Corticotropin Releasing Factor: A Key Role in the Neurobiology of Addiction

Abstract

Drug addiction is a chronically relapsing disorder characterized by loss of control over intake and dysregulation of stress-related brain emotional systems. Since the discovery by Wylie Vale and his colleagues of corticotropin-releasing factor (CRF) and the structurally-related urocortins, CRF systems have emerged as mediators of the body’s response to stress. Relatedly, CRF systems have a prominent role in driving addiction via actions in the central extended amygdala, producing anxiety-like behavior, reward deficits, excessive, compulsive-like drug self-administration and stress-induced reinstatement of drug seeking. CRF neuron activation in the medial prefrontal cortex may also contribute to the loss of control. Polymorphisms in CRF system molecules are associated with drug use phenotypes in humans, often in interaction with stress history. Drug discovery efforts have yielded brain-penetrant CRF₁ antagonists with activity in preclinical models of addiction.. The results support the hypothesis that brain CRF-CRF₁ systems contribute to the etiology and maintenance of addiction.

Introduction

According to a 2012 report by the Substance Abuse and Mental Health Services Administration, within the past 12 months, approximately 15% of the population aged 12 and older experienced substance use disorders on alcohol, cigarettes, or an illegal drug. Alcohol use disorders alone have an annual prevalence of approximately 10% and account for 4.6% of all disability-adjusted life-years in developed countries (Rehm et al., 2009). Available pharmacotherapies for substance use disorders have only modest long-term efficacy and are underutilized (Heilig et al., 2011). Since the successive discovery by Wylie Vale and his colleagues of corticotropin-releasing factor (CRF) (Vale et al., 1981), the structurally-related urocortins (Ucn 1, Ucn 2, Ucn 3), and their cognate receptors (CRF₁, CRF₂) (Bale and Vale, 2004;Fekete and Zorrilla, 2007), CRF systems have emerged as therapeutic targets for substance abuse.

CRF binds with high and moderate potency to CRF₁ and CRF₂ receptors, respectively. Ucn 1 is a high-affinity agonist at both of these G-protein coupled receptors, whereas the type 2 urocortins (Ucn 2 and Ucn 3) are selective CRF₂ receptor agonists (Bale and Vale, 2004;Zorrilla and Koob, 2004;Fekete and Zorrilla, 2007). Vale and colleagues first demonstrated that CRF initiates the hypothalamic-pituitary-adrenal (HPA) axis neuroendocrine stress response by binding CRF₁ receptors in the anterior pituitary after release into portal blood. In addition, however, CRF₁ receptors are widely distributed in stress-responsive brain regions, including the neocortex, central extended amygdala, medial septum, hippocampus, thalamus, cerebellum, and autonomic midbrain and hindbrain nuclei (Grigoriadis et al., 1996; Primus et al., 1997;Sanchez et al., 1999;Van Pett et al., 2000). The brain CRF₁ receptor distribution resembles the distribution of its natural ligands CRF (Fig. 1) and Ucn 1 and accounts for the dissociable, non-endocrine role of extrahypothalamic CRF₁ systems (i.e., outside the HPA axis) to mediate behavioral and autonomic stress responses (Swanson et al., 1983;Kozicz et al., 1998;Bale and Vale, 2004;Zorrilla and Koob, 2004;Fekete and Zorrilla, 2007).

Figure 1. Brain CRF mediates the facilitation of compulsive-like drug use.

As shown in the sagittal brain schematic, corticotropin-releasing factor (CRF), first isolated by Professor Wylie Vale (photo), is expressed in neuronal cell bodies (filled circles) and projections (blue arrows) that subserve behavioral, autonomic and neuroendocrine responses to stress. As summarized from left-to-right by the arrows, CRF systems play an integral role in regulating the intersection between drug self-administration and stress systems. For example, drug or alcohol withdrawal elevates CRF activity in the central extended amygdala, including the central nucleus of the amygdala (CeA), leading to a negative emotional state that motivates resumption of and maintenance of drug-taking. Pharmacological studies with CRF antagonists show that increased CRF-CRF₁ system activation underlies several withdrawal-induced behavioral phenotypes, including anxiety-like behavior, aversion, and elevated drug self-administration. CRF₁ antagonists also reduce the effect of acute stressors on drug-related behaviors, including stress-induced reinstatement. In contrast, CRF₁ antagonists do not alter non-stress mechanisms that reinstate drug-seeking, such as drug primes, cues or contexts, reflecting the distinct neuroanatomical substrates of relapse behavior. Blockade of CRF-CRF₁ systems does not have intrinsic rewarding (or aversive) properties in place conditioning models and has little effect on baseline intake in nondependent individuals.

Extensive preclinical data suggest that extrahypothalamic CRF₁ systems subserve negative emotional states. Accordingly, small-molecule CRF₁ antagonists are being developed as potential treatments for affective-like disorders, including posttraumatic stress disorder, irritable bowel syndrome, anxiety disorders, and major depression (Zorrilla and Koob, 2004;Holsboer and Ising, 2008;Zorrilla and Koob, 2010;Koob and Zorrilla, 2012;Zorrilla et al., 2013a). Indeed, Dr. Vale was a major force in the pharmaceutical development of drug-like small-molecule CRF₁ antagonists. He co-founded Neurocrine Biosciences, which successfully developed a wide range of such compounds, spanning multiple patents.

