CRF modulation of central monoaminergic function: Implications for sex differences in alcohol drinking and anxiety

Abstract

Decades of research have described the importance of corticotropin-releasing factor (CRF) signaling in alcohol addiction, as well as in commonly co-expressed neuropsychiatric diseases including anxiety and mood disorders. However, CRF signaling can also acutely regulate binge alcohol consumption, anxiety, and affect in non-dependent animals, possibly via modulation of central monoaminergic signaling. We hypothesize that basal CRF tone is particularly high in animals and humans with an inherent propensity for high anxiety and alcohol consumption, and thus these individuals are at increased risk for the development of alcohol use disorder and comorbid neuropsychiatric diseases. The current review focuses on extrahypothalamic CRF circuits, particularly those stemming from the bed nucleus of the stria terminalis (BNST), found to play a role in basal phenotypes and examines whether the intrinsic hyperactivity of these circuits is sufficient to escalate the expression of these behaviors and steepen the trajectory of development of disease states. We focus our efforts on describing CRF modulation of biogenic amine neuron populations that have widespread projections to the forebrain to modulate behaviors including alcohol and drug intake, stress reactivity, and anxiety. Further, we review the known sex differences and estradiol modulation of these neuron populations and CRF signaling at their synapses to address the question of whether females are more susceptible to the development of comorbid addiction and stress-related neuropsychiatric diseases because of hyperactive extrahypothalamic CRF circuits compared to males.

1) Introduction

Alcohol consumption has profound negative effects across organs in the body that contribute to the incidence of many diseases including cardiovascular disease, hepatic and renal failure, cancers, diabetes, and neuropsychiatric disorders such as addiction and anxiety (Fan, Russell, Stranges, Dorn, & Trevisan, 2008; George, Nutt, Dwyer, & Linnoila, 1990; van de Wiel & de Lange, 2008). The negative impact of alcohol is highly governed by the amount and pattern of its consumption, with repeated cycles of high volume intake to intoxication followed by abstinence-induced withdrawal being the most problematic for both acute and chronic health effects. Recurrent binge drinking and withdrawal have been shown to produce maladaptive plasticity in brain circuits that drive reward-seeking behavior and also to recruit stress circuits, leading to states of anxiety or negative affect that facilitate further alcohol intake (Breese et al., 2005; G. F. Koob, 2003, 2008; Sinha et al., 2008). Because of this close tie between emotional state and alcohol drinking behavior, individuals with an alcohol use disorder are at increased risk for the development of an anxiety or mood disorder, and vice versa. Women are at a particularly high risk for the comorbid expression of these neuropsychiatric diseases (Kessler et al., 1997), suggesting that the neural circuit mechanisms underlying these phenotypes are highly sensitive to the effects of alcohol and stress in females. Therefore, understanding the circuits driving these behaviors in males and females and the plasticity occurring within them that contributes to disease states is critical to understanding the mechanisms of these diseases and to developing successful treatment strategies.

Conceptual models of alcohol and drug abuse posit that early drug seeking behavior is driven by the positive rewarding effects achieved by activation of midbrain dopamine neurons that project to the limbic system, and that plasticity occurs in this circuit across repeated use and withdrawal to decrease DA release and reward. At the same time, because alcohol use is a physiological stressor, neural circuits that underlie the stress response become simultaneously recruited and hyperactivated to produce behavioral states of anxiety and negative affect. Thus even in the absence of the positive reinforcing effects of alcohol, repeated consumption is maintained by the anxiolytic properties of alcohol that provide relief from the sustained negative behavioral state that exists during abstinence (negative reinforcement). Many extrahypothalamic stress circuits have been shown to be involved, including those dense with biogenic amines and “stress” neuropeptides like corticotropin-releasing factor (CRF)(G. Koob & Kreek, 2007b; G. F. Koob, 2004; G. F. Koob & Le Moal, 1997; G. F. Koob, Rassnick, Heinrichs, & Weiss, 1994; G. F. Koob & Volkow, 2016).

CRF is not only critical for the maintenance of neuropsychiatric disease phenotypes but can promote both reward and aversion in naïve animals, including basal binge alcohol drinking and anxiety (Lowery-Gionta et al., 2012; Lowery & Thiele, 2010). CRF-mediated modulation of central monoamine systems may play a role in driving these behavioral phenotypes. For example, recent studies show that CRF neurons in the bed nucleus of the stria terminalis (BNST) within the extended amygdala that project to the midbrain modulate alcohol consumption and anxiety (Pleil, Rinker, et al., 2015; Rinker et al., 2016). CRF signaling modulates the activity of dopamine and norepinephrine neurons to enhance reward and stress reactivity, respectively (Lavicky & Dunn, 1993; Stenzel-Poore, Heinrichs, Rivest, Koob, & Vale, 1994; Wanat, Hopf, Stuber, Phillips, & Bonci, 2008; Weiss et al., 1994), and CRF-mediated activation of serotonergic neurons has been shown to modulate behavioral stress responses (Hammack et al., 2002; Hammack, Schmid, et al., 2003). Interestingly, the BNST projects to the major dopaminergic, serotonergic, and noradrenergic producing nuclei of the brain (Lebow & Chen, 2016); although it is not know whether these all of projections carry CRF, the existence of the pathways raises the intriguing possibility that increased activation of BNST CRF neurons may alter central monoaminergic signaling to drive alcohol drinking and anxiety.

Here, we review and synthesize this emerging literature. We hypothesize that high intrinsic excitability of BNST CRF neurons drives the expression of excessive alcohol consumption and anxiety in non-dependent animals, increasing the likelihood of and leading to earlier and more severe expression of neuropsychiatric disease states, and that differential modulation of central monoaminergic signaling by BNST CRF neurons may facilitate this transition. As women are at increased risk for comorbid alcohol/substance abuse and anxiety disorders (Kessler et al., 1997), we posit that females have upregulated CRF systems compared to males at baseline, conferring vulnerability. We discuss the existing evidence, when available, about known sex differences in and estrogen modulation of limbic CRF circuits and signaling relevant to alcohol drinking and anxiety.

