Sex differences in cognition following variations in endocrine status

Abstract

Spatial memory, mediated primarily by the hippocampus, is responsible for orientation in space and retrieval of information regarding location of objects and places in an animal's environment. Since the hippocampus is dense with steroid hormone receptors and is capable of robust neuroplasticity, it is not surprising that changes in spatial memory performance occur following a variety of endocrine alterations. Here, we review cognitive changes in both spatial and nonspatial memory tasks following manipulations of the hypothalamic–pituitary–adrenal and gonadal axes and after exposure to endocrine disruptors in rodents. Chronic stress impairs male performance on numerous behavioral cognitive tasks and enhances or does not impact female cognitive function. Sex-dependent changes in cognition following stress are influenced by both organizational and activational effects of estrogen and vary depending on the developmental age of the stress exposure, but responses to gonadal hormones in adulthood are more similar than different in the sexes. Also discussed are possible underlying neural mechanisms for these steroid hormone-dependent, cognitive effects. Bisphenol A (BPA), an endocrine disruptor, given at low levels during adolescent development, impairs spatial memory in adolescent male and female rats and object recognition memory in adulthood. BPA's negative effects on memory may be mediated through alterations in dendritic spine density in areas that mediate these cognitive tasks. In summary, this review discusses the evidence that endocrine status of an animal (presence or absence of stress hormones, gonadal hormones, or endocrine disruptors) impacts cognitive function and, at times, in a sex-specific manner.

Broadly, cognitive function refers to the information processing ability of an animal. Cognition includes all aspects of knowledge, including attention and the ability to learn, retain, and recall information. This review focuses on the basic cognitive functions of acquisition (learning) and/or memory (information recall) primarily in rodents in both spatial and nonspatial tasks, with special attention to sex differences and sex-specific alterations in learning and/or memory following experiences such as exposure to stress, gonadal hormones, or endocrine disruptors. Additionally, because experiential effects on cognition vary as a function of age at time of exposure and gonadal state (Bowman 2005; Bowman et al. 2004, 2006), this review includes effects of adrenal and gonadal hormones on spatial and nonspatial memory at various development times of exposure and assessment.

Spatial and nonspatial memory: behavioral assessment and sex differences

Spatial memory, critical to species survival, is responsible for orientation in space and retrieving information about the locations of objects and places in the environment (Eichenbaum 2017; Chen et al. 2020). Spatial memory is the most widely assessed form of memory in rodents, and a variety of tasks have been developed to measure it, including the Morris water maze (MWM), eight-arm radial maze (RAM), radial arm water maze (RAWM) object placement (OP), and the Y-maze. Behavioral assessments of spatial memory are dependent on an intact hippocampus and also require interaction with the prefrontal cortex (PFC) (Luine 2015). In addition, the tasks rely on the innate ability of rodents to know and defend a territory by using salient environmental landmarks to establish a cognitive map that resides within hippocampal CA1 networks (Mumby et al. 2002). In the RAM, subjects receive a food reward at the end of arms (eight to 17 arms) that radiate out from a central, round hub. Completing the RAM task takes advantage of natural foraging strategies of rats, and the ability to complete the task without re-entering arms previously visited depends on building a cognitive map of the cues in the area around the maze. This is termed reference memory, while remembering the arms entered using those cues is working memory.

Recognition memory tasks, developed in the 1990s, have been advantageously applied to further investigate memory processes and to determine effects of drugs and hormones (Ennaceur and Meliani 1992). These tasks use the natural, novelty-seeking exploratory nature of rodents and thereby mitigate possible confounding influences of task requirements, experience, reinforcements (positive or negative, as occur in radial arm and water mazes), or psychological performance variables. Thus, recognition memory tasks were an important addition to learning and memory testing repertoires. First, subjects receive a sample or training trial where they explore two identical objects, and after an intertrial delay ranging from 1 to 24 h, subjects are returned to the arena for a test trial (see Fig. 1). In the testing or retention trial, subjects view the same objects, but one of the objects either is moved to a new location (place memory [OP]) or is replaced with a new or novel object (object memory or object recognition [OR]) in the retention trial. Both object and place memory rely on the hippocampus and the PFC, but memory for objects (OR task) relies less on the hippocampus and more on the PFC than place memory, while the opposite relationship is true for place memory (OP task) (Luine 2015; Luine et al. 2017). However, this anatomical distinction may be a question of degree, as there are interconnections between the hippocampus and PFC. Direct projections from the hippocampus to the PFC have been demonstrated (Churchwell et al. 2010; Churchwell and Kesner 2011), but the reciprocal projection has not been shown. Rather, it was recently demonstrated that both the hippocampus and PFC project to the paraventricular nucleus of the thalamus, which in turn projects to both the hippocampus and PFC (Viena et al. 2022). Therefore, the two areas critical to both spatial and nonspatial memory are connected both directly and indirectly.

Figure 1.

