Prepubertal ovariectomy confers resilience to stress-induced anxiety in adult female mice

Abstract

The increased vulnerability to stress-induced neuropsychiatric disorders in women, including anxiety disorders, does not emerge until pubertal onset, suggesting a role for ovarian hormones in organizing sex-specific vulnerability to anxiety. Parvalbumin (PV) interneurons in the prefrontal cortex are a potential target for these ovarian hormones. PV+ interneurons undergo maturation during the adolescent period and have been shown to be sensitive to stress and to mediate stress-induced anxiety in female mice. To test the idea that ovarian hormones at puberty are necessary for the acquisition of sensitivity to stress, hypothetically driving the response of PV+ interneurons to stress, we performed ovariectomy or sham surgery before pubertal onset in female mice. These mice then were exposed to four weeks of unpredictable chronic mild stress in adulthood. We then assessed anxiety-like behavior and PV/FosB colocalization in the medial PFC. Additionally, we assessed stress-induced anxiety-like behavior in female mice following ovariectomy in adulthood to determine if puberty is a sensitive period for ovarian hormones in mediating vulnerability to stress. We found that prepubertal ovariectomy protects against the development of anxiety-like behavior in adulthood, an effect not found following ovariectomy in adulthood. This effect may be independent of ovarian hormones on prefrontal PV+ interneurons response to stress.

1. Introduction

Stress-induced neuropsychiatric disorders, including major depressive disorder and anxiety disorders, are approximately twice as common in women as in men (McLean et al., 2011; Altemus et al., 2014; Maeng and Milad, 2015). This sex-specific vulnerability, however, does not emerge until pubertal onset (Bale and Epperson, 2015), suggesting that pubertal development in females could shape brain circuits underlying response to stress during the adolescent period.

Adolescence is a developmental period characterized by the maturation of neural circuitry regulating cognitive, social, and emotional functions (Dahl 2004; Sturman and Moghaddam, 2011; Fuster 2001). Critical to the maturation of these behaviors is the structural and functional remodeling of the prefrontal cortex (PFC), a brain region important for the regulation of the stress response, that undergoes a protracted maturation during the adolescent period in humans, in non-human primates, and in rodents (Sowell and Jernigan, 1998; Sowell et al., 2003; Swartz et al., 2014; Alexander et al., 1982; van Eden 1986). In adolescence, the inhibitory GABAergic neurotransmitter system matures, driven in part by the maturation of parvalbumin (PV)-expressing GABAergic inhibitory interneurons. This leads to an increase in inhibitory control (Hashimoto et al., 2009; Silveri et al., 2013; Caballero et al., 2014; Caballero et al., 2016).

PV+ interneurons, characterized by their fast-spiking phenotype, play a major role in maintaining the excitatory-inhibitory balance in cortical circuits to contribute to cognitive, social, and emotional regulation (Lim et al., 2018; Ferguson and Gao, 2018; Zou et al., 2016; Murray et al., 2016). Disruptions in PV+ interneuron expression have been linked to the pathology of numerous neuropsychiatric disorders, including major depressive disorder and bipolar disorder (Rajkowska et al., 2007; Ruden et al., 2021). Furthermore, in preclinical models, PV+ interneurons are also responsive to treatment with antidepressant drugs such as fluoxetine or ketamine. Chronic fluoxetine treatment decreases PV+ interneuron expression in the frontal cortex (Takeuchi et al., 2013), while the antidepressant effects of ketamine are mediated in part by binding to NMDA receptors localized on PV+ interneurons, and both acute and chronic ketamine administration results in downregulation of PV mRNA expression (Gerhard et al., 2020; Zhou et al., 2014; Okine et al., 2020). Changes in PV+ expression have also been suggested as a potential underlying mechanism in anxiety disorders, as a greater density of PV+ interneurons in the cortex associated with increased anxiety-like behaviors in rodents (Ravenelle et al., 2014; Shepard and Coutellier, 2017; Shepard et al., 2016; Lee and Lee, 2021). Anxiogenic drugs increase PV+ interneurons in the amygdala, suggesting a role for these cells in the genesis of anxiety (Hale et al., 2010). PV+ interneurons are also sensitive to the effects of chronic stress, which is a major risk factor for the development of anxiety disorders (McEwen 2004; Duman and Monteggia, 2006; Page and Coutellier; 2019).

Though PV+ interneuron vulnerability to stress has been noted across various brain regions, , including the PFC, hippocampus, and basolateral amygdala (McKleeven et al., 2016; Guadagno et al., 2020; Hu et al., 2010), the directionality of these changes is highly variable depending on other important variables, including sex and period of the lifespan at which stress was applied (for review, please see Woodward and Coutellier 2021; Perlman et al., 2021). It is important to assess how these variables interact with chronic stress to influence changes in PV+ interneurons and affect brain function and emotional regulation. Our lab has previously found that prefrontal PV+ interneurons contribute to increased susceptibility to stress-induced anxiety-like behaviors specifically in female mice. We reported that prefrontal PV+ interneurons in female mice are particularly sensitive to chronic stress (Shepard et al., 2016), and that chronic stress induces hyperactivity of prefrontal PV+ neurons resulting in increased anxiety-like behaviors in female mice (Shepard et al., 2016; Page et al., 2019). Additionally, chemogenetic activation of prefrontal PV+ interneurons in the absence of stress induces anxiety-like behavior in female mice (Page et al., 2019). These findings identify the sex-specific sensitivity of prefrontal PV+ neurons as a potential contributor to heightened vulnerability to stress-induced anxiety in females. However, the mechanisms underlying the increased vulnerability of females PV+ interneurons to stress are unknown. Based on the protracted maturation of prefrontal PV+ neurons starting at puberty, there may be a role for ovarian hormones in conferring stress sensitivity.

