Estrogen Receptor α Functions in the Regulation of Motivation and Spatial Cognition in Young Male Rats

Estrogenic functions in regulating behavioral states such as motivation, mood, anxiety, and cognition are relatively well documented in female humans and animals. In males, however, although the entire enzymatic machinery for producing estradiol and the corresponding receptors are present, estrogenic functions have been largely neglected. Therefore, and as a follow-up study to previous research, we sub-chronically applied a specific estrogen receptor α (ERα) antagonist in young male rats before and during a spatial learning task (holeboard). The male rats showed a dose-dependent increase in motivational, but not cognitive, behavior. The expression of hippocampal steroid receptor genes, such as glucocorticoid (GR), mineralocorticoid (MR), androgen (AR), and the estrogen receptor ERα but not ERβ was dose-dependently reduced. The expression of the aromatase but not the brain-derived neurotrophic factor (BDNF) encoding gene was also suppressed. Reduced gene expression and increased behavioral performance converged at an antagonist concentration of 7.4 µmol. The hippocampal and blood serum hormone levels (corticosterone, testosterone, and 17β-estradiol) did not differ between the experimental groups and controls. We conclude that steroid receptors (and BDNF) act in a concerted, network-like manner to affect behavior and mutual gene expression. Therefore, the isolated view on single receptor types is probably insufficient to explain steroid effects on behavior. The steroid network may keep motivation in homeostasis by supporting and constraining the behavioral expression of motivation.


Introduction
In recent years, estrogen receptors have increasingly been identified as involved in modulating motivation and cognition in female human development, postmenopausal mood disorders, and corresponding animal models [1], [2], [3]. The effects in male subjects, however, have been largely neglected, although the entire enzymatic machinery for locally producing estrogens as well as both estrogen receptors (ERa and ERb) are present in male brains. Moreover, there is evidence that cognitive deficits can be rescued by estrogens [4]. Most studies focused on sexual and aggressive behavior [5], [6], [7]. The large body of evidence of estrogenic effects on neuronal plasticity, such as long-term potentiation, spine plasticity, and neurogenesis [8], [9], [10], [11], is contrasted by only a few studies on the effects of more general states such as motivation and mood and their outcome in behavioral performance. Aggression and modulations of the stress axis activity have been reported to be affected by estrogenic mechanisms in male mice and rats [12], [13], [14]. Aggression, therefore, may be stimulated by ERa and suppressed by ERb activation in male rats [15]. Estrogenic effects in spatial learning have also been reported in both sexes [16], [17], [18], [19], [20], [21].
Hippocampal synthesis of estradiol in male rats is realized by the enzyme aromatase that converts testosterone into estradiol. Aromatase as well as synaptic and nuclear ERa have been identified in all subregions of the hippocampus and in the dentate gyrus. Estradiol can induce rapid upregulation of spine number and fast modulation of hippocampal synaptic plasticity [22]. Accordingly, rapid alternation and control of various behaviors in males, including learning, are controlled by brain-derived estrogens [23].
In a previous study [24], we found a positive correlation between hippocampal ERa gene expression and the behavioral performance of young post-pubertal male rats in a spatial holeboard paradigm. The motivational (i.e.,task readiness) component, which was extracted with principal component analysis from numerous behavioral elements, in particular was strongly correlated with the expression of ERa and weakly correlated with the testosterone binding AR, whereas ERb, MR, and GR receptor gene expression was uncorrelated with components representing motivation, spatial cognition, and emotion. Therefore, in the present study, we applied the ERa-specific antagonist methylpiperidino-pyrazole (MPP) in four dosages to reveal more specific functions of ERa activity in motivation and spatial cognition during our holeboard paradigm in male rats. Steroid receptors are assumed to act in concert and mutual interactions via heterodimerization [25] and other protein-protein interactions [26], [27]. Ligand-activated cytosolic ERa translocates into the nucleus where, similar to all steroid receptors, this receptor acts as a transcription factor, activating or repressing the expression of target genes including those of other steroid receptors. Thus, we measured the expression of hippocampal corticosterone binding receptors, the MR and GR genes, as well as the AR and both estradiol binding estrogen receptors. In addition to these slow genomic functions, membrane-bound steroid receptors can mediate fast non-genomic functions by activating different intracellular pathways [28]. Membrane receptors that are at least closely related to ERa and ERb due to their ability to be activated by ERa-or ERb-specific agonists have been identified [29]. Membrane-bound estrogen receptors are also involved in hippocampal-dependent object memory consolidation [30].
Therefore, while the MPP binds to cytosolic ER, we cannot rule out that membrane-bound receptors are blocked as well. The blood serum and hippocampal stress and sex hormone concentrations were also measured. In addition, aromatase and BDNF gene expression was measured. The latter was considered because rat hippocampal BDNF has been reported to interact with estrogens such that the administration of estradiol enhances hippocampal BDNF-mRNA [31], [32], by activation of extranuclear ER [33], whereas corticosterone [34] reduces BDNF-mRNA, and thus have opposite effects on BDNF expression and possibly on BDNF effects on learning and memory.
Based on our previous data, we hypothesized reducing the motivation of males treated with an ERa antagonist, whereas the spatial cognitive aspect should be unaffected. The latter hypothesis was confirmed, whereas MPP treatment resulted in increased motivation. Possible reasons for these results are discussed, and advanced hypotheses are outlined.