2) CRF regulates alcohol drinking and anxiety

a) CRF is hyperactivated in addiction

CRF is a 41 amino acid neuropeptide with two primary G_s-coupled G protein-coupled receptors, CRFR1 and CRFR2 (Bale & Vale, 2004; Herman et al., 2003). Endogenous levels and polymorphisms in the genes coding for CRF and its receptors are associated with risk for the development of neuropsychiatric diseases including alcohol abuse and anxiety disorders in humans, as well as with many behavioral phenotypes including high stress responsivity and excessive alcohol consumption across mammalian species (Bale & Vale, 2004; Barr et al., 2009; Barr et al., 2008; A. C. Chen et al., 2010; Hayes, Knapp, Breese, & Thiele, 2005; Heilig, 2004; Heilig & Koob, 2007; M. B. Muller et al., 2003; Shekhar, Truitt, Rainnie, & Sajdyk, 2005).

While CRF and its receptors are located throughout the body and modulate stress responsivity in peripheral organs, the primary effects of CRF occur via receptor activation at neural loci (as reviewed in Gilpin, 2012; Lebow & Chen, 2016; Quadros, Macedo, Domingues, & Favoretto, 2016; Valentino & Bangasser, 2016; Valentino, Van Bockstaele, & Bangasser, 2013; Waters et al., 2015). The CRF neuron population in the hypothalamus is known to play a large role in stress responsivity as the initiator of the hypothalamic-pituitary-adrenal (HPA) axis during acute stress responses, but CRF in extrahypothalamic regions can regulate drinking and stress behaviors independent of the HPA axis (Le et al., 2000; Lowery et al., 2010; Pastor et al., 2008). Interestingly, while hypothalamic CRF and HPA axis function is suppressed in neuropsychiatric diseases states including alcohol addiction (Richardson, Lee, O’Dell, Koob, & Rivier, 2008), extrahypothalamic CRF is upregulated. For example, repeated alcohol exposure and withdrawal alter CRF expression within extrahypothalamic regions including the amygdala, prefrontal cortex and BNST, which are key sites for the integration of reward and stress information (Funk, O’Dell, Crawford, & Koob, 2006; Lack, Floyd, & McCool, 2005; Merlo Pich et al., 1995; Olive, Koenig, Nannini, & Hodge, 2002; Sommer et al., 2008; Zorrilla, Valdez, & Weiss, 2001).

Several reviews over the past decade have elegantly described the role of extrahypothalamic CRF signaling in alcohol dependence and addiction (Heilig & Koob, 2007; Lowery & Thiele, 2010; Phillips, Reed, & Pastor, 2015; Zorrilla, Logrip, & Koob, 2014). CRF is recruited during binge alcohol drinking and altered in rodent models of alcohol dependence (Ayanwuyi et al., 2013; Barr et al., 2009; Heilig & Koob, 2007; Kaur, Li, Stenzel-Poore, & Ryabinin, 2012; Lowery-Gionta et al., 2012; Lowery et al., 2010; Lowery & Thiele, 2010; Roberto et al., 2010; Sparta et al., 2008; Valdez & Koob, 2004). Pharmacological manipulations have shown a causal relationship between CRF signaling and behaviors associated with alcohol use disorder, from excessive alcohol consumption and alcohol-induced aggression (Quadros et al., 2014) to withdrawal-induced anxiety, negative affect, and relapse behavior (Baldwin, Rassnick, Rivier, Koob, & Britton, 1991; Breese, Knapp, & Overstreet, 2004; Huang et al., 2010; Le et al., 2000; Rassnick, Heinrichs, Britton, & Koob, 1993; Valdez et al., 2002; Valdez, Zorrilla, Roberts, & Koob, 2003) including stress-induced reinstatement (Funk et al., 2006; Le et al., 2000; Liu & Weiss, 2002). In particular, CRF signaling at CRFR1 and CRFR2 in extrahypothalamic regions in the brain are involved in heavy alcohol drinking during the transition to and maintenance of alcohol dependence (Chu, Koob, Cole, Zorrilla, & Roberts, 2007; Funk & Koob, 2007; Funk, Zorrilla, Lee, Rice, & Koob, 2007; Gehlert et al., 2007; Gilpin, 2012; Hansson et al., 2006; Le et al., 2000; Sommer et al., 2008).

These data suggest that plasticity in CRF signaling is crucial in the development of alcohol use disorder and comorbid anxiety (reviewed in Heilig & Koob, 2007; G. F. Koob, 2008; G. F. Koob & Nestler, 1997; Phillips et al., 2015). Taken together, these and other studies have demonstrated that CRF is recruited across repeated alcohol use and withdrawal and is critical for the maintenance of neuropsychiatric disease phenotypes. This relationship suggests that individuals with elevated basal CRF system reactivity may, upon initiation of alcohol drinking, transition to dependence more quickly. Indeed, evidence suggests that CRF can promote excessive alcohol consumption, anxiety, and aversion in the absence of prior alcohol exposure (discussed in the next section). As females tend to develop alcohol dependence more rapidly than males, we hypothesize that sex-dependent differences in CRF signaling may underlie this relationship.

b) CRF regulates basal reward and anxiety: systemic evidence

The negative health effects of alcohol consumption in humans are thought to result from “binge drinking”, intake marked by high volume alcohol consumption (achieving a blood ethanol concentration (BEC) of at least 80 mg/dl), and subsequent withdrawal. Binge drinking behavior is a major risk factor for alcohol use and anxiety disorders in humans (Blazer & Wu, 2009; Courtney & Polich, 2009; Enoch, 2006) and is modeled well in rodents (Lange & Voas, 2000; Lyons, Lowery, Sparta, & Thiele, 2008; Rhodes, Best, Belknap, Finn, & Crabbe, 2005; Rhodes et al., 2007; Thiele & Navarro, 2013). Initial alcohol/drug use is thought to elicit increases in mid- and hindbrain dopamine, norepinephrine, and opioid signaling that encodes drug reward, followed by withdrawal-induced negative affect and anxiety driven by stress neuropeptides including CRF (J. B. Becker, Perry, & Westenbroek, 2012; Sarnyai, Shaham, & Heinrichs, 2001), which in turn promote intake. While in some brain regions CRF seems to become further and further recruited across repeated consumption (G. F. Koob, 2013, 2014), CRF itself can elicit binge drinking and anxiety in naïve animals.

