluR) and the neurotrophin (BDNF and/or NT3) receptor, TrkB [16–19]. In a recent study, it was shown that the acute effects of E₂ on SERT function are mediated by different as well as common signaling pathways. Estradiol- or DPN-induced slowing of 5-HT clearance mediated by ERβ was blocked after inhibition of MAPK/ERK1/2 but not that of PI3K/Akt signaling pathways. Estradiol’s or PPT’s inhibition of the fluvoxamine-induced slowing of 5-HT clearance mediated by ERα was blocked after inhibition of either MAPK/ERK1/2 or PI3K/Akt signaling pathways [19]. In addition, estrogen and BDNF seem to share common targets, effects, and mechanisms of action [20]. Further, there is evidence for the neurotrophin TrkB receptor being a mediator of E₂’s effects in the hippocampus [21]. Consistent with this, blockade of the TrkB receptor abolished the acute effect of E₂ and DPN on 5-HT clearance but had no effect on E₂’s or PPT’s inhibition of the fluvoxamine-induced slowing of 5-HT clearance [19].

The brain’s serotonergic system is complex, in part due to the diffuse distribution of its multiple receptor subtypes and the various signal transduction pathways regulated by these receptors [22]. Modulation of 5-HT neurotransmission has long been a primary pharmacological target for the treatment of depression and more recently anxiety disorders. SSRIs enhance synaptic 5-HT action by blocking the reuptake of 5-HT [23]. Global enhancement of serotonin neurotransmission may activate all subtypes of serotonin receptors in brain, whereas each 5-HT receptor subtype has different and specific functions in defined brain regions [24, 25]. Similar to estradiol’s signaling, two major signaling pathways have been implicated in responses underlying antidepressant effects; the PI3K/Akt/Glycogen synthase kinase-3 (GSK3) signaling pathway and the MAPK/ERK signaling pathway [24, 26]. In addition to its interaction with estrogen, BDNF and its TrkB receptor is also involved in the mechanism(s) of action of ADs [27]. Phosphorylation of the TrkB catalytic domain (tyrosine residue Y705) is considered to be the initial step in TrkB receptor activation, which further regulates the phosphorylation and activation of other tyrosines of which Y515 (the docking site of the Shc adaptor protein) and Y816 have been most extensively studied [28]. Antidepressant treatment increased TrkB phosphorylation at Y705 as well as at Y816 but not at Y515 [29–31].

Previous studies primarily examined the acute effects of estradiol or subtype selective agonists. However, estradiol is often given chronically, particularly for adjunctive treatment of postmenopausal depressed patients [32, 33]. Therefore, this study examined the behavioral consequences of longer-term treatments with E₂ or ER subtype selective agonists and/or an SSRI in the forced swim test (FST). A second part of this study investigated whether behavioral effects of long-term E₂ are associated with changes in the signaling pathways and /or modulatory receptors (TrkB) in the hippocampus accompanying ER activation and/or SSRI-induced antidepressant-like effects. As stated above, the signaling pathways (MAPK/ERK1/2 and PI3K/Akt) and TrkB receptor, selected in this study, are involved not only in signaling through estrogen receptors but also are prominently involved in many effects produced by antidepressants.

METHODS

Animals

Ovariectomized (OVX) rats (Sprague-Dawley; 250–280g, 4 months old, Harlan, Indianapolis, IN) were housed on a 12:12h light/dark cycle with lights on at 07:00 and with food and water provided ad libitum. All animal procedures were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the local Institutional Animal Care and Use Committee. All efforts were made to minimize the number of animals used, or stress and discomfort to the animals during the experimental procedure. OVX rats were allowed 2–3 weeks recovery after surgery before the start of the experiment.

Drug treatment

For the FST experiments, in experiment one, OVX rats were administered for two weeks the following compounds subcutaneously via implantation of one osmotic minipumps containing either: estradiol (E₂, 5µg/day, Sigma-Aldrich, St. Louis, MO); the ERβ agonist diarylpropionitrile (DPN, 5 or 10µg/day, Bio-Techne, Minneapolis, MN); the GPR30 agonist (G1 at 5–100µg/day, Bio-Techne) or the SSRI sertraline (10mg/kg/day). For experiment two, examining the effect of the ERα agonist PPT (Bio-Techne) on the AD-like effect of sertraline, two osmotic minipumps were subcutaneously implanted into rats. One was filled with either PPT or vehicle and the other with sertraline or vehicle as shown in the following table. Control group received vehicle which consisted of 25% EtOH/H₂O.