Gonadal hormones play a role in brain development, with some studies suggesting that gonadal hormones in females play disproportionate roles in the maturation of late-developing brain regions. For instance, studies assessing the role of gonadal steroid hormones on adolescent brain development find that gonadal hormones play a role in mediating adolescent pruning of neurons and glia in the medial PFC (mPFC) of female, but not male rats (Koss et al., 2015). Additionally, prepubertal ovariectomy blocks the adolescent increase in inhibitory neurotransmission in the frontal cortex (Piekarski et al., 2017). Likewise, estradiol is required for PV+ interneuron maturation in the hippocampus of female, but not male, rats (Wu et al., 2014). PV+ interneurons express estrogen receptor beta (Blurton-Jones and Tuszynski, 2002), providing a potential mechanism for ovarian hormones to facilitate the development of inhibitory transmission. However, whether ovarian hormones contribute to the development of the stress-sensitive phenotype of prefrontal PV+ interneurons in females remains unknown.

Here, we hypothesize that ovarian hormones, including estradiol, mediate the development of brain circuitry responsible for stress-induced anxiety in adulthood. Specifically, we test the idea that female PV+ interneurons develop their sensitivity to stress during puberty through an ovarian hormone-mediated mechanism, leading to increased activity of these cells in response to stress exposure in adulthood. To determine this, we performed ovariectomy or sham surgery in prepubertal female mice and assessed both anxiety-like behaviors and markers of activity of prefrontal PV+ neurons following chronic stress exposure in adulthood in presence of 17β-estradiol supplementation. Our findings provide a novel understanding of the effects of pubertal maturation on anxiety-like behavior as well as new insight on the relationship between ovarian hormones, prefrontal PV+ neurons, and sensitivity to stress.

2. Methods

2.1. Animals

All experiments were conducted in C57Bl6/J mice from Jackson Laboratory (Maine). For the prepubertal ovariectomy experiment, mice were bred in-house and weaned at postnatal day (P) 21 into groups of 3–5 mice per cage. 1–2 mice were used within a single experimental group per litter. For the adult ovariectomy experiment, 10- and 11-week female C57Bl6/J mice were ordered and allowed to habituate to the colony room for one week. All mice were maintained on a 12-hour reverse light/dark cycle and had access to food and water ad libitum. All procedures related to animal maintenance and experimentation were approved by the Institutional Animal Care and Use Committee of the Ohio State University and conformed to the principles outlined by the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

In Experiment One, mice underwent prepubertal ovariectomy (Section 2.2) at P25, and immediately began hormone replacement via peroral estradiol (Section 2.3). Three weeks after surgery (P46), mice began four weeks of unpredictable chronic mild stress (UCMS) (Section 2.4). Based on our study design, mice in Experiment One were divided into 8 groups (2 surgery groups × 2 hormone replacement groups × 2 stress groups) with N=9–11 mice/group. In Experiment Two, mice underwent ovariectomy in adulthood (12–13 weeks). Ovariectomy in adulthood was not followed by any estradiol replacement. Two weeks following surgery, mice began four weeks of UCMS (Section 2.4). Based on our study design, mice in Experiment Two were divided into four groups (2 surgery groups × two stress groups) with N = 7–8 mice/group. In both experiments, mice underwent a series of behavioral tests (Section 2.5) beginning twenty-four hours following completion of the UCMS protocol prior to transcardial perfusion (Section 2.6) (Figure 1).

Figure 1.

Experimental design. Created with BioRender.com.

2.2. Ovariectomy

Ovariectomies were performed at P25 in prepubertal mice (Experiment One, n = 9–11 per group) and at 12–13 weeks in adult mice (Experiment Two, n = 7–8 per group). P25 was chosen to eliminate the physiological or behavioral impacts of increased female gonadal hormones released during puberty, which has been observed at a mean age of P29 in female C57Bl6 mice (Bell et al., 2018). Prior to ovariectomy, all adolescent mice were visually inspected to confirm lack of vaginal opening, the primary external indicator of pubertal onset in rodents (Rodriguez et al., 1997.

Ovariectomy was conducted under deep anesthesia (1.5–5% isoflurane). A dorsal midline incision was made through the skin and muscle layer, and the ovarian fat pad was extended through the incision. The fallopian tubes and surrounding blood vessels were clamped with a hemostat, followed by removal of the ovary. The fat pad was then tucked back into the muscle layer. This procedure was repeated on the bilateral side. Skin incisions were closed using sutures. Sham animals underwent the same procedure, but the ovaries were simply visualized, with no clamping or removing of any tissue. Buprenorphine (0.1mg/kg) and carprofen (5mg/kg) were given as analgesic.

