The Post-Ovariectomy Interval Affects the Antidepressant-Like Action of Citalopram Combined with Ethynyl-Estradiol in the Forced Swim Test in Middle Aged Rats

The use of a combined therapy with low doses of estrogens plus antidepressants to treat depression associated to perimenopause could be advantageous. However the use of these combinations is controversial due to several factors, including the time of intervention in relation to menopause onset. This paper analyzes whether time post-OVX influences the antidepressant-like action of a combination of ethynyl-estradiol (EE2) and citalopram (CIT) in the forced swim test (FST). Middle-aged (15 months old) female Wistar rats were ovariectomized and after one or three weeks treated with EE2 (1.25, 2.5 or 5.0 µg/rat, s.c.; −48 h) or CIT (1.25, 2.5, 5.0 or 10 mg/kg, i.p./3 injections in 24 h) and tested in the FST. In a second experiment, after one or three weeks of OVX, rats received a combination of an ineffective dose of EE2 (1.25 µg/rat, s.c., −48 h) plus CIT (2.5 mg/kg, i.p./3 injections in 24 h) and subjected to the FST. Finally, the uteri were removed and weighted to obtain an index of the peripheral effects of EE2 administration. EE2 (2.5 or 5.0 µg/rat) reduced immobility after one but not three weeks of OVX. In contrast, no CIT dose reduced immobility at one or three weeks after OVX. When EE2 (1.25 µg/rat) was combined with CIT (2.5 mg/kg) an antidepressant-like effect was observed at one but not three weeks post-OVX. The weight of the uteri augmented when EE2 was administrated three weeks after OVX. The data suggest that the time post-OVX is a crucial factor that contributes to observe the antidepressant-like effect of EE2 alone or in combination with CIT.