GROUPS:	control	PPT-5	PPT-50	sertraline	PPT-5/ sertraline	PPT-50/ sertraline
osmotic minipump 1	vehicle	PPT (5µg)	PPT (50µg)	vehicle	PPT(5µg)	PPT(50µg)
osmotic minipump 2	vehicle	vehicle	vehicle	sertraline	sertraline	sertraline

In these experiments, 6 to 14 rats were used in each group.

For the Western blot analysis experiments, OVX rats were implanted with two osmotic minipumps subcutaneously for two weeks. The treatments were as shown in the following table.

GROUPS:	control	E2	DPN	PPT	PPT/sertraline	sertraline	G1
Osmotic minipump 1	vehicle	E2 (5µg)	DPN (10µg)	PPT (50µg)	PPT	vehicle	G1 (10µg)
Osmotic minipump 2	vehicle	vehicle	vehicle	vehicle	sertraline	sertraline	vehicle

In these experiments, 6 rats were used in each group.

The dose of E₂ used in this study was shown to produce serum estradiol levels seen in proestrus [34]. The initial doses used for ER agonists selected are based on that same study. However when no effect was obtained with a low dose, higher doses were also tried to make sure that the lack of effect was not due to a dose that was too low. Control groups received vehicle which consisted of 25% EtOH/H₂O.

The behavioral experiments shown in Figures 1 & 2 consisted of 2 cohorts. In the first cohort a lower dose of E₂ and ER agonists was used. In a second cohort a higher dose of the agonists was used when necessary. Each cohort had its own control group (n=7). The results, with low and high doses were combined in the same figures and the two separate control groups were also to provide a larger n of 14.

Figure 1.

Effect of long-term treatment with E₂, ERβ, GPR30 agonists and sertraline in the FST. OVX rats were treated for 2 weeks with E₂, DPN, G1 or sertraline via subcutaneous osmotic minipumps, as described in the methods section. Mean counts for immobility, swimming and climbing behaviors were sampled every 5 sec of the swim test period. Bars and brackets represent the mean value ± SD, n= 6–14/group. None of the treatment groups altered climbing behavior (data not shown).

a- Immobility: One way ANOVA revealed that there was a significant main effect between groups (F_(6,48)=17.11, p<0.001). Dunnett’s post-hoc analysis was carried out; *p<0.001, comparing all groups with vehicle.

b- Swimming: One way ANOVA revealed that there was a significant main effect between groups (F_(6,48)=14.26, p<0.001). Dunnett’s post-hoc analysis was carried out; *p<0.001, comparing all groups with vehicle.

Figure 2.

Effect of long-term treatment with an ERα agonist (PPT) alone or combined with an SSRI in the FST. OVX rats were treated for 2 weeks with PPT and/or sertraline via subcutaneous osmotic minipumps, as described in the methods section. Mean counts for immobility, swimming and climbing behaviors were sampled every 5 sec of the swim test period. Bars and brackets represent the mean value ± SD, n= 6–14/group. None of the treatment groups altered climbing behavior (data not shown).

a- Immobility: Two way ANOVA showed a significant main effect for sertraline (F_(1,54)=33.56, p<0.001) and for PPT (F_(3,54)=4.58, p<0.006); as well as a significant interaction between sertraline X PPT (F_(3,54)=10.32, p<0.001); Dunnett’s post-hoc analysis was carried out *p<0.001, comparing PPT within vehicle and sertraline within vehicle and #p<0.001, comparing PPT alone with PPT and sertraline.

b- Swimming: Two way ANOVA showed a significant main effect for sertraline (F_(1,54)=44.52, p<0.001) and for PPT (F_(3,54)=6.43, p<0.001); as well as a significant interaction between sertraline X PPT (F_(3,54)=12.80, p<0.001); Dunnett’s post-hoc analysis was carried out *p<0.001, comparing PPT within vehicle and sertraline within vehicle and #p<0.001, comparing PPT alone with PPT and sertraline.

The dose of sertraline was selected as it was shown not only to downregulate the SERT but also to produce an AD-like effect in the FST when given chronically [35–37].

