Abstract
Sex differences in motivation for food rewards, gambling, and drugs of abuse are modulated by multiple factors, including sensory stimuli, gonadal hormones, and cognitive bias. Cues, drugs of abuse, and a high-fat diet can significantly impact neural signaling in the reward system and functioning of neural systems that regulate executive functions differentially in males and females. Additionally, sex differences in risky decision-making, cognitive bias, and motivation for food and drugs of abuse are mediated by gonadal hormones in both sexes. As neuroscientists analyze data from both sexes, it is becoming apparent that these differences are not simply mediated by hormones in females, but involve sex differences in the specific neural responses to stimuli, including both external stimuli and internal hormonal signals. Understanding sex differences in the mechanisms underlying reward-seeking behaviors and the development of substance use disorders will help uncover potential therapies and treatments that will benefit both men and women. Based on these observations, it is essential that females are included in neuroscience research.
Introduction
Addiction and obesity are major public health concerns, and understanding the neural mechanisms mediating reward and addiction in both females and males is important for improving prevention and enhancing treatment for these conditions. For too long, female subjects have been excluded from scientific research, although it is now clear that female rodents are not more variable than males (Beery and Zucker, 2011; Becker et al., 2016; Beery, 2018; Smarr and Kriegsfeld, 2022). Understanding sex differences and the mechanisms underlying reward-seeking behaviors and the development of substance use disorder will help uncover potential therapies and treatments that will benefit both men and women.
In this review, we discuss sex differences in the reward system that mediate the likelihood of substance abuse, gambling, and obesity and the potential role of sex differences in cognitive bias in these disorders. We define “sex differences” as those differences between males and females which are determined from the gonadal and chromosomal composition at birth, characteristic of biological sex. We do not refer to “gender differences” in this review, defined as culturally acquired or attributed characteristics or social structures pertaining to gender identity.
It is becoming increasingly apparent that sex differences in the reward system are not simply mediated by hormones in females but involve specific neural responses within the reward system to stimuli, including both external and internal signals, that differ by sex and may be modulated by gonadal hormones in either sex. It is beyond the scope of this article to go into extensive detail about the reward system circuitry. The reader is referred to other excellent reviews for additional details (Kauer and Malenka, 2007; Berridge, 2012; Millan et al., 2017; Yoest et al., 2018; Nestler and Lüscher, 2019; Kokane and Perrotti, 2020; Doncheck et al., 2021). For the purposes of this discussion, we focus on the ascending dopaminergic pathways from the midbrain (cell bodies in the ventral tegmental area and substantia nigra) that project to the nucleus accumbens (NAc) and dorsal striatum, respectively, as well as other forebrain regions, including the medial prefrontal cortex (mPFC). Within the regions of the dopamine (DA) cell bodies and terminals, the GABAergic medium spiny neurons (MSNs) maintain inhibitory control over the DA neuronal activity (Yoest et al., 2018; Kokane and Perrotti, 2020). There is further inhibitory regulation of the NAc and dorsal striatum by the mPFC (Stefani and Moghaddam, 2006; Richard and Berridge, 2013). Finally, MSNs in the NAc also receive excitatory input from the paraventricular thalamus (PVT) (Hoover and Vertes, 2007; Dong et al., 2017). Sex differences and hormonal modulation of these brain regions will be discussed.
Previous research into the biological bases for sex differences in reward and addiction has primarily focused on how estradiol (E2) alters the reward system in females. For example, clinical models have found that, when E2 levels are greatest, women report an enhanced euphoria or “high” with cocaine smoking (Sofuoglu et al., 1999). In preclinical studies, female rats show enhanced cocaine craving and motivation for cocaine when E2 levels are elevated (Roberts et al., 1989; Hu et al., 2004; Becker and Hu, 2008; Nicolas et al., 2019; Martin et al., 2021). While these data suggest that E2 plays a major role in facilitating motivation and other behaviors in females specifically, the role of gonadal hormones in modulating males' addiction-like behaviors also needs to be understood.
Not all sex differences are created equal (Becker and Koob, 2016; Becker and Chartoff, 2019; Beltz et al., 2019). In some cases, as is true for reproductive functions, males and females exhibit two distinct phenotypes that cannot be measured on the same scales, referred to as “qualitative differences.” In other cases, both sexes can exhibit a response or phenotype, such as verbal fluency where one sex is faster than the other (Hampson and Kimura, 1988; Hampson, 1990). These are “quantitative differences.” In many cases, there is variation in a trait, but more males or more females exhibit a particular phenotype as a result of “population differences.” As an example, more men than women are addicted to alcohol (Becker and Koob, 2016). There are also sex differences in which both males and females exhibit a trait, but the neural mechanisms or neural circuits that mediate the trait are different, as is the case for corticotropin releasing factor where different intracellular signaling systems mediate the response to corticotropin releasing factor in males and females; this is a “mechanistic or latent difference” (Valentino et al., 2013; Salvatore et al., 2018). These are not mutually exclusive categories.