2.3. Hormone Replacement

Following prepubertal ovariectomy or sham surgery, mice received estradiol as a form of hormone replacement, as estradiol has been shown to play a role in the adolescent maturation of PV+ interneurons and to regulate their firing in adulthood (Wu et al., 2014; Clemens et al., 2019). Estradiol replacement following prepubertal ovariectomy was achieved through peroral administration of 17β-estradiol in hazelnut cream (Nutella; Ferro Scandinavia AB, Malmo, Sweden) (Strom et al., 2012). Beginning at weaning, mice were given 60 mg of hazelnut cream once a day for four days for habituation. For the remainder of the experimental period, mice were placed into individual cages once daily in the morning and served 60 mg of hazelnut cream containing 1.12 μg 17β-estradiol mixed with corn oil (dose based on Ingberg et al. 2012). Mice were returned to their original housing upon finishing the hazelnut cream, on average 1–2 minutes later. Mice receiving vehicle were given 60 mg of hazelnut cream with corn oil daily for the duration of the experiment.

2.4. Unpredictable Chronic Mild Stress

The unpredictable chronic mild stress (UCMS) protocol started two (adult) and three (prepubertal) weeks following ovariectomy or sham surgery. Control mice remained group housed and handled daily for 1–2 minutes over a 4-week period. Stressed mice were single housed and underwent UCMS as previously described (Page et al. 2019; Shepard et al. 2016). Briefly, mice were exposed to alternating mild stressors once daily for four weeks, a period which has previously been shown to be sufficient to induce an anxiety-like phenotype in female mice as well as increase activity in prefrontal PV+ cells in both males and females (Page et al. 2019). Stressors included removal of nesting material for 24 hours, absence of bedding in the cage for 8 hours, 20° cage-tilt for 6 hours, restraint under bright light for 4 minutes, and restraint in the dark for eight minutes. Stressors were applied on an unpredictable schedule during the dark phase of the light cycle. (Figure 1).

2.5. Behavioral Testing

Behavioral testing began 24-hours after the end of the UCMS or handling period to measure emotional behavior. Testing was performed during the dark phase of the light cycle. Mice were acclimated to the testing room for at least one hour prior to testing. The following tests were performed in this order (with at least 24 hours between each test): elevated plus maze (EPM), open field test (OFT), splash test, marble burying test (MBT), and social interaction test.

Briefly, the EPM consists of a raised arena with four arms (each 35 cm long), with two closed arms surrounded by 22-cm-high black walls. Mice were placed in the center of the maze and allowed to freely explore under red light for 5 minutes, after which they were returned to their home cage and the arena was cleaned with 70% ethanol. Increased time spent in the closed arms of the EPM is interpreted as increased anxiety-like behavior. The OFT consists of a 40×40 cm arena with white opaque walls. Mice are placed in the arena and allowed to freely explore for 10 minutes under dim white light before being returned to their home cage, Again, the arena was cleaned with 70% ethanol prior to testing the next mouse. Increased time spent near the walls of the open field is interpreted as increased anxiety-like behavior. The EPM and OFT are recorded with an overhead camera for offline analysis. Automated, unbiased analyses were conducted using the EthoVision XT software from Noldus Information Technology (Wageningen, Netherlands).

The marble burying test consists of mice being placed in a clean cage with 20 marbles placed on top of 5 cm of fresh bedding. Mice are left undisturbed in the dark for 30 minutes before being returned to their home cage. The number of marbles buried (over two-thirds covered with bedding) is recorded; an increased number of marbles buried in the MBT is indicative of increased anxiety-like behavior, particularly following prior stress exposure (Keida and Chattarji, 2014; Torok et al., 2019).

The splash test and social interaction test were used to assess other domains of emotional behavior, particularly depressive-like behavior. In the splash test, mice were allowed to habituate to a new cage for 10 minutes prior to being splashed with a 10% sucrose solution. Latency to groom and total time spent grooming over a five-minute period were scored by a blinded experimenter, with increased latency to groom and decreased total time spent grooming as increased apathy and decreased self-care behavior (Isingrini et al., 2010). In the social interaction test, mice are habituated to a clean cage for 10 minutes. After the habituation period, two cups are introduced into the cage, one of which containing an unfamiliar same sex mouse. The time spent sniffing each cup over a five-minute period is scored by a blinded experimenter. These times are then used to calculate a social preference ratio, as described below:

$$\mathit{\text{social}}\mathit{\text{preference}}\mathit{\text{ratio}}=\frac{\left(\mathit{\text{Time}}\mathit{\text{sniffing}}\mathit{\text{novel}}\mathit{\text{mouse}}\right)-\left(\mathit{\text{Time}}\mathit{\text{sniffing}}\mathit{\text{empty}}\mathit{\text{cup}}\right)}{\mathit{\text{Total}}\mathit{\text{Time}}\mathit{\text{sniffing}}}$$

A lower social preference ratio indicates that the subject prefers the empty cup, indicating increased social anxiety and anhedonia (Scheggi et al., 2018; Yang et al., 2011).