The efficacy of the treatment paradigms were confirmed by body and uterine weight. It has been shown that ERα mediates the attenuating effect of E₂ on body weight gain as well as that of increased uterine weight after E₂ treatment [38–44]. Changes in body weight after different treatment were measured (rats weight at the end of the treatment was subtracted from their weight just before the start of treatment). In addition, at the end of long-term treatment with E₂ or ER agonists, uteri were collected and weighed immediately after removal.

Forced swimming test

A modified FST procedure from the one described by Cryan et al., (2002) was used. On day 15 of chronic treatment (while hormones and/or SSRIs were still administered), a swim test session of 5min was carried out with no pretest session. This approach is used in order to avoid the extended delay (2 weeks) between the first (pretest) and the second exposure (test) or the impact of hormone being present during the pretest session if pretest and test are separated by the usual 24h. In addition, it has been shown that positive results with the drugs could be obtained even if no training session was carried [45]. Rats were placed individually into a Plexiglas cylinder (21 × 46 cm) filled with 25°C water to a depth of 35-cm for a 5min swim and behaviors were recorded by a video camera positioned above the tank. A time sampling technique was employed whereby the predominant behavior in each 5 sec bin of the test was analyzed. Climbing was defined as upward-directed movements of the forepaws against along the side of the swim chamber. Swimming was defined as active movements (usually horizontal) throughout the swim chamber, which also include crossing into another quadrant. Immobility is assigned when no active movement other than that necessary to keep the rat’s head above the water. The rater was blind with respect to the experimental conditions being scored.

Western blot analysis

To prevent changes in phosphorylation level of proteins due to stress from the FST experiment, another set of rats was used for these experiments. OVX rats were treated as described above. After completion of drug treatment (14 days), the hippocampus was dissected and homogenized in lysis buffer containing 50mM Tris-HCl, 1mM EDTA, 0.35% sodium deoxycholate, 150mM NaCl, 1% Igepal, phosphatase inhibitors and protease inhibitors. After incubation on ice for 30min, samples were centrifuged for 15min at 13,000g. Protein levels of the collected supernates were determined using the Bradford assay. Each sample was run on SDS-PAGE gel and blotted onto a nitrocellulose membrane. The blot was incubated overnight at 4°C with the following primary antibodies: anti-phospho TrkB (from Abcam, Cambridge, MA): anti-phospho-Y515 (1:1000), anti-phospho-Y705 (1:1000), antiphospho-Y816 (1:4000), anti-TrkB (1:10,000, Neuromics, Edina, MN), and the following primary antibodies (from Cell Signaling Technology, Danvers, MA) all used at 1:1000: antiphospho-Akt (S473), anti-phospho-Akt (T308), anti-Akt, anti-phospho-ERK1/2) and anti-ERK1/2. Equal loading was confirmed using anti-β-actin (1:200,000, Sigma-Aldrich) and used for normalization. After three washes, the membranes are incubated with the appropriate horseradish peroxidase conjugated secondary antibody (1:20,000, Sigma-Aldrich) for 1hr at RT. Enhanced chemiluminescence (ECL) detection is used followed by an exposure to X-ray film for ECL detection. Relative densities were measured using the NIH imageJ program. Phospho- protein values are normalized and calculated as a ratio against total protein values. The results are shown graphically as a percentage of the control group.

Statistical analysis

Data were analyzed using SigmaStats (Systat Software Inc, San Jose, CA). One Way Analysis of Variance (ANOVA) followed by Dunnett’s analysis was used when comparisons were made with the control group (Tables 1, 2, 4; Figure 1). Two Way ANOVA followed by Dunnett’s analysis was used to compare (1) effect of PPT within the vehicle and within the sertraline groups, and (2) effect of sertraline within the vehicle and within the PPT groups (Table 3; Figure 2). For all the Western blot analysis, where the data are expressed as percentage of control, Kruskal-Wallis test followed by Dunnett’s post-hoc analysis was used to compare all groups with vehicle and the sertraline group with the PPT + sertraline group (Figures 3–6). Only when there was a significant main effect and/or interaction effect in the ANOVA’s were post-hoc analyses carried out. Significance was determined at p < 0.05.

Table 1.