CRF signaling has been shown to modulate drinking in various animal models of escalated alcohol self-administration, including the well-established “drinking in the dark” (DID) and two-bottle choice paradigms, among others. For example, systemic or central administration of CRF and CRFR1 agonists enhance, and CRFR1 antagonists blunt, binge alcohol consumption without altering low, non-binge levels of alcohol intake (Cippitelli et al., 2012; Funk et al., 2007; Lowery-Gionta et al., 2012; Lowery et al., 2010; Richardson, Zhao, et al., 2008; Sparta et al., 2008; Valdez et al., 2002). Similarly, CRFR1 KO mice consume less of a standard high concentration 20% alcohol solution, but not a low concentration solution, than wild type mice in two-bottle choice (Pastor et al., 2011) and DID paradigms (Kaur et al., 2012). Thus, CRF signaling, particularly at CRFR1, appears to be important for excessive alcohol consumption in both dependent and non-dependent states. In humans, CRFR1 polymorphisms are associated with binge drinking and alcohol intake (A. C. Chen et al., 2010; Treutlein et al., 2006), including the ability of stressful life events to predict heavy alcohol drinking (Blomeyer et al., 2008; Schmid et al., 2010). CRF’s role in alcohol consumption is also mediated by signaling at CRFR2; for example, CRF-deficient mice do not develop a conditioned place preference for alcohol as readily as control mice (Olive et al., 2003), and loss of CRFR2 signaling apparently decreases alcohol’s rewarding properties (Giardino, Cocking, Kaur, Cunningham, & Ryabinin, 2011; Ryabinin et al., 2012). Given the role of the CRF system in both the positive and negative reinforcement of alcohol drinking behavior, high basal CRF signaling may be both more rewarding and more aversive, affording CRF a role in positive and negative reinforcement-driven alcohol consumption even in animals that do not have a history of alcohol exposure.

We hypothesize that elevated basal CRF signaling in addiction susceptible individuals may produce a hyperactive stress response, which in turn stimulates excessive drinking and accelerates the frequency and severity binge drinking episodes, thereby hastening the transition to alcohol dependence. In support of this, a line of rats selectively bred to prefer alcohol, termed “P rats”, exhibit increased basal stress reactivity and alcohol consumption, as well as high basal central CRF signaling, compared to non-preferring (NP) rats; specifically, some strains of P rats display high CRF and CRFR1 expression (Hansson et al., 2006; Zhou, Colombo, Gessa, & Kreek, 2013) and potentiated responses to central CRF administration across brain regions (Ehlers et al., 1992). Conversely, stress reactivity and anxiety are blunted in transgenic mice that do not express CRFR1 receptors (Bale & Vale, 2004; M. B. Muller et al., 2003). In humans, there is some evidence that elevated CRF and CRFR1 levels may contribute to anxiety and affective disorders in clinical populations (Risbrough & Stein, 2006). The anxiogenic effect of central CRFR1 activation has been extensively reviewed (Dautzenberg & Hauger, 2002; Reul & Holsboer, 2002; Shekhar et al., 2005; Zorrilla & Koob, 2004). Interestingly, recent evidence suggests that there are sex differences in CRF receptor activation and adaptation in response to stress (Bangasser et al., 2010); these differences may help explain the increased vulnerability of females to developing stress-related pathology, including comorbid alcohol use and anxiety disorders.

c) Sex differences in addiction: are na ve females in a post-dependent state?

Women are more likely than men to have anxiety and mood disorders, as well as comorbid alcohol/substance abuse disorders (Back, 2011; Breslau, 2001; Conway, 2006; Kessler, Chiu, Demler, Merikangas, & Walters, 2005; Steiner et al., 2005; Wilcox & Yates, 1993). For example, co-expression of severe alcohol use disorder (AUD) and anxiety disorders is almost twice as high in women (61%) as men (36%) (Brady & Randall, 1999; Kessler et al., 1997). Given this staggering statistic, it is unsurprising that women have higher stress levels and are more likely to consume alcohol in response to acute stress (Rice & Stuart, 2010; Timko, Finney, & Moos, 2005). While epidemiological data suggest that men are currently more likely to suffer from alcohol abuse disorder, changing social factors that have led to increased public alcohol consumption in women (Slade et al., 2016). These changes are evidenced by the recent report that alcohol use disorder is already more prevalent in teenaged girls than boys (White et al., 2015), which suggests that this gender gap in adults may reverse within the next 20 years. Females across mammalian species also display higher levels of basal stress reactivity/anxiety and binge alcohol consumption than males, suggesting that even prior to the exposure of acute or chronic stressors the mechanisms driving these phenotypes may be engaged to a greater degree in females compared to males. Similarly to the differences observed P rats compared to NP rats, females may have higher basal CRF signaling in critical neural circuits than males and be more sensitive to the effects of both acute and chronic stressors, including alcohol. Furthermore, there is evidence that estradiol interacts with estrogen receptors to increase transcription of the Crh gene, which encodes CRF protein (X. N. Chen, Zhu, Meng, & Zhou, 2008; Lalmansingh & Uht, 2007). Altogether, this increased stress responsivity may confer an augmented trajectory toward the development of neuropsychiatric disease states, the so-called “telescoping effect”. While this vulnerability for disease is known from epidemiological and other observational data, females have largely been excluded from pre-clinical and clinical studies. As such, understanding the role of central CRF circuits and the mechanisms by which this system is modulated in both males and females is imperative to understanding, preventing, and treating these diseases.

Previous work has suggested that telescoping of addiction in women may be driven by the greater likelihood of females to self-medicate anxiety or negative affect with alcohol or drugs (J. B. Becker et al., 2012). In contrast, men are more likely to use drugs initially to achieve positive rewarding/euphoric effects in the absence of anxiety or negative affect, particularly in young people (Kuntsche & Muller, 2012; S. Muller & Kuntsche, 2011). However, the euphoric effects of many drugs are greater in females than males, even in those not previously exposed. For example, female rats develop a CPP at lower doses of cocaine and morphine than males (Karami & Zarrindast, 2008; Russo, 2003b; Zakharova, Wade, & Izenwasser, 2009), and have a more pronounced reinstatement of extinguished cocaine CPP (Bobzean, 2010) as well as greater locomotor responses to acute cocaine administration (Sircar & Kim, 1999). Estradiol potentiates euphoria felt from stimulant use in cycling women (Justice & de Wit, 1999, 2000). Furthermore, ovariectomy decreases cocaine and amphetamine CPP and sensitization, and replacement of ovarian hormones restores these effects in female rats (Russo, 2003a; Silverman & Koenig, 2007; Sircar & Kim, 1999). These findings demonstrate the role of estradiol in modulating drug reward. Although this relationship has yet to be carefully characterized for alcohol reward, there is some evidence that ovarian hormones increase alcohol reward in female rats (Torres, Walker, Beas, & O’Dell, 2014). Furthermore, estradiol has been shown to increase ethanol consumption by female rats and mice (Ford, Eldridge, & Samson, 2002, 2004; Quirarte et al., 2007; Rajasingh et al., 2007). In addition, central CRF signaling increases grooming in females to a greater extent than males and this effect is potentiated in high estrogen states, suggesting that estrogen facilitates the anxiogenic effects of CRF (Wiersielis et al., 2016). More research is needed to understand the role of estradiol in modulating alcohol reward and to determine whether CRF mediates or potentiates any observed effects of ovarian hormones on binge drinking behavior.