The middle box shows a rat during the training or sample trial where two identical objects are explored, typically for 3–10 min, or until a set exploration time is obtained. An intertrial delay is given (depicted by red arrows), typically 1–24 h, and then, shown at the left and right, the retention or testing trial is given (usually for 3–5 min). The left box shows OP, when one of the objects is moved to a new location. The right box shows OR, when an identical object is replaced with a new object. Chronic treatments can be given pretraining (hours to days) or immediately posttraining to assess rapid changes in memory. Posttraining injections test effects of agents on memory consolidation. Data can be analyzed as time at old and new locations where significantly more time spent at the new location or object indicates the formation or consolidation of memory during the intertrial interval (see Fig. 2). Results can also be presented as discrimination ratio (percentage time at new location or object), where ratios of 50% indicate chance performance and higher ratios indicate formation of memory for the object or location (see Fig. 3).

Males perform better than females on spatial learning and working memory tasks (for review, see Jonasson 2005). There is general consensus that sex differences in learning and memory are due to navigational strategies used rather than an innate male behavioral advantage and, with practice, females can perform as well as males (Luine et al. 2017; Fertan et al. 2019; Yagi and Galea 2019). In recent years, our laboratories have shifted away from the traditional spatial memory task such as the RAM due to the long training/learning time involved (i.e., typically at least 10 training sessions over 5 d, during which many experiential alterations, such as stress effects, begin to attenuate) to the OP and OR tasks, which require no training per se. We have shown sex differences in performance in OP, but not OR. For example, on the OP task, control male rats can discriminate between the old and new object locations following 2.5- and 4-h intertrial delays, while females cannot (Beck and Luine 2002; Bisagno et al. 2003a).

It is also important to note that sex differences in hippocampal structure in rodents (McLaughlin et al. 2009; McEwen and Milner 2017; Scharfman and MacLusky 2017; Brandt et al. 2020) and humans (van Eijk et al. 2020) have been reported. In addition, there are sex differences in neurogenesis in the hippocampus of rats (Duarte-Guterman et al. 2015), and the effects of estradiol on long-term potentiation in the hippocampus are sexually dimorphic (Gall et al. 2021). Therefore, there are numerous differences between the sexes that may impact the circuits involved in mediating sex differences in cognition.

Spatial and nonspatial memory: sex-specific responses to chronic restraint stress

Stress is a common experience in the lives of all animals across the life span, and the term stress generally refers to any perceived threat to homeostasis. Stressful experiences trigger a sequence of highly coordinated physiological responses, the fight or flight reaction, mediated by the hypothalamic–pituitary–adrenal axis (Chowdari et al. 2002), which ultimately releases glucocorticoids (GCs), primarily corticosterone (CORT) (González-Ramírez et al. 2014), in rodents and cortisol in primates from the adrenal glands (Dallman 2007; Leistner and Menke 2020). The release of CORT from the adrenal glands in response to stress is sexually differentiated (Galea et al. 1997; Figueiredo et al. 2002; Bowman et al. 2004). Following GC release, adaptive or maladaptive physiological changes occur, and this distinction has largely been attributed to the nature (i.e., intensity and duration) of the perceived stressor (Selye 1976). More recently, the concept of allostasis and allostatic load has been used to describe this stress relationship (McEwen and Stellar 1993; McEwen 1998). In essence, acute or short-term stress in males enhances cognition as an effective coping mechanism, while continued stress imparts too large an allostatic load on the organism and stress becomes debilitating to both peripheral and central nervous systems, resulting in impaired learning and memory and other deleterious outcomes (McEwen 2016). Recent studies show that neural functions such as cognition/memory also follow this temporal pattern of response but are highly dependent on the sex, age, and gonadal hormone status of the subjects (Bowman 2005).

The effect of stress exposure on cognitive responses is well established in male rats. Chronic stress (using a variety of paradigms, including days to weeks of daily restraint, immobilization, loud noise, lights, cold water, or a combination of different stresses) generally results in impaired learning and memory. On the RAM, stressed male rats make more errors per trial and take longer to reach a learning criterion than unstressed rats (Luine and Rodriguez 1994; Luine et al. 1994). A similar pattern of impaired learning in rats is shown after chronic stressors in the MWM (Kitraki et al. 2004a,b) as well as on the water maze version of the RAM (Ortiz et al. 2015; Peay et al. 2020). In addition to these stress-induced male learning impairments, spatial memory is also impaired in males. Following chronic stress in male rodents, memory in OP (Beck and Luine 1999; Bowman et al. 2009; Gomez et al. 2012) and the Y-maze task (Conrad et al. 2003; Wright and Conrad 2005; Gomez et al. 2012; Woo et al. 2018; Peay et al. 2020) is impaired. See Table 1 for a summary of the effects of chronic stress on spatial learning and memory in male rodents.

Table 1.

Effects of chronic stress on learning and memory function by sex

Additionally, stressed male rats and mice have impaired OR memory (nonspatial memory), as they are unable to significantly discriminate between known and new objects (Beck and Luine 1999, 2002; Peay et al. 2020) in a test of temporal order recognition memory (Wei et al. 2014). Thus, chronic stress impairs learning and memory in male rodents in a variety of tasks. Subjects recover from impairments if stress is not maintained (Ortiz et al. 2015; Luine et al. 2017; Ortiz and Conrad 2018). See Table 1 for a summary of these results, and see Conrad (2010) and Luine et al. (2017) for in-depth reviews.