Ethics Statement
All experiments were performed in accordance with the European Communities Council Directive of 24th Nov. 1986 (86/609/EEC) and the German guidelines for the care and use of animals in laboratory research. The experimental protocols were approved by the ethics committee of Saxony-Anhalt. All efforts were made to reduce the number of rats used in this study and their suffering.

Animal Keeping
Wistar rats from the institute's breeding colony were weaned at post-natal day 21 and then housed in groups of 5 males in standard cages (59 cm638 cm625 cm). They were maintained under a 12 h: 12 h light regimen with lights on at 6:00 a.m.; the ground was covered with commercial bedding material (wood spans, ssniff, Soest, Germany) and food pellets (ssniff,/M-H), and tap water was given ad libitum. After the implantation of a cannula when the rats were 7 weeks old, they were transferred to individual cages (40 cm625 cm618 cm).

Behavioral Tests
Holeboard. The holeboard apparatus consisted of a black board (1 m61 m) with 16 regularly arranged holes, 7.5 cm in diameter and 7 cm deep. The board was surrounded with Plexiglas walls 50 cm high. The walls were covered with white paper on the outside and marked with a different black cue at each side. Four out of 16 holes were baited in a fixed pattern with standard food pellets (dustless precision pellets, 45 mg, BioServ).
Path trajectories were recorded with the tracking system BiObserve Viewer software (Version 3.0.0.92), and behavioral parameters were measured. The head-dips of the animals were registered by photobeams at the middle of each hole. The photobeam signals were detected and counted by the holeboard-plugIn of the BiObserve Viewer software.
Thus, the time to find all pellets (the latency for rats that did not find all four pellets was scored as 120 s), the average velocity (given as the average speed in cm/s during a given trial), the number of pellets found, the hole dips/s, the total hole dips, the mean distance to the wall (the distance from the animal's body to the nearest wall, given as the mean over all time points of a trial), the working memory errors (a rat revisits a hole that it already took bait from during a specific trial), and the reference memory errors (visiting an unbaited hole) were recorded. We calculated an index of performance for both types of memory. This calculation helps to divide the animals that made no errors, because they did not move at all on the holeboard from those that made no errors because they performed the task correctly. The index was calculated as total visits/(error+total visits). Thus, an index of 0 indicates no hole visit at all (0 errors because of 0 attempts were set to 0), 0.5 the number of hole visits equals the number of errors, and 1 indicates no error (four dips coincides with four pellets found).
Animals stayed in their home cages in the testing room during the entire experiment. Animals were transferred from the cage to the test arena for each training trial. A trial was automatically stopped after 2 min or when the animal had found all 4 pellets. All experimental animals were familiarized with the test set-up 1 day before training. Then they received spatial training on a fixed pattern of baited holes over 10 trials (day 1: five trials, day 2: four trials, and the retention trial on day 3), with a 15 min inter-trial interval [35]. Training started at 9:00 am, and the retention trial (day 3) started 24 h after the last trial of day 2 (10:00 am). The board was cleaned after each trial with 20% ethanol.
Open-field test and elevated plus maze test. Animals were tested for 5 min in each test twice: open field at 9:00 a.m. followed by the elevated plus test 1 h later. This procedure was repeated 24 h later. Animals stayed in the test room overnight. For these tests, we used non-food-deprived animals. The arenas were cleaned after each trial with 20% ethanol.
Open-field test. The holeboard served also as an open-field arena, with the exception that the floor with the holes was covered with a black plastic plate. The arena was, via the tracking system, divided into different zones: four corner zones (25 cm625 cm), four wall zones (25 cm650 cm), and one center zone (50 cm650 cm). The percentage of time spent in each zone, the mean velocity, and the track length were measured.
Elevated plus maze. The elevated plus maze consisted of two closed and two open arms of black plastic at a height of 80 cm above floor. The arms were 10 cm wide and 50 cm long. The closed arms were equipped with black walls (30 cm high). From a center arena (10 cm610 cm), the animals started to explore the maze. The percentage of time spent in each arm, the mean velocity, and the track length were measured via the BiObserve Viewer software.