One proposed clinical indication for CRF₁ antagonists is drug addiction (Fig. 1), where the brain stress systems are hypothesized to impact key elements of the addiction cycle. Drug addiction is a chronically relapsing disorder characterized by loss of control over drug intake and emergence of a negative emotional state during abstinence. Drug addiction has been conceptualized as a cycle progressing through three stages—binge/intoxication, withdrawal/negative affect, and preoccupation/anticipation—that become worse over time and ultimately lead to a severe neurobiological disorder. CRF systems are hypothesized to play a key role in all three stages of the addiction cycle but particularly in the withdrawal/negative affect stage. Chronic use of a drug of abuse, even if initiated for its rewarding effects, increasingly leads to negative emotional symptoms and negatively reinforced substance use. An extension of the “opponent process theory of affective regulation” (Solomon and Corbit, 1974), this hypothesis of addiction proposes that drugs of abuse initially activate brain structures that subserve positive emotional states (e.g., pleasure, contentment). The positive reinforcing effects of drugs are regulated in part by the ventral striatum and extended amygdala reward system, as well as by dopaminergic and opioid inputs from the ventral tegmental area (VTA) and arcuate nucleus of the hypothalamus, respectively. To maintain emotional homeostasis, however, a counter-regulatory opponent process then decreases mood and increases vigilance/tension via downregulation of brain reward systems (e.g., ventral striatum) and upregulation of brain stress systems, including CRF and norepinephrine systems in the extended amygdala (Heilig and Koob, 2007;Heilig et al., 2010a;Heilig et al., 2010b;Koob and Zorrilla, 2010;Breese et al., 2011;Heilig et al., 2011;Logrip et al., 2011;Koob and Zorrilla, 2012). With continued cycles of intoxication/withdrawal, the opponent process allostatically predominates over the primary rewarding process (Fig. 2). As a result, more substance of abuse is needed simply to maintain euthymia. If drug use stops, negative emotional symptoms emerge (i.e., acute withdrawal: anxiety, dysphoria, irritability). With a sufficient drug use history, stress-like symptoms of dysphoria may episodically and spontaneously resurge even weeks or months after detoxification (i.e., protracted withdrawal). Furthermore, exaggerated responses to otherwise mild stressors may be seen despite continued abstinence. Under this conceptualization of addiction, substance abuse escalates because the drug of abuse mitigates the counter-regulatory negative emotional symptoms of acute and protracted withdrawal (Heilig and Koob, 2007;Koob and Zorrilla, 2010;Zorrilla et al., 2013a).

Figure 2. The progression to compulsive drug use alters emotional homeostasis via opponent-process upregulation of CRF activity in multiple brain nuclei.

The top-panel illustrates the opponent-process, emotional allostasis model of drug addiction. In the model, acute drug use initially elicits positive shifts in mood (“a” process). This is subsequently countered, however, by homeostatic decrements in mood (“b” opponent-process). With repeated drug use, the “b” opponent-process manifests earlier and more prominently, such that each drug experience elicits a smaller, briefer positive shift in mood. The associated neuroadaptations alter the individual’s emotional set points, yielding allostatic states of decreased reward function and increased stress function. The individual’s mood in a drug-free state does not return to the drug-naïve baseline (homeostatic point). Rather, new, stable allostatic states result, in which the drug-free baseline mood (allostatic points) becomes increasingly more negative, experienced as dysphoria in the absence of drug. With continued use, further drug-taking can no longer reattain the baseline, drug-naïve homeostatic set point, let alone a subjective positive “high” (Koob and Le Moal, 2001). Compulsive, escalating drug use is motivated by the negative emotional allostatic state in negative reinforcement fashion in a futile attempt to reduce dysphoria and regain euthymia. The bottom panel hypothesizes the successive recruitment of CRF systems across different brain regions during the transition to compulsive drug-taking and loss of control. During the initial phases of drug self-administration, when drug-taking still effectively elicits positive mood states and maintained by positive reinforcement, brain regions central to reward processes (ventral tegmental area, nucleus accumbens) are recruited. With continued drug use, neuroadaptations in CRF systems in the ventral tegmental area and prefrontal cortex may contribute to the generation of compulsive drug seeking. Finally, opponent-process activation of CRF systems in the extended amygdala and recruitment of circuitry linked to the dorsal striatum underlie the negative emotional state and habitual drug-seeking, respectively, observed in individuals with severe addiction.

The reviewed opponent process putatively of otherwise silent brain CRF₁ receptor stress systems of the extended amygdala. For example, in dependent rat models, acute alcohol withdrawal activates CRF systems in the central nucleus of the amygdala (CeA) (Merlo Pich et al., 1995;Zorrilla et al., 2001;Funk et al., 2006;Roberto et al., 2010) and bed nucleus of the stria terminalis (Olive et al., 2002). Extracellular CRF in rats also increased in the CeA during precipitated withdrawal from chronic nicotine (George et al., 2007), withdrawal from binge cocaine self-administration (Richter and Weiss, 1999), and precipitated withdrawal from opioids (Weiss et al., 2001) and cannabinoids (Rodriguez de Fonseca et al., 1997). Nicotine withdrawal in rats also increased CeA CRF mRNA and, especially in females, NAc CRF mRNA levels (Aydin et al., 2011;Torres et al., 2013). Conversely, amygdala CRF tissue content is reduced during acute withdrawal from ethanol exposure (Zorrilla et al., 2001;Funk et al., 2006;Wills et al., 2010) and from binge cocaine self-administration (Zorrilla et al., 2001;Zorrilla et al., 2012), suggesting degradation and depletion after sustained secretion. Supporting a functional role for central extended amygdala CRF₁ receptor activation in the negative affect/withdrawal stage, site-specific injections of CRF receptor antagonists into the central amygdala reduce anxiety-like behavior, motivational deficits for other reinforcers, and excessive self-administration of addictive substances during acute withdrawal (Heilig and Koob, 2007;Heilig et al., 2010b;Koob and Zorrilla, 2010;Logrip et al., 2011;Parylak et al., 2011).