Finally, sex differences are dynamic, in that they can be invoked or activated under selective circumstances or revealed because of the way in which a test is conducted (Figure 1). An example of this is cognitive bias, the perception of ambiguous stimuli as either negative or positive. A cognitive bias allows an animal or individual to rapidly respond to cues in its environment and is influenced by emotional state in species from rodents to primates (Nguyen et al., 2020). In rodents, negative affect, usually related to stressful conditions, results in a negative cognitive bias, while nonstressed rodents are more positively biased (Nguyen et al., 2020). There are sex and age differences in rat cognitive bias, and in functional connectivity involved in this bias, although animals may exhibit similar abilities to discriminate between two contexts.
Sex differences in cognitive bias
Cognitive bias can be tested to determine the affective state, the expectation or perception of events, and biases in the attention and memory of individuals (Hertel and Mathews, 2011; Jones and Sharpe, 2017). Increased motivation is associated with a positive interpretation of ambiguous cues (positive cognitive bias) while a negative cognitive bias is associated with a lack of motivation or a dysfunctional processing of reward (anhedonia) (Pizzagalli, 2014; Blackwell et al., 2015). Cognitive bias involves pattern separation (Leal et al., 2014; Magaraggia et al., 2021) and the ability to discriminate between highly similar situations (Yassa and Stark, 2011). Interestingly, there are sex differences in pattern separation (Yagi et al., 2016). Further, hippocampal neurogenesis is required for pattern separation (Sahay et al., 2011; Déry et al., 2013; Fang et al., 2018; Boldrini et al., 2019; Riphagen et al., 2020), which implicates the hippocampus in negative cognitive bias.
Although there are several studies that have examined cognitive bias in rodents and humans (Harding et al., 2004; Joormann et al., 2007; Dearing and Gotlib, 2009; Enkel et al., 2010; Boleij et al., 2012; Papciak et al., 2013; Rinck et al., 2018; Drozd et al., 2019), very few have examined sex differences in cognitive bias and the mechanisms underlying this bias. In a study by Hodges et al. (2022), a novel shock-based cognitive bias task was used to assess sex and age differences in cognitive bias and related neural function. Rats were trained to discriminate between a negative context (paired with a footshock, high freezing behavior) and a neutral context (paired with no footshock, low freezing behavior), which differed by lighting, wall patterns, and transportation methods to the context. After training, cognitive bias was measured in response to an ambiguous context, which resembled both the neutral and negative contexts, via freezing behavior and quantified using a discrimination index. From adolescence to adulthood, negative cognitive bias increased in female and male rats, but male rats exhibited greater negative cognitive bias than female rats in middle age (Hodges et al., 2022). These data are similar to previous findings of an increased fixation on negative stimuli from young to old age and an immature assessment of risk in adolescence in humans (Rodham et al., 2006; Bucher et al., 2020). However, these human studies did not examine sex; these findings by Hodges et al. in rats suggest a greater negative cognitive bias with age in males compared with females.
Intriguingly, although there were no sex differences in the greater positive cognitive bias of adolescents and the increased negative cognitive bias of young adult rats, there were sex differences in the neural areas activated in response to these cognitive biases (Hodges et al., 2022). In females, greater neural activation in the NAc, amygdala, and hippocampal regions was correlated with freezing in response to the ambiguous context, as demonstrated by enhanced expression of the protein product of the immediate-early-gene c-fos, compared with males.
Regardless of similar cognitive bias behavior, functional connectivity between brain regions in response to cognitive bias also depended on sex and age (Hodges et al., 2022). Negative correlations among neural activity in regions of the hippocampus and the frontal cortex or basolateral amygdala (BLA) were only found in males and only emerged in adulthood (young adults, middle-aged adults) in response to cognitive bias. In females, positive correlations between the NAc and the hippocampus in adolescence shifted to positive correlations between the NAc and the frontal cortex in adulthood.