Because estrogen levels are higher in females than males and estrogens exhibit neurotrophic, antioxidant, and antiapoptotic effects, and also promote some aspects of cognitive function (Luine 2016), we hypothesized that female rodents may be less sensitive than males to the impairing effects of chronic stress on cognition. Indeed, when female rodents are given the same stress regimens and tested on the same tasks as males, a different pattern emerges. Our laboratory was the first to show that chronic restraint stress (CRS) enhanced performance of female rats on the RAM (Bowman et al. 2001). Also, female performance was enhanced in the MWM following chronic stress (Kitraki et al. 2004b). In recognition memory tasks, stressed females show better performance than unstressed females in OP (Beck and Luine 1999; Bowman et al. 2009), and memory is either not affected or is enhanced on the Y-maze (Conrad et al. 2003; Gomez and Luine 2014), in OR (Beck and Luine 2002; Bisagno et al. 2004; Bowman et al. 2009; Gomez and Luine 2014), and on temporal order recognition memory tests (Wei et al. 2014). A recent study showed this same pattern of sex differences in stress effects on memory using chronic unpredictable stress (Ortiz et al. 2015; Peay et al. 2020). Based on the theoretical frameworks of Selye (1976) and McEwen (1998), we hypothesized that females may take longer than males to transition from adaptive to maladaptive stress effects (Bowman et al. 2003). However, impairments in RAM, OR, and OP were not found in female rats that were stressed for 28 d (Bowman et al. 2002) or 35 d (Bowman and Kelly 2012). A summary of these sex-dependent stress effects is shown in Table 1.

Thus, it appears that female rodents, as compared with males, show resilience to chronic stress in terms of learning and memory. While it is possible that stress effects or sex differences in response to stress are mediated via nonmnemonic processes, a detailed review of these factors is beyond the scope of this review, but see Luine et al. (2017) and Luine et al. (2018) for discussion. Overall, the preponderance of current evidence suggests that CRS causes changes in mnemonic factors. It is important to note that while CRS has provided an excellent model for the understanding of the effects of stress on cognition, there is a pressing need to better understand the context of the stress itself and subsequent consequences. Studies investigating different stressors, such as chronic social isolation or overcrowding, are becoming increasingly popular as translatable models that offer the promise of increased understanding of human stress-related pathology (Mumtaz et al. 2018).

Another important aspect of stress research that needs further investigation is the impact of stress on early development. Developmental stress exerts deleterious effects on behaviors both during development and extending into adulthood. However, the bulk of such studies has assessed anxiety and depressive-like behaviors, not cognition, and few have included females. For example, stressing adolescent male rats increased anxiety in both adolescent and adult subjects and decreased performance on the MWM in adolescence but not adulthood (Avital and Richter-Levin 2005). Male mice stressed during adolescence show decreased Y-maze and MWM, but not OR, performance at adulthood (Sterlemann et al. 2010). Both sexes were examined following chronic variable adolescent stress, and both showed depressive-like symptoms at adulthood (Bourke and Neigh 2011). A recent study of female adolescent stress showed impairments in cognitive flexibility (Hyer et al. 2021). Thus, at this time, it is unclear whether cognitive resilience to stress exhibited by adult females is also present in adolescent females, and further research is required.

Accounting for female resistance to stress: the role of estradiol

The interaction between stress and the estrous cycle is unclear, with some studies in the literature finding an effect following stress (do Nascimento et al. 2019; Hokenson et al. 2021) and others not (Conrad et al. 2004; ter Horst et al. 2013). A discussion of the influence of estrous cycle on the stressed-induced changes in cognition is beyond the scope of this review (see ter Horst et al. 2012).

Several lines of evidence suggest that estradiol, of both ovarian and neural origin, may be responsible for female rodents’ cognitive resilience to chronic stress. First, females have higher circulating estradiol (E₂) levels as compared with males, and estrogens exert neurotrophic effects on the brain. Stress also activates the hypothalamic–pituitary–gonadal (HPG) axis, causing release of E₂ from the ovaries, which may mitigate the effects of stress (Shors et al. 1999; Liu et al. 2011; Lennartsson et al. 2012). Consistent with this evidence, we hypothesized that if higher circulating levels of E₂ in adult females, compared with males, is responsible for the female resistance to deleterious effects of stress on cognitive function, then ovariectomized (OVX) females would be more sensitive to (impaired similar to males) or respond differently to chronic stress than intact controls. In support of the hypothesis, chronically stressed, OVX females were no longer enhanced on the RAM but they were not impaired, and E₂ treatment restored stress-induced enhancements (Bowman et al. 2002). However, McLaughlin et al. (2009) reported that chronically stressed OVX female rats still showed improvements on the Y-maze as compared with nonstressed subjects following OVX. When these experiments were completed in the early 2000s, it was hypothesized that resilience to stress in females was due to the activating effects of E₂ on cognitive function and neurotrophic changes in neurons. However, more recent experiments showed that significant amounts of E₂ are synthesized locally in discrete regions of the brain by the enzyme aromatase from androgen precursors such as testosterone (T) and directly from cholesterol, and, as such, levels in the hippocampus, hypothalamus, and PFC are higher than present in the circulation (for review, see Yuen et al. 2016). In addition, E₂ is still present in the hippocampus following OVX (Kato et al. 2013). Thus, the in situ production of E₂ suggests that neural E₂ can contribute to cognitive stress resilience in females such that E₂ is available in the brain even after OVX. Hence, intrahippocampal E₂ synthesis is well poised to contribute to cognitive stress resilience in females, though contributions of peripheral estrogens cannot be discounted. Further evidence that the resilience of females to stress depends on E₂ was provided by Wei et al. (2014), who explored stress effects in the temporal order recognition (TORM) task, an explicit working memory task relying on the medial PFC and similar to OR. CRS impairs males, but not females, in the TORM task. Interestingly, impairments of performance in stressed males were prevented by E₂ administration, and stressed females were impaired following aromatase inhibition, suggesting that peripheral, possibly gonadally derived, E₂ may also contribute to resilience to stress.