The opponent process also may involve pituitary CRF₁-dependent activation of the HPA-axis, reflected by elevated ACTH and corticosteroids, because withdrawal from all drugs of abuse studied to date leads to an activated HPA stress response. Interestingly, glucocorticoids, effectors of the HPA-axis, can activate and sensitize CRF-CRF₁ systems of the extended amygdala, causally linking the neuroendocrine and extrahypothalamic CRF system stress responses. Consistent with a functional role for the HPA-axis component of the opponent process, glucocorticoid receptor antagonists can reduce the development and expression of excessive alcohol self-administration that results from repeated, intermittent ethanol intoxication (Vendruscolo et al., 2012).

Of preclinical relevance, systemic injections of small molecule CRF₁ receptor antagonists that block pituitary and brain CRF₁ receptors can also reduce the heightened anxiety-like behavior in dependent rodents acutely withdrawn from alcohol at doses that do not alter the anxiety-like behavior of non-dependent animals (Knapp et al., 2004;Overstreet et al., 2004;Breese et al., 2005a;Breese et al., 2005b;Gehlert et al., 2007;Sommer et al., 2008). Similarly, withdrawal from the repeated administration of cocaine, nicotine, cannabinoids, opiates, and benzodiazepines produces an anxiogenic-like response that can be reversed by intracranial administration of non-selective peptide CRF receptor antagonists (Sarnyai et al., 1995;Rodriguez de Fonseca et al., 1997;Basso et al., 1999;Tucci et al., 2003) or systemic administration of brain-penetrant CRF₁-selective nonpeptide receptor antagonists (George et al., 2007;Skelton et al., 2007;Park et al., 2013). Furthermore, the aversive state of opiate withdrawal and the decreased brain reward function associated with nicotine withdrawal are both CRF₁ receptor-dependent (Contarino and Papaleo, 2005;Stinus et al., 2005;Bruijnzeel et al., 2007;Bruijnzeel et al., 2009;Bruijnzeel et al., 2012;Garcia-Carmona et al., 2012). Likewise, intracerebroventricular administration of a nonselective CRF_1/2 antagonist ameliorated the decreased brain reward function resulting from ethanol withdrawal (Bruijnzeel et al., 2010). Supporting the motivational significance of these effects for addiction, systemic injections of small-molecule CRF₁ antagonists reduced the increased alcohol intake of dependent or postdependent rodents (Sabino et al., 2006;Chu et al., 2007;Funk et al., 2007;Gehlert et al., 2007;Gilpin et al., 2008;Richardson et al., 2008) as well as the increased intravenous self-administration of cocaine (Specio et al., 2008), nicotine (George et al., 2007), and heroin (Greenwell et al., 2009) in rats with a history of extended access to the drug of abuse. Similarly, both global (Chu et al., 2007) and conditional brain-specific Crhr1 knockout (Crhr1[NestinCre]) mice (Molander et al., 2012) show reduced ethanol intake during withdrawal in the postdependent state compared with their wildtype littermates.

CRF₁ receptor knockout mice also drink less 20% v/v ethanol under basal conditions (Pastor et al., 2011). Moreover, both CRF and CRF₁ knockout mice show reduced ethanol intake and blood ethanol concentrations in a murine model of scheduled, limited access to ethanol (“drinking-in-the-dark”) that can produce binge-like intake (Kaur et al., 2012), suggesting an early role for CRF in neuroadaptations associated with the binge/intoxication stage of the addiction cycle. Perhaps accordingly, systemic administration of small-molecule CRF₁ antagonists can reduce binge-like but not non-binge-like ethanol intake in C57BL/6J mice and outbred rats (Lowery et al., 2010;Cippitelli et al., 2012;Simms et al., 2013) (but see (Giardino and Ryabinin, 2013) for additional findings suggesting that these effects may not be specific for ethanol). Site-specific infusion of CRF₁ antagonists into the CeA or VTA likewise could reduce heightened ethanol intake under intermittent access schedules (Lowery-Gionta et al., 2012;Hwa et al., 2013).

Many individuals who suffer from symptoms of anxiety or depression may turn to a substance of abuse for its potential anxiolytic (e.g., alcohol) or mood-enhancing (e.g., cocaine) effects (Pohorecky, 1991). By reducing dysphoria, CRF₁ receptor antagonists may help treat individuals who “self-medicate” their anxious or depressed state with a drug of abuse. Consistent with this hypothesis, small-molecule CRF₁ receptor antagonists reduce alcohol drinking in rodent models with high innate anxiety, including genetically selected Marchigian Sardinian alcohol-preferring rats (Ciccocioppo et al., 2006;Hansson et al., 2006;Hansson et al., 2007;Heilig and Koob, 2007;Sommer et al., 2008) and isolation-reared Fawn-hooded rats (Lodge and Lawrence, 2003) at doses that do not alter the intake of normal, outbred rodents.