2.6. Immunohistochemistry

Twenty-four hours after the last behavioral test, mice were anesthetized with isoflurane and transcardially perfused with 25 mL of 0.1 M phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (PFA). The brains were then removed and placed in fresh PFA for 2 days at 4°C, after which they were placed in 30% sucrose. Brains were sectioned at 50 μM with a cryostat in a manner to obtain 3 sets of the PFC, including the prelimbic and infralimbic regions, according to the Mouse Brain in Stereotaxic Coordinates (Paxinos and Franklin, 2004, reference figures 14–18). After sectioning, the thin slices were stored in cryoprotectant at −20°C until immunostaining. After a series of washes in PBS with 0.1% Triton, sections were blocked with 5% normal donkey serum and 1% bovine serum albumin for one hour. Sections were then incubated overnight at 4°C with a guinea pig anti-PV antibody (1:500, Synaptic Systems, 195004) and a rabbit recombinant anti-FosB/delta FosB antibody (1:2000, Abcam ab184939). FosB is a marker of chronic activity which accumulates following long-term stimulation (Garcia-Perez et al., 2012), with levels of FosB positively correlating with chronic stress in the PFC (Vialou et al., 2015). After another series of washes in PBS with 0.1% Triton, sections were incubated for two hours at room temperature using Alexa Fluor donkey anti-guinea pig 488 (1:250) and Alexa Fluor donkey anti-rabbit 555 (1:500) secondary antibodies. Sections were then washed, mounted, and cover slipped with DAPI mounting media. The quantitative analysis of PV cells that have FosB in the prelimbic and infralimbic PFC was performed using the unbiased stereology method with StereoInvestigator software from MBF Bioscience (Willston, VT). Cells were counted in every three sections from four animals in each group. Accuracy of the estimate of the total number of positive-stained cells based on our counting was assured by verifying that the mean coefficient of error (CE) of Gundersen (Gundersen et al., 1999) was below 0.10.

2.7. Statistics

Behavioral and molecular data were analyzed using RStudio and Prism 9.01. For Experiment One, three-way ANOVAs (with surgery, estradiol replacement, and stress as independent factors) were used to determine if ovarian hormones at puberty influence anxiety-like behavior to chronic stress in adulthood. Two-way ANOVAs (with ovariectomy and stress as independent factors) with multiple comparisons were used for Experiment Two to determine if ovarian hormones influence anxiety-like behavior in response to chronic stress in adulthood. When appropriate, Tukey’s postdoc analyses were conducted. All statistics from the ANOVAs are reported (Supplementary Tables 1–3).

3. Results

3.1. Experiment One: prepubertal ovariectomy prevents adult UCMS-induced anxiety-like behaviors.

Mice underwent ovariectomy or sham surgery prior to the onset of puberty (P25). Following ovariectomy, mice were supplemented with 17-β estradiol or corn oil and exposed to four weeks of UCMS or daily handling (Figure 1).

In the EPM, we found a main effect of chronic stress inducing hyperlocomotion (F_1,65 = 6.284, p = 0.0147; figure 2a). We did not find an effect of ovariectomy, estradiol replacement, or chronic stress exposure on number of entries into the open arms (figure 2b). Chronic stress also increased anxiety-like behavior, as evidenced by reduced time spent in the open arms (F_1,65 = 7.093, p = 0.0097; figure 2c), and decreased distance traveled in the open arms (F_1,65 = 25.986, p < 0.0001; figure 2d). Estradiol supplementation led to an overall reduction of anxiety-like behaviors as shown by more time spent (F_1,65 = 13.064, p = 0.0097; figure 2c) and longer distance traveled in the open arms (F_1,65 = 11.768, p = 0.0010; figure 2d) compared to vehicle-treated animals.

Figure 2.

Chronic stress exposure induces hyperlocomotion (a) and increases anxiety-like behavior in the elevated plus maze, an effect mitigated by estradiol supplementation (c,d). In the open field test, chronic stress increases hyperlocomotion (e), and estradiol supplementation following prepubertal ovariectomy increases anxiety like-behavior in animals (f-h). * p < 0.05, ** p < 0.01, *** p < 0.001 (mean ± SEM).

In the OFT, chronic stress induced hyperlocomotion (F_1,69 = 7.554, p = 0.0076; figure 2e). Additionally, we found a three-way interaction between ovariectomy, estradiol supplementation, and chronic stress inducing hyperlocomotion, however there were no significant post-hoc effects (F_1,69 = 4.174, p = 0.04487; figure 2e). Estradiol increased anxiety-like behavior in the OFT, with mice that received estradiol supplementation making fewer entries into the center of the open field (F_1,69 = 6.160, p = 0.0155; figure 2f), spending less time spent in the center of the open field (F_1,69 = 4.860, p = 0.0308; figure 2g), and traveling a shorter distance into the center of the open field (F_1,69 = 9.210, p = 0.0034; figure 2h). We also found a significant two-way interaction between estradiol and surgery on number of entries into the center of the open field (F_1,69 = 4.796, p = 0.0320; figure 2f), time spent in the center of the open field (F_1,69 = 4.083, p = 0.0472; figure 2g), and distance traveled in the center of the open field (F_1,69 = 5.747, p = 0.0192; figure 2h). Tukey’s post hoc test revealed that OVX/estradiol mice displayed more anxiety-like behavior than OVX/vehicle mice in these measures (p = 0.0080, p = 0.0472, and p = 0.0014, respectively). Moreover, we find that ovariectomy, estradiol supplementation, and chronic stress interact to increase the number of entries into the center of the open field (F_1,69 = 6.104, p = 0.0160). Tukey’s post-hoc analyses reveal that OVX/estradiol mice that were not exposed to chronic stress displayed significantly increased anxiety-like behavior compared to sham/vehicle (p = 0.0158) and OVX/vehicle (p = 0.0304) mice that underwent UCMS and OVX/vehicle control mice (p = 0.0408), as evidenced by making fewer entries into the center of the open field (figure 2f).