These findings demonstrate a “mechanistic sex difference,” where the behavior is similar between the sexes, but there are differences in underlying neural function. Furthermore, these data provide evidence for how neural function underlying behavior might change with development, and this can also interact with sex. A “quantitative sex difference” also emerges with age, as middle-aged males display a greater negative cognitive bias than middle-aged females. In addition, these data suggest that the NAc, a brain region involved in reward, motivation, and salience (for review, see Salgado and Kaplitt, 2015), may play a greater role in the cognitive bias of females compared with males.
As we think about the neural mechanisms mediating motivation for rewards, we must also consider that: (1) adolescents may have a greater positive cognitive bias and impaired risk assessment; (2) sex and age differences in the neural mechanisms may underlie these behaviors; and (3) cognitive bias may influence sex differences in reward and motivation across the lifespan. For example, sex differences in reward approach have been demonstrated in rodent cognitive bias tasks (Brown et al., 2016). In humans, the relationship between gambling disorder and greater cognitive bias differs for men and women (Jiménez-Murcia et al., 2020). More broadly, cognitive bias and its underlying neural connectivity may play a role in the mechanisms mediating sex differences in motivation for food, gambling, and drugs of abuse (Fig. 1).
Factors that contribute to sex differences in reward-related cognition and behavior. Sex differences are dynamic in nature and can manifest differently depending on factors that span various domains (e.g., biological, psychological, and environmental). The confluence of these factors can modulate organizational and functional activity within the brain. In addition to (or in combination with) the potential impact of gonadal hormones (estradiol and testosterone), the influence of these factors can sculpt neural activity to produce sex differences in reward-related behavior and cognition that are highly dependent on the individual and the surrounding environment.
Sex differences in models of risky decision-making
Increased sensitivity to rewards, and/or reduced learning from punishments, can impact risky decision-making and impulsivity. High levels of impulsivity and risky choice can contribute to the generation and maintenance of chemical and behavioral addictions (Gowin et al., 2013; Carroll and Smethells, 2015; Chen et al., 2020; Orsini et al., 2021). Whether sex differences in risk-taking and impulse control contribute to the distinct trajectory of addiction disorders in men and women is an open research question.
To understand the neurobehavioral mechanisms underlying drug-induced changes in decision-making, rodent animal models have been developed that recapitulate many aspects of decision-making observed in humans, including the existence of sex differences in risky decision-making (van den Bos et al., 2013). Across such tasks, males display a greater preference for rewards associated with greater risks relative to females (Orsini et al., 2016; Pellman et al., 2017; Liley et al., 2019; Islas-Preciado et al., 2020), and preliminary data suggest that such a difference may be because of reduced sensitivity to negative consequences in males.
Although such a diminished sensitivity to negative consequences would at first seem at odds with the greater negative cognitive bias in males relative to females observed by Hodges et al. (2022; described above), this behavioral discrepancy may be because of the nature of the tasks involved in the studies and the learning inherent in these tasks. Most decision-making tasks involve response learning (e.g., a lever press will result in a specific outcome); the cognitive bias task used by Hodges et al. (2022) relies on associative learning (e.g., context A signals outcome, with no action required from the subject). Comparisons across such tasks are therefore informative as they reveal that sex differences in behavioral tasks can vary depending on cognitive load and learning requirements associated with the tasks.
Reward-paired cues play a critical role in animal models of drug addiction (Robinson and Berridge, 1993; Grimm et al., 2001), and adding casino-inspired sound and light cues to laboratory-based gambling tasks that are presented concurrent with winning outcomes increases risky decision-making in both humans and rats (Barrus and Winstanley, 2017; Spetch et al., 2020; Hynes et al., 2021). Behavioral pharmacology studies in male rats suggest that these cues render decision-making much more sensitive to DA drugs, specifically DA ligands acting at D2 and D3 receptors (Bortolozzi et al., 2002; Heidbreder et al., 2005; Barrus and Winstanley, 2016). D3 receptors play a prominent role in regulating drug self-administration, particularly under heavily cued reinforcement schedules (Le Foll et al., 2005; Beninger and Banasikowski, 2008). As such, the ability of sensory cues to enhance drug-taking and risky decision-making may be regulated by similar neurochemical mechanisms. However, these conclusions are largely based on experiments that exclusively used male rats.
Given that reward-predictive cues increase DA activity (Schultz et al., 1997), and that this increase in DA is necessary for the cues to acquire incentive salience (Flagel and Robinson, 2017), it was reasoned that chemogenetically inhibiting DA projections from the ventral tegmental area (VTA) to the NAc would decrease cue-induced risky choice in the rat gambling task (Barrus and Winstanley, 2017). This was true for male, but not female, rats (Hynes et al., 2020), indicating that dopaminergic regulation of this form of decision-making differs across sex.