We and others also found evidence that addiction-like activation of CRF systems may play a role in the motivational properties of palatable food. Specifically, rats acutely withdrawn from intermittent access to a high-sucrose, chocolate-flavored diet showed increased anxiety-like behavior (Cottone et al., 2009). As has been seen with substances of abuse, the increased anxiety-like behavior was accompanied by increased mRNA and peptide expression of CRF in the CeA; similar molecular changes were seen by Bale and colleagues in mice withdrawn from high-fat diet (Teegarden et al., 2009). Systemic pretreatment with the selective CRF₁ antagonist R121919 blocked food withdrawal-associated anxiety at doses that did not alter the behavior of chow-fed controls. CRF₁ antagonist pretreatment also decreased the magnitude of overeating of the palatable sucrose-rich diet by diet-cycled animals at doses that did not alter the food intake of chow-fed controls or of animals fed the sucrose-rich diet, but without a history of diet cycling. Moreover, R121919 reduced evoked inhibitory postsynaptic potentials in the CeA more in diet-cycled rats than in chow-fed controls, suggesting greater control over CeA GABAergic neurotransmission by CRF₁ receptors. The findings resemble the enhanced modulatory influence of CRF₁ antagonists on CeA GABAergic synaptic transmission that is seen during withdrawal from alcohol (Roberto et al., 2010). When diet-cycled animals had access to the preferred, sucrose-rich diet, both their anxiety-like behavior and CeA CRF levels normalized, supporting the hypothesis that activation of the amygdala CRF-CRF₁ system helped subserve the palatable food withdrawal-like state.

Protracted withdrawal

Symptoms of negative affect can persist for weeks and months after detoxification from drugs of abuse (Alling et al., 1982). These negative emotional symptoms of protracted withdrawal, including anger, frustration, sadness, anxiety and guilt, are subacute and appear to be key precipitants of relapse (Hershon, 1977;Lowman et al., 1996;Zywiak et al., 1996;Annis et al., 1998) in the preoccupation/anticipation stage of addiction. Neuroadaptations in amygdala CRF₁ systems have been proposed to promote such protracted abstinence syndromes. Consistent with this hypothesis, increased levels of CRF and the CRF₁ receptor have been reported weeks after detoxification from repeated cycles of alcohol intoxication/withdrawal in animal models (Zorrilla et al., 2001;Sommer et al., 2008), and CRF₁ receptor antagonists reduce the potentiated anxiogenic-like and ethanol intake behavior responses to otherwise ineffectual stressors seen during protracted withdrawal (Rimondini et al., 2002;Valdez et al., 2002b;Valdez et al., 2003b;Sommer et al., 2008). CRF₁ antagonists also attenuate the increased spontaneous anxietylike behavior (Overstreet et al., 2002;Valdez et al., 2002b;Breese et al., 2005a;Breese et al., 2005b;Zhao et al., 2007b;Sommer et al., 2008) and alcohol intake that can be seen in postdependent rats even under low exteroceptive stress conditions (Rimondini et al., 2002;Valdez et al., 2002b;Sommer et al., 2008). Both sets of findings are significant because the resurgence of negative emotional states during protracted withdrawal is a major predictor of relapse in alcoholics (Mossberg et al., 1985;Pickens et al., 1985).

Stress-induced reinstatement

Exposure to external stressors can also increase drug craving (Childress et al., 1994;Cooney et al., 1997;Sinha et al., 2000) and lead to relapse. This stress-induced relapse is hypothesized to be motivated by self-medication of the associated negative emotional symptoms of stress (Lowman et al., 1996;Zywiak et al., 1996), in a fashion similar to self-medication of symptoms of protracted withdrawal, but here elicited by external stimuli. Consistent with this hypothesis, systemic injection of yohimbine, an α₂ adrenoceptor antagonist that induces stress-and anxiety-like responses (Bremner et al., 1996), can induce alcohol and heroin craving in human drug addicts (Stine et al., 2002;Umhau et al., 2011). Moreover, yohimbine can reinstate drug-seeking behavior in rats (Shepard et al., 2004;Feltenstein et al., 2012) and monkeys (Lee et al., 2004), as well as the seeking of alcohol (Marinelli et al., 2007) or palatable food in rats (Ghitza et al., 2006). Supporting a role for CRF systems in this relapse mechanism, yohimbine-induced reinstatement of substance-seeking can be prevented by systemic pretreatment with brain-penetrant CRF₁ receptor antagonists or intracranial pretreatment with peptide CRF receptor antagonists (Le et al., 2002;Marinelli et al., 2007;Shalev et al., 2010).

