Introduction

There is evidence that various neurodegenerative alterations are more frequent in hypoestrogenic women (Gibbs 1998; Henderson 2008). According to some authors, estrogen depletion due to aging, surgical or pathological events is partially responsible for the cognitive decline associated with neurodegenerative disorders such as Alzheimer’s disease (Craig and Murphy 2010; Zhao and Brinton 2006) and other forms of dementia (Vegeto et al. 2008). Estrogens, besides regulating endocrine processes, stimulate brain regions related to cognitive and learning integration, such as the neocortex, hippocampus and amygdala (Paris and Frye 2008). In fact, hormonal replacement therapy has been proven to be capable of attenuating some symptoms of Alzheimer’s (Gibbs and Aggarwal 1998; Ryan et al. 2008; Wu et al. 2008).

Since menopause is defined as the permanent cessation of menstruation resulting from the surgical or natural loss of ovarian follicular activity (Utian 2004), the removal of ovaries (OVX) in rats has been proposed, with restrictions, as a model for the study of some disorders associated with human postmenopause (Bosse and Di Paolo 1995). Rats under this hormonal condition have shown high levels of anxiety and depression (Estrada-Camarena et al. 2011; Picazo et al. 2006; Rodríguez-Landa et al. 2009), alterations in body temperature (Li and Satinoff, 1996), bone loss (Gurkan et al. 1986; Kalu 1991), a low threshold of pain (Kobayashi et al. 2000; Okada et al. 1997) and a deficit in memory function detectable with an autoshaping learning task (ALT) 18 weeks after surgery (Espinosa-Raya et al. 2011). Regarding the latter, replacement therapy with E2 or the synthetic steroid tibolone (Espinosa-Raya et al. 2011, 2012), initiated immediately after OVX, seems to overcome this cognitive impairment. Contrarily, when the hormonal treatment is initiated a relatively long time after the removal of ovaries or cessation of their activity, such protective actions are not observed (Daniel et al. 2006; Frick 2009; Gibbs 2010; Maki 2006). This is referred to as the critical period hypothesis (Resnick and Henderson 2002; Zhang et al. 2011) and is currently under discussion.

Another treatment that has reportedly produced improved cognitive abilities is an enriched environment (EE), a complex set of external factors in the immediate surroundings (Nithianantharajah and Hannan 2006). Different schedules of EE induce an improvement in learning and memory through the development of new nerve cells in the hippocampus and enhanced dendritic growth (Bruel-Jungerman et al. 2005; Diamond 2001; Kempermann et al. 1997; Leggio et al. 2005; Saito et al. 1994). Compared to standard conditions, EE facilitates a greater stimulation of sensorial, cognitive and motor circuits of the brain. For instance, non-spatial (Sun et al. 2010) and spatial memory impairment produced by cerebral ischemia or scopolamine can be restored when rats are exposed to a complex environment (Dahlqvist et al. 2004; Lima et al. 2014; Sun et al. 2010; Wadowska et al. 2014).

These findings have been confirmed by clinical studies showing that EE increases the dendrite branching complexity (Jacobs et al. 1993), reduces the risk of Alzheimer’s disease (Wilson et al. 2002) and protects against hippocampal lesion atrophy in the chronic stages of traumatic brain injury (Miller et al. 2013; White et al. 2013). In the same sense, it has been shown that for aging humans, aerobic exercise training increases brain volume (Colcombe et al. 2006), reduces brain tissue loss (Colcombe et al. 2003) and improves cognitive function (Colcombe and Kramer 2003).

Even though the majority of studies have reported positive results with EE, some studies have found contradictory data. This could be explained by differences in the schedule of EE exposure, aging and gender of the animals, housing conditions, and so on (Gresack et al. 2007; Lambert et al. 2005). For instance, compared to control animals housed under standard conditions, a few hours of EE exposure improves spatial working memory but not spatial reference memory of female mice (Redolat and Mesa-Gresa 2012), while a 4-week EE exposure of aged, but not young, OVX mice improves object recognition memory (Gresack et al. 2007). These discrepancies about the estrogens–EE interaction are difficult to interpret due to the methodological variables involved, but also because this issue has not been extensively explored so far. To our knowledge, this is the first report using a noninvasive method of hormonal administration for studying the role of E2 and EE on cognition.