In the MBT, chronic stress increased the number of marbles buried (F_1,59 = 4.145, p = 0.0463; figure 3a), particularly in sham animals, as evidenced by a two-way interaction between surgery and stress (F_1,59 = 5.819, p = 0.0190). Post-hoc analyses find that animals that underwent UCMS following sham surgery buried more marbles compared to those that underwent handling (p = 0.0143).

Figure 3.

Prepubertal ovariectomy protects against stress-induced anxiety-like behavior in the marble burying test (a), while UCMS in adulthood induces apathy in the splash test (b). Neither stress nor hormones influence behavior in the social interaction test following prepubertal ovariectomy (c). * p < 0.05, *** p < 0.001 (Mean ± SEM).

In the splash test, stress decreased total time spent grooming following splash with 10% sucrose solution (F_1,66 = 14.957, p = 0.0003; figure 3b), as well as a main effect of estradiol supplementation decreasing total time grooming (F_1,66 = 13.301, p = 0.0005; figure 3b). No effects of surgery, estradiol supplementation, or stress were found to effect latency to begin grooming in the splash test (data not shown). In the social interaction test, there were no significant effects of ovariectomy, estradiol supplementation, or stress exposure on the social preference ratio (figure 3c).

Overall, we observed that prepubertal ovariectomy protects against increased anxiety-like behaviors in adulthood, particularly as measured in the OFT and MBT. Additionally, we found that estradiol supplementation had test-dependent effects on UCMS-induced anxiety- and depressive-like behavior.

3.2. Experiment One: effect of prepubertal ovariectomy on adult UCMS-induced increased activity of prefrontal PV+ neurons

To assess the potential effect of prepubertal ovariectomy and estradiol supplementation on prefrontal PV+ neurons UCMS-induced increased in activity (Page et al, 2019), we used double immunohistochemistry with PV and FosB as a marker of chronic activity (figure 4a). We did not find a significant effect of prepubertal ovariectomy, estradiol supplementation, or chronic stress on prefrontal PV expression (figure 4b). We observed a significant two-way interaction between surgery and estradiol on number of PV+ interneurons expressing FosB (F_1,38 = 6.321, p = 0.0163; figure 4c), however no significant post-hoc effects were found. Additionally, we observed a significant three-way interaction between surgery, estradiol, and stress on number of PV+ interneurons expressing FosB (F1_,38 = 5.440, p = 0.0251; figure 4c). Post-hoc analysis found that OVX/vehicle mice tended to have a higher number of PV+ interneurons expressing FosB at baseline compared OVX/estradiol mice (p = 0.0673; figure 4c). There was also a trend of surgery and estradiol interacting to impact the percent colocalization of PV+ interneurons with FosB (F_1,38 = 3.786, p = 0.0593; figure 4d).

Figure 4.

Representative image of PV (green) and FosB (red) within the prelimbic prefrontal cortex taken at 63X (a). We do not identify any effects of prepubertal ovariectomy, estradiol supplementation, or chronic stress on overall PV+ interneuron number (b), however these factors interact to influence PV/FosB colocalization, as Sham/E2 and OVX/vehicle animals display increased levels of PV/FosB in the PFC (c). Disruption to normal levels of estradiol at puberty tended to increase PV+ activation in the PFC, as evidenced by higher levels of colocalization in Sham/E2 and OVX/vehicle groups (d). * p < 0.05 (Mean ± SEM)

3.3. Experiment Two: effect of adult ovariectomy on UCMS-induced anxiety-like behaviors.

To determine whether the effects of ovariectomy on UCMS-induced anxiety-like behaviors are age-specific, we performed ovariectomy in adulthood on a second cohort of mice. Overall, we observed that ovariectomy in adulthood did not prevent the increase in anxiety-like behaviors induced by chronic stress exposure.