DA plays a particularly prominent role during learning of stimulus-outcome contingencies, yet the former study was performed once rats had already acquired the task. Again, using chemogenetics, DA cells within the VTA were inhibited while rats learned the cued rat gambling task. Whereas this manipulation significantly attenuated the ability of the cues to drive risky choice in male rats, the opposite effect was observed in female rats: these animals took significantly more risks when DA signaling was dampened (Hynes et al., 2021). Yet unpublished findings show that chemogenetic activation of VTA DA cells during acquisition amplifies cue-induced risky choice in males, with a similar effect in females. Thus, another “mechanistic sex difference” is that the DA system of females is more finely balanced, with both increases and decreases in activity leading to less optimal decision-making.
Sex differences in gonadal hormonal modulation of risky decision-making
Accumulating evidence has revealed that sex differences in risky decision-making are mediated by circulating gonadal hormones (Pellman et al., 2017; Wallin-Miller et al., 2017; Orsini et al., 2021). For example, using a task in which rats choose between a small, safe food reward and a larger food reward accompanied by varying risk of footshock punishment, Orsini et al. (2021) found that ovariectomies (OVX) shifted choice preference in females from the small, safe reward to the large, risky reward (i.e., increased risk taking), whereas orchiectomies (ORCH) shifted choice preference in males from the large, risky reward to the small, safe reward (i.e., decreased risk taking). Exogenous administration of E2 to OVX females decreased risk taking; surprisingly, exogenous administration of testosterone did not reinstate greater risk taking in ORCH males. Notably, E2 administration reduced risk taking in both ORCH and intact males. This indicates that E2 has a specific effect even in the presence of testosterone. Given that testosterone did not reinstate greater risk taking in ORCH males, it seems unlikely that phenotypical risk taking in males is because of aromatization of testosterone to E2 in the brain (if this were the case, one would expect E2 to attenuate ORCH-induced effects on risk taking). It is possible, however, that conversion of testosterone to its metabolite dihydrotestosterone mediates male risk-taking behavior. Indeed, dihydrotestosterone is a more potent androgen than testosterone and is capable of altering avoidance-related behavior in ORCH males (Edinger et al., 2004; Celec et al., 2015; Tobiansky et al., 2018). Future studies in the Orsini laboratory will conducted to test this hypothesis.
These findings demonstrate a role for circulating gonadal hormones in mediating phenotypical risk-taking behavior in males and females, a “quantitative sex difference.” They also reveal that E2 is critical for biasing choice away from risky choices, regardless of biological sex. This modulatory role for gonadal hormones, however, appears to vary depending on the risk or cost associated with the larger reward. For instance, when the risk consists of the possibility of omission of the larger reward, gonadectomies have no effect on decision-making. Further, although exogenous testosterone exerts a moderate effect on choice behavior in males, exogenous E2 does not alter behavior in females (Islas-Preciado et al., 2020). In contrast, when the cost consists of physical effort to obtain the larger reward, OVX increases choice of this option and E2 mitigates this effect (Uban et al., 2012). Considering these studies together, gonadal hormones, and E2 in particular, seem to have a specific and important regulatory role in decision-making when the risks associated with the more rewarding of the options involve a physical cost (e.g., effort, punishment) to the individual.
Given the involvement of gonadal hormones in modulating risky decision-making, it is conceivable that, in addition to altering the neural mechanisms that underlie reward-related behavior, chronic exposure to drugs disrupts the ability of hormones to regulate risky decision-making, resulting in elevated and maladaptive risk taking. In women who are currently or have a history of using cocaine, opioids, or alcohol, there is evidence of disruptions in their menstrual cycle (Mello, 1998; Emanuele et al., 2002; Schmittner et al., 2005). Similarly, chronic exposure to cocaine results in persistent irregularities in hormonal cyclicity in nonhuman primates and rats (King et al., 1990, 1993; Mello et al., 1997). Preliminary unpublished data from the Orsini laboratory reveal that, concurrent with the increase in risk taking that occurs during abstinence from cocaine self-administration (Blaes et al., 2022), there is also a disruption in estrous cyclicity in that persists for at least 6 weeks.
With respect to testosterone, studies have shown that acute exposure to drugs, such as cocaine, alter circulating testosterone levels, with an initial increase and then a subsequent decline (Gordon et al., 1980; Berul and Harclerode, 1989; Kohtz et al., 2019). It is, however, currently unclear what the long-term effects of drug use are on testosterone levels in adult males; as a consequence, it is unknown whether drug-induced impairments in risky decision-making in males are caused by a dysfunction in hormonal regulation of decision-making. Studies are currently underway in the Orsini laboratory to extend and further examine the effects of drugs of abuse on hormonal regulation of risky decision-making in males and females and to determine whether selective hormonal manipulations (e.g., activation or inhibition of E2 receptor [ER] signaling) can rectify decision-making impairments that occur because of drug use.