Uterine weight (in mg) after 14 days treatment with vehicle, E₂ or ER agonists

Vehicle (n=14)	E2-5µg (n=7)	DPN-5µg (n=7)	DPN-10µg (n=6)	G1–5µg (n=7)	G1–10µg (n=7)	PPT-5µg (n=7)	PPT-50µg (n=7)
103±4.7a	418*±12.7	98±1.6	104±4.4	103±5.2	108±2.2	350*±10.3	508*±14.9

OVX rats were treated for 2 weeks with drugs via osmotic minipumps, as described in methods section. At the end of treatment/experiment, the uterine horn was dissected and weighted.

^arepresent mean value ± SD.

One way ANOVA showed a significant main effect (F_(7,53)=3339.38, p < 0.001). Dunnett’s posthoc analysis was carried out,

^*p < 0.001, comparing all groups with the control group.

Table 2.

Change in body weight (in g) after 14 days of treatment with vehicle, sertraline, E₂, ERα or GPR30 agonists

Vehicle (n=14)	Sertraline (n=7)	E2-5µg (n=7)	DPN-5µg (n=7)	DPN-10µg (n=6)	G1–5µg (n=7)	G1–10µg (n=7)
+36±7.5a	+31±9.2	−1*±15.0	+29±6.4	+41±9.6	+29±9.0	+38±4.5

^aAll the numbers are change in body weight ± SD after chronic treatment with drugs: the weight of the rat at the end of treatment was subtracted from the weight of the rat at the beginning of the treatment.

One way ANOVA was carried out. There was a significant main effect between groups (F_(6,48)=17.04, p < 0.001). Dunnett’s post-hoc analysis was carried out

^*p < 0.001, comparing all groups with vehicle.

Table 4.

Effect of long-term treatment with G1 (10µg/day, 2 weeks) on the phosphorylation levels of Akt, ERK and TrkB in the hippocampus of OVX rats

groups	pAkt (T308)	pERK2 (Y204)	pTrkB (Y816)
Control (n=6)	99.6±13a	99.5±9	100.4±22
G1 (n=6)	134.0±15*	141.2±15*	159.8±17*

OVX rats were treated for 2 weeks with G1 or vehicle (control group) via osmotic minipumps, as described in methods section. Western blot analysis of hippocampal levels of total-protein and that of phospho-protein for Akt, ERK and TrkB (Y816) were carried out.

Phospho-protein values are normalized and calculated as a ratio against total protein values. Total proteins levels for Akt, ERK and TrkB were not altered by G1 treatment when compared to values in the vehicle treatment group (data not shown).

^arepresent mean value of a percentage of the control group ± SD.

One way ANOVA followed by Dunnett’s post-hoc analysis were carried out.

^*p < 0.05, comparing each group with their control groups.

Table 3.

Change in body weight (in g) after 14 days treatment with the ERα agonist PPT and/or sertraline

veh/veh (n=11)	PPT-5µg/veh (n=7)	PPT-50µg/veh (n=6)	veh/sertr (n=11)	PPT-5µg/sertr (n=7)	PPT-50µg/sertr (n=7)
+45±6.7a	+37±5.9	−1*±4.9	+40±10.8	+40±10.4	+32.9#±8.3

^aAll the numbers are change in body weight ± SD after chronic treatment with drugs: the weight of the rat at the end of treatment was subtracted from the weight of the rat at the beginning of the treatment.

Two way ANOVA was carried out for weight change in sertraline and PPT groups. There was a significant main effect for sertraline (F_(1,43)=18.27, p < 0.001) and for PPT(F_(2,43)=42.76, p < 0.001), as well as a significant interaction between SSRIxPPT (F_(2,43)=22.46, p < 0.001).

Dunnett’s post-hoc analysis was carried out *p < 0.001, comparing PPT within vehicle and within sertraline groups and #p < 0.001, comparing sertraline within vehicle and within PPT groups.

Figure 3.

Effect of long term treatment with E₂, DPN, PPT and/or sertraline on phosphorylation levels of Akt (T308) in the hippocampus. OVX rats were treated for 2 weeks with drugs via osmotic minipumps, as described in methods section. Western blot analysis of hippocampal levels of total Akt and that of phospho-Akt (T308) were carried out. Representative Western blotting of phospho- and total Akt are shown in the upper insert panel. A single band was detected at approximatively 60 kDa that correspond to phospho-Akt, or total Akt. Bars represent mean value ± SD of phospho to total protein ratio (n=6).