Stress effects on spatial and nonspatial memory in aged rats

Additional evidence for the role of E₂ in cognitive sex differences in response to stress comes from data in aged rats. When 21-mo-old male and female rats were exposed to CRS, the performance on memory tasks differed from that of young adult rats, and no sex differences were observed in the aged rats. Interestingly, spatial memory sex differences on the OP task were no longer present in aged unstressed males and females, and stress did not alter performance in either sex (Bowman et al. 2006). The maladaptive effects of CRS observed in young male rats are no longer present and the adaptive effects of CRS on spatial memory in young female rats are lost in aged females. For OR, both stressed males and females performed better than nonstressed controls. This result is in contrast to stress in young adulthood, which impairs male, but does not alter female, OR performance. While few studies have considered the effects of aging on the stress response, these data suggest that chronic stress was beneficial/adaptive in aged rats because they performed some tasks better. The age-dependent changes are more pronounced in males than females, as the deleterious effects of CRS on memory observed in young adult males are absent in aged males. Since circulating E₂ is higher in the old males versus females, E₂ may also contribute to these age-related changes in responsivity (Bowman et al. 2006).

Sex, stress, and neuronal changes

The hippocampus contains high levels of CORT receptors, which can mediate rapid nongenomic and slow gene-mediated neuronal actions (Joëls 2018). Thus, chronic stress, with concomitant CORT release in rodents, is associated with sex-specific changes in neuronal structures and chemicals mediating cognition.

One possible neurochemical that may mediate sex-dependent behavioral stress effects is brain-derived neurotrophic factor (BDNF), as this neurotrophin is present in high amounts in the hippocampus and promotes memory consolidation through synaptic plasticity (Scharfman and Maclusky 2005; Notaras and van den Buuse 2020). Hippocampal BDNF levels are reversibly decreased in male rats following chronic stress exposure (Luo et al. 2004), while hippocampal infusions of BDNF protect against spatial memory impairments on the MWM (Radecki et al. 2005). Taken together, it appears that lower BDNF levels in males may contribute to their stress impairments in spatial memory. In addition, the fact that E₂ increases hippocampal BDNF levels (Blurton-Jones et al. 2004; Zhou et al. 2005) may account for the sex-specific stress enhancements observed in females.

Monoaminergic neurotransmitters are sensitive to both acute and chronic stress exposure, and sex-specific alterations in monoamines may contribute to the sex-specific stress effects. Decreased norepinephrine, dopamine, and serotonin levels are observed in the hippocampi of chronically stressed male rodents (Beck and Luine 1999, 2002; Sunanda and Raju 2000; Torres et al. 2002; Hasegawa et al. 2018). However, in female rats, serotonin and norepinephrine levels are increased following chronic stress exposure (Beck and Luine 1999, 2002; Bowman et al. 2002, 2003). In addition to the hippocampus, dopaminergic activity is decreased in male rats and increased in females in the medial PFC following chronic stress (Beck and Luine 2002; Luine 2002; Bowman et al. 2003), although this effect can depend on the age of stress and time of testing (for review, see Perry et al. 2021).

Stress-induced structural remodeling in the CA3 region of the male rat hippocampus is well documented (Watanabe et al. 1992b; Ortiz and Conrad 2018). Following 21 d of CRS, structural changes in the CA3 region of the male rat hippocampus include decreases in both apical dendritic branching and total dendritic length (Watanabe et al. 1992b; Galea et al. 1997). In contrast, stressed female rats do not show atrophy of apical dendritic branches. However, a significant decrease in the number of branch points within the basal dendritic area is observed (Galea et al. 1997). The atrophy is mediated by GC levels and N-methyl-D-aspartate receptor-mediated excitatory input (Watanabe et al. 1992a,b) and can be prevented by inhibiting GC secretion. Phenytion administration blocks the stress-induced atrophy of CA3 dendrites (Watanabe et al. 1992a) and leads to reversal of RAM impairments (Luine et al. 1994). Furthermore, similar to the behavioral effects, the stress-induced hippocampal alterations are temporally constrained, and stress-induced atrophy attenuates by 5–10 d after stress (McLaughlin et al. 2009). McLaughlin et al. (2009) showed that E₂ treatment to OVX-stressed female rats prevented CA3 dendritic retractions and increased CA1 spine density. These results show that stress alters hippocampal neurons and provides morphological evidence that the female hormone E₂ confers cognitive resilience to stress in females.