Pharmacology
The ERa antagonist MPP binds to extranuclear receptors with very high specificity [36] and has been proved to antagonize estradiol-induced gene transactivation and -repression with no effect on these processes mediated through ERb [37]. MPP, dissolved in water, was administered in four concentrations, 1.5 mmol (n = 18), 3.7 mmol (n = 13), 7.4 mmol (n = 18), and 12.6 mmol (n = 12), and applied over 7 consecutive days (starting 5 days before the experiments and ending at training day 2) via a Hamilton syringe (5 ml volume over 5 min). A flexible tube allowed the animals to move freely during administration. Two food-deprived, vehicle-treated groups, one trained (n = 13) and one untrained (n = 7), served as time-matched controls. The untrained control group remained in the testing room throughout the experiments. All animals in the various groups were killed euthanized at the same time point (15 min after the retention trial, i.e., between 10:15 and 10:30 a.m.).
Similar as in previous studies, we investigated the right hippocampus (genes, hormones) as well as the hormone levels in blood serum.
The rats used for the open-field and elevated plus maze tests were treated with 7.4 mmol MPP i.c.v. (the most effective dose in the holeboard task) or vehicle. In accordance with the holeboard procedure, daily injections started 5 days before the first test day and continued during the two test days.

Tissue Sampling and Hormone Assaying
Animals were decapitated 15 min after the last trial, and trunk blood was collected in vials containing clot activator (REF 41.1500.005, Sarstedt; Germany). Tissue from the right hippocampus was rapidly dissected and frozen (220uC) until measurements. Samples were homogenized (Biovortexer No. 1083; BioSpec products), diluted (Sample diluent; IBL Hamburg; REF KLZZ731) to reach a final volume of 25 ml/mg tissue weight, and centrifuged (10 min, 10000 rpm). The supernatant was stored at 220uC. For the hormone assays, we used the enzyme-linked immunosorbent assay (ELISA). Briefly, samples were thawed and diluted (brain samples 1:3, 1:2, and 1:1, serum samples 1:10, 1:2, and 1:5 for the testosterone, 17b-estradiol, and corticosterone assays, respectively). Samples and standards were applied in duplicate. OD values were measured at 450 nm in a micro-plate reader (Thermo Scientific MultiSkan FC ELISA Reader) and calculated via a standard four-parameter logistics plot. For the testosterone assay (Testosterone Saliva ELISA by IBL Hamburg; Germany), the limit of detection (LOD) was 2.0 pg/ml, and the intra-assay and inter-assay coefficients of variation were 8.2% and 5.5%, respectively. The estradiol assay (17beta-Estradiol Saliva ELISA by IBL Hamburg) had an LOD of 0.4 pg/ml. The intraassay and inter-assay coefficients of variation were a maximum of 9.9% and 11.1%, respectively. For the corticosterone kit (Corticosterone ELISA by IBL Hamburg), the LOD was 1.631 nmol/L, and the intra-assay and inter-assay coefficients of variations were 2.77% and 6.14%, respectively. Randomly chosen subsets of the animals used in the behavioral experiments were analyzed.