3) Critical extrahypothalamic sites for CRF signaling: BNST projections and monoaminergic targets

a) The Bed Nucleus of the Stria Terminalis (BNST): a critical source of extrahypothalamic CRF neurons

CRF and CRF receptor-containing neurons are highly expressed not only in the hypothalamus to initiate the HPA axis but throughout the mammalian brain including the cortex, extended amygdala, thalamus, cerebellum, midbrain and hindbrain (G. F. Koob, 2013; Swanson, Sawchenko, Rivier, & Vale, 1983). Data over the past two decades has illustrated the important role of CRF neurons and signaling in the central nucleus of the amygdala (CeA) in the extended amygdala in addiction and anxiety disorders (Gilpin, Herman, & Roberto, 2015; Silberman & Winder, 2015), and more recently evidence has characterized a critical role for a dense population of CRF neurons in the BNST in regulating behavioral state and driving intake behavior.

The BNST is a critical hub in the neural circuitry mediating stress and reward (G. F. Koob, 2009; Silberman & Winder, 2013). Seated in the extended amygdala, it receives synaptic input from cortical, limbic, thalamic and hindbrain structures and sends projections to downstream regions responsible for coordinating the behavioral stress response and implicated in drug and alcohol self-administration (H. Dong, Petrovich, & Swanson, 2000; H. W. Dong, Petrovich, & Swanson, 2001; H. W. Dong, Petrovich, Watts, & Swanson, 2001; Georges & Aston-Jones, 2001; Kash, 2012; Kash & Winder, 2006; G. Koob & Kreek, 2007a; G. F. Koob et al., 2004; Myers, Mark Dolgas, Kasckow, Cullinan, & Herman, 2013; Shin, Geerling, & Loewy, 2008). Up to 90% of neurons in the BNST are GABAergic (Cullinan, Herman, & Watson, 1993; Kudo et al., 2014; Le Gal LaSalle, Paxinos, & Ben-Ari, 1978; Marcinkiewcz et al., 2016; Silberman, Matthews, & Winder, 2013) and many co-release one or more neuropeptides and express their receptors, including CRF (Ju & Han, 1989; Poulin, Arbour, Laforest, & Drolet, 2009; Rainnie, 1999; Rainnie et al., 2004; Walter, Mai, Lanta, & Gorcs, 1991). The roles of these various afferents and efferents, as well as neurochemical subtypes within the BNST, in addiction, anxiety, and affect has recently been reviewed (Lebow & Chen, 2016; Vranjkovic, Pina, Kash, & Winder, 2017).

BNST CRF neurons have been shown to play important roles in alcohol/drug-seeking behavior and stress responsivity in models of relapse, such as stress-induced reinstatement for cocaine seeking (Vranjkovic et al., 2017). CRF mRNA is increased during alcohol withdrawal and decreased with subsequent acute alcohol, suggesting dynamic modulation of CRF production in these neurons across phases of alcohol intake (Olive et al., 2002). In addition, BNST CRF neurons drive both binge drinking and anxiety in non-dependent male and female mice (Pleil, Lowery-Gionta, et al., 2015; Rinker et al., 2016). While further evidence specifically for the role of CRF neurons in such phenotypes in non-dependent animals is scarce, the BNST has been shown to be involved in many aspects of alcohol/drug seeking behaviors that CRF mediates. For example, the BNST is activated by alcohol-associated cues (Dayas, Liu, Simms, & Weiss, 2007; Hill, Ryabinin, & Cunningham, 2007) and is involved in the expression of CPP for alcohol and other drugs of abuse (Buffalari & See, 2011; Pina, Young, Ryabinin, & Cunningham, 2015; Rogers, Ghee, & See, 2008; Sartor & Aston-Jones, 2012), suggesting it encodes some aspects of alcohol/drug reward. Together, the BNST as a region, and possibly CRF neurons within this region, may be involved in both the positive and negative reinforcement of alcohol consumption.

CRF neurons in the BNST can serve as interneurons or principle neurons with long range projections (Marcinkiewcz et al., 2016; Silberman et al., 2013); as such, CRF signaling elicited by activation of BNST CRF neurons can occur both within the BNST as well as at synapses of long range projections at other sites. CRF signaling within the BNST facilitates anxiety and aversion (Sahuque et al., 2006), startle behavior (Y. Lee & Davis, 1997), and stress reactivity, including stress-induced reinstatement of drug seeking (Cummings 1988; Dong and Swanson 2006a,b; Erb and Stewart 1999). Thus, it is unsurprising that chronic alcohol exposure increases CRF tone in the BNST (Francesconi et al., 2009; Silberman et al., 2013). These studies indicate a potentially substantive role for CRF signaling from local BNST CRF neurons on alcohol and anxiety-related behaviors. However, these studies have yet to be performed using female subjects. It is possible that females have higher basal alcohol consumption and anxiety and stress reactivity than males due to higher CRF tone from tonic activation of BNST CRF neurons.

Interestingly, while BNST volume is larger in male rats, the density of CRF neurons is greater in females (Fukushima, Furuta, Kimura, Akema, & Funabashi, 2013; Funabashi, Kawaguchi, Furuta, Fukushima, & Kimura, 2004). Characterized projection sites of BNST CRF neurons in male rats and mice include the VTA, lateral hypothalamus, and locus coeruleus, however to our knowledge these have not been investigated in females to the same degree. In humans, structural connectivity between the BNST and connected regions is 76% greater in females than males (Avery et al., 2014), suggesting that there is greater density and/or diversity in synaptic connections within the BNST in women. Together, this higher density of CRF neurons and greater connectivity could have profound effects on the function of BNST CRF neurons and the functional and behavioral consequences of their activation.

