Obsessive-compulsive disorder (OCD) is characterized by intrusive thoughts (obsessions) and/or repetitive behaviors (compulsive rituals) in response to the obsessions (American Psychiatric Association, 2013). OCD has a lifetime prevalence of around 2.3% in the United States (Ruscio et al., 2010) and it has been listed as a common mental disorder in adults (Eaton et al., 2008). The obsessive beliefs lead to compulsive symptoms among patients. For example, contamination obsessions can result in compulsive cleaning (Wheaton et al., 2010). Moreover, OCD can negatively impact cognitive and affective functions in humans. Human studies involving neurocognitive tests and image analysis showed impairments in non-verbal (Kashyap et al., 2013), spatial working (van der Wee et al., 2007; Nakao et al., 2009) and visual memories (Dirson et al., 1995). Associated comorbidities like depression (Peris et al., 2010; Remijnse et al., 2013) and anxiety disorders (Nestadt et al., 2001) are also very common in the OCD condition.

Clinical and genetic data for OCD corroborate the hypothesis of sexual dimorphism, which reveals differences in clinical manifestations between males and females (Labad et al., 2008; Torresan et al., 2009). Obsessions for cleaning and compulsive contamination are more prevalent in females than males, while males have higher rates of symmetrical and sexual obsessions when compared to females (Noshirvani et al., 1991; Lensi et al., 1996; Bogetto et al., 1999; de Mathis et al., 2008; Labad et al., 2008). There is also a sex difference to treatment response (Mundo et al., 1999). Women typically have a later onset when compared to men and display a bi-modal distribution with the first peak occurring between 13–16 years of age and the second peak around 22–32 years. These are puberty and child bearing stages in a women’s life, respectively (Brandes et al., 2004) when sex hormone (estrogen and progesterone, P4) levels are known to fluctuate.

It is well established that a plausible cause of OCD is abnormal cortical-striatal-thalamic circuitry activation (Ahmari et al., 2013) and altered serotonergic (Schilman et al., 2010), glutamatergic (Arnold et al., 2004; Egashira et al., 2013; Porton et al., 2013) and GABAergic (Egashira et al., 2013) systems. Interestingly, female hormones, such as estrogen and P4, regulate various neurotransmitter signaling pathways in brain regions implicated in OCD (Dreher et al., 2007; Karakaya et al., 2007; Benmansour et al., 2009; Alonso et al., 2011; Quinlan et al., 2013; Barth et al., 2015). During the estrous phase, circulating estrogen levels are higher and serotonin release is lower in striatal neurons (Yang et al., 2015), while in the frontal cortex, estrogen depletion by ovariectomy (OVX) decreases 5-HT2A receptor density and mRNA levels (Cyr et al., 1998). On the other hand, P4 increases dopamine release mediated by NMDA receptor activation in striatal neurons (Cabrera and Bregonzio, 1996) and decreases NMDA binding density in the frontal cortex after OVX (Cyr et al., 2000). Therefore, ovarian sex hormones may account for the sex differences observed in OCD.

Women are subjected to hormonal fluctuations during their entire life span, which may lead to significant alterations in mood and cognition (Soares and Zitek, 2008). However, remarkable challenges are encountered during the menopause transition due to the natural decline in ovarian function, the primary source of estrogen and P4 (Luine, 2014). Natural menopause is also associated with cognitive deficits and mood disorders (Weber et al., 2012; Dumas et al., 2013). Such dysfunction in mood and cognitive functions has also been reported in women with surgical menopause (Chen et al., 2013; Faubion et al., 2015). Physiological challenges during surgical menopause are much more drastic due to a sudden depletion of ovarian sex steroids as compared to progressive menopause which follows fluctuating patterns of steroid levels (Bachmann, 2001; Rodriguez and Shoupe, 2015; Rodríguez-Landa et al., 2015). This results in greater predisposition to mood and anxiety disorders when compared to natural menopause (Rodríguez-Landa et al., 2015).

The impact of acute ovarian dysfunction during surgical menopause on compulsive behaviors and comorbid affective behaviors in females are currently poorly understood. In addition to younger women, one out of eight women after the age of 55 undergoes bilateral oophorectomy (surgical removal of ovaries) before reaching natural menopause due to benign diseases, prophylaxis against cancer and autoimmune disorders (Shuster et al., 2010; Erekson et al., 2013; Cox and Liu, 2014). Existing studies have investigated obsessions and compulsions only during and after progressive menopause with contradictory evidence (Vulink et al., 2006). One such study showed that, OCD is not a rare comorbidity during post menopause (Uguz et al., 2010), while another study demonstrated that the symptoms are more related with menarche and decreases during menopause (Guglielmi et al., 2014).