Implications of sex differences in stress responses for mental health and disease

The sex differences in behavior and neural responses to stress described here may contribute to well-described sex differences in humans for coping with chronic stress and in the sexually divergent incidence of stress-related diseases. For example, women have a higher incidence of anxiety disorders, posttraumatic stress disorder, and major depression while they have a lower incidence of alcohol and drug abuse, and all of these disorders can be precipitated by stress (Bangasser and Valentino 2014). Moreover, sex differences are also present in incidence and progression of neuropsychiatric disorders like schizophrenia and in neurodegenerative diseases like Alzheimer's and Parkinson's (Gillies and McArthur 2010). An understanding of the neural underpinnings of sex differences in stress responses may lead to more effective treatments for these disorders. In support of this idea, the gender medicine movement is gaining momentum and posits that we need to do away with the assumption that men and women are similarly affected by diseases and treatments, and it advocates for sex-based diagnoses and treatments (Mauvais-Jarvis et al. 2020).

Cognitive responses to gonadal hormones show less sexual specificity than adrenal hormones

In contrast to the robust sexually dimorphic effects of adrenal steroids on cognition in rodents, responses to gonadal hormones generally do not differ greatly between males and females. E₂, the female hormone, has a long history of cognitive study in rodents and humans (Luine 2014; Luine and Frankfurt 2020), and it enhances performance of a wide variety of learning and memory tasks in OVX females. While effects of estrogens in orchidectomized (ORX) males have not been widely studied, enhancements have been found after chronic treatments on the RAM (Luine and Rodriguez 1994), Barnes maze (Locklear and Kritzer 2014), and T-maze (Gibbs 2005) but not in OR (Aubele et al. 2008). Given immediately after T1, E₂ also enhances OP in ORX males (see Fig. 2; Jacome et al. 2016). Thus, the effects on learning and memory by estrogens in males mimic changes seen following E₂ treatments to females.

Figure 2.

Effects of gonadal hormones on recognition memory in male and female rats. A and B show results of OP testing in male rats. ORX male rats were given vehicle (Veh, corn oil), 750 µg/kg testosterone (T), or 20 µg/kg estradiol (E₂) S.C. immediately after the training trial, and place recognition was tested 2 h later. Entries are mean ± SEM. (**) P < 0.01, by paired t-test. Vehicle-injected ORX males spent the same amount of time exploring objects in the old and new locations, suggesting poor memory, but when either T or E₂ was given, subjects spent more time exploring at the new location, which suggests better memory. A and B are adapted with permission from Jacome et al. (2016) by permission of Oxford University Press. C and D show results of OR testing in female rats. OVX rats received two subcutaneous injections of the vehicle (sesame oil), 500 µg of testosterone propionate (TP), or 50 µg/kg estradiol benzoate (EB). Forty-eight hours later, subjects were tested for OR. Entries are mean ± SEM. (***) P < 0.001, by paired t-test. C is adapted with permission of Elsevier from Luine et al. (2022); permission conveyed through Copyright Clarance Center, Inc. Vehicle-injected OVX rats spent the same amount of time exploring old and new, suggesting poor memory, but when either TP or EB was given, subjects spent more time exploring the new object, which suggests better memory. Data in D redrawn from Jacome et al. (2010).

In addition, the principal male hormone (T) and other androgens enhance learning and memory in gonadectomized male and female rodents. Castration (CAS) impairs acquisition (Gibbs 2005; Locklear and Kritzer 2014) and working memory (Gibbs and Johnson 2008) on the RAM, decreases spatial recognition memory on the Y-maze (Hawley et al. 2013), and impairs memory on the novel object (Aubele et al. 2008) and OP (McConnell et al. 2012) tasks. Chronic treatment of CAS males with T restores novel object memory (Aubele et al. 2008), OP memory (McConnell et al. 2012), retention deficits on the spatial Barnes maze (Locklear and Kritzer 2014), spatial recognition memory on the Y-maze (Hawley et al. 2013), working memory errors on the RAM (Spritzer et al. 2011), and working memory on the MWM, when configured as a delayed match to place task (Sandstrom et al. 2006). Acute T treatment for 4 h also enhances OP in CAS males (Jacome et al. 2016).

Effects of androgens in females have been less investigated. However, T or dihydrotestosterone (DHT) treatment increased the percentage of correct choices in the Y-maze 1 h later, and percentage of time exploring novel objects in an OR task 24 h later, compared with OVX females (Frye and Lacey 2001). Another androgen, dihydroepiandosterone (DHEA), administered to OVX rats immediately after posttraining trials enhanced spatial memory on the MWM 24 h later but did not alter OR (Frye and Lacey 1999). Recently, we found that a 2-d treatment with T enhanced OR (Fig. 2C), but not OP, and DHT and DHEA treatments enhanced OP in OVX female rats (Luine et al. 2022). Thus, androgens appear to enhance memory in OVX female rats, both acutely and chronically, in the few studies that have been completed. Figure 2 shows that castrated males respond to E₂ or T with enhanced place memory, and OVX females respond to E₂ or T with enhanced object memory. Further investigations are clearly warranted, but thus far, the accumulated data suggest that cognitive responses to gonadal hormones are more similar than different in male and female rodents.