Introduction
Women report more depression than men (by a factor of 2:1), particularly associated with their reproductive life span [1]. In fact, the onset or exacerbation of depressive symptoms has been associated to perimenopause [2,3]. Some reports indicate that the antidepressant response could be modified by the endocrine condition, since premenopausic women are more responsive than post-menopausic females [4][5][6]. However, other reports do not show differences in the response to antidepressants during the perimenopause transition [7,8]. Important methodological differences could contribute to explain these controversial results; for example, the time of intervention in relation to menopause onset, called the "critical period" [9,10] as well as whether menopause is natural or induced. In this line, recent reports suggest that an opportune intervention during early menopause is more effective than during late-menopause [11][12][13][14][15]. However, Henderson and Popat [16] in a systematic review, reported that the literature is not enough to support or decline this hypothesis.
Estrogen replacement therapy (ERT) is an effective treatment for depression associated to perimenopause, along with its beneficial effects on other climacteric symptoms like osteoporosis and hot flashes [17]. The use of a combined therapy of antidepressants plus ERT to treat depression and climacteric symptoms seems useful. For example, ERT may shorten the onset of the therapeutic effects of antidepressants [5,[18][19][20] and may reduce vasomotor symptoms, prevent osteoporosis and cardiovascular insults, as well as promoting neuroprotection [21][22][23]. However, controversial results have restricted their utilization [17,24], among them the efficacy of ERT to reduce depression [8,25] versus their putative adverse effects. For example, the combined use of ERT plus antidepressants has been related to a higher risk of developing cancer, particularly in long-term users and in women with a history of breast cancer [26,27]. In contrast, short-term treatment with a low-dose of conjugated equine estrogens and fluvoxamine is effective and safe for oophorectomized women in relieving hot flashes and depression [28]. Preclinical data in models that have predictive value, using young rats and at a constant post-OVX interval, showed that estrogens shorten and potentiate the antidepressant-like action of selective serotonin reuptake inhibitors (SSRIs) [21,29]. However there are no studies exploring whether age and the post-OVX interval modify the antidepressant-like action of the combination of estrogens and antidepressants.
Interestingly, the antidepressant-like effects of estrogens appear to depend on both the age and post-OVX interval [30][31][32][33]. For instance, 5 or 10 µg/rat of 17 β-estradiol (E 2 ) produced an antidepressant-like action in 3 month old OVX rats, while in 15 month old rats, only 10 µg/rat had an effect; suggesting that E 2 sensitivity decreases with age [34]. Furthermore, no action was reported in senescent rats [30,31]. Interestingly, aging also abbreviates the intervention window with E 2 , since the point of restitution in relation to time after OVX is three months after OVX in young rats, while in aged rats it is restricted to one week [13,[32][33][34]. The action of antidepressants is also modulated by age, because their effect decreases in middle-age rats and may vanish at senescence [34][35][36].
Therefore, the aim of the present study was to explore whether the post-OVX interval influences the antidepressant-like action of 17α-ethynyl estradiol (EE 2 ) and citalopram (CIT) alone or in combination in the forced swim test in rats aged 15 months. As an index of EE 2 's activity in peripheral tissues, the uterus weight was evaluated [32,37]. Citalopram was chosen because of its wide use in clinical practice due to less pharmacokinetic interactions with other drugs [38,39], particularly in middle-aged women that frequently consume several drugs [23]. EE 2 was selected because it is an estrogen commonly used in ERT, is more potent than E 2 in reducing immobility behavior in the FST [40] and possesses a wider window of effect in the FST after OVX [32]. Figure 1 shows the