Kruskal-Wallis test showed a significant main effect (F_(5,30)=7.52, p<0.001). Dunnett’s post-hoc was carried out, *p<0.01, comparing all groups with vehicle.

Figure 6.

Effect of long term treatment with E₂, DPN, PPT and/or sertraline on phosphorylation levels of TrkB (Y816) in the hippocampus. OVX rats were treated for 2 weeks with drugs via osmotic minipumps, as described in methods section. Western blot analysis of hippocampal levels of total TrkB and that of phospho-TrkB (Y816) were carried out. Representative Western blotting of Y816- and total TrkB are shown in the upper insert panel. A single band was detected at approximatively 92kDa that correspond to Y816, or total TrkB. Bars represent mean value ± SD of phosphoprotein to total protein ratio (n=6).

Kruskal-Wallis test showed a significant main effect (F_(5,30)=14.07, p<0.001). Dunnett’s post-hoc analysis was carried out, *p<0.001, comparing all groups with control group, and ^#p<0.05, comparing sertraline group with PPT + sertraline group.

RESULTS

Long-term effects of treatment with E₂ and ER agonists on uterine weight

Uterine weights were investigated as a peripheral markers of estrogenic activity. As expected, treatment with estradiol and PPT (5 as well as 50µg) increased significantly uterine weight, as compared to the weights in vehicle-treated rats. By contrast, treatment with either DPN or G1 (each at 5 or 10 µg) did not have any effect (Table 1).

Long-term effects of treatment with E₂, ER agonists and/or sertraline on the change in body weight

The initial body weights of the various groups before treatments were not significantly different (data not shown). Shown in Tables 2 and 3 are the changes in body weight produced by the various treatments. After 2 weeks of treatment with vehicle, there was a 36–45g gain in body weight. Long-term treatment with an ERβ or a GPR30 agonist also resulted in a weight gain that was not statistically different from that obtained after treatment with vehicle. However treatment with E₂ resulted in a small weight loss (−1g) that was statistically different from the weight gain obtained after treatment with vehicle (Table 2). Long-term treatment with sertraline also induced a similar increase in body weight (31–40g) as that measured in control rats (Table 2 and Table 3). Long term treatment with the ERα agonist PPT had a dose-dependent effect on body weight. At a low dose (5µg), weight gain was equivalent to that obtained in control rats. However at a higher dose (50µg), PPT treatment decreased significantly the body weight compared to the body weight gain seen in vehicle rats. It is interesting to note that in the presence of sertraline the effect of the higher dose of PPT on body weight was not observed (Table 3).

Long-term effects of E₂, ER agonists and/or sertraline in the FST

The behavioral consequences of longer-term treatment with E₂ or ER subtype specific agonists and/or sertraline were examined in the FST. As expected, chronic treatment with sertraline induced AD-like behavior in the FST, i.e., decreased immobility and increased swimming behavior as compared to that in control rats. Similarly, AD-like effects were found after chronic treatment with E₂, DPN (5 or 10µg) or G1 (at 10 but not at 5 µg) (Figure 1). Increasing doses of DPN (from 5 to 10µg) or G1 (from 10 to 100µg) did not produce greater behavioral effects (data not shown for G1 above 10µg). None of the treatments altered climbing behavior (data not shown).

Chronic treatment with the ERα agonist, PPT at doses up to 100 µg had no effect on immobility or swimming behaviors compared to those measured in control rats. However, the decreased immobility and increased swimming behaviors induced by chronic treatment with sertraline were blocked by long-term treatment with PPT (at either 50 or 100µg). PPT at 5µg did not alter the sertraline-induced decreased immobility but significantly lowered the sertraline-induced increased swimming behavior in the FST (Figure 2). Again, none of the treatments altered climbing behavior (data not shown).

Long-term effects of E₂, ER agonists and/or sertraline on phosphorylation levels of TrkB and signaling proteins