Since a clear cognitive deficit in both tests used here has previously been shown following ovariectomy (Bastos et al. 2015; Fonseca et al. 2013; Espinosa-Raya et al. 2012), here only OVX females were used, sham controls being not necessary to demonstrate the impaired performance induced by estrogen depletion. Thus, taking into account the advantages represented by noninvasive procedures and considering the critical period idea, the current study explored whether or not an 18-week protocol of EE alone or in combination with E2 treatment improved this ovariectomy-induced impairment of memory.

Rats with E2 treatment initiated immediately after OVX were exposed to an EE following a modified protocol reported by Leal-Galicia et al. (2008) and then submitted to two memory tests: the ALT and the novel object recognition test (NORT). ALT has been proposed for measuring associative learning (Meneses 2003) and requires an intact neuronal system in the hippocampus, septum and cortex, as well as the participation of the cholinergic signaling transduction pathways in the integration of the acquisition–consolidation process (Steckler et al. 1993; Meneses 1999; Meneses and Terrón 2001; Espinosa-Raya et al. 2007, 2011). On the other hand, NORT has been used for the study of short- and long-term memory (Taglialatela et al. 2009) and requires participation of the perirhinal cortex for acquisition, consolidation and retrieval of object information. Accordingly, glutamatergic, serotoninergic and cholinergic neurotransmission is necessary for the integration of these processes (Dere et al. 2007). Finally, to discard that changes in motor activity could influence performance on these behavioral tasks, all rats were evaluated with an automatic activity counter immediately after the ALT test.

Materials and methods

Animals

The Animal Care and Use Committee of the Escuela Superior de Medicina at the Instituto Politécnico Nacional approved all experimental protocols. Forty female Sprague–Dawley rats (200–250 g) were purchased from Harlan Laboratories (Harlan Mexico, S.A. de C.V.). Before beginning the treatments, the animals were housed in groups (n = 10) in polycarbonate cages for 2 weeks under an inverted 12/12-h light–dark cycle (lights on at 9 p.m.) in a temperature-controlled (22 °C) room. All animals had free access to food (Purina Rat Chow) and water throughout the experiments. Animal care and handling were in accordance with internationally accepted procedures and approved by our Institutional Committee (NOM-062 ZOO, 1999). Special care was taken to minimize animal suffering. All experiments were carried out between 11:00 and 15:00, and independent groups were used.

Environmental enrichment

This procedure was carried out following a modified version of the conditions proposed by Leal-Galicia et al. (2007). It has been reported that exposure to this EE schedule elicits an increase in the hippocampal neurogenesis, modifies rat emotionality and improves the recognition memory in old rats (Leal-Galicia et al. 2007, 2008). Briefly, this apparatus consists of an exploration chamber (1.5 × 1.5 m; illuminated by red light) containing different elements such as plastic balls, tunnels, nesting materials, mesh wire ladders and running wheels. For EE exposure, rats in the EE and E2 + EE groups (kept under standard housing conditions) were placed in this chamber 2 h daily (for 2 weeks). To avoid behavioral habituation, for the remaining of the 18-week treatment period exposure was carried out only on Saturdays for 3 h. All moveable objects used for enrichment were placed in a different arrangement inside of the exploration chamber before each session.

Treatments

At the time of surgery, rats were randomly assigned to one of the following groups (n = 10 each): (a) a control group not treated with E2 nor exposed to enrichment (CONT); (b) a group chronically treated with E2 but not exposed to enrichment (E2); (c) a group exposed to enrichment but not treated with E2 (EE); and (d) a group treated with E2 and exposed to enrichment (E2 + EE).