effect of several EE 2 doses (panel A) administered 1 or 3 weeks post-surgery on behavioral immobility in the FST. The two-way analysis of variance indicated differences for the factor time [F(1,51) = 23.81, p < 0.001] and dose of EE 2 [F(3,51) = 4.23, p = 0.009] without a significant interaction between them [F(3,51) = 1.25, ns]. Post-hoc comparisons revealed that at 1 week post-OVX (Panel A), EE 2 reduced the immobility behavior compared to the control group at doses of 2.5 µg/kg (p = 0.008) or 5.0 µg/kg (p = 0.04). However, at 3 weeks after OVX, no dose of EE 2 induced changes on immobility. Differences between one and three weeks were observed at the doses of 2.5 or 5.0 µg/kg (p = 0.002 and p = 0.001, respectively). h before the test; n = 6-8 per group) at one or three weeks after ovariectomy in the forced swim test. The data represent the mean ± S.E.M. of the number of immobilities scored in intervals of 5-s during a 5-min test session. Holm-Sidack test * p < 0.05; ** p < 0.005 versus respective control group; && p < 0.005 versus three weeks. Figure 1B shows that no CIT dose tested reduced immobility at one or three weeks after OVX. The two-way analysis of variance showed no significant effect of time [F(1,63) = 0.02, ns]; dose of CIT [F(4,63) = 1. 64, ns] or the interaction between these factors [F(4,63) = 0.14, ns]. Table 1 shows the effect of EE2 or CIT on the active behaviors scored in the FST. EE2, at 5.0 µg/rat, increased swimming only one week after OVX (p < 0.001). This increase differed from the value of the control group and the group treated with EE2 at this dose but tested three weeks post-OVX (p < 0.001).  Figure 1B shows that no CIT dose tested reduced immobility at one or three weeks after OVX. The two-way analysis of variance showed no significant effect of time [F(1,63) = 0.02, ns]; dose of CIT [F(4,63) = 1.64, ns] or the interaction between these factors [F(4,63) = 0.14, ns]. Table 1 shows the effect of EE 2 or CIT on the active behaviors scored in the FST. EE 2 , at 5.0 µg/rat, increased swimming only one week after OVX (p < 0.001). This increase differed from the value of the control group and the group treated with EE 2 at this dose but tested three weeks post-OVX  The data are shown as mean˘S.E.M of the number of counts for swimming and climbing behaviors scored in 5-sec intervals in a 5-min session of forced swimming. † SAL = saline. Holm-Sidack test; ** p < 0.001 versus the control group; ## p < 0.001 versus three weeks. Table 2 shows the weight of uteri in 15 month old rats treated with different doses of EE 2 (1.25, 2.5 or 5.0 µg/rat, s.c.´48 before FST) at one or three weeks post-OVX. The uteri from rats treated acutely with oil three weeks after surgery were lighter than those from rats treated with oil after one week (p = 0.03). The groups that received EE 2 at one week post-OVX did not show significant changes in their uteri weight. In contrast, the groups that received a single dose of EE 2 at 2.5 or 5.0 µg/rat had significantly increased uteri weight compared to the control group (p < 0.05) and to the groups that received the same dose one week after OVX (p < 0.001). The two-way ANOVA revealed no significant differences for the factors time   Figure 2 shows the effect of a combined low dose of EE 2 (1.25 µg/kg) and CIT (2.5 mg/kg) administered at one or three weeks post-OVX on the FST. EE 2 or CIT alone did not modify the immobility behavior in the FST at one or three weeks post-OVX. In contrast, the combination EE 2 plus CIT tested one week after OVX decreased immobility compared to the control group (p < 0.05) and versus the same treatment at three weeks post-OVX (p < 0.001), when it failed to have an effect.   Table 3 shows the effect of the combination of low doses of EE2 plus CIT on swimming and climbing scored in the FST. EE2 in combination with CIT increased swimming at one, but not at three, weeks post-OVX compared to the control group and the group treated three weeks after OVX.  The data are presented as mean ± S.E.M of the number of counts for swimming and climbing scored in 5-sec intervals in a 5-min session of forced swimming. Holm-Sidack test ** p < 0.001 versus the control group; ## p < 0.001 versus three weeks.