A large literature has demonstrated that the BNST regulates many behaviors in a sex-specific manner, including aggression, sociability, reproductive and parental behaviors, and pain (Hagiwara, Funabashi, Akema, & Kimura, 2013; Hagiwara et al., 2009; He, Wu, & Yu, 2014; He, Yu, & Wu, 2013; Klampfl, Brunton, Bayerl, & Bosch, 2014, 2016; Martinez, Levy, & Petrulis, 2013; Martinez & Petrulis, 2011; Trainor, Finy, & Nelson, 2008; Trainor, Greiwe, & Nelson, 2006; Trainor, Lin, Finy, Rowland, & Nelson, 2007). We hypothesize that the organization and function of BNST CRF neurons may be fundamentally different in males and females, particularly as related to the described sex differences in alcohol consumption and anxiety behaviors driven by this circuit. Further, the modulation of BNST CRF circuits by sex hormones may play a large role in these differences, evidence for which is discussed in the succeeding sections.

b) The Locus Coeruleus (LC): a critical locus of CRF signaling

The locus coeruleus (LC) is the major source of central norepinephrine (NE). The LC is activated by stress and is an integral part of the neural stress response system (F. J. Chen & Sara, 2007). Afferents from the LC project diffusely across the brain, and increased noradrenergic signaling has been implicated in many neuropsychiatric disease states including anxiety disorders and addiction (Sara, 2009). CRF signaling likely plays a role in driving LC excitability, as the LC is innervated by CRF-containing fibers originating in the central amygdala and BNST (Sakanaka, Shibasaki, & Lederis, 1987; Swanson et al., 1983; Valentino, Page, Van Bockstaele, & Aston-Jones, 1992) and NE-producing LC afferents express CRFR1 (De Souza, 1987; De Souza et al., 1985; Reyes, Glaser, & Van Bockstaele, 2007; Van Bockstaele, Colago, & Valentino, 1996). In fact, local or i.c.v. CRF administration increases the excitability of LC neurons (Curtis, Lechner, Pavcovich, & Valentino, 1997; Prouty, Waterhouse, & Chandler, 2017; Valentino & Foote, 1987, 1988), and CRF administration promotes NE release in LC efferent terminal regions (Lavicky & Dunn, 1993). Intra-LC CRF administration has long been known to promote anxiety (Stenzel-Poore et al., 1994; Weiss et al., 1994), and stress-induced increases in LC CRF levels may alter LC activation states to drive the behavioral stress response (Curtis, Bello, & Valentino, 2001; Valentino, Foote, & Page, 1993; Valentino, Page, & Curtis, 1991). More recently, McCall and colleagues (2015) demonstrated that activation of CRF inputs to the LC increases tonic firing, and this activity is correlated with increased expression of anxiety-like behaviors in mice (McCall et al., 2015). Given the importance of CRF in driving LC excitability and anxiety, this may be one neural locus where individual differences in CRF responsivity promote increased basal stress responses and alcohol addiction vulnerability.

Disrupted noradrenergic signaling has long been implicated in excessive alcohol intake. Alcohol acutely increases plasma NE in non-dependent humans (Howes & Reid, 1985) and mice (Hawley, Major, Schulman, & Lake, 1981; Kovacs, Soroncz, & Tegyei, 2002), and NE depletion may reduce the rewarding effects of alcohol intake (Ventura, De Carolis, Alcaro, & Puglisi-Allegra, 2006). Interestingly, there is some evidence that alcohol withdrawal increases NE levels (Kovacs et al., 2002), and noradrenergic therapeutics that decrease LC excitability have shown some efficacy in reducing drinking and the negative symptoms of alcohol withdrawal in clinical populations (Muzyk, Fowler, Norwood, & Chilipko, 2011; Simpson et al., 2009) as well as in preclinical models of excessive alcohol drinking (Gilpin & Koob, 2010; Le, Harding, Juzytsch, Funk, & Shaham, 2005; Skelly & Weiner, 2014; B. M. Walker, Rasmussen, Raskind, & Koob, 2008). Dysregulated LC excitability is thought to contribute to the anxiogenic state induced by alcohol withdrawal, and CRF signaling may potentiate this excitability; in fact, protracted involuntary alcohol exposure has been shown to increase anxiety, LC excitability, and the distribution of CRF receptors in LC noradrenergic neurons (Retson, Reyes, & Van Bockstaele, 2015). Together, these findings suggest that activation of CRF inputs to the LC may drive drinking acutely and increase the risk of developing alcohol dependence. As such, individuals that are more sensitive to the excitatory effects of CRF signaling on central noradrenaline levels may be more likely to engage in binge drinking and therefore at greater risk of developing alcohol dependence. Understanding the role of CRF in driving the activity of this region could prove important in preventing the progression of alcohol-related disease states. Several studies have demonstrated that the BNST sends CRF producing afferents to the LC (Curtis, Bello, Connolly, & Valentino, 2002; Van Bockstaele, Bajic, Proudfit, & Valentino, 2001; Van Bockstaele et al., 1996; Van Bockstaele, Colago, & Valentino, 1998; Van Bockstaele, Peoples, & Valentino, 1999). Although CRF-containing projections from the central amygdala (CeA) to the LC have been implicated in driving alcohol drinking and anxiety (Gilpin et al., 2015; Valentino & Van Bockstaele, 2008), BNST CRF inputs to this region have been less well characterized but may also play a critical role in driving alcohol drinking and related anxiety. Evidence suggests potential sex differences in the responsivity of the LC neurons to CRF input, although these effects are independent of estrogen signaling. Specifically, morphological differences in the dendritic structure of LC cells may increase the sensitivity of these cells to CRF input from the BNST and other limbic regions in females (Bangasser, Wiersielis, & Khantsis, 2016). Additionally, stress has been shown to increase the firing rate of LC projections to a greater extent in females, and this effect is mediated by CRF (Curtis, Bethea, & Valentino, 2006b). Females also exhibit increased basal sensitivity to CRF and reduced sensitization of CRF-mediated LC excitability following chronic stress when compared to males (Bangasser, 2010; Curtis, Bethea, & Valentino, 2006a). Similarly, although in males CRFR1s apparently internalize following prolonged activation (Bangasser, 2010; Reyes, Fox, Valentino, & Van Bockstaele, 2006; Reyes, Valentino, & Van Bockstaele, 2008), no such compensatory mechanism exists in females (Bangasser et al., 2010; Bangasser et al., 2013). Chronic exposure to alcohol increased activation of LC neurons and the expression of CRF receptors in this region to a greater extent in females than in males (Retson et al., 2015). Together, these findings suggest that CRF-mediated potentiation of LC excitability may be increased in females compared to males, and support our hypothesis that increased sensitivity of the CRF system augments vulnerability to stress-related disorders and alcohol addiction in a sex-dependent manner.