Developmental programming of adult cognition by gonadal hormones

Influences of gonadal hormones on cognition begin in utero when the testes of males actively secrete T, which directly masculinizes the male external genitalia. In the central nervous system, T is aromatized to E₂, and it is this locally produced E₂ that programs the expression of male-like behaviors, including sexual behavior. E₂ also underlies the expression of some nonsexual behaviors like aggression, activity, and motor abilities as well as some aspects of learning and memory (Gillies and McArthur 2010; Watson et al. 2010). The organizational events induced by E₂ leave long-lasting imprints that are manifested in sex differences at adulthood.

Sex behaviors are strikingly different between the sexes, but sex differences in nonsexual behaviors, including cognition, are usually smaller. However, male rodents outperform females on tasks requiring spatial memory, RAM (Williams et al. 1990; Luine and Rodriguez 1994), MWM, and OP (Beck and Luine 2002; Bisagno et al. 2003b). It should be noted that with further training and practice, females can perform RAM at the same level as males (Williams and Meck 1991). Fewer studies have examined memory tasks that are not dependent on the hippocampus, but Wood and Shors (1998) investigated the classical conditioning paradigms of eyeblink response, and females acquire this task in fewer trials than males. Equally, in active avoidance tests, intact or gonadectomized females acquired the behavior in fewer trials than gonadectomized males (van Haaren et al. 1990).

Sex differences are also prominent in the strategies by which cognitive tasks are accomplished in rodents. For example, in learning to obtain rewards on the T-maze, two strategies can be used: a response or place strategy. Korol et al. (2004) reported that female rats were more likely to use a place strategy (but could alter strategy based on estrus cycle state), while Packard and McGaugh (1996) reported that 90% of male rats use a response strategy. In solving the spatial memory task, RAM, female and male rats also often apply different strategies that are reflected in males having a faster rate and accuracy in learning the task, but over time both sexes reach the same level of performance (Williams and Meck 1991; Tropp and Markus 2001). Similar strategy differences between the sexes have also been noted in the water maze (Daniel and Lee 2004; Korol et al. 2004). Little research has directly assessed developmental effects of hormones, but Williams and Meck (1991) showed that neonatal CAS of males and E₂ treatment to females mimicked the demonstrated sex differences in adult rats. Neonatal exposure to T also modifies sex differences in RAM and MWM performance and hippocampal morphology (Roof 1993), and neonatal treatment of females with androgens reversed the female advantage in active avoidance (van Haaren et al. 1990). Thus, research suggests a strong organizing effect of estradiol, derived from testosterone, on adult sex differences in the acquisition and performance of predominantly spatial, but other, memory tasks.

Gonadal hormones and neuronal changes

There is a well-documented interaction between memory, estrogen, and dendritic spine density in both the hippocampus and PFC (Luine 2014; Frankfurt and Luine 2015) in adult animals. When adult female rats are OVX, both memory performance and dendritic spine density in CA1 and the mPFC decline (Wallace et al. 2006). An E₂-induced increase in OP is accompanied by an increase in dendritic spine density in CA1 in OVX rats (Luine and Frankfurt 2013) and mice (Li et al. 2004). Enhanced performance on the Y-maze is also associated with increased dendritic spine density in the mPFC following E₂ (Velazquez-Zamora et al. 2012). Aged female Fischer 344 rats have memory impairments that are accompanied by decreased circulating estrogen and a marked decrease in dendritic spine density in pyramidal cells in both CA1 (Luine et al. 2011) and the mPFC (Wallace et al. 2007). Last, acute E₂, 30-min treatment increases CA1 and mPFC, but not dentate gyrus (DG), dendritic spine density in OVX rats (Tuscher et al. 2016; Luine et al. 2018).

Few studies have examined spine changes following gonadal hormone treatment in male rats, but acute E₂, 2-h treatment to CAS rats is associated with enhanced OP memory in CAS male rats and increased CA1, but not DG, dendritic spine density (Jacome et al. 2016; Avila et al. 2017). In the same study, acute T given to gonadectomized male rats was also associated with an increase in CA1, but not DG, spine density and enhanced OP. Interestingly, in a study where T effects on spine density were studied in different hippocampal sublayers, T increased spine density in the strata radiatum (SR) but not in the stratum lacunosum moleculare (SLM) layer of CA1 in castrated adult mice (Li et al. 2012), and chronic T treatment to SAMP8 mice (Alzheimer's disease model) also increased CA1 spine density (Jia et al. 2016). Although information is limited, androgens alter spine density in female rodents. Both DHEA and TP, given subchronically, enhance recognition memory and increase spine density on pyramidal cells in CA1 and the PFC in OVX rats (Luine et al. 2022).