Activation of signaling proteins was assessed by increases in their phosphorylation status in the hippocampus after long-term (14 days) treatment with E₂, ER agonists and/or sertraline, using Western blot analysis. Total protein levels for Akt, ERK and TrkB were not altered by any of the treatments when compared to values in the vehicle treatment group (Figures 3–6). Akt was examined at two phosphorylation sites Threonine (T) 308 and Serine (S) 473. Phospho -Akt (T308) was significantly increased after treatment with E₂, G1, PPT, sertraline and the combination of PPT with sertraline; however, treatment with DPN did not alter the phosphorylation level of Akt as compared to control values (Figure 3, Table 4). Similar results were obtained with phospho-Akt at S473 (data not shown). Phospho-ERK2 (at tyrosine (Y) 204) was significantly increased in all treatment groups as compared to the values in the control group (Figure 4, Table 4). Western Blot analysis of TrkB phosphorylation revealed that phosphorylation levels at tyrosine 705 were increased after long-term treatment with E₂, DPN, and sertraline alone or when the SSRI was combined with PPT. While long-term treatment with PPT alone did not alter phosphorylation at Y705 as compared to that seen in the control group, it is interesting to note that the combination of sertraline and PPT produced an increase in Y705 that was significantly smaller than that obtained after sertraline alone (Figure 5). A similar pattern of results were obtained at Y816 on TrkB as treatment with E₂, DPN, G1 and sertraline alone or sertraline + PPT increased the phosphorylation levels at tyrosine 816, whereas treatment with PPT alone did not have any effect as compared to that seen in the control group (Figure 6 and Table 4). Again, the combination of sertraline and PPT produced an increase in Y816 that was significantly smaller than that obtained after sertraline alone. There were no changes in TrkB phosphorylation at Y515 (data not shown).

Figure 4.

Effect of long term treatment with E₂, DPN, PPT and/or sertraline on phosphorylation levels of ERK2 (Y204) in the hippocampus. OVX rats were treated for 2 weeks with drugs via osmotic minipumps, as described in methods section. Western blot analysis of hippocampal levels of total ERK and that of phospho-ERK (Y204) were carried out. Representative Western blotting of phospho- and total ERK1/2 are shown in the upper insert panel. Bands of phospho-p42 and phospho-p44, ERK-like immunoreactivity at approximately 42 and 44 kDa, respectively were detected. Only data from phospho-ERK2 are quantified, pERK1 was not quantified due to poor immunoreactivirty. Bars represent mean value ± SD of phospho to total protein ratio (n=6).

Kruskal-Wallis test showed a statistically significant difference among the treatment groups (p=0.005). Dunnett’s post-hoc analysis was carried out, *p<0.05, comparing all groups with control group.

Figure 5.

Effect of long term treatment with E₂, DPN, PPT and/or sertraline on phosphorylation levels of TrkB (Y705) in the hippocampus. OVX rats were treated for 2 weeks with drugs via osmotic minipumps, as described in methods section. Western blot analysis of hippocampal levels of total TrkB and that of phospho-TrkB (Y705) were carried out. Representative Western blotting of Y705- and total TrkB are shown in the upper insert panel. A single band was detected at approximatively 92kDa that correspond to Y705, or total TrkB. Bars represent mean value ± SD of phospho to total protein ratio (n=6).

Kruskal-Wallis test showed a significant main effect (F_(5,30)=18.61, p<0.001). Dunnett’s post-hoc analysis was carried out, *p<0.001, comparing all groups with control group, and ^#p<0.001, comparing sertraline group with PPT + sertraline group.

DISCUSSION

These results demonstrate that long-term treatment of young adult OVX rats with estradiol, ERβ or GPR30 selective agonists induced an AD-like behavior in the FST, i.e., decreased immobility and increased swimming behavior. This effect is similar to that obtained after chronic treatment with the SSRI sertraline. Long-term treatment with the ERα agonist (PPT) had no effect of its own in the FST; however, it blocked the AD-like effect of sertraline. Furthermore long-term treatment with estradiol, GPR30 agonist or sertraline increased the phosphorylation levels of signaling proteins such as Akt, ERK and that of the TrkB receptor in the hippocampus. Long-term treatment with ER-subtype selective agonists induced similar changes in the levels of phosphorylation of signaling proteins as that obtained with estradiol, except that treatment with DPN did not alter phospho-Akt, whereas PPT did not change TrkB phosphorylation levels.

The lack of a gain in body weight associated with increased uterine weight after long term treatment with E₂ validates the efficacy of the dose of E₂ used (Tables 1 and 2). ER subtype selective agonists had different effects on body and uterine weight (Tables 1–3). PPT both increased uterine weight and at a dose of 50 µg maintained body weight at a level similar to E₂ consistent with the idea that ERα plays a predominant role in uterine proliferation and maintenance of body weight [38, 43, 46, 47]. DPN and G1 did not alter uterine weight, in agreement with previous reports [39, 48]. Weight gain also increased with these agonists to an extent similar to that seen in the vehicle-treated rats.