c) The Ventral Tegmental Area (VTA): a critical locus of CRF signaling

A great deal of research over many decades has demonstrated the importance of the mesolimbic dopamine system in driving reward seeking behaviors, including drug and alcohol consumption (Bromberg-Martin, Matsumoto, & Hikosaka, 2010). The ventral tegmental area (VTA) in particular is an important source of dopaminergic afferents to downstream regions involved in driving behavioral responses to reward and aversion. Interestingly, the dopamine system is also directly activated by stressful stimuli (Horger & Roth, 1996); for example, stress increases the firing rate (Anstrom & Woodward, 2005) and synaptic strength (Saal, Dong, Bonci, & Malenka, 2003) of VTA dopamine neurons. Acute stress exposure also increases DA release in brain regions where VTA afferents terminate, including the striatum, prefrontal cortex, nucleus accumbens and amygdala (Abercrombie, Keefe, DiFrischia, & Zigmond, 1989; Adler et al., 2000; Gao & Wan, 2008; Hauger, Risbrough, Brauns, & Dautzenberg, 2006; Inglis & Moghaddam, 1999; McFarland, Davidge, Lapish, & Kalivas, 2004; Soliman et al., 2008; Tidey & Miczek, 1996; B. Wang, 2005).

CRF modulation of VTA DA neurons may enhance the response of these cells to acute stress exposure. The VTA receives a dense CRF projection from the BNST (Rinker et al., 2016; Rodaros, Caruana, Amir, & Stewart, 2007; Vranjkovic et al., 2017), and stress induces CRF release in the VTA possibly via activation of these afferents. For instance, stress increases DA release in brain regions receiving input from VTA afferents (Wanat, 2008), as does central CRF administration (Lavicky & Dunn, 1993), while CRF receptor antagonists reduce cocaine-induced dopamine release (Lodge & Grace, 2005). Additionally, administration of CRF directly into the VTA increases the activation of VTA DA neurons (Wanat et al., 2008), downstream DA release (Lemos et al., 2012; Wanat, 2008), and the expression of dopamine-dependent behaviors (Kalivas, Duffy, & Latimer, 1987; B. Wang et al., 2005). Stress-induced increases in VTA CRF have been shown to increase DA neuron activity in a similar fashion to drugs of abuse (Holly & Miczek, 2016; Sinha, 2008), and CRF can increase inhibition in VTA DA neurons in a manner similar to that observed following chronic stress or protracted cocaine administration (Beckstead, 2009; Creed et al., 2016).

Numerous preclinical studies have demonstrated that VTA DA signaling contributes to the reinforcing effects of ethanol (Gonzales, Job, & Doyon, 2004). Alcohol directly activates VTA DA neurons (Brodie, Pesold, & Appel, 1999) and cross-sensitization of VTA DA neurons by stressors and alcohol has been reported (Roberts, Lessov, & Phillips, 1995). Inhibition of CRF receptors in the VTA decreases alcohol consumption in a voluntary self-administration model of alcohol consumption (Holly et al., 2016), modulates reinstatement of drug seeking following extinction (Holly et al., 2016; Holly & Miczek, 2016; B. Wang, 2005), and drives stress-induced reinstatement of cocaine self-administration via signaling at DA neuron synapses (Wise & Morales, 2010). CRFergic projections from the BNST to the VTA drive reward-seeking behaviors (Jennings et al., 2013; Pina et al., 2015; Rinker et al., 2016), likely by disinhibiting DA neurons. This DA neuron disinhibition by BNST CRF/GABA neurons may be the mechanism by which this projection contributes to CPP for alcohol (Pina et al., 2015) and facilitation of binge alcohol consumption in non-dependent mice (Rinker et al., 2016). Blocking CRF receptors in the VTA produces the same alcohol consumption phenotype as inhibiting BNST CRF neuron terminals in the VTA (Rinker et al., 2016), and activation of BNST CRF projections to the VTA are involved in binge drinking and stress-induced reinstatement of alcohol intake (Rinker et al., 2016; Vranjkovic, Gasser, Gerndt, Baker, & Mantsch, 2014). After 4 weeks of chronic alcohol exposure, CRFR1 inhibition in the VTA decreased alcohol consumption VTA (Albrechet-Souza et al., 2015; Hwa, Debold, & Miczek, 2013; Sparta et al., 2013). Thus, increased basal excitability of BNST CRF projections to the VTA might confer increased addiction vulnerability in certain populations (for example, in females).

Although sex differences in the responsivity of VTA neurons to CRF input has not been directly studied, many groups have uncovered evidence that estrogen modulates VTA DA neuron activity and the rewarding effects of drugs of abuse. For example, estradiol has been shown to modulate dopamine release in response to ethanol in female rats (Dazzi et al., 2007). In humans, females report that cocaine administration produces a more robust feeling of euphoria while in a high-estrogen state (Evans & Foltin, 2010; Evans, Haney, & Foltin, 2002), and preclinical work has demonstrated that estradiol stimulates dopamine release to a greater extend in females than males (J. Becker & Cha, 1989; Robinson & Becker, 1982). Drug-induced dopamine release is potentiated in females during high estrogen states (Q. D. Walker, Ray, & Kuhn, 2006), and assessment of cocaine-induced VTA DA neuron excitability has been shown to be modulated by naturally fluctuating estrogen levels (Calipari et al., 2017). Furthermore, cocaine self-administration is greater in females when circulating estrogen levels are high (Long, Yang, Faingold, & Steven Evans, 2007). Given the suspected role of CRF projections from the BNST to the VTA in driving alcohol intake and anxiety (Pleil, Lowery-Gionta, et al., 2015; Rinker et al., 2016), future work should directly assess whether estrogen-induced modulation of this input contributes to the known effects of estrogen state on drug intake and related behaviors.

d) The Dorsal Raphe Nucleus (DR): a critical locus of CRF signaling

The dorsal raphe nucleus of the brainstem is the primary source of central serotonin and sends afferents broadly, including to limbic and cortical targets implicated in stress responding, affect regulation, and addiction (Hale, Shekhar, & Lowry, 2012; Maier & Watkins, 2005). The DR is innervated by CRF containing fibers, and CRF release in this region impacts the excitability of DR afferents via actions at postsynaptic CRFR1 and CRFR2 (Austin, Rhodes, & Lewis, 1997; Kirby, Rice, & Valentino, 2000; Ruggiero, Underwood, Rice, Mann, & Arango, 1999; Sakanaka et al., 1987; Swanson et al., 1983). Disrupted DR CRF signaling has been implicated in mood and affective disorders in clinical and pre-clinical research (Gold & Chrousos, 2002; Holsboer, 1999; Nemeroff, 1996; Waselus, Valentino, & Van Bockstaele, 2005); for example, CRF-immunoreactivity is increased in the DR of suicide victims who suffered from depression (Austin, Janosky, & Murphy, 2003). As such, the DR is one region where individual differences in basal CRF signaling in response to alcohol and other stressors may contribute to anxiety and addiction vulnerability.