Sex differences in E2 modulation of reward
Another question that is important to consider is as follows: of the three known ERs, which mediates the response to E2 in reward? E2 acts by binding to ERs α (ERα), β (ERβ), and/or G-protein coupled ER 1 (GPER1). For example, risk aversion in females seems to be dependent on ERβ, although ERα has not yet been tested (Orsini et al., 2022). The traditional ERs, α and β, are located either in the cell nucleus, where they bind ER response elements to activate genomic responses, or in the extracellular plasma membrane, in association with caveolin proteins, where they mediate rapid intracellular signaling via activation of metabotropic glutamate receptors (Razandi et al., 1999; Boulware et al., 2007; Luoma et al., 2008; Meitzen and Mermelstein, 2011; Fuentes and Silveyra, 2019). GPER1 is also localized on the plasma membrane, as well as in the endoplasmic reticulum, and it activates intracellular signaling cascades mediated by cAMP, ERK, and PI3K (Otto et al., 2008).
In females, but not males, E2 rapidly enhances stimulated DA release in the dorsolateral striatum (DLS) and NAc (Thompson and Moss, 1994a,b; Cummings et al., 2014; Shams et al., 2016, 2018; Yoest et al., 2018, 2019; Song et al., 2019). ERα and GPER1 in the DLS are localized to GABAergic MSNs as well as cholinergic interneurons and glia (Almey et al., 2012, 2016). The cellular localization of ERα and GPER1 in the DLS of males remains unknown, and the NAc has not been characterized at the electron microscopic level in either sex (Quigley et al., 2021b), although in both sexes ERα and GPER1 in the DLS and NAc transition from a nuclear location during early development to being primarily extranuclear in the adult (Krentzel et al., 2021). When looking at effects of selective activation of ERs on NAc and DLS function in OVX female rats, ERβ modulates stimulated DA release in NAc, while ERα and ERβ have been shown to modulate neural activity in DLS (Schultz et al., 2009; Grove-Strawser et al., 2010; Yoest et al., 2019).
Motivation for drugs of abuse is enhanced by E2. Female rats in estrus exhibit greater drug-primed reinstatement compared with females not in estrus and males (Kippin et al., 2005). Female rodents express signs of enhanced drug craving during estrus compared with nonestrus (Nicolas et al., 2019). In OVX females, E2 acting at ERβ potentiates reinstatement of drug-seeking (Larson and Carroll, 2007; Becker and Hu, 2008; Doncheck et al., 2018). The effect of E2 on reinstatement of cocaine self-administration has been shown to be mediated by the effects of ERβ and GPER1, but not ERα, in the mPFC (Doncheck et al., 2021). Previous work also suggests that, during drug-primed reinstatement, females who are in estrus display greater cocaine-seeking behavior than nonestrous females and males, indicating that these effects of E2 can be seen in female rats across the estrous cycle (Kerstetter et al., 2008). Further, females take longer to extinguish cocaine-seeking behaviors compared with males (Kerstetter et al., 2008). These studies suggest that E2 plays a role in enhanced drug cravings in females by enhancing DA activity in the NAc and DLS, while suppressing inhibition of motivational circuitry via its effects in the mPFC. Together, these effects of E2 are proposed to contribute to the enhanced cocaine-seeking behavior in females.
Activation of GPER1, via intra-DLS administration of the GPER1 agonist G1, potentiated female rats' motivation to self-administer cocaine, but not that of males. Furthermore, G1 treatment resulted in greater drug-induced reinstatement in females, but not in males (Quigley et al., 2021a). Activation of GPER1 did not alter preference for cocaine or saccharin in females (Quigley and Becker, 2021). In males, on the other hand, activation of GPER1, via administration of GPER1 agonists ICI 182780 or G1 directly into the DLS, attenuated conditioned place preference for 10 mg/kg cocaine, while inhibition of GPER1, via G15 in the DLS, enhanced conditioned place preference at a 5 mg/kg cocaine dose (Quigley and Becker, 2021). Gene knockout studies in mice reveal that GPER1 can modulate preference for morphine in males (Sun et al., 2020), indicating that this effect is not selective for cocaine. Similarly, GPER1 activation, via G1 in the DLS, prevented males from forming a preference for 0.1% saccharin over plain water (Quigley and Becker, 2021). Together, these results demonstrate another “mechanistic sex difference”: E2 acting selectively at ERs plays different roles in males and females, enhancing drug preference in females through its actions in the NAc, DLS, and mPFC, but attenuating drug and saccharin preference in male rats by acting in the DLS.