In the EPM, chronic stress did not impact locomotor activity, but adult ovariectomy decreased the total distance traveled (F_1,26 = 5.534, p = 0.0265; figure 5a). Chronic stress significantly increased anxiety-like behavior as shown by reduced number of entries into the open arms (F_1,26 = 7.580, p = 0.0106; figure 5b), time spent in the open arms (F_1,26 = 21.96, p < 0.001; figure 5c), and distance traveled in the open arms (F_1,26 = 16.74, p = 0.0004; figure 5d). In the OFT, stress increased total distance traveled (F_1,26 = 5.733, p = 0.0241; figure 5e). This effect was driven by a significant interaction between ovariectomy and stress (F_1,26 = 6.487, p = 0.0171), with mice that underwent ovariectomy in adulthood displaying significantly more activity within the OFT following chronic stress than those that underwent sham surgery (p = 0.0065). There were no significant effects of ovariectomy or stress exposure on number of entries (figure 5f) or distance traveled in the center of the arena (figure 5h), however stress exposure tended to decrease time spent in the center of the arena (p = 0.073, figure 5g).

Figure 5.

Adult ovariectomy decreases distance traveled in the elevated plus maze, while chronic stress-induces overall anxiety-like behavior (a-d). In the open field, adult ovariectomy and stress interact to increase overall locomotor activity (e), but do not influence the number of entries into the center (f), the time spent in the center (g), or the distance in the center of the open field (h). * p < 0.05, *** p < 0.001 (Mean ± SEM)

In the MBT, chronic stress exposure increased the number of marbles buried (F_1,24 = 7.314, p = 0.0124; figure 6a). In the splash test, chronic stress exposure significantly decreased the latency to groom ((F_1,25 = 4.867, p = 0.0368; figure 6b), and we observed a near-significant effect of ovariectomy on total time grooming (F_1,25 = 3.924, p = 0.0582; figure 6c). No effects were observed in the social preference test (figure 6d).

Figure 6.

Ovariectomy in adulthood does not protect against the stress-induced increase in marble burying behavior following chronic stress (a). In the splash test, stress exposure significantly decreases latency to groom (b), while ovariectomy has a tendency to increase total time grooming (c). We did not find an effect of ovariectomy or stress in the social interaction test (d). * p < 0.05 (Mean ± SEM)

4. Discussion

The mechanisms underlying increased risk to stress-induced anxiety in females are unknown. Here, we tested the hypothesis that the increase of ovarian hormones during puberty drives this female-specific susceptibility to stress by shaping the responsivity of prefrontal PV+ neurons to chronic stress. We performed prepubertal ovariectomies on female mice with the prediction that the absence of ovarian hormones during puberty would prevent the development of stress-induced anxiety-like behaviors and increased activity of prefrontal PV+ neurons in adulthood as previously reported (Shepard et al, 2016; Page et al, 2019). We found that prepubertal ovariectomy in female mice prevents the appearance of anxiety-like behaviors in adulthood after chronic stress exposure in certain subdomains of anxiety, as measured by the OFT and MBT, The behavioral phenotype of increased anxiety-like behavior following chronic stress exposure has been associated in previous studies with increased number of prefrontal PV+ neurons expressing cFos, a marker of increased neuronal activity (Page et al., 2019). Overall, this suggests that according to our hypothesis, pubertal ovarian hormones shape vulnerability to stress in adult females, however this might be independent of their effects on prefrontal PV+ neurons, as we did not find any direct effects of prepubertal ovariectomy on the number of PV+ interneurons in the PFC.

By conducting ovariectomy in adulthood, we were able to demonstrate that the effects of ovarian hormones on stress-induced anxiety are specific to the pubertal or adolescent period. Indeed, adult ovariectomy does not prevent the appearance of anxiety-like behaviors following stress in the EPM or MBT. This mirrors findings from numerous other studies, which report that ovariectomy in adulthood increases anxiety-like behavior in rodents, an effect which can be recovered with supplementation of estrogen and progesterone (Bowmen et al., 2002; Renczés et al., 2020).

We found mixed effects of hormonal manipulation and stress among different behavioral assays following prepubertal ovariectomy, suggesting that certain behavioral assays may be more sensitive to ovarian status than others. Mice that were exposed to chronic stress showed increased anxiety in the EPM and the MBT, and behavior in both of these assays was impacted by hormonal status. In the EPM, we find that estradiol supplementation decreased anxiety-like behavior, as evidenced by increased time spent and increased distance traveled in the open arm. In the MBT, we found that UCMS only increased marble burying behavior in animals that underwent sham surgery, suggesting that prepubertal ovariectomy may be protective against stress-induced anxiety-like behavior. This effect was not recovered by estradiol supplementation following prepubertal ovariectomy, it may be mediated by another ovarian hormone, such as progesterone. Multiple studies have found that progesterone administration reduces marble burying behavior in ovariectomized and intact female rats (Llanzea and Frye, 2009; Schneider and Popik, 2007).

In the OFT, however, we found that prepubertal ovariectomy and estradiol supplementation impacted anxiety-like behavior independently of stress exposure. Mice that undergo prepubertal ovariectomy without estradiol replacement show significantly less anxiety-like behavior in the open field compared to those who received estradiol replacement following prepubertal ovariectomy, suggesting that the absence of estradiol at puberty may protect against the development of anxiety in adulthood. Estradiol supplementation influenced anxiety-like behavior in an inverse manner in the OFT and the EPM; while estradiol supplementation decreased anxiety-like behavior in the EPM, it increased anxiety-like behavior in the OFT.