These neuronal changes appear to be mediated by receptors for estrogens and androgens, since they are abundant in the hippocampus and PFC. Both nuclear receptors, which mediate long-term changes, and extranuclear/membrane receptors, which mediate rapid effects for estrogens (Torres-Revereron et al. 2020) and androgens (Sar et al. 1990; Simerly et al. 1990; Clancy et al. 1992; Kritzer 2004), are present in both sexes. Surprisingly, no major differences in the distribution of gonadal hormone receptors between males and females have been reported (Simerly et al. 1990; Weiland et al. 1997), which may account for the positive effects of male and female hormones on memory and spine density in both sexes. However, further verification and expansion of these results are clearly warranted.

Gonadal hormones in health and disease

Because women undergo menopause with the associated drastically lowered estradiol and concomitant symptoms of osteoporosis, hot flashes, cognitive loss, and mood changes, hormone replacement, without cancerous side effects in the breasts and uterus, has long been sought. While this topic is beyond the scope of the current review, it is notable that selective estrogen receptor modulators (SERMs) are being developed and hold the potential for treating certain cancers, as well as endometriosis, inflammatory diseases like rheumatoid arthritis, and cardiovascular and CNS conditions (Minutolo et al. 2011). Since the data presented here indicate that the cognitive responses of male rodents to estrogens are similar to responses of female rodents, the possibility of men also being treated by SERMs is raised. In addition, a novel phytoestrogen-based (plant estrogens) drug, referred to as the phyto-β-SERM formulation, prolonged survival, improved spatial recognition memory, and slowed progression of amyloid pathology in a female mouse model of Alzheimer's disease (Zhao et al. 2013). In a recent trial in women, this formulation was well tolerated, and no safety issues were raised, which suggests moving to a phase 2 efficacy trial (Schneider et al. 2019). Since male rodents responded to estradiol similarly to female rodents in spatial memory tasks (Fig. 2), treatments could be expanded to men.

Sex-specific effects of endocrine disruptors on spatial and nonspatial memory

In addition to stress, our laboratories have examined how cognitive function is altered following exposure to bisphenol A (BPA), an endocrine disruptor that has been shown to modulate estrogenic, androgenic, and antiandrogenic effects (https://researchfeatures.com/wp-content/uploads/2021/05/Maya-Frankfurt.pdf). BPA has been extensively used in the manufacturing of hard plastics, and detectable levels of BPA have been reported in body fluids of humans and animals (Biedermann et al. 2010; Geens et al. 2011; Rubin 2011), indicating that BPA exposure is ubiquitous and has potential health hazards (for a review, see Prins et al. 2019).

Most research investigating the effects of BPA in animal models has focused on exposure during the perinatal period. Our studies examined the effects of BPA exposure during adolescence because this developmental period is characterized by profound hormonal changes on the brain and subsequently behavior (Sisk and Romeo 2019). In addition, many studies have used a chronic BPA administration paradigm at doses far greater than what humans encounter during daily living. In contrast, we have used a dose of 40 µg/kg/d given during adolescence (postnatal days 38–49) and measured performance on the OP and OR tasks in male and female rats. This dose is lower than what is considered acceptable by the United States Environmental Protection Agency.

We have shown that following this low-dose adolescent BPA exposure in intact males and females, OP was impaired in BPA-treated animals of both sexes as compared with controls on a 2-h delay trial (Fig. 3A; Diaz Weinstein et al. 2013). In addition, while not reaching significance (P = 0.08, η^² = 0.075), the ability to discriminate between the old and new locations appears more disrupted in BPA-treated males than BPA-treated females (Diaz Weinstein et al. 2013), and this sex-specific BPA trend is consistent with previously reported sex-specific behavioral alterations following low-dose BPA exposure (Gonçalves et al. 2010). Similarly, Xu et al. (2011) reported spatial memory effects from chronic BPA exposure in adolescent mice. Specifically, adolescent males exposed to chronic BPA (40 µg/kg) exhibited impaired MWM performance compared with females (Xu et al. 2011). Additional evidence for sex-specific effects of BPA was shown following prenatal exposure in which the male offspring were impaired on both the novel OR and T-maze, an effect not observed in the female offspring (Thongkorn et al. 2021). Interestingly, as with stress effects on cognitive performance that attenuate with recovery from stress exposure, impairments on spatial memory following adolescent BPA exposure did not persist when OP performance was measured in adulthood (Bowman et al. 2015). However, adolescent BPA exposure did lead to lasting impairments on OR performance when tested in adulthood in males but not females (Fig. 3B; Bowman et al. 2015).

Figure 3.

Effects of adolescent exposure to the endocrine disruptor bisphenol A on object placement and recognition memory. In all panels, rats were exposed to a subcutaneous low-dose injection of BPA (40 µg/kg/d) given across 12 d of adolescent development. All entries are the mean ± SEM. (A) Object placement testing in gonadally intact juvenile male and female rats. The ability to discriminate between the old and new locations was impaired in BPA-treated animals, regardless of sex, compared with control subjects. (**) P < 0.01 by two-way ANOVA. (B) Object recognition testing in intact male and female rats tested in adulthood. The ability to discriminate between the old and new locations was impaired by BPA in males but did not alter female recognition memory. (*) P < 0.05 by two-way ANOVA. (C) Object placement testing in ovariectomized juvenile female rats. BPA-treated OVX females had impaired spatial memory, as evidenced by less time exploring the object in the new location. (*) P < 0.05 by ANOVA. (D) Object recognition testing in ovariectomized adult female rats. Adolescent BPA exposure impaired recognition memory in adulthood. (*) P < 0.05 by ANOVA. Data in A redrawn from Diaz Weinstein et al. (2013). Data in B redrawn from Bowman et al. (2015). Data in C and D redrawn from Bowman et al. (2019).