Sex differences in mPFC regulation of reward
The mPFC is essential for regulation of reward-related behavior and associated reward circuitry (Ferenczi et al., 2016). There are sex differences in mPFC activity during a task when the response is likely to be punished, with females exhibiting a greater mPFC response to probabilistic punishment, compared with males (Jacobs and Moghaddam, 2020).
Within the mPFC, the primary interneurons are GABAergic and they regulate output from the mPFC (Hu et al., 2014). Parvalbumin (PV) GABAergic interneurons play a key role in maintaining the excitatory:inhibitory balance within the mPFC by establishing and maintaining coordinated and synchronized neuronal firing. This ultimately contributes to the generation of γ oscillations between brain regions (Hu et al., 2014; Wingert and Sorg, 2021). In addition, ERβ-positive neurons are predominantly PV-immunoreactive and can influence neuronal excitability (Blurton-Jones and Tuszynski, 2002). Both preclinical and clinical studies have shown sex differences in the GABAergic system under both basal and pathological conditions in many brain regions, and are influenced by gonadal hormones (Searles et al., 2000; Joh et al., 2006; Skilbeck et al., 2008; Kirkovski et al., 2018; Cornez et al., 2020b; Woodward and Coutellier, 2021).
A hallmark of PV interneuron maturity in the cortex is the formation of perineuronal nets (PNNs) around their cell body and proximal dendrites, which have been shown to have sex- and brain region-specific differences after exposure to rewarding stimuli (Dingess et al., 2020). These PNNs dynamically regulate PV interneuron activity, ultimately influencing excitatory:inhibitory balance within the PFC and PFC output. PNNs appear during experience-dependent “critical periods” of neuronal refinement and restrict plasticity in adulthood (Pizzorusso et al., 2002; Sigal et al., 2019). We now know that PNNs play key roles in maintaining fast and precise neuronal transmission (Balmer, 2016; Wingert and Sorg, 2021), protection against oxidative stress (Morawski et al., 2004; Cabungcal et al., 2013), and restricting plasticity (Pizzorusso et al., 2002, 2006). An appreciation for the dynamic regulation of PNNs in both healthy brain physiology as well as numerous brain diseases has caused a resurgence in studying PNNs (Lasek et al., 2018; Fawcett et al., 2019; Reichelt et al., 2019; T. E. Brown and Sorg, 2022; Chelyshev et al., 2022). However, even with the relatively newfound excitement within the neuroscience community to better understand PNN dynamics, our awareness of potential sex differences in PNNs has lagged.
Much like humans, rats prefer palatable foods to their home cage diets, with female rats showing greater preference for a diet high in fat than males (Darling et al., 2016; Maric et al., 2022). Work from the Brown laboratory has shown that consumption of a high-fat diet (60%) induced changes in PNNs within the mPFC in a sex- and region-specific manner (Dingess et al., 2020). That is, adult (60 d) male and female Sprague Dawley rats fed a high-fat diet for 21 d showed differential changes in PNN staining intensity, suggestive of shifts in PNN maturity. Brains from male rats had an attenuation in PNN intensity within the prelimbic (PL) PFC and ventromedial orbitofrontal cortex independent of weight gain (Dingess et al., 2017). However, there was no effect in PL PFC or ventromedial orbitofrontal cortex in female rats, but instead a significant increase in PNN intensity within the infralimbic (IL) PFC, illustrating an additional mechanistic sex difference. The PL- and IL-PFC have been postulated to have dichotomous functions, with the PL being important for driving rewarding behaviors and the IL being important for suppressing rewarding behaviors. This oversimplified view, however, has numerous exceptions (for a review on PL and IL roles, see Moorman et al., 2015).
The sex differences in PNNs reported by the Brown laboratory are consistent with results of other studies that have observed sex differences in PNNs. Female mice fed a high-fat diet show a reduction in PNN intensity in the hypothalamus, and gonadectomy reduces PNN intensity in both sexes in the arcuate (Zhang et al., 2021). Studies in songbirds have revealed region-specific sex differences in PNNs in nuclei responsible for song production, which is regulated by testosterone (Cornez et al., 2015, 2020a,b). Within the amygdala, there is a reduction in PNN intensity from ER-containing neurons of females, relative to male mice (Ciccarelli et al., 2021). Sex differences have also been reported between juvenile and adult rats: juvenile females show a reduction in PNNs within the hippocampus that does not persist into adulthood and is not seen in males (Griffiths et al., 2019). Finally, several studies have shown that rodent models of early life trauma result in sex-specific effects on PNNs (Page and Coutellier, 2018; Guadagno et al., 2020; Gildawie et al., 2021).