Animals in the E2 and E2 + EE groups were subcutaneously implanted with a pellet that time-releases E2 for 60 days. This pellet, containing 0.05 mg of E2 (1 μg/rat/day) (Innovative Res. America, FL, USA), was placed at the dorsal portion of the neck during the ovariectomy. In order to complete 18 weeks of treatment, the pellet was replaced 9 weeks after implantation. It has been continuously confirmed that these pellets supply physiological hormonal levels beginning 2 weeks after their insertion (Singh et al. 2008; Strom et al. 2008a, b; Nordell et al. 2005; Aubele and Kritzer 2012). In order to eliminate variations due to daily handling, animals in the control group were manipulated in a similar manner as those in the experimental groups.

Surgery

Animals were bilaterally ovariectomized (OVX) under 2,2,2-tribromoethanol anesthesia (0.2 g/kg, i.p.), by a well-trained technician, through a dorsal incision that allows for the location and removal of the ovaries (Picazo et al. 2006) and whose complete extraction was corroborated by visual inspection. Immediately after surgery, rats were placed for approximately 2 h in a harm platform until their complete recovery. Rats were then gently transported and housed in plastic cages and randomly assigned to the different treatments.

ALT

The ALT is used to measure associative learning (Meneses 2003). Briefly, it consists of a standard Skinner box (Med Associates Inc., USA) with an illuminated lever in the middle of one wall (4 cm above the floor) and a food tray located 5 cm to the right of the lever. Animals submitted to this test are previously fasted for 24–36 h. On the training day, each rat is placed in the experimental chamber and given time to acclimate to it (~15 min). During this time, animals find and eat 20 food pellets (45 mg Mach) that are in the tray. Afterward, the rat can obtain another pellet each time the illuminated lever is pressed. An inter-trial interval of 5 s is used as the standard. Thus, an increase or decrease in the number of conditioned responses (CRs) is considered an enhancement or impairment in the consolidation of learning, respectively. The protocol consists of three sessions, one every 24 h, with the first and second sessions comprising 20 trials and the third session only 10. As it is usual practice for this paradigm, data are expressed as the mean number of CRs (i.e., times the lever is pressed) during the last session (Espinosa-Raya et al. 2011).

Spontaneous activity test

To assess the potential effects of the drugs on motor activity, all rats were tested in an automatic activity counter, which consists of an acrylic cage measuring 51.1 × 9.5 × 69.2 cm with two arrays of 15 infrared beams (each spaced 2.5 cm apart) that are perpendicular to each other. The interruption of a beam generates an electric impulse, which is processed and presented as a count (Opto-Varimex; Columbus Instruments, OH, USA). This test was carried out immediately after the third session in the ALT. Ambulation, vertical activity (climbing) and total activity are automatically registered during a 5-min session.

NORT

The NORT is used to study short- and long-term memory. This test, described by Ennaceur and Delacour (1988), is based on the natural tendency of rodents to explore new objects and compare them to ones with which they are already familiar. Thus, when a rat is exposed to a new object, it is expected that the animal will take longer exploring it than when exposed to a familiar one.

Briefly, the task consists of an open-field arena (30 × 30 × 15 cm) with two identical objects (F + F′) located in opposite and symmetrical corners. During the first phase (acquisition), the rat was placed in the arena for 5 min. To avoid forcing the animal to explore some specific object, it was always placed in the center of the arena. During this phase, animals that did not explore the objects for at least 30 s were not included in the current study. At the end of the session, the rat was returned to its housing cage.

Given that this task is useful for the study of memory at different time points (Taglialatela et al. 2009), animals were tested again 1 h (short-term memory) and 24 h (long-term memory) after the acquisition phase. Before the latter two sessions, one of the familiar objects (F) was replaced by a different one (N). All sessions were videotaped for later scoring by a single observer, who was blind to the treatment conditions.

In this test, data are frequently presented as the preference index (PI), which expresses the difference between the time spent with the familiar object and the novel one [time spent with novel object/(time spent with novel object + time spent with familiar object) × 100] (Wang et al. 2007). A PI above 50 % (i.e., the chance level) indicates a novel object preference and can be interpreted as normal cognitive performance (Navarrete et al. 2008; Walf et al. 2006).