In order to differentiate whether the observed effects in gonadally intact animals were due to BPA interactions with circulating gonadal hormones, we investigated the effects of adolescent BPA exposure in OVX female rats. Spatial memory was significantly impaired following 30-min and 1-h OP delays in BPA-treated OVX females compared with controls, when measured in adolescence (Fig. 3C; Bowman et al. 2019). When measured in adulthood, the BPA spatial memory impairments persisted at short (i.e., 10-min) but not long (30-min or 1-h, although the BPA effect trended P = 0.07, η^² = 0.23) intertrial delays (data not shown; Bowman et al. 2019). Similar to intact BPA females who showed no OR performance changes, BPA-treated OVX females performed the same as controls when measured in adolescence. However, adult OR performance was decreased in OVX females by adolescent BPA treatment (Fig. 3D; Bowman et al. 2019). The difference observed in BPA effects on OVX female adolescent and adult OR performance supports the idea that maturation of the PFC may be important for nonspatial working memory.

In sum, adolescent BPA exposure in OVX females has differential effects on cognitive function depending on the nature of the task (OP vs. OR), and these effects are similar across intact and OVX female rats. Additionally, in intact rats, BPA seems to have more profound behavioral effects in adolescence, while in OVX rats BPA seems to have greater effects in adulthood. This highlights the importance of the interaction between gonadal steroids and neural development and suggests that gonadally produced estrogen in adolescence may mitigate the effects of BPA.

Possible mechanisms for BPA effects: dendritic spine density

BPA's disruptive effects on spatial learning and memory are not well understood but may be related to BPA's ability to decrease NMDA receptor binding in areas known to underlie spatial learning and memory (CA1, CA3, and DG regions of the hippocampus) (Tian et al. 2010) as well as BPA's down-regulation of estrogen receptor β expression in the hippocampus. Furthermore, BPA's impairing effects on spatial memory may be due to dendritic and synaptic plasticity changes. It has been shown that a single, low dose of BPA in adulthood decreased apical and basal dendritic spine density in both the hippocampal CA1 region and medial PFC in adult male rats (Eilam-Stock et al. 2012) and that BPA can block gonadal steroid induction of dendritic spines in both CA1 and medial PFC in OVX rats (Inagaki et al. 2012). Moreover, BPA might disrupt spatial memory via morphological changes by blocking cell signaling pathways (Eilam-Stock et al. 2012).

We have demonstrated that short-term, low-dose exposure to BPA during the critical period of adolescence (PND 42–49) decreases spine density and that, in some cases, these treatment effects are dependent on sex (Bowman et al. 2014). In intact male and female rats, adolescent BPA exposure led to a substantial decrease in the spine density in CA1 and the medial PFC of all BPA-treated subjects compared with controls (Bowman et al. 2014). Furthermore, BPA exposure significantly interacted with sex whereby the BPA-dependent basal dendritic spine density decrease in male rats was greater than that observed in females (Bowman et al. 2014).

Adolescent BPA treatment also altered dendritic spine density in OVX females. In the dentate gyrus, BPA decreased dendritic spine density compared with control, but there were no changes in CA1. However, in OVX females, BPA exposure decreased spine density in the medial PFC (Bowman et al. 2019). Taken together, these findings demonstrate that BPA's negative effects on spatial memory may be mediated through neural mechanisms in areas that underlie these cognitive functions.

In sum, our research has shown that exposure to BPA decreases memory function in both intact male and female rats and OVX female rats in late adolescence and adulthood. In general, it appears that decreased memory is accompanied by decreased dendritic spine density in both the hippocampus and medial PFC. While the use of BPA has been limited in some cases (e.g., banned from baby bottles), there continues to be widespread use of and exposure to bisphenols, and continued understanding of their effects on cognition, neuronal development, and overall health is warranted.

Conclusion

There is compelling evidence that endocrine status of an animal impacts cognitive functioning on both spatial and nonspatial behavioral assessments. Exposure to adrenal steroids via chronic restraint stressors has sex-dependent effects whereby males are impaired, and females are either enhanced or not affected. E₂ and T appear to promote cognitive function in both sexes, whereas endocrine disruptors lead to behavioral detriments in cognition that are more pronounced in males than females. These behavioral alterations in cognitive function appear to be the functional consequences of sex-specific changes in neurochemicals and morphology that occur following stress, gonadal hormone, or endocrine disruptor manipulations. Accounting for the mechanisms of steroid action on learning and memory and the importance of endocrine status is important for better understanding and development of treatments for neurodegenerative disease- and stress-related losses in cognition.