Likewise, CRF₁ antagonists can reduce stressor-induced reinstatement of substance-seeking in animal models, suggesting a key role for CRF in the preoccupation/anticipation stage of the addiction cycle. For example, systemic injections of CP154,526 attenuated footshock-induced reinstatement of alcohol seeking in nondependent rats (Le et al., 2000b), and subsequent studies showed that antalarmin and MTIP, brain-penetrant nonpeptide CRF₁ antagonists, likewise attenuated footshock-induced reinstatement of alcohol seeking, particularly in alcohol-dependent rats and genetically selected, Marchigian Sardinian alcohol-preferring rats (Gehlert et al., 2007), each of which show increased activity of extended amygdala CRF systems (Hansson et al., 2006;Francesconi et al., 2009;Roberto et al., 2010). Mixed CRF₁/CRF₂ antagonists injected intracranially into the extended amygdala, median raphe, and VTA (see below for details) and small-molecule CRF₁ antagonists administered systemically also blocked stressor-induced reinstatement of cocaine-, opiate-, nicotine-, and methamphetamine-seeking behavior (Shaham et al., 1998;Le et al., 2000a;Lu et al., 2003;Shaham et al., 2003;Bruijnzeel et al., 2009;Nawata et al., 2012). A CRF₁ antagonist also reduced social defeat stress-induced locomotor sensitization to cocaine and escalated “binge” operant self-administration of cocaine (Boyson et al., 2011) as well as stress-induced reinstatement of cocaine-seeking behavior in rats with a history of extended access to cocaine (Blacktop et al., 2011). A selective CRF₁ antagonist reduced shock-induced reinstatement of nicotine-seeking behavior (Plaza-Zabala et al., 2010). Finally, CRF₁ knockout mice were resistant to the ability of repeated forced swim stress to increase ethanol intake following deprivation as compared to wildtype mice (Pastor et al., 2011), and conditional brain-specific Crhr1 knockout (Crhr1[NestinCre]) mice were partly resistant to the ability of social defeat stress and forced swim stress to increase ethanol intake (Molander et al., 2012).

CRF₁ antagonists are ineffective in blocking cue-, substance-, and context-induced reinstatement, however, indicating specificity of their actions on the stress component of the addiction cycle. This specificity of action is consistent with the unique neuroanatomical and neuropharmacological bases for stress-induced reinstatement as contrasted from overlapping, but distinct, neurocircuits that differentially subserve drug prime- or drug cue- (discrete, discriminative or contextual) induced reinstatement (see (Koob, 2008;Steketee and Kalivas, 2011;Bossert et al., 2013) for reviews).

Data indicate that stressor-induced reinstatement is not mediated by activation of the HPA axis. For example, adrenalectomy did not alter fooshock-induced reinstatement (Le et al., 2000b), and antalarmin had no effect on yohimbine-induced corticosterone secretion at doses that reduced “relapse” behavior (Marinelli et al., 2007). Rather, CRF₁ receptor antagonists can reduce stress-induced reinstatement via extrahypothalamic brain sites, as has been shown for alcohol (Le et al., 2000b), cocaine (Erb et al., 1998), heroin (Shaham et al., 1997), nicotine (Zislis et al., 2007), and methamphetamine (Nawata et al., 2012). For example, site-specific blockade of CRF receptors in the median raphe nucleus (Le et al., 2002;Le et al., 2013) was sufficient to attenuate footshock- and yohimbine-induced reinstatement of alcohol seeking. Likewise, intracerebral injections of a mixed CRF₁/CRF₂ antagonist or a small-molecule CRF₁ antagonist into the bed nucleus of the stria terminalis or VTA, but not the amygdala or nucleus accumbens (NAc), reduced reinstatement of substance-seeking behavior (Lu et al., 2003;Shaham et al., 2003). The results agree with other evidence that activation of extrahypothalamic CRF sites subserves stress-induced reinstatement behavior (Shaham et al., 1997;Erb et al., 1998;Wang et al., 2005;Wang et al., 2006;Blacktop et al., 2011). Consistent with this hypothesis, intra-VTA (Wang et al., 2005) or intracerebroventricular (Blacktop et al., 2011;Kupferschmidt et al., 2011;Brown et al., 2012;Buffalari et al., 2012;Kupferschmidt et al., 2012) infusion of CRF can reinstate cocaine-seeking behavior in rats.

Some evidence supports a role for NAc CRF₁ systems in promoting substance consumption, if not seeking, during stress. For example, restraint stress during deprivation from ethanol can stimulate increased ethanol intake upon renewed access in ethanol-preferring P rats, an effect that can be blocked by intra-NAc, but not intra-amygdala, intra-dorsal raphe, or intra-VTA, administration of a CRF₁ antagonist (Knapp et al., 2011). Consistent with above, CRF injection into the NAc, but not VTA, raphe, or amygdala, can potentiate ethanol intake following a deprivation period (Knapp et al., 2011). More research is needed to determine conditions under and mechanisms via which CRF₁ systems of the VTA and NAc may differentially influence reinstatement of seeking vs. consummatory behavior, respectively.

Corticotropin-releasing factor, stress, and the frontal cortex

Converging evidence may link the impairment of medial prefrontal cortex (mPFC) cognitive function, activation of CRF in the PFC, and overactivation of the CeA with the development of compulsive-like responding for drugs of abuse, again suggesting a role for CRF in the binge/intoxication stage of the addiction cycle (Briand et al., 2008a;Briand et al., 2008b;George et al., 2008). Extended access to drugs of abuse, such as cocaine self-administration, can induces a compulsive-like pattern of intake taking that associates with impaired working memory (George et al., 2008). Whereas long-access (LgA) and short-access (ShA) rats both exhibit mostly correct responses in a delayed non-matching-to-sample task under low cognitive demand (delay < 10 s), working memory performance of LgA rats is substantially impaired by increasing the delay. The magnitude of escalation of cocaine intake correlates directly with the impaired working memory performance at the long delay. Furthermore, deficits in working memory were accompanied by a decreased density of dorsomedial PFC (dmPFC) neurons or oligodendrocytes that persisted for months despite cocaine cessation. Thus, dmPFC dysfunction might contribute to the loss of control associated with compulsive drug use and facilitate the progression to drug addiction.