The AD-like effects of estradiol mediated via ERβ have been shown in rodents in several studies, using subtype-selective agonists as well as knockout (KO) animals. In acute studies, administration of E₂ or ERβ-selective estrogen receptor modulators (SERMs) systemically or directly into the hippocampus, decreased anxiety and depressive behavior whereas administration of ERα-specific SERMs did not [9, 49]. Similarly, the AD-like effect of estradiol or DPN in the FST was shown in mice and this effect was lost in ERβ-KO mice [7].Long-term treatment of OVX rats with E₂ or DPN induced an AD-like effect in the FST whereas treatment with PPT had no effect [50, 51]. Consistent with these data, the results obtained in this study indicate that E₂, DPN as well as G1 have an antidepressant-like effect in the FST after long-term treatment. Most importantly, PPT by itself had no effect in the FST, but it blocked the AD-like effect of sertraline (Figures 1 and 2). Even though the lower dose of PPT increased uterine weight, PPT at this dose had no effect on the AD-like effect of sertraline (Table 2, Figure 2). It is possible that this low dose was not physiologically effective since it was not associated with a lack of body weight gain as was the case for the higher dose (50µg).

The long term treatment data from the FST are consistent with the results obtained previously using chronoamperometry to measure the rate of clearance of 5-HT from the CA₃ region of the hippocampus [10, 13]. In these studies, the AD-like effect of E₂ (i.e., slower 5-HT clearance) was due to ERβ and/ or GPR30 activation as their specific agonists mimicked the effect of E₂. By contrast, E₂’s blockade of the SSRI fluvoxamine’s ability to slow serotonin clearance was mimicked only by an ERα agonist [10].

The regulation of serotonergic system by estradiol leading to the behavioral responses is mediated by ER subtypes that can be located in 5-HT neurons and/or non-5-HT neurons. In general, it is thought that ERβ but not ERα is present in serotonin neurons. However, the patterns for expression of the two ER subtypes in 5-HT neurons in the dorsal raphe nucleus (DRN) are species dependent as well as controversial. In rats it was shown that 5-HT neurons of the DRN contain mRNA for ERβ but not ERα [52]. Distinct groups of 5-HT immunoreactive neurons in the DRN were found to show ERα and/or ERβ expression in mice but not in rat [53]. In mice ERβ, but not ERα was found located within 5-HT neurons [54, 55]. In non-human primates, ERβ, but not ERα, seems to be the predominant ER expressed in raphe 5-HT neurons [56–58]. Although GPR30 immunoreactivity was found throughout the brain, it is however not known if GPR30 is localized in 5-HT neurons [59, 60]. Local activation of ERβ by either E₂ or DPN in the DRN increased the expression of tryptophan hydroxylase-2 (TPH2), the brain specific rate limiting enzyme for 5-HT synthesis, and decreased despair-like behavior in the FST; however, it failed to decrease anxiety like behavior [61]. That study demonstrated that estradiol’s regulation of despair-like behavior may primarily involve ERβ located in the DRN whereas its effect in regulating anxiety-like behavior could involve ER located in other brain areas such as the hypothalamus or the amygdala. Further studies will be necessary to investigate if the ERs mediating the effect of estradiol on serotonergic function obtained in the present study are located on 5-HT and/or non-5-HT neurons.

Although we have concentrated our effort on the effect of estradiol in the hippocampus, it is important to note that effects of estradiol occur throughout the brain. In addition to the hippocampus, affective behaviors are also modulated by other limbic structures including the amygdala. This brain region contains high amount of estrogen receptors [62]. The amygdala is well known as a brain region important for modulating fear and anxiety behavior [63, 64] and has also been investigated for its importance for affective behavior through a kindling model [65]. The importance of the amygdala for the sex-specific or hormone dependent modulation of neurobiological changes underlying anxiety disorders has been demonstrated in several studies. For example, lesions of the central nucleus of the amygdala lead to decreased anxiety type behaviors and decreased neuroendocrine responses to a spectrum of stressors [66, 67]. Systemic as well as intra-amygdala estradiol administration increased antianxiety behavior in OVX rats [49].