This may be due to differences in the subdomains of anxiety that these behavioral assays assess. For instance, the EPM has been used as a measure of state anxiety, particularly after subjects have been exposed to repeated stress (Belzung and Griebel, 2001; Jakovcevski, Schachner, and Morellini, 2007), while the OFT has been used as a measure of trait anxiety (Hawley, Grissom, and Dohanich, 2011; de Kort et al., 2021). It is possible that circulating estradiol may differentially regulate trait and state anxiety. Indeed, a 2018 study found that levels of estradiol in women moderated the relationship between state and trait anxiety, with higher levels of circulating estradiol being associated with higher attentional bias to threat (Graham and Shin, 2018). Previous studies in rodents have also reported a link between ovarian hormones and anxiety-like behavior in these behavioral assays. Female rodents are more active in the open field compared to males, an effect eliminated by ovariectomy in adulthood and recapitulated by estrogen and progesterone replacement (Blizard et al., 1975); similar effects have also been reported to influence anxiety-like behavior in the EPM (Zimmerberg and Farley, 1993). Anchan and colleagues (2014) found that administration of an agonist to the estrogen receptor GPR30 in adult female rats decreased anxiety-like behavior in the OFT but had no effect on behavior in the EPM (Anchan et al., 2014), suggesting that there may be differential roles for estrogens in regulating sub-domains of anxiety-like behavior, perhaps in a receptor-dependent manner.

To our knowledge, this study is the first to assess stress-induced anxiety following prepubertal ovariectomy. Other recent experiments assessing basal anxiety-like behavior following prepubertal gonadectomy also find that prepubertal gonadectomy reduces anxiety-like behavior in females. For instance, Delevich et al. finds that prepubertal gonadectomy decreases basal levels of anxiety in female mice as measured by the approach-avoidance task (Delevich et al., 2020). Likewise, prepubertal gonadectomy increases baseline social interaction in female rats (Kim and Spear, 2016). In Siberian hamsters, prepubertal ovariectomy increased exploration and novelty seeking in females (Kyne et al., 2019). Not all studies eliminating the effects of ovarian hormones prior to puberty find changes in anxiety-like behavior, however. Female rats administered a gonadotropin-releasing hormone antagonist during puberty did not display long-term changes in basal anxiety-like behavior or basal levels of c-Fos staining in the medial or basolateral amygdala (Hodgson et al., 2020). This suggests that ovarian hormones mediate specific organizational effects on the development of neural circuitry involved in the regulation of anxiety-like behaviors during puberty in a potentially region-specific manner.

This idea is supported by findings showing that estradiol is required for PV+ interneuron maturation in the hippocampus of female, but not male, rats (Wu et al., 2014). Here, we observed that while total number of PV+ neurons in the PFC was not impacted by ovariectomy or estradiol replacement, prepubertal ovariectomy without estradiol replacement in control mice tended to increase the number of PV+ interneurons expressing FosB in the mPFC at baseline compared to OVX/estradiol mice. This suggests that estradiol plays a role in the regulation of PV+ interneuron activation, as supported by previous findings (Piekarski et al., 2017; Clemens et al., 2019). These results parallel behaviors in the EPM, where OVX/vehicle mice displayed significantly higher levels of anxiety-like behavior compared to OVX/estradiol mice. This may be due to the increased activation of PV+ interneurons in this group, which has previously been associated with increased anxiety-like behavior in female rodents (Page et al, 2019). However, because our approach removed all ovarian hormones at a prepubertal timepoint and subjects remained hormone-deprived throughout adulthood, it is not possible to distinguish between the potential organizational effects of pubertal hormones on PV+ neurons vs. the activational effects that might occur during adulthood.

While we do not report the increase in PV+ activation that our lab has previously shown following chronic stress in females, there does appear to be an increase in PV expression following chronic stress in our control animals which would align with our results in previous studies (Shepard et al, 2016; Page et al, 2019). This increase in PV expression does not reach significance, however. Additionally, as levels of PV expression are activity dependent, this may reflect an increase in activity of PV+ interneurons that is not captured by FosB double staining. In future studies, a different marker of neuronal activation such as c-Fos could be used to assess levels of PV+ activity.

In this experiment, we used a peroral method of estradiol supplementation as previously described by Ingberg et al (2012) as a method of hormone replacement following prepubertal ovariectomy. The dose of estradiol used in this experiment was based on physiological levels at the body weight of 8–10-week-old female mice, however estradiol administration at this dose began immediately following prepubertal ovariectomy at approximately 3.5 weeks of age before body weight stabilization and remained constant, independent of individual body weight. Mice that underwent prepubertal ovariectomy, therefore, may have been exposed to supraphysiological levels of estradiol during early life. Therefore, we cannot conclude that the physiological dose of estradiol at the pubertal period would not have led to the development of chronic stress-induced anxiety-like behavior in adulthood. This may support behavioral similarities between subjects that underwent prepubertal ovariectomy with estradiol supplementation and those that underwent sham surgery and received estradiol supplementation. Additionally, estradiol supplementation in animals that underwent prepubertal ovariectomy continued beyond the pubertal period and into adulthood. Considering that puberty is a sensitive period for the organizational effects of gonadal steroid hormones including estradiol on brain circuitry (Schulz and Sisk, 2016), the continuation of estradiol supplementation into the adult period may have obfuscated the potential recovery of behaviors that develop during this period. The timing and dosage of estradiol administration should be carefully considered in future experiments.