CRF is released in the DR in response to stress (Hammack et al., 2002; Roche, Commons, Peoples, & Valentino, 2003; Valentino & Commons, 2005) and produces a biphasic effect of DR neuron excitability. Specifically, low doses of CRF inhibit neuronal activity and serotonin levels in downstream targets via CRFR1 receptor activation, while at high doses CRF potentiates DR afferent excitability via activation of CRFR2s, resulting in increased serotonin release at terminal sites (Hammack, Pepin, DesMarteau, Watkins, & Maier, 2003; Hammack, Schmid, et al., 2003; Kirby et al., 2000; Pernar, Curtis, Vale, Rivier, & Valentino, 2004; Price, Curtis, Kirby, Valentino, & Lucki, 1998; Price & Lucki, 2001; Valentino & Commons, 2005). Thus, it seems possible that individual differences in the strength of stress-induced CRF release in the DR could result in divergent behavioral stress responses. Importantly, CRF has been shown to modulate serotonin release at terminals of afferents originating in the DR (Price & Lucki, 2001), although modulation of these effects by circulating sex hormones or other variables has yet to be investigated. The DR also contains a small population of CRF-producing neurons; these cells appear to co-localize with serotonin and project to the extended amygdala, including the central amygdala and BNST (Commons, Connolley, & Valentino, 2003; Valentino, Lucki, & Van Bockstaele, 2010).

CRF signaling in the DR has been implicated in promoting ethanol self-administration. For instance, both systemic and intra-DR infusion of a CRFR1 antagonist have been shown to decrease binge alcohol drinking in mice and rats (Hwa et al., 2013; Hwa et al., 2016). In another study, combined alcohol withdrawal and acute stress exposure prompted increased alcohol self-administration and anxiety in rats, and CRFR1 antagonists infused into the DR prevented the anxiogenic effects of this treatment (although without affecting alcohol intake) (Knapp et al., 2011). As reviewed above, stress increases CRF release in the DR, and the net effect of this increased CRF signaling appears to be CRFR-mediated inhibition of DR serotonin projections. This CRF-induced reduction in serotonin signaling has been suggested to promote the initiation of drinking by increasing impulsivity (Valentino et al., 2010; Virkkunen & Linnoila, 1990). In support of this, decreased serotonin signaling has been reported in animal models of alcohol self-administration, and in alcohol naive rats selectively bred for high ethanol preference (LeMarquand, Pihl, & Benkelfat, 1994; McBride & Li, 1998). Likewise, clinical research suggests that decreased serotonergic signaling increases vulnerability to alcohol dependence in humans (Johnson, 2000; Johnson & Ait-Daoud, 2000; LeMarquand et al., 1994). Although the source of CRF release in the BNST in response to alcohol and other stressors is not known, BNST CRF neurons do project to the DR (Holstege, Meiners, & Tan, 1985; H. S. Lee, Kim, Valentino, & Waterhouse, 2003; Weissbourd et al., 2014); more work is needed to determine whether this projection plays a role in CRF-mediated stress and alcohol responses of DR neurons, whether sex differences exist in the responsivity of this system, and whether such differences might confer increased risk of developing comorbid alcohol use and anxiety disorders in women.

The DR may be one region where sex differences in CRF signaling determine differential responses to stress. Several authors have reported that females exhibit a potentiated stress response to acute administration of a serotonin reuptake inhibitor (SSRI) (Carlsson & Carlsson, 1988; Goel & Bale, 2010; McEuen, Semsar, Lim, & Bale, 2009; Nishizawa et al., 1997), and Howerton and colleagues (2014) hypothesized that sex-dependent CRF modulation of DR serotonergic projections may contribute to this phenotypic difference. Indeed, this group demonstrated that intra-DR infusion of a CRFR1 antagonist blunts HPA axis activation and related anxiety in response to acute stress exposure in male, but not female, mice (Howerton et al., 2014). This finding was linked to decreased CRF receptor-mediated neuronal activation in females (Howerton et al., 2014). Another group demonstrated that CRFR1 and CRFR2 levels were higher in the DR of female adolescent rats as compared to males (Lukkes, Norman, Meda, & Andersen, 2016). To our knowledge, sex differences in the responsivity of DR neurons to CRF input have not been studied directly. Future studies should investigate whether there are sex differences in estrogen-dependent modulation of DR excitability or central serotonin signaling, and whether estrogen might interact with CRF to alter the excitability of DR neurons, thereby contributing to the increased binge drinking and rate of transition to alcohol dependence observed in females.

e) Overview of sex differences in CRF modulation of monoaminergic signaling

As discussed above, there is strong evidence that CRF modulates neuronal excitability in several brain regions implicated in anxiety and alcohol addiction, and that these effects may be sex-dependent (Bangasser et al., 2010; Curtis et al., 2006a; Valentino & Bangasser, 2016; Valentino et al., 2013). Although the neurophysiological mechanisms of these sex-related differences vary, taken together these findings suggest that stress and alcohol increase CRF signaling to a greater effect in females than in males. As CRF hypersecretion contributes to stress-related psychopathologies in humans (Austin et al., 2003; Bremner et al., 1997; De Bellis, Gold, Geracioti, Listwak, & Kling, 1993; Heuser et al., 1998; Nemeroff, 1996; Raadsheer, Hoogendijk, Stam, Tilders, & Swaab, 1994; S. S. Wang, Kamphuis, Huitinga, Zhou, & Swaab, 2008), and individuals afflicted with these disorders are more likely to engage in binge alcohol intake (Keyes, Hatzenbuehler, & Hasin, 2011), this difference in CRF signaling may bias females toward increased alcohol addiction vulnerability. In fact, heightened sensitivity of female brains to CRF signaling has been proposed to contribute to the increased incidence of stress-related psychiatric conditions in females (Bangasser, Eck, Telenson, & Salvatore, 2017). This group recently demonstrated that CRF increases BNST activation in females to a greater extent than males, and that plasma estradiol levels are negatively correlated with CRF-induced activation of this region in females (Salvatore et al., 2017). Given the known role of CRF in driving alcohol drinking (Heilig & Koob, 2007; Logrip, Koob, & Zorrilla, 2011; Quadros et al., 2016), we are extending this hypothesis to propose that sex-dependent differences in stress-related CRF release and subsequent signaling may potentiate binge alcohol drinking and promote alcohol addiction vulnerability in females.