Discussion
The antidepressant-like action of EE2, in contrast to that of CIT, depends on the post-OVX interval. Remarkably, the combination of a non-effective dose of EE2 plus CIT promotes an antidepressant-like action in the FST only one week post-OVX. In middle-aged female rats, the antidepressant-like effect of EE2 was observed at one but not at three weeks post-OVX. These results are partially in line with previous data showing that ERT produces an antidepressant-like effect in  Table 3 shows the effect of the combination of low doses of EE 2 plus CIT on swimming and climbing scored in the FST. EE 2 in combination with CIT increased swimming at one, but not at three, weeks post-OVX compared to the control group and the group treated three weeks after OVX.

Discussion
The antidepressant-like action of EE 2 , in contrast to that of CIT, depends on the post-OVX interval. Remarkably, the combination of a non-effective dose of EE 2 plus CIT promotes an antidepressant-like action in the FST only one week post-OVX. In middle-aged female rats, the antidepressant-like effect of EE 2 was observed at one but not at three weeks post-OVX. These results are partially in line with previous data showing that ERT produces an antidepressant-like effect in FST if given closer to the estrogen decline [30,32]. However, the present findings are in contrast with a previous report where EE 2 exerted an antidepressant-like effect in the FST after twelve weeks of OVX [32]. An explanation for this divergence is the age of the females: young (3 months) versus middle-aged (15 months) rats. In support, some reports indicate that ovariectomized young (3 months) and adult (7 months) rats are more responsive to E 2 restitution than middle-aged animals (12 months) [30,31]. At the age of twelve months none of the female rats shows regular estrous cycles most have irregular cycles and a few show persistent estrus or diestrus [41][42][43]. After reproductive senescence, there could be an adjustment in estrogen receptor sensitivity [44][45][46] that affects the critical period for an intervention with estrogens. Importantly, most studies perform OVX before periestropause and evaluate the critical window for intervention considering the period of estrogen decline [12,33,45,47], but not the age at which OVX is performed. In this sense, recently García et al. [47] showed that changes in the expression of genes involved in social and affiliative behaviors-such as vasopressin and oxitocin-varied according to the age when OVX was performed as well as the interval after OVX.
CIT lacked an effect on the FST independently of the post-OVX interval. Several reports indicate that CIT produces an antidepressant-like effect in the FST [48,49]; however these studies were performed in 3 month old rats. It could be argued that the doses we tested were insufficient to promote an antidepressant-like effect. However, Flores Serrano et al. [48] showed that CIT reduced immobility on the FST at doses of 1 or 3 mg/kg, in 2-3 month old female rats independently of the estrous cycle. In the present study, doses in the range of 1.25 to 10 mg/kg had no action, making it unlikely that a dose-related problem explains the absence of an effect. Most likely, age and endocrine condition are factors that contribute to decrease the sensitivity to CIT. A similar result was reported by Olivares-Nazario et al. [35] who showed that senescent (23-25 months) male rats did not respond to the antidepressant-like effect of desimipramine or fluoxetine in the FST. In contrast, middle-aged females (14 to 18 month old), subjected to the chronic mild stress depression model, showed a transitory antidepressant-like effect of 10 mg/kg CIT [50]. Differences in the model and treatment schedule (chronic versus sub-acute) could contribute to the divergent results. However, the transitory effect of CIT in the chronic mild stress paradigm [50] suggests some degree of age-dependent insensitivity.
CIT is a highly selective serotonin transporter (SERT) inhibitor [49]. A previous study showed that SERT expression increased in response to stress by FST in young but not in middle-aged rats [51]. Indeed, studies in non-human primates have shown that aging is associated with decreased specific SERT binding [52]. Accordingly, it is possible that CIT lacked an antidepressant-like effect due to reduced SERT sensitivity caused by aging [53] and the endocrine condition. Supporting this notion, the present study showed that the combination of a non-effective dose of EE 2 plus a low dose of CIT produced an antidepressant-like action after one week of OVX, suggesting that EE 2 modulates SERT activity. In agreement, earlier reports indicated that the activity and affinity of SERT are modulated by estrogens, such as EE 2 [54][55][56], even in males. Interestingly, the antidepressant-like action of EE 2 requires presynaptic elements, possibly the SERT [57].
Alternatively, the combination of EE 2 plus CIT maybe is effective one week after OVX because EE 2 acts at estrogen receptors, which indirectly modulate the serotonergic system through 5-HT1A receptors [57][58][59]. The present data agree with a previous report showing that the combination of EE 2 plus a SSRI mainly affected swimming behavior [49], which is regulated by the serotonergic system [60].
Importantly, the modulation that EE 2 exerts on the effect of CIT has a critical period after which it is ineffective. The factors that underlie this response are unknown; however, it has been reported that estradiol requires the ERα-mGlutR1, PI3K/Akt and MAPK/ERK1/2 pathways to modulate the activity of SSRIs on the SERT [61]. Interestingly, ERα loses sensitivity with aging in the same manner than SERT [44,45,51,53]. Furthermore, ERα-IGF-R1 sensitivity decreased with aging if an intervention with estradiol did not occur close to OVX [62]. Consequently, it is feasible to consider that both age and time after OVX contribute to explain the lack of sensitivity of females after three weeks of OVX. Taken together it is possible that a complex mechanism of action, which requires further examination, underlies the antidepressant-like action of EE 2 alone and in combination with CIT. EE 2 increased the weight of uteri even though it lacked an effect on the FST. The present data suggest that the beneficial effect of estrogens on the brain is restricted to a specific period close to the natural decline of estrogen levels and this is independent of the peripheral effect of estrogens on different tissues. A shortcoming of the present data is the absence of histological analyses to evaluate the structural changes induced by EE 2 in the uterus. Furthermore, the differences in the weight of uteri between control groups could be explained by the stimulatory effect of the remaining hormones after OVX. Thus, one week post-OVX, the control uteri were heavier than those obtained after three weeks and in the former EE 2 did not modify their weight at any dose tested. In contrast, after three weeks, acute EE 2 administration increased their weight to control values, suggesting that EE 2 has a physiological action. Specific experiments to analyze a putative stimulatory effect of EE 2 on uterus weight are warranted.

Animals
Female Wistar rats of 15 months were group-housed (five to six per cage) in polycarbonate cages. All animals had free access to food and water. They were maintained on a 12:12-h light:dark cycle with lights on at 10:00 h and a room temperature of 23˘2˝C. All procedures observed the Mexican Official Norm for animal care and handling (NOM-062-ZOO-1999) and were approved by the Local Institutional Ethics Committee.