Future studies will determine whether gonadal hormones play a key role in diet-induced adaptations in PNNs because some studies have suggested that gonadal hormones can contribute to reorganization of PNNs (Cornez et al., 2020b; Uriarte et al., 2020). To truly grasp how diet-induced changes between males and females modulate neuronal plasticity and food-seeking behaviors, a more comprehensive framework needs to be established that incorporates the interplay of PNNs at the cellular level between presynaptic and postsynaptic terminals, glial cells, and the loose extracellular matrix (“tetrapartite synapse”) (Hodebourg et al., 2022).
Ovarian hormones and motivation for food
Ovarian hormones and the estrous cycle have been shown to affect drug-intake and drug-seeking as discussed above. E2 alone is sufficient to reduce food intake and body weight in OVX rats. Ovarian hormones also affect transient changes in motivation for food and drugs of abuse. Administration of E2 and progesterone to OVX rats transiently reduces motivation for food (Yoest, 2018). Selective activation of ERα, Erβ, or GPER1 does not fully replicate the effect of E2 + progesterone on motivation for food, indicating that multiple ER subtypes are involved in changes in food reward. Further, while administration of E2 alone reduces the number of rewards that animals earned over the course of each session, it was not sufficient to attenuate motivation for food (Yoest, 2018).
In naturally cycling female rats conditioned to associate an auditory stimulus with the delivery of a sucrose pellet in a food cup, there was a reduction in conditioned approach to the food cup when females were in the proestrus/estrus phases of the estrous cycle (characterized by elevated levels of E2 and progesterone), consistent with the discussion above (Alonso-Caraballo and Ferrario, 2019). Exogenous treatments of E2 and progesterone in OVX rats also decreased the conditioned approach behaviors. Together, these results demonstrate that motivational responses to Pavlovian food cues are also hormone-dependent in female rats.
To understand the effects of the ovarian cycle on the underlying neuronal mechanisms that regulate motivational responses to food cues, the effects that circulating ovarian hormones on the intrinsic properties of MSNs in the NAc were examined. Consistent with the behavioral findings, and the role of NAc MSN excitability in motivated behaviors, rats in proestrus/estrus had decreased MSN excitability compared with rats in diestrus (Proaño et al., 2018; Alonso-Caraballo and Ferrario, 2019). These data suggest that, throughout the cycle, ovarian hormones act on the NAc motivational circuitry to regulate the behavioral responses to food cues.
Additionally, the effects of abstinence from a junk-food diet on NAc glutamatergic transmission in an obesity-susceptibility rat model was examined. Glutamatergic transmission through calcium-permeable AMPARs in the NAc mediate cue-induced motivation for food (sugary-high-fat diets) (Derman and Ferrario, 2018) and the psychostimulants cocaine (Wolf and Tseng, 2012) and methamphetamine (Scheyer et al., 2016) in male rats. However, in females, junk-food diet exposure followed by abstinence does not enhance calcium-permeable AMPAR-mediated glutamatergic transmission as it does in males (Alonso-Caraballo et al., 2021). The effects of different periods of abstinence from a junk-food diet on glutamatergic transmission and NAc MSN excitability in females are currently being studied.
Sex differences in motivation for opioid drugs
In addition to sex differences in motivation for food and cocaine, there are also sex differences in pain and opioid use (Koons et al., 2018). Sex and gender differences are important factors for both the experience of pain and opioid use disorder. Women experience and perceive pain at a greater scale than men. Pain and opioid use combine with biological sex differences to create a greater risk of misusing and abusing prescription opioids in women than in men (Mogil and Bailey, 2010; Pieretti et al., 2016; Koons et al., 2018). However, we know very little about the neurobiological mechanisms underlying these sex differences.
In preclinical models, glutamatergic synaptic transmission within the NAc underlies cue-driven reward-seeking behaviors, and recent evidence shows that synaptic plasticity within projections from the PVT to the NAc shell is required for the expression of morphine withdrawal in male mice (Zhu et al., 2016). Here it was determined whether short (24 h) or long (14 d) abstinence from oxycodone self-administration results in similar changes in PVT-NAc shell plasticity and glutamatergic transmission in male and female rats.