Statistics

Data were analyzed with a one-way ANOVA or with a Kruskal–Wallis analysis (depending on whether populations were normally distributed or not) followed by the corresponding post hoc comparisons. Other statistical analyses were performed using paired t tests. A p value <0.05 was considered significant.

Results

ALT

Hormonal treatment and/or exposure to EE restored cognitive performance in the ALT of OVX rats, compared to the control group. Overall, there was a significant difference across the four groups tested (H = 15.885; n = 40; p = 0.001). Mann–Whitney U tests showed that groups exposed to E2 and EE treatments performed significantly better than the control group (E2 vs C: U = 7.50, n1 = 10, n2 = 10, p = 0.001; EE vs C: U = 24.0, n1 = 10, n2 = 10, p = 0.047; EE + E2 vs C: U = 11.0, n1 = 10, n2 = 10, p = 0.003; Fig. 1). The combination of both treatments did not improve the number of CRs obtained under EE treatment (EE + E2 vs EE: U = 45.0, n1 = 10, n2 = 10, p = 0.733). Contrarily, such combination decreased the number of CRs produced only by E2 administration (EE + E2 vs E2: U = 18.0, n1 = 10, n2 = 10, p = 0.017).

Fig. 1
figure 1

Number of conditioned responses, in an autoshaping learning task, for rats chronically treated with estradiol (E2), exposed to an enriched environment (EE) for 18 weeks or given a combination of both treatments (EE + E2). Each column represents the mean ± SE of ten rats. The asterisk denotes a difference versus the control group (C) or between those groups indicated by the bracket (Mann–Whitney U test; p < 0.05)

Full size image

Spontaneous activity test

As aforementioned, all rats were evaluated with an automatic activity counter immediately after the ALT. It is clear from this evaluation that no treatment altered the motor ability of the animals (Table 1).

Table 1 Spontaneous activity of ovariectomized rats under chronic estrogen treatment (E2) and exposed to an enriched environment (EE)
Full size table

NORT

All rats in the present study spent a similar amount of time on object exploration during the acquisition phase in the NORT, where F and F′ represent identical objects (Fig. 2, upper panel), as reflected by a lack of significant difference in the PI across groups (lower panel; one-way ANOVA; F = 0.249; p = 0.861).

Fig. 2
figure 2

Behavior of rats chronically treated with estradiol (E2), exposed to an enriched environment (EE) for 18 weeks or given a combination of both treatments (EE + E2) during the acquisition phase in the novel object recognition test. The upper panel shows the amount of exploration time spent by rats during a 5-min period, while the corresponding preference index is depicted below. In this figure, F and F′ denote identical (familiar) objects

Full size image

Although the test carried out 1 h after the acquisition phase showed that rats belonging to each group spent more time exploring the novel object (N) compared to the familiar one (F) (Fig. 3, upper panel; paired t tests; all ps < 0.05), the PI did not significantly differ across groups (Fig. 3, lower panel; one-way ANOVA, F = 1.115, p = 0.356).

Fig. 3
figure 3

Exploration time and preference index for rats chronically treated with estradiol (E2), exposed to an enriched environment (EE) for 18 weeks or given a combination of both treatments (EE + E2) 1 h after the acquisition phase in the novel object recognition test. The familiar object is represented by F and the novel one by N. The asterisk denotes the significant difference between the time spent by the rats exploring the novel and the familiar object (paired t test; p < 0.05)

Full size image

In the test carried out 24 h after the acquisition phase, rats belonging to all the experimental groups, but not to the control group, spent more time exploring the novel object than the familiar object (Fig. 4, upper panel; paired t tests, all ps < 0.05). Moreover, the PI significantly differed across groups (Fig. 4, lower panel; Kruskal–Wallis; H = 17.039; n = 40; p = 0.001); all experimental groups significantly differed compared to the control group (Tukey’s test, all ps < 0.01), but did not significantly differ between each other.