Ovariectomy
In the present study all rats were ovariectomized (OVX) one or three weeks before behavioral assessments in order to simulate menopause [63][64][65]. A ventral incision was made to remove the ovaries in rats anesthetized with tribromoethanol (2%; dose: 0.1 mL/kg. i.p.). Care was taken to totally excise the ovaries, which was corroborated by visual inspection. A recovery period of one or three weeks was allowed before rats were randomly assigned to the experimental groups [32].

Drugs
Citalopram clorhidrate (kindly donated by Psicofarma ® , México, México) was dissolved in physiological saline solution to prepared doses of 1.25, 2.5, 5.0 or 10 mg/kg that were administered in a sub-acute schedule (3 injections/´23,´5 and´1 h before FST) in a volume of 2 mL/kg. 17 α-ethynyl estradiol (Sigma-Aldrich, Toluca, México) was dissolved in corn oil and administered in a volume of 0.2 mL/rat 48 h before the FST at doses of 1.25, 2.5 or 5 µg/rat. The drugs were freshly prepared. The doses and latencies were chosen from previous studies .

Forced Swim Test
The FST was conducted by placing rats inside individual Plexiglasscylinders (height: 46 cm and diameter: 20 cm) filled with 30 cm of water at 23˘2˝C [66][67][68]. Two swim sessions were conducted: a pretest of 15-min followed 24 h later by a 5-min test, which was videotaped for scoring by an observer unaware of the treatments. After each swim session, rats were dried with a towel and placed in heated cages for 30 min before returning them to their home cages.
Three behavioral variables were scored during the test: (1) immobility, defined as the minimal movements done by the animal to keep its head above the water; (2) swimming, identified as gentle movements executed by the rat around the cylinder; and (3) climbing, characterized by vigorous movements of the forepaws directed against the wall of the cylinder [40,67,68].

Experiment 1: Effect of Post-OVX Interval on the Effect of EE2 or CIT in Middle-Aged Rats on the FST
A dose-response curve for each compound was performed to determine the effective dose of EE 2 or CIT to decrease immobility in middle-aged OVX rats. Independent groups were tested one or three weeks after the surgery; for CIT the groups were: saline, 1.25, 2.5, 5.0 and 10 mg/kg (n = 6-8 per group) and for EE 2 the groups included were: oil, 1.25, 2.5 and 5.0 µg/rat (n = 7-9 per group). CIT was administered´23,´5 and´1 h before the FST. EE 2 was dissolved in corn oil, prepared 72 h before its administration and injected acutely´48 h before the behavioral test. Doses and latencies were taken from previous data [29,36,65]. At the end of the experiment six rats of each group that received EE 2 were sacrificed and their uterus was removed, cleaned of fat and weighted in order to obtain an index of the peripheral effect of the steroid [32,37] In order to evaluate the effect of the post-OVX interval on the combination of EE 2 plus CIT, independent groups of OVX rats received: oil/saline, EE 2 (1.25 µg/rat,´48 h/saline), CIT (2.5 mg/kg; 23,´5,´1 h/oil) and EE 2 (1.25 µg/rat;´48 h) plus CIT (2.5 mg/kg;´23,´5,´1 h). All groups were tested one (n = 6-8 per group) or three weeks (n = 6-8 per group) after OVX. The dose of EE 2 was established in experiment one. Due to the fact that all CIT doses assayed were ineffective in experiment one, the dose chosen for this experiment was 2.5 µg/rat because other SSRIs are effective at this dose when combined with estrogens [21][22][23][24][25][26][27][28][29].

Statistical Analysis
The data are presented as mean˘S.E.M. The data were analyzed using two-way analysis of variance considering time after OVX and treatment as factors. Holm-Sidack tests were used as a post-hoc method of paired comparison. In all cases a p < 0.05 was considered significant.

Conclusions
In conclusion, the present data are in agreement with clinical reports showing that hormonal replacement therapy improves the effects of selective serotonin reuptake inhibitors in the post-menopausal condition [69,70].Importantly, the time of EE 2 restitution in relation to menopause should be considered in order to reach maximal beneficial effects and reduce the risk of adverse events.