Estradiol is not the only ovarian hormone contributing to changes in the brain during the pubertal period, and our finding that estradiol administration does not reverse the stress-resilient phenotype following prepubertal ovariectomy may suggest that other ovarian hormones contribute to the development of stress sensitivity at puberty in females. A likely contributor to these effects is progesterone and its metabolite, allopregnanolone. Like estradiol, progesterone is synthesized both in the periphery and the brain. Allopregnanolone plays a role in modulating GABA_A receptors to increase inhibitory neurotransmission and has been implicated in regulating neuronal firing and mood during several periods of hormonal change, including puberty, pregnancy, and the postpartum period (Mackenzie & Maguire, 2014). Notably, administration of allopregnanolone during the pubertal period has anxiogenic effects in female mice, whereas administration of allopregnanolone during juvenility or adulthood is anxiolytic (Shen et al., 2007). While allopregnanolone is synthesized in the brain even after ovariectomy (Paul and Purdy, 1992), the ovaries are the primary source of progesterone and allopregnanolone, and removal of the ovaries may lead to the effects we see on GABAergic activation. Future experiments should consider the systemic effects of ovariectomy and the impact of eliminating all ovarian hormones on GABAergic activation, particularly through investigating the effects of removing progesterone and allopregnanolone during the pubertal period.

Overall, we find that prepubertal ovariectomy confers some protection to stress-induced anxiety-like behavior in adulthood, an effect not seen following ovariectomy in adult mice. This supports the idea that ovarian hormones at puberty organize brain circuitry that mediate stress-induced anxiety-like behavior in adulthood. This effect, however, may be specific to individual behavioral assays or certain subdomains of anxiety, as we observed differences in the effects of ovariectomy and estradiol administration across our behavioral paradigm. Our molecular analyses of prefrontal PV+ neurons under baseline and stressful conditions show that prepubertal ovariectomy without estradiol replacement and estradiol supplementation after sham surgery leads to increased PV+ interneuron activation as measured by PV/FosB colocalization but neither ovariectomy nor estrogen replacement changes their response to stress. This suggests that ovarian hormones do play a role in regulating PV+ interneuron activity but might not influence their stress responsivity. Our immunohistochemistry approach, however, provides limited information, and other approaches, including electrophysiology, could provide more detailed information on the behavior of these neurons in response to stress and in presence/absence of gonadal hormones. These findings may also be limited by further differences in prepubertal and adult subjects in this study. While subjects that underwent prepubertal ovariectomy were bred in-house, subjects that were examined following adult ovariectomy were shipped to our animal facility. Shipping of animals has been identified as a ‘hidden variable’ that can lead to unexpected outcomes in behavioral testing (Butler-Struben et al., 2022); in this experiment, however, mice were shipped at 10 weeks of age, and therefore may have been less susceptible to the effects of transport stress (Laroche et al., 2009). Additionally, mice that underwent prepubertal ovariectomy began UCMS three weeks following surgery, whereas adult mice began UCMS two weeks following surgery. These differences in animal handling and experimental design between the two groups may have contributed to some of the differences we report in anxiety-like behavior following juvenile and adult ovariectomy. In addition, future studies should also aim to target the pubertal period more precisely to elucidate the activational versus organizational effects of ovarian hormones on stress sensitivity at the behavioral and neuronal level. Further work on this topic should broaden the scope of this experiment and examine other ovarian hormones able to modulate GABAergic signaling, particularly progesterone and its metabolite allopregnanolone as potential mediators of anxiety and PV+ interneuron activity. As we found that ovarian hormones during puberty may influence certain domains of anxiety-like behavior, this work has important implications for the use of hormonal contraceptives in early adolescence. Some studies assessing the impacts of hormonal contraceptive use on adolescent brain development have found that oral contraceptive use across adolescence is associated with increased vulnerability to depression (Anderl et al., 2020; de Wit et al., 2020). A recent study, however, found that oral contraceptive users may be protected against the increase in anxiety and depressive symptoms in adolescence, even after controlling for other lifestyle factors (Doornweerd et al., 2022). This aligns with our findings in which prepubertal ovariectomy without estradiol replacement prevents some domains of anxiety-like behavior. The nature of ovarian hormones on influencing susceptibility to anxiety and depression needs to be further assessed to fully understand how hormonal contraceptive use in the adolescent population may contribute to risk for neuropsychiatric disease. Furthermore, identification of how these hormones interact with the developing GABAergic system, particularly at puberty, has the potential to influence future pharmacological targets for treatment for anxiety disorders.

Highlights.

• Prepubertal ovariectomy prevents aspects of stress-induced anxiety-like behavior in adulthood.

• Ovariectomy in adulthood did not prevent the development of anxiety-like behaviors following four weeks of chronic stress.

• Prepubertal ovariectomy without estradiol replacement may increase baseline levels of PV+ interneuron activation in the mPFC of female mice.