Quantitative Real-time RT-polymerase Chain Reaction
Tissue from the right hippocampus, added with an mRNA stabilizing agent (RNA later, Qiagen, Hilden; Germany), was stored at 280uC. For the analysis, mRNA was isolated (RNeasy Plus MiniKit, Qiagen) and transcribed to cDNA (high-capacity cDNA reverse Transcription Kit from Applied Biosystems (Carlsbad, CA, USA, now Life Technologies).

Statistical Analyses
All statistical analyses were conducted with SPSS (V. 20). The distribution of all data was tested with the Shapiro-Wilk-test. Differences in gene expression data and hormone concentrations were analyzed with the univariate general linear model (GLM) with the treatment as a factor followed (if significant) by post-hoc tests (Tukey). Data that were not normally distributed (reference and working memory errors) were analyzed with the Kruskal-Wallis test for k-group comparisons followed (if significant) by the Mann-Whitney-U-test for pairwise comparisons. Behavioral data over the trials were tested with a linear mixed model with trials, learning phases (days), treatment, and phase6treatment interaction as fixed factors and trials and learning phases as repeated measure variables. The linear mixed model, in contrast to the general linear model, compares the phases with different numbers of trials and handles missing values. Behavior in the open-field and elevated plus maze tests was analyzed with the GLM for repeated measures. All tests were two-tailed, and the level of significance was set at p#0.05.

Behavior
We analyzed six behavioral parameters: the latency to find all pellets, the average velocity, the number of pellets found, the average dips per second, the total number of hole dips and the mean distance of the rats to the wall (Figure 1). The outcome of the statistical analysis is summarized in Table 1. We found significant overall differences (test of fixed effects) for all parameters regarding the treatment, trial effects could be determined only in the time to find all pellets, the number of found pellets and the number of hole dips/s. A significant interaction of dose and learning phase appeared only for the time to find all pellets. The estimates of the fixed effects that animals treated with 3.7 mmol MPP showed significantly decreased slopes at acquisition phases 1 and 2 but not during retention, whereas the decrease in the group treated with 7.4 mmol MPP was significant in all phases. Pairwise comparisons of estimated marginal means revealed significantly faster times to find all pellets, higher velocity, more pellets found, more total hole dips and hole dips/s and a larger wall distance for the group treated with 7.4 mmol MPP (Video S1), faster times to find all pellets and higher velocity for the animals treated with 3.7 mmol MPP, and larger wall distances of rats treated with 12.6 mmol MPP compared to vehicle-treated rats (Video S2).
In addition to these behavioral parameters, we analyzed the working and reference memories. For investigating and comparing different memory states, we calculated the error indices for the end of the acquisition (trial 9) and for retention trial 10 ( Figure 2). We found a significant overall difference in the reference memory (chi 2 = 9.75, df = 4, p = 0.045) but not working memory (chi 2 = 1.19, df = 4, p = 0.880) indices in trial 9. However, single group comparisons revealed no significant differences in reference
The post hoc tests revealed higher hippocampal corticosterone concentrations in the animals treated with 3.7 mmol MPP compared to the rats treated with 1.5 mmol (p = 0.044) and 12.6 mmol (p = 0.026) MPP.
The single group comparisons revealed that animals treated with 7.4 mmol MPP showed the highest suppression of relative gene expression of ERa compared with animals treated with 1.5 mmol (p = 0.008), 3.7 mmol (p = 0.017), and vehicle (p = 0.004) and untrained vehicle-treated control rats (p = 0.002). Furthermore, animals administered 12.6 mmol MPP had a significantly lower ERa-mRNA level than the vehicle-treated trained (p = 0.036) and untrained (p = 0.014) controls. The relative gene expression of ERb, however, differed significantly only in the 7.4 mmol MPP-treated rats compared to the untrained vehicletreated rats (p = 0.043).