There is an interplay between ERs, certain signaling cascades and the TrkB receptor in the behavioral effects of estradiol. Activation of dorsal hippocampal PI3K/Akt and ERK signaling pathways is necessary for E₂ to enhance object recognition memory in young as well as middle aged female mice [68, 69]. Similarly, the activation of ERα, ERβ or GPR30 results in induction of ERK-Akt signaling pathways mediating neuroprotection and preservation of cognitive function [70, 71]. Both MAPK and PI3K are required in the neuronal actions of estrogen in facilitating female reproductive behavior as lordosis induced by estradiol in rats was abolished in the presence of specific inhibitors of the two signaling pathways [72]. In addition to the importance of these two signaling pathways in the effects of estradiol, the activation of the TrkB receptor plays a critical role in E₂-induced neuroprotection from NMDA toxicity [73]. Furthermore, evidence for the TrkB receptor as a mediator of E₂’s effects in the hippocampus has been demonstrated [18].

In this study, long-term treatment with estradiol increased phosphorylation levels of Akt, ERK and TrkB in the hippocampus and these effects were similar to those observed after chronic treatment with sertraline (Figures 3–6). Long-term treatment with a GPR30 agonist produced effects similar to those of E₂ (Table 4). Interestingly, long-term treatment with ERα and ERβ agonists produced a different pattern of results such that Akt phosphorylation was not changed after DPN treatment whereas PPT treatment did not alter TrkB phosphorylation (Figures 3, 5 and 6). PPT, unlike DPN, did not display an AD-like effect in the FST (Figures 1 and 2). Taken together, one might hypothesize that TrkB activation is necessary for an AD-like effect to occur in response to SSRIs, estradiol, ERβ and GPR30 selective agonists. By contrast, activation of Akt may not be required for the DPN-induced AD-like effect in the FST, and perhaps that of estradiol and G1 as well. Recent published findings from acute studies strengthen this hypothesis. Using selective inhibitors of signaling pathways or that of interacting receptors showed that the acute E₂-induced slowing of serotonin clearance via activation of ERβ required MAPK/ERK1/2 signaling pathways and involved interactions with the TrkB receptor. However, the E₂-induced prevention of the ability of an SSRI to slow serotonin clearance, via activation of ERα, required MAPK/ERK1/2 and PI3K/Akt signaling pathways but not TrkB receptor activation [19]. Perhaps consistent with this idea also is that long-term treatment with the combination of PPT and sertraline reduced significantly the level of phosphorylation of TrkB as compared to sertraline alone (Figures 5 and 6) and it also blocked the AD-like effect of sertraline (Figure 2). However, further studies using inhibitors of signaling pathways or that of interacting receptors may offer a better perspective on the impact of signaling pathways and the TrkB receptor in the behavioral effects observed after long term-treatment with E₂ or ER selective agonists. Nonetheless, these data provide insight into the intracellular activation of specific signaling proteins that accompany these treatments and that could be involved in the antidepressant-like behavioral responses obtained with estradiol.

In addition to its interaction with estrogen, BDNF/TrkB is also involved in the mechanism(s) of action of ADs [27] and perhaps the pathogenesis of depression [74]. Evidence that activation of the TrkB receptor is required for a behavioral response typically induced by ADs has been shown [27]. Numerous studies have also implicated MAPK and PI3K pathways in the etiology and treatment of mood disorders. Pharmacological manipulations of these signaling pathways also affect behavior in models of depression and antidepressant response [24, 26, 75]. These pathways have become major pharmacological targets for putative neuropsychiatric treatments. Our results strengthen the idea that estradiol activation of ERα, ERβ or GPR30, lead to behavioral effects that could be due at least in part to activation of ERK, Akt and/or the TrkB receptor in the hippocampus. Such information could be useful in identifying novel drug targets that would bypass ERs.

In conclusion, estradiol has both an antidepressant-like effect as well as an effect that interferes with the efficacy of SSRIs. Different estrogen receptors (ERs) and differences in ER-coupled signaling pathways produce these two different effects of estradiol. Understanding the way in which estrogens affect SERT function in young adult animals is a first step to defining age-related changes in response to hormones. Whether estradiol has effects in a rodent model of peri- or post-menopausal depression similar to effects seen in younger adult rats is under investigation. And if so, then ERβ selective agonists, or novel drugs that would target their signaling pathways, are likely to be more efficacious in postmenopausal depressed patients than estradiol.