Fig. 4
figure 4

Exploration time and preference index for rats chronically treated with estradiol (E2), exposed to an enriched environment (EE) for 18 weeks or given a combination of both treatments (EE + E2) 24 h after the acquisition phase in the novel object recognition test. The familiar object is represented by F and the novel one by N. Each column represents the mean ± SE of ten rats. The asterisk denotes a significant difference between the time spent by rats exploring the novel and the familiar object (upper panel, paired t test; p < 0.05) and between the control group and each experimental group (lower panel, Tukey’s test; p < 0.05). Paired comparisons among experimental groups did not show a significant difference (Mann–Whitney U test; p > 0.05)

Full size image

Discussion

Data from the current study can be summarized as follows: (a) The impairment of memory, putatively caused by removal of the ovaries, could be overcome by with estrogens or EE; (b) hormonal treatment was better than the combination EE + E2 for improving the performance of rats in the ALT; (c) the combination of these two treatments showed different results depending on the test used for measuring cognition.

As aforementioned, an estrogen deficiency has profound effects on cognition. Accordingly, deterioration of the natural ovarian function or removal of the ovaries has been associated with cognitive deficits (Markowska and Savonenko 2002; Walf et al. 2006). For instance, several studies have reported that during induced or natural proestrus/estrus, rodents perform better on learning tasks than during diestrus or than counterparts without ovaries (Paris and Frye 2008; Van Goethem et al. 2012). The results of the current contribution are in accordance with this evidence. Adult OVX rats evaluated 18 weeks post-surgery showed virtually no response in the ALT, whereas sham-operated females displayed approximately 6–10 CRs/min (Espinosa-Raya et al. 2012). This cognitive deficit due to OVX can be restored by chronic estrogenic treatment initiated post-surgery (Gresack and Frick 2004; Markowska and Savonenko 2002) as long as the beginning of treatment is within 10 weeks post-OVX (McLaughlin et al. 2008). This phenomenon has been observed in several tests including ALT and NORT. In the present study, the NORT showed that E2 treatment reinstated the object recognition memory of castrated females, reaching levels of performance similar to those reported for sexually receptive rats (Fernandez and Frick 2004; Luine et al. 2003; Walf et al. 2006). This evidence is in line with the critical period hypothesis, which considers that estrogen treatment confers optimal performance benefits for women when initiated close in time to the onset of menopause (Resnick and Henderson 2002).

Interestingly, the ALT showed that OVX rats with E2 administration plus EE exposure experienced less cognitive benefits than animals given E2 alone. Contrarily, the NORT demonstrated that OVX rats in these two groups—the combined treatment and the E2 treatment alone—performed in a similar fashion. In this regard, Gresack and Frick (2004) reported that with the NORT, OVX mice administered with E2, but without EE exposure, had a significantly enhanced memory, while the simultaneous application of both treatments led to reduced cognitive ability, an effect only observed in young but not aged mice (Gresack et al. 2007).

There were obvious methodological differences between the present study and that by Gresack and Frick (2004), including the use of mice instead of rats, acute daily injections of high doses of E2 and a delay in post-OVX hormonal replacement. In such study, the reduced cognitive ability found with the NORT in regard to the combined EE + E2 versus E2 treatment seems to indicate a role of stress factors. Regarding this, we took into account the critical period hypothesis and began the E2 release immediately after the OVX through hormone-releasing pellets instead of injecting rats for 15 days, diminishing in this way the role of stress produced by injections that could have masked the observations of Gresack and Frick (2004).