CRF may be implicated in the reviewed frontal cortex dysfunction. Rats receiving chronic intermittent access to two-bottle choice alcohol drinking, a paradigm that leads to escalated alcohol intake (Wise, 1973;Simms et al., 2008), show increased activation of GABA and CRF neurons in the mPFC during abstinence. Working memory impairments in these rats correlate directly with greater alcohol drinking during acute abstinence (George et al., 2012). Abstinence was also associated with a functional disconnection of the mPFC and CeA, but not hippocampus or NAc. The results show a recruitment of a subset of GABA and CRF neurons in the mPFC during alcohol withdrawal and suggest that disconnection of the PFC-CeA pathway may be key to impaired executive control over motivated behavior. In summary, dysregulation of mPFC interneurons may be an early index of the transition to alcohol dependence.

Clinical trials of CRF₁ receptor antagonists in addiction

Collectively, the reviewed studies demonstrate a key role for brain CRF₁ receptors in three addiction-related domains: (1) negative emotional symptoms of acute and protracted withdrawal that can occur sans exteroceptive stress, (2) escalated, compulsive-like substance intake (e.g., with substance dependence), and (3) stress-induced relapse to substance seeking. Accordingly, small-molecule CRF₁ antagonists are currently in clinical trials for stress-related aspects of the addiction process. GlaxoSmithKline and the National Institutes of Health (NIH) are evaluating whether verucerfont can reduce stress-induced alcohol craving in anxious, stress-reactive alcoholic women (NCT01187511). Bristol Myers Squibb and NIH are testing whether pexacerfont can prevent stress-induced craving for palatable food in dieters (NCT01656577), stress-induced craving for tobacco in abstaining smokers (NCT01557556), and stress-induced craving for alcohol in anxious alcoholic women (NCT01227980) (Zorrilla et al., 2013a).

Functional and genetic heterogeneity of substance use disorders

The effectiveness of medications for substance use disorders differs across individuals and even within individuals at different times of their disease process (Heilig et al., 2010b;Koob and Zorrilla, 2010;Heilig et al., 2011;Logrip et al., 2011;Logrip et al., 2012). Based on the reviewed evidence, CRF₁ receptor antagonists would be expected to be most effective if substance use has transitioned to use driven by negative reinforcement (withdrawal/negative affect) and to protect against stress-induced relapse (stress-related craving). On the other hand, CRF₁ antagonists would be predicted to less effectively reduce reward-motivated, recreational substance abuse earlier in the addiction process (Heilig and Koob, 2007) or to mitigate relapse episodes that were precipitated by cues or contexts associated with previous drug taking (Liu and Weiss, 2002).

Given that substance use disorders are partly heritable, pharmacogenetic differences also may be relevant to CRF₁ antagonist pharmacotherapy, as is true of other treatments (Heilig et al., 2011;Sinha, 2011). Indeed, animal models support the hypothesis that gene variants for CRF system molecules may promote negatively-reinforced alcohol intake. For example, many msP alcohol-preferring rats carry two G-to-A polymorphisms in allelic identity with one another in the distal promoter of the Crhr1 gene. These mutations are not seen in other alcohol-preferring lines or outbred rats (Hansson et al., 2006; Logrip, Walker, Ayanwuyi, Sabino, Ciccocioppo, Walker, Koob and Zorrilla, unpublished observations). Perhaps as a result, the msP line exhibits increased CRF₁ receptor expression in several stress-related brain regions, increased anxiety-like behavior, and increased sensitivity to the ability of CRF₁ receptor antagonists to reduce alcohol self-administration and stress-induced reinstatement of alcohol seeking (Ciccocioppo et al., 2006;Hansson et al., 2006;Gehlert et al., 2007;Ayanwuyi et al., 2013; Logrip, Walker, Ayanwuyi, Sabino, Ciccocioppo, Walker, Koob and Zorrilla, unpublished observations). Similarly, rhesus monkeys that carry a C-to-T single nucleotide polymorphism in the promoter of the Crh gene do not show normal glucocorticoid feedback inhibition of CRF peptide expression. This gene variant is associated with two-fold greater alcohol consumption in monkeys exposed to early life stress, without altering the basal drinking of unstressed monkeys (Barr et al., 2009).

Several polymorphisms in human CRF system molecules have also been associated with alcohol use phenotypes, often in interaction with stress history. Several Crhr1 haplotype variants in adolescents predict binge drinking and lifetime prevalence of intoxication and alcohol dependence (Treutlein et al., 2006). Crhr1 single-nucleotide polymorphisms (SNPs) also predicted greater alcohol consumption in already dependent individuals (Treutlein et al., 2006). A study of 1,049 Caucasian subjects from 209 families in the Collaborative Study on the Genetics of Alcoholism (COGA) also found significant associations between the P3 amplitude and alcohol dependence with multiple SNPs in the Crhr1 gene (Chen et al., 2010). Stress history produces greater increases in future alcohol intake (Blomeyer et al., 2008;Schmid et al., 2010) and an earlier onset of drinking (Schmid et al., 2010) in adolescents homozygous for the C allele of the rs1876831 SNP of the Crhr1 gene. Adolescent carriers of the A allele of the rs242938 Crhr1 SNP similarly reported more drinking when exposed to stress in some (Schmid et al., 2010) but not other (Blomeyer et al., 2008) studies. Conversely, adolescents homozygous for the H2 haplotype containing the rs1876831 minor allele are protected against early abuse-associated increases in alcohol intake and dependence (Nelson et al., 2010).