Dose Effects
To compare dose effects on behavior and gene expression, we calculated a dose-effect curve ( Figure 5) at the individual level. We followed the trend indicated by the between-group statistical effects and calculated the curve as the percentage of treated individuals that showed measurements higher (behavior) or lower (gene expression) than the mean of the control animals. The  percentage measures were plotted against the logarithm of the dosages. It became immediately apparent that all effects (except total hole dips) were highest at the 7.4 mmol dosage. Thus, only when almost all individuals showed suppression in the expression of all genes was a behavioral effect observed. In addition, ERa, ERb, and AR gene expression was the most sensitive to the treatment. Effects of more than 50% were observed at the lowest concentration and reached a plateau at 3.7 mmol MPP. Ninety to hundred percent of the individuals responded at 3.7 mmol MPP in MR and GR expression and maintained the effects at 7.4 mmol, whereas the aromatase and BDNF gene expression was less sensitive but also reached 100% responders at 7.4 mmol MPP. Interestingly, with the decay in responders at 12.6 mmol in behavior, the expression of MR, GR, and BDNF increased, whereas ERa, ERb, and AR expression remained at the 100% level.

Open-field and Elevated Plus Maze Tests
To test whether the effects observed in the holeboard merely resulted from unspecific effects on locomotor activity or differences in anxiety induced by the pharmacological treatment, separate groups of animals underwent open-field and elevated plus maze tests ( Figure 6). These animals were treated with 7.4 mmol MPP (the most effective dose in the holeboard experiments, n = 12, one animal with missing values at day 1) or vehicle (n = 12; due to technical reasons, we lost the data for one animal in the elevated plus maze test at day 2). We found no statistically significant interaction between treatment and trial and no treatment effects in the open field (time spent in the center F 1,22 = 0.06, p = 0.815;