Some of the main factors that alter the effect of enrichment are the exposure schedule, the size of the enrichment boxes and the number of animals in the housing cages (Diamond 2001; Simpson and Kelly 2011), showing that this manipulation does not always produce the same consequences. For instance, although the majority of studies report a beneficial effect of EE (Costa et al. 2007; Freret et al. 2012; Frick and Fernandez 2003; Hu et al. 2010; Jankowsky et al. 2005; Laviola et al. 2008), some housing conditions can produce stress and as a consequence result in high levels of corticosterone (CORT) (Girbovan and Plamondon 2013). For instance, in intact female rats or mice exposed to EE for 6 weeks, higher levels of plasmatic CORT are found in animals housed four per cage than in those housed individually (Arndt et al. 2009; Konkle et al. 2010; Martin and Brown 2010). It is currently known that this hormone improves performance when found at optimal levels (Gresack and Frick 2004; Sampedro-Piquero et al. 2014), essentially during the consolidation memory phase (Roozendaal 2000), but produces amnesia at high concentrations (Meaney et al. 1988; Schwabe et al. 2012).

Surprisingly, E2 also increases the levels of CORT (Chan et al. 2014; Farid et al. 2013), suggesting that with either E2 administration or EE exposure alone, the mnemonic benefits could be due to the level of CORT produced by each treatment. However, with combined treatment there would seem to be an additive effect of the CORT secretion induced by each individual treatment, which if true would explain the reduction in the CRs produced by this combination. Thus, it is possible that the combined hormone and enrichment treatment result in high levels of CORT and therefore a mnemonic deficit (Schwabe et al. 2012).

In contrast to data derived from the ALT, results from the NORT evidenced the effectiveness of E2, EE and E2 + EE, each producing very similar effects on object recognitive function in the long but not short run. The reason for this finding could lie in the fact that the ALT and NORT stimulate different brain regions: the first excites the amygdala–NAcc–hippocampus pathway (Correa 2007), while the NORT seems to affect hippocampal cortices (Antunes and Biala 2012), specifically the perirhinal cortex (Winters and Bussey 2005).

It is also known that E2 administration and EE exposure, besides producing changes in CORT, increase the expression of the brain-derived neurotrophic factor (BDNF). The role of BDNF in cognition has been well demonstrated (Ozawa et al. 2014), and its increase occurs, under different enrichment protocols, in central amygdala, hippocampus, and/or visual and entorhinal cortices (Farmer et al. 2004; Ickes et al. 2000; Pham et al. 1999; Ramírez-Rodríguez et al. 2014; Ravenelle et al. 2014; Rossi et al. 2006). The increase in this neurotrophin is prevented by an OVX (Berchtold et al. 2001) and can be restored by E2 treatment (Berchtold et al. 2001; Gibbs 1999). When E2 is combined with exercise, as part of EE, the level of BDNF is above the value reached by hormone replacement alone (Berchtold et al. 2001). This evidence suggests that the memory consolidation observed at 24 h, but not at 1 h, after the acquisition phase in the NORT could be due to the BDNF expression induced by either E2 or EE treatment. Finally, although both types of memory, short- and long-term memory, are integrated in hippocampus (Hammond et al. 2004), it is important to address that short-term memory does not require protein synthesis (Balderas et al. 2014; Rossato et al. 2007), an observation that could help to understand the current data.

To sum up, a general hypothesis to explain the present results should include the role of CORT and BDNF as modulators of the learning and memory displayed in the two tests herein used. It should also take into account that recognition memory is integrated in extrahipocampal cortices (Antunes and Biala 2012; Balderas et al. 2008; Spada et al. 2006; Winters and Bussey 2005), where the corticoid receptors are scarce in comparison with the hippocampus (Sánchez et al. 2000) and that autoshaping learning depends on the amygdala–NAcc–hippocampus pathway (Correa 2007), where there is a high density of CORT receptors (Reul and De Kloet 1986; Roozendaal and McGaugh 1997) modulating the actions of CORT on the associative learning (Beylin and Shors 2003; Roozendaal 2000).

Conclusions

The current results show that chronic treatment with estrogen or EE exposure has a beneficial effect on two types of memory. Regarding the effect on associative learning, herein measured with the ALT, E2 treatment was found to be better than certain schedules of EE. Finally, the combination of both treatments did not produce perceptible advantages.