Of note, while some of the sex-dependent differences in CRF signaling reviewed herein are dependent on estrogen signaling, others are independent of estrogen release. For example, the observed increased sensitivity of LC noradrenergic neurons to CRF in females appears to occur via an estrogen-independent Gs-coupled mechanism (Bangasser, 2010; Bangasser, Dong, et al., 2017). Similarly, observed sex differences in CRF sensitivity in the DR appear to be explained by differences in the expression of CRF receptors in this region (Howerton et al., 2014). Conversely, although sex differences in VTA CRF responsivity have not been directly studied, many existing studies have implicated estrogen modulation of activity in this region in drug self-administration, raising the intriguing possibility that cyclic alterations in circulating estrogen levels may also impact the effects of CRF signaling on VTA dopamine neuron activity. Interestingly, two recent studies have demonstrated that estrous cycle-dependent fluctuations in ovarian hormones alter the connectivity of CRF-activated brain regions in response to CRF administration (Salvatore et al., 2017; Wiersielis et al., 2016). Together, these findings suggest that both cyclic and structural differences in CRF sensitivity may predispose females toward increased stress-related behaviors, including increased likelihood of alcohol self-administration.

4) Conclusions

Here, we have synthesized evidence regarding CRF modulation of biogenic amine function related to alcohol consumption and anxiety, and we have described how CRF neurons in the BNST may be the source of CRF signaling at those critical sites. Interestingly, some of the many synaptic inputs into the BNST include those from the LC, VTA, and DR, suggesting at least indirect reciprocal modulation of BNST CRF and these amine projection neurons (Marcinkeiwcz 2016; Meloni 2006). For example, natural rewards and drugs of abuse, including alcohol, increase DA release in the BNST (Carboni 2000; Park 2012 or Park 2013), and the rewarding effects of alcohol and sucrose (as measured by CPP) can be attenuated with antagonism of DA D1 receptors in the BNST (Eiler 2003). VTA (and periaqueductal gray) DA neurons have been shown to synapse directly onto BNST CRF neurons (Meloni 2006), so BNST CRF-VTA DA interactions may be critical for the positive reinforcing effects of alcohol that drive binge drinking and potentiated in females.

While to date these DA inputs are the only known to monosynaptically innervate CRF neurons, future studies may reveal that the many GABA, glutamate, and neuropeptide inputs to the BNST (recently reviewed in Vranjkovic et al., 2017) directly modulate CRF neurons. These studies will be illuminating regarding sex differences in BNST CRF neuron function, as estrogen is known to modulate both glutamatergic and GABAergic synaptic transmission via postsynaptic membrane-bound estrogen receptors in a sex-dependent manner. For example, in the hippocampus, estrogen signaling increases postsynaptic glutamatergic function in both sexes but via different mechanisms (Oberlander & Woolley, 2016; Wong & Moss 1992) and retrogradely suppresses inhibitory synaptic transmission via postsynaptic membrane-bound ERα in females but not males (Huang and Woolley 2012; Rudick Woolley 2001). In addition, estrogen can rapidly potentiate neuronal excitability (Wong & Moss, 1991), perhaps as a consequence of these combined mechanisms. Given the known anxiogenic effects of ERα activation (Borrow and Handa 2017), its high expression in the BNST (Trainor et al., 2007; Laflamme et al., 1998), and evidence that it mediates other rapid behavioral effects of estrogen in the BNST (Trainor et al., 2007; 2008), it could also regulate alcohol drinking and anxiety behaviors via modulation of BNST CRF neuron function. Sex differences in the organization of BNST CRF neuron circuits and estradiol modulation of BNST CRF neurons and the CRF signaling that happens at their target loci could have large-scale implications for the observed sex differences in basal behavioral phenotypes and risk for rapid and severe development of neuropsychiatric disease states. BNST CRF neurons could be the source of high extrahypothalamic CRF tone that predispose females to the co-expression of alcohol use and anxiety disorders.

It is important to note that there are many other brain regions where CRF signaling may potentially modulate alcohol intake in a sex-dependent manner. For example, CRF receptor binding is increased in the cortex and amygdala of female rats (Weathington & Cooke, 2012; Weathington, Hamki, & Cooke, 2014), and sex-dependent differences in the distribution of CRF receptors have been reported in the hippocampus (Williams, Akama, Knudsen, McEwen, & Milner, 2011). In general, CRF has been shown to increase neuronal activation across wide swaths of the hindbrain, limbic system, and neocortex, although sex-dependent effects have not always been considered when investigating these effects (Wiersielis et al., 2016). Additionally, CRF signaling has been proposed to promote binge drinking by altering the excitability of extrahypothalamic regions not discussed in this review, including the central amygdala (Lowery-Gionta et al., 2012; Silberman & Winder, 2015; Vendruscolo et al., 2015) and nucleus accumbens (Knapp et al., 2011), and alcohol-induced alterations in glucocorticoid signaling have been shown to alter extrahypothalamic CRF signaling and promote alcohol intake (Edwards, Little, Richardson, & Vendruscolo, 2015). In addition to CRF, other stress-related hormones may differentially contribute to binge drinking in males and females; for example, intriguing new evidence suggests that sex differences exist in the responsivity of amygdala subnuclei to corticosterone which may promote sex differences in stress-induced alcohol intake (Logrip, Oleata, & Roberto, 2017). Taken together with the studies reviewed elsewhere in this article, these findings highlight the importance of identifying the specific and likely complex mechanisms underlying sexually dimorphic brain responses to stress and alcohol. Such research will assist in the development of new tools to prevent and treat alcohol addiction and related disorders.

Highlights.

• Women are more susceptible to comorbid expression of alcohol use and anxiety disorders, neuropsychiatric diseases that are associated with CRF signaling.

• Individuals with inherently high central CRF tone may binge drink more and have higher basal anxiety, contributing to disease risk.

• The bed nucleus of the stria terminalis (BNST) in the extended amygdala is a dense source of extrahypothalamic CRF neurons known to regulate binge alcohol drinking and anxiety, and to project to sites containing biogenic amine-containing neurons.

• CRF signaling modulates the function of biogenic amine signaling in a sex-dependent manner.