CRF system polymorphisms in the CRF binding protein (CRF-BP), which the Vale laboratory discovered to moderate the ability of CRF to interact with its receptors (Potter et al., 1991;Potter et al., 1992;Behan et al., 1993;Behan et al., 1995), also have been linked to human alcohol phenotypes. For example, Crhbp gene SNPs are associated with an endophenotype of alcoholism (i.e., decreased electroencephalographic alpha wave power) and are more prevalent in alcohol use disorders (Enoch et al., 2008), including in alcoholics with comorbid anxiety disorders (Enoch et al., 1999). A Crhbp polymorphism (rs10055255) has also been associated with greater stress imagery-induced alcohol craving and dysphoria (Ray, 2011). Furthermore, SNPs in the Crhbp (rs3811939) and Crhr1 (the widely studied rs110402 polymorphism) genes jointly predicted comorbid alcohol use disorder in patients with schizophrenia (Ribbe et al., 2011), with greater ratios of mononuclear Crhr1/Crhbp mRNA seen in dual carriers of the polymorphism (Ribbe et al., 2011). Finally, a recent prospective study showed that CRF-BP genotype moderated the relationship between stress-induced negative affect and the self-reported negative consequences of drinking (Tartter and Ray, 2012). Further work is needed to determine whether similar gene variants are associated with phenotypes for other substance use disorders. Perhaps pharmacogenomic profiling could identify patients for whom CRF₁ receptor antagonist pharmacotherapy would be especially useful to prevent relapse (Sinha, 2011).

CRF₂ systems and addiction

Whereas CRF₁ systems are generally recognized to exert an overarching, pro-stress-like effect, CRF₂ receptor activation may, in addition to suppressing food intake (Spina et al., 1996;Inoue et al., 2003;Fekete et al., 2007), decrease stress responsiveness (Valdez et al., 2002a;Valdez et al., 2003a). In the context of addiction-related behavior, intracerebroventricular infusion of Ucn 3, a selective CRF₂ agonist, reduced the heightened anxiety-like behavior and increased ethanol self-administration in dependent rats during acute withdrawal (Valdez et al., 2004) as well as the alcohol intake of mice that receive limited, binge-like access to alcohol (Sharpe and Phillips, 2009;Lowery et al., 2010). Site-specific infusion of Ucn 3 into the CeA likewise reduced the heightened alcohol self-administration in withdrawn, dependent rats (Funk and Koob, 2007), and microinjection of Ucn 1 into the CRF₂ receptor-rich lateral septum potently reduced the acquisition and expression of alcohol intake in rats (Ryabinin et al., 2008). There is controversy in this area, however, because pharmacological activation of CRF₂ receptors has been found to decrease anxiety-like behavior, increase anxiety-like behavior, or not alter stress-related behavior, depending on the neuropharmacological probe, dose, and brain site (Ho et al., 2001;Takahashi et al., 2001;Fekete and Zorrilla, 2007;Zhao et al., 2007a). Accordingly, (any) influence of the CRF₂ system on addiction-related behavior may also be brain-region specific.

CRF₂ receptors in stress-induced reinstatement

CRF₂ receptors outside the extended amygdala have been suggested to act in concert with CRF₁ receptors to facilitate aspects of compulsive-like drug seeking. VTA CRF₂ receptors have been proposed to facilitate stress-induced reinstatement of cocaine seeking, in which footshock exposure in cocaine-experienced rats elicited the release of extended amygdala-derived CRF into the VTA (Wang et al., 2005). The CRF-mediated stress-induced reinstatement of cocaine seeking is putatively mediated by the sensitization of VTA glutamate release. The activation of VTA CRF₂, rather than CRF₁, receptors, was proposed to mediate these actions (Wang et al., 2007), but it should be noted that the putative CRF₂ antagonist used, anti-sauvagine-30, is not highly selective for CRF₂ receptors (Zorrilla et al., 2013b). Indeed, others who used the more selective CRF₂ antagonist astressin₂-B did not observe a reversal of stress-induced reinstatement of substance seeking (Bruijnzeel et al., 2009), and negligible CRF₂ receptor mRNA is detected in the rodent VTA by in situ hybridization under basal conditions (Chalmers et al., 1995;Van Pett et al., 2000). Nevertheless, the possible role for CRF₂ receptors in stress-induced relapse to drug seeking also implicates a possible role for the CRF peptide family members that have high affinity for CRF₂ receptors (Gysling, 2012), including Ucn 1, 2 and 3.

CRF₂ receptor and opioid withdrawal

Recent studies have obtained evidence that the CRF₂ receptor may play a key role in opiate withdrawal. First, whereas knockout of the CRF₁ receptor exacerbated the somatic signs of opiate withdrawal (Papaleo et al., 2007), genetic deletion of the CRF₂ receptor blocked the somatic signs of opiate withdrawal (Papaleo et al., 2008). Moreover, CRF₂ knockout mice did not manifest the dysphoria-like or anhedonia-like behaviors of opiate withdrawal (Ingallinesi et al., 2012), whereas they showed normal neuroendocrine responses, as indexed by HPA activation (e.g., CRF in the paraventricular nucleus the hypothalamus). The findings potentially implicate CRF₂ receptors in driving extrahypothalamic CRF and dynorphin responses that subserve opiate withdrawal distress, perhaps via presynaptic positive regulation of CRF synthesis and release (Ingallinesi et al., 2012).