Data from unpublished studies show that short abstinence from oxycodone self-administration increased somatic withdrawal symptoms (e.g., wet dog shakes, ptosis, and grooming) but did not affect PVT-NAc shell synaptic plasticity or drug seeking behavior for oxycodone. In contrast, prolonged abstinence from oxycodone self-administration enhanced PVT-NAc shell synaptic plasticity in both males and females. Interestingly, drug seeking was greater in females compared with males despite both having similar drug intake during self-administration. GABAergic transmission was not affected by abstinence in males or females. In addition, evidence suggests that the effects of oxycodone abstinence on glutamatergic transmission are likely driven by presynaptic site of action.
Collectively, these findings suggest that the effects of abstinence from oxycodone self-administration on PVT-NAc shell glutamatergic transmission are time- and sex-dependent. These results show that inputs to the NAc core and shell mediate regulation of responses of the reward system and that there are sex differences in this regard.
In conclusion, mechanistic sex differences in motivation for food, gambling, and drugs of abuse and in risky decision-making are modulated by multiple factors, including sensory stimuli, gonadal hormones, and cognitive bias as well as psychosocial and other biological factors (Fig. 1). Furthermore, drugs of abuse, negative cognitive bias caused by stress, and a high-fat diet can significantly impact neural signaling in the reward system and functioning of neural systems that regulate executive functions differentially in males and females. Additionally, drugs of abuse, such as cocaine and opioids, may disrupt gonadal hormonal function, thus having indirect effects on these behaviors. Finally, there are age and sex differences in cognitive bias and in the neural mechanisms underlying this bias, highlighting the importance of examining age and sex when investigating the mechanisms involved in motivation and risk assessment.
It is now clear that there are sex differences in the neural circuitry of the reward system as well as sex differences in the roles that gonadal hormones play in modulating neurotransmitter activity and neural regulatory factors (Fig. 2). We have shown that there are “mechanistic sex differences” in the neural mechanisms and circuitry mediating cognitive bias, risky decision-making, motivation for drugs of abuse, and mPFC regulation of food motivated behavior. There are also sex differences in risky decision-making and in motivation for food and drugs of abuse that are mediated by gonadal hormones in both sexes. Finally, E2 has specific effects in females to enhance neural activity in the NAc and DLS, while downregulating the inhibitory effects of the mPFC on motivated behaviors.
Sex differences at various levels of the brain can lead to sex differences in motivated behavior. Sex differences in motivated behavior may arise from sex differences in neural function at multiple levels of the brain. Sex-dependent cognitive bias, food-seeking, and decision-making in the presence of salient cues are because of differential activity and connectivity between brain regions within the mesocorticolimbic system, such as the NAc and VTA (a major source of DA in the brain) as well as input from thalamic regions (e.g., PVT). Gonadal hormones contribute to phenotypical male and female reward-based decision-making and can facilitate aspects of drug-seeking behavior. Not only can these hormones influence such motivated behavior by modulating connectivity between brain regions, they can also interact with neurotransmitters, such as dopamine and glutamate, and/or receptors (receptors for hormones themselves or receptors for neurotransmitters) to promote sex-specific reward-related behavior (involving natural or drug rewards). Finally, recent evidence shows that regulatory factors, such as PNNs in the extracellular matrix, can bias neural activity in the mPFC in a sex-dependent manner. Although graphically depicted separately, it is likely these mechanisms interact and influence one another to promote sex differences in motivated behavior. E2, estradiol; T, Testosterone; P, progesterone; DHT, dihydrotestosterone.
Understanding sex differences in the mechanisms underlying reward-seeking behaviors and the development of substance use disorder will help uncover potential therapies and treatments that will benefit both men and women. Based on studies of sex differences in the neural and hormonal substrates of cognitive bias, food reward, gambling, and drugs of abuse, it can be concluded that it is of utmost importance to include females in neuroscience research.
Footnotes
This work was supported in part by National Institutes of Health Grants R01 DA049795 and R01 DA046403 to J.B.B., R01 DA040965 to T.E.B., K00DA053527 to Y.A.-C., and R00 DA41493 and R21 DA05346 to C.A.O.; and Canadian Institutes of Health Research Project Grant PJT-162312 to C.A.W. T.E.H. was supported by Institute of Mental Health Marshalls Scholars Program (University of British Columbia, Canada) and Natural Sciences and Engineering Research Council of Canada operating grant (Liisa A.M. Galea, 2018-04301).
The authors declare no competing financial interests.
- Correspondence should be addressed to Jill B. Becker at jbbecker{at}umich.edu