Discussion
We found dose-dependent changes (up to 7.4 mmol MPP) in behavioral performance with most significant effects in animals treated with 7.4 mmol MPP. These animals needed less time to find all pellets, moved faster, had more hole dips, and found more pellets than the control animals. In contrast to our hypothesis, the MPP-treated animals performed better regarding motivation (velocity, hole dips/s), whereas, in line with our hypothesis, the cognitive component of the task was not affected by the treatment. Treatment effects on the general locomotor activity and changes in anxiety, which could have contributed to the observed differences, can be ruled out by our open-field and plus maze tests. Regarding exploration and anxiety, the MPP-treated rats did not behave differently from the controls.
The effects of the MPP treatment are particularly surprising, because we found previously that the upregulation of hippocampal ERa gene expression is a learning-related holeboard effect in males at the age tested here. We also found, although not significant, upregulation of ERa gene expression in vehicle-treated males compared to untrained animals. However, the MPP-treated animals showed significantly reduced levels of gene expression not only for ERa but also for MR, GR, AR as well as the BDNF gene, whereas ERb and aromatase gene expression was unaffected in the trained animals. These results support two conclusions: i. ''Specific'' steroid receptor functions can be revealed only against the steroid network background, characterized by a stronger relation of ERa activity to its own gene expression and to that of glucocorticoid and testosterone binding androgen receptors than to ERb, and an effect on the BDNF but not on the aromatase encoding gene. ii. The training-induced increased expression of the ERa encoding gene does not positively correlate with the receptor function, because the receptor blockade results in increased rather than impaired performance of motivationindicating behavior. Activation of extranuclear ER induces hippocampal BDNF signaling, thus mediating its effects on neuroprotection and plasticity [33], [39]. Because of these known interactions, we included the BDNF gene in the present study, and interestingly, we found the MPP treatment affected BDNF gene expression. Hippocampal BDNF and the related receptors are involved in learning and memory [40], [41]; thus, an ER-BDNF interaction may partly play a role in mediating behavior in the present study.
Thus, at least parts of the observed system of steroid receptors, and there are many others [42], react in a coordinated and dosedependent manner to the MPP treatment, which suggests that compensating mechanisms exist within the network to keep the system in homeostasis and to provide a mechanism for reacting quickly to external signals. The apparent discrepancy between the previously observed positive correlation of ERa-mRNA with motivational behavior and the failure to reduce motivation by ERa blockade on the protein level in the present study, may be explained by the translation of preexisting, training-induced, silent ERa-mRNA that could be quickly and locally translated into functional receptors [43], [44] by extracellular signaling. MR, that has been identified as the crucial acute stress-related receptor in our holeboard paradigm [45] may be involved in the transduction of external signals. These mechanisms could provide effects upon behavior independently from rapid changes in the availability of locally produced estrogens [23]. Future studies should prove this with in situ-hybridization and optical methods for identifying the hippocampal steroid receptor-mRNA distribution and trafficking in trained and untrained animals. Membrane-and cytosol-specific westernblots can detect training-induced changes in site-specific functional steroid receptors within the hippocampus. The behavioral effect that motivation increased with the ERa-blockade is likely a result of the MPP treatment that reduces ligand binding with functional receptors, which occupation with a natural ligand otherwise leads to a decrease in motivation. Thus, ERa activation does not promote motivational behavior but may constrain overmotivated, high-risk behavior preserving the animal from lifethreatening situations. This may be adaptive especially in young post-pubertal rats that ontogenetically are in the phase of migrating from their social groups and exploring new environments. Importantly, we did not find upregulation of ERa gene expression in older individuals in the same situation [24].
The network characterization is further supported by the fact that we found significant effects in behavior only when the expression of all genes (except the aromatase and ERb encoding genes) was affected, although there were differences in the sensitivity to MPP treatment between genes, as suggested by the dose response curve. Second, changes in hormone concentrations, due to the treatment, were absent. Therefore, compensatory modulations of receptor expression resulting from altered hormonal states cannot explain the results. Last, the effects of MPP on gene expression are dose dependent (up to 7.4 mmol MPP), thus indicating not unspecific treatment-induced downregulation of gene expression. Further, not all genes were affected.
We noted dose-dependent effects of the MPP treatment on all behavioral elements and only an interaction of dose and learning phase effects for the times to find all pellets. Thus, the differences remained stable over the training for most of the behavioral parameters. The partial but not completely similar behavioral results in animals treated with 3.7 and 7.4 mmol MPP (especially for the learning phase6dose interactions) along with the differences in the effects on gene expression between the two dosages suggest complex effects of structural changes in the receptor network on motivation and behavior and may indicate a structural change in the receptor network over training. Training phasespecific changes in correlations of hippocampal steroid hormone concentrations with behavior has been described for the same learning procedure [46]. In that study, hippocampal testosterone concentrations switched from a negative correlation with reference memory errors at acquisition phase 1 to no correlation during acquisition 2 and back to a negative correlation during retention. Corticosterone concentrations in the prefrontal cortex negatively correlated with reference memory errors during acquisition and positively during retention. This may be related to brain regionspecific changes in steroid functions in behavioral regulation during training or reflect network adjustments over different brain regions. Surprisingly, the dose-dependent effects of MPP on behavioral performance disappeared when the animals were treated with a higher dose than 7.4 mmol MPP. Animals treated with 12.6 mmol MPP showed only a larger wall distance when compared to the vehicle-treated group. Notably, this group also showed similar MR and GR expression as the vehicle-treated animals and significantly higher expression of these receptor genes compared to the group treated with 7.4 mmol MPP. Thus, higher concentrations of MPP may result in unspecificity of the receptor blockade with a subsequent rescue of MR and GR expression and behavior.
Due to the intra-cerebroventricular administration of MPP, resulting in brain-wide distribution, other brain regions such as the amygdala and hypothalamus, which contain estrogen receptors [47], [48], [49], are probably involved in the observed behavioral effects. Although no effects of ERa knockdown in the medial amygdala on sexual or aggressive behavior in 16-week-old male mice were observed, the same procedure in the ventromedial nucleus of the hypothalamus reduced both types of behavior [7]. The effects, however, may be age dependent [50], because different interactions and contributions of ERa and ERb in maintaining hippocampal-dependent memory during aging in female mice have been reported [51]. Future studies, including local drug administration, should reveal the specific contributions of these brain regions to the motivational states and behavioral performance in young rats and the possible rules underlying the experience-dependent structural changes in the steroid network.

Supporting Information
Video S1 Representative example of the behavioral performance of a rat treated with 7.4 mmol MPP (trial 10). (MPG) Video S2 Representative example of the behavioral performance of a rat treated with vehicle (trial 10). (MPG)