Urocortin 1 and alcohol intake

In the brain, Ucn 1, which has equal activity at CRF₁ vs. CRF₂ receptors, is synthesized principally in cell bodies of the stress-responsive (Korosi et al., 2005) perioculomotor Ucn 1-containing area (pIIIu), now also known as the nonpreganglionic Edinger-Westphal nucleus (Weitemier et al., 2005), and, to a lesser extent, in the lateral superior olive (Vaughan et al., 1995;Ryabinin et al., 2005). Ucn 1 was first implicated in alcohol consumption based on the preferential induction of Fos expression in Ucn 1-containing neurons (Bachtell et al., 2002;Ryabinin et al., 2003;Spangler et al., 2009) of the pIIIu following voluntary alcohol drinking in multiple species (Topple et al., 1998;Ryabinin et al., 2001;Weitemier et al., 2001;Bachtell et al., 2003;Sharpe et al., 2005;Turek and Ryabinin, 2005;Kaur and Ryabinin, 2010;Anacker et al., 2011) and because basal Ucn1 immunoreactivity in the pIIIu is elevated in several inbred or selectively bred rodent lines that demonstrate high alcohol drinking phenotypes (Bachtell et al., 2002;Bachtell et al., 2003;Turek and Ryabinin, 2005;Fonareva et al., 2009) or that show increased sensitivity to the rewarding, hypothermic, or sedative actions of ethanol (Bachtell et al., 1999;Kiianmaa et al., 2003;Ryabinin and Weitemier, 2006;Turek et al., 2008). Accordingly, electrolytic lesions of the centrally projecting Edinger-Westphal nucleus in C57BL/6J mice attenuated ethanol intake and preference in a two-bottle choice paradigm without influencing the consumption of sucrose, quinine, saccharin, or saline (Bachtell et al., 2004;Weitemier and Ryabinin, 2005). Consistent with a role for Edinger-Westphal Ucn 1 in ethanol preference, either electrolytic lesion of the Edinger Westphal nucleus or Ucn 1 knockout reduces ethanol preference, but neither treatment produces an additional effect above and beyond the other (Giardino et al., 2011). Furthermore, ethanol did not promote a conditioned place preference in mice deficient in Ucn 1 or the CRF₂ receptor (Giardino et al., 2011). The reviewed results support the hypothesis that centrally projecting Ucn 1 neurons from the Edinger-Westphal nucleus may promote ethanol intake, preference, or reward, perhaps via actions on CRF₂ receptors. Limiting this conclusion, however, bilateral administration of Ucn 1 into the mouse lateral septum selectively attenuated alcohol self-administration during both the acquisition and expression of a limited access alcohol drinking procedure (Ryabinin et al., 2008). More work is needed to elucidate which specific Ucn 1-containing circuits promote vs. oppose ethanol use and the applicability of reviewed findings to other substances of abuse.

Concluding remarks

On a personal note, Wylie Vale was the impetus and driving force behind our early work on the role of CRF in behavioral responses to stressors. We (Floyd Bloom, George Siggins, and myself, George Koob) joined Vale on the NIH program project grant from the National Institute on Diabetes and Digestive and Kidney Diseases when CRF was discovered in 1981. We subsequently injected CRF into the brain of rats and charted their behavioral responses in collaboration with Vale and colleagues (Sutton et al., 1982). The resulting behavioral profile and subsequent studies led to the overall hypothesis that CRF, in addition to its effects on the HPA axis, also mediated behavioral responses to stressors. This work proceeded in parallel, but independent of, our addiction work until we conceptualized between-system neuroadaptations as explanations for opponent process (Koob and Bloom, 1988) and invoked the hypothesis that activation of the brain CRF systems was a key player in such between-system neuroadaptations. During this period, as data implicating the role for CRF in addiction consolidated, Eric Zorrilla joined our joint Salk/Scripps project extending our work to protracted abstinence and compulsive eating.

Wylie Vale was the guiding light to our work conceptually, innovatively, and motivationally. He embraced our overall hypothesis that CRF drove the “dark side” of addiction, provided the molecular advances and tools to test this hypothesis (knockout mice, discovery of the CRF₁ and CRF₂ receptors, powerful antibodies, discovery of urocortins, etc.), and delighted us with his constant enthusiasm for each new finding. Wylie Vale was a force of nature who had a tremendous impact on the fields of neuroendocrinology and neuroscience because he embraced any reasonable hypothesis that was supported by the data and let the work on CRF be guided not by preconceived notions. Not to mention his wonderful humor, incredible curiosity and intellect. We miss him terribly.

Highlights.

CRF systems have emerged as mediators of the body’s response to stress and, relatedly, the pathophysiology of addiction.

CRF systems have a prominent role in driving the addiction via actions in the central extended amygdala.

Addiction-related actions include anxious behavior, brain reward deficits, excessive drug use and stress-induced drug-seeking.

Polymorphisms in CRF system molecules are associated with drug use phenotypes in humans, often with stress history.