In animal studies, acute administration of estradiol (E2) in pre-pubertal female rats exerted an anti-compulsive-like effect (Flaisher-Grinberg et al., 2009), while male mice with estrogen deficiency (aromatase enzyme knockout) displayed compulsive-like behavior (Hill et al., 2007). In OVX rats, concurrent administration of E2 and P4 was able to reduce compulsive-like lever pressing behavior (Fernández-Guasti et al., 2006). Most of these studies were conducted on induced (drug or gene knockouts) models and did not investigate the associated comorbidities like anxiety and cognitive impairments, while only one study looked at the effect of P4 and E2 in the ovariectomized condition (Fernández-Guasti et al., 2006). Moreover, though OCD has a compelling genetic basis (Nestadt et al., 2010) the role of genetic background in influencing steroid actions in OCD condition during menopause has never been explored.

How do already existing compulsions in females get affected during acute sex hormone deprived conditions when compared to non-compulsive females are not clearly known. The co morbid anxiety and cognitive functions associated with OCD during such a physiological state and the role of genetic background in influencing steroid actions demands investigation. According to Maio et al. (2014), our mice developed through selective breeding for phenotypes of increased or decreased amounts of compulsive-like behavior can be a heuristic tool for studying OCD, especially the replicate BIG strains (BIG1 and BIG2). An unpublished study from our lab has shown that there is variation in compulsive-like and affective behaviors between the two replicate BIG strains that mimics heterogeneity as seen in subgroups of OCD patients. This study however did not look into hormonal deprivation and manipulations. We therefore investigated the hypothesis that acute deprivation of estrogen and P4 through OVX for 7 days will increase the compulsive-and anxiety-like behavior and impair novel object recognition memory in compulsive-like mouse strains. We also hypothesized that the administration of estrogen (E2) and P4 will attenuate the exacerbation in compulsive-like, anxiety-like and cognitive behaviors in compulsive-like strains. Though rodents do not have menopause, surgical removal of the ovaries can cause depletion of E4 and P4 (Kato et al., 2013). We therefore used bilateral OVX as the sex hormone deprived surgical menopause model to achieve the experimental endpoints in this study.

The University of Alaska Fairbanks Institutional Animal Care and Use Committee approved the animal care and experimental procedures (IACUC assurance numbers 568518 and 631126).

In the current study we showed that acute OVX for 7 days resulted in a significant increase in the compulsive-like nesting and marble burying behaviors of BIG1 and BIG2 female mice. No increase in nesting and marble burying was observed for the Control and SMALL strains, which shows the specificity of the OVX effects for the compulsive-like condition. The exacerbations in compulsive-like behaviors in BIG mice were attenuated by acute subcutaneous administration of E2, but not P4. Human studies have shown that gonadal steroids trigger or precipitate mood disorders in women with a history of an already existing disease condition when compared to women without it (Hay et al., 1994; Schmidt et al., 1998; Clayton and Ninan, 2010). Onset and exacerbation of OCD associated with pregnancy and postpartum has been shown in human studies (Neziroglu et al., 1992; Williams and Koran, 1997; Labad et al., 2005; Uguz et al., 2007; Forray et al., 2010) establishing a strong link between reproductive events and OCD. However, there is lack of literature on how obsessions and compulsions and associated affective and cognitive behaviors vary during induced menopause. According to a review (Forray et al., 2010) a large variation exists in human studies on onset and exacerbation of OCD during reproductive events and one of the contributing factors could be innate differences in patient populations. In congruence with this we found that BIG1 sham females had higher nesting scores when compared to BIG2 sham females. This variation in compulsive-like nesting behavior was also seen post acute OVX but was abolished in the E2 treatment regimen. What is more intriguing is the fact that BIG1 and BIG2 females did not exhibit variation in compulsive-like marble burying. This is an interesting finding indicating heterogeneity in the BIG strains based on compulsive-like traits and genetic background as often seen in subgroups of OCD patients (Fontenelle et al., 2005; Grados and Riddle, 2008; Leckman et al., 2009). Whether E2 might be more effective compared to P4 in reducing OCD symptoms in postmenopausal females, as seen in our OVX compulsive-like mice, remains to be elucidated.

A prior study has shown that acute E2 administration along with P4 to OVX rats reduced compulsive perseverance in the T-maze (Fernández-Guasti et al., 2006). This is similar to our findings in the compulsive-like mice though, E2 alone had an attenuating effect in the OVX state. Interestingly, previous findings show that acute P4 administration reduced compulsive-like marble burying behavior in male rats (Umathe et al., 2009). However, in the current study we did not see an anti-compulsive effect of P4 treatment in BIG mice. This could be due to various factors, including using mouse strains and females in our study compared to male rats in the (Umathe et al., 2009) study.

In the anxiety-like measures, the OVX BIG strains spent less time in the center of the open field and also explored the open arm less in elevated plus maze when compared to the sham groups. The SMALL and the Control OVX strains showed no significant changes in open field and elevated plus maze when compared to their sham counterparts. Therefore, OVX worsened anxiety-like behaviors in compulsive-like condition specifically. Acute administration of E2 resulted in increased time spent in the central square of the open field in BIG1, but not BIG2 strains, indicating a strain dependent effect in the E2 treatment response. The acute dosage and the time frame of administration of E2 in our study previously also showed anxiolytic effects in the open field in acutely OVX mice (Walf et al., 2008b) and rats (Walf and Frye, 2009). In the plus maze no significant effect of E2 was observed in the compulsive-like mice, which could be due to the fact that the open field and the elevated plus maze tests measure different aspects of emotionality associated with anxiety (Ramos, 2008; Anchan et al., 2014).

P4 treatment, on the other hand, decreased anxiety-like behavior in open field for both BIG1 and BIG2 strains. For the elevated plus maze, however, P4 was effective only in the BIG2 and not the BIG1 strain. These results indicate a strain specific response to anxiety-like behavior due to P4 administration. Strain specific effects of E2 and P4 on behavioral responses have been sparsely explored in rodent studies. Only one study has shown significant strain specific effects of E2 on depressive-like forced swim behavior (Koss et al., 2012). Behavioral responses to alterations in gonadal steroids have been found to vary in women with and without premenstrual syndrome (Schmidt et al., 1998), which might be similar to the differences in behavioral responses to E2 and P4 in the BIG1 and BIG2 strains.

The association of memory impairment with OCD is not clear. Many clinical studies have failed to find any evidence that OCD is associated with memory deficits (McDonald, 1991; Dirson et al., 1995; Radomsky and Rachman, 1999). In addition, no impairment in declarative and short-term memory has been found in OCD patients compared to normal controls (Roth et al., 2004; Demeter et al., 2013). However, many others have reported working memory impairments in patients (Martin et al., 1995; Nakao et al., 2009). Our BIG mice showed a larger object recognition memory deficit than the Control mice, both in the sham and OVX groups. However, the SMALL mice also showed a similar memory deficit and, therefore, whether this memory deficit in the BIG mice was due to a genetic correlation between compulsive-like behaviors and object recognition memory, or was due to genetic differences between the two strains caused by founder effects or random drift (Bult and Lynch, 1996, 2000), remains to be elucidated. Also as the BIG and SMALL strains had a RI below 0.5 they appeared to avoid the new object, while the control mice had an index above 0.5 and appeared to favor the new object.

Acute OVX did not result in significant object recognition memory impairment in the compulsive-like condition. Contradictory evidence exists as to whether acute OVX leads to object recognition memory impairment in normal mice (c57 strain). While few studies have shown memory loss during acute OVX (Gresack and Frick, 2006; Rhodes and Frye, 2006), others show the opposite (Willard et al., 2011; Fonseca et al., 2013; Bastos et al., 2015). However, chronic OVX has consistently caused poor performance in the novel object recognition test (Fonseca et al., 2013; Bastos et al., 2015). Though there was no overall effect in novel object recognition memory among sham and OVX strains, E2 treatment improved the RI in only the BIG2 but not the BIG1 strain. P4 treatment however improved the RI in both the strains. In the current context of investigation object recognition was performed to evaluate short term or working memory impairments associated with OCD in the mouse strains. However, object recognition encompasses just one aspect of assessing otherwise very complex memory consolidation and cognition process in rodents. In future studies we aim to include a more robust assessment of both short-term and spatial memory components.

The current study supports a complex interplay of genetic background and sex steroids during acute ovarian dysfunction in the compulsive-like condition. We report exacerbation of compulsive-like behaviors with trait specific intra-strain variation during acute ovarian failure in the spontaneously compulsive-like mouse strains only, which was restored by E2 and not P4. This effect was similar for the spontaneously compulsive-like BIG1 and BIG2 strains, unlike the associated anxiety-like and cognitive-like behaviors, which displayed differences between the BIG1 and BIG2 strains for both E2 and P4 effects on these behaviors. We therefore hypothesize that the associated comorbidities in the surgical menopause state in the OCD condition might vary among individuals because of genetic differences. Future studies will focus on investigating effects of E2 and P4 on the chronic OVX state and also the potential signaling pathways in the brain of compulsive-like mice. Overall, the results presented here strengthen the face, predictive and construct validities of the mouse model for investigating heterogeneity associated with OCD during ovarian failures in females.

SM, CPB and KB conducted all experiments and performed data analysis. SM and CPB lead manuscript writing efforts. AB-I and GSP made significant contributions to research design, data interpretations and manuscript preparation.

Research reported in this publication was supported by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number P20GM103395 to SM. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the NIH. The work was also supported by College of Natural Sciences and Mathematics (CNSM) and the Office of the Vice-Chancellor for Research to SM and AB-I. CPB received a fellowship from CAPES/Brazil for this project. These funding sources did not have a role in the study design, data collection, analysis, interpretation